English Historical Linguistics: Volume 1 9783110251593, 9783110202205

Table of contents :
Acknowledgments
In memoriam
General abbreviations
I Periods
1 Pre-Old English
2 Old English
3 Middle English
4 Early Modern English
5 Late Modern English
6 Contemporary English
II Linguistic Levels
7 Phonology
8 Prosody
9 Morphology
10 Syntax
11 Semantics and lexicon
12 Idioms and fixed expressions
13 Pragmatics and discourse
14 Onomastics
15 Orthography
16 Styles, registers, genres, text types
III Old English
17 Phonology
18 Morphology
19 Syntax
20 Semantics and lexicon
21 Pragmatics and discourse
22 Dialects
23 Language contact
24 Standardization
25 Literary language
IV Middle English
26 Phonology
27 Morphology
28 Syntax
29 Semantics and lexicon
30 Pragmatics and discourse
31 Dialects
32 Language contact
33 Standardization
34 Sociolinguistics
35 Literary language
36 The language of Chaucer
V Early Modern English
37 Phonology
38 Morphology
39 Syntax
40 Lexicon and semantics
41 Pragmatics and discourse
42 Dialects
43 Language contact
44 Standardization
45 Sociolinguistics
46 Pronouns
47 Periphrastic DO
48 The Great Vowel Shift
49 Relativization
50 Literary language
51 The language of Shakespeare
VI Late Modern English
52 Phonology
53 Morphology
54 Syntax
55 Semantics and lexicon
56 Pragmatics and discourse
57 Dialects
58 Standardization
59 Sociolinguistics
VII Standardization
60 Prescriptive tradition
61 The complaint tradition
62 Standards in the history of English
63 Codifiers
64 English language regard
65 Bible translations
66 Dictionaries and the standardization of English
VIII English and the Media
67 Newspapers
68 Television
69 Radio
70 Internet
Index

Citation preview

English Historical Linguistics HSK 34.1

Handbücher zur Sprach- und Kommunikationswissenschaft Handbooks of Linguistics and Communication Science Manuels de linguistique et des sciences de communication Mitbegründet von Gerold Ungeheuer (†) Mitherausgegeben 1985−2001 von Hugo Steger

Herausgegeben von / Edited by / Edités par Herbert Ernst Wiegand Band 34.1

De Gruyter Mouton

English Historical Linguistics An International Handbook

Edited by Alexander Bergs Laurel J. Brinton Volume 1

De Gruyter Mouton

ISBN 978-3-11-020220-5 e-ISBN 978-3-11-025159-3 ISSN 1861-5090 Library of Congress Cataloging-in-Publication Data A CIP catalog record for this book has been applied for at the Library of Congress Bibliographic information published by the Deutsche Nationalbibliothek The Deutsche Nationalbibliothek lists this publication in the Deutsche Nationalbibliografie; detailed bibliographic data are available in the Internet at http://dnb.dnb.de © 2012 Walter de Gruyter GmbH & Co. KG, Berlin/Boston Cover design: Martin Zech, Bremen Typesetting: Apex CoVantage Printing: Hubert & Co. GmbH & Co. KG, Go¨ttingen ⬁ Printed on acid-free paper s Printed in Germany www.degruyter.com

Preface to Historical Linguistics of English

The study of the English language has a lengthy history. The second half of the 18th century saw a phenomenal increase in the number of published grammars of the vernacular language, while the field of comparative linguistics arising in the 19th century was concerned in large part with the Germanic languages, including English. However, it is in the field of theoretical linguistics that English has played a truly central role. While there are no reliable statistics, it seems safe to say that the majority of studies in contemporary linguistics deal at least in part with English, and are also written in English. During the 20th century, monumental works concerned with the English language, both synchronic and diachronic, were produced, following historical/comparative and more contemporary linguistic approaches. In keeping with developments in the field of general linguistics, today it is possible to find descriptions and analyses of the history and development of English from virtually any linguistic perspective: external, internal, generative, functional, sociolinguistic, pragmatic, comparative, phonological, morphological, syntactic, lexical, semantic. There are numerous “Histories of English” to cater to just about every (theoretical) taste, as well as detailed descriptions of historical periods, language levels, or theoretical frameworks of English and specialized studies of individual topics in the development of the language. Work on the history of English has culminated most recently in the seven-volume Cambridge History of the English Language, edited by Richard M. Hogg (1992–2001). Study of the history of any language begins with its texts. Increasingly, however, scholars are turning to dictionaries and corpora of English that are available online or electronically. The pioneer historical corpus of English, the Helsinki Corpus of English Texts, was first released to scholars in 1991. The third edition of the Oxford English Dictionary online is now fully integrated with the Historical Thesaurus. The searchable Middle English Dictionary, completed in 2003, is available online along with the Corpus of Middle English Prose and Verse. The Dictionary of Old English Web Corpus is also searchable online. ARCHER, A Representative Corpus of Historical English Registers 1650–1990, accessible at a number of universities, provides a balanced selection of historical texts in electronic form. COHA, a 400-million-word, balanced Corpus of Historical American English 1810–2009, was launched online in 2010. Smaller corpora, such as the Corpus of English Dialogues 1560–1760, the Lampeter Corpus of Early Modern English Tracts, the Corpus of Early English Correspondence, the Corpus of Early English Medical Writing, and the Old Bailey Corpus, have made more specialized corpora available to scholars. Archives of historical newspapers online, including the Zurich English Newspaper Corpus, provide another source of electronic data. Finally, syntactically annotated corpora for historical stages of English are being produced, including the The York-Helsinki Parsed Corpus of Old English Poetry, The York-Toronto-Helsinki Parsed Corpus of Old English Prose, The Penn-Helsinki Parsed Corpus of Middle English, and The Penn-Helsinki Parsed Corpus of Early Modern English.

vi

Preface to Historical Linguistics of English Taking into account the important developments in the study of English effected by the availability of electronic corpora, this Handbook of English Historical Linguistics offers a comprehensive, interdisciplinary, and theory-neutral synopsis of the field. It is meant to facilitate research by offering overviews of all the relevant aspects of the historical linguistics of English and by referring scholars and students to more indepth coverage. The handbook is intended primarily for researchers in the field of (historical) linguistics generally, as well as for researchers in allied fields (such as history, literature, and culture). The handbook comprises two volumes, each volume consisting of approximately 70 articles written by a wide variety of authors from a number of different countries world-wide, representing a variety of theoretical approaches, and including both younger scholars as well as more established experts.

Volumes 1 and 2 The sequencing of material in the two volumes of the Handbook of English Historical Linguistics is bottom-up, beginning with detailed studies of the periods, levels, and linguistic components of each period. The second volume moves to a higher level, with a focus on general underlying concepts, theories, and methods as well as new and hitherto rather neglected approaches to the history of English. While the two volumes form a set, with cross-reference as far as possible in order to facilitate reader-guidance, they are also capable of standing alone. Following this essentially inductive approach, then, the first volume (edited by Laurel J. Brinton) is focused on the details of English language history. After overviews of the recognized periods of English (Section I), the volume then treats the linguistic levels. These are broadly understood to include newer components such as prosody, pragmatics, phraseology, discourse, styles, registers, and text types as well as more traditional areas such as orthography and onomastics in addition to the fully acknowledged areas of phonology, morphology, syntax, and semantics (Section II). These summaries will be useful both to students and to those not working directly in the field of English historical linguistics, such as typologists. Sections III–VI contain detailed descriptions of the different periods – Old English, Middle English, Early Modern English, and Late Modern English – in respect to the range of linguistic levels; discussions of language contact, standardization, sociolinguistics, and literary language are included for most periods. Moreover, for each period, selected important phenomena (such as the development of do-periphrasis, the Great Vowel Shift, pronoun usage, or relativization) have been chosen for more detailed study. Following the treatment of the different periods, the volume addresses a variety of questions of standardization (Section VII), such as the effects that dictionaries, the Bible, language attitudes, and codifiers have on normalizing the language. The last section (VIII) brings the handbook into the 21st century by treating the effects of new media (radio, television, computer) on forms of the language, as well as the longer established effects of newspapers. The second volume (edited by Alexander Bergs) then abstracts away from these details and moves outward to address theoretical concerns raised by the topics covered in Volume 1. Volume 2 first surveys resources for the studying and teaching of English (Section IX). Section X on interdisciplinarity (in particular literature and music) and historiography explores some of the debates involved in writing a history of English, questioning, for example, how the continuum of history is divided into accepted

Preface to Historical Linguistics of English

vii

“periods”, how oral and written forms of the language are accommodated in a history of English, and how new and perhaps “alternative histories” relate to the more established stories. This is followed by a history of the discipline of English historical linguistics itself, as it has developed in different parts of the world (Section XI). A significant part of Volume 2 covers changes in the English language as they have been theorized in various linguistic fields in the 20th century (Section XII). As Neogrammarian and Structuralist approaches are, to a great extent, embodied in the treatment of topics in Volume 1, this volume begins with later 20th century theories, including Generative Grammar, Construction Grammar, Lexical Functional Grammar, Rates of Change, Frequency, Lexical Diffusion, Grammaticalization, Lexicalization, and Language Acquisition. Related to the theoretical perspectives are new approaches which have been developed in the analysis of the history of English, including Historical Dialectology, Historical Sociolinguistics, Historical Pragmatics, Corpus Linguistics, Information Structuring, and Actuation/ Change from Below. Another important aspect of Volume 2 is its focus on the effects of language contact and the often neglected history of different varieties of English. It offers a section on language contact in the history of English, organized by contact languages, and supplemented by discussions of pidginization and creolization in the history of English and its varieties (Section XIII). Section XIV comprises historical sketches of more than ten varieties of English, and complementary theoretical discussions of dialect contact, diffusion, and supra-regionalization. The history of several second-language varieties is treated in Section XV, ending with a discussion of Global English. The beginning of a new millennium seems the right time for taking stock of the long span of scholarship in English historical linguistics and for surveying the field as a whole. Furthermore, the availability of electronic resources has changed the study of the history of English in fundamental ways, and it is important that a new handbook recognize this turning point in the study of English. Alexander Bergs, Osnabru¨ck (Germany) Laurel J. Brinton, Vancouver (Canada)

Contents

Volume I Acknowledgments. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xv In memoriam . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xvii General abbreviations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xix

I Periods 1 2 3 4 5 6

Pre-Old English  Jeannette K. Marsh . Old English  Ferdinand von Mengden. Middle English  Jeremy J. Smith . . . . . Early Modern English  Arja Nurmi. . . Late Modern English  Joan C. Beal . . Contemporary English  Jim Miller . . .

. . . . . .

. . . . . .

. . . . . .

. . . . . .

. . . . . .

. . . . . .

. . . . . .

. . . . . .

. . . . . .

. . . . . .

. . . . . .

. . . . . .

. . . . . .

. . . . . .

. . . . . .

. . . . . .

. . . . . .

. . . . . .

. . . . . .

. . . . . .

. . . . . .

. . . . . .

. . . . . .

. . . . . .

. . . . . .

. . . . . .

. 1 19 32 48 63 78

7 Phonology  Janet Grijzenhout . . . . . . . . . . . . . . . . . . 8 Prosody  Donka Minkova . . . . . . . . . . . . . . . . . . . . . 9 Morphology  Dieter Kastovsky . . . . . . . . . . . . . . . . . 10 Syntax  Graeme Trousdale . . . . . . . . . . . . . . . . . . . . 11 Semantics and lexicon  Elizabeth Closs Traugott. . . . 12 Idioms and fixed expressions  Gabriele Knappe . . . . 13 Pragmatics and discourse  Andreas H. Jucker . . . . . . 14 Onomastics  Carole Hough . . . . . . . . . . . . . . . . . . . 15 Orthography  Hanna Rutkowska . . . . . . . . . . . . . . . 16 Styles, registers, genres, text types  Claudia Claridge.

. . . . . . . . . .

. . . . . . . . . .

. . . . . . . . . .

. . . . . . . . . .

. . . . . . . . . .

. . . . . . . . . .

. . . . . . . . . .

. . . . . . . . . .

. . . . . . . . . .

. . . . . . . . . .

. . . . . . . . . .

. . . . . . . . . .

. . . . . . . . . .

. . . . . . . . . .

. . . . . . . . . .

97 113 129 148 164 177 197 212 224 237

. . . . . . . . .

. . . . . . . . .

. . . . . . . . .

. . . . . . . . .

. . . . . . . . .

. . . . . . . . .

. . . . . . . . .

. . . . . . . . .

. . . . . . . . .

. . . . . . . . .

. . . . . . . . .

. . . . . . . . .

. . . . . . . . .

. . . . . . . . .

. . . . . . . . .

255 272 294 313 325 340 362 373 385

II Linguistic Levels

III Old English 17 18 19 20 21 22 23 24 25

Phonology  Robert Murray . . . . . . . . . . . . . . Morphology  Ferdinand von Mengden . . . . . . Syntax  Rafał Molencki . . . . . . . . . . . . . . . . . Semantics and lexicon  Christian Kay . . . . . . . Pragmatics and discourse  Ursula Lenker . . . . Dialects  Hans Sauer and Gaby Waxenberger. Language contact  Gernot R. Wieland . . . . . . Standardization  Lucia Kornexl . . . . . . . . . . . Literary language  Robert D. Fulk . . . . . . . . .

. . . . . . . . .

. . . . . . . . .

. . . . . . . . .

. . . . . . . . .

. . . . . . . . .

x

Contents

IV Middle English 26 27 28 29 30 31 32 33 34 35 36

Phonology  Nikolaus Ritt . . . . . . . . . . . . . . . . . . . . Morphology  Jerzy Wełna. . . . . . . . . . . . . . . . . . . . Syntax  Jeremy J. Smith . . . . . . . . . . . . . . . . . . . . . Semantics and lexicon  Louise Sylvester . . . . . . . . . Pragmatics and discourse  Elizabeth Closs Traugott Dialects  Keith Williamson. . . . . . . . . . . . . . . . . . . Language contact  Herbert Schendl . . . . . . . . . . . . Standardization  Ursula Schaefer . . . . . . . . . . . . . . Sociolinguistics  Alexander Bergs . . . . . . . . . . . . . . Literary language  Leslie K. Arnovick . . . . . . . . . . The language of Chaucer  Simon Horobin . . . . . . .

. . . . . . . . . . .

. . . . . . . . . . .

. . . . . . . . . . .

. . . . . . . . . . .

. . . . . . . . . . .

. . . . . . . . . . .

. . . . . . . . . . .

. . . . . . . . . . .

. . . . . . . . . . .

. . . . . . . . . . .

. . . . . . . . . . .

. . . . . . . . . . .

. . . . . . . . . . .

. . . . . . . . . . .

. . . . . . . . . . .

. . . . . . . . . . .

399 415 435 450 466 480 505 519 534 551 576

V

Early Modern English

37 38 39 40 41 42 43 44 45 46 47 48 49 50 51

Phonology  Julia Schlu¨ter . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Morphology  Claire Cowie . . . . . . . . . . . . . . . . . . . . . . . . . . . . Syntax  Elena Seoane. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Lexicon and semantics  Ian Lancashire. . . . . . . . . . . . . . . . . . . Pragmatics and discourse  Dawn Archer. . . . . . . . . . . . . . . . . . Dialects  Anneli Meurman-Solin . . . . . . . . . . . . . . . . . . . . . . . Language contact  Laura Wright . . . . . . . . . . . . . . . . . . . . . . . Standardization  Lilo Moessner . . . . . . . . . . . . . . . . . . . . . . . . Sociolinguistics  Helena Raumolin-Brunberg . . . . . . . . . . . . . . Pronouns  Ulrich Busse . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Periphrastic DO  Anthony Warner . . . . . . . . . . . . . . . . . . . . . . The Great Vowel Shift  Manfred Krug . . . . . . . . . . . . . . . . . . . Relativization  Christine Johansson . . . . . . . . . . . . . . . . . . . . . Literary language  Colette Moore . . . . . . . . . . . . . . . . . . . . . . The language of Shakespeare  Ulrich Busse and Beatrix Busse .

. . . . . . . . . . . . . . .

. . . . . . . . . . . . . . .

. . . . . . . . . . . . . . .

. . . . . . . . . . . . . . .

. . . . . . . . . . . . . . .

. . . . . . . . . . . . . . .

. . . . . . . . . . . . . . .

589 604 621 637 652 668 685 698 714 731 743 756 776 791 808

52 Phonology  Charles Jones. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53 Morphology  Britta Mondorf . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54 Syntax  Bas Aarts, Marı´a Jose´ Lo´pez-Couso, and Bele´n Me´ndez-Naya . 55 Semantics and lexicon  Marina Dossena . . . . . . . . . . . . . . . . . . . . . . 56 Pragmatics and discourse  Diana M. Lewis . . . . . . . . . . . . . . . . . . . . 57 Dialects  Susanne Wagner . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 58 Standardization  Anita Auer. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59 Sociolinguistics  Erik Smitterberg . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

827 842 869 887 901 915 939 952

VI Late Modern English

VII 60 61 62

Standardization Prescriptive tradition  Edward Finegan. . . . . . . . . . . . . . . . . . . . . . . . . . 967 The complaint tradition  Tony Crowley . . . . . . . . . . . . . . . . . . . . . . . . . 980 Standards in the history of English  Claudia Lange. . . . . . . . . . . . . . . . . 994

Contents

63 64 65 66

xi

Codifiers  Carol Percy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . English language regard  Dennis R. Preston and Jon Bakos . . . . Bible translations  Thomas Kohnen . . . . . . . . . . . . . . . . . . . . . . Dictionaries and the standardization of English  John Considine.

. . . .

. . . .

. . . .

. . . .

. . . .

1006 1020 1039 1050

. . . .

. . . .

. . . .

. . . .

. . . .

1063 1075 1089 1105

VIII English and the Media 67 68 69 70

Newspapers  Udo Fries . . . . . Television  Jane Stuart-Smith . Radio  Ju¨rg Rainer Schwyter . Internet  Theresa Heyd . . . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1119

xii

Contents

Volume 2 IX Resources 71 72 73 74 75 76

X 77 78 79 80 81

Early textual resources  Kathryn A. Lowe Electronic/online resources  Oliver M. Traxel Lexicographic resources  Philip Durkin Teaching perspectives  Michael Adams Textbooks  Mary Blockley Online resources for teaching  Beatrix Busse

Interdisciplinarity and Historiography Literature  Andrew Johnston Music as a language – the history of an idea  Nadja Hekal Periodization in the history of the English language  Anne Curzan Myths of the English language; or, alternative histories of “English”  Richard J. Watts Spoken and written English – orality and literacy  Ursula Schaefer

XI History of English Historical Linguistics 82 83 84 85 86 87 88 89 90

XII 91 92 93 94 95 96 97 98 99 100 101 102

Overview  Alexander Bergs and Laurel J. Brinton The historiography of the English language  Jeremy J. Smith North America  Thomas Cable Germany and the German-speaking countries  Ilse Wischer The Netherlands and Belgium  Peter Petre´ Northern Europe  Risto Hiltunen East-Central and Eastern Europe  Ire´n Hegedu˝s Southern Europe  Teresa Fanego Asia  Minoji Akimoto

New Perspectives, Theories and Methods Historical dialectology  I. Keith Williamson Historical sociolinguistics  Terttu Nevalainen Historical pragmatics  Irma Taavitsainen Information structure and syntax in the history of English  Bettelou Los and Ans van Kemenade The actuation problem revisited  Richard J. Watts Corpus linguistics  Merja Kyto¨ Frequency and language change  K. Aaron Smith Lexical diffusion  Betty S. Phillips Grammaticalization  Lieselotte Brems and Sebastian Hoffmann Lexicalization  Laurel J. Brinton Diachronic change and language acquisition  Holger Diessel Generative approaches to English historical linguistics  Bettelou Los

Contents

103 104

Construction Grammar  Alexander Bergs Lexical Functional Grammar  Cynthia L. Allen

XIII English in Contact 105 German and Dutch  Jennifer Hendriks 106 French  Janne Skaffari 107 Celtic and Celtic Englishes  Markku Filppula and Juhani Klemola 108 Latin  Letizia Vezzosi 109 Greek  Brian D. Joseph 110 Norse  Richard Dance 111 English in contact with “other” European languages  Cristina Sua´rez-Go´mez 112 Native American Languages  Keren Rice 113 Pidgins and creoles  Suzanne Romaine 114 Middle English creolization  David M. Trotter 115 African American English (AAE) early evidence  Alexander Kautzsch

XIV Varieties of English 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131

Standard American English  Richard W. Bailey Re-viewing the origins and history of African American Language  Sonja L. Lanehart Regional varieties of American English  Luanne von Schneidemesser Canadian English in real-time perspective  Stefan Dollinger Standard British English  Pam Peters Received Pronunciation  Lynda Mugglestone Estuary English  Ulrike Altendorf Regional varieties of British English  Bernd Kortmann and Christian Langstrof Scots  Robert McColl Millar English in Ireland  Jeffrey L. Kallen English in Wales  Colin H. Williams English in Australian/New Zealand English  Marianne Hundt Cockney  Sue Fox Diffusion  David Britain Dialect contact  Peter Trudgill Supraregionalization  Raymond Hickey

XV Second-Language Varieties 132 133 134 135 136

English in India  Devyani Sharma English in Africa  Rajend Mesthrie Second-language varieties of English  Daniel Schreier English-based Creoles  Andrea Sand Global English  Joachim Grzega

xiii

Acknowledgments

Foremost, the editors wish to thank the nearly 150 experts in English historical linguistics worldwide who contributed chapters, without whom these volumes would not exist. We are particularly grateful to those who wrote two chapters or who stepped in to fill gaps that arose late in the process of assembling the contributions. We would also like to thank our Advisory Board – Cynthia Allen, Merja Kyto¨, Donka Minkova, and Elizabeth Closs Traugott – who gave us invaluable advice in the initial stages of this project. Thanks too to Anne Curzan, who helped in the planning stages. Our student assistants provided invaluable assistance in the editing stage: Slade Stolar and Martin McCarvill of the University of British Columbia; Jens Bonk, Lisa Gratzke, Barbara Hagenbrock, Claudia Ko¨mmelt, Mona Matzke, Meike Pentrel, and Lena Probst of the University of Osnabru¨ck. At De Gruyter Mouton, we are grateful to former Publishing Director Anke Beck for inviting us to develop this project and to Uri Tadmour to seeing it to completions, to Barbara Karlson for her encouragement, patience, gentle prodding, and expert guidance, and to Ulrike Swientek for her production expertise. For her keen eye and soft touch in copy-editing, we are most appreciative of Catherine Every (of EveryWord), and for her meticulous indexing, we thank Vicki Low (of Scholar’s Cap). We extend our gratitude to all of the following scholars, who generously contributed their time and expertise in serving as referees for the articles contained in these volumes. Some went well beyond the call of duty and reviewed more than one article or both wrote and reviewed an article: Sylvia Adamson John Algeo Ulrike Altendorf Leslie K. Arnovick Richard Bailey Jo´hanna Barðdal Joan Bresnan Derek Britton Ulrich Busse Joan Bybee Deborah Cameron Ruth Carroll Jack Chambers Claudia Claridge Eve Clark Sandra Clarke Richard Coates Juan Camilo Conde-Silvestre John Considine Nikolas Coupland

Jonathan Culpeper Hubert Cuyckens Mary Catherine Davidson Hendrik De Smet Dagmar Deuber Hans-Ju¨rgen Diller Stefan Dollinger Bridget Drinka Edwin Duncan Stefan Evert Edward Finegan Olga Fischer Susan Fitzmaurice Robert Fulk Heinz Giegerich Eugene Green Peter Grund Trinidad Guzma´n Gonza´lez Martina Ha¨cker Antonette diPaolo Healey

xvi

Acknowledgments Lena Heine Juan Manuel Herna´ndez-Campoy Susan Herring Raymond Hickey Gary Holland Richard Ingham Matti Kilpio¨ John Kirk Marina Kolokonte Lucia Kornexl William Kretzschmar Barbara Kryk-Kastovsky Merja Kyto¨ Roger Lass Gerhard Leitner Christian Liebl Michael Linn Angelika Lutz T. W. Machan Michael K. C. MacMahon Christian Mair Murray McGillivray Daniel McIntyre Donka Minkova Marianne Mithun Rosamund Moon Bruce Moore Colette Moore Susanne Mu¨hleisen Pieter Muysken Robert Murray Minna Nevala

Hans Frede Nielsen Arja Nurmi Stephen B. Partridge Meike Pfaff Joanna Przedlacka Matti Rissanen Juhani Rudanko Mats Ryde´n Pingali Sailaja Joseph Salmons Holger Schmidt Anne Schro¨der Elena Seoane K. Aaron Smith Dieter Stein Merja Stenroos Patrick Studer Sali Tagliamonte Sanna-Kaisa Tanskanen Sarah G. Thomason Ingrid Tieken Boon von Ostade Carola Trips Uwe Vosberg Susanne Wagner Terry Walker Gregory Ward Brita Wa˚rvik John Wells Gernot R. Wieland Walt Wolfram Alison Wray Nuria Ya´n˜ez-Bouza

In memoriam

We commemorate those friends and colleagues who passed away since this project came into being. Without them, English historical linguistics will not be the same: Richard Bailey, Derek Britton, and Richard Hogg.

General abbreviations

ACC ACT ADJ ADV AN

Angl. AUX AP

C C COMPR DAT CP DEM

DM DU

EModE EWSax. FEM

Fr. GEN

Ger. Gk. Go. Grmc. IE IMP IND INF INFL INSTR

IP Kent. Lt. LModE LWSax. MASC

ME MED ModE NEG NEUT

accusative case active adjective adverb Anglo-Norman Anglian auxiliary adjective phrase consonant complementizer comparative dative case complementizer phrase demonstrative discourse marker dual Early Modern English Early West Saxon feminine French genitive case German Greek Gothic Germanic Indo-European imperative indicative infinitive inflected instrumental case inflection phrase Kentish Latin Late Modern English Late West Saxon masculine Middle English Middle English Dictionary Modern English negative neuter

xx

General abbreviations noun nominative case NP noun phrase O object OBJ objective case OE Old English OED Oxford English Dictionary OFr. Old French OFris. Old Frisian OHG Old High German ON Old Norse OSax. Old Saxon OV object-verb word order P person PASS passive PAST past PDE Present-day English PGrmc. Proto-Germanic PIE Proto-Indo-European PL plural PP prepositional phrase PREP preposition PRON pronoun PRTC participle PRES present PRET preterit S subject SG singular SUBJ subjunctive mood SUP superlative SOV subject-object-verb word order SV subject-verb word order SVO subject-verb-object word order SVX subject-verb-other parts of sentence word order T tense THM thematic vowel TMA tense-modality-aspect TVX topic-verb-other parts of sentence word order V verb V2 verb second V vowel VO verb-object word order VP verb phrase WGrmc. West Germanic WSax. West Saxon XP variable phrase XSV others parts of sentence-subject-verb word order N

NOM

General abbreviations XVS

> < Ø *

other parts of sentence-verb-subject word order changes to, becomes derives from no ending reconstructed form, ungrammatical form spelling

xxi

I Periods 1 Periods: Pre-Old English 1 2 3 4 5 6 7 8

Introduction Origins of English The phonology of Proto-Germanic Phonology: Proto-Germanic to Pre-Old English Morphology Syntax Summary References

Abstract The topic of this chapter is the pre-historic stage of English as it developed between the 5th-century Germanic migrations to Britain and its first attestations in the 7th century. The beginnings of Old English are situated with respect to the language’s closest West Germanic relatives as well as to its Indo-European linguistic heritage. The phonological system is traced from Indo-European through Proto-Germanic and West Germanic stages with a focus on those innovations that occurred during the pre-Old English period. Brief descriptions of Indo-European and Proto-Germanic morphological structure provide the basis of the sketch of pre-Old English morphology, while both phonological and morphological changes that later obscured these systems in the early development of English are illustrated. The chapter ends with a brief discussion of the development of pre-Old English syntax.

1 Introduction Though this chapter is titled “Pre-Old English”, there was, of course, no clear-cut division between the attested Old English (OE) language and what came before. What is meant by “Pre-Old English” here is the pre-historic stages of English, that is, the Germanic language spoken in Britain after the migrations of Germanic speakers (Germani) from their continental home, but prior to the language’s first textual transmission, i.e. from the 5th to the 7th century. The Northwest branch of Germanic from which English descends is only meagerly attested prior to and during this period in the form of runic inscriptions. We must therefore base our sketch of pre-Old English on comparative reconstruction of other Germanic and even Indo-European languages and then interpolate the specific features of this stage of the language using the first attestations of Old English. Thus the role of the Germanic linguistic inheritance on Old English will figure prominently in this chapter.

2 Origins of English The English language owes much of its character to its ancestry in the Indo-European (IE) family of languages. The Germani’s ultimate homeland is attested by a number of Alex Bergs and Laurel Brinton 2012, Historical Linguistics of English (HSK 34.1), de Gruyter, 1–18

2

I Periods classical sources including Caesar, Tacitus, and Jordanes, all of whom describe Germanic tribes living in northern Europe and along the North Sea coast. The Goths’ late 2nd century migration toward the Black Sea left the remaining northern and western branches of Germanic to develop separately. It was the tribes that remained along the shores of present-day Germany, southern Denmark, and the Netherlands after more southerly West Germanic tribes had pushed toward the Danube and the Alps that formed the linguistic stock of what would become Old English, Old Frisian (first attested from the 13th century), and probably some of Old Saxon as well (attested from the 9th century). This broad dialect group is referred to as North Sea Germanic or Ingvaeonic and the term “Anglo-Frisian” refers to the Ingvaeonic sub-grouping from which English, or at least dialects of it, derived. Archaeologists have observed a continuity of cultural artifacts between areas of Germanic settlement in Britain and those in the settlers’ original homelands on the continent and there are linguistic parallels as well which link Anglian, Kentish, and, to a lesser degree, Northumbrian dialects with Old Frisian (see Nielsen 1989: 53–65 for an overview of scholarship on these parallels). Generally speaking, this group of dialects was more innovative than the rest of West Germanic, likely due, at least in part, to the social upheaval and ensuing linguistic contact that was precipitated by the migrations and subsequent settlement of Britain. While the North Sea linguistic ancestry of the OE dialects is undisputed, there has been a more recent scholarly movement to trace some of the innovations seen in OE and Middle English (ME) texts to Celtic influence in Britain. Only about a dozen Celtic loanwords survive in Old English, these being mostly place names and names of geographical features. Traditional scholarship held that it was only under limited linguistic contact that Celtic could have had so little influence on the Grmc. dialects. But new scholarship suggests that Romanized Celts and Germanic people probably lived in close contact, sharing cultural items and communicating with each other in the languages of the invading Germanic tribes. Some scholars propose a contact situation in which the Celts, though far outnumbering the Germanic settlers, learned the language of the Germanic speakers imperfectly. The large ratio of bilingual Celts to invading Germanic speakers, in conjunction with the Celts’ limited access to Germanic, would have resulted in the Celts imposing a number of features of their native language onto their second language (pre-Old English) (see, for example, van Coetsem 1988: 7–45, 83–91; Thomason and Kaufman 1988: 35– 63; Guy 1990: 48–54, for discussions of the social and linguistic circumstances that foster imposition of first-language features onto the second language, instead of borrowing). The sheer number of partially bilingual Celts would have nearly ensured transmission of those features into the following generations of British Germanic speakers. For an introduction to the current scholarship in this area, see Flippula et al. (2002: 5–26). Many of the Old English grammatical handbooks treat the phonological and morphological development of Germanic from its IE ancestor. Among these are Luick (1964a [1914–21], 1964b [1929–40]), Wright and Wright (1925), Campbell (1959), Brunner (1965), Hogg (1992), and Hogg and Fulk (2011). More detail on the sound changes and morphological structures of the early Germanic stages are presented in Prokosch (1939), Krahe and Meid (1969), the essays in van Coetsem and Kufner (1972), and, more recently, in Ringe (2006). In the sections to follow we present an overview of both the features that English inherited from its Germanic ancestors and the changes

1 Periods: Pre-Old English

3

which occurred in the intervening periods that gave Old English its particular character. The phonology section consists of a description of the features which West Germanic inherited from its IE ancestor, followed by a description of the specific developments which occurred during the pre-OE period. Since changes in the phonology also had an impact on the morphology of the language, those changes will be introduced in the phonology section. The morphology section will provide an overview of the development of morphological categories and structure from the ancestors of Old English. Syntax, being considerably more difficult to reconstruct without substantial attestation, will comprise a final, brief, section of the chapter.

3 The phonology of Proto-Germanic 3.1 The consonant system The linguistic change that is most commonly used as a marker of the Germanic (Grmc.) language family is the First Germanic Consonant Shift – also referred to as Grimm’s Law – in which the entire system of IE stop consonants is alleged to have shifted. The version of the shift presented here is the traditional one and that most commonly assumed today. For details of an alternative reconstruction of the IE stop system and its ensuing shift into Germanic, the reader is referred to the Glottalic Theory proposed separately by Gamkrelidze and Ivanov (1995) and by Hopper (1973). The consonant system of late western Indo-European is traditionally reconstructed as having had the stops shown in Table 1.1. Table 1.1: Late western Indo-European stops

voiceless voiced voiced aspirates

labials

coronals

velars

labiovelars

p (b) bh

t d dh

k g gh

kw gw gw h

The First Germanic Consonant Shift shifted the IE voiceless stops, *p, *t, *k, *kw, to fricatives, f, θ, x, xw. The IE voiced stops, *b, *d, *g, *gw, then shifted into the vacated position of the voiceless stops, p, t, k, kw, and the voiced aspirates, *bh, *dh, *gh, *gwh, shifted to voiced fricatives, β, ð, ɣ, ɣ w. The consonants of later Latin loanwords did not undergo the same shifts as the native Grmc. consonants and therefore often demonstrate a more transparent reflex of the original IE stop. For example, the native Germanic development of the IE roots *de´km ‘ten’ and *bhra¯´ter ‘brother’ illustrate the Germanic consonant shift, while Latinate loanwords from these same roots do not show these shifts: IE *de´km > PGrmc.*texun (OE tı¯en) vs. Latin decimal; IE *bhra¯´ter > PGrmc. *bro¯´θar (OE bro¯þor) vs. fraternity (< Old French < Latin, in which IE *bh > Lt. f ). The accent of Indo-European was a pitch accent whose placement was morphologically and lexically determined. When the original IE pitch accent had preceded a medial voiceless stop, the stop spirantized to a voiceless fricative as predicted by Grimm’s Law, e.g. IE *bhra¯´ter > PGrmc. *bro¯´θar. But when a high pitch accent had followed the stop, a major exception to the expected outcome occurred. Presumably

4

I Periods due to the slack vocal folds and comparatively low pitch of the preceding unaccented syllable (D’Alquen 1988: 17–20; Page 1998: 186–188; Petrova 2004: 376–381), the fricative was perceived as voiced instead of voiceless as in IE *pate¯´r > PGrmc. *fa´ðe¯r. This exceptional voicing, known as Verner’s Law, affected all voiceless fricatives including */s/. Thus, IE *ge´us appears in OE ce¯osan ‘to choose’ (with /s/), but in the PGrmc. 1P ´ m, where the accent had followed the fricative, the /s/ was voiced to /z/ PL PRET *gusu and ultimately rhotacized to /r/ in Northwest Grmc. (cf. Section 4.1.5), thus OE curon ‘we chose’. The effects of Verner’s Law are evident in all of the Grmc. languages, though its appearance is much more restricted in Gothic. The resulting system is shown in Table 1.2. Following the application of Verner’s Law, the IE accent shifted to the root syllable. This increase in energy and duration of the root syllable would be responsible for enormous changes from the inception of the Grmc. languages through the Modern period. A number of these are described in the sections which follow. Table 1.2: The early Germanic consonant system

stops, voiceless fricatives voiceless voiced nasals liquids glides

labial

dental

alveolar

p f

θ

t s

k x

ð n l, r

ɣ

b m

palatal

j

velar

w (labiovelar)

In addition to the singleton consonants, geminate consonants could also appear in postvocalic environments. These developed in Proto-Germanic through contact assimilations of adjacent consonants and resulted in a system that included geminate versions of all of the stops, nasals, liquids, /s/, and probably of both of the glides as well. The voiced fricatives that developed from both IE voiced aspirated stops and from the voiced output of IE voiceless stops through Verner’s Law hardened into voiced stops (b, d, g, gw) at various times according to dialect and phonological environment. When following nasals, ß and ð probably became stops within the Grmc. period. The process would have continued in later periods with word-initial and perhaps post-liquid environments. The fricative articulation was preserved the longest for *ɣ, while *ð eventually developed a stop articulation in all environments in the West Grmc. branch of languages. Goblirsch (2003: 111–119) provides a detailed review of the scholarship on the development of the voiced fricatives in English and Frisian.

3.2 Indo-European to Proto-Germanic vowels The late IE vowel system consisted of long and short i, e, a, o and u. Non-syllabic high vowels could combine with preceding vowels to produce the diphthongs a¯̆ i, a¯̆ u, e¯̆ i, e¯̆ u, o¯̆ i, and o¯̆ u. Liquids and nasals could also function as syllabic nuclei in Indo-European but were reinterpreted by the early Germanic speakers as short u + resonant, e.g. IE *wṛg- > PGrmc. *wurk-. These reflexes are highly visible in the third and fourth principal parts

1 Periods: Pre-Old English

5

of Grmc. strong verb classes III and IV where u + resonant developed from the earlier syllabic resonant of the root syllable (Murray, Chapter 17: Section 2.4).

3.2.1 Long vowels Germanic preserved the distinction between long and short vowels from IndoEuropean with some shifting of the quality of those vowels within their respective systems. In the long vowel system Indo-European *a¯ moved up to merge with the existing *o¯ and IE *e¯ moved downward toward [æ¯] (also called “e¯1”). The vacated e¯ position was filled in Germanic by a monophthongization of IE *e¯i and by a front vowel with a relatively limited distribution that demonstrated *ı¯~*e¯ alternations in North Sea and North Germanic. The resulting e¯ is often referred to as e¯2. Original IE *ı¯, *o¯, and *u¯ remained as phonemes into Proto-Germanic. Original *ı¯ was reinforced by a monophthongization of IE *ei. Thus the PGrmc. long vowel system was as in Figure 1.1. ¯i

u¯

e¯ 2

o¯

¯ (e¯ 1) æ

Figure 1.1: Proto-Germanic long vowels

3.2.2 Short vowels In the short vowel system, movement was in the other direction so that IE *o and *a merged unconditionally into the existing *a, which continued into West Germanic. IE *e shifted to *i, and *u to *o in Proto-Germanic, but the application of these shifts was dependent on the following segments, demonstrating the developing preference for harmony between the stressed root vowel and the following vowels. Clusters of nasal + C preferred preceding high vowels, so nasal clusters facilitated the shift of *e to *i, but blocked the fall of *u to *o. We also see the effect of two kinds of distance vowel assimilations at this stage, a raising umlaut, “i-umlaut”, conditioned by a following high vowel or glide and a lowering umlaut (often called “a-umlaut”) conditioned by non-high back vowels. Thus the PGrmc. shift of *e to *i occurred unless an *a or *o followed in the next syllable. PGrmc. *u shifted to *o unless a nasal or *i followed and PGrmc. *i sometimes shifted to *e under similar conditions. The outcomes of some of these shifts are particularly evident in the principal parts of the OE strong class III verb. When a nasal follows the root vowel, as in PGrmc. *bendan-, *band, *bundun, *bundan, it is responsible for raising the *e of the present stem to i as well as for preventing the u of the past participle from being pulled to o by the a in the following syllable. Thus class III pre-OE principal parts bindan, band, bundun, bundan, but helpan (with no raising of e to i before l ), healp (cf. Section 4.2.3 for a discussion of the diphthongized vowel), hulpun, holpan (with lowered root vowel). The resulting system of short vowel phonemes appears in Figure 1.2.

6

I Periods i

u e a

Figure 1.2: Proto-Germanic short vowels

3.2.3 Diphthongs The Grmc. reflexes of the IE diphthongs suggest that the first element of the IE long diphthongs had generally shortened prior to subsequent Grmc. developments. Both IE *o¯i and *oi, for instance, developed into PGrmc. *ai. IE *e¯i and ei are exceptions to this pattern. As described above, *e¯i became e¯2 and *ei became PGrmc. *ı¯. IE *eu was retained into Proto-Germanic and was joined by a new diphthong, *iu. Thus, the PGrmc. diphthongs were *ai, *au, *eu, and *iu.

4 Phonology: Proto-Germanic to Pre-Old English The North Sea Germanic dialect which would develop into English and Frisian was differentiated from surrounding dialects by a number of phonological and morphological features. The consonantal features of the North Sea dialects that differentiated them from the rest of Germanic were the seeds of velar palatalization and a generalized loss of nasals before voiceless fricatives. The vowel system of this group also developed differently both through shifts in the quality of inherited vowels and in how vowels were affected by neighboring sounds. The most significant of these developments are outlined below.

4.1 Consonantal changes 4.1.1 Geminates The Grmc. inventory of geminate consonants was bolstered by the output of West Germanic consonant gemination, an innovation of the West Grmc. branch (with traces in North Germanic) that resulted from the effects of the resonants *j, *w, *l, and *r, on preceding consonants (except *r) following a short vowel, e.g. PGrmc. *lagjan > West Grmc. *laggjan; *wilja > *willja; *bitr- > *bittr-; *apl- > *appl-; *nakw- > *nakk-. As unstressed final vowels were reduced and lost, some originally medial geminates came to be word final. These tended to be simplified gradually in the early West Grmc. dialects. Degemination continues throughout the OE period where we find doublets with both geminate and simple final consonants, e.g., cynn ~ cyn, will ~ wil.

4.1.2 Palatalization Palatalization of velar consonants in the environment of adjacent (originally) front vowels is a feature shared by all of the North Sea Grmc. dialects, but whether the process began during an early period of their relative unity continues to be debated. If it did, then it is probable that pre-palatal *k developed an allophonically palatalized

1 Periods: Pre-Old English

7

articulation during this period and only later developed fully phonemicized assibilated phonemes in Old English, Old Frisian, and Old Saxon. (The reader is referred to the discussion of the OE palatalization of velars in Murray, Chapter 17: Section 3.3.)

4.1.3 The Pre-Old English consonant inventory PGrmc. ð was closed to d in all of West Germanic and β had become a stop in most positions by prehistoric English, while the fricative articulation of ɣ persisted intervocalically. Thus, the PGrmc series of voiced fricatives had become a series of voiced stops at this stage. The voiceless velar fricative had also begun to change. It had weakened to [h] word-initially and between sonorants and vowels, where it ultimately was lost. Its effect on preceding vowels (described in Section 4.2.3) suggests that it may also have been weakening in other coda positions during this period. The early English inventory had been enriched by West Germanic consonant gemination which produced geminates of all original (i.e. non-palatalized) stops, fricatives, nasals, and liquids except r. Geminate f appeared as in Old English and geminate glides usually combined with the preceding vowel to form sequences of diphthong + singleton glide. Thus the pre-OE system would have had the singleton consonants shown in Table 1.3. Table 1.3: Pre-Old English simple consonants labial stops voiceless voiced fricatives nasals liquids glides

p b f m

dental

alveolar

θ

t d s n l, r

palatal

velar k ɡ x

j

w (labiovelar)

4.1.4 Nasal loss and compensatory lengthening All of West Germanic underwent a loss of postvocalic nasals before PGrmc. *x (< IE *k). The North Sea dialects extended it to apply to postvocalic nasals before any voiceless fricative. The nasal cluster had the usual raising effect on preceding vowels, but the loss of the nasal conditioned a compensatory lengthening of the preceding short vowel. In Old English the nasalized, now lengthened, a¯˜ appears as o¯. Thus we have from late PGrmc. *sanft, gans, kunþs > OE so¯ft, go¯s, and cu¯þ with nasal loss, compensatory lengthening, and raising of the original a to o¯. This change is responsible for the now opaque relationship between PDE bring and brought, the latter having gone through the intermediary stages PGrmc. *branx-te > West Grmc. *bra¯˜ xte > OE bro¯hte.

4.1.5 Rhotics and their effects West Grmc. languages are also marked by a rhotacism of the IE *s that had undergone voicing to /z/ as a result of Verner’s Law. Proto-Germanic already had a rhotic which

8

I Periods was presumably coronal in articulation (see Denton 2003: 15–16, 19–30 for a discussion of the articulatory qualities of early Grmc. rhotics and their articulatory effects in OE dialects). The rhotacism of z eventually led to a merger with the original r which is visible in the third and fourth principal parts of strong verbs that had originally had medial s voiced through Verner’s Law (see examples in Section 3.1). But in word-final or unstressed position, rhotacized z was lost. This loss makes for a difference between the 1/2P SG personal pronouns in North Sea Germanic and those in the rest of West Germanic: OE DAT SG me, þe compared to OHG mir and dir.

4.2 The vowel system By the end of the Grmc. period, the short vowel system had only one low vowel phoneme and the mid back vowel only existed as an umlauted allophone of /u/ before mid and low vowels. Following this period the short vowels were further modified by their phonological environments and the long vowel system was enriched by the monophthongization of the PGrmc. diphthongs. By the start of the OE period PGrmc. *ai had become long a¯ and *au had become long æ¯a, written . Reflexes of PGrmc. *eu and *iu (long or short) remained largely distinct in Mercian and Northumbrian dialects (appearing as eo and io, respectively), but had merged in most environments in the earliest West Saxon (WSax.) texts, though both spellings remained. The long high and mid vowels (ı¯, e¯2, u¯, and o¯) continued into the Pre-OE period, but e¯1 underwent different developments in the various OE dialects, appearing as æ¯ in West Saxon and as e¯ elsewhere. When followed by a nasal, it appeared as o¯ in all of Old English. The WSax. æ¯ was irregularly retracted before mid and high back vowels in the following syllable (referred to as u-umlaut), but otherwise continued as a front vowel.

4.2.1 Anglo-Frisian brightening and retraction A similar process of allophonic split before nasals occurred with PGrmc. short *a (< IE *a and *o). While it shows up as a in the rest of West Germanic, in the Anglo-Frisian area it originally developed two allophones, a back variant before nasals and a front variant everywhere else. The back variant is alternately spelled either or in Old English as in the doublet mann ~ monn. Since this nasalized vowel was distinct from the existing o and eventually merged with a in most of Old English, it probably had a quality similar to [ɔ]. The non-nasalized variant, [æ], is the output of the process called “Anglo-Frisian brightening” (AFB) or “First Fronting” in the entire Ingvaeonic area. This variant behaved as a true front vowel, diphthongizing in Old English breaking environments (cf. Section 4.2.3) and palatalizing some velar consonants, e.g. PGrmc. gastiz > (AFB) gæste > (palatalization) g˙easte > (i-umlaut, cf. Section 4.2.4) g˙ieste. A subsequent process of retraction, however, pulled the front vowel back to a before an immediately following w in all dialects, while Anglian also retracted the front vowel before rC, and Northumbrian before lC.

4.2.2 Restoration of a The result of Anglo-Frisian brightening was subject to the early mutating effects of back vowels in the following syllable. This process, known as “restoration of a”, foreshadows

1 Periods: Pre-Old English the back vowel umlauts of the early OE period and is visible in the OE masculine and neuter a-stem paradigms. When the root vowel of the nominative/accusative, genitive, and dative singular is æ, as in dæg˙ ‘day’, dæg˙es, dæg˙e, it appears as a in the plural due to the retracting effect of the suffixes’ back vowels: NOM/ACC dagas, GEN daga, DAT dagum. The phonological conditioning and particular interactions of Anglo-Frisian brightening, retraction, restoration of a, and breaking are responsible for much of the dialectal variation in Old English and are discussed in detail, in the modern Old English grammatical handbooks (e.g. Brunner 1965: 38–46, 54–60; Luick 1964a [1914–21]: 122– 166; Campbell 1959: 50–64; Hogg 1992: 76–101; Lass and Anderson 1975: 59–69; Lass 1994: 39–44; Wright and Wright 1925: 38–68).

4.2.3 Breaking Breaking was a process of front vowel mutations conditioned by the following consonantism. Though preserved in its most regular form in the WSax. dialect of Old English, breaking appears to have applied with slight variation in all of pre-Old English and was even shared, in part, by Old Frisian and Old Norse. The graphic realization of the output of breaking in Old English was , , and from the high, mid, and low front vowels, respectively. Prior to breaking, the Grmc. languages had no short diphthongs which are typologically marked. Whether those sounds represented by short digraphs in Old English were true short diphthongs phonemically is part of a larger controversy, though OE metrics confirm that they were indeed distinguished from long diphthongs quantitatively. Stockwell and Barritt (1951: 14) first questioned the literal reading of these digraphs and many scholars since have argued that the output of breaking was monophthongal and that the addition of a central or back vowel grapheme indicated either a retraction of the original vowel quality or a secondary articulation on the following consonant. The entire controversy is nicely presented in brief in Lass and Anderson (1975: 75–83) (see also Murray, Chapter 17: Section 3.2), while more recently, White (2004: 58–59) takes up Daunt’s (1939) argument that the OE spelling pattern had Irish origins. Breaking appears to have applied only in stressed syllables in the pre-OE period and was more regular the lower and the shorter the vowel. Early textual evidence suggests that the breaking environments originally caused the front vowels to develop central- or back-vowel off-glides at the corresponding height, /i/ > [iu], /e/ > [eo], /æ/ > [æa]. The factors which conditioned breaking were the reflex of PGrmc. /x/, whether alone or followed by another consonant, r + C, w, and, to a lesser degree, l + C. The conditioning factors of breaking are generally assumed to have had back articulations that spurred a transitional glide between the palatal front vowel and the following consonant. Thus x and w are velar, the l of lC clusters are presumed velarized, and the r in rC clusters were also presumed to have had a back articulation similar, perhaps, to the Scots “burred” r. Howell (1991: 83–105) has refuted this assumption, demonstrating that the kinds of diphthongizations seen in breaking rarely occur with sequences of stressed front vowel + velar liquid or x in modern Grmc. dialects, though analogous diphthongizations are quite common when the coda contains a weakened, non-velar, articulation of these same segments. Howell adduces dialectal evidence that all breaking factors had less constricted, even vocalic, articulations in coda

9

10

I Periods positions and Denton (2003: 21–30) makes a similar argument about OE r on articulatory grounds. The preserved effects of breaking with limited retraction and restoration of a in West Saxon contribute to the marked difference in sound and appearance of this dialect compared to that of Anglian and Northumbrian. This fact was compounded by Anglian’s later monophthongization or “smoothing” of many of these diphthongs. The output of breaking has morphological consequences, perhaps the most obvious being the subclass of class III strong verbs which are characterized by a short root vowel followed by a liquid-consonant cluster, e.g., Pre-OE. *werþan, *wærþ; *helpan, *hælp > OE weorþan, wearþ; helpan, h(e)alp. Breaking before h ( y, o > œ, a > æ, though it also raised the low and mid front vowels with less regularity. (See Section 5.1 in Murray, Chapter 17 as well as the OE grammatical handbooks for a description of the conditioning and output of i-umlaut in Old English). The effects of i-umlaut are responsible for the root vowel change in the mutated plurals: NOM SG mann, go¯s, mu¯s; NOM PL menn, gœ̅ s ~ ge¯s, mȳ s < earlier *mann-i, go¯s-i, mu¯s-i Even more visible are the mutated vowels in the 2/3P PRES IND of strong verbs which originally had the suffixes -is and -iþ, respectively: c˙e¯osan 2/3P SG PRES IND c˙¯ıest < *ce¯osis, *ce¯osiþ; faran, færst < *farist, færþ < *fariþ. The entire system of class I weak verbs is subject to i-umlaut from the j of the -jan suffix that had originally marked the class. This fact is only evident when compared to the unumlauted root from which a class I weak verb was derived, e.g., early OE dœ̅ man ‘to judge’ < early West Grmc. *do¯m-jan-.

4.2.5 Stress and its immediate effects A dominating stress accent on the root syllable was one of the hallmarks of the Grmc. languages, but a secondary stress accent on polysyllabic words was also likely, though scholars don’t agree on the rules of its placement. What we can surmise about the placement of all stresses in early Germanic comes from the metrics of the first attested languages and from the manner and order in which medial and final syllables were reduced in the period prior to the first transmission of texts. This reduction began in word-final position with the loss of short unstressed vowels. Final unstressed diphthongs were monophthongized and unstressed medial vowels in open syllables were often lost prior to the earliest OE texts. Final nasals were also lost eventually, leaving the nuclear vowels exposed to a new round of reductions. Generally more resistant to loss were unstressed high vowels. Their ultimate loss in both unstressed final and medial syllables

1 Periods: Pre-Old English

11

was usually conditioned by a preceding heavy foot, i.e. a single stressed heavy syllable (CV̅ , CVV, or CVC(C)) or a stressed light syllable followed by another syllable. This conditioned loss was responsible for -u~-Ø and -i~-Ø alternations in the West Grmc. u- and i-stem nominal paradigms as illustrated by the OE examples in Table 1.4. The nominative singular of feminine o¯-stems also shows an alternation between -u after light stems and -Ø after heavy, e.g. giefu vs. la¯r. Table 1.4: High vowel loss in nominal paradigms i-stem, SG NOM ACC GEN DAT PL NOM/ACC GEN DAT

u-stem,

MASC

MASC

light

heavy

light

heavy

win-e (< -i) win-e (< -i) win-es win-e win-e, -as win-e, -as win-um

giest (< -i) giest (< -i) giest-es giest-e giest-as giest-as giest-um

sun-u sun-u sun-a sun-a sun-a sun-a sun-um

feld (< -u) feld (< -u) feld-a feld-a feld-a feld-a feld-um

While unstressed vowels tended toward reduction, the stressed root vowels of the North Sea Grmc. and pre-OE periods were affected in a different manner, being further differentiated through the assimilatory effects of following sounds. Sound changes of this type included breaking, retraction, and the umlauts described above.

5 Morphology 5.1 Morphological structure Pre-Old English inherited a large percentage of its word stock from Indo-European, though at least 30% of its lexicon may have come from other sources. Vennemann (2003: xiii–20) provides an introduction to some of the issues inherent in identifying IE origins for both Germanic lexemes and morphosyntactic structures while he provides background for innovative theories of early Germanic contact with non-IE languages. Germanic did inherit much of its morphological structure from its IE ancestor along with its primary word-formation processes of compounding and derivation. IE roots were monosyllabic CVC structures with slots for a nasal, liquid, or glide on either side of the vowel. While the consonants of the root remained fairly stable across morphological categories, the radical vowel could vary between the e- or “full” grade, the o-grade, and the zero-grade, according to the word’s function. This kind of vowel alternation is the source of many of the ablaut, or root vowel substitution, patterns in the Grmc. languages. Various types of suffixes could be added to the root and these also could display different vowel grades within the same consonantal context. The position of the IE accent was partially dependent on the grade of the root and of its suffix. Thus, the overall structure of all IE morphology was that of root + suffix + inflectional ending, the combination of root + suffix constituting the stem. This structure continued to be the foundation of the nominal, adjectival, and verbal classes of the Grmc. languages.

12

I Periods

5.2 Nominal morphology 5.2.1 The case system The IE nominal categories were of three types: the nouns and adjectives, the demonstrative pronouns, and the personal pronouns. All of these were marked for case and number. Indo-European probably had eight cases which indicated the relationship of the noun phrases in the sentence both to each other and to the action of the verb. These cases were the nominative, accusative, dative, ablative, locative, instrumental, and vocative. Four cases were preserved in West Germanic with a vestige of a fifth. The nominative remained the case of the subject of the sentence (as well as of predicate nominals), but was also used for direct address. The accusative remained for the direct object, duration of time, and extent of space. The dative expressed a less direct impact or reference of the action of the verb on a noun such as with the indirect object, motiontoward, and many of the functions of the original locative and often the instrumental. The genitive was used for possession, for a part of a larger whole, and with certain adjectives and prepositions. In Old English and Old Frisian the instrumental forms were only preserved in the demonstrative and interrogative pronouns and in the strong adjective declension.

5.2.2 Gender Late Indo-European had three genders which were preserved into the OE period: masculine, feminine, and neuter. While these categories probably had some connection to real-world physical and/or cultural characteristics at one time, by early Germanic they largely served only a grammatical function. Masculine and neuter nouns, pronouns, and adjectives are closely related and always share a number of endings. The difference between the two genders often lies in the way in which the nominative and accusative singular are marked. While these categories may be formally differentiated in the masculine, they are always identical in the neuter singular, a feature that goes back to Indo-European (e.g. NEUT NOM and ACC SG hit ‘it’, scip ‘ship’). In the most common declensions the IE feminine differed from the masculine and neuter in the presence of the feminine’s *a¯ theme vowel, which developed into PGrmc. *o¯. Forms of other feminine declensions were also heavily influenced by the *o¯-stem endings. Though in the earliest OE paradigms the endings of the feminine are wholly different from those of the masculine and neuter, the endings of all three genders of all stem types derived from a single set of inflectional endings.

5.2.3 Nouns and adjectives Nouns and adjectives were indistinguishable in Indo-European and continued to share many of their endings into the OE period. It was the particular quality of the theme or stem vowel (or the lack thereof) which determined a noun’s or adjective’s class. Each class had a particular set of inflectional endings associated with it which likely derived from a single set of endings for all noun classes in Indo-European. The differences in inflectional endings that we see across the early Grmc. paradigms are the result of a combination of factors among which are

1 Periods: Pre-Old English

13

(1) variations in the original placement of the IE accent which resulted in different grades of the stem vowel and different developments of IE *s (which could appear as either OE s or r, depending on the application of Verner’s Law) (cf. Section 3.1); (2) the coalescence of stem vowels with the endings; and (3) the Grmc. reductions of final syllables. An example of the masculine a-stems, feminine o-stems and masculine and feminine n-stems illustrate this development below. The reconstructions have been simplified somewhat for clarity in Table 1.5. Table 1.5: Reflexes of the Proto-Germanic case inflections in Old English declensions PGrmc.

a-stem

o-stem

i-stem

u-stem

n-stem

SG

MASC

MASC

FEM

MASC

MASC

MASC

NOM ACC

-V-z -V-m

stan < az stan < am

gief-u < o¯´ gief-e < o¯´n

win-e < iz win-e < im

sun-u < uz sun-u < um

GEN

-V-so

stan-es < a´so

gief-e < o¯z

win-es < iza

sun-a < auz

DAT

-V-i

stane- < ai

gief-e < ai

win-e < ı¯

sun-a < au

nam-a < o¯n nam-an < anam nam-an < in(e/a)z nam-an < ini

NOM

-V-z(ez)

gief-e < o¯z

win-e < ı¯z

sun-a < iuiz

nam-an < anez

ACC

-V-nz

gief-e < o¯´nz

-V-n

gief-e < o¯n

win-e < (NOM PL) win-a < io¯n

sun-a < uns

GEN

stan-as < o¯z(ez) stan-as < (NOM PL) stan-a < o¯n

DAT

-V-miz

stan-um < amiz

gief-um < o¯miz

win-um < imiz

sun-um < umiz

nam-an < anunz nam-an < anon nam-um < anmiz

PL

sun-a < o¯n

The classes that became dominant in Germanic were the a-stems, o¯-stems, and the weak n-stems. The Grmc. i- and u-stems, though still viable, were no longer robust since many of their former members had moved over to a- and o¯-stem declensions. Other minor classes could be marked by reflexes of a consonantal suffix added to the root or by a lack of theme as in the athematic or root nouns, the latter marked in Old English by i-umlauted root vowels. (See von Mengden, Chapter 18: Section 1.2.8 for a description of these.) The adjectives originally had the same thematic classes as the nouns, including a- and o¯-stems (as well as ja- and jo¯-stem subtypes), i-stem, and u-stems, but these were heavily influenced by the pronominal declension in the transition from Indo-European to Germanic. The resulting strong adjectival endings are consequently a mixture of the strong nominal and the demonstrative pronominal declensions. By the early OE period most adjectives of the minor declensions, as well as many nouns of the minor declensions, had moved over to those of the a- and o¯-stems and the ja- and jo¯-stems. A major innovation of the Grmc. languages was the development of a weak adjectival declension similar in form to the weak n-stems nouns. The weak adjectival suffix derived from IE *-en-/-on- which appears in Latin and Greek nicknames Cato (GEN Catonis) ‘smart/shrewd (one)’ < catus ‘smart, shrewd’ and Greek Strabo¯n ‘squint-eyed (one)’ < strabos ‘squint-eyed’. This suffix was probably used in Grmc. phrases like the precursor of OE se blinda mann to mean ‘the blind one, a man’. The individualizing character of this suffix was gradually grammaticalized into the form of the adjective that

14

I Periods appeared with definite noun phrases, while the strong adjectival suffixes came to be used with indefinite noun phrases (Jasanoff 2008: 205; Krause 1968: 175; Osthoff 1876: 120–133). All nominal elements of the noun phrase agreed in case, number, and gender. Comparative and superlative adjectives were formed from suffixation to the adjectival stem. Two West Grmc. suffixes, *-iz- and *-o¯z-, were responsible for the OE comparatives which both developed into OE -r- via Verner’s Law and subsequent rhotacism (cf. Section 3.1 and Section 4.1.5). The original -i- of the first of these caused umlaut of the preceding vowel. The OE superlatives -est and -ost/-ast also developed from two different suffixes, *-ist- and *-o¯st-, respectively, the first of which also caused i-umlaut of the preceding vowel. Thus Old English has root vowel alternations in some adjectives like eald/ieldra/ieldest and long/lengra/lengest, but not in others earm/earmra/ earmost. Adverbs were also derived from adjectives. (See von Mengden, Chapter 18: Section 1.3.4 for an overview of these and for a description of numerals.)

5.2.4 Pronouns Old English had two demonstrative pronouns and an interrogative pronoun that agreed with the other members of their noun phrase in case, number, and gender, though there was no gender distinction in the plural. These three pronominal paradigms derive from demonstrative pronouns in the earlier IE language and show substantial similarities in their endings. The main demonstrative was the se/þæt/se¯o ‘the, that (one)’ paradigm which served as both definite article and the unmarked demonstrative pronoun. Its suppletive merging of the IE *s- and *t- pronominal bases in a single paradigm was evident as far back as Greek and Sanskrit. A second, derivative, pronoun came to serve as the proximal demonstrative, þes/þis/þe¯os ‘this (one)’, which derived from the same IE *t- base as the other demonstrative, but with the addition of an *-s(s)- suffix following the vowel. This pronoun could also function both pronominally and as a determiner. Only Northwest Germanic has a paradigm of this particular construction. Both demonstratives preserved a fifth case form in the masculine and neuter singulars only. This was þy¯~þon in the se/þæt/se¯o paradigm and þy¯s in the þes/þis/þe¯os paradigm. These forms, though labeled “instrumental” in the grammars and handbooks, neither derive directly from an earlier instrumental case nor are they used only to express instrumentality. Rather they are used for adverbial and idiomatic expressions such as þy¯ geare ‘in that year’, þy¯ læ¯s ‘lest’, æ¯r þon ‘before that’, þy¯ ma¯re þy¯ … ‘the more, the…’ The Grmc. interrogatives have clear cognates in other IE languages, descending from IE roots in *kw-. Though the inflectional endings of this paradigm are quite similar to those of the demonstrative pronouns, the interrogatives differ in that they have no plural forms and the singular forms combine the masculine and feminine into a single, animate, category. Outside of the nominative in hwa¯, the endings of the masculine/feminine forms resemble those of the demonstratives. The neuter has a separate nominative/accusative singular form, but is otherwise identical to the masculine-feminine paradigm. The Grmc. first and second personal pronouns are also derived from IE material. Though Gothic demonstrates that the dual pronouns originally required agreement with dual verbal forms in early Germanic, by Old English the dual forms of verbs had been lost and dual pronouns therefore agreed only with plural verbs. Unlike the

1 Periods: Pre-Old English first and second person pronouns, the Grmc. third person pronouns are marked for gender as well as for case and number. This fact may be due to their origins in demonstrative bases that were, themselves, marked for gender. Indeed, the personal pronouns’ heritage is heterogeneous. Four separate bases fed into the development of the various Grmc. third person pronominal paradigms. Those responsible for the English system were first the IE deictic in *k- (PGrmc. *x), which developed into the OE singular forms in h-, i.e. he, hine, heo, his, etc. The second base derived from an IE demonstrative pronoun in *ei- ~ -i- that formed the base of the OE plurals in h- (the initial h- was probably added later by analogy to the singular forms in h-) (Lass 1994: 141). ProtoGermanic had no common pronoun for introducing relative clauses, but the indeclinable pronoun þe developed as the most common means of expressing this kind of syntactic relationship in Old English.

5.3 Verbal morphology The ablauting verbs of Indo-European are divided into seven classes of strong verbs in Old English, all of which employ ablaut in conjunction with suffixation to express differences in tense, mood, and number. A small group of anomalous verbs also derived from IE origins. Perhaps the most significant innovation in the Grmc. verbs was the shift of what was primarily an aspectual system to a binary one of present versus past tense. As with the nouns, the dual was lost from the verbs. The indicative was preserved as the mood of declaratives and the second person imperatives continued to express commands, but only in the present tense. The infinitive was a present tense verbal noun which retained case marking in the early Grmc. inflected infinitive forms of which Old English preserves the original dative case. The present participle was a verbal adjective whose form is cognate with that of other IE languages. A Grmc. subjunctive, used in both present and preterit tenses, was formed from the IE optative, which had expressed the wish of the speaker. In the Ingvaeonic languages the three persons of the indicative plural were collapsed into one and the subjunctive had only a single form in each of the singular and plural of both tenses. West Germanic retained only the active voice with vestiges of the passive. The passive/past participle was a passive adjective when transitive, but was simply preterit when intransitive. OE ha¯tte/ha¯tton ‘is or was/ are or were called’ preserves the original passive, while all other uses of a passive in Old English were periphrastic constructions using weorþan or be¯on/wesan with the passive participle.

5.3.1 Strong verbs In the West Grmc. languages each strong verb had four principal parts: a present stem, a preterit first and third person singular stem, a preterit plural and second person singular stem, and a passive participle stem. For at least the first five classes it is presumed that there was root stress on the first two principal parts and suffix stress on the third and fourth. This assumption is supported by the output of Verner’s Law visible in the alternation between root-final voiceless fricatives in the first two principal parts and voiced fricatives or stops in the third and fourth (Section 3.1). For a detailed discussion of the development of the PGrmc. strong verbs from both Indo-European and non-IE origins, the reader is referred to Mailhammer (2007).

15

16

I Periods

5.3.2 Weak verbs The weak verbs, though an innovation of the Grmc. languages, formed their stems from the same morphological material as the strong verbs, i.e. a root – often in the o-grade – followed by a stressed suffix. What identified the weak verbs was a dental suffix marking the preterit rather than a root vowel alternation. One of the more popular hypotheses regarding the origin of the dental suffix was that it was a grammaticalization of the preterit forms of the verb to do suffixed to the verbal stem. The formal similarity between the Gothic forms of the weak preterit suffix and those of the PGrmc. verb *don constitute the basis for this claim, e.g. PGrmc. 3P PL *da¯dun/dedun ‘they did’ compared to the Go. 3P PL PRET IND suffix -de¯dun. Germanic probably had four classes of weak verbs all of which contained verbs derived from other categories. The first class suffix was from IE *-eja- which became *-ja in Germanic. The palatal glide was responsible for both i-umlaut of the root vowel and for gemination of the preceding consonant (cf. Section 4.2.4 and Section 4.1.1, respectively) when the j was preserved into West Germanic. In the second and third singular present indicative, the imperative, and the entirety of the preterit system, the original glide merged with following vowels to form a high front vowel rather than a glide. Because i did not spur gemination of a preceding consonant as j did, we find no gemination in these forms of the verb, though i-umlaut is present in all forms of the weak class i. In the second class the suffix derived from IE *-a¯je-/-a¯jo- which developed to *-o¯i- in Germanic. At this stage there was no longer a trigger for gemination of the root consonant since the original *j had collapsed into a diphthong with the preceding vowel. Unlike in class one, there is therefore no gemination in class two and also no i-umlaut in the root. If the following inflection began with a vowel, then the thematic -o¯i- sequence generally became -ia- in Old English as in the infinitive PGrmc.*lufa¯je-onom > OE lufian. If the suffix were followed by a consonant, the glide of the diphthong was lost and the long, unstressed, o¯ developed regularly to a. Thus we find 2/3P SG PRES IND lufast, lufaþ in class two where we found fremest, fremeþ in class one. OE class three verbs have no reflex of a stem vowel, though one does appear in this class elsewhere in Germanic. Thus, the OE personal endings are added directly to the root. There is no trace of a fourth class of weak verbs in Old English and even the third class has been reduced through the migration of its verbs into the first two classes.

5.3.3 Preterit-present verbs The preterit-present verbs constitute a third system of verbs in the Grmc. languages. They began as strong verbs whose past tense forms developed stative, present-tense, meaning. The strong class I preterit singular, wa¯t < *wı¯tan ‘to see’ (cf. Lt. vı¯dere), for instance, came to mean ‘I know’ (presumably through the development ‘I saw’ > ‘I saw, therefore I know’). In order to express a past tense of this new meaning, speakers created a new preterit form using a dental suffix which has sometimes been associated with the dental preterit marker of the weak verbs, though there are other theories of its origin (see, for example, Prokosch 1927: 334–335). Thus the preterit-presents have a present tense which resembles a strong preterit and a new weak preterit tense. It preserves some of the more archaic endings of the IE perfect seen in the present singular forms in which the earlier -e of the 1/3P SG was lost in West Grmc., e.g., sceal, and in

1 Periods: Pre-Old English the -t of the 2P SG which predated the -st that predominates elsewhere, e.g., scealt. The preterit marker was added to the original perfect stem with no intervening stem vowel: wiste/wisse and wiston (< *wit-te and *wit-ton, respectively). Old English had preterit-present verbs for each of the first six classes of strong verbs.

6 Syntax Though case marking on noun phrases allowed early Grmc. word order considerable flexibility, the unmarked word order for main clauses was OV, an order inherited from Indo-European. Modifiers most commonly occurred after the phrasal heads. Prepositions far outnumbered postpositions and were closely related to verbal particles which also preceded their heads. In Pre-Old English and later these particles were often written as separate words, although by the OE period we consider them verbal prefixes. Between the Northwest Grmc. runic inscriptions of the 3rd-7th centuries and the first OE texts, SVO word order became somewhat more common and modifiers also commonly appeared before their heads (Lass 1994: 218–222). Pre-Old English must have experienced substantial variation in word order as the language shifted from OV, which dominated the Northwest Grmc. runic inscriptions, to allowing verb-second order, which was quite common in Old English (Lass 1994: 225).

7 Summary The history of the Germanic speakers who migrated to Britain is characterized by migration, ongoing social upheaval, and heavy linguistic contact. It is therefore not surprising that the Pre-OE period was one of substantial linguistic change. Indeed, Old English may be the most innovative of the early Grmc. languages in terms of sound change alone. Though social instability and linguistic contact continued to spur innovation in the history of English, the process of rapid differentiation began with the first waves of migration to Britain and can be seen in the first OE texts.

8 References Brunner, Karl. 1965. Altenglische Grammatik. Nach der Angelsa¨chsischen Grammatik von Eduard Sievers. Dritte, neubearbeitete Auflage. Tu¨bingen: Niemeyer. Campbell, Alistair. 1959. Old English Grammar. Oxford: Oxford University Press. van Coetsem, Frans. 1988. Loan Phonology and the Two Transfer Types in Language Contact. Dordrecht/Providence, RI: Foris. van Coetsem, Frans and Herbert L. Kufner. 1972. Toward a Grammar of Proto-Germanic. Tu¨bingen: Max Niemeyer. Curzan, Anne and Kimberly Emmons (eds.). 2004. Studies in the History of the English Language II: Unfolding Conversations. Berlin/New York: Mouton de Gruyter. D’Alquen, Richard. 1988. Germanic Accent, Grammatical Change and the Laws of Unaccented Syllables. New York: Peter Lang. Daunt, Marjorie. 1939. Old English sound changes reconsidered in relation to scribal tradition and practice. Transactions of the Philological Society 1939: 108–137. Denton, Jeannette M. 2003. Reconstructing Germanic *r. Diachronica 20(1): 11–43. Flippula, Markku, Juhani Klemola, and Heli Pitka¨nen. 2002. Introduction: Early contacts between English and the Celtic languages. In: Markku Flippula, Juhani Klemola, and Heli Pitka¨nen (eds.), The Celtic Roots of English, 1–26. Joensuu: University of Joensuu, Faculty of the Humanities.

17

18

I Periods Gamkrelidze, Thomas V. and Vjacheslav V. Ivanov. 1995. Indo-European and the Indo-Europeans. Trans. by Johanna Nichols. 2 vols. Berlin/New York: Mouton de Gruyter. Goblirsch, Kurt Gustav. 2003. The voicing of fricatives in West Germanic and the partial consonant shift. Folia Linguistica Historica 24(1–2): 111–152. Guy, Gregory. 1990. The sociolinguistic types of language change. Diachronica 7: 47–67. Hogg, Richard M. 1992. A Grammar of Old English. Vol. I: Phonology. Oxford/Cambridge, MA: Blackwell. Hogg, Richard M. and R. D. Fulk. 2011. A Grammar of Old English. Vol. II: Morphology. Oxford/ Cambridge, MA: Wiley-Blackwell. Hopper, Paul J. 1973. Glottalized and murmured occlusives in Indo-European. Glossa 7(2): 141–166. Howell, Robert B. 1991. Old English Breaking and its Germanic Analogues. Tu¨bingen: Max Niemeyer. Jasanoff, Jay. 2008. Gothic. In: Roger D. Woodard (ed.), The Ancient Languages of Europe, 189– 214. Cambridge: Cambridge University Press. Krahe, H. and W. Meid. 1969. Germanische Sprachwissenschaft. 7th edn. Berlin: Walter de Gruyter. Krause, Wolfgang. 1968. Handbuch des Gotischen. 3rd edn. Munich: Beck. Lass, Roger. 1994. Old English: A Historical Linguistic Companion. Cambridge: Cambridge University Press. Lass, Roger and John M. Anderson. 1975. Old English Phonology. Cambridge/New York: Cambridge University Press. Luick, Karl. 1964a [1914–21]. Historische Grammatik der englischen Sprache, Vol. I, Part 1. Stuttgart: Bernhard Tauchnitz. Luick, Karl. 1964b [1929–40]. Historische Grammatik der englischen Sprache, Vol. I, Part 2. Stuttgart: Bernhard Tauchnitz. Mailhammer, Robert. 2007. The Germanic Strong Verbs. Berlin/New York: Mouton de Gruyter. Nielsen, Hans Frede. 1989. The Germanic Languages: Origins and Early Dialectal Interrelations. Tuscaloosa/London: University of Alabama Press. Osthoff, Hermann. 1876. Zur Geschichte des schwachen deutschen Adjecktivums. Forschung im Gebiet der indogermanischen nominalen Stammbildung, Vol. II. Jena: Pa¨ltz. Page, Richard B. 1998. Verner’s Law. Beitra¨ge zur Geschichte der deutschen Sprache und Literatur 122(2): 175–193. Petrova, Olga. 2004. The role of perceptual contrast in Verner’s law. In: Curzan and Emmons (eds.), 371–408. Prokosch, Eduard. 1927. The Old English weak preterits without medial vowel. Publications of the Modern Language Association 42(2): 331–338. Prokosch, Eduard. 1939. A Comparative Germanic Grammar. Philadelphia: Linguistic Society of America. Ringe, Don. 2006. A Linguistic History of English: From Proto-Indo-European to ProtoGermanic, Vol. I. Oxford: Oxford University Press. Stockwell, Robert P. and C. Westbrook Barritt. 1951. Some Old English Graphemic-Phonemic Correspondences – ae, ea and a. (Studies in Linguistics, Occasional Papers, 4.) Norman, OK: Battenburg Press. Thomason, Sarah, Gray and Terrence Kaufman. 1988. Language Contact, Creolization, and Genetic Linguistics. Berkeley: University of California Press. Vennemann, Theo. 2003. Bemerkung zum fru¨hgermanischen Wortschatz. In: Theo Vennemann (ed.), Europa Vasconica – Europa Semitica, 1–20. Berlin/New York: Mouton de Gruyter. White, David L. 2004. Why we should not believe in short diphthongs. In: Curzan and Emmons (eds.), 57–84. Wright, Joseph and Elizabeth Mary Wright. 1925. Old English Grammar. Oxford: Oxford University Press.

Jeannette K. Marsh, Waco (USA)

2 Periods: Old English

19

2 Periods: Old English 1 2 3 4 5 6

Preliminaries A chronological delimitation of Old English The external history and internal development of Old English Language-internal development: the decline of inflections Summary References

Abstract This chapter offers a survey of the main linguistic changes that took place during the Old English period – from the Anglo-Saxon migration around 450 CE to the beginning of the Norman rule of England. Considering that the major features and developments on all linguistic levels will be presented in Section III in more detail, the present article sketches the most salient and important linguistic features of Old English and otherwise focuses on political and cultural events of the period, which had an impact on the development of the English language. Predominantly, these are events that lead to the emergence of new contact situations – such as the Christianization (Latin), the Viking raids (Old Norse) and the emerging Norman influence on the English court in the 11th century (French) – and the most important waves of literary productivity – e.g. King Alfred’s educational program and the increasing book production following the Benedictine Reform.

1 Preliminaries The term “Old English” refers to those varieties of Germanic which were spoken in Great Britain from the Anglo-Saxon migration around 450 up until the end of the 11th century. While the geographical delimitation of the Old English language is unproblematic, the chronological limits are more difficult to determine and to some degree based on convention. Before I describe some of the main developments and characteristics of the Old English period, the most relevant approaches to and motivations for defining a chronological starting point and end point of Old English will be discussed briefly (Section 2). Taking into consideration that the main characteristics of the different domains of linguistic description are discussed in more detail in Section III, they will remain in the background in this article. The largest part of this chapter (Section 3) will focus on those aspects or developments of Old English that are related with or influenced by the non-linguistic history of its speakers. However, Section 4 will deal with languageinternal developments. It will be shown in this context that, while the choice of external dates for period boundaries may of course be associated with salient linguistic features during the development of a language, the relation between the internal and external factors is nevertheless mutual: once a choice of period boundaries has become conventional, the typological characterization of a language (in a given period) is dependent on this choice, which may be useful, but by no means necessary (cf. Curzan, Volume 2, Chapter 79). Alex Bergs and Laurel Brinton 2012, Historical Linguistics of English (HSK 34.1), de Gruyter, 19–32

20

I Periods

2 A chronological delimitation of Old English 2.1 The beginning of Old English In the case of Old English, it is easier to determine the beginning of the period than its end. Nevertheless, there are three historical events that can reasonably be interpreted as starting points in different respects. One is the settlement of Germanic tribes in England in the middle of the 5th century, the second is their Christianization around 600, and the third is the date of the earliest surviving written records around 700. With the arrival of Germanic settlers in England in the middle of the 5th century, their varieties of Germanic develop independently from the varieties of the cognate tribes that have remained on the Continent. Although the differences between the varieties of the settlers and those on the Continent cannot have been too great at the time of the migration, it is the settlers’ geographic and political independence as a consequence of the migration which constitutes the basis for the development of English as a variety distinct and independent from the continental varieties of the West Germanic speech community (cf. Section 3.1). Because close relations with the Continent persist for a relatively long time after the arrival of the Anglo-Saxons, it takes another one and a half centuries until their conversion to the Christian religion constitutes the first landmark of an independent AngloSaxon history. The Christianization and its impact on the Old English language will be discussed below in Section 3.2. At this point it may suffice to mention that, because the conversion is the first major change in the society and culture of the Anglo-Saxons that is not shared by the related tribes on the Continent, it is similarly significant for (the beginning of) an independent linguistic history of English as the settlement in Britain. Moreover, the immediate impact of the conversion on the language of the AngloSaxons is much more obvious than that of the migration: first, the Latin influence on English grows in intensity and, perhaps more crucially, enters new domains of social life; second, a new writing system, the Latin alphabet, is introduced, and third, a new medium of (linguistic) communication comes to be used – the book. Finally, one could approach the question of the starting point of Old English from a modern perspective. It is only indirect evidence that gives us a clue about the linguistic consequences of the two aforementioned events. Our direct evidence of any characteristic of (Old) English begins with the oldest surviving written sources containing Old English. Apart from onomastic material in Latin texts and short inscriptions, the earliest documents written in Old English date from the early 8th century. A distinction between a reconstructed “pre-Old English” before 700 and an attested “Old English” after 700, as drawn e.g. in Thomason and Kaufman (1988: 265), therefore does not seem implausible. While this criterion for determining the beginning of Old English is mainly based on a change in our modern access to the earliest stages of English (reconstruction vs. written evidence), some aspects of Anglo-Saxon history may in fact play a role here: as Hogg (1992b: 6) points out, it is feasible that the shift from a heptarchy of more or less equally influential Anglo-Saxon kingdoms to the cultural dominance of Northumbria in the time after Christianization may be connected with the fact that texts are produced not exclusively in Latin, but also in the vernacular. In other words, we may speculate (but no more than that) that the emergence of the earliest Anglo-Saxon

2 Periods: Old English cultural and political centre in Northumbria in the 8th century may lead the AngloSaxons to view themselves as one people rather than as different Germanic tribes, and, accordingly, to view their language as English (or, Anglo-Saxon) rather than as the Saxon, Anglian, Kentish, Jutish, etc. varieties of Germanic (but cf. an opposing point raised below in Section 3.1). Evidence for this change in attitude may be the composition of the Historia ecclesiastica gentis anglorum by Bede in 731, in which the gens anglorum of the title comprises all Germanic inhabitants of Britain and not only the Angles. Yet, as we can see from the earliest written evidence, the English language is not only sufficiently distinct from its closest cognates on the Continent around the year 700 with respect to its structure, its lexicon, and its phonology, it is at this point also considerably heterogeneous in itself – and continues to be so. I would therefore propose that, whatever happens to the language of the AngloSaxon settlers in Britain and for whatever reason it happens, any development after 450 should be taken as specifically English and before 450 should be taken as common (West) Germanic. That our knowledge of the underlying developments is necessarily based on a different method of access before and after around 700 is ultimately secondary to the relevant linguistic changes themselves and for any categorization of Old English.

2.2 The end of Old English The end point of Old English is marked by the Norman Conquest of 1066. The accession to the throne by William of Normandy in December of that year is considered a landmark in the history of England and, thus, of the community of speakers of English. It should, however, be questioned whether, or to what extent, the events of the year 1066 have only a symbolic value for the history of England rather than constituting an actual break. As will be discussed below (Section 3.6), in terms of the development of the contact situation between (Norman) French and English, the immediate relevance of the Norman Conquest is by far smaller than the prominence of this date in both the history books and in the handbooks on the history of English might suggest. The major linguistic changes that may be taken as relevant for a distinction between Old and Middle English take place in different linguistic domains – and, accordingly, at different times. The prototypical morphosyntactic features of Middle English – increasing syncretism of inflectional distinctions – begin to emerge as early as in the 10th century, whereas the typical Old English lexicon – largely Germanic with a moderate share of Latin borrowings but hardly any Romance elements – continues to exist up until the end of the 12th century and is attested even in Early Middle English texts such as The Owl and the Nightingale or Lagamon’s Brut; cf. Lutz (2002). In sum, if we define the Old English period as ranging from 450 to 1100 we mainly follow conventions. The distinction between Old and Middle English cannot be said to be motivated by sufficiently significant linguistic criteria and it is largely arbitrary to refer to the Norman Conquest in this context. Whether we determine 1066, 1100, or 1150 as the endpoint of Old English, it has become conventional to draw the line between Old and Middle English around this time. Strictly speaking, i.e. if we follow purely linguistic criteria, the transition from Old to Middle English expands over the period from the end of the 10th century to the end of the 12th century.

21

22

I Periods

3 The external history and internal development of Old English 3.1 The Anglo-Saxon migration The 7th-century historian Bede reports that in the year 449 Angles, Saxons, and Jutes, “de tribus Germaniae populis fortioribus” – ‘from the strongest tribes of the Germanic people’ (Bede, Historia Ecclesiastica 1.15; see Colgrave and Mynors [eds.] 1969: 50) – come to Britain and settle there. It has been mentioned above that the migration of the Anglo-Saxons means the beginning of a new speech community whose linguistic and non-linguistic history is now independent of the related tribes and their language(s). The main significance of the Anglo-Saxon migration therefore lies in the geographical reorganization of the Germanic speech community leading to the emergence of a new society and hence to a new, separate continuum of varieties. Although we do not know much about the social and political organization of these early Germanic settlers, some considerations are possible in this context. It is, on the one hand, reasonable to assume that the people involved in the migration speak varieties that originate in one and the same Continental Germanic dialect continuum. On the other hand, we do not know how close their varieties are or even to what extent there is mutual intelligibility among the settlers. But it is largely irrelevant how heterogeneous the language(s) of the Anglo-Saxons are at the time of their conquest. What we can reasonably assume is that, upon arrival in Britain, all the different groups involved in the settlement view themselves as speaking the same language in contradistinction to the Celtic inhabitants they encounter on the island. The identity of the settlers in their new homelands is necessarily based, among other things, on their common (albeit probably not quite uniform) linguistic background. The migration leads, moreover, to a new group identity of a subset of speakers of Germanic, who, irrespective of the heterogeneity of their own language(s), distinguish themselves both from the earlier inhabitants of Britain and from their relatives on the Continent. It is plausible to argue that it is this sociopsychological aspect which, more than anything else, constitutes the birth of the English language. Accepting this, we can assume, in turn, that it does not take too long after the settlement before the Anglo-Saxons view their version of Germanic as noticeably distinct from other Germanic varieties spoken by those who have stayed behind on the Continent. The migration itself does not immediately trigger any major change in the linguistic system. The earliest linguistic changes that English does not share with the cognate Germanic languages seem to be, at a first glance, independent of non-linguistic events. In the earliest period of Old English, a larger rearrangement of the phonological system takes place that affects mainly, but not exclusively, stressed vowels. (For a detailed description see Murray, Chapter 17; also cf. Campbell [1959: 50–112] and Hogg [1992a: 76–218] and for a shorter overview cf. Hogg [1992c: 100–119].) Two circumstances are employed in dating these sound changes. One is that other Germanic languages do not seem to have been affected by these changes. And secondly, the earliest written sources of Old English provide evidence that the relevant sound changes must have been completed by the date of their composition. Both these facts together suggest that all these sound changes take place in the time between 450 and 700.

2 Periods: Old English Although this dating is undisputed, the question nevertheless arises whether it is plausible to assume that, within the relatively narrow time frame between 450 and 700, the phonological system is considerably rearranged whereas it remains relatively stable for significantly more than 250 years both before and after this period. Any attempt to answer this question would have to remain speculative to some degree. But it is feasible to assume that the new social and political identity of the settlers leads to some cross-adaptations in the linguistic usage of the settlers. That dialectal differences among the settlers persist and, as we know, never cease to exist, does not exclude the possibility that some of the heterogeneity of these dialects is levelled out as a consequence of the formation of a new community. And since many of the obvious dialectal differences of a continuum are phonological, it is plausible that these adjustments take place predominantly on the phonological level. Bearing in mind the narrow time frame of the early Old English sound changes, it is therefore not implausible to assume that they have been enforced, if not triggered, by the formation of a new speech community as a consequence of the migration.

3.2 Christianization From around 600 onward, the Christian religion is spread quickly across the country both from the north, under the influence of the Irish Church, and from the south, by St. Augustine of Canterbury and his missionaries, sent to England by Pope Gregory I in 597. As indicated above, the Christianization of the Anglo-Saxons has immediate consequences on two domains of the English language – the lexicon and the writing system. Language contact with Latin is not a new phenomenon in those days. There has always been a moderate-to-intensive exchange of words between Latin and Germanic, ever since the two languages were neighbors on the Continent. Even after the migrationto England, the Anglo-Saxons adopt some Latin words, although the exact settings of these bilingual contacts are unclear. Plausible contexts are continuing relations with the other Germanic tribes on the continent, scattered speakers of Latin who have stayed in Britain after the Roman armies withdrew in 410, or speakers of Celtic who either use Latin as a lingua franca in communication with the Anglo-Saxons or whose language contains itself words borrowed from Latin which are then passed on to the Anglo-Saxons. But the words that are imported into the English language in the course of and after Christianization are of a considerably different kind than earlier, predominantly common-Germanic loans from Latin. While earlier Latin loans are words related to trading, to the military, or expressions for every-day concepts like household devices, the vocabulary imported with Christianization mostly denotes either concepts immediately related to the new religion and its institutions (e.g. abbod ‘abbot’, alter ‘altar’, munuc ‘monk’) or, generally, more abstract concepts. Another difference from earlier Latin loanwords is the medium through which they are introduced. Because many words come into English via books rather than via oral communication, the words are transferred from written registers of Medieval Latin rather than from Vulgar Latin, which is the source language of the pre-Christian borrowings. For the linguistic observer (rather than for the speaker of Old English), there is another notable difference. Earlier loanwords participate in most or all of the major sound changes that take place in the earliest period of Old English. That is, with respect to their phonology, the pre-Christian words from Latin have adapted to the

23

24

I Periods phonological system of Old English in their written attestations. By contrast, the more recent Latin words, introduced through Christianization, often retain a Latinate shape. The different phonological characteristics are best exemplified by Latin words that are borrowed both in the early and in the later period and that, accordingly appear in different forms in the Old English records. Thus, Lt. calix ‘chalice’ occurs in Old English in an older form, celc and a later form, calic. Likewise, OE leahtrice and OE lactuce are both borrowed from Lt. lactuca ‘lettuce’ at different times. (See Wieland, Chapter 23.) A far-reaching side-effect of Christianization is the introduction of the Latin alphabet in Anglo-Saxon England. Originally introduced as the medium on which the new religion is based, it soon comes to be used for the composition of other texts not immediately related to the Christian faith. Moreover, from about 700 onwards texts written in Latin script are composed in or translated into the English vernacular. Although never widespread in the early Middle Ages, through the introduction of the book and of Latin writing literacy enters new domains of social life which have not been reached by the use of the older writing system, the runic Futhorc. The first vernacular texts are mostly written in the north of England, in the monastic centers of Northumbria and Mercia. More than in the south, Christianity in the north is influenced by its Irish variety. The Irish use the Latin script themselves, although with a considerable number of stylistic alterations. Consequently, some of the Latin characters used in England in the early Middle Ages are of a slightly different shape than those used on the Continent. The ‘yogh’ or ‘insular G’, shaped 〈ᴣ〉, is used in Anglo-Saxon manuscripts to represent the voiced velar stop /ɡ/, the velar fricative /ɣ/ and the glide /j/. Only in the late Old English literature do we find instances of the ‘Carolingian G’, shaped 〈ɡ〉, then used to distinguish the stop from the fricative. The phonemes /æ/ and /æː/ are represented by an ‘ash’ 〈æ〉, a character shaped as a ligature of Latin 〈a〉 and 〈e〉. Other Latin letters used in Anglo-Saxon manuscripts differ considerably in shape from the variants we use today in printing. Three other characters are either not or indirectly taken from the Latin alphabet. Two of them certainly originate in the Futhorc, the set of Runic letters that is used by the Anglo-Saxons before the introduction of Latin writing: 〈ƿ〉 (‘wynn’) represents the bilabial glide /w/. Although the character 〈w〉 is usually employed in modern editions of Old English texts, Anglo-Saxon manuscripts use 〈ƿ〉 almost throughout, and only occasionally 〈u〉 or 〈uu〉 for /w/. When Old English is represented by the Latin alphabet, ‘thorn’ 〈þ〉 and ‘eth’ 〈ð〉 are used indistinguishably as allographs for both the voiced and the unvoiced dental fricatives, /θ/ and /ð/. While the former is also adopted from the Futhorc, the origin of the latter is uncertain (cf. Hogg 1992c: 75).

3.3 The Vikings in England The raid on the Lindisfarne monastery in 793 is the first known instance of a series of increasingly intense attacks on England by Vikings. While initially local plundering rather than attempts at gaining political influence motivates the raids, their quality and purpose change, and by the middle of the 9th century, a large area of England comprising almost all parts of Northumbria and Mercia is under Danish control. When King Alfred of Wessex comes to the throne in 871, his West Saxon kingdom is the only autonomous area of what once used to be the Anglo-Saxon heptarchy. Alfred succeeds in protecting his own territory from Danish rule and also in re-conquering the western

2 Periods: Old English parts of Mercia. Moreover, as a consequence of his military success he manages to negotiate a truce resulting in the Treaty of Wedmore in 878. In this treaty Alfred and the Danish leader Guthrum agree on a borderline between an area of Danish legislation (the Danelaw) and an independent Wessex. The linguistic consequences of the Viking rule in England are difficult to measure. Because of the division of England into a Danish and an English political sphere, the situation in the 10th century is as follows: in the south, a relatively stable and peaceful political situation allows Alfred to instigate an educational reform. He supports the import of new books from the continent and the production of new books in England, and he also initiates the translation of Latin books into English and the production of books written in the vernacular. Thus, for the remaining two centuries of the AngloSaxon period, the vast majority of sources comes from the area of the West Saxon kingdom. In the north, by contrast, Viking influence on English is naturally much stronger. Therefore, judging from surviving Old English texts alone, the evidence of language contact between Old Norse and Old English in England is quite small – about 150 words of Scandinavian origin are attested in the Old English sources. Judging however from Middle English evidence, we may assume that there must have been a very intensive contact situation, at least in the area of the Danelaw (cf. Kastovsky 1992: 320). This can be seen from the type of loanwords and their features, rather than from the mere number of words of Scandinavian origin in the English of today. In contrast to the Latin loanwords introduced through the influence of Christianization, Scandinavian loanwords in English are less technical and much more part of the basic vocabulary of English. Words like skirt, egg, or sky, for example, denote everyday concepts rather than more abstract and learned concepts as represented by the Latin borrowings. Much more efficiently than Latin words, Old Norse loanwords seem to have been integrated into the basic vocabulary with a high token frequency. An example is Old Norse tacan ‘take’ which replaces the highly frequent Old English niman and which is one of the few borrowed verbs that is (even today) inflected as a strong verb. Similarly, the verb get is, if not borrowed from Old Norse, at least influenced in its phonological shape, as the Old English equivalent gietan was pronounced with an initial glide, i.e. /'jetɑn/, and would have resulted in PDE *yet rather than get. Moreover, more than the French influence during the Middle English period, Old Norse at some points enters grammatical structures of English, as we have traces of Old Norse in the pronominal paradigms (e.g. they, them, their) and in the inflectional paradigms (e.g. -s 3P SG for southern -[e]þ). It should also be noted that Old Norse and Old English are quite close cognates. It is impossible to say whether or to what extent the two languages are mutually intelligible in the 9th and 10th centuries, but many words seem close enough. Townend argues that there is no full comprehensibility between Old Norse and Old English, but that there is what he calls “adequate” or “pragmatic intelligibility” (Townend 2002: 181–183), i.e. a degree of comprehension that allows for basic conversation, but that does not cover the full morphosyntactic complexities of the two systems. Whether or not two etymologically equivalent lexemes can be confused, it is more crucial that, in some cases, their phonological differences cannot be distinguished by the spelling conventions of Old English. What matters therefore is to what degree, given the close relatedness of the two languages, the spelling conventions of Old English allow us to identify traces of

25

26

I Periods linguistic transfer. In other words: how can an Old Norse loanword be distinguished from its Old English cognate in a nearly contemporary source? For instance, the fricative /ʃ/ is represented in Old English manuscripts by the digraph 〈sc〉, but this is at the same time the combination of letters which serves best to represent the Old Norse cluster /sk-/. As a consequence cognate pairs like Old Norse /'skyrte/ (PDE skirt) and Old English /'ʃyrte/ (PDE shirt) are impossible to distinguish by the Old English spelling: both forms would appear as 〈scyrte〉, and it is therefore impossible to say when the Scandinavian loanword skirt entered the English language.

3.4 Alfred’s educational reform King Alfred’s contribution to the history of the English language is twofold. One aspect is that, as a result of his military success, the Danish conquest of England comes to a halt (Section 3.3). The claim that Old Norse would have become the major language of England if the Danish troops had also occupied the south of the country must necessarily remain speculative, but it can at least be assumed that Old English would have been a threatened language and that, had Alfred not defended a stable (in political and military terms) English-speaking area, English would have lost its role as a predominant language in England. There is another achievement of King Alfred which may not influence the development of English with the same intensity as the contact with Old Norse does, but which influences considerably our knowledge of the English language in the early Middle Ages. Alfred’s educational reform is the impetus for a considerable increase in the production (and import) of books in general, and in the production of written literature in Old English in particular. Alfred himself initiates the translation of a number of important and influential Latin texts: Gregory the Great’s Cura Pastoralis and Dialogi, Augustine of Hippo’s Soliloquia, Boethius’s De consolatione philosophiae, Paulus Orosius’s Historiae adversus paganos and, perhaps most important of all, Bede’s Historia ecclesiastica. The degree of Alfred’s personal participation in the translation process varies (and remains disputed), but it can be said that all these translations result from his education policy. Moreover, a number of vernacular texts are composed in the same context, again with varying degrees of Alfred’s personal involvement, e.g. the Anglo-Saxon Chronicle, which is continued in several versions up until the 12th century, and a Martyrology. As a result, up until the end of the Anglo-Saxon period the vast majority of texts of various genres is produced in the royal court or in the monastic centers of the West Saxon kingdom, most of all in the capital Winchester. For the first time in the history of Old English a large corpus of long prose texts is produced, and for the first time there is evidence of a more standardized written language mainly based on features of the West Saxon dialect, but not without traces from the variety attributed to Mercia (cf. Section 3.5). Up to the end of the Old English period, a large number of Old English documents originate in West Saxon or are heavily infiltrated by features of the West Saxon dialect. The evidence from the Old English sources for our knowledge of the history of the English (spoken) language is therefore clearly misleading, because the varieties spoken in the Midlands (i.e. what in Anglo-Saxon times was the Mercian variety) contribute much more to the development of Present Day (Standard) English than those of West Saxon do.

2 Periods: Old English

3.5 The Cluniac reform Half a century after Alfred’s reign, the flourishing book production receives a further impulse. This is triggered by a movement that affects the ecclesiastical history of England rather than, as in the case of Alfred’s contribution, the political history, although the two are tightly connected. A monastic reform movement aiming for a stricter and more ascetic interpretation of the Benedictine Rule initiated in the monasteries of Cluny and Fleury (later Saint-Benoıˆt-sur-Loire) in France spreads to England in the middle of the 10th century. One of the central figures in the Cluniac reform movement in England is Æthelwold, Bishop of Winchester 963–984. Educated at the court of King Æthelstan (reigned 921–939), one of Alfred’s brothers and successors on the West Saxon throne, Æthelwold continues what Alfred has begun. Although Alfred and Æthelwold carry out two different reform projects, both instigate a revival of literary and intellectual productivity after a period of a constant Viking threat has, at least in this region of England, come to an end. While Æthelwold’s commitment is religious rather than political, he can be sure of royal support for his work. The school founded by Æthelwold at the New Minster in Winchester produces not only books in high numbers. The texts that emerge from the scriptorium at Winchester also have a remarkable stability in orthography and also, as far as we can judge, in their vocabulary. From this, it has been deduced that the language of many documents composed or copied in Winchester and in related scriptoria represents the first attempt at a standardization of the English language. It is, on the one hand, obvious that West Saxon texts from the end of the 10th century onwards often show a remarkable uniformity in the choice and orthographic form of words. But it should also be taken into consideration that we are dealing with a set of texts covering a limited range of scholarly fields. It would be problematic to deduce the existence of a genuine standard language from the relative homogeneity of the Winchester texts alone. Indeed, the very fact that the documents representing “Standard Old English” all derive from a tight network of authors and instigators in a predominantly monastic context – all in all a rather small, albeit influential, group of people – speaks against rather than in favor of the wider use of their linguistic features outside these circles. It is therefore justified to speak of orthographic conventions characteristic of the Winchester school, perhaps of a West Saxon Schriftsprache, but it is difficult, if not impossible to make judgments about the scope and influence of the Winchester conventions. The impression that the “Winchester standard” spread widely is not only owed to the political situation described in the Section 3.4, i.e. to the fact that other dialect areas were to a larger or smaller extent excluded from the production of books – at least judging from the material that has survived until today – it is also due to the fact that the most productive single author of the Anglo-Saxon period, Ælfric of Eynsham (c.955–c.1010), is a disciple of Æthelwold. Therefore, it is perhaps more appropriate to say that a large share of West Saxon texts of the later Old English period (and thus of the corpus of Old English texts in general) has some roots in the Winchester tradition rather than that the Winchester tradition contributes to the standardization of Old English. A few general caveats should be expressed in this context. First, not many of the texts that have survived until today can be said to be pure representatives of one specific variety. Perhaps such a statement would apply most to prototypical representatives of the variety of 10th century Northumbria, such as the Lindisfarne Gospels and the

27

28

I Periods Durham Ritual. But pure West Saxon texts are rare, for several reasons, such as the fact that many of the West Saxon texts, including Winchester documents, are copies of older, lost versions and attest to an earlier layer of Mercian linguistic features. Second, West Saxon cannot have been a dialect as uniform as the descriptions of the Old English dialects often suggest. Kastovsky (1992: 346) points out that the traditional distinction between “Early West Saxon” for the language prototypically represented by the Alfredian translations and “Late West Saxon”, prototypically represented by Ælfric’s texts, is misleading, as it suggests a mere diachronic distinction between two varieties within only a little more than a hundred years. Rather, it should be assumed that the differences between the two groups of documents are diatopic at least to the same extent as they are diachronic. Third, as indicated above, the idea of an Old English standard language presupposes not only a process of deliberate regulation; it also requires a broader distribution of a standard language in larger parts of the population. As to the last point it should be noted that perhaps the term “standard”, introduced in this context by Gneuss (1972), is the main problem in this context. If a standard is understood as an institutionalized variety that, among other things, serves as a means of communication bridging several local and social differences in the usage of a language, the hypothesis of a Late West Saxon standard involves two problems. First, it is not falsifiable, because we have no clues as to how widely a deliberately regulated variety may have made its way outside the scriptoria. And second, the idea is implausible because it is not clear how a variety attested in a number of specialized scholarly texts should have spread into other areas of society given that literacy was limited to a rather small elite. What is plausible, though, and for this we do have evidence, is that there is an influential intellectual elite which has an enormous impact on the literary productivity in late AngloSaxon England, and who seem to have used the language of their works in a deliberate and comparatively uniform way. (See further Kornexl, Chapter 24.).

3.6 Old English and Old French In addition to the impetus for the production of literature, the import of the Cluniac reform into England must also be seen in a different context of the history of English. It has briefly been discussed earlier (Section 2.2), that French influence on English begins gradually and not abruptly with the Norman Conquest. The monastic reform is in fact the first instance of contact between speakers of French and English. That the Gallo-Romance vernacular is perceived as sufficiently distinct from Latin can be deduced from the explicitly trilingual character of the Oaths of Strasbourg of 842. Although the Oaths can hardly be employed as a data source for the Old French language, it nevertheless attests to the fact that the French vernacular is considered an idiom independent from any variety of Medieval Latin. That this applies not only to the perspective of the speakers of early Old French but also to that of the Anglo-Saxons around the turn of the millennium is confirmed by an English source composed in 1011 by Byrhtferth of Ramsey, i.e. his Manual or Enchiridion, in which Byrhtferth makes a remark on the correct use of Latin versus French (Byrhtferth, Enchiridion 2.1. 449–454; see Baker and Lapidge [eds.] 1995: 88–90). In spite of the fact that French and Latin were without doubt two distinct idioms in the late 10th century, it is nevertheless impossible in many cases to distinguish clearly whether Romance material in English documents is of Latin or of French origin. For

2 Periods: Old English this reason alone there are hardly any clear traces of contact between French and English. Yet, traces do exist: particularly in 11th-century (but pre-Conquest) glossaries, we do occasionally find French words among the Old English interpretations. Most of them, provided they are unambiguously of French rather than of Latin origin, are attested only once, so that we cannot assume that they have ever been part of the English lexicon. Only two such words are clearly French and are attested more than once: capun ‘capon’ in the Antwerpen and Brussels glossaries and iugelere ‘magician’ occurring several times in different glossaries to Aldhelm’s De laude virginitatis and, notably, once with a different spelling in an anonymous homiletic text, i.e. not as a gloss but in a prose text (cf. von Mengden 1999). The period of the monastic reform is certainly the earliest date from which contact between speakers of the two languages is attested. It is plausible, therefore, to assume at least a slight degree of lexical transfer. The contact situation continues in the early 11th century when the relations of the Crown with the Duke of Normandy intensify, at the latest under the reign of Æthelred II (reigned 978–1016), who married Emma, the sister of Richard II, Duke of Normandy. That is, in both royal and ecclesiastical circles, there are tight connections between French and English speaking people around the year 1000. It is difficult to determine how far-reaching and how widespread contacts between these two groups are in England in the first half of the 11th century. But there is some evidence of Norman influence in pre-Conquest England. The occupation of official positions by native speakers of (Norman) French began, albeit on a small scale, as early as the 1040s, with Robert of Jumie`ges being appointed Bishop of London in 1044 (and promoted to Archbishop of Canterbury in 1051). And both of the two separate entries for the year 1052 in the Worcester Chronicle refer to a Norman castle in England (ChronD 1052.1 15 and 1052.2 2; see Cubbin [ed.] 1996: 70, l. 19 and 71, l. 6). These passages clearly imply that the Normans must have been numerous and powerful enough to erect their own military fortifications on English soil – some fifteen years before the Conquest. But this passage in the Chronicle at the same time attests to ongoing language contact: the expression used to refer to the fortification is in fact the first instance of the English word castle, in its Old English form castel. Its meaning and its grammatical gender (MASC) reveal that it must be a borrowing from Norman French and that it cannot be identical with the homonymous Latin loanword castel (NEUT) ‘town, village’. One should, of course, not overstate the linguistic transfer between French and English before the Conquest. But these aspects may suffice to support the point made above in Section 2.2, i.e. that the Norman Conquest as such did not have any immediate consequences for the English language. The events of the year 1066 seem to have been the consequence of a series of steps by the Norman nobility to gain political influence in England – a development always accompanied by support from an influential pro-Norman party in the Anglo-Saxon aristocracy. It is therefore feasible to assume that the intensity of French influence, although traceable, is not considerably greater in the years immediately following the Norman Conquest than it is before. From this perspective, the Norman Conquest stabilizes, but by no means ignites or reinforces, the growing intensity of Anglo-Norman relations. As such, William’s victory at Hastings may be seen as one of several important events that pave the way for the enormous influence that French exerts on English in the 13th and 14th centuries. The beginnings

29

30

I Periods of this development are clearly part of the history of Old English rather than of Middle English.

