english prosodic morphology...nar on prosodic morphology at the university of marburg, taught by...

ENGLISH PROSODIC MORPHOLOGY

This book has been published in cooperation with the editors of the journalMorphology, also published by Springer. The aim of Morphology is to publish highquality articles that contribute to the understanding of morphological phenomena andto the articulation of morphological theory and linguistic theory in general. The editorsalso function as intermediaries in the publication of monographs sharing the aims andambitions of the journal. This monograph is the first result of this arrangement.

Geert BooijJennifer HayIngo Plag

ENGLISH PROSODICMORPHOLOGY

SABINE LAPPEUniversity of Siegen

Germany

Library of Congress Control Number: 2008920293

ISBN 978-1-4020-6005-2 (HB)ISBN 978-1-4020-6006-9 (ebook)

Published by Springer,P.O. Box 17, 3300 AA Dordrecht, The Netherlands.

www.springer.com

Printed on acid-free paper

All Rights Reserved© 2007 Springer Science+Business Media B.V.

No part of this work may be reproduced, stored in a retrieval system, or transmitted in any form or by any means, electronic, mechanical, photocopying, microfilming, recording

or otherwise, without written permission from the Publisher, with the exception of any material supplied specifically for the purpose of being entered

and executed on a computer system, for exclusive use by the purchaser of the work.

This work is a revised version of the author’s doctoral dissertation submitted to theUniversity of Siegen in August 2005 in partial fulfillment of the requirements for the degree

of Doctor of Philosophy (Dr. phil.).

CONTENTS

PREFACE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . XI

SYMBOLS AND ABBREVIATIONS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . XIII

INTRODUCTION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1

1. ENGLISH PROSODIC MORPHOLOGY: MORPHOLOGY . . . . . . . . . . . 31.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31.2 A Phonological Theory of a Morphological Process . . . . . . . . . . . . . . . 5

1.2.1 Prosodic Morphology – The Theory . . . . . . . . . . . . . . . . . . . . . . . 51.2.2 Problems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71.2.3 English Prosodic Morphological Processes . . . . . . . . . . . . . . . . . 101.2.4 English y-hypocoristics: Further Issues . . . . . . . . . . . . . . . . . . . . . 14

1.3 English Prosodic Morphology and Word Formation . . . . . . . . . . . . . . . 151.3.1 The Form of the Output . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 161.3.2 The New-word Status of Truncated Words . . . . . . . . . . . . . . . . . . 20

1.4 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29

2. ENGLISH PROSODIC MORPHOLOGY: PROSODY . . . . . . . . . . . . . . . . 312.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 312.2 Descriptive Approaches to Name Truncation and Word

Clipping . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 312.2.1 Sundén's (1904) Analysis of Name Truncation . . . . . . . . . . . . . . . 322.2.2 Kreidler's (1979) Analysis of Word Clipping . . . . . . . . . . . . . . . . 342.2.3 Simpson (2001): Truncated Place Names in

Australian English . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 352.3 Recent Accounts with a Focus on Linguistic Modelling:

Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 352.3.1 Output Structure in Unsuffixed Truncation . . . . . . . . . . . . . . . . . . 362.3.2 Output Structure in Suffixed Truncation . . . . . . . . . . . . . . . . . . . . 37

2.4 The MinWd Approach . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 402.4.1 The Central Tenets . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 402.4.2 Problems with the MinWd Approach:

Quantity-sensitivity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 442.4.3 Problems with the MinWd Approach:

Extrametricality . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 472.5 Correspondence: Truncated Words and their Bases . . . . . . . . . . . . . . . 502.6 Accounting for the Structure of English Truncations:

The Program . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54

v

VI CONTENTS

3. THE PATTERNS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 593.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 593.2 Methodology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59

3.2.1 The Basic Rationale of the Project . . . . . . . . . . . . . . . . . . . . . . . . 593.2.2 The Main Corpus on Name Truncation:

Phillips (2002) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 613.2.3 Supplementary Data on Name Truncation . . . . . . . . . . . . . . . . . . 613.2.4 The Corpus on Word Clipping . . . . . . . . . . . . . . . . . . . . . . . . . . . 63

3.3 The Inventory: Name Truncation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 663.4 The Inventory: Word Clipping . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 693.5 How many Bases, How many Derivatives? . . . . . . . . . . . . . . . . . . . . . . 723.6 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 74

4. THE STRUCTURE OF MONOSYLLABIC TRUNCATED NAMES . . . . 774.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 774.2 Basic Syllable Structure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77

4.2.1 Onset, Nucleus, Coda . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 774.2.2 VV or VVC? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 78

4.3 Consonant Clusters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 804.3.1 General Cluster Variability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 804.3.2 Mixed Clusters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 84

4.3.2.1 [r]-initial Clusters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 854.3.2.2 [l]-initial Clusters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 874.3.2.3 Nasal-initial Clusters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 89

4.3.3 Obstruent Clusters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 914.3.3.1 Stop-stop and Fricative-stop Clusters . . . . . . . . . . . . . . . . . . 924.3.3.2 Stop-fricative Clusters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 94

4.3.4 Intermediate Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 954.4 Truncated Forms and Their Bases – Segment Stability . . . . . . . . . . . . . 97

4.4.1 Vowel Changes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 974.4.2 Consonant Changes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 97

4.5 Truncated Names and Their Bases – the Material Copied . . . . . . . . . . 1014.5.1 The Contiguity of Segments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1014.5.2 Anchoring . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 103

4.6 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 106

5. THE STRUCTURE OF Y-HYPOCORISTICS . . . . . . . . . . . . . . . . . . . . . . 1075.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1075.2 Basic Syllable Structure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1075.3 Consonant Clusters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 109

5.3.1 General Cluster Variability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1095.3.2 Mixed Clusters Rising in Sonority . . . . . . . . . . . . . . . . . . . . . . . . 1125.3.3 Mixed Clusters Falling in Sonority . . . . . . . . . . . . . . . . . . . . . . . . 114

5.3.3.1 [r]-initial Clusters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 115

CONTENTS VII

5.3.3.2 Nasal-initial Clusters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1165.3.3.3 [l]-initial Clusters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 118

5.3.4 Obstruent Clusters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1195.3.4.1 Stop-stop and Fricative-stop Clusters . . . . . . . . . . . . . . . . . . 1205.3.4.2 Stop-fricative Clusters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 122

5.3.5 Intermediate Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1235.4 Truncated Forms and Their Bases – Segment Stability . . . . . . . . . . . . . 126

5.4.1 Vowel Changes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1265.4.2 Consonant Changes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 131

5.5 y-Hypocoristics and Their Bases – The Material Copied . . . . . . . . . . . 1325.5.1 The Contiguity of Segments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1335.5.2 Anchoring . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 134

