The Handbook of the History of English

Edited by

Ans van Kemenade and Bettelou Los

The Handbook of the History of English

Blackwell Handbooks in Linguistics

This outstanding multi-volume series covers all the major subdisciplines within linguisticstoday and, when complete, will offer a comprehensive survey of linguistics as a whole.

Already published:

The Handbook of Child LanguageEdited by Paul Fletcher and Brian MacWhinney

The Handbook of Phonological TheoryEdited by John A. Goldsmith

The Handbook of Contemporary Semantic TheoryEdited by Shalom Lappin

The Handbook of SociolinguisticsEdited by Florian Coulmas

The Handbook of Phonetic SciencesEdited by William J. Hardcastle and John Laver

The Handbook of MorphologyEdited by Andrew Spencer and Arnold Zwicky

The Handbook of Japanese LinguisticsEdited by Natsuko Tsujimura

The Handbook of LinguisticsEdited by Mark Aronoff and Janie Rees-Miller

The Handbook of Contemporary SyntacticTheoryEdited by Mark Baltin and Chris Collins

The Handbook of Discourse AnalysisEdited by Deborah Schiffrin, DeborahTannen, and Heidi E. Hamilton

The Handbook of Language Variation andChangeEdited by J. K. Chambers, Peter Trudgill,and Natalie Schilling-Estes

The Handbook of Historical LinguisticsEdited by Brian D. Joseph and Richard D. Janda

The Handbook of Language and GenderEdited by Janet Holmes and Miriam Meyerhoff

The Handbook of Second Language AcquisitionEdited by Catherine J. Doughty andMichael H. Long

The Handbook of BilingualismEdited by Tej K. Bhatia and William C. Ritchie

The Handbook of PragmaticsEdited by Laurence R. Horn and GregoryWard

The Handbook of Applied LinguisticsEdited by Alan Davies and Catherine Elder

The Handbook of Speech PerceptionEdited by David B. Pisoni and Robert E. Remez

The Blackwell Companion to Syntax, Volumes I–VEdited by Martin Everaert and Henk vanRiemsdijk

The Handbook of the History of EnglishEdited by Ans van Kemenade and Bettelou Los

The Handbook of World EnglishesEdited by Braj B. Kachru, Yamuna Kachru,and Cecil L. Nelson

The Handbook of the History of English

Edited by

Ans van Kemenade and Bettelou Los

© 2006 by Blackwell Publishing Ltd

BLACKWELL PUBLISHING350 Main Street, Malden, MA 02148-5020, USA9600 Garsington Road, Oxford OX4 2DQ, UK550 Swanston Street, Carlton, Victoria 3053, Australia

The right of Ans van Kemenade and Bettelou Los to be identified as the Authors of the Editorial Material in this Work has been asserted inaccordance with the UK Copyright, Designs, and Patents Act 1988.

All rights reserved. No part of this publication may be reproduced,stored in a retrieval system, or transmitted, in any form or by anymeans, electronic, mechanical, photocopying, recording or otherwise,except as permitted by the UK Copyright, Designs, and Patents Act1988, without the prior permission of the publisher.

First published 2006 by Blackwell Publishing Ltd

1 2006

Library of Congress Cataloging-in-Publication Data

The handbook of the history of English / edited by Ans van Kemenade and Bettelou Los.

p. cm. — (Blackwell handbooks in linguistics)Includes bibliographical references and index.ISBN-13: 978-0-631-23344-2 (hardcover : alk. paper)ISBN-10: 0-631-23344-X (hardcover : alk. paper) 1. English language—

History. I. Kemenade, Ans van, 1954– II. Los, Bettelou. III. Series.

PE1075.H335 2006420.9—dc22

2005033628

A catalogue record for this title is available from the British Library.

Set in 10/12pt Palatinoby Graphicraft Limited, Hong KongPrinted and bound in the United Kingdomby TJ International, Padstow, Cornwall

The publisher’s policy is to use permanent paper from mills that operatea sustainable forestry policy, and which has been manufactured frompulp processed using acid-free and elementary chlorine-free practices.Furthermore, the publisher ensures that the text paper and cover boardused have met acceptable environmental accreditation standards.

For further information onBlackwell Publishing, visit our website:www.blackwellpublishing.com

Contents

Editors’ Introduction viiNotes on Contributors x

Part I Approaches and Issues 1

1 Change for the Better? Optimality Theory versus History 3April McMahon

2 Cuing a New Grammar 24David W. Lightfoot

3 Variation and the Interpretation of Change in Periphrastic Do 45Anthony Warner

4 Evolutionary Models and Functional-Typological Theories of Language Change 68William Croft

Part II Words: Derivation and Prosody 93

5 Old and Middle English Prosody 95Donka Minkova

6 Prosodic Preferences: From Old English to Early Modern English 125Paula Fikkert, Elan B. Dresher, and Aditi Lahiri

7 Typological Changes in Derivational Morphology 151Dieter Kastovsky

8 Competition in English Word Formation 177Laurie Bauer

Part III Inflectional Morphology and Syntax 199

9 Case Syncretism and Word Order Change 201Cynthia L. Allen

vi Contents

10 Discourse Adverbs and Clausal Syntax in Old and Middle English 224Ans van Kemenade and Bettelou Los

11 The Loss of OV Order in the History of English 249Susan Pintzuk and Ann Taylor

12 Category Change and Gradience in the Determiner System 279David Denison

Part IV Pragmatics 305

13 Pathways in the Development of Pragmatic Markers in English 307Laurel Brinton

14 The Semantic Development of Scalar Focus Modifiers 335Elizabeth Closs Traugott

15 Information Structure and Word Order Change: The Passive as an Information-rearranging Strategy in the History of English 360Elena Seoane