4 Language-internal development: the decline of inflections The previous section focused on those developments that were either triggered by language-external events or, at least, should be seen and explained in the context of the history of the speech community. A selection of instances of internal change taking place during the Old English period will be discussed briefly in the following. Old English is often described as an inflecting language. This label follows a particular classification according to morphosyntactic types of languages, which can be observed cross-linguistically. It is particularly prominent in the descriptions of Old English because it is motivated by the contrast with the analytic character of Present Day English. In this context, the classification goes back to Henry Sweet (1874: 160) who took the “full inflections” of Old English as the main defining criterion for his periodization of the English language (cf. the discussion in Curzan, Volume 2, Chapter 79). According to this approach, Middle English is the period with a “limited” set of inflectional categories. The decline of a complex system of case inflections begins, however, long before the period that we have defined as Old English. During Old English times, the dual number is retained only in some pronominal forms and disappears almost completely by the end of the Old English period. Case syncretism has been a continuous process at all times. While for Proto-Indo-European eight cases are reconstructed, all daughter languages of IndoEuropean have less than eight cases even in their earliest attested stages. At the beginning of the literary period of Old English, the merger of instrumental and dative has almost been completed, with distinctions retained only in some pronominal forms and in a few adjectives. But to say that during the Old English period a system of formerly five cases reduces to four cases, would again be a simplification, because in many forms of masculine and feminine nouns (neuter nouns never encode the distinction), nominative and accusative are not any longer distinguishable in 11th-century sources. Thus, in spite of the categorization as the “period of full inflections”, during the Old English period as many distinctively encoded case values get lost as in Middle English. The causes of this particular stage in the reduction of the case system are predominantly phonological. The fixed, initial stress characteristic of the Germanic languages generally entails unstressed final syllables. The result is that first front vowels begin to merge in final syllables in the earlier stages of Old English, a process naturally affecting many inflectional endings. Only in the 11th century does the phonetic reduction also affect back vowels. Thus the dative plural suffix -um is comparatively stable and even in its reduced forms /-on/ or /-ən/, it is still distinguishable from the other case/gender suffixes in most noun classes because of the nasal; cf. Hogg (1992a: 3n. 2). If we consider that the loss of case distinctions necessitates the (ultimately Middle English) replacement of the predominant V2 word order in Old English by a rigid SVO order, we can observe a long term development of cross-influences of various linguistic domains: intonation (Germanic initial stress) → phonology → inflectional morphology → syntax. Of these, Old English particularly observes the phonological changes, the syncretism of some of the inflectional markers and also the reduction (or loss) of some inflectional categories.

2 Periods: Old English The decline of morphological values during the Old English period is more substantial in the nominal paradigms than in the verbal forms. While verbal endings are affected by the phonological reduction too, the main difference between the nominal and the verbal system is that the reduction of verb endings does not result in a syncretism of inflectional values to the same extent to which it does in the nominal paradigms. Throughout the Old English period three person values are distinguished in the singular, but not in the plural. There is a general tense distinction between past and present. Finally an indicative and a subjunctive mood are distinguished morphologically. The most salient feature of the Old English verbal system is shared with all other Germanic languages: the distinction between strong and weak verbs. While weak verbs mark their inflectional values by suffixes, strong verbs use a combination of suffixes and systematic vowel alternations. In this context, the typical descriptions are again more idealized for Old English than they are for later stages of English. The relatively clear set of seven classes of strong verbs that we often find in handbook descriptions is, naturally, full of idiosyncrasies. Moreover, the traditional class distinctions are based on a set of criteria which are not completely consistent (cf. the more detailed discussion in von Mengden, Chapter 18, Morphology, Section 2.4). Again, the impression that the Old English system of strong verbs appears to be more regular and systematic than the Middle English system is certainly not wrong. Yet, it should be noted that handbook descriptions of the verbal system of Middle English tend to include variation among and within the paradigms of the verb classes, whereas equivalent descriptions of Old English focus more on their regularity. Therefore, although we can generally assign the label ‘inflectional language’ to Old English – irrespective of its wide diachronic and diatopic variation – it should at the same time be borne in mind that the decline of inflectional categories has been a continuous process since long before the Anglo-Saxon migration. If we refer to Old English as the period of “full inflections” (Sweet 1874: 160), the attribute “full” wrongly implies that the Old English system has reached the highest possible degree of morphological complexity – both from the point of view of the history of English and from a cross-linguistic perspective. Rather, the factors which justify assigning a new label – “English” – to a language whose history does not really have a beginning, and on the basis of which we have defined a starting point of Old English in Section 2.1, are of a sociolinguistic nature. But neither the beginning nor the end of Old English coincide with any salient changes in the inflectional system (cf. the discussion in Curzan, Volume 2, Chapter 79).

5 Summary Bearing in mind the various problems involved in both periodization and categorization discussed above, we may say that Old English is, particularly in contrast to later stages of English, a typically Germanic language in many respects. The share of inherited Germanic words in the vocabulary is much greater than it is today, and even the moderate share of Latin loanwords is a feature that applies to all Germanic languages in the early Middle Ages. Its syntactic (V2) and morphological (inflectional) features are similarly characteristic of Germanic. Finally, in spite of a major rearrangement of stressed vowels at an early stage of the Old English period, the phoneme inventory, too, is basically the same as that of the other early Germanic languages.

31

32

I Periods

6 References Baker, Peter S. and Michael Lapidge (eds.). 1995. Byrhtferth’s Enchiridion. Oxford: Oxford University Press. Campbell, A. 1959. Old English Grammar. Oxford: Clarendon. Colgrave, Bertram and R. A. B. Mynors (eds.). 1969. Bede’s Ecclesiastical History of the English People. Oxford: Clarendon. Cubbin, G. P. (ed.). 1996. The Anglo-Saxon Chronicle: A Collaborative Edition. Vol. 6: MS D. Cambridge: Brewer. Gneuss, Helmut. 1972. The origin of Standard Old English and Æthelwold’s school at Winchester. Anglo-Saxon England 1: 63–83. Hogg, Richard M. 1992a. A Grammar of Old English. Vol. 1: Phonology. Oxford/Cambridge, MA: Blackwell. Hogg, Richard M. 1992b. Introduction. In: Hogg (ed.), 1–25. Hogg, Richard M. 1992c. Phonology and Morphology. In: Hogg (ed.), 67–167. Hogg, Richard M. (ed.). 1992. The Cambridge History of the English Language. Vol. 1: The Beginnings to 1066, 1–25. Cambridge: Cambridge University Press. Kastovsky, Dieter. 1992. Semantics and vocabulary. In: Hogg (ed.), 290–408. Lutz, Angelika. 2002. When did English begin? In: Teresa Fanego, Bele´n Me´ndez-Naya, and Elena Seoane (eds.), Sounds, Words, Texts and Change: Selected Papers from 11 ICEHL, Santiago di Compostela, 7–11 September 2000, 145–171. Amsterdam/Philadelphia: John Benjamins. Sweet, Henry. 1874. A History of English Sounds from the Earliest Period. Oxford: Oxford University Press. Thomason, Sarah Grey and Terrence Kaufman. 1988. Language Contact, Creolization, and Genetic Linguistics. Berkeley: University of California Press. Townend, Matthew. 2002. Language and History in Viking Age England: Linguistic Relations between Speakers of Old Norse and Old English. Turnhout: Brepols. von Mengden, Ferdinand. 1999. Franzo¨sische Lehnwo¨rter im Altenglischen. In: Wolfgang Schindler and Ju¨rgen Untermann (eds.), Grippe, Kamm und Eulenspiegel: Festschrift fu¨r Elmar Seebold zum 65. Geburtstag, 277–294. Berlin/New York: Mouton de Gruyter.

Ferdinand von Mengden, Berlin (Germany)

3 Periods: Middle English 1 2 3 4 5 6 7

What is Middle English? Issues of evidence Writing and speech Grammar Lexicon Summary References

Abstract Middle English is the period in the history of English when variation is most thoroughly recorded in the spoken mode, and the body of surviving material is very large. In this chapter, the extralinguistic reasons for the survival of variation in Middle English are Alex Bergs and Laurel Brinton 2012, Historical Linguistics of English (HSK 34.1), de Gruyter, 32–48

3 Periods: Middle English given, focusing on the relationship between linguistic form and socio-cultural function. Middle English has come down to us directly through manuscripts, and indirectly through reconstruction and through the study of residualisms in Present-day English; the reliability of the evidence for Middle English is assessed. Each level of language in turn is then discussed: writing- and speech-systems (= transmission), grammar and lexicon. New resources (especially online) for the study of Middle English are flagged, offering exciting possibilities for future research. The aim of the article is to offer both a characterization of Middle English itself and a sketch of the resources available for its study.

1 What is Middle English? Middle English is the form of English spoken and written roughly between 1100 and 1500 CE. Its beginning roughly corresponds to the Norman Conquest of 1066, while its end roughly corresponds to the first book printed in English (1475) and the arrival of printing in England (1476): these two historical events, though of very different kinds, have implications for the status of the language during the Middle English period. Up until the Norman Conquest, English (i.e. Old English) had a distinct status as a language of record and for literary expression; it developed a written standardized form, classical Late West Saxon, which was used by scribes outside its area of origin, i.e. Wessex in south-west England (see Kornexl, Chapter 24). After the Norman Conquest, Middle English was displaced in prestige for literary and documentary manuscripts by Latin, and as a language of literary culture (and later record) by varieties of French: by Norman French (which in England developed into Anglo-Norman) up to the middle of the 13th century, and by Central French after that. English remained the speech of the majority of the population and continued to be widely written; but the written mode had a restricted function. If writers wished to communicate beyond their area of origin, or wished to leave a record for subsequent generations, they tended to use other languages than English. These restricted functions for English had linguistic implications, as we shall see. Towards the end of the Middle English period, English became more elaborated in social function, i.e. used more commonly as a language of record and literary function. The process was gradual: Latin remained common as a language of record and of “serious” literature well into the Early Modern period, and there are still remnants of Anglo-Norman in certain set phrases in the British parliament, e.g. La Reyne le veult ‘the Queen wishes it’ for public bills on receiving royal assent (even though parliamentarians began using English from the middle of the 14th century). But the efflorescence of English literature which we associate with (say) the Elizabethan and Jacobean theater could only have come about because it became possible to use English in a more elaborated way. Reasons for this change include the Reformation, which gave the vernacular a new role in religious expression, and the rise in power of the “middling” classes, notably the bourgeoisie, for whom English was the usual mode of expression (see Schendl, Chapter 32). Defining Middle English in linguistic terms is in some ways straightforward, with reference to every level of language traditionally identified: in lexicon, in grammar, in sound-system, in writing-system (see Section IV). Middle English differs from Old English in that it manifests large-scale borrowing of vocabulary from French, and in that Norse vocabulary, which had been mostly a feature of the spoken language in

33

34

I Periods Old English, appears in the written record. In grammar, Middle English is often referred to as the period of “reduced” inflexion, whereby the relationship between words and phrases is expressed to a greater extent than in Old English through element order, the use of prepositions, and through markedly distinctive pronouns; however, there is a wider range of inflexions than in Present-day English, notably with regard to the verb but also (in many varieties) in the adjective. A phonological distinction between voiced and voiceless fricatives, not made in Old English, appears in Middle English; and there are changes in the vowel-system, most notably in the loss of quantity as a feature distinguishing “short” and “long” vowels and in the configuration of the system of diphthongs. These developments are in many cases reflected in the writing system of Middle English, which is highly variable, in that spelling usages seem to have varied from parish to parish; variation in every level of language is reflected in writing. The variety of written Middle English clearly relates to its social function; if English had a local rather than a national function, it makes sense for written forms of Middle English to reflect the spoken mode quite closely, to help readers who would encounter it primarily at the first stage of literacy. Middle English similarly differs from Modern English. In vocabulary, the transition from Middle to Modern English saw the transfer of large numbers of items not only from French and Latin but also the beginnings of a flood of words from non-European languages, the result of trade and the beginnings of imperial expansion. During the transition from Middle English to the Early Modern English period, usually dated 1500–1700, the present-day pattern of inflexion emerged, while the sound-system underwent a radical change, most notably with regard to the long vowels of late Middle English: the Great Vowel Shift (see Krug, Chapter 48). Finally, variety in writing systems was gradually suppressed as the Middle English period segued into Early Modern English, first of all in writings designed for public consumption and eventually in private letters as well. Again, this suppression of variety related to the changing function of English; it made sense for a written language used for the purposes of record or national communication to develop a form which was less prone to variation. But such a straightforward characterization of Middle English, though useful as an initial outline, is a massive over-simplification of a highly complex period in the evolution of the English language, omitting the extralinguistic contexts which triggered these changes and the dynamic interaction of levels of language which took place at different speeds, and for different reasons, in different parts of the country. Old English did not become Middle English at midnight in the year 1100; Middle English did not become Modern English at midnight in the year 1500. Moreover, it is in many ways more accurate to consider the history of Middle English – as arguably of all English – as the history of Englishes, the history of (in Roger Lass’s useful phrase) “a population of variants moving through time” (Lass 2006: 45). The over-simple characterization offered at the beginning of this section also omits something else crucial: the fact that the primary evidence for Middle English is, though extensive, nevertheless limited. We know a great deal about Middle English – indeed, we are learning more about this state (these states?) of the language all the time – but there is also a great deal which we clearly do not know, and which it is likely we never will. In sum, the characterization just offered is a useful listing of prototypical features but not a history, i.e. a narrative which offers at least a partial explanation for – or, in April McMahon’s phrase, a “relief from puzzlement” about – the phenomena described

3 Periods: Middle English (McMahon 1994: 45). This chapter is designed to engage with current research in this field.

2 Issues of evidence The primary evidence for Middle English is much restricted in comparison with the massive resources available for students of Modern English. There are, of course, no soundrecordings of Middle English speech, which have been available to students only since the end of the 19th century; and there are no contemporary attempts at structured linguistic descriptions, as are available for scholars working in the Early Modern period. Students of Middle English interested in engaging with contemporary linguistic scholarship on the English language have to make do with the somewhat vague and general comments about linguistic differences made by literary authors, mocking the usages of folk from parts of the country other than their own, e.g. Chaucer’s representation of Northern speech in the Reeve’s Tale, or the humorous representation of Southern speech in the Wakefield Second Shepherd’s Play. There is, unfortunately, nothing in Middle English equivalent to the Old Icelandic First Grammatical Treatise, the sophisticated treatise on pronunciation of the vernacular produced in Iceland in the 13th century. Evidence for Middle English consists of material remains, most notably manuscripts written by scribes on parchment and latterly on paper, and (at the very end of the Middle English period) early printed books. Some texts – more than have until recently been recognized – were written on the walls of churches, especially towards the end of the period. Thousands of these objects have survived, even though it is likely that very many have been lost. Such texts were produced not only in the scriptoria of monasteries or cathedrals but also in private households, country vicarages and – increasingly so towards the end of the period – in workshops in towns, notably London. Many are now stored in the great libraries of the world: the British Library in London; the college libraries of Oxford and Cambridge; the Bodleian Library in Oxford and Cambridge University Library; the great private libraries of the United States, such as the Huntington and the Pierpont Morgan, and the libraries of universities such as Princeton and Yale; the libraries of Trinity College Dublin and the University of Durham; the John Rylands and Chetham Libraries in Manchester; the Bibliothe`que Nationale in Paris; Glasgow’s Hunterian Collection and Edinburgh’s National Library of Scotland. But there are also many manuscripts in less obvious places: in (e.g.) the private Takamiya collection in Tokyo, or in the muniment rooms of local councils, or, even today, of “great houses” such as Alnwick or Petworth. These manuscripts and early printed books contain texts ranging from major literary works (poems such as Chaucer’s Canterbury Tales or Langland’s Piers Plowman, or works of religious instruction such as sermons and saints’ lives) to more pragmatic and/or ephemeral material (e.g. works of medical reference, wills, letters). Such texts, of course, survive not through conscious selection of representative material, but by chance, often because of the idiosyncratic interests of the 16th- or 17th-century antiquaries who collected them, such as Sir Robert Cotton (whose manuscripts are now in the British Library) or Franciscus Junius (whose collection is now in the Bodleian). Whereas modern sociolinguists or dialectologists can select their informants on the basis of age or social class or precise locality, students of Middle English have to make do with what the vagaries of time have bequeathed to them.

35

36

I Periods Thus, the first task for any student of Middle English is always to assess the status of the primary evidence presented: the philological enterprise of textual analysis must precede engagement with larger questions of linguistic enquiry. Traditionally the philological study of Old and Middle English has been expressed through the editing of texts: “critical” and “diplomatic”. The critical edition has dominated textual scholarship since at least the middle of the 19th century, notably in the editing of Biblical and classical texts, and is generally referred to as “textual criticism”. The goal of textual criticism was the creation of the “critical edition”, whereby the “threads of transmission” were “follow[ed] back […] to restore the texts as closely as possible to the form which they originally had” (Reynolds and Wilson 1974: 186). The principles of textual criticism, which emerged first in the editing of Biblical and classical texts, were transferred to the editing of vernaculars, and some of the most sophisticated textual criticism has been carried out by editors of Middle English texts. Thus, the editors of the great Athlone edition of Piers Plowman, one of the principal (if controversial) achievements of 20th-century Middle English editorial scholarship, refer to the outcome of their enterprise as follows: “a theoretical structure, a complex hypothesis designed to account for a body of phenomena in the light of knowledge about the circumstances which generate them” (Kane and Donaldson 1974: 212). The focus of the critical edition, therefore, is the author’s intention; it is no coincidence that textual criticism flourished as the authorial voice became reified in Romantic and Victorian cultures, and critical editions, published in prestigious series such as those issued by the Early English and Scottish Text Societies (EETS, STS), continue to dominate medieval English literary studies. Textual criticism has developed a battery of techniques whereby extant materials are compared (“collation”), and whereby putative errors – more properly, non-original readings – are identified and used to establish relationships between texts. Once these relationships have been expressed through an appropriate modeling process, traditionally the “stemma”, the archetypal text is deemed to be established and can then be corrected in the light of the critical editors’ judgement, based on their knowledge of the linguistic and cultural contexts of the period in question. There is, of course, a clear parallel with the tree-diagrams of the 19th-century historical linguists, and it is no surprise that many such linguists combined the editorial and linguistic enterprise (thus the traditional meaning of the term “philology” in Anglophone countries). With the rise of new approaches to literary criticism from the 1930s onwards, however, there was a decoupling of the literary and linguistic traditions, and most recent EETS and STS publications are critical editions (albeit “conservative” ones), designed primarily for literary scholars and focused on the authorial intention for the work in question. Such editions are potentially problematic for the linguistic student of Middle English, since they are avowedly an abstraction from the surviving data, which is in almost all cases mediated through scribes. It is therefore often necessary for linguists to return to the original materials, but medieval manuscripts are of course restricted in terms of access; their fragility means that they cannot be easily transported from place to place, and there are also conservation issues to which librarians are rightly sensitive. Photographic images are thus commonly used, traditionally on microfilm or in published facsimiles but increasingly on the internet (for a useful resource, see the relevant section of the Middle English Compendium [McSparran (ed.) 2006]). However, photographic reproductions present

3 Periods: Middle English linguists with further problems. Publication of images, however accurate, does not offer any interpretation of the facts presented. For such an interpretation, “diplomatic” editing, demanding high-level hermeneutic skills (paleographical, codicological, linguistic) are needed: “What is necessary is that a single editor should spend the time necessary to solve the problems of the manuscript, even those that are themselves trivial and unimportant, and should find a means of presenting his [sic] results so that others may benefit from his [sic] pains; the job should be done thoroughly once, not superficially by each individual user of a facsimile” (Dobson 1972: xii–xiii). Diplomatic editions, especially when presented in machine-readable form, also allow for “text-mining”, whereby texts are searched for forms of interest and the results analysed using various statistical tools. Diplomatic editions are not often favored by literary scholars, and it is noticeable that EETS, for instance, has not produced diplomatic editions to follow on from the Ancrene Riwle series. Yet there are modern developments in what might be termed “computational philology” which are beginning to recuperate the diplomatic edition in Middle English studies, at least for those whose scholarly orientation is primarily linguistic. Computational philology applied to English historical linguistics has become particularly associated with the University of Helsinki, whose VARIENG project (Nevalainen, Taavitsainen, and Leppa¨nen 1998–) continues to make a massive contribution to the historical study of English and related languages, particularly in the area of grammar, and with the University of Michigan, whose Middle English Compendium (McSparran [ed.] 2006) is a text-resource linked to the Middle English Dictionary (Kurath et al. 1952–2001). But the Helsinki and Michigan projects have tended to utilise critically edited rather than diplomatic texts, which – though often suitable for broad-brush syntactic studies, or for establishing the Middle English lexicon – have limitations when used for detailed study of, say, spelling practices. It is therefore no coincidence that the most significant study yet published of written language usage during the Middle English period, the Linguistic Atlas of Late Mediaeval English (LALME) (McIntosh et al. 1986), derived its information not from critical editions but largely directly from manuscript sources, supported where possible by input from diplomatic editions (an online and revised version of the Linguistic Atlas, e-LALME is forthcoming; see http://www.ling.ed.ac.uk/research/ihd/index.shtml). Perhaps the most impressive new direction in fresh diplomatic editing in relation to computational philology is The Canterbury Tales Project (Robinson and Bordalejo 1996–), which began as an attempt to develop a new kind of critical edition but which ended up with arguably a much more valuable resource: not only up-to-date transcriptions and bodies of associated apparatus, which have replaced the old Chaucer Society diplomatic editions produced by F. J. Furnivall and his Victorian associates, but also their presentation in machine-readable form. Because machine-readable, such editions allow for highly sophisticated text-mining, identifying spelling-systems and systems of punctuation which are of interest not just for linguists but also for book historians and students of changing literacy practices. Such online electronic corpora, based on diplomatic editing, are now an established part of linguistic enquiry, underpinning (e.g.) the Edinburgh Linguistic Atlas of Early Middle English (Laing and Lass 2007) and Linguistic Atlas of Older Scots (Williamson 2007), more recent Helsinki initiatives such as the online Corpus of Early English Correspondence (Nevalainen et al. 1998) and Corpus of Scottish Correspondence (Meurman-Solin 2007), or the Stavanger-Glasgow Middle English Grammar Corpus (Stenroos et al. 2011).

37

38

I Periods Diplomatic editions of Middle English texts are focused not so much on the author of the text as on the transmission of the text; i.e. they engage with scribes as informants. This focus on scribal output underpinned the development of the LALME, for instance. Researchers have showed that texts, when appropriately analysed, can be used as evidence for the language not of authors but of scribes, and, since scribes are, like authors, native users of Middle English, and of course comparatively much more numerous, the information they provide can be investigated for the purposes of linguistic (as opposed to literary) study.

3 Writing and speech The textual resources described at the end of the previous section are, and will continue to be, invaluable for the study of what might be termed the “transmission” of Middle English, viz. the systems of writing and speech. The mapping between writing and speech has been a concern of scholars since antiquity. In the classical and medieval west, authorities such as Donatus and Priscian adopted the “doctrine of littera”, which was developed to correlate speech with alphabetic writing-systems of the kind used for Greek or Latin. This doctrine distinguished between nomen (‘name [of the letter]’), potestas (‘power’ = ‘sound-value’), figura (‘representation’ = ‘written-symbol’), with the term littera (‘letter’) as a superordinate classificatory term. Although this ancient usage has been recently recuperated, most notably by scholars associated with LALME and its successor-projects, most students of Middle English use the accepted terminology to discuss transmission which has been developed in the 20th century. Broadly speaking, written languages are either “phonographic”, where there is a mapping (however conventional) between grapheme and phoneme, or “logographic”, where there is a mapping between a conventional symbol and a word or morpheme. The relationship between these different systems is of course clinal. Towards the logographic end is Chinese, whose conventionalized characters derive ultimately from pictorial representations of certain key concrete concepts, though this practice was rapidly modified to deal with more abstract notions: “Modern Chinese characters hold few really firm clues as to their pronunciation” (Newnham 1971: 44). Written Middle English, on the other hand, represents the opposite end of the cline, whereas Present-day English, with its various conventionalizations, is, while remaining broadly phonographic, rather closer to the logographic pole (see Rutkowska, Chapter 15). It is sometimes said that Middle English was written at a time when folk “wrote as they spoke”. Such a statement, of course, oversimplifies a complex situation; given that writing-systems are designed to give comparative permanence to something evanescent, i.e. speech, a degree of conventionalisation is to be expected. Nevertheless, the statement does summarise, albeit in broad terms, the phonographic status of Middle English. Given the historic primacy of speech over writing, it is no surprise that, until comparatively recently, the focus of scholarship has been on the reconstruction of Middle English speech-systems, i.e. the phonology of Middle English. Significant recent discussions include survey articles written for the relevant volumes of the Cambridge History of the English Language, especially the second volume (Blake 1992), and there are also useful introductory outlines in (e.g.) Horobin and Smith (2002) or Wright and Wright (1923). The most comprehensive outlines remain those by Luick (1964) and Jordan

3 Periods: Middle English (1974); however, both of these date in their essentials from the early 20th century, and both are seriously in need of updating. Despite this established scholarly tradition, there are approaches which have focused on the written mode as an object of enquiry in its own right. The best-known of these approaches is associated with the creation and exploitation of LALME and its successor projects, treating graphological features as objects of study without necessary reference to their spoken-language equivalents. A simple example makes the point. There is some evidence, supplied by LALME, that the form itt ‘it’ has a geographical distribution in Middle English; it is a dominant form in several texts localized to northern England, notably Cumberland. But it seems very doubtful that the -spelling signifies any sound-difference from . A more subtle difference, again identified as a result of work for the Linguistic Atlas (Benskin 1982), is to do with the allographic representation of the grapheme . (In this context, the term “grapheme” is used as the written language equivalent of the phoneme, i.e. the symbolic unit being aimed at by the scribe, while an “allograph” is the realisation of the grapheme in writing. Replacement of one grapheme by another changes the meaning of the word in which it occurs, but replacement of one allograph by another realisation of the same grapheme does not change the word’s meaning. It is conventional to place graphemes in angle brackets, thus: .) In general terms, in northern varieties of Middle English, seems to have been written identically with , whereas in southern varieties the two graphemes are realized distinctly. Again, there does not seem to be a sound-difference here, although the restriction of = in some northern texts to “function” words such as yat ‘that’, ye ‘the’ (cf. southern þat, þe) may correlate with the development of initial-fricative voicing in such cases. Another example of a spelling-usage which can be localized dialectally is initial in xal ‘shall’, xuld ‘should’, which is diagnostically East Anglian. How this feature maps onto pronunciation is again problematic; a plausible argument can be made that such forms represent a pronunciation-difference, but alternatively the usage could simply be a convention with only a local currency. The “ = ” example, of course, problematizes some of the conventional terminology used for the discussion of the writing/speech relationship. Briefly put, the problem is as follows: assignment of the allograph y to either the grapheme or the grapheme depends on the mapping of the allograph onto the sound-equivalent: is the form in question, in context, to be interpreted as a vowel or a semi-vowel, or is it to be interpreted as a fricative? Seeing y and þ as allographs of a grapheme raises some theoretical issues about the assignment of allographs to graphemes: can one allograph, in this case y, belong to two different graphemes? It is an axiom of phonological theory, of course, that by definition one allophone cannot belong to two different phonemes (although cf. neutralization of phonemes, for which see e.g. Gimson 1989: 50). Similar problems arise with (e.g.) and ; as is well-known, is used prototypically, in Middle English texts, to map onto either a vowel or a consonant when in medial position, whereas is used initially, e.g. vpon ‘upon’, loue ‘love’. Which mapping is made is determined by context. In Present-day English, the grapheme is similarly differentiated, mapping onto a consonant [j] in initial position but a vowel in final position, cf. yacht, dizzy. Michael Benskin (1982) simplifies this problem of categorization by recuperating the ancient doctrine of littera, and setting aside as unnecessarily complex the apparently

39

40

I Periods obvious parallelisms between grapheme/phoneme, allograph/allophone. According to Benskin’s approach, the figura , in some varieties, may be assumed simply to map onto two potestates, or sound-equivalents, with the nomen “thorn”. Developed discussions of these issues, using this terminology, have been adopted by others working in the LALME tradition (e.g. Laing and Lass 2009). Although it is true that there are many features of Middle English writing-systems which do not seem to map onto sound-differences, there are also features where such a mapping is much more likely. When a Northern scribe writes stane for ‘stone’ and a Southern or Midland scribe writes stone it seems likely that some sound-difference is being addressed; when a Southern scribe writes voules ‘birds’ (cf. PDE fowls) and a Northern or Midland scribe writes foules, again it seems likely that a sound-difference is being flagged. The issue, of course, is what these sound-differences really are. Our understanding of Middle English phonology is based on the interpretation of the following: a) “reconstruction”, both comparative (dealing with cognate languages) and internal (dealing with paradigmatic variation); b) analysis of “residualisms” surviving in modern accents of English; c) analysis of the writings of spelling-reformers and phoneticians from the Early Modern English period, supplying information about usages closer to the Middle English period than now; d) analysis of contemporary verse-practices, based on the analysis of rhyme, alliteration and meter; and e) analysis of spellings. None of these approaches, of course, provides direct evidence – there are no taperecordings from the Middle Ages – but they allow for the recuperation of the phonologies of Middle English in fairly broad terms. Details are given in standard handbooks on the subject, to which further reference should be made (e.g. Jordan 1974; see also Horobin and Smith 2002 and references there cited). Given the restrictions of space, an example might suffice to show not only what can be achieved using these various resources, but also the kind of problems faced by students of Middle English interested in reconstructing the sound-systems of the period. The example chosen here is from a 14th-century Middle English romance, Sir Orfeo. Sir Orfeo survives in three medieval manuscripts, of which the best known and the most authoritative is the Auchinleck Manuscript (Burnley and Wiggins [eds.] 2003); a facsimile, transcription and associated apparatus are available as an online resource. The Auchinleck Manuscript is a large miscellany in several hands; Sir Orfeo is now the thirty-eighth item contained within it. There is good evidence from external sources that the Auchinleck Manuscript was copied in London, but the rhyming practices in Sir Orfeo indicate that it was composed in the West Midlands. Evidence for such a localization includes “mistaken” rhymes such as frut ‘fruit’ (from Anglo-Norman fru¯t): lite ‘little’ (from OE ly¯t), man ‘man’ (OE mann): opan (OE uppon). The form opan seems to be an artificial formation, not attested elsewhere in Middle English and used here to sustain the eye-rhyme with the form man; forms for ‘upon’ all otherwise end in –on, and the rhyme would only work where ‘man’ appears as mon, prototypically a Western form. With the rhyme frut:

3 Periods: Middle English lite, the scribe has not bothered to sustain the eye-rhyme, and this too is suggestive. Old English ly¯t is reflected in lute, again, in Western dialects. A. J. Bliss, in his edition of Sir Orfeo, argues against a West Midland provenance for the poem on the basis of the form owy ‘away’ rhyming with cri ‘cry’ and fairy ‘fairy’; cf. mid-line oway elsewhere in the text; Bliss considers the form owy to be “unmistakably Kentish” (Bliss [ed.] 1966: lii). However, the Kentish associations of the form owy have been challenged (see comments by E. G. Stanley, cited by Bliss), and the remaining rhyming evidence, along with certain literary characteristics in the text, suggests a West Midland association. (See further Samuels 1955: 59.) The mapping of these rhymes onto speech can be reconstructed from an examination of residualisms in later accents of English. For instance, a rounded back vowel in ‘man’ was recorded in (inter alia) Warwickshire, Worcestershire, Shropshire, etc., by Joseph Wright, an excellent observer, at the beginning of the 20th century (see Wright 1905: 520); similar usages are noted in the mid-20th century (see e.g. Wakelin 1977: 96), where the form is seen as characteristic of the West Midlands. It therefore seems to be reasonable to suppose that this situation also applied in Middle English times, as suggested by spellings such as mon ‘man’. Such complex interpretative strategies as the one just illustrated exemplify the procedures used for reconstructing Middle English pronunciation; current developments in Middle English studies, most notably the completion of LALME and its successor projects, make, as indicated above, a comprehensive new interpretative survey a major scholarly desideratum.

4 Grammar Just as Middle English accents vary diatopically and diachronically, so do Middle English grammars. In broad terms, northern grammars tend to early innovation, while southern grammars tend to be more conservative, but this broad characterization misses the complex set of grammatical systems which existed in Middle English, and the various constraints which underpin them. Traditionally, the history of English grammar has been described in terms of the shift from synthesis to analysis, i.e. from a language which expresses the relationship between words by means of inflexional endings joined to lexical stems to one which maps such relationships by means of function-words such as prepositions. This broad characterization, of course, needs considerable nuancing, and can better be expressed as a comparatively short shift along a cline. Old English, in comparison with Present-day English, is comparatively synthetic, but nowhere near as synthetic as (say) non-Indo-European languages such as Present-day Finnish or Zulu, older Indo-European languages such as classical Latin – or even earlier manifestations of Germanic such as 4th-century Gothic, which, unlike OE, regularly distinguished nominative and accusative plural forms of the noun; cf. OE hla¯fas ‘loaves’ (both NOM and ACC), Go. hla´ibo¯s (NOM), hla´ibans (ACC). Present-day English is comparatively analytic, but not as analytic as (say) Present-day Mandarin Chinese; a 21st-century English-speaker still marks person, number and case, and sustains grammatical cohesion, with concord between verbal and pronominal inflexions, for instance, e.g. I love bananas beside she hates bananas. As might be expected, Middle English grammars occupy a half-way position between Old and Present-day English, although there is both diachronic and diatopic variation. Here in (1), for instance, is a short passage from one of the writings ascribed

41

42

I Periods to the Yorkshire writer Richard Rolle, as it survives in the Thornton Manuscript held in the Cathedral Chapter Library at Lincoln, and localized to Lincolnshire: (1)

Arestotill sais þat þe bees are feghtande agaynes hym þat will drawe þaire hony fra thayme. Swa sulde we do agaynes deuelis, þat afforces thame to reue fra vs þe hony of poure lyfe and of grace. For many are, þat neuer kane halde þe ordyre of lufe ynence þaire frendys, sybbe or fremmede. Bot outhire þay lufe þaym ouer mekill, settand thaire thoghte vnryghtwysely on thaym, or þay luf thayme ouer lyttill, yf þay doo noghte all as þey wolde till þame (Sisam 1921: 41) ‘Aristotle says that the bees are fighting against him who wishes to take their honey from them. So ought we to do against devils, who endeavor [literally, endeavor them, cf. the French reflexive verb s’afforcer] to take from us the honey of humble life and of grace. For there are many, who never know how to observe moderation in [literally, never know how to hold the order of] towards their friends, akin or not akin. But either they love them too much, focusing their mind unrighteously on them, or they love them too little’ (my translation)

Aspects of the spelling of this passage defamiliarize the text for a modern reader, e.g. swa ‘so’ and sulde ‘should’, both of which are distinctively northern in dialectal distribution (the former with northern as the reflex of Old English a¯, the latter with in ‘shall’, ‘should’ – demonstrating, incidentally, the similarity between northern Middle English and usages in West Germanic languages beyond the British Isles, cf. Presentday German and Dutch cognates). But in many grammatical features there are features which correlate, broadly, with Present-day English; thus verbs such as sais, are appear much as modern English, cf. PDE says, are, and the pronominal system is also similar to modern usage, e.g. þaire ‘their’, tha(y)me ‘them’, cf. OE hı¯e, him. There are some differences in the passage in inflexional morphology in comparison with Present-day Standard English. The form afforces, for instance, has an inflected form in –es despite being governed by a relative pronoun þat ‘who’ postmodifying a plural noun deuells ‘devils’, and the present participle ending –ande in feghtande might be compared with the Present-day Standard English -ing. However, such forms are only seen as deviant from the perspective of standard usage; -s-type endings in plural verbs, and a present participle inflexion in –an (derived from –ande) are both features of Present-day Scots, for instance. It is noticeable that, in the Rolle passage, the -s is dropped when the verb is immediately preceded by a third-person plural pronoun, e.g. þay luf(e); this structure, the so-called Northern Personal Pronoun Rule, is still a characteristic feature of Scots. (The origins of the Northern Personal Pronoun Rule have been much debated. Several scholars have suggested that the construction is derived through contact with a similar pattern in Celtic languages, and this view has recently been developed by Klemola 2000; Benskin 2011.) The Rolle passage might be compared with another 14th-century text from a quite distinct dialect area: The Ayenbite of Inwyt by Dan Michel of Northgate, which survives in a holograph manuscript (London, British Library Arundel 57), localized to Canterbury in Kent, in the extreme south of England, and dated to 1340. The following passage gives some flavor of the language of the text (2): (2)

Efterward þer wes a poure man, ase me zayþ, þet hedde ane cou; and yhyerde zigge of his preste ine his prechinge þet God zede in his spelle þet God wolde yelde an

3 Periods: Middle English hondreduald al þet me yeaue uor him. Þe guode man, mid þe rede of his wyue, yeaf his cou to his preste, þet wes riche. Þe prest his nom bleþeliche, and hise zente to þe oþren þet he hedde. Þo hit com to euen, þe guode mannes cou com hom to his house ase hi wes ywoned, and ledde mid hare alle þe prestes kun, al to an hondred. Þo þe guode man yseз þet, he þoзte þet þet wes þet word of þe Godspelle þet he hedde yyolde; and him hi weren yloked beuore his bissope aye þane prest. (Sisam 1921: 34–35) ‘Afterwards there was a poor man, as it is said to me, who had a cow; and [he] heard say from his priest in his preaching that God said in his gospel that God would repay a hundredfold all that one gave for him. The good man, according to the advice of his wife, gave his cow to his priest, who was rich. The priest took his [cow] happily, and sent her to the others that he had. When it came to evening, the good man’s cow came home to his house as she was accustomed, and led with her all the priest’s cattle, up to a hundred. When the good man saw that, he thought that that was the word of the Gospel that he had restored; and they were adjudged [literally, looked] to him before the bishop against the priest’ [i.e. the bishop ruled that a poor man should have the property in question, rather than the priest; see further Sisam 1921: 213]’ (my translation) Again, there are distinctive features of spelling in this passage, such as z, u/v where would appear in Present-day English, e.g. zayþ ‘says’, zente ‘sent’, uor ‘for’, wyue ‘wife’; this usage seems to reflect “Southern voicing”, still a feature of South-Western accents but recessive in Kent by Joseph Wright’s day (see Wright 1905: 226, 241). But in grammatical terms, the passage demonstrates southern retention of the older pronouns in h-, e.g. hi ‘they/she’, and even a relic of the inflected determiner, viz. þane ‘the/that’ (DAT). In phrases such as þe guode man ‘the good man’, adjectival inflexion is retained. In verbs, the Old English ge-prefix, which was apparently in origin an aspectual marker, is retained in reduced form as y-, in ywoned ‘accustomed’, yloked ‘adjudged’ etc., while the third person singular inflexion appears as –þ in the impersonal construction me zayþ ‘it is said to me’. In sum, the inflexional morphology of this passage is conservative in comparison with the northern usage represented by the Rolle text (see 1). Syntactically, Middle English retains some features characteristic of Old English which are no longer current in present-day usage. The following passage (3) is from a Western text, John of Trevisa’s translation of Higden’s Polychronicon, as it survives in London, British Library Cotton Tiberius D.vii, a manuscript dated to around 1400 and localized to Berkeley in Gloucestershire, on the English side of the Severn Estuary. (3)

Yn Britayn buþ meny wondres. Noþeles foure buþ most wonderfol. Þe furste ys at Pectoun. Þar bloweþ so strong a wynd out of þe chenes of þe eorþe þat hyt casteþ vp aзe cloþes þat me casteþ yn. Þe secunde ys at Stonhenge bysydes Salesbury. Þar gret stones and wondur huge buþ arered an hyз, as hyt were зates, so þat þar semeþ зates yset apon oþer зates. Noþeles hyt ys noзt clerlych yknowe noþer parceyuet houз and wharfore a buþ so arered and so wonderlych yhonged. (Sisam 1921: 146) ‘There are many wonders in Britain. Nonetheless four are most wonderful. The first is at “Pectoun” [a mistake for the Peak of Derbyshire; see Sisam 1921: 252]. There such a strong wind blows out of the fissures of the earth that it throws back up clothes that are thrown in [literally, that one throws in]. The

43

44

I Periods second is at Stonehenge besides Salisbury. There great and wonderfully huge stones are raised up high, in the manner of gates [literally, as it were gates], so that there seem to be gates set upon other gates. Nevertheless it is not clearly known or perceived how and wherefore they are raised up in this way and so wonderfully hung’ (my translation) The passage contains several features characteristic of a south-west Midland dialect, e.g. forms such as buþ ‘are’, with vowel in and plural inflexion in -þ, past participles with y-, e.g. yhonged, yknowe, or the form a ‘they’. But there are also syntactic archaisms; thus, the main verb is in “second position” after an initial adverbial, as in Old English and Present-day German, in constructions such as Þar bloweþ so strong a wynd out of þe chenes of þe eorþe, while gret stones and wondur huge ‘great and wonderfully huge stones’ is an example of the characteristically Old English construction known as the “splitting of heavy groups” (see Mitchell 1985: 612–616, Sections 1464–1471). The grammatical differences between the three passages analysed so far seem to derive from the different interaction each dialect has had with the language with which late Old English had had most intimate contact, viz. Old Norse. The impact of Norse on English was not really reflected in the written language in Anglo-Saxon times, since “classical” Late West Saxon Schriftsprache – based on the usage of the parts of England furthest from the Viking invasions – is by far the best recorded variety of English during the period, and was a language of official record (and thus liable to standardization) (see Kornexl, Chapter 24). Only during the Middle English period, with the loss of the national currency of the Schriftsprache, is the full impact of Norse to be discerned. It might be expected that the influence of Norse would be strongest in those areas where Viking settlement was most dense, viz. Northern England; and the fact that grammatical innovation is earliest and most thorough in northern varieties indicates that the impact of Norse encouraged the continuing shift, in English, from synthesis to analysis that characterizes – albeit with varying speeds and to varying extents – all the Germanic languages. Barbara Strang famously argued that the process “came first in the [North] because of its long exposure to bilingualism”: Those of us who have tried, as adults, to fit into a new language-community know that in such circumstances one does not retain any more of the morphology of the new language than is strictly necessary to make oneself understood. Where a community is unilingual, old patterns remain from inertia, regardless of their functional obsolescence. In this indirect way, far more than through direct syntactic borrowing, the presence of a Scandinavian community affected the syntax of English (Strang 1970: 281).

5 Lexicon Norse left its mark on the English lexicon as well as (it seems) its grammar. Most scholars hold that the third-person plural pronouns already cited as characteristic of Northern Middle English and now standard in Present-day English, viz. they, them, their have a Norse etymology, although their adoption may have been encouraged by congruence of the initial sound in Norse þeir, þeim, þeira with that of the so-called “simple” demonstrative of Old English, viz. þa¯, þæ¯m, þa¯ra, all of which could be used with pronominal function. A more complex example is the form she, which seems to demonstrate the

3 Periods: Middle English impact of Norse in a more roundabout way. Most scholars hold that this form arose from a blend of the Old English he¯o with a Norse pronunciation, yielding *hjo¯; such forms seem to be represented in spelling by (e.g.) ʒho. The /hj/-cluster, being of low yield in the English lexicon, has a tendency to shift to [ ʃ ], cf. the form Shug, a modern nickname for the common Scottish personal name Hugh, or the commonly-cited example of Shetland from older Hjaltland (see further Britton 1991; see also Samuels 1972: 114–116). Adoption of the Norse forms of the pronouns seems to have been encouraged by general inflexional loss; more distinctive pronouns were required as syntactic tracking devices, and it is noticeable that the Present-day English third-person pronouns he, she, it and they are much more phonetically distinctive than the Old English equivalents he¯, he¯o, hit and hı¯e. It is probably not a coincidence, therefore, that it was in Kent, where inflexional distinctions were best preserved, that the new distinctive pronouns were slowest to develop – although of course this slowness would have been reinforced by the fact that the new forms, because derived from Norse, were not available in that part of the country for early selection. Norse also left its mark on the set of open-class words commonly occurring in English; common words such as ill, wing, egg, take, skill are all from Norse. Some words derived from Norse but with Old English cognates have, interestingly, developed distinct meanings, e.g. Norse-derived skirt beside Old English-derived shirt. Because of the closeness of the relationship between the two languages – particularly between Norse and the most northerly dialect of Old English, Old Anglian – it is occasionally hard to work out whether the Present-day English word is derived from Norse or from a dialect of English; a good example of such an uncertain form is call, which is probably derived from Norse kalla but is also recorded once in Old English as ceallian (see the etymology in OED for call v.). Norse-derived place-names often indicate Norse settlement, even if the forms involved are no longer part of the local dialect; thus Ormskirk in Lancashire contains the Norse generic kirk ‘church’, flagging a time when the local word for ‘church’ was kirk. However, the form is recessive, and is now largely restricted to Scottish use, referring to the established Church of Scotland (see Volume 2, Dance, Chapter 110). Whereas Norse-derived words seem to fall for the most part into everyday domains of language during the Middle English period, other words taken from other languages are of more restricted currency. Into this group fall what are by far the largest body of new items entering the English lexicon, viz. words derived from varieties of French. The earliest French-derived forms which seem to have become part of the English lexicon were taken, as might be expected, from Norman French, e.g. prison, grace, mercy, baron etc.; later on during the Middle English period many words are taken from Central French. Sometimes an early loan from Norman French can be distinguished in English by its pronunciation; thus Present-day English war comes from Norman French werre rather than from Central French guerre. Sometimes Norman and Central French cognates have both entered the language, cf. warden alongside guardian. Many French stems have been combined with English affixes to produce new words, e.g. gentleness (see Volume 2, Skaffari, Chapter 106). The influx of French words also had a structural effect on the semantics of the English lexicon: new words took over slots hitherto occupied by native forms, and a reorganisation followed. For instance, the French-derived word memory, first recorded in

45

46

I Periods English in the 13th century, took over from Old English (ge)mynd, which in turn replaced the obsolete and largely poetical form hyge with the meaning ‘mind’ (see Smith 1996: 138). The study of the shifting semantic structure of English is currently in its infancy. The recent completion of the Historical Thesaurus of the Oxford English Dictionary (Kay et al. 2009) will allow for a major leap forward in our understanding of semantic change with reference to the lexicon; the possibilities for such work have already been demonstrated by the online Thesaurus of Old English (Edmonds et al. 2005; also Roberts et al. 2000). The primary resources for the study of the Middle English lexicon remain, however, not only the Oxford English Dictionary (Simpson [ed.] 2000–), on which the Historical Thesaurus draws and with which it is now linked online, but also the Middle English Dictionary (Kurath et al. 1952–2001). The Middle English Dictionary, whose first fascicule appeared in 1952 but which derived originally from William Craigie’s call, back in 1919, for a series of “period” dictionaries to complement the Oxford English Dictionary, was completed in print form in 2001. It was transferred online shortly afterwards as part of the larger Middle English Compendium (McSparran [ed.] 2006; http://quod. lib.umich.edu/m/mec/digitMSS.html) which also includes a corpus of machine-readable texts and a hyper-bibliography, greatly increasing its functionality. The online version of the Middle English Dictionary makes it possible for complex searches to be performed; it also enhances massively our understanding of the range of variation that existed in Middle English.

6 Summary It will be clear from the above that the study of Middle English is at an exciting stage. Middle English is, for the reasons given above, the period in the history of English when variation in the spoken mode is reflected (however partially) most thoroughly in writing, and the variationist paradigms now dominant in theoretical studies offer scholars many exciting avenues for future research. The combination of the new resources now available alongside major developments in linguistic theory means that a reconciliation of “old” philology and “new” linguistics is now possible (see the very apposite comments in Rissanen 1990, which prefigured many current developments in the field).

7 References 7.1 Online resources Burnley, David and Alison Wiggins (eds.). 2003. Auchinleck Manuscript. National Library of Scotland. See: http://www.nls.uk/auchinleck/ Edmonds, Flora, Christian Kay, Jane Roberts, and Irene´ Wotherspoon. 2005. Thesaurus of Old English. University of Glasgow. http://libra.englang.arts.gla.ac.uk/oethesaurus/ Kurath, Hans, Sherman M. Kuhn, John Reidy, and Robert E. Lewis. 1952–2001. Middle English Dictionary. Ann Arbor: University of Michigan Press. http://quod.lib.umich.edu/m/med/ Laing, Margaret and Roger Lass. 2007. LAEME: A Linguistic Atlas of Early Middle English, 1150–1325. University of Edinburgh. http://www.lel.ed.ac.uk/ihd/laeme1/laeme1.html McSparran, Frances (ed.). 2006. Middle English Compendium. Ann Arbor: University of Michigan Press. http://quod.lib.umich.edu/m/mec/. See also http://quod.lib.umich.edu/m/mec/digitMSS.html

3 Periods: Middle English Meurman-Solin, Anneli. 2007. Corpus of Scottish Correspondence. University of Helsinki. http:// www.helsinki.fi/varieng/CoRD/corpora/CSC/; for manual, see: http://www.helsinki.fi/varieng/ csc/manual/ Nevalainen, Terttu, Irma Taavitsainen, and Sirpa Leppa¨nen. 1998– The Research Unit for Variation, Contacts and Change in English (VARIENG). Department of English, University of Helsinki. http://www.helsinki.fi/varieng/index.html Nevalainen, Terttu, Helena Raumolin-Brunberg, Jukka Kera¨nen, Minna Nevala, Arja Nurmi, and Minna Palander-Collin. 1998. Corpus of Early English Correspondence (CEEC). Department of English, University of Helsinki. http://www.helsinki.fi/varieng/CoRD/corpora/CEEC/index.html Robinson, Peter and Barbara Bordalejo. 1996–. The Canterbury Tales Project. Institute of Textual Scholarship and Electronic Editing, University of Birmingham. See: http://www. canterburytalesproject.org/ Simpson, John (ed.). 2000–. The Oxford English Dictionary. 3rd edn. online. Oxford University Press. www.oed.com/ Stenroos, Merja, Martti Ma¨kinen, Simon Horobin, and Jeremy Smith. Version 2011.2. The Middle English Grammar Corpus (MEG-C). http://www.uis.no/research/culture/the_middle_english_ grammar_project/meg-c/ Williamson, Keith. 2007. A Linguistic Atlas of Older Scots, Phase 1: 1380-1500 (LAOS). http:// www.lel.ed.ac.uk/ihd/laos1/laos1.html

7.2 Printed resources Benskin, Michael. 1982. The letter and in later Middle English, and some related matters. Journal of the Society of Archivists 7(1): 13–30. Benskin, Michael. 2011. Present indicative plural concord in Brittonic and Early English. Transactions of the Philological Society (Special Issue on Languages of Early Britain, ed. by Stephen Laker and Paul Russell) 109(2): 158–185. Blake, Norman (ed.). 1992. The Cambridge History of the English Language. Vol. II: 1066–1476. Cambridge: Cambridge University Press. Bliss, A. J. (ed.). 1966. Sir Orfeo. 2nd edn. Oxford: Clarendon Press. Britton, Derek. 1991. On ME she/sho: A Scots solution to an English problem. NOWELE 17: 3–51. Dobson, Eric J. (ed.). 1972. The English Text of the Ancrene Riwle: The Cleopatra Text. Oxford: Early English Text Society. Gimson, A. C. 1989. An Introduction to the Pronunciation of English. London: Arnold. Horobin, Simon and Jeremy Smith. 2002. An Introduction to Middle English. Edinburgh: Edinburgh University Press. Jordan, Richard. 1974. A Handbook of Middle English Grammar: Phonology. Rev. and trans. by Eugene Crook. The Hague: Mouton. Kane, George and E. Talbot Donaldson (eds.). 1974. Piers Plowman: The B-Text. London: Athlone Press. Kay, Christian, Jane Roberts, Michael Samuels, and Irene´ Wotherspoon (eds.). 2009. Historical Thesaurus of the Oxford English Dictionary. Oxford: Oxford University Press. Klemola, Juhani. 2000. The origins of the northern subject rule: A case of early contact? In: Hildegard L. C. Tristram (ed.), The Celtic Englishes. Vol. II, 329–346. Heidelberg: Winter. Laing, Margaret and Roger Lass. 2009. Shape-shifting, sound change and the genesis of prodigal writing systems. English Language and Linguistics 13(1): 1–31. Lass, Roger. 2006. Historical Linguistics and Language Change. Cambridge: Cambridge University Press. Luick, Karl. 1964. Historische Grammatik der englischen Sprache. Oxford: Blackwell. McIntosh, Angus, Michael Samuels, and Michael Benskin, with Margaret Laing and Keith Williamson. 1986.ALinguisticAtlasofLateMediaevalEnglish(LALME).Aberdeen:AberdeenUniversityPress. McMahon, April. 1994. Understanding Language Change. Cambridge: Cambridge University Press.

47

48

I Periods Mitchell, Bruce. 1985. Old English Syntax. Oxford: Clarendon Press. Newnham, Richard. 1971. About Chinese. Harmondsworth: Penguin. Reynolds, Leighton and Nigel Wilson. 1974. Scribes and Scholars: A Guide to the Transmission of Greek and Latin Literature. Oxford: Clarendon Press. Rissanen, Matti. 1990. On the happy reunion of English philology and historical linguistics. In: Jacek Fisiak (ed.), Historical Linguistics and Philology, 353–370. Berlin/New York: Mouton de Gruyter. Roberts, Jane and Christian Kay with Lynne Grundy. 2000. A Thesaurus of Old English in Two Volumes. 2nd revised edn. Amsterdam/Atlanta: Rodopi. Samuels, Michael. 1955. Review of Sir Orfeo, ed. by Alan J. Bliss. Medium Aevum 24: 56–60. Samuels, Michael. 1972. Linguistic Evolution. Cambridge: Cambridge University Press. Sisam, Kenneth (ed.). 1921. Fourteenth-Century Verse and Prose. Oxford: Clarendon Press. Smith, Jeremy. 1996. An Historical Study of English. London: Routledge. Strang, Barbara. 1970. A History of English. London: Methuen. Wakelin, Martyn. 1977. English Dialects. London: Athlone Press. Wright, Joseph. 1905. An English Dialect Grammar. Oxford: Frowde. Wright, Joseph and Elizabeth M. Wright. 1923. An Elementary Middle English Grammar. Oxford: Clarendon Press.

Jeremy J. Smith, Glasgow (UK)

4 Periods: Early Modern English 1 2 3 4 5 6

Introduction Historical and social background Printing and vernacularization Resources for the study of Early Modern English Changes in Early Modern English References

Abstract The two most notable changes in the Early Modern English period (1500–1700) were standardization and the growth of the lexicon. Changes in the cultural and political climate, such as the spread of printing and increasing availability of education and subsequent growing literacy among the population, were linked to these changes. The process of vernacularization in many areas (science, religion, law, government) produced new uses for English, and the Renaissance ideals of writing produced new styles and registers. Increased mobility, particularly towards London, contributed to the spread of linguistic innovations. The progressively more global trade brought contacts with new languages, and the spread of English world wide took its first steps in the colonialization of North America.

1 Introduction Early Modern English is perhaps most commonly said to range from 1500 to 1700, but since language change is gradual rather than abrupt, such demarcation lines are Alex Bergs and Laurel Brinton 2012, Historical Linguistics of English (HSK 34.1), de Gruyter, 48–63

4 Periods: Early Modern English naturally abstractions. Late Middle English during the 15th century increasingly shows features typical of Modern English, becoming more easily understandable even to the untrained present-day reader. The spread of the printing presses, one of the shaping forces on the development of Early Modern English, started towards the end of that century. Similarly, the cut-off date at 1700 marks the approximate time when most great changes during the Early Modern English period had run their course, and leaves the heyday of the prescriptive and normative tradition of the 18th century outside this stage in the development of the language. The social, political, and cultural changes associated with the Renaissance all influenced the development of the English language in the early modern period. These changes were in many ways interrelated, and reinforced each other. The preference for studying classical sources instead of the medieval authorities’ commentaries on them (Ad fontes) led to an educational reform, benefiting from the new appreciation of learning. The new schools provided literacy for an increasingly large part of the population. The changes in the intellectual climate and educational opportunities were both tied to the advancement of science and the concomitant vernacularization process of scientific writing. The Reformation, with a gradual break from the Latin traditions of the Catholic Church and the vernacularization of religious life, showed a parallel trend in stressing the authority of the original source, the Bible, and the need for people to have the ability to study it firsthand. Similar trends of vernacularization can be found in other areas of life, such as politics and law. Finally, the Age of Discovery provided contacts with new cultures as well as the beginnings of colonialization and, on the linguistic front, the first stages of American English. Many changes had a direct impact on the daily lives of the population, and the printing press was instrumental in disseminating these trends to the reading public. On the level of linguistic change, the two most notable processes are the standardization of written language and the vast increase in the lexicon. Much attention has been paid to the standardization process, which tends to provide an overly narrow view of the language as a whole, since dialects continued to be spoken (and in some cases written) by a vast majority of the population, even if this is disguised in the evidence remaining to us. While standardization of particularly printed sources tends to mask existing linguistic variation, there are also sources which give us a new perspective on the language. From the point of view of the linguist studying the period, the most important difference with earlier centuries is the wealth of new evidence on the linguistic practices of the population, providing us with English that was never written down before, or not in such quantity. Not only are there more types of texts (such as scientific and religious writing) being written in English, there are also more people than ever leaving written evidence concerning their lives. “Ego documents”, such as letters, autobiographies, wills, and travelogues, all have first person singular in common, but they also all give first-hand evidence of the linguistic practices of people in often quite private and personal, informal circumstances. We are still at the mercy of what has been preserved, but because more texts of all kinds were produced, there are also more kinds of writing remaining. There are few general descriptions of Early Modern English. The most extensive of these is the third volume of the Cambridge History of English (Lass 1999). There are also three book-length introductions aimed primarily at an undergraduate readership: Barber (1997), Go¨rlach (1991), and Nevalainen (2006). Each represents a particular

49

50

I Periods stage in the scholarship on Early Modern English, as can be most clearly seen in the evidence they use to support their description of the language. Barber illustrates his volume mostly through literary texts, following the tradition of stressing the importance of Shakespeare, and the value of drama as evidence of spoken language. Go¨rlach includes an ample selection of texts from a number of genres, bringing a wider perspective by the inclusion of more formal, non-literary types of writing. Nevalainen is a representative of the present-day paradigm of corpus-based research and draws her examples from electronic corpora and databases (see Section 4 below). All approaches have their merits, and together they provide a fuller picture of what the English language was like. Together they also illustrate the varied approaches it is possible to take when studying the language of the early modern period. This chapter has its main focus on the social, cultural, and political contexts in which Early Modern English was produced (Section 2) and on the production of English language texts (Section 3). A brief introduction to resources for the study of Early Modern English is provided in Section 4. The last section gives a brief overview of some of the changes in the language of the period, but leaves the more detailed discussion of all the linguistic aspects of Early Modern English to be found in the relevant chapters elsewhere in these volumes.

2 Historical and social background Language variation and change never take place in isolation. The connections between language and society mean that historical events need to be taken into account also when discussing the overall developments of Early Modern English. While Lass (1999: 5) is certainly correct in asserting that “[t]he story of a language ‘itself’ must be carefully distinguished from the story of its changing uses, users and social context – just as the changes themselves (as results) must be distinguished from the mechanisms by which they came about (e.g. lexical and social diffusion)”, it would still be remiss of us to overlook the influence of political and social changes taking place in the society where Early Modern English was being spoken and written. These changes required language users to adapt to new situations by creating new words and new styles of writing, they facilitated or hindered the dissemination of changes, and they influenced the variety of texts produced and preserved to us for study. Arguably, the most significant political events in the history of the period (at least when looked at from a linguistic point of view) were probably the Reformation and the consequent dissolution of the monasteries in the 16th century and the Civil War in the 17th century. Both episodes led to increasingly loose network ties because of the increased mobility of the population, and these in turn sped up linguistic change for some variables (Milroy 1992; Raumolin-Brunberg 1998). Both also produced, directly or indirectly, new types of texts, which allowed English to be written down in ways unlike those of previous centuries. The main development in social history relevant for language change was the advancement of educational opportunities for a wider range of social strata and the subsequent increased literacy rate of the population. This, in turn, gives us more linguistic first-hand evidence from a larger proportion of the population. The more widespread literacy also tied in with the advancement of printing, since there was a more extensive reading public than before. The availability of new genres for a lay audience and the

4 Periods: Early Modern English growing vernacularization of genres such as science and religion, which had previously been mainly the province of Latin, were all part of the larger picture. The population of England increased rapidly during the 16th century, followed by a time of stagnation before a further increase began in the 18th century (Coleman and Salt 1992: 2). Since there were no reliable statistics or census data created at the time, estimates of population size have been made based on such divergent data as muster rolls, lay subsidy rolls, ecclesiastical censuses, and parish registers. The reliability of population information increases when “a modernizing mercantilist state” required accurate information of resources and security, but also increasing literacy and numeracy and a more settled social and political order contributed to the development (Coleman and Salt 1992: 7). A summary of different population estimates suggests that the mid-16th-century population of England was somewhere around 2.8 million, rising to 4.1 million in the early 17th century, and showing a reasonably steady 5.0–5.2 million in the latter half of the 17th century (Coleman and Salt 1992: 5). The population of Scotland around 1600 has been estimated at one million, stagnating after that, while Ireland went from the same one-million population in 1600 to twice that in 1700; the North American English-speaking population started from a few thousand and reached quarter of a million by 1700 (Kishlansky 1996: 8). The population of London increased at a much more rapid pace than that of the country in general, from 50,000 in 1500 to 200,000 in 1600 and 575,000 in 1700 (Coleman and Salt 1992: 28). This shows the growing importance of the metropolis as a hub of government, commerce, and culture, and points to a special position also from the point of view of linguistic development (see Nevalainen, Volume 2, Historical Sociolinguistics). Already in medieval Britain there was a fair amount of geographical mobility and a later age of marriage than seems to have been common elsewhere in Europe. This was due to a free market in land, labor, and food. During the early modern period the average household consisted of 4.7 people, who were members of the nuclear family, and only rarely were there three generations under one roof. Up to 30% of households included servants, i.e. resident household or farm workers, who were typically single, aged between 15 and 30, and both men and women. As many as three quarters of boys and half of girls were in service at some point of their life. The common practice of service increased the geographical mobility of the population even further. The average age of first marriage in the 17th century for men was around 28 and for women 26 (Coleman and Salt 1992: 7, 14–15). This pattern of population movement had an impact on linguistic change in promoting dialect contact. As the population increased, there was even greater pressure for migration. While there was a great deal of subsistence migration by the unskilled and poor, particularly after 1650 there was also a large number of skilled people migrating to better themselves. Especially Scotland and Ireland produced a constant stream of migrants both to England and later to the New World (Kishlansky 1996: 13). We can only speculate how different the linguistic patterns of the mostly unlettered subsistence migrants and the at least minimally educated skilled migrants were, and how much influence either group would have had in the new location they settled at. The effect of social ambition on linguistic patterns has been established, so it is plausible that skilled workers would have more resources for linguistic adaptation. The major population crises during the early modern period were epidemics of the sweating sickness in the mid-16th century and recurring plague. In London, over 15%

51

52

I Periods of all deaths between 1580 and 1650 were caused by the plague. There were significant epidemics in 1563, 1593, and 1603, with minor outbreaks in 1578 and 1582 (Rappaport 1989: 72). The epidemics increased mobility among the population in two ways. On the one hand, the number of deaths meant there was more room for newcomers from all over the country. On the other hand, people fleeing London because of the danger probably took their new city ways, including any linguistic innovations, with them. In the 17th century, the Civil War had its own cost in loss of life: the estimated number of deaths was 80,000 in combat and 100,000 from disease (Coleman and Salt 1992: 24). Overall, the death rate in towns was higher than the birth rate, which made constant migration necessary for their growth. The continuous stream of migrants to London meant that English from all over the country could be heard in the streets, even if some areas (particularly North England) were over-represented (Coleman and Salt 1992: 27). During the early modern period, unlike the medieval times, there was no major, linguistically significant, influx of immigrants from abroad, but there were foreign craftsmen who moved to England to stay. Many of these were Protestants escaping religious persecution on the continent, but also skilled craftsmen from various countries seeking a livelihood. Around 1500 one in ten craftsmen in London were immigrants, and by 1540 they numbered one in six. In the rest of the country they were found in smaller numbers (Youings 1984: 128). Again, we can only speculate on the influence of these people, but arguably they would have had some influence on the professional language of their particular trade if nothing else. In contrast to earlier periods, there was more emigration from England, Ireland, and Scotland. The colonization of North America began, and after the first wave of migrants the surplus population of Ireland and Scotland was over-represented among those seeking opportunities in the New World. This obviously had an impact on how the new variety began to be shaped. As a result of new trade routes being discovered, English merchants, such as the East India Company, could be found trading at far distant places from the late 16th century onwards. While the original intent was to trade mostly with the East Indies, the company founded trading posts all over the Asian coastline, including India and China, but also e.g. Japan. There were also other trading companies in the West Indies and West Africa, but none was quite as long-lived or influential. The trading contacts led not only to new vocabulary for previously unknown peoples, cultures, and merchandise, but also to yet another new genre of popular writing, the travelogue. Contact was maintained not only with the indigenous peoples but also with other European traders working in the same areas, which led to the creation of trading jargons, and numerous letters were sent home, describing conditions of trade.

2.1 Education The humanist ideas of the Renaissance led to an increased appreciation of education. The aristocracy began to maintain the ideal of the well-rounded gentleman, which included learning. At the same time, education was increasingly seen as the means of providing the country with competent public servants. This was a trend that had already started in the 15th century, but it became increasingly important in the early modern period. Since the reorganization of the Tudor state and the expansion of government

4 Periods: Early Modern English activities, as well as the increase in diplomacy and foreign trade, came with a concomitant need for voluminous correspondence and detailed record keeping, there was a constant demand for literate and learned civil servants (Briggs 1994: 97; Cressy 1975: 5). Education was increasingly a secular business (Youings 1984: 17). While the dissolution of the monasteries had led to the end of schools in connection with monasteries, it provided the country with unemployed monks, who were often able to work as freelance school teachers in informal schools over the country. At the same time, new secular schools were being founded. Henry VIII alone is linked to at least eighteen schools founded or re-established during his reign, and many of the nobility followed suit. Schoolmasters needed a license to teach, and before the Civil War these licenses were under ecclesiastical control, with the purpose of preventing Catholic and Puritan teaching. During the Commonwealth control was shifted to the Parliament, but the success of any authority on imposing their demands on individual teachers is likely to have been limited at best. Education, like everything else in the society, depended on social status. The number of schools increased by at least 300 in 1500–1620, but the type of schools was extremely varied. At the one end, there were small private schools kept by a single master, at the other, grammar schools with wealthy patrons (Briggs 1994: 123). In 1647, educational reformer Samuel Hartlib envisioned four different types of education for the different social strata: one for the “vulgar, whose life is mechanical”, another for the gentry, “who are to bear charges in the commonwealth”. The third kind of school should be for scholars, who would go on to be teachers, and the fourth for the ministry (cited from Cressy 1975: 23). For the highest ranks of society, education was a value in itself, often initially received at home from tutors, while the lower ranks saw education as a means of social advancement, and were more typically educated at the various types of schools (Youings 1984: 119–120). Apart from social rank, gender was another major influence on the type and breadth of education available for individual people. While there were exceptional women at any given time, highly educated and well-read, they were definitely in the minority. Women like Margaret Roper, Katherine Astley, Queen Elizabeth, Ann Conway, or Dorothy Osborne were the exception, not the rule. Formal education was mostly unavailable, although some girls attended petty schools, small elementary schools often run by a single teacher. With higher social status came the possibility of private tutoring, at times including even classical learning. Since women did not work outside the home in professions where literacy and learning were needed, teaching them anything beyond basic skills was not considered a priority. In 1581 Richard Mulcaster, when discussing the education of women, suggested that “[r]eading if for nothing else (…) is very needful for religion, to read that which they must know and ought to perform” (cited from Cressy 1975: 110). A century later, in 1673, Bathusa Makin started a private academy for young ladies at Tottenham, but was very conscious of the resistance she was likely to meet: “I expect to meet with many scoffs and taunts from inconsiderate and illiterate men, that prize their own lusts and pleasure more than your [= that of ‘all ingenious and virtuous ladies’] profit and content” (cited from Cressy 1975: 113). Increased educational opportunities appeared at all levels of schooling: the number of students attending universities rose, and new groups of people, such as parish priests, had a university degree. For lay people, universities were a means of social advancement in the administration, but many depended on a wealthy patron or a scholarship

53

54

I Periods to pay for their expensive education. Legal training at the Inns of Court was almost entirely beyond people below the rank of gentry because of the prohibitive cost (Briggs 1994: 124). In the end, sources describing the realities of education leave much for conjecture, and a great deal of what we know is based on estimations, but there is a body of writing on the theory. A great many handbooks describing the kind of education that was desirable were published, ranging from Elyot’s Book Named the Governor (1531) to Ascham’s Schoolmaster (1570) and Mulcaster’s Positions (1581). What the existence of these books shows us is that the content of a suitable education was in general agreed upon. The handbooks range from the philosophical to the practical, discussing the education of all strata of society (Brink 2010: 31).

2.2 Literacy In the beginning of the early modern period literacy in England was restricted to the elites of society, but the transition to mass literacy began during this time (Cressy 1980: 175). Around 1500, the estimated rate of illiteracy for men was approximately 90%, while for women it was still very close to 100%. The literate people belonged to the highest strata of the population, and literacy was, for talented young men lucky enough to find a patron to support their education, a way of social advancement. Around 1600, illiteracy had clearly decreased, close to 30% of men being literate, but still only 10% of women. By 1700, the change is remarkable, since nearly half of all men could read and write and a quarter of women as well (Cressy 1980: 177). It should be remembered that these statistics are based on estimates, and that some of the assumptions at the basis of them are not completely reliable. It may well be, for example, that a person may sign a document with a mark and yet be able to write. Particularly the literacy of gentlewomen may be underestimated. There were many occasions when it was vitally important for a woman de facto looking after the estates in the absence of her husband to be able to keep account of household matters, to oversee the work of scribes employed by the family and the like, and this could be achieved more reliably if the mistress of the house was herself literate. At the same time, when scribes were available, women may simply have preferred to make use of their services rather than writing themselves (Brink 2010: 28–29). Literacy in early modern England was taught as two separate skills, reading and writing. This means that those who could read were not necessarily able to write. Reading was advocated by religious and secular writers alike. The ability to read the Bible was considered to be a spiritual benefit of great value to the general public. At the same time, education was seen as having both a moral and a civic value (Cressy 1980: 186). As mentioned above, different strata of the population did not have equal access to education, which also leads to literacy being unequally represented among them. It should be remembered, however, that literacy was not necessarily learned at school, but could also be taught by a family member or employer. Boys apprenticed to craftsmen and merchants were usually expected to have an elementary command of literacy and numeracy (Youings 1984). They would then be further instructed by their masters in the skills specific to their trade. Letter writing, for example, was often learned by copying old letters. This transferred not only the spelling conventions of the writing community, but also the textual practices involved in that particular genre. In addition

4 Periods: Early Modern English to factual literacy, being able to transfer one’s thoughts to paper, it was often necessary to be familiar with genre conventions and the requirements of a particular author-audience relationship. While some social conventions of spoken language (such as forms of address) could be more or less directly transferred to written form, there were other practices in the areas of, for example, style shifting and deferential discourse which were probably more tied to the written expression of social relations. In addition, much depended on a writer’s command of the “rules” of written language (see e.g. Palander-Collin 2009).

3 Printing and vernacularization One of the greatest changes in the early modern period when compared to the Middle Ages was the proliferation of all kinds of writing in English. As mentioned above, this was linked to the cultural developments related to the Renaissance and to the greater number of literate readers as a lay audience for new genres. Both entertainment and information of all kinds were reaching its readership. This new proliferation of different types of texts was possible because there was a reading public willing to pay. The mass-production of books, the increased literacy, and the relative affluence of middle ranks gave rise to a new audience for the more popular sorts of writing. Fiction of all kinds – prose, verse, and drama – was published in increasing amounts, ranging from broadside ballads to multivolume collections of plays. Pamphlets were produced to enlighten the public, to present political views, and to introduce new ideas and inventions. For example, sermons, the pros and cons of tobacco smoking, and new scientific discoveries were all topics suited for this form of publication. Handbooks providing instruction on many fields from medicine and culinary recipes to letter writing and proper conduct in polite society were increasingly made available to the lay readership. For the linguist, there is also much more surviving data than from earlier centuries, probably because so much more was written, both for publication and for private audiences. Many examples of private writing, particularly ego documents such as private letters and journals, remain unedited in archives, but the wealth of edited data is significantly more varied than in previous centuries. On the one hand, English was being used in new kinds of written language, presenting us with registers which either did not exist in earlier centuries or were curtailed to spoken language. On the other hand, because of the increased proportion of literate people, we are gaining direct access to the language of an ever widening part of the population. We are no longer solely reliant on fiction for the language of the middle and even lower ranks, since they – or at least some of them – are able to put pen to paper themselves. Similarly, women’s voices are more clearly heard during the early modern period than ever before. As the legal system increasingly functioned in English, court proceedings were also beginning to be recorded (and published) in that language, giving further voice to the previously silent. This means that our understanding of the full range of English in use is more complete than during earlier centuries. We are still far from actual spoken language, but we are getting a better idea of private and informal language from the actual speakers themselves. Personal correspondence is one obvious genre, made necessary, for example, by the mobility of people, as they entered service, moved to London to find their fortunes, or married outside their own immediate locality. There are also more personal journals,

55

56

I Periods commonplace books, and household accounts, which all reveal the more private and often informal side of people. On the more official side, the number of documents prepared by the growing number of civil servants increases notably during the period, and these documents are more typically written in English than during the Middle Ages, when they were more often written in Latin and Norman French. Because we have very little corresponding material from earlier periods, it is often difficult to estimate whether some words or forms of expression are new to the age or have simply never been written down before – or at least not in a form that has survived to us.

3.1 Printing Printing was a way of disseminating ideas, but also a way of disseminating the emerging written standard language. Printed books had a wider circulation than manuscripts (and many genres still circulated largely in that form), but it is notoriously difficult to estimate how great a difference this made to the actual size of the reading public. Ownership of books was certainly fashionable, and a way of displaying wealth (Youings 1984: 194). The number of books printed each year increased steadily, and in addition to books, there were pamphlets and broadsheets (Briggs 1994: 123). The book trade had its centre in London, but was by no means confined to it. Major towns had their established book sellers, and books were available at markets and fairs, by traveling peddlers along with other merchandise. During the third quarter of the 16th century nearly 4000 books were published, and during the last quarter this nearly doubled. In the 17th century, nearly ten thousand books were published in each quarter century (O’Callaghan 2010: 165). One of the consequences of the educational system becoming more regulated was a greater degree of shared background amongst the educated, and a widening of areas of interest. The ideal of the “Renaissance Man” included both literary culture and the visual arts, but also physical skills such as fencing, shooting, riding, and dancing (Briggs 1994: 124). This led to an interest in guides and handbooks in the various areas of expertise deemed necessary for the perfect courtier. Also other books, ostensibly aimed at a more common readership, were in actual fact aimed at the highest ranks. For example, Fitzherbert’s Book of Husbandry, published in 1523, was quite expensive and had a small print run of a few hundred copies (Youings 1984). There was also an increasing interest in news, which led to the publication of newsbooks from the early 1620s onwards. There were newsletters that readers could subscribe to, and these were often distributed in manuscript form, but printing was eventually the way for news as well. By the end of the 17th century, there were numerous news sheets being published, as well as twice- or thrice-weekly newspapers such as the London Gazette or the British Mercury (Briggs 1994: 165–166). Cheap, popular writing of the era included ballads, chapbooks, almanacs, and jestbooks, as well as other types of fairly ephemeral writing, which has often been regarded as representing the literary tastes of the lower ranks of society (Barry 1995: 73). On the other hand, escapist literature in the form of chivalric romances continued to be quite popular, and it is more than likely that the readership of these books went far beyond the highest strata of literate society (Barry 1995: 74). Texts were translated, abridged, rewritten, and sampled for the benefit and pleasure of those not able to read them in the original (Barry 1995: 80). Snippets were published as unbound books and in

4 Periods: Early Modern English newspapers and magazines, which made them available at a lower price. Texts would be shared by several people by reading aloud in places where people gathered, which further lowered the cost for each reader (Barry 1995: 81).

3.2 Vernacularization Early modern England was no longer a multilingual country in the way medieval England was. While Celtic languages continued to be spoken in the west, the Norman French aristocracy had seemingly lost their language by the 16th century. Despite the loss of societal multilingualism, functional multilingualism continued in many ways. Latin was still the language of higher and upper-class education, and people would learn other languages according to the necessities of their trade. So, for example, merchants involved in foreign trade would know a variety of languages depending on the direction of trade. For trade with continental Europe, French, Dutch, and Italian could be useful, while the more far-reaching trade of the East India Company, for example, made it useful for traders to learn at least a smattering of the languages of people traded with, as well as trading languages and jargons. The knowledge and use of languages other than English was reflected in the codeswitching patterns of different genres. There was a greater variety of languages switched into than in medieval times (see Raumolin-Brunberg, Chapter 45), reflecting the changes in society and types of texts, but Latin was still the most frequently occurring language. Particularly, scientific and religious texts show a high incidence of passages in languages other than English, especially when the intended audience was professional (Pahta and Nurmi 2006). This seems to indicate that there was still an expectation of Latin being known by the readers, even if the main body of the text was in English. Code-switching in these two domains can be seen as a bridge phenomenon in one of the processes that had a great influence on the development of Early Modern English, vernacularization (see Wright, Chapter 43). This was a progression that could be seen in many types of texts. While scientific and religious writing are often cited as examples, the same development could also be seen in e.g. administrative documents. English was now being used in registers and domains which had previously been performed in another language, most typically Latin and French. Go¨rlach (1999: 462) estimates that around 1500 legal texts were already mostly produced in English, although there were still remnants of Latin and French. In the realm of literature the rise of standard English is most evident, while some Latin and also dialects of English are still used. Scholarly texts are the area where Latin is still most frequently used. In Go¨rlach’s estimate, approximately half of scientific writing was in English at the beginning of the early modern period, the other half being mostly in Latin. The vernacularization process continued through the 16th and 17th centuries, and by the mid-17th century English was the primary language for scholarly texts in England. The language of religion also went through a gradual change. With the Reformation, Bible translations were ever more widely spread, and the language of liturgy changed from Latin to English. Because of the constant tension between Anglicans and Catholics and later also the Puritans and other groups, not to mention the rising Quakerism and other minor groups, there was also a constant need of discussion and writing on religion, and this was carried out in English, outlining the particularly English context in which these debates were carried out.

57

58

I Periods The expansion of English into new registers placed requirements on the language, and the influences can be seen on many levels of language. Not only does the lexicon constantly expand to accommodate the expression of new ideas, but new rhetorical styles had their influence on ideals of writing. The fact that education was still very much on the pattern of classical Latin meant that “the English style used in many formal text types was apparently praised according to how close it came to Latin models” (Go¨rlach 1999: 464). Even on the level of individual linguistic items the influence of education can be suspected. So, for example, epistemic uses of may and must spread first in the language of university-educated high-ranking men, which would suggest that the thought styles taught at Oxford and Cambridge included the use of epistemic modality (Nurmi 2003, 2009).

4 Resources for the study of Early Modern English Because of the proliferation of different types of text, we are also able to benefit from a larger variety of electronic corpora as sources for studying Early Modern English. On the one hand, we have multi-genre general-purpose corpora, and on the other, there are also more specialized corpora of a single genre or domain of writing. Of the first type, the Helsinki Corpus (Rissanen et al. 1991) covers the years 1500–1710 and consists of 500,000 words in 18 genres. The ARCHER corpus (Biber and Finegan 1990–93/2002/ 2007/2010) focuses more on the late modern period, but it starts from 1650, and has 11 genres. Also, other varieties are covered: ARCHER includes a corresponding selection of both British and American English, and the Helsinki Corpus of Older Scots (Meurman-Solin 1995) brings a possibility of contrastive studies. With the digitization of more and more materials, large commercial databases offer an ever increasing selection of the early printed sources in massive archives, such as Early English Books Online (Chadwyck-Healey 2003–2011) and the Literature Online (Chadwyck-Healey 1996–2011) database. The more specialized corpora focus on a single genre, topic domain, or publication type. A good example of the last is the Lampeter Corpus (Claridge et al. 1999). It contains tracts published between 1640 and 1740, and has six topic domains, which are represented for each decade of the timeperiod, numbering over a million words. The Zurich English Newspaper Corpus (ZEN) (Fries, Lehmann et al. 2004) covers early English newspapers between 1661–1791, giving access to 1.6 million words of whole newspaper issues with their varied content types. Newdigate Newsletters (Hines 1995) presents the precursor and competitor of the newspaper, written between 1674–1715 (750,000 words). The single-genre Parsed Corpus of Early English Correspondence (PCEEC) (Nevalainen et al. 2006) consists of personal letters written between 1410 and 1681, altogether 2.2 million words. There is also a short version containing a selection of the texts, the Corpus of Early English Correspondence Sampler (CEECS) (Kera¨nen et al. 1998), with 450,000 words. A Corpus of English Dialogues 1560–1760 (Kyto¨ and Culpeper 2006) focuses on speech-based texts, containing both authentic dialogue in trial proceedings and witness depositions, and constructed dialogue in drama and prose fiction, again reaching over a million words. There are also corpora focusing on a special domain of writing. The Corpus of Early English Medical Writing (Taavitsainen et al. 1995–) presents various text categories aimed at both expert and lay readership and covering multiple types of writing from

4 Periods: Early Modern English the purely academic to health guides for the general public. The Corpus of English Religious Prose (Kohnen et al. forthc.), which is being prepared, tackles the domain of religion and the various genres of writing that are connected with it. New corpus projects arise all the time, and the variety of these projects and the types of corpora they aim to build are a testimony to the multiplicity of material available for scholars of Early Modern English.

5 Changes in Early Modern English The two most striking changes taking place in the early modern period were the standardization, particularly of orthography, in published writing (with a gradual spread of similar spelling conventions to private texts as well) and the explosive growth of the lexicon. The early stages of descriptive and prescriptive writing on language were also seen during this period, even if the main developments only arrived in the 18th century. Many of the other linguistic developments of the age were continuations of long-term trends which had their origins in Middle and – in some cases – Old English. Standardization is most often viewed on the level of orthography, and certainly the changes there were remarkable during the two centuries in question, but also other levels of language can be argued to have developed some form of standard. Printing and the growing and developing civil service spread the particular type of writing of literate people in London and at the universities to a more varied readership than before, and provided a model to aim towards. Many linguistic features which become an established part of the new general dialect did not necessarily spread to the spoken regional forms of language, but, since our remaining sources are written, they tend to obscure the richness of local variation which must have existed all through the centuries in order to have survived to the present day (see Moessner, Chapter 44). The vocabulary of English was increasing as more types of texts were produced, and this led to hard word dictionaries being published. These often took the form of wordlists, which might contain words invented by the compiler of the lexicon, never seen outside these compilations, but they also presented many words which have since established their place in the English lexicon. Some dictionaries were aimed at translators, others specialized in a given field, such as legal or medical terminology or the language of thieves (see Considine, Chapter 66). The orthoepists discussed ways of improving the English spelling system and as an unintended side product gave us a clearer idea of how the language was pronounced. Early grammars were heavily based on the Latin model, and were often not very succesful in describing English in those terms. Because of the newly literate middle ranks of society, there was a welcoming readership for these works, although the age of the autodidact did not properly begin before the 18th century. How much influence any of these volumes had on the English actually used is an open question, but they give us an indication of the increasing interest in codifying, analyzing and teaching English. The English lexicon increased in size in several ways and for several reasons. New words were borrowed for new concepts, both scientific and cultural, from any number of languages. While Latin was the most influential source, the influx of new words from both European and world languages is notable. As the English became more familiar with the world, they introduced new words to describe the flora and fauna, the artifacts and merchandise, the peoples and cultures they encountered. At the same time, the

59

60

I Periods Renaissance ideal of expressing an idea in as many ways as possible contributed to the borrowing of Latin words in order to introduce variety. New words did not come solely from borrowing: also word formation through prefixes and suffixes, as well as compounding, was frequent. Many of the elements included in these processes were borrowed themselves, but there was eventually also mixing of native and borrowed elements. Many near synonyms were introduced, but only a selection of those has survived (see Lancashire, Chapter 40). On the level of morphology, the loss of nominal case endings that had been going on for a long time reached its culmination, with only the genitive -s remaining in the nominal system. The use of apostrophes to signify the genitive as distinguished from plural (or to distinguish genitive singular and plural) arose only gradually, and did not reach present usage before the 18th century. In the case of personal pronouns the most notable changes appeared with regard to second person. The singular pronoun thou became increasingly marked, and was used less and less except for highly specific contexts (intimacy, status difference, religious language). As thou disappeared, the corresponding verbal inflection disappeared as well. In the plural, the object form you replaced the old subject form ye. In the case of relative pronouns, subject pronoun who became established in human reference (see Busse, Chapter 46). Adjectives and adverbs showed more variation in the formation of comparative and superlative forms than Present-Day English, and the rules governing the use of inflections or the periphrastic forms were still in flux, leaving room for double forms (most happiest). Adverbs had variant forms without the suffix -ly, so that smooth/smoothly could be used interchangeably. The verbal system saw a rise of auxiliaries. Periphrastic do established itself in questions and negative statements. For a while, it seemed that do was also making inroads in affirmative statements, but this development was cut short (see Warner, Chapter 47). Verbal inflections followed a similar trend as nominal case endings, and the early modern period saw the loss of all but the third-person singular suffix, which changed from the earlier -th to -s. Since the loss of inflections made the subjunctive scarcer, modal auxiliaries took some of its functions. The meanings of modal auxiliaries shifted more towards the present model with the increasing frequency of epistemic meanings. The progressive be + -ing form started increasing, although the real development of this construction took place in the 18th and 19th centuries. As for word order, the long-term change towards a fixed pattern of subject-verbobject in declarative statements saw the last stages of formalization. Sentence-initial adverbs could still cause subject-verb inversion in the early modern period, but, apart from the greater liberties taken by verse, this pattern was notably less frequent by 1700. Many syntactic patterns typical of Latin could be seen in high styles of writing, whether legalese or ornate literature (see Seoane, Chapter 39). On the level of pronunciation, the Great Vowel Shift was perhaps the most notable development. The raising of long vowels took place over three centuries, and was a series of local developments (see Krug, Chapter 48). All parts of the shift did not run their course in all dialects, and there was variation in how individual words were affected. Local dialects continued as the main spoken form, but the beginnings of Received Pronunciation appeared in the cultural hub that was London (see Mugglestone, Volume 2, Chapter 121).

4 Periods: Early Modern English All in all, developments in Early Modern English levelled much of the earlier variation as the new standard language was formed. The place for standard was in official, published and formal kinds of writing, but private, unpublished, and informal language continued to show much more regional and stylistic variation. Being able to command the standard register was one of the requirements of inclusion in the elites of the country, but large parts of the population could lead successful lives without the requirements of shaping their language to this new pattern.

6 References 6.1 Online resources Biber, Douglas and Edward Finegan. 1990–93/2002/2007/2010. A Representative Corpus of Historical English Registers (ARCHER). Version 3.1 http://www.llc.manchester.ac.uk/research/ projects/archer/ Chadwyck-Healey. 2003–2011. Early English Books Online, 1475–1700 (EEBO). Ann Arbor: ProQuest. http://eebo.chadwyck.com/home Chadwyck-Healey. 1996–2011. Literature Online (LION). http://lion.chadwyck.com/ Claridge, Claudia, Josef Schmied, and Rainer Siemund. 1999. The Lampeter Corpus of Early Modern English Tracts. In: ICAME Collection of English Language Corpora (CD-ROM), 2nd edn., Knut Hofland, Anne Lindebjerg, and Jørn Thunestvedt (eds.), The HIT Centre, University of Bergen, Norway For manual see http://khnt.hit.uib.no/icame/manuals/LAMPETER/ LAMPHOME.HTM Fries, Udo, Hans Martin Lehmann et al. 2004. Zurich English Newspaper Corpus (ZEN). Version 1.0. Zu¨rich: University of Zu¨rich. http://es-zen.unizh.ch (see also: http://www.helsinki.fi/ varieng/CoRD/corpora/ZEN/index.html) Hines, Philip, Jr. 1995. Newdigate Newsletters. In: ICAME Collection of English Language Corpora (CD-ROM), 2nd edn., Knut Hofland, Anne Lindebjerg, and Jørn Thunestvedt (eds.), The HIT Centre, University of Bergen, Norway. For manual see http://icame.uib.no/newdigateeks.html Kera¨nen, Jukka, Minna Nevala, Terttu Nevalainen, Arja Nurmi, Minna Palander-Collin and Helena Raumolin-Brunberg. 1998. Corpus of Early English Correspondence Sampler (CEECS). In: ICAME Collection of English Language Corpora (CD-ROM), 2nd edn., Knut Hofland, Anne Lindebjerg, and Jørn Thunestvedt (eds.), The HIT Centre, University of Bergen, Norway. For manual see http://khnt.hit.uib.no/icame/manuals/ceecs/INDEX.HTM (see also: http://www. helsinki.fi/varieng/domains/CEEC.html) Kohnen, Thomas, Sandra Boggel, Tanja Ru¨tten, Dorothee Groeger, Ingvilt Marcoe, and Kirsten Gather. forthc. Corpus of English Religious Prose (COERP). http://www.helsinki.fi/ varieng/CoRD/corpora/COERP/index.html Kyto¨, Merja and Jonathan Culpeper. 2006. A Corpus of English Dialogues 1560–1760. With the assistance of Terry Walker and Dawn Archer. Uppsala University and Lancaster University. http://www.engelska.uu.se/Research/English_Language/Research_Areas/Electronic_Resource_ Projects/A_Corpus_of_English_Dialogues/ Meurman-Solin, Anneli. 1995. Helsinki Corpus of Older Scots. Department of English, University of Helsinki. http://www.helsinki.fi/varieng/CoRD/corpora/HCOS/index.html Nevalainen, Terttu, Helena Raumolin-Brunberg, Jukka Kera¨nen, Minna Nevala, Arja Nurmi, and Minna Palander-Collin. 2006. Parsed Corpus of Early English Correspondence (PCEEC). Annotated by Ann Taylor, Arja Nurmi, Anthony Warner, Susan Pintzuk, and Terttu Nevalainen. Helsinki: University of Helsinki and York: University of York. Distributed through the Oxford Text Archive. Rissanen, Matti, Merja Kyto¨, Leena Kahlas-Tarkka, Matti Kilpio¨, Saara Nevanlinna, Irma Taavitsainen, Terttu Nevalainen, and Helena Raumolin-Brunberg. 1991. The Helsinki Corpus of English Texts. In: ICAME Collection of English Language Corpora (CD-ROM), 2nd edn.,

61

62

I Periods Knut Hofland, Anne Lindebjerg, and Jørn Thunestvedt (eds.), The HIT Centre, University of Bergen, Norway. For manual, see http://khnt.hit.uib.no/icame/manuals/hc/index.htm Taavitsainen, Irma, Pa¨ivi Pahta, Martti Ma¨kinen, Turo Hiltunen, Ville Marttila, Maura Ratia, Carla Suhr, and Jukka Tyrkko¨. 1995–. Corpus of Early English Medical Writing (CEEM). University of Helsinki. http://www.helsinki.fi/varieng/CoRD/corpora/CEEM/index.html

6.2 Printed resources Barber, Charles. 1997. Early Modern English. 2nd edn. Edinburgh: Edinburgh University Press. Barry, Jonathan. 1995. Literacy and literature in popular culture: Reading and writing in historical perspective. In: Tim Harris (ed.), Popular Culture in England, c. 1500–1850, 69–94. London: Macmillan. Briggs, Asa. 1994. A Social History of England: From the Ice Age to the Channel Tunnel. 2nd edn. London: Weidenfeld and Nicolson. Brink, Jean R. 2010. Literacy and education. In: Hattaway (ed.), Vol. 1, 27–37. Coleman, David and John Salt. 1992. The British Population: Patterns, Trends, and Processes. Oxford: Oxford University Press. Cressy, David. 1975. Education in Tudor and Stuart England. London: Arnold. Cressy, David. 1980. Literacy and the Social Order: Reading and Writing in Tudor and Stuart England. Cambridge: Cambridge University Press. Go¨rlach, Manfred. 1991. Introduction to Early Modern English. Cambridge: Cambridge University Press. Go¨rlach, Manfred. 1999. Regional and social variation. In: Lass (ed.), 459–538. Hattaway, Michael (ed.). 2010. A New Companion to English Renaissance Literature and Culture. Vols. 1–2. Oxford: Wiley-Blackwell. Kishlansky, Mark. 1996. A Monarchy Transformed: Britain 1603–1714. London: Penguin. Lass, Roger. 1999. Introduction. In: Lass (ed.), 1–12. Lass, Roger (ed.). 1999. The Cambridge History of the English Language. Vol. III. 1476–1776. Cambridge: Cambridge University Press. Milroy, James. 1992. Linguistic Variation and Change. Oxford: Blackwell. Nevalainen, Terttu. 2006. Introduction to Early Modern English. Edinburgh: Edinburgh University Press. Nurmi, Arja. 2003. The role of gender in the use of MUST in Early Modern English. In: Sylviane Granger and Stephanie Petch-Tyson (eds.), Extending the Scope of Corpus-based Research: New Applications, New Challenges, 111–120. Amsterdam/Atlanta: Rodopi. Nurmi, Arja. 2009. May: The social history of an auxiliary. In: Andreas H. Jucker, Daniel Schreier, and Marianne Hundt (eds.), Corpora: Pragmatics and Discourse. Papers from the 29th International Conference on English Language Research on Computerized Corpora (ICAME 29). Ascona, Switzerland, 14–18 May 2008, 321–342. Amsterdam/Atlanta: Rodopi. O’Callaghan, Michelle. 2010. Publication: Print and manuscript. In: Hattaway (ed.), Vol. 1, 160– 176. Pahta, Pa¨ivi and Arja Nurmi. 2006. Code-switching in the Helsinki Corpus: A thousand years of multilingual practices. In: Nikolaus Ritt, Herbert Schendl, Christiane Dalton-Puffer, and Dieter Kastovsky (eds.), Medieval English and its Heritage: Structure, Meaning and Mechanisms of Change, 203–220. Frankfurt: Peter Lang. Palander-Collin, Minna. 2009. Patterns of interaction: Self-mention and addressee inclusion in the letters of Nathaniel Bacon and his correspondents. In: Arja Nurmi, Minna Nevala, and Minna Palander-Collin (eds.), The Language of Daily Life in England (1400–1800), 53–74. Amsterdam/Philadelphia: John Benjamins. Rappaport, Steve. 1989. Worlds within Worlds: Structures of Life in Sixteenth-century London. Cambridge: Cambridge University Press.

5 Periods: Late Modern English

63

Raumolin-Brunberg, Helena. 1998. Social factors and pronominal change in the seventeenth century: The Civil-War effect? In: Jacek Fisiak and Marcin Krygier (eds.), Advances in English Historical Linguistics (1996), 361–388. Berlin: Mouton de Gruyter. Youings, Joyce. 1984. Sixteenth-century England. London: Penguin.

Arja Nurmi, Helsinki (Finland)

5 Periods: Late Modern English 1 2 3 4 5 6 7 8 9

Introduction: definitions of Late Modern English The growth of Late Modern English studies External history Syntax and morphology Phonology Lexis Normative grammarians Summary References

Abstract This chapter provides an overview of research into the history of English in the Late Modern period. It begins with an account of how this period came to be defined as a distinct period lasting roughly from 1700–1900 and goes on to discuss the reasons why this period has, until relatively recently, received less scholarly attention than earlier ones. An overview of the external history of the period follows, concentrating on factors such as urbanization, industrialization, and the growth of transport and communications technology, all of which contributed to the social and geographical mobility which characterize this period. The remainder of the chapter provides brief accounts of research into the syntax and morphology, phonology, and lexis of this period. The chapter concludes with a short account of recent research which re-evaluates the normative grammarians of this period and some suggestions for future directions for the study of Late Modern English.

1 Introduction: definitions of Late Modern English The phrase “Late Modern English” seems to have been first used by Poutsma in the title of his Grammar of Late Modern English (first published in 1914), but he was referring here not to a historical period so much as to what was, to him, contemporary English, “the English Language as it presents itself in the printed documents of the last few generations” (Poutsma 1928: viii). As such, his study was synchronic rather than diachronic, dealing with the language as it was in his time and in the very recent past. The tripartite division of the history of English into Old, Middle and Modern English can probably be attributed to Sweet, who proposed this division in a lecture to the Philological Society (1873–1874). Sweet saw “Modern English” as a unity stretching from the 16th century to his own time and characterized it as the period of lost inflections. Alex Bergs and Laurel Brinton 2012, Historical Linguistics of English (HSK 34.1), de Gruyter, 63–78

64

I Periods The recognition of Late (or Later) Modern English as a specific period in the history of English appears to have followed much later. Charles Barber uses the term “Later Modern English” to distinguish this period from Early Modern English: I am taking eModE to be the English language between 1500 and 1700. All such divisions are arbitrary, for linguistic change is continuous; but there are a number of features in the language of that period which mark it off fairly clearly from Middle English (ME) and Later Modern English (LModE) (Barber, 1976: 1).

Barber provides no end-point for the Later Modern period, but a consensus has since emerged that the Late Modern period lasts roughly from 1700 to 1900, though Beal (2004) extends the end-point to 1945. (See Curzan, Volume 2, Periodization.)

2 The growth of Late Modern English studies Charles Jones refers to the 18th and 19th centuries as the “Cinderellas of English historical linguistic study” (Jones 1989: 279). This situation lasted until the final decade of the 20th century, when Bailey (1996) and Go¨rlach (1999) both published monographs on 19th-century English. Until this point, studies of 18th- and/or 19th-century English tended to concentrate on specific areas. The agenda for research on this period had been set by Leonard’s (1929) volume on “correctness” in 18th-century English, so that general histories of English tended to describe this period as characterized by the appearance of prescriptive grammars and authoritative dictionaries. These earlier works express the view that there were no linguistic changes worthy of investigation after 1700. A typical statement is that made by Bloomfield and Newmark, who assert that “after the period of the Great Vowel Shift was over, the changes that were to take place in English phonology were few indeed” (Bloomfield and Newmark 1963: 293), and, even as late as 1992, Freeborn stated that “the linguistic changes that have taken place from the eighteenth century to the present day have been relatively few” (Freeborn 1992: 180). That any language could be spoken for 200 years without any change taking place is highly unlikely, but scholars were referring here to the kind of major, structural changes, such as the Great Vowel Shift or the introduction of do-support, which dominate much discussion of earlier periods. Some scholars, such as Strang, suggest that the nature of change in the Late Modern period was different from that in Middle and Early Modern English. Some short histories of English give the impression that changes in pronunciation stopped dead in the 18c, a development which would be inexplicable for a language in everyday use. It is true that the sweeping systematic changes we can detect in earlier periods are missing, but the amount of change is no less. Rather, its location has changed: in the past two hundred years changes in pronunciation are predominantly due, not, as in the past, to evolution of the system, but to what, in a very broad sense, we may call the interplay of different varieties, and to the complex analogical relationship between different parts of the language (Strang 1970: 78–79).

Although Strang acknowledges that changes have taken place since 1700, her suggestion that “sweeping systematic changes” were completed by then and that the later changes were caused by “the interplay of different varieties” would appear to flout

5 Periods: Late Modern English the “uniformitarian principle” that the mechanisms governing linguistic variation and change operated in past times and societies as they do today. Research in the fields of socio-historical linguistics and historical sociolinguistics, such as Romaine (1982) and Nevalainen and Raumolin-Brunberg (2003), has demonstrated that “the interplay of different varieties” was as much a factor in linguistic change in the Early Modern period as Strang claims it to be for “the past two hundred years”. The perceived difference between linguistic changes of earlier and later periods is more likely to be a result of the scholar’s perspective: changes appear simpler and more abstract as chronological distance increases, and as the amount of detailed evidence for variation decreases. Jones suggests as much when he writes: There has always been a suggestion … especially among those scholars writing in the first half of the twentieth century, that phonological and syntactic change is only properly observable at a great distance and that somehow the eighteenth, and especially the nineteenth centuries, are “too close” chronologically for any meaningful observations concerning language change to be made (Jones, 1989: 279).

Perhaps one reason for the upsurge of interest in Late Modern English in the 1990s is that, as the new millennium hove into sight, scholars felt sufficiently distanced from the 18th and 19th centuries to be able to observe the linguistic changes that occurred in this period. However, chronological distance is not the only reason for the growth of studies in Late Modern English in the late 20th and early 21st centuries. The development of corpus studies and the availability of a wide range of texts and text types in electronic form have enabled scholars both to identify overall patterns of change and to interrogate individual texts in ways which were simply not possible prior to the digital revolution. Historical linguists can now search for syntactic patterns or lexical collocations in corpora consisting of millions of words, or they can access individual texts for close study. David Denison’s comment on the nature of syntactic change in Late Modern English comes from this perspective: Since relatively few categorical losses or innovations have occurred in the last two centuries, syntactic change has more often been statistical in nature, with a given construction occurring throughout the period and either becoming more or less common generally or in particular registers. The overall, rather elusive effect can seem more a matter of stylistic than syntactic change (Denison 1998: 93).

The area of linguistics in which constructions becoming “more or less common” is of primary interest is, of course, sociolinguistics, where quantitative methods have been used to investigate variation and change in language since the 1960s. The application of models taken from sociolinguistics has allowed scholars to make sense of the variability of data from the Late Modern period. The notion of “stylistic change” has also been given more prominence in sociolinguistics than in formal linguistics, and the application of sociolinguistic models, together with the availability of texts, has led to the investigation of the “styles” of social networks and of individuals in the Late Modern period. So, we could say that a confluence of circumstances has led to the increasing prominence of Late Modern English studies in the late 20th and early 21st centuries. There

65

66

I Periods are now two monographs entirely devoted to this period, Beal (2004) and Tieken-Boon van Ostade (2009); a conference on Late Modern English initiated by Charles Jones in 2001 has now had four meetings; and the proceedings of the first three of these conferences have been published as Dossena and Jones (eds.) (2003), Bueno Alonso et al. (eds.) (2007), and Tieken-Boon van Ostade and van der Wurff (eds.) (2009). The contents of these volumes provide a good indication of the agenda that is developing for those working in the field of Late Modern English. Dossena and Jones divide their volume into three sections entitled “The Late Modern English Grammatical Tradition”, “The Syntax of Late Modern English”, and “Language and Context in the Late Modern Period”. The third section is fairly eclectic, dealing with the language of individuals, and with pragmatics and sociolinguistics. Bueno Alonso et al. present individual contributions rather than grouping them into sections, but of the fifteen papers, seven are devoted to syntax or morphology, three to pragmatics, and five to areas such as variation and change and the development of genres. Tieken and van der Wurff divide their volume into four sections entitled “Prescriptive and Normative Concerns”, “Late Modern Work on the English Language”, “Studies in Grammar and Lexis”, and “Studies on Letters”. This indicates that the main areas of research in Late Modern English are: morpho-syntactic change; the normative tradition; historical pragmatics; and what we might broadly refer to as historical sociolinguistics. This is not to say that other areas have been totally neglected: for instance, two monographs have been published (Beal 1999 and Jones 2006) which are entirely devoted to the phonology and phonetics of English in this period. As the quotes from Strang (1970) and Denison (1998) above might have predicted, the study of Late Modern English has more recently been informed by the methods and frameworks of sociolinguistics and corpus linguistics (see, for instance, Mair 2002; Rohdenburg 2007; and De Smet 2007), so that research is often difficult to categorize in conventional linguistic terms. In the sections that follow, I shall provide a brief account of the external history of English-speaking communities in the Late Modern Period, then go on to summarize the main developments in morpho-syntax, phonology, and lexis, before discussing the ways in which scholars of Late Modern English have interrogated the normative texts of this period to arrive at a more nuanced view of the “Age of Authority” than had hitherto been put forward.