5.6 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 137

6. THE STRUCTURE OF CLIPPINGS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1396.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1396.2 The Structure of Monosyllabic Clippings . . . . . . . . . . . . . . . . . . . . . . . . 139

6.2.1 Basic Syllable Structure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1396.2.2 Consonant Clusters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 141

6.2.2.1 Mixed Clusters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1426.2.2.2 Obstruent Clusters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 145

6.2.3 s-suffixation: Morphological Effects on Syllable Structure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 147

6.2.4 The Relation Between Clippings and Their Bases: Segment Stability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 148

6.2.5 Anchoring in Monosyllabic Clippings . . . . . . . . . . . . . . . . . . . . . 1516.2.6 Intermediate Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 152

6.3 The Structure of y-suffixed and o-suffixed Clippings . . . . . . . . . . . . . . . 1526.3.1 Basic Syllable Structure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1526.3.2 Consonant Clusters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 154

6.3.2.1 Clusters in y-suffixed Clippings . . . . . . . . . . . . . . . . . . . . . . . 1546.3.2.2 Clusters in o-suffixed Clippings . . . . . . . . . . . . . . . . . . . . . . 156

6.3.3 Suffixed Clippings and Their Bases: Segment Stability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 157

6.3.4 Suffixed Clippings and Their Bases: Anchoring . . . . . . . . . . . . . . 1576.3.5 Intermediate Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 158

6.4 The Structure of Unsuffixed Disyllabic Clippings . . . . . . . . . . . . . . . . . 1586.4.1 General Properties of the Process . . . . . . . . . . . . . . . . . . . . . . . . . 1586.4.2 The Stress Pattern of the Base and the Clipping: Anchoring . . . . 1606.4.3 The Structure of the Final Syllable: Segmental Restrictions . . . . 1626.4.4 Intermediate Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 163

6.5 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 163

VIII CONTENTS

7. CONSTRAINT INTERACTION: WORD STRUCTURE . . . . . . . . . . . . . . 1677.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1677.2 The Preservation of Prominent Material: Faithfulness . . . . . . . . . . . . . 171

7.2.1 Effects in Truncation and Related Processes . . . . . . . . . . . . . . . . . 1717.2.2 Formalisation: MAX-P . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 174

7.3 Word Structure Characterised by Prominence: Markedness . . . . . . . . . 1787.4 Unsuffixed Truncation – the Typology . . . . . . . . . . . . . . . . . . . . . . . . . . 1807.5 Suffixed Truncation: Prominence and Affix Retention . . . . . . . . . . . . . . 1847.6 Further Restrictions: Anchoring in Name Truncation . . . . . . . . . . . . . . 1907.7 Further Restrictions: Stress in Unsuffixed Disyllabic Clipping . . . . . . . 1977.8 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 201

8. CONSTRAINT INTERACTION: SYLLABLE STRUCTURE . . . . . . . . . . 2078.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2078.2 Basic Syllable Structure in Monosyllabic Forms . . . . . . . . . . . . . . . . . . 209

8.2.1 CONSONANT and Resultant Gaps in Truncation . . . . . . . . . . . . . . . 2098.2.2 Syllable-weight and Consonants . . . . . . . . . . . . . . . . . . . . . . . . . . 2168.2.3 Extrasyllabicity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 220

8.3 Suffixed Disyllabic Forms: Complex Onsets . . . . . . . . . . . . . . . . . . . . . 2268.3.1 Mixed-cluster Onsets in the Final Syllable

of y-suffixed Truncation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2278.3.2 Medial Obstruent Clusters in y-suffixed Truncation:

Complex Onsets? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2288.3.3 The Peculiarities of o-suffixed Clippings . . . . . . . . . . . . . . . . . . . 231

8.4 Restrictions on Syllable Structure in Unsuffixed Disyllabic Clippings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 233

8.5 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 236

9. CONSTRAINT INTERACTION: CLUSTER PHONOTACTICS . . . . . . . . 2419.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2419.2 Mixed Clusters: the General Hierarchy . . . . . . . . . . . . . . . . . . . . . . . . . 244

9.2.1 The Generalisations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2449.2.2 Analysis: Markedness Hierarchies

and Faithfulness Constraints . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2479.3 Mixed Clusters: Additional Constraints on Monosyllables . . . . . . . . . . 250

9.3.1 The Avoidance of Voiced C2s . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2519.3.2 The Absence of the Voice Effect in Clipped Words . . . . . . . . . . . 255

9.4 Mixed Clusters: Additional Constraints on Suffixed Disyllables . . . . . . 2559.4.1 The Sonority Effect . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2579.4.2 The Homorganicity Effect . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 260

9.5 Mixed Clusters with a Special Status: [rC] and Nasal-Stop Clusters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2639.5.1 Exceptional Clusters, Part 1: The Antics of [rC] . . . . . . . . . . . . . 263

9.5.1.1 The Special Status of [r] . . . . . . . . . . . . . . . . . . . . . . . . . . . . 263

CONTENTS IX

9.5.1.2 [r]'s Faithfulness to its Structural Role . . . . . . . . . . . . . . . . . 2669.5.2 Exceptional Clusters, Part 2: The Antics of Nasal-stop . . . . . . . . 267

9.5.2.1 The Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2689.5.2.2 Analysis: The Unmarked Status

of Homorganic Nasal-stop Clusters . . . . . . . . . . . . . . . . . . . . 2719.6 Intermediate Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2749.7 Obstruent Clusters in Monosyllables and Disyllables . . . . . . . . . . . . . . 2779.8 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 284

10. CONSTRAINT INTERACTION: SEGMENTAL CHANGES . . . . . . . . . 28710.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28710.2 Schwa-Deneutralisation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28910.3 Avoidance of the Dental Fricative . . . . . . . . . . . . . . . . . . . . . . . . . . . . 291

10.3.1 Data and Generalisation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29110.3.2 Analysis: Markedness Versus Featural Identity . . . . . . . . . . . . . 292

10.4 Avoidance of Word-final [r] . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29510.4.1 Data and Generalisation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29510.4.2 Analysis: Markedness vs. CONTIGUITY . . . . . . . . . . . . . . . . . . . . 297

10.5 Intermediate Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29810.6 Vowel Laxing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30010.7 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 302

CONCLUSION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 303

BIBILIOGRAPHY . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 307

AUTHOR INDEX . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 321

SUBJECT INDEX . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 325

PREFACE

In the past years I have spent much of my time writing what in the end has becomea very long book about very short words. What has emerged is a complex, but alsoa clear and empirically founded picture of the structural properties of English trun-cated words as well as an optimality-theoretic model of these properties in theresearch tradition of Prosodic Morphology. These will be of use to linguists inter-ested in the structure of English as well as to those interested in phonological and /or morphological theory. I have deliberately kept separate the empirical analysis andthe theoretical account of the data, so that the book can be used by scholars workingwithin OT as well as by those who do not. Finally, the style as well as the structureof the text are such that the book may not only have an academic readership, but mayalso be recommended as supplementary reading in pertinent university courses atboth undergraduate and graduate levels.