Part V Pre- and Post-colonial Varieties 393

16 Old English Dialectology 395Richard Hogg

17 Early Middle English Dialectology: Problems and Prospects 417Margaret Laing and Roger Lass

18 How English Became African American English 452Shana Poplack

19 Historical Change in Synchronic Perspective: The Legacy of British Dialects 477Sali A. Tagliamonte

20 The Making of Hiberno-English and Other “Celtic Englishes” 507Markku Filppula

Part VI Standardization and Globalization 537

21 Eighteenth-century Prescriptivism and the Norm of Correctness 539Ingrid Tieken-Boon van Ostade

22 Historical Sociolinguistics and Language Change 558Terttu Nevalainen

23 Global English: From Island Tongue to World Language 589Suzanne Romaine

Appendix: Useful Corpora for Research in English Historical Linguistics 609

Index 612

Editors’ Introduction

The format of this Handbook is not a traditional one, as will immediately beclear from the Contents list. Its chapters are not arranged chronologically, moving from Old to Middle to Modern English, or linguistically, moving fromphonology to morphology to syntax. We realized from the start that it was going to be impossible to achieve any degree of exhaustiveness in the choiceof topics in the format of a single volume, especially in the case of a languagewhose history is as well studied as that of English – a fact borne out by the existence of the Cambridge History of the English Language (CHEL),which requires six volumes with virtually book-length chapters to give com-prehensive survey treatment of the work done on relevant aspects of theEnglish language. We decided to concentrate instead on other typical hand-book functions: those of providing shortcuts to current thinking for readers who want to become familiar with subjects that are outside their own areas ofinterest, and of providing a “state of the art” overview of current research. Thisled us to concentrate on the kind of work which is grounded in philologicalexpertise and care but which is also informed by recent theoretical advancesand issues in the study of English historical linguistics and of historical lin-guistics and language change more generally. The term “theoretical advances”should be taken in a broad sense here, including theoretical models of linguisticcompetence and language use, the theory and methodology of social andregional variation in language, quantitative corpus-driven work and corpus-making, or a combination of any of the above. Given this chosen slant of theHandbook, we also specifically requested a number of our contributors to gobeyond the survey approach that is so readily associated with handbooks, and to try as much as possible to present chapters in which the work for which they are best known is argued and illustrated, while also presenting new material and suggesting new avenues of research. This has led to a fairlyunorthodox chapter division.

The first part presents a discussion of four very different approaches to thestudy of the history of language change in the context of the English language.

They all represent ways of researching language change that have become part of the research agenda relatively recently: April McMahon evaluates theviability of Optimality Theory for historical work in English phonology. DavidW. Lightfoot presents the case for approaching syntactic change from the per-spective of the language learner, and highlights the role of grammatical cuesin the acquisition of the grammatical rule system by new generations of learn-ers. Anthony Warner presents a theoretically informed variationist approachwhich casts new light on the classical problem of the diffusion of a new syn-tactic feature, in this case that of periphrastic do in early Modern English. Finally,William Croft discusses evolutionary models in combination with functional-typological theories as they apply to the phenomenon of language change.

The second part presents historical work done at the level of the word. Thechapters by Donka Minkova, and Paula Fikkert, Elan B. Dresher, and Aditi Lahiripresent work done in prosody, an exciting area of research which bringstogether changes in phonology and morphology. The chapters by DieterKastovsky and Laurie Bauer provide overviews of historical word formation.

The chapters in Part III, by Cynthia L. Allen, Ans van Kemenade andBettelou Los, Susan Pintzuk and Ann Taylor, and David Denison, are concernedwith various aspects of word order change, with the perennial problem of theinteraction between inflectional change and syntactic change, and with the prob-lem of how to model this interaction. They also illustrate the value of corpuswork in historical syntax. The use of corpora in English diachronic syntax isan excellent, and possibly unique, example of how formal theories and per-formance data can be brought together to provide us with a better understandingof the syntax of early English and the phenomenon of syntactic change.

In Part IV, “Pragmatics,” the chapters by Laurel Brinton and Elizabeth ClossTraugott represent approaches that focus on how word-meanings may changein non-arbitrary ways. Particularly interesting is the development of degreeadverbs, focus modifiers and pragmatic markers, which have acquired thesespecial functions via pragmatic implicatures and inferences that can bededuced from the uses of lexemes in specific constructions. Elena Seoane’s chap-ter discusses the interaction of syntax and information structure that is evidentin the changing role of the passive construction in the history of English.

The chapters in Part V, “Pre- and Post-colonial Varieties,” discuss regionalvariation and the origin of varieties of English and, conversely, how historicalchange is reflected in present-day dialects. The fine-grained philological con-cerns of Old and Middle English dialectology and the construction of corporasuited for such research, as presented in the chapters by Richard Hogg, and Mar-garet Laing and Roger Lass, are complemented by the detailed arguments emerg-ing from recent exciting work on the British dialectal origins of African AmericanEnglish in the chapter by Shana Poplack, the spin-offs of older regional dialectsin present-day language features in the chapter by Sali A. Tagliamonte, andthe development of Celtic Englishes in the chapter by Markku Filppula.

Finally, Part VI contains three chapters that have as their common denom-inator various types of external influence on the language: a chapter on the

viii Editors’ Introduction

Editors’ Introduction ix

impact of eighteenth-century prescriptivism by Ingrid Tieken-Boon vanOstade, one on sociolinguistic influences as brought to light by careful corpuswork in the chapter by Terttu Nevalainen, and one on the development of Englishas a world language, by Suzanne Romaine.

Although we have been much concerned with putting together a book withwork that attempts to give new and theoretically informed answers to old ques-tions, we have been reluctant to make divisions in terms of theoretical issues.The most important reason for this is that we were worried that any theoret-ical division would construct as much as reflect theoretical ideologies. Not onlywas this never our concern, we also feel that it would have been entirely inappropriate as an approach toward a handbook on a field that is as empiricalas “the history of English.” This attitude is also reflected in the chapters by ourcontributors, who on the whole seemed a good deal more concerned with pre-senting theoretically informed work that resolves puzzles in their variousempirical domains than with engaging in polemics with each other. The divi-sion of the book into parts has, in the end, been guided by the empirical domainsin which theoretical progress of various kinds has been made since the mid-1990s, as reflected in detailed case studies. There is a good deal of emphasisoverall on language variation of various kinds, and this chimes well with therecent general booming in the field of sociolinguistics and language variation.The strength of the book lies, we hope, in applying recent insights to old problems.

We wish to conclude with a work of thanks to most of our contributors fortheir prompt delivery, and for their willingness to engage in a dialogue on theirchapters, and to exchange views with each other. Many thanks are also due toSarah Coleman at Blackwell for her cheerful and practical guidance of the wholeproject.