3 External history We saw in Section 1 that the Late Modern period is usually described as beginning around 1700. In his introduction to Volume III of the Cambridge History of the English Language covering the period 1476–1776, Lass gives a dramatic summary of the events leading up to this point in history: “By the eighteenth century, the nation had been through a religious reformation, a regicide, a commonwealth, the flight of the hereditary monarch, and the accession of a foreign king who signed away much of his power” (Lass 1999: 3). Lass also agrees with the historian Roy Porter (2000) in viewing the publication of Newton’s Principia (1687) as marking the beginning of the English Enlightenment, an age of “reason” and science. The most obvious linguistic consequence of scientific progress in the Late Modern period was to stimulate lexical innovation as new inventions, processes, and whole disciplines required names. Newton wrote his Principia in Latin, but the Royal Society was to foster a style of scientific writing in English as writing in Latin declined.

5 Periods: Late Modern English The scientific discoveries of the late 17th and early 18th centuries led to the technological innovations which drove the Industrial Revolution of the late 18th and 19th centuries. Britain became an industrial and urban nation as workers moved from the countryside to the newly-expanding towns and cities. The historian Michael Rose writes: At the beginning of the [nineteenth] century there were only fifteen towns in England and Wales with more than 20,000 inhabitants; by 1851, there were sixty-three, and one half of the population could be described as town dwellers compared to about one third in the late eighteenth century. By 1900, almost 80 per cent of the population lived in urban districts with populations of 10,000 or more. In the space of a hundred years, Britain had transformed into an urban society (Rose 1985: 277).

The main linguistic consequence of this urbanization and movement of populations was dialect contact, leading to levelling and the formation of new, urban dialects. Dialect contact was also facilitated by advances in transport and communications in this period. In the course of the 18th century, the Turnpike Trusts funded a substantial number of new roads, cutting the length of a journey by carriage from York to London from three days to one, and opening up the possibility of travel for leisure as well as business, at least for the reasonably well-to-do. The development of the railway network in the 19th century made affordable leisure travel possible for the lower and middle classes. The introduction of the Penny Post in 1840 and the electric telegraph in 1837 increased the possibilities for written communication, and bequeathed a rich legacy of data to scholars of Late Modern English, who have been able to create corpora of letters from this period, such as the Corpus of late 18c Prose (Denison et al. 2002) and the 18th century extension to the Corpus of Early English Correspondence (CEECE) (Nevalainen et al. 1998–2006). The invention of the phonograph in 1877 made it possible for speakers of English to hear the disembodied voices of speakers from distant places. All these developments had the effect of increasing dialect contact between speakers (and writers) from different parts of Britain (and beyond). Dialect contact of a different kind was occasioned by the introduction of compulsory, free elementary schooling in 1870. One consequence of this was that every child would come into contact with Standard English because every village would have at least one schoolteacher who, along with the clergyman, would act as models of “correct” usage. Scholarly interest in philology led to the formation of the English Dialect Society in 1873, the objective of which was to collect material for a comprehensive dictionary of English dialects. This objective was achieved with the publication of Wright’s English Dialect Dictionary (1898–1905), but in the interim many glossaries of individual dialects were published under the auspices of the English Dialect Society. The authors of these glossaries all express a sense of urgency: they have collected their material “just in time”, and all blame universal education and the railways for the imminent demise of dialects. An example of such a statement is the following, from Morris’s Yorkshire Folk-Talk: Railways and certificated schoolmasters, despite their advantages, are making sad havoc of much that is interesting and worth preserving in the mother tongue of the people. This is to be regretted. It is with the object of collecting any such relics of the past, which would otherwise be doomed to oblivion, that I make the following appeal to my brother Yorkshiremen (Morris 1892: v).

67

68

I Periods What few of these 19th-century antiquarians and philologists recognized was that dialect contact can create new varieties as well as destroying old ones. In the towns and cities of the Industrial Revolution, new, urban dialects were being forged in the crucible of contact between speakers of different regional and national varieties of English. In some places the impact of Irish immigration during and immediately after the Great Famine was considerable: in Liverpool and in the 19th-century “new town” of Middlesbrough, approximately one-fifth of the population recorded in censuses of this period was Irish-born. Elsewhere, the influx of population was mainly from the rural hinterland, but in places such as Newcastle and the West Riding of Yorkshire, dialect writing and performance testifies to a growing consciousness of, and pride in, new urban identities and varieties of speech (Beal 2000, 2011). These urban varieties have since become the main object of dialect study in the late 20th and early 21st centuries, and are now in their turn seen as potentially threatened by levelling and supralocalization. The Late Modern period also saw the beginning of the “great divide” between British and American English. Although the first English-speaking colonies in what is now the USA were founded in the early 17th century, the development of American English as a national variety with its own prescribed norms was precipitated by the American Revolution (1775–1783). In 1789, Webster asserted that “customs, habit and language, as well as government, should be national. America should have her own, distinct from all the world” (Webster 1789: 179). His American Dictionary of the English Language (Webster 1828) provided norms for spelling which were deliberately differentiated from those of British English, and legitimized “Americanisms” by including them in the dictionary. Once America became independent, Americans loyal to the crown moved to Canada and British colonial expansion diverted to Australia in 1788 and, in the 19th century, to South Africa and New Zealand. The development of distinct national norms of English in these countries was perhaps more a phenomenon of the 20th century, but the history of these Englishes belongs entirely in the Late Modern period (see Volume 2, Section XV). Thus, the Late Modern period is one in which English ceases to be predominantly the “property” of speakers in Britain, and any history of English in this period needs to take account of this. Another linguistic consequence of the British Empire was the introduction into English of loan words from a wide variety of languages, as flora, fauna, topographic features, and customs hitherto unknown to speakers of English required names. This, along with the scientific discoveries and inventions referred to above, accounts for the dramatic increase in lexical innovation during the 19th century. Within Britain, the commercial opportunities of the Industrial Revolution and the expansion of educational provision led in the course of the Late Modern period to the emergence of an ambitious and influential middle class, who were now able to gain the wealth, power, and influence which had formerly been the preserve of the landed gentry. Indeed, the most prominent industrialists and entrepreneurs could be ennobled to the peerage, build their own “great houses” in the country, socialize with the gentry, and send their sons to the public schools which became the cradle of Received Pronunciation. Social mobility led to linguistic insecurity amongst the upwardly-mobile and so created a market for the normative texts for which this era is famous. As has already been noted, Leonard’s (1929) work on the grammarians of the 18th century was so influential that, for much of the 20th century, the received view of these grammars was that they imposed arbitrary rules to suppress variation in English. In recent years, the wider availability of these texts facilitated by Eighteenth-Century

5 Periods: Late Modern English Collections Online (Gage Cengage Learning 2009) and other digitization projects, has allowed scholars to develop a more nuanced account of these grammars, and to compare precept with practice. Nevertheless, it cannot be denied that the Late Modern period marks the final stages of the standardization process for British English and that the prescriptive attitudes of this era are still with us today (Beal 2008, 2009).

4 Syntax and morphology As Denison noted (see Section 2), the Late Modern English period saw very little in the way of categorical change in syntax or morphology. Some constructions and patterns which had been used variably in Early Modern English became regulated in this period, so that variants formerly found in Standard English texts would be restricted to nonstandard usage and/or “marked” in some way. Many scholars have blamed the prescriptive grammarians of the 18th century for this regulation, but recent research involving the comparison of precept and practice in this period suggests that they were simply codifying what was already best practice. I shall therefore discuss these changes in Section 7, and concentrate here on the more categorical changes. As far as morphology is concerned, there was very little change within Standard English in the Late Modern period. We saw in Section 1 that Sweet defined the Modern English period as a whole as the period of lost inflections. However, the only inflection to be lost after 1700 is the second person singular -st ending. This in turn depends on the loss, from Standard English, of the distinction between second person singular thou, thee, thy, thine and the formerly plural ye, you, your, yours. The singular forms had become marked in the Early Modern period, and by 1700 “survived only in dialects, among Quakers, in literary styles, as a device of heightening […] and in its present religious function” (Strang 1970: 140). Some 18th-century authors, notably Sheridan and Richardson, put thou forms into the mouths of upper-class males, but this usage was not universally accepted. Although 18th-century grammarians often include thee, thou, thy in their paradigms of pronouns in order to demonstrate the singular-plural distinction, Greenwood tells his readers that “it is ungentile [sic.] and rude to say, Thou dost so and so” (Greenwood 1711: 110). Later in the century, Lowth states that “Thou in the Polite, and even in the Familiar style, is disused, and the plural You is employed instead of it” (Lowth 1762: 48). In this case, the statements of the grammarians accord with the evidence from 18th-century texts: by the middle of the century, use of thou had declined from “ungentile” to “disused”, at least in Standard English. (See further Busse, Chapter 46.) The only categorical innovation in Late Modern English syntax concerns what has variously been called the “be + -ing” construction, the “progressive” and the “expanded form”, as in She is reading a book. Although this construction had been used before 1700, its use in Early Modern English was optional in contexts where today it would be required. In Hamlet (II.ii.190), Polonius asks Hamlet “What do you read my Lord?” Today this would be interpreted as an inquiry into the prince’s reading habits, but Polonius was referring to the book that Hamlet had in his hands at the time. Today, the required construction in this pragmatic context would be “what are you reading, my Lord?” There was a remarkable increase in usage of this construction throughout the Late Modern period, both in terms of the sheer numbers of such constructions, and the types of clause in which it can occur, such that it becomes fully grammaticalized in the course of this period. From the second half of the 18th century, there is a rise

69

70

I Periods in its use with stative verbs such as love, wish, etc., with verbs denoting “instant” actions such as explode, fall etc., and with nominal and adjectival complements (e.g. You’re being a fool/foolish). The extension to the passive is likewise an innovation of the Late Modern period. Until the late 18th century, passive voice and progressive aspect could not both be marked in a clause, so that, if somebody wished to say that a house was under construction but not yet finished, the most acceptable way of expressing this would be “the house is building”. The passive was understood and the sentence unambiguous because houses can’t build themselves. The first examples of the passive with be + -ing are found in late 18th-century letters (1): (1)

I have received the speech and address of the House of Lords; probably, that of the House of Commons was being debated when the post went out. (Letters “to” Ser. Lett. 1st Earl Malmesbury; Denison 1998: 152).

Letters are, of course, the most informal genre of writing, and innovations are likely to be recorded earlier here than in more formal styles. It took some time for this construction to become accepted in all styles: even in the early 20th century Curme and Kurath seem to begrudge it: From 1825 on […] the form with being + perfect participle began to lead all others in this competition, so that in spite of considerable opposition the clumsy is being built became more common than is building in the usual passive meaning, i.e. where it was desired to represent a person or thing as affected by an agent working under resistance vigorously and consciously to a definite end: “The house is being built”, “My auto is being repaired” (Curme and Kurath 1931: 444).

The extension of the construction to longer verb phrases involving perfective and modal verbs came later. They appear in the late 19th century, but are roundly condemned by grammarians. By the early 20th century, such constructions are found in literary texts: they were rare then as now, simply because the pragmatic circumstances in which they might be used are rare. An early example is from Galsworthy (2): (2)

She doesn’t trust us: I shall always be being pushed away from him by her. (1915 Freelands; Denison 1998: 158).

A much fuller account of this and other syntactic changes of the Late Modern period can be found in Aarts, Lo´pez-Cuoso, and Me´ndez-Naya, Chapter 54. I have singled out the “be + -ing” construction here because it is one of very few areas in which the grammar of English has changed between 1700 and the present day to the extent that an earlier usage would seem ungrammatical to today’s speakers and, conversely, it is difficult for us to comprehend what 18th- and 19th-century grammarians found so “unnatural” in a sentence such as the house is being built.

5 Phonology The phonology of Late Modern English has, until very recently, had much less scholarly attention paid to it than that of earlier periods. This is probably because, as MacMahon

5 Periods: Late Modern English suggests, “superficially, the period under consideration might appear to contain little of phonetic and phonological interest, compared with, for example, earlier changes such as the transition from Old to Middle English, and the Great Vowel Shift” (MacMahon 1998: 373). However, from the sociohistorical point of view, it is a very rich period. The 18th century saw the beginning of the elocution movement and the publication of pronouncing dictionaries intended as guides to the “correct” pronunciation of English, i.e. that of educated, well-bred Londoners, whilst the 19th century witnessed the rise of Received Pronunciation (RP) (see Mugglestone, Volume 2, Chapter 121). Although the pronouncing dictionaries written by elocutionists such as Walker (1791) and Sheridan (1780) were undoubtedly normative, their detailed descriptions of sounds and transcriptions of every word in their dictionaries provide a wealth of evidence for the prestigious pronunciation of the period. A more detailed account can be found in Beal (1999) and Beal (2004: 125–167); see, also, Jones, Chapter 52: here, I shall briefly describe the main changes in the pronunciation of “received” English between 1700 and 1900. As far as consonants are concerned, the most striking development, at least as far as the pronunciation of English in England is concerned, is the loss of rhoticity. This is one of the most salient perceived differences between British and American English today (even though many varieties of British English are rhotic and several varieties of US English are non-rhotic). The detailed evidence from 18th- and 19th-century sources reveals that this change came “from below” and that it took a century for it to become accepted in the usage of RP speakers. Walker is often cited as the earliest source of evidence for the loss of rhoticity: “In England, and particularly in London, the r in lard, bard, card, regard, is pronounced so much in the throat, as to be little more than the middle or Italian a lengthened into baa, baad, caad, regaad” (Walker 1791: 50). But a careful reading reveals that he considered this to be a marker of lower-class London usage. Moreover, Walker describes the Irish pronunciation of /r/ as too harsh, but says that the pronunciation at the beginning of a word should be more “forcible” than at the end, so that “Rome, river, rage, may have the r as forcible as in Ireland, but bar, bard, card, hard &c. must have it nearly as soft as in London” (Walker 1791: 50). This suggests Walker was recommending a weakened /r/ in final and preconsonantal positions rather than total loss of rhoticity. As Mugglestone (1995: 98– 103) demonstrates, “dropping” of /r/ continued to be overtly stimatized until the late 19th century. Today, of course, it is the rhotic accents of England that are stigmatized as “rustic”. The other main consonantal changes in Late Modern English are not so much changes in the system, or even the distribution of phonemes, as the regulation of variants. Two of the greatest shibboleths of non-standard pronunciation in the 20th and 21st centuries are popularly known as “dropping” of and . In the latter case, the term “dropping” is not at all accurate, since the stigmatized variant is /n/ as opposed to /ŋ/ in, e.g. hunting, shooting and fishing. In both cases, the stigmatized variants had been attested at least from the Early Modern period, but are not labelled as “vulgar” or “incorrect” before the 18th century. Whilst “h-dropping” became the greatest social shibboleth of the 19th century, the alveolar pronunciation of has a more complex history. This pronunciation was a marker both of lower-class and upper-class usage throughout the Late Modern period and it can still be heard in the speech of very elderly, very conservative RP speakers. The stigmatization of alveolar thus

71

72

I Periods provides an intriguing example of a linguistic change of the Late Modern period in which middle-class speakers were in the vanguard, as has so often been seen to be the case in sociolinguistic studies of the 20th century. Such vowel changes as occurred in the Late Modern period largely involve the continuation of processes begun in the 16th and 17th centuries. In all cases, the earlier variants are still found in English regional accents, with many of the innovations still confined to RP and southern English varieties. The innovation known as the STRUT– ¯ and u˘ were unrounded to /ʌ/ in southern FOOT split, whereby some reflexes of ME o accents of (English) English, can be traced back to the 17th century. However, the lack of this split became stigmatized as “provincial” in the 18th century (and remains so to this day). Walker writes: If the short sound of the letter u in trunk, sunk etc., differ from the sound of this letter in the northern parts of England, where they sound it like the u in bull […] it necessarily follows that every word where this letter occurs must by these provincials be mispronounced (Walker 1791: xiii).

The other highly salient marker of the “north-south divide” in present-day accents of England, the pronunciation of the vowel in bath, laugh, grass, etc., has a more complex history. The lengthening and backing of this vowel seems, like the loss of rhoticity, to have been a change “from below”. Evidence of lengthening of /a/ in certain environments can be found in late 17th-century sources, but, throughout the 18th and 19th centuries, the pronunciation with /ɑː/ was not universally accepted. Walker tells us that, although “Italian a” had previously been heard in words such as glass, fast: “this pronunciation seems to have been for some years advancing to the short sound of this letter, as heard in hand, land, grand etc. and pronouncing the a in after, answer, basket, plant, mast, etc. as long as in half, calf etc. borders very closely on vulgarity” (Walker 1791: 10). This change seems to have begun as a lengthened /æː/ in the 17th century, and not to have become stigmatized until the lengthened vowel was retracted to [ɑː] The latter pronunciation is described as “drawling” throughout the 19th century, and there is evidence of a pronunciation with [æ] or even [ɛ] by young ladies wanting to avoid the “vulgar” [ɑː]. Those who wished their pronunciation to be beyond reproach had to avoid both the “drawling” [ɑː] and the “mincing” [æ] at least until the beginning of the 20th century, when Daniel Jones’s use of cardinal [ɑ] seems to have established this as the RP pronunciation. The other vowel changes to be considered here could be regarded as the tail-end of the Great Vowel Shift. In words such as face and goat, the ME vowels had been raised to /eː/ and /oː/ respectively, and these are the pronunciations recommended by Walker, who describes them as the “long, slender” sound of and the “long, open” sound of respectively, both of which are unequivocally monophthongal. However, the first evidence for diphthongal pronunciations of both these vowels comes very soon after Walker’s first edition: MacMahon (1998: 459) points out that the first evidence for diphthongization of /oː/ comes from the Scottish orthoepist William Smith in 1795 and it is generally accepted that the first attestation of a diphthongal pronunciation of /eː/ comes from Batchelor (1809). (In both cases, the diphthongal pronunciations are widely accepted in the 19th century, and are still found in RP and many other accents of present-day English.)

5 Periods: Late Modern English

73

6 Lexis The picture that we have of lexical innovation in the Late Modern period has largely been determined by the policies and practices of the Oxford English Dictionary (OED) (Simpson [ed.] 2000–). Until very recently, any research of a statistical nature into lexical innovation in this period (and earlier ones) was based on the Chronological English Dictionary (CED) (Finkenstaedt et al. 1970), which in turn took its data from the Shorter Oxford English Dictionary [SOED] [Onions 1964]). The CED has proved immensely useful to historians of English, as it reorganizes entries in the SOED by year of first citation and provides a numerical code to indicate the etymology of each word. For each year, the total number of first citations is given at the end of the table. Figure 5.1 is based on the total number of first citations for each decade of the Late Modern Period. 3000 2500 2000 1500 1000 500

0

0 –1

90

0

89

–1 18

91

0

88

–1 18

81

0

87

–1

61

18

18

71

0 –1

86

0

85

–1

41

18

18

51

0 –1

84

0

83

–1 18

31

0

82

–1

11

18

18

21

0 –1

81

0

80

–1

91

18

17

01

0 –1

79

0

78

–1

71

17

17

81

0 –1

77

0

76

–1 17

61

0

75

–1

41

17

17

51

0 –1

74

0

73

–1

21

17

17

31

0 –1

72

0

71

–1 17

11

0

70

–1 17

01

0

69 16

91

–1

68 81

–1 16

71

16

16

61

–1

67

0

0

Figure 5.1: Numbers of first citations in SOED by decade 1661–1900

What is immediately apparent from Figure 5.1 is that lexical innovation is at a low rate through the late 17th and 18th centuries, and then rises to a peak in the mid-19th century before a further sharp decline towards the end of the century. However, the authors of the CED acknowledge that “the vocabulary of the twentieth century is less systematically represented than that of the preceding periods” (Finkenstaedt et al. 1970: xi). This almost certainly accounts for the steep drop shown in Figure 5.1 from the end of the 19th century onwards: since the SOED is based on the first edition of the OED, it reflects the coverage of the latter, work on which began in earnest in 1879, but which was not published in its entirety until 1928. It is hardly surprising that innovations of the late 19th and early 20th centuries were not recorded, as the earlier fascicles, covering earlier letters of the alphabet, would have been completed by then. The apparent “trough” in the 18th century can also, to some extent, be accounted for by the practices of the OED, as the compilers of the first edition excerpted texts from the 18th century much less exhaustively than those from earlier periods or from the 19th century. This is now being remedied by the compilers of the third edition of the OED, so that research based on the OED online will give a more accurate picture

74

I Periods of lexical innovation in this period. Nevertheless, it still seems to be the case that there was less innovation in the 18th century than in the early 17th or mid-19th centuries. It is tempting to attribute this decline in innovation to the conservatism of the age. Celebrated authors such as Swift and Addison ridiculed the pretentious innovations of the age in the periodicals of the time such as the Tatler and the Spectator, but the same kind of discourse can be found in today’s newspapers whenever a dictionary releases a list of new words. If there was no innovation, they would have nothing to satirize. It is more likely that the rate of lexical innovation in the 18th century seems low by comparison with the earlier period, when the move towards printing in English created a gap in learned vocabulary that needed to be filled, and the 19th century, when the sheer pace of scientific progress demanded new words for new discoveries. Even in the “conservative” 18th century, it was acknowledged that everything in nature must have a name, so innovations which filled gaps in the vocabulary and which added to the clarity of nomenclature were accepted. Many of these were introduced via works such as Chambers’ Cyclopedia (Chambers et al. 1753): a search of OED online for all words with first citations from 1753 yields a total of 538, 273 or just over 50% of which were first cited in that year’s supplement to the Cyclopedia. Not surprisingly, these were mainly learned, scientific words with Latin and/or Greek etymologies. Examples are: archivist, cotoneaster, eczema, hydrangea, linguiform, phosphorical and trifoliate. Along with these classical forms, Chambers introduced words from “exotic” languages to describe the findings of explorers and plant hunters, and eponyms which name the object after its discoverer or inventor. Examples of the former are jacaranda and jacare (an alligator), both from Tupi, whilst the eponyms include camellia (named after the Moravian Jesuit botanist Kamell) and fuchsia (from the German botanist Fuchs). There are also examples of compounds, such as boat-fly, bull-fight, and butter-nut and of introductions from French, such as ballon. These trends continued in the 19th century, but on a much larger and wider scale. The vast majority of lexical innovations in this century were scientific terms formed from Latin and Greek elements, such as the medical terms conjunctivitis, myelitis, and synovitis, the botanical terms bicrenate, bifoliate, brachiate, and campylotropous, and the less technical abnormal (replacing French anormal), intensifier, paraffin, and revolver, all of which are first cited in 1835. The influx of classical formations became a cause for complaint, just as it had during the “inkhorn wars” of the 16th century. R. Chenevix Trench, one of the founders of the OED, wrote of such scientific terms: These, so long as they do not pass beyond the threshold of the science for whose use they were invented, have no proper right to be called words at all. They are a kind of shorthand, or algebraic notion of the science to which they belong, and will find no place in a dictionary constructed upon true principles (Trench 1870: 120–121).

New fields of science opened up in the course of the 19th century, requiring their own technical vocabularies. In the field of geology, newly-discovered minerals were often formed by adding the suffix -ite to the name of the discoverer or the place where they were first found. Examples are bromlite, lanarkite, leadhillite, proustite, smithsonite, stromeyerite, troostite, uralite, and voltzite, all first cited in 1835. Likewise the -itis suffix found in conjunctivitis, myelitis, and synovitis was to become very productive as a means of

5 Periods: Late Modern English naming inflammations of various body parts. Thus the scientific vocabulary snowballed in this century as the proliferation of terms provided new building-blocks for future innovation. The 19th century was also, of course, the era of British colonial expansion and imperialism. This is reflected in the wide range of sources from which words were taken in this century: just as British plant-hunters, archaeologists, and explorers plundered the world’s resources and brought them home, so the names of these trophies brought elements from “exotic” languages into the vocabulary. Examples, again all first cited in 1835, from every continent outside Europe are kiwi, rata, and tui from Maori; chacma from the African language Nama; fulwa, the Bengali word for the butter tree; nandu from Tupi/Guarani; and tepee from Lakhota. I have only been able to give a flavor of lexical innovation in the Late Modern period here: much more detail can be found in Dossena, Chapter 55. The most important point to be borne in mind is that our knowledge of lexical innovation in this, and indeed any, period, can only be as extensive and accurate as the sources available for our research. As the third edition of the OED reaches completion, we can look forward to new research findings in this area.

7 Normative grammarians In this section, I shall give a brief overview of research on what, to earlier scholars, was the main area of interest in Late Modern English: the rise of prescriptivism in the 18th century and the effect of normative texts on standardizing and regulating the language. As I have indicated above (Section 2), the seminal text in this field was S. A. Leonard’s The Doctrine of Correctness in English Usage 1700–1800 (1929). This was so influential that it led a whole generation of scholars to describe the 18th century as the age of correctness and to present 18th-century grammarians as unenlightened prescriptivists who imposed arbitrary rules on the language on the basis of logic, analogy, or conformity to the rules of Latin. In the past decade, a more nuanced view of these grammarians has come to the fore, facilitated by the wider availability of texts from this period provided by databases such as Eighteenthcentury Collections Online (Gage Cengage Learning 2009) and corpora of Late Modern texts. This has enabled scholars to scrutinize more carefully what the grammarians actually wrote and to compare their precepts with actual usage of the period. For instance, the myth that Robert Lowth “invented” the rule that a preposition should not appear at the end of a sentence still endures in cyberspace, but Ya´n˜ez-Bouza has demonstrated that he did no such thing. After examining a wide range of statements from grammarians on this topic she concludes that “Lowth was neither the first one, nor (ipso facto) was he the only one, nor was his stricture proscriptive” (Ya´n˜ez-Bouza 2008: 277). The role of prescriptive grammarians in relegating constructions such as the “double negative” to non-standard usage in this period has also received attention. Greenwood’s statement that “two Negatives or two Adverbs of Denying, do in English affirm” (Greenwood 1711: 160) is often cited as an example of mathematical logic inappropriately applied to language. Yet Nevalainen and Raumolin-Brunberg (2003) demonstrate that multiple negation was already subject to social stratification in the Early Modern period, its use largely confined to the lower classes. Thus, Greenwood may well have been rationalizing a stigma that already existed, rather than imposing an arbitrary rule on the language. Research in this area is ongoing, but the collection of papers in

75

76

I Periods Tieken-Boon van Ostade (ed.) (2008) provides a good introduction, and a salutary riposte to Leonard (1929). (See further Percy, Chapter 63.)

8 Summary This brief overview of scholarship in the field of Late Modern English has demonstrated that this is a young and growing area of research. The ever-increasing availability of corpora and electronic editions of texts from this period has led to an upsurge of interest from scholars in a number of fields. Their work challenges the previously received view that the Late Modern period was a time of relative linguistic stasis and normative attitudes to language and reveals the “complex analogical relationships” predicted by Strang (1970: 79).

9 References 9.1 Online resources Denison, David, Linda van Bergen and Joana Soliva. 2002. The English Language of the North-west in the late Modern English Period: A Corpus of late 18c Prose. http://www.llc.manchester.ac.uk/temp/ lel/david-denison/corpus-late-18th-century-prose/ Gage Cengage Learning. 2009. Eighteenth Century Collections Online (ECCO). Gale Digital Collections. http://mlr.com/DigitalCollections/products/ecco/ Nevalainen, Terttu, Samuli Kaislaniemi, Mikko Laitinen, Minna Nevala, Arja Nurmi, Minna Palander-Collin, Helena Raumolin-Brunberg, Anni Sairio (ne´e Vuorinen), and Tanja Sa¨ily. 1998–2006. Corpus of Early English Correspondence Extension (CEECE). Department of English, University of Helsinki. http://www.helsinki.fi/varieng/CoRD/corpora/CEEC/ceece.html Simpson, John (ed.). 2000–. The Oxford English Dictionary. 3rd edn. online. Oxford University Press. www.oed.com/

9.2 Printed resources Bailey, Richard W. 1996. Nineteenth-century English. Ann Arbor: University of Michigan Press. Barber, Charles. 1976. Early Modern English. London: Deutsch. Batchelor, Thomas. 1809. An Ortho¨epical Analysis of the English Language. London: Didier and Tebbett. Beal, Joan C. 1999. English Pronunciation in the Eighteenth Century: Thomas Spence’s “Grand Repository of the English Language” (1775). Oxford: Clarendon Press. Beal, Joan C. 2000. From Geordie Ridley to Viz: Popular literature in Tyneside English. Language and Literature 9/4: 343–359. Beal, Joan C. 2004. English in Modern Times 1700–1945. London: Arnold. Beal, Joan C. 2008. Shamed by your English? The market value of a “good” pronunciation. In: Joan C. Beal, Carmela Nocera, and Massimo Sturiale (eds.), Perspectives on Prescriptivism, 21–40. Bern: Peter Lang. Beal, Joan C. 2009. Three hundred years of prescriptivism (and counting). In: Ingrid Tieken-Boon van Ostade and Wim van der Wurff (eds.), 35–66. Beal, Joan C. 2011. Levelling and enregisterment in Northern dialects of Late Modern English. In: David Denison, Ricardo Bermu´dez-Otero, Chris McCully, and Emma Moore (eds.), Analysing Older English 126–139. Cambridge: Cambridge University Press. Bloomfield, M. W. and L. Newmark. 1963. The English Language: A Historical Introduction. Westport, CT: Greenwood.

5 Periods: Late Modern English ´ lvarez, Javier Pe´rez-Guerra, and Esperanza RamaBueno Alonso, Jorge L., Dolores Gonza´lez A Martı´nez (eds.). 2007. “Of Varying Language and Opposing Creed”: New Insights into Late Modern English. Bern: Peter Lang. Chambers, Ephraim, John L. Scott, and John Hill (eds.). 1753. A Supplement to Mr. Chambers’s Cyclopedia: or Universal Dictionary of Arts and Sciences. London: printed for W. Innys and J. Richardson et al. Curme, George O. and Hans Kurath. 1931. A Grammar of the English Language. Boston: D. C. Heath & Co. De Smet, Hendrik. 2007. For … to infinitives as verbal complements in Late Modern and Presentday English: Between motivation and change. English Studies 88(1): 67–94. Denison, David. 1998. Syntax. In: Romaine (ed.), 92–329. Dossena, Marina and Charles Jones (eds.). 2003. Insights into Late Modern English. Bern: Peter Lang. Finkenstaedt, Thomas, Ernst Leisi, and Dieter Wolff. 1970. A Chronological English Dictionary: Listing 80,000 Words in Order of their Earliest Known Occurrence. Heidelberg: Carl Winter. Freeborn, Derek. 1992. From Old English to Standard English. 1st ed. London: Macmillan. Go¨rlach, Manfred. 1999. English in Nineteenth-century England. Cambridge: Cambridge University Press. Greenwood, James. 1711. An Essay Towards a Practical English Grammar. London. Jones, Charles. 1989. A History of English Phonology. London: Longman. Jones, Charles. 2006. English Pronunciation in the Eighteenth and Nineteenth Centuries. Basingstoke: Palgrave Macmillan. Lass, Roger. 1999. Introduction. The Cambridge History of the English Language. Vol. III. 1476– 1776, 1–12. Cambridge: Cambridge University Press. Leonard, Stirling A. 1929. The Doctrine of Correctness in English Usage 1700-1800. Madison: University of Wisconsin Press. Lowth, Robert. 1762. A Short Introduction to English Grammar. London: R. & J. Dodsley. MacMahon, Michael. 1998. Phonology. In: Romaine (ed.), 373–535. Mair, Christian. 2002. Three changing patterns of verb complementation in Late Modern English: A real-time study based on matching text corpora. English Language and Linguistics 6(1): 105–131. Morris, Marmeduke C. F. 1892. Yorkshire Folk-Talk: with Characteristics of those who Speak it in the North and East Ridings. London: Henry Frowde. Mugglestone, Lynda. 1995. “Talking Proper”: The Rise of Accent as a Social Symbol. Oxford: Clarendon Press. Nevalainen, Terttu and Helena Raumolin-Brunberg. 2003. Historical Sociolinguistics: Language Change in Tudor and Stuart England. London: Longman. Newton, Isaac. 1687. Philiosophiae Naturalis Principae Mathematica. London: Jussu Societatis Regiae ac typis Josephe Streatii. Onions, C. T. (ed.). 1964. The Shorter Oxford English Dictionary on Historical Principles. 3rd edn. Revised with addenda. Oxford: Clarendon Press. Porter, Roy. 2000. Enlightenment: Britain and the Creation of the Modern World. London: Penguin. Poutsma, Hendrik. 1928. A Grammar of Late Modern English. For the Use of Continental Students. 2nd edn. Groningen: Noordhof. Rohdenburg, Gu¨nter. 2007. Functional constraints in syntactic change: The rise and fall of prepositional constructions in early and Late Modern English. English Studies 88(2): 217–233. Romaine, Suzanne. 1982. Socio-historical Linguistics: Its Status and Methodology. Cambridge: Cambridge University Press. Romaine, Suzanne (ed.). 1998. The Cambridge History of the English Language. Vol. IV. 1776– 1997. Cambridge: Cambridge University Press. Rose, Michael E. 1985. Society: The emergence of urban Britain. In: Christopher Haigh (ed.), The Cambridge Historical Encyclopedia of Great Britain and Ireland, 276–281. Cambridge: Cambridge University Press.

77

78

I Periods Sheridan, Thomas. 1780. A General Dictionary of the English Language. London: R. & J. Dodsley, C. Dilly and J. Wilkie. Smith, William. 1795. An Attempt to Make the Pronunciation of the English Language More Easy to Foreigners. London: T. Gillet. Strang, Barbara M. H. 1970. A History of English. London: Methuen. Tieken-Boon van Ostade, Ingrid. 2009. An Introduction to Late Modern English. Edinburgh: Edinburgh University Press. Tieken-Boon van Ostade, Ingrid (ed.). 2008. Grammars, Grammarians and Grammar-writing in Eighteenth-century England. Berlin/New York: Mouton de Gruyter. Tieken-Boon van Ostade, Ingrid and Wim van der Wurff (eds.). 2009. Current Issues in Late Modern English. Bern: Peter Lang. Trench, R. Chenevix. 1870. History in English Words. 7th edn. London: Macmillan. Walker, John. 1791. A Critical Pronouncing Dictionary. London: G. G. J. and J. Robinson and T. Cadell. Webster, Noah. 1789. Dissertations on the English Language. Boston: Isaiah Thomas and Company. Webster, Noah. 1828. An American Dictionary of the English Language. New York: S. Converse. Wright, Joseph. 1898–1905. The English Dialect Dictionary. London: Henry Frowde. Ya´n˜ez-Bouza, Nuria. 2008. Preposition stranding in the eighteenth century: Something to talk about. In: Tieken-Boon van Ostade (ed.), 251–277.

Joan C. Beal, Sheffield (UK)

6 Periods: Contemporary English 1 2 3 4 5 6 7

Introduction Phonology Morphology Vocabulary Syntax Summary References

Abstract This chapter sketches current changes in phonetics-phonology, morphology, vocabulary, and syntax, focusing on data from British, New Zealand, and Australian English. Central themes are the interplay of spoken and written, standard and non-standard English, changes in linguistic practices as opposed to changes in phonology, grammar, and vocabulary, the many changes affecting all geographical varieties of English versus changes confined to particular varieties. The changes in vocabulary illustrate the usual close connection between language and cultural change. Many of the apparent changes in grammar consist of old constructions hitherto used only in speech but now spreading into written texts. They raise questions of analysis bearing on classification (middles or passivals?), constituent structure, and particular areas of grammar, such as the tense-aspect system and modality. As usual, lying behind the changes in language and practices are issues of social and cultural capital. Alex Bergs and Laurel Brinton 2012, Historical Linguistics of English (HSK 34.1), de Gruyter, 78–96

6 Periods: Contemporary English

1 Introduction 1.1 Content This chapter presents an overview of changes in progress in English. Section 1 discusses general issues, Sections 2, 3, 4, and 5 deal with, respectively, changes in phonology, morphology, vocabulary, and syntax. Section 5 on syntax, morpho-syntax and lexico-syntax is much longer than the others. One reason is that anyone looking for further information about the phonology or morphology of any variety of English will find it in Kortmann and Schneider et al. (2004–08). Another reason is that the author of this chapter has for many years worked on the syntax of English, spoken and written, standard and non-standard and can comment on constructions not usually mentioned in corpus-based work. He also believes that syntax raises most questions about changing norms and language practices, the relationship between spoken and written language, what counts as standard and non-standard, and whether perceived changes involve new structures or a wider domain for already existing ones.

1.2 Changes in grammar and changes in language practices The question of ongoing change in contemporary English is complex, encompassing changes in the grammatical code and in language practices. One example of change in language practices is what Mair and Leech (2006: 336) call “colloquialization”, the use in written language of features associated with spoken language: semi-modals (e.g. you want to do it this way), the get passive, that or zero relative clauses, and singular they. None of these is new; in fact the last three are attested in the early 17th century. Some of them have their own functions and meanings: singular they enables writers to avoid the clumsy he or she, the get passive is dynamic and is frequently, but by no means exclusively, used for adverse events, and that is the norm in restrictive relative clauses in American English. Some of the changes presented in Section 5 relate to the use in writing of structures common in spoken English and might be counted by Mair and Leech as colloquialization (N.B. “written English” and “spoken English” are convenient simplifications; there are many genres of both). However, these structures are found in both formal and informal spoken English, and raise various questions: How old are they? What is their constituent structure? In what types of written and spoken text are they found? Which categories of users use them? These are important questions but answers to them require intensive research and detailed discussion and cannot be accommodated here. These spoken English structures are central to Section 5 because they are used by speakers of both standard and non-standard English and in combination with both standard and non-standard syntax. A major question is the extent to which regional, nonstandard norms will be followed by speakers and writers producing more formal, public texts other than dialogue in novels and plays. Much informal Internet writing already follows the norms of informal speech, norms that may be standard or non-standard but which all relate to spoken, especially spontaneous spoken, English and represent colloquialization.

79

80

I Periods

1.3 Recognizing recent or on-going changes Linguistic practices and usages are changing but typically involve the extension of existing constructions and words. Mair and Leech (2006) warn that long-term variability can be mistaken for change, that apparent changes may simply be the abandoning of “marginal shibboleths” (less books vs. fewer books), and that long-term changes may be missed. Denison (1998: 95) points out that an aberrant usage might become a successful change, might be adopted but not generally, or might remain an error not established in anybody’s usage. Denison also warns us that apparently new usages might be old ones previously confined to informal spoken English, standard or non-standard, but now appearing in formal speech and in writing. Denison’s point is supported by the examples from Mair and Leech (2006) listed in the opening paragraph, which have long been typical of spontaneous speech. A further point, emphasized in Miller (2006), is that it is difficult to classify constructions as general spoken or regional non-standard. Some apparent changes might simply be once-general usages that were restricted for a time and are now being reintroduced. (See the discussion in Section 5.1.1.) The use in British English (BrE) sat/stood instead of sitting/standing illustrates the difficulty. Carter and McCarthy (1997: 34) attribute the utterance the pilot was sat in one of the seats to “Yorkshire dialect”. In contrast, Cheshire et al. (1993: 70–71) show that BrE sat/stood is widespread and characteristic of “a general non-standard or semi-standard variety of English”. Trask (2001) reflects the uncertainty, describing was sat as “colloquial British English” and was stood as “regional British English”. Burchfield (1981), writing for the BBC, declared was sat/ stood there unacceptable in any circumstances, but almost thirty years on, the structure is used by, e.g., reporters on the BBC News at Ten (though not by the presenters). Not only is it possible that many structures considered “non-standard” are actually spoken standard, but changes in formal spoken norms and practices may be evolving much more quickly than previously thought.

1.4 The data The syntactic and morpho-syntactic data in this chapter come mainly from the author’s own database, collected over the past thirty years. Some examples were collected on the hoof from informal conversation, university meetings, and radio and television programs (some of which were recorded); some are from transcriptions of conversation and business meetings; and there are written examples from newspapers, e-mails, minutes of meetings, students’ dissertations and examination answers, and books. The discussion focuses on British and New Zealand English (NZE), which the author has experienced directly, but with supplementary data from the Australian component of International Corpus of English (ICE; for information, see http://ice-corpora.net/ice/) in the Macquarie Archive and the 4-volume Varieties of English (Kortmann and Schneider et al. 2004–8). With respect to syntax, the BrE and NZE data are typical of many other varieties of English. Many recent discussions are based on exhaustive computer searches of corpus data, which are invaluable for information about lexical collocations and morpho-syntax but not necessarily about syntax, especially the syntax of spoken language (see the discussion in Miller and Weinert 1998: Chapter 1). The trained specialist with native or

6 Periods: Contemporary English near-native knowledge of English and handling attentively all sorts of spoken and written texts on a daily basis can pick up many clues as to variant usages and possible ongoing changes. Hypotheses about changes can then be checked against corpus data (which may be behind the times) or via elicitation tests. The data presented here can confidently be taken as representing on-going changes in linguistic code or linguistic practice because of the time span, the range of speakers and writers and text types, and comparisons with other databases.

2 Phonology All geographical varieties of English evince rich phonological variation and change. Some changes are confined to particular varieties while others affect many. One example of the former is the Northern Cities Shift, six changes affecting the vowel system in the variety spoken in Chicago, Buffalo, Cleveland, and Detroit (Gordon 2008; Labov 1994, 2001). A second example is the NZE vowel in dress, which has become as close as the vowel in fleece, whereas in Australian English (AustrE) the dress vowel has become more open. The NZE vowel in trap has raised and fronted and the kit vowel has become backed and lowered (kit is allegedly perceived by Australians as cut). In modern Received Pronunciation (RP) the dress vowel has become more open, the trap vowel less forward and more open, the price diphthong more central and the square vowel a long Cardinal 3 and not a diphthong, according to Upton (2008: 241), who also mentions the falling together of paw and poor. For NZE, Bauer and Warren (2008) mention the on-going merger of the near and square diphthongs. For AustrE Horvath (2008: 105) lists ten changes in the vowel system between the 1960s and the 1990s. Most vowel changes are variety-specific, but one change affecting most if not all varieties of English is the fronting of the foot vowel. A widespread change affecting consonants is l-vocalization, observed in London and South East England, and potentially causing pairs such as meal and mill and pool and pull to fall together (Altendorf and Watt: 2008). The same l-vocalization after front vowels and in syllabic /l/ has been studied in Glasgow by Stuart-Smith et al. (2006), who note that the new vocalization exists alongside the enduring results of the old Scots l-vocalization after back vowels which gave ba’ for ball. L-vocalization is widespread in AustrE (Horvath: 2008) and is present in NZE, where it is associated with the neutralization of the vowels in pairs such as fill and full and full and fool. Another widespread change is t-glottalisation, occurring throughout the UK, as attested by the articles in Foulkes and Docherty (1999), even among speakers of modern RP (see Fabricius 2002). For varieties in London and the South East, Altendorf and Watt (2008: 212) refer to a “dramatic rise” in the frequency of /r/ realized as a labio-dental approximant in the speech of young, working-class speakers. A prosodic change affecting all varieties of English is the appearance of the High Rising Terminal (HRT). Horvath (2008) and Bauer and Warren (2008) comment on AustrE and NZE, where it is a feature of many speakers under forty. By 1990 the HRT was spreading among speakers of British English under thirty, and it is thoroughly established in the USA and Canada. The vowel mergers mentioned above are phonological in nature, but changes such as the fronting of the foot vowel, t-glottalization, and the realization of /r/ as a labio-dental

81

82

I Periods approximant are phonetic. Also phonetic, but altering the phonotactics of the affected varieties, are changes in NZE (and observed elsewhere, Laurie Bauer, p.c.) such as combinations of palato-alveolar affricate plus /r/ in try and dry, and the use of an initial palato-alveolar fricative in words such as strong and stripe. It is tempting to see some of the NZE and AustrE changes as part of a general differentiation from BrE, as is happening with vocabulary (see Section 4).

3 Morphology Observers can find many changes in linguistic practices and new uses for old forms, but recently created forms are rare apart from new plurals and singulars. Collins and Peters (2008: 342) report that standard AustrE has the past tense forms shrunk and sunk, which they consider unacceptable. However, sunk is an old form that still occurs in speech and writing in BrE and both forms are widely used in North America (Laurel Brinton, p.c.). Users of AustrE are not inventing new past tense forms but favoring one existing form over another. Changes in practice are affecting noun plurals, witness the fashion for plural forms such as fora, referenda, stadia, and corpora; Trask (2001: 130) says “fora is now increasingly used, especially by academics”. Another change affects adjectives used as nouns: plural forms such as renewables and sharps (in medical usage, sharp instruments, including needles) are common (but note examples of long standing like shallows and deeps and the names of mountains ranges such as The Remarkables and The Rockies; instructions on the use of washing machines refer to delicates and woolens, and cottons and linens). Some plural nouns are beginning to be used as singulars. Examples are bacteria, now regularly so used, and premises – notices in several shop windows in Edinburgh declare This premises is under CCTV surveillance. There are plausible reasons for the new uses. The singular bacterium is rarely seen outside technical literature, and a collection of internal spaces is most frequently referred to by singular nouns such as shop, office, house. A number of originally mass nouns are now also used as count nouns. Rugby players are said (not) to get a lot of ball, university administrators are interested in how much grant they receive, and a notice in the Western General Hospital in Edinburgh advises medical staff to avoid taking unnecessary bloods (equivalent to samples of blood not types of blood, unlike wines, butters, and so on). (Laurie Bauer, p.c., observes that the increase in the expressions such as amount of people vs. number of people indicates a more general change involving the concepts of count and mass.) Changes in derivational morphology typically consist of increased usage of a particular affix and its attachment to a larger set of words. Bauer (1994: 40–47) discusses the increasing use of -ee with non-passive meaning – escapee, absentee vs. payee, nominee, and its new use with nouns denoting non-human entities – advancee, controllee. Denison (2008) refers to a fashion among students at Manchester University for using the suffix -age, often humorously. He cites There was some general sleepage/chattage/faffage. Jespersen (1961: 436) cites chattage from a novel by Marjory Allingham (but since her novels are out of fashion and print, we can assume that the students’ use is an independent invention). In Britain outage began to replace power cut in the 1980s in connection with power cuts in North America but the Oxford English Dictionary (OED) supplies an example from 1900, also in relation to power cuts in the United States.

6 Periods: Contemporary English Jespersen (1961: 173–183) exemplifies different types of what he calls “reduplicative compounds”, such as those many many bodies (Shakespeare) or an old, old man (Dickens and Shelley). The construction is apparently becoming more frequent, sometimes with an intensifying meaning, as in It’s a big big fish (television presenter) or That’s a long long street. Note too Do you really really like me? (from the film Restless Natives). An interviewee on Channel 4 News in the UK, discussing poverty, said they’re not poor poor; that is, the persons referred to were not poor in an absolute sense but relatively poor, receiving less than 60% of the median wage. Nouns too combine, signalling ‘this X is (not) really something that is properly classed as an X’. Examples heard recently by the writer are It’s not a problem problem (i.e. no need to worry, it’s easily solved) and In spite of his age he’s a student student (i.e. he has signed up for courses and is genuinely doing the work).

4 Vocabulary The preface to the 2nd edition of the OED (Simpson and Weiner 1989: xxiv) distinguishes words of unquestioned “Anglicity” and words of doubtful “Anglicity”, including “local dialect and slang”, the technical terms of trades, processes, and science. As will be seen below, technical terms and local dialect are tenacious. Copious borrowing from other languages is usually considered a hallmark of English, but Minkova and Stockwell (2006: 483) suggest that “English has turned inwards to its own resources for new words and new readings”, and appeal to the fact that of the new words in English between 1963 and 1981 only 7.5% were borrowed. This general picture is supported by the Cambridge Advanced Learner’s Dictionary (Good [ed.] 2008), which analyzes newly included words. Out of 120 items, 20 are borrowings, much more than 7.5%, but, significantly, all under the rubric “food and drink”: e.g. basmati rice, blini, chorizo, nigiri. The other new words are either compounds from existing words or roots (grey water, energy efficient, biofuels, biomass) or old words with a new meaning (troll, hybrid). Many additions come from technology, especially computing and the Internet. Technical terms in business and commerce also find their way into general use. The dictionary lists brand aware, cool hunter, flick factor, and flexible working (but just missed credit crunch, toxic debts, and subprime). Many new items are compounds (see Mair and Leech’s comment [2006: 333] about a possible resurgence of the Germanic noun + noun sequences). Borrowing has not been abandoned by all users of all varieties of English. Consider NZE and the practices of certain users in Scotland. New Zealand society is establishing its own norms (see Schneider 2003). Cultural independence is signalled very clearly by vocabulary. New Zealanders use Antipodean vocabulary of English origin but, strikingly, Maori vocabulary too. Macalister (2006) demonstrates an early wave of borrowing into NZE from Maori (1841–1880) and, from 1970, a second period of borrowing. The latter coincides with a reviving Maori society and a growing interest in Maori culture among white New Zealanders, who had to break free of cultural (and economic) dependence on Britain when the latter joined the Common Market in 1973. Non-standard varieties too undergo lexical change, and Scots shows how complex the process can be. Macafee (1994) charts the serious loss of traditional Scots vocabulary among working-class Scots, whereas Hardie (1996) refers to the knowledge among professionals of Scots vocabulary in literature, the law, and local authority documents.

83

84

I Periods Against this background the Scottish Parliament, reconstituted in 1999 after a break of 292 years, has on its website documents in classic Scots, the former language of government. An excerpt from a document on language awareness in schools is given in (1). (1)

Amang the ettles o the study is: Tae speir oot the kind o wittins anent the languages o Scotland wantit … by education (www.scottish.parliament.uk/msp/crossParty Groups/groups/scots/SoP%20version%202.PDF; accessed 4 May 2008)

Whit ‘what’, maist ‘most’, mak ‘make’ are still in current, regular use among many speakers in Scotland. Not in regular use is speir ‘ask’; fairly esoteric are wittins ‘knowledge’, ettles ‘aims, intentions’, and anent ‘concerning’, though the last occurs regularly in legal documents. The authors of this and other documents on the same web site have gone back to written texts and revived vocabulary that otherwise was in disuse. To sum up, in the English-speaking world current developments in vocabulary are more complex and varied than might be thought, with lapses into disuse, revivals, borrowings and extensions of meaning. And in various locations all these developments accompany attempts by groups of speakers to use “non-standard” vocabulary in text types and contexts in which it is not accepted by the owners of social capital and cultural power.

5 Syntax 5.1 Middles The middle construction, as in Nothing drives like a Ford Falcon (New Zealand advertising slogan), is reviving and spreading. The middle is generally thought to exclude the progressive, to have a generic or habitual interpretation, and to require adverbs of manner. However, (2a, b) refer to single events and are identical in form and meaning with an example such as The story told well (1815 Jane Austen, Emma Chapter 23, paragraph 4). (2)

a. […] a 1912 Silver Ghost sold for £1.5m in California (2007 The Herald Scotland, Dec. 4) b. skylarks […] soon established throughout the country (Gill et al. 1994: Bird B)

There is controversy over what counts as an instance of the middle construction. In her exhaustive investigation Hundt (2007: 141–147) distinguishes mediopassives – nonprogressive, generic, focusing on inherent properties – from passivals, possibly progressive, definitely non-generic, and not focusing on inherent properties. The distinction is difficult to apply. She proposes that generic mediopassives focus on inherent properties by means of manner adverbs, modal verbs and negatives. She treats (3) as a middle/mediopassive, but a temporary transmission failure is not an inherent property; (4) she treats as a marginal middle because it describes an inherent property but is in the progressive. (3)

The fax may send if you tried again. (Hundt 2007: 142)

(4)

The 1971s, at only £1000 a bottle, are drinking so much better at the moment. (1998 Private Eye, Feb. 20, p. 7; Hundt 2007: 143)

6 Periods: Contemporary English The proposed distinction, more semantic than syntactic, is here disregarded. We propose a single middle construction with clear semantic and discourse properties which is reclaiming grammatical properties it possessed earlier and extending its range of lexical items. Note that the middle and the passive differ distinctly in syntax and meaning. The passive has its own passive morphosyntax, while the middle requires active verb forms. The passive allows optional agent phrases, the middle does not. The passive presupposes an agent, the middle does not. The middle presents an entity as controlling (Kemmer 1993) a given situation, but the passive does not. Bolinger (1968: 130) talks of “self-propelled” activities, as in The coffee is making, which applies neatly to a British Army news briefing in 2003: there were three bombs that didn’t guide for one reason or another some of them went short. The blame fell on the bombs (just as the child saying The vase just broke is blaming the vase, and might add all by itself ). Since the properties described above apply both to “mediopassives” and “passivals”, the distinction is not worth drawing. Hundt’s impressive range of middles is supported by the author’s modest data, 60 examples with 54 different lexical verbs from BrE and NZE showing that the middle is regularly used in speech and writing wherever a non-agent participant is perceived as the controller, as in (5). (5)

a. if the features are privative and require no value, then they simply check in the way that we have already seen […] (Adger 2003: 169) b. The lawsuit […] claims that the nano scratches “excessively during normal usage”. (2006 New Zealand Herald) c. that’s processing for us now (Scottish Gas employee referring to an invoice being prepared by the computer)

5.2 Indirect questions In Fiona asked if we were going to France the clause if we were going to France is an indirect question. It has the word order of a declarative clause, is introduced by if, and conveys the content of a question. The speaker’s direct question, Are you going to France?, is not reproduced. This is the construction typical of formal written English. An alternative construction is Fiona asked were we going to France, in which the complement clause has the subject-auxiliary inversion of direct questions. Quirk et al. (1985: 1052) say the construction is “common in Irish English and dialectally”. Denison (1998: 246), following Henry (1995), asserts that the construction is normal in Ulster English, Welsh English, recent American English, and the New Englishes. In fact, the “alternative” construction is typical of spontaneous spoken English in all varieties, but now regularly occurs in newspapers and in written texts not subject to copy-editing. It is not new; witness this example from Bleak House: I had thought beforehand that I knew its purport, and I did. It asked me would I be the mistress of Bleak House. (1852–53 Dickens, Bleak House, Part 14, chapter 44). The author’s private database contains written and spoken examples from final examination scripts, dissertations, newspapers (British and New Zealand), web sites, university meetings, and business meetings. The construction functions as the complement of verbs such as ask, find out (6a, b), nouns such as issue (7), adjectives such as

85

86

I Periods sure (8), and prepositions (9). Miller and Weinert (1998: 83) report that a sample of 14 Scottish English conversations contained 3 instances of the “classic” construction as in asked if they were going to France and 22 instances of the alternative construction. (6)

a. You have to ask why is it necessary to raise this very delicate and difficult subject [..] (2005 The Herald, Scotland, Feb. 17 [written]). b. Log on at the BBC World Service Aids site to find out how much do you know about condoms. (BBC webpage)

(7)

this issue about how are we preparing students to flow on seems to me quite important (Associate Deans Meeting, University of Auckland, [unplanned]).

(8)

No one is sure how long are the passages leading off from this centre. (1988 Taylor [travel article], Scotland on Sunday, Nov. 13)

(9)

The question centres on where did this come from? (Final Honours examination script, University of Edinburgh, June 2002)

5.3

NP-clause

(left dislocation)

Quirk et al. (1985: 1416–1417) assign the example in (10) to loose, informal speech. (10) This man I was telling you about – well, he used to live next door to me. They analyze This man I was telling you about as setting the “point of departure” for the utterance as a whole and as enabling speakers to avoid the tricky processing of clauses containing complex phrases. Speakers do avoid complex subject NPs in spontaneous speech, but many left-dislocated NPs are simple. Consider (11a–c): (11) a. this film it does give a real close-up of what goes on behind the scenes. (2007 Brook, “Talking movies”, BBC World, Apr. 19). b. “I like his [Keith Floyd’s] style of cookery,” adds Fenton’s wife Patricia. “He just throws everything in. Rick Stein – he’s only copying Floyd, isn’t he?” (2007 The Independent, Extra, Oct. 11, 2–3) c. “My youngest daughter gets embarrassed when she sees me on television,” says Stewart. “My eldest, she doesn’t mind so much […]” (2007 “Preparing to rock your world”, The Herald, Scotland, Nov. 13, p. 17 [Dr Iain Stewart being interviewed by Susan Swarbrick]) The construction can simply signal a contrast, as in (11b, c), or simultaneously ease the production of complex utterances and signal a contrast, as in (12)–(13). (12) “What struck me was that people who behaved the way my ex and I did, their children were fine, but those who made more mistakes, their children suffered more.” (2007 “Divorce doesn’t have to be a disaster”, The Herald, Scotland, Dec. 3, p. 15)

6 Periods: Contemporary English (13) “You know, it’s an amazing building. The one that was never built, that would have been even more amazing. It was going to be over 550 feet in height, an unbelievable sight”. (2007 The Tablet, Dec. 22–29, pp. 12–13 [Sir Terry Leahy, interview by Chris Blackhurst]) Three important points arise. The construction is not new; (14) is from Bleak House. Mr Jarndyce is not a vocative but is in apposition to Your cousin. (14) “Your cousin, Mr Jarndyce. I owe so much to him. Would you mind describing him to me?” (1852–1853 Dickens, Bleak House Chapter 4) The construction has a contrastive discourse function: Rick Stein, in contrast with Keith Floyd, the one that was never built in contrast with the one that was, the eldest daughter in contrast with the youngest. Finally, it raises interesting questions about the typological difference between subject-prominent and topic-prominent languages and the typology of spontaneous spoken English in contrast with the typology of formal written English (see Miller and Weinert 1998: 363–366).

5.4 Direct object

NP

+ complement clause

The King James Bible translates Matthew VI.28 as “Consider the lilies of the field how they grow”. Imitating the Greek (Miller and Weinert 1998: 362), the construction has the merit of making central the lilies of the field, as the referent of the direct object in a simple clause, and stating their relevant property in a separate complement clause. The constituent structure is not clear, though the lilies of the field and How they grow both modify consider. Other translators have been unhappy with the construction: the Revised English Bible has “Consider how the lilies of the field grow”, with the lilies of the field as subject of the complement clause and removed from the center of attention. The construction has been noticeable in spontaneous spoken English for some time; (15) was uttered in 1978 by a seventeen-year old Scottish male in Edinburgh, with no pause between religion and the damage … (15) i was brought up a catholic and i hate religion the damage it does to human people […] (Miller-Brown corpus of Scottish conversations, mbc2–m87, University of Edinburgh) Here the direct object of hate is religion and the complement clause conveying the relevant property of religion is the damage it does to human people. Example (16) is from the New Zealand component of ICE, again with no pause. (16) i can never remember any of my family how old they are The direct object of remember is any of my family and the relevant property is conveyed by how old they are.

87

88

I Periods Example (17) was uttered in November 2008 by the prosecuting counsel in a court case in Britain (The transcription of the words appeared on the TV screen and were spoken by actors). (17) Did you threaten Michael X at any time that you would have him killed? The direct object of threaten is Michael X and the clause that you would have him killed conveys, not a property of Michael X, but the content of the putative threat. A written version might be Did you threaten to have Michael X killed…? (The OED lists a superficially similar construction from the Wycliffe Bible of 1380: And he threatenyde hem, that thei schulden not seie to any man of him. However the clause that thei… is a clause of purpose. The people were threatened so that they would not say anything.) This sub-section closes with an example of the construction from writing (18). (18) everyone turned to [Stoichev] with a smile […] I remembered Rossi, how he’d listened so modestly to the cheers and speeches (2005 Kostova, The Historian)

5.5 Tense and aspect Two oppositions constitute the core of the English tense-aspect system, past vs. present and simple vs. progressive. These oppositions yield, e.g. writes vs. wrote, is writing vs. was writing, writes vs. is writing, and wrote vs. was writing. Early English originally had a simple present and a simple past (Elsness 1997). Appearing in the 14th century, the progressive came into regular use in subordinate clauses by the 18th century and by the late 20th century, in British English, had become very frequent in main clauses, especially in speech. Smitterberg (2000) demonstrates that the progressive became more frequent in the genres of Letters, Drama, Fiction, and History but less frequent in the genre of Science. The databases of Mair and Leech (2006: 323) show the progressive becoming generally more frequent but also spreading into new parts of the verb system such as the passive. Collins and Peters (2008: 346) observe the same phenomenon in AustrE and NZE. Here we focus on main clauses and the spread of the progressive to all lexical or situation aspects. Imperfectives express single actions, states, and habitual events. Examples of the progressive collected by the author from written texts in Britain and New Zealand suggest that the English progressive is acquiring imperfective traits. It occurs with stative verbs, as in (19), and is used for habitual events, as in (20). The stative progressives are familiar to the author, an older speaker, but the habitual progressives are either peculiar or unacceptable. Examples (19a–b) offer two stative verbs, understand and see, in the progressive, (19c) is both stative and generic, and (19d) is generic. (19) a. I am sorry to have to worry you again with […] X’s resubmission. However Department Y is still not really understanding what it is that X needs to do. (University of Auckland, e-mail from a committee chair) b. “And there is an older generation who are seeing NCEA as lowering the standards […]” (2007 New Zealand Listener, June 9–15, p. 23)

6 Periods: Contemporary English c. She lives in a house which is dating back 200 years. (2007 [photography program] BBC, June 17). d. […] it may be that internal linguistic factors […] are governing the choice between have to and have got to […] (Tagliamonte 2004: 43) These examples are significant; (19a, b, d) were written by contributors who care about their grammar; the e-mail was formal; the New Zealand Listener is a heavyweight periodical; academic texts are scrutinized by referees. Performance errors are unlikely. Examples (20a–d) occurred in final degree examination scripts. In examinations students have little time for planning and editing and produce constructions typical of speech or informal writing but unusual in formal writing. Examples (20a, b) are generic but have stative verbs in the progressive – precede, understand, and depend. Example (20c) has a progressive in a clause denoting a repeated event: the students repeatedly forget the new numbers. (20) a. The first vowel in [complaints] is short as it is preceeding [sic] the nasal bilabial /m/. (Final degree examination script, University of Lancaster, June 2002) b. Naturally a child is depending on his parents, or other adults to provide an environment were he can learn new words. (Final Honours examination script, University of Edinburgh, June 1983) c. The code is often changed and students are forgetting the new number (Minutes of Staff-Honours Students Liaison Committee Meeting, University of Edinburgh, Feb. 1998 [written by a 4th year student]) Examples (20a, b) may reflect a choice of perspective. In examinations students discuss examples given in the question paper and write down their analysis as it proceeds. They may use the progressive to metaphorically put their readers in the middle of on-going events. If correct, this explanation does not contradict the comments on the increasing frequency of the progressive but provides one of its causes. The simple present in main clauses is already isolated as a marker of certain texttypes such as sports commentaries and stage directions and of temporal and conditional subordinate clauses: When we go to London, we avoid the Underground; As soon as she arrives, we will have lunch; If you see them, please pass on my best wishes. The simple past, however, is in regular and frequent use, since speakers can use it or the present perfect to refer to past events. The simple past focuses on past situations while the perfect focuses on their results (Michaelis 1994). Miller (2000, 2004a, 2004b) argues that there are different sets of speakers with different usages of the perfect and simple past. Elsness (1997) demonstrates that the simple past with just for recent events and ever for experiential meaning (Were you ever in Paris?) is the majority use in American English, but this is also the regular pattern in the spontaneous speech of many, perhaps most, speakers in Britain. To complicate matters, there is evidence that speakers who use the perfect are beginning to combine it with specific past-time adverbs, a usage not recognized by reference grammars of standard English but mentioned in Comrie (1985: 33). (Bauer [1989] notes the same development in NZE.

89

90

I Periods He observes [p.c.] that police in New Zealand and Australia regularly use perfects in their reports to the media, even with definite time adverbs, as in an accident has occurred last night.) Elsness (1997: 250) supplies examples from Shakespeare, Pepys, and Galsworthy. As so often, an apparently new usage is an old usage spreading.

5.6 (Apparently) recent constructions 5.6.1 Latinate prefixes and Germanic particles The past thirty years have seen a growing tendency in speech and writing for Germanic particles to be combined with both prefixed and prefix-less Latin verbs, such as extend out, project out, restore back, reintroduce back, reduce down, circulate round, amplify up. In (21) reduce combines with down. Re- is a prefix, in Latin, but for most speakers of English reduce has no prefix; the particle down signals the essential “down” component of the meaning. (A noticeboard in the Botanic Gardens in Edinburgh, an eminently academic context, talks of plants being restored back into their original habitat.) (21) Deglaze the pan with the wine and reduce down to approximately 2 tablespoons. (2002 Christensen-Yule and McRae, The New Zealand Chef, p. 29)

5.6.2 Modal verbs Brown and Miller (1982) report that in their Scottish English conversations must was very infrequent and had only epistemic meaning. Deontic obligation was expressed by have to and have got to, the latter used when the obligation was imposed by someone in authority or by circumstances. Have to was neutral in that respect. May was missing altogether, whether for permission or epistemic possibility. Shall and ought were missing. Tagliamonte (2004), analyzing data collected in York (Britain) in the late 90s, reports that only 15% of the deontic forms are of must, and that must is hardly used by speakers under 30 and mostly by speakers over 70. This supports Brown and Miller’s finding, but with respect to have (got) to. Tagliamonte says that grammaticality judgments confirm that there is little to choose between I have to go shopping and I’ve got to go shopping. Grammaticality judgments are not the strongest of tests; Brown and Miller used other elicitation techniques and it would be interesting to see the results of any similar investigation in York. One major change affecting all varieties of English in the past thirty years is the use of may instead of might for the expression of remote possibility. (As a result, may has reappeared in Scottish English – see the first paragraph of Section 5.6.2.) Example (22) is spoken NZE, cited in the New Zealand Herald, and (23) is from a book by one of the UK’s leading historians, Norman Davies. (22) A St John medical adviser acknowledged that Mr Boonen may have lived had he not waited for the ambulance. (2007 New Zealand Herald, Feb. 28) (23) a. The witness said that days later he heard media reports that someone had gone missing. He believed what he had seen that night may be connected and he and his wife met with police.

6 Periods: Contemporary English The trial, before Judge Lord Matthews, continues. (2009 The Herald, Scotland June 23, p. 3) b. If the style had matched the content, it may have had more success in crossing the sectarian divide. (1999 Davies, The Isles: A History, pp. 512–513) Trask (2001:183) declares that Standard English absolutely requires might but the change is well-established in speech and writing.

5.6.3 Argument structure The argument structure of transfer verbs is susceptible to change, that is, verbs denoting the movement of something to someone and involving a person transferring, the entity transferred (typically inanimate), and the recipient. The recipient is typically human and speakers tend to place other humans at the center of events. The changes in argument structure all involve a change whereby recipient NPs are not (possibly optional) oblique objects but obligatory direct objects central to the clause syntax. The syntactic centrality parallels the semantic centrality of recipients. Verbs such as donate, attribute, and forward take the construction V NP1 to NP2, as in donated the treasure to the museum, attributed the painting to Raphael and forwarded the letter to Susan. Current dictionaries show these verbs excluding the construction V NP2 NP1, as in donated the museum the treasure. Over the past twenty years, however, each set of students taking the author’s courses has contained two or three individuals for whom donated the museum the treasure, attributed Raphael the painting and forwarded Susan the letter are perfectly acceptable. Verbs such as issue, confer, bequeath, parachute, circulate, confer, grant, and inflict typically take the construction V NP1 P NP2, where P is to or on, as in issue/bequeath something to someone and inflict something on someone. Data from written English, such as they issued him with a deadline to resign, indicate that these verbs are being used in the construction V NP2 with NP1, as in (24)–(26). (24) PM Thaksin Shinawatra is facing a showdown with critics after they issued him with a deadline to resign […] (2006 New Zealand Herald, Mar. 24) (25) It is odd that the whiskered traffic managers [of the railway companies] have bequeathed Scotland with her most pronounced token of nationhood […] (1993 Clarke, “A sporting chance for Scottish Tories” The Scotsman, Aug. 13) (26) – if it’s a severe winter we’ll have to parachute the island with food supplies like Bosnia (1994 Scotland on Sunday, Apr. 24) A passive example such as they will be conferred with honorary degrees (The Herald, Scotland) suggests the active construction confer someone with a degree. It is possible that the passive construction appeared first and gave rise to the active. Denison (1998: 215) cites a passive example from 1917 I’m issued with her [a horse]. For many speakers the pair rob and steal are clearly distinguished; you steal valuables by taking them away illegally, but you rob the owner of the valuables (Trask 2001: 250). Rob is now being used by many speakers for the action applied to the

91

92

I Periods valuables, as in the sentence They’ll probably end up going to jail or something or probably robbing stuff (BA dissertation, University of Lancaster, 2002). And of course Trask’s entry on rob and steal indicates that the usage is frequent enough to catch the attention of conservative speakers. In fact, this use of rob is an old usage that has survived in speech and in non-standard English. In Johnson’s Dictionary the third definition for rob is “to take away unlawfully”. Examples are fashion a carriage to rob love from any (Shakespeare) and Double sacrilege […] to rob the relick and deface the shrine (Dryden). Prevent occurs in three constructions: We prevented Susan from marrying a monster; We prevented Susan marrying a monster; We prevented Susan’s marrying a monster. The last construction is now archaic. Stop, which is similar in meaning to prevent, also takes the first two constructions, as in We stopped Susan from marrying a monster and We stopped Susan marrying a monster. Both verbs have to do with situations in which somebody is (metaphorically) kept away from or moved away from an action. Burchfield (1996: 622) notes the three prevent constructions, as does Trask (2001: 228) but neither mentions that the construction is attracting other verbs, such as thwart in(27) and intimidate in (28). (27) Her parents, financially thwarted from education themselves, were “adamant that we would succeed […]” (2007 New Zealand Listener, June 9–15, p. 27) (28) Witnesses are being intimidated from coming forward. (2007 New Zealand Herald, Feb.) The usual constructions are, e.g. She was thwarted in her attempt to defraud the bank and The hooligans intimidated them into handing over their cash. However, from is semantically motivated; someone thwarted or intimidated into not doing something is metaphorically removed from that activity.

5.7 Syntax and the organization of text We conclude with changes affecting conjunctions and the arrangement of clauses into sentences. Quirk et al. (1985: 667–668) classify plus as a marginal preposition, but add that it “can even be used as a conjunction”, as in You can get what you want, plus you can save money. A New Zealand example is For centuries people used liquorice […] It has anti-inflammatory, anti-bacterial, expectorant and laxative properties plus it tastes good (2007 Weekend Herald, Canvas, Sept. 9). In the above examples plus is clause-initial. Example (29) contains a sentence-initial example, with plus separated from the rest of the sentence by a comma and being treated as a sentence adverb. (29) […] it avoids the ‘duplication problem’ […] Plus, the lack of ordered rules means that OT analyses are not burdened with various intermediate levels […] (Ph.D. thesis, University of Edinburgh, 2002) Although is a subordinating conjunction introducing adverbial clauses of concession. Also is an adverb which typically occurs to the left of the main verb as in She also writes

6 Periods: Contemporary English poetry, I am also writing a book. Both have begun to function as sentence adverbs, as in (30). (30) Also, since I did not have access to the original photocopies […], I was unable to establish whether certain transformations were simply caused by transcription mistakes. (BA dissertation, University of Lancaster, June 2003) The Australian English component of ICE yielded eight examples in unscripted speech, as in (31) and (32). (31) Mmm That’s so true That’s so true Also you spent enough money on drinks on Friday night I think (S1A-094(B):333) (32) Although, English has been the most successful language in an attempt of [sic JM] becoming a lingua franca. (Final Honours examination script, University of Edinburgh, June 2001) We conclude this section with a change and a non-change. Quotative like is relatively new, at least in British English. It is absent from the corpus of Scottish English collected by the author and Keith Brown in 1977–1978 (Brown and Miller 1982) but abundant in the corpus of Scottish English collected by Jane Stuart-Smith et al. (1996) and in the corpus of English English that appears in Carter and McCarthy (1997). An excellent account of previous work and analysis of original data is offered by Buchstaller (2004). The non-change is the use of the discourse marker like. Contrary to Anderwald (2008: 459–460), it was not imported to Britain from the United States and is not recent. Utterance-final like occurs in dialogue in works by James Hogg and Sir Walter Scott from the early 19th century and utterance-medial like occurs in a mid-19th century example in Grant and Main-Dixon’s Grammar of Scots. D’Arcy (2007: 401) reports the occurrence of utterance-initial and -medial like in recordings of New Zealand speakers made in 1946–1948. Their parents were from various parts of the UK, not from the USA. Although not a new usage, the discourse-marker like deserves to be taken seriously in view of its longevity, its widespread use and the lack of detailed knowledge about its origin (see Miller 2009).

6 Summary The phonological changes described in Section 2 pertain to issues in theories of linguistic change such as the interplay between social networks, social and national identities, and properties of language systems. Space does not permit a discussion of these issues (but see Labov 1994, 2001). The lexical changes mentioned in Section 4 exemplify wellknown phenomena: creation of new vocabulary for new phenomena, (metaphorical) extension of meaning as in troll and grey water, and so on. Just as importantly and interestingly, in some varieties of English, changes in vocabulary signal the assertion or re-assertion of cultural independence. In contrast, the syntactic changes covered in Section 5 raise familiar and unfamiliar issues. Grammaticalization is exemplified by plus, which has acquired the more grammatical function of conjunction, and by both the quotative and the discourse marker like. The relation between rob and steal exemplifes the

93

94

I Periods phenomenon of persistence, whereby old structures persist in use alongside new structures to which they have given rise. The revival of the middle raises questions about transitivity in English and about basic and non-basic constructions. Context is important, of course. Mair and Leech (2006) comment that whom is far from dead; it thrives in formal texts, especially written. Shall, which few speakers of Scottish English use in speech or writing, thrives in formal notices such as This shop shall be open on Thursday evening till 9 pm (the notice was not a legal document issued, say, by the City Council but a notice produced by the shop in question). Constructions that raise a number of questions are the indirect question (Section 5.2), the NP-clause (Section 5.3), and the direct object + complement clause (Section 5.4). The constituent structure of the last construction is not clear. Dealing with the NPclause construction, one analysis puts the NP outside the clause (and has no sentences), while another one has the NP and clause as constituents of a single sentence. The constructions are unusual, too, in not fitting the theory that changes happen from below, or filter into standard English from non-standard varieties. Speakers (and writers) of standard and non-standard English alike use the constructions. The constructions raise questions about the typology of (spontaneous) spoken English, as discussed in Miller and Weinert (1998: 363–366), and about social capital and norms – who decides them and attitudes to them in informal everyday practice, in schools, and in employment. Acknowledgments: Many thanks to Laurie Bauer for his comments on an earlier draft of Section 2 in particular but also the whole chapter.

7 References Aarts, Bas and April McMahon (eds.). 2006. The Handbook of English Linguistics. Oxford: Blackwell. Adger, David. 2003. Core Syntax: A Minimalist Approach. Oxford: Oxford University Press. Altendorf, Ulrike and Dominic Watt. 2008. The dialects in the south of England: Phonology. In: Kortmann and Upton (eds.), 194–222. Anderwald, Lieselotte. 2008. English in the Southeast of England: Morphology and syntax. In: Kortmann and Upton (eds.), 440–462. Bauer, Laurie. 1989. The verb have in New Zealand English. English World-Wide 10(1): 69–83. Bauer, Laurie. 1994. Watching English Change. London/New York: Longman. Bauer, Laurie and Paul Warren. 2008. New Zealand English: Phonology. In: Burridge and Kortmann (eds.), 39–76. Bolinger, Dwight. 1968. Aspects of Language. New York: Harcourt, Brace, and World. Brown, Keith and Jim Miller. 1982. Aspects of Scottish English syntax. English World-Wide 3(1): 1–17. Buchstaller, Isabelle. 2004. The Sociolinguistic Constraints on the Quotative System: US English and British English Compared. Unpublished Ph.D. thesis, University of Edinburgh. Burchfield, Robert. 1981. The Spoken Word: A BBC Guide. London: BBC Publications. Burchfield, Robert. 1996. The New Fowler’s Modern English Usage. 3rd edn. Oxford: Clarendon Press. Burridge, Kate and Bernd Kortmann (eds.). 2008. Varieties of English 3: The Pacific and Australasia. Berlin/New York: Mouton de Gruyter. Carter, Ronald and Michael McCarthy. 1997. Exploring Spoken English. Cambridge: Cambridge University Press.

6 Periods: Contemporary English Cheshire, Jenny, Viv Edwards, and Pamela Whittle. 1993. Non-standard English and dialect levelling. In: James Milroy and Lesley Milroy (eds.), Real English, 52–96. London: Longman. Collins, Peter and Pam Peters. 2008. Australian English: Morphology and syntax. In: Burridge and Kortmann (eds.), 341–361. Comrie, Bernard. 1985. Tense. Cambridge: Cambridge University Press. D’Arcy, Alexandra. 2007. Like and language ideology: Disentangling fact from fiction. American Speech 82(4): 386–419. Denison, David. 1998. Syntax. In: Suzanne Romaine (ed.), The Cambridge History of the English Language. Vol. IV. 1776–1997, 92–329. Cambridge: Cambridge University Press. Denison, David. 2008. Patterns and productivity. In Susan M. Fitzmaurice and Donka Minkova (eds.), Empirical and Analytical Advances in the Study of English Language Change, 207– 230. Berlin/New York: Mouton de Gruyter. Elsness, Johan. 1997. The Perfect and the Preterit in Contemporary and Earlier English. Berlin/ New York: Mouton de Gruyter. Fabricius, Anne. 2002. On-going change in modern RP: Evidence for the disappearing stigma of t-glottalling. English World-Wide 23(1): 115–136. Foulkes, Paul and Gerry Docherty (eds.). 1999. Urban Voices: Accent Studies in the British Isles. London: Arnold. Gill, Brian, Juliet Hawkins, and Les McPherson. 1994. New Zealand Songbirds, Bird 8. Auckland: Godwit Press Limited. Good, Melissa (ed.). 2008. Cambridge Advanced Learner’s Dictionary. 3rd edn. Cambridge: Cambridge University Press. Gordon, Matthew J. 2008. New York, Philadelphia, and other northern cities: Phonology. In: Edgar W. Schneider (ed.), Varieties of English 2: The Americas and the Caribbean, 67–86. Berlin/New York: Mouton de Gruyter. Hardie, Kim. 1996. Lowland Scots: Issues in nationality and identity. In: Charlotte Hoffman (ed.), Language, Culture and Communication in Contemporary Europe, 61–74. Clevedon: Multilingual Matters. Henry, Alison. 1995. Belfast English and Standard English: Dialect Variation and Parameter Setting. Oxford: Oxford University Press. Horvath, Barbara M. 2008. Australian English: Phonology. In: Burridge and Kortmann (eds.), 89–110. Hundt, Marianne. 2007. English Mediopassive Constructions. A Cognitive, Corpus-based Study of their Origin, Spread, and Current Status. Amsterdam/New York: Rodopi. Jespersen, Otto. 1961. A Modern English Grammar on Historical Principles. Part VI. London/ Copenhagen: George Allen and Unwin and Ejnar Munksgaard. Kemmer, Suzanne. 1993. The Middle Voice. Amsterdam/Philadelphia: John Benjamins. Kortmann, Bernd and Edgar Schneider, in collaboration with Clive Upton, Kate Burridge, and Rajend Mesthrie. 2004–08. Varieties of English. 4 vols + 1 CD ROM. Berlin/New York: Mouton de Gruyter. Kortmann, Bernd and Clive Upton (eds.). 2008. Varieties of English 1: The British Isles. Berlin/ New York: Mouton de Gruyter. Labov, William. 1994. Principles of Linguistic Change. Vol. 1. Internal Factors. Oxford/Malden, MA: Blackwell. Labov, William. 2001. Principles of Linguistic Change. Vol. 2. Social Factors. Oxford/Malden, MA: Blackwell. Macafee, Caroline I. 1994. Traditional Dialect in the Modern World: A Glasgow Case Study. Frankfurt: Peter Lang. Macalister, John. 2006. The Maori presence in the New Zealand English lexicon, 1850-2000. English World-Wide 27(1): 1–24. Mair, Christian and Geoffrey Leech. 2006. Current changes in English syntax. In: Aarts and McMahon (eds.), 318–342.

95

96

I Periods Michaelis, Laura A. 1994. The ambiguity of the English present perfect. Journal of Linguistics 30: 111–157. Miller, Jim. 2000. The perfect in spoken and written English. Transactions of the Philological Society 98(2): 323–352. Miller, Jim. 2004a. Problems for typology: Perfects and resultatives in spoken and non-standard English and Russian. In: Bernd Kortmann (ed.), Dialectology Meets Typology: Dialect Grammar from a Cross-Linguistic Perspective, 305–334. Berlin/New York: Mouton de Gruyter. Miller, Jim. 2004b. Perfect and resultative constructions in spoken and non-standard English. In Olga Fischer, Muriel Norde, and Harry Perridon (eds.), Up and Down the Cline: The Nature of Grammaticalization, 229–246. Amsterdam/Philadelphia: John Benjamins. Miller, Jim. 2006. Spoken and written English. In: Aarts and McMahon (eds.), 670–691. Miller, Jim. 2009. Like and other discourse markers. In: Pam Peters, Peter Collins, and Adam Smith (eds.), Comparative Studies in Australian and New Zealand English, 317–338. Amsterdam/Philadelphia: John Benjamins. Miller, Jim and Regina Weinert. 1998. Spontaneous Spoken Language: Syntax and Discourse. Oxford: Clarendon Press. Minkova, Donka and Robert Stockwell. 2006. English words. In: Aarts and McMahon (eds.), 461–482. Quirk, Randolph, Sidney Greenbaum, Geoffrey Leech, and Jan Svartvik. 1985. A Comprehensive Grammar of the English Language. London: Longman. Schneider, Edgar W. 2003. The dynamics of new Englishes: From identity construction to dialect birth. Language 79(2): 233–281. Simpson, J. A. and E. S. C. Weiner (eds.). 1989. Oxford English Dictionary. 2nd edn. Oxford: Oxford University Press. Smitterberg, Erik. 2000. The progressive form and genre variation during the nineteenth century. In: Ricardo Bermu´dez-Otero, David Denison, Richard Hogg, and C. B. McCully (eds.), Generative Theory and Corpus Studies. A Dialogue from 10 ICEHL, 283–297. Berlin/New York: Mouton de Gruyter. Stuart-Smith, Jane, Claire Timmins, and Fiona Tweedie. 2006. Conservation and innovation in a traditional dialect. L-vocalization in Glaswegian. English World-Wide 27(1): 71–87. Tagliamonte, Sali. 2004. Have to, gotta, must: Grammaticalization, variation and specialization in English deontic modality. In: Hans Lindquist and Christian Mair (eds.), Corpus Approaches to Grammaticalization in English, 33–55. Amsterdam/Philadelphia: John Benjamins. Trask, R. Larry. 2001. Mind the Gaffe: The Penguin Guide to Common Errors in English. London: Penguin. Upton, Clive. 2008. Received Pronunciation. In: Kortmann and Upton (eds.), 237–252.

Jim Miller, Edinburgh (UK)

II Linguistic Levels 7 Linguistic Levels: Phonology 1 Introduction to some basic terms and developments in phonological theory 2 Stability and instability in consonant inventories 3 The future of looking back in time on sound changes 4 Summary 5 References

Abstract Changes in the sound system of a language may involve many different aspects. First, phonemes may be added to an inventory, they may become obsolete, or they may change their shape. Second, allophonic rules may emerge (e.g. “voice intersonorant fricatives”), disappear, or change. Third, phonotactic restrictions may be added (e.g. no syllable-initial /kn/ sequences, so that such clusters are reduced to one sound), or change their effect (e.g. the ban on /kn/ sequences may be resolved by epenthesis). Fourth, prosodic structure may change (resulting in, for instance, stress shift) and, fifth, morphophonological alternations (e.g. ablaut and umlaut) may start to play a different role or they may vanish (e.g. when morphology is regularized). This chapter will first briefly introduce the field of phonology as envisaged by structuralists and generative linguists. Section 2 discusses some changes in the consonant inventory, the allophonic variations, and the phonotactic restrictions in the history of the English language and shows that many of these changes are interrelated in the sense that a change in one component triggers an effect in another component. Section 3 expresses some ideas on the future of looking back on changes in the sound system, and Section 4 summarizes the chapter.

1 Introduction to some basic terms and developments in phonological theory The ability of speakers to distinguish separate segments in a string of speech sounds is an important part of the knowledge that language users have about their native language. Speakers of English realize that the words pin, thin, bin, and fin differ only in the first sound and they may use this knowledge in rhyme, alliteration and in games with nonsense words (e.g. “Annie the pannie the thanny”, “Simon the bimon the fimon”). The sounds represented by p, th, b, and f have a function in English, i.e. these speech sounds – together with approximately 35 others – are the minimal units that can distinguish meaning in English. The smallest elements that can cause a change in meaning are called “distinctive sounds” or “phonemes”. Jones (1967: vi) mentions that the French word “phone`me” appears to have been invented by the Frenchman L. Havet, who used it in 1876 to mean “speech sound”. At the beginning of the 20th century the term “phoneme” acquired a more abstract meaning and referred to the minimal unit in speech that can function to distinguish meaning. Alex Bergs and Laurel Brinton 2012, Historical Linguistics of English (HSK 34.1), de Gruyter, 97–113

98

II Linguistic Levels A theory of the phoneme began to be developed at the turn of the 20th century – especially in the works of Baudouin de Courtenay (1895) and de Saussure (1916) – and became one of the important research interests of structural linguistics in the first half of the 20th century. In a statement submitted to the First International Congress of Linguistics meeting in The Hague in 1928, Roman Jakobson, Sergej Karcevskij, Nikolaj Sergeevicˇ Trubetzkoy, and other members of the Prague School emphasized that a scientific description of a language must include a characterization of its phonological system, i.e. the repertory, pertinent to that language, of the distinctive contrasts among its speech sounds (see Jakobson 1971 [1962]). Thus, an important goal of structural phonology in Europe – as well as in North America – was to establish the phoneme inventories of languages. The method used to establish the phoneme inventory of a language is to systematically compare words with different meanings that differ in one sound only. Such word pairs are called “minimal pairs”. Furthermore, the international congress held that the field of language change should not be confined to studying isolated changes; rather, changes should be considered in terms of the linguistic system which undergoes them. As Waugh (1976: 21) puts it, according to the Prague School, “we must understand the structure before the change begins, the structure after it takes place, the sense of the change undergone in respect to the undergoing system, the level at which the change takes place (e.g. distinctive feature or phoneme) and the effects of the change on the system”. One of the main concerns of Roman Jakobson became to develop a theory of phonology that would predict exactly those distinctive sounds that can be found in the world’s languages, and he hypothesized that there is a limited number of phonological (or “distinctive”) features – approximately 15 – that characterize the sounds of human languages (see, e.g., Jakobson 1939; Jakobson et al. 1951; Jakobson and Halle 1956). In the system that Jakobson and his colleagues developed, each phoneme is represented by a set of features such as [grave]/[acute] and [flat]/[nonflat] which are unrelated (i.e. not grouped into smaller sets) and binary. For Jakobson, speech sounds are characterized by features and for him it followed that sound changes should involve features or phonemes too: a sound change is a change in the distribution of a phonological feature or phoneme within a system. With Sapir (1921) before him, Jakobson posited that if in a system /p, t, k/, one segment has changed (e.g. /p/ to /b/), the outcome is asymmetric (/b, t, k/). The simplest way to restore the symmetry is an analogous change of the other members in the system (e.g. /t, k/ → /d, ɡ/). Should the resulting pattern already exist, then the system of oppositions can only survive if the older series (in our example /b, d, g/) itself undergoes a change (e.g. spirantization). In this hypothetical example, the stops specified as being voiceless gradually changed into voiced stops – i.e. the feature that expresses voicing in stops changed – and the older voiced stops became fricatives – i.e. for voiced stops the feature that specified complete obstruction in the vocal tract changed into a feature that specifies incomplete obstruction ([+continuant]). In this case, the phonological sound change alters the relationship between two elements from /p/ versus /b/ to /b/ versus /v/. Jakobson also allowed for sound changes that lead to the elimination of a phonological contrast or to the formation of a contrast and we will encounter English examples of modification, elimination, and creation of phonological contrasts in the remainder of this chapter. Speech sounds may be articulated differently depending on the position in the word. For example, the initial sound in the word pin (and the non-word pannie) is pronounced

7 Linguistic Levels: Phonology with a puff of air (called “aspiration”), which is lacking in the word spin. If we were to replace the sound p in pin and pannie by the corresponding p-sound in spin or spaniel, or by the unaspirated p-sound used in Dutch pin, the words would sound odd to native speakers of English, but pin would still mean ‘a short thin piece of metal used for fastening things together’. Aspiration thus does not change the meaning of the words in question, but it is part of the grammar that speakers of Present-day English employ. The phonological rule that applies in this case would be “in word-initial position, add the feature [spread glottis] (for aspiration) to segments specified as [stiff vocal folds] underlyingly”. Speech sounds that do not cause a change in meaning, but are realizations of one sound in different contexts (e.g. aspirated [ph] in absolute word-initial position versus unaspirated [p] preceded by /s/) are called “allophones” (a term that was invented about 1934 by Benjamin Lee Whorf). To explain sound alternations in particular contexts, it is useful to distinguish the underlying representation from the actual realization (or “surface representation”): the former is put between slashes (“/ /”) and the latter is put between square brackets (“[ ]”). For example, the underlying form of the English regular past tense marker spelled -ed is /d/ (as in fail[d], love[d]). Underlying /d/ is realized with an epenthetic vowel after a stem that ends in an alveolar stop (need[ɪd], want[ɪd]) and it is realized without voicing – i.e. as [t] – after a stem-final voiceless obstruent (as in kiss[t], walk[t]). In generative frameworks, the surface form (or “output”) is derived from the underlying form (or “input”) by phonological rules and in Optimality Theory (OT), the surface form that violates the least highly ranked constraints compared to alternative surface forms is selected as the “optimal” one for phonetic realization. In generative phonology, changes in allophonic variation may thus involve a change in rule applications, whereas in OT, allophonic changes involve a reranking of universal constraints in the language specific grammar. Apart from the knowledge about which sounds are part of the sound system of a language and the knowledge about the realization of sounds in particular contexts, speakers also have clear intuitions about how sounds are organized to form words, i.e. how some sounds may combine with other sounds in the language. Speakers of English know that /p/ may combine with /s/ – as in spin – whereas the th-sound may not (*sthin). Changes in the phonotactics of the English language have occurred throughout its history and we will consider a few of them in this chapter. Some phonological phenomena have scope over larger domains than one segment, e.g. a syllable, or take effect at a morphosyntactic boundary. An example of the first kind of phenomenon is word stress (see Minkova, Chapter 8, which discusses the development of word and phrasal stress from Old to Present-day English). As an example of a phonological phenomenon that applies across a morphosyntactic boundary, English consonant-intrusion can be mentioned here. To avoid two vowels becoming adjacent, some varieties of English have the option of inserting a sonorant consonant. Which particular sound is inserted depends on the preceding vowel: the glide /j/ is inserted after a front high vowel (e.g. I see [j] it), /w/ is inserted after a back high vowel (e.g. too [w] old) and an r-sound is inserted when a non-high vowel-final word is followed by a suffix or a word that is vowel-initial (e.g. I saw [r] it), for instance in the Eastern Massachusetts dialect (McCarthy 1993: 170–171). Thus, the epenthetic consonant is a sonorant and the preceding vowel determines whether it is realized as /j/, /w/, or rhotic /r/. The present chapter is mainly concerned with segmental phonology, i.e. the study of the function, behavior, and organization of speech sounds in one language and across

99

100

II Linguistic Levels languages. The three factors mentioned immediately above are part of the knowledge that speakers have about the sound system of their native language (see, e.g., Goldsmith 1995: 1–13; Lass 1984). The problem for historical linguistics is of course that there are no speakers of earlier stages of the language alive. Fortunately, in many cases they left traces of their speech in written records (manuscripts, grammars, etc.) and where written records are absent, it is often possible to reconstruct phonological systems by comparing different languages and language varieties that derived from the language period we are interested in. The present chapter examines some changes in the phoneme inventory, the allophonic variations, and the phonotactic restrictions in the history of the English language. It is not our ambition to investigate all phonological properties; we will concentrate on only a few properties that concern speech sounds and that have changed over time in the history of the English language as spoken in England. The most well-known phonological changes in the history of English are changes in the vowel system. Since these changes are discussed at length in other chapters of this volume (e.g. the Great Vowel Shift in Krug, Chapter 48), we will refrain from repeating the story of vowel changes here and concentrate on changes in the consonant inventory, allophonic variation of consonants, and phonotactic restrictions on consonant clusters. The chapter on prosody (see Minkova, Chapter 8) complements this one and considers phenomena that cannot be restricted to single speech sounds (e.g. word and phrasal stress assignment).

2 Stability and instability in consonant inventories At the turn of the 20th century, the major goal of historical linguistics was to reconstruct phonological systems and rules of language stages of which we have no direct evidence. According to an extreme version of the Neogrammarian doctrine, historical study was the only genuinely scientific approach to the facts of language. By examining sound changes, the Neogrammarians tried to find out why modern languages have developed the way they have and how they relate to other languages. The Neogrammarians maintained that language change is systematic and takes place without exceptions. For example, it was already observed by e.g. Friedrich von Schlegel in 1801 that Latin words that start with a labial plosive (e.g. ped, pisces) correspond to words that start with a labial fricative in Germanic languages (English foot, fish; German Fuss, Fish) and that words beginning in (/t/) or (/k/) in Latin would have (/θ/) or (/x/ or /h/) in early stages of Germanic languages. The Neogrammarians deduced from these observations that the voiceless stops /p, t, k/ in the common ancestor of these languages (i.e. Indo-European) remained unchanged in Latin, but became voiceless fricatives in Germanic. By applying their method of systematically comparing sounds in certain positions of similar words in different languages, the Neogrammarians constructed family trees of related languages. Even though it is still common to refer to language family trees and say that, for instance, English, German, Dutch and Frisian belong to the “branch” of West-Germanic languages and are “sister” languages, few people today would support the idea that the origin of the languages that we now know can be traced back to one common ancestor (see e.g. Aitchison 2001: 23–36, Wunderlich 2008). Nevertheless, the important contribution of Neogrammarians to linguistic research is the insight that some sound changes do not take place arbitrarily, but that classes of sounds in a particular linguistic context may undergo a certain change.

7 Linguistic Levels: Phonology To illustrate how consonant inventories may change over time, consider as an example the inventory of obstruents that the Neogrammarians assumed for Indo-European (1a) and its descendant Proto-Germanic (1b): (1)

a. Indo-European obstruents b. Proto-Germanic obstruents /f, θ, xj, x, xw/ /p, t, kj, k, kw/ /b, d, ɡj, ɡ, ɡw/ /p, t, kj, k, kw/ /b, d, ɡj, ɡ, ɡw/ /bh, dh, ɡhj, ɡh, ɡhw/ /s, h/ /s, h/ and geminate stops after short vowels geminate stops after short vowels

What remained “stable” in the transition from the Indo-European obstruent inventory to the Proto-Germanic one is the number of phonemes: both systems distinguish 17 obstruents. What changed was the set of phonemes. Proto-Germanic has fewer laryngeal contrasts, but more contrasts in manner of articulation for the different places of articulation as a result of a set of consonant shifts – commonly referred to collectively as “Grimm’s Law” – whereby the Indo-European voiceless stops spirantized and became voiceless fricatives, the unaspirated stops underwent a strengthening process and became voiceless aspirated stops and, finally, the so-called “breathy voiced” stops were deaspirated and became voiced unaspirated stops (e.g. Harbert 2007: 41–88; for a phonological account of parts of this consonant shift see, e.g. Iverson and Salmons 1995). This type of sound change is an example of what Jakobson would consider a modification of a phonological contrast: it altered the relationship from voiceless stop to voiced stop and from voiced stop to breathy voiced stop into the opposition voiceless fricative versus voiceless stop and voiceless stop versus voiced stop. We can only speculate as to why the system changed the way it did. The internal change from voiceless stops to voiceless fricatives in one branch of a language family may perhaps be attributed to the fact that voiceless /p, t, k/ are perceptually close to voiced /b, d, ɡ/ whereas the fricatives /f, θ, x/ and the stops /b, d, ɡ/ are more distinct. One hypothesis may thus be that in order to make the two classes more perceptually distinct, one class was “enhanced” by spirantization. However, Proto-Celtic kept the /p, t, k/ versus /b, d, ɡ/ contrast and conflated (or “merged”) /b, d, ɡ/ and /bh, dh, ɡh/, so that two perceptually close classes emerged as in (2a, b): (2)

a. Indo-European obstruents b. Proto-Celtic obstruents /(p), t, (kj), k, kw/ /p, t, kj, k, kw/ j w , ɡ, ɡ / /b, d, (ɡj), ɡ, (ɡw)/ /b, d, ɡ h h hj h hw /b , d , ɡ , ɡ , ɡ / /s, h/ /s, h/ and geminate stops and geminate stops

In this case, what remained “stable” (i.e. the “pertinent” properties) are the different manners of articulation. The unstable or “transient” property is the reduction in the number of phonemes: Proto-Celtic exhibits fewer laryngeal contrasts and – due to depalatalization and delabialization – fewer places of articulation. This type of sound change is an example of what Jakobson would consider elimination of a phonological contrast. The change altered the three-way opposition from voiceless stop to voiced

101

102

II Linguistic Levels stop to breathy voiced stop into a two-way contrast between voiceless stop and voiced stop. The question why Proto-Germanic opted to enhance the contrast by spirantization of one class, whereas Proto-Celtic kept the /p, t, k/ versus /b, d, ɡ/ contrast and conflated /b, d, ɡ/ and /bh, dh, ɡh/ is notoriously difficult to answer. In the framework of Optimality Theory as introduced in the early 1990s in different works by McCarthy, Prince, and Smolensky (McCarthy and Prince 1993a, 1993b, 1995; Prince and Smolensky 1993), the suggestion is made that the desire to satisfy “perceptual distance” to accommodate the needs of the hearer was valued more in Proto-Germanic than the constraint “avoid marked elements” (i.e. fricatives are universally more marked than stops and thus disfavored in phoneme inventories). In contrast, the constraints “faithfulness to manner” and “minimize phoneme inventory” (i.e. the so-called “principle of economy”) gradually gained more weight and thus became more important than “perceptual distance” in Proto-Celtic. Of course this only says what happened in different language communities and not why it happened. As an autonomous reviewer pointed out, we cannot answer the question why “perceptual distance” was more valued in one language family and less in another and why the “principle of economy” came to play a more important role in some language communities, because we do not know the reasons for variation preferences in speech communities.

2.1 Stability and instability in consonant inventories related to phonotactics Proto-Germanic is the ancestor of Old English (OE). In (3) we compare the inventory of obstruents assumed by the Neogrammarians for Proto-Germanic and the inventory of Old English as proposed by Murray (Chapter 17; see also Marsh, Chapter 1). (3)

a. Proto-Germanic obstruents /p, t, kj, k, kw/ /b, d, ɡj, ɡ, ɡw/ /f, θ, xj, x, xw/ /s, h/ and geminate stops after short vowels

b. Old English obstruents /p, t, kj, k/ /b, d, ɡj, ɡ/ /f, θ, xj, x/ /s, ʃ, h/ and geminate consonants (except /h/, /ʃ/ and /r/) after short vowels

The effect of OE consonant gemination in intervocalic position was to create a closed syllable. Geminate consonants were lost before the 13th century. Note first of all that Murray (Chapter 17) does not assume labiovelar stops and fricatives for Old English. The spelling of OE cwe¯n ‘queen, wife of a king’ and cwæþ ‘said, spoke, called, named, proclaimed’ suggests that /kw/ was a legitimate sound or sound sequence. Rather than being analyzed as one phoneme, the spelling suggests that we are dealing with a consonant cluster of a velar stop followed by a labiovelar approximant /w/ (due to the influence of Anglo-Norman scribes, the spelling changed into in the Middle English [ME] period). Murray does assume palatalized velar stops and fricatives in the Old English consonant inventory, but these were most probably fronted realizations of the velar obstruents before front vowels.

7 Linguistic Levels: Phonology We will now focus on the fact that before the change from Proto-Germanic to Old English, there are some asymmetries in the consonant system. The ones we will have a closer look at are: (i) only velar obstruents have a labialized counterpart (e.g. /k/ contrasts with /kw/, but /t/ does not contrast with /tw/) and (ii) only alveolar stops and fricatives are followed by /w/ (cf. OE twa ‘two’, twelf ‘twelve’, twentig ‘twenty’, betwix ‘between’, and sweord ‘sword’; /w/ disappeared in some initial /tw/ and /sw/ clusters preceding a round back vowel in late Middle English and Early Modern English, so that is a “silent” letter in the words two and sword today). If it is true that the labiovelar segments were reanalyzed as consonant clusters of a velar stop followed by a labiovelar approximant at an early stage of Old English, the phoneme inventory changed in that labiovelar obstruents gradually became obsolete and the phonotactics changed as a consequence: the language now allowed alveolar and velar obstruents followed by a labiovelar approximant in the onset of a syllable. Clusters with labial obstruents followed by /w/ did not emerge, presumably due to a ban on identical places of articulation in two adjacent onset consonants, i.e. a so-called “OCPeffect” (Obligatory Contour Principle), which prohibits /pw/, /fw/, /tl/, /dl/, etc. as possible word-initial clusters. The change described above from a complex phoneme to a consonant cluster is one example of how the loss of a phoneme was compensated for by relaxing the phonotactics. The reverse state of affairs is also attested: at least in one case, a phonotactic restriction (“disallow sequences of /s/ immediately followed by /k/”) resulted in the emergence of a new phoneme. Some sequences of alveolar /s/ followed by the velar stop /k/ were at some point no longer pronounced as /sk/ by the Anglo-Saxons, but rather as the alveopalatal sound /ʃ/ (as in the words ship, sheep, shoe and fish). It is often argued that in Old English, the change from /sk/ to /ʃ/ was the result of palatalization (see below). However, palatalization occurred in the context of front vowels, whereas /sk/ clusters seem to have become alveopalatal fricatives in more environments, for instance, word-initially independent of the quality of the following vowel, word-medially (except where the cluster is not tautosyllabic before back vowels, e.g., *aisko¯jan → a¯scian ‘to ask’ where represents /sk/) and word-finally. A similar assimilation process must have applied in the German language as well, since words of the same origin – so-called “cognates” – are also pronounced with an alveopalatal fricative in this language (as in the Ger. words Schiff ‘ship’, Schaff ‘sheep’, Schuh ‘shoe’, and Fisch ‘fish’). Other West-Germanic languages such as Frisian and Dutch still have initial consonant clusters in these words (e.g. Frisian [skip] and Dutch [sxɪp] for ‘ship’). In different Germanic systems, /sk/-clusters are more or less stable. For instance, in the history of Icelandic, /sk/ remained fairly stable, whereas /sk/ was less stable in the history of English and very unstable in the history of German, where such clusters are now extremely rare. Even though these three languages derived from a common origin, they put different restrictions on consonant clusters in their respective histories. We conjecture here that Icelandic retained /sk/ clusters because they were highly frequent and could be syllabified as coda-onset clusters in most environments (e.g. fis.