The project had its origins in a small analysis of a few structural aspects ofEnglish truncated names that I did as part of the requirements for a graduate semi-nar on Prosodic Morphology at the University of Marburg, taught by Ingo Plag andBirgit Alber. Back then, truncated names to me were more or less just 'short', and Ithought I could describe them in terms of their number of syllables and, in the rarecases in which there happened to be more than one, in terms of stress assignment.This has changed over the past few years. In this book, which is a revised (andabridged) version of my doctoral dissertation, not much is left of the original analy-sis. Instead, thousands of truncated words have taught me that, as everywhere in life,things are not as simple as they may seem. Working with the data has left me inamazement at the richness of the story that short words are able to tell us about theirexistence, about the nature of the language that they come from, about the intrica-cies of research in language, and, last but not least, about linguistic theory.

Although only my name is given as the author's name on the front cover, thisbook would not have been written without the support and encouragement of a lotof people who deserve a whole-hearted thank-you. I was very lucky to have learntfeeling enthusiastic about English linguistics from Rüdiger Zimmermann, and to bedragged into phonological and morphological theory by Birgit Alber and Ingo Plag.To Ingo and Birgit I owe a lot for constant encouragement, patience, and a tremen-dous amount of time and care spent on reading and discussing my drafts. Ingo hasbeen an awsome supervisor and friend, from whom I have learnt a lot. ChristianGrau, Christina Kellenter, Gero Kunter, Henner Metz, Taivi Rüüberg, MareileSchramm, Ute Wagner, Josephine Thomschke and Linda Zirkel, all from our Siegenteam, have carefully proofread and discussed various draft versions. I would espe-cially like to thank Taivi for the amount of time and care spent on the manuscript,much beyond the call of duty, as well as for her insightful comments on many of thepassages, which have helped improve the text both in form and in content.

xi

XII PREFACE

Furthermore, I am grateful for the constructive feedback I have received from par-ticipants of various linguistic colloquia and conferences both at home and abroad,especially from Geert Booij, Armin Mester, Iggy Roca, Christian Uffmann, andRichard Wiese, as well as from the members of my dissertation committee, RüdigerZimmermann, Gerhard Augst, and Franz-Josef Klein. Needless to say, all remainingerrors are my own. Finally, I would like to thank those people who have provided mewith loads of moral support, helping me to keep my chin up in rough times as wellas sharing moments of joy, always showing an interest in (or at least patiently lis-tening to) stuff that other people may consider completely abstract: Ute, Mareile,Gero, Maria, Sabine, Christian, Holger, Jörg, Martin, Hartwig, Taivi, Robert, and thewhole 'lunch gang', as well as my family, who have always known that this will takelong, but also that everything will go well in the end. Last but not least, Lutz, whostarted out as my boyfriend and is now my husband, was always there, was incredi-bly patient, and is now half a linguist himself. This is his book, too.

Sabine Arndt-Lappe

Siegen, December 2006

SYMBOLS AND ABBREVIATIONS

A. General Terms and Symbols

BT (correspondence) Base – Truncation (correspondence)CHAID Chi-squared Automatic Interaction Detector

(a model used to build classification trees for nominaldata, implemented in the SPSS Answer Tree®

software package)IO (correspondence) Input – Output (correspondence)IT (correspondence) Input – Truncation (correspondence)OED Oxford English DictionaryOT Optimality Theorythe Prosodic Hierarchy:

- Wd prosodic word- MinWd minimal prosodic word- ft foot- � syllable- � mora

UR Underlying RepresentationYate’s �2 Yate’s correction for continuity, applied in �2 tests

with less than 2 degrees of freedom

B. Transcription Symbols for Segmental and Suprasegmental Structure

[�l] phonetic transcription['�l] stressed syllable�i� orthographic representation(�) metrical foot�� extrametrical syllableH, L heavy, light syllableS, w phonetically strong, weak syllableC consonantC1, C2 first, second consonant in a clusterV lax vowelVV tense vowel or diphthong

xiii

INTRODUCTION

It is not by chance that the quantitative method of modern science wasborn and reared in Anglo-Saxon countries. The Anglo-Saxon mind isin the main a practical one, and such a method as word-clipping is pri-marily the work of a practical-minded nation.

(Marchand 1960: 363f.)

The purpose of this book is twofold: On a broad empirical basis, it will present a sys-tematic structural analysis of what Marchand refers to as word clipping in English.Secondly, it will introduce an optimality-theoretic model that accounts for theseproperties.

Other terms that scholars have used to refer to outputs of word clipping includestump words (Jespersen 1965repr.), truncations, diminutives, and, if the base form isa personal name, hypocoristics. Word clipping in Marchand’s terms is in fact a coverterm for a variety of different truncatory processes that exist in English. As is clearfrom the title of this book, I will consider these processes to be instances of ProsodicMorphology (McCarthy & Prince 1986, 1993a et seq.). These comprise, for exam-ple, hypocoristics like Barty (� Bartholemew) and clippings of common nouns likelab (� laboratory), comfy (� comfortable), and legit (� legitimate).

In most standard accounts of English word-formation, not much attention isdevoted to truncation (e.g. Marchand 1960, Jespersen 1965repr., Adams 1973, Bauer1983, Adams 2001, Carstairs McCarthy 2002; but cf., Plag 2003 for a differentview). One of the reasons is that outputs of truncation are traditionally held to bestructurally unpredictable. The little attention, however, stands in sharp contrast tothe fact that, as is universally acknowledged, truncation is extremely frequent andproductive (Bauer & Huddleston 2002: 1635) in modern English. Already Jespersen(1965repr) points out that stump words are

probably nowhere quite so numerous as in ME [Modern English, S.L.]and in fact constitute one of the most characteristic traits in the devel-opment of the English language in its recent stage.

(Jespersen 1965repr.: 550f.)

A detailed empirical investigation of the structural basis for the productivity of trun-cation in English is, however, still a desideratum.

In contrast to morphological theory, recent phonological theory has developed alarge interest in truncation. In particular, reduplication and truncation have becomeimportant test cases for the discipline of Prosodic Morphology (McCarthy & Prince1986, 1993a et seq.). For truncated forms, studies in Prosodic Morphology claim thatsize restrictions are most adequately described in terms of prosodic categories. Withrespect to English, however, no detailed account of the structural properties has asyet emerged from Prosodic Morphology. So far studies have only focussed on thegeneralisation that (some!) truncatory patterns in English conform to the minimal

1

2 ENGLISH PROSODIC MORPHOLOGY

word (McCarthy & Prince 1986, 1998, Weeda 1992, Benua 1995, 1997). Only veryfew other aspects have been touched.