Ans van Kemenade and Bettelou LosMay 2005

Notes on Contributors

Cynthia L. AllenCynthia L. Allen is a Reader in the School of Language Studies at the Austra-lian National University and the Director of ANU’s Centre for Research onLanguage Change. She is a Fellow of the Australian Academy of the Human-ities. She specializes in history of English morphosyntax and her publicationsinclude Case Marking and Reanalysis: Grammatical Relations from Old to Early ModernEnglish (1995).cindy.allen@anu.edu.au

Laurie BauerLaurie Bauer holds a personal chair in Linguistics at Victoria University ofWellington, New Zealand. He is the author of a number of works on morphologyand English word-formation, including, most recently, in collaboration withRodney Huddleston, the chapter on English word-formation for the CambridgeGrammar of the English Language (eds. R. Huddleston and G. Pullum, 2002). Healso writes on international varieties of English.laurie.bauer@vuw.ac.nz

Laurel BrintonLaurel Brinton is Professor of English at the University of British Columbia,Canada, specializing in the history and structure of the English language. Herresearch areas include verbal aspect, grammaticalization, lexicalization, and his-torical pragmatics, with particular attention to the development of pragmaticmarkers. Most recently, she has co-authored Lexicalization and Language Change(with Elizabeth Closs Traugott, 2005) and The History of English: A LinguisticIntroduction (with Leslie Arnovick, 2005).brinton@interchange.ubc.ca

William CroftWilliam Croft is Professor of Linguistics at the University of New Mexico (USA).His main areas of specialization are typology, including grammaticalization;

Notes on Contributors xi

evolutionary models of language change; construction grammar; and cognitivelinguistics. His books include Explaining Language Change: An EvolutionaryApproach (2000), Radical Construction Grammar: Syntactic Theory in TypologicalPerspective (2001), Typology and Universals (2nd edition, 2003), and CognitiveLinguistics (with D. Alan Cruse, 2004).wcroft@unm.edu

David DenisonDavid Denison is Professor of English Linguistics at the University ofManchester, one of the founding editors of the journal English Language andLinguistics, author of English Historical Syntax (1993) and of the “Syntax” chap-ter in vol. 4 of The Cambridge History of the English Language (1998), one of theeditors of Fuzzy Grammar (2004), and joint editor of a forthcoming History ofthe English Language (2006).david.denison@manchester.ac.uk

Elan B. Dresher Elan B. Dresher is Professor of Linguistics at the University of Toronto. He haspublished on a wide variety of topics, including Hebrew phonology and mor-phology, historical linguistics, and learnability. His work on historical linguisticsincludes Old English and the theory of phonology and joint work with AditiLahiri on historical West-Germanic. He is also well known for his work on learn-ability, published in major journals. His current research interest concerns the role of contrast in phonological theory. He has been a columnist for GlotInternational.Dresher@chass.utoronto.ca

Paula FikkertPaula Fikkert is a researcher at the Radboud University of Nijmegen. She receivedher PhD in 1994 at Leiden University, and from 1994 to1999 she was aresearcher at the University of Konstanz, returning to the Netherlands as aresearch fellow of the Royal Dutch Academy of Science. Recent publicationsinclude “Acquisition of Phonology,” in The First Glot International State-of-the-Article Book (2000); (with N. Schiller and C. Levelt) “Stress priming in picturenaming: an SOA study,” in Brain and Language (2004); and (editor with H. Jacobs)Development in Prosodic Systems (2003).p.fikkert@let.ru.nl

Markku FilppulaMarkku Filppula is Professor of English at the University of Joensuu, and Docentin English Philology at the University of Helsinki. He is also currently Directorof LANGNET, the Finnish Graduate School in Language Studies. He wasawarded his PhD by the National University of Ireland (Dublin) in 1986, andhas published widely on Hiberno-English, other “Celtic Englishes” and lan-guage contacts in general. He is the author of The Grammar of Irish English:Language in Hibernian Style (1999), English and Celtic in Contact (with Juhani

xii Notes on Contributors

Klemola and Heli Pitkänen, forthcoming), and he wrote the chapter on IrishEnglish morphology and syntax for A Handbook of Varieties of English, vol. 2:Morphology and Syntax (eds. B. Kortmann et al., 2004). He was also one of theeditors (with Klemola and Pitkänen) of The Celtic Roots of English (2002).markku.filppula@joensuu.fi

Richard HoggRichard Hogg has been Smith Professor of English Language and MedievalLiterature at Manchester since 1980. His doctorate is from the University ofEdinburgh, and he held further appointments in Amsterdam and Lancaster.His research interests are English historical linguistics, particularly in dialec-tology and Old English phonology and morphology. His publications includeOld English Grammar (1992). He is General Editor of The Cambridge History of the English Language, the fifth and final volume of which was completed in2001. He edited volume 1 (The Beginnings to 1066), and wrote the chapter onphonology and morphology for that volume. He has recently published AnIntroduction to Old English Language.r.m.hogg@man.ac.uk

Dieter KastovskyDieter Kastovsky is Professor of English Language and Translation Studies atthe University of Vienna. His doctorate is from the University of Tuebingenand he has held a further appoint at Wuppertal, and guest professorships atStockholm, Cape Town, Georgetown and Lublin (Poland). His major researchinterest is in early English word formation, and he has published widely onthis and related topics, including his contribution on Old English word for-mation to Volume I of the Cambridge History of the English Language (1992).Dieter.Kastovsky@univie.ac.at

Ans van KemenadeAns van Kemenade is Professor of English Language and Linguistics at theRadboud University of Nijmegen. After her doctorate in 1987 at the Universityof Utrecht, she held posts at the University of Leiden and the Vrije UniversiteitAmsterdam. Her main research interests are in historical syntax and in syn-tactic variation and change. She has published Syntactic Case and MorphologicalCase in the History of English. She has edited special issues of Lingua (1993,vol. 89/1–2 with Aafke Hulk), Linguistics (1999, vol. 37.6) and the Yearbookof Morphology 2003 (with Geert Booij), and, with Nigel Vincent, Parameters ofMorphosyntactic Change (1997). She has published numerous articles, mainly onOld and Middle English syntax.a.v.kemenade@let.ru.nl

Aditi LahiriAditi Lahiri has taught at University of California at Los Angeles and SantaCruz, and has been a research scientist at the Max-Planck-Institute for Psycho-linguistics in the Netherlands. She is currently Professor of General Linguistics

Notes on Contributors xiii

at the University of Konstanz in Germany. Her research interests includephonology and morphology (synchronic and diachronic), phonetics, experi-mental psycholinguistics, and neurolinguistics. Recent publications on Germanicinclude Analogy, Leveling, Markedness: Principles of Change in Phonology andMorphology (editor, 2000), as well as papers in Language 75 (1999; with Elan B.Dresher), English Language and Linguistics 3 (1999; with Paula Fikkert), andTransactions of the Philological Society 102 (2004; with Astrid Kraehenmann).aditi.lahiri@uni-konstanz.de

Margaret LaingMargaret Laing is Research Fellow in the Institute for Historical Dialectology,English Language, University of Edinburgh. She contributed to the productionof A Linguistic Atlas of Late Mediaeval English (1986). Since then she has beenengaged in research towards the creation of a Linguistic Atlas of Early MiddleEnglish (LAEME) in collaboration (from 2002) with Roger Lass. She has pub-lished extensively on early Middle English, including A Catalogue of Sources fora Linguistic Atlas of Early Medieval English (1993) and many articles arising outof the investigation of early Middle English dialects and scribal systems.M.Laing@ed.ac.uk