103

104

II Linguistic Levels kur ‘fish-NOM’, fis.kinn ‘the fish-ACC’); they are retained in environments where other languages would ban them because of analogy (e.g. in fisk ‘fish-ACC’) – here the cluster is not modified in analogy to the form fiskur ‘fish-NOM’). In English, /sk/ clusters were modified at a particular stage in the history of the language and they were reintroduced when a large number of loanwords from Old Norse, Latin, etc. entered the language which had /sk/-clusters. Finally, in German, /sk/ clusters underwent a change that also affected most loans (presumably because German never borrowed as many words which involved /sk/-clusters as English did, so that the driving force to reintroduce them – i.e. the renewed relative frequency of /sk/-sequences – did not apply here; only a very few loanwords, all of Greek origin, retained the /sk/ cluster, e.g. Ger. Skelett ‘skeleton’). Theoretically, there are many ways to resolve the ban on /sk/ clusters, e.g. metathesis (/sk/ → /ks/), prosthesis (/sk/ → /ɛsk/), epenthesis (/sk/ → /sәk/), simplification (/sk/ → /s/ or /sk/ → /k/), gemination (/sk/ → /ss/ or /kk/), change of the place or manner of articulation of one of the consonants in the cluster (/sk/ → /sx/), or merger, so that the place of articulation of one segment survives and the manner of articulation of the other segment survives (/sk/ → /t/ or /sk/ → /ʃ/). The open question is why a language opts for which solution. The first possibility, metathesis, is not favored in word-initial positions, because the result does not make a better word-onset than the original cluster. Prosthesis is often found in a context of a preceding consonant, so that the preceding consonant can fill the onset position and the prosthesized vowel can form the nucleus of a new syllable of which /s/ can be the coda consonant. Prosthesis is attested in texts from the 2nd century onwards in word-initial /s/ plus voiceless stop clusters after consonants in Romance languages. This process gradually spread to all contexts in which word-initial /s/ is followed by a stop (e.g. Fr. esprit ‘spirit’ and Spanish estado ‘state’). English has borrowed a considerable number of words with initial /ɛs/ plus stop clusters from Romance (e.g. escape). Epenthesis is an option that second language users often employ to remedy syllable structures that do not occur in the native language. Simplification of /s/ plus consonant clusters is an option that is often found in early child speech. Kiparsky (2003: 329) points out that assimilation of consonant clusters resulting in gemination seems to happen in languages that already have geminates, whereas languages without pre-existing geminates prefer to simplify clusters. A change of manner of articulation of the second consonant (/sk/ → /sx/) has taken place in Dutch. The merger of the segments /s/ and /k/ into one that shares properties with both consonants such as /t/ – which like /s/ is alveolar and which like /k/ is a stop – is unlikely, because the single consonant /t/ does not reflect the perceptually distinct stridency of the original /sk/ cluster. In Old English, the strident nature of the first element in the cluster is preserved and the alveolar fricative is retracted in the context of the following back consonant which is left unrealized (/sk/ → /ʃ/). This process introduces a new segment: the system did not accommodate alveopalatals before the change took place and the effect of the change is thus an extension of the consonant inventory. Note that we do not need to assume the introduction of a novel distinctive feature for this change if we assume that velar sounds are specified for the feature [back]; alveopalatal fricatives may be then specified for [strident] as well as [back], so that the contrast between /s/ [strident], /ʃ/ [strident, back] and /x/ [back] can be expressed by features that were already distinctive in the phonology of speakers of English. One of the most interesting aspects of linguistic change is the fact that some units resist change. In Romance languages, prosthesis affected not only /sk/, but also /sp/

7 Linguistic Levels: Phonology and /st/. In English, only /sk/ underwent a change. The fact that /sk/ is modified whereas /sp/ and /st/ do not results in an asymmetry in the system: now only labial and alveolar stops follow /s/; velar stops no longer do. However, the change from word-initial /sk/ → [ʃ] did not apply across the board and was not permanent. Consider in this respect that under the influence of Scandinavian invaders and Latin scholars, loanwords were introduced into the language which started with /sk/ clusters and these clusters were not modified, but borrowed as such into the English language (e.g. skill and sky, which are loans from Old Norse). The temporary ban on /sk/ clusters resulted in a change in the linguistic system. The change /sk/ → [ʃ] itself was short lived, but it had a long-lasting effect. In Jakobson’s terminology, this sound change is the “formation of a contrast”: the effect of the change on the system is the creation of a new contrast within the class of fricatives (such that alveolar /s/, for instance, now contrasts with alveopalatal /ʃ/).

2.2 Stability and instability in consonant inventories related to allophonic variation In many varieties of Old English, most notably West Saxon and Northumbrian, the Germanic voiceless velar stop /k/ gradually developed into the voiceless palatal affricate [tʃ] in the following three contexts: (i) if initial followed by a front vowel or /j/ (e.g. OE cild [kild] → [kjild] → [tʃı¯ld] ‘child’; OE short vowels lengthened sometime around the 10th century if they were preceded by /ld/ clusters. Long vowels later underwent the Great Vowel Shift (e.g. OE /kı¯ld/ → Early Modern English /tʃaild/) (see Krug, Chapter 48). (ii) if medial preceded and followed by a front vowel (e.g. OE cwice ‘quitch’) and (iii) if final preceded by front vowel (e.g. OE ic [ɪk] → [ɪtʃ] ‘I’). The voiced velar stop /ɡ/ underwent a similar process: if followed by a front vowel it was realized as /j/ (e.g. *georn → [j]ern ‘eager’). The voiced stop /ɡ/ was palatalized if preceded by a front vowel (OE bricg [brɪɡ] → [brɪɡj] → [brɪdʒ] ‘bridge’), or between a nasal and /j/ (e.g. *sangjan → sen[dʒ]an ‘to singe’), or when geminated before /j/ (e.g. *laggian → lecgan ‘to lay’, where represents /dʒ/). At the time when palatalization of /k/ and /ɡ/ to [kj] and [ɡj]/[j] took place, the consonant inventory was stable and no phonological contrast was introduced. Rather, the segments /k/ and /ɡ/ had allophonic variants which they did not have before. In later stages in the history of English, we find velar stops before front vowels both in words of Germanic origin and in borrowed words. With respect to words of Germanic origin consider that for instance PGrmc. *kunningaz had a velar stop before a back vowel. At the time that palatalization applied in Old English, the velar stop was still followed by a back vowel. Later, the vowel in question was affected by a process commonly referred to as “i-mutation”, by which stressed long and short back vowels were fronted in the context of a following high front segment ([i], [j], [y]): PGrmc. *kunningaz > OE kynning (mostly written as ) ‘king’. After i-mutation, the process of palatalization did not apply and for this reason, many words of Germanic origin have a velar stop followed by a front vowel. Also note that palatalization of velar stops did not take place in Scandinavian borrowings (e.g. Present-day English kid, kettle, dike,

105

106

II Linguistic Levels give, get, egg). Thus, [tʃ] and [dʒ] are not allophonic variants of the phonemes /k/ and /ɡ/ in Modern English. Instead, after the rule of palatalization ceased to play a role in the phonology of Old English and after borrowings from Scandinavian and other languages and after the change from [kj, gj] to [tʃ, dʒ], the consonant inventory is expanded, i.e. a new system emerged in which the phonemes /k/ and /ɡ/ contrast with the phonemes /tʃ/ and /dʒ/, respectively. (For a more detailed account of palatalization, the reader is referred to Murray, Chapter 17, Section 3.3.) There is a lot of debate over the realization of velar fricatives. Most textbooks maintain that OE /x/ was realized as [h] word-initially before vowels and before the sonorants /n, l, r, w/, it was pronounced as [x] medially and finally except after front vowels and it was realized as palatal [c¸] after front vowels. The variant [c¸] gradually vocalized in all southern English dialects, resulting in compensatory lengthening of a preceding short vowel, which was later affected by the Great Vowel Shift (e.g. [ɪc¸] → [iː] → [ai] in words like knight, and night). In late Middle English, the variant [x] was labialized and changed into [f] when following round vowels (mostly /u/) in some dialects (cf. ModE enough where = /f/) and gradually disappeared entirely in other phonological contexts. With respect to the direction of phonological change, it is interesting to reflect on the following developments: in some variants of English, /xt/ changed into /ft/ after a short back vowel (cf. Modern English draught), whereas in the early history of Dutch, monomorphemic /ft/ was disfavored and realized as /xt/ (compare Ger. Luft ‘air’ and Kraft ‘power’ to Dutch lucht and kracht and English soft and after to Dutch zacht ‘soft’ and achter ‘after, behind’ where represenets [x]). It is thus impossible to say that postvocalic /xt/ is generally less favored than postvocalic /ft/. Rather, both /xt/ and /ft/ are “unstable” clusters and the one may turn into the other, depending on the local conditions within a language at a certain time: English developed in such a way that velar fricatives became disfavored (i.e. they became obsolete by replacement by other segments or by omission), whereas at a certain time, Dutch extended the distribution of velar fricatives (e.g. the Germanic voiced velar plosive is realized as /x/ in most dialects of Modern Dutch). Another consonantal innovation in Old English concerns allophonic variation among the fricatives: the singleton (or “short”) voiceless fricatives /f, θ, s/ became voiced word-medially in the context of other voiced phonemes, as in (4): (4)

SG

PL

wulf smiþ ( = /θ/) hu¯s

wulfas ( = /v/) ‘wolf’ smiþas ( = /ð/) ‘smith’ hu¯sa ( = /z/) ‘house’

After this allophonic rule had taken effect, no singleton voiceless fricatives were realized in word-medial position after a vowel. Old English intervocalic geminate consonants were degeminated later and this change gave rise to the phonemic opposition between short voiced and voiceless fricatives in word-medial positions. Thus, the addition of the phonological rule of degemination to the grammar generated segments (in this case singleton fricatives) that might have been the input of the phonological rule of intervocalic voicing. The fact that the degeminated fricatives did not undergo voicing indicates that the former allophonic rule ceased to have an effect. The result of these developments was that in word-medial positions, the opposition geminate versus singleton fricative was replaced by the opposition voiceless versus voiced fricative.

7 Linguistic Levels: Phonology The word-final contrast between voiced and voiceless fricatives developed through schwa-reduction of full vowels in unstressed position and eventual loss of final schwa between 1100 and 1300, i.e. the fricatives that were voiced by an allophonic rule before, were no longer followed by a vowel, but occurred in word-final position. The initial voiced fricatives /v/ and /z/ originated primarily from Norman-French borrowings in Middle English (e.g. valour ‘valor’, veel ‘veal’, and zele ‘zeal’). The result of the allophonic variation in Old English between voiceless and voiced fricatives combined with the loss of geminate consonants and the loss of word-final schwa and the introduction of /v/ and /z/ in initial position with borrowings from French resulted in a system where voiceless and voiced fricatives are distinctive.

2.3 Changes in phonotactics without effects on the consonant inventory In describing phonotactic restrictions, it is generally agreed in phonological theory that the concept of “sonority” plays a crucial role. Sievers (1881) and Jespersen (1904) introduced the Sonority Sequencing Principle to explain the fact that, within a syllable, the less sonorous segments are found at the periphery and the most sonorous sounds are found in the syllable peak. The idea that sonority is not an absolute property, but rather a relative one, gave rise to the notion of “strength hierarchies” (e.g. Lass 1970) and a “sonority scale”. In the sonority scale presented in (5), the degree of sonority of segments increases from left to right: (5)

Sonority scale for some segments p, t, k, b, d, ɡ f, s, ʃ, x m, n l r j, w i, u e, o a stops fricatives nasals liquids glides vowels of different height

In Old English, possible syllable-initial consonant clusters only needed to show a slight increase of sonority. In stressed syllables, initial clusters of voiceless /k/ and /x/ or voiced /ɡ/ followed by a nasal (/kn/, /ɡn/, /xn/) or clusters of the glide /w/ followed by a sound which was of the same degree of sonority (a rolled /r/) were as common as clusters with a steeper increase of sonority such as stop-liquid clusters. Lutz (1992) suggests that initial clusters with unfavorable phonotactics may result in different changes such as (i) the loss of the initial consonant of the cluster (e.g. ME wlispen → 14th/15th centuries lisp and ME fne¯sen → 14th/15th centuries neeze ‘sneeze’), (ii) the replacement of the initial consonant by a consonant that forms a more favorable consonant sequence (e.g. OE wlott → 14th century blot), and (iii) loss of words that start with such unfavorable clusters (e.g. OE wlank ‘proud’, wrabble ‘squirm’, and gnede ‘misery’ became obsolete in the 15th and 16th centuries). The syllable-initial clusters /xn/, /xl/, and /xr/ were lost within a short time span between the OE and the ME periods. The initial clusters /kn/, /ɡn/, and /wr/ were spelled and probably still pronounced as such when the writing system became more and more standardized after the introduction of Caxton’s printing press in the late 15th century. The

107

108

II Linguistic Levels reduction of /wr/ to /r/ probably took place relatively early (possibly starting in late Middle English). Shakespeare’s puns on knight–night, knot–not and wring–ring indicate that the initial sound in such clusters was either no longer pronounced during his time, or that there was variation between the pronunciation with and without the initial consonant. In the course of the 16th century, the phonotactics of the language gradually changed in such a way that sonority distances in syllable-initial clusters became larger: except for word-initial clusters with /s/ (spy, stop, sky, sneeze), the only permissible onset clusters after Shakespeare’s time are those in which a stop or fricative is followed by a liquid or a glide (as in the words dry, fly, cue, queen). In 17th-century educated English, the reduction of /kn/ and /ɡn/ to /n/ was completed and other dialectal varieties followed. In Present-day English, words that are borrowed from other languages that start with such impermissible clusters are modified by English speakers in such a way that both sounds of the cluster are realized. However, the ban on having syllable-initial /kn/ or /ɡn/ clusters still exists. In order to realize both members of the clusters, the strategy that speakers use today is to insert a vowel between the two consonants. Thus, the Hebrew word Knesset and the German name Knopf, for instance, are both pronounced with initial /kәn/ in Present-day English (see Green 1997: 25–28). Thus, the Present-day repair strategy for unfavorable phonotactics is vowel epenthesis.

2.4 Stability and instability in the vowel inventory and the effect on phonotactics Middle English diphthongs /iʊ/ (occurring in words like chew, due, and hue) and /eʊ/ (occurring in words like beauty, dew, and few) eventually collapsed under /juː/ (see Schlu¨ter and Jones, this volume). In the diphthong /iʊ/, the first part was reanalyzed as the glide /j/ and assigned to the onset, while /ʊ/ turned into /u:/ to compensate for the loss of vowel quantity in the late 16th century. In the diphthong /eʊ/, the first member was gradually raised to /i/ after the 16th century. In the 18th century, the element /i/ was reanalysed as a glide and the element /ʊ/ was tensed and lengthened, so that the original contrast between /iʊ/ and /eʊ/ was lost by the end of the 19th century in most dialects of English. The question is whether the original diphthongs palatalized the preceding consonant (as in djue and djew), or whether the original diphthongs are now realized as sequences of the palatal glide /j/ followed by the high back vowel /u:/. Consider in this respect that consonants with and without a following /j/ are not allophones of a phoneme, because – as the (6) illustrates – there are also instances of consonants followed by a high back vowel without the intervening palatal glide /j/: (6)

Syllable-initial sonorant a. music [mjuːzɪk] b. nude [njuːd] c. lurid [ljuːrɪd]

+ /j/ clusters in Present-day British English versus moose [muːs] versus noon [nuːn] versus lunatic [luːnətɪk]

Interestingly, even though the palatal glide [j] may follow [l] in word-initial position in Present-day British English, we never find that a cluster of a non-strident obstruent plus [l] is followed by [j]. Hence, there is no [j] between [l] and a following vowel in words like plumage, blue, clue, glue, but, according to Harris (1994: 61), some speakers have

7 Linguistic Levels: Phonology one in slew. This fact has important consequences for the analysis of English syllable structure. In particular, we might conclude from this observation that /j/ should not be analyzed as a secondary place of articulation. If it were, the explanation for the fact that a palatal glide may be present in single consonants, but not in some consonant clusters would be cumbersome. Instead, the phoneme /j/ is a consonantal segment which may occupy one position in the onset. Apart from /s/-initial clusters, the English onset may be filled by at most two positions and three-consonantal clusters such as */plj/, */blj/, */klj/ or */ɡlj/ would violate this restriction. Note that American English introduced a further restriction on initial clusters with /j/: it bans clusters in which a coronal consonant is followed by a palatal glide /j/ (as in the words tune, dune, suit, new). Since a palatal place of articulation is a sub-class of coronals, this ban may be interpreted as an OCP-effect (see Section 2.1).

3 The future of looking back in time on sound changes For Roman Jakobson and many phonologists after him, “phonology is the study of the properties of the sound systems which speakers must learn or internalize in order to use language for the purpose of communication” (Hyman 1975: 1). The word “systems” implies the notion of “stability” and suggests that the properties of sound systems are stable. However, as we saw in this chapter, little in phonology seems to be stable. For communication to be successful, speakers and hearers must share linguistic knowledge (e.g. knowledge of how sounds are used by the speaker). When we compare different stages of a language, we find that the properties of sounds that speakers use are not stable at all but nevertheless, communication between speakers and hearers functions well. How is this possible? Part of the answer lies in the fact that phonological properties of human languages are flexible and constrained only by what our vocal tract allows us to produce and by what our ears allow us to perceive. For instance, to differentiate voiceless stops from voiced ones, the following properties may vary: the length of the preceding vowel, the formants of a preceding and following vowel, the duration of the stop’s closure phase (i.e. closure duration), the presence/absence of vocal fold vibration during the closure phase (so-called “closure voicing”), the voicing lag or “VOT”, i.e. the time between the release of the closure phase and the onset of vocal fold vibration (with a large scale to choose from), burst intensity, burst duration, and more. In English, the major cues to distinguish laryngeal classes of stops are the length of the preceding vowel and the VOT values of the stops in question: voiced stops are preceded by a relatively long vowel duration and have a relatively short voicing lag whereas voiceless stops are preceded by a comparatively shorter vowel and have a longer voicing lag when initial in a word. Pohl and Grijzenhout (2010 and references therein) suggest that in German, a difference in voicing lag is the major cue to maintaining the contrast, whereas in Dutch closure voicing and the duration of the closure phase seem to matter most to identify a voiceless or voiced stop and in Swiss German closure duration alone contrasts two classes of stops. Thus, in the modern Germanic languages, each language favors a different cue or set of cues (and the speakers have different mental representations for the contrast in question). Children must “learn” to pay attention to the relevant cues in their language and then “internalize” that information – i.e. to build mental representations – to use it to convey meaning. Later generations may shift the attention to another cue or cues and this shift may

109

110

II Linguistic Levels be very small and hardly noticeable at first. When this shift is used by more speakers of a community, it may cause temporal or permanent change. For example, Dutch seems to shift from closure voicing and closure duration to differences in VOT as the major cue to maintain the distinction; the contrast between /p/ and /b/ remains, but its instantiation is different and it may lead to different mental representations for future generations. Thus, one source of change in the sound system is its inherent variability and flexibility which allows speakers to highlight relevant phonological properties and vary the irrelevant properties of sounds. Hearers too are able to ignore irrelevant phonetic detail (i.e. they abstract away from much information in the acoustic signal) and they pick up on just a few properties of the speech signal that make a distinction between two sounds. The system may change to enhance, create, or omit a contrast (e.g. a system with positive VOT distinctions of 10ms versus 80ms may become a system with negative VOT and positive VOT, or there may be room to add another VOT contrast or the contrast may vanish over time). Part of the answer why change occurs is thus: because the system is flexible, change is pre-programmed. The next question is when does change occur? The Neogrammarians advocated the view that change is “spontaneous” and others also find an answer in sociological factors (e.g. peer group pressure, the need to be different from your neighbor, the need to borrow words from your neighbor). Human beings live in groups. In order for the group to be successful, it has either to blend in with other groups (so that a language contact situation may result in language change), or to compete with other groups (so that variations emerge within one dialect or language which may eventually lead to language split) or – when the group becomes too big – to form subgroups with their own identity (and, hence, their own language variety). This is a continuing or “dynamic” process. Languages will always change and there is enough flexibility in the speech signal to allow them to do just that. The question that many people raise is why some languages seem to resist change or change relatively slowly, whereas others seem to change more rapidly. To illustrate the point, we would here like to compare the histories of Icelandic and English. Icelandic is spoken in a relatively isolated part of the world where it did not come in contact with many other languages (one possible source of language change). Moreover, as Kristja´n ´ rnason (p.c.) once pointed out to me, there has been a long tradition of a conscious A effort to maintain the language and not to allow innovations or to “imitate” the Danes and the Germans. Dialectal variations that emerged in smaller communities in Iceland were stigmatized and gave way to the standard dialects spoken by the majority. Moreover, the population is small enough for language planning committees to be able to impose their recommendations and social pressure has a major effect on the relative stability of the language. In contrast, ever since the Anglo-Saxons landed on the British Isles, the English language has been exposed to foreign influences and this has had a major effect on the relative instability of some of its properties. There have been many attempts to preserve or “improve” the language, but such attempts apparently have been less successful than in Iceland (see Percy, Chapter 63). As late as at the turn of the 20th century, phoneticians like Daniel Jones attempted to develop a theory of sound systems that would help second language learners to acquire the accurate pronunciation of words in a foreign language (Jones 1917). Moreover, phonetics as the study of sound systems was considered to be useful in helping to improve the pronunciation of native speakers. George Bernard Shaw certainly contributed to this view in

7 Linguistic Levels: Phonology 1912 when he made Professor Higgins teach Eliza Doolittle the “proper” pronunciation of words in his popular play Pygmalion. Such attempts have proven to be futile: there is no proper pronunciation of English and the many speakers of the language will never employ exactly the same pronunciation of all sounds. English has always had many dialects: some are conservative and have archaic features, whereas others are more progressive and are open to innovations. Both types of dialects have considerable numbers of speakers for whom it is advantageous (from a sociolinguistic perspective) to use the dialectal variety. Also, since the 16th century, the language has expanded to new territories and the distance between the British Isles may have lead to independent developments between varieties of English spoken throughout the world. There are thus many reasons why English has changed so much over the centuries and why there is so much diversity. The future of looking back on English sound change lies in the fact that English is a global language with a rapidly growing number of speakers and for this reason alone, interest in its ever changing sound system will not vanish. Even though this handbook is voluminous, it will still be incomplete in that not all changes which have occurred in the English language will be illustrated and discussed. Improved techniques, the availability of more extensive corpora, etc. will in the near future certainly result in extending our knowledge of which aspects triggered which kind of change in the history of the English language. Linguists like Aditi Lahiri and Frans Plank (p.c.), however, hold that it is actually more surprising that some aspects of a language are stable and do not change (or change relatively slowly or only within a relatively small area), i.e. the future of looking back on linguistic change also lies in the interest in which properties of language are unstable and highly variable compared to properties that are stable or invariable in the sense that they resist change or change slowly. This is a relatively underexplored field.

4 Summary Phonology is the study of the properties of sounds which speakers use to convey meaning. As pointed out in Section 1, a primary goal of structural and generative phonology is to investigate the function of speech sounds, i.e. to establish and compare phoneme inventories of languages. In this chapter, we have seen examples of phonemes that changed their shape (e.g. voiceless stops becoming fricatives according to Grimm’s Law), phonemes that were added to the English inventory (e.g. alveopalatal fricatives), and phonemes that became obsolete (e.g. /kw/, /x/). We have also looked into the behavior of speech sounds: some allophonic rules (e.g. voicing of fricatives in the context of other voiced segments) emerged at some stage and disappeared again later. Section 2.2 discussed two examples of allophonic variants that developed into phonemes (alveopalatal affricates and voiced fricatives, respectively). Another goal of phonology is to account for the distribution of speech sounds. For English, we have seen that changes in phonotactics may be temporal (e.g. the ban on /sk/) and that the effect of a phonotactic restriction may give rise to a new phoneme (e.g. /ʃ/). Some phonotactic restrictions may be added relatively late in the history of a language (e.g. the ban against /kn/ clusters) and may change their effect (from deletion of a segment in the cluster to insertion of a vowel to break up the cluster). Sometimes, a change in the phoneme inventory may result in the addition of new

111

112

II Linguistic Levels phonotactics (e.g. changes from the diphthongs /iʊ/ and /eʊ/ to the monophthong /uː/ preceded by /j/ which can form a complex onset with any single consonant). In many cases, a merger of two consonants into one may result in a new phoneme for which the distinctive features are already part of the system (/sk/ → /ʃ/), allophonic rules and borrowings may also create new phonemes (e.g. voiced fricatives) and a change from a complex phoneme into a simplex one may result in new phonotactics.

5 References Aitchison, Jean. 2001 [1981]. Language Change: Progress or Decay? 3rd edn. Cambridge: Cambridge University Press. Baudouin de Courtenay, Jan Niecisław. 1895. Versuch einer Theorie phonetischer Alternationen. Strasbourg: Tru¨bner. Goldsmith, John A. 1995. Phonological theory. In: John A. Goldsmith (ed.), The Handbook of Phonological Theory, 1–23. Cambridge, MA/Oxford: Blackwell. Green, Anthony Dubach. 1997. The Prosodic Structure of Irish, Scots Gaelic, and Manx. Ph.D. dissertation, Cornell University. Harbert, Wayne. 2007. The Germanic Languages. Cambridge: Cambridge University Press. Harris, John. 1994. English Sound Structure. Cambridge, MA: Blackwell. Hyman, Larry M. 1975. Phonology: Theory and Analysis. New York: Holt, Rinehart and Winston. Iverson, Gregory K. and Joseph C. Salmons. 1995. Aspiration and laryngeal representation in Germanic. Phonology 12: 369–396. Jakobson, Roman. 1939. Observations sur la classement phonologique des consonnes. Proceedings of the 3rd International Congress of Phonetic Sciences: 34–41. Jakobson, Roman, Gunnar Fant, and Morris Halle. 1951. Preliminaries to Speech Analysis. Cambridge, MA: MIT Press. Jakobson, Roman and Morris Halle. 1956. Fundamentals of Language. The Hague: Mouton. Jakobson, Roman. 1971 [1962]. Retrospect. In: Selected Writings I: Phonological Studies, 631–658. The Hague: Mouton. Jespersen, Otto. 1904. Lehrbuch der Phonetik. Leipzig: Teubner. Jones, Daniel. 1917. An English Pronouncing Dictionary. London: Dent. Jones, Daniel. 1967. The Phoneme; its Nature and Use. 3rd edn. Cambridge: W. Heffer & Sons Ltd. Kiparsky, Paul. 2003. The phonological basis of sound change. In: Brian D. Joseph and Richard D. Janda (eds.), The Handbook of Historical Linguistics, 313–343. Malden, MA: Blackwell. Lass, Roger. 1970. Boundaries as obstruents: Old English voicing assimilation and universal strength hierarchies. Journal of Linguistics 7: 15–30. Lass, Roger. 1984. Phonology: An Introduction to Basic Concepts. Cambridge: Cambridge University Press. Lutz, Angelika. 1992. Lexical and morphological consequences of phonotactic change in the history of English. In: Matti Rissanen, Ossi Ihalainen, Tertta Nevalainen, and Irma Taavitsainen (eds.), History of Englishes, 156–166. Berlin/New York: Mouton de Gruyter. McCarthy, John J. 1993. A case of surface constraint violation. Canadian Journal of Linguistics 38: 169–195. McCarthy, John J. and Alan S. Prince. 1993a. Generalized Alignment. In: Geert Booij and Jaap van Marle (eds.), Yearbook of Morphology 1993, 79–153. Dordrecht: Kluwer. McCarthy, John J. and Alan S. Prince. 1993b. Prosodic Morphology I: Constraint Interaction and satisfaction. Ms., University of Massachusetts, Amherst and Rutgers University. McCarthy, John J. and Alan S. Prince. 1995. Faithfulness and reduplicative identity. In: Jill N. Beckman, Suzanne Urbanczyk, and Laura W. Dickey (eds.), University of Massachusetts Occasional Papers 18: 249–384.

8 Linguistic Levels: Prosody

113

Pohl, Muna and Janet Grijzenhout. 2010. Phrase-medial bilabial stops in three West Germanic languages. Linguistische Berichte 222: 141–167. Prince, Alan and Paul Smolensky. 1993. Optimality Theory: Constraint interaction in Generative Grammar. Ms., Technical Report no. 2, Rutgers University Center for Cognitive Science, Cambridge, MA, MIT Press. Sapir, Edward. 1921. Language: An Introduction to the Study of Speech. New York: Harcourt, Brace; downloaded from http://www.bartleby.com/186/. de Saussure, Ferdinand. 1916. Course de linguistique ge´ne´rale. Paris: Payot. Sievers, Eduard. 1881. Grundzu¨ge der Phonetik. Zur Einfu¨hrung in das Studium der Lautlehre der Indogermanischen Sprachen. Leipzig: Von Breitkopf & Ha¨rtel. Waugh, Linda R. 1976. Roman Jakobson’s Science of Language. Lisse: The Peter de Ridder Press. Wunderlich, Dieter. 2008. Spekulationen zum Anfang von Sprache. Zeitschrift fu¨r Sprachwissenschaft 27: 229–265.

Janet Grijzenhout, Konstanz (Germany)

8 Linguistic Levels: Prosody 1 2 3 4 5 6 7

Definition of terms Syllable structure and syllable weight Historical sources of information for prosodic reconstruction Old English meter and prosody Middle English meter and prosody Post-Middle English prosodic innovations References

Abstract The chapter traces the development of word and phrasal stress from Old to Present-day English. Section 1 and Section 2 define the terms needed to describe the prosodic patterns of speech and address the notions of syllable structure and syllable weight. Section 3 surveys the methodological bases for prosodic reconstruction, focusing specifically on the interplay between meter and language in the recovery of rhythmic patterns in speech. Old English meter and prosody are covered in Section 4, where the basic principles of Old English alliterative versification provide the foundation for reconstructing word and phrasal stress. Middle English meter and prosody are covered in Section 5, again with specific references to metrical form, word stress, and phrasal stress. The section includes a discussion of the effect of lexical borrowing from French and Latin on the prosody of English. Section 6 is devoted to the major prosodic changes in English during and after the Renaissance.

1 Definition of terms The term “prosody”, as used in this chapter, refers to the properties and the organization of syllables into words, phrases, and sentences in speech. Outside of linguistics, the term prosody can also be used with reference to the study of verse and its properties; Alex Bergs and Laurel Brinton 2012, Historical Linguistics of English (HSK 34.1), de Gruyter, 113–128

114

II Linguistic Levels for the conventionalized rhythmic structures of verse we reserve the term “meter”. The prosodic properties of speech are “suprasegmental”: their domain is larger than individual speech sounds, which are organized into higher-level units that are independently pronounceable, namely “syllables”. The ability to divide an utterance into syllables is part of the intuitive knowledge that speakers have of their language. Very importantly, syllables are the carriers of “stress”, the contrastive intensity that marks some syllables as more or less prominent. Phonetically, stress is associated with the use of a greater amount of respiratory energy on a syllable, increased tension of the vocal folds, and loudness. In terms of metrical structure, the prominent position is called an “ictus” (S); ictic positions are usually, but not always, filled by stressed syllables, while “non-ictic” (W) positions attract unstressed syllables. Stress is binary in the sense that syllables are either stressed or unstressed. Further, a stress may range from a full primary/main stress, here marked with ´ (acute), to various levels of non-primary stress, here marked with ` (grave). Although informally we speak of “stressed” and “unstressed” vowels, and we place the stress marks over the vowels for typographic convenience, it is important to bear in mind that stress is a property of the entire syllable.

2 Syllable structure and syllable weight The syllable is the smallest pronounceable prosodic unit, but it is also structurally complex in that it is further decomposable. At the core of the syllable is its “nucleus” or “peak”, the segment of highest sonority in the string. Every syllable has to have one and only one nucleus, usually a vowel or a diphthong, but sometimes also a syllabic sonorant /r̩ , l̩, m̩ , n̩ /. Consonants or consonant clusters to the left of the nucleus constitute the syllable “onset”, and the consonants following the nucleus make up the “coda”. The onset and the coda are not obligatory elements of the syllable. Universally, a filled onset is preferred to a filled coda. A coda consonant can contribute to the “weight” of a syllable, whereas an onset is commonly considered weight-neutral. The division of a string of sounds into syllables follows the “Maximal Onset Principle”. According to that principle, a single consonant between two vowels fills the onset of the syllable to the right (syllable divisions are marked with a period): rea.son, e.ne. my, de.hu.mi.di.fy. A two-consonant cluster is either divided or not, depending on whether the resulting onset is also a possible word-initial cluster: com.post, prag.ma. tic, fic.tion, but hi.sto.ry, pa.tri.ot. Three consonants between vowels are split again depending on the nature of the resulting cluster: emp.ty, friend.ly, coun.try, um.bre.lla, a. strin.gent, o.sprey. The Maximal Onset Principle does not apply across prosodically independent words, so boil eggs is not *boi.leggs. Syllable weight is a prosodic property tightly associated with stress: universally, heavy syllables attract stress and syllables that carry stress are likely to become heavy. In English a “heavy syllable” is any syllable whose peak is a long vowel or a diphthong: see.saw, pay.ee, or any syllable that ends in a consonant: com.pul.sion, prac.tice. A “light syllable” has a short vowel in the peak and no coda: A.me.ri.ca, re.pli.ca. In practice, in Present-day English all syllables except those ending in /ɪ, ɛ, æ, ʊ, ʌ, ə/ are heavy. Monosyllabic major class words (clue, club, day, fry, three, wet) cannot have a light syllable; it follows that /*clɛ, *frɪ, *sʊ/ would not be possible

8 Linguistic Levels: Prosody English words, while /clɛs, frɪn, sʊg/ are possible words which are accidental gaps in our vocabulary.

3 Historical sources of information for prosodic reconstruction Reconstructing the prosodic properties of the earlier stages of English is a complex task. The types of segmental changes that occur in stressed and unstressed syllables are very dissimilar. Vowel lengthening, vowel shifting, and gemination typically occur in stressed syllables, while vowel reduction and loss and consonant lenition are expected in unstressed syllables. If we find textual evidence of such processes, we can make prosodic inferences: the progressive reduction and loss of the prefix ge- (OE geriden > ME iriden > PDE ridden ‘ridden’) is good evidence that ge- was unstressed in Old English and Middle English. Similarly, Middle English spellings luved, luvd for earlier luvede ‘loved’ indicate reliably that the form was initially stressed. Our most direct source of information about the prosodic structure of earlier English, however, comes from the way in which the forms of speech are matched to the structural positions in verse. The greatest challenge for the use of verse as the primary evidence for prosodic reconstruction is circularity: since there are no records of instructions on what is permitted in early versification, we rely on templates extrapolated from the surviving poetic corpus. Our understanding of how the metrical templates worked is thus founded on a web of typological inferences about language and meter with no possibility of direct verification. The way we avoid ignotum per ignotius, explaining ‘the unknown by means of the more unknown’ is by applying testable quantitative and typological criteria to the formulation of the rules of meter and the reconstruction of prosodic patterns. The statistical data on some features, e.g., in 26,088 verses of OE poetry, only 36, or 0.001%, lack alliteration (Hutcheson 1995: 169), justify reliance on alliteration as the binding principle in the alliterative long line, here marked in boldface. Moreover, testably unstressed syllables, such as inflectional syllables, never alliterate, which makes the co-occurrence of stress and alliteration a solid source of prosodic reconstruction. No matter what theory of Old English meter one adopts, there can be no doubt that in Beowulf (henceforth Beo) 102: wæs se grı´mma gæst / Gre´ndel haten ‘was the grim ghost / Grendel called’, the words grı´mma and Gre´ndel are initially stressed. Typologically too, all Germanic languages, including Present-day English, stress native unprefixed words on the first syllable; we can safely project that back to Old English and posit ´ ðen ‘heathen’, so´þe ‘truly’. root-initial stress on cy´ning ‘king’, de´maþ ‘they judge’, hæ The alignment of the main stress with the left edge of a simplex word in early English is known as the “Germanic Stress Rule” (GSR). Statistical and typological grounding of prosody-meter correspondences is our best recourse in spite of some inherent uncertainties. The historical poetic corpus presents cases where deviations from an established norm may be interpreted as deliberate creative choices. A poet may force an unstressed syllable into an ictic position to fit the expectations of the template: thus Chaucer rhymes felawe : awe, biddyng : thing. This convention of versification is of no use to us in trying to reconstruct the prosodic contour of Germanic felawe or biddyng in speech – the words were always initially stressed. On the other hand, Chaucerian rhymes such as honour : flour, servise : wyse have, all too freely, been taken as evidence for non-initial stress on the Romance borrowings honour

115

116

II Linguistic Levels and servise. Such evidence has to be evaluated carefully and compared to the evidence of the placement of such words in line-medial position. Similarly, placing re´ady, u´nder, ma´keth at the left edge of an iambic (W S) line is a metrical inversion which breaks the monotony of repeated identical structures, but it tells us nothing new or special about the prosody of these native words. However, placing loans such as citees, justice, poynaunt line-initially is open to both WS and SW meterical scansion and can be considered good evidence that such words maintain their Romance stress contour. We will return to these metrical issues in Section 5.1. For now we just note that decisions on the prosodic history of loanwords will have to be based on fine-grained and comprehensive coverage of the placement of individual items in the verse.

4 Old English meter and prosody Germanic and Old English versification is notoriously difficult to model. Although new theories of Old English meter continue to appear, most recently in Getty (2002) and Bredehoft (2005), no new approach rivals the descriptive adequacy and scholarly acceptance of the observations and patterns in Sievers (1893); see Stockwell and Minkova (1997), Minkova (2008a).

4.1 Basic principles of Old English alliterative verse Sievers’s hypothesis about the metrical structure of Old English verse rests on the following configurations: – a line consists of two verses, the “on-verse” (“a-verse”) and the “off-verse”, (“b-verse”), linked by alliteration; – each verse contains two feet and at least four positions; – each foot contains an ictus (S), also known as a “lift”, and at least one non-ictic position (W), also known as a “dip”. This allows us to represent the structure of the line as in Figure 8.1, where the numbers at the bottom stand for positions: LONG LINE

ON-VERSE/a-VERSE

FOOT

1 ne

OFF-VERSE/ b-VERSE

FOOT

2 leof

‘not friend

3 ne

FOOT

4 lað

nor enemy

1 be

FOOT

2 lean

dissuade

3 mih

4 te

could’ (Beo 511)

Figure 8.1: The structure of the Old English verse line

8 Linguistic Levels: Prosody The binary representation in Figure 8.1 is an abstraction based on the minimal line structure in terms of syllable count. The prominence relations are unspecified; within the feet, lifts and dips can appear in either order. Each position is ideally filled by a single syllable, and an S position must be filled by at least one syllable. Unlike the familiar notion of classical metrical feet, positions and feet in Old English verse may be of uneven size, due mainly to the expandability of the non-final weak positions in each ´ re ‘thanes are united’ (Beo 1230a) the template S verse, thus in þe´gnas syndon geþwæ W S W has the first W position filled by four unstressed syllables: -nas syn.don ge-. The one-to-one correspondence between a syllable and a position may be disregarded for S-positions under special metrical conditions; this is known as “resolution”. Resolution is a metrical equivalence: one and only one heavy syllable can fill a lift, but a light syllable and any other syllable may jointly fill a lift to avoid an unacceptable metrical violation, such as an expanded dip at the right edge of the verse. Thus in the S W S W verse re´ceda under ro´derum ‘of halls under heavens’ (Beo 310a), the syllables ro´. de- are metrically subsumed under the second S position to avoid the unacceptable matching of the last W to -de.rum. The conventions of alliteration which help us separate relevant from irrelevant metrical information are: – in the on-verse both S positions may alliterate. – in the off-verse only the first S position is allowed to alliterate. Nearly all verses are complete syntactic units. The smallest linguistic units that occupy a verse are compounds, e.g. þeodcyninga ‘of tribe-kings’ (Beo 2a), wilgesiþas ‘willing companions’ (Beo 23a), landgemyrcu ‘shore-boundaries’ (Beo 209b). Most often, however, a verse is coextensive with a clause or a syntactic phrase: Hi hine þa ætberon / to brimes faroðe ‘they him then carried / to the sea’s current’ (Beo 28). An intriguing convention, not fully understood, describes the hierarchy of syntactic elements within the verse with respect to alliteration. In a verse where the S-positions are filled by a noun and a verb, the noun will consistently be strong, whether it is an NP-VP string as in Him ða Scyld gewat ‘Then Scyld departed’ (Beo 26a), or a VP-NP: Gebad wintra worn ‘Lived to see winters many’ (Beo 264a). This alliterative regularity is known as Sievers’s “Rule of Precedence” (Sievers 1893: Sections 22–29); it states that if an inflected verb precedes a noun it does not have to alliterate, that it must not alliterate if the noun does not alliterate too, and that a non-alliterating noun can never be followed by an alliterating finite verb. The rule does not exclude double alliteration: þenden wordum weold ‘when with words ruled’ (Beo 30a), geafon on garsecg ‘gave in ocean’ (Beo 49a), so projecting the Rule of Precedence on to the prosody of Old English speech is not always straightforward. The significance of alliteration in the reconstruction of phrasal and utterance prosodic contours will be discussed further in Section 4.3.

4.2 Old English word stress As noted in Section 3, Old English word stress falls on the first stressed syllable of word roots. The acoustic prominence of stress is thus, unsurprisingly, an important and consistent morphological boundary signal. All root-initial syllables are stressed. The weight of the root-initial syllable is irrelevant; both heavy and light syllables can be stressed: drı¯´.fan

117

118

II Linguistic Levels ‘drive’, fu´l.tum ‘help’, me´n.gan ‘mix’, so´¯ .na ‘soon’ (heavy), and cy´.ning ‘king’, ga´.fol ‘tax’ me´.du ‘mead’, sca´.mu, ‘shame’ (light). A very important difference between Old English and Present-day English is the stability of stress on the first root syllable in a derivational set: while suffix-induced stress-shifts in Present-day English can leave root-initial syllables completely stressless: chronic–chrono´logy, ´ıdiot–idio´tic, so´lid–solı´dity, Old English word roots are always marked by the presence of stress: ge´ogoð ‘youth’ hla´ford ‘lord’

ge´ogoðhad ‘youth-hood’ hla´fordscipe ‘lordship’

wo´ruld ‘world’ wu´ldor ‘glory’

wo´ruldlic ‘worldly’ wu´ldorfull ‘glorious’

The addition of suffixes in Old English never affects the primary prominence. The suffixes themselves can bear some degree of non-primary stress because they can be ictic, but they are automatically excluded from the positions of obligatory alliteration, the first ictic positions in each verse. Inflectional suffixes are always unstressed, while derivational suffixes exhibit complex behavior in the verse and it is likely that their prosodic realization in speech was gradient, ranging from non-primary stress to absence of stress. The variability is attested both synchronically and diachronically. The position of the suffix with regard to the word boundary is of relevance, and so is vowel quality and quantity. When inflected, heavy suffixes with long non-high vowels (-le¯as- ‘-less’, -do¯m- ‘-dom’, -fæst- ‘-fast’ -ha¯d- ‘-hood’) are regularly scanned as lifts, e.g. wı´sdo`me heold ‘with wisdom ruled’ (Beo 1959b), of cı´ldha`de ‘from childhood’ (Elene 914a), but uninflected -do¯m ‘-dom’, -fæst ‘-fast’ are not ictic: word ond wı´sdom ‘word and wisdom’ (Andreas 569a), wı´sfæst wo´rdum ‘wise with words’ (Beo 626a). The placement of the word linearly in the verse is also significant: the suffixes -sum ‘-some’, -scipe ‘-ship’, -ian ‘-en’ (V) occupy ictic positions only in the coda of the verse; for full coverage see Fulk (1992: 197–216). Further indeterminacies arise from the difficulty of assigning suffixal status to morphological units which are also attested as independent words: do¯m, fæst, full, ha¯d, le¯as are separate lexical entries, and their semantic autonomy may be related to the preservation of stress. Additionally, as demonstrated in Minkova and Stockwell (2005), the full prosodic history of native suffixes has to refer to rhythmic factors linked to the types and frequency of derived words in the lexicon. Thus the equally productive OE suffixes -ha¯d and -do¯m would be expected to emerge either both with a full vowel, or both with a reduced vowel in Early Modern English. However, in Middle English close to 70% of the -dom derivatives followed a monosyllabic root (earldom, freedom, kingdom, wisdom), where stress-clash avoidance resulted in de-stressing of the suffix to [-dəm/-dm̩ ], while during the same period 73% of -hood derivatives had a disyllabic stem (bishophood, maidenhood, womanhood), allowing the preservation of secondary stress on the affix and raising of the long vowel to [uː] prior to 17th-century shortening to [-hʊd]. In summary, all factors identified above – syllable weight, vowel quality or quantity, semantic independence, and rhythmically induced changes – must be considered in the account of Old English suffixal stress. Derivational affixes often have their diachronic roots in independent words. Within the larger family of affixes, suffixes are cross-linguistically more likely to lose their

8 Linguistic Levels: Prosody

119

independent word status than prefixes, and therefore one would expect more word-like behavior from prefixes. Identifying the exact range of prefixes in Old English is a widely recognized problem, precisely because outside of the invariably bound forms: æf-, and-, be-, ed-, fær-, for-, ge-, mis-, etc., there is no clear-cut divide between prefixes such as ofer-, on-, wiþ-, ymb- and words. Moreover, the metrical treatment of both bound and free prefixal forms may differ for nouns and adjectives, where main stress aligns with the left edge of the whole word, leaving the root with secondary stress, and verbs and adverbs, where the main stress is kept on the root: compare swylce o´ncy`þðe ‘such grief’ (Beo 830a) to he onfe´ng hraþe ‘he seized quickly’ (Beo 748b) As argued in Minkova (2008b), both syllable weight and the grammatical nature of the base are determiners of stress in OE prefixation. Light prefixes behave like clitics; they do not form independent prosodic words and are consistently unstressed, while prefixes capable of forming independent prosodic words get stressed in accord with the word class of the derivative. NO

V, ADV Morphological Representation

Prefix ≠ PRWD

NO

Prefix = PRWD

YES

N, ADJ

Figure 8.2: Prefixal stress in Old English (adapted from Minkova 2008b: 36)

The principle of root-initial stress persists in compounding, where roots get their first syllables stressed as if they were independent words. Within the larger domain of compounds the stress to the left is primary, marking off the left boundary of the entire word, while compound-internal stresses are secondary. In the verse, the obligatory alliteration is consistently placed on the first stressed syllable onset, e.g. ofer hro´nra`de ‘over whaleroad’ (Beo 10a), wo´rolda`re forgeaf ‘worldly honor gave’ (Beo 17b). The second stressed syllable may alliterate only if the first stress alliterates too: wið þe´odþre`aum ‘against people’s calamity’ (Beo 178a), he´ardhı`cgende ‘hard-minded’ (Beo 394a). Such selfalliterating compounds are restricted to the on-verse by definition, since alliteration is prohibited from the second ictus in the off-verse. This restriction does not extend to affixal elements, thus la´ðlı`ce ‘hatefully’ is found at the right edge of the off-verse. The inference is clear: in þe´odþre`a ‘people-calamity’, he´ardhı`cgend ‘hard-minded’, both roots retain their semantic independence and strong prosodic prominence. Such forms present an analytical problem: they are interpretable both as compounds and as freely formed syntactic phrases. Another difficulty comes from the fact that many of the self-alliterating compounds in the Old English corpus are hapax legomena, single-instance forms: be´arngeby`rdo ‘child-bearing’, e´all-ı`ren ‘all of iron’, fe´n-fre`oðo ‘marsh refuge’, gry´re-ge`atwe ‘terrifying armor’, gry´re-gı`est ‘terrible visitor’, he´ardhı`cgend ‘hard-minded, he´oro-ho`cyht ‘savagely hooked’, hı´lde-hlæ` mm ‘battle crash’, swa´t-swa`ðu ‘bloody track’, sy´n-snæ` d ‘huge cut’, þe´odþre`a ‘people-calamity’ are some examples of such unique forms in Beowulf. The status of these constructions is an area deserving further inquiry; cf. Giegerich (2009) who shows that end-stress on noun-noun compounds in Present-day English: steel bridge, apple pie, Madison Avenue, may reflect the syntactic provenance of incompletely

120

II Linguistic Levels lexicalized forms, and that nominals of the form attribute-head can be both lexical and syntactic.

4.3 Old English phrasal stress In connected speech words are grouped together in larger prosodic constituents: clitic and phrasal groups. Clitic groups are made up of a fully stressed head-word and a clitic, an unstressed function word such as an article, a preposition, a conjunction, or a pronoun: the bo´ok, at scho´ol, etc., are clitic groups. Such groups behave in the same way in Old English: se rı´ca ‘the ruler’, on be´arme ‘on bosom’. The stressed words in a sentence are syntactically organized into noun-, verb-, adverb- and adjective phrases, coordinate phrases, and clauses. In Present-day English such syntactic units are right-prominent; i.e., the highest prominence is on the rightmost stressed syllable, while other stresses are secondary: ca`reful respo´nse, drı`ve ca´refully, ve`ry ca´reful, qu`ick and ca´reful, Be`n ca´res. Recovering the corresponding prosodic features of Old English from the existing textual records is challenging, and therefore the issue of phrasal prosody is underresearched and controversial. We know with certainty that the poets treated finite verbs differently from nouns; see Section 4.1 for Sievers’s Rule of Precedence. ´ l ala`mp ‘until time came’ (Beo 622b)) are Clause-final intransitive verbs (oþþæt sæ metrically weaker, but this convention may not match speech prosody. Clause-initial finite verbs may be skipped by the alliteration, e.g. Co`m þa to re´cede ‘came then to building’ (Beo 720a), Forge`af þa Be´owulfe ‘gave then to Beowulf’ (Beo 1020a), although the verbs also occupy ictic positions. Like the second elements of compounds, finite verbs may alliterate only if the other stressed word in the verse alliterates: we`ox under wo´lcnum ‘waxed under the clouds’ (Beo 8a), … wo´rdum we`old ‘… with words ruled’ (Beo 30a). Throughout the modern Continental West Germanic languages and in older Germanic, complements are stronger than their verbs, irrespective of the linear order. This typological comparison and the consistency with which complement-verb prosodic relations are observed in the verse – the complement always alliterates – is a good argument for projecting this prosodic contour to Old English. Prominence in noun and adjective phrases and coordinate phrases is not directly recoverable from the verse. In this area the rules of alliteration may be more of a handicap than help. The frequent assumption that obligatory alliteration on the first word in such phrases (lange hwile ‘a long while’ (Beo 16a), … hond ond rond ‘… hand and shield’ (Beo 656a)) translates directly into left prominence in the prosody is unfounded. The linear alliterative arrangement is a purely metrical convention, as can be seen from the freedom with which the poet switches components to fit the scheme in the line: Geata dryhten ‘lord of the Geats’ (Beo 2561b) vs. dryhten Geata (Beo 2901a); madma fela ‘of treasures many’ (Beo 36a) vs. fela missera ‘many of half-years’ (Beo 153b), manig oðerne ‘many other (men)’ (Beo 1860b) vs. æþeling manig ‘hero many’ (Beo 1112b). Some other facts also prompt skepticism about the link between alliteration and linguistic prominence: the default contour (no special focus) for noun phrases and coordinate phrases in the modern Germanic languages is right-prominent; for German see Selkirk (1984: 225–230). Right-hand prominence is attested also in copulative combinations of the type Anglo-Sa´xon, Native Cana´dian; they also typically align with

8 Linguistic Levels: Prosody syntactically coordinated phrases. The density of double alliteration in on-verses coextensive with noun + prepositional phrase (bat under beorge ‘boat under cliff’ [Beo 211a]) and in conjoined phrases (word ond wı´sdom ‘word and wisdom’ [Andreas 569a]), exceeds by far the overall 47% ratio of double alliteration in the on-verse, as reported in Hutcheson (1995: 112). This asymmetrical distribution precludes a linguistic bias towards left-prominence, but does not rule our equal or right-hand prominence. The absence of double alliteration in the off-verse can only be metrically determined; see Russom (1987: 114), Hutcheson (1995: 271). The most economical account that does not require a historical shift, therefore, is that the right-prominent prosodic contour of phrasal stress has been in the language since Old English times.

5 Middle English meter and prosody The Norman Conquest of 1066 coincides roughly with the abandonment of the structural principles of Classical Old English alliterative versification. The last surviving pieces of alliterative poetry that conform to the norms outlined in Section 4.1 are two short poems: Durham, c.1100, and The Grave, c.1150. Early Middle English compositions such as The Proverbs of Alfred, The Worcester Fragments of the Soul’s Address to the Body, The Bestiary, and Lagamon’s Brut, are “hybrid” compositions, mixing rhyme, alliteration and syllable-counting in often erratic patterns. Being grounded in the prosodic pattern of stress on the first root syllable, alliteration as a cohesive device survived, and a significant portion of the literary activity in the 14th century was channeled into the reinvention and composition of alliterative verse, culminating in masterpieces like Sir Gawain and the Green Knight, Piers Plowman, and The Alliterative Morte Arthure. At the same time new modes of versification based on rhyme, stress alternation, and syllable counting were gaining popularity. The relative rigidity of the new forms provides a solid basis for reconstructing the prosodic properties of Middle English.

5.1 Middle English metrical innovations: isosyllabicity, rhyme, iambic feet Verses of equal numbers of syllables – “isosyllabic verses” – are not uncommon in Old English poetry, but the recurrence was not structurally regulated; a verse could have from a minimum of four to fourteen syllables. Isosyllabism is an imported metrical feature in Middle English. Schemes based on the iteration of isosyllabic lines – the octosyllabic line, the septenarius, and, with Chaucer, the decasyllabic iambic pentameter – are at the core of Middle English verse composition. All of these forms allow an unstressed syllable after the last ictus; such “extrametrical” syllables are outside the metrical template and their presence or absence does not affect the isosyllabicity of the line. The lines were often linked in couplets or larger groups by end-rhymes. Rhyming did appear occasionally as an ornamentation in Old English verse, but the influence of Anglo-Norman made it the verse-line marker of choice. The third component of the new type of versification is the “iamb”, a binary sequence of a weak and a strong position (W S). Iambic feet could occur in Old English verse as a subset of a larger right-strong metrical type: the first three feet in ne le´of ne la´ð / bele´an mihte ‘not friend nor enemy / dissuade could’ (Beo 511; see Figure 8.1),

121

122

II Linguistic Levels happen to be iambs. In Middle English isosyllabic verse, however, the iamb became the dominant metrical foot. The earliest post-Conquest long non-alliterative compositions, The Owl and the Nightingale and The Ormulum, both late 12th century, are strictly iambic. Chaucer’s poetic works are also iambic. (All Middle English verse examples in this chapter are from Chaucer; abbreviations are from The Riverside Chaucer [Benson ed. 1987: 779].) The reconstruction of stress based on the new type of versification is most reliable line-medially. The interplay between prosody and meter at the two edges of the line is complicated by specific properties of the first and the last foot. The left edge of the line is rhythmically malleable, so that the expected W S / W S metrical cadence of the first two feet may be filled by: – a prosodic /s w w s/. An inverted foot S W followed by a regular W S foot is known as ´ n.der his belt … (GP 105). Triples may a “triple”: Tho´n.ked be Go´d … (WBT 5), U appear elsewhere in the line, but the probability of a triple decreases sharply from left to right. – a prosodic /s w s w/, resulting in trochaic inversion in both feet (Spo´ones and sto´oles and … (WBT 288), Ha´rdy he wa´s and … (GP 405). Occasionally whole lines can be trochaic: Ble´ssinge ha´lles, cha´mbres, kı´chenes, bo´ures (WBT 869). – by /Ø s w s/, where the W of the first foot is unfilled and the line is headless: Twe´nty bo´okes, clad in blak or reed (GP 294), Swe´re and ly´en, as a womman kan (WBT 228). The strong position in the rightmost foot of the line, where the rhyme is located, is metrically demanding in that it enforces prominence on the syllable filling that position. This is a verse convention, possibly observed in recitation, but it does not carry over into the prosody of speech. In Middle English rhyming practice, some suffixes appear to acquire metrically-induced secondary stress: bo´ldely`, dro´nkene`sse. The metrical strictness of the last strong position is such that it can even invert the prosodic contour of a native derived word by suppressing the primary stress and using the suffix as the single carrier of prominence, as in: … and make a thyng : … at his writy`ng (GP 325– 326), … in hir dro´nkene`sse : … that I took witne`sse (WBT 381–382). The convention is linguistically motivated only to the extent that derivational suffixes, but not grammatical suffixes, are subject to such metrical promotion. The fashion for iambic versification in Middle English was a cultural import from the Continent, but it could not have been adopted with such ease if the prosodic conditions had not been favorable. The gradual loss of final and inflections in Middle English resulted in a growing number of words realized as monosyllables, allowing flexibility in the prosody-to-meter matching. Increased use of prepositions compensating for inflectional loss created new W S clitic groups: at nı´ght, to re´st, with che´er. Prefixed verbs and adverbs supplied another set of natural iambic structures: befo´re, forgı´ve, perfo´rm, asle´ep. Phrasal stress continued to be right-strong; phrases made up of stressed monosyllables easily match an iambic foot: five bo´oks, tall me´n, full gla´d, God kno´ws. The poets also draw from an inventory of handy “fillers”, semantically dispensable monosyllabic words, e.g. and, now, for, some, and the grammatically redundant “pleonastic” this, that. Thus, although individual underived words retained root-initial stress, in connected speech metrical W S cadences were frequent and easy to construct; this permits an effortless “fit” between language and meter.

8 Linguistic Levels: Prosody Except for the metrical conventions at the line edges, iambic verse provides a reliable framework for reconstructing the stress of Middle English words on the basis of meter-to-prosody correspondences.

5.2 Native and non-native word stress in Middle English The continuing stability of the GSR, aligning primary stress with the left edge of all words and with the left edge of the root for prefixed verbs and adverbs, is easily demonstrated in verse, as in dro´ppyng, ho´uses, smo´ke, chı´dyng, wy´ves, ma´ken in Chaucer’s Wife of Bath’s Tale (WBT ), 278–279: Thow seyst that dro´ppyng ho´uses, and eek smo´ke, And chı´dyng wy´ves ma´ken men to flee

Words derived by suffixation also show the expected main stress on the leftmost root syllable, as kı´ngship, wı´sdom, wı´tness, ho´ly, blı´ssful: The ho´oly blı´sful … (GP 17). Again predictably, the first syllable of compounds is regularly aligned with a metrical S: … for a´ny le´checra`ft (KnT 2745), with wı´lde tho´nder-dy`nt … (WBT 276), to be´ me wa´rde-co`rs … (WBT 359). Compounds usually start in even (S) positions, but since both roots carry a degree of prominence, if the first part is monosyllabic, it can be ´ f clooth-ma´kyng she hadde placed in W, while the second root is in S, e.g. O ´ ´ swich an haunt (GP 447, headless), He was short-sholdred, brood, a thikke knarre (GP 549). Phrasal stress is not testable in iambic verse if there is a buffer weak syllable between the stressed syllables: of so´ndry fo´lk (GP 25), and ma´de fo´rward (GP 33), ty´me and spa´ce (GP 35). Monosyllabic adjectives in noun phrases do provide some corroboration for continuing right-prominent phrasal stress: ne po´lax, ne´ short kny´f (KnT 2544), Gret swe´ryng is (PardT 631), but the stress-alternating nature of the verse and the availability of optional -e and metrical slot-fillers obscure the picture. As argued in Minkova and Stockwell (1997), there is no good reason to posit any dramatic changes in the prosody of phrasal stress from Old English to Present-day English. Even if we assume a more level phrasal stress in Old English than in Present-day English, the right-hand prominence of Old French and Anglo-Norman would have contributed to the present contour. The introduction of a large non-native component into the vocabulary of Middle English is a central theme in any account of the history of English word stress. The non-native vocabulary of Old English never exceeded about 3%, while the portion of the Romance vocabulary at the end of the Middle English period is estimated at about 25%. Once again, attestations in verse provide our best test for the realization of loanwords in the spoken language. Thus we can safely posit initial stress on seson in: Bifil that in that se´son on a day (GP 19), And eek the lusty se´son of that May (KnT 2484). The word was first attested in English 1340–1370 (OED, Simpson [ed.] 2000–), roughly during Chaucer’s lifetime (1343–1400), yet out of the 15 times Chaucer uses the word in The Canterbury Tales and in Troilus and Criseyde, there is one single attestation of the word in rhyme (… thy declinacion : … tyme and his seson (FrT 1033– 1034)) where one could possibly posit right prominence. Such evidence suggests that a

123

124

II Linguistic Levels metrical promotion to seso`n is not different from the treatment of native writy`ng, witne`sse discussed in Section 5.1, i.e., there is no reason to differentiate between native and borrowed words at the line end. The 13th-century loanword country is used 45 times in CT and in Tr, 21 of which are in rhyme position and are realized as end stressed. Of the 24 line-internal attestations, however, there is not a single example of end stress on the word; they are all of the type illustrated by SumT 1710: A mersshy co´ntree called holdernesse. Such findings lead to a serious methodological amendment to the way of collecting verse data for prosodic reconstruction. As argued in Minkova (2000, 2006), the blanket assumption that the verse-final foot provides reliable information on stress is flawed. When we take rhyme position out of the picture, the rate of assimilation of the foreign prosodic contours to the native stem-initial prominence is significantly faster than has been previously acknowledged. The new Romance words coming into the language after the Conquest could be direct loans from the Classical languages, or they could be coming via Anglo-Norman or Old French. Latin (and Greek-via-Latin) disyllabic words would be stressed initially by default: a´xle, e´rgo, hy´mnal, he´rpes, mo´rtar, stu´por, o´nyx were all borrowed in Middle English. According to the Latin Stress Rule, in words of more than two syllables stress falls on the penultimate syllable if it is heavy, otherwise, on the antepenultimate syllable. The Latin Stress Rule in polysyllabic words is thus weight-sensitive, but since many early Latin borrowings lost their inflectional markers (-a, -(t)is, -us, -um, etc.), the picture was often obscured, thus ju´ncture < junctu¯ra, hu´man < huma¯nus. Anglo-Norman and Old French words were stressed depending on the weight of the final syllable: if heavy, the final syllable attracts stress: author, chaplain, jargon, merchant. Light final syllables are unstressed: able, chambre, piece. Since final syllables containing schwa are unstressable, in initially polysyllabic words like bataille, folye, justice, servise, visage the stress was on the penultimate, as in Latin. The extent to which weight-sensitivity at the right edge of the new words affected the prosody of Middle English has often been overestimated, mostly because of misinterpretation of the verse evidence; see Section 5.1. Both disyllabic and trisyllabic preRenaissance borrowings show a strong tendency of leftward stress-shifting, in conformity with the GSR, as in juncture, human, chaplain, merchant, battle, folly, novice, service. The leftward stress-shift disregards syllable weight; indeed in many cases the stress shifts leftwards from a heavy to a light syllable, as in chaplain, battle, folly, justice, novice. Table 8.1 shows the stress profiles of borrowed disyllabic simplex nouns and adjectives in alliterative and syllable-counting verse; the search ignores attestations in the final foot of iambic verse: Table 8.1: Romance loans in Middle English verse (from Minkova 2006: 114) Text

Tokens

Initial Stress

Sir Gawain and the Green Knight The Siege of Jerusalem Troilus and Criseyde Henryson’s poetry

283 87 266 151

276/97.5% 84/96.5% 223/84% 137/90.7%

Non-initial Stress 7/2.5% 3/3.4% 43/16% 14/9.3%

It is evident that the initial wave of borrowing did not upset the stem demarcation on the left. Verbs in which the prefixation is transparent behave like the native prefixed

8 Linguistic Levels: Prosody verbs discussed in Section 4.2.: Perfo´urme it out … (Tr III 417), ye na´t disco´vere me (MerT, 1942). Prefixed nouns and adjectives vary. Chaucer uses both initial and final stress on proverb, a word first recorded in his works (OED, Simpson [ed.] 2000–): Wel may that be a pro´verbe … (WBT 284), And therfore this prove´rbe is … (RvT 4319). Etymologically non-transparent prefixed nouns and adjectives tend to follow the native rule: Ben humble su´bgit … (Tr II 828), … in joye and pe´rfit heele (KnT 1271). The history of stress on prefixed loanwords in Middle English is an area which has not been fully researched yet – it is an inquiry that promises to throw light on the continuity and/or reintroduction of functional stress-shifts in English of the type a´bstract (N, ADJ)–abstra´ct (V); re´cord (N)–reco´rd (V). In iambic verse, polysyllabic words may be hard to fit to a metrical frame of alternating prominences. In Section 5.1 we noted how native suffixes appear to acquire metrically induced secondary stress: bo´ldely`, dro´nkene`sse, do´utele`es, ma´rtyrdo`m. The combination of dominant word-initial stress and the rhythmic preference for stress alternation in borrowed words produces a comparable effect in the new Romance vocabulary: the linguistic /w w s/ in the source language is realized in English as /s w s/: a`moro´use, cha`rite´e, la`xatı´f, o`pposı´t, o`riso´un, ple`ntevo´us, re`gio´un, only in this case it is probably the left edge of the word that carried secondary stress at first, judging from the strong preference for placement of such words in rhyme position: wro´oth was she´ : cha`rite´e (GP 451–452), whı´t : o`pposı´t (KnT 1893–1894), ho´us : ple`ntevo´us (GP 343–344), ado´un : re`gio´un (KnT 2081–2082). The switch from word-initial secondary to primary stress on such trisyllabic words probably started during Middle English, but the precise dating is not recoverable from iambic verse, where both primary and secondary stresses may fill S-positions. The preservation of some degree of stress on the final syllable in Romance loans comfortably beyond Middle English is well documented in Early Modern English, see Dobson (1957: 830–860). The placement of secondary stress on the initial syllable of four-syllable words with an unstressable final syllable: dı`gestı´ble : Bı´ble (GP 437–438) sa`crifı´ce : wı´se (KnT 2369– 2370), dy´e : of bı`gamy´e (WBT 85–86) is also attributable to the principle of rhythmic stress alternation enhanced by the native left-edge prominence pattern. If the final syllable is stressable, the additional stress appears on the second syllable to the left: relı`gio´un : to´un (GP 477–478), coma`ndeme´nt : yse´nt (KnT 2869–2870). The Middle English stress alternation and the eventual demotion and loss of the original primary stress in the foreign vocabulary was attributed to the school pronunciation of Latin in Middle English by Danielsson (1948: 26–29, 39–54) who used the term countertonic accentuation to describe the shift of e.g. post-Classical Latin melancholı´a (1375) to me´lancho`ly, in line with the native model of ma´idenho`d, dro´nkene`sse.

6 Post-Middle English prosodic innovations The rise of literacy in Early Modern English was accompanied by a parallel rapid expansion of the lexicon. Barber (1997: 219–220) estimates that as many as 95 new words were recorded in English during each decade between 1500 and 1700; his counts are based on sampling entries in the OED. This exceeds by far the rate of borrowing in Middle English, which he estimates at 17 new words per decade, using the same methodology. Two-thirds of the new forms in Early Modern English were based on already recorded roots and affixes and about one third were straight borrowings. The large

125

126

II Linguistic Levels majority of these words were coined or adopted by English speakers who were proficient in Latin and Greek and who would therefore automatically apply the Latin Stress Rule to the novel “English” forms: abla´tion, cathe´dral, demo´cracy, mea´nder, te´rminus. The density of these forms and the shared literate understanding of their prosody gave rise to a new, parallel model of stress in English, which is weight sensitive, and ´ regon (1765, possibly Connecticut pidgin which can apply to new words such as O Algonquian, OED, Simpson [ed.] 2000–), kaı´nga ‘village’ (Maori, 1820), palachı´nka ‘pancake’ (Slavic, 1884). The tenacity of the GSR continued during the Early Modern English period in spite of the unprecedented influx of foreign loans, however. Consolidation of the primary stress on the initial syllable of the stem went beyond the disyllabic shifts recorded in Section 5.2 and affected trisyllabic nouns and adjectives: a´morous, cha´rity, la´xative, o´pposite, o´rison, ple´nteous, re´gion have changed their Chaucerian pronunciation in accord with the GSR, similarly ´ınfantry, me´rcury, o´rient, ca´lendar, ge´nial. Stress shift to the initial syllable often proceeds in spite of the etymological heaviness of the penultimate syllable, as in the early loans a´morous, fo´rtunate, ´ınfantry, ´ınterval, o´rient and many post-Middle English forms such as ve´rtebra (1615), ta´lisman (1638), sy´nergy (1660), Ca´vendish (1839), ba´dminton (1845), a´llergy (1911), bo´ondo`ggle (1935). The emerging picture is complexly layered: the prosody of native words follows the Old English left-alignment of word or stem with the main stress. The non-native vocabulary displays hybrid patterns, and no single model covers all realizations without exceptions, so we can only define strong tendencies. New words can fall in with the native left-strong Germanic model, or they can follow a weight-sensitive model whereby stress in non-derived words is assigned by word class and by syllable weight. Verbs with heavy final syllables are generally end-stressed, e.g. para´de, deny´, mainta´in, oblı´ge, prote´ct. Nouns may be stressed depending on the weight of the penultimate syllable in accord with the Latin Stress Rule: age´nda, ca´nopy, horı´zon, ´ınfidel, Toro´nto. Although the considerable overlap between the patterns noted in Section 5.2. for disyllabic nouns and adjectives continues, end-stressed nouns like aby´ss, baro´que, caba´l, cana´l, dure´ss, elı´te, mala´ise, ravı´ne do occur. The extent to which such words retain their prosodic “foreignness” may vary in British and American English. Table 8.2 shows some examples with first attestation dates from the OED; some of these are simply “majority” pronunciations in variation with the alternative pronunciation. Table 8.2: Stress differences between American and British English American English

Date

British English

´ınquiry po´lice (also polı´ce) fru´strate premı´er mo´ustache debrı´s cafe´ gara´ge

(1440) (1450) (1447) (1500) (1585) (1708) (1802) (1902)

inquı´ry polı´ce frustra´te pre´mier mousta´che de´bris ca´fe ga´rage

8 Linguistic Levels: Prosody The hybridity of the Present-day English stress system is also evident in the variability of stress patterns within the last century. Bauer (1994: 96–103) records items which have undergone a recent shift to penultimate stress, e.g. a´bdomen, a´cumen, a´nchovy, e´tiquette, moly´bdenum, pre´cedence, qua´ndary, se´cretive, so´norous, va´gary. He notes a further complicating factor: stress placement in derived words can ignore the nature of the suffix and preserve the prosody of a pre-existing and frequently used base, thus ca´pital, prefe´r are the bases which trigger the change of old capı´talist to current ca´pitalist, and of old pre´ferable to prefe´rable. As noted in Section 4.2, suffixation in Old English was never associated with mainstress reduction; the highest level of prominence for derivational suffixes was secondary stress. The adoption of a large number of foreign affixed words along with their prosodic contours changed this situation. Present-day English suffixes can attract main stress themselves or they can trigger the placement of main stress on one or two syllables to the left of the suffix. Among the suffixes attracting primary stress and reducing the original stress of the base to secondary stress are: -ette (1849), as in ma`jore´tte, -een (1551) as in ve`lvete´en, -ese (1898), as in jo`urnale´se, -eer (1704) as in mo`untaine´er. Final main stress appears also on word endings that may not be etymologically productive suffixes: -ade as in le`mona´de, -ique as in boutı´que, -oo as in ka`ngaro´o. Main stress usually falls on the syllable immediately preceding the suffixes -ic, -id, -ion, -ity/-ety: nume´ric, caro´tid, rebe´llion, tranquı´lity. Among the borrowed suffixes that place the main stress on the antepenultimate syllable of the derived word are -acy, -ast, -ose, -tude: demo´cracy, ico´noclast, co´matose, simı´litude. The antepenultimate is stressed also in combining forms such as -o´logy, -o´sophy, -o´graphy, -o´latry, -o´cracy etc. These new patterns of stress-assignment extend to native roots under foreign suffixation as in Icela´ndic (1674), weathero´logy (1823), speedo´meter (1904), Cha`pline´sque (1921). The placement of stress in derived words has been the subject of much linguistic research. A good descriptive coverage is found in Fudge (1984); the analytical problems are addressed in Giegerich (1999). Another innovation in the post-Renaissance period is the growing productivity of functional stress-shifting in homographic pairs, the a´ddict (N)–addı´ct (V), pre´sent (N)– prese´nt (V), po´lice (N)–polı´ce (V) model, where the shift of stress from one part of speech to another is no longer a matter of prefixation, as in the native shifts in u´pset (N), o´verhang (N). The new stress-shifts do not require compositionality; on the other hand, they are directional (right-to-left) and subject to syllabic and segmental restrictions on the base, not applicable to the native pairs (Minkova 2009). In conclusion, the prosody of Present-day English presents a mixture of word-stress patterns, some inherited from Old English, some introduced in Early Modern English. What we share with Old English is an uninterrupted line of left-edge marking of compounds, unstressable function words and head-prominence in clitic groups, and righthand phrasal prominence. Many relevant details in the prosodic history of English remain under-researched: we lack good documentation of the prosodic behavior of borrowings in Middle English and we still need to evaluate the relevance of competing factors such as phonological composition, frequency, morphological marking and transparency, social prestige, spelling. The relationship between innovations in verse form and prosodic innovations is also of considerable linguistic and cultural interest. Other areas that invite further inquiry are the prosodic patterns in the regional and ethnic varieties of English, and the contact-induced changes in the English spoken in countries where it is an official second language.

127

128

II Linguistic Levels

7 References Barber, Charles. 1997. Early Modern English. Edinburgh: Edinburgh University Press. Bauer, Laurie. 1994. Watching English Change. London: Longman. Benson, Larry D. (ed.). 1987. The Riverside Chaucer. 3rd edn. Boston: Houghton Mifflin. Bredehoft, Thomas. 2005. Early English Metre. Toronto: University of Toronto Press. Danielsson, Bror. 1948. Studies on the Accentuation of Polysyllabic Latin, Greek, and Romance Loan-Words in English: with special reference to those ending in -able, -ate, -ator, -ible, -ic, -ical, and -ize. Stockholm: Almqvist and Wiksell. Dobson, Eric J. 1957. English Pronunciation 1500–1700. Vol. II. Oxford: Clarendon Press. Fudge, Eric. 1984. English Word-Stress. London: George Allen & Unwin. Fulk, Robert D. 1992. A History of Old English Meter. Philadelphia: University of Pennsylvania Press. Getty, Michael. 2002. The Metre of Beowulf. A Constraint-Based Approach. Berlin/New York: Mouton de Gruyter. Giegerich, Heinz. 1999. Lexical Strata in English. Morphological Causes, Phonological Effects. Cambridge: Cambridge University Press. Giegerich, Heinz. 2009. Compounding and Lexicalism. In: Rochelle Lieber and Pavol Sˇtekauer (eds.), Handbook of Compounding, 178–200. Oxford: Oxford University Press. Hutcheson, Bellenden Rand. 1995. Old English Poetic Metre. Cambridge: D. S. Brewer. Minkova, Donka. 2000. Middle English prosodic innovations and their testability in verse. In: Irma Taavitsainen, Terttu Nevalainen, Pa¨ivi Pahta, and Matti Rissanen (eds.), Placing Middle English in Context, 431–461. Berlin/New York: Mouton de Gruyter. Minkova, Donka. 2006. Old and Middle English Prosody. In: Ans van Kemenade and Bettelou Los (eds.), The Handbook of the History of English, 95–125. Oxford: Blackwell. Minkova, Donka. 2008a. Review of Thomas Bredehoft, Early English Metre. Speculum 83(3): 673–675. Minkova, Donka. 2008b. Prefixation and stress in Old English. Word Structure 1(1): 21–52. Minkova, Donka. 2009. Continuity or re-invention in functional stress-shifting. Paper presented at ICEHL 15, Munich. Minkova, Donka and Robert Stockwell. 1997. Against the emergence of the nuclear stress rule in Middle English. In: Jacek Fisiak (ed.), Studies in Middle English, 301–335. Berlin/New York: Mouton de Gruyter. Minkova, Donka and Robert Stockwell. 2005. Clash avoidance in morphologically derived words in Middle English. (Why [-hʊd] but [-dəm])? In: Nikolaus Ritt and Herbert Schendl (eds.), Rethinking Middle English. Linguistic and Literary Approaches, 263–280. Bern: Peter Lang. Russom, Geoffrey. 1987. Old English Meter and Linguistic Theory. Cambridge: Cambridge University Press. Selkirk, Elisabeth O. 1984. Phonology and Syntax: The Relation Between Sound and Structure. Cambridge, MA: MIT Press. Sievers, Eduard. 1893. Altgermanische Metrik. Halle: Niemeyer. Simpson, John (ed.). 2000–. The Oxford English Dictionary. 3rd edn. online. Oxford University Press. www.oed.com/ Stockwell, Robert and Donka Minkova. 1997. Prosody. In: Robert Bjork and John Niles (eds.), A Beowulf Handbook, 55–85. Lincoln: University of Nebraska Press.

Donka Minkova, Los Angeles (USA)

9 Linguistic Levels: Morphology

129

9 Linguistic Levels: Morphology 1 2 3 4 5

Introduction Morphological typology Modern English History References

Abstract Modern English morphology is the result of a long-range typological restructuring, triggered by phonological changes in connection with the emergence of the Germanic language family, leading to an erosion of unstressed final syllables. As a result, the originally root-based morphology became stem-based and finally word-based. Also morphology was originally characterized by pervasive phonologically conditioned morphophonemic alternations, which gradually became morphologically conditioned, because of phonological changes. This was replaced by a simplified system with base invariancy and phonologically conditioned alternations of inflectional endings as a default case characterizing the regular inflection of nouns, verbs and adjectives. The irregular patterns continue properties of the original system and can be interpreted as stem-based with morphologically conditioned alternations of the base form. This is also true of many non-native word-formation patterns, which have been borrowed from stem-based languages such as French, Latin or Greek and have re-introduced base alternation into English derivational morphology.

1 Introduction Modern English morphology is the result of several millennia of linguistic change, which has transformed its ancestor, Indo-European morphology, into a completely different morphological type. These changes often left relics behind, which like tombstones commemorate earlier stages, as e.g. irregular verbs of the type write : wrote : written, which go back to Indo-European ablaut alternations, or irregular noun plurals like mouse : mice, goose : geese, oxen, which reflect old Indo-Germanic and Germanic inflectional classes. Similarly, the stress alternation between verbal o`verflo´w and nominal o´verflo`w, etc. can be traced back to a combination of the Germanic innovation of initial stress and syntactically governed stress distribution in Proto-Germanic (cf. Minkova 2008a, 2008b), eventually acting as a landing site for Romance loans like reco´rd V and re´cord N. This new Germanic stress system is also held responsible for the loss of unstressed syllables, many of which were carriers of morphological information, the result being a language with little inflection left. On the other hand, there are also typological innovations like the vowel and/or consonant alternations in sane : sanity, serene : serenity, Japa´n : Ja`pane´se, hı´story : histo´ric : hı`storı´city, ele´ctric : ele`ctrı´city, close : closure, resulting from the integration of non-native (Romance, Latin and Neo-Latin) wordformation patterns into English with a concomitant variable stress system, which reverses the original typological drift towards a non-alternating relation between bases and derivatives and is reminiscent of Indo-European, where variable stress/accent Alex Bergs and Laurel Brinton 2012, Historical Linguistics of English (HSK 34.1), de Gruyter, 129–147

130

II Linguistic Levels produced variable vowel quality/quantity (ablaut). It is such long-range developments and their typological consequences that will be dealt with in the following. The existing literature, e.g. Brunner (1960–62), Wełna (1996), the respective chapters in Hogg (1992), Blake (1992), or Hogg and Denison (2006), pay only cursory attention to these typological changes. Moreover, they reflect the Neo-grammarian tradition of looking backwards to the Indo-European and Germanic morphological structures, interpreting later patterns in terms of older structures as having undergone losses rather than as acquiring structural innovations. As a consequence, morphological restructuring is often not topicalized enough or assumed to have occurred much later than it actually did. A good example is Brunner’s (1962: 3–10) treatment of Old English nominal inflection. He makes a point of describing this in terms of Indo-European inflectional classes involving stem-formatives like IE -o- (= Grmc. -a-), IE -o¯- (= Grmc. -a¯-), etc., arguing that reference to the Indo-European categories rather than the Germanic ones is preferable in order to emphasize the fact that the Germanic declinations correspond exactly to the declinations of other Indo-European languages. The synchronic relevance of classifying Old English nouns in terms of Indo-European o- or a¯-stems, etc. is simply taken for granted and never questioned, which obscures a profound typological change between Indo-European and Old English, viz. the shift from root-based to stem-based and even word-based morphology and the demise of the stem-formatives as a functional category. Even though this approach allows for restructurings, it would tend to place them at a later period than when they actually occurred. What is needed, therefore, is a Janus-like approach: this should take into account both where a synchronic linguistic stage comes from (= retrospective), but also in which direction it is moving (= prospective).

2 Morphological typology 2.1 Traditional typology Morphological typology as introduced by the Schlegels (Schlegel, Friedrich 1808; Schlegel, August W. 1818) and Humboldt (1827–29) covered only inflectional morphology and was based on two overlapping scales with one parameter each. The first parameter resulted in the distinction between analytic languages, where grammatical functions are expressed word-externally by prepositions, auxiliaries, etc., and synthetic languages, where grammatical functions are expressed word-internally by inflection, incorporation, etc. The second parameter was based on how grammatical functions are represented formally and resulted in the distinction between isolating, agglutinating, inflectional, and incorporating languages. In isolating languages, grammatical functions are expressed word-externally; i.e. this type coincides with the analytic one, and languages belonging to this type have no inflectional morphology. The other three are sub-types of the synthetic type. In agglutinating languages, there is a one-to-one correspondence between a grammatical category and its exponent, as, e.g., in Turkish, Finnish or in English, where number and case are expressed consecutively, if the plural is irregular, cf. (1): (1)

ox en s

geese s

PL GEN PL

GEN

boy s

Ø

PL GEN

9 Linguistic Levels: Morphology In inflectional languages, one morphological exponent represents more than one grammatical category, functioning as a portmanteau morph, cf. Lt. am-o¯ ‘I love’ (-o¯ = 1P [person] SG [number] PRES [tense] IND [mood] ACT [voice]), or OE cyning-as (-as = PL [number] + NOM/ACC [case]). In incorporating languages, grammatical functions such as subject, object, adverbial complement are integrated into the predicate itself together with their nominal carriers; this might be postulated for recent English formations such as machine-translate, thought-read, flight-test, etc. (cf. Section 3.2.2.)

2.2 Extended typology These parameters only partially characterize the overall gestalt of the morphology of a language and therefore have to be supplemented by further aspects, which will lead to a more differentiated picture. The following additional parameters have proved useful (cf. Kastovsky 1997, 2006a, 2006c): – – – – – –

morphological status of the input to the morphological processes (word, stem, root) number and status of inflectional classes formal representation of inflectional and derivational markers status and function of morphophonemic/allomorphic alternations position of affixes/position of the head existence and status of morphological levels (e.g. native vs. foreign)

The first parameter requires a comment. I will adopt the distinction between “lexeme” (= dictionary entry), “word-form” (inflectional form of a lexeme), and “word” (actual representation of a lexeme via word-form in an utterance as independent syntactic element or free form) from Matthews (1974: 20–26) and Lyons (1977: 18–25). “Word-formation” (“derivational morphology”) is therefore actually “lexeme-formation”, but I will retain the former term for convenience’s sake. What is crucial is the status of the lexeme representation which acts as input to inflectional and derivational processes, the “base form”. This is defined as that lexeme representation from which all word-forms and the result of word-formational processes can be derived. Thus, word-formation and inflectional morphology are interdependent: the input to word-formation processes, and the demarcation between inflection and derivation, depends on the typological status of inflection. For this, the following distinction is suggested: a) word-based morphology: The base form can function as a word (free form) in an utterance without the addition of additional morphological (inflectional or derivational) material, e.g. ModE cat(-s), cheat(-ed), beat(-ing), sleep(-er). b) stem-based morphology: The base form does not occur as an independent word, but requires additional inflectional and/or derivational morphological material in order to function as a word. It is a bound form (= stem), cf. OE luf-(-ian, -ast, -od-e, etc.), luf-estr-(-e) ‘female lover’, Grmc. *dag-(-az) ‘day, NOM SG’, ModE scient-(-ist) vs. science, dramat-(-ic) vs. drama, astr-o-naut, tele-pathy; thus luf-, luf-estr-, *dag-, dramat-, astr-, -naut, tele-, -pathy are stems. c) root-based morphology: Here the input to morphological processes is even more abstract and requires additional morphological material to become a stem, to which the genuinely inflectional endings can be added in order to produce a word.

131

132

II Linguistic Levels Such roots can either be affiliated to a particular word-class, or they can be word-class neutral. In this case the word-class affiliation is added by a word-formative process, cf. IE roots like *wVr- ‘bend, turn’ (cf. Lt. uer-t-ere ‘turn’, OE weor-þ-an ‘become’, wyr-m ‘worm’, etc.), with V standing for the ablaut vowel, whose shape is determined by the morphological process in question. Whether Indo-European roots were word-class specific or neutral is not quite clear, but not relevant for our purposes. Terminology varies considerably in this respect, cf., e.g., Bauer (1983: 20–21; 1992: 252– 253), Giegerich (1999) or Huddleston and Pullum (2002: 1624–1625). The terminology used here is geared to the history of Indo-European morphology, with its shift from a root-based to a stem-based and a (partly) word-based system. I will therefore use the term “stem” for what some linguists call “root” also with reference to English, i.e. for a lexical element which is bound and which can only occur as a word with additional morphological material, as, e.g., scient-, dramat-, tele-, -pathy. The term “root” will be restricted to the Indo-European period and is not relevant for the Indo-European daughter languages. The ultimate starting point of English was a root-based morphology (IndoEuropean), which became stem-based in the transition to Germanic. In the transition from Germanic to Old English, inflection became partly word-based, and this eventually became the dominant typological trait of Modern English.

3 Modern English 3.1 Inflectional morphology 3.1.1 Inflectional morphology vs. word-formation In Modern English, there is a fairly neat division between inflectional morphology and word-formation, except for adverb formation in -ly, which is ambivalent. Since -ly produces a change of word-class (ADJ > ADV), it is usually treated as derivational. But on account of the complementary distribution of adverbs and adjectives as in (2) (2)

He smokes heavily : He is a heavy smoker.

it could also be treated as inflectional, parallel to nominal case inflection, cf. Hockett (1958: 211). This is probably why Marchand (1969) did not include adverb formation in his handbook.

3.1.2 Regular vs. irregular inflectional morphology Inflectional morphology of verbs and nouns is based on a default system consisting of two inflectional classes: regular and irregular. Regular inflection is fully predictable. It is word-based, base-invariant, and the morphophonemic alternations of the inflectional endings are phonologically conditioned: 3P SG, plural and genitive have the allomorphs: /z/ (underlying form and default), e.g. kids /kɪdz/; /ɪz/ (vowel insertion after base-final coronal sibilants), e.g. kisses /kɪsɪz/; and /s/ (devoicing after base-final voiceless consonants), e.g. bets /bets/. Preterit and past participle have the allomorphs: /d/ (underlying and default), e.g. loved /lʌvd/; /ɪd/ (vowel insertion after base-final alveolar stops), e.g. hated /heɪtɪd/; and devoicing after base-final voiceless consonants, e.g. kissed /kɪst/.

9 Linguistic Levels: Morphology Irregular verbs and nouns can be grouped into various sub-classes (classification in the grammars varies considerably); they have either variable bases or phonologically unpredictable inflectional allomorphs, or both, and are relics of historically earlier stages, cf. (3a, b): (3)

a. keep /ki:p/ : kep-t /kep-t/ (variable base, regular inflectional allomorph), deal /di:l/: deal-t /del-t/ (variable base and unpredictable inflectional allomorph /t/) shut : shut-Ø (invariable base, unpredictable inflectional allomorph Ø), write : wrote : writt-en, sing : sang : sung, bleed : bled : bled (unpredictable inflectional allomorphs, viz. replacives, -en); b. leaf /li:f/ : leav-es /li:v-z/ (variable base, regular inflectional allomorph), ox : ox-en (invariable base, unpredictable inflectional allomorph), mouse : mice, child : child-ren, fung-us : fung-i (unpredictable inflectional allomorphs)

The status of /kep/, /del/, /li:v/ /ʧɪld/ is arguable. On the one hand, these are allomorphs of the lexeme representations /ki:p/, /di:l/, /li:f/, /ʧaɪld/, which are words. But since they only occur together with inflectional morphemes, these allomorphs might be regarded as stems, i.e. these lexemes have both a word-based and a stem-based morphological representation. This is similar to the derivational pattern science : scient-ist, drama : dramat-ic, where the lexical base is represented as a word, but also has an allomorph which occurs in derivatives and which has to be regarded as a stem. Write : wrote : written, or mouse : mice are word-based with a replacive morph representing the inflectional morpheme, which in the case of writt-en might be interpreted as discontinuous, consisting of the replacive /aɪ →ɪ/ and the suffix -en. In this case, the morphophonemic alternation is morphologically relevant and expresses a morphological contrast, whereas in the cases of keep : kept, leaf : leaves, etc. the morphophonemic alternation is non-functional. Synchronic classifications do not always reflect historical developments. Thus, Modern English sing : sang, write : wrote and bleed : bled all have replacives as tense and participle markers, but with sing and write the replacive reflects an Indo-European ablaut alternation, whereas with bleed : bled the alternation has the same origin as that occurring in keep : kept, deal : dealt, viz. shortening of long vowels before certain consonant clusters in late Old English. But with bleed : bled the alternation has become morphologically distinctive because of the lack of any other overt exponent of tense/participle, whereas the alternation in keep : kept might be treated as morphologically non-distinctive. On the other hand, bleed : bled and shut : shut share the same history with regard to the loss of the original past tense/past participle ending, cf. OE ble¯d-d-e, ge-ble¯d-(ed), scyt-t-e, ge-scyt-t-(ed): the ending was lost (= replaced by zero, cf. Kastovsky 1980), when gemination was lost in Middle English; but with bleed : bled, a formal contrast had arisen thanks to vowel shortening in the preterit/ past participle, which was morphologized, whereas with shut no formal contrast remained, which results in a zero allomorph for preterit and past participle.

3.1.3 Inflectional morphology of nouns For nouns, the following morphological categories are relevant: number, case, class. Gender does not figure as an inflectional category but is determined by the nature of the referent (= natural gender), in contradistinction to Old English, where it was a

133

134

II Linguistic Levels relevant grammatical category in the organisation of inflectional paradigms. For its loss, cf. Kastovsky (1999). The category of number is primary; the category of case (genitive) is secondary and dissociated from number, i.e. has a separate exponent, in contradistinction to Old English, where the two categories were fused, cf. cyning-as NOM/ACC PL, cyning-um DAT PL. In view of structures such as the Queen of England’s castles, where the genitive attaches to the whole NP rather than to its head Queen, the genitive might be regarded as a clitic, and not as an inflection. The category of class relates to the formal expression of plural, which involves a default system: regular plurals with base invariance and phonologically determined plural allomorphs, and irregular plurals, which deviate from this, cf. (3b).

3.1.4 Inflectional morphology of verbs For verbs, the following inflectional categories are relevant: tense, number, person, mood, finiteness, class. On the basis of the degree to which these categories are exploited, the verb system can be subclassified into three subcategories, each with distinct properties: 1) modal verbs, 2) lexical verbs, 3) the copula be. Modal verbs are the continuation of the Germanic preterit presents, which have undergone a shift towards a separate category in Middle and Early Modern English. They only exploit the tense dimension, i.e. they have no infinitive or participle forms, no person/number marking and no mood contrast. The verbs in question are can, may, shall, will, must, ought to and, to a certain extent, dare, need. Semantically speaking, the tense contrast is problematic, since the morphological opposition can : could, may : might, etc. most of the time does not really signal ‘past’ vs. ‘non-past’, but rather a different degree of modality. In view of their specific morphological and semantic properties, the modal verbs are treated as a separate category. Lexical verbs are characterized by the contrast between an unmarked base form (which also functions as infinitive and non-preterit), the 3P SG, and the marked preterit, present participle, and past participle. The copula be has the richest morphology with a person/number contrast in the preterit, and person distinctions marking first, second and third person in the non-past. With the lexical verbs, person and number are expressed jointly, but they only have a person/number marker in the 3P SG non-preterit, which might better be regarded as an agreement morpheme rather than a person/number morpheme proper. The morphological dimension of mood involves the categories indicative, subjunctive, and imperative, where the morphological marking is reduced to the absence of the agreement morpheme, whereas in Old English there were full-fledged separate paradigms. Inflectional class is an inherent property of the verb, determines how it forms the preterit and the second participle, and is based on a default system. Regular verbs are the default case: they are base-invariant and select the phonologically conditioned allomorphs /d/, /t/ and /ɪd/. The irregular verbs (c.200) involve a modification of the base form (allomorphy), or select a non-phonologically conditioned representation of the past tense and past participle morphemes or both.

9 Linguistic Levels: Morphology

3.2 Word-formation Modern English word-formation includes the following processes: compounding, affixation (prefixation, suffixation and zero-derivation/conversion), clipping, blending, acronyms, sound symbolism; for a comprehensive, somewhat dated but still relevant survey cf. Marchand (1969). It can be argued (Kastovsky 2009b) that most of these word-formation processes are prototypical patterns arranged on a scale of progressively less independent constituents, cf. (4): (4)

word compounding > stem compounding > clipping and clipping compounds > affixation (word-/stem-based) > blending (splinters)) > acronyms

3.2.1 Compounding English compounds are conventionally defined as combinations of full words resulting in a new lexical unit, e.g. bird cage, flatfish, earthquake, house-keeping, law-breaker, sunrise, writing table, dance hall; color-blind, icy-cold, heart-breaking, easy-going, manmade, etc. These illustrate nominal and adjectival compounds, most of which have been in the language since pre-Old English. The status of verbal compounds is problematic. Marchand (1969: 96) accepts as compounds only verbs with a locative particle like overdo, underestimate, and outdo, but with some reservations because of semantic problems with the first constituents, which make them look like semi-prefixes. These continue Old English formations with inseparable particles originally having a purely locative meaning, which was later extended metaphorically to ‘excess’, ‘deficiency’. For the type stagemanage, playact, spoonfeed, newcreate, Marchand (1969: 104) assumes backderivation from nouns such as stage-manager, playacting, spoonfed, newcreated, since the Germanic languages never had a genuinely productive verbal compound pattern. But in the last decades, the situation has changed, and a verbal compound pattern based on incorporation is apparently taking over, which no longer needs a nominal base as a source, cf. quick-march, new-form, slow-step, slow-kill, dark-adapt, quick-frost, surekill, hard-learn, quick-check; quantum-teleport, knee-jerk, machine-translate, thoughtread, flight-test, depth-bomb, finger-tap, hull-walk, flash-vaporize, flash-freeze, randomly culled from science fiction novels published in the last 15 years. Even V + V compounds are catching on, cf. think-hiss, whisper-hiss, touch-share, wobble-hop, glide-walk, skimglide, bend-swivel (in the form bent-swivelled), strip-search, hop-step, strip-mine from the same kind of source (cf. Wald and Besserman 2002). The compound principle as such is old and goes back to Indo-European, probably even to its oldest layer before the development of certain inflectional and syntactic categories like full relative clauses (cf. Kastovsky 2009a). Therefore, at this stage compounds do not seem to have consisted of full-fledged words, but of roots and stems (with stemformatives). At a later stage syntactic groups consisting of full-fledged words with case marking of the first member expressing its semantic-syntactic relation to the second member seem to have coalesced to compound patterns, carrying along their original first-member inflectional endings. These eventually lost their syntactic function and became mere morphological linking elements like the original stem-formatives, cf. ModE driver’s seat, bull’s eye or Ger. Universita¨tsbibliothek, Kindergarten, Frauenkirche.

135

136

II Linguistic Levels

3.2.2 Stem-compounds Formations such as astr-o-naut, Mars-naut, astr-o-physics, agr-o-chemical, agr-i-culture; hyper-active, hyper-aemia, omn-i-pre´sent, omn-ı´-scient, hepat-itis, wakeup-itis, astr-ology, ex-o-bio-logy, megal-o-mania, star-mania belong to the non-native level of English word-formation. Their constituents have been dubbed “combining forms” in the Oxford English Dictionary for want of a better term at the time of its inception, but there has never been a satisfactory definition of this heterogeneous category (cf., e.g., Prc´ic´ 2005, 2007, Kastovsky 2009b). This term is misleading, covering a number of different phenomena, and should therefore be discarded. Instead, we should recognize the existence of stem-compounds in English, i.e. compounds whose constituents are lexeme representations without word-status, a pattern inherited from languages like Latin or Greek. Formations such as astr-o-naut, agr-i-culture, omn-i-scient, etc. thus contain stems, and the middle element is – historically speaking – a stem-formative, but today functions as a linking element as is the case with -s- in word-compounds. Elements like -logy also go back to such stems, but have developed into suffixes, just as OE do¯m in kingdom, stardom or OE ha¯d in childhood, statehood, where do¯m, ha¯d had originally been independent words which developed into suffixes. Similarly, elements like ante-, anti-, auto-, mono-, etc., originally independent lexemes in the source languages, have developed into prefixes. And the elements Ameri- or Euro- in Euro-City can be regarded as clipped stems (< Ameri(ca), Euro(pa)), which can also be used in compounding.

3.2.3 Affixation Let me now turn to affixation. There are three basic sources for affixes: a) bleaching and grammaticalization of lexemes as compound constituents, as in -dom, -hood, or for-, be- and many other prefixes going back to locative particles; b) borrowing from other languages, as was the case with the majority of ModE prefixes (e.g. ante-, auto-, co-, counter-, de-, dis-, ex-, hyper-, in-, mono-, non-, post-, pre-, re-, sub-, and many others) and also suffixes (e.g. -able, -age, -al, -ance, -ation, -ery, -ess, -ify, -ism, -ist, -ity, -ize, -ment, -ory, and many others); c) secretion, as in -teria caketeria, washeteria), -gate < Watergate (> Irangate, zippergate), -burger < hamburger (> beef-burger > burger), etc. The latter two sources, borrowing and secretion, are closely related, because the borrowing of a non-native affix also involves secretion, since the borrowed affixal element has to be recognized as a separate morphological unit before it can become productive. The demarcation between compounding and affixation is fluid because of the diachronic shift from lexeme to affix, resulting in a bridge category “semi-affix”, e.g. -monger, -like, -wise (Marchand 1969: 356–358), or out-, over-, under- in their metaphorical meaning. On the other hand, affixes might also turn into lexemes, cf. burger, mini, extra.

3.2.4 Zero-derivation There is one word-formation process whose status is controversial, but which I interpret as a sub-type of suffixation, i.e. derivation by means of a zero morpheme, as in salt N >

9 Linguistic Levels: Morphology salt V, clean ADJ > clean V, cheat V > cheat N, walk V > walk. Several approaches have been suggested to account for this process; the following two seem to be the most prominent ones. The older is “conversion”, which assumes that this process is different from normal suffixation and simply consists of converting a lexeme belonging to one word class, e.g. salt N, to a lexeme belonging to another word class, salt V. Connected with this is the view that this is a recent phenomenon in English, which only became really productive after the loss of inflectional endings (cf., e.g., Biese 1941). The second approach assumes that this is a regular derivational process following the binary determinant/determinatum schema of word-formation, except that the derivational element is not overtly present, but is there in content; this is formally expressed by assuming a zero morpheme as the derivational element, which allows one to account for the semantic addition involved in this process. It is parallel to suffixal derivation, cf. carbon N > carbon-ize V ‘provide with carbon’ = salt N > salt-Ø V ‘provide with salt’, black ADJ > black-en V ‘make black’ = clean ADJ > clean-Ø V ‘make clean’, teach V > teach-er N ‘someone who teaches’ = cheat V > cheat-Ø N ‘someone who cheats’, walk V > walk-ing N ‘action of walking’ = walk V > walk-Ø N ‘act of walking’ (Marchand 1969: 359–364). This interpretation presupposes a neat distinction between inflection and derivation, where inflectional endings have no derivational force, and therefore allows us to treat as zero-derivations also instances where the derived word is accompanied by inflectional affixes, as in OE a¯rN ‘honor’ > a¯rN-ØV-ianINF ‘to honor’, de¯opADJ ‘deep’ > de¯opADJ-ØV-ianINF ‘to deepen’, feohtV-anINF ‘to fight’ > feohtV-ØN-eNOM.SG‘fight’. The latter approach, apart from taking into consideration the semantics of the derivational process, also has the advantage of reflecting the historical development, because in all the instances where today we would have to assume zero-derivation, there originally was some derivational marker in the guise of a stem-formative, which was lost due to phonetic attrition and was “replaced by zero” (cf. Kastovsky 1980, 1996).

3.2.5 Blending The end of the scale is made up by blending, which is a heterogeneous category and consists of formations where both parts consist of curtailed lexemes, e.g (5a, b, c). (5)

a. Chunnel = Ch-(annel) + (t)-unnel motel = mo-(tor[ist]) + (ho)-tel b. Oxbridge = Ox-(ford) + (Cam)-bridge transceiver = trans-(mitter) + (re)-ceiver c. smog = sm-(oke) + (f)-og brunch = br-(eakfast) + (l)-unch slithy = sli-(m) + (li)the + y chortle (chuckle + snort)

Examples (5a, b) echo the structure of regular determinative compounds, except that both constituents are clipped. Oxbridge, transceiver are clipped dvandva compounds, which denote a combination (union) of the respective referents, i.e. Oxbridge ‘the universities of Oxford and Cambridge’ is comparable to Austro-Hungary ‘an entity which consists of both Austria and Hungary’, similarly concavo-convex. This compound type is an innovation in English, borrowed from Latin and Greek. The formations in (5a, b) are

137

138

II Linguistic Levels usually treated on the same footing as (5c), but this does not seem to be correct. While the examples in (5a, b) echo regular compound structures, those in (5c) do not. They denote a referent which is a mixture or crossbreed of the referents of the constituent lexemes. They have no direct counterpart in compounding and seem to be a relatively recent development alongside the tremendous rise of acronyms like aids, laser, NATO, FBI, CIA, which form the endpoint of the scale in (4).

3.2.6 Typological properties Let me conclude this section with a few general remarks on the typological properties of Modern English word-formation. The input to Modern English word-formation can be words (at the native level), stems (at the non-native level), both of which can be curtailed (clipped) in clipping-compounds and blends. This implies that there are different morphological levels with different properties. The native level is characterized by base invariancy (no alternations), affix-invariancy (except for phonologically determined alternations) and stress always located on the first constituent (= German stress rule) except for verbal prefixes as in u`ntı´e. The non-native level allows for both stem and affix variation as well as variable stress position coupled with vowel alternation, cf. intolerable : impossible : illegible, Japa´n : Ja`pane´se, etc. As in all Germanic languages, the standard constituent sequence in word-formations is determinant/determinatum (modifier/head). There is one controversial area, however, viz. prefix formations of the type defrost, disarm, unbutton. Some linguists (e.g. Lieber 1983: 253) assume that the prefixes here act as heads in order to account for the change of word-class involved, thus reversing the standard sequence of modifier/ head. This analysis coincides, interestingly, with the rejection of zero derivation as a productive derivational process; rather the relationship between the involved lexical items is described as a static lexical correspondence. But as has been argued above, zero-derivation might be regarded as a sub-type of suffixal derivation, and in the case of the prefixal verbs defrost, disarm, unbutton, the second part might also be regarded as zero-derived, functioning as a head, which would make the prefix a modifier, as befits.

4 History 4.1 Indo-European Some features of Modern English morphology go back to Indo-European, e.g. ablaut like sing : sang : sung, but also the division into strong and weak verbs and the various inflectional classes of nouns. Therefore, a look at Indo-European might be useful. Its central morphological category was the root, a monosyllabic consonantal skeleton with a vowel slot (conventionally assumed to have been /e/ with its ablaut alternants) and certain restrictions on the co-occurrence of consonants in onset and coda position (cf. Clackson 2007: 69–71). The actual nominal, adjectival, or verbal paradigms were derived by adding stem-forming elements and/or other derivational elements again followed by stem-formatives, to which the inflectional endings proper were added, as in (6): (6)

root + stem-formative (+ derivational affix + secondary stem-formative elements) + inflection proper

9 Linguistic Levels: Morphology From such roots primary nouns and primary verbs were formed. There was no directional connection between verb and derived noun, or noun and derived verb; both were indirectly related via a common root, from which they were derived by independent morphological processes. This relationship characterizes the ancestors of the strong verbs and the related nouns and adjectives. From these primary derivatives further, secondary derivatives could be formed (see Figure 9.1). Root Primary stem-formation Noun

Verb

Secondary stem-formation Verb

Noun

Figure 9.1: Derivational relationship between IE roots, noun stems, and verb stems

The Indo-European verb system was characterized by a mixture of aspectual and temporal categories such as present, imperfect, perfect, aorist, etc., and mode of action categories such as iterative, intensive, durative, inchoative, etc. In Modern English or German, these categories are separated as either belonging to inflection (tense, aspect) or derivation (mode of action/aktionsart). In Indo-European, however, no such separation was possible, and these aspect/mode of action categories were derived directly from the root, resulting in stems (see Figure 9.2). Root

Present stem

Perfect stem

Aorist stem

Figure 9.2: Verbal stems in Indo-European

Aspect formation and the person/number inflection of these primary verbs were characterized by morphologically-governed stress alternations, which produced nonfunctional differences of vowel quality and quantity, called ablaut. This is the source of the Germanic system of strong verbs, consisting of six classes (and one additional class of reduplicative verbs fitting into the same pattern), which are based on the syllable-structure of the root, viz. ablaut vowel + syllable coda in (7): (7)

Class 1: -ViTClass 2: -VuTClass 3: -VL/NTClass 4: -VL/NClass 5: -VTClass 6: -VH(V = ablaut vowel, T = obstruent, L = liquid, N = nasal, H = laryngeal)

The secondary derived verbs, i.e. those derived from primary nouns, verbs, or adjectives, but not directly from roots, only occurred in the non-perfective (= present) aspect. They are the ancestors of the Germanic weak verbs containing stem-formatives like -j- and -o¯j- characterizing class 1 and 2 weak verbs.