On a theoretical level, it is an open question in how far the theory can beextended to cover more aspects of truncation than minimal-word size. This is par-ticularly true for more recent developments of Prosodic Morphology, where it isargued that output structures are not determined by templates, but by optimality-theoretic constraint interaction, which crucially involves high-ranking markednessconstraints (McCarthy & Prince 1994a, b et seq.). An optimality-theoretic model ofEnglish Prosodic Morphological processes is not only an important test case for thishypothesis; it also has implications for views on English phonology. If English trun-catory processes are a product of constraint interaction, the study of truncated formscan contribute important insights to our understanding of marked word structure inEnglish.

Based on a systematic empirical analysis comprising more than 3,000 truncatedforms and their bases, I will show in this book that the unpredictability claim raisedin the word-formation literature is untenable. Specifically, the inventory of Englishtruncatory patterns consists of several structurally distinct classes which can bedescribed in terms of distinct phonological characteristics. In the optimality-theoreticanalysis I will argue that the structure of English truncated forms indeed reveals muchabout word-level markedness in English.

The text is subdivided into three large sections. The first section (chapters 1 and 2)provides the theoretical basis for the analysis. Chapter 1 is concerned with the status oftruncatory processes in Prosodic Morphology and word-formation theory. Chapter 2provides a review of previous studies of the phonological characteristics of Englishtruncated forms. In particular, I will show that the central tenet of OT-accounts ofProsodic Morphological phenomena, which views size restrictions on truncated formsas the effect of a specific set of prosodic markedness constraints, is not viable forEnglish. The final section of chapter 2 contains a detailed program for the empiricalinvestigation and the theoretical model to be proposed in the chapters to follow.

Chapters 3–6 are concerned with the empirical analysis. Chapter 3 introducesthe corpora and the methodology, and provides an overview of the basic truncatorypatterns as they are represented in the corpora. Chapters 4–6 are each devoted to aparticular class of truncations (chapters 4–5: name truncation, suffixed and unsuf-fixed; chapter 6: other clipped words). Chapters 4–6 have a parallel structure, so asto facilitate a comparison between truncatory patterns.

In chapters 7–10 I present the optimality-theoretic model. The analysis is struc-tured in terms of the representational level concerned (chapter 7: the whole word;chapter 8: the syllable; chapter 9: consonant clusters; chapter 10: the segment).

1. ENGLISH PROSODIC MORPHOLOGY:MORPHOLOGY

3

1.1 INTRODUCTION

The aim of this chapter is to introduce and define the class of phenomena that willbe analysed in this book. We will adopt the perspective of morphological theory,focussing on two major questions:

• What can the research discipline of Prosodic Morphology (McCarthy &Prince 1986, 1993a et seq.) contribute to a definition of truncatory patternsin English?

• What is the status of truncated words in the English word formation system?

With respect to the former question, we will see in section 1.2 that English, likeother languages, does not draw a clearcut distinction between 'ordinary' morpho-logical processes and Prosodic Morphological processes. Instead, we find thatmorphological processes differ in terms of the degree to which their outputstructure is determined by prosodic wellformedness constraints. Truncated wordsare at the extreme end of the scale. The term Prosodic Morphology will be used inthis book as a descriptive category to refer to this end of the scale. The finding thatthere is no clear dividing line between Prosodic Morphological processes andother morphological processes is in line with recent developments in the researchdiscipline.

The second question posed above directly addresses the view widely held inthe literature that truncated forms are both structurally and semantically unpre-dictable. Results from recent empirical studies render this view questionable.Evidence comes from the investigation of the structural properties of truncatedforms in this book as well as from recent studies of the meaning and function ofsuch forms (most notably, Schneider 2003). The alleged structural and semanticunpredictability of truncated forms have led scholars to assume that truncationis not word formation. We will conclude that also this view is in need of revision.

The data in (1) illustrate the processes we will be concerned with: truncatedpersonal names, y-hypocoristics, unsuffixed clippings, y-clippings, and o-clippings.

(1) English Prosodic Morphological processes

a. unsuffixed truncated names:Pete (� Peter)Trish (� Patricia)


b. y-hypocoristics:Pety (� Peter)Trishy (� Patricia)

c. unsuffixed clippings:fab (� fabulous)exec (� executive)

d. y-clippings:daffy (� daffodil)veggie (� vegetable)

e. o-clippings:delo (� delegate)intro (� introduction)

The terminology adopted here largely follows the terminology used within theresearch tradition of Prosodic Morphology (McCarthy & Prince 1986 et seq.), wherethese processes are across the board referred to as truncation processes. Other termsused for truncated forms in the literature are stump words (Jespersen 1965repr) orclippings (e.g. Marchand 1960, Kreidler 1979). For reasons of convenience I willreserve the term clipping for truncated words which are not personal names. I will usethe term truncated names to refer to the forms exemplified in (1.a). The distinctionbetween clippings and truncated names is well grounded for English in the fact that,as we will see in this book, the two classes of processes exhibit important structuraldifferences. The label y-hypocoristics has been chosen for the forms in (1.b) on thebasis of the terminology found in descriptive grammatical accounts (e.g. Bauer &Huddleston 2002).

The processes exemplified in (1.b), (1.d), and (1.e) involve the suffixes -y and -o.Apart from -y and -o, English also has other suffixes which may be classified asProsodic Morphological. -s and -[ ] (spelled -a or, in nonrhotic dialects, -er) are casesin point (Schneider 2003: 109ff. on British English and Simpson 2001 on AustralianEnglish). Examples are provided in (2).

(2) a. s-suffixed forms:Gabs (� Gabrielle)Pabs (� Pablo)Cuts (� Cutler)

b. [ə]-suffixed forms:Jimma (� Jimmy)Micka (� Mickey)Ecker (� Eric)

(data from Simpson 2001: 106f.)

e

MORPHOLOGY 5

For reasons of space, the analysis in this book will be confined to y-suffixed ando-suffixed processes.

1.2 A PHONOLOGICAL THEORY OF A MORPHOLOGICAL PROCESS

To state it, provocatively, the field is ultimately self-destructive

(Kager & Zonnefeld 1999: 14)

1.2.1 Prosodic Morphology – The Theory

The origins of Prosodic Morphology as a research discipline lie in the study of non-concatenative morphology, most notably in the study of root-and-pattern morpho-logical systems and reduplication (McCarthy 1979, McCarthy & Prince 1986, 1990).Contrary to previous, essentially segmentalist, approaches, Prosodic Morphologyput forward the claim that what is constitutive of these processes is morphologicalgrammar making crucial reference to prosodic categories (i.e. prosodic word, foot,syllable, and mora), but not to the segmental content of the morpheme. The ProsodicHierarchy, which comprises these categories and which forms the basis of work inProsodic Morphology, is schematised in (3).