Roger LassRoger Lass is Professor Emeritus of Linguistics and Honorary Research Fellowin English at the University of Cape Town. He has published numerous booksand technical papers in the fields of theoretical and historical linguistics andthe history of English. His most recent books are Historical Linguistics and LanguageChange (1997) and volume 3 of the Cambridge History of the English Language(1999), which he both edited and contributed a major chapter to. Most of hiscurrent research is dedicated to the LAEME project for which, among otherthings, he is responsible for the etymological corpus.lass@iafrica.com

David W. LightfootDavid Lightfoot works on language acquisition, change, and syntax. His majorappointments have been at McGill University, the University of Utrecht, theUniversity of Maryland, where he established a new department of linguistics,and at Georgetown University, where he was Dean of the Graduate School of Arts and Sciences. He is now Assistant Director of the National ScienceFoundation. He is the author of ten books, including The Language Lottery: Towarda Biology of Grammars (1982), How to Set Parameters: Arguments from LanguageChange (1991), and (with Stephen R. Anderson) The Language Organ: Linguisticsas Cognitive Physiology (2002), and most recently How New Languages Emerge (2005).lightd@georgetown.edu

Bettelou LosBettelou Los is a Lecturer in Linguistics at the Vrije Universiteit Amsterdam.She graduated from the University of Amsterdam in 1986 and has since held

xiv Notes on Contributors

teaching and research positions at the University of Amsterdam, the VrijeUniversiteit Amsterdam (where she obtained her doctorate in 2000), theUniversity of Nijmegen, the University of Leuven and other colleges of highereducation. She contributes, with Wim van der Wurff, to the morphology andsyntax section of The Year’s Work in English Studies. She has published The Riseof the to-infinitive (2005) and many articles on diachronic syntax. blj.los@let.vu.nl

April McMahonApril McMahon is Forbes Professor of English Language at the University of Edinburgh. She previously worked in the Department of Linguistics at the University of Cambridge and held a Chair in English Language andLinguistics at the University of Sheffield. Her research interests involve the inter-action between phonological theory and historical evidence, as well as issuesof language comparison and classification. Her books include UnderstandingLanguage Change (1994), Lexical Phonology and the History of English (2000),Change, Chance, and Optimality (2000), and Language Classification by Numbers(with Robert McMahon, forthcoming). She is currently the President of theLinguistics Association of Great Britain.April.McMahon@ed.ac.uk

Donka MinkovaDonka Minkova is Professor of English at University of California, LosAngeles. She has published widely in the areas of English historical phono-logy, meter, dialectology, and syntax. She is the author of The History of FinalVowels in English (1991), English Words: History and Structure (with RobertStockwell, 2001), and Alliteration and Sound Change in Early English (2003).She has co-edited Studies in the History of the English Language: A MillennialPerspective (with Robert Stockwell, 2002) and Chaucer and the Challenges ofMedievalism (with Theresa Tinkle, 2003).minkova@humnet.ucla.edu

Terttu NevalainenTerttu Nevalainen is Professor of English Philology at the Department ofEnglish, University of Helsinki, Finland, and the Director of the Research Unitfor Variation and Change in English, co-funded by the Academy of Finlandand the University of Helsinki. She is the leader of the project “Sociolinguisticsand Language History,” which has produced the Corpus of Early EnglishCorrespondence (CEEC) and a number of publications on the topic, includingHistorical Sociolinguistics: Language Change in Tudor and Stuart England (co-author with Raumolin-Brunberg, 2003). Her other research interests include phonetics and historical lexicology.Terttu.Nevalainen@Helsinki.Fi

Susan PintzukSusan Pintzuk is Professor of English Language and Linguistics at theUniversity of York (UK). Her research interests include syntactic variation and

Notes on Contributors xv

change, particularly in the history of English and other Germanic languages;statistical models of language change; and corpus linguistics. She has par-ticipated in the construction of corpora in the English Parsed Corpora series(the York–Toronto–Helsinki Parsed Corpus of Old English Prose, with AnnTaylor, Anthony Warner, and Frank Beths; the York–Helsinki Parsed Corpusof Old English Poetry, with Leendert Plug; a parsed version of the Corpus ofEarly English Correspondence, with Taylor and Warner, in collaboration withthe Research Unit for Variation and Change in English, Helsinki) and theBrooklyn–Geneva–Amsterdam–Helsinki Parsed Corpus of Old English, withEric Haeberli, Ans van Kemenade, Willem Koopman, and Frank Beths.sp20@york.ac.uk

Shana PoplackShana Poplack is Distinguished University Professor and Canada Research Chair in Linguistics at the University of Ottawa, where she also directs theSociolinguistics Laboratory. An expert in linguistic variation theory and its appli-cation to diverse areas of language contact, she has published widely on code-switching, Hispanic linguistics, Canadian French and numerous aspects of AfricanAmerican Vernacular English.spoplack@aix1.uottawa.ca

Suzanne RomaineSuzanne Romaine has been Merton Professor of English Language at theUniversity of Oxford since 1984. Her research interests lie primarily in his-torical linguistics and sociolinguistics, especially in problems of societal multi-lingualism, linguistic diversity, language change, language acquisition, and language contact in the broadest sense. She edited volume IV (1776 to 1997) ofThe Cambridge History of the English Language (1998).suzanne.romaine@ling-phil.ox.ac.uk

Elena SeoaneElena Seoane is Associate Professor of English in the Department of English at the University of Santiago de Compostela, and a member of the researchproject “Variation and Linguistic Change,” directed by Professor TeresaFanego. In 1996 she received her PhD in English historical linguistics from theUniversity of Santiago. Her main areas of specialization are syntactic changein the history of English, historical pragmatics, and grammaticalization theory.iaelena@usc.es

Sali A. TagliamonteSali A. Tagliamonte is Associate Professor of Linguistics and Director of theSociolinguistics Laboratory at the University of Toronto, Canada. Her exper-tise is language variation and change in synchronic corpora. She has publishedon African American English, British and Irish dialects, teen language and tele-vision. Ongoing research focuses on morph-syntactic developments in Englishthrough cross-variety and apparent time comparisons.sali.tagliamonte@utoronto.ca

xvi Notes on Contributors

Ann TaylorAnn Taylor is a Research Fellow at the University of York (UK). Her researchinterests include syntactic variation and change, particularly in Old andMiddle English. She is co-creator with Anthony Kroch of the Penn–HelsinkiParsed Corpus of Middle English; with Anthony Warner, Susan Pintzuk, andFrank Beths, of the York–Toronto–Helsinki Parsed Corpus of Old EnglishProse; and is currently producing a parsed version of the Corpus of Early EnglishCorrespondence with Warner and Pintzuk, in collaboration with the ResearchUnit for Variation and Change in English, Helsinki.at9@york.ac.uk