139

140

II Linguistic Levels

4.2 Germanic to Old English: verbal system One major Germanic innovation was a shift from an aspectual to a tense system. This coincided with the shift to initial accent, and both may have been due to language contact, maybe with Finno-Ugric. Initial stress deprived ablaut of its phonological conditioning, and the shift from aspect to tense required a systematic marking of the new preterit tense. For this, two types of exponents emerged. One is connected to the secondary (weak) verbs, which only had present aspect/tense forms. They developed an affixal “dental preterit”, together with an affix for the past participle. The source of the latter was the Indo-European participial -to-suffix; the source of the former is not clear (cf. Tops 1974). The most popular theory is grammaticalization of a periphrastic construction with do (IE *dhe-), but there are a number of phonological problems with this. The second type was the functionalization of the originally non-functional ablaut alternations to express the new category, i.e. the making use of junk (Lass 1990). But this was somewhat unsystematic, because original perfect forms were mixed with aorist forms, resulting in a pattern with over- and under-differentiation. Thus, in class III (helpan : healp : hulpon : geholpen) the preterit is over-differentiated, because the different ablaut forms are non-functional, since the personal endings would be sufficient to signal the necessary distinctions. But in class I (wrı¯tan : wra¯t : writon : gewriten), there is under-differentiation, because some preterit forms and the past participle have the same vowel. This situation proved unstable and was levelled out in Middle English, along with a loss of many strong verbs or their shift to the regular (= originally weak) class. Moreover, the transparent syllable structure on which this system had been based had lost its transparency in Old English due to numerous sound changes. It is therefore arguably possible to treat the Old English strong verbs already as irregular, i.e. to assume that the switch from a system based on the distinction between weak verbs characterized by stem-formatives and strong verbs characterized by ablaut without a stem-formative to a system based on the dichotomy of regular and irregular verbs had already occurred in Old or even in pre-Old English (cf. Kastovsky 1997). The weak verbs were originally derived from nouns, adjectives or verbs by various stem-formatives, which are still recognizable in Gothic, cf. (8): (8)

Class Class Class Class

1: 2: 3: 4:

-j-/-i-o¯j-/-o¯-e¯(j)-no¯-

(Go. (Go. (Go. (Go.

sat-j-an, OE sett-an) salb-oˆ-n, OE sealf-ian) hab-an, OE habb-an) full-n-an, ?OE beorht-n-ian)

Classes 1 and 2 survive in Old English. Remnants of class 3, e.g. libban, lifde ‘live’, habban, hæfde ‘have’, hyc˙g˙an, hogode ‘think’ are best treated as irregular. Class 4 seems to have survived with -n- having been re-interpreted as a derivational suffix, e.g. beorht-nian < OE beorht ADJ, the antecedent of the -en-suffix in blacken, redden, fasten, etc. At this stage, the stem formative had a dual function: it acted as a derivational morpheme like -ize, -ify, -en, but it also had an inflectional function characterizing a particular inflectional class. Gradually, the weak verbs lost their status as a derived category (cf. Kastovsky 1996). Many of them were reinterpreted as basic, partly because their derivational base was lost, partly due to semantic reinterpretation. The ablaut nouns related to strong verbs

9 Linguistic Levels: Morphology

141

usually denoted agents, actions, results or instruments related to the underlying verbal action. These lexical-semantic categories are typically deverbal. Now these deverbal nouns often served as the basis for secondary verbal derivatives, e.g. faran ‘to go, travel’ > fo¯r ‘going, journey’ > fe¯ran ‘go, come, depart’ (= ‘make a journey’). In this way, many pairs of the type action noun > verb ‘perform action’, agent noun > verb ‘act as agent’ came into existence. For these, however, the direction of derivation was just the opposite of the one relating the basic nouns to their strong verbs and in many instances came to be reinterpreted in the same way as in back-derivations such as peddler > peddle ‘act like a peddler’ → peddle > peddler ‘someone who peddles’ because of write > writer ‘someone who writes’. Such reinterpretations eventually established a general pattern of nominal derivation from weak verbs, so that these lost their exclusively derived character, cf. the following examples of clearly deverbal nouns in (9): (9)

hwistlian ‘to whistle’ > hwistle ‘whistle’ cnyllan ‘to strike, knock, ring a bell’ > cnyll ‘clang, stroke of a bell’ huntian ‘to hunt’ > hunta ‘hunter’

While in Germanic and pre-Old English the stem formatives were clearly recognizable throughout the respective paradigms, this was no longer true of Old English, where they have fused with the person/number endings, cf. trymm-an ‘to strengthen’ < trum ‘strong’, lufian ‘to experience love’ < lufu ‘love’, de¯man ‘to pass judgement’ < do¯m ‘judgement’, nerian ‘to save’; see Table 9.1. Table 9.1: Typical paradigms of Old English class 1 and class 2 weak verbs

Class 1

Stem

+

Person/Number/Inf.

trymm trymm trym trym trymm de¯m de¯m de¯m de¯m de¯m ner ner ner ner ner

+ + + + + + + + + + + + + + +

anINF e1P SG est2P SG eð3P SG aðPL anINF e1P SG st2P SG þ3P SG aðPL ianINF ie1P SG est2P SG (e)þ3P SG iaðPL

Class 2

Stem

+

Person/Number/Inf.

luf luf luf luf luf

+ + + + +

ianINF ie1P SG ast2P SG að3P SG iaðPL

The stems of class 1 may exhibit morphologically conditioned allomorphy caused by West Germanic Consonant Lengthening (e.g. trymm-an : trym-est : trym-ed-e); so did the person/number morphemes in both classes. The choice of the latter was determined by the class membership of the verb, which was an inherent feature of the stem, no longer linked to an overt class-defining stem-formative, i.e. external class characterization had become holistic, i.e. paradigm-dependent.

142

II Linguistic Levels In the past tense and the past participle, the two classes still seemed to be kept apart by the reflexes of the original stem formatives /e/ and /o/, as in (10): (10) a. Class 1: trym+e+dPAST+e1P SG trym+e+d+est2P SG trym+e+d+onPL gePP+trym+e+dPP b. Class 2: luf+o+dPAST+e1P SG luf+o+dPAST+est2P SG luf+o+d+onPL gePP+luf+o+dPP

This analysis we find implicitly or explicitly in the traditional handbooks, and it is the primary basis for the distinction of the two classes. But it is questionable synchronically, especially in view of the increased number of class shifts, beginning with the overlap of the nerian- and the lufian-type, but eventually going far beyond it. The morphological structure postulated for the present without a stem-formative must have also been extended to the preterit already in the course of the early Old English period. This was certainly a gradual process, and there was a transition period of morphological indeterminacy. But by the end of the 9th century, if not even earlier (the earliest class shifts might make this date more precise), the change had been complete, and the phonological relics of the stem-formatives had been reinterpreted as part of the underlying representation of the preterit/past participle morpheme, i.e. (11): (11) trym+e+dPAST+e > trym+edPAST+e luf+o+dPAST+e > luf+odPAST+e This was facilitated by the pre-Old English syncopation of /i/ in class 1 when preceded by a heavy stem or dental stops (i.e. Siever’s Law/High Vowel Deletion), cf. de¯m-d-e, set-t-e, as against ner+e+d+e. With fæstan, vowel deletion resulted in an ungrammatical cluster /fæst+t+e/, which was simplified to /st/ with concomitant loss of the representation of the preterit and past participle morpheme, i.e. fæst+Ø+e. This is the first zero allomorph of the preterit/past participle, i.e. the antecedent of the type cut : cut : cut. More instances arose through the loss of geminates in Middle English, when forms like OE set+t+t+e developed into set+Ø(+e). This development had a very important morphological side effect. The stem formatives had had a dual function, derivational and inflectional. With their loss, inflection became lexically determined (implicational), and the derivational function was no longer expressed overtly: overt derivation became zero-derivation. At the same time, the fuzzy delimitation between inflection and derivation became clear-cut. Incidentally, zero-derivation was the only process for creating denominal verbs, since the -n-suffix only derived deadjectival verbs. It was only in Middle and Early Modern English that new suffixes were introduced through borrowing, viz. -ate, -ify, and -ize. Towards the end of the Old English period, the only class whose morphological behavior was fully predictable was class 2. Verbs of class 1 exhibited morphophonemic/morphological alternations, partly old, partly created by the Late Old English vowel lengthening and shortening processes, which are the germ of a subset of the Modern English irregular verbs (type keep : kept). The underlying form of the past tense and past participle morpheme originally had not contained a vowel but was preceded by the stem-formative, but the re-interpretation postulated in (11) above created a new underlying form with a vowel, /ed/ and /od/. While /od/ was stable, /ed/ allowed deletion of the vowel in certain, originally predictable environments. This deletion process was gradually extended and became phonologically unpredictable. At the same time, /ed/ and /od/ merged in /әd/, which elevated

9 Linguistic Levels: Morphology

143

vowel-deletion to a class distinctive feature. We now have two classes: one that deletes the vowel, and one that does not. In the course of Middle English, vowel deletion was gradually generalized, at first controlled by stylistic and rhythmic factors, until it became the rule except after base-final alveolar stops. At this stage, the vowel deletion process must have been re-interpreted as a vowel-insertion process (rule inversion). There is one further important development, viz. the loss of the infinitive ending, which is already attested in the North at a relatively early stage, and gradually spreads to the South. This introduced word-based morphology also to the verb morphology, whereas it had evolved in noun morphology already in pre-Old English.

4.3 Germanic to Old English: nominal system Let me now turn to nominal inflection. A typical reconstructed Indo-European paradigm, e.g. the masculine o-stems, would have the structure shown in Table 9.2. Table 9.2: Reconstructed IE masculine o-stem (= Grmc. a-stem) paradigm of ‘day’ root SG

NOM

* h ´ hw

ACC

* h ´ hw

NOM

d og d og * h d oghw * h d oghw * h ´ hw d og *dhoghw

ACC

* h ´ hw

GEN

* h

GEN DAT PL

stem-formative

inflection

o o e´/o´ Ø/o´ː oː o´ː o o´ː/(?e´ː) o

s m so ı´/i s(es) s(es) ns m mis

d og d oghw * h ´ hw d og

DAT

Again, the stem-formatives had both derivational and inflectional function and exhibited allomorphic ablaut alternations. At this stage, the stem-formatives had a consistent, predictable exponent. In Germanic, this variable system was replaced by non-variable initial stress, which destroyed the phonological conditioning of ablaut and led to a gradual weakening and loss of inflectional syllables. This resulted in the reconstructed Germanic paradigm shown in Table 9.3 based on Voyles (1982). Table 9.3: Reconstructed Grmc. a-stem (= IE o-stem) paradigm of ‘day’ SG

NOM

*

ACC

*

GEN DAT

da´g da´g * ´g da * ´g da

+ + + +

az a es eː

PL

NOM

*

ACC

*

GEN DAT

da´g da´g * ´ dag * ´ dag

+ + + +

oːs / da´g + oːz a˜ o˜ː amz

Here it is no longer plausible to segment a stem-formative separate from the inflectional endings using conventional methods of morphological analysis, i.e. the original morphological structure had been destroyed. In forms such as SG ACC *dag-a, SG DAT *dag-eː, PL ACC *dag-a ˜ , PL GEN *dag-o˜, the original case/number endings had been lost, and in the interest of paradigm symmetry it can be assumed that the

144

II Linguistic Levels stem-formative was re-interpreted as a new case/number ending, thereby losing its original function. In the other forms, there is also no consistent representation of the old stem-formative. This resulted in a new set of inflectional endings with a concomitant shift from a ternary root + stem-formative + inflection to a binary stem + inflection structure. Further restructuring took place in the transition to Old English, as seen in the paradigm given in Table 9.4. Table 9.4: OE a-stem SG

NOM ACC GEN DAT

dæg dæg dæg + es dæg + e

PL

NOM ACC GEN DAT

dag dag dag dag

+ + + +

as as a um

With these masculine nouns and the corresponding neuters, e.g. scip, the nominative/ accusative singular had lost its exponent, resulting in a base form which could function as a word without any additional morphological material. This part of nominal inflection thus had become word-based. Since these paradigms contained the majority of nouns, they attracted more and more nouns from other paradigms, e.g. from the original -ja- and -i-stems, where a form like end-e came to be re-interpreted as ende, with -e losing its inflectional function as nominative singular and becoming part of the unmarked base form, parallel to cyning. At this stage, however, morphology was still inflectional: the two inflectional categories of number and case were fused in one exponent.

4.4 Post Old English At the end of the Old English period a number of new developments happened. These have primarily to do with the merger of unstressed vowels in schwa, i.e. luf-u > luf-e /lufə/ (original “-o¯-stem”), gum-a >gum-e /ɡumə/ (original “-n-stem”), etc., which had still been stem-based with -u and -a functioning as nominative singulars. This development makes these forms look like the continuation of the original strong -jaand -i-masculines, whose (originally inflectional) ending -e > /ə/ had already been reinterpreted as part of the base. These originally stem-based forms underwent the same reinterpretation and were integrated into the emerging generalized word-based noun morphology. Another development concerned the other inflectional forms, especially the singular genitive and dative as well as the plural nominative and accusative of these stem-based nouns. With the reduction of the final vowels to /ə/ these forms were now identical with the unmarked singular nominative, i.e. they had become endingless. It is therefore not surprising that the salient singular -s-genitive was gradually transferred to these nouns and also to other inflectional classes. At the same time, these nouns also adopted -s-forms in the plural nominative/accusative. This suggests that the ending -es had come to be reinterpreted as a general marker of the category plural only, without any additional case function. The same apparently happened to the ending -en characterizing the weak masculines of the type gum-a(n). This brings about a crucial typological realignment: the categories of number and case became dissociated,

9 Linguistic Levels: Morphology because otherwise the analogical transfer of a case ending in the singular and of a plural ending independent of a case function (plural nominative/accusative) cannot be explained. This means that already at the end of the Old English period noun inflection had become completely word-based. The gradual marginalization of the dative (both singular and plural) must have started at this period as well. It had become a basically preposition-controlled ending, whose presence or absence was more and more governed by metrical-stylistic requirements (cf. Lass 1992: 110). This indicates that its actual grammatical function had been lost at this stage (or had at least been considerably weakened), cf. comparable 19th century Ger. dem Hund-e > dem Hund. The Middle English period is characterized by the gradual generalization of the -splural, since eventually most of the -n-plurals (exceptions such as ox-en, childr-en, brethr-en are the last remnants today) adopted this plural form. Similarly, the -s-genitive singular came to be generalized. The genitive plural poses a problem, because there had never been an -s-genitive in the plural, but the irregular plural nouns – as well as the group genitive – adopted this form. For a possible explanation, cf. Kastovsky (2008). The only other remarkable restructuring between Middle English and Modern English (probably in the transition from Middle English to Early Modern English) concerns the relationship between the regular allomorphs of the plural and genitive morphemes. Originally, i.e. until Middle English, their underlying representation contained a vowel, probably /e/ or /ə/, which was raised to /ɪ/ in certain varieties. During the Middle English period, this vowel could be deleted for rhythmic reasons, so that its presence or absence became optional, except before a coronal sibilant. Eventually, this vowel deletion became the rule, just as with verbs. This generalization of the vowel deletion led to rule-inversion: the vowelless allomorph came to be interpreted as underlying, and vowel-insertion replaced vowel-deletion, controlled by phonotactic requirements. In Old English, adjectives agreed with the nominal head with regard to number, gender and case. Moreover, there were two types of inflection, so-called “weak” inflection after demonstratives, and “strong” inflection without a determiner, which was a Germanic innovation. But adjectival inflection became unstable towards the end of the Old English period and was lost in Middle English, clearly as a result of the general breakdown of NP inflection.

4.5 Word-formation Let me conclude with a few remarks on word-formation. One striking feature is the almost total loss of ablaut nouns and adjectives related to strong verbs with corresponding ablaut alternations of the type write : writ, sing : song. These had formed part of the Old English core vocabulary, as they still do in German. In English, however, they were gradually replaced by base-invariant derivatives, since native word-formation – like regular inflection – adopted base invariancy as its basic principle. The massive borrowing of Romance and Latin vocabulary in Middle and Early Modern English, however, reintroduced base alternations at a new non-native level. Another remarkable phenomenon is the demise of most Old English verbal prefixes, many of which had become semantically opaque already in Old English (cf. Hiltunen

145

146

II Linguistic Levels 1983). They were gradually replaced by non-native patterns, which dominate much more in prefixation than in suffixation. Finally, the rise of a new compound type, viz. washing machine, swimming pool in Middle English should be mentioned. This has to be seen in connection with the rise of the new -ing-participle/gerund, which replaced the old -ende-participle. But, as many other Middle English developments in this area, this still needs more detailed investigation (cf. Kastovsky 2007).

5 References Bauer, Laurie. 1983. English Word-Formation. Cambridge: Cambridge University Press. Bauer, Laurie. 1992. Introducing Linguistic Morphology. Edinburgh: Edinburgh University Press. Biese, Y. M. 1941. Origin and Development of Conversions in English. Helsinki: Finnish Academy of Sciences. Blake, Norman (ed.). 1992. The Cambridge History of the English Language. Vol. II. 1066–1476. Cambridge: Cambridge University Press. Brunner, Karl. 1960–62. Die englische Sprache. Ihre geschichtliche Entwicklung. 2 Bd. 2. Aufl. Tu¨bingen: Niemeyer. Clackson, James. 2007. Indo-European Linguistics. An Introduction. Cambridge: Cambridge University Press. Giegerich, Heinz J. 1999. Lexical Strata in English: Morphological Causes, Phonological Effects. Cambridge: Cambridge University Press. Hiltunen, Risto. 1983. The Decline of the Prefixes and the Beginnings of the English Phrasal Verb. The Evidence from Some Old and Early Middle English Texts. Turku: Turun Yliopisto. Hockett, Charles. 1958. A Course in Modern Linguistics. New York: MacMillan. Hogg, Richard M. (ed.). 1992. The Cambridge History of the English Language. Vol. I. The beginnings to 1066. Cambridge: Cambridge University Press. Hogg, Richard and David Denison (eds.). 2006. A History of the English Language. Cambridge: Cambridge University Press. Huddleston, Rodney and Geoffrey K. Pullum. 2002. The Cambridge Grammar of the English Language. Cambridge: Cambridge University Press. Kastovsky, Dieter. 1980. Zero in morphology. A means of making up for phonological losses? In: Jacek Fisiak (ed.), Historical Morphology, 213–250. The Hague: Mouton. Kastovsky, Dieter. 1996. Verbal derivation in English: A historical survey. Or: Much ado about nothing. In: Derek Britton (ed.), English Historical Linguistics 1994, 93–117. Amsterdam/ Philadelphia: John Benjamins. Kastovsky, Dieter. 1997. Morphological classification in English historical linguistics: The interplay of diachrony, synchrony and morphological theory. In: Terttu Nevalainen and Leena KahlasTarkka (eds.), To Explain the Present. Studies in the Changing English Language in Honour of Matti Rissanen, 63–75. Helsinki: Socie´te´ Ne´ophilologique de Helsinki. Kastovsky, Dieter. 1999. Inflectional classes, morphological restructuring, and the dissolution of Old English grammatical gender. In: Barbara Unterbeck and Matti Rissanen (eds.), Gender in Grammar and Cognition. Vol. 2. Manifestations of Gender, 709–727. Berlin/New York: Mouton de Gruyter. Kastovsky, Dieter. 2006a. Historical morphology from a typological point of view: Examples from English. In: Terttu Nevalainen (ed.), Types of Variation: Diachronic, Dialectal and Typological Interfaces, 53–80. Amsterdam/Philadelphia: John Benjamins. Kastovsky, Dieter. 2006b. Vocabulary. In: Hogg and Dennison (eds.), 199–270. Kastovsky, Dieter. 2006c. Typological changes in derivational morphology. In: Ans van Kemenade and Bettelou Los (eds.), The Handbook of the History of English, 151–176. Oxford: Blackwell.

9 Linguistic Levels: Morphology

147

Kastovsky, Dieter. 2007. Middle English word-formation: A list of desiderata. In: Gabriella Mazzon (ed.), Studies in Middle English. Forms and Meanings, 41–56. Frankfurt: Peter Lang. Kastovsky, Dieter. 2008. The genesis of the Modern English genitive plural: Structural and phonostylistic factors. In: Jo´zef Andor, Be´la Hollo´sy, Tibor Laczko´, and Pe´ter Pelyva´s (eds.), When Grammar Minds Language and Literature. Festschrift for Prof. Be´la Korponay on the Occasion of his 80th Birthday, 263–273. Debrecen: Institute of English and American Studies, University of Debrecen. Kastovsky, Dieter. 2009a. Diachronic aspects. In: Rochelle Lieber and Pavol Sˇtekauer (eds.), The Oxford Handbook of Compounding, 323–340. Oxford: Oxford University Press. Kastovsky, Dieter. 2009b. English word-formation, combining forms and neo-classical compounds: A reassessment. In: Current Issues in Unity and Diversity of Languages. Selected Papers from the CILT 18, Held at Korea University in Seoul on July 21–26, 2008, 724–734. Seoul: Korea University. Lass, Roger. 1990. How to do things with junk: Exaptation in language evolution. Journal of Linguistics 26: 79–102. Lieber, Rochelle. 1983. Argument linking and compounds in English. Linguistic Inquiry 14: 251–286. Lyons, John. 1977. Semantics. 2 vols. Cambridge: Cambridge University Press. Marchand, Hans. 1969. The Categories and Types of Present-Day English Word-Formation. 2nd rev. edn. Mu¨nchen: Beck. Matthews, Peter. 1974. Morphology. Cambridge: Cambridge University Press. [2nd edn. 1991]. Minkova, Donka. 2008a. Prefixation and stress in Old English. Word Structure 1: 21–52. Minkova, Donka. 2008b. Continuity or re-invention in functional stress-shifting. Paper presented at the 15th International Conference on English Historical Linguistics, Munich, August 24–30, 2008. Prc´ic´, Tvrtko. 2005. Prefixes vs initial combining forms in English: A lexicographic perspective. International Journal of Lexicography 18: 313–334. Prc´ic´, Tvrtko. 2007. Headhood of suffixes and final combining forms in English word formation. Acta Linguistica Hungarica 54: 381–392. Schlegel, August Wilhelm. 1818. Observations sur le langage et la litte´rature provenc¸ales. Paris. ¨ ber die Sprache und Weisheit der Indier: Ein Beitrag zur Begru¨ndung Schlegel, Friedrich. 1808. U der Altertumskunde. Heidelberg: Winter. Tops, Guy A. J. 1974. The Origin of the Germanic Dental Preterit. A Critical Research History since 1912. Leiden: Brill. ¨ ber die Verschiedenheit des menschlichen Sprachbaues. Bonn. von Humboldt, Wilhelm. 1827–29. U Voyles, Joseph B. 1992. Early Germanic Grammar. Pre-, Proto- and Post-Germanic Languages. San Diego: Academic Press. Wald, Benji and Lawrence Besserman. 2002. The emergence of the verb-verb compound in twentieth century English and twentieth century linguistics. In: Donka Minkova and Robert Stockwell (eds.), Studies in the History of the English Language. A Millennial Perspective, 417–447. Berlin/New York: Mouton de Gruyter. Wełna, Jerzy. 1996. English Historical Morphology. Warszawa: Wydawnictwa Uniwersytetu Warszawskiego.

Dieter Kastovsky, Vienna (Austria)

148

II Linguistic Levels

10 Linguistic Levels: Syntax 1 2 3 4 5

Introduction: the syntactic history of English and English historical syntax The syntactic history of English English historical syntax Summary References

Abstract Syntactic change in English can be considered in a number of ways, depending on whether the aim of the research is to find out more about the ways in which English has changed over time, or whether it is to find out more about general constraints on change in the syntactic component of human language by using evidence from diachronic varieties of English. Adapting a proposal from Honeybone (2008), on phonological change in English, this article is concerned with both the syntactic history of English and English historical syntax, and discusses evidence from a range of varieties of the language (both contemporary and historical). Changes discussed in the article include the development of word order patterns, the evolution of auxiliaries, and the category “subject” in English. Both formal and functional theories of language change are considered, along with some discussion of the roles played by borrowing, reanalysis, and analogy in shaping change in English syntax.

1 Introduction: the syntactic history of English and English historical syntax This section presents an overview of the ways in which the syntax of English has changed. Other contributions to this volume consider the syntax of particular periods, so this is a rather global approach to syntactic change. The overview is divided into two parts, the syntactic history of English (Section 2) and English historical syntax (Section 3), following a similar distinction regarding the evolution of English phonology proposed by Honeybone (2008). To illustrate this distinction, consider the development of the tense auxiliary do in the history of the language (see further Warner, Chapter 47). There are a number of questions one might ask about this particular change, including: a. based on the textual evidence available to us, what do we know of the regional and textual provenance of this change? b. to what extent do current varieties of English display similar or divergent patterns with respect to the use of do-support? c. how can we relate the development of do-support to other changes affecting auxiliary verbs in English, such as the modals and the aspect and voice auxiliaries? d. is this change typologically odd? e. what does this change tell us about properties of verbs and their dependents? f. how does this affect our understanding of constraints on syntactic change, or of more general issues such as reanalysis and analogy/extension? Alex Bergs and Laurel Brinton 2012, Historical Linguistics of English (HSK 34.1), de Gruyter, 148–164

10 Linguistic Levels: Syntax The first three questions relate to the syntactic history of English, since answers to such questions will tell us something about a particular development in a particular set of varieties, based on a particular set of evidence. By contrast, the last three questions relate to English historical syntax, since answers to such questions will tell us something about general properties of human language undergoing change, comparing data from (varieties of) English with other varieties as evidence for universal tendencies. It is important to keep the syntactic history of English distinct from English historical syntax, because different questions need to be asked in each case. The importance of the distinction does not lie in a difference between description and explanation: it is wrong to say that work in the syntactic history of English does not concern itself with explaining patterns of change, and equally wrong to say that matters of description are irrelevant to those working on English historical syntax. Both approaches require the appropriate use of appropriate data, and both require “theorizing”, but not necessarily in the same way. In a discussion of contemporary dialect variation and theoretical syntax, Adger and Trousdale (2007: 261) note that dialect syntax in its sociolinguistic context provides a challenge for syntactic theory because it “raises important questions regarding what the theory is actually modelling”: and the same holds true for English historical syntax. Furthermore, although the syntactic history of English and English historical syntax are distinct, they are related. This can be exemplified by situations of language contact. An accurate account of the historical syntax of English will describe the many ways in which contact with other languages (from the Celtic substratum which has been said to have had a significant effect on the evolution of British English [see various papers in Filppula et al. 2008], to the new Asian varieties which have emerged in the later modern period); but contact linguistics more generally interfaces with linguistic theory, so these developments also have a role to play in English historical linguistics (not least in the way in which phonological, morphological, and syntactic changes are themselves related, and even accelerated, in contact situations). The various inputs to a child acquiring a particular variety may differ significantly depending on the degree of imperfect second language learning which may characterize the adult’s output. Exogenous changes, brought about by borrowing from one variety into another, serve as part of the input system to subsequent generations acquiring the language (who may have no knowledge that a particular form is the consequence of borrowing), which may prompt a series of endogenous changes. However one wishes to describe this phenomenon in theoretical terms – parameter resetting, constraint reranking, constructional entrenchment, and so on – the relation between the syntactic history of English and English historical syntax is at the heart of much of the new research in this field of linguistic enquiry. My reason for detailing this distinction at some length is because much research remains to be done, despite advances made in both fields. Some work which (in part or in whole) addresses grammatical change in English is more clearly aligned with the syntactic history of English (e.g. Mitchell 1985); similarly a lot of work on historical syntax uses English as a source of data, an understandable decision given the wealth of historical material available for analysis (e.g. Lightfoot 1999); the vast majority of research includes some aspects of both (e.g. the various syntax chapters in the Cambridge History of the English Language [Hogg (ed.) 1992–2001]). However, there is still much in the domain of the syntactic history of English which remains to be uncovered; similarly, work in English historical syntax is also rapidly evolving. Some, perhaps

149

150

II Linguistic Levels most, research into English historical syntax follows mainstream work on Chomskyan generative syntax, though there is increasing work within Optimality Theory, and construction grammars of various kinds, as discussed in Section 3 below. Furthermore, some research counts as an excellent representation of both kinds of historical work, in which we discover more about both the historical development of English syntax, and the interface of these developments with more general issues in syntactic (and linguistic) theory (e.g. Warner 1993 and Allen 1995). The remainder of this introduction is concerned with some other issues of relevance to syntactic change in English. Some of these issues concern categorizations of various kinds. First, there are the rather problematic issues of what we mean by terms such as “the English language” or an “Old English dialect” when considering the changes that have occurred. Who counts as a native speaker of contemporary English, and at what arbitrary point in time do we separate Old English (OE) from Middle English (ME), for instance? Second, what evidence do we use to document the changes we observe, and how coherent are these (spoken or written) text types as instances of a single category? Comparing a Kentish charter from the 9th century with a contemporary text message written in Singlish may well indicate some of the ways in which English syntax has changed in the course of over a millennium, but we need to bear in mind not simply diachronic change, but also provenance, text type, intended communicative function, and the like. Matters such as these are perhaps more directly relevant to the syntactic history of English than to English historical syntax, but recent work on the earlier syntax of English, using the various parsed corpora, has associated different dialect patterns with both endogenous constraints and instances of language contact. A good example concerns the effects of Viking invasions on the more northerly dialects of Old English and Middle English. In terms of the particular history of English, we can witness not only place name evidence of settlements and the like, but also the structural consequences of prolonged contact with Old Norse, in the more rapid loss of inflections witnessed in texts from northern parts of England, in comparison with texts with a more southerly provenance. The syntactic consequences of this contact are far reaching, including variation in the position of finite verbs (the V2 constraint), with northern varieties of Middle English as a COMPLEMENTIZER PHRASE-V2 type, and southern varieties as INFLECTION PHRASE-V2 (Kroch et al. 2000), and, following from that, the identification of the functional projection under CP in Old English as AgrSP (Subject Agreement Phrase) (Haeberli 2000). This snapshot is intended to illustrate how it is possible to connect, via a series of steps, particular features of the syntactic history of English to highly theory-specific accounts in English historical syntax. While it is helpful to separate the syntactic history of English from English historical syntax, it is also useful to see how the two relate to one another. Very little mention is made in what follows of the methodologies adopted in accounts of syntactic change, from the primarily philological to the primarily theoretical; equally, I have decided for reasons of space not to discuss many of the ways in which the use of computerized corpora has revolutionized how work on grammatical change in English is conducted. However, both of these issues are addressed briefly in the conclusion. For further discussion of good practice in historical syntax research, see Fischer (2007: 11–52). In an overview chapter such as this, it is possible only to deal with a small subset of syntactic changes. There are three main changes I concentrate on

10 Linguistic Levels: Syntax (OV/VO word order, the loss of impersonals and the establishment of the category “modal”), linking these where possible to other changes in the system, and showing how various theoretical issues can be addressed by considering these particular changes.

2 The syntactic history of English The syntactic history of English, based on the extant materials available to us, reveals some issues of continuity throughout the history of language, e.g. the availability of preand post-modification of nouns in both Old English and Present-day English (PDE), and some aspects of change, e.g. the use of the sequence [mə] with the first person singular pronoun to indicate futurity in some non-standard varieties of English, as in I’m a get a drink. Before I go on to discuss some particular features of linguistic change, it is important to consider some aspects of the way in which the story of English is told in many historical accounts. Typically, the story is that of the evolution of forms which constitute the standard variety – explanations are given to work out why it is that the most geographically generalized and conventionalized of forms have come to take the shape that they have. There are a number of reasons for this. For instance, much work in syntactic theory has used English as a data source, and particularly, has used the idiolect of the researcher as a way of judging grammaticality, which typically reflects the middleclass, educated variety of English that typifies the standard. This is of particular relevance to English historical syntax, but still features in accounts of the history of a particular language. By way of example, we can consider the history of word order in English.

2.1 Word order changes Most research on word order in clauses has been concerned with three particular phenomena, the V2 constraint, verb raising, and OV word order, illustrated by (1), (2), and (3) respectively: (1)

On his dagum sende Gregorius us fulluht In his day-DAT.PL send-3P.SG.PAST Gregory 1P.PL.DAT baptism ‘In his time, Gregory sent us Christianity’ (ChronA2 18.565.1; Haeberli 2002: 88)

(2)

locige ic buton to ðæm eaðmodum To hwæm To who-DAT look-1P.SG.PRES I except to the-DAT humble-DAT.PL ‘To whom do I look except to the humble?’ (CP 41.299.18; Fischer et al. 2000: 67)

(3)

He ne mæg his agne aberan He NEG can his own support-INF ‘He cannot support his own’ (CP 7.53.1; Moerenhout and van der Wurff 2005: 85)

The loss of V2, OV word order and V-to-T raising has been central in much work in English historical syntax (for useful summaries within a principles and parameters model, see Fischer et al. 2000). The first two changes are typically used to explain why English word order has become more “fixed”. Exceptions to the rule in contemporary

151

152

II Linguistic Levels standard English are noted (e.g. in clauses beginning with negative adverbs, as in Never have I heard such rubbish), but typically, the story ends there. However, word order in “English” continues to evolve. The syntactic history of English is a history of contact, and in some contact varieties we see a continuation of variation in terms of V2 and verb raising in (4) and (5): (4)

if you want we can go earlier because # at four thirty starts the quiz (GermanEnglish bilingual; Eppler 1999: 303)

(5)

What he has eaten? (Indian vernacular English; Bhatt 2000: 74)

Do we include such patterns as part of the syntactic history of English? To a large extent the answer to that question depends on the degree of conventionalization in different communities. Example (4) might be a “one-off” case of interference between the competing grammars of a single bilingual (though as we will see in Section 3, the notion of competing grammars among monolinguals has been highly influential in theorizing earlier stages of word order variation and change), thus an innovation rather than a change (since the latter requires spread). But (5) is rather well established as a feature of a particular variety of English spoken by a substantial number of people. Because so much of the syntactic history of English is concerned with the tracing of what is now the standard for inner circle varieties of the language, many of the “big stories” in the syntactic history tend to come to the same conclusion, by relating what has happened to give us the contemporary standard forms.

2.2 The English modals There are also a number of questions regarding the extent of continuity in the history of the language. A useful example here comes from another of the more widely discussed changes, the story of the English modals (again a topic of relevance to both the syntactic history of English, and English historical syntax: for a useful overview of both the relevant data and different theoretical accounts, see Denison 1993: 292–339). The story of the modals is one of divergence, in which a series of (formally) slightly anomalous verbs became (formally and functionally) even more anomalous over time. Although part of a set of preterit-presents in Old English, the ancestors of modern modals had more verblike formal properties than the Present-day English modals do, such as the ability to take an object complement and to appear in non-finite form, with these properties gradually being lost over time. Allied to the formal development is a functional change, such that these verbs come to encode aspects of the speaker’s viewpoint, whether that be the intention to lay down some sort of obligation on the hearer (in deontic modality) or the assessment of the veracity of the proposition encoded elsewhere in the clause (in epistemic modality), functions which had previously been coded by subjunctive inflectional endings on verbs. The story of the modals does not involve an orderly transition from lexical verb to auxiliary. This is shown most clearly by the findings of Warner (1993), who emphasizes the apparent messiness of the development. For example, instead of developing more auxiliary-like properties in the transition from Old English to Middle English, the modals in some ways become more verb like (e.g. OE *sculan and cunnan develop new

10 Linguistic Levels: Syntax present tense forms sculeþ and cunneþ), and different dialects of Middle English show changes at different times and in different ways. What Warner’s study suggests is that the category of auxiliary (and the subcategory of modals) becomes strengthened in the history of English (see also Hudson 1997), by showing bonds between the various forms that were to emerge as auxiliaries (e.g. negative contraction in nolde ‘would not’, næs ‘was not’, nabbe ‘have not’). Warner argues that the key development is that the auxiliary category becomes more well-defined (a more basic level category), in part in terms of its interaction with other phenomena (e.g. the appearance of tag questions, negative clitics, and the position of unstressed adverbs). Towards the end of Middle English (roughly at the turn of the 15th century), the differentiation between auxiliaries and verbs becomes sharper. A further issue regarding continuity concerns the existence of “double modal” constructions in some contemporary varieties of English, as illustrated by (6) and (7): (6)

Could you might possibly use a teller machine? (Southern US English; Mishoe and Montgomery 1994: 11)

(7)

Oh no, they’re double-glazed. They wouldn’t could [break] (Tyneside English; McDonald 1981)

Such forms are well-attested in earlier varieties of English, as we see in (8): (8)

& hwu muge we þone weig cunnen and how may we the-MASC.SG.ACC way can-INF ‘And how can we know the way?’ ( Jn [Warn 30]; Fischer and van der Wurff 2006: 147)

and there is some debate regarding the extent to which the contemporary forms should be seen as continuations from earlier stages of the language, or as independent developments. Of relevance here is the fact that the syntax and semantics of modal combinations in those varieties of British English which allow them (i.e. dialects of north-eastern England and central Scotland) are rather different from those in the southern United States dialects. So it may be that we have continuation in the case of the British dialects, but independent development in the American ones (see further Nagle 1994).

2.3 Subjects and the impersonal construction My final example to illustrate the centrality of variation in the syntactic history of English concerns the category of “subject”. While it is certainly the case that there are some instances of “subjectless” clauses in early English: (9)

norþan sniwde from-north snow-3P.SG.PAST ‘It snowed from the north’ (Seafarer 31)

153

154

II Linguistic Levels such unambiguous examples are restricted either to a particular lexical set (weather verbs) (9) or to a particular kind of information packaging (impersonal passives) (10): forðy to ungemetlice (10) ðætte that therefore too greatly geliðod ðæm the-DAT let-off-PAST.PRTC ‘that therefore it must be let off too greatly et al. 2000: 39)

ne sie not be-3P.SG.PRES.SUBJ scyldgan guilty-DAT to the guilty’ (CP 20.149.24; Fischer

In cases of co-ordination such as (11): (11) and him and 3P.SG.MASC.DAT and him and 3P.SG.MASC.DAT ‘and angels came to 2000: 39)

comon englas to, come-PL.PAST angels-PL.NOM/ACC to ðenodon served him, and served him’ (ÆCHom I, 11.174.17; Fischer et al.

the absence of the subject in the second clause is predictable from the discourse context, and such VP coordination is equally common in Present-day English. More crucial are the well-known impersonal constructions such as in (12): mannes (12) him ofhreow þæ s 3SG.MASC.DAT rue-3P.SG.PAST the-GEN man-GEN ‘He pitied the man’ (ÆCHom I, 8.192.16; Denison 1993: 85) in which neither argument is marked as nominative. Such constructions have been thoroughly discussed in the literature (most comprehensively by Allen 1995), and highlight the problematic notion of “subject” as a category for earlier English. One widely held view is that the loss of impersonal constructions is a consequence of case loss (see e.g. Lightfoot 1991), where assignment of inherent/lexical (dative) case to experiencers prevents structural case assignment. The question remains, however, as to whether dative or genitive noun phrases in such constructions constitute subjects. Evidence from raising and co-ordination, as in (13) and (14) respectively:

ælmihtigum (13) him sceal sceamian ætforan gode 3P.SG.MASC.DAT shall shame-INF before God-DAT almighty-DAT ‘he shall be ashamed before God almighty’ (ÆLS [Ash Wednesday] 12.169 Visser 1970: 23) (14) ac gode ne licode na heora geleafleast … but god-DAT NEG please-3P.SG.PAST NEG 3P.PL.GEN faithlessness … ac asende him to fyr of heofnum but send-3P.SG.PAST 3P.PL.DAT to fire-ACC from heavens-DAT ‘But their faithlessness did not please God … but [he] sent them fire from heaven’ (ÆCHom II, 20.644.71; Denison 1993: 89)

10 Linguistic Levels: Syntax suggests that dative noun phrases had some subject properties; however, the category of subject was not as fully grammaticalized as it is in Present-day English, which may be associated with a more general grammaticalization of the transitive construction (Trousdale 2008). Anderson (1997: 216–224) proposes a different analysis, in which the change involves a gradual coalescence of morphosyntactic and syntactic subject. The distinction between the two can be illustrated by existential clauses such as there are books on the shelf, where books is the syntactic subject, there the morphosyntactic subject (controlling concord). In Anderson’s analysis, earlier English optionally marked morphosyntactic subjecthood, though raising in (13) suggests that the dative NP is a syntactic subject. Turning to the situation in contemporary English, as Anderson (1997: 224) observes, instances such as there’s books on the shelf suggest the wide systemic spread of the coalescence of morphosyntactic and syntactic subject. But even subjectlessness has not disappeared entirely. We find it in particular (albeit restricted) kinds of written or spoken discourse, such as diary entries of the kind went home, ate dinner, fell asleep (Haegeman 1997), or in casual dialogue (A: How many people were at the party? B: Dunno, couldn’t say), where it is understood that the subject is first person singular. There are many other aspects of the syntactic history of English which I do not have the space to deal with here (for an excellent summary, see Fischer and van der Wurff 2006; for authoritative treatment of individual periods, see the various syntax chapters of the Cambridge History of the English Language [Hogg (ed.) 1992–2001]). There are also many other topics which are still awaiting detailed treatment. Some of these are highly specific. For instance, what is the precise history of the ditransitive construction where both objects are pronouns (e.g. he sent it her vs. he sent her it)? What do we know about the spread of the progressive with stative verbs (e.g. I’m loving your new look), and how widespread is this in the new Englishes? Why does þa trigger V2 more consistently than þonne in OE? What is the spread of the indirect passive in the later history of English (and why was the spread so slow)? There are also general questions relating to on-going changes: for instance, what features characterise the syntax of emergent contact varieties involving English? Answers to these and other questions will clarify even further the various features of change in the syntactic history of English.

3 English historical syntax In this section, I look at some of the ways in which data from the history of English has been used to explore more general issues in syntactic change. I will use some of the changes discussed in Section 2 to indicate how the syntactic history of English and English historical syntax overlap, and bring in some further data to highlight some of the other relevant issues. In discussing cross-linguistic patterns, Harris and Campbell (1995: 50) state that there are three mechanisms of syntactic change: borrowing, reanalysis, and extension. I deal with each in turn, though for reasons of space I say less about borrowing than about the other mechanisms of change, and I give generally accepted definitions of each mechanism. Borrowing, in its strictest sense, occurs when the “replication of the syntactic pattern is incorporated into the borrowing language through the influence of a host pattern found in a contact language” (Harris and Campbell 1995: 51); reanalysis involves “change in the structure of an expression or class of expressions that does not involve any immediate or intrinsic modification of its surface

155

156

II Linguistic Levels manifestation” (Langacker 1977: 58), while extension is the reverse, involving changes “in the surface manifestation of a pattern […] which does not involve immediate or intrinsic modification of its underlying structure” (Harris and Campbell 1995: 51). Harris and Campbell see extension as part of analogy, but avoid using this second term because it has a range of meanings in the literature. The critical issue is whether the analogy involves some exemplar or not, as discussed below: exemplar-based analogy may be equated with extension, but “non-exemplar based analogy” is rather different, as Kiparsky (forthc.) has argued.

3.1 Borrowing Research on borrowing in the earlier history of the language (i.e. on English spoken in Britain) has considered both borrowing from other languages and borrowing from different dialects of English. In terms of language contact, for earlier English, syntactic borrowing from the Celtic substratum (Filppula et al. 2008) and from Old Norse (McWhorter 2002) has been suggested, and the effects of long-term contact with Vikings in the north of England have been shown to be widespread (see further below). An interesting case regarding borrowing from Latin concerns the development of accusative with infinitive constructions in Middle English: Fischer (1989) suggests that Latin borrowing alone was not responsible for the change, though the existence of such constructions in Latin may have facilitated some aspects of its spread. The later history of language contact involving English has more to do with English beyond Britain, particularly in the development of African American English, and a range of pidgin and creole languages. Dialect contact has also been influential in the development of particular morphosyntactic features of both British English, and other Englishes, e.g. in the spread of do-support, and the variation between -th, -s and zero as inflections on third person singular indicative verbs (Nevalainen and Raumolin-Brunberg 2003).

3.2 Reanalysis Reanalysis and extension/analogy have been at the centre of much of the work in English historical syntax. In research on many of the changes discussed here (e.g. OV/ VO word order, loss of the impersonal construction, and the establishment of the category modal), two rather different claims have been made. One is that change is catastrophic, the other is that it is gradual. This raises the question of what we mean by syntactic change, and different kinds of grammarians are likely to give different answers to that question. I deal first with formal approaches to syntactic change, and how these approaches deal with the issue of reanalysis. Later in this section, I consider functional accounts of reanalysis in syntactic change. For many formal grammarians, what changes is the system, the set of parameters (or features on lexical items) which determine well-formedness in a particular manifestation of human language. In this approach to change, the primary focus is on acquisition, on how a child acquiring a particular language comes to set parameters in a particular way, given a particular set of inputs. Acquisition is the locus of reanalysis, as the language learner sets the parameters of his or her grammar, based on the primary linguistic data. The notion of parameters has changed as Chomskyan theory has developed, with specific consequences for this particular theory of syntax and our understanding of

10 Linguistic Levels: Syntax syntactic change more generally, and parametric change has been suggested by some (e.g. Pintzuk et al. 2000) to be reduced to the creation and combination of feature bundles in lexical items. This diachronic notion of the locus of change correlates with other Minimalist work on synchronic dialectal variation (e.g. Adger 2007 on variation in Buckie Scots). Increased use of corpora in formalist accounts of change, combined with particular discussions of dialectal differences in the implementation of changes (e.g. Kroch and Taylor 1997 on verb movement in Middle English) has rather altered some of the perceptions of the importance of E(xternalized)-language data in accounting for I(nternalized)-language changes. Formalist attempts to associate statistical patterns in corpora/E-language to changes in a particular individual, mental system/ I-language have re-evaluated some of the claims regarding the ways in which reanalysis is actualized. Nonetheless, reanalysis is the primary mechanism of change in most formal accounts, a mechanism of change which is also discussed in the more functionalist grammaticalization literature. Indeed Roberts and Roussou (2003), adopting a Minimalist theory of language structure, argue that grammaticalization is an epiphenomenon, since it is simply a particular case of parametric change involving reanalysis – the production of functional material from lexical material (primary grammaticalization) or other functional material (secondary grammaticalization) – and structural simplification. This focus on functional heads – and the notion that functional heads are the “magnets” which trigger Move (Chomsky 1995) – is also used to explain patterns of word-order variation, a linguistic feature which Meillet (1958 [1912]) also considered to be associated with grammaticalization. However, while Roberts and Roussou identify similarities between patterns of word order variation and instances of grammaticalization, they also show how the two are distinct. For instance, in their Minimalist model, all of the word order changes mentioned in the previous section – the loss of V2, the loss of verbraising/V-to-T, and the shift from OV to VO – involve “loss of movement to a higher functional position” (Roberts and Roussou 2003: 206). Loss of movement also characterizes the development of the modals (Roberts and Roussou 2003: 195), which unlike the others is an instance of grammaticalization. They therefore identify five significant differences between the two instances of loss of movement, one of which entails grammaticalization, the other of which does not, for example: a. only the grammaticalization of the modals creates a “new realization for T (T* Merge)” (Roberts and Roussou 2003: 207) b. the word order changes involve a reanalysis to a lower functional head (e.g. C(OMPLEMENTIZER)-to-T), the modals to a higher one (i.e. V-to-T) c. loss of V-to-T affects all lexical verbs, but the changes in the modals affect only a subset (but see Hudson 1997 for an alternative analysis which dispenses with V-to-T movement) d. upward reanalysis only is associated with bleaching and phonetic attrition (though this seems rather a stipulation than anything explanatory) e. upward reanalysis can be cyclical. Associated with this account of parameter resetting is the notion of grammar competition (Kroch 1989) as an explanation of language change. In grammar competition, an individual speaker is said to display patterns of variation that cannot be the product

157

158

II Linguistic Levels of operations of a single grammar. In a principles and parameters model, it is impossible for an individual speaker to have one grammar that both allows verb raising and simultaneously disallows it: the “switch” for the V-to-T parameter cannot be set at both “on” and “off” in one individual grammar. In more recent Chomskyan theory, as noted above, the focus of grammar competition has shifted to features of lexical items (and in cases of variation, how the same lexical item surfaces with apparently contradictory feature markings). A further issue associated with grammar competition is the Constant Rate Effect (Kroch 1989), a phenomenon used to link frequency with rates of change: for any change involving grammar competition, while the change may occur more frequently in one syntactic context than in another, the rate of change across different contexts remains the same. Different rates of change indicate that there is likely to be more than one change involved. Grammar competition has been used as an explanation for word order changes such as the loss of OV. In an elegant study (which also illustrates the usefulness of computerized corpora in syntactic change), Pintzuk and Taylor (2006) provide quantitative evidence that, although VO order occurs with different kinds of objects (positive, quantified, and negative) all affected by different kinds of factors (such as length and thematic role), the rates of change are different, and so cannot be explained simply in terms of grammar competition; rather, there is both (a) grammar competition between head initial and head final VPs and (b) addition stylistic motivations for objectmovement. Crucially, the corpus data provide no evidence of negative objects postposing with OV grammars, and little evidence of positive objects preposing with VO grammars. The rise of VO and the demise of OV, in this account, is due to not only the loss of the OV grammar, but also the loss of the movement rule in VO generated grammars. Given that the rate of loss of the preposing movement rule is different for different kinds of objects (i.e. positive or quantified), Pintzuk and Taylor account for the gradual nature of the actualization of the changes in the history of the language. In such an approach to change, the actualization of change may be gradual, and this gradualness may be systemic (changes may occur in a particular syntactic context before spreading to another), spatial (affecting the idiolects of one geographical area before another), stylistic (originating in a particular speech context, or register) or social (occurring in the language of a particular subgroup in the speech community before being transmitted to another subgroup), or any combination of these. However, the reanalysis itself is “abrupt and catastrophic” (Lightfoot 1999: 88), and what changes is grammar. By contrast, in some functionalist analyses of grammatical change, what changes is use. Indeed, in usage-based models (Kemmer and Barlow 2000) the relationship between form and function is something of a feedback loop, where “usage feeds into the creation of grammar, just as much as grammar determines the shape of usage” (Bybee 2006: 730); the frequency of use affects the mental representation of language, such that “grammar is the cognitive organization of one’s experience with language” (Bybee 2006: 711). In models of language structure such as these, reanalysis occurs, but it is not restricted to changes of features or resetting parameters; instead, what is involved in change is form-function reanalysis, alterations to “the form-meaning mapping in a grammatical construction” (Croft 2000: 118). This notion of reanalysis is central to Hollmann’s work on the development of the have causative in English (e.g. my boss had me work late) arising from an “affecting event” construction (e.g. he wolde haue his reign endure and last), in which Hollmann argues that the form-function

10 Linguistic Levels: Syntax reanalysis is based on alternative construals: the pragmatics of an experiential event allow for a causative interpretation where the experiencer is understood to be more powerful than the other participant in the process (Hollmann 2003: 87): compare (15) and (16): (15) I often have my boss come in when I’m sleeping (= ‘It often happens that my boss comes in when I’m sleeping’) (16) My boss often has me come in when he’s sleeping (= ‘My boss forces me to come in when he’s sleeping’) While both examples could be interpreted as either causative or as “affecting event” constructions, (15) seems more likely as an affecting event, and (16) as a causative, because of the different power relations inherent in an employer-employee relationship. Hollmann’s account shows how a cluster of constructions involving have with a different (formal) complement had a similar function to the emerging have causative in late Middle English. Reanalysis here, then, is conceived as accommodation attempts: speakers and hearers negotiating the relationship between the forms and functions of constructions while trying to maintain their conventionalized uses (see further Croft 2000: 118). Change in this sense is most likely to be gradual, because it may involve incremental changes at all parts of a construction. We can consider such an instance of gradual change in the development of the English determiner. In their discussion of the elements of the noun phrase, Fischer and van der Wurff (2006: 114) argue that there has been little change in the order of the various dependents, suggesting that there has been no new functional slot in the history of the language. This position runs counter to that of Denison (2006: 288), who suggests that, rather, parallel to the development of the category Modal in English, the “evidence for the existence of D[eterminer] is much shakier in earlier English”, and somewhat problematic even in contemporary English. Categorization – both the general cognitive process and the language-specific outcomes – has been, and continues to be, an important issue in general linguistics, but it is of particular concern in historical work, where issues of synchronic gradience at time 1, and diachronic gradualness, either temporally (i.e. between time 1 and time 2), or structurally (e.g. between a change in some aspect of meaning and a change in some aspect of form), or sociolinguistically (e.g. in the spread between groups of speakers, or registers), are of considerable importance, as noted above. Specifically in the syntactic history of English, we certainly see a strengthening of the category “determiner” over time. We see this development most sharply, perhaps, in the development of articles, the most grammaticalized of determiners. In this development, English is like many other languages, in that the definite article the has arisen from a demonstrative, while the indefinite article a(n) has arisen from the numeral “one” (OE an). In less formal contemporary English, the use of some with a singular count noun (as in Some guy called earlier) we witness a further typical grammaticalization path, through which an indefinite article emerges from and becomes layered with an earlier quantifier (OE sum). The absence of such clearly grammaticalized articles in the Old English period does not, of course, mean that speakers of the language at that time could not express definiteness; the difference was simply in

159

160

II Linguistic Levels the means of expression: strong forms of adjectives typically indicated indefiniteness, as in (17): (17) ðurh boclic-e lare through book-like teaching ‘by teaching with books’ (ÆCHom Pref. 175.68) while weak forms (often with a demonstrative, suggesting incipient grammaticalization even at the earliest stages of the language) typically indicated definiteness (18): (18) se

frumsceapen-a mann first-created-NOM.SG.MASC man ‘the first man’ (ÆCHom I 7.240.250)

DEM.NOM.SG.MASC

In other words, what has remained stable is the capacity to mark particular “things” as (in)definite; what has changed is the means by which this is marked. Put another way, speakers of English have always had some way of grounding nominals, but over time they have chosen to do this using different linguistic strategies: the function is constant, the form changes. This observation from history is of relevance not just to our understanding of the categories of Present-day English, but more generally for our understanding of the nature of categories. Particularly, it brings into question whether categories may be determined by syntax alone (see Anderson 1997 and Aarts 2007 for different views on this), and it helps us to understand the relationship between gradience and gradualness in grammatical change (Traugott and Trousdale 2010). The crystallization of articles forms part of a more general development in the gradual evolution of the category “determiner” in Denison’s account; many of these crystallizations may be considered as instances of grammaticalization, including the deictification of post-determiners such as various and several (Davidse and Breban 2006).

3.3 Analogy Analogy has been an equally debated concept in syntactic change. Both reanalysis and analogy have been widely debated in the grammaticalization literature (see Hopper and Traugott 2003 for a useful summary); recently Fischer (2007) has called for a greater focus on the role of analogy in grammaticalization (arguing, for example, that in early Modern English be going to joins the token set of [AUX V] based on analogy with other future markers like will), while Kiparsky (forthc.) has equated non-exemplar based analogy with grammaticalization. Fischer’s model presents an interesting synthesis between iconic and indexical relations, between types and tokens (and relevant sets for each), and between the various paradigmatic and syntagmatic processes in grammaticalization; the model also raises the important issue of whether reanalysis results in a totally new structure, but see also Meillet (1958 [1912]) on analogy, grammaticalization and the creation of new forms. Kiparsky’s distinction between exemplar based and non-exemplar based analogy make very interesting predictions for unidirectionality in syntactic change. Because exemplar-based analogy is local (language-specific), there are exceptions to it, but it is nonetheless an attempt by

10 Linguistic Levels: Syntax the language user to simplify the grammar by abandoning redundancies; by contrast, the only “model” for non-exemplar analogy has to be the most general of grammatical constraints, universal grammar itself. It is this non-exemplar based analogy that is more commonly known as grammaticalization in the sense of Meillet (1958 [1912]): the creation of new forms. For Kiparsky, then, analogy and grammaticalization both instantiate types of grammar optimization. From a functionalist perspective, a rather different way of unifying reanalysis and analogy comes from grammatical constructionalization, the entrenchment of schematic constructions through a series of discrete reanalyses, motivated by pattern matching – recall the definition of reanalysis provided by Croft (2000) above, as a “change to the form-meaning mapping of a grammatical construction” – this series of discrete reanalyses give the appearance of gradual change, and can be exemplified by a series of changes in English grammar, including the development of degree modifiers (Denison 2002; Traugott 2008a) and cleft constructions (Traugott 2008b; Patten 2010).

4 Summary The one thing that is constant about the history of English syntax is variation. But questions still remain regarding the best way of capturing and modelling that variation optimally. Certainly work using computerized corpora has revolutionized what can be done in the quantitative (and qualitative) analysis of syntactic change. In relation to the history of English syntax, this has been most firmly established in work that intersects with the generative approach to diachronic syntax, as we have seen; but more recently, in cognitive linguistics, we see the use of corpora to explain patterns of collostructional variation and change in new Englishes (Mukherjee and Gries 2009) and in other languages (Hilpert 2008). This has the potential to be a very illuminating way of exploring many of the major issues in syntactic change from a cognitive, usage-based perspective. In the overall history of the syntax of the language, we see constant processes of renewal, as similar functions come to be coded in different ways. Particularly, we see renewal as part of grammaticalization, such that terms expressed by means of inflection at one stage in the language come to be expressed periphrastically, by means of the syntax, at a subsequent stage. The importance of contact on particular developments in English syntax cannot be understated. First, it played a significant role in the development of the determiner and auxiliary system, and on the establishment of particular word orders, in the earlier history of English; but second, and equally importantly, as English continues to be used in different communities around the world, speakers of the new varieties which emerge from contact with other languages conventionalise new patterns, so a subsequent wave of variation begins. Such external factors combine with system-internal changes – loss of phonological contrasts, and increased morphological syncretism – and lead to a series of restructurings. Thus histories of the syntax of English must take into account both the particular social and linguistic context in which speakers have used, and continue to use, varieties of English, and the more general patterns of syntactic change which can be witnessed in any human language. Both kinds of change require theorization and explanation, and neither can be achieved by appeal solely to local customs of use, or solely to cross-linguistic tendencies.

161

162

II Linguistic Levels Acknowledgments: I would like to thank Laurel Brinton and Elizabeth Traugott for their very helpful comments on an earlier version of this article.

5 References Aarts, Bas. 2007. Syntactic Gradience. Oxford: Oxford University Press. Adger, David. 2007. Combinatorial variability. Journal of Linguistics 42: 503–530. Adger, David and Graeme Trousdale. 2007. Variation in English syntax: Theoretical implications. English Language and Linguistics 11: 261–278. Allen, Cynthia. 1995. Case Marking and Reanalysis. Oxford: Clarendon. Anderson, John M. 1997. A Notional Theory of Syntactic Categories. Cambridge: Cambridge University Press. Bhatt, Rakesh M. 2000. Optimal expressions in Indian English. English Language and Linguistics 4: 69–95. Bybee, Joan. 2006. From usage to grammar: The mind’s response to repetition. Language 82: 711–733. Chomsky, Noam. 1995. The Minimalist Program. Cambridge, MA: MIT Press. Croft, William. 2000. Explaining Language Change. London: Longman. Davidse, Kristin and Tine Breban. 2006. Deictification: The Development of Postdeterminer Uses of Adjectives. Preprint no. 250. Department of Linguistics, KU Leuven. Denison, David. 1993. English Historical Syntax: Verbal Constructions. London: Longman. Denison, David. 2002. History of the sort of construction family. Paper presented at the International Conference on Construction Grammar 2, Helsinki, 7 September 2002. Denison, David. 2006. Category change and gradience in the determiner system. In: van Kemenade and Los (eds.), 279–304. Eppler, Eva. 1999. Word order in German English mixed discourse. UCL Working Papers in Linguistics 11: 285–308 Filppula, Markku, Juhani Klemola, and Heli Paulasto (eds.). 2008. English and Celtic in Contact. London: Routledge. Fischer, Olga. 1989. The origin and spread of the accusative and infinitive construction in English. Folia Linguistica Historica 8: 143–217. Fischer, Olga. 2007. Morphosyntactic Change: Functional and Formal Perspectives. Oxford: Oxford University Press. Fischer, Olga, Ans van Kemenade, Willem Koopman, and Wim van der Wurff. 2000. The Syntax of Early English. Cambridge: Cambridge University Press. Fischer, Olga and Wim van der Wurff. 2006. Syntax. In: Richard Hogg and David Denison (eds.), A History of the English Language, 109–198. Cambridge: Cambridge University Press. Haeberli, Eric. 2000. Adjuncts and the syntax of subjects in Old and Middle English. In: Pintzuk et al. (eds.), 109–131. Haeberli, Eric. 2002. Inflectional morphology and the loss of verb second in English. In: David Lightfoot (ed.), Syntactic Effects of Morphological Change, 88–106. Oxford: Oxford University Press. Haegeman, Liliane. 1997. Register variation, truncation, and subject omission in English and French. English Language and Linguistics 1: 233–270. Harris, Alice C. and Lyle Campbell. 1995. Historical Syntax in Cross-linguistic Perspective. Cambridge: Cambridge University Press. Hilpert, Martin. 2008. Germanic Future Constructions: A Usage-based Approach to Language Change. Amsterdam/Philadelphia: John Benjamins. Hogg, Richard M. (ed.). 1992–2001. The Cambridge History of the English Language. Cambridge: Cambridge University Press Hollmann, Willem. 2003. Synchrony and Diachrony of English Periphrastic Causatives: A Cognitive Perspective. Ph.D. dissertation, University of Manchester.

10 Linguistic Levels: Syntax Honeybone, Patrick. 2008. Historical phonology: Phonological variation and language change. Keynote seminar, Nordic Language Variation Network. University of Bergen/Fla˚m. Hopper, Paul and Elizabeth Closs Traugott. 2003. Grammaticalization. 2nd edn. Cambridge: Cambridge University Press. Hudson, Richard. 1997. The rise of auxiliary do: Verb-non-raising or category strengthening? Transactions of the Philological Society 95: 41–72. Kemmer, Suzanne and Michael Barlow. 2000. Introduction: A usage-based conception of grammar. In: Michael Barlow and Suzanne Kemmer, (eds.), Usage-based Models of Language, 7–28. Stanford: CSLI. Kiparsky, Paul. forthc. Grammaticalization as optimiziation. In: Dianne Jonas and John Whitman (eds.), Grammatical Change: Origins, Nature, Outcomes. Oxford: Oxford University Press. Kroch, Anthony. 1989. Reflexes of grammar in patterns of language change. Language Variation and Change 1: 199–244. Kroch, Anthony and Ann Taylor. 1997. Verb movement in Old and Middle English: Dialect variation and language contact. In: Ans van Kemenade and Nigel Vincent (eds.), Parameters of Morphosyntactic Change, 297–325. Cambridge: Cambridge University Press. Kroch, Anthony, Ann Taylor, and Donald Ringe. 2000. The Middle English Verb-Second constraint: A case study in language contact and language change. In: Susan C. Herring, Pieter van Reenen, and Lene Schøsler (eds.), Textual Parameters in Older Languages, 353–391. Amsterdam/ Philadelphia: John Benjamins. Langacker, Ronald W. 1977. Syntactic reanalysis. In: Charles Li (ed.), Mechanisms of Syntactic Change, 57–139. Austin: University of Texas Press. Lightfoot, David. 1991. How to Set Parameters: Arguments from Language Change. Cambridge, MA: MIT Press. Lightfoot, David. 1999. The Development of Language: Acquisition, Change, and Evolution. Oxford: Blackwell. McDonald, Christine. 1981. Variation in the Use of Modal Verbs with Special Reference to Tyneside English. Ph.D. dissertation, University of Newcastle upon Tyne. McWhorter, John. 2002. What happened to English? Diachronica 19: 217–272. Meillet, Antoine. 1958 [1912] L’evolution des formes grammaticales. Reprinted in: Linguistique Historique et Linguistique Ge´ne´rale, 130–158. Paris: Champion. Mishoe, Margaret and Michael Montgomery. 1994. The pragmatics of multiple modal variation in North and South Carolina. American Speech 69: 3–29. Mitchell, Bruce. 1985. Old English Syntax. Oxford: Clarendon. Moerenhout, Mike and Wim van der Wurff. 2005. Object-verb order in early sixteenth-century English prose: An exploratory study. English Language and Linguistics 9: 83–114. Mukherjee, Joybrato and Stefan Th. Gries. 2009. Collostructional nativisation in New Englishes: Verb-construction associations in the International Corpus of English. English World Wide 30: 27–51. Nagle, Stephen J. 1994. The English double modal conspiracy. Diachronica 11: 199–212. Nevalainen, Terttu and Helena Raumolin-Brunberg. 2003. Historical Sociolinguistics. London: Longman. Patten, Amanda L. 2010. Grammaticalization and the it-cleft construction. In: Traugott and Trousdale (eds.), 221–243. Pintzuk, Susan and Ann Taylor. 2006. The loss of OV order in the history of English. In: Ans van Kemenude and Bettelou Los (eds.), The Handbook of the History of English, 249–278. Oxford: Blackwell. Pintzuk, Susan, George Tsoulas, and Anthony Warner. 2000. Syntactic change: Theory and method. In: Pintzuk et al. (eds.), 1–22. Pintzuk, Susan, George Tsoulas, and Anthony Warner (eds.). 2000. Diachronic Syntax: Models and Mechanisms. Oxford: Oxford University Press.

163

164

II Linguistic Levels Roberts, Ian and Anna Roussou. 2003. Syntactic Change: A Minimalist Approach to Grammaticalization. Cambridge: Cambridge University Press. Traugott, Elizabeth Closs. 2008a. Grammaticalization, constructions and the incremental development of language: Suggestions from the development of degree modifiers in English. In: Regine Eckardt, Gerhard Ja¨ger, and Tonjes Veenstra (eds.), Variation, Selection, Development: Probing the Evolutionary Model of Language Change, 219–252. Berlin/New York: Mouton de Gruyter. Traugott, Elizabeth Closs. 2008b. “All that he endeavoured to prove was …”: On the emergence of grammatical constructions in dialogic contexts. In: Ruth Kempson and Robin Cooper (eds.), Language in Flux: Dialogue Coordination, Language Variation, Change and Evolution, 143– 177. London: Kings College Publications. Traugott, Elizabeth Closs and Graeme Trousdale. 2010. Gradience, gradualness and grammaticalization: how do they intersect? In: Traugott and Trousdale (eds.), 19–44. Traugott, Elizabeth Closs and Graeme Trousdale (eds.). 2010. Gradience, Gradualness and Grammaticalization. Amsterdam/Philadelphia: John Benjamins. Trousdale, Graeme. 2008. Words and constructions in grammaticalization: The end of the English impersonal construction. In: Susan Fitzmaurice and Donka Minkova (eds.), Studies in the History of the English Language IV: Empirical and Analytical Advances in the Study of English Language Change, 301–326. Berlin/New York: Mouton de Gruyter. Visser, F. Th. 1970. An Historical Syntax of the English Language. Part One: Syntactical Units with One Verb. Leiden: E. J. Brill. Warner, Anthony. 1993. English Auxiliaries: Structure and History. Cambridge: Cambridge University Press.

Graeme Trousdale, Edinburgh (UK)

11 Linguistic Levels: Semantics and lexicon 1 2 3 4 5 6 7 8 9

Introduction Cognitive semantics and metaphor Invited inferencing and conceptual metonymy Collocation and collostructional analysis Productivity of semantic changes at specific periods Differences between lexical and grammatical changes Changes in the lexicon Future prospects References

Abstract Selected topics in research on semantic change are discussed with focus on work based in Cognitive Linguistics and neo-Gricean pragmatics. Recent work on metaphor and metonymy, grammaticalization, subjectification, and collostructional analysis are highlighted and shown to provide theoretical underpinnings for some traditional taxonomies of semantic change. As the inventory of form-meaning pairs in a language, the historical lexicon reflects semantic change; it also reflects how vocabulary size has changed in English, and how borrowings have affected the typological ways in which meanings are packaged into words. Alex Bergs and Laurel Brinton 2012, Historical Linguistics of English (HSK 34.1), de Gruyter, 164–177

11 Linguistic Levels: Semantics and lexicon

1 Introduction While meaning or “semantics” has been a central concern of philosophy since early Greek times, semantic change has been a subject of investigation for little more than a hundred years. A landmark work is Bre´al (1897), in which a taxonomy of semantic changes was developed (and revised in structuralist terms by Ullmann 1964) that has in essence been repeated in most textbooks and handbooks on language change (e.g. Hock and Joseph 1996; Campbell 2004; Fortson 2003). Key concepts include change due to metaphor (e.g. tissue ‘woven cloth’ > ‘aggregation of cells in animals or plants’) and metonymy (e.g. board ‘table’ > ‘people sitting around a table, governing body’), pejoration (conceit ‘(self)concept’ > ‘overestimation of one’s qualities’) and amelioration (e.g. ME nice ‘foolish’ > ‘pleasant’), narrowing (e.g. OE mete ‘(solid) food’ > ‘food derived from animals’) and broadening (e.g. ME bridde ‘nestling’ > bird), and taboo avoidance (e.g. 16th century toilet ‘cloth’ > 17th century ‘cloth covering for dressingtable’ > 18th century ‘dressing-table’ > 19th century ‘lavatory’ (a euphemistic use) > 20th century ‘bathroom fixture’). Often several types of change affect one item, e.g. the changes to toilet involve first narrowing (restriction to a certain type of cloth), metonymy (object for object covering it, a subtype of whole for part), taboo avoidance, and then further narrowing. Such examples suggest that semantic change is haphazard and unpredictable. Central to Bre´al’s theory of semantics and semantic change was the importance not only of reference to objects in the world but also of what we now call “sense” (meaning defined in terms of linguistic relations), and of polysemy (related meanings associated with the same form). With the advent of structuralism, differences in sense came to be thought of in terms of “lexical fields” consisting of tightly-knit sets of words with similar meaning, such as terms for intellectual cleverness, colors, or kinship (for examples and references, many of them German, see Ullmann 1964: 243–253). In much of this work it was assumed that there were relatively fixed components of meaning, and that they could be organized in different ways across languages and times. In the latter part of the 20th century, semantic change came to be rethought in terms of more flexible sets of semantic properties. For example, one can think of how the term car fits into a “semantic space” devoted to vehicles like tank, plane, with respect to such factors as their constituent parts (metal, tires), shape, purpose, and how they are driven. Such factors are called “qualia” in Pustejovsky’s (1995) Generative Lexicon theory. The flexibility of categories is highlighted in work on prototypes and categories with fuzzy boundaries that change over time (see Geeraerts 1997; Grondelaers et al. 2007). Research on cognitive semantics, Gricean pragmatics, and grammaticalization showed that semantic change was more frequently replicated and “regular” than had often been assumed, and the advent of electronic corpora made fine-grained analysis of change in context possible. Meaning change can be conceptualized along two dimensions. One is “semasiological”: attention is paid to how meaning changes, while form remains reasonably constant (but subject to phonological change). The questions are, what meanings does a word have, how are they related, and how did they arise over time? Most of the changes listed in the taxonomies were thought of in terms of semasiology. The other dimension is “onomasiological”: attention is paid to relations that hold between the lexical items in semantic space and which forms come to express a certain meaning. This is the

165

166

II Linguistic Levels principle behind Buck’s (1949) dictionary of synonyms in Indo-European languages. It is also one of the principles behind the Historical Thesaurus of the Oxford English Dictionary (Kay et al. 2009). Current work on onomasiology concerns a wide-ranging set of concepts from expressions of emotion such as anger and pain (see Dı´az Vera 2002), to intelligence and stupidity (Allan 2008), adjectives of difference like distinct, several (Breban 2008), and modality, especially as expressed by semi-modals like have to, dare (to) (Krug 2000). The meaning of a word is only part of what we know about it. We also know its morphosyntactic structure, its phonology, what register it belongs to, perhaps what language it was borrowed from, etc. These complex sets of information are specified in the “lexicon”, the inventory of form-meaning pairs in a language. This chapter ends with some discussion of the lexicon from a historical perspective (see Section 7). In this chapter I focus on aspects of recent research on semantic change, mostly from a cognitive perspective.

2 Cognitive semantics and metaphor Cognitive Linguistics, as developed in the 1970s and 1980s, theorized a view of linguistic structures “not as if they were autonomous, but as reflections of general conceptual organization, categorization principles, processing mechanisms, and experiential and environmental influences” (Geeraerts and Cuyckens 2007: 3). A fundamental claim is that words do not have stable meanings. Rather, they are cues to potential meaning, or instructions to create meanings as words are used in context (Warren 1992, and especially 1999). These meanings are non-discrete and have prototypical properties, with core and peripheral readings (Geeraerts 1997). Furthermore, they may have rich polysemic structures (in contrast to monosemous views of semantics such as Relevance Theory, Sperber and Wilson 1995). In Cognitive Linguistics a sharp distinction between semantics and pragmatics is rejected. Although details of the theory vary considerably from author to author, there has been relative convergence among those who do work on historical semantics. The central research questions have concerned metaphor and metonymy, and ways in which semantic change occurs in linguistic contexts. Sweetser (1990) proposed a theory of metaphor and metaphorical change drawing on theories of embodiment (e.g. Lakoff 1987) and force-dynamics including exertion of force and blockage by barriers (e.g. Talmy 1988). Sweetser argued, for example, that a metaphor such as KNOWING IS SEEING developed in Indo-European languages from embodied perceptual capacities such as seeing, hearing, and grasping, and that mapping from the socio-physical world of embodiment to the abstract epistemic one of reasoning accounted for the directionality of such cross-linguistically attested meaning changes as PIE *weid ‘see’ > wit, and Gk. oı˜da perfective of eidon ‘to see’ > idea, or must ‘be required’ (compelled by socio-physical force; deontic) > ‘can be inferred’ (compelled by reasoning; epistemic). One of Sweetser’s hypotheses was that since meaning change is not random, there must be constraints on mapping from one domain to another. These constraints involve “certain abstract and topological aspects of semantic structure, which we have termed image-schematic structure, […] which must be preserved across metaphorical mappings” (Sweetser 1990: 59, italics original). Such image-schematic structures may involve barriers, e.g. may of permission represents a potential barrier that is not yet in place whether in the socio-physical domain (permission), or the domain of reasoning (possibility).

11 Linguistic Levels: Semantics and lexicon Much of the initial work on metaphor was synchronic and lexical. Particularly influential in this respect was Reddy’s (1993 [1979]) study of the conduit metaphors used for language, which accounts for extensive use of expressions concerned with language conceptualized as a physical pipe-line, e.g. put something into words, the letter contains many typos. In historical work, however, metaphor came to be noted especially in studies of grammaticalization, understood at the time as the use of lexical expressions to serve a grammatical function (see Hoffmann and Brems, Volume 2, Chapter 99). Crosslinguistically, temporal expressions derive from spatial ones (Heine et al. 1991; Haspelmath 1997), e.g. after ‘behind+COMPR’; (cf. aft of a ship); prepositions and adverbs from body parts (e.g. behind, ahead), etc. In this work, the two dimensions of semasiology and onomasiology were fruitfully combined. Heine and Kuteva’s (2002) World Lexicon of Grammaticalization provides an appendix with cross-linguistically attested source > target meaning change; this is a semasiological approach tracking changes of meaning (e.g. BODY > reflexive, cf. self ). A second appendix outlines target < source meaning change, an onomasiological approach tracking where expressions of a particular meaning derive from (e.g. FRONT < FACE, HEAD, MOUTH). Metaphor operates on the dimension of choice from among related meanings, and therefore of analogy, iconicity, and paradigmaticity. On the other hand, metonymy, being associative, operates on the dimension of indexicality, linear production, and perception (Anttila 1989: 142). Metonymy and metaphor often intersect at the conceptual level, and indeed the metaphors of embodiment ultimately derive from metonymic associations. Barcelona hypothesizes that “the target and/or source must be understood or perspectivized metonymically for the metaphor to be possible” (Barcelona 2000: 31). On this view, metonymy activates mental access to another domain, e.g. the metaphor SADNESS IS DOWN derives from experiential association with the downward bodily posture that people tend to adopt when they are sad (see also Ko¨vecses and Radden 1998; Panther and Thornburg 2003).

3 Invited inferencing and conceptual metonymy Sweetser’s work engendered tremendous interest in semantic change as metaphorical change. However, challenges and alternative proposals abounded as well. For example, Bybee (2007: 978) points out that image-schema preservation cannot apply to some changes, such as the change from perfect/anterior > present in inchoative (change of state) or state verbs, since there is no plausible image-schema associated with the perfective. In a language like Island Carib, some perfective verbs are used to express present state. Bybee suggests that an expression meaning ‘It has turned ripe’ is relevant only if it is still ripe at the time of utterance. It is the meaning of “present relevance” that underlies the change. Metaphor cannot be involved here. Rather, what is in operation is the associative, metonymic process called “invited inferencing”. We may note that most of the “core” modals in English were once “preterit-presents”, i.e. state verbs with past morphology; Gk. oı˜da cited above is the perfective form; and must is a past tense form of mot- ‘be able’ (an originally preterit-present verb). In fact, when semantic changes are considered in context, not in the decontextualized format of the examples I have cited above, they may be construed rather differently. What are schematically presented as see > ‘understand’, must ‘obligation’ > ‘epistemic conclusion’, be going to ‘motion with a purpose’ > ‘future’ are abstractions over many

167

168

II Linguistic Levels centuries of micro-changes in very specific contexts. What appears to be a metaphor may be the outcome of a number of changes in which pragmatic inferencing is activated in the flow of speech, as attested by textual data in historical corpora. The Invited Inferencing Theory of Semantic Change (IITSC) is a usage-based approach founded in the investigation of textual evidence (Traugott and Ko¨nig 1991; Traugott and Dasher 2002). Although sharing many of the assumptions of Cognitive Linguistics, it diverges in several respects, most notably in distinguishing between pragmatics and semantics. The term “invited inferencing” was borrowed from Geis and Zwicky (1971), but used in an extended way to evoke negotiation of meaning between speakers producing or even intending meanings beyond what is said, and hearers inferring such meaning. Likewise it draws on neo-Gricean pragmatics (Grice 1989; Horn 1984; Levinson 2000), and the distinction between particularized and generalized conversational implicatures, but appeals to partially different maxims. The Gricean Maxim that is considered key to change is his Quantity 2 (“Do not make your contribution more informative than is required”), combined with Relevance and rephrased as “Say no more than you must, and mean more thereby” (Horn 1984). Use of this maxim leads to rich interpretations. The hypothesis is that invited inferences that arise on the fly may become conventionalized (commonly activated) as generalized invited inferences. They may be “salient” in the community in that they can be drawn on consciously, cf. the causal implicature of after, but for the most part they are used unconsciously (Keller 1994). These generalized invited inferences may continue to be available over centuries, even millennia (cf. after). A regularly occurring context which “supports an inference-driven contextual enrichment” of one meaning to another has been called a “bridging context” (Evans and Wilkins 2000: 55). Conventionalization as a bridging context is a pragmatic development. Sometimes such inferences may be absorbed into the meaning of an expression with which they were formerly only pragmatically associated. In this case semantic reanalysis has occurred (Eckardt 2006) and a new coded meaning has become available, as evidenced by the use of an old form with the new meaning in a context which was not available before; e.g., siþþan ‘since’ was originally restricted to temporal ‘after’ and later came to be used with a causal meaning as well. In other words, semanticization of a formerly pragmatic meaning has occurred. Although originally discussed mainly with reference to grammaticalization, invited inferencing is conceived as a major motivation for semantic change in general (Traugott and Dasher 2002). It encompasses the changes associated with metonymy and metaphor, but also pejoration and amelioration. In the latter cases, the invited inferences are not only conceptually but also socially motivated. For example, those considered blessed and innocent may be evaluated as ignorant or foolish, cf. OE selig ‘blessed’ > ‘silly’, or Lt. nescius ‘unknowing, innocent’ (< ne + scius ‘not knowing’) > Middle Fr. nice ‘foolish’. Semanticization of a speaker’s beliefs and/or attitudes to what is being said is called “subjectification” (Traugott 1989, 2010; see also Davidse et al. [eds.] 2010). A partially overlapping, but more restricted, view of subjectification associates it with changes in the cognitive construal of vantage-point (e.g. Langacker 1990, 2006; see also Athanasiadou et al. [eds.] 2006). Subjectification in the first sense encompasses shifts not only from the perspective of the sujet d’e´nonce´ ‘syntactic subject’ to the sujet d’e´nonciation ‘speaking subject’ (Benveniste 1971), but also a range of meaning developments based

11 Linguistic Levels: Semantics and lexicon in the speaker’s perspective: spatial, temporal, metalinguistic, etc. Subjectification is extensively evidenced by the use of phrases like after all, anyway as discourse markers, and of adjectives like very (< ‘true’, see Fr. vrai) as scalar degree modifiers. It is also evidenced by the development of epistemic meanings, and of performative uses of speech act verbs. Many of the latter derive ultimately from past participles of Latin verbs, such as promise (< Lt. pro + miss- ‘forward sent’), suggest (< Lt. sub + gest‘under carried’) (note the conduit metaphors in the Latin, and the past participle form which shows these were originally stative and non-agentive).

4 Collocation and collostructional analysis Cognitive Linguistics and the Invited Inferencing Theory of Semantic Change are both conceptualized as “usage-based” theories. For the most part, the first is exemplified by constructed data, the latter by empirically attested data found in historical texts. Bybee has said of cognitive, usage-based linguistics in general that it is a “framework that allows change to be gradual and specific on various dimensions, such as the lexical, phonetic, and morphosyntactic, while at the same time providing general principles of linguistic organization that explain why change moves in certain directions and not others” (Bybee 2007: 981). Bybee draws attention especially to frequency effects on these dimensions. Particularly valuable in work on semantic change is the notion of collocation, or relationships among words or groups of words that go together. In a contextualized approach to the change in the meaning of conceit, for example, we find that from the beginning it was often associated with negative meanings. For example, of the five examples of conceit in Chaucer’s work, two are modified by wrong, and another is embedded in a context that suggests the opinion held is wanting: (1)

O sely preest! O sely innocent! With coveitise anon thou shalt be blent! O graceless, ful blynd is thy conceite (Canon’s Yeoman’s Tale 1076–1078) ‘Oh foolish priest! Oh foolish innocent! With covetousness you shall be blinded! Oh lacking God’s grace, fully blind is your opinion/mind’

Over time, although conceit could also be used in the sense of ‘good judgment’, the negative meaning became semanticized into conceit, presumably due to frequency of use in negative contexts. One of the interesting research questions is how some words become associated with either negative or positive contexts and whether or not these contexts become semanticized into the word. The phenomenon has been called “semantic prosody” (Stubbs 2001) and takes several shapes. One is illustrated by conceit, where a word is used so frequently in a negative context that the negative evaluation becomes part of the meaning. Others are preferred collocations, e.g. a shred of in the abstract quantifier sense ‘some’ has come to be used almost exclusively with not (hence it is understood as ‘not any’, and is often analyzed as a negative polarity item). Speakers often have preconceptions about such collocations, which may or may not be accurate. Invaluable for testing such preconceptions are computer-assisted

169

170

II Linguistic Levels approaches to corpora that provide “collostructional analysis”. Originally developed for synchronic work (Stefanowitsch and Gries 2003), this kind of analysis has been adapted for historical work on grammaticalization (Hilpert 2008), but could be extended to work on semantic change in general. It involves the exhaustive extraction of all tokens of a construction from a corpus while keeping one element constant, e.g. in the construction be going to+V, be going to with its verbal complements is kept constant. The objective is to determine not only which verbal complements it came to collocate with (this is “host-class expansion”, which Himmelmann 2004 considers criterial for grammaticalization), but also the strength of the association between them. It reveals changing selectional restrictions, hence fine-grained meaning change. It can be used to compare similar changes across languages, and to test schematic hypotheses about paths of change. For example, Hilpert (2008) tests Bybee et al.’s (1994: 270) hypotheses about the paths by which futures develop, one of them being (2): (2)

motion → intention → future

Collostructional evidence shows that Swedish komma at ‘come to’, although a motion verb that becomes a future marker, does not do so via intention. It also shows that in English be going to was initially used most often with speech act verbs like answer and begin, and became increasingly attracted to verbs with punctual meaning, especially those that are transitive with agentive meaning (e.g. get, marry), but also intransitive (e.g. die, leave). On the other hand, Dutch gaan ‘go’, which also became a future marker, was initially used primarily with motion verbs like laupen ‘walk’, and became increasingly attracted to intransitive, durative meanings (e.g. denken ‘think’, and voelen ‘feel’). Although both be going to and gaan exemplify (2), they have distinctly different semantic micro-histories, and hence different meanings.

5 Productivity of semantic changes at specific periods Not many studies have been conducted on semantic changes that affect a large class of items at a specific period of time. However, a few may be mentioned here. In one of the earliest attempts to demonstrate that semantic change can be patterned, Stern (1964 [1931]) showed that around 1300 adverbs meaning RAPIDLY developed the polysemy IMMEDIATELY in the context of perfective, i.e. punctual, verbs “denoting the action as a unit” (185–191), e.g. ME georne ‘rapidly’ > ‘immediately’. This change ceased, he said, around 1400, e.g. fleetly (1598) and rapidly (1727) itself did not undergo this change. Stern’s study is a non-quantified precursor of collostructional analysis, showing how change of meaning occurs metonymically in context. The change itself (which he calls a kind of “permutation”) is conceptualized as a case of what we would now call invited inferencing and unidirectionality of change: “it is evident that if a person rides rapidly up to another, the action is soon completed; but we cannot reverse the argument and say that if a person soon rides up to another, then the action is also rapidly performed” (Stern 1964: 186). A collostructional analysis would be needed to verify that this change was indeed as particular to the time as Stern claims. It is certainly crosslinguistically attested; Buck (1949: 964) commented that “the majority of words for ‘soon’ are, or were once, simply ‘quickly’ ” (for examples in Japanese see Traugott and Dasher 2002: 69). But that does not mean that there was a cluster of changes

11 Linguistic Levels: Semantics and lexicon that was highly productive at a particular point in time involving this particular semantic shift in any other language. Another example of a semantic change among adverbs that has been noted as having occurred particularly frequently at a given time is the development in Early Modern English of “boosters” or degree adverbs out of qualitative and highly evaluative adverbs (e.g. terribly, horribly, villainous – as now, some adverbs occurred without -ly). Peters (1994: 271) argues that most earlier boosters had developed from “local, dimensional or quantitative adverbs”, such as highly (dimensional) and vastly (quantitative). He hypothesizes that the development of boosters out of qualitative adverbs was associated with the development of more colloquial styles at the period. A third adverbial domain which is said to have become highly productive at a particular period is that of epistemic adverbs like probably, alledgedly, reportedly from the 18th century on. In recent work exploring the relationship between “cultural scripts” and semantics, Wierzbicka (2006) proposed that an increase in the number of epistemic adverbs is correlated with the advent of empiricism, especially under the influence of the philosopher John Locke’s An Essay Concerning Human Understanding (1690). Likewise, words like fairness and reasonableness underwent a significant shift. Using collocations as her evidence, Wierzbicka hypothesizes that our modern understanding of reasonableness includes common sense, a standard by which anyone’s behavior can be judged (108), and compatibility with reason (134), cf. reasonable doubt, reasonable care. However, prior to the enlightenment it meant ‘required by reason’ (134), or ‘having reason’, as in (3): (3)

Man is a resonable two foted beest (c.1380–87 Chaucer, Boece V. pr. iv. 128; Wierzbicka 2006: 109)

Testing Wierzbicka’s hypothesis that prior to the British enlightenment there was a cultural script of faith and certainty, Bromhead (2009) investigates a number of expressions like I think, in truth, verily, some of which, like verily, are no longer used, or recessive (in truth). She supports Wierzbicka’s conclusions in general, by providing evidence that the meanings of these expressions prior to the 18th century predominantly expressed certitude and confidence, rather than the doubt associated with the modern empiricist ethos. However, she also shows that the meaning changes Wierzbicka identifies were developing prior to the 18th century and the appearance of Locke’s book. This suggests that he was a synthesizer of current views, as well as a catalyst for their spread.

6 Differences between lexical and grammatical changes While there are similarities with respect to metaphor and metonymy between changes affecting lexical and grammatical (or grammaticalizing) expressions, there are also differences. Most notably, lexical semantic change concerns contentful change, whereas the development of grammatical expressions (see Section 2) is associated with “bleaching” or loss of contentful, lexical meaning: “[a]s grammatical morphemes develop, they lose specific features of meaning and thus are applicable in a wider range of environments” (Bybee 2007: 975). As Peters’s (1994) examples of boosters deriving from qualititative adverbs illustrates, grammaticalization may result in such collocations as

171

172

II Linguistic Levels terribly happy, or more recently pretty ugly. Bleaching is not found exclusively in grammaticalization. Occasionally lexical items may also lose in substantive content, e.g. OE þing ‘law court, assembly’ by metonymy > ‘thing, matter of concern’. But grammaticalization does not only involve loss of lexical meaning. There is enrichment of grammatical meaning as the original abstract implicature is semanticized, thus terribly lost the lexical meaning of ‘terror’, but gained abstract scalar meaning placing its complement high on the scale; pretty lost the lexical meaning ‘good-looking’, but gained scalar meaning, serving to place its complement above the median on a scale of intensity. Other respects in which semantic change associated with grammaticalization differs from that associated with lexical items is that it is more frequently replicated, and usually cross-linguistically attested. Subjectification occurs in lexical change (see Section 3 above) but it is particularly closely associated with grammaticalization because grammatical markers serve to indicate the speaker’s perspective on who does what to whom (case), how the situation is related to speech time (deictic tense) or to the temporality of a reference point other than speech time (relative tense), whether the situation is perspectivized as continuing and open-ended or not (aspect), whether the situation is relativized to the speaker’s beliefs (modality, mood), and how utterances are connected to each other (connectives, discourse markers), among other things (Traugott 2010).

7 Changes in the lexicon As indicated at the end of Section 1, the lexicon is the inventory of form-meaning pairs in a language. These pairs may be conceptualized as on a continuum from substantive, contentful, or “lexical”, like car, shoot, to “grammatical”, procedural, and indexical, like but, although. In the lexicon, meaning is often accounted for in terms of lexical relations, such as synonymy, antonymy, hyponymy/hyperonymy (taxonomic relations of member/superordinate set, e.g. carrots/vegetables), and meronymy/holonymy (part of/ whole of, e.g. finger/hand). Since the inventory of lexical and grammatical items that constitutes the lexicon consists of abstract structures without any inflections, the term “word” is not usually used in discussion of the lexicon. Instead, “lemma” is used for abstract lexemes, and sometimes “gram” for abstract grammatical items. It is usually assumed that the lexicon contains only items that are conventionalized in the sense that they are used by more than one speaker. However, it is particularly difficult in historical work on the earlier periods to know when this criterion is met, since if an item occurs only once, we cannot know whether this is because it was a nonce-item used by just one speaker, or whether it happens to appear in only one surviving manuscript. The debate has been particularly lively with respect to kennings in Old English. Some, especially those related to the sea, recur and may be considered candidates for the lexicon of Old English, e.g. swa¯nra¯d ‘swan road’ (see Kay, Chapter 20); many others may be nonce-forms. Kennings are usually thought of as metaphors constructed as compounds, but Broz (forthc.) analyzes them as complex metonymy-metaphor combinations. He illustrates with a productive process of forming kennings that denote the concept of ‘body’. In these, ba¯n ‘bone’ is the first element, as in ba¯ncofa ‘bone-chamber’, ba¯nfæt ‘bone-container’ or ba¯nhu¯s ‘bonehouse’. Broz suggests that several cognitive operations are at work here: a containment image schema which gives rise to conceptual metaphors such as BODY IS

11 Linguistic Levels: Semantics and lexicon CONTAINER (Lakoff 1987), which combines with the PART FOR WHOLE metonymy, bone being the essential part of the body. Changes affecting items in the lexicon clearly include semantic change, but research into the lexicon also concerns changes to the form of lexemes (e.g. hla¯f weard ‘loaf guardian’ > lord) (see Brinton, Volume 2, Chapter 100), and such factors as changes in the size of the lexicon and the consequences of borrowing. The latter two topics provide insight into what meanings were salient at particular periods, and in particular text-types, and will be the focus of this section. The size of the English lexicon at any period is hard to measure. The only criterion is the number of lemmas and grams listed in dictionaries, a factor highly dependent on tradition, and the purposes of a dictionary (see Kay, Chapter 20 and Dossena, Chapter 55 on some of the problems attendant on estimating increases in vocabulary size). Practices differ with respect to whether and when polysemies are counted as separate entries. When there is an academy that regulates language practice, as in France, a dictionary is less likely to include new words and meanings than when there is no such academy. That said, English speakers have been more willing than speakers of many other languages to borrow large amounts of vocabulary. The influence of Scandinavian, French, Latin, and Greek is widely known, but Arabic, Spanish, and more recently Japanese are among the many others that have also contributed to the inventory. No one speaker and no one register makes use of more than a tiny subset of all the possible lexical lemmas recorded in a dictionary. Dictionaries represent the aggregate of items used or usable by the totality of those regarded as speakers of a language. Since English speakers have been borrowing and inventing words for over a millennium and a half, and surprisingly few words are dropped from collective, as opposed to individual, inventories, the size of the vocabulary has been increasing since the beginning. As might be expected, such increases are not independent of periods of contact or new discourses and lifestyle changes (see borrowing in the 19th century of jihad for political and religious discourse, and of sushi for life-style discourse). Perhaps the most interesting increase, as shown by the Chronological English Dictionary, occurred in the period 1570–1630 (see discussion in Nevalainen 1999 and in Lancashire, Chapter 40). In this case it resulted not from contact, but from a conscious effort to shape English as a national language, no longer second to French or Latin, and from the explosion of new literary works, including the plays of Shakespeare. Many of the words he used never became conventionally used in English e.g. offendress ‘woman who offends’. Words were borrowed, especially if they were technical (cerebellum, specimen); derivational affixes were added to native or borrowed words (uncertitude; see in contemporary English u¨berlame); converted from one part of speech to another, e.g. from noun to verb (calendar ‘enter on one’s calendar’), or coined (giggle). When borrowing occurs, various kinds of consequences can be noted. Narrowing of one of the terms may occur if a native term already exists: compare the generic term cow (OE cu) with beef ‘kind of meat’ (Fr. boeuf ). Another is that the vocabulary may become multidimensionally stratified. Dimensions of variation that have received particular attention are region, social group, field of discourse (transaction, homily, letter), medium (spoken or written), and attitude (Kastovsky 2006). An example of social and age stratification is some contemporary teenagers’ use of intensifiers like dead, often in collocations that differ from those of standard varieties of English such as dead healthy, pure funny, enough funny (e.g. Stenstro¨m 2000; Macaulay 2005).

173

174

II Linguistic Levels Yet another consequence of borrowing is the coexistence of typologically different layers of vocabulary. Because many words have been borrowed into English from French, the lexicon abounds in words that have phonological alternations in the base (e.g. hı´story, histo´ric, historı´city), whereas native words do not (Kastovsky 2006: 212). The lexicon also includes verbs that have been said to be typologically very different from a semantic point of view. Different languages package semantic material into words in different ways (Lehrer 1992: 249). Talmy (e.g. 2000) proposed that there are two basic ways of expressing semantic path and manner/cause of motion in the verb. He distinguishes Chinese and all branches of Indo-European except Romance as languages that encode Motion and Manner/Cause together and treat Path as satellite. In Jill floated into the cave, float conflates Motion and Manner, in The napkin blew off the table, blow conflates Motion and Cause (wind). By contrast Romance, Semitic, Polynesian, and Navajo encode Motion and Path together and treat Manner as satellite, as in Spanish La botella entro´ a la cueva flotando ‘The bottle entered into the cave floating’. Because French words have been borrowed, we can use both enter (French), which treats Manner as satellite, and swim, which treats Path as satellite. Very interestingly, Latin, like English, encoded Motion and Manner/ Cause. Stolova (2008) attributes the change from Latin to Romance largely to the loss in Late Latin of verb prefixes that encoded Path, cf. abire ‘out-go’, and replacement of them in the Romance languages by verbs derived from nouns that had connotations of direction in them, e.g. Fr. monter ‘go up, climb’ < mons ‘mountain’. Verbs like entrare (< intra ‘into’) that had already become univerbated in Late Latin were not understood as encoding path morphologically, and so were treated like monter, as inherently including direction within them. This means that borrowings into English from Latin such as exit ‘outgo’, ascend (ad-scandere ‘at climb’) are typologically like English go up, while mount, enter from French are typologically different – we do not usually mount up a hill, though debts may mount up (a cumulative, not directional Path use of up), nor do we usually enter into a cave, although we may enter into an agreement, a conduit metaphor).

8 Future prospects Theoretical work on semantics, pragmatics, and the lexicon is increasing exponentially from many perspectives, among them computational, quantificational, typological, socio-cultural, and rhetorical. We can therefore expect substantial advances in research on semantic and lexical change in English in the near future. As these developments occur, the question of methodology in the use not only of electronic corpora, but also of dictionaries such as the OED will come to be of ever-increasing importance (see Allan and Robynson forthc.).

9 References Allan, Kathryn L. 2008. Metaphor and Metonymy: A Diachronic Approach. Malden, MA: WileyBlackwell. Allan, Kathryn and Justyna Robinson (eds.). forthc. Current Methods in Historical Semantics. Berlin/New York: De Gruyter Mouton. Anttila, Raimo. 1989 [1972]. Historical and Comparative Linguistics. Amsterdam/Philadelphia: Benjamins. Athanasiadou, Angeliki, Costas Canakis, and Bert Cornillie (eds.). 2006. Subjectification: Various Paths to Subjectivity. Berlin/New York: Mouton de Gruyter.

11 Linguistic Levels: Semantics and lexicon Barcelona, Antonio. 2000. On the plausibility of claiming a metonymic motivation for conceptual metaphor. In: Antonio Barcelona (ed.), Metaphor and Metonymy at the Crossroads: A Cognitive Perspective, 31–58. Berlin/New York: Mouton de Gruyter. Benveniste, Emile. 1971 [1958]. Subjectivity in language. In: Problems in General Linguistics, 223– 230. Trans. by Mary Elizabeth Meek. Coral Gables, FL: University of Miami Press. (Originally published as ‘De la subjectivite´ dans le langage’. Journal de psychologie 55 [1958]: 267–276.) Bre´al, Michel. 1897. Essai de Se´mantique (Science des Significations). Paris: Hachette. (English trans. by Mrs. Henry Cust, Semantics: Studies in the Science of Meaning. New York: Dover, 1900). Breban, Tine. 2008. Grammaticalization, subjectification and leftward movement of English adjectives of difference in the noun phrase. Folia Linguistica 42: 259–306. Bromhead, Helen. 2008. The Reign of Truth and Faith: Epistemic Expressions in 16th and 17th Century English. Berlin/New York: Mouton de Gruyter. Broz, Vlatko. forthc. Kennings: Riddles of metonymy or metaphor? In: Allan and Robynson (eds.). Buck, Carl Darling. 1949. A Dictionary of Selected Synonyms in the Principal Indo-European Languages. Chicago: University of Chicago Press. Bybee, Joan. 2007. Diachronic linguistics. In: Geeraerts and Cuyckens (eds.), 945–987. Bybee, Joan, Revere Perkins, and William Pagliuca. 1994. The Evolution of Grammar: Tense, Aspect, and Modality in the Languages of the World. Chicago: University of Chicago Press. Campbell, Lyle. 2004. Historical Linguistics. Cambridge, MA: MIT Press. Davidse, Kristin, Lieven Vandelanotte, and Hubert Cuyckens (eds.). 2010. Subjectification, Intersubjectification and Grammaticalization. Berlin/New York: De Gruyter Mouton. Dı´az Vera, Javier E. (ed.). 2002. A Changing World of Words: Studies in English Historical Lexicography, Lexicology and Semantics. Amsterdam/New York: Rodopi. Eckardt, Regine. 2006. Meaning Change in Grammaticalization: An Enquiry into Semantic Reanalysis. Oxford: Oxford University Press. Evans, Nicholas and David Wilkins. 2000. In the mind’s ear: The semantic extensions of perception verbs in Australian languages. Language 76: 546–592. Fortson, Benjamin W., IV. 2003. An approach to semantic change. In: Brian D. Joseph and Richard D. Janda (eds.), The Handbook of Historical Linguistics, 648–666. Oxford: Blackwell. Geeraerts, Dirk. 1997. Diachronic Prototype Semantics: A Contribution to Historical Lexicology. Oxford: Clarendon Press. Geeraerts, Dirk and Hubert Cuyckens. 2007. Introducing Cognitive Linguistics. In: Geeraerts and Cuyckens (eds.), 3–21. Geeraerts, Dirk and Hubert Cuyckens (eds.). 2007. The Oxford Handbook of Cognitive Linguistics. Oxford: Oxford University Press. Geis, Michael L. and Arnold M. Zwicky. 1971. On invited inferences. Linguistic Inquiry 2: 561–566. Grice, H. Paul. 1989. Logic and conversation. Studies in the Way of Words, 22–40. Cambridge, MA/ London: Harvard University Press. (first published in Peter Cole and Jerry L. Morgan [eds.], Syntax and Semantics III: Speech Acts, 41–58. New York: Academic Press, 1975). Grondelaers, Stefan, Dirk Speelman, and Dirk Geeraerts. 2007. Lexical variation and change. In: Geeraerts and Cuyckens (eds.), 988–1011. Haspelmath, Martin. 1997. Temporal Adverbials in the World’s Languages. Munich: LINCOM EUROPA. Heine, Bernd, Ulrike Claudi, and Friederike Hu¨nnemeyer. 1991. Grammaticalization: A Conceptual Framework. Chicago: University of Chicago Press. Heine, Bernd and Tania Kuteva. 2002. World Lexicon of Grammaticalization. Cambridge: Cambridge University Press. Hilpert, Martin. 2008. Germanic Future Constructions: A Usage-based Approach to Language Change. Amsterdam/ Philadelphia: John Benjamins. Himmelmann, Nikolaus P. 2004. Lexicalization and grammaticalization: Opposite or orthogonal? In: Walter Bisang, Nikolaus P. Himmelmann, and Bjo¨rn Wiemer (eds.), What Makes Grammaticalization – A Look from its Fringes and its Components, 19–40. Berlin/New York: Mouton de Gruyter.