(3) the Prosodic Hierarchy

(Selkirk 1980a,b, McCarthy & Prince 1986, 1993a et seq.)

prosodic word

foot

syllable

mora

Prosodic categories are assumed to play a role not only in delineating the size andstructure of (reduplicative and truncatory) templates, but also in determining thelocus of affixation, leading to infixational patterns (analysed as effects of ProsodicCircumscription in McCarthy & Prince 1990). Truncatory patterns in English aswell as in other languages have attracted much interest in Prosodic Morphology.Thus, for example, the invariant monosyllabic structure of English simple trun-cated names has been used in McCarthy & Prince (1986) as one of the key exam-ples to show that the prosodic category of the minimal word is attested to delineatethe structure of a template. Analyses of truncatory patterns in other languages have


been proposed, for example, by Féry (1997, German), Itô & Mester (1997,German), (Wiese 2001a, German), Piñeros (1998 et seq., Spanish), Scullen (1997,French), and Nelson (1998, French).

In the initial phase of the project (McCarthy & Prince 1986, 1990), the mainthrust of the argument has been to advocate the role of prosodic categories, and toargue against segmentally-based analyses. Furthermore, the focus of both researchmethodology and interest is clearly on phonological, not on morphological theory.The data introduced in McCarthy & Prince 1986 as core areas of interestfor Prosodic Morphology reflect this focus, comprising not only 'unproblematic'morphological processes like reduplication and templatic morphology, but alsoprocesses whose status as morphological processes is at least debated; among thoseprocesses we find some truncatory processes (nickname formation, abbreviation,Argot) and – particularly – language games and secret languages (cf., e.g., Itô et al.1996 for Japanese, Scullen 1997 for French, Piñeros 1998 for Spanish).

The adoption of an optimality-theoretic framework by Prosodic Morphology(McCarthy & Prince 1993a) led to an interesting new conception of the nature ofthe processes studied in the field. Under parallel constraint evaluation in OT, thebasic insight that in templatic processes morphological categories make referenceto prosodic categories is translated into a problem of alignment of morphologicaland phonological structure. Alignment is determined by constraint interaction, i.e.the interaction of morphological constraints and phonological constraints onprosodic structure. The role of prosodic categories in template and reduplicantformation is modelled as the effect of the domination of some phonologicalconstraint P over a morphological constraint M. In reduplicative and truncatorytemplatic morphology, the interaction between the P and M constraints is an inter-action of constraints making conflicting claims about the size of the reduplicant,with 'M' being the MAX-constraint and 'P' being a constraint on the alignment of acategory of the Prosodic Hierarchy and the morphological category 'reduplicant'. Incircumscriptional morphology, the P-constraint can be either a constraint whichdemands the alignment of the morphological category 'affix' with a phonologicalcategory in the base word (cf., e.g., the Ulwa possessive, where the locus of inser-tion for the possessive affix is determined by the foot structure of the base word,McCarthy & Prince 1993a: 106), or it can be a general phonological constraint ofthe language which makes reference to a phonological category of the ProsodicHierarchy (e.g. NOCODA in Tagalog um- infixation, which makes reference tosyllable structure, McCarthy & Prince 1993a: 120). The M-constraint aligns a givenaffix with the edge of its base; as a result, the ranking P �� M determines the locusof insertion of the affix.

Crucially, with the adoption of OT the defining characteristic of ProsodicMorphological processes is a specific ranking schema (cf., McCarthy & Prince1993a et seq.):

P (honological / prosodic constraint) �� M (orphological constraint)

MORPHOLOGY 7

It is this core ranking which is shared by the phenomena studied in the framework.These phenomena are templatic and circumscriptional morphology. ProsodicMorphology itself is claimed to be a general theory about the interaction of mor-phological and phonological factors in the processes studied:

Prosodic Morphology (McCarthy and Prince 1986 et seq.) is a theory ofhow morphological and phonological determinants of linguistic forminteract with one another in a grammatical system. More specifically,it is a theory of how prosodic structure impinges on templatic andcircumscriptional morphology, such as reduplication and infixation.

(McCarthy & Prince 1993a: 1)

Further developments in the framework have focussed on the more detailed investi-gation of the nature of the phonological constraints that determine the structure oftemplates. While templates are first conceived of as effects of alignment constraintsaligning morphological and phonological categories, later versions of the theory(Generalised Template Theory, McCarthy & Prince 1994a, b et seq.) largely focuson the substitution of the earlier 'templatic' alignment constraints through generalphonological markedness constraints which are not specific to a morphological cat-egory anymore. The structure of the realisation of a morphological category is nowassumed to be determined by a specific correspondence relation between the tem-platic morpheme (e.g. the reduplicant or the truncated form) and its base. What hadbeen modelled as alignment of phonological categories and morphological cate-gories before, is now interpreted as an effect of a correspondence relation that is spe-cific to the morphological category and its base (BT- or OO-correspondence;McCarthy & Prince 1994a, 1995, 1997, Benua 1995, 1997).

This approach, however, constitutes a fundamental change in the conception oftemplates: Their structure is not anymore determined by a process making refer-ence to prosodic categories, but they are effects of prosodic markednessconstraints being free to exert their influence on the structure of the morpheme. Inother words: The factor that determines the structure of templates is not anyprosodic category, but a prosodically unmarked category; templates are unmarkedprosodic categories (word, foot, syllable) of their language and, thus, instantia-tions of the Emergence of the Unmarked (TETU, McCarthy & Prince 1994a). Thisdevelopment also led to a loss of some of the original empirical basis of ProsodicMorphological phenomena – most notably, clearly templatic processes like root-and-pattern morphology ceased to be core processes to be accounted for withinProsodic Morphology.

1.2.2 Problems

Seen against the development of the research discipline and the changes it has under-gone, an exact definition of what should count as a Prosodic Morphological processseems problematic. In fact, given the more recent conception of ProsodicMorphology as a general theory of the phonology-morphology interface, such a


definition does not even seem desirable.1 If, however, we accept this generalisedconcept of Prosodic Morphology, we run into different problems. These problemsconcern the relationship between Prosodic Morphology and morphological theory.In what follows I will discuss two issues which I think are problematic for anchor-ing Prosodic Morphology in morphological theory:

a. The Emergence of the Unmarked is nothing new in morphological theory.However, TETU effects in morphology are not confined to the categoriesof the Prosodic Hierarchy.

b. Not all phonology-morphology interaction can be accounted for in termsof TETU effects.