Ingrid Tieken-Boon van OstadeIngrid Tieken-Boon van Ostade is Senior Lecturer in the Department ofEnglish at the University of Leiden. She specializes in the field of historical socio-linguistics (social network analysis), has published on periphrastic do andmultiple negation, and has recently obtained a major research grant for her pro-ject “The Codifiers and the English Language.” The aim of this project is totrace the origins of the norm of standard English as well as to study the influenceof prescriptivism on actual usage.i.m.tieken@let.leidenuniv.nl

Elizabeth Closs TraugottElizabeth Closs Traugott is Professor Emeritus of Linguistics and English atStanford University, California. Early work included research in historicalsyntax, socio-historical linguistics, and linguistics and literature. In the last twodecades her prime research focus has been on grammaticalization, lexicaliza-tion, and semanticization of conversational inferences and subjectification, andshe is currently studying the syntax–pragmatics interface in Old English. Her publications include A History of English Syntax (1972); Linguistics forStudents of Literature (with Mary L. Pratt, 1980); On Conditionals (1986; co-editedwith Alice ter Meulen, Judith Snitzer Reilly, and Charles A. Ferguson);Grammaticalization (with Paul Hopper, 1993, 2nd much revised edn. 2003); andLexicalization and Language Change (with L. Brinton, forthcoming). She has beena Guggenheim Fellow, a Fellow at the Center for Advanced Study in theBehavioral Sciences, and is a Fellow of the American Association for theAdvancement of Science.traugott@csli.stanford.edu

Anthony WarnerAnthony Warner is Professor of English Linguistics at the University of York(UK). He has written on the history of English syntax and on monostratal syntax. His interests include the study of variation and the creation and useof electronic parsed corpora, both seen as major components of a historical syntactician’s tool kit.aw2@york.ac.uk

I Approaches and Issues

1 Change for the Better?Optimality Theory versusHistory

APRIL McMAHON

Change is not made without inconvenience, even from worse to better.Dr Johnson, quoting Richard Hooker

1.1 Introduction

Optimality Theory (OT) may have been described by Archangeli (1997: 1) as“THE Linguistic Theory of the 1990s,” but it is clearly making a bid, at leastin phonology, for continuation of that favored status into the twenty-first cen-tury. OT analyses have now been developed for a vast range of phonologicalphenomena in the languages of the world; and as the architecture of OT haschanged, and the edifice has grown, so competing possibilities have been proposed. The OT phonological community has also reached a point of gen-erally acknowledging the strengths of the model (mainly in prosody, in theenlightening modeling of the interaction of competing motivations, and in theintegration of typological evidence into phonology), and also its weaknesses(mainly the handling of opacity, specifically in morphophonological alterna-tions; and in some areas of persistent inclarity about central tenets of the the-ory, notably the role of Gen, the shape of the input, and whether constraintsare universal).

In this chapter, I shall try to highlight some of these issues as they affect theOT analysis of sound change. Historical linguists, who may well begin from asituation of relative theoretical neutrality, will be persuaded to make use of aparticular formal model only if they are convinced that it offers the possibil-ity of providing more enlightening analyses of change than would be possiblein the absence of the model. It is therefore important to show what OT claimsto be able to say about change, and how the OT approach has changed duringthe (relatively short) life of the theory.

This chapter consequently has two goals, though their development will beinterwoven rather than sequential. On the one hand, I shall show that the OTapproach to sound change has altered considerably since the first applicationsof the model to diachronic questions. Early OT accounts tended to assume thatchange was, in some sense, for the better, notably in the analysis of emergence

4 April McMahon

of unmarked effects; and these analyses tended also to be presented as a changefor the better over other possible accounts in different models, in providingdirect explanations of change. However, more recently, some of these claimshave been diluted, so that not all historical linguists now working in OT neces-sarily expect or attempt explanations of change, but rather a neat method ofmodeling conflicting factors and motivations. In a sense, this means the OTapproach to change is now less concerned with the details of, and theimpulses behind individual changes, and more involved with a global view:hence, Bermúdez-Otero and Hogg (2003) focus on OT modeling of the “life-cycle” of phonological processes, while Minkova and Stockwell (2003) demon-strate that various rankings of four constraints give rise to different, thoughperhaps related, types of change.

This concentration on macroanalysis of language history might itself be seenas a change for the better; however, in the second theme of this chapter, I shallshow that some of the areas of inclarity in the current formulation of OT areof particular relevance and concern for analyses of change. For instance, thewell-known question over the acceptability of different constraints and con-straint types means it can be hard to evaluate alternative accounts of the samechange, and this overlaps with the unresolved question of the universality ofconstraints. Perhaps even more seriously, recent questions over the possiblephonetic grounding of phonological constraints are of particular relevance tosound change, if we accept that changes often have their inception in phono-logizations of automatic, phonetically motivated processes. Many of thesepoints can be made with reference to alternative OT analyses of a single, cen-tral change, namely the Great Vowel Shift, which will therefore be the focusof much of the discussion below.

1.2 Optimality Theory and Sound Change: The Basic Ingredients

For the benefit of those readers who have not yet acquainted themselves withthe (sometimes slightly arcane) paraphernalia and impedimenta of OT, a verybrief outline of the model might be helpful (and see also Further Reading, below).OT is based around the idea of a set of constraints, Con: initially, and still insome current versions, these constraints are seen as universal and innate, butcrucially violable. For each input, or mental lexicon form postulated on the basisof primary linguistic data, all possible parses, or outputs, are generated by amechanism called Gen. These alternative, competing parses are then evaluatedby the ranked set of constraints; the parse which violates fewest high-rankingconstraints will win and be produced. However, although the constraints areuniversal, their ranking is not, and differences between languages therefore resultfrom a different prioritization of the same constraints.