175

176

II Linguistic Levels Hock, Hans Henrich and Brian D. Joseph. 1996. Language History, Language Change, and Language Relationship: An Introduction to Historical and Comparative Linguistics. Berlin/New York: Mouton de Gruyter. Horn, Laurence R. 1984. Toward a new taxonomy for pragmatic inference: Q-based and R-based implicature. In: Deborah Schiffrin (ed.), Meaning, Form, and Use in Context: Linguistic Applications, 11–42. (Georgetown University Round Table ’84.) Washington, DC: Georgetown University Press. Kastovsky, Dieter. 2006. Vocabulary. In: Richard Hogg and David Denison (eds.), A History of the English Language, 199–270. Cambridge: Cambridge University Press. Kay, Christian, Jane Roberts, Michael Samuels, and Irine´ Wotherspoon (eds.). 2009. Historical Thesaurus of the Oxford English Dictionary. Oxford: Oxford University Press. Keller, Rudi. 1994. On Language Change: The Invisible Hand in Language. London: Routledge. (Originally published as Sprachwandel: Von der unsichtbaren Hand in der Sprache. Tu¨bingen: Francke, 1990.) Ko¨vecses, Zolta´n and Gu¨nter Radden. 1998. Metonymy: Developing a cognitive linguistic view. Cognitive Linguistics 9: 37–77. Krug, Manfred G. 2000. Emerging English Modals: A Corpus-based Approach to Grammaticalization. Berlin/New York: Mouton de Gruyter. Lakoff, George. 1987. Women, Fire, and Dangerous Things: What Categories Reveal about the Mind. Chicago: University of Chicago Press. Langacker, Ronald W. 1990. Subjectification. Cognitive Linguistics 1: 5–38. Langacker, Ronald W. 2006. Subjectification, grammaticalization, and conceptual archetypes. In: Athanasiadou, Canakis, and Cornillie (eds.), 17–40. Lehrer, Adrienne. 1992. A theory of vocabulary structure: Retrospectives and prospectives. In: Mario Pu¨tz (ed.), Thirty Years of Linguistic Evolution: Studies in Honour of Rene´ Dirven on the Occasion of his Sixtieth Birthday, 243–256. Berlin/New York: Mouton de Gruyter. Levinson, Stephen C. 2000. Presumptive Meanings: The Theory of Generalized Conversational Implicature. Cambridge, MA: MIT Press. Macaulay, Ronald K. S. 2005. Talk that Counts: Age, Gender, and Social Class Differences in Discourse. Oxford: Oxford University Press. Nevalainen, Terttu. 1999. Lexis and semantics. In: Roger Lass (ed.), The Cambridge History of the English Language. Vol. III. 1476–1776, 332–458. Cambridge: Cambridge University Press. Panther, Klaus-Uwe and Linda L. Thornburg. 2003. Metonymies as natural inference and activation schemas: The case of dependent clauses as independent speech acts. In: Klaus-Uwe Panther and Linda L. Thornburg (eds.), Metonymy and Pragmatic Inferencing, 127–147. Amsterdam/ Philadelphia: Benjamins. Peters, Hans. 1994. Degree adverbs in Early Modern English. In: Dieter Kastovsky (ed.), Studies in Early Modern English, 269–288. Berlin/New York: Mouton de Gruyter. Pustejovsky, J. 1995. The Generative Lexicon. Cambridge, MA: MIT Press. Reddy, Michael J. 1993 [1979]. The conduit metaphor: A case of frame conflict in our language about language. In: Andrew Ortony (ed.), Metaphor and Thought, 164–201. Cambridge: Cambridge University Press. Sperber, Dan and Dierdre Wilson. 1995. Relevance: Communication and Cognition. 2nd edn. Oxford: Blackwell. Stefanowitsch, Anatol and Stefan Th. Gries. 2003. Collostructions: Investigating the interaction between words and constructions. International Journal of Corpus Linguistics 8: 209–243. Stenstro¨m, Anna-Britta. 2000. It’s enough funny, man: Intensifiers in teenage talk. In: John M. Kirk (ed.), Corpora Galore: Analyses and Techniques in Describing English, 177–190. Amsterdam/Atlanta, GA: Rodopi. Stern, Gustaf. 1964. Meaning and Change of Meaning: With Special Reference to the English Language. Bloomington/London: Indiana University Press. (Reprint of Elanders Boktryckeri Aktiebolag, Go¨teborg, 1931)

12 Linguistic Levels: Idioms and fixed expressions

177

Stolova, Natalya I. 2008. From satellite-framed Latin to verb-framed Romance: Late Latin as an intermediate stage. In: Roger Wright (ed.), Latin vulgaire, latin tardif: Actes du VIIIe`me Colloque International sur le Latin Vulgaire et Tardif, Oxford, 6–7 Septembre 2006, 253–262. Hildesheim: Georg Olms Verlag. Stubbs, Michael. 2001. Words and Phrases: Corpus Studies of Lexical Semantics. Oxford: Blackwell. Sweetser, Eve E. 1990. From Etymology to Pragmatics: Metaphorical and Cultural Aspects of Semantic Structure. Cambridge: Cambridge University Press. Talmy, Leonard. 1988. Force dynamics in language and cognition. Cognitive Science 2: 49–100. Talmy, Leonard. 2000. Toward a Cognitive Semantics. 2 vols. Cambridge: Cambridge University Press. Traugott, Elizabeth Closs. 1989. On the rise of epistemic meanings in English: An example of subjectification in semantic change. Language 65: 31–55. Traugott, Elizabeth Closs. 2010. (Inter)subjectivity and (inter)subjectification: A reassessment. In: Davidse, Vandelanotte, and Cuyckens (eds.), 29–70. Traugott, Elizabeth Closs and Richard B. Dasher. 2002. Regularity in Semantic Change. Cambridge: Cambridge University Press. Traugott, Elizabeth Closs and Ekkehard Ko¨nig. 1991. The semantics-pragmatics of grammaticalization revisited. In: Elizabeth Closs Traugott and Bernd Heine (eds.), Grammaticalization, Vol. 1, 189–218. Amsterdam/Philadelphia: John Benjamins. Ullmann, Stephen. 1964. Semantics; An Introduction to the Science of Meaning. Oxford: Blackwell. Warren, Beatrice. 1992. Sense-Developments: A Contrastive Study of the Development of Slang Sense and Novel Standard Sense in English. Stockholm: Almqvist & Wiksell. Warren, Beatrice. 1999. Laws of thought, knowledge and lexical change. In: Andreas Blank and Peter Koch (eds.), Historical Semantics and Cognition, 215–243. Berlin/New York: Mouton de Gruyter. Wierzbicka, Anna. 2006. English: Meaning and Culture. Oxford: Oxford University Press.

Elizabeth Closs Traugott, Palo Alto (USA)

12 Linguistic Levels: Idioms and fixed expressions 1 2 3 4 5 6 7 8

Introduction The concept and scope of phraseology and phraseological research Approaching historical English phraseology Metalinguistic sources and their value for the identification of phraseological units in historical texts Origin, change, and earlier uses of English phraseological units The impact of phraseology and collocation / string frequency on language change Summary References

Abstract In order to establish a framework of reference for an approach to historical English phraseology, a brief discussion of the concept and scope of phraseology and phraseological scholarship is provided in Section 2. On this basis, Section 3 approaches historical phraseology in two consecutive steps – always paying special regard to English: an overview of the state of historical phraseology leads to major approaches, which establish the main Alex Bergs and Laurel Brinton 2012, Historical Linguistics of English (HSK 34.1), de Gruyter, 177–196

178

II Linguistic Levels procedure in Sections 4 to 6. In Section 4, the role of metalinguistic sources for historical phraseology is pointed out, especially with relation to the identification of phraseological units in historical texts. Theories of the origin, rise, and kinds of change of phraseological units as well as the employment of English phraseological units in literary texts are the subjects of Section 5. While Section 5 is thus concerned with change in the phraseological system over time, Section 6, vice versa, finishes off the article with an outlook on the influence of frequent word combinations and phraseology on language change in a broader perspective.

1 Introduction Since about the middle of the 20th century, a branch of linguistics that is concerned with the study of “idioms and fixed expressions” has developed in Europe under the name of “phraseology”. Being the oldest and most comprehensive systematic branch of linguistics dealing with “idioms and fixed expressions”, or “formulaic language” – which, however, still lacks a fully developed historical approach – it will provide the point of departure for this article. Owing to the developments in American and British linguistics in the past century, the branch of phraseology has only recently started to gain a footing there and to complement existing studies on specific aspects of formulaic language, such as idioms and conversational routines. In fact, the year 2007 stands out with a number of pertinent publications. However, only very few historical studies of English phraseology have been published as yet, while some publications which were not explicitly conceived as belonging to this branch can certainly be viewed from this perspective.

2 The concept and scope of phraseology and phraseological research The concept and scope of phraseology will be discussed by first approaching phraseology through formulating a description of phraseological units as its objects of study, giving some examples, and pointing out defining criteria and major properties of these units, in particular highlighting the factor of variation. In the main, this subchapter proceeds along lines which now seem to be accepted among phraseologists (cf., e.g., Cowie 1998) after a long period of heated discussion (cf., e.g., Welte 1992). Having thus outlined its content, one major classification of English phraseology that has been suggested will briefly be referred to in order to be able to locate findings of individual (historical) studies within a descriptive system of phraseology. Finally, a rough outline of established approaches of phraseological research will be briefly indicated in order to be able to determine a place for historical (English) phraseology.

2.1 Defining criteria and major properties of phraseological units “Phraseological unit” seems to be one of the most widely accepted English terms for the linguistic units studied by scholars of phraseology. A description may be given as follows: Phraseological units are semantically and/or pragmatically fixed units in a language which consist of two or more smaller units that are either lexemes of that language

12 Linguistic Levels: Idioms and fixed expressions outside these expressions or recognized as unique lexemes within them, and that together do not exceed sentence length. Therefore, high frequency structural units such as the phraseological unit as well in example (1), and pragmatic markers as for instance in (2), belong to phraseology just like infrequent idioms as in (3), or proverbs as in (4). (1)

as well ‘too; also’ (All meaning equivalents quoted in this article to explain phraseological units are taken from Cowie et al. (1983) and Cowie and Mackin (1993), if not stated otherwise.)

(2)

you know ‘I am informing, or reminding, you’ (but cf. also Brinton [1996: 42], who points out the following functions of you know: ‘indicating knowledge shared between speaker and hearer’, ‘indicating general knowledge’, and ‘presenting new information as if it were old information in order to improve its reception’)

(3)

jobs for the boys ‘the provision of paid employment for favoured groups within a hierarchy, profession, administration etc. (the implication being that the work of these groups is not really necessary)’

(4)

Let sleeping dogs lie ‘do not provoke, disturb or interfere with somebody/something that is giving no trouble though he/it might, or could, do so’

In linguistic analysis, phraseological units may be marked with regard to their semantics as in the much-quoted opaque expression (5), their lexis (6), syntax (7), or pragmatics (8). (5)

to kick the bucket ‘(informal) die’ (not equivalent to ‘to move one’s foot violently against a pail’)

(6)

to put the kibosh on something/somebody ‘dispose of finally, finish off, do for’ (cf. Oxford English Dictionary (Simpson [ed.] 2000–; henceforth OED), s.v. kibosh n.; the word kibosh (of obscure origin) is not used in general English)

(7)

to trip/dance/tread the light fantastic ‘(facetious) dance’ (< Come, and trip it as ye go / On the light fantastic toe, from John Milton’s L’Allegro; the phraseological unit lacks the noun)

(8)

an eye for an eye (and a tooth for a tooth) ‘(a warning that) an act of aggression will be met with retaliation of the same kind (especially in personal or national conflicts)’ (Exodus xxi. 22–24; the meaning is situationally determined)

Phraseological units may carry connotations and may have intensifying and emphatic functions in a text, as Gla¨ser (1986, 1998) points out from the point of view of stylistic

179

180

II Linguistic Levels analysis, and they can be fruitfully studied from a cultural perspective, too (cf. the contributions in Skandera 2007; for a recent study of the special features of phraseological units, cf. Dobrovo 0skij and Piirainen 2009). For their linguistic study it is important to consider that phraseological units, once they are implemented in a speech community, are reproduced in language use rather than appearing as newly produced structures each time they are employed. What makes them difficult to describe and may render them very hard to identify by means of computer processing in, especially historical, texts are at least four major factors: Phraseological units are idiomatic to varying degrees; idiomaticity is here understood semantically, referring to the relation of the meanings of the parts to the meaning of the whole expression (semantic [non-]compositionality). Cowie (1983: xiii) distinguishes between three broad categories of phraseological units primarily on the basis of semantic compositionality. Firstly, there are fully idiomatic phraseological units (pure idioms) such as example (5); they have a fixed form and no literal meaning. Secondly, figurative idioms such as the first example in (10) possess a literal and a figurative meaning. Thirdly, restricted collocations such as example (9) are characterized by a combination of one word which appears in a literal sense and another whose meaning is determined by the context. While some members of the third category allow lexical variation, it is rare in the second. (9)

to jog one’s/somebody’s memory ‘remind somebody of / about something; help or stimulate somebody to recall something’ (to jog is used in a transferred sense, memory is used literally)

The frequency of use differs for various types of phraseological units (cf. especially Moon 1998b: 88–89). Phraseological units may show unpredictable transformational behavior, for instance with regard to passivization (10). However, the syntactic behavior of individual expressions follows (strong) tendencies rather than strict rules (cf., e.g., Moon 1998b: 88–90). Thus, while the expression (10a) can be transformed (“⇒”) into (10b), for example, there is a strong tendency that the phraseological unit (10c) cannot be transformed into the passive (10d). (10) a. to spill the beans ‘(informal) give away information, deliberately or unintentionally’ ⇒ b. the beans were spilled; BUT c. to drive a hard bargain ‘(have the means, power, or cunning to) force a bargain; contrive an exchange of goods and services that is either unfair or to one’s own advantage’ ⇒ d. *a hard bargain is driven (the passive transformation is not possible according to Cowie et al. 1983; however, an unmonitored Google search in December 2011 for the expression in the affirmative, including different grammatical forms of the verb phrase, yielded over 3 million hits for the active construction, but also some 560 hits for the passive one, among them the following quotation from The Telegraph of 8 November 2006, “‘A hard bargain was driven over appearances,’ said Jennings of the transfer which paid an initial £5 million.” http:// www.telegraph.co.uk/sport/columnists/henrywinter/2349920/Walcott-well-upto-speed-on-his-learning-curve.html)

12 Linguistic Levels: Idioms and fixed expressions Above all, phraseological units possess different degrees of formal fixedness (see especially Moon 1998a: 120–177 and Moon 1998b: 90–96; see also the paragraph on idiomaticity, above): in addition to the realization of, e.g., one’s/somebody’s in (9), which depends on the context, and lexical variation and choice in (7) and (11), for instance, phraseological units may also be open to modification and show a striking breadth of creative employment for particular purposes. This may happen to such an extent that even an almost wholly deviant set of words may still be related to one particular phraseological unit (12) (cf. Fiedler 2007: 95–96; Gla¨ser 1998: 142–143). Along similar lines, finally, phraseological patterns such as the combination of subject and predicative constituent in the “Incredulity Response Construction” as exemplified in (13) are constructions with highly idiosyncratic meaning but not much lexical fixation (cf. Fleischer 1997: 130; Sailer 2007: 1069; Huddleston and Pullum 2002: 890, who term these “bare predication polar echo constructions”). This lack of fixedness may make phraseological units difficult to handle by both the speaker and the linguist. In addition, scholars of phraseology have to deal with individual, diastratic, and diatopic variation in a language’s store of phraseological units (“phrasicon”). (11) cap/hat in hand ‘uncovering the head as a sign of reverence, respect, courtesy’ (cf. OED, s.v. cap n.1, def. 4h and hat n., def. 5a; cap in hand in use since the 16th century, hat in hand attested in the 19th century) (12) Others’ Trash Is Now An Architect’s Treasure (The New York Times / Su¨ddeutsche Zeitung 13 June 2005; example taken from Fiedler 2007: 96. The sentence’s semantics, formal structure, and rhythm unambiguously relate it to the proverb One man’s meat is another man’s poison ‘what seems good or pleasing to one person may be bad or unsuitable for another’.) (13) Me worry?! Kim resign?! Her a genius?! (response construction expressing incredulity)

2.2 Classifying phraseological units In the light of such a wide and complex range of criteria pertaining to the characterization of phraseological units, it follows that in order to ascertain the scope of phraseology, a comprehensive, “neutral” descriptive framework of reference is called for. Rosemarie Gla¨ser’s attempt of 1986 will be introduced at this point (for brief versions in English, cf. Gla¨ser 1998; Knappe 2004: 15–22). Following the earlier Russian tradition and the discussion based on it, Gla¨ser adapts the system to English and distinguishes (beyond the polar notion of idiomaticity and non-idiomaticity as the primary division) between “word-like” units, called “nominations”, which form the centre of English phraseology, and “sentence-like” units, called “propositions”, which are found on the periphery. The nominations are classified according to word class and include, for instance, nominal phraseological units such as (3) and verbal phraseological units as found in (5), (6), (7), (9), and (10). An example of a proposition is, for instance, the proverb (4), but slogans, commandments, maxims, routine formulae, and quotations belong in this category, too. Between the nominations and the propositions, Gla¨ser establishes a class of “reduced propositions” which are nominations in form but on a

181

182

II Linguistic Levels deeper level are propositions, such as the fragment of a proverb (14) and the stereotyped comparison (15). These can overlap with the nominations, as in the case of irreversible binomials with and, which in (16) should be viewed as one nominal phraseological unit, and sometimes a fragment of a proverb may be hard to identify for native speakers. Although this system is rather comprehensive, further categories can be suggested (cf., e.g., Burger 2010: 33–58). (14) a new broom (< a new broom sweeps clean) ‘somebody recently appointed to office or a responsible post’ (starts with an energetic programme of reform and change, sometimes not welcomed by those already there) (15) as blind as a bat [is] ‘unable to see, or read, very easily’ (but usually not completely blind); (figuratively) unable to see, or perceive, something that is obvious to other people (two underlying propositions in, e.g., He is as blind as a bat: He is blind and A bat is blind) (16) bread and butter ‘livelihood’ (Gla¨ser 1986: 46) (Sub-)classifications of phraseology may of course fruitfully be tailored to the scholars’ research objectives. In order to give just one example, which is also interesting for historical lexicography, Rosamund Moon’s (1998a, 1998b) approach will be briefly referred to at this point. She works from a lexicographer’s perspective and uses corpus linguistic methods to study “fixed expressions and idioms”. Her typology answers the question of which phraseological entries should be included in a monolingual dictionary and to this end addresses three different kinds of non-compositionality in phraseological units: problems of lexico-grammar are inherent in anomalous collocations, such as the “cranberry collocation” in (6), dependence on the discoursal or situational interpretation marks the class of formulae, which is exemplified, for instance, by the “saying” (8), and finally three degrees of semantic transparency are distinguished, (5) being an example of an “opaque metaphor”. General difficulties of classifying phraseological units beyond their structure or syntactic function are apparent in Moon’s finding that one fourth of the c.6,700 phraseological units which she considers fall into more than one of these categories.

2.3 The study of phraseology, with particular regard to English Within the framework of phraseological scholarship, phraseological units have been investigated from a wide array of perspectives, which cannot be aptly summarized in this article. A handy overview of established approaches may be found in the 20 chapters of the two-volume Phraseologie/Phraseology, published in the series Handbooks of Linguistics and Communication Science (Burger et al. 2007). One important observation is that the number of English contributions in handbooks and collections of papers has increased considerably in recent years. This will certainly further research in English phraseology, as all scholars of English are addressed by the use of this metalanguage (from this perspective, compare Burger et al. 2007 to Burger et al. 1982; cf. also the collections of English articles by Cowie 1998 and Skandera 2007).

12 Linguistic Levels: Idioms and fixed expressions Norrick (2007) has summarized the state of phraseological research within the main linguistic theories and approaches of modern Anglo-American linguistics (cf. also Sick 1993: 17–54): He found that systematic treatment of phraseological units is rare (Norrick only mentions Makkai 1972; but systematic classifications can also be found in some other studies); American research discusses the issue of idioms within a generative framework; in the British tradition, collocation studies and corpus linguistics have contributed to phraseological research. On the other hand, the studies of English phraseology by Cowie (1983), Gla¨ser (1986, 1998) and Moon (1998a), as well as several contributions in Cowie (1998) such as those by Altenberg, Mel 0cˇuk, and Howarth, and also the monograph Idiom Structure in English by Makkai (1972), are to varying degrees indebted to the (Eastern) tradition of phraseology. Further important publications include, for example, studies from the perspectives of (first and second) language acquisition (e.g. Wray 2002 from a theoretical standpoint, and Nattinger and DeCarrico 1992 from the point of view of language teaching), proverbs (paremiology), pragmatics, metaphor, and psycholinguistics as well as cognitive linguistics, the matter of irreversible binomials and factors influencing the fixedness of word groups (cf. Hudson 1998), to mention but a short selection. Some of these studies show that different theoretical approaches such as stratificational grammar and idiom studies (Makkai), theories of the mental lexicon and formulaic language (Wray), or practical tools such as corpus linguistics and the investigation of phraseology (Moon, Altenberg, Mel 0cˇuk) can be combined in trying to explore the nature and use of idioms and fixed expressions in English.

3 Approaching historical English phraseology While the phraseological units of present-day English have been given increasing attention, the exploration of historical phraseology has not increased in like degree. In fact, it is still correct to speak of a Cinderella status of historical (English) phraseology: As far as the English collections are concerned, they do not include a single historical article apart from one contribution on the development of English proverb collections (Doyle 2007b; this is the sole focus of Doyle 2007a, too). In book-length studies of English phraseology, concern with the origin and diachronic change of phraseological units is expressed only briefly if at all: thus Welte (1992: 578) judges the historical aspect to be “fascinating” but does not elaborate on it, and Gla¨ser (1986: 51–53) devotes three pages to “diachronic aspects” of phraseology which owe their inclusion in her study of Phraseologie der englischen Sprache exclusively to their relevance for present-day usage, with emphasis on relic forms, idiomatization, and word-formation. In Fiedler’s (2007) coursebook, historical phraseology has no place at all. Where historical phraseology is treated in its own right in comprehensive surveys and collections – as in Burger et al. (1982: 315–382), Fleischer (1997: 244–246), Ha¨cki Buhofer and Burger (2006: 413– 465) and also in the new collection by Burger et al. (2007: 1078–1145, with contributions on English, German, French, Italian and the Slavic languages) – it is striking that it tends to be the final chapter (not, however, in Burger 2010: 129–154 and Thun 1978: 75–84). The systematic inclusion of historical phraseology in overviews may, however, be interpreted as a sign that scholars are getting ready to explore the place of phraseology in the history of a language in its own right. Historical phraseology with its two branches

183

184

II Linguistic Levels of synchronic investigation of past systems and diachronic investigation into the development of the language data both re-opens the panorama of phraseological research and at the same time demands methods particularly tailored to historical study. As far as studies specifically devoted to historical English phraseology are concerned, published studies include recent papers such as Howarth’s (2000) contribution, which highlights pertinent aspects of phraseological variation and change around a model of conventionality, concentrating on the period 1800–1900, and Cowie (2003) tests the possibilities of the OED for the study of English (historical) phraseology. Dictionaries of idioms and proverbs have also contributed to the study of English historical phraseology (cf. Section 4.2, below). Monographs and collected volumes on aspects of historical English phraseology are rare: examples are Hiltunen (1983), Brinton and Akimoto (1999), Claridge (2000), Moralejo Ga´rate (2003), Trousdale (2008) within the framework of construction grammar – all on aspects of “multi-word verbs”, in particular on the rise of “complex” and phrasal verbs; Prins (1952) on French influence on English phrases; Sontheim (1972) on proverbs and proverbial expressions; Weinstock (1966) on the use of formulaic language in Shakespeare; Knappe (2004) on phraseology in English language study to 1800. A large-scale survey of historical English phraseology has not been published yet. In fact, a great number of research questions are waiting to be explored. These are too numerous to be listed explicitly and comprehensively in this article. Rather, it is hoped that the following systematic (but necessarily not comprehensive) account of possibilities of English historical phraseology, which are presented together with selected references to existing studies, will help to lay a basis for its future exploration within an accepted system of “idioms and fixed expressions” (phraseology). Three major approaches to historical phraseology have been singled out in the past (cf. in particular Sialm et al. 1982; Burger and Linke 1998; also Howarth 2000), which will be looked at in the following sections: a. the study of metalinguistic sources such as proverb collections, dictionaries, and grammars and their contribution to the identification of phraseological units in historical texts (Section 4); b. the investigation of the origin and change of phraseological units (Section 5.1); and c. their usage and function in earlier texts (Section 5.2). Another promising area for future research is the study of the impact of phraseological units and frequent collocations on larger issues of language change (Section 6).

4 Metalinguistic sources and their value for the identification of phraseological units in historical texts Historical metalinguistic sources are relevant for historical English phraseology in a twofold way: first, as representing the scholarship of past ages they are the objects of the historiography of English phraseology both in its own right and with regard to the judgment of the reliability of the works as data sources. Second, both historical and modern collections of historical phraseological language data are, if they are judged to reflect actual language use, major aids in the identification of phraseological units in historical texts.

12 Linguistic Levels: Idioms and fixed expressions

4.1 The historiography of English phraseology A book-length survey of the place of phraseology in pre-19th-century English language study is Knappe (2004). Included are: English proverbs in – – special collections (cf. now also Doyle 2007a, who goes beyond 1800); and – – textbooks; as well as English phraseological units in – – stylistic treatises; – – handbooks for foreign-language teaching and contrastive language study; – – bilingual and multilingual lexicography with English as the first language entered; – – translation theory; – – philosophical and universal language schemes (including shorthand writing); – – monolingual English lexicography; – – monolingual English grammars; and – – early classifications of phraseological units. The analysis of the English material showed that particular types of phraseological units stand out as being especially interesting to the early textbook authors. These are proverbs, routine formulae, stereotyped comparisons, binomials, restricted collocations, phrasal verbs, figurative idioms, and idiomatic phraseological units in general. As early as from the time of the Restoration, and partly owing to the endeavours of the Royal Society (especially members of the Society such as John Ray and John Wilkins), scholars writing in the British Isles started approaching the topic in a more analytical way, developing classifications of the phraseological material (cf. also Knappe 2006c) which culminated in George Campbell’s Philosophy of Rhetoric (1776). Joseph Priestley in the 18th century can be credited with the emancipation of the notion of an “idiom” from its roots in language comparison. For the identification of phraseological units in historical texts it is important to know that the material discussed or entered in these textbooks and dictionaries reflects actual language use (cf. also Section 4.2, below). It is hoped that future studies will extend and complement the findings in Knappe (2004) and join them up with the beginning of the scholarly study of phraseology in the 20th century. In the long run, the history of English phraseology will have to be viewed in a European perspective (for German, cf. also Burger et al. 1982: 360–382 and Weickert 1997).

4.2 The identification of phraseological units in historical texts Several criteria have been singled out that may lead to a proper judgment of whether a given historical word combination – found by targeted reading, corpus search, or a combination of the two – may claim phraseological status (cf., e.g., Friedrich 2007: 1093; Burger in Sialm et al. 1982: 346-360). These criteria may be deduced both from linguistic considerations and from metalinguistic sources. Thus, a linguistic criterion may be the fact that a modern parallel of the expression exists, as in (11), the communicative function may betray a routine formula (17), formal criteria such as alliteration are typical of some expressions (18), word-formations are sometimes based on earlier

185

186

II Linguistic Levels phraseological units, such as the dephraseological formation in (19), the degree of idiomaticity (semantic non-compositionality) of the whole expression is a strong indicator of phraseological status, as in (20), and there are cases where the original composition of the phraseological unit has become opaque (21). The exact meaning of the components at particular times can be difficult to determine, and thus it may be hard to distinguish idioms from restricted collocations in past stages of the language (22). Finally, in the case of translated texts, word combinations which are (recurrently) different from the original may be further investigated for their phraseological status, and the high frequency of occurrence of structural units or pragmatic markers is an indication of phraseological status, too, as it is in present-day language. (17) How do you? ‘How are you?’ “God be thanked for you, How do you?”, quotation 1570 (cf. OED, s.v. do v., def. 19) > How do you do? (cf. OED, s.v. how adv., def. 2a and how-do-you-do, how-d’ye-do phr. and n.; first quotation: 1697) (18) kith and kin ‘country and kinsfolk’ > ‘acquaintance and kinsfolk’ > ‘relatives’ (cf. OED, s.v. kith n., def. 5) (19) fence-sitter < to sit on the fence (cf. Gla¨ser 1986: 52; cf. OED, s.v. fence n., def. C2 and sit v., def. B3c, in quotation 1887: “Those who sit ‘on the fence’ – men with impartial minds, who wait to see … ‘how the cat will jump’ ”; first occurrences: fence-sitter 1905, slightly earlier: fencesitting 1904) (20) to bear in/on hand ‘bring forth something wrong against someone’ (14th–16th century; cf. OED, MED, early dictionaries) (21) good-bye < God be with you/ye (cf. OED, s.v. good-bye) (22) at non hond ‘in no way, not at all’, figurative use of Middle English hond as ‘a position or direction to one side or the other’? (cf. MED, s.v. ho¯nd(e n., def. 7; last quotation of at no hand in OED, s.v. hand n1., def. 25g: 1690) Metalinguistic indications, such as the comments “literally” or “proverbial” in a text (e.g. 1835: “We came in literally neck and neck”; 1880: “The Burials Bill is thought to resemble the proverbial chip in porridge, which does neither good nor harm”; Howarth 2000: 226), or the inverted commas in the quotation in (19), are helpful for the identification of phraseological units to a certain extent, but reliable larger collections, which we possess from the 17th century onwards (and for proverbs also earlier than this), are certainly profitable resources for historical English phraseology: it was not only the analytic spirit which developed in Restoration England, but this period may also be seen as the time when collectors started to restrict themselves to the inclusion of authentic

12 Linguistic Levels: Idioms and fixed expressions English language material. Thus, from the 17th century onwards contrastive handbooks for foreign language teaching and dictionaries stand out as rich and to a large extent unexploited storehouses of language data, which in the future will be highly useful in the investigation of the historical development of English phraseological units. Late 19th, 20th, and early 21st-century historical lexicography, in particular the OED and the Middle English Dictionary (Kurath et al. 1952–2001; henceforth MED), which both are now searchable online, as well as efforts of phraseography (dictionaries of idioms and fixed expressions) and especially paremiography (dictionaries of proverbs) (cf. Apperson 1929; Tilley 1950; Whiting 1968; Mieder et al. 1992; Speake 2003) provide information on language data which is of paramount importance for the study of the historical development of phraseological units. As far as the great OED is concerned, Cowie (2003) found that much phraseological material can be retrieved from its text but that this material is often neither prominently displayed nor consistently handled. Earlier alphabetic and topical dictionaries can usefully complement these modern lexicographical endeavours, as has for instance been shown in a pilot study on Roget’s Thesaurus (cf. esp. Knappe 2006a, but also 2006b) – although here the exact meaning and potential formal variations of a given phraseological unit may not be found in the collections, but will have to be deduced from actual texts. Reliable and comprehensive collections of phraseological units from all periods of English are most important for the study of historical phraseology. Such an historical database of phraseological units can then fruitfully be employed as the point of departure for corpus-based investigation along the lines of the approach adopted in Moon’s research (1998a, 1998b), who gathered her material on the basis of the phraseological entries in the first edition of the Collins Cobuild English Language Dictionary of 1987 (cf. also Moon 2007). Even if researchers choose not to adopt this “consultation paradigm” which starts from a prefabricated list of phraseological units but prefer an “analysis paradigm” (on these terms, cf. Sailer 2007: 1064) by retrieving hypothetical phraseological units from text corpora by various means, a database of accepted forms will finally be necessary for the evaluation of the potentially phraseological findings. As far as the “analysis paradigm” is concerned, a large amount of work on the available electronic corpora and their annotations will still be necessary before phraseological units which do not match the criterion of frequent lexical co-occurrence can be retrieved from them (cf., e.g., also Degand and Bestgen 2003 on the automatic retrieval of idioms). The problems and challenges of corpus-based phraseological study outlined by Sailer (2007: 1067–1069) for the analysis of Present-day German, for instance with regard to the variation in phraseological units, will have to be reviewed from the point of view of historical (English) corpora. Problems relating to the rare occurrence of many phraseological units and the comparatively small size of the databases are of course even more severe for historical than for present-day databases.

5 Origin, change, and earlier uses of English phraseological units Next to the identification of phraseological units in older texts, historical phraseological research has addressed questions of the origins and change of phraseological units through time. Many of these points are connected to key issues of historical linguistics. In addition, the uses of phraseological units in literary texts have been addressed, mainly with the aim to assess the literary language of individual authors or the style

187

188

II Linguistic Levels of particular text types in the past. These studies, however, also contribute to our knowledge of the forms, meanings and functions of phraseological units at particular points in time.

5.1 Origin and change of phraseological units as linguistic signs It seems safe to say that all phraseological units, once they are accepted parts of the language, share the feature of “conventionality”. It seems also safe to say that at the basis of all phraseological units lies the pragmatic feature of “usefulness”. These “useful units”, then, may develop from several origins (cf. Howarth 2000). Some idiomatic phraseological units may have started out from a free or restricted collocation, such as suggested by the semantic change of the complex predicate in (23). In these cases, classification criteria in typologies which have been suggested for the description of the rise of a phraseological unit are the structure from which the unit developed and the kind of semantic change of the basis compared to the result (cf. Munske 1993). Other phraseological units such as proverbs (4), oaths or the metaphorical phraseological unit in (24) are probably coined expressions with no original literal use. Some phraseological units which are in common use today originate from literary texts, that is, they can be traced back to one creative mind, as in (7). Borrowing, too, is another source of phraseological units – a source not to be underestimated for English as a mixed language – such as the foreign phrase in (25) or, more commonly, adapted or translated ones (26). However, one has to be aware that it can be difficult to distinguish borrowed units from common-source and polygenetic ones. (23) to take steps ‘walk’ > ‘perform a move or moves in a course of action’ (17th century) > ‘take action or measures towards attaining an end’ (from 18th century) (cf. also OED, s.v. step n.1, def. 6d) (24) to be in (another person’s) shoes ‘be in his position or place’ (from 18th century according to OED, s.v. shoe n., def. 2k) (25) vice versa ‘contrariwise, conversely’ (first attestation: 1601; cf. OED, s.v. vice versa adv.) (26) to keep company ‘give a person one’s company’ (first attestation: 1509; < tenir compaignie, according to Prins 1952) In cases when phraseological units develop from free collocations, the question of how such a phraseological “fixation” arises must be addressed. For the explanation of the rise of English adverbial phraseological units, for example, Jean Hudson (1998) has suggested a cyclical model of fixation which draws on discussions in cognitive linguistics, pragmatics, discourse analysis, grammaticalization theory, and other research in language change. Her model is unidirectional and circular. Put in a nutshell, Hudson

12 Linguistic Levels: Idioms and fixed expressions proposes that at the level of discourse, an ad hoc expression, that is, an expression formed on the principle of open choice, turns into a unit with a different (new?) meaning through pragmatic inferencing. This unit is not yet fixed. Due to this pragmatic inferencing, however, salience reduction of the constituent parts occurs on the second level, which is conceptualization. Hudson replaces the notion of “semantic decompositionality” by “cognitive analyzability”, which depends on salience. Reduced salience, in its turn, fixes the expression to some degree on the level of realization. Symptoms of fixedness are unexpected syntactic constraints on the constituent parts, such as the other day but *the other days, and unexpected collocational restrictions within the expression, e.g. first of all but *second of all. On a less theoretical level, signs of a fixation process have, among other things, been named by Burger and Linke (1998: 747–750) as the reduction of lexical variants, of possibilities of modification, and of variety in morphosyntactic structure (cf. also Forga´cs 2004 on aspects of lexicalization). With regard to the choice and establishment of a phraseological unit’s canonical form in the history of English, Voitl (1969) has tentatively suggested influence of prescriptive linguistics. However, the tension between convention and variation, as well as creative exploitation, can readily be seen as typical of living phraseological units. It is obviously a major challenge for historical phraseology to distinguish synchronic variation from diachronic change. The investigation of the formal and semantic change of English phraseological units, too, addresses key issues of historical linguistics. In Burger and Linke (1998) and Friedrich (2007), several distinctive changes are discussed. These changes affect, for example, the order of the elements within the unit (27), the reduction or increase in the number of elements (28), the exchange of lexemes (29), the change in syntactic structure (17), the change in meaning and use (23), and finally the death of a phraseological unit (20). While examples for changes of these kinds can partly be suggested by the entries in the OED and MED, these entries are, after all, highly selective and not intended to give precise phraseological details. In-depth studies focussing on English phraseological developments, based on a wide coverage of lexicographical sources and the use of a large corpus of historical texts, are wanting. (27) heels over head (from 14th century) ‘upside down’ > head over heels (from 18th century, with development of figurative sense; cf. OED, s.v. heel n.1, def. 15a, head n.1, def. 46b) (28) a bird in hand [is better than / worth two in the wood / bush] (15th –18th century) > a bird in the hand […] (from 15th century) (cf. OED, MED, other dictionary sources) (29) to know (best) where one’s shoe wrings (one) ‘know where (a person’s) difficulty or trouble is’ (from 14th century; last quotation 1887) > to know (best) where one’s shoe pinches (one) (from 16th century; cf. OED, s.v. pinch v., def. 5b, shoe n., def. 2f, wring v., def. 4b) Restrictions of space forbid a detailed review of further studies pertaining to the origin of phraseological units and their change over time. Therefore, a selection of the

189

190

II Linguistic Levels published ones will only be briefly mentioned, and one unpublished study will be referred to in some more detail. Thus, for example, foreign influence on the rise of English phraseological units is the subject of Prins (1952), who devoted a book-length study and several articles to the French Influence in English Phrasing. The traces left by Latin phrases in Old English were investigated by Gneuss (1955), and Gustaf Stern (1931: 224) in his model of semantic change included fixed expressions which are influenced by foreign languages. Several formal subtypes of phraseological units (see Section 2.2, above) that are of specific interest to English linguistics have become the object of study, too. Thus, the structural and semantic rise and historical development of phrasal verbs (e.g. to give up) and complex predicates (e.g. to take a bath) have been given a rather large amount of attention (cf. the monographs cited in Section 3, above). The problem of differentiating noun + noun compounds from syntactic groups with phraseological status has been tackled by Marchand (1969: 20–30), also in a historical perspective. Among studies investigating subtypes are Ross’s (1975) article on “alliterative phrases” and Sontheim’s (1972) monograph addressed to the historical development of English proverbs and proverbial phrases. Apart from individual formal subtypes, the potential of smaller phraseological units, such as the adverbial phraseological units in hand and on hand, for example, as elements in the formation of larger phraseological units (Fleischer’s 1997 phraseologische Reihe ‘phraseological chain’), such as in to take in hand or a bird in hand [is better than / worth two in the wood / bush] (28), for example, can fruitfully be explored in the history of English, too. An as yet unpublished study by the present author based on lexicographical sources has uncovered 62 phraseological units in which in hand and on hand appear in the history of English. Two of the results will be briefly mentioned here: First, a “phraseological semantic force” (my own term) of in hand and on hand can be found, for instance in the development of take in hand. All main semantic features of in hand seem to have gained control over the unit take in hand almost simultaneously until the feature [PROCESS] prevailed from the 16th century on. Thus, from the literal meaning ‘in the hand’ the feature [PROXIMITY] took hold (e.g. ‘to take with one’), also [POWER, CONTROL] ‘to bring under control’, and finally [PROCESS] ‘to carry out’, ‘to undertake’. The second conclusion relates to the “phraseological binding force” (my own term) of adverbial phraseological units such as in hand, which may vary over time. According to my data, most of the new units with in hand and on hand were formed in the 16th century, a period marked by unprecedented interest in the creative potential and rhetorical force of the mother tongue, aided above all by a desire for rhetorical copia, that is, full and variable English expression (on the latter aspect, cf. Knappe 2004: 49–111). The historical development of phraseological units can also be explored from an onomasiological point of view. Thus Anders (1995) has studied lexical and phraseological realizations of the concept of “dying”, also in a historical perspective. And finally, connecting this section up to Section 5.2, the historical change in form and function of English routine formulae can be studied. Thus, Wyld (1936) included as the tenth chapter of his history of modern colloquial English a discussion of the historical development of routine formulae and finds, for instance, that greetings and farewells in the 16th century were less “stereotypical” than in his own day. A useful synchronic collection and classification from which systematic studies of routine formulae

12 Linguistic Levels: Idioms and fixed expressions from a diachronic perspective could start is the monograph on Conversational Routines in English by Karin Aijmer (1996).

5.2 The use of phraseological units in historical texts: aspects of pragmatics Phraseological units have always been employed by writers in a variety of functions, and canonical poetry, vice versa, can itself act as the source of phraseological units, as in (7) (cf. also MacKenzie 2003). In addition to finding out what the employment of formulaic language means for the literary text, it can also give evidence on the (creative) use of phraseology at a given time in the history of a language. The stylistically motivated employment of phraseological units in literary texts has been emphasized in the Russian and (Eastern) European study of historical phraseology. English writers of the past have selectively been studied, too. To start at the beginning, it is well-known that the oral style in medieval poetry is characterized by formulaic expressions (cf. the classical article by Magoun 1953). These have received much attention in scholarship, and are examples for the construction and transmission of artistic texts by help of pre-set formulas, sometimes employed creatively. In later texts, the use of proverbs in particular has been investigated (cf. especially the bibliography on international proverb scholarship by Mieder 1982). To mention just a few examples, Weinstock (1966) looked at Shakespeare’s use of proverbs and found that in his later works, the great writer employed proverbs to mark dramatic turns, and they also served to characterize persons. Reuter (1986) studied Deloney from this perspective, and both Brewer (1986: 229–232) and Windeatt (1992: 332–335, 345–354) comment on Chaucer’s use of set phrases (e.g. bold as blind Bayard), oaths (e.g., God so my soule save) and proverbs (e.g. The blind man cannot judge in colors) in the discussion of his poetic style. Thus, for instance, one of Chaucer’s favourite lines (adapted from the Italian) is For pitee renneth soone in gentil herte ‘because pity flows swiftly in a noble heart’, which he uses in the Canterbury Tales and in the Legend of Good Women to reinforce “our sense of humane values”, according to Brewer (1986: 230). Brewer continues to say that in using a large number of familiar phrases Chaucer created a traditionalist diction and aimed at familiar effects, ready communication, and sympathetic attention. The sententiousness in Chaucer’s Troilus and Criseyde, particularly the use of proverbs, however, which cluster at important points of persuasion and self-persuasion, creates distinct ambiguity which questions the limitations and value of prudential wisdom (cf. Windeatt 1992: 345–354). These are only a few examples of employment of phraseological language in literary texts. Again, more work needs to be done along these lines.

6 The impact of phraseology and collocation / string frequency on language change By way of an outlook one further interesting line of research in historical English phraseology will be mentioned: this is the effect that multi-word combinations can have on language change. To this end, frequently repeated word combinations (frequent

191

192

II Linguistic Levels collocations) and phraseological units will be discussed together. First, the common ground of both concepts will be addressed. A description of a phraseological unit was given above (Section 2.1). The use of “collocation” as addressed here needs further explanation. As defined by Sinclair (1991: 170), collocation is “the occurrence of two or more words within a short space of each other in a text”. Typically, these co-occurrences are frequently repeated or statistically relevant in a corpus (cf. also Bartsch 2004: 65; Moon 2007: 1046–1047). Research into conceptual structures may help us understand the common ground of both collocations as described above and phraseological units in language processing and use, which is important for the question of their influence on language change. Both frequent word combinations and phraseological units are characterized by their degree of entrenchment (cf. Harris 1998): it is either based on the form only or on the semantic/pragmatic unity of a string with certain formal characteristics. The effect of frequently recurring word form combinations (string-frequency) on language change has received attention in scholarship, such as seen in coalescence phenomena and the phonological effect of word-boundary liaison as the trigger for the “Great Vowel Shift” (cf. Krug 1998 and 2003 and Chapter 48). So if the frequent co-occurrence of lexical items affects language change, as research in string frequency claims, it may be suggested that phraseological units, too, may have an impact on it. Thus, Knappe and Schu¨mann (2006) have investigated the influence of both collocations (as described above) and phraseological units on the sudden switches of pronouns of address between the singular and the plural in the address of a single person in dialogues in Chaucer’s Canterbury Tales. About 90 such switchings occur in the text. The use of the plural pronoun for formal address was introduced on the model of French but was by the time of Chaucer not yet vigorously applied. For about one third of the cases of sudden pronoun switch we claim that a “collocational-phraseological” force was the trigger. This force is so strong that it may even override the pragmatically preferred choice of the pronouns of address in the particular situation (cf. Jucker, Chapter 13). For instance, terms for body parts (e.g. thy eyen) tend to co-occur with the singular pronoun of address and may provoke a pronoun switch. The verb prayen ‘I pray you/thee’ occurs more often with the plural (62 instances) than with the singular pronoun (14 instances). To give an example, the more common combination of pray + you may account for the shift from the singular to the plural in Absolon’s speech to Alisoun in the Miller’s Tale (3361–3362), as seen in (30): (30) “Now, deere lady, if thy wille be, I praye yow that ye wole rewe on me …” ‘ “Now, dear lady, if it be your will, I beseech you to have mercy on me …” ’ In (31) the formula used for the opening of sermons Heere may ye se ‘By this may you see, i.e. understand’ is a phraseological unit. In the Friar’s Tale (1567) it can account for the insertion of the plural form in the yeoman’s (i.e. the devil’s) address to the summoner (1566–1568):

12 Linguistic Levels: Idioms and fixed expressions (31) “Lo, brother,” quod the feend, “what tolde I thee? Heere may ye se, myn owene deere brother, The carl spak oo thing, but he thoghte another.” ‘ “Lo, brother”, said the devil, “what did I tell you? By this may you see, my own dear brother, the fellow spoke one thing, but he thought another” ’ (The quotations in [30] and [31] are taken from Benson [ed.] 1987.) Starting from here, the role of entrenched structures in the introduction of a distinctive system of pronouns of address in Middle English could be studied.

7 Summary To sum up, a great amount of work in the retrieval and analysis of phraseological material from the history of English still needs to be done. Some first studies and research from other languages and in different traditions of scholarship can provide a basis from which to start with the systematic investigation of the phraseological past of the English language. In the long run, the aim of scholarship in historical English phraseology should be the description of all facets of the origin and historical development of the phraseological system, or systems, of English. Moreover, at a time when the impact of “formulaic language” on language acquisition, use, and change is increasingly being studied in linguistic theory, historical English phraseology in its interaction with other related approaches of scholarship and methods such as collocation studies, corpus linguistics, cognitive linguistics, grammaticalization and lexicalization theories, and also construction grammar, will open up large, promising fields of research possibilities.

8 References Aijmer, Karin. 1996. Conversational Routines in English: Convention and Creativity. London: Longman. Anders, Heidi. 1995. “Never say die”– Englische Idiome um den Tod und das Sterben. Frankfurt: Peter Lang. Apperson, George Latimer. 1929. English Proverbs and Proverbial Phrases: A Historical Dictionary. 2 vols. London/Toronto: J. M. Dent and Sons; New York: Dutton & Co. Bartsch, Sabine. 2004. Structural and Functional Properties of Collocations in English: A Corpus Study of Lexical and Pragmatic Constraints on Lexical Co-occurrence. Tu¨bingen: Narr. Benson, Larry D. (ed.). 1987. The Riverside Chaucer. 3rd edn. Oxford: Oxford University Press. Brewer, Derek. 1986. Chaucer’s poetic style. In: Piero Boitani and Jill Mann (eds.), The Cambridge Chaucer Companion, 227–242. Cambridge: Cambridge University Press. Brinton, Laurel J. 1996. Pragmatic Markers in English: Grammaticalization and Discourse Functions. Berlin/New York: Mouton de Gruyter. Brinton, Laurel and Minoji Akimoto (eds.). 1999. Collocational and Idiomatic Aspects of Composite Predicates in the History of English. Amsterdam/Philadelphia: John Benjamins. Burger, Harald. 2010. Phraseologie: Eine Einfu¨hrung am Beispiel des Deutschen. 4th edn. Berlin: Erich Schmidt. Burger, Harald, Annelies Buhofer, and Ambros Sialm. 1982. Handbuch der Phraseologie. Berlin/ New York: Mouton de Gruyter.

193

194

II Linguistic Levels Burger, Harald, Dmitrij Dobrovo 0skij, Peter Ku¨hn, and Neal R. Norrick (eds.). 2007. Phraseologie/ Phraseology: Ein internationales Handbuch der zeitgeno¨ssischen Forschung / An International Handbook of Contemporary Research. 2 vols. Berlin/New York: Mouton de Gruyter. Burger, Harald and Angelika Linke. 1998. Historische Phraseologie. In: Werner Besch, Anne Betten, Oskar Reichmann, and Stefan Sonderegger (eds.), Sprachgeschichte: Ein Handbuch zur Geschichte der deutschen Sprache und ihrer Erforschung. 2nd ed. Vol. I, 743–755. Berlin/ New York: Walter de Gruyter. Claridge, Claudia. 2000. Multi-word Verbs in Early Modern English: A Corpus Based Approach. Amsterdam/Atlanta, GA: Rodopi. Cowie, A. P. 1983. General introduction. In: Cowie, Mackin, and McCaig, Vol. II, x-xvii. Cowie, A. P. (ed.). 1998. Phraseology: Theory, Analysis, and Applications. Oxford: Clarendon Press. Cowie, A. P. 2003. Some aspects of the treatment of phraseology in the OED. In: Cornelia Tschichold (ed.), English Core Linguistics: Essays in Honour of D. J. Allerton, 205–224. Bern/Berlin: Peter Lang. Cowie, A. P., R. Mackin, and I. R. McCaig. 1983. Oxford Dictionary of Current Idiomatic English. Vol. II: Phrase, Clause & Sentence Idioms. Oxford: Oxford University Press. [Repr. 1993 under the title: Oxford Dictionary of English Idioms] Cowie, A. P. and R. Mackin. 1993. Oxford Dictionary of Phrasal Verbs. Oxford: Oxford University Press. [= 2nd edn. of Oxford Dictionary of Current Idiomatic English, Vol. I. Verbs with Prepositions & Particles] Degand, Liesbeth and Yves Bestgen. 2003. Towards automatic retrieval of idioms in French newspaper corpora. Literary and Linguistic Computing 18: 249–259. Dobrovo 0skij, Dimitrij, O. and Elisabeth Piirainen. 2009. Zur Theorie der Phaseologie: Kognitive and kulterelle Aspekte. Tu¨bingen: Staufferburg. Doyle, Charles Clay. 2007a. Historical phraseology of English. In: Burger, Dobrovo 0skij, Ku¨hn, and Norrick (eds.), 1078–1092. Doyle, Charles Clay. 2007b. Collections of proverbs and proverb dictionaries: Some historical observations on what’s in them and what’s not (with a note on current “gendered” proverbs). In: Skandera (ed.), 181–204. Fiedler, Sabine. 2007. English Phraseology: A Coursebook. Tu¨bingen: Narr. Fleischer, Wolfgang. 1997. Phraseologie der deutschen Gegenwartssprache. 2nd edn. Tu¨bingen: Niemeyer. Forga´cs, Tama´s. 2004. Grammatikalisierung und Lexikalisierung in phraseologischen Einheiten. In: Christine Palm-Meister (ed.), Europhras 2000: Internationale Tagung zur Phraseologie vom 15.–18. Juni 2000 in Aske / Schweden, 137–149. Tu¨bingen: Stauffenburg. Friedrich, Jesko. 2007. Historische Phraseologie des Deutschen. In: Burger, Dobrovo 0skij, Ku¨hn, and Norrick (eds.), 1092–1106. Gla¨ser, Rosemarie. 1986. Phraseologie der englischen Sprache. Leipzig: VEB Verlag Enzyklopa¨die. Tu¨bingen: Niemeyer. Gla¨ser, Rosemarie. 1998. The stylistic potential of phraseological units in the light of genre analysis. In: Cowie (ed.), 125–143. Gneuss, Helmut. 1955. Lehnbildungen und Lehnbedeutungen im Altenglischen. Berlin: Erich Schmidt. Ha¨cki Buhofer, Annelies and Harald Burger (eds.). 2006. Phraseology in Motion I: Methoden und Kritik. Akten der Internationalen Tagung zur Phraseologie (Basel, 2004). Baltmannsweiler: Schneider-Verlag Hohengehren. Harris, Catherine L. 1998. Psycholinguistic studies of entrenchment. In: Jean-Pierre Koenig (ed.), Conceptual Structure, Discourse and Language. Vol. II. Stanford, CA: CSLA. http://kybele. psych.cornell.edu/~edelman/Psych-426/Harris97-psycholinguistic-entrenchment.pdf Hiltunen, Risto. 1983. The Decline of the Prefixes and the Beginnings of the English Phrasal Verb: The Evidence from Some Old and Early Middle English Texts. Turku: Turun Yliopisto. Howarth, Peter. 2000. Describing diachronic change in English phraseology. In: Gloria Corpas Pastor (ed.), Las lenguas de Europa: Estudios de fraseologı´a, fraseografı´a y traduccio´n, 213– 230. Granada: Editorial Comares.

12 Linguistic Levels: Idioms and fixed expressions Huddleston, Rodney and Geoffrey K. Pullum. 2002. The Cambridge Grammar of the English Language. Cambridge: Cambridge University Press. Hudson, Jean. 1998. Perspectives on Fixedness: Applied and Theoretical. Lund: Lund University Press. Knappe, Gabriele. 2004. Idioms and Fixed Expressions in English Language Study before 1800: A Contribution to English Historical Phraseology. Frankfurt: Peter Lang. Knappe, Gabriele. 2006a. The treasury of phrases in Peter Mark Roget’s Thesaurus of English Words and Phrases (1852). In: Christoph Houswitschka, Gabriele Knappe, and Anja Mu¨ller (eds.), Anglistentag 2005 Bamberg: Proceedings, 475–487. Trier: Wissenschaftlicher Verlag. Knappe, Gabriele. 2006b. Peter Mark Roget’s Thesaurus of English Words and Phrases: A midnineteenth century example of the place of phraseology in the history of linguistic theory and practice. In: Christian Mair and Reinhard Heuberger (eds.), in collaboration with Josef Wallmannsberger, Corpora and the History of English: Papers Dedicated to Manfred Markus on the Occasion of His Sixty-Fifth Birthday, 205–220. Heidelberg: Winter. Knappe, Gabriele. 2006c. Phraseology in English language study before 1800 and Lewis Chambaud’s Idioms of the French and English Languages (1751). In: Ha¨cki Buhofer and Burger (eds.), 413–423. Knappe, Gabriele and Michael Schu¨mann. 2006. Thou and ye: A collocational-phraseological approach to pronoun change in Chaucer’s Canterbury Tales. Studia Anglica Posnaniensia 42: 213–238. http://ifa.amu.edu.pl/sap/files/42/17Knappe.pdf Krug, Manfred. 1998. String frequency: A cognitive motivating factor in coalescence, language processing and linguistic change. Journal of English Linguistics 26: 286–320. Krug, Manfred. 2003. (Great) Vowel Shifts present and past: Meeting ground for structural and natural phonologists. Penn Working Papers 9.2: Selected papers from NWAV 31: 107–122. Kurath, Hans, Sherman M. Kuhn, John Reidy, and Robert E. Lewis. 1952–2001. Middle English Dictionary. Ann Arbor: University of Michigan Press. http://quod.lib.umich.edu/m/med/ MacKenzie, Ian. 2003. Poetry and formulaic language. In: Christine Michaux and Marc Dominicy (eds.), Linguistic Approaches to Poetry, 75–86. Amsterdam/Philadelphia: John Benjamins. Magoun, Francis P. 1953. The oral-formulaic character of Anglo-Saxon narrative poetry. Speculum 28: 446–467. Makkai, Adam. 1972. Idiom Structure in English. The Hague/Paris: Mouton. Marchand, Hans. 1969. The Categories and Types of Present-Day English Word-Formation: A Synchronic-Diachronic Approach. 2nd edn. Munich: Beck. Mieder, Wolfgang. 1982. International Proverb Scholarship: An Annotated Bibliography. New York: Garland. Mieder, Wolfgang, Stewart A. Kingsbury, and Kelsie B. Harder (eds.). 1992. A Dictionary of American Proverbs. New York: Oxford University Press. Moon, Rosamund. 1998a. Fixed Expressions and Idioms in English: A Corpus-Based Approach. Oxford: Clarendon Press. Moon, Rosamund. 1998b. Frequencies and forms of phrasal lexemes in English. In: Cowie (ed.), 79–100. Moon, Rosamund. 2007. Corpus linguistic approaches with English corpora. In: Burger, Dobrovo 0skij, Ku¨hn, and Norrick (eds.), 1045–1059. Moralejo Ga´rate, Teresa. 2003. Composite Predicates in Middle English. Munich: LINCOM. Munske, Horst Haider. 1993. Wie entstehen Phraseologismen? In: Klaus J. Mattheier, KlausPeter Wegera, Walter Hoffmann, Ju¨rgen Macha, and Hans-Joachim Solms (eds.), Vielfalt des Deutschen, 481–515. Frankfurt am Main: Peter Lang. Nattinger, James R. and Jeanette S. DeCarrico. 1992. Lexical Phrases and Language Teaching. Oxford: Oxford University Press. Norrick, Neal R. 2007. English phraseology. In: Burger, Dobrovo 0skij, Ku¨hn, and Norrick (eds.), 615–619. Prins, A. A. 1952. French Influence in English Phrasing. Leiden: Universitaire Pers.

195

196

II Linguistic Levels Reuter, O. R. 1986. Proverbs, Proverbial Sentences and Phrases in Thomas Deloney’s Works. Helsinki: Societas Scientiarum Fennica. Ross, Alan S. C. 1975. “Run and Reve” and similar alliterative phrases. Neuphilologische Mitteilungen 76: 571–582. Sailer, Manfred. 2007. Corpus linguistic approaches with German corpora. In: Burger, Dobrovo 0skij, Ku¨hn, and Norrick (eds.), 1060–1071. Sialm, Ambros, Harald Burger, and Angelika Linke. 1982. Historische Phraseologie. In: Burger, Buhofer, and Sialm, 315–382. Sick, Christine. 1993. Adverbiale Phraseologismen des Englischen. Tu¨bingen: Narr. Simpson, John (ed.). 2000–. The Oxford English Dictionary. 3rd edn. online. Oxford University Press. www.oed.com/ Sinclair, John. 1991. Corpus, Concordance, Collocation. Oxford: Oxford University Press. Skandera, Paul (ed.). 2007. Phraseology and Culture in English. Berlin/New York: Mouton de Gruyter. Sontheim, Kurt. 1972. Sprichwort, sprichwo¨rtliche und metaphorische Redewendungen: Synchronische und diachronische Studien zu semantisch-idiomatischen Konstruktionen im Englischen. Ph.D. dissertation, University of Erlangen-Nu¨rnberg. Speake, Jennifer (ed.). 2003. The Oxford Dictionary of Proverbs. 4th edn. Oxford: Oxford University Press. Stern, Gustaf. 1931. Meaning and Change of Meaning: With Special Reference to the English Language. Westport, CT: Greenwood Press. Tilley, Morris Palmer. 1950. A Dictionary of the Proverbs in England in the Sixteenth and Seventeenth Centuries: A Collection of the Proverbs Found in English Literature and the Dictionaries of the Period. Ann Arbor: University of Michigan Press. Thun, Harald. 1978. Probleme der Phraseologie: Untersuchungen zur wiederholten Rede mit Beispielen aus dem Franzo¨sischen, Italienischen, Spanischen und Romanischen. Tu¨bingen: Niemeyer. Trousdale, Graeme. 2008. Constructions in grammaticalization and lexicalization: Evidence from the history of a composite predicate construction in English. In: Graeme Trousdale and Nikolas Gisborne (eds.), Constructional Approaches to English Grammar, 33–67. Berlin/New York: Mouton de Gruyter. Voitl, Herbert. 1969. Probleme der englischen Idiomatik. Germanisch-romanische Monatsschrift, N.F. 19: 194–212. Weickert, Rainer. 1997. Die Behandlung von Phraseologismen in ausgewa¨hlten Sprachlehren von Ickelsamer bis ins 19. Jahrhundert: Ein Beitrag zur historischen Phraseologie. Hamburg: Kovac. Weinstock, Horst. 1966. Die Funktion elisabethanischer Sprichwo¨rter und Pseudosprichwo¨rter bei Shakespeare. Heidelberg: Winter. Welte, Werner. 1992. On the properties of English phraseology: A critical survey. In: Claudia Blank (ed.), Language and Civilization: A Concerted Profusion of Essays and Studies in Honour of Otto Hietsch. Vol. II, 564–591. Frankfurt: Peter Lang. Whiting, Bartlett Jere, with the collaboration of Helen Wescott Whiting. 1968. Proverbs, Sentences, and Proverbial Phrases: From English Writings Mainly Before 1500. Cambridge, MA: Belknap Press of Harvard University Press. Windeatt, Barry. 1992. Troilus and Criseyde. Oxford: Oxford University Press. Wray, Alison. 2002. Formulaic Language and the Lexicon. Cambridge: Cambridge University Press. Wyld, Henry Cecil. 1936. A History of Modern Colloquial English. 3rd edn., with additions. Oxford: Basil Blackwell.

Gabriele Knappe, Bamberg (Germany)

13 Linguistic Levels: Pragmatics and discourse

197

13 Linguistic Levels: Pragmatics and discourse 1 2 3 4 5 6 7

Introduction Pragmatic explanations in language change Pragmatics as the study of performance phenomena Discourse as dialogue Discourse as a domain of communication Summary References

Abstract Pragmatics studies the processes of language use, while discourse analysis is devoted to its product, i.e. discourse. Pragmatics can be understood in a narrow sense focussing on cognitive-inferential aspects of information processing, and it can be understood in a wider sense in which it also includes social aspects of interaction. In historical pragmatics, the former conceptualization lies behind work on pragmatic explanations in language change, while the latter conceptualization studies earlier language use from a social and interactional perspective, including such aspects as inserts (e.g. interjections and discourse markers), speech acts, and terms of address. Discourse, as the product of language use, can be seen as a stretch of conversation (dialogue) or as a domain of communication. In the former conceptualization, research focuses on the structural properties of the dialogue, and in the latter, it deals with the linguistic practices pertaining to particular fields of knowledge or interaction, e.g. courtroom discourse, the discourse of science, and news discourse.

1 Introduction In a very general sense pragmatics can be defined as the study of language use, while discourse analysis, in an equally general sense, can be defined as the analysis of the result of human communication, viz. discourse. It has been suggested that discourse analysis is more text-centered, more static, more interested in product (in the well-formedness of texts), while pragmatics is more user-centred, more dynamic, more interested in the process of text production. Discourse analysis is frequently equated with conversational analysis, and pragmatics with speech act theory. It would seem difficult to distinguish the two with any conviction, however (Brinton 2001: 139).

There is certainly a great deal of overlap between the two fields. A large range of topics can be dealt with under either heading. Speech acts, such as greetings and farewells, or discourse markers, such as well, so, or you know have both interactional (pragmatic) functions and text-structuring or discourse functions. As a field of study, pragmatics has grown very considerably over the last thirty years or so. Traditionally, linguists were mainly concerned with an analysis of language structure at the levels of phonology, morphology, and syntax, but with the pragmatic turn in the late 1970s and early 1980s some of the interest shifted from the structure of Alex Bergs and Laurel Brinton 2012, Historical Linguistics of English (HSK 34.1), de Gruyter, 197–212

198

II Linguistic Levels language to the language user. At the beginning of this development, pragmatics was often seen as the ragbag of linguistic description (see Mey 1998: 716). As such it covered performance phenomena that could not be handled at the traditional levels of linguistic description, such as speech acts, conversational implicature, deixis, and politeness, but also the structure of conversations. On the other hand, even in the early days of pragmatics, the discipline was also seen as a perspective. As such it was not a level of linguistic description but a different way of analyzing language. Language was not seen as a system of signs but as a means of communication. “Pragmatics is a perspective on any aspect of language, at any level of structure” (Verschueren 1987: 5, italics in original; see also Verschueren 1999: 2). Under the former view, pragmatics was a separate level of linguistic description, parallel to other levels, such as syntax or semantics. Under the latter view, pragmatics was a particular way of doing linguistics that could be applied to all other levels of linguistic description from phonology and morphology to syntax, semantics and, indeed, discourse. These positions have developed into a more restricted cognitive-inferential conceptualization of pragmatics (adhered to, generally speaking, by Anglo-American researchers) and a broader socio-interactional conceptualization (common among European researchers). Cruse (2000), for instance, gives the following narrow definition of pragmatics: For present purposes pragmatics can be taken to be concerned with aspects of information (in the widest sense) conveyed through language which (a) are not encoded by generally accepted convention in the linguistic forms used, but which (b) none the less arise naturally out of and depend on the meanings conventionally encoded in the linguistic forms used, taken in conjunction with the context in which the forms are used (Cruse 2000: 16).

In this conceptualization, people routinely understand more than what is explicitly communicated. They read between the lines, as it were, and this is the field of the pragmaticist. In her handbook article on historical pragmatics, Traugott (2004: 539) also takes pragmatics “to be non-literal meaning that arises in language use”, and Sperber and Noveck (2004: 1) define pragmatics as “the study of how linguistic properties and contextual factors interact in the interpretation of utterances.” In their view, pragmatics is not restricted to a study of implicit meanings. In fact, they are at pains to demonstrate that there are many aspects of explicit meaning that require access to contextual information for their interpretation, but they exclude the wider social issues of language use from the scope of pragmatics. The European tradition adopts a broader, more sociologically based view of pragmatics that includes social and cultural conditions of language use. Trosborg (1994: 37), a representative of this broader European tradition, for instance, states that “sociopragmatics is concerned with the analysis of significant patterns of interaction in particular social situations and/or in particular social systems. For example, speech acts may be realized differently in different social contexts and situations as well as in different social groups within a speech community,” while Blakemore, a representative of the Anglo-American tradition, finds it “misleading to include phenomena like politeness, face-saving and turn taking […] under the general heading of pragmatics” (Blakemore 1992: 47). The two conceptualizations of pragmatics, obviously, have consequences for the interaction of pragmatics and historical linguistics. The former conceptualization

13 Linguistic Levels: Pragmatics and discourse suggests a range of specific performance-related topics, while the latter suggests a specific way of investigating earlier stages of a language and its development. The term “discourse” is perhaps even more open to different definitions. On the one hand, it can be seen as the spoken equivalent of a text. A (written) text is made up of sentences while a (spoken) discourse is made up of utterances. In this sense, the term “discourse” is more or less synonymous with the term “dialogue” (see below, Section 4). Brinton (2001: 139–140) distinguishes between three discourse analytical approaches to historical data. First, the discourse analyst may use forms, functions, and structures of discourse at historical stages of a language. She calls this approach “historical discourse analysis proper”. Second, the discourse analyst may study the discourse-pragmatic factors and motivations behind language change. This approach is called “discourseoriented historical linguistics”. And third, the discourse analyst may focus on the diachronic development of discourse functions and discourse structures over time. She calls this third approach “diachronic(ally oriented) discourse analysis”. However, the term “discourse” can also be used in a much wider sense, not just for a linguistic unit larger than utterances, but as a domain of language. In such a view, a discourse is a collection of linguistic practices characterized by a distinct group of people and a distinct group of genres and text types, e.g. the discourse of science, or more specifically the discourse of medical science or the discourse of modern linguistics. In the following I shall evaluate how these conceptualizations of the terms “pragmatics” and “discourse” can be applied to the analysis of historical data and in particular to English historical data.

2 Pragmatic explanations in language change In the Anglo-American conceptualization of pragmatics, pragmatics is mainly a tool to describe and explain patterns of language change. Language is a means of communication and, therefore, the communicative forces that are at work when people use language must be taken into consideration when we analyse, for instance, the syntax of a language and indeed when we analyse diachronic changes in the syntax of a language. Thus, pragmatics becomes a principle of explanation in language change. In Brinton’s (2001) terminology this would be “discourse-oriented historical linguistics”. If pragmatics is seen as one level of linguistic description on a par with other levels such as phonology, morphology, syntax and semantics, it is largely restricted to nontruth-conditional aspects of language, and to aspects of language that depend on the context of utterance. Deictic elements, for instance, depend on the situation of use for their interpretation. Speech acts in their early conceptualization of doing things with words were also restricted to non-truth-conditional aspects. Speech act theory took its starting point from Austin’s (1962) observation that speech acts are regularly used for purposes other than stating facts that are assessable in terms of true or false. Meanings are not abstract entities that pertain to linguistic expressions but the result of negotiations between speaker/writer and addressee/reader, which – through repetition of use – have become conventionalized. A theory of meaning change, therefore, must take into account the communicative situation of speaker/writer and addressee/ reader. Traugott and Dasher (2005), for instance, argue that it is ad-hoc negotiations of meanings that may lead to meaning change if they are invoked repeatedly until they become conventionalized in the entire speech community. They call such ad-hoc

199

200

II Linguistic Levels meanings “invited inferences”, a term borrowed from Geis and Zwicky (1971). However, Traugott and Dasher use it in a broader sense and do not restrict it to generalized implicatures. It signals the speaker/writer’s role in inviting the addressee to infer the intended ad-hoc meaning. As an example they cite the case of as/so long as (Traugott and Dasher 2005: 36–37). In Old and Middle English the spatial meaning (‘of the same length as’) co-existed with the temporal meaning (‘for the same length of time as’). In some contexts, the meaning invited the conditional meaning ‘provided that’, as for instance in (1). (1)

wring þurh linenne clað on þæt eage swa lange swa him ðearf sy. (850–950 Lacnunga, p. 100; Traugott and Dasher 2005: 36, ex. 19) wring through linen cloth on that eye as long as him need be-SUBJ ‘squeeze (the medication) through a linen cloth onto the eye as long as he needs.’

The medicine is to be applied for the duration that it is needed, which invites the inference that it is to be applied only if it is needed. According to Traugott and Dasher all examples of as/so long as in Old and Middle English are either spatial or temporal, and while some allow a conditional reading, the conditional reading is never predominant. This changes in Early Modern English, when examples occur in which the invited inference of conditionality has been generalized to contexts of reasoning and cognition in which a temporal reading does not make sense or is at least not salient as in (2). (2)

They whose words doe most shew forth their wise vnderstanding, and whose lips doe vtter the purest knowledge, so as long as they vnderstand and speake as men, are they not faine sundry waies to excuse themselues? (1614 Hooker, p. 5; Traugott and Dasher 2005: 37, ex. 20)

Here the conditional reading is salient, while the temporal meaning is still available. Traugott and Dasher paraphrase the temporal meaning as “for the time that they understand and speak as men”, i.e. “as long as they live”. From the mid-19th century there are examples in which the conditional is the only possible meaning as in (3). (3)

“Would you tell me, please, which way I ought to go from here?” “That depends a good deal on where you want to get to,” said the Cat. “I don’t much care where-” said Alice. “Then it doesn’t matter which way you go,” said the Cat. “- so long as I get somewhere,” Alice added as an explanation. (1865 Carroll, Chapter 6, p. 51; Traugott and Dasher 2005: 37, ex. 21a)

Thus meaning change is the result of the interaction between speakers/writers and addressees/hearers in communicative situations. Speakers/writers use established coded meanings (e.g. the temporal reading of so/as long as) in creative ways to invite inferences. Through repeated use, such invited inferences become conventionalized and ultimately they become new coded meanings (Traugott and Dasher 2005: 38). Thus language change is seen as the result of what Keller (1994) has called an “invisible hand process”. Language change comes about as a causal effect of the accumulation of individual speakers’ action, who – individually – did not intend this effect.

13 Linguistic Levels: Pragmatics and discourse

3 Pragmatics as the study of performance phenomena Performance phenomena pertain mostly to the spoken language, i.e. to language that is produced under the constraints of online production. Such phenomena were shunned as irrelevant for a long time. For historical linguists they were doubly irrelevant. They were irrelevant because they were not part of the language system itself, and they were irrelevant because historical linguists did not have access to the spoken language of the past. The communicative turn in the ’70s and ’80s of the 20th century turned performance phenomena into legitimate objects of investigation for synchronic linguistics. Pragmaticists focused their attention on transcriptions of spoken interaction. They studied the minutiae of the turn-taking system, the form and function of individual utterances (speech acts), and so on. But these studies were restricted to present-day data. Pragmaticists saw written language as secondary and therefore as uninteresting for pragmatic analyses. Today performance phenomena have made their way into standard descriptions of the English language (e.g. Biber et al. 1999, who spend a considerable amount of space on such phenomena within the confines of a structural description of the English language), and within the last decade or so, significant progress has been made on the description of performance phenomena from a diachronic perspective. I shall briefly mention three examples which have received a considerable amount of attention from historical pragmaticists, inserts, speech acts and terms of address. To the extent that the analyses of these elements rely on references to social conditions of their use, they clearly go beyond the narrow Anglo-American conceptualization of pragmatics.

3.1 Inserts Biber et al. (1999: 1082) use the term “inserts” to refer to “stand-alone words which are characterized in general by their inability to enter into syntactic relations with other structures. […] They comprise a class of words that is peripheral, both in the grammar and in the lexicon of the language”. They distinguish nine different types of inserts: interjections (oh, ah), greetings and farewells (hi, hello, goodbye), discourse markers (well, right), attention signals (hey, yo), response elicitors (right?, eh?), response forms (yeah, yep), hesitators (um, er), various polite speech-act formulae (thanks, sorry), and expletives (shit, good grief!). Not all of these are equally amenable to a historical analysis. Biber et al. (1999: 1096–1098) provide some statistics about their distribution in American English and British English conversations, but they do not say anything about their occurrence in written genres. It seems reasonable to assume that some of them are relatively infrequent in the texts that have survived from earlier centuries. While some inserts, such as interjections or discourse markers, have been analyzed in their own right, others, like thanks and sorry, have been investigated in larger contexts of speech act studies of thanking and apologizing (e.g. Jacobsson 2002; Jucker and Taavitsainen 2008b), and expletives have been investigated in the context of the language of insults (e.g. Craun 1997). Taavitsainen (1995) investigates the form, function and distribution of exclamations, such as alas, ey, ah, harrow, and O in Late Middle and Early Modern English (see also Hiltunen 2006; Person 2009). Their distribution is clearly genre specific. In the Helsinki

201

202

II Linguistic Levels Corpus, which was used for the investigation, exclamations were particularly frequent in the genres comedy and fiction. They also occurred in trials and in Bible texts. In other genres they were rare. Exclamations were used more widely and with a broader variety of functions than in Present-day English. They were regularly used as vocatives and as appeals to the addressee. The interjection O, for instance, is often prefixed to an exclamatory sentence and it often combines with a vocative as in example (4), which is taken from a sermon. (4)

O my God, my God, why haste thou forsaken me? (1614 Hooker, Two Sermons Upon Part of S. Judes Epistle, 1614, p. 7; Helsinki Corpus, Taavitsainen 1995: 453)

Discourse markers have received considerable attention in historical pragmatics. Brinton (1996), for instance, analyzed a broad range of discourse markers, or “pragmatic markers”, as she calls them, including Old English hwæt, Middle English gan, and Middle and Early Modern English anon. She is interested not only in the developing discourse functions of these elements but also in the grammaticalization processes that they instantiate. In more recent publications she has added analyses of only (Brinton 1998), I say (Brinton 2005) and I mean (Brinton 2007) (see also Jucker 1997, 2002; Fischer 1998; Brinton 2006).

3.2 Speech acts Speech acts are not easily amenable to historical investigations because the traditional research methods developed for present-day languages cannot be applied to historical data. Originally the concept was developed by philosophers who investigated the nature of speech acts on the basis of careful considerations of what it means to name a ship, to make a promise, to issue a command, to ask a question, or to greet somebody (Austin 1962; Searle 1969). Later, empirical methods, such as discourse completion tests and role-plays, were developed to investigate speech acts and their realisations by different groups of speakers (e.g. Blum-Kulka et al. 1989; Trosborg 1994). For obvious reasons, none of these methods can be applied to historical data. More recently, corpus-based research methods have been improved and developed to such an extent that various avenues of investigations of historical speech act material have become available. It is, of course, possible to search for verbs denoting specific speech acts. Such speech act verbs are sometimes used performatively to carry out the speech act they denote. Kohnen (2008a), for instance, argues that in Old English explicit performatives were typically used to issue requests and commands as in (5): (5)

Ic bidde eow þæt ʒe ʒymon eowra sylfra, swa eowere bec eow wissiað. (Ælfric, Letter to Wulfsige, 26; Helsinki Corpus, Kohnen 2008a: 30) ‘I ask you to take care of yourselves, as your books teach you.’

The Old English verb biddan ‘ask, bid’ is here used performatively. By saying Ic bidde eow ‘I ask you’ the speaker carries out the speech act of asking or requesting (see in particular Kohnen 2000). However, many verbs that describe a speech act are not normally used performatively. They are used to talk about the speech act they name. They may occur in

13 Linguistic Levels: Pragmatics and discourse narratives with an account that a particular speech act had been performed, or in negotiations when the precise speech act value of an utterance is being discussed. (6)

If eny man wolde challenge a frere of Seint Frauncessis ordre and seue … Frere, thou louest money as myche as othere men […] (c.1449 Pecock Repr.; Taavitsainen and Jucker 2007: 113) ‘If any man were to challenge a friar of the order of St. Francis and to say … “Friar, you love money as much as other men […]’

In (6) the speech act verb “challenge” is used together with an example of an utterance with this speech act value. Many speech acts, perhaps most, are carried out without the relevant speech act verb. In order to locate relevant speech acts, the researcher has to rely on the philological method of actually reading the source texts. Jucker and Taavitsainen (2000) have used this method to describe insults in the history of English. But the method obviously precludes any statistical results. The findings can only be very selective based on the available research time. Some speech acts show recurrent surface patterns. Deutschmann (2003), for instance, has shown that apologies in English are mostly formulaic. They can be traced with corpus-linguistic tools by searching for a small number of expressions that typically occur in apologies, such as sorry, pardon, and excuse together with related and expanded forms. The same method has recently been used to trace apologies (Jucker and Taavitsainen 2008b), promises (Valkonen 2008) and compliments (Jucker et al. 2008).

3.3 Terms of address In the 13th century under the influence from French, English started to use the second person plural pronoun ye not only for two or more addressees but – under certain circumstances – also for one single addressee. Many Indo-European languages still have this distinction between two pronominal forms of address for a single addressee. On the basis of Latin tu and vos, the pronoun choices are usually abbreviated as T and V (Brown and Gilman 1960: 254). The conditions under which one pronoun or the other is chosen have been the object of extensive research in recent years (see, for instance, the volume by Taavitsainen and Jucker 2003). Brown and Gilman (1960) in their seminal article on the topic tried to find a common denominator for all languages with such a system. They argue that this common denominator is the semantics of power and solidarity. In medieval Europe, according to this theory, the power semantics accounted for a non-reciprocal use of T from the more powerful to the less powerful. The more powerful received V in return from the less powerful. Equals of the upper classes exchanged mutual V, while equals of the lower social classes exchanged mutual T. The power semantics of medieval Europe has been replaced by the solidarity semantics in which mutual V signals distance and mutual T solidarity. A significant body of research has shown that social conditions for the choice of T or V in specific situations are considerably more complex. Mazzon (2000), Honegger (2003), and Jucker (2006), for instance, have shown that Chaucer’s system of pronoun choices is much more situationally governed than the usual present-day systems in languages such as German, French or Italian. In the present-day forms of these languages,

203

204

II Linguistic Levels choices are more or less fixed for any given dyad of speakers, and a switch from mutual V to mutual T is a noticeable event, often accompanied by some kind of ritual (a switch from mutual T to mutual V, i.e. from informal to formal, would be very unusual). In Chaucer’s English, the characters of his fictional work used a more complex system that was based not only on social status between the characters but also on the basis of situational dominance or subjugation. Such approaches have replaced the earlier accounts of Chaucer’s use of personal pronouns by such scholars as Nathan (1959), Wilcockson (1980), and Burnley (1983), who tried to explain the choices largely on the basis of fixed social relationships. By the time of Shakespeare, it does no longer seem possible to provide an account that explains individual pronoun choices. Researchers, therefore, generally focus on frequencies and on co-occurrence patterns of nominal and pronominal terms of address. U. Busse (2002, 2003), for instance, shows that titles of courtesy, such as Your Grace, Your Ladyship, (my) liege or sir, are more likely to occur together with a V pronoun than any of the other categories of nominal terms of address, while terms of endearment, such as bully, chuck, heart, joy or love are most likely to occur together with a T pronoun (see also Stein 2003; B. Busse 2006).

4 Discourse as dialogue Discourse can be seen as a stretch of conversation or as a domain of language. In this section, I will use the term “dialogue” to refer to the former and the term “domain of discourse” for the latter. The terms “discourse” and “dialogue” imply an interaction between a speaker or writer and a recipient. Written texts, although there is no regular exchange of roles between speaker/writer and hearer, do have an addressee, even if the addressee is only a recipient and cannot actively contribute to the interaction. They are what Kilian (2005: 102) identifies as a “functional” dialogue. Fritz (1995: 469) distinguishes three stages of what he calls “historical dialogue analysis”. The first stage is characterized by analysis of the pragmatic structure and function of a historical dialogue in its social and historical context. The second stage is characterized by a contrastive comparison of earlier dialogue forms with later dialogue forms. The third and most advanced stage is characterized by an investigation of the evolution and dissemination of specific forms of dialogue. In the first stage, the researcher can use the same conversation analytical or dialogue analytical tools that are employed in modern data in order to investigate older forms of dialogue. The analysis can either adopt a macro perspective or a micro perspective. Under the macro perspective, the researcher focuses on the structure of the dialogue under analysis. Levinson (1983) reserved the term “discourse analysis” for such macro analyses of dialogue structures. Under the micro perspective, the researcher focuses on individual pragmatic elements, such as greetings, address terms, discourse markers and so on; or on local structures, e.g. adjacency pairs, such as question – answer sequences. Levinson (1983) used the term “conversation analysis” for this type of investigation. An analysis of individual pragmatic elements in individual dialogues of earlier periods coincides with the pragmatic research interests sketched out above. And indeed, a considerable amount of research has been published, e.g. on address terms in Chaucer’s narratives or in Shakespeare’s plays (see Section 3.3). But researchers have also

13 Linguistic Levels: Pragmatics and discourse adopted the larger perspective of looking at the inventory of pragmatic elements making up a specific type of historical dialogue. Watts (1999), for instance, investigates in detail two dialogues that were printed in 16th-century English language coursebooks for the benefit of learners of English as a foreign language. However, in practice it is not always easy to distinguish between the different stages envisaged by Fritz. Jucker and Taavitsainen (2000), for instance, investigate the use of insults in the history of English. The aim is to show a development or an evolution from the earlier forms to the later forms, but at present all that seems to be possible is a contrastive analysis of selected examples at different periods in the history of English. It is not yet possible to trace a continuous evolution of specific speech acts, such as insults. Archer in various publications (e.g. Archer 2005, 2006, 2007) gives a detailed picture of Early Modern English courtroom dialogue and thus carries out research at the first stage of historical dialogue analysis, but she also compares these findings to the present-day courtroom, representing the second stage. And finally she also draws attention to developments within the period under investigation, and thus contributes to stage three of historical dialogue analysis. She focuses mainly on the question-answer sequences in the courtroom dialogues and uses these to pinpoint the (changing) discursive roles of the active participants in the English courtroom, i.e. the judges, lawyers, witnesses and defendants. Taavitsainen (1999) also investigates the evolution of a particular form of dialogue. She assesses medical dialogues in Late Middle and Early Modern English, and traces the evolution of these dialogues between 1375 and 1750. She describes two traditions that are evident in Early English medical dialogues: the scholastic formula, based on the format of debates by Greek philosophers, and the mimetic dialogues, in which material is presented in fictional conversations between the author and the reader or between fictional characters. Taavitsainen shows how these traditions develop over the centuries and how, in the 18th century, medical dialogues merge with the new pamphlet tradition, in which social matters, such as health-care for the poor or polite conversations, are treated.

5 Discourse as a domain of communication As pointed out in Section 1, the term “discourse” can also be used in a more general sense as the totality of linguistic practices that pertain to a particular field of knowledge or to a particular occupation. Such discourses consist not of utterances but of typical text types, characterized by specific lexical items, idiosyncratic syntax, and particular routinized patterns of interaction. In such a context, researchers also ask more general questions about the dissemination of information within groups of speakers. Three such domains of communication in particular have received a fair amount of scholarly attention for the Early Modern English period: courtroom discourse, the discourse of science and news discourse.

5.1 Courtroom discourse A considerable amount of research has appeared on courtroom discourse in the Early Modern English period. The Early Modern English courtroom differed considerably from its modern equivalent. While modern courts presume a defendant to be innocent

205

206

II Linguistic Levels until proven guilty, the Early Modern courtroom expected the defendants to prove their innocence. Archer (2005: 85) demonstrates how this leads to a more active involvement on the part of the defendant. It was only in the later part of the Early Modern period that courtrooms introduced defence counsels who started to speak on behalf of the defendant. Koch (1999: 410–411), in his analysis of excerpts of three early Romance court records, draws attention to the communicative complexity of such records. The records written by a court scribe and addressed to a future reader are legal documents with appropriate formality of expression especially in the ritualistic elements pertaining to the formalities of the proceedings. These parts of the court records are characterized by the “language of distance” as Koch calls it. Embedded in this formal document there is a transcription of the verbal interaction taking place in the courtroom between the judge, the witnesses, the defendants and the lawyers. These utterances, even if they are written down, are closer to spoken language, or the “language of immediacy”. There may even be further embeddings, especially if the court cases dealt with libel, in which courtroom interactants report utterances that were spoken outside the courtroom. Such reported utterances are even closer to the language of immediacy. In her work on the Early Modern English courtroom Archer (2005, 2006, 2007) draws a detailed picture of the strategies adopted by the judge, the lawyers, the defendants and the witnesses. She concludes that the frequency of questions, their function and their interactional success depended on a number of sociopragmatic factors, such as the speech event, the position of the question and the discursive roles of the speaker and the addressee as well as the date of the trial (2005: 281). Culpeper and Semino (2000) extend the scope of courtroom discourse. They use two types of data, learned treatises on the topic of witchcraft and courtroom witness depositions. In their analysis, they deal with speech act verbs, such as to curse and they show how such verbs could be used to reinterpret trivial arguments within a village community into a witchcraft event. The witch trials that took place in 1692 in the Puritan village of Salem in the colony of Massachusetts have attracted a considerable amount of research into the discourse strategies adopted by the participants and the functional and structural properties of the trials as such. Kahlas-Tarkka and Rissanen (2007), for instance, investigated the discourse strategies of “successful” and “unsuccessful” defendants in the Salem witch trials, while Hiltunen and Peikola (2007) focus on the material evidence of these trials, i.e. the handwritten records and the printed editions. Their contribution demonstrates vividly how important it is not to forget the communicative role of the scribe who commits the spoken words in the courtroom to writing and thus makes it available for future generations (see also Doty and Hiltunen 2002; Hiltunen 2004; Doty 2007).

5.2 The discourse of science In the late medieval world, the discourse of science was multilingual. The main language for written texts was Latin, but texts started to be translated into the vernacular and the Greco-Roman tradition provided a model for scientific writing in the vernacular. In modern linguistics, “medical discourse” refers collectively to the communicative practices of the medical profession, both written and spoken. In the late medieval period, the medical profession consisted of heterogeneous groups of practitioners, including

13 Linguistic Levels: Pragmatics and discourse physicians, surgeons, barbers, midwives, itinerant specialists (e.g., bonesetters and oculists), herbalists, apothecaries, wisewomen, and others. They can be roughly divided into clerical and elite practitioners and tradespeople or ordinary practitioners; literacy was restricted mostly to the elite group (Taavitsainen 2006: 688).

Taavitsainen (2006) gives an overview of genres that were important for this discourse community. Compilations and commentaries of earlier studies were important for the dissemination of scholastic knowledge. Texts in question-and-answer format and pedagogical dialogues were also popular genres of scientific and medical writing that were adopted from Latin models into the vernacular. The volume edited by Taavitsainen and Pahta (2004) contains a range of detailed studies of medical and scientific writing in Late Medieval English. Ma¨kinen (2004), for instance, describes Middle English herbal recipes and recipes in manuals for medicinal plants and shows the textual traditions that link them together. Valle (1999: vii) takes the view that “science has at least since the seventeenth century taken place within a knowledge-producing discourse community, and that this community will in some way be ‘represented’ in scientific texts, in forms which can be identified and studied”. The totality of texts produced by this discourse community is, therefore, the discourse of science. In her study, Valle describes the discourse community of the Royal Society on the basis of a corpus of texts drawn from the Philosophical Transaction, spanning the three centuries from the beginning of publication in 1665 to 1965 (see also Valle 1997, 2006). Gotti (2006), too, deals with the discourse community of the Royal Society in London and illustrates some of the methods that were used by this community to spread the news about new discoveries and other scientific findings. Letters exchanged between scholars played an important role. They were not only exchanged between individuals but they were frequently copied and passed on to new recipients. Some influential scholars at the centre of scientific networks regularly received, sent and resent a large number of letters and thus had the role of clearing houses.

5.3 Early English news discourse With the invention of the printing press it became possible to publish accounts of recent events and to disseminate them to a large audience. In the 16th and 17th centuries pamphlets and newsbooks were used for this purpose (Raymond 2003). The first newspapers in the modern sense appeared in the early 17th century, first on the continent but soon also in England (Brownlees 1999; Studer 2008). The first newspapers or corantos, as they were originally called, consisted mainly of dispatches from correspondents from important places throughout Europe. These letters were inserted into the newspaper in the order in which they arrived at the editorial office in London. There was no other structural principle. It took another century for the first daily newspapers to be published in the early 18th century. As Sommerville (1996) has pointed out, the revolutionary aspect of this kind of news discourse consisted in the fact that newspapers appeared in regular intervals, weekly at first, twice or three times a week later and then daily. Thus, news was no longer reported in response to important events but a certain amount of space had to be filled with news on a regular basis. The early news discourse has attracted a fair amount of research recently not only in collections of articles, such as Ungerer (2000), Herring (2003), Raymond (2006) or

207

208

II Linguistic Levels Brownlees (2006) but also in monographs. Studer (2008), for instance, develops a larger picture of the development of news discourse on the basis of the Zurich English Newspaper Corpus (ZEN). He argues that news discourse is shaped by such external factors as the historical context and technological innovations. News discourse both adopted and adapted generic conventions; that is to say, it used existing genres, e.g. in the form of the letters from correspondents in the early newspapers, and it transformed and shaped them for its own needs. (See further Fries, Chapter 67.)

6 Summary It is not possible to draw a principled distinction between historical topics that are treated with pragmatic tools of investigation and those that are treated with discourse analytical tools. Traditionally, those approaches that focus on the interactional and dynamic aspects of language belong to pragmatics while those that focus on the structural aspects of dialogues, conversations or discourses belong to discourse analysis. The application of pragmatic and discourse analytical tools to historical data has uncovered a rich area of investigation and thrown new light on much familiar data. But a lot still needs to be done. At present, three areas of research appear to be particularly promising. First, the research on the history of speech acts has only just started to attract more than just occasional research efforts. In the volume edited by Jucker and Taavitsainen (2008a) a number of researchers have joined forces to investigate a range of different speech acts in the history of English and to develop the necessary methodologies. Recent advances in corpus technology have made it increasingly possible to locate some speech acts automatically. Second, the research of the evolution of forms of dialogue is still in its infancy. Kilian (2005) has presented an introduction into historical dialogue research, in which he develops a detailed typology of historical types of dialogues and some methodologies to investigate a broad range of such dialogues, i.e. dialogues in which speakers and addressees take turns in their roles. Culpeper and Kyto¨ (2010: 2) ask: “what was the spoken face-to-face interaction of past periods like?” in a systematic way and approach this question from various angles. In particular they look at the structure of conversations, at what they call “pragmatic noise”, i.e. pragmatic interjections or discourse markers, and social roles and gender in interaction. And third, the evolution of domains of discourse appears to be a very promising field of research. The existing work on courtroom discourse, the discourse of science and news discourse needs to be continued, and other domains should be tackled. The discourse of religion, for instance, would be an obvious candidate because there is wealth of historical material available consisting of many different text types, such as sermons, prayers, treatises and saints’ lives. The compilation at the University of Cologne of a Corpus of English Religious Prose is very likely to be a first significant step in this direction (see Kohnen 2007). Thus it seems that the new corpora and advances in corpus linguistics have had and are having a considerable impact on historical pragmatics and historical discourse analysis. The cooperation between corpus linguists and historical pragmaticists/discourse analysts has only just started, but it promises considerable advances in our understanding of human interaction and communication from a historical perspective.

13 Linguistic Levels: Pragmatics and discourse Acknowledgments: I thank Thomas Kohnen, Daniela Landert, and Elizabeth C. Traugott for valuable comments on a draft version of this paper. The usual disclaimers apply.

7 References Archer, Dawn. 2005. Questions and Answers in the English Courtroom (1640–1760). Amsterdam/ Philadelphia: John Benjamins. Archer, Dawn. 2006. (Re)Initiating strategies: Judges and defendants in Early Modern English courtrooms. Journal of Historical Pragmatics (Special Issue on Historical Courtroom Discourse, ed. by Barbara Kryk-Kastovsky) 7(2):181–211. Archer, Dawn. 2007. Developing a more detailed picture of the English courtroom (1640–1760): Data and methodological issues facing historical pragmatics. In: Fitzmaurice and Taavitsainen (eds.), 185–217. Austin, J. L. 1962. How to Do Things With Words. Oxford: Oxford University Press. Biber, Douglas, Stig Johansson, Geoffrey Leech, Susan Conrad, and Edward Finegan. 1999. Longman Grammar of Spoken and Written English. London: Longman. Blakemore, Diane. 1992. Understanding Utterances. An Introduction to Pragmatics. Oxford: Blackwell. Blum-Kulka, Shoshana, Juliane House, and Gabriele Kasper (eds.). 1989. Cross-Cultural Pragmatics: Requests and Apologies. Norwood, NJ: Ablex. Borgmeier, Raimund, Herbert Grabes, and Andreas H. Jucker (eds.). 1998. Anglistentag 1997 Giessen. Proceedings. Trier: Wissenschaftlicher Verlag. Brinton, Laurel J. 1996. Pragmatic Markers in English. Grammaticalization and Discourse Functions. Berlin/New York: Mouton de Gruyter. Brinton, Laurel J. 1998. “The flowers are lovely; only, they have no scent.”: The evolution of a pragmatic marker in English. In: Borgmeier, Grabes, and Jucker (eds.), 9–33. Brinton, Laurel J. 2001. Historical discourse analysis. In: Deborah Schiffrin, Deborah Tannen, and Heidi E. Hamilton (eds.), The Handbook of Discourse Analysis, 138–160. Oxford: Blackwell. Brinton, Laurel J. 2005. Processes underlying the development of pragmatic markers: The case of (I) say. In: Janne Skaffari, Matti Peikola, Ruth Carroll, Risto Hiltunen, and Brita Wa˚rvik (eds.), Opening Windows on Texts and Discourses of the Past, 279–299. Amsterdam/Philadelphia: John Benjamins. Brinton, Laurel. 2006. Pathways in the development of pragmatic markers in English. In: Ans van Kemenade and Bettelou Los (eds.), The Handbook of the History of English, 307–334. Oxford: Blackwell. Brinton, Laurel J. 2007. The development of I mean: Implications for the study of historical pragmatics. In: Fitzmaurice and Taavitsainen (eds.), 37–79. Brown, Roger, and Albert Gilman. 1960. The pronouns of power and solidarity. In: Thomas A. Sebeok (ed.), Style in Language, 253–276. Cambridge, MA: MIT Press. Brownlees, Nicholas. 1999. Corantos and Newsbooks: Language and Discourse in the First English Newspapers (1620–1641). Pisa: Edizioni ETS. Brownlees, Nicholas (ed.). 2006. News Discourse in Early Modern Britain. Selected Papers of CHINED 2004. Bern: Peter Lang. Burnley, David. 1983. A Guide to Chaucer’s Language. London: Macmillan. Busse, Beatrix. 2006. Vocative Constructions in the Language of Shakespeare. Amsterdam/ Philadelphia: John Benjamins. Busse, Ulrich. 2002. Linguistic Variation in the Shakespeare Corpus. Morpho-Syntactic Variability of Second Person Pronouns. Amsterdam/Philadelphia: John Benjamins. Busse, Ulrich. 2003. The co-occurrence of nominal and pronominal address forms in the Shakespeare corpus: Who says thou or you to whom? In: Taavitsainen and Jucker (eds.), 193–221. Craun, Edwin D. 1997. Lies, Slander, and Obscenity in Medieval English Literature. Pastoral Rhetoric and the Deviant Speaker. Cambridge: Cambridge University Press.

209

210

II Linguistic Levels Cruse, Alan. 2000. Meaning in Language. An Introduction to Semantics and Pragmatics. Oxford: Oxford University Press. Culpeper, Jonathan, and Elena Semino. 2000. Constructing witches and spells: Speech acts and activity types in Early Modern England. Journal of Historical Pragmatics 1(1): 97–116. Culpeper, Jonathan and Merja Kyto¨. 2010. Early Modern English Dialogues: Spoken Interaction as Writing. Cambridge: Cambridge University Press. Deutschmann, Mats. 2003. Apologising in British English. Umea˚: Institutionen fo¨r moderna spra˚k, Umea˚ University. Doty, Kathleen L. 2007. Telling tales: The role of scribes in constructing the discourse of the Salem witchcraft trials. Journal of Historical Pragmatics 8(1): 25–41. Doty, Kathleen and Risto Hiltunen. 2002. “I will tell, I will tell”: Confessional patterns in the Salem witch trials, 1692. Journal of Historical Pragmatics 3(2): 299–336. Fischer, Andreas. 1998. Marry. From religious invocation to discourse marker. In: Borgmeier, Grabes, and Jucker (eds.), 35–46. Fitzmaurice, Susan M. and Irma Taavitsainen (eds.). 2007. Methodological Issues in Historical Pragmatics. Berlin/New York: Mouton de Gruyter. Fritz, Gerd. 1995. Topics in the history of dialogue forms. In: Jucker (ed.), 469–498. Geis, Michael and Arnold M. Zwicky. 1971. On invited inferences. Linguistic Inquiry 2: 561–566. Gotti, Maurizio. 2006. Disseminating Early Modern science: Specialized news discourse in the Philosophical Transactions. In: Brownlees (ed.), 41–70. Herring, Susan C. 2003. Media and language change: Introduction. Journal of Historical Pragmatics (Special Issue on Media and Language Change, ed. by Susan C. Herring) 4(1): 1–17. Hiltunen, Risto. 2004. Salem, 1692: A case of courtroom discourse in a historical perspective. In: Risto Hiltunen and Shinichiro Watanabe (eds.), Approaches to Style and Discourse in English, 3–26. Osaka: Osaka University Press. Hiltunen, Risto. 2006. “Eala, geferan and gode wyrhtan”: On interjections in Old English. In: John Walmsley (ed.), Inside Old English: Essays in Honour of Bruce Mitchell, 91–116. Oxford: Balckwell. Hiltunen, Risto and Matti Peikola. 2007. Trial discourse and manuscript context: Scribal profiles in the Salem witchcraft records. Journal of Historical Pragmatics 8(1): 43–68. Honegger, Thomas. 2003. “And if ye wol nat so, my lady sweete, thanne preye I thee, […].”: Forms of address in Chaucer’s Knight’s Tale. In: Taavitsainen and Jucker (eds.), 61–84. Jacobsson, Mattias. 2002. Thank you and thanks in Early Modern English. ICAME Journal 26: 63–80. Jucker, Andreas H. 1997. The discourse marker well in the history of English. English Language and Linguistics 1(1): 91–110. Jucker, Andreas H. 2002. Discourse markers in Early Modern English. In: Richard Watts and Peter Trudgill (eds.), Alternative Histories of English, 210–230. London/New York: Routledge. Jucker, Andreas H. 2006. “Thou art so loothly and so oold also”: The use of ye and thou in Chaucer’s Canterbury Tales. Anglistik 17(2): 57–72. Jucker, Andreas H., Gerold Schneider, Irma Taavitsainen, and Barb Breustedt. 2008. Fishing for compliments: Precision and recall in corpus-linguistic compliment research. In: Jucker and Taavitsainen (eds.), 273–294. Jucker, Andreas H. and Irma Taavitsainen. 2000. Diachronic speech act analysis: Insults from flyting to flaming. Journal of Historical Pragmatics 1(1): 67–95. Jucker, Andreas H., and Irma Taavitsainen. 2008b. Apologies in the history of English: Routinized and lexicalized expressions of responsibility and regret. In: Jucker and Taavitsainen (eds.), 229–244. Jucker, Andreas H. (ed.). 1995. Historical Pragmatics. Pragmatic Developments in the History of English. Amsterdam/Philadelphia: John Benjamins. Jucker, Andreas H., Gerd Fritz, and Franz Lebsanft (eds.). 1999. Historical Dialogue Analysis. Amsterdam/Philadelphia: John Benjamins. Jucker, Andreas H. and Irma Taavitsainen (eds.). 2008a. Speech Acts in the History of English. Amsterdam/Philadelphia: John Benjamins.

13 Linguistic Levels: Pragmatics and discourse Kahlas-Tarkka, Leena and Matti Rissanen. 2007. The sullen and the talkative: Discourse strategies in the Salem examinations. Journal of Historical Pragmatics 8(1): 1–24. Keller, Rudi. 1994. On Language Change. The Invisible Hand in Language. London: Routledge. Kilian, Jo¨rg. 2005. Historische Dialoganalyse. Tu¨bingen: Niemeyer. Koch, Peter. 1999. Court records and cartoons: Reflections of spontaneous dialogue in Early Romance texts. In: Jucker, Fritz, and Lebsanft (eds.), 399–429. Kohnen, Thomas. 2000. Explicit performatives in Old English: A corpus-based study of directives. Journal of Historical Pragmatics 1(2): 301–321. Kohnen, Thomas. 2007. From Helsinki through the centuries: The design and development of English diachronic corpora. In: Pa¨ivi Pahta, Irma Taavitsainen, Terttu Nevalainen, and Jukka Tyrkko¨ (eds.), Studies in Variation, Contacts and Change in English, Vol. 2. http:// www.helsinki.fi/varieng/journal/volumes/02/ Kohnen, Thomas. 2008a. Directives in Old English: Beyond politeness? In: Jucker and Taavitsainen (eds.), 27–44. Kohnen, Thomas. 2008b. Tracing directives through text and time: Towards a methodology of a corpus-based diachronic speech-act analysis. In: Jucker and Taavitsainen (eds.), 295–310. Levinson, Stephen C. 1983. Pragmatics. Cambridge: Cambridge University Press. Ma¨kinen, Martti. 2004. Herbal recipes and recipes in herbals: Intertextuality in early English medical writing. In: Taavitsainen and Pahta (eds.), 144–173. Mazzon, Gabriella. 2000. Social relations and form of address in the Canterbury Tales. In: Dieter Kastovsky and Arthur Mettinger (eds.), The History of English in a Social Context. A Contribution to Historical Sociolinguistics, 135–168. Berlin/New York: Mouton de Gruyter. Mey, Jacob L. 1998. Pragmatics. In: Jacob L. Mey (ed.), Concise Encyclopedia of Pragmatics, 716– 737. Amsterdam: Elsevier. Nathan, N. 1959. Pronouns of address in the Canterbury Tales. Mediaeval Studies xxi: 193–201. Person, Raymond R., Jr. 2009. Oh in Shakespeare: A conversation analytic approach. Journal of Historical Pragmatics 10(1): 84–107. Raymond, Joad (ed.). 2006. News Networks in Seventeenth Century Britain and Europe. London: Routledge. Raymond, Joad. 2003. Pamphlets and Pamphleteering in Early Modern Britain. Cambridge: Cambridge University Press. Sommerville, John. 1996. The News Revolution in England. Cultural Dynamics of Daily Information. New York/Oxford: Oxford University Press. Sperber, Dan, and Ira A. Noveck. 2004. Introduction. In: Ira A. Noveck and Dan Sperber (eds.), Experimental Pragmatics, 1–22. Houndmills: Palgrave Macmillan. Stein, Dieter. 2003. Pronominal usage in Shakespeare: Between sociolinguistics and conversational analysis. In: Taavitsainen and Jucker (eds.), 251–307. Studer, Patrick. 2008. Historical Corpus Stylistics. Media, Technology and Change. London: Continuum. Taavitsainen, Irma. 1995. Interjections in Early Modern English: From imitation of spoken to conventions of written language. In: Jucker (ed.) 439–465. Taavitsainen, Irma. 1999. Dialogues in Late Medieval and Early Modern English medical writing. In: Jucker, Fritz, and Lebsanft (eds.), 243–268. Taavitsainen, Irma. 2006. Medical discourse: Early genres, 14th and 15th centuries. In: Keith Brown (ed.), Encyclopedia of Language and Linguistics, 2nd edn., 688–694. Oxford: Elsevier. Taavitsainen, Irma and Andreas H. Jucker (eds.). 2003. Diachronic Perspectives on Address Term Systems. Amsterdam/Philadelphia: John Benjamins. Taavitsainen, Irma and Andreas H. Jucker. 2007. Speech act verbs and speech acts in the history of English. In: Susan M. Fitzmaurice and Irma Taavitsainen (eds.), Methods in Historical Pragmatics, 107–138. Berlin: Mouton de Gruyter. Taavitsainen, Irma and Pa¨ivi Pahta (eds.). 2004. Medical and Scientific Writing in Late Medieval English. Cambridge: Cambridge University Press. Traugott, Elizabeth Closs. 2004. Historical pragmatics. In: Laurence R. Horn and Gregory Ward (eds.), The Handbook of Pragmatics, 538–561. Oxford: Blackwell.

211

212

II Linguistic Levels Traugott, Elizabeth Closs and Richard B. Dasher. 2005. Regularity in Semantic Change. Cambridge: Cambridge University Press. Trosborg, Anna. 1994. Interlanguage Pragmatics. Requests, Complaints and Apologies. Berlin/New York: Mouton de Gruyter. Ungerer, Friedrich (ed.). 2000. English Media Texts Past and Present. Amsterdam/Philadelphia: John Benjamins. Valkonen, Petteri. 2008. Showing a little promise: Identifying and retrieving explicit illocutionary acts from a corpus of written prose. In: Jucker and Taavitsainen (eds.), 247–272. Valle, Ellen. 1997. A scientific community and its texts: A historical discourse study. In: BrittLouise Gunnarsson, Per Linell, and Bengt Nordberg (eds.). The Construction of Professional Discourse, 76–98. London/New York: Longman. Valle, Ellen. 1999. A Collective Intelligence. The Life Sciences in the Royal Society as a Scientific Discourse Community, 1665–1965. Turku: University of Turku. Valle, Ellen. 2006. Reporting the doings of the curious: Authors and editors in the Philosophical Transactions of the Royal Society of London. In: Brownlees (ed.), 71–90. Verschueren, Jef. 1987. The pragmatic perspective. In: Marcella Bertuccelli-Papi and Jef Verschueren (eds.), The Pragmatic Perspective: Selected Papers from the 1985 International Pragmatics Conference, 3–8. Amsterdam/Philadelphia: John Benjamins. Verschueren, Jef. 1999. Understanding Pragmatics. London: Arnold. Watts, Richard J. 1999. Refugiate in a strange countrey: Learning English through dialogues in the 16th century. In: Jucker, Fritz, and Lebsanft (eds.), 215–241. Wilcockson, Colin. 1980. Thou and ye in Chaucer’s Clerk’s Tale. The Use of English 31(3): 37–43.