The first issue is closely related to the fact that Prosodic Morphology has its originsin phonological theory. Interestingly, however, the ranking 'P �� M' is not onlycharacteristic of Prosodic Morphological processes like reduplication or templaticmorphology, but also of other affixational processes. Crucially, however, the phono-logical constraints involved in affixation are not confined to constraints that makereference to categories of the Prosodic Hierarchy. They can also be constraints whichoperate on the segmental or subsegmental level of representation, i.e. levels whichProsodic Morphology says nothing about. Assimilation processes in English nomi-nal plural formation are a good example (cf., e.g., Russell 1997 for an analysis).

Alongside segmental markedness constraints, however, also prosodic marked-ness constraints have been shown to play an important role in affixation (cf., e.g.,Plag 1999, Rosenthall 1999, Rowicka 1999, Wiese 2001d). Strikingly, the label ofProsodic Morphology is used in some of these studies, while in others it is not.

For example, Plag (1999: 172ff.) shows that stem-allomorphic effects that areobservable in English -ize derivatives are triggered by high-ranking phonologicalmarkedness constraints on the rhythmic structure of the output. Concatenation of -izeleads to the deletion of base-final vowels if the following conditions are met: The basemust be dactylic and it must end in a vowel other than schwa. The prosodic constrainttriggering the deletion of the base-final vowel is *LAPSE, militating against stresslapses in the derivative. Examples are provided in (4).

(4) stem allomorphy in -ize derivatives with dactylic bases: (Plag 1999: 172ff.)

memory memorize *memoryizesummary summarize *summaryize

1 Note in this context that work in Prosodic Morphology has never explicitely claimed that there exists aclear conceptional difference between two classes of morphological processes – instead, scholars havestressed the nature of the theory as a theory about the interaction of phonology and morphology in inter-face phenomena.

MORPHOLOGY 9

Rhythmic optimisation of the output of a morphological process has also beenclaimed to be a determining factor in the selection of the appropriate allomorph invarious derivational and inflectional processes in German (e.g. -n vs. -en plural for-mation, ge- vs. zero prefixation in verbs, cf. Wiese 2001d). While, however, inWiese (2001d) the above processes are explicitely categorised as ProsodicMorphological processes (cf. the title of the relevant section of the article: Prosodicmorphology in complex words – allomorphy in affixation), the concept is neithermentioned nor discussed in Plag (1999). This situation is symptomatic of the wayin which interface phenomena have been viewed in the literature from differentdisciplines, phonological or morphological. Investigating the role of prosody in thespecification of lexical units in German, the goal of the investigation in Wiese(2001d) is clearly phonological in nature. The analysis of what Wiese callsProsodic Morphological phenomena – i.e. phenomena in which allomorphic effectsare sensitive to prosodic structure – serves as evidence in favour of the generalclaim of the paper, which is a claim about the nature of the phonological represen-tation of lexical items.2 By contrast, Plag (1999) looks at stem-allomorphic effectsof -ize derivatives in the context of a general investigation of the phonological prop-erties of these derivatives (pp. 145–194). The phonological constraints which heproposes to interact with morphological constraints in forming -ize derivativesmake reference to prosodic as well as to segmental categories. Thus, for example,base truncation as a stem-allomorphic effect can be triggered not only by prosodicmarkedness (rhythmic optimisation, cf. above), but also by segmental markednessconstraints. For example, a segmental OCP constraint is active in -ize suffixation,which prevents ungrammatical formations like *feminin-ize in favour of attestedfemin-ize (Plag 1998, 1999: 185ff.). The argumentation focusses on showing thatphonological wellformedness influences the structure of the derivatives rather thanon establishing on which level of phonological representation – segmental orprosodic – this interaction takes place.

Thus, while Wiese's argumentation is in line with the original formulation of thegoal of Prosodic Morphology as a research program that set out to investigate the natureof phonological representation (McCarthy & Prince 1986; cf. above), the analysispresented in Plag 1999 shows that, when it comes to the morphological analysis ofallomorphic effects, there may not necessarily be a qualitative difference between thephonological constraints interacting in the processes: Constraints making reference toprosodic categories are found here alongside those making reference to segmentalcategories. It seems that, while Prosodic Morphology constitutes an innovation withinphonological theory in the sense that it proposes that the categories of the ProsodicHierarchy all constitute valid categories of phonological representation (contrary toearlier segmentalist approaches and similar to what Kahn 1976 had already proposedfor the syllable as a category), the status of Prosodic Morphological processes as acategory of processes within morphological theory appears to be rather questionable.

2 For another example of the diversity of phenomena treated within Prosodic Morphology cf. a collectionof papers from a workshop bearing the title 'Prosodic Morphology' (held in June 1994 at UtrechtUniversity), which was published under the title of 'The Prosody-Morphology Interface' (Kager &Zonneveld eds. 1999).


The second problematic issue that was listed above concerns the claim that tem-plates constitute unmarked prosodic categories of their language (McCarthy & Prince1994a, b). The theory is based on the assumption that in template formation themarkedness ranking of the prosodic constraints which come to exert their influence isthe same as in the language as a whole. Furthermore, Emergence-of-the-Unmarkedeffects observable in templatic morphology are due to the existence of a particularcorrespondence-relationship between the templatic output and its base. But as wehave just seen, also nontemplatic, affixational processes which exhibit allomorphicalternations may refer to prosodic markedness constraints. Some of the proposals putforward in the literature to account for these alternations have, however, deviatedfrom the idea that the general markedness ranking of the language is at the heart ofthe alternation. For example, it has been assumed that alternations serve as signals toa morphological category (cf., e.g., Alderete 2001, Plag 2002), that complex wordsare subject to their own markedness ranking which constitutes a cophonology partic-ular to the paradigm (Inkelas 1996, Orgun 1996), or that these alternations are simplyidiosyncratic and are encoded in base-specific constraints in the grammar of the lan-guage (Russell 1999). While for allomorphic effects we find this variety of differentproposals, for reduplicative and truncatory templatic morphology it is an open ques-tion whether or not the original claim made in Generalised Template Theory is borneout crosslinguistically. If this is the case, then this would mean that templaticprocesses differ substantially from other morphological processes in their respectivelanguages in that they always trigger Emergence-of-the-Unmarked effects, whereasamong other morphological processes, some do and some do not.

1.2.3 English Prosodic Morphological Processes

In this section we will compare English truncatory processes to other English morpho-logical processes; we will see that we can use the term Prosodic Morphology only as adescriptive category to refer to one end of a scale of phenomena that differ in terms ofthe degree to which constraints on prosodic structure determine the shape of the output.As examples, I will consider -ize derivation (Plag 1999, cf. above), comparativeformation (McCarthy & Prince 1986: 2), and so-called 'expletive infixation' (e.g.Minne-fuckin-'sota, McCarthy 1982, Hammond 1997a, 1999: 161–164). Examples areprovided in (5). Examples of truncatory patterns are for convenience repeated in (6).