We might as historical linguists be faced with two sister languages, one ofwhich has an onset in every syllable, while the other often has onsets, but not

Change for the Better? 5

in absolutely every case. These onset consonants routinely match; but where thesecond language has no onset consonant, the first consistently has a glottal stop. In OT terms, we conclude that in the first language (1a), the universalconstraint Onset (which requires all syllables to have onsets) is ranked so high that it will always be obeyed on the surface. If there is no onset in thecommon ancestral form, or in a borrowed form, an onset is inserted ex nihiloin the form of a glottal stop. In the second sister (1b), although the same con-straint is inherited by children, it is less important than another constraint (normally stated as Dep-io in current versions of OT), which requires the out-put to be faithful to (that is, essentially the same as) the input, and whichspecifically forbids the insertion of segments. So, in this case, onsets are welcomed wherever they are already available in the input, but the higher ranking of Dep-io as compared to Onset means it is more important to keep the one-to-one relationship between input and output than to resort toepenthesis and satisfy the requirement for an onset consonant. In the first language, the priorities are the other way around.

(1) a./atel/ Onset Dep-io

[?atel] E *[atel] *!

b./atel/ Dep-io Onset

[?atel] *![atel] E *

The very great advantage of this constraint-ranking approach is its resolu-tion of a venerable problem in linguistics, namely the existence of universalpressures which do not always create absolutely universal results. We can saythat languages typically prefer to have filled onsets; but there is a continuumfrom those languages (like Arabic) that absolutely require them, through thosewhere onsets will be allowed, although speakers do not go out of their way toproduce them in every case. The status of universal tendencies of this kind wasunclear in earlier models, but they are a straightforward and predictable partof OT, where the violability and rankability of constraints create different out-puts depending on language-specific priorities. We can also model the rangeof results provided by all possible rankings of a set of constraints, and shouldideally find that each ranking provides an attested output: we shall return belowto an illustration of this so-called factorial typology from Minkova andStockwell (2003).

Given that different rankings will produce different outputs, a neat andstraightforward means of modeling change presents itself, and it is not

6 April McMahon

surprising that OT phonologists seized on this relatively early in the develop-ment of the theory: “Under OT, the formal characterization of language changethrough time is that constraints are reranked” (Archangeli 1997: 31). That is,taking our “toy” example of the onset-requiring language, and its sister whichcan take onsets or leave them, we might assume that the latter represents thehistorical situation, and that the first sister has subsequently undergone a reranking whereby Onset is now more highly ranked than Dep-io. In otherwords, at some point a discontinuity arose between generations, such that parents arguably had a grammar with one ranking, and their children insteadacquired an alternative grammar with the two relevant constraints ranked the other way round. Although this is undoubtedly a pleasing analysis, andhas the added advantage of following very directly from the OT account ofsynchronic, typological variation, it does raise the problem of explanation; andthis is a key issue to be explored in the next section. Armed with a very basicoutline of the architecture of OT, and the consequent model of change, we nowturn to a more complex change and some alternative means of modeling it inconstraint-based terms.

1.3 Optimality Theory and the Great VowelShift

1.3.1 The Great Vowel Shift: one thing after another,or just one thing?

A diagram bearing at least a close resemblance to (2) below is almost an expectedingredient of survey courses on the history of English, and the Great VowelShift, which it portrays, is highly likely to figure on the average top ten list ofsound changes in English, or indeed Germanic.

(2) i: u:↑ ↑

ei e: o: ou↑ ↑E: O:↑a:

As presented, this change involves the diphthongization of the high vowels(which subsequently continue to lower to their present-day values of /aI/ and/aU/, and the stepwise raising of the long monophthongs. Later changesobscure aspects of this pattern, since Middle English (ME) /a:/ does notremain at /E:/, but continues to raise to /e:/ (now often diphthongal [eI], asin name), while ME /E:/ itself raised two steps to /i:/ in most cases (sea, beat;exceptions are great, steak), merging therefore with ME /e:/ (green, sleep) which

Change for the Better? 7

had itself raised by one step to /i:/. Other changes following the Great VowelShift seem to fit into the same pattern – so, as ME /u:/, having diphthongizedin the early stages of the Shift, continued to lower to [aU], earlier /au/ raisedand monophthongized in words like law, filling the low mid slot vacated by/O:/ when it earlier raised to /o:/ (boat, coat).

As this brief description has already shown, it is very hard to write aboutthe Great Vowel Shift (GVS), or indeed any chain shift, without thinking interms of systems containing specific slots. From these slots some elements willshift, leaving space for others to move in (the classical description of a dragchain); alternatively, elements already moving may put pressure on others, whichare forced into a consequent shift away to avoid merger (giving rise to the traditionally termed push chain). Either way, the core of the GVS seen in thisway is “a particular kind of chain shift, in which segments in a given phonolo-gical subspace play musical chairs and don’t go anywhere” (Lass 1988: 397).Indeed, Lass (1976) used the GVS specifically to argue for the inclusion andrecognition of systems and inventories in phonological theory. Intriguingly, healso proposed a series of metarules and output conditions to analyze large-scaleshifts of this kind; a “metarule” which effectively says vowels should raise seemspresciently close to the “vowels are high” type of OT constraint.

Lass notes that “probably the majority of historians of English . . . havecome to accept the GVS as something that can be talked about, on a par withother reifications of diachronic correspondence-sets like Grimm’s, Verner’s,Grassman’s Laws, the High German obstruent shift, and the like” (1988: 396);elsewhere, and indulging in a little reification himself, he characterizes such“major sound changes” as having “left the languages they affected very dif-ferent from what they were before, either in inventory (Grimm’s Law, the HighGerman Consonant Shift), or in morphophonemic complexity (the GreatVowel Shift)” (1976: 52). But although this acceptance of the unity of the GVSmight be fairly general, it is not uncontested. Indeed, a slightly unorthodoxbut closely fought game of academic tennis has been going on over this pointfor the last 20 years: on one side of the net is Lass (1976, 1988, 1992, 1999), whoargues for the unity of at least part of the conventionally described GVS, andon the other are Stockwell and Minkova (Stockwell 1975; Stockwell andMinkova 1988a, 1988b, 1990, 1997, 1999; Minkova 1999), who believe that the“GVS” is a misleading label for a series of changes which are essentiallyunconnected except in being examples of the kind of changes that routinelyhappen in the history of English, because English is the sort of language wherethat kind of change routinely happens.

Obviously, a challenge of this kind to the integrity of the GVS is a majorissue (not least for this chapter, where I propose to discuss several accounts ofa possibly non-existent change): as Lass (1992: 145) puts it: “For nearly a cen-tury the exegesis of the GVS has been one of our major cottage industries,”but “Certainly questions about the causes, beginnings, and structure of some-thing that did not happen are rather thin stuff for serious scholars to be spend-ing their time on.” In this context, it is reassuring to note that Stockwell and

8 April McMahon

Minkova do not deny the individual changes often grouped together as theGVS: “The facts, indeed, seem quite clear; the question is how to make senseout of them” (Stockwell 1975: 333). In other words, all parties to the discus-sion would agree that the changes in (3) can reasonably be postulated for the history of English; the difference is that Lass would also accept the geo-metrical representation in (2), which shows (some of) the same individual changes as linked.