Andreas H. Jucker, Zu¨rich (Switzerland)

14 Linguistic Levels: Onomastics 1 2 3 4 5 6

Introduction Toponyms Anthroponyms Transmission of names Summary References

Abstract Names provide evidence for language history in two main respects: firstly, as regards lexical and semantic content when first coined; and secondly, as regards phonological and morphological development over the course of time. In neither respect is there widespread agreement as to the extent to which evidence from names can be extrapolated to other areas of language. On the one hand, both place-names and personal names testify to areas of vocabulary and registers of language sparsely represented in other sources; on the other, it is sometimes unclear whether these reflect ordinary language or a specialized onomastic usage. Factors pertaining to the formation and transmission of names are in some respects unique, and will be outlined in this chapter alongside a discussion of the main types of linguistic evidence preserved in the onomasticon. Alex Bergs and Laurel Brinton 2012, Historical Linguistics of English (HSK 34.1), de Gruyter, 212–223

14 Linguistic Levels: Onomastics

1 Introduction Onomastics is the study of names, its two main branches being toponymy (the study of place-names) and anthroponymy (the study of people’s names). Traditionally regarded as a sub-class of nouns having reference but no sense, names occupy a special position within language in that they can be used without understanding of semantic content. Partly for this reason, they tend to have a high survival rate, outlasting changes and developments in the lexicon, and easily being taken over by new groups of speakers in situations of language contact. Since most names originate as descriptive phrases, they preserve evidence for early lexis, often within areas of vocabulary sparsely represented in other sources. Many place-names, and some surnames, are still associated with their place of origin, so the data also contribute to the identification of dialectal isoglosses. Moreover, since names are generally coined in speech rather than in writing, they testify to a colloquial register of language as opposed to the more formal registers characteristic of documentary records and literary texts. Much research has been directed towards establishing the etymologies of names whose origins are no longer transparent, using a standard methodology whereby a comprehensive collection of early spellings is assembled for each name in order to trace its historical development. These spellings themselves can then be used to reveal morphological and phonological changes over the course of time, often illustrating trends in non-onomastic as well as onomastic language. The relationship between the two is not always straightforward, however, since the factors pertaining to the formation and transmission of names are in some respects unique. This chapter will discuss the main types of linguistic evidence preserved in names of various kinds, and will also consider the relevance of this evidence to other areas of language.

2 Toponyms 2.1 The origins of place-names The names of most villages, towns and cities in England were coined during the Anglo-Saxon period from Old English or (in areas of Scandinavian settlement) Old Norse. Others derive from the Celtic languages more strongly represented in Scotland, Wales and Ireland, while survivals from pre-Celtic linguistic strata are mainly found in the names of large topographical features. The names of major rivers are among the most ancient toponyms, and parallels between British and European river-names appear to reflect a system of hydronymy in use on the continent and brought to Britain by pre-Celtic immigrants. These river-names preserve evidence for the earliest form of language spoken in the British Isles, although it remains controversial whether this language was Indo-European – the majority view – or non-Indo-European. The names of smaller features tend to be later, dating from the medieval or early modern periods, and the same applies to the field-names given to units of cultivated land. Names for new urban developments continue to be created up to the present day, with street-names still being coined in large numbers. Nevertheless, even streetnames can be more than a thousand years old in the medieval parts of cities such as Derby, London, Nottingham and York.

213

214

II Linguistic Levels Research into the origins of English place-names has been carried out systematically on a county-by-county basis since the 1920s by the English Place-Name Society (1924–), and is published in a series of annual volumes known collectively as the English PlaceName Survey (EPNS). Whereas the Survey initially focused on names of historical significance such as medieval settlement-names, a growing recognition of the linguistic interest of the material led to coverage being expanded to include other types of names and those dating from later periods. This means that early volumes are not only less up-to-date than those currently being produced, but more limited in scope. Supplementary publications, including a field-name series, aim to redress the balance; and regional studies and dictionaries are also underway to provide coverage of the major names of all parts of England in advance of completion of the full Survey. Volumes 25–26 of the EPNS series comprise a dictionary of place-name terminology, a successor to which is currently in progress (Parsons, Styles, and Hough 1997–), and EPNS collections also form the basis for Field’s (1972) dictionary of field-names.

2.2 The structure of place-names Most English settlement-names are made up of one or (more usually) two elements, identical or closely related to vocabulary words. Most represent a description of a landscape feature or man-made structure, as with the single element or “simplex” names Dean or Deane (OE denu ‘main valley’), Ford (OE ford ‘ford, river-crossing’), Ham (OE hamm ‘hemmed-in land’), Hope (OE hop ‘small enclosed valley’), Lea or Leigh (OE le¯ah ‘wood, clearing’), Stoke (OE stoc ‘outlying farmstead’), Wick (OE wı¯c ‘specialized farm’) and Worth (OE worth ‘enclosure’). The corpus is highly repetitive, and each of the above examples occurs several times in different parts of England. Compound names provide a more precise description by including information on such aspects as appearance, ownership, usage, flora, or fauna. Examples include Abbotsley ‘Ealdbeald’s clearing’, Bagley ‘badger clearing’, Bradford ‘broad ford’, Bulwick ‘bull farm’ and Cotterstock ‘dairy farm’. Here, as in most non-Celtic place-names, the descriptive element or “specific” precedes the defining element or “generic”. Whereas the generic usually identifies a topographical feature or habitation, specifics have a much wider range, including personal names, older place-names, descriptive adjectives, animal, bird, and plant names, and occupational terms. Some names also contain an additional element generally known as an “affix” (although the term “distinguisher” might be more accurate), as with Stoke Mandeville, from the Mandeville family, and Stoke on Trent, from its position on the River Trent. In field-names and street-names, the generic is often a term for a field or street, with specifics again being more varied. In London, Chancery Lane refers to the chancellor’s office (ME chauncerie), Mincing Lane to nuns (OE myncen), and Sherborne Lane to a privy (ME *shite-burgh), while Mansfield Hill and Markfield Road preserve earlier field-names meaning ‘common field’ (OE (ge)mæ¯ne) and ‘boundary field’ (OE mearc) respectively (Mills 2001). However, non-literal formations are also common. Bare Arse, Labour in Vain, and Small Gains are recurrent field-names referring to unproductive land; irony may be suspected in some of the many fields called Paradise; the ubiquitous Hundred Acres almost invariably designates a very small field; and at least one Mount Pleasant refers to a refuse tip. A tendency for landscape features to receive metaphorical names is illustrated by the recurrent Cow and Calf Rocks (large and

14 Linguistic Levels: Onomastics small), and by various hill-names from the Scandinavian-derived carline ‘old woman’, including Carling Howe, Carling Knott and Fishcarling Head.

2.3 The language of place-names Many place-names are coined from terms also on record elsewhere, while others contain elements that are otherwise unknown, and for which the only evidence is the toponymicon. The repetitive nature of the corpus makes it possible to assemble a collection of names containing the same term in order to analyze its range of use. This applies particularly to generics. The most common element in English settlement-names, OE tu¯n ‘farmstead, village’, occurs in hundreds of names, and some of those mentioned above in more than a hundred. Many topographical generics can still be compared directly with the features described, and this has led to one of the main insights of place-name scholarship in recent years. The examination of places named from the same generic has identified subtle distinctions between the use of terms previously thought to be synonyms, revealing that the Anglo-Saxons had a more extensive and nuanced vocabulary for landscape features than has survived into later stages of English (Gelling and Cole 2000). Definitions given in Section 2.2 are summary only, and do not fully reflect the distinction between, for instance, OE denu, the standard term for a main valley, and OE hop, a small and often remote enclosed valley. Both differ from OE cumb ‘short, broad valley with three fairly steep sides’, OE slæd ‘flat-bottomed valley’, OE halh ‘nook’ (often referring to a less firmly-shaped valley than cumb or denu), and so on. Similarly precise usages have been established for other areas of topographical lexis, including terms unattested outside the place-name corpus such as OE *hlenc ‘extensive hill-slope’ and OE *ofer ‘flat-topped ridge with a convex shoulder’. Evidence from other Germanic languages may assist interpretation. A meaning of OE halh as ‘slightly raised ground in marsh’ is suggested by some place-name occurrences and supported by a similar use of the North Frisian cognate. Specifics represent a wider cross-section of open class vocabulary than generics, and contain a higher proportion of unattested words. Again the profile of use throws light on meaning. The specific of Bagley is an unattested OE *bagga. The range of generics with which it combines, including references to natural habitat and snares, as in Bag Hill (OE hyll ‘hill’), Bagshot (OE sce¯at ‘projecting land’) and Bawdrip (OE træppe ‘trap’), suggests a wild animal such as the badger, and this is supported by Germanic cognates. The specific of Grazeley (OE sol ‘wallowing-place’), Gresty (OE stı¯g ‘path’) and Greywell (OE wella ‘spring, stream’) is a substantive use of OE græ¯g ‘grey (animal)’ thought to refer to the wolf, as does an OE *wearg in place-names such as Warnborough and Wreighburn (both OE burna ‘stream’). Other words are attested in the place-name corpus earlier than in literary sources. OE *bagga may be the root of PDE badger, although this is uncertain. The specific of Bulwick is an OE *bula, the etymon of PDE bull, on independent record only from c.1200. Carling Howe is first recorded c.1170, whereas the earliest occurrence of carline ‘old woman’ recorded in the OED dates from c.1300. The compound *fish-carline ‘fish-wife’, attested in Fishcarling Head, is otherwise unknown. In some instances, place-names support other types of evidence for word meanings. The Dictionary of Old English (Cameron et al. 2008) tentatively suggests a meaning “ ‘churn, or even ‘dairy’ ” for three gloss occurrences of OE corþer, and this is confirmed

215

216

II Linguistic Levels by a toponymic use as the specific of Cotterstock (OE stoc ‘farm’). A substantive use of OE bru¯n ‘brown’ recorded as a nonce occurrence within a riddle appears to refer to a brown animal, possibly the pig, and again this is confirmed by its use in place-names such as Broomden (OE denn ‘woodland pasture, especially for swine’) and Brownwich (OE wı¯c ‘specialized farm’). Certain types of words, including animal names, bird names and topographical terms, are particularly well represented in place-names. So too are plant names (Hough 2003). The toponymic corpus considerably extends our knowledge of these areas of lexis, as of the colloquial range of vocabulary. Several of the above examples may represent demotic terms as opposed to the more formal registers of language represented in written sources. However, it is unclear to what extent the place-name corpus as a whole is representative of ordinary vocabulary. Some elements, particularly generics, are thought to belong to an onomastic register common to the northwest Germanic languages, which may have diverged from ordinary lexis at an early date (Nicolaisen 1995). Parallels between OE halh and its North Frisian cognate are mentioned above, and there are similar instances where pairs of generics in the West and North Germanic toponymica have more in common with each other than with the corresponding lexical terms. OE ha¯m ‘homestead, estate’, one of the earliest habitative generics in English place-names, is cognate with ON heimr, and this may help to explain why it appears to have been in use as a place-name forming element earlier than as a common noun. Other generics also have a different chronological or geographical profile from their lexical counterparts, while the fact that some elements (e.g. OE le¯ah) occur as generics only, apparently not being available for use as specifics, further supports the theory that they were not selected freely from the lexicon. It may therefore be appropriate to regard such elements as cognate with, rather than identical to, vocabulary items, deriving from a common ancestor but developing along different lines. When they began to develop separately is uncertain, but evidence from river-names suggests that a so-called “onomastic dialect” may already have emerged early in the history of Indo-European (Kitson 1996). Specifics may be closer to ordinary lexis, but again develop uses in place-names which are not necessarily the same as those in non-onomastic language. Analogy is common in place-name formation, and this limits the value of toponymic evidence for word geography as well as for historical semantics. The place-name distribution of carline, for instance, suggests a wider currency in northern dialects of Middle English than is supported by lexical evidence, extending to the north-west of England outside the Danelaw, but it is possible that this is due to the conventional use of the term for landscape features. Another factor is the high level of repetition within the name stock, which suggests that even some compounds may have been drawn from an existing pool. Ekwall (1960) includes 40 occurrences of Burton from OE burh-tu¯n ‘settlement by a fortification’, 29 of Charlton from OE ce¯orla-tu¯n ‘settlement of freemen’, and 20 of Easton from OE e¯ast-tu¯n ‘east settlement’, and it is unlikely that they were coined afresh on each occasion. Field-names and street-names preserve much vocabulary from the medieval and later periods, and comprise the bulk of the evidence for Middle English terms attested uniquely or earliest in place-names (Hough 2002). Street-names are a particularly rich source of occupational terms. Birchin Lane in London, recorded from the 12th century,

14 Linguistic Levels: Onomastics derives from an unattested ME *berdcherver ‘beard-cutter, barber’; Felter Lane in York, recorded from the 13th century, contains the term felter ‘felt-maker’ attested in the OED only from 1605; and Fletcher Gate in Nottingham, recorded in 1335 as Flesshewergate, is the earliest occurrence cited by the OED of the obsolete term flesh-hewer ‘butcher’. Again there is much repetition within the corpus. Some names are transferred directly from one place to another (e.g. Piccadilly, now found in Manchester and York as well as London), while some are coined by analogy with others. Street-names in -gate are common in areas of urban settlement within the former Danelaw, where they often derive from ON gata ‘street’; but in other instances they have simply been modelled on the original Viking names (Fellows-Jensen 2007). They cannot therefore be used as evidence for the currency of the term gata itself, nor for Scandinavian influence on the language. Nonetheless, microtoponyms make an important contribution to the study of Old and Middle English dialectology by offering a large quantity of data for analysis. On a national scale, they facilitate the mapping of synonyms, an approach taken in Kitson’s (1995) investigation of Old English dialect isoglosses, and in studies of the distribution of complementary terms such as whin, gorse, and furze (e.g. Cameron 2008). On a regional scale, they reflect a mix of languages more accurately than the smaller number of settlement names. This has been exploited in studies using the ratio between fieldname elements of Old English and Old Norse origin to measure the extent of Scandinavian impact on local dialect (e.g. Cameron 1996, Watts 2002, Parsons 2006).

3 Anthroponyms 3.1 Personal names Like place-names, Anglo-Saxon personal names are made up of one or (more usually) two elements drawn from a corpus corresponding closely to a subset of the lexicon. Animal names, abstract nouns, and descriptive adjectives are particularly well represented, as is the semantic field of warfare. Masculine and feminine names are formed along parallel lines, with the grammatical gender of the second element or “deuterotheme” often coinciding with the gender of the person. Examples include Æthelgar ‘noble-spear’, Guthfrith ‘battle-peace’, and Wulfstan ‘wolf-stone’ (all masculine), Æthelflæd ‘noblebeauty’, Æthelthryth ‘noble-strength’, and Wulfgifu ‘wolf-gift’ (all feminine). These and other names do not make literal sense, but may have signalled family relationships. Members of a kin group characteristically have names with alliterating (sometimes identical) first elements or “protothemes”, while the deuterotheme of a parent’s name may be passed on to children of the same sex. The naming system goes back to Common Germanic. Links have been identified with the vocabulary of heroic poetry, and here even more than with place-names there is reason to regard name elements as cognate with, rather than identical to, their lexical counterparts. Whereas most place-names were literally descriptive of their referents in Anglo-Saxon England, personal names clearly were not, and this lack of sense suggests that elements were drawn from an anthroponymicon rather than from the lexicon. The impression of an older system reflected but no longer motivated in the Anglo-Saxon naming tradition is strengthened by instances where the correlation between natural and grammatical gender is not consistently maintained.

217

218

II Linguistic Levels Elements corresponding to feminine nouns such as OE mund ‘protection’ and OE no¯th ‘boldness’ form masculine names (e.g. Byrhtnoth ‘bright-boldness’, Sigemund ‘victoryprotection’), as do elements corresponding to neuter nouns such as OE cild ‘young person’ (e.g. Leofcild ‘dear-child’) and to adjectives such as OE beald ‘bold’ and OE beorht ‘bright’ (e.g. Æthelbeald ‘noble-bold’, Ealdbeorht ‘old-bright’). Although some elements function both as protothemes and as deuterothemes, others appear to be restricted to one position only, and this lack of flexibility – together with the fact that not all vocabulary words appear to have been available for use in coining names – again points to some fossilization of the system. This means that where a name element is unattested in literary sources from the Anglo-Saxon period, as with *flæd, whose meaning is reconstructed from Germanic cognates as “elegance or daintiness as of a courtly lady” (Kitson 2002: 97), it cannot be assumed that it was ever in use as a lexical item in Old English. Anglo-Saxon personal names were largely replaced by Continental Germanic and Biblical names after the Norman Conquest, and are sparsely represented in the present-day name stock. Although personal names have occasionally been coined from vocabulary words, as with the “virtue” names associated with the Puritan movement in post-Reformation England (e.g. Hope, Joy, Patience), and others referring to precious stones, plants, and the like (e.g. Ruby, Heather, Holly), the vast majority of names in current use are of Biblical or classical origin, and even those that remain semantically transparent are generally used without reference to etymological meaning. Recent research has emphasized the importance of social and cultural factors in present-day and early modern name giving, with linguistic origin being a lesser consideration.

3.2 Bynames Prior to the evolution of surnames, bynames were used to differentiate between people with the same personal name. Already in use in Anglo-Saxon England, they became much more common during the early Middle English period, partly because of the dwindling number of baptismal names in general use. By the late 13th century, most personal names were routinely qualified by a byname. The four main categories of bynames are local, familial, occupational, and characteristic. Examples from Clark’s (1983) study of King’s Lynn bynames are atte Ling ‘beside the heather’ (local), Lellesmai ‘Lelle’s kinsman’ (familial), Le Blekestere ‘the bleacher’ (occupational) and Le Longe ‘the tall’ (characteristic). Bynames differ from the other types of names discussed above in being literally descriptive, and drawn freely from the lexicon. Despite having much in common with surnames, they were not hereditary (nor shared with siblings), and hence did not lose their lexical meaning. Because of this, they can be taken to represent contemporary language rather than fossilized forms. Nonetheless, they may be difficult to interpret. Bynames such as Ioie ‘joy’ and Trouthe ‘truth’ may be understood literally or ironically, and Clark discusses ambiguities surrounding others such as Baril ‘barrel’ and Peper ‘pepper’, which may be occupational referring to a barrel-maker and spicer, or characteristic referring to corpulence and hot temper. Like place-names, bynames preserve a number of words unattested elsewhere or on record only from a later date. Tengvik’s seminal study of Old English bynames identifies 58 antedated words (Tengvik 1938: 24–26), and others have since been added. Some are animal- or bird-names used as characteristic bynames. Otherwise unknown terms

14 Linguistic Levels: Onomastics include OE *gor-pyttel ‘dung-hawk’, recorded as the byname of Gotselin Gorpittel (c.1100–1130). Occupational bynames are an important source of Middle English occupational terms, and several of the King’s Lynn bynames either antedate or are close to antedating the earliest documentation in literary sources. These include Oylman (a worker at the town’s oil-mills) and Habertasker ‘haberdasher’. Previously unrecorded occupational terms from the same corpus include *candelwif and *chesewoman, referring to women involved in the candle- and cheese-making industries. Topographical terms are preserved in local bynames, again often providing antedatings (e.g. Carlsson 1989: 146–147). The languages represented in bynames also make it possible to trace linguistic influences on local dialect. Le Blekestere is based on a Scandinavian loanword, while trading links with France are reflected in the definite article and in French forms such as Baril.

3.3 Surnames Like bynames, surnames evolved to differentiate between people with the same personal name; but unlike bynames, they were passed on from one generation to the next, functioning as markers of relationship rather than as descriptions of individuals. Surnames began to come into use in England during the late 11th century, partly as a result of changes in society following the Norman Conquest, but the system was not fully established until about 1400. The main types of surnames correspond to the main types of bynames: local, familial, occupational, and characteristic. The earliest were local surnames held by landowning families, some imported from Normandy and others taken from the names of estates in England. The use of surnames gradually spread down the social scale, with occupational surnames associated particularly with skilled craftsmen, and characteristic surnames with the lower classes. Local surnames – those referring to place of origin, or to ownership of land – can themselves be divided into two categories: locative and topographical. Locative surnames derive from place-names, topographical surnames from lexical words. Thus the surname Glasgow is locative, from the Scottish city, while the surname Hough is topographical, from OE ho¯h ‘spur of land’. The latter type can reflect regional variation in morphology as well as lexis. Examples such as Bridger and Bridgeman ‘dweller by the bridge’ and Weller and Wellman ‘dweller by the stream’ show the morphemes -er and -man suffixed to a topographical term to indicate place of residence. This type of surname is common in central and southern England, but rare in the north, where the morphemes do not seem to have been productive in this sense. Familial surnames, or surnames of relationship, most commonly derive from parents’ forenames. Those from a father’s name are known as “patronymics”; those from a mother’s name as “metronymics”. The earliest derive from unaffixed personal names, whether from Old English, as with Godwin, or from the continental name stock introduced to England after the Conquest, as with Allen, Maude, and Thomas. Surnames with the suffix -son are slightly later and often formed from shortened, or hypochoristic, forms of personal names, as with Dobson and Robson (Robert), Ibbotson and Ibson (Isabel), and Wilson and Wilkinson (William). Again there are regional variations. Most surnames in -son originated in the north and north midlands of England, and in southern Scotland. In southern and central England, the suffix -s was more common, as in Andrews, Roberts and Williams.

219

220

II Linguistic Levels Occupational surnames reflect lexical variation in different areas (McKinley 1990: 143–147). Surnames relating to the same trade within the cloth industry are Fuller (south and south-east), Tucker (south-west), and Walker (north); while Barker (north) and Tanner (south) are synonyms. This group is also a primary source of information on Middle English occupational terms. Thuresson (1950) deals with about 850 words, of which 271 were not recorded in the OED, and Fransson (1935) also provides a number of new words and predatings. Some surnames from characteristics are based on English or French phrases which appear to have had a wide currency in Middle English despite being rarely if at all recorded elsewhere. Examples such as Fairwether, Goodall ‘good ale’, Makehate ‘make joy’, Parlebien ‘well spoken’, Passavaunt ‘go before’, Proudfoot, and Spendlove are suggested by McKinley (1990: 166) to “perhaps convey some flavour of the ordinary man’s spoken English as it was in the 13th and 14th centuries, when most of these names arose”.

4 Transmission of names The lack of sense that is widely considered a defining characteristic of names allows them to preserve fossilized forms of words that have long since disappeared from the lexicon. Moreover, the fact that most place-names are locatable in space, and many personal names locatable in time, gives them an advantage over texts preserved in manuscripts of uncertain provenance. Forms of moneyers’ names on Anglo-Saxon coins can be ordered chronologically by coin-types, providing closely datable material for studies of Old English phonology (e.g. Colman 1984, 1992), while the range of spellings relating to individual place-names and place-name elements can reveal dialectal and morphological variation. In East Anglia, spellings in place-names from OE stre¯t/stræ¯t ‘Roman road’ (e.g. Stradsett, Stradbroke, Stradishall) testify to a development from Saxon æ¯-forms rather than Anglian e¯-forms in early Old English (Kristensson 2001); variation between Scandinavian and English lexical terms in spellings of individual topographical names from the 14th to 16th centuries throws light on local dialect during the Middle English period (Sandred 2001: 51); and 13th- and 14th-century forms of the field-names Hanging Furlong and Hanging Wong with the present participle in reflect Scandinavian morphological influence (Sandred 2001: 51–52). This type of evidence needs to be handled with care, however, as it is also widely recognized that names may behave differently from vocabulary items. Lass (1973) points out that many surnames could not have developed their modern form through standard processes of phonological change, and illustrates this through a detailed analysis of the surname Shuttlebotham and its variants Shipperbottom, Shuflebotam, Shovelbottom, Shoebotham, Shoebottom, and Shubotham. Colman (1988) too urges caution in the use of names as evidence for linguistic reconstruction and historical dialectology. Some name spellings become fossilized, failing to reflect phonological variation, while others reflect the influence of more than one etymon. Here as in the initial formation of names, analogy plays a key role. Following the loss of semantic meaning, element substitution may occur, with familiar elements tending to replace less familiar ones. The common element wulf ‘wolf’, often spelled in Old English personal names, replaces the original deuterotheme col ‘coal’ of the moneyers’ names Sæcol

14 Linguistic Levels: Onomastics and Swartcol in the later spellings and ; the topographical placename generic OE le¯ah, often surviving as , influences the development of names such as Hawksley and Notley from hafoces hlewe ‘hawk’s tumulus’ and hnut clyf ‘nut slope’ (Gelling 1997: 202–205), and of Grazeley from grægsole; and the development of Warnborough from weargeburnan shows interference from the habitative generic OE burh ‘fortification’. It is particularly common for place-names to be influenced by neighbouring ones, and Coates (1987: 329) cites 24 pairs of place-names, mostly less than five miles apart, where one has developed along non-standard lines through the influence of the other. Analogy also operates at the level of the whole name. The ubiquitous Burton, from OE burh-tu¯n, has attracted to itself place-names from different specifics. Burton Bradstock contains the name of the River Bride, Burton in Sussex contains an Old English personal name Budeca, and Burton Salmon contains an adjective OE bra¯d ‘broad’. Similarly, a Middlesex place-name containing a personal name Ceolred has developed into another Charlton, and a Devon place-name containing a personal name Ælfric or Æthelric into another Easton. In each case, the influence of analogy has overridden standard processes of phonological change. Spelling patterns may also be affected by the medium in which the names are recorded, an area explored by Anderson and Colman (2004) in relation to wrap-around conventions and the transposition of graphs in the representation of personal names on Anglo-Saxon coins. Another factor is the role of government in regulating coinage and documenting names. The recurrence of certain forms on coins may reflect not pronunciation but an attempt to regularize spelling (Smart 1983); while students of placenames discriminate carefully between local spellings and those emanating from a centralized administration. The compound structure characteristic of both personal and place-name formations may also lead them to develop differently from their lexical cognates. In some instances, phonological changes are reflected earlier than in written texts. The loss of /d/ from word-final /nd/, common in written texts from the 15th century onwards, is attested as early as the late 12th and 13th centuries in the stressed (first) syllable of placenames and surnames, where it is attributed to the influence of the consonant initial in the second part of the compound (Wełna 2005). Consonant loss at the juncture of compound personal names is discussed alongside other issues specific to compounds by Colman (1984: 126–136), who demonstrates the value of Old English name-spellings as evidence for phonological developments in obscured compounds (i.e. compounds that have undergone phonological reduction and semantic obscuration). In general, motivated change affects names more than ordinary lexis. Personal names and surnames are subject to idiosyncratic choices as regards both pronunciation and spelling, and place-names too may be “improved”, as with Sherborne Lane in London mentioned above. Early spellings show that it preserves an unattested ME *shite-burgh ‘privy’, but this has been changed for reasons of delicacy. In other instances, an element that has become semantically opaque is reshaped to conform with known lexis, or simply replaced by a word that appears to make more sense. Mincing Lane in London derives from the obsolete word myncen ‘nun’, and Fletcher Gate in Nottingham from flesh-hewer ‘butcher’ – apparently associated with the later occupational term fletcher ‘arrow-maker’, itself now obsolete but plausible as a street-name specific. The same process of folk etymology accounts for the development (from a

221

222

II Linguistic Levels Lancashire place-name) of the surnames Shovelbottom, Shoebottom, and so on. So too the loss of much of the Anglo-Saxon personal name stock following the Norman Conquest led to the names becoming unfamiliar and being reshaped where they appear in surnames or place-names. The surname Freelove derives from an Old English personal name Frithulaf, while the first element of Abbotsley is an Old English personal name Ealdbald. Because names are so often affected by folk etymology, they do not always follow standard patterns of phonological or morphological development. For the same reason, however, they provide evidence for the operation of folk etymology itself.

5 Summary In sum, the relationship between names and lexis is not straightforward, but this does not diminish the value of onomastic material in the study of historical linguistics. Although neither the initial formation nor the subsequent transmission of names directly parallels the lexicon, the differences are themselves enlightening and reveal information unavailable from other sources. Handled with appropriate caution, onomastic evidence provides insights which both supplement and extend those offered by other areas of language.

6 References Anderson, John and Fran Colman. 2004. Non-rectilinear name-forms in Old English and the media of language. In: Gunnar Bergh, Jennifer Herriman, and Mats Moba¨rg (eds.), An International Master of Syntax and Semantics: Papers Presented to Aimo Seppa¨nen on the Occasion of his 75th Birthday, 31–42. Go¨teborg: Acta Universitatis Gothoburgensis. Cameron, Angus, Ashley Crandall Amos, Antonette diPaolo Healey et al. 2008. Dictionary of Old English: A-G on CD-ROM. Fascicle G and Fascicles A to F (with revisions). Toronto: Dictionary of Old English Project. http://www.doe.utoronto.ca/pages/pub/fasc-a-g.html Cameron, Jean. 2008. The distribution of whin, gorse and furze in English place-names. In: O. J. Padel and David N. Parsons (eds.), A Commodity of Good Names: Essays in Honour of Margaret Gelling, 253–258. Donington: Shaun Tyas. Cameron, Kenneth. 1996. The Scandinavian element in minor names and field-names in northeast Lincolnshire. Nomina 19: 5–27. Carlsson, Stig. 1989. Studies on Middle English Local Bynames in East Anglia. Lund: Lund University Press. Clark, Cecily. 1983. The early personal names of King’s Lynn: an essay in socio-cultural history. Part II – by-names. Nomina 7: 65–89. Coates, Richard. 1987. Pragmatic sources of analogical reformation. Journal of Linguistics 23: 319–340. Colman, Fran. 1984. Anglo-Saxon pennies and Old English phonology. Folia Linguistica Historica 6: 91–143. Colman, Fran. 1988. What is in a name? In: Jacek Fisiak (ed.), Historical Dialectology: Regional and Social, 111–137. Berlin/New York: Mouton de Gruyter. Colman, Fran. 1992. Money Talks. Berlin/New York: Mouton de Gruyter. Ekwall, Eilert. 1960. The Concise Oxford Dictionary of English Place-Names. 4th edn. Oxford: Clarendon Press. English Place-Name Society. 1924–. The Survey of English Place-Names. Vols. 1–. Cambridge/ Nottingham: Cambridge University Press and English Place-Name Society. Fellows-Jensen, Gillian. 2007. The Scandinavian element gata outside the urbanised settlements of the Danelaw. In: Beverley Ballin Smith, Simon Taylor, and Gareth Williams (eds.), West Over

14 Linguistic Levels: Onomastics

223

Sea: Studies in Scandinavian Sea-Borne Expansion and Settlement Before 1300. A Festschrift in Honour of Dr Barbara E. Crawford, 445–459. Leiden/Boston: Brill. Field, John. 1972. English Field-Names. A Dictionary. Newton Abbot: David & Charles. Fisiak, Jacek and Peter Trudgill (eds.). 2001. East Anglian English. Cambridge: D. S. Brewer. Fransson, Gustav. 1935. Middle English Surnames of Occupation 1100–1350. Lund: Gleerup. Gelling, Margaret. 1997. Signposts to the Past: Place-Names and the History of England. 3rd edn. Chichester: Phillimore. Gelling, Margaret and Ann Cole. 2000. The Landscape of Place-Names. Stamford: Shaun Tyas. Hough, Carole. 2002. Onomastic evidence for Middle English vocabulary. In: Peter J. Lucas and Angela M. Lucas (eds.), Middle English from Tongue to Text. Selected Papers from the Third International Conference on Middle English: Language and Text, held at Dublin, Ireland, 1–4 July 1999, 155–167. Frankfurt: Peter Lang. Hough, Carole. 2003. Place-name evidence for Anglo-Saxon plant-names. In: C. P. Biggam (ed.), From Earth to Art. The Many Aspects of the Plant-World in Anglo-Saxon England, 41–78. Amsterdam/New York: Rodopi. Kitson, Peter R. 1995. The nature of Old English dialect distributions, mainly as exhibited in charter boundaries. In: Jacek Fisiak (ed.), Medieval Dialectology, 43–135. Berlin/New York: Mouton de Gruyter. Kitson, Peter R. 1996. British and European river-names. Transactions of the Philological Society 94: 73–118. Kitson, Peter R. 2002. How Anglo-Saxon personal names work. Nomina 25: 91–131. Kristensson, Gillis. 2001. Language in contact: Old East Saxon and East Anglian. In: Fisiak and Trudgill (eds.), 63–70. Lass, R. 1973. Review of P. H. Reaney, The Origin of English Surnames, Routledge and Kegan Paul, London, 1967. Foundations of Language 9: 392–402. McKinley, Richard. 1990. A History of British Surnames. London/New York: Longman. Mills, A. D. 2001. A Dictionary of London Place Names. Oxford: Oxford University Press. Nicolaisen, W. F. H. 1995. Is there a Northwest Germanic toponymy? Some thoughts and a proposal. In: Edith Marold and Christiane Zimmermann (eds.), Nordwestgermanisch, 102–114. Berlin/New York: Walter de Gruyter. Parsons, David N. 2006. Field-name statistics, Norfolk and the Danelaw. In: Peder Gammeltoft and Bent Jørgensen (eds.), Names Through the Looking-Glass: Festschrift in Honour of Gillian Fellows-Jensen July 5th 2006, 165–188. Copenhagen: C. A. Reitzels. Parsons, David N. and (for Vols. 1 and 2) Tania Styles with (for Vol. 1) Carole Hough. 1997–. The ´ –Box); 2 (Brace–Cæster); 3 (Ceafor–Cock-pit). NotVocabulary of English Place-Names: 1 (A tingham: Centre for English Name Studies (Vols. 1 and 2) and English Place-Name Society (Vol. 3). Sandred, Karl Inge. 2001. East Anglian place-names: Sources of lost dialect. In: Fisiak and Trudgill (eds.), 39–61. Smart, Veronica. 1983. Variation between Æthel- and Ægel- as a name-element on coins. Nomina 7: 91–96. Tengvik, Go¨sta. 1938. Old English Bynames. Uppsala: Almqvist & Wiksells. Thuresson, Bertil. 1950. Middle English Occupational Terms. Lund: Gleerup. Watts, Victor. 2002. Medieval field-names in two South Durham townships. Nomina 25: 53–64. Wełna, Jerzy. 2005. “Now you see it, now you don’t” once more: The loss and insertion of dental stops in medieval English. Studia Anglica Posnaniensia 41: 71–84.

Carole Hough, Glasgow (UK)

224

II Linguistic Levels

15 Linguistic Levels: Orthography 1 Theoretical approaches to language, speech, and writing: between referentiality and autonomy 2 Definitions of orthography and related terms 3 Classification of writing systems and principles governing English orthography 4 Units of writing systems: terminological evolution 5 The inventory and distribution of English graphemes from the synchronic and diachronic perspective 6 Orthography as the source of phonological evidence 7 Sociolinguistic aspects of orthography 8 Summary 9 References

Abstract This chapter offers a critical overview of some of the most influential ideas concerning writing in general, and orthographic systems, with particular attention paid to English orthography and its intricate structure. It also presents and explains the terminology which can be found in literature dealing with this subject. Section 1 gives an account of different attitudes to writing and the two main theoretical approaches, relational and autonomistic, which they have motivated. It is followed by a summary of selected definitions of orthography and related terms in Section 2. Section 3 examines various types of writing systems, places the English orthographic system in that taxonomic context, and discusses its governing principles. Section 4 deals with the evolution of terms used to denote units of writing systems. In the Section 5 all the graphemes of English orthography are listed and selected historical aspects of their evolution are mentioned. Section 6 contains a brief evaluation of orthography as the source of evidence for phonological change. Finally, the sociolinguistic aspects of orthography are identified, including its role as a binding norm within a language community, and the significance of orthographic variation.

1 Theoretical approaches to language, speech, and writing: between referentiality and autonomy In order to provide a definition of orthography and put it in the appropriate theoretical context, one must first consider and define the general terms to which it is related. These are, most importantly, “language”, “speech”, and “writing”. It seems quite clear at present that speech and writing differ in numerous ways, including, for instance, the purely physical properties, as well as the communicative situations and the purposes for which each of them is used. Thus, speech is time-bound, fleeting, spontaneous, used to express opinions and emotions, whereas writing is space-bound, permanent, planned, edited, and used to record and convey information (detailed accounts of these differences can be found, e.g., in Crystal 1995: 291–293 and Cook 2004: 31–53). In fact, both speech and writing can be treated as complementary ways of using language, where the latter is understood as “a complex system residing in our brain which allows Alex Bergs and Laurel Brinton 2012, Historical Linguistics of English (HSK 34.1), de Gruyter, 224–237

15 Linguistic Levels: Orthography us to produce and interpret utterances” (Rogers 2005: 2). However, there is no general agreement among linguists as regards the relation between speech and language on one hand, and writing and language, as well as writing and orthography, on the other. Since the beginning of the 20th century, these notions have been viewed from two main points of view labelled as “relational” and “autonomistic” by Sgall (1987: 2–3; see also Ruszkiewicz 1976: 37–44). The structuralists, representing the relational perspective, equated speech with language, and considered writing as an extra-linguistic phenomenon: – “Language and writing are two distinct systems of signs; the second exists for the sole purpose of representing the first” (Saussure 1993: 41a). – “The written forms are secondary symbols of the spoken ones – symbols of symbols” (Sapir 1921: 20). – “Writing is not language, but merely a way of recording language by means of visible marks” (Bloomfield 1933: 21). – “The linguist distinguishes between language and writing” (Hockett 1958: 4). These quotations, to some extent, echo the classical theories of writing, which assumed a close dependence of written language on spoken language. For example, “[i]n De Interpretatione, Aristotle (1963, 43) states that ‘spoken sounds are symbols of affections of the soul, and written marks [are] symbols of spoken sounds’ ” (Liuzza 1996: 35). Also Quintilian claimed that “spelling should follow pronunciation because the text is a repository for the vox” (1920, 1:144; quoted in Liuzza 1996: 35). Because of the influence exerted by Saussure, Bloomfield, and other early structuralists, the validity of their statements has mostly been taken for granted. According to Venezky (1970: 27), “Bloomfield is responsible probably more than any other contemporary linguist for the view that writing is secondary and subservient to speech”, because he consistently expressed this opinion in several publications. In contrast to Saussure, Bloomfield, and others, scholars adhering to the autonomistic point of view, e.g., Vachek, Bolinger, Stetson, McIntosh, McLaughlin, and Venezky, have claimed that “[m]uch is written that is not pronounced” (Stetson 1981 [1937]: 35), and that “writing does more than represent speech” (Cook 2004: 32). More precisely, according to them “[w]riting is any manifestation of language in visible signs; a written language is a code that may not need preliminary decipherment into speech to be understood.” (Vachek 1982: 38). In this view, the written language “has to a great extent become an instrument for the direct expression of meaning, co-ordinate with audible language. The result of this has been that the written language has in part been developed on lines of its own, independent of the development of oral speech.” (Bradley 1919 [1913]: 14–15; quoted by Liuzza 1996: 27). Although the preoccupation with spelling as the reflection of speech was the dominant feature of discussions on orthography in the 20th century, the adherents of the autonomistic approach to writing also had some forerunners to draw upon. Already in the 16th and 17th centuries, some scholars (e.g. William Bullokar, Alexander Gil, Alexander Hume, and John Wallis) realized that orthography carried some non-phonemic information. Venezky (1970: 18, 23–25) mentions also the writings of Goold Brown (especially his comprehensive grammar published in 1850), which clearly expressed the idea that “words are not mere sounds, and in their orthography more is implied than in phonetics

225

226

II Linguistic Levels or phonography. Ideographic forms have, in general, the advantage of preserving the identity, history, and lineage of words” (quoted in Venezky 1970: 18, 24). Brown did not deny the close relationship between letters and sounds, but he also emphasized the relative autonomy of the writing system: “The deaf and the dumb, also, to whom none of the letters express or represent sounds, may be taught to read and write understandingly. […] Hence it would appear that the powers of the letters are not, of necessity, identified with their sounds” (both quotations found in Venezky 1970: 23–24). The second citation is reminiscent of a much earlier paper written by Wallis (on teaching the deaf to speak), where the author claims that “there is nothing, in the nature of the Thing it self, why Letters and Characters might not as properly be applyed to represent Immediately, as by Intervention of Sounds, what our Conceptions are” (Wallis: 1670: 1091). The linguists’ varying attitudes to written language clearly correspond to different expectations of its aims and functions. The early structuralists’ disparaging view of writing has influenced the definitions of orthography, with which writing has commonly been confused (Vachek 1976 [1945–49]: 128). It motivated the treatment of orthography as an imperfect device for obtaining information about speech, and popularized and somewhat fossilized the perception of orthography as a non-linguistic subject. As a result, the English orthographic system has rarely been devoted separate sections, let alone chapters, in historical grammars of English and handbooks of linguistics. Obviously, as it has been impossible to ignore orthography completely, elements of it have been mentioned in discussions of English phonology (see Kniezsa 1991 for an overview of histories of English spelling). Summing up, the main differences between linguists expressing their views about the written language and its connection to speech concern the level of autonomy and the linguistic status of writing systems. Because the purely relational perspective seems biased and limited, in what follows, I have taken the more balanced, autonomistic view towards writing, and consequently towards orthography, but I will also refer to the existing competing views wherever necessary (see, especially, Section 4 on the units of writing systems).

2 Definitions of orthography and related terms According to the OED (Simpson [ed.] 2000–), the word orthography came into English from Old French ortografie, but it originates in Greek. Since its appearance in English in the mid-15th century, it has retained the prescriptive sense of a spelling norm, where spelling is ‘the manner of expressing words with letters’. In the 16th century, it also acquired the descriptive meaning of ‘that part of grammar which treats of the nature and values of letters and of their combination to express sounds and words’. The term orthography has also been defined in various other ways, and its meaning at least partly overlaps with that of other expressions. For example, although in everyday usage orthography is often understood as synonymous only with spelling, in a more technical sense, the orthography of a specific language comprises spelling as well as the capitalization, punctuation, and word division permitted in that language. Spelling is most closely connected with the levels of phonology, morphology, and lexicon, whereas capitalization, punctuation, and word division show some correspondences also with syntax, semantics, and stylistics. Thus, in accordance with this broader sense, one can assume the following definition: “[t]he standardized writing system of a language is known as its orthography” (Crystal 1995: 257).

15 Linguistic Levels: Orthography Another term associated with orthography is graphology, introduced by McIntosh (1961: 107) to denote the study of orthographic systems. The advantage of that term is the parallelism to other linguistic levels, such as phonology and morphology. That term was defined further by Halliday et al. (1964: 50) as “orthography, punctuation, and anything else that is concerned with showing how a language uses its graphic resources to carry its grammatical and lexical patterns”. In this sense, graphology and orthography can also be considered synonymous to the writing system. Graphemics, first recorded in 1951, by analogy to phonemics (Pulgram 1951: 19, see also Stockwell and Barritt 1951 on the relational view of graphemics) is another synonym of orthography. It is defined in the OED as ‘the study of systems of written symbols (letters etc.) in their relation to spoken languages’. However, some linguists have suggested that “the term graphemics should be confined to the study of systems of writing only” (Bazell 1981 [1956]: 68), as well as postulated the introduction of the term graphophonemics for “[t]he discipline concerned with the study of the relationship between graphemics and phonemics” (Ruszkiewicz 1976: 49). The term graphotactics, in turn, usually refers to the syntax of graphemes (understood as units of an orthographic system) (see Haas 1970: 59, Carney 1994: 66–69), defined also as “the laws governing [the] combination of graphemes” (Vachek 1973: 9). All the expressions mentioned above are still found today in descriptions of writing systems, and the choice of particular terminology usually depends on the specific focus and theoretical approach of the writer.

3 Classification of writing systems and principles governing English orthography It is currently assumed that writing systems (or “scripts”) can be classified according to what linguistic level is represented by the written symbols (or graphemes). Thus, the writing systems of natural languages can be divided into “morphographic” and “phonographic” (Rogers 2005: 272). In morphographic systems, the symbols are related primarily to morphemes. Some of those morphemes may constitute words, and therefore such systems have been often referred to as “logographic” (from Greek logos ‘word’, OED; see Sampson 1985; Venezky 1999). Chinese and Sumerian are typical examples of writing systems showing this type of relationship. In phonographic systems, the symbols relate to phonological units. Depending on the type of unit represented, they are divided into “syllabic”, “moraic”, and “phonemic” (or “alphabetic”) (Rogers 2005: 272). In moraic writing, graphemes correspond to morae, where a “mora” is “either a syllableinitial CV sequence or a codal (final) consonant” (Rogers 2005: 250). Moraic systems, such as the Japanese katakana, and Cherokee, have traditionally (though rather imprecisely) been called “syllabic” systems (Sampson 1985; Venezky 1999). In phonemic systems, graphemes represent mainly phonemes, or segments, including consonants and vowels. This type of representation has also been referred to as “alphabetic” (Venezky 1999: 4). The phonemic (or alphabetic) principle implies “biuniqueness”, which “requires not only that a given phoneme is represented by a constant symbol but also that the symbol involved does not represent other phonemes” (Carney 1994: 15; see also Lass and Laing 2007: Section 2.2.1). However, total observance of such one-to-one correspondence between sound and symbol is not recorded among the writing systems of natural languages. Scripts classified as the closest to this principle are classical Greek and Finnish (Rogers 2005: 274).

227

228

II Linguistic Levels The problem with the classification of writing systems is that they are never purely morphographic or phonographic. In fact, the longer a given script is used by a speech community or a nation as an everyday means of communication, the more it evolves, and the more mixed characteristics it acquires (Sampson 1985: 42; Rogers 2005: 272). Examples of taxonomical mixtures “regularly using both morphographic and phonographic symbols” are Egyptian, Maya, and the mixed use of morphographic kanji and two types of phonographic kana in Japanese (Rogers 2005: 272). Even such systems as Sumerian and Chinese, where the morphographic principle is clearly the dominant one, show a limited level of phonography (Sampson 1985: 54; Rogers 2005: 274–275). Writing systems differ not only by their amount of morphography, but also by their “orthographic depth”. The more different morphemes are distinguished in spelling, the higher the level of morphography of a given system (Rogers 2005: 275). For example, English is classified as a phonemic system, but it shows the distinctions among different homophonous morphemes, e.g., you – yew – U – ewe, right – rite – wright – write (Bradley 1904: 214; Rogers 2005: 273; see also Craigie 1928: 1, Bolinger 1946: 335; Vachek 1973: 43–44 for discussions on the distinction between homophones). That property of English spelling has also been referred to as the “lexical principle” (Chomsky and Halle 1968: 49). On the other hand, “[o]rthographic depth is greater if different allomorphs of the same morpheme are written the same” (Rogers 2005: 275). Accordingly, a final in English can be related directly to {past tense}, because it does not show systematic differentiation between its allomorphs /ɪd/ (pointed), /d/ (played) and /t/ (washed). Constant spelling for a morpheme, irrespective of its spoken variants, can also be found in heterophonous, but etymologically and semantically related words, such as child-children, south-southern, and sign-signature (Francis 1958: 562; Venezky 1970: 42–43, 108; Vachek 1973: 25). Such a preservation of morpheme constancy may be considered consistent with the “morphophonemic principle” (Hall 1981 [1960]: 74). Morphemic and morphophonemic spellings breach the “alphabetic principle”, because they lead to a multiplication of sound-symbol (or grapho-phonemic) correspondences, and by the same measure contribute to the opacity, or orthographic depth, of a phonographic system. On the basis of the pairs of etymologically related words provided above, may correspond to /aɪ/ and /ɪ/, to /aʊ/ and /ʌ/, and to /ɡ/ or zero, and these sets are by no means exhaustive lists of all the permissible correspondences. Thus, English orthography can be called “deep”, in contrast to, e.g., the Spanish orthographic system, which shows less divergence from the alphabetic principle, and consequently can be referred to as “shallow”. In fact, “the deep/ shallow contrast is a gradient rather than all-or-none distinction” (Sampson 1985: 45), and morphemic and morphophonemic spelling is not an exceptionless rule in English. Rather, English orthography “represents a level intermediate in depth between the phonemic and the morphophonemic level” (Sampson 1985: 44; see also Hockett 1958: 542). For example, the vowel in the plural suffix (as in pitches) is represented by , but the difference in voicing between /s/ and /z/ (as in cats vs. dogs) is not represented at all. A maximally deep orthography would then have *pitchs instead of pitches. However, the sequence is not permitted in English, and would violate the “graphotactic principle”, determining the permissible letter sequences (Carney 1994: 67). The English writing system also contains a few marginal, non-phonemic, and nonmorphemic, but rather iconic (or pictographic) elements, including, e.g., the symbols ,

15 Linguistic Levels: Orthography and , which represent the words and, at, and the phrase per cent, respectively (Sampson 1985: 34; Lass and Laing 2007: Section 2.2.1). It is noteworthy that the identification of different principles governing the English writing system, and orthographic systems in general, is not a recent discovery. Already in the late 19th and early 20th centuries Jan Baudouin de Courtenay elaborated a set of rules governing orthographic systems. They included three principles: the phonetic, etymological, and historical. Thus, he related orthography to what he believed were its three determining factors, pronunciation, origin (referring to morphology), and tradition (Ruszkiewicz 1981: 24–25; Sgall 1987: 2–3). The idea of multi-level principles and correspondences according to which orthography operates has been further elaborated upon and modified by numerous linguists. For example, Firth (1935: 61) wrote about the notion of “polysystem”, or system of systems, referring to the co-existence of and interaction between phonological, grammatical, and lexical systems of orthographic representation. Overviews of the investigations on the interaction of principles in English orthography and in other orthographic systems can be found, e.g., in Ruszkiewicz (1976), Sgall (1987), and Liuzza (1996). They are summarized by Sgall thus: “In the literature one often speaks about an orthography being based on several principles, the main among which is the phonemic one, while the others underlie the deviations from this basic principle and can be classed more or less exactly in accordance with the levels of the language system” (Sgall 1987: 12).

4 Units of writing systems: terminological evolution In the course of time and investigation into the nature of relationships between written characters and other levels of language, not all linguists have been satisfied with the simple, traditional terms letter and sound, and consequently, there has gradually developed a whole set of terms and definitions in order to make descriptions of the writing system more precise. The terminology used by particular linguists has been developed in close connection with their specific approaches to orthography and their perception of its structure and functions. The following sections discuss some of the most important terms and the ideas behind them.

4.1 The doctrine of littera The most characteristic feature of the early attempts at defining the unit of the writing system is the apparent lack of a clear distinction between letters and sounds. The term letter was used to refer both to written alphabetic marks and to sounds. One can consider, for example, the statement “The Elements of Language are Letters, viz. Simple discriminations of Breath or Voice” (Holder 1669; quoted by Abercrombie 1981 [1949]: 10). What at first sight may seem to be the result of utter confusion, can, in fact, show the impact of the classical notion of littera, introduced by the Stoic grammarians, and described by Aelius Donatus in his Latin grammar in the 4th century CE. Littera referred to the smallest element of language, combining three attributes, nomen (name used for identification), figura (shape or visual configuration), and potestas (power to signify sound). (Abercrombie 1981 [1949]: 14; Henderson 1985: 142). Thus, the visual mark and the sound could be viewed as different aspects of the same entity.

229

230

II Linguistic Levels In spite of its significant and long-lasting influence on numerous generations of grammarians, the concept of littera, in the course of time, proved insufficient for the description of both spoken and written systems of language, at least partly due to its potential for misinterpretation. Quotations such as: “Letters are Signes of Sounds, not the Sounds themselves” (Brightland 1711; quoted by Abercrombie 1981 [1949]: 11) demonstrate that the complexity of the concept of littera was not transparent and could lead to ambiguity. However, after years of abandonment in the 20th century, the notion has recently been revived by linguists and it has inspired the development of such concepts as “litteral substitution sets” (LSS), which have proved useful in the description of early Middle English writing systems (see, e.g., Benskin 1991: 226; Laing 1999; Laing and Lass 2003, 2009; Lass and Laing 2007: Sections 3.3.1 and 2.3.2). The word letter is still in use both in everyday language with reference to a character of an alphabet, or in more technical descriptions, usually as part of a definition of the distinctive unit of the orthographic system.

4.2 Grapheme Since the introduction of the term grapheme by Baudouin de Courtenay in 1901 (Ruszkiewicz 1976: 24–37, 1981 [1978]: 20–34), it has been defined in various ways: – “the class of graphs which denote the same phoneme” (Hammarstro¨m 1981 [1964]: 97); – “the class of letters and other visual symbols that represent a phoneme or cluster of phonemes” (OED, Simpson [ed.] 2000–); – “any minimal letter string used in correspondences” (Carney 1994: xxvii); – “[t]he unit of writing” (Stetson 1981 [1937]: 35); – “the minimal functional distinctive unit of any writing system” (Henderson 1984: 15); and – “a purely distinctive visual unit, part of an autonomous semiotic system” (Liuzza 1996: 28). As can be seen from these quotations, the available definitions can be divided into two groups, corresponding to two main senses, and reflecting “conflicting linguistic views of the status of writing” (Henderson 1985: 142): 1. a letter or cluster of letters referring to or corresponding with a single phoneme; 2. the minimal distinctive unit of a writing system. The former sense, evident in the first three quotations above, assumes that orthography is mainly, if not exclusively, a means of notation for representing speech, and is usually adopted by the proponents of the relational approach to orthography. The latter sense, in turn, is more typical of the adherents to the autonomistic approach to writing (see Section 1 above). It emphasizes the contrastiveness of the grapheme within the orthographic system as one of its necessary features. Moreover, definitions associated with this sense also occasionally point to the analogy which can be drawn between the status of graphemes and that of phonemes within their respective systems: “the graphemes of a given language – like its phonemes – remain differentiated from one another, i.e. […]

15 Linguistic Levels: Orthography they do not get mixed up” (Vachek 1976 [1945–49]: 128–129). Pulgram (1951: 15) even proposed a comprehensive list of correspondences between the graphemes and phonemes of a language. He was also one of the first linguists to use the term allograph, by analogy to allophone. According to him, “all graphs identifiable as members of one grapheme are its allographs”. This definition was later developed by McLaughlin: “an allograph or allographic set which contrasts significantly with all other allographs or allographic sets or with zero will be called a GRAPHEME”, where an allograph is “a group of similar characters, modifications, or features [i.e. graphs] classed together […] in graphemic analysis” (McLaughlin 1963: 29). Within the graphemic system, graphs can be defined as “each hic et nunc realization of a grapheme” (Pulgram 1951: 15). It was cogently argued by Henderson (1985) that the autonomistic definition of the grapheme, in comparison to the relational one, “lends itself to the most coherent and principled use. It also allows the term to be applied across the range of writing systems, including irregular alphabetic systems, syllabaries and logographic systems” (Henderson 1985: 146).

5 The inventory and distribution of English graphemes from the synchronic and diachronic perspective The inventory of graphemes used in modern English comprises “alphabetic” and “nonalphabetic” graphs. The former refer to the twenty-six letters of the alphabet . Each letter has also a capitalized counterpart, e.g., , so altogether we can distinguish 52 alphabetic characters. Opinions about the graphemic significance of capital letters are divided. According to Haas (1970: 22–23), since the lower-case and upper-case alphabetic graphs occur in complementary distribution, they can be described as allographs of a particular grapheme, e.g., or . Taking this point of view, we could distinguish 26 (not 52) graphemes in English. However, Henderson points to the fact that in sentences such as “The archer was called Archer” archer and Archer are different lexical identities and “so the contrast expressed by capitalization has graphemic status” (Henderson 1985: 144). The non-alphabetic graphic symbols appear in combination with the alphabetic ones, and are subdivided into four types (McLaughlin 1963: 30): 1. “punctuation marks”, usually helping to indicate grammatical structure, including, e.g., < , . : ; ’ ? ! - – … ‘ ’ “ ” ( ) < > [ ] { } >; 2. “graphic components” referring to diacritics, conjoined to other graphs and used to distinguish different values of the same letter, including accent marks, dots, cedillas, subscript hooks, macrons, e.g. (these appear in English only occasionally, in borrowings); 3. “tachygraphs” (found mainly in medieval manuscripts), used only with alphabetic symbols, and standing for one or more of them, e.g., a bar through the descender of the letter

, functioning as an abbreviation for or in, e.g., person; and 4. “word signs”, e.g. (an ampersand), standing for the word and, or the symbol , which stands for at. Moreover, Francis (1958: 436) suggests including a space in the writing system, as a sort of zero grapheme (see also Venezky 1970: 47 and Carney 1994: 5, cf. Sgall 1987: 7 who uses the expression “blank”).

231

232

II Linguistic Levels Viewed from a diachronic perspective, the inventory of English graphemes should also include several additional symbols, particularly thorn , eth , wynn , ash , and yogh . Thorn and wynn were “borrowed” from the runic alphabet (futhorc) in medieval times. Ash was a ligature derived from the Roman alphabet, eth was an Irish modification of to represent, along with thorn, the dental fricatives, and yogh originated as the insular open g (ᵹ), borrowed from the Irish. All these characters had been replaced by other symbols by modern times. On the other hand, the differentiation between and , as well as between and as separate letters of the alphabet is a comparatively new invention in the English inventory of graphemes. (OED, Simpson [ed.] 2000–; Scragg 1974: 2, 8, 10 passim). Over the centuries, changes affected not only the composition of the inventory of symbols used for writing in English. Graphotactics and grapho-phonemic correspondences have also evolved over time, making English orthography more and more inconsistent with the alphabetical principle governing phonemic writing systems. For example, the Old English corresponded to /ʧ/, and /k/, but never to /ʃ/, and corresponded to /ʃ/, but never to /sk/. Likewise, punctuation, capitalization, and word division patterns have been modified over the centuries. The reasons for the changes have been multiple and diverse, but a few important events and influences are particularly noteworthy in the development of English graphemics. For example, the arrival of Christianity at the end of the 6th century CE brought the adoption of the Latin alphabet (with some modifications), and the Norman Conquest resulted in the appearance of numerous French-derived spelling conventions, which partly replaced the previous practices. For example, the Norman scribes introduced the doubling of letters to indicate long vowels in words such as goose and meet, and also replaced , , , and with , , , and , respectively. The introduction of printing contributed to the fixing and standardization of English spelling, or at least to the dissemination of standardized practices; and extensive borrowing from various languages explains the diversity of spelling patterns. Readers interested in more detail on the structure and the rules governing English orthography, can consult a number of books outlining the synchronic and diachronic aspects of English orthography. The most recent synchronic summaries include Rollings (2004), Carney (1994), Venezky (1970, 1999), and Wełna (1982). However, the only relatively reliable comprehensive study devoted to the history of English spelling published so far is Scragg (1974).

6 Orthography as the source of phonological evidence In view of the depth of English orthography, due to the variety and complexity of principles governing it (discussed in the previous sections), a linguist studying the phonology of an older text cannot take orthographic evidence at face value as the source of information about pronunciation, while one studying the morphology or syntax will find relevant evidence much easier to obtain from a written text. Penzl (1957: 197–200) provides a classification of the types of evidence available for students of previous pronunciation, dividing it into orthographic, orthoepic (referring to the comments of phoneticians, grammarians, and spelling reformers), metrical, comparative, and contact evidence. The evaluation of various sources of information about the

15 Linguistic Levels: Orthography phonemic change can be found, apart from Penzl (1957), also in Ko¨keritz (1953), Dobson (1957–68), Wrenn (1967), Wolfe (1972), Liuzza (1996), Stenroos (2002, 2004, 2006), and in earlier works, e.g., Zachrisson (1913) and Wyld (1936). Orthography is thus one of several sources of evidence for historical phonologists, and should at best be analyzed in conjunction with the information gathered from the examination of other types of evidence. Orthographic evidence comprises occasional spellings (sometimes referred to as naive spellings), e.g. douter for daughter, or ruff for rough, and back spellings (also called reverse spellings or hypercorrections), e.g. quight for quite (by analogy to light, night). Unfortunately, the interpretation of such orthographic evidence is fraught with problems: 1. not every change in spelling means a change in pronunciation; e.g. the etymologizing insertion of in doubt did not result in assigning any sound value to that letter; 2. not all changes in pronunciation are reflected in spelling; e.g. the diphthongization of Middle English [iː] and [uː], as in mice and house, is not; 3. occasional spellings may be due to (typo)graphical errors, especially if one deals with a single instance of a given form; and 4. the value of the same spellings may vary according to the period and regional dialect. Nevertheless, a linguist can minimize the risks connected with such types of evidence by gathering a large number of examples before drawing conclusions, by comparing orthographic evidence with other types of evidence, and, most importantly, by examining thoroughly the orthographic rules of the relevant writing system with due attention paid to the period and regional variety, before starting the analysis (see, e.g., Penzl 1957: 197–203; Wolfe 1972: 110–130; Stenroos 2002: 445–468; Smith 2006: 136). Liuzza (1996: 33) emphasizes particularly the last measure mentioned above: “Only by reconstructing the orthographic rules practiced by a writer can one determine anything about the spoken language behind the text; the evidence, otherwise, is mute”.

7 Sociolinguistic aspects of orthography As has been indicated in the previous sections, orthography can be treated as one of the subjects of linguistic research. Nevertheless, in the estimation of an average language user, orthography functions nowadays mainly as a spelling norm. Such a norm is binding within a language community and non-compliance with it can have social consequences. It applies particularly to public domains, such as education and official written media. Someone using non-standard forms and making spelling mistakes in these contexts is likely to be socially punished by being judged uneducated or even unintelligent. The notion of spelling error is a relatively recent development. In the Middle Ages, there was no orthographic standard in English in the modern sense. Instead, dialectal variation was abundant. In the 15th century the situation started to evolve towards standardization. This process took more than two centuries to complete, including the stages of selection, acceptance, functional elaboration, and codification (Haugen 1972: 110). The process of codifying those orthographic forms which were viewed as proper was prompted and advocated by the spelling reformers, grammarians, and lexicographers. Their proposals were essential for the regularization of spelling in English (Brengelman 1980). Spelling books, grammars, and dictionaries, which were printed and

233

234

II Linguistic Levels disseminated by a growing number of publishing houses, contributed to the elimination of most orthographic variants and to the considerable increase in orthographic consistency. This increase was so substantial that it is usually recognized that by the end of the 17th century English orthography had become standardized (Scragg 1974: 80, Go¨rlach 2001: 78). However, it has recently been emphasized that the completion of standardization can be applied mainly to printed documents, whereas “[d]eviant spelling continued in letters and diaries, even among the educated” (Go¨rlach 2001: 78; see also Salmon 1999: 44; Tieken-Boon van Ostade 2009: 46–50). Furthermore, also in the 19th-century English, “while public printed texts manifest greater stability, even these are not devoid of change” (Mugglestone 2006: 278). In present times, with well-established orthographic standards, orthographic variation still exists. In English, there is the codified and politically sanctioned diatopic division into two spelling systems, British and American, including the respective use of such spelling variants as those in honour/honor, centre/center, monologue/monolog, and aeroplane/ airplane. The spelling can differ also depending on the particular reference of the word in a given context. For example, in British English, the form program is reserved for computingrelated contexts, and programme is used in the other meanings, whereas American English uses program in all senses of that lexical item (OED, Simpson [ed.] 2000–). In today’s languages, the use of non-standard orthographic variants can also be deliberate (Beaugrande 2006: 43), or even constitute “a powerful expressive resource” capturing some of the “immediacy, the ‘authenticity’ and ‘flavor’ of the spoken word in all its diversity” (Jaffe 2000: 498). One can experience the effectiveness of non-standard orthography in daily contexts, for example in trade-names, television commercials and in computer-mediated communication (Crystal 2001) (cf. Heyd, Chapter 70). It can also be found in printed texts, for instance in poetry (e.g. Tom Leonard), or popular fiction (e.g. Irvine Welsh), where it is used for stylistic as well as ideological reasons.

8 Summary Orthography differs in many ways from the other levels of linguistic description. It does not enjoy a stable and generally accepted terminology, partly because in numerous natural languages, such as in English, it is governed by a set of heterogeneous principles, and partly due to the fact that linguists differ significantly as regards the aims and functions which they ascribe to orthographic systems. The existence of diverse definitions of the term orthography itself reflects the complexity typical of writing systems. This term is employed to denote, on one hand, a system of graphemes as contrastive units, and, on the other, a set of rules governing the correspondences between the graphemes, phonemes, and morphemes. It also functions as a social code which needs to be observed under the pain of stigmatization. In the outline offered above, I have attempted to present some of the important aspects of the controversy concerning the relationship between speech and writing, discuss the different levels of linguistic representation in orthography, and describe the main principles according to which the English writing system operates. The characteristics of this system are adequately expressed by Venezky (1999: 4): “English orthography is not a failed phonetic transcription system, invented out of madness or perversity. Instead, it is a more complex system that preserves bits of history (i.e. etymology), facilitates understanding, and also translates into sound”.

15 Linguistic Levels: Orthography

9 References Abercrombie, David. 1981 [1949] What is in a “letter”? In: Ruszkiewicz (ed.), 9–19 [Reprinted from Litera 3: 43–46]. Bazell, Charles E. 1981 [1956] The grapheme. In: Ruszkiewicz (ed.), 66–70 [Reprinted from Lingua 2: 54–63]. de Beaugrande, Robert. 2006. Speech versus writing in the discourse of linguistics. Miscela´nea: A Journal of English and American Studies 33: 31–45. Benskin, Michael. 1991. In reply to Dr Burton. Leeds Studies in English: New Series 22: 209–262. Bloomfield, Leonard. 1933. Language. New York: Holt, Rinehart, and Winston. Bolinger, Dwight. 1946. Visual morphemes. Language 22: 333–340. Bradley, Henry. 1904. The Making of English. London: Macmillan. Bradley, Henry. 1919 [1913] On the Relations between Spoken and Written Language, with Special Reference to English. Oxford: Clarendon Press. [Reprinted] Brengelman, Fred H. 1980. Orthoepists, printers, and the rationalization of English spelling. Journal of English and Germanic Philology 79: 332–354. Carney, Edward. 1994. A Survey of English Spelling. London/New York: Routledge. Chomsky, Noam and Morris Halle. 1968. The Sound Pattern of English. New York: Harper and Row. Cook, Vivian. 2004. The English Writing System. London: Hodder Arnold. Craigie, William A. 1928. English Spelling: Its Rules and Reasons. London: George G. Harrap. Crystal, David. 1995. The Cambridge Encyclopedia of the English Language. Cambridge: Cambridge University Press. Crystal, David. 2001. Language and the Internet. Cambridge: Cambridge University Press. Dobson, E. J. 1957–68 . English Pronunciation: 1500–1700. Vols. I and II. Oxford: Clarendon Press. Firth, John R. 1935. The technique of semantics.Transactions of the Philological Society, 36–72. Francis, W. Nelson. 1958. The Structure of American English. New York: The Ronald Press Company. Go¨rlach, Manfred. 2001. Eighteenth-century English. Heidelberg: C. Winter. Haas, William. 1970. Phono-graphic Translation. Manchester: Manchester University Press. Hall, Robert A. 1981 [1960] A theory of graphemics. In: Ruszkiewicz (ed.), 71–80 [Reprinted from Acta Linguistica 8: 13–20]. Halliday, Michael A. K., Angus McIntosh, and Peter Strevens. 1964. The Linguistic Sciences and Language Teaching. Bloomington: Indiana University Press. Hammarstro¨m, Go¨ran. 1981 [1964] Type and typeme, graph and grapheme. In: Ruszkiewicz (ed.), 89–99 [Reprinted from Studia Neophilologica 36: 332–340]. Haugen, Einar. 1972. Dialect, language and nation. In: John Pride and Janet Holmes (eds.), Sociolinguistics, 11–24. Harmondsworth: Penguin Modern Linguistics Readings. Henderson, Leslie. 1984. Writing systems and reading processes. In: Leslie Henderson (ed.), Orthographies and Reading: Perspectives from Cognitive Psychology, Neuropsychology, and Linguistics, 11–24. London: Lawrence Erlbaum. Henderson, Leslie. 1985. On the use of the term “grapheme”. Language and Cognitive Processes 1(2): 135–148. Hockett, Charles F. 1958. A Course in Modern Linguistics. New York: The Macmillan Company. Jaffe, Alexandra. 2000. Introduction: Non-standard orthography and non-standard speech. Journal of Sociolinguistics 4(4): 497–513. Kniezsa, Veronika. 1991. “The due order and reason”: On the histories of English spelling. Folia Linguistica Historica 12(1–2): 209–218. Ko¨keritz, Helge. 1953. Shakespeare’s Pronunciation. New Haven: Yale University Press. Laing, Margaret. 1999. Confusion wrs confounded: Litteral substitution sets in early Middle English writing systems. Neuphilologische Mitteilungen 100: 251–270. Laing, Margaret and Roger Lass. 2003. Tales of the 1001 nists: The phonological implications of litteral substitution sets in some thirteenth-century South-West Midland texts. English Language and Linguistics 7: 257–278.

235

236

II Linguistic Levels Lass, Roger and Margaret Laing. 2007. Introduction. Part I: Background. Chapter 2: Interpreting Middle English. In: Margaret Laing and Roger Lass, LAEME: A Linguistic Atlas of Early Middle English, 1150–1325. University of Edinburgh. http://www.lel.ed.ac.uk/ihd/laeme1/laeme1. html (date of access: 18 November 2009) Laing, Margaret and Roger Lass. 2009. Shape-shifting, sound-change and the genesis of prodigal writing systems. English Language and Linguistics 13(1): 1–31. Liuzza, Roy M. 1996. Orthography and historical linguistics. Journal of English Linguistics 24(1): 25–44. McIntosh, Angus. 1961. “Graphology” and meaning. Archivum Linguisticum 13: 107–120. McLaughlin, John C. 1963. A Graphemic-phonemic Study of a Middle English Manuscript. The Hague: Mouton. Mugglestone, Lynda. 2006. English in the nineteenth century. In: Mugglestone (ed.), 274–304. Mugglestone, Lynda (ed.). 2006. The Oxford History of English. Oxford: Oxford University Press. Penzl, Herbert. 1957. The evidence for phonemic changes. In: Ernst Pulgram (ed.), Studies Presented to Joshua Whatmough on his Sixtieth Birthday, 193–208. The Hague: Mouton. Pulgram, Ernst. 1951. Phoneme and grapheme: A parallel. Word 7: 15–20. Rogers, Henry. 2005. Writing Systems: A Linguistic Approach. Malden: Blackwell. Rollings, Andrew G. 2004. The Spelling Patterns of English. Mu¨nchen: Lincom. Ruszkiewicz, Piotr. 1976. Modern Approaches to Graphophonemic Investigations in English. Katowice: Uniwersytet S´la˛ski. Ruszkiewicz, Piotr. 1981 [1978] Jan Baudouin de Courtenay’s theory of the grapheme. In: Ruszkiewicz (ed.), 20–34 [Reprinted from Acta Philologica 7: 117–135]. Ruszkiewicz, Piotr (ed.). 1981. Graphophonemics: A Book of Readings Katowice: Uniwersytet S´la˛ski. Salmon, Vivian. 1999. Orthography and punctuation. In: Roger Lass (ed.), The Cambridge History of the English Language. Vol. III. 1476–1776, 13–55. Cambridge: Cambridge University Press. Sampson, Geoffrey. 1985. Writing Systems: A Linguistic Introduction. Stanford: Stanford University Press. Sapir, Edward. 1921. Language: An Introduction to the Study of Speech. New York: Harcourt, Brace & World. de Saussure, Ferdinand. 1993. Troisie`me Cours de Linguistique Generale (1910–1911): d’apre`s les cahiers d’Emile Constantin./Third Course of Lectures on General Linguistics (1910–1911): From the Notebooks of Emile Constantin. Eisuke Komatsu (ed.) and Roy Harris (trans.). Oxford: Pergamon Press. Scragg, Donald G. 1974. A History of English Spelling. New York: Manchester University Press and Barnes & Noble. Sgall, Petr. 1987. Towards a theory of phonemic orthography. In: Philip A. Luelsdorff (ed.), Orthography and Phonology, 1–30. Amsterdam/Philadelphia: John Benjamins. Simpson, John (ed.). 2000–. The Oxford English Dictionary. 3rd edn. online. Oxford University Press. www.oed.com/ Smith, Jeremy J. 2006. From Middle to Early Modern English. In: Mugglestone (ed.), 120–146. Stenroos, Merja. 2002. Free variation and other myths: Interpreting historical English spelling. Studia Anglica Posnaniensia 38: 445–468. Stenroos, Merja. 2004. Regional dialects and spelling conventions in Late Middle English: Searches for (th) in the LALME data. In: Marina Dossena and Roger Lass (eds.), Methods and Data in English Historical Dialectology, 257–285. Bern: Peter Lang. Stenroos, Merja. 2006. A Middle English mess of fricative spellings: Reflections on thorn, yogh and their rivals. In: Marcin Krygier and Liliana Sikorska (eds.), To Make his Englissh Sweete upon his Tonge, 9–35. Frankfurt a. Main: Peter Lang. Stetson, Raymond H. 1981 [1937] The phoneme and the grapheme. In: Ruszkiewicz (ed.), 35– 44 [Reprinted from Me´langes de linguistique et de philologie offerts a Jacq. van Ginneken, 353–356].

16 Linguistic Levels: Styles, registers, genres, text types

237

Stockwell, Robert P. and C. Westbrook Barritt. 1951. Some Old English Graphemic-Phonemic Correspondences – ae, ea and a. (Studies in Linguistics, Occasional Papers, 4.) Norman, OK: Battenburg Press. Tieken-Boon van Ostade, Ingrid. 2009. An Introduction to Late Modern English. Edinburgh: Edinburgh University Press. Vachek, Josef. 1973. Written Language: General Problems and Problems of English. The Hague: Mouton. Vachek, Josef. 1976 [1945–49] Some remarks on writing and phonetic transcription. In: Vachek (ed.), 127–133 [Reprinted from Acta Linguistica 5: 86–93]. Vachek, Josef (ed.). 1976. Selected Writings in English and General Linguistics. The Hague: Mouton. Vachek, Josef. 1982. English orthography: A functional approach. In: William Haas (ed.), Standard Languages: Spoken and Written, 37–56. Manchester: Manchester University Press. Venezky, Richard L. 1970. The Structure of English Orthography. The Hague: Mouton. Venezky, Richard L. 1999. The American Way of Spelling. New York: The Guildford Press. Wallis, John. 1670. A Letter of Doctor John Wallis to Robert Boyle Esq. concerning the said Doctors Essay of Teaching a person Dumb and Deaf to speak, and to understand Language, together with the success thereof, made apparent to his Majesty, the Royal Society, and the University of Oxford. Philosophical Transactions 61: 1087–1099. Wełna, Jerzy. 1982. English Spelling and Pronunciation. Warszawa: Pan´stwowe Wydawnictwo Naukowe. Wolfe, Patricia. 1972. Linguistic Change and the Great Vowel Shift in English. Berkeley: University of California Press. Wrenn, Charles L. 1967. The value of spelling as evidence. In: Charles L. Wrenn (ed.), Word and Symbol: Studies in English Language, 129–149. London: Longmans. Wyld, Henry. 1936. A History of Modern Colloquial English. Oxford: Basil Blackwell. Zachrisson, Robert. 1913. Pronunciation of English Vowels 1400–1700. Go¨teborg: W. Zachrissons boktryckeri.

Hanna Rutkowska, Poznan´ (Poland)

16 Linguistic Levels: Styles, registers, genres, text types 1 2 3 4 5 6

Introduction A selective history of registers Stylistic developments in English Approaches to historical texts Summary References

Abstract The chapter surveys research on registers, styles, text types and genres in the history of English. The presence of registers is connected with socio-cultural conditions, such as the structure of society, multilingualism, the practice of translation, academic traditions, and technological progress. The legal and the scientific registers are described in more detail. The stylistic development of English is treated here as linked to standardization Alex Bergs and Laurel Brinton 2012, Historical Linguistics of English (HSK 34.1), de Gruyter, 237–253

238

II Linguistic Levels (elaboration in syntax and lexicon) and the orality-literacy continuum (increasing development of more literate characteristics). The curial style of the 14th/15th century and the plain style of the 17th century are highlighted. Three approaches to genre and to text types are presented. Inventories of genre labels highlight the presence, nature and development of genres throughout English history. Changes in texts, and genres across periods can also be studied through looking at their conventional structure (e.g. letters) or their linguistic features (e.g. discourse deixis). Lastly, text types/genres can also play a role in language change.

1 Introduction Looking through books on the history of English, one will as a rule not find separate sections on texts and registers (with the exception of Go¨rlach 1999a, 2001a). Looking at language in use, in contrast, it is obvious that it “exists in texts” (Diller 2001: 3). Individual linguistic features are realized in texts and for the sake of creating coherent and effective texts, so that language history can not only be investigated by studying texts but is actually only constituted by texts and changing textual needs. Any linguistic history that is textless, therefore, ignores a very important perspective on the development of English. The inter-relationship between the development of linguistic features and their textual uses has only relatively recently received more attention on a firmer theoretical and methodological basis than before. The collection edited by Diller and Go¨rlach (2001) and the special issue of the European Journal of English Studies (ed. Moessner 2001) dealing with genre and text types in a historical perspective bear witness to this. Unfortunately, the definitions of the terms in the title are quite varied in the literature (cf. Diller 2001 for a discussion). This article is based on the following understanding of the terminology. While registers and styles are inventories of linguistic devices, genres and text types are classes of texts. Genre is primarily based on text-external considerations (cf. Biber 1989: 5–6) and refers to aspects such as the functions, conventional shape, and structure of texts. Genres are linked to expectations on the part of text users about the (proto-)typical functions and (surface) features of texts belonging to the genre; thus competent speakers have fairly clear ideas about what a fairy tale, a letter, a prayer, or a weather forecast is like. As folk categories, genres are not necessarily defined by a strict and homogeneous set of criteria. Text types, on the other hand, are defined by text-internal linguistic criteria, which to a certain extent go along but do not completely overlap with genre distinctions (cf. Taavitsainen 2002: 220). Texttypological approaches (e.g. Longacre 1996; Werlich 1983) have presented broad categories such as narrative, descriptive, expository, instructive, or procedural/behavioral, and argumentative, which are characterized by a typical (co-)occurrence of linguistic features, e.g. narrative: past tense verbs, time adverbials; procedural: imperatives. Such internal text-type features can be present in a given text to varying extents, thus making it more or less expository/narrative/etc. in nature, and features from different types can combine, producing, for example, an expository-argumentative hybrid. Register is a more general term, comprising both oral and written productions based in particular on situational, social, and professional contexts and the field or domain of discourse (cf. also Lenker, Chapter 21). Thus the domains of religion, law, science, journalism, etc., constitute the religious, legal etc. registers, all of which exhibit a certain cohesion in terms of possible interaction types, aims, and contents, producing

16 Linguistic Levels: Styles, registers, genres, text types lexico-grammatical similarities on a more general level than text types. Registers usually comprise various genres. Style is the vaguest of all these terms and potentially cuts across all the other distinctions. Diller (1998: 155–156) defines style as used by textual stylistics as the idiosyncractic “characteristic linguistic features of a text”, and that of linguistic stylistics as “different ways of saying the same thing”. From the point of view of the language user, style implies aspects like choice between linguistic items, perceptions of appropriateness going beyond register conventions, ideas of norm vs. deviance, and (potentially prescriptive) aesthetic notions. A given style can go with an individual, a group of people, or a time period. A note on the application of some of these concepts in a commonly used source in historical linguistics, namely the Helsinki Corpus (HC) (Rissanen et al. 1991), may be in order here. The corpus encodes what it calls “text type” () and “prototypical text category” (). The latter refers to (broad) text types as explained above, containing such labels as “expository”, “instruction religious”, and “narration imaginative”, to which it adds register information (religious) or the more general distinction of (non-)fictionality in some cases. Helsinki Corpus text types, in contrast, in some cases are rather genres (e.g. handbook, sermon, preface, comedy) and in others are reminiscent of registers (e.g. science medicine, history). The emphasis of this article will be on non-literary texts (cf. Fulk, Chapter 25, Arnovick, Chapter 35, and Moore, Chapter 50). There will be some overlap with the notion of discourse as treated in Lenker, Chapter 21.

2 A selective history of registers Some registers are attested throughout the history of English, albeit perhaps in varying strength and internal variety. The religious register is one of those with a long history, for example, and also with a certain breadth of representation. Apart from the Bible (cf. Kohnen, Chapter 65), the register is represented by – both original and translated – sermons and homilies, texts related to the liturgy like prayers, the Creed, the psalter or hymns, lives of saints, hortative and instructional writing, and academic theological/exegetical writing. Except for the last type, all are attested from Old English onwards. Other registers, in contrast, have only emerged at some later point in history (e.g. newspaper language, cf. Fries, Chapter 67). Late evolution may have to do with extralinguistic developments, such as the possibilities offered by the printing press or the rise of modern natural science, but also with the different status of English vis-a`-vis other languages in different periods. As to the latter aspect, in domains like religion, the law, and scientific/academic writing, Latin in particular, and French played important roles during the Middle Ages and also beyond, thus competing with, sometimes dominating, English in the respective registers (cf. Go¨rlach 1999b: 462). This will also be visible in the short outline of the registers of law and science which follows.

2.1 Legal English Like the religious register, legal (and administrative) English is a long-standing variety with a remarkable functional stability and a very distinctive form of English. Hiltunen (1990) provides a historical survey of legal English, but without covering all periods equally (Early Modern English is particularly neglected). The oldest extant legal

239

240

II Linguistic Levels texts are the laws of King Ethelbert of Kent (635 CE), which are followed by various other law codes up to the reign of Cnut in the 11th century (cf. Liebermann 1903–16). Old English legal language already shows the complex structures present in the modern variety, even if to a lesser extent, such as conditional, relative and adverbial subordination, and multiple embedding. But it also lacks some modern characteristics, such as the emphasis on unambiguous reference and precision, and thus also its repetitiveness. With the Norman Conquest, however, English ceased to exist as a language of the law for about four centuries, until the 1362 Statute of Pleading re-established English as the oral legal language, and the first Act of Parliament to be written in English was passed in 1483. During the Middle English (ME) period, legal writing had used first “Law Latin”, later French, while pleading had taken place in French, thus adopting not only foreign legal procedures, but also the linguistic patterns of legal Latin and French. Similar linguistic mixtures are also found in administrative or business records, whose macaronic writing style (mixing Latin, French, and English) has been termed a “deliberate, formal register” by Wright (1992: 769). The full establishment of English in all spheres of law was gradually carried through during the Early Modern English period, involving again translation of important texts into English. From about the 16th century there is no shortage of legal texts in English. Through trial transcripts and proceedings we also have an insight into the oral forms of historical legal discourse; studies on Early Modern English courtroom language have been carried out by Archer (2005) on British data and by, e.g., Grund (2007) on the Salem Witchcraft Trials. As a consequence of its history, Hiltunen (1990: 52) has characterized modern legal language as “essentially a kind of ‘creole’, where the formative elements go back to an amalgamation of native resources and extensive borrowing”. One of the noteworthy characteristics of legal English is its lexicon reflecting the influence of the various legal traditions throughout its history, from Anglo-Saxon terminology (which has mostly disappeared, often together with the concepts, e.g. wergeld), via Norse terms, to French and Latin words (Mellinkoff 1963). Modern law language is further characterized by archaic lexical usages, such as aforesaid, theretofore and similar elements, which are fossils from the Early Modern English period. Throughout its English history, the register seems to have followed a trend towards ever more specificity and explicitness, thus increasing its linguistic complexity. Partly, this complexity has been made more accessible by structure-building visual arrangements, which had not been used in the past; this last point shows that the visual aspect of texts should also play a role in textual studies.

2.2 Scientific language The “vernacularisation boom of the fourteenth century” (Taavitsainen 2001: 189) was certainly important for law, and so it was for another register, i.e. scientific writing. In contrast to law, science is a younger discipline and thus also a newer register. Anglo-Saxon science and an Old English scientific register in a proper sense do not exist; what is extant is rather texts of a practical nature, such as astrological texts, herbals, and medical recipes (e.g. Bald’s Leechbook). In general the scientific or academic community of the Middle Ages, which was an international one, used Latin, not only as a written but also as a spoken language, e.g. within the universities. This custom was

16 Linguistic Levels: Styles, registers, genres, text types slow to die, extending into the 17th and 18th centuries in England, when Latin works were still being written, although to an ever-decreasing extent. From an extralinguistic point of view, science – in the modern understanding of an empirical (also experimental) and rationalist undertaking in search of new knowledge, which is prototypically represented by the natural sciences – is a product of the 17th and 18th centuries (Hunter 1981). What we find before that is the more traditional and conservative scholastic tradition as well as various precursors of modern science from the late Middle Ages onwards, both with regard to thinking, method, and language use. This development is being charted by the “Scientific Thought-styles” project centering on the exemplary investigation of medical writing, which revealed a shift from argumentation founded on established authorities to more evidence-based argumentation. The split of the domain into a learned-popular continuum, emerging from late ME (Taavitsainen 2005), can be seen in this connection; while the learned end is characterized by expository and argumentative texts, the popular side tends to be practically oriented and instructive in style. Taavitsainen (2001) presents a survey of the early development of the scientific register. Vernacular writing in medicine, followed by other scientific fields (e.g. astronomy), emerged in the 14th century. Some forms of writing in this field had a continuous vernacular history, e.g. recipes, rules of health, charms, prognostications, and remedy books (Taavitsainen 2005). Generally, however, the developing English register took as its model the Latin academic and scientific register (termed “modelling from above” by Taavitsainen 2001: 188), which was aided by the fact that many texts were either translations (e.g. by Trevisa in the 14th century) or adaptations of foreign sources. Newly introduced genres, such as specialized treatises or surgical texts, often made their way into English through translations. In particular the conventions for the learned genres were apparently taken over from the Greco-Roman tradition, but research in this area is complicated by the fact that the relevant genres in Latin and Greek have not been sufficiently investigated (Taavitsainen 2005). Taavitsainen mentions that the classical format of questions and answers, which was simplified in English, developed from fairly irregular early attempts to fixed and regular structures in the 16th century, and remained in use in handbooks for centuries. The scientific vocabulary is thus also based on extensive borrowing from Romance sources (increasing from the 14th century onwards), with the respective greater use of native vs. foreign lexical resources corresponding to more practical/popular writings vs. more theoretical/academical writing in the medical field, for example. While the Helsinki research group around Taavitsainen and Pahta concentrates on medical writing, Halliday (1988) charted the development of physics writing as a representative of the scientific register. He shows how the major propositional points are over time increasingly presented in nominalized forms, easily allowing both objectivization, categorization, and fore-/backgrounding, while verbs are progressively more restricted to expressing relational or existential aspects of the nominal arguments. In parallel, though starting somewhat later than the just mentioned aspects, the depersonalisation of scientific discourse is in progess. Let me end this section with two general points. First, translations have played a role in various registers. This means that register and textual studies will have to pay attention to how foreign models influenced English developments, and thus, more crosslinguistic studies and also more text-oriented studies on languages such as Medieval Latin are necessary. Second, a link can be made between register development and

241

242

II Linguistic Levels standardization in English (cf. also Section 3), as different registers had a differing impact on the standard and also were receptive to the standard to various extents. The administrative register in the form of Chancery writings was one input into the emerging English standard (Fisher 1996), Lollard texts within the religious register have been connected with another important strand of standardizing varieties (Samuels 1963), with which ME scientific writing also had certain affinities, although this register seems to have resisted the standard somewhat more (Taavitsainen 2001). In EModE times, the Book of Common Prayer and the Authorized Version also had a standardizing impact.

3 Stylistic developments in English Styles are at least partly linked to the orality-literacy continuum, with oral and literate features mixing in particular ways in texts. In this connection, standardization, especially the process of elaboration (Haugen 1972), is important for stylistic developments, as standardization will emphasize the literate end of the continuum. Another important aspect is the varying impact of foreign models, French and in particular Latin, which goes together with the long-standing but ultimately waning influence of (classical) rhetoric. Furthermore, specific stylistic changes are sometimes embedded in certain registers and genres, literary as well as non-literary, which also accounts for the substantial contributions to historical stylistics from literary studies. A last aspect of note is the fact that most studies of style deal with the (early) modern period, at the most extending into late Middle English. Gordon (1966), still the only large-scale historical study of prose style extant, shows this bias as well, the Middle Ages being dealt with in a mere 35 pages.

3.1 Standardization processes and their stylistic effects The syntactic elaboration of English proceeding especially during the 16th to 18th centuries ultimately produced more complex, ordered, and explicit structures, all of them principally useful for complex writing and thus essentially literate (e.g. Rissanen 1999 for an overview of syntactic developments). Such stylistically relevant structures include greater and more sophisticated use of hypotactic and embedded structures (adverbial and relative clauses, non-finite constructions, innovations in the conjunction class) and emerging and/or expanding use of topicalizing constructions (e.g. passive, clefting), as well as more discriminate use of prepositions and prepositional phrases, enabling a tighter information structure. In Chafe’s (1982) terms, these characteristics make for the more integrated and detached characteristics of modern writing. Despite the fact that it is difficult to clearly attribute individual changes to foreign, especially Latin, models, the general attitude towards Latin as a model of grammatical precision and stylistic elegance fostered the stylistic “improvement” of English in order to make it “equal” to Latin (cf. the varying assessment of English reported e.g. by Rusch 1972). What also played a role in this context is the humanistic rediscovery of the original classical texts and the influence of classical rhetoric in English education (cf. the works of Erasmus and, in English, Wilson; cf. Plett 2004). The lexical elaboration of English, going on since the early Middle Ages but consciously intensified particularly in the 16th and 17th centuries (cf. Nevalainen 1999; Barber 1997), produced stylistic levels within the lexicon, with the Romance elements tending towards more impersonal, abstract, and formal expression versus the native

16 Linguistic Levels: Styles, registers, genres, text types elements’ potentially more down-to-earth, emotive, and colloquial impact. While the lexical division is partly linked to register (e.g. cardio-/cardiac in the medical register vs. heart elsewhere), it can also be exploited for stylistic effects. Hughes (2000) illustrates how (quasi-) synonymous doublets and triplets differ stylistically depending on etymology and time of borrowing. In general one can say that the style of a given piece of writing or passage from at least about the 15th century onwards is in no small part determined by the percentage of Romance and classical vocabulary used (Gordon 1966). The so-called “aureate” style of Lydgate and others in the 15th century is created by a conscious use of elevated Romance lexis, and partly also by complex noun phrases and Latinate syntax. How Shakespeare exploited these lexical distinctions for characterizing persons and situations as well as changing relationships and generally for stylistic effect was shown by Scha¨fer (1973). Besides expanding the means of the language in general, lexical growth had a stylistic aim from the start in being crucially linked to the rhetorical concept of copiousness, i.e. lexical variety through amplification, synonymy, repetition, and paraphrasing.

3.2 Oral vs. literate styles The connection between standardization, writing, texts and the orality-literacy cline was investigated in particular by Biber (1995; cf. also Biber and Finegan 1989, 1992). He examined the development of eight English genres, called “registers” by him (personal essays, medical research articles, science research articles, legal opinions, fiction, personal letters, and dialogue from drama and from fiction) from the 17th to the 20th century based on the three dimensions: (1) Involved vs. Informational Production, (2) Situation-dependent vs. Elaborated Reference, and (3) Non-abstract vs. Abstract Style. These dimensions are characterized by the significant presence or absence of the following features: (1) private verbs, 1P and 2P pronouns, present tense verbs, demonstrative pronouns, be as main verb etc., (2) wh-relative clauses, pied piping, nominalisations, phrasal coordination, (3) conjuncts, passives, past participles, adverbial subordinators (cf. Biber 1988 for a comprehensive feature listing). The features of (1) represent the oral/spoken end of the continuum and those of (2) and (3) the written/literate end. Biber’s results show that all prose registers are clearly non-oral in character already at the beginning of his time frame and that most of them become increasingly more literate during the next 100 or 200 years, especially more informational and elaborated. The 18th century in particular is characterized by very pronounced literate textual realizations. According to Biber (1995: 298), such a development is typical of the early stages of the introduction of writing in a language. As this does not really apply to English in the 17th century, it is rather the combined result of ongoing standardization, in particular the culmination point of its normative phase in

243

244

II Linguistic Levels the 18th century, and perhaps the requirements of the more “public” genres. McIntosh’s (1998) “gentrification” thesis confirms and complements Biber’s results, by tracing a development towards a more literate and elegant prose in the course of the 18th century, a process that is driven by class-consciousness and a feeling of propriety. The resulting “gentrified” style is marked by complex but very orderly sentences, passives, nominalizations, as well as polysyllabic and abstract vocabulary. While Biber’s genres develop largely in parallel until the 18th century, they start to diverge from the 19th century onwards. Medical, scientific, and legal prose consistently proceeded to develop towards even more literate styles up to the 20th century, while the popular, non-expository genres (essays, fiction, drama, letters) gradually and increasingly reversed towards more oral, i.e. involved, non-abstract, and situation-dependent, realizations. In the modern period there is thus a clear stylistic split between the two groups. While the research just summarized deals with a long-range perspective and uses the stylistic dimensions originally produced on the basis of 20th-century material, Biber (2001) reapplied factorial analysis to 18th-century data, thus producing the stylistic dimensions typical for that time. The results point to the fact that there was a more pronounced distinction between spoken and written registers in the past than there is today, with drama but none of the other types being marked as extremely oral. It is also possible that this result is an artefact of the missing spoken dimension, thus making drama stand out in peculiar ways.

3.3 Historical styles: two examples Any literary history will abound with styles, be it the aureate diction of the 15th century, Euphuism in the 16th century or Milton’s Latinate style. Some of these may be of literary interest only, while others are relevant to both literary and linguistic scholars, among which are the clergial or curial style, the so-called “plain style” and the stylistic shift taking place in the 18th century mentioned above. The clergial style, found in the 14th and 15th centuries, derives ultimately from medieval chancellary Latin, the dictaminal arts in general, and more directly from French vernacular models, which were imitated and developed by English writers. It eventually turned into a courtly prose style. It can be found in Chaucer’s prose (Bornstein 1978, with a long list of examples), in writings of the English royal administration and in Caxton’s works (Burnley 1986; Blake 1992). Features of this style include Latinate words and constructions, extensive clausal qualifiers, and long complex sentences, synonymous doublets, anaphoric cohesive devices, epithets, and a preference for the use of the passive. The plain style emerging in the 17th century needs to be looked at in a larger context, as it is partly a reaction against other stylistic models. According to Sprat’s wellknown formulation about the discourse of the Royal Society the motivation for the plain style was to reject all the amplifications, digressions, and swellings of style: to return back to the primitive purity, and shortness, when men deliver’d so many things, almost in an equal number of words. […] a close, naked, natural way of speaking; positive expressions; clear senses; a native easiness: bringing all things as near the Mathematical plainness, as they can: and preferring the language of Artizans, Countrymen, and Merchants before that, of Wits, or Scholars (Sprat 1667: 113).

16 Linguistic Levels: Styles, registers, genres, text types In spite of Sprat’s testimony neither the genesis, the characteristics, nor the type of users of the plain style are entirely clear; partly this may also be due to the fact that the 17th century itself used the term to refer to opposing styles (Adolph 1968: 130). Taavitsainen (2001: 196) characterizes the “house style” of the Philosophical Transactions, the organ of the Royal Society, as marked by first-person narration, subjective point of view, and expressions of low modality. Hu¨llen’s (1989) analysis of some of the descriptive adjectives Sprat uses (e.g. plain, naked, easy) shows that apart from conveying the senses ‘generally understandable’ and ‘unadorned’ they have a number of meanings linking linguistic style to thought, and more generally life styles (e.g. ‘not intellectual, objective, frank, theoretical, contentedness’). For Gordon (1966: 127), this style embodied a rejection of Latinate syntax (Ciceronian and Senecan), of rhetorical figures, and of metaphor and simile, as well as a return to Anglo-Saxon sentence structures – i.e. it is one form of the speech-based prose using ordinary vocabulary and the “grammar of spoken English” (Gordon (1966: 122) that he identified in the 17th century. It is, however, the genteel and polite form produced by educated gentleman, which is what Sprat’s early scientists were; it is not necessarily the speech of merchants or artisans (Gordon 1966: 128; Atkinson 1996: 362–364; Hu¨llen 1989: 84). While Sprat linked the new style to the Royal Society and thus to natural science, it is better seen as a larger and manifold process. A plain style is also found with authors such as John Wilkins, John Webster, and Joseph Glanvill, and in fields such as law, religion, and travel literature, which do not belong to the new science (Hu¨llen 1989: 70), and needs to be seen in the context of a larger development from a rhetoric of persuasion to one of exposition (thus linking up with text types) from the end of the 16th century onwards (Howell 1956: 388). Various authorities, while agreeing on the overall stylistic shift, have emphasized different aspects and times as being decisive for the new style. Fish (1952) stressed the importance of puritan preaching style, shifting the emergence back to as early as 1570, and Warner (1961: 97) saw the likelihood of an influence of the Authorized Version of the Bible, science figures prominently in other explanations. Adolph (1968), placing the shift around 1600, saw the scientific and puritan notions of utilitarianism as decisive for its development. Picking the same time frame, Croll (1921) opted for Francis Bacon’s personal antiCiceronian style as the foundation of the new plain style. Jones (1953) came closest to Sprat by identifying the Restoration period and the emerging sciences and rationalism as the crucial aspects for the new style. Despite the disagreements, some things seem clear: new communicative needs, based on diverse religious practices and on an emerging natural science, but also on newly arising economic and social conditions (e.g. rise of the “middle classes”), led to a new style. This style was less overtly influenced by Latin rhetorical models, broke with “scholastic” (thought)styles (cf. also Taavitsainen 2001), and was potentially more focused on the author. Gordon’s characterization of this style as speech-based, however, conflicts with Biber’s (1995) and Biber and Finegan’s (1989, 1992) findings of the already largely literate character of this period’s prose (cf. above).

4 Approaches to historical texts Kohnen (2008) distinguishes three approaches to the investigation of historical texts: (1) historical text linguistics proper, (2) diachronically oriented text linguistics, and (3) text-oriented historical linguistics.

245

246

II Linguistic Levels The first of these approaches is the study of texts, genres and discourse domains in historical periods of a language, which crucially includes genre inventories. The second approach studies changes in texts, genres and discourse domains across periods, e.g. conventional text structure, or the link between macro-categories and specific text types. The third approach deals with the functions of texts, genres and discourse domains in language change.

4.1 Historical text linguistics: genre histories Go¨rlach (1992, 2001b, 2002, 2004) takes an external approach to the question of genres, although he uses the term “text type”. His suggestion is to collect genre names existing at a given time or throughout the history and to conduct a componential analysis on the lexical field found. Go¨rlach (2001b, 2004) gives the following 24 parameters (thought of as equivalent to semantic markers) for distinguishing text types, to which I add in brackets his characterization of the genre “contract”: field (law), intention (binding), act/ action connected to text (none), accompaniment by music or visual material (no), conglomerate (no), composite (yes), boundness (free), cohesion/coherence (yes), original (yes), nativeness (yes), general/regional (standard), time (present-day), medium (written), style (formal), form (prose), formulaicness (yes), orientation (content-oriented), specialization (technical), truth (non-fictional), spontaneity (no: revised), publicness (yes), length (n.a.), and official (yes). Go¨rlach (2004) provides two alphabetical lists of English genre terms based on entries in the Shorter Oxford English Dictionary (with dates) and on the Thesaurus of Old English. Examples of genre terms from Go¨rlach’s list are eulogy, handbook, invoice, joke, lecture, parable, report, and small talk. He also illustrates his approach with an exemplary componential analysis of genre terms beginning with , from which the contract example above was taken. The presence or absence of a genre term at a given time, extractable from Go¨rlach’s list, and its understanding by contemporaries themselves highlight the cultural determinacy of the concept genre. Genres and text types may remain fairly stable over a long time (e.g. the sonnet), but equally they may be falling into disuse (e.g. those marked by a dagger in Go¨rlach’s list, or the telegram at present) or be newly emerging (e.g. the essay in the 16th century) (cf. Go¨rlach 2002). Some types evolve only within (newly established) media, registers, or publication types, such as dedications within books (15th century), and a whole variety within newspapers (e.g. editorial 1830, obituary 1828, weather forecast 1883; Go¨rlach 2004; see Fries Chapter 67), though sometimes taking existing genres as models (e.g. early dedications and news reports being connected to letters).

4.2 Diachronically oriented text linguistics Conventional and changing textual structures have been investigated for the genre of letters, for example. Richardson (1984) describes Chancery and other letters, based on Latin style, to be structured as follows: (1) address, (2) salutation, (3) notification,

16 Linguistic Levels: Styles, registers, genres, text types (4) (5) (6) (7) (8) (9)

exposition, disposition/ disjunction, final clause, valediction, attestation and date.

Davis (1965), in contrast, argues for a French model to be followed in correspondence, consisting of (1) (2) (3) (4)

address, commendation formula, health formula at the beginning of the letter, and closing formula.

Nevalainen (2001) explores the use of these formulae in the letters of the Johnson merchant family written 1542–1552. She finds that the Johnson letters follow six of the nine medieval letter-writing conventions, namely date (and place), salutation, address, “health” formula, notification, and (a form of) valediction, but they do this to varying degrees, with great individual variation and with additions and modifications. Okulska (2006) deals with structural letter elements, but also with topic development and information structure in the sub-genre of the diplomatic narrative report letter, from the 15th to the 18th century. With regard to the latter aspect, the letters represent discursive hybrids combining narrative (past orientation, topic-based) and reporting (present tense, person-centered) elements, the latter also frequently triggering evaluative-argumentative comments in the early, but not the later, part of the period. As to the presentation of information in these letters, there is a shift from inductive topic-delayed to deductive topic-first thesis presentation from the 16th century onwards, which puts the focus more clearly on the information itself than on the persons interacting. Another approach with regard to text types is to follow the diachronic development of the prototypical text types, such as narrative or exposition, or, as there are hardly pure types, of genres which are typically dominated by one of these, e.g. works on language as representing the instructive type. This has not been done for any of the types on a broad scale so far, but smaller-scale treatments exist. Taavitsainen (2004), for example, deals with genres of secular instruction from Old English to Early Modern English, in particular with how they are made more pleasant and entertaining for the readers. This means the inclusion of other than purely instructive features, such as the use of verse form, narrative passages, dialogues, typical involvement features, and a conventional fictionalized frame. Taavitsainen found both considerable variation within the text category and interesting diachronic continuities, as well as a “circular movement” of features moving from learned texts to more popular texts, thus again leaving space for innovation at the learned end of the continuum. With regard to narrative discourse, for example, Wa˚rvik (1990a and 1990b) describes the history of grounding markers in English narratives. She finds that English has undergone a change from a foreground-signalling language in Old English (by use of the more or less obligatory marker þa) to a background-signalling language in modern English

247

248

II Linguistic Levels (by means of, e.g., subordination, non-finite forms). Whereas marking in Old English is fairly specific and, in a certain sense, mono-dimensional, the modern system is more fuzzy, being characterized by a variety of non-obligatory, partly stylistic options. A further avenue of research investigates the aspect that certain text categories or genres seem to prefer certain text-level features rather than others. One good example for this is discourse deixis, which has been investigated by Fries (1993, 1994), Claridge (2001), and Kilpio¨ (1997). Fries (1993, 1994) found deictic elements marking text location very or most commonly in the text categories instruction (religious and secular), non-imaginative narration, and exposition in Old English and Early Modern English. The markers afore and the said very clearly dominate in statutory texts. Statutory and similar texts also figure prominently in Kilpio¨’s (1997) diachronic investigation, which concentrates on participial adjectives of the type (a) forementioned in the entire Helsinki Corpus (HC) (thus using the Helsinki Corpus text-labelling terminology). While these discourse-deictic elements are rare in Old English and early Middle English, they become more frequent from 1350 onwards with genres like official letters, documents, and law heading the frequency lists (joined by history and science in Early Modern English). In contrast, Old English law does not use such features. Of interest are also the kinds of genres which make do without these discourse-deictic elements, namely Middle English and early Modern English rule, Bible, homily, drama, comedy, and those that use them very rarely, namely sermons, fiction, and education. Sermons and religious treatises also show the lowest instance of discourse-deictic terms in Claridge’s (2001) study, while the legal register, represented by many different genres in the corpus, has the highest number. Early Modern English scientific genres show a high number of items indicating present location and forward-indicating items, pointing to the importance of commenting on the ongoing discourse procedure of the author. The study of connective devices is another case in point for text-level feature preferences insofar as it illuminates the kinds of explicit cohesive relationships typical for certain genres. Kohnen (2007) investigated what he termed “connective profiles”, i.e. the overall use of the whole range of coordinating and subordinating conjunctions/complementizers. Comparing 17th-century sermons and statutes he found quite distinct profiles: a fairly high and heterogeneous number of connectives in sermons and a much lower, less varied use in statutes. Partly this seems to be due to a stylistic preference for conjunctionless non-finite subordination in statutory language. But it is also based on the communicative functions common of statutes. Rare but and non-existent for are expendable because contrastive viewpoints and the provision of justifications are hardly found in these texts, whereas moderately common and is useful for enumerations. All subordinators are less common in statutes than in sermons, except for manner/comparison markers, which is due to the legal formulaic use of introductory whereas/as clauses. A notable frequency is furthermore only exhibited by nominal clause connectives (especially that) and by conditional clauses. Kohnen also looked at the diachronic development of sermons over the 15th, 16th, 17th, and late 20th century, and sums up the general pattern as a decrease in oral features and simultaneous increase in literate features in Early Modern English, followed by the reverse development in the 20th century.