(5) a. -ize derivation (Plag 1999, repeated from (4))memory memorize *memoryizesummary summarize *summaryize

b. comparative suffixation (McCarthy & Prince 1986: 2, modified)red redder *more redyellow yellower ?more yellowobtuse *obtuser more obtuseexpensive *expensiver more expensiveconventional *conventionaller more conventional

MORPHOLOGY 11

c. expletive infixation (Hammond 1997a: 46)

'Minne'sota

'Minne-fuckin-'sota *Mi-fuckin-nesota

'Winnepe'saukee

'Winne-fuckin-pe'saukee *Winnepesau-fuckin-kee

'Winnepe-fuckin-'saukee

'Tim

'buk'tu

'Tim-fuckin-

'buk'tu

(6) a. truncated namesAlfreda AlfCamille CamElisabeth Liz

b. y-hypocoristicsAlfreda Alfy *AlfredyCamille Cammie *CamillyElisabeth Lizzy *Elisabethy

In all processes exemplified, prosodic categories influence the way in which mor-phological categories are realised. In truncated names (6), prosodic influence takesthe shape of a restriction on the structure of the output, which is invariant for all expo-nents of the morphological categories 'truncated name' and 'y-hypocoristic'. Thisinvariant output structure has repeatedly been noted by scholars as a characteristic oftruncated words in English (cf., e.g., Sundén 1904: 51, Jespersen 1965repr.: 539,Marchand 1960: 351), but it has only quite recently been given a formal treatment,when a framework became available to describe invariant output structures in termsof prosodic templates (McCarthy & Prince 1986, Weeda 1992). While truncatednames are monosyllabic in English, y-hypocoristics are disyllabic, with stress on thefirst syllable. This invariant output structure is realised regardless of the prosodicstructure of the names that form the bases of the truncated forms. Whereas, forexample, Camille has two syllables, Alfreda has three syllables and Elisabeth has foursyllables, the truncated forms are all mono- (6.a) or disyllabic (6.b), respectively.3

Prosodic categories also influence the shape of the outputs of the affixationalprocesses in (5) above. As in name truncation, the prosodic categories that play a roleare the syllable and stress. In -ize derivation (5.a), prosodic restrictions ban forms whichcontain a stress lapse and vowel hiatus across morpheme boundaries (*memory-ize),whereas in comparative suffixation (5.b) the function of prosody is to determine whichbases qualify as grammatical bases for -er. The suffix -er is attached only to bases that

3 The fact that the truncated form is largely independent of its base in terms of prosodic structure becomesparticularly obvious if one compares the syllabification of consonants in the base and in the truncatedform. The syllable role of a consonant in the base name is to a large extent irrelevant for its syllable rolein the truncated form. Thus, for example, [m] in Camille forms the onset of the second syllable' ([kə.'mil], but is syllabified in the coda in the truncated word ([kæm]); in Alfreda, [f] is also part of theonset of the second syllable ([æl.fri.də]), but word-final in the truncated word ([ælf]). The independenceof the truncated form from its base in terms of prosodic structure stands of course in a sharp contrast tothe relation of total dependence between the two forms in terms of segmental structure; every truncatedform contains segmental material from its base.


are either monosyllabic (e.g. red � red-er) or disyllabic (e.g. stupid � stupid-er). A fur-ther restriction on disyllabic bases is that main stress must be on the initial syllable(hence, stupid � stupid-er, but obtuse � *obtus-er). With all other bases a periphrasticconstruction is used (for a more detailed account of the complexities involved in Englishcomparative formation cf., e.g., Mondorf 2003 and references therein).

Expletive infixation (5.c) exemplifies yet another way in which prosody-levelphonological categories can influence the realisation of a morphological category. Theanalysis sketched here follows the accounts proposed by McCarthy (1982) andHammond (1997a, 1999). In Hammond's analysis (1999: 163f.), expletive infixation issensitive to the foot structure of the base: the infix must not be inserted within a footin the base. Hence, for example, the base Winnepesaukee, for which Hammond postu-lates the foot structure (

'Winne)pe('saukee), offers two possible loci of insertion for the

infix. The infix may either precede or follow the unfooted syllable -pe-, rendering theoutput either (

'Winne)-(

'fuckin)-pe('saukee) or (

'Winne)pe-(

'fuckin)-('saukee).4

Apart from the similarities just outlined, the processes exemplified in (5) and (6)also differ in important aspects. Crucial differences concern

a. the degree to which prosody shapes the morphological category, and

b. the representation (base or derived form) to which prosodic structuralrequirements relate.

Prosodic categories are most influential in the formation of simple truncated names(6.a). In these cases prosodic structure serves as the only signal to the morphologicalcategory of the word; without the prosodic restriction to a monosyllable, the outputwould be indistinguishable from the base name. By contrast, in -ize derivation, com-parative suffixation, and expletive infixation, prosody is accompanied by other elementswhich serve to signal morphological category. All three processes are concatenative –the status of the output as a complex word is signalled by the concatenation of elements.

Furthermore, the processes illustrated in (5) and (6) differ in that only truncationleads to an invariant output structure. Unlike truncated forms, suffixed comparativeforms may, for example, be disyllabic (redd-er) or trisyllabic (stupid-er), and prod-ucts of expletive infixation not only vary greatly in terms of their number of syllables,but also in terms of stress patterns (cf., e.g., Minne-fuckin-'sota vs. in-fuckin-flu'enza,which are both derived from quadrusyllabic bases, Hammond 1999: 164).

4 Note that, of course, the analysis is heavily dependent on the metrical framework adopted. Thus, forexample, the admission of trisyllabic feet into the system would render Hammond's generalisation inad-equate; if -pe- in (Winnepe)('saukee), is part of the initial foot and not unparsed, the form (Winne)-(fuckin)-pe('saukee) constitutes a violation of Hammond's generalisation that the infix should not interruptmetrical feet in the base. In the same way a foot structure that would assign a monosyllabic foot to themain-stressed syllable -sau- would pose problems for Hammond's account: If the foot structure ofWinnepesaukee is (Winnepe)('sau)kee (whereby the final syllable may be either extrametrical or footed),Hammond's generalisation cannot explain why the infix may not be inserted after the main-stressed syl-lable, producing the unattested form *Winnepesau-fuckin-kee.