(3) ME EModE ModEi: → ei → aIe: → i: → i:E: → e:a: → E: → e: → eIu: → ou → aUo: → u: → u:O: → o: → oU

In such an extensive series of arguments and counterarguments, there areof course further detailed points of disagreement, some of them fundamentalto the analysis of the individual changes. For instance, while both Stockwelland Minkova and Lass agree that the original ME high vowels diphthongizedto some intermediate form before lowering to /aI aU/, they propose differentintermediate forms. Furthermore, Lass (following Luick’s original suggestion)argues that the first step in the GVS was raising of the high mid vowels /e:/and /o:/, and provides dialectal data in evidence: wherever /o:/ had alreadymoved out of the back vowel system through a northern change of /o:/-Fronting,leaving an “empty space” in the back column, /u:/ failed to diphthongize, giving Scots hoose, moose rather than house, mouse. Lass and Luick take the linethat, since there is no parallel disparity with the front vowels, and no soundchange which specifically targeted /e:/, the fronting of /o:/ and the failure of/u:/ to diphthongize must be connected. In other words, the GVS began as apush chain, as the raising high mid vowels put pressure on the high ones, whichconsequently diphthongized: in cases where no pressure was applied from below,the diphthongization did not happen. Stockwell and Minkova, on the other hand,argue that the change of diphthongization (which for them is a much lower-level process, a merger of possibly pre-existing diphthongal variants of the highvowels with pre-existing diphthongs from other sources) came first. It is alsopossible to detect a degree of rapprochement in later contributions to the debate,with Lass (1992: 149) maintaining that the “top half” of the traditional GVS(that is, the combination of high vowel diphthongization and high mid vowelraising) shows an inner cohesion and unity, but accepting that the “bottom half”might well be a series of independent vowel raisings which should be referredto as “pseudo-GVS” or “Post-GVS Raising” rather than Phase II of the GVSproper (a term he reintroduces, however, in Lass 1999). But the core of the dis-agreement remains the same now as in the earliest contributions. On one view,

Change for the Better? 9

either part or all of the GVS is a unified series of events, which “can be andhas been shown to have what Luick . . . called an ‘innere Zusammenhang’ ” (Lass1992: 148). On the other, there are several relatively minor and self-containedchanges, which are due to the same underlying motivations of maximizing theperceptual optimality of diphthongs, or maintaining contrast; but these takeplace over too long a period to be recognized as forming a unitary shift, andthe generally accepted GVS pattern is simply a “linguist’s creation through hind-sight” (Stockwell and Minkova 1988a: 376). In that case, “the massive GVSbecomes a classic example of structuralist overgeneralization more suited fora history of linguistics than for a history of the English language” (Minkova1999: 84).

It would be presumptuous to assume that I could reach any definite con-clusion on this debate, and more so in the space available here. On the otherhand, a contested change, or series of changes, of this kind does provide anexcellent focus for a discussion of the OT approach to change. As we shall see,there are several alternative analyses of the GVS in the literature, and it mightbe instructive to see whether modeling these change(s) in OT terms helps usreach a view on their relatedness. That is, we shall be concerned below withthe efficacy of OT in modeling sound change; the several different approachesto the GVS each highlight a number of key issues on this matter. However, weshall also consider what might justify a selection of OT over other possible models in the analysis of sound change, and a relevant issue here is whetheror not OT might help us decide on the most reasonable view of the “GVS.”

1.3.2 Miglio (1998)Miglio (1998), like most early attempts to apply OT to sound change, isengaged with explanations: “Optimality Theory . . . can be . . . an importantmeans to explain language change” (1998: 1). A typical early analysis of soundchange in OT, simplifying for the sake of argument (see, further, McMahon2000a, 2000b, 2002), might model the loss of a relatively marked segment viaa markedness or well-formedness constraint disfavoring that segment . At StageI, this constraint is ranked low, and speakers do not attend to it; they get bypronouncing the difficult segment somehow. But at Stage II, the markednessconstraint has been reranked higher, and the difficult segment is no longer pro-nounced. Sound change therefore involves a gradual prioritization of marked-ness constraints over time, and its consequence is a shift towards less markedpronunciations (though changes may have effects elsewhere in the grammarwhich let other cases of markedness in). As Minkova and Stockwell (2003: 169)put it:

Optimality Theory is the ultimate capitalist economy. It conceives not just of changebut of all phonetic realizations as the result of competition. The basic competitionfor change that has been envisaged is competition between markedness constraintsand faithfulness constraints. The scenario is, the faithfulness constraints are

10 April McMahon

constantly battered by the markedness constraints; the latter are constantly climb-ing the corporate ladder and dislodging the entrenched corporate management.

The most significant difficulty for such early OT accounts of sound changeinvolves the nature of reranking, the process of advancement of markednessconstraints up the ladder in the quote above, and crucially whether this mechanism can be seen as explanatory: this is the chicken and egg problem of McMahon (2000a: 126–7). Why, for instance, did the loss of [k] in [kn-] clusters (know, knee, knight) happen? Does reranking happen in responsemode, allowing children to build the simplest grammar, since they (and quitepossibly their parents) were already not pronouncing [k]? Or is the rerank-ing causal? Does it take place first, for whatever reason, so that the next generation has no choice but to produce [n-], even where they hear [kn-] in their primary linguistic data? From the point of view of language acquisition and what we know about the sociolinguistic input to language change, the latter is, of course, much less plausible than the former: the causalaccount of reranking might be appealing since it allows phonological theoryto explain change, rather than having to rely on factors outside the grammaritself, but what it gains in this respect it loses very decisively in its reliance onmystical grammatical determinism. This seems to be the consensus in currentOT treatments of change: better to see reranking as post hoc grammar tidyingand as an effect of acquisition, however prosaic this might seem, than claimreranking is explanatory where this really only pushes the real explanation onestep back.