16 Linguistic Levels: Styles, registers, genres, text types

4.3 The role of texts in language change Kohnen (2001) sees genres (“text types” in his terminology) as catalysts in language change, i.e. they facilitate change and are responsible for the spreading of a construction, his example being the adverbial first participle construction. He showed how this construction spread across text types over a period of two centuries, its textual frequency first increasing in religious treatises around 1340, then about 1390 in both homilies/sermons and petitions/statutes, by roughly 1470 in chronicles, and finally by around 1520 in narrative prose and private letters. In each instance the construction fulfils an important functional requirement of the text type, e.g. introducing explanatory passages in Biblical exegesis (treatises), providing vivid descriptions in sermons, petitions, and narrative prose, or encoding formulaic speech acts in the introductory sections of letters. According to Kohnen, the particular chronological adoption of the feature by the genres is due to the following three aspects. Early or well-established vernacular types, such as religious treatises and sermons, make use of the participle construction earlier than types which are found only later as vernacular forms, e.g. letters. So do genres close to Latin and/or French traditions (both of which had the construction in question) and those that are more formal in character, such as religious prose and petitions/ statutes. Lastly, the linguistic features adopted by genres with considerable social relevance and a prestigious status, i.e. religious treatises, statutes, and documents, are more likely to spread to other text types later than features adopted first by less prestigious texts.

5 Summary The above presentation has somewhat artificially separated things that intimately belong together. A style is found in a text which belongs both to a register and to a genre. Ideally, these aspects should be treated together then, but this would increase the complexity to such an extent that larger-scale and diachronic investigations would be difficult. Thus, particular research efforts tend to concentrate on only one of these aspects, even if this necessarily means simplifying. Another complicating factor, also visible in the extant research, is the variable usage of the basic terminology – what is one researcher’s style may be another’s register, which makes comparisons across works difficult. Furthermore, despite Kohnen (2008) quoted above, there is as yet no historical or diachronic text linguistics as an established field. Researchers in this field may at the moment tend to place themselves vaguely in the paradigm of historical pragmatics. In contrast to the theoretical difficulties, actually researching historical texts may be getting increasingly easier. Many of the researchers quoted above approach the field with the help of corpus linguistics. A range of historical corpora (Kyto¨, Chapter 96, Volume 2,) is available by now, all of which use some kind of register, text type or genre coding system. A Representative Corpus of Historical English Registers (ARCHER) (Biber and Finegan 1990–93/2002/2007/2010) has been compiled and is being extended by Biber and various affiliates. The Helsinki Corpus (HC) (Rissanen et al. 1991) contains texts from all the registers mentioned. As more general corpora often do not contain individual registers in sufficient numbers, single-register corpora are a particular asset. The Zurich English Newspaper Corpus (ZEN) (Fries et al.

249

250

II Linguistic Levels 2004), the Lampeter Corpus (pamphlets) (Claridge et al. 1999) and the Lancaster Newsbook Corpus (McEnery and Hardie 2001–07) can be used together to investigate the origins of the press register. Similarly, the Corpus of Early English Medical Writing (CEEM) (Taavitsainen et al. 1995–), compiled at Helsinki University, charts the development of a scientific writing based on the medical prototype. The Corpus of Early English Correspondence Sampler (CEECS) (Kera¨nen et al. 1998) could be called a single-genre corpus. Not all of the above-mentioned corpora contain complete texts, however, which would be of especial importance for the genre or text-type approach.

6 References Adolph, Robert. 1968. The Rise of Modern Prose Style. Cambridge, MA: MIT Press. Archer, Dawn. 2005. Questions and Answers in the English Courtroom (1640–1760). Amsterdam/ Philadelphia: John Benjamins. Atkinson, Dwight. 1996. The Philosophical Transactions of the Royal Society of London, 1675– 1975. Language in Society 25: 333–371. Barber, Charles. 1997. Early Modern English. Edinburgh: Edinburgh University Press. Biber, Douglas. 1988. Variation Across Speech and Writing. Cambridge: Cambridge University Press. Biber, Douglas. 1989. A typology of English texts. Linguistics 27(1): 3–43. Biber, Douglas. 1995. Dimensions of Register Variation. Cambridge: Cambridge University Press. Biber, Douglas. 2001. Dimensions of variation among 18th-century registers. In: Diller and Go¨rlach (eds.), 89–109. Biber, Douglas, and Edward Finegan. 1989. Drift and the evolution of English style: A history of three genres. Language 65: 487–517. Biber, Douglas and Edward Finegan. 1990–93/2002/2007/2010. A Representative Corpus of Historical English Registers (ARCHER). Version 3.1. Consortium of fourteen universities, see http:// www.llc.manchester.ac.uk/research/projects/archer/ Biber, Douglas, and Edward Finegan. 1992. The linguistic evolution of five written and speechbased English genres from the 17th to the 20th centuries. In: Rissanen et al. (eds.), 688–704. Blake, Norman. 1992. The literary language. In: Norman Blake (ed.), Cambridge History of the English Language, Vol. II. 1066–1476, 500–541. Cambridge: Cambridge University Press. Bornstein, Diane. 1978. Chaucer’s Tale of Melibee as an example of “style clergial”. The Chaucer Review 12: 236–254. Burnley, J. David. 1986. Curial prose in England. Speculum 61(3): 593–614. Chafe, Wallace. 1982. Integration and involvement in speaking, writing, and oral literature. In: Deborah Tannen (ed.), Spoken and Written Language: Exploring Orality and Literacy, 35–53. Norwood, NJ: Ablex. Claridge, Claudia, Josef Schmied, and Rainer Siemund. 1999. The Lampeter Corpus of Early Modern English Tracts. In: ICAME Collection of English Language Corpora (CD-ROM), 2nd edn., Knut Hofland, Anne Lindebjerg, and Jørn Thunestvedt (eds.), The HIT Centre, University of Bergen, Norway For manual see http://khnt.hit.uib.no/icame/manuals/LAMPETER/ LAMPHOME.HTM Claridge, Claudia. 2001. Structuring text: Discourse deixis in Early Modern English Texts. Journal of English Linguistics 29/1: 55–71. Croll, Morris W. 1921. “Attic Prose” in the seventeenth century. Studies in Philology 18: 79–128. Davis, Norman. 1965. The Litera Troili and English letters. Review of English Studies, New Series 16: 233–244. Diller, Hans-Ju¨rgen. 1998. Stylistics: linguistic and textual. European Journal of English Studies 2(2): 155–174.

16 Linguistic Levels: Styles, registers, genres, text types Diller, Hans-Ju¨rgen. 2001. Genre in linguistic and related discourses. In: Diller and Go¨rlach (eds.), 3–43. Diller, Hans-Ju¨rgen and Manfred Go¨rlach (eds.). 2001. Towards a History of English as a History of Genres. Heidelberg: Winter. Fish, Harold. 1952. The Puritans and the reform of prose-style. A Journal of English Literary History 19: 229–248. Fisher, John. 1996. The Emergence of Standard English. Lexington: University Press of Kentucky. Fries, Udo. 1993. Towards a description of text deixis in Old English. In: Klaus R. Grinda and Claus-Dieter Wetzel (eds.), Anglo-Saxonica: Beitra¨ge zur Vor- und Fru¨hgeschichte der englischen Sprache und zur altenglischen Literatur, 527–540. Munich: Fink. Fries, Udo. 1994. Text deixis in Early Modern English. In: Dieter Kastovsky (ed.), Studies in Early Modern English, 111–128. Berlin/New York: Mouton de Gruyter. Fries, Udo, Hans Martin Lehmann et al. 2004. Zurich English Newspaper Corpus (ZEN). Version 1.0. Zu¨rich: University of Zu¨rich. http://es-zen.unizh.ch http://www.helsinki.fi/varieng/CoRD/ corpora/ZEN/index.html (see also: http://www.helsinki.fi/varieng/CoRD/corpora/ZEN/index. html). Gordon, Ian A. 1966. The Movement of English Prose. London: Longman. Go¨rlach, Manfred. 1992. Text-types and language history: the cookery recipe. In: Rissanen et al. (eds.), 736–761. Go¨rlach, Manfred. 1999a. English in Nineteenth-century England. Cambridge: Cambridge University Press. Go¨rlach, Manfred. 1999b. Regional and social variation. In: Lass (ed.), 459–538. Go¨rlach, Manfred. 2001a. Eighteenth-century English. Heidelberg: Winter. Go¨rlach, Manfred. 2001b. A history of text types: A componential analysis. In: Diller and Go¨rlach, 47–88. Go¨rlach, Manfred. 2002. What’s in a name? Terms designating text types and the history of English. In: Andreas Fischer, Gunnel Tottie, and Hans-Martin Lehmann (eds.), Text Types and Corpora. Studies in Honour of Udo Fries, 17–28. Tu¨bingen: Narr. Go¨rlach, Manfred. 2004. Text Types and the History of English. Berlin/New York: Mouton de Gruyter. Grund, Peter. 2007. From tongue to text: The transmission of the Salem Witchcraft Examination Records. American Speech 82: 119–150. Halliday, M. A. K. 1988. On the language of physical science. In: Mohsen Ghadessy (ed.), Registers of Written English, 162–178. London: Pinter. Haugen, Einar. 1972. Dialect, language, nation. In: J. B. Pride and Janet Holmes (eds.), Sociolinguistics, 97–111. Harmondsworth: Penguin. Hiltunen, Risto. 1990. Chapters on Legal English. Aspects Past and Present of the Language of the Law. Helsinki: Suomalainen Tiedeakatemia. Howell, W. S. 1956. Logic and Rhetoric in England, 1500–1700. New York: Russell & Russell. Hughes, Geoffrey. 2000. A History of English Words. Oxford: Blackwell. Hu¨llen, Werner. 1989. “Their Manner of Discourse”: Nachdenken u¨ber Sprache im Umkreis der Royal Society. Tu¨bingen: Narr. Hunter, Michael. 1981. Science and Society in Restoration England. Cambridge: Cambridge University Press. Jones, Richard Foster. 1953. The Triumph of the English Language. Stanford: Stanford University Press. Kera¨nen, Jukka, Minna Nevala, Terttu Nevalainen, Arja Nurmi, Minna Palander-Collin and Helena Raumolin-Brunberg. 1998. Corpus of Early English Correspondence Sampler (CEECS). In: ICAME Collection of English Language Corpora (CD-ROM), 2nd edn., Knut Hofland, Anne Lindebjerg, and Jørn Thunestvedt (eds.), The HIT Centre, University of Bergen, Norway. For manual see http://khnt.hit.uib.no/icame/manuals/ceecs/INDEX.HTM (See also: http://www.helsinki.fi/varieng/domains/CEEC.html)

251

252

II Linguistic Levels Kilpio¨, Matti. 1997. Participial adjectives with anaphoric reference of the type the said, the (a) forementioned from Old to Early Modern English: The evidence of the Helsinki Corpus. In: Terttu Nevalainen and Leena Kahlas-Tarkka (eds.), To Explain the Present: Studies in the Changing English Language in Honour of Matti Rissanen, 77–100. Helsinki: Socie´te´ Ne´ophilologique. Kohnen, Thomas. 2001. Text types as catalysts for language change: The example of the adverbial first participle construction. In: Diller and Go¨rlach (eds.), 111–124. Kohnen, Thomas. 2007. “Connective profiles” in the history of English texts. In: Ursula Lenker and Anneli Meurman-Solin (eds.), Connectives in the History of English, 289–308. Amsterdam/Philadelphia: John Benjamins. Kohnen, Thomas. 2008. Historical text linguistics: Investigating language change in texts and genres. Paper delivered at ICEHL 15, Munich. Lass, Roger (ed.). 1999. Cambridge History of the English Language, Vol. III. 1476–1776. Cambridge: Cambridge University Press. Liebermann, Felix. 1903–16. Die Gesetze der Angelsachsen. 3 vols. Halle an der Saale: Max Niemeyer. Longacre, Robert E. 1996. The Grammar of Discourse. 2nd edn. New York: Plenum Press. McEnery, Tony and Andrew Hardie. 2001–07. Lancaster Newsbooks Corpus. UCREL and Linguistics and English Language, University of Lancaster. http://www.ling.lancs.ac.uk/activities/ 695/. Available through the Oxford Text Archive: http://ota.ahds.ac.uk/headers/2531.xml McIntosh, Carey. 1998. The Evolution of English Prose, 1700–1800: Style, Politeness, and Print Culture. Cambridge: Cambridge University Press. Mellinkoff, David. 1963. The Language of the Law. Boston: Little, Brown. Moessner, Lilo (ed.). 2001. Special Issue of European Journal of English Studies 5(2). Nevalainen, Terttu. 1999. Early Modern English lexis and semantics. In: Lass (ed.), 332–458. Nevalainen, Terttu. 2001. Continental conventions in early English correspondence. In: Diller and Go¨rlach (eds.), 203–224. Okulska, Urszula. 2006. Textual strategies in the diplomatic correspondence of the Middle and Early Modern English periods: The narrative report letter as a genre. In: Marina Dossena and Susan Fitzmaurice (eds.), Business and Official Correspondence: Historical Investigations, 47–76. Bern: Lang. Plett, Heinrich F. 2004. Rhetoric and Renaissance Culture. Berlin/New York: Mouton de Gruyter. Richardson, Malcolm. 1984. The dictamen and its influence on fifteenth-century English prose. Rhetorica 2: 207–226. Rissanen, Matti. 1999. Syntax. In: Lass (ed.), 187–331. Rissanen, Matti, Merja Kyto¨, Leena Kahlas-Tarkka, Matti Kilpio¨, Saara Nevanlinna, Irma Taavitsainen, Terttu Nevalainen, and Helena Raumolin-Brunberg. 1991. The Helsinki Corpus of English Texts (HC). In: ICAME Collection of English Language Corpora (CD-ROM), 2nd edn., Knut Hofland, Anne Lindebjerg, and Jørn Thunestvedt (eds.), The HIT Centre, University of Bergen, Norway. For manual, see http://khnt.hit.uib.no/icame/manuals/hc/index.htm Rissanen, Matti, Ossi Ihalainen, Terttu Nevalainen, and Irma Taavitsainen (eds.). 1992. History of Englishes: New Methods and Interpretations in Historical Linguistics. Berlin/New York: Mouton de Gruyter. Rusch, Ju¨rg. 1972. Die Vorstellung vom Goldenen Zeitalter der englischen Sprache im 16., 17., und 18. Jahrhundert. Bern: Francke. Samuels, M. L. 1963. Some applications of Middle English dialectology. English Studies: A Journal of English Language and Literature 44: 81–94. Scha¨fer, Ju¨rgen. 1973. Shakespeares Stil: germanisches und romanisches Vokabular. Frankfurt: Athena¨um. Skaffari, Janne, Matti Peikola, Ruth Carroll, Risto Hiltunen, and Brita Wa˚rvik (eds.). 2005. Opening Windows on Texts and Discourses of the Past. Amsterdam/Philadelphia: John Benjamins.

16 Linguistic Levels: Styles, registers, genres, text types Sprat, Thomas. 1667. The history of the Royal-Society of London for the improving of natural knowledge by Tho. Sprat. London: Printed by T. R. for J. Martyn …, and J. Allestry … (accessible via EEBO: http://eebo.chadwyck.com/home). Taavitsainen, Irma. 2001. Language history and the scientific register. In: Diller and Go¨rlach (eds.), 185–202. Taavitsainen, Irma. 2002. Historical discourse analysis: Scientific language and changing thoughtstyles. In: Teresa Fanego, Bele´n Me´ndez-Naya, and Elena Seoane (eds.), Sounds, Words, Texts and Change, 201–226. Amsterdam/Philadelphia: John Benjamins. Taavitsainen, Irma. 2004. Genres of secular instruction: A linguistic history of useful entertainment. Miscela´nea: A Journal of English and American Studies 29: 75–94. Taavitsainen, Irma. 2005. Genres and the appropriation of science: Loci communes in English in the late medieval and early modern period. In: Skaffari et al. (eds.), 179–196. Taavitsainen, Irma, Pa¨ivi Pahta, Martti Ma¨kinen, Turo Hiltunen, Ville Marttila, Maura Ratia, Carla Suhr, and Jukka Tyrkko¨. forthc. Corpus of Early English Medical Writing (CEEM). University of Helsinki. http://www.helsinki.fi/varieng/CoRD/corpora/CEEM/index.html Warner, Alan. 1961. A Short Guide to English Style. Oxford: Oxford University Press. Wa˚rvik, Brita. 1990a. On grounding in English narratives: A diachronic perspective. In: Sylvia Adamson, Vivien Law, Nigel Vincent, and Susan Wright (eds.), Papers from the 5th International Conference on English Historical Linguistics, Cambridge, 6–9 April 1987, 559–575. Amsterdam/ Philadelphia: John Benjamins. Wa˚rvik, Brita. 1990b. On the history of grounding markers in English narrative: Style or typology? In: Henning Andersen and Konrad Koerner (eds.), Historical Linguistics 1987, 531–542. Amsterdam/Philadelphia: John Benjamins. Werlich, Egon. 1983. Text Grammar of English. 2nd edn. Heidelberg: Quelle & Meyer. Wright, Laura. 1992. Macaronic writing in a London archive, 1380–1480. In: Rissanen et al. (eds.), 762–779.

Claudia Claridge, Duisburg-Essen (Germany)

253

III Old English 17 Old English: Phonology 1 2 3 4 5 6

Introduction Terminology, evidence, methods Synchrony Phonological–orthographic correspondences Diachrony References

Abstract The investigation of Old English phonology has been incessant over the decades and carried out from the vantage point of many different theoretical perspectives, but it remains a remarkable fact of our neogrammarian legacy that fundamental aspects of their Old English analyses have weathered the changing theoretical winds particularly well. This high degree of consensus allows us to present a very broad, relatively uncontroversial overview of many fundamental aspects of OE phonology. At the same time, of course, problematic areas remain, and some controversies of perennial interest are indicated.

1 Introduction Old English (OE) phonology has been the subject of scientific investigation for well over 100 years. At the end of the 19th century, an informal group of linguists and philologers based in Leipzig and known as the “neogrammarians” (Ger. Junggrammatiker) constituted the dominant force in linguistic science. Their primary interest was the study of language change – especially sound change in light of the budding science of phonetics (see Sievers 1901) – and a significant amount of their scholarly attention was directed at the investigation of the earliest stages of the Germanic languages. Much of this work was codified in grammars and historical handbooks, which – in keeping with the neogrammarian emphasis – focused primarily on phonological reconstruction and sound change, as well as inflectional morphology. As an early Germanic dialect with a relatively long documented history that included significant literary works such as Beowulf, Old English was of natural interest to the neogrammarians. Eduard Sievers – perhaps the most accomplished and renowned neogrammarian – published his Angelsa¨chsische Grammatik (‘Old English Grammar’) in 1882, and other grammars and classic works strongly influenced by the neogrammarian perspective followed, such as Sweet (1888), Bu¨lbring (1902), Wright and Wright (1925), Campbell (1959), and Brunner (1965 – the last revision of Sievers’s grammar). Luick (1964a [1914–21], 1964b [1929–40]) is of special note in its copious treatment of the sound changes through the entire history of English from its Proto-Indo-European origins to the modern stages. (Luick 1964a [1914–1921]: 94–320 treats phonology and sound changes up to the end of the Old English period.) Although there were many areas of disagreement, for the main part our classic grammars and handbooks reflect a kind of rough neogrammarian consensus Alex Bergs and Laurel Brinton 2012, Historical Linguistics of English (HSK 34.1), de Gruyter, 255–272

256

III Old English achieved by the early 20th century after decades of intensive research. These works, and especially the much more comprehensive studies on which they are built, remain invaluable research tools to the present day, which in many respects have not been superseded.

2 Terminology, evidence, methods Before presenting our overview of OE phonology, a brief discussion of terminology is in order. In fact, the designation “Old English” is a very broad cover term that abstracts away from the dialectal and chronological realities. The OE period is traditionally set at 700–1100 CE, and four major dialect groups are recognized: Northumbrian, Mercian, West Saxon, and Kentish (where Northumbrian and Mercian together form Anglian). As such, it is sometimes not particularly meaningful to speak of OE phonology without qualification, given the inherent dialectal diversity and the significant changes that occurred over the four hundred year time span. In fact, most traditional descriptions focus on the best-documented dialect group, West Saxon, in which one variant became a kind of standard language around the end of the 10th century. This classical Old English, best represented in the works of Ælfric, is the default reference point for our treatment. It is also worth noting here that there is no direct line of descent from this classical Old English to any of the modern standard varieties, which are based primarily on the Mercian dialect group (see Hogg 1992a: 83–84 and Hogg 2006, Lass 1994: 1–5; see also Sauer and Waxenberger, Chapter 22; Kornexl, Chapter 24). We also briefly consider the types of evidence used in the reconstruction of OE phonology (see Lass 1992: 27–32 for a good discussion of the role of various kinds of evidence). It is an unfortunate fact of the Old and Middle English periods that we have no direct descriptions of the spoken form of any dialect, since such descriptions of English only begin in the 16th century. Accordingly, the most direct evidence available for the OE period is primarily of three types: spelling, poetic metrical conventions, and borrowing. In the case of writing, although it is an exaggeration to state that the “scribes wrote as they spoke”, spelling during the OE period was more or less phonologically based, so generally there was a fairly close match between the respective phonological and written forms. This fact, along with the use of the Roman alphabet, means that the first level of analysis – that is, a very broad phonological rendering – is relatively straightforward at the segmental level, although of course not entirely without controversy. At the same time, orthographic systems are inevitably deficient in various ways, and some properties, such as vowel length, were not indicated by the OE scribes. Vowel length, however, can be reconstructed on the basis of the metrical principles derived from such poetic works as Beowulf. The meter of this type of poetry is based on a crucial distinction between light and heavy syllables, which depends in part on the contrast between a short and a long vowel in an open syllable. For example, the first syllables in the words cwe˘.ne ‘woman’ and de¯.man ‘to judge, deem’ function metrically as light and heavy, respectively (see Section 3.6 below). Further corroboration of a length contrast is found in the borrowing of words from Latin into Old English, which although quite limited in scope allows for some cross-referencing in light of our more detailed knowledge of Latin phonology. For example, the borrowing of Lat. nō na as OE nō n ‘noon’, beyond suggesting a phonetic similarity in the quality of

17 Old English: Phonology

257

Lat. and OE ō , dovetails more generally with the reconstruction of a vowel length contrast for both languages. Most importantly, though, the task of reconstructing the synchrony and diachrony of OE phonology is governed by the general principles and methods of linguistic reconstruction (Fox 1995). Fundamental principles here are uniformity (reconstructed systems must be compatible with our knowledge of present-day systems), plausibility (all assumed sound changes must be well motivated), and regularity (generalized regular sound change is assumed, all things being equal).

3 Synchrony 3.1 Vowels In the neogrammarian tradition, a system consisting of seven vowels and contrastive length is reconstructed for Old English (Table 17.1). An additional vowel, [ø] (usually written ), occurred in very early West Saxon, but by the classical OE period it had undergone derounding and merger with e ([ø] survived longer in Mercian and Northumbrian); for example, twoelf ~ twelf ‘twelve’, fo¯et ~ fe¯t ‘feet’ (Hogg 1992b: 124–126). From a typological perspective, the posited Old English vowel system is not particularly unusual, finding a close modern parallel, for example, in Finnish (Maddieson 1984: 275). Table 17.1: Old English vowel system

High Mid Low Long ı¯s ‘ice’ fe¯dan ‘to feed’ dæ¯d ‘deed’ hy¯dan ‘to hide’

Unrounded Front

Rounded

i e æ

y

Unrounded Back

Rounded u o

a Short fı˘sc ‘fish’ le˘þer ‘leather’ ɡlæ̆ d ‘ɡlad’ dy̆ ppan ‘to dip’

Long hlu¯d ‘loud’ bro¯þor ‘brother’ sa¯da ‘snare’

Short hu˘nd ‘hound, dog’ bo˘ɡa ‘bow’ sa˘dol ‘saddle’

Although traditionally duration is considered the primary property distinguishing the Old English short and long vowel pairs, it is common – especially in textbooks on the history of English – to find confident statements about tense/lax distinctions. For example, Pyles and Algeo (1993: 103) state that the short vowels “were approximately [ɛ], [ɪ], [ɔ], and [ʊ] respectively, as in net, nit, nought, and nut”. A more systematic and comprehensive type of challenge to the traditional position was initiated by Stockwell (1952), who analyzed the contrast between the vowel pairs in terms of simple vs. complex (diphthongal) nuclei rather than in terms of duration (Lass and Anderson 1975: 201–205 remains a good discussion of the general issues; see also Hogg 1992a: 85–86 for a discussion of possible special properties of the low vowels). Nevertheless, although the reconstruction of some type of tense/lax contrast is intuitively appealing from the perspective of most present-day standard varieties of English, the phonetic and phonological interplay of vowel length

258

III Old English and quality is an extremely complex area (see, for example, Rosner and Pickering 1994). While it is not uncommon for quality differences (usually described in terms of tenseness or peripherality) to accompany a vowel length contrast, modern languages such as Finnish attest to the fact that salient quality differences are not necessary concomitants of vowel length. In fact, a comparison of Old and Middle English sound changes supports the traditional reconstruction’s focus on the durational – as opposed to qualitative – basis of the OE vowel pairs. In late Old English, in a change known as Homorganic Cluster Lengthening (see also Section 5.2 below), an original short vowel lengthened before clusters such as -ld, -nd, and -mb. In this lengthening, no vowel quality change is evident, as in fe¯ld (< fe˘ld) ‘field’, gru¯nd (< gru˘nd) ‘ground’, and clı¯mban (< clı˘mban) ‘to climb’, thus implying the qualitative matching of the short/long pairs. By contrast, in Middle English, when another set of vowel lengthenings occurred (Open Syllable Lengthening), quality differences are apparent; for example, [e˘] > open [ɛ̄ ] (not [e¯]), as in bre˘ken > bre˛̄ ken ‘to break’); [u˘] > [o¯], du˘res > dō res ‘doors’; and [ı˘] > [e¯], wı˘kes > wē kes ‘weeks’. ˙ ˙ This later Middle English treatment implies significant quality differences between the short and long counterparts, and justifies the assumption that the vowel pairs were no longer paired strictly in terms of duration. In sum, although by early Middle English an increasing qualitative differentiation between the vowel pairs is evident, for Old English the default assumption is a contrast built primarily on duration. Phonologically, the difference can be represented in terms of mono- versus bimoraicity; that is, there are seven vowels, each participating in the suprasegmental length contrast (where μ represents a mora), as in (1): (1)

Vowel length contrast μ

μ

e

μ e

3.2 Diphthongs Classical Old English had only two diphthongs, usually written and , as in de¯op ‘deep’ and de¯aþ ‘death’. These falling or off-gliding diphthongs, like the vowels, carried a length contrast, as confirmed, for example, by their historical development and participation in the syllable weight conventions of OE poetic meter. The length difference was a property of the overall diphthong, not of the individual segments; that is, e͝o vs. e͞o, not e˘o vs. e¯o or eo˘ vs. eo¯ (following convention we continue to use e˘o and e¯o below). The quality of the off-glide is difficult to determine, but the practice of writing and in older manuscripts suggests an original high rounded [w]-like element, which was subject to further assimilatory and reductive changes over time (Lass 1994: 50). In fact, although not represented orthographically, the primary difference between the two diphthongs lay in the quality of the initial portion; that is, began with a mid e-type vowel and with a more open æ-type. This qualitative difference is suggested by the orthographic variation in older manuscripts in which

17 Old English: Phonology alternates with , and with both and . It is also suggested by the fact that some instances of eo and ea derive historically from earlier e and a, respectively, through sound changes. For example, in a change traditionally known as “breaking”, diphthongization of the vowel occurs when it is followed by specific consonants (l, r, h), as in e˘orþe (e˘o < e˘) ‘earth’ and eahta (e˘a < a) ‘eight’. Accordingly, the standardized use of (as opposed to a more phonologically accurate rendering, or ) can be considered an artifact of a scribal preference for avoiding the repetition of certain graph sequences. Most of the confident description presented in the previous paragraph can be gleaned from our earliest grammars and reference works (for example, Sievers 1898: 14; 1901: 194–195, 293; and Luick 1964a [1914–21]: 138). However, one seemingly innocuous aspect of the traditional view – namely, the assumption of a length contrast – has proved very troubling to succeeding generations of linguists and resulted in a massive amount of scholarly attention apparently disproportionate to the importance of two humble diphthongs destined to disappear without a trace in Middle English (Hogg 1992a: 104). In fact, however, the scholarly effort is justified in that the reconstruction of diphthongal length raises important theoretical and typological issues, given the apparent rarity – or absence according to White (2004) – of such a contrast in the languages of the world. From a uniformitarian perspective, the failure to find a parallel in any modern language would strongly suggest the inappropriateness of the reconstruction for an obsolescent language. Skepticism relating to phonological plausibility has been at the root of numerous attempts to revise the traditional analysis, beginning with Daunt (1939) whose interpretation is developed on the assumption that OE scribes adopted the Irish scribal practice of using vowels as diacritics to indicate specific qualities of an immediately following consonant; that is, the primary function of the second part of the short digraph was not to indicate diphthongal qualities, but rather the backness of the following consonant. In other words, the OE phonological system did not contain short diphthongs at all, and the digraphs and , when “short”, were simply an orthographic convention (the existence of the long diphthongs is not questioned). Intense debate on this topic continued over the decades – especially under the rubric of American structuralism; see, for example, Stockwell and Barritt 1955 and Hockett 1959 – but no consensus was ever reached. (For a good overview of the issues and literature, see Hogg 1992b: 16–24.) Part of the difficulty in reaching a consensus is that the plausibility issues are not as easily resolved as one might naively expect. The empirical waters are muddy in part because diphthongs are theory-dependent entities. For example, length contrasts involving apparent diphthongal pairs such as e˘w vs. e¯w are relatively common in languages. However, if for example e¯w is analyzed as a sequence of V̅ + C – that is, not as a “true” diphthong involving a complex nucleus – then the existence of such pairs becomes irrelevant to the OE situation. Furthermore, in cases where diphthongal length is assumed, it is often argued that the contrast rests on an opposition of long vs. overlong, rather than on short vs. long as assumed for Old English. For example, in a language such as standard Thai (which has been analyzed with a diphthongal length contrast) the short/long diphthongs are indeed phonetically longer than the simple short/long vowel pairs. At the same time, however, both vowels and diphthongs pattern in the same way in terms of the length contrast – that is, there is close to a 1:2 durational difference between short and long pairs regardless of whether simple vowels or diphthongs

259

260

III Old English are involved (Noss 1964: 15). Thus, given that a diphthongal length contrast is possible in principle, for a language such as Old English – with its strict syllable weight system – there would seem to be nothing implausible about pairs of diphthongs conforming phonologically to the short/long system, as opposed to building some other independent sub-system such as long vs. overlong. Although some scholars (for example, White 2004) have attempted to reject the possible existence of short diphthongs on theoretical grounds, the arguments are difficult to sustain. If the primary difference between a diphthong and a vowel is expressed in terms of a branching vs. non-branching nucleus, the default assumption is that diphthongs and vowels can, in principle, pattern phonologically in the same way as in (2), regardless of the phonetic details. This is the type of interpretation implied in the traditional descriptions of Old English (see also Lass 1994: 45–48), and its rejection as a theoretical possibility would have to derive from some – yet to be proposed – independently well-motivated principle. (2)

Vowel/diphthongal length contrast a. Short

b. Long

μ

μ

N

N

e

e

μ

μ

μ

N o

e

μ N

e

o

In sum, there is no doubt that the typological and theoretical issues relating to the Old English diphthongs are worthy of continued discussion, but in the meantime Hogg’s (1992b: 20) conclusion remains the most reasonable: “The evidence from both OE and ME suggests very strongly that the traditional position is in essence correct […]”.

3.3 Consonants With some exceptions, the general traits of the OE consonantal system are not particularly controversial. The basic inventory is provided in Table 17.2 (see Section 3.7 below for a set of phonological generalizations). In comparison with present-day standard varieties, there are differences in the inventory (for example, OE [c¸, x, ɣ]), suprasegmentals (almost all segments participate in a length contrast), and phonotactics (word edge clusters such as [kn-], [hl-], and [-mb], as in cne¯o ‘knee’, hlu¯d loud’, lamb ‘lamb’). One striking difference involves pairs of segments that are contrastive in modern varieties but only distributional variants in Old English (distributional pairs are boxed in Table 17.2). That is, the OE voiced fricatives [v, ð, z] occur only in a surrounding voiced environment – otherwise the voiceless counterparts occur – and [ŋ] occurs only preceding a velar. As is typically the case in writing systems, such phonetic detail was not usually indicated by the scribes, as in [f], [v] ‘wolf, wolves’. In the case of the velar voiceless fricative pair (both written as ), it is usually assumed that [c¸] occurred in a palatal environment ( ‘boy’), whereas [x] occurred elsewhere ( ‘brought’, ‘plough’). A reflection of this type of pattern can

17 Old English: Phonology

261

still be seen in modern standard German pairs such as ich [c¸] ‘I’ and acht [x] ‘eight’. Since the distribution of [h] (also written as ) was restricted (see Section 3.7, generalization l below), it is often treated as a distributional variant along with [c¸, x]. Further phonetic detail for individual segments beyond what is indicated in Table 17.2 can sometimes be reconstructed. For example, in the diphthongization process traditionally known as breaking (e > eo and a > ea triggered by specific post-vocalic consonants), the patterning of r with h [x] suggests a velarized variant of r in coda position, as in eorþe (< erþe) ‘earth’ and feohtan (< fehtan) ‘to fight’ (see Howell 1991). Particularly controversial areas of reconstruction involve the three segments in parentheses in Table 17.2; namely, [ ʃ ] (< [sk]), [tʃ ] (< [kj]), and [dʒ] (< [ɡj]). The difficulties lie not so much in the understanding of the general developments, but rather in attempting to determine a precise chronology of events. Table 17.2: Old English consonant system (late 10th century) Bilabial Labiodental Plosives

voiceless p voiced b

Fricatives

voiceless voiced

Nasals

t d f v

m

Inter- Alveolar Alveo- Palatal Velar Glottal dental palatal

θ ð

s z

(ʃ)

kj ɡj

k ɡ

c¸

x ɣ ŋ

n

Affricates

h

(tʃ ) (dʒ)

Liquids Lateral Central Approximants

l r j

w

In the case of [ ʃ ], there is no doubt of an assimilatory change involving coarticulatory effects by which early OE sk (usually written ) became [ ʃ ] by Middle English, a segment that is still maintained in Present-day English (ship < OE scip). The question that arises, though, is whether this change should be ascribed to the Old or early Middle English period. In fact, although [ ʃ ] is commonly reconstructed for Old English, the correctness of this assumption can be challenged in various ways. For example, in most OE poetry, as in Germanic generally, a cluster consisting of sp-, st-, or sk- can only self-alliterate. In late Old English, however, this strict system begins to break down, and there are cases of sp-, st-, and sk- alliterating with s-. Although the issues are complex, the alliteration of sk- with s- suggests that even in this later period we are still dealing with an sk- cluster, since alliteration of [ ʃ ] and [s] would be entirely unexpected (for detailed discussion, see Minkova 2003: 130–133). The evolution of the original voiceless velar plosive is a particularly problematic area. Again, from a bird’s eye diachronic perspective, the situation is relatively straightforward. We are dealing with phonological split arising from a very common type of sound change whose beginning and end points are clear; that is, a front vowel or j

262

III Old English environment results in a palatalized [kj], which ultimately evolves into an affricate, [tʃ ] (as in OE cild ‘child’ with original [k]). From a synchronic perspective, however, determining the chronology of the intermediate stages is a complex issue, and difficulties arise in reconstructing both the phonological status and the phonetic details of the evolving segments during the OE period. Although it is fair to say that the reconstruction of [tʃ ] already for early Old English has become a kind of standard theory (see, for example, Hogg 1992a: 95; Lass 2006: 54), in fact, robust evidence for the assumption of an affricate is strikingly absent. Indeed, the alliterative evidence once again suggests a more conservative progression of palatalization, since even in late OE verse, no distinction is made between k and the palatalized k for the purposes of alliteration, a practice that seems unlikely if palatalized k had already fully progressed to an affricate (for detailed discussion, see again Minkova 2003: 71–113, who argues further that the palatalized segment was not yet phonologized even in late Old English). The voiced velar obstruents also display a complex history, and the chronological and phonetic details are especially difficult due to the orthographic indeterminacy. The scribes did not consistently differentiate [ɡ], [ɣ], [ɡj], and [j], all of which could be written as (although the palatal quality could be indicated by adding an , as in [ɡj] ‘to singe’ and [j] ‘yoke’). In addition, the velar [ɡɡ] and palatal [ɡɡj] geminates could be written as or , as in ~ [ɡɡ] ‘dog’ and ~ [ɡɡj] ‘lay’. In a nutshell, early Old English displays the following phonological distribution of [ʝ], [ɣ], and [ɡ]: (a) the palatal fricative [ʝ] occurred in the front vowel environment (g(e)arn ‘yarn’, nægel ‘nail’), (b) the velar fricative [ɣ] occurred in a back vowel environment (ga¯st ‘spirit, ghost’, fugol ‘bird’), and (c) the plosive [ɡ] occurred only following a nasal and as a geminate (tunge ‘tongue’, dogga ‘dog’). By the classical OE period, the palatal fricative had merged with original [j], and [ɣ] had become [ɡ] word initially, so [ɣ] only occurred in the back vowel environment. Palatalized [ɡj] was found in the environment between a nasal and an original front vowel or j (seng(e)an < *sangjan ‘to singe’) or as the result of earlier West Germanic gemination (see OE lecgan vs. Go. lagjan ‘to lay’). This segment ultimately evolved into the affricate [dʒ] by Middle English, but for Old English an intermediate stage of palatalization can be assumed (represented here as [ɡj], in parallel with the case of [kj] discussed above). Even some modern dialects have forms such as brig ‘bridge’ and rig ‘ridge’ (OE brycg and hrycg), which suggest that in these dialects at least the affrication stage was never reached (Wright and Wright 1925: Section 319, note). In sum, the segments [ ʃ ], [tʃ ], and [dʒ] arguably do not belong to the OE inventory (hence their parenthetic status in Table 17.2), and in the case of the palatalized plosives it is plausible to assume an intermediate stage, which we represent here as [kj] and [ɡj].

3.4 Stress Although the ultimate details of OE stress are a complex matter, the location of primary word stress is relatively straightforward. In fact, for all words belonging to a major lexical category, except verbs, the primary generalization is simply: stress the initial syllable (regardless of whether it is a prefix or root syllable), as in a´nd-saca ‘adversary’, wæ´ ter ‘water’, ma´nigum ‘many-DAT.SG’, and also compounds such as bry¯´dguma ‘bridegroom’. At the same time, though, morphological factors come into play, and certain prefixes, such as ge- and be-, are never stressed, as in gewı´der ‘storm’ and bega´ng ‘practice’.

17 Old English: Phonology Verb stress, however, is somewhat more complicated. Prefixes are ignored for stress purposes – so the generalization here is: stress the initial root syllable of a verb, as in ofwu´ndrian ‘to be astonished’, on-sa´can ‘to dispute’ – unless the prefix is an adverbial or the verb derives from a noun carrying initial stress, in which case the stress falls on the prefix, as in ´ıncuman ‘to come in’, a´ndswarian ‘to answer’, cf. a´ndswaru ‘answer’. Although traditionally it is assumed that such verbal prefixes receive primary stress, Minkova (2003: 24–34) argues that stress on the prefix is subordinate to that of the root syllable. In fact, the topic of secondary stress in morphologically complex forms is a particularly difficult chapter of OE studies, especially since our primary source of evidence for stress patterns derives from the study of the complex interplay of stress and the metrical conventions of Old English poetry. For detailed discussion, see Lass 1994: 83–95 and Minkova 2006.

3.5 Unstressed syllables The segmental inventories presented in Section 3.2 and Section 3.3 above reflect the full set of contrasts found in the stressed environment. In many languages, unstressed syllables tolerate less complexity than stressed syllables. For example, unstressed syllable heads in Old English tend to be more restricted than stressed ones, so clusters such as kn- and hl- do not occur in the unstressed environment. In the case of vowels, unstressed syllables do not allow a vowel length contrast (although such a contrast can be reconstructed for pre-Old English; see Section 3.6 below). In addition, increasing vowel quality restrictions develop in unstressed syllables throughout the OE period. Although for classical Old English it is still possible to identify five vowels in unstressed syllables (that is, i, e, a, o, u; original æ had merged with e, and diphthongs are not possible), many mergers had in fact already taken place in specific environments. The general diachronic trend is for high vowels to lower to mid vowels, with local environmental factors facilitating or hindering the process. For example, -i becomes -e word finally, as in wine (< *wini) ‘friend’, but is preserved preceding palatal consonants and -ng, as in hefig ‘heavy’ ( = [j]) and cyning ‘king’. Similarly, -u generally becomes -o, although specific environments – for example, a following m – favor retention of u, as in heofon (< heofun) ‘heaven’ and fato (< fatu) ‘vats’, but sunum ‘son-DAT.PL’. Synchronically, given the relatively straightforward distribution of i ~ e, some argument can be made for treating them as distributional variants in unstressed syllables, but this is less likely in the case of u and o (see Hogg 1992a: 88). An important general issue here, however, is the possible effect of orthographic conservatism – particularly in the classical Old English manuscripts – since the written forms likely lag behind changes in pronunciation and do not reflect the full extent of phonological reduction (see Hogg 1992a: 121). Regardless, although once again it is difficult to determine with precision the Old English intermediate stages, the original vowel contrasts were neutralized in unstressed syllables by early Middle English, as reflected in the orthographic conflation of suffixal vowels to , as in name (OE nama) and tale (OE talu). Old English, like other early Germanic languages, allowed liquids and nasals to form the nucleus of an unstressed syllable, as in [hrı¯ðr̩ ] ‘head of cattle’ and [a¯dl̩] ‘disease’ (similarly, Go. akrs [akr̩ s] ‘acre’ and tagl [taɣl̩] ‘hair’). The diachronic source of such forms is found in vowel loss and, judging from OE poetry, the reduced forms were originally monosyllabic (that is, *akr < *akra- ‘acre’). However, they

263

264

III Old English regained a syllable probably first through the nuclearization of the liquid or nasal with later insertion of an anaptyctic vowel -i (later e) if the preceding vowel was front, and u (later o) if the preceding vowel was back. In fact, determining whether the nucleus of the unstressed syllable consisted of a sonorant or an anaptyctic vowel is difficult. Orthographic , for example, could be adopted by convention to represent nuclear l, just as writing might simply reflect a conservative use of orthography and not necessarily the absence of an anaptyctic vowel; thus [a¯tr̩ ~ a¯tor] ‘poison’, [næ¯dl̩] ~ [næ¯del] ‘needle’, [wæ¯pn̩ ~ wæ¯pen] ‘weapon’ (Campbell 1959: 151, Wright and Wright 1925: 100f). However, using the – not necessarily dependable – orthographic trends as a guide, it would appear that l, m, and n, which are often written in syllables without a vowel, were most likely to be nuclear. By contrast, r is seldom written alone, suggesting consistent anaptyxis. This would also reflect the pattern of earlier nuclearization, since the metrical evidence suggests that r was the first sonorant to nuclearize (Sievers 1893: Section 79,4a; Hogg 1992b: 237).

3.6 Quantity Old English, like the other early Germanic languages, is classified as a quantity language. Although all quantity languages – by definition – distinguish syllables in terms of weight, the phonological details of the weight contrast can vary from language to language. In fact, a common type of contrast is displayed in early Germanic: a codaless syllable with a short nucleus is light (monomoraic), all other syllables are heavy (bi- or polymoraic). This distinction is reflected in both sound change and poetic meter. For example, pre-Old English apocope of high vowels occurred after a heavy syllable but not after a light syllable, as in de¯or (< *de¯o.ru) ‘deer-PL’ but su˘nu (< *su˘.nu) ‘son’. In the case of Old English poetry, a property called resolution is displayed in which a sequence consisting of a light stressed syllable plus any immediately following syllable is treated as equivalent to a single heavy syllable, thus bisyllabic forms such as scipu ‘ships’ and werod ‘army’ can fill the same metrical position as heavy monosyllables such as wı¯f ‘woman’ and word ‘word’ (see, for example, Russom 1987: 12). In addition, quantity languages typically contrast both vowel and consonant length, but again there can be significant language-specific differences in the interplay of length, syllable structure, and stress. For example, a quantity language such as Finnish demonstrates complete independence of the three variables: the vowel length contrast is found in both stressed and unstressed syllables and in open and closed syllables, and geminate consonants can follow a long or short vowel, regardless of whether the vowel’s syllable is stressed or not (Becker 1998: 61–65). “Syllable-based quantity language” is the cover term for languages in which the weight contrast is found in both accented and unaccented syllables. Although Proto-Germanic – as well as the Indo-European classical languages such as Latin – belonged to this type, the attested early Germanic languages reflect a very strong tendency to begin restricting the full set of length contrasts and the weight contrast itself to stressed syllables only, leading to the designation “stress-based quantity language” (for full discussion, see Vennemann 1994, 1995). Late Old English, in its complete elimination of vowel and consonant length contrasts from unstressed syllables, had moved much further along this path than other Germanic languages such as Gothic and Old High German; for example, OE mihtı˘g, Go. mahteigs

17 Old English: Phonology ( = [ı¯]), OHG mahtı¯g ‘mighty’; OE sealfu˘de ~ sealfo˘de, Go. salbo¯da, OHG salbo¯ta ‘I anointed’ (see also Section 5.2 below). In sum, Old English can be described as a stress-based quantity language with morphologically-determined stress assignment (for further theoretical discussion, see also Dresher and Lahiri 1991 and Hayes 1995).

3.7 Phonological generalizations We present here a set of phonological generalizations that includes some of the main properties discussed above, as well as some other properties that distinguish Old English from present-day standard varieties. It is intended to be representative, not exhaustive. Below read “following” as “immediately following.” Standard orthographic forms are used (see Section 4). (For similar generalizations for Gothic, see Vennemann 1985.) a. Length contrasts occur only in a stressed environment: hw毴 las ‘whales’, 毴 las ‘eels’; -be´de ‘prayer-DAT’, be´dde ‘bed-DAT’. b. All segments except voiced fricatives, approximants, [ŋ], and [h] participate in the length contrast. c. Long consonants occur only in intersonorant environment following a stressed, short vowel: cynnes ‘kin-GEN’, bettra ‘better’. d. Long consonants are geminates; that is, they close and give weight to the preceding syllable; pyffan [pyf.fan] ‘puff’, dogga [doɡ.ɡa] ‘dog’. e. [v, ð, z] occur only in a surrounding voiced environment: seofon ‘seven’, wulfas ‘wolves’, lifde ‘s/he lived’. f. Non-geminate [f, θ, s] do not occur in a surrounding voiced environment (see e). g. Non-geminate [b] occurs only word-initially or following a nasal: blo¯d ‘blood’, climban ‘to climb’. h. Non-geminate [ɡ] occurs only following a nasal: singan ‘to sing’. In late Old English, it also occurs word initially; gu¯þ ‘combat, war’. i. [ɣ] does not occur following a nasal or a front vowel (compare fugol ‘bird’, swelgan ‘to swallow’, both with [ɣ]). In late Old English, it also does not occur word initially (see h). j. [j] does not occur following a back vowel (compare dæg˙ [j] ‘day’). k. Non-geminate [ɡj] occurs only following [n] and word finally: seng(e)an ‘singe’, ecg ‘edge’. l. [h] occurs only in word initial position or following certain prefixes such as be- and ge-: hof ‘enclosure, court’, behindan ‘behind’, geheald ‘keeping custody’. m. [c¸], [c¸c¸] occur only following a front vowel or diphthong, and non-geminate [c¸] occurs only in syllable coda: hliehhan ‘to laugh’, cnihtas ‘boys’, riht ‘right’. n. [x] does not occur word initially, and [x], [xx] do not occur following a front vowel or diphthong (see (m); compare pohha ‘pocket’, dohtor ‘daughter’, to¯h ‘tough’, sulh ‘plow’, all with [x]). o. [ŋ] occurs only preceding a velar: drincan ‘to drink’, singan ‘to sing’. p. The sequences below can form syllable heads (onsets) under stress: [k] or [h] plus one of [n, l, r, w]: cna¯wan ‘to know’, clæ¯ne ‘pure, clean’, cre¯da ‘belief, creed’, cwe¯n ‘woman, queen’; hnutu ‘nut’, hlu¯d ‘loud’, hring ‘ring, fetter’, hwæt ‘what’. [ɡ] plus one of [n, l, r]: gnagan ‘to gnaw’, glı¯dan ‘to glide’, gre¯tan ‘to greet’. [f] plus one of [n, l, r]: fne¯san ‘to sneeze’, flo¯d ‘flood’, friþa ‘protector’.

265

266

III Old English [w] plus one of [l, r]: wlitig ‘radiant, beautiful’, wræc ‘misery’. q. [mb], [nɡ] can form syllable codas: lamb ‘lamb’, lang ‘long’. r. Only vowels and diphthongs can form the nucleus of a stressed syllable. s. Only (short) vowels (specifically [i, e, a, o, u]), m, n, l, and probably r can form the nucleus of an unstressed syllable: ma¯þm̩ ~ ma¯þum ‘gift’, be¯acn̩ ~ be¯acen ‘sign, beacon’, segl̩ ~ segel ‘sail’, a¯tr̩ ~ a¯tor ‘poison’.

4 Phonological–orthographic correspondences As mentioned above, for the most part OE orthography is phonologically based and, in fact, most graphs can be roughly interpreted in terms of their equivalent IPA values; for example, , represent [æ], [y], respectively, as in fæder ‘father’ and hyll ‘hill’. At the same time, a consistent indication of vowel and diphthongal length is notably absent in the manuscripts, so , for example, represents both the short and long vowel; [ky̆ niŋɡ] ‘king’, [bry¯d] ‘bride’. In addition, orthographic geminates are often maintained word finally, although phonological geminates are not found in this position, as in bedd [bed], bedde [bedde] ‘bed-NOM/DAT’. Some primary conventions that are not transparent or self-resolving are listed below. Although we abstract away from the significant variation evident in the manuscripts, the correspondences given in Table 17.3 provide a reasonable reflection of classic OE practices. (We include [ø], although in fact it underwent early derounding in West Saxon; see Section 3.1 above.) Table 17.3: Old English phonological–orthographical correspondences Phonology

Orthography

Example

Phonological Form

Gloss

(a)

[f] ~ [v]

(b)

[ff] [θ] ~ [ð]

~

(c)

[θθ] [s] ~ [z]

~

(d)

[ss] [k]

, also

(e)

[kj] (> [tʃ ])

, also

(f)

[ɡ] ~ [ɣ]

wulf wulfe pyffan þencan ~ ðencan broþor ~ broðor moððe ~ moþþe sæ nosu is cyssan cræft weorc kyning cinn bec þeccean ʒast brinʒan laʒu

wulf wulve pyffan θeŋkjan bro¯ðor moθθe sæ¯ nozu ı¯s kyssan kræft weork kyniŋɡ kjinn be¯kj θekkjan ɡa¯st briŋɡan laɣu

‘wolf’ ‘wolf-DAT’ ‘puff’ ‘to think’ ‘brother’ ‘moth’ ‘sea’ ‘nose’ ‘ice’ ‘to kiss’ ‘skill’ ‘work’ ‘king’ ‘chin’ ‘books’ ‘to cover ‘spirit’ ‘to bring’ ‘law’ (Continued )

17 Old English: Phonology

267

Table 17.3: Continued Phonology

Orthography

(g)

[j]

(h) (i) ( j) (k)

[ɡj] (> [dʒ]) [ɡɡj] (> [(d)dʒ]) [ɡɡ] [h], [x] ~ [c¸]

(l)

[n] ~ [ŋ]

(m)

[w]

(n)

[æ¯̆ ]

(o)

[y¯̆ ]

(p)

[ø̄̆ ]

(q)

[æ¯̆ a̯ ]

(r)

[e¯̆ o̯ ]

(s)

[ks]

ʒift ʒeong dæʒ , also brycʒ ~ lecʒan ~ leʒʒan ~ froʒʒa ~ frocʒa

hamor behindan dohtor crohha flyht

hnutu spinnan tunʒol

ƿolcen saƿol

fæstan sæd

þyncan yþ

oele cwoen

eall deaþ

ʒeolo deop

axian , also

Example

Phonological Form

Gloss

jift joŋɡ dæj briɡj leɡɡjan froɡɡa hamor behindan doxtor kroxxa flyc¸t hnutu spinnan tuŋɡol wolkn̩ sā wol fæ̆ stan sæ¯d θy̆ ŋkjan y¯θ ø̆ le kwø̄ n æ̆ a̯ l dæ¯a̯ θ je˘o̯ lo de¯o̯ p a¯ksian

‘marriage gift’ ‘young’ ‘day’ ‘bridge’ ‘to lay’ ‘frog’ ‘hammer’ ‘behind’ ‘daughter’ ‘crock pot’ ‘flight’ ‘nut’ ‘to spin’ ‘star’ ‘cloud’ ‘soul’ ‘to fast’ ‘seed’ ‘to seem’ ‘wave’ ‘oil’ ‘queen’ ‘all’ ‘death’ ‘yellow’ ‘deep’ ‘to ask’

The scribes typically did not make any phonological distinction in their use of and (Hogg 1992b: 33–34). The graph was occasionally used instead of , especially preceding , as in kyning ‘king’. If the sound change *sk > [ ʃ ] is posited for Old English (see Section 4.3 above), then [ ʃ ] is represented by . By the time of Middle English, [ ʃ ] was usually written or . For the diphthongs note that we give here a literal transcription ([eo̯ and [æa̯ ]), although various off-gliding values can be assumed (see Section 3.2). Old English texts are subjected to modern editorial conventions. For example, although the symbol was used by OE scribes, most modern works transliterate using . Note, however, that by early Middle English and were used contrastively for [j] and [ɡ], respectively. Similarly, the symbols (for [kj], later [tʃ]), (for [j]), and / (for [(ɡ)ɡj], later [dʒ]), are not found in the manuscripts; they are used by modern editors to distinguish the palatal and velar counterparts; see (d)–(j) above. Finally, although both and were used in early texts, in fact [w] was typically represented by runic wynn, ƿ. In modern works, the rune symbol is invariably transliterated as .

268

III Old English

5 Diachrony 5.1 The “age of harmony”: Umlaut No outline of Old English phonology would be complete without mention of what Lass (1994: 59) labels “the age of harmony”, which subsumes the various types of umlaut (vowel harmony changes) that took place from the Proto-Germanic to early Old English periods. Germanic umlaut is a type of partial regressive assimilation in which the target vowel – typically the stressed vowel – takes on qualities of a following trigger vowel. Umlaut is pervasive in all the early Germanic languages with the exception of Gothic, and its effects include all the logical possibilities of lowering, backing, fronting, and raising. Lowering of u to o, for example, is evident in West Germanic a-umlaut, as in OE gold (< *gulda) ‘gold’. Backing of *æ to a in the environment of a following back vowel is found in later pre-Old English in a change that is often called Restoration, since preOE æ, which arose through an earlier general fronting of original a, was “restored” to a under back umlaut conditions, as reflected, for example, in paradigmatic allomorphy of the type dæg, dagas ‘day, days’ and fæt, fatu ‘vat, vats’. Back umlaut can also produce diphthongs (assuming the existence of short diphthongs; see Section 3.2 above), although this change had only limited effect in West Saxon; eofor (< efor) ‘boar’, heorut (< herut) ‘hart’. In Old English, the most generalized and systematic subtype is i-umlaut, which involves primarily the fronting of vowels and diphthongs under the influence of a following i or j, although raising in the case of the short low vowel can also occur (see Table 17.4). (Note that e had already been raised to i in earlier Germanic. Also, umlauted a˘ preceding a nasal is found as æ̆ in early texts, and then usually e˘ in later ones.) Table 17.4: Old English i-umlaut (vowels)

æ̆ > e˘ ă > æ̆ > e˘ a¯ > æ¯ o˘ > ø̆ (> e˘) o¯ > ø̄ (> e¯) u˘ > y̆ u¯ > y¯

i

y˘¯

e˘

ø˘¯

u˘¯ o˘¯

˘ æ

a˘

æ¯

a¯

bedd senda dæ¯lan dehter se¯can cynn dy¯stig

(Go. badi) (Go. sandjan) (*da¯ljan) (*dohtri) (Go. so¯kjan) (Go. kuni) (compare du¯st ‘dust’)

‘bed’ ‘to send’ ‘to divide’ ‘daughter-DAT’ ‘to seek’ ‘race, generation’ ‘dusty’

In general, i-umlaut in Old English does not display the complexity it does in the other Germanic languages. In Old High German, for example, the intervening consonantal environment plays a significant role in facilitating or hindering umlaut, and in Old Norse there is a complex interaction between umlaut and other sound changes, especially i-syncope (Howell and Salmons 1997; Iverson and Salmons 2004). By contrast, the West Saxon situation reflects a relatively straightforward, highly generalized application in which long and

17 Old English: Phonology short u, o, and a regularly undergo umlaut. At the same time, though, there are two primary restrictions. First, raising umlaut of short æ (> e) can be blocked by (non-geminate) clusters; for example, umlaut is found in hebban ‘to raise’ but not in fæstan ‘to make firm’ (both with an original *-jan suffix). Second, long æ¯ resists umlaut altogether, as in læ¯ce ‘physician’, where æ¯ remains in spite of the original *-ja- (> e) suffix. In fact, these restrictions constitute a pattern in conformance with other Germanic languages and general principles governing umlaut. Howell and Salmons (1997: 89) show that the propensity to undergo umlaut increases in accordance with the degree of qualitative difference between trigger and target vowels, which in the case of Old English relates primarily to the back–front dimension. Thus, although fronting of the back vowels (u, o, a) is regular, short æ, which is already front, can resist raising umlaut in the cluster environment. Similarly, the even greater resistance to umlaut displayed by the long vowel æ¯ is typical, as umlaut preferentially affects short vowels. For example, in Old High German, short a is most susceptible to umlaut (known as primary umlaut), and in Dutch only short vowels undergo umlaut. These assimilatory changes had a significant impact on OE phonology and morphology. At first, the umlaut vowels [ø] and [y] were only distributional variants of o and u occurring under specific conditions, but already in pre-Old English they were phonologized, yielding two new segments. Although ø underwent early derounding in West Saxon, y was relatively stable. Along with this phonologization came a dramatic increase in allomorphy, which is evident throughout the lexicon; for example hnutu, hnyte ‘nut-NOM, nut-DAT’, ic do¯, he de¯þ ‘I do, he does, bra¯d, bræ¯dra ‘broad, broader’. Although most of the allomorphy was leveled out in later stages, some traces remain even into Present-day English. Modern pairs displaying lexical split such as older, elder and brothers, brethren have developed in accordance with Kuryłowicz’s (1947) fourth “law” governing leveling; that is, the original form (umlauted elder, brethren) took on a specialized meaning, while the new form carries the primary meaning. Finally, i-umlaut was morphologized as a plural marker, although only a handful of these umlaut plurals remain in present day English, as in man, men; foot, feet; and mouse, mice (compare OE a¯c, æ¯c ‘oak, oaks’; bo¯c, be¯c ‘book, books’, and cu¯, cy¯ ‘cow, cows’).

5.2 Quantity changes Both the pre-Old English and late Old English/early Middle English stages were robust periods of quantity change. The early Germanic and pre-Old English changes primarily affected unstressed syllables, while stressed syllables were targeted in the later changes. In comparison with these early and late stages, the OE period itself was relatively stable, especially with regard to the stressed syllable. The pre-Old English period involved a severe reduction of unstressed syllables, including various kinds of vowel and consonant deletions and loss of vowel length, as in bend (Ø < -i; compare Go. bandi ‘band, ribbon’); hand (Ø < -uz; Go. handus ‘hand’), and sealfude (u˘ < o¯; Go. salbo¯da ‘he anoints’). Traditionally, this reduction is commonly linked to the shift from pitch to stress accent and the fixing of stress on the root syllable that occurred in Proto-Germanic. In addition, although difficult to quantify, it is also sometimes claimed that the intensity of the stress accent gradually increased throughout the history of English with concomitant weakening of unstressed syllables. (See Luick 1914–21: 267–362; Lass 1994: 95–102; Lutz 1991: 281–282.) In fact, the amount of diachronic reduction that unstressed syllables undergo can vary significantly from one language to the next. For example,

269

270

III Old English Finnish (with its initial stress accent) is extremely conservative in maintaining contrasts in unstressed syllables, a fact strikingly apparent in its treatment of very early borrowings from Germanic; for example, Finnish kuningas, PGrmc. *kuningaz ‘king-NOM’. (For further discussion, see van Coetsem et al. 1981; Salmons 1992: 166–168; Boutkan 1995.) The late Old English and Middle English changes are particularly significant, since they involve the complete breakdown of the original quantity system, an event that occurred in almost all Germanic dialects at one point or another during the medieval period. Three of the main quantity changes assumed are Closed Syllable Shortening (CSS), which eliminated the vowel length contrast in closed disyllables (ke¯pte > ke˘pte ‘kept’); Degemination, which eliminated the contrast between short and long consonants (OE æppel [pp], ME apel [p], where usually an ambisyllabic consonant is assumed for the latter); and Open Syllable Lengthening (OSL), which eliminated the vowel length contrast in open syllables (OE na˘ma, ME na¯me ‘name’). Although these changes are properly ascribed to the late Old English (CSS) and Middle English periods (Degemination, OSL), Luick’s (1898, 1964a [1914–21]) view that they represent a continuation of a process that began in preOld English, and even in West Germanic, has been extremely influential (see, for example, Ritt 1994). According to Luick (1898), there was a rhythmic tendency operative throughout early English in which syllable weight was gradually being standardized according to a set of prosodic weight templates. For example, in the case of disyllables, the ideal stressed syllable was assumed to be bimoraic. Thus, CSS (through shortening and loss of a mora) and OSL (through lengthening and addition of a mora) yielded this ideal type (ke˘p.te, na¯.me), while syllables already bimoraic remained unchanged. In fact, though, for the Old English period, evidence for a standardization of quantity is not particularly robust, and in the late Old English period, at least one important change moved the stressed syllable away from the bimoraic ideal. In the change known as Homorganic Cluster Lengthening (see Section 3.1), in which vowels lengthened before clusters such as -ld, -mb, and -nd, an already bimoraic syllable became overlong, as in ME clı¯m.ben (< OE clı˘m.ban) ‘to climb’. Further, even in the case of Middle English, Degemination does not conform to the bimoraic preference, since the stressed syllable is already bimoraic at the pre-Degemination stage, as in OE æp.pel. For detailed discussion and an alternative interpretation, see Murray 2000 (and references there) and Mailhammer 2007.

6 References Becker, Thomas. 1998. Das Vokalsystem der deutschen Standardsprache. Frankfurt: Peter Lang. Brunner, Karl. 1965. Altenglische Grammatik. Nach der angelsa¨chsischen Grammatik von Eduard Sievers. 3rd edn. Tu¨bingen: Max Niemeyer. Boutkan, Dirk. 1995. The Germanic “Auslautgesetze”. Amsterdam: Rodopi. Bu¨lbring, Karl D. 1902. Altenglisches Elementarbuch. Vol. I: Lautlehre. Heidelberg: Winter. Campbell, Alistair. 1959. Old English Grammar. Oxford: Oxford University Press. Daunt, Marjorie. 1939. Old English sound changes reconsidered in relation to scribal tradition and practice. Transactions of the Philological Society 1939 38: 108–137. Dresher, B. Elan and Aditi Lahiri. 1991. The Germanic foot: Metrical coherence in Germanic. Linguistic Inquiry 22(2): 251–286. Fox, Anthony. 1995. Linguistic Reconstruction: An Introduction to Theory and Method. Oxford: Oxford University Press. Hayes, Bruce. 1995. Metrical Stress Theory: Principles and Case Studies. Chicago: University of Chicago Press.

17 Old English: Phonology Hockett, Charles F. 1959. The stressed syllabics of Old English. Language 35: 575–597. Hogg, Richard M. 1992a. Phonology and morphology. In: Richard M. Hogg (ed.), The Cambridge History of the English Language. Vol. I. The Beginning to 1066, 67–167. Cambridge: Cambridge University Press. Hogg, Richard M. 1992b. A Grammar of Old English. Oxford: Blackwell. Hogg, Richard M. 2006. Old English Dialectology. In: van Kemenade and Los (eds.), 395–416. Howell, Robert B. 1991. Old English Breaking and its Germanic Analogues. Tu¨bingen: Max Niemeyer. Howell, Robert B. and Joseph C Salmons. 1997. Umlautless residues in Germanic. American Journal of Germanic Linguistics 9: 83–111. Iverson, Gregory K. and Joseph C Salmons. 2004. The conundrum of Old Norse umlaut: Sound change versus crisis analogy. Journal of Germanic Linguistics 16: 77–110. Kemenade, Ans van and Bettelou Los (eds.). 2006. The Handbook of the History of English. Oxford: Blackwell. Kuryłowicz, Jerzy. 1947. La nature des proce`s dits analogiques. Acta Linguistica 5: 15–37. (Reprinted in Eric P. Hamp, Fred W. Householder, and Robert Austerlitz (eds.), Readings in Linguistics 2, 158–174. Chicago: Chicago University Press, 1966.) Lass, Roger. 1992. Phonology and morphology. In: Norman Blake (ed.), The Cambridge History of the English Language. Vol. II. 1066–1476, 23–155. Cambridge: Cambridge University Press. Lass, Roger. 1994. Old English: A Historical Linguistic Companion. Cambridge: Cambridge University Press. Lass, Roger. 2006. Phonology and morphology. In: Richard Hogg and David Denison, A History of the English Language, 43–108. Cambridge: Cambridge University Press. Lass, Roger and John M. Anderson. 1975. Old English Phonology. Cambridge: Cambridge University Press. Luick, Karl. 1898. Beitra¨ge zur englischen Grammatik III: Die Quantita¨tsvera¨nderungen im Laufe der englischen Sprachentwicklung. Anglia 20: 335–362. Luick, Karl. 1964a [1914–1921]. Historische Grammatik der englischen Sprache, Vol. 1, Part 1. Stuttgart: Bernhard Tauchnitz. Luick, Karl. 1964b [1929–1940]. Historische Grammatik der englischen Sprache, Vol. 1, Part 2. Stuttgart: Bernhard Tauchnitz. Lutz, Angelika. 1991. Phonotaktisch gesteuerte Konsonantenvera¨nderungen in der Geschichte des Englischen. Tu¨bingen: Max Niemeyer. Maddieson, Ian. 1984. Patterns of Sounds. Cambridge: Cambridge University Press. Mailhammer, Robert. 2007. On syllable cut in the Orrmulum. In: Christopher M. Cain and Geoffrey Russom (eds.), Studies in the History of the English Language III – Managing Chaos: Strategies for Identifying Change in English, 37–61. Berlin/New York: Mouton de Gruyter. Minkova, Donka. 2003. Alliteration and Sound Change in Early English. Cambridge: Cambridge University Press. Minkova, Donka. 2006. Old and Middle English prosody. In: van Kemenade and Los (eds.), 95–124. Minkova, Donka and Robert P. Stockwell. 1992. Homorganic clusters as moric busters in the history of English: The case of -ld, -nd, -mb. In: Matti Rissanen, Ossi Ihalainen, Terttu Nevalainen and Irma Taavitsainen (eds.), History of Englishes: New Methods and Interpretations in Historical Linguistics, 191–207. Berlin/New York: Mouton de Gruyter. Murray, Robert W. 2000. Syllable cut prosody in Early Middle English. Language 76: 617–681. Noss, Richard B. 1964. Thai Reference Grammar. Washington, DC: Foreign Service Institute. Pyles, Thomas and John Algeo. 1993. The Origins and Development of the English Language. 4th edn. Orlando, FL: Harcourt Brace Jovanovich. Ritt, Nikolaus. 1994. Quantity Adjustment: Vowel Lengthening and Shortening in Early Middle English. Cambridge: Cambridge University Press. Rosner, B. S. and J. B. Pickering. 1994. Vowel Perception and Production. Oxford: Oxford University Press.

271

272

III Old English Russom, Geoffrey. 1987. Old English Meter and Linguistic Theory. Cambridge: Cambridge University Press. Salmons, Joseph. 1992. Accentual Change and Language Contact: Comparative Survey and Case Study of Early Northern Europe. Stanford: Stanford University Press. Sievers, Eduard. 1893. Altgermanische Metrik. Halle: Max Niemeyer. Sievers, Eduard. 1898 [1882]. Angelsa¨chsische Grammatik. Halle: Max Niemeyer. Sievers, Eduard. 1901. Grundzu¨ge der Phonetik zur Einfu¨hrung in das Studium der Lautlehre der indogermanischen Sprachen. 5th edn. Leipzig: Breitkopf & Ha¨rtel. Stockwell, Robert P. 1952. Chaucerian Graphemics and Phonemics: A Study in Historical Methodology. Unpublished doctoral dissertation. University of Virginia, Charlottesville. Stockwell, Robert P. and C. Westbrook Barritt. 1951. Some Old English Graphemic-Phonemic Correspondences - ae, ea and a. (Studies in Linguistics, Occasional Papers, 4.) Norman, OK: Battenburg Press. Stockwell, Robert P. and C. Westbrook Barritt. 1955. The Old English short digraphs: Some considerations. Language 31: 373–389. Stockwell, Robert P. and Donka Minkova. 1990. The Early Modern English vowels, more o’ Lass. Diachronica 7: 199–214. Sweet, Henry. 1888. History of English Sounds. 2nd edn. Oxford: Clarendon Press. van Coetsem, Ronald, Hendricks and Susan McCormick. 1981. Accent typology and sound change. Lingua 53: 295–315. Vennemann, Theo. 1985. Phonologically conditioned morphological change: Exceptions to Sievers’ Law in Gothic. In: Edmund Gussman (ed.), Phonomorphology: Studies in the Interaction of Phonology and Morphology, 193–219. Lublin: Katholische Universita¨t Lublin. Vennemann, Theo. 1994. Universelle Nuklearphonologie mit epipha¨nomenaler Silbenstruktur. In: Karl Heinz Ramers, Heinz Vater, and Henning Wode (eds.), Universale phonologische Strukturen und Prozesse, 7–54. Tu¨bingen: Max Niemeyer. Vennemann, Theo. 1995. Der Zusammenbruch der Quantita¨t im Spa¨tmittelalter und sein Einfluß auf die Metrik. Quantita¨tsproblematik und Metrik. Amsterdamer Beitra¨ge zur a¨lteren Germanistik (special volume ed. by Hans Fix) 42: 185–223. White, David L. 2004. Why we should not believe in short diphthongs. In: Anne Curzan and Kimberly Emmons (eds.), Studies in the History of the English Language II. Unfolding Conversations, 57–84. Berlin/New York: Mouton de Gruyter. Wright, Joseph and Elizabeth Mary Wright. 1925. Old English Grammar. Oxford: Oxford University Press.

Robert Murray, Calgary (Canada)

18 Old English: Morphology 1 Inflectional morphology of the noun phrase 2 Inflectional morphology of verbs 3 References

Abstract Old English is in many respects a typical Indo-European language. This is particularly true of its morphological categories and its complex inflectional systems. It is mainly due to this complexity that this chapter cannot treat all aspects of OE Alex Bergs and Laurel Brinton 2012, Historical Linguistics of English (HSK 34.1), de Gruyter, 272–293

18 Old English: Morphology morphology in full detail. It therefore focuses on the most important inflectional systems of Old English. Morphological word-formation patterns are necessarily treated only marginally (Kastovsky, Chapter 9). Moreover, there is a considerable degree of dialectal variation in Old English which is also manifested in the morphological paradigms. This variation cannot be covered here comprehensively. This article therefore has a strong bias towards the later stages of the West Saxon variety – the dialect and period from which the greatest share of our extant sources is transmitted. For more comprehensive accounts, including the details of the diachronic and diatopic variation, I refer the reader to the relevant sections in Hogg and Fulk (2011) as well as to the older, but still valuable works by Campbell (1959) and Brunner (1965).

1 Inflectional morphology of the noun phrase 1.1 The inflectional categories of the noun phrase Case/number is marked by inflection (as opposed to agglutination), that is, there is no distinctive marker encoding a value of only one of the two categories. Case/number is generally marked on any element of the noun phrase, i.e. on all modifiers and on the head noun. The noun paradigms are the least distinctive of all, so that in many instances only the case marking on the adjective and/or determiner can unambiguously indicate the case/number value of the entire noun phrase. Old English (OE) has a typically Indo-European (IE) case system. Therefore any general descriptions of the functions of IE cases can well be applied to Old English. Of the eight IE cases, five survived: nominative, accusative, dative, genitive, and instrumental. The former three are used predominantly to mark grammatical relations. The genitive case marks a noun phrase as a modifier of a superordinate noun phrase. The instrumental is formally distinguished from the dative only in some adjectival and pronominal paradigms. If it is encoded, it marks complements that take the semantic role of an Instrument. For good descriptions of the functions of the five OE cases cf. Blake (2001: Section 2.3) or, specifically for Old English, Mitchell (1985: I, Sections 1240–1427). There are two number values for English nouns: singular and plural. A third number value, the dual, can be encoded in some pronominal paradigms. There are three gender values, all inherited from Proto-Indo-European, masculine, feminine, and neuter. The gender value of a noun determines agreement patterns on pronouns and on dependent adjectives and determiners so that, while inherent in the lexical entry of a noun, gender is morphologically encoded on any noun modifier. As in Indo-European languages in general, there is no distribution along the lines of natural gender. Particularly expressions denoting female persons can have masculine (wı¯fman ‘woman’) or neuter (wı¯f ‘woman’) gender. Expressions for inanimate referents do not necessarily have neuter gender. Only some expressions that may refer to either a female or a male person, such as e.g. names of occupations, can have both a masculine and a feminine form, which is then used to semantically distinguish between a female and a male member of that class (e.g. munuc ‘monk’ vs. mynecen ‘nun’).

273

274

III Old English

1.2 The inflectional paradigms of nouns In every noun class there is some degree of syncretism of forms. As in all IE languages, the following values are not distinguished formally in any noun class of Old English: nominative and accusative of neuter nouns for both number values, and the genitive plural and the dative plural of all three genders. The traditional labels of the major noun classes refer to the thematic morpheme reconstructed for Proto-Indo-European, which was inserted between the lexical root and the inflectional suffixes. Although the thematic elements “are not synchronically transparent and reflect the product of historical reconstructions” (Blake 2001: 4; cf. also Lass 1994: 123), they are still in common use in the description of Indo-European daughter languages. There is a major division into vocalic classes (classes 1 to 4) and consonantal classes (classes 5 and 6) according to whether the thematic element was a vowel or a consonant. Outside the vocalic/consonantal distinctions, there are other classes that originally did not have a thematic element and are therefore called “athematic” nouns. None of these terms refer to properties that are synchronically transparent in Old English. In the following, I simply specify the noun classes by numbers from 1 to 8; but I add the traditional labels in the section headings.

1.2.1 Class 1 – Germanic a-stems, Indo-European o-stems The nouns of class 1 are all either masculine or neuter. There are three main paradigms of this class, according to whether the noun is masculine (class 1a), neuter with a short root syllable (class 1b), or neuter with long root syllables, i.e., either with a long root vowel or with a V(ː)C-cluster in the nucleus of the root (class 1c); cf. Tables 18.1–18.3. Phonological cross-influences created a number of variant paradigms, an overview of which will be presented in the following. Table 18.1: Class 1a (sta¯n ‘stone’)

NOM ACC DAT GEN

SG

PL

sta¯n-∅ sta¯n-∅ sta¯n-e sta¯n-es

sta¯n-as sta¯n-as sta¯n-um sta¯n-a

Table 18.2: Class 1b (scip ‘ship’)

NOM ACC DAT GEN

SG

PL

scip-∅ scip-∅ scip-e scip-es

scip-u scip-u scip-um scip-a

18 Old English: Morphology

275

Table 18.3: Class 1c (word ‘word’)

NOM ACC DAT GEN

SG

PL

word-∅ word-∅ word-e word-es

word-∅ word-∅ word-um word-a

Apophonic variation: Because /ɛ/ was lowered to /a/ before palatal vowels in pre-Old English, nouns with the root vowel /ɛ/ show an apophonic alternation between the singular (uninflected or with suffixes containing a back vowel) and the plural forms (with suffixes all containing a front vowel). This applies to 1a and 1b nouns, so that there is a root variation as in dæg- ‘day-SG’, fæt- ‘vessel-SG’ vs. dag- ‘day-PL’, fat- ‘vessel-PL’. Root-final /-χ/: If followed by a vowel, root-final /-χ/ was lost. It was retained only in the uninflected forms, in which it was represented as 〈-h〉. In the relevant nouns we find, mearh-∅ ‘horse-NOM/ACC.SG’, but mear- for all other values. Syncope and epenthesis: Two types of paradigm-internal variation may look synchronically like one and the same phenomenon, but result from two opposing processes: there is both a syncope of an unstressed vowel in a bisyllabic root, e.g. engel-∅ ‘angel-NOM/ACC.SG’ vs. engl-es ‘angel-GEN.SG’, and a vowel epenthesis in a cluster consisting of a stop and a liquid, e.g. fugol-∅ ‘bird-NOM/ACC.SG’ formed from fugl- ‘bird’. The Germanic (Grmc.) ja-stems: This variant is considered a distinct subclass in Proto-Germanic. Originally, the thematic element preceding the inflectional suffixes was */-ja-/. In many forms originally belonging to this subclass, there are no differences from the default paradigms of classes 1a–1c in Tables 18.1–18.3. In some cases, however, the glide may still be represented as a root-final /-j-/ (usually spelled 〈-g-〉), or as /-ə/ following the root. Such traces can be found, for example, in her-e-∅ ‘army-THM-NOM/ACC.SG’ vs. her-g- (/ˈher-j-/) for all other values. (THM=thematic element.) The Grmc. wa-stems: Similar to the previous group there are some nouns descending from the PGrmc. sub-class with */-wa-/ as thematic element. It has been retained in some OE nouns as a root-final /-u/ in the uninflected forms and as a glide between root and suffix in the inflected forms. This is the case, for instance, in bear-u-∅ ‘grove-THM-NOM/ACC.SG’ vs. bear-w- for all other values.

1.2.2 Class 2 – Germanic o¯-stems, Indo-European a¯-stems The nouns of this class are all feminine. The nominative/accusative plural forms differ in West-Saxon from the other regional varieties of Old English. As in the neuter nouns of class 1, there is a distinction between nouns with short syllables (class 2a) and with long syllables (class 2b). In class 2a, later West Saxon (WSax.) texts show the genitive plural ending -ena for -a. Cf. Tables 18.4 and 18.5. Phonological crossinfluences are less numerous than those of class 1. The most important ones are briefly presented below:

276

III Old English Table 18.4: Class 2a (talu ‘tale’)

NOM ACC DAT GEN

SG

PL

tal-u tal-e tal-e tal-e

tal-a (WSax.), tal-e (non-WSax.) tal-a (WSax.), tal-e (non-WSax.) tal-um tal-a, tal-ena (late WSax.)

Table 18.5: Class 2b (wund ‘wound’)

NOM ACC DAT GEN

SG

PL

wund wund-e wund-e wund-e

wund-a (WSax.), wund-e (non-WSax.) wund-a (WSax.), wund-e (non-WSax.) wund-um wund-a

The Grmc. jo¯-stems: As in class 1, there is a subclass postulated for those nouns of class 2, in which the thematic element was */-joː-/. The glide is usually no longer present, but it caused the nominative singular suffix -u to be reduced to /-ə/ and to be dropped subsequently. Otherwise the paradigm does not differ from the default ones shown above in Tables 18.4 and 18.5. The Grmc. wo¯-stems: The thematic morpheme */-woː-/ is usually retained in inflected forms as /-u/ after short root syllables and before inflectional suffixes, for instance in sin-u-∅ ‘sinew-THM-NOM.SG’. In forms with an inflectional suffix it remained a glide, e.g. sin-w- ‘sinew-THM-’. After long root syllables the thematic element was dropped.

1.2.3 Class 3 – Germanic i-stems The forms of class 3 merged to a large extent with those of class 1a, class 2, and classes 1b or 1c, depending, respectively, on whether they are masculine, feminine, or neuter. The original thematic vowel of class 3, */-i-/, is no longer retained in most OE forms. Only in the nominative and accusative singular forms of masculine and neuter nouns with short syllables does final /-ə/ attest to the former i-suffix, e.g. win-e ‘friend-NOM/ACC.SG(MASC)’, sper-e ‘spear-NOM/ACC.SG(NEUT)’. However, the suffix left a trace in all members of this class as it caused i-umlaut in the root vowel. So, in spite of the fact that a number of inflectional suffixes were transferred from other noun classes, the umlauted root vowel constitutes the major formal difference between class 3 and classes 1 and 2. Yet, since there is no alternation of the root vowels within the paradigms of any subgroup of class 3, this is basically a diachronic feature and has no significance synchronically. Where there are original i-stem suffixes, they usually occur in doublets with corresponding forms of classes 1 and 2. For instance, in masculine nouns with short syllables inherited suffixes still compete with those of class 1. In the nominative and accusative plural we find both a reflex of the original -i (reduced to /-ə/) and the class 1 suffix -as. In the genitive plural, the original suffix /-ija/ (usually spelled 〈-iga〉) is still attested in poetry. This results in forms like win-e next to win-as ‘friend-NOM/ACC.PL’, win-iga (poetic) next to win-a ‘friend-GEN.PL’.

18 Old English: Morphology

277

In the plural forms of masculine nouns with long syllable the pattern is generally that of class 1a. Only exceptionally do we find suffixes of the original class 3 in some tribal names as e.g. Engl-e ‘Angle-NOM/ACC.PL’. In the paradigm of feminine nouns of class 3 with long root syllables, original forms in /-ə/ (< /-i/) and those of class 2a compete only in the accusative singular, so that both dæ¯d-∅ and dæ¯d-e ‘deed-ACC.SG’ can occur. All other forms correspond with those of class 2a. The neuter i-stems were influenced less strongly by other paradigms. The original suffix for the nominative and accusative singular, /-ə/ (< */-i/), has been retained in nouns with short root syllables and was dropped in nouns with long root syllables. The same holds for the nominative/accusative plural suffix /-u/. This results in forms like sper-e ‘spear-NOM/ACC.SG’ and sper-u ‘spear-NOM/ACC.PL (short root) vs. flæ¯sc-∅ ‘flesh-NOM/ACC.SG/PL’ (long root).

1.2.4 Class 4 – Germanic u-stems The nouns of class 4 were either masculine or feminine (but cf. Brunner 1965: Section 275; Hogg and Fulk 2011: Section 2.71 for neuter relics). There is no formal distinction between the two gender values, but there is again a distinction between short and long root syllables. In the former the old stem syllable in the nominative/accusative singular has been retained (cf. Table 18.6), whereas the suffixes for the same values were dropped in the latter (hand-∅ ‘hand-NOM/ACC.SG’). As with most OE nouns, there is some conflation with classes 1 and 2. However the pattern of class 4 has been preserved much better than that of class 3. Table 18.6: Class 4 with short root syllables (sunu ‘son’)

NOM ACC DAT GEN

SG

PL

sun-u sun-u sun-a sun-a

sun-a sun-a sun-um sun-a

1.2.5 Class 5 – Germanic n-stems In Proto-Indo-European, the thematic element of this noun class was formed by a nasal and had the structure /-Vn-/, with the vowel being subject to ablaut alternation. (Here and in the following, I distinguish between “apophony” as a term for any morphological or morphophonemic alternation of a vowel creating an allomorphic distinction irrespective of how it came into being. “Ablaut” only refers to those apophonic alternations that go back to the gradation patterns of Proto-Indo-European; i.e., zero-grade, full grade, etc.) The grammars distinguish three different paradigms for the three gender values, but the patterns differ only slightly in the nominative and accusative singular forms. All other values show the same forms for all three gender values. Cf. Tables 18.7–18.9.

278

III Old English Table 18.7: Class 5 masculine (guma ‘man’)

NOM DAT ACC GEN

SG

PL

gum-a gum-an gum-an gum-an

gum-an gum-um gum-an gum-ena

Table 18.8: Class 5 feminine (tunge ‘tongue’)

NOM DAT ACC GEN

SG

PL

tung-e tung-an tung-an tung-an

tung-an tung-um tung-an tung-ena

Table 18.9: Class 5 neuter (e¯age ‘eye’)

NOM DAT ACC GEN

SG

PL

e¯ag-e e¯ag-an e¯ag-e e¯ag-an

e¯ag-an e¯ag-um e¯ag-an e¯ag-ena

If the root ended with a vowel, the vocalic elements of the suffixes were dropped, as in gefe¯-a ‘joy-NOM.SG’, gefe¯a-na ‘joy-GEN.PL’, gefe¯a-m ‘joy-DAT.PL’ and gefe¯a-n for all other values. Also, some traces of an alternation in the vowel preceding the nasal survived into Old English, as in ox-(e)na ‘ox-GEN.PL’, ox-num ‘ox-DAT.PL’.

1.2.6 Class 6 – Germanic s-stems The original thematic element */-s-/, which, according to Verner’s Law, became */-z-/ and subsequently /-r-/, has been retained in class 6 in the plural forms and occasionally also in the singular where it was lost in most words. Cf. Table 18.10 below.

1.2.7 Class 7 – Germanic r-stems; kinship terms Class 7 consists of only a small number of kinship terms. They had */-r-/ as a thematic element. Except in fæder ‘father’ and sweostor ‘sister’, the dative singular form has an umlauted root vowel. In particular, the masculine nouns of this class show strong influence from class 1, so that the nominative/accusative plural forms fæd(e)r-as can be said to be the common forms in some nouns. Cf. Table 18.11.

18 Old English: Morphology

279

Table 18.10: Class 6 (lamb ‘lamb’)

NOM ACC DAT GEN

SG

PL

lamb lamb lamb-e lamb-es

lamb-ru lamb-ru lamb-rum lamb-ra

Table 18.11: Class 7 (dohtor ‘daughter’)

NOM ACC DAT GEN

SG

PL

dohtor dohtor dehter dohtor

dohtor dohtor dohtrum dohtra

1.2.8 Class 8 – Root nouns The nouns of Class 8 are labelled “athematic” or “root nouns” in traditional descriptions because in earlier stages they lacked a thematic element between root and inflectional suffix. These nouns are either masculine or feminine. The paradigms of the two gender values differ from each other slightly, both because they showed different paradigms already in Proto-Germanic, which later caused different patterns of apophonic alternations, and because they were influenced by the forms of classes 1a and 2 respectively. The nominative and accusative forms of masculine root nouns contained */-i(-)/ in the inflectional suffix in pre-Old English. The respective forms therefore can be identified by the umlauted root vowel (cf. Table 18.12). In northern documents and in early texts, if /-oː-/ is the root vowel, the umlauted vowel is /-œː-/ rather than /-eː-/, represented in the documents by 〈-oe-〉 or occasionally by 〈-œ-〉. The feminine nouns of class 8 originally followed the same pattern as the masculine nouns. However, influence from class 2 was stronger on feminine than that of class 1a on masculine nouns. Particularly in the genitive and dative singular, umlauted root forms competed with non-umlauted affixed forms (cf. Table 18.13). Table 18.12: Class 8: masculine root nouns ( fo¯t ‘foot’)

NOM ACC DAT GEN

SG

PL

fo¯t fo¯t fe¯t fo¯t-es

fe¯t fe¯t fo¯t-um fo¯t-a

280

III Old English Table 18.13: Class 8b: feminine root nouns (bo¯c ‘book’)

NOM ACC DAT GEN

SG

PL

bo¯c bo¯c be¯c, bo¯c be¯c, bo¯c-e

be¯c be¯c bo¯c-um bo¯c-a

1.2.9 Marginal paradigms Some paradigms do not fit in any of the above patterns, either because they are historically derived from extinct noun classes or because several patterns were conflated randomly. Two groups should be mentioned here: Nouns like fre¯ond ‘friend’ or hettend ‘enemy’ are derived from present participles and show the formative element -nd- (cf. below Section 2). However, present participles, if inflected, use the suffixes of the ja-stems (cf. Section 1.2.1), whereas the nouns of this group may follow several different, though not uniform, patterns. Not only is there some inconsistency in the inflectional endings of one word (e.g. frı¯end-∅, fre¯ond-e ‘friend-DAT.SG’; hettend-∅, hettend-e, hettend-as ‘enemy-NOM.PL’), there is also no uniform pattern among these nouns so that there is no point in postulating an independent class for these nouns. Ultimately, these lexemes have to be taken as idiosyncratic. Remainders of the PGrmc. dental stems sometimes still display the dental fricative following the root. Yet, this is again far from resulting in a consistent pattern: ealu exists, for instance, beside ealoþ ‘ale’ and monaþ ‘month’ shows the dental fricative consistently, but its inflection follows that of class 1a. Again, the small number of lexemes to which this applies and the lack of regularity within and across the respective paradigms does not justify postulating an independent noun class for these cases.

1.3 Adjectives OE adjectives agree with the noun they modify in case, number, and gender. For these three categories, there are two different types of declensions of adjectives in Old English. These two patterns are commonly referred to as “strong declension” and “weak declension” and for want of a better term I use these labels here. However, while the same labels are used for different verb classes (Section 2.1), the strong/ weak-distinction of adjectives is functional rather than lexical: the weak declension is used when preceded by a demonstrative or a possessive pronoun, the strong declension is generally used in any other case. In the predicative use, adjectives can be either strong or uninflected. Hogg and Fulk (2011) employ the labels “indefinite” and “definite” adjectives for the two classes, respectively. However, that the pattern of strong and weak adjectives does not quite follow the indefinite/definite distinction, can be seen by the fact that inherently definite pronouns like ælc ‘every’ are followed by a strong adjective. So is the ordinal o¯þer ‘second’ even when it is preceded by a demonstrative. It is therefore difficult to determine a specific function of this distinction since neither [± demonstrative] nor [± definite] reflect the actual use in Old English properly. Cf. the discussion in Mitchell (1985: I, Sections 136–141).

18 Old English: Morphology

281

In poetry, there was a more liberal distribution of the two declensions. If Campbell (1959: 261, Section 638) is right, that “the later the verse the less it diverges from the syntax of prose in this matter”, then this suggests that the distribution between strong and weak adjectives was quite stable up until the end of the OE period and that in an early stage the necessities of the meter could overrule the constraints of adjective inflection. Different from the noun paradigm, there are distinct instrumental forms for strong masculine and neuter singular adjectives. Like in the nominal paradigms, there is some syncretism of forms as can be seen in Tables 18.14 and 18.15 below.

1.3.1 Strong adjectives Strong adjectives follow a pattern which is often referred to as “a- and o¯-declensions”. However, because a number of values are marked by suffixes different from those of noun classes 1 or 2, this label is rather misleading particularly from a synchronic point of view. Those suffixes that do not correspond to the nominal forms are mostly the same as in some pronominal paradigms (cf. below Section 1.5). The default forms for strong adjectives are displayed in Table 18.14. Table 18.14: Strong adjectives (go¯d ‘good’) SG

NOM ACC DAT GEN INSTR

PL

MASC

FEM

NEUT

MASC

FEM

NEUT

go¯d-∅ go¯d-ne go¯d-um go¯d-es go¯d-e

go¯d-e go¯d-e go¯d-re go¯d-re –

go¯d-∅ go¯d-∅ go¯d-um go¯d-es go¯d-e

go¯d-e go¯d-e go¯d-um go¯d-ra –

go¯d-e, go¯d-a go¯d-e, go¯d-a go¯d-um go¯d-ra –

go¯d-∅ go¯d-∅ go¯d-um go¯d-ra –

Some values can differ according to whether the root syllable is long or short. The resulting variation parallels the differences between the noun classes 1a and 1c (Section 1.2.1) for neuter forms and between the noun classes 2a and 2b (Section 1.2.2) for feminine forms. There is also some diachronic variation in the paradigm. Final 〈-e〉 can appear as 〈-æ〉 in early texts. Late texts in West Saxon and Kentish can insert an epenthetic vowel in the endings for the genitive plural (-era) and for the dative and genitive singular of feminine adjectives (-ere). There are a number of phonologically conditioned modifications of the root in some forms of some (groups of) adjectives. These correspond in general with those alternations described in Section 1.2.1 for nouns. Slight variations in the paradigms depending on the existence of a root-final glides /-j-/ and /-w-/ (historically the -ja-/-jo¯- and the -wa-/-wo¯-stems) correspond to those described above for nouns in Section 1.2.1 and in Section 1.2.2.

1.3.2 Weak adjectives The endings of the weak adjectives correspond to those of the nominal class 5 (cf. Section 1.2.5). Only the forms of the genitive plural correspond with the pronominal

282

III Old English paradigms. However, while most forms use the pronominal -ra, some early West Saxon texts attest to the alternative suffix -ena for the genitive plural. In contrast to the strong adjectives, there is no form distinguishing the instrumental case from the dative. Gender distinctions exist only in the nominative and accusative singular. The default forms are shown in Table 18.15: Table 18.15: Weak adjectives (go¯d ‘good’) SG

NOM ACC

PL

MASC

FEM

NEUT

go¯d-a go¯d-an

go¯d-e go¯d-an

go¯d-e go¯d-e

DAT GEN

go¯d-an go¯d-an

go¯d-an go¯d-an go¯d-um go¯d-ra, go¯d-ena (early WSax.)

1.3.3 Comparison of adjectives As in most IE languages, the only independent morphological category of adjectives is comparison with the positive as the default value and the marked values comparative and superlative. The comparative is formed with the suffix -ra. Because this is homophonous with the genitive plural suffix, the forms for this value usually have only one of the two suffixes, so that the form for the comparative in the nominative singular is identical with that for the positive in the genitive plural. An alternative strategy is used in Northumbrian where the suffix -rena is used for the genitive plural of comparative forms. The suffix for the superlative varies between -ost, -ast, and -est. More archaic variants of the same morpheme are -ust (whence -ost). The variation is probably owed to two competing sets of suffixes in Proto-Germanic: *-o¯zan-/*-o¯sta- and *-izan-/*-ista(COMPR/SUP). The fact that some adjectives show i-umlaut in the comparative and superlative forms whereas most adjectives do not (cf. Table 18.16) is further evidence for the different sets of suffixes. The predominance of the superlative -est (< *-ista-) in Old English (and in the later history of English) must therefore have developed after i-umlaut affected the roots. While some adjectives show irregular formations due to phonological cross-influences of suffixes and roots, but use etymologically the same root (labelled “idiosyncratic” in Table 18.16), some frequent adjectives employ hybrid paradigms, i.e., their forms are based on an etymologically different root in the comparative and superlative forms than the positive form. Cf. Table 18.16. Table 18.16: Comparison of adjectives POSITIVE

COMPARATIVE

SUPERLATIVE

regular without i-umlaut

earm ‘poor’ glæd ‘glad’

earm-ra glæd-ra

earm-ost, -ast, -est glæd-ost, -ast, -est

regular with i-umlaut

eald ‘old’ lang ‘long’

ield-ra leng-ra

ield-est leng-est (Continued )

18 Old English: Morphology

283

Table 18.16: Continued

idiosyncratic

hybrid

POSITIVE

COMPARATIVE

SUPERLATIVE

micel ‘great’ ly¯tel ‘little’

ma¯ra læ¯ssa

mæ¯st læ¯st

bet(t)(e)ra se¯lra, sella wiersa

bet(e)st, best se¯lest wierrest, wierst

go¯d ‘good’ yfel ‘bad’

1.3.4 The formation of adverbs There are two (sets of) suffixes to form adverbs from adjectives. Most frequent is the suffix -e, as in clæ¯n ‘pure’, clæ¯n-e ‘purely’. The ultimate source of this suffix is, according to Brunner (1965: 249, Section 315), a PIE ablative marker. Because adverbs from adjectives formed with the adjectivizer -lic frequently occur, the two suffixes (-lic- and -e) are reanalysed as one adverbial marker, so that sometimes adjectives form their adverbial forms with -lice (whence PDE -ly ADV). Moreover, two suffix variants -unga and -inga form adverbs out of any other word class. Hence, a¯n-inga ‘entirely’ from a¯n ‘one’, fæ¯r-inga ‘suddenly’ from fæ¯r ‘attack’, eall-unga ‘entirely’ from eall ‘all’. Some inflected forms of adjectives – mostly genitive forms – can be used adverbially and have been lexicalized as adverbs, e.g. eall-es ‘entirely’ from ‘all-GEN.MASC’, micl-es ‘very’ from ‘great-GEN.MASC’, singal-es ‘always’ from ‘constant-GEN.MASC’. In the same way case/number forms of nouns can be used as adverbial adjuncts from which some forms became lexicalized as adverbs. Examples are dæg-es ‘day-GEN. SG’ > ‘by day’, þonc-es ‘gratitude-GEN.SG’ > ‘willingly’, hwe¯n-e ‘small amount-INSTR. SG’ > ‘a little (ADV)’, fa ¯ cn-e ‘fraud-DAT.SG’ > ‘deceitfully’, hwı¯l-um ‘while-DAT.PL’ > ‘sometimes’. Adverbs formed from adjectives and a number of lexicalized adverbs form their comparative with -or and their superlative with -ost, with some variation in the vowel of the suffix.

1.4 Cardinal Numerals Cardinal numerals are uninflected if they immediately precede a quantified noun. If not, cardinal numerals from ‘1’ to ‘12’ inflect for case and gender. The numerals a¯n ‘1’, and those from ‘4’ to ‘12’ use adjectival endings. The numerals for ‘2’ and ‘3’ have their own paradigms, displayed in Tables 18.17 and 18.18, respectively. They contain partly independent forms, and partly forms that correspond roughly with the pronominal paradigms. Multiples of ‘10’ may or may not be inflected. If they are, the relevant suffixes are those of noun class 1b or of 3. In the genitive they may also use the suffix -ra. The same holds for hund ‘100’ and for þusend ‘1000’. For a detailed description of the Old English numeral system, the numeral forms and their morphological variants, and for the principles of the formation of complex numerals see von Mengden (2010: Chapters II and III).

284

III Old English Table 18.17: Inflectional paradigm of twa¯ ‘2’

NOM/ACC

MASC

FEM

NEUT

twe¯gen

twa¯

twa¯, tu¯

twa¯m, twæ¯m twe¯g(e)a, twe¯g(e)ra

DAT GEN

Table 18.18: Inflectional paradigm of þre¯o ‘3’

NOM/ACC

MASC

FEM

NEUT

þrı¯e

þre¯o

þre¯o

DAT GEN

þrim þre¯ora

1.5 Pronouns The four main types of pronouns in Old English – demonstrative pronouns, personal pronouns, possessive pronouns, and interrogative pronouns – all show slightly different patterns with respect to inflectional categories. When modifying a noun or when used anaphorically, pronouns generally agree with the co-referential noun in number, case, and gender. There is no gender distinction in the plural. If appropriate, differences in the agreement patterns among the four types are mentioned individually in the following sections.

1.5.1 Demonstrative pronouns There are two sets of demonstrative pronouns, often labelled “simple” vs. “composite” (Brunner 1965: Section 337–338: “einfach” vs. “zusammengesetzt”) and “definite article” vs. “demonstrative” (Lass 1994: Section 6.2.2). The latter distinction is justified in so far as the one demonstrative develops into a definite marker. However, although towards the end of the period marking definite reference increasingly becomes its main function, this is never a necessary use and even in late texts definiteness is not systematically encoded. The former distinction – simple vs. composite – is motivated by the etymology of the composite demonstrative, which emerged from the fusion of the inflected form of the simple demonstrative form and a particle *-si. Neither of these sets of labels describes the synchronic situation of the OE period adequately. Because the distribution of the two sets of demonstratives roughly corresponds to that of the PDE pronouns this and that, the terms “distal” and “proximal” appear more adequate. As the distal demonstratives are clearly more frequent, they can be said to be the default set of demonstrative pronouns (cf. Hogg and Fulk 2011: Section 5.3). Both sets of demonstratives distinguish five case/number values, including instrumental forms for masculine and neuter singular. The most frequent forms of the two paradigms are shown in Tables 18.19 and 18.20.

18 Old English: Morphology

285

Table 18.19: Distal demonstratives SG

NOM DAT ACC GEN INSTR

PL

MASC

FEM

NEUT

se þa¯m þone þæs þon, þy¯

se¯o þæ¯re þa¯ þæ¯re –

þæ t þa¯m þæ t þæ s þon, þy¯

þa¯ þa¯m, þæ¯m þa¯ þa¯ra, þæ¯ra –

Table 18.20: Proximal demonstratives SG

NOM ACC DAT GEN INSTR

PL

MASC

FEM

NEUT

þes þisne þissum þisses þy¯s

þe¯os þa¯s þisse þisse –

þis þis þissum þisses þy¯s

þa¯s þa¯s þissum þissa –

1.5.2 Personal pronouns Personal pronouns distinguish case/number as described above. Additionally there are distinct forms for the first, second, and third person. There is no distinction between an inclusive and an exclusive use of the first person plural. The pronoun of the third person also distinguishes gender in the singular. There are four case values (no instrumental) and three number values (singular, dual, plural) for the first and second person and two (no dual) for the third person. The first and second person paradigms show no formal distinction between dative and accusative, except in Anglian texts where the accusative forms mec ‘1P.SG’, uncet ‘1P.DU’, u¯sic ‘1P.PL’ and ðec ‘2P.SG’, incit ‘2P.DU’ and e¯owic ‘2P.PL’ occur. Cf. Tables 18.21–18.23. Table 18.21: Personal pronoun of the first person

NOM ACC/DAT GEN

SG

DU

PL

ic me mı¯n

wit unc uncer

we u¯s u¯re

Table 18.22: Personal pronoun of the second person

NOM ACC/DAT GEN

SG

DU

PL

þu þe þı¯n

git inc incer

ge e¯ow e¯ower

286

III Old English Table 18.23: Personal pronoun of the third person SG

NOM ACC DAT GEN

PL

MASC

FEM

NEUT

he hine him his

he¯o hı¯(e) hire hire

hit hit him his

hı¯(e), he¯o hı¯(e), he¯o him hira, heora

Possessive pronouns are derived from the genitive forms of personal pronouns. In addition to their original case value they can inflect for cases using the suffixes of the strong adjectives (Section 1.3.1). For the third person, there is an independent possessive sı¯n, but it is clearly less frequent than the forms based on his.

1.5.3 Interrogative pronouns Interrogative pronouns have a reduced gender distinction. Nominative and accusative forms distinguish between common and neuter, whereas the other case forms have no case distinction at all. There are forms for five cases, i.e., including the instrumental. Plural forms do not exist. The forms are shown in Table 18.24. Table 18.24: Interrogative pronoun

NOM ACC

MASC/FEM

NEUT

hwa hwone

hwæt hwæt

DAT GEN INSTR

hwa¯m, hwæ¯m hwæs hwı¯, hwy¯

2 Inflectional morphology of verbs 2.1 The morphological categories of verbs All verbs generally encode two tense values, past and present. For both tenses, there are two mood values, indicative and subjunctive. The indicative is the default value and the subjunctive is mainly used when the predication represents the wish of the speaker rather than a real event. In some handbooks it is therefore referred to as “optative”. A third mood value, the imperative, does not distinguish tense. As in all Grmc. languages, there are two main groups of verbs in Old English, traditionally called “weak” and “strong”, which do not differ in the inflectional categories/values they encode, but do differ considerably in the morphological strategy used for encoding the values. While weak verbs employ affixes only (and one circumfix), strong verbs use a system of transfixes (on the distinction between the various types of affixes and the terminology involved cf. Melˈcˇuk 2000: Sections 3.2.2–5; Melˈcˇuk 2006: Sections 3.3.2–5.) The transfix-patterns go back to morphophonemic ablaut alternations in Proto-Indo-European. The affix system of weak verbs emerged during the common Germanic period. Weak and strong verbs will be treated differently in the main sections 2.2 and 2.3 below.

18 Old English: Morphology The verb agrees with the subject in person and number. Number is distinguished between singular and plural (the latter also used when the subject is in the dual). Only in the indicative singular are first, second, and third person distinguished. Generally, the degree of syncretism is considerably smaller among the inflectional markers of verbs than in the nominal paradigms. There are three types of infinite verb forms for each OE verb, the infinitive, a present participle and a past participle, all inflecting for case/number and gender like adjectives (cf. Section 1.3). Analytic constructions with auxiliaries (most commonly beon/wesan ‘be’, habban ‘have’ and weorðan ‘become’) and participles are attested, but they do not seem to be used as systematically for particular tense/aspect/mood values as they are in Present-day English; for details cf. Mitchell (1985: I, Sections 682–743). Finally, the combination of an auxiliary (beon/wesan or weorðan) with the past participle can be used to form past constructions; cf. Mitchell (1985: I, Sections 744–858). On the functions of the two participles cf. further Mitchell (1985: I, Sections 972–989) and on the infinitive cf. Mitchell (1985: I, Sections 920–971).

2.2 Weak verbs 2.2.1 Preliminaries Synchronically, the three classes of weak verbs postulated in traditional descriptions do not differ in the paradigmatic use of inflectional endings. The only differences are morphophonemic alternations in a number of forms affecting either suffix or root which result from different derivational suffixes in Proto-Germanic. Such alternations can, however, also be observed within some of the three classes – resulting from later morphophonemic cross-influences between suffix and root – so that the traditional class distincti