MORPHOLOGY 13

Given these differences, I will use the term Prosodic Morphology to refer to mor-phological processes whose prosodic output structure is invariant. Note that this isexactly the original defining characteristic of classic templatic morphology (e.g.McCarthy 1979, Marantz 1982). However, I will show in the chapters to follow thatEnglish truncatory processes are not necessarily templatic. We can explain their struc-ture without assuming the existence of templates. This is parallel to what studies withinProsodic Morphology have found to be true for a large group of truncatory processes inother languages (cf., e.g., Féry 1997, Itô & Mester 1997, Wiese 2001a for German,Piñeros 1998 et seq. for Spanish, Scullen 1997, Nelson 1998 for French). ProsodicMorphology will serve as a term to refer to morphological processes which exhibit thekind of invariant output structure that has traditionally been associated with templates.

Note that in this sense the term Prosodic Morphology cannot be more thandescriptive. I do not claim that Prosodic Morphological processes share a generalunderlying phonological principle which sets them apart from other morphologicalprocesses. As shown above, they share with other morphological processes theP �� M ranking schema. They also share with other morphological processes thatthey may be concatenative. This is illustrated in the y-hypocoristics in (6.b), whichexhibit an invariant output structure, but which additionally signal morphologicalcategory through the suffix -y. We will see in section 1.3 that the formal composi-tionality of the process is mirrored by its meaning; y-hypocoristics do not have thesame meaning as their unsuffixed counterparts.

A final argument against a categorical dividing line between English ProsodicMorphology and other morphology can be derived from the fact that the invariantoutput structure of English truncatory processes is not always the result of theEmergence of the Unmarked in word structure. As we will see in chapter 7 of thisbook, word structure of English truncations is often the result of a compromisebetween phonological markedness and faithfulness. I will argue that suffixed trun-catory patterns like those in (6.b) are a case in point; an even clearer case is found inthe patterns illustrated below.

(7) English unsuffixed disyllabic clippingsceleb (� celebrity)exec (� executive)ident (� identity)

Main stress is on the final syllable of the truncated form in all examples in (7).This word structure cannot be explained in terms of unmarked prosodic wordstructure in English, but can be accounted for if we incorporate high-ranking faith-fulness constraints into our analysis. If, however, the output structure of truncatedforms is a compromise between phonological markedness and faithfulness, thentruncation does not fundamentally differ from other morphological processes.Within morphological theory, similar compromises have long been known as basetruncation (cf., e.g., Aronoff's 1976 analysis of English -ee suffixation on Englishverbs ending in -ate, as in nominate � nominee, *nominate-ee). As the analysis of -ize


derivatives in Plag (1999) shows, base truncation can be due to markedness con-straints on prosodic structure (cf. above). We conclude that English ProsodicMorphological processes do not form a homogeneous class of processes in termsof the phonological principles that determine their structure.

1.2.4 English y-hypocoristics: Further Issues

With respect to suffixed truncatory patterns, two further issues are in need of clari-fication. These two issues concern the suffix -y. First of all, -y will be referred to asa suffix in this book in spite of the fact that some scholars have expressed doubt that-y is a conventional suffix. The issue has mainly been taken up in the literature on aparallel, i-suffixed truncation process in German, but the argument can be trans-ferred to English. Thus, for example, the observation that truncated forms corre-spond to an invariant output structure leads Féry (1997) to assume that i-suffixationin German is not a case of derivation in the narrow sense ('Derivation im engstenSinne des Wortes', Féry 1997: 465). This assessment is based on the assumption thatmorphological derivation does not usually impose restrictions on the phonologicalstructure of the output in German. A second argument can be found in Wiese (2001a:149f.), who points out that -i is different from other suffixes in that it does not func-tion as the grammatical head of the derived form. In my view, both arguments do notsuffice to exclude English -y or German -i from the inventory of the languages' suf-fixes. The fact that -y and -i do not assume all grammatical head functions is explain-able if we analyse -y and -i as diminutive suffixes (cf. section 1.3.3 of this chapter).Lack of grammatical head function is one of the properties standardly noted aboutdiminutive suffixes in various languages (cf., e.g., Dressler & Merlini Barbaresi1994 on Italian).

The second issue to be clarified here concerns the orthographic representation ofEnglish -y. Following Schneider (2003) and the earlier literature, I assume that -y hasthree major orthographic variants. Examples of the major orthographic variants of -yare provided in (8).

(8) orthographic variants of -y (examples from Schneider 2003: 87f.)

�y�Charly (� Charles)botty (� bottle)

�ie�Charlie (� Charles)bottie (� bottle)

�ey�Charley (� Charles)lovey (� love)

MORPHOLOGY 15

Schneider (2003: 87f.) shows that differences in orthographic representation are sub-ject to preferences in different varieties of English. Crucially, however, all variantsare represented in all varieties. For reasons of convenience I will use �-y� as thedefault orthographic representation in this book. Thus, if the �-y� variant of a trun-cated form is represented in my corpora, I will use this form as the citation form.Other forms will only be used if �-y� is not attested.

1.3 ENGLISH PROSODIC MORPHOLOGY AND WORD FORMATION

Talkshow Host: Now Frank – Fran – Frannie – little Frannie-pooh . . .Guest: No. I'm leaving. I'm off. I'm going. I've never . . .(exits)

(Monty Python's Flying Circus, episode 1 (1969): 'It's the Arts')

Whereas truncatory processes have been investigated in Prosodic Morphology in aresearch tradition which primarily focuses on problems of phonological representa-tion, models of word formation within morphological theory have been rather reluc-tant to incorporate truncatory processes into their frameworks. I will argue in thissection that this reluctance is largely unnecessary.

In their most general form, models of word formation make reference to threefactors which are assumed to be constitutive of word formation processes. A changeof word form serves as a signal to a change of word meaning. The changes in formand meaning are mediated through a particular relation between the base form andthe derived form. This relation may be a rule (e.g. Aronoff's 1976 Word-FormationRule, WFR), or, in Optimality Theory, a set of ranked constraints. The factorsinvolved in word formation are broadly schematised in (9). Note that I abstract awayfrom major issues which are debated in the literature, such as the divide betweenword-based and morpheme-based models.

(9) Word-formation – schema (broadly modelled on Plag 2003: 21, 184ff.)

form

meaning

form

meaning

base form: derived form:

WFR

With respect to all three factors (form, meaning, relation between base and derivative),truncatory processes have been argued to differ from other word formation processes.On these grounds scholars have assumed that truncation processes are outside mor-phological grammar. The main reasons for excluding truncatory patterns from wordformation are listed in (10). I have grouped them according to the relevant factor.


(10) Problems for a word formation approach to truncation

a. problems related to the form of the derivative

• Truncations do not exhibit a complex (i.e. multi-morphemic) internal word structure(item-and-arrangement models of word formation, Hockett 1958, Marchand 1960).

• The output of truncatory processes is unpredictable, whereas word formation asa grammatical process produces predictable outputs (cf., e.g., Jespersen1965repr., Marchand 1960, Kreidler 1979, Bauer 1983, Dressler & MerliniBarbaresi 