Miglio (1998) attempts to extend analyses of this kind to the GVS, which initself is a tricky proposition: chain shifts are a notorious locus of opacity, a majordifficulty for OT accounts of synchronic phonology, and as changes they arelogically impossible to deal with using only markedness, since by definitiondisfavoring a particular vowel, which might therefore be encouraged to raiseor diphthongize, would lead us to expect that vowel to remain disfavored –not for the same slot to be refilled by another vowel shifting in from some-where else. Miglio’s solution (without going into full details on this or the otheraccounts below, for reasons of space) is to identify “weak spots” in the systemas probable triggers of change; “weakness” here might equate to markednessor to perceptual difficulty. In the case of the GVS, Miglio identifies /E: O:/ asthe “weak spots,” because long lax mid vowels are typologically rare, and there-fore highly marked. In fact, these vowels had been in the system for some time,but Miglio argues that they became more frequent and therefore more notice-able as outputs of the preceding sound change of Middle English OpenSyllable Lengthening. The higher frequency of these vowels then triggered aconstraint reranking promoting the constraint *[–ATR]mm, which disfavors longlax vowels (ATR, or Advanced Tongue Root, is here used to signify thelax–tense opposition). One might expect this to have caused vowel mergers;but the constraint Distance, which is also quite high-ranked, favors the main-tenance of contrasts.

Change for the Better? 11

Miglio’s identification of the long lax low mid vowels as “weak spots,” which consequently shift, and her invocation of constraints discouragingmergers, account for the raisings from low mid to high mid, and high mid to high, and for the diphthongization of the ME high vowels. However, Miglio also includes /a:/ > /E:/ as part of the shift – although /E:/ is one of the disfavored vowels which started the whole process. Here, Miglio must assume a further re-reranking, demoting *[–ATR]mm, and promoting a local conjunction of faithfulness constraints for [ATR] and height (we return to the issue of local conjunction in the next section); /a:/ then raises to low mid.

This is potentially an interesting result: after all, Stockwell and Minkova seethe “GVS” as composed of a series of independent changes, and Miglio’s inabil-ity to model the whole sequence without further reranking might square withtheir position. However, Stockwell and Minkova, and Lass, agree that the ele-ments of the traditional Shift which are most likely to show unity are thoseinvolving the high and high mid vowels: neither suggests that diphthongiza-tion plus the raisings of both mid vowel series qualify as part of a unitary change,without the raising of /a:/, yet this is what Miglio’s model would suggest. Itis worth noting here that Ahn (2002) provides a closely parallel account toMiglio’s, proposing that “the overall chain shift could have been initiated bythe raising of lax vowels for phonetic reasons. Then, the subsequent raisingand diphthongisation were consequences of Maintain Contrast requiring thatthe input vowel contrast be maintained in the output” (Ahn 2002: 153). Ahnalso divides the Great Vowel Shift into two shifts, each stage characterized bya different constraint ranking; again, the second stage involves the raising of /a:/, and this time also the further lowering of the diphthongs from ME /i: u:/, to [aI aU].

There are several difficulties with Miglio’s account. First, seeking the moti-vation for the GVS in the output of Middle English Open Syllable Lengthening(MEOSL), which, Miglio asserts, produced a superfluity of long lax low midvowels, is potentially problematic given that MEOSL preceded the GVS byapproximately 300 years – one of Stockwell and Minkova’s objections to theGVS as a unitary change involves the more than 150 years between the topand bottom halves. Reranking also seems to be used fortuitously, and in par-ticular there seems little motivation for the temporary rehabilitation of /E:/,which has to be seen as less marked purely so low /a:/ can shift to it; thoughsince, of course, there was a further step later, post-GVS, which raised this new/E:/ to high mid /e:/, Miglio is going to have to re-re-rerank the constraintsagain. It is hard to see reranking as in any sense explanatory under these circumstances. Finally, if /E: O:/ are the weak points in the system which motivate the change, we must ask why only the front one is lost through theGVS, while its highly marked back congener continues into the present-daysystem. Further issues, involving novel mechanisms like local conjunction, andthe “grounded” nature of general constraints like Distance, will be pursuedbelow.

12 April McMahon

1.3.3 Miglio and Morén (2003)

Miglio and Morén (2003) carries over two main issues from Miglio’s earlier ana-lysis of the GVS. Miglio and Morén also deal only with the diphthongizationsof the high vowels, and the single-step raisings from low mid to high mid andhigh mid to high. They do not include the raising of /a:/; but as with Miglio’saccount, this seems to produce a strange split on either the Lass or Stockwelland Minkova view: if we attempt to model only Phase I or the “top half” ofthe traditional GVS as a unitary change, then the shifts of low mid /E: O/ shouldnot be included; but if we include those later raisings, as Miglio and Moréndo, there seems no good reason not to include the raising of /a:/ too.

Miglio and Morén argue that the English vowel system, leading up to theGVS, developed complete predictability of vowel length for all but the longmid vowels, which had developed a tense/lax distinction. Like Miglio (1998),though proposing different constraints, they ascribe the vowel shift to the disfavoring of long lax mid vowels, and an avoidance of merger. Essentially,Miglio and Morén see the GVS as both a lengthening and a raising change.Contextual lengthening takes place in certain environments, including openmonosyllables (unfortunately, Miglio and Morén decline to discuss the rationale for this predisposing change: “The motivation for lengthening is notimportant here and is not shown” (2003: 206)). For the most part, this lengthen-ing only involves tense vowels; but there is also “a combination of constraintsforcing some mid vowels to be long” (2003: 206), so that both lax and tensemid vowels will undergo conditioned lengthening. Miglio and Morén then invokea family Deplink-mora of moraic faithfulness constraints, which spell out whathappens when a segment of a particular type is forced to carry an extra mora,as in the coerced lengthening (“motivation . . . not important”) of lax midvowels. Assuming that Deplink-mora [rtr, high, low] (where rtr isRetracted Tongue Root) outranks Deplink-mora [high, low], “it is worse toadd a mora to a surface lax mid vowel than it is to add a mora to a surfacetense mid vowel” (2003: 209). Consequently, lengthened /E/ must also tenseto /e:/, raising perceptually. In turn, the pre-existing /e/, when lengthened,has to raise to /i:/ – this time because it is worse to add a mora to any midvowel than to add a mora to a high vowel, as Deplink-mora [high, low]outranks Deplink-mora [high]. Raising rather than lowering is guaranteedbecause Ident [high] outranks Ident [low] among the faithfulness constraints.

There is one final piece of the puzzle, which is the fact that /E/ becomes/e:/, rather than raising all the way to high – the usual chain shift paradox.Miglio and Morén assume a Local Conjunction of (Ident [rtr] & Ident[low]). As Kager (1999: 392) puts it, “Under Local Conjunction, two constraintsare conjoined as a single constraint which is violated if and only if both of itscomponents are violated in the same domain.” This conjoined constraint uni-versally outranks its component constraints. In a chain shift, a double raisingwill violate more faithfulness constraints, like the Ident ones considered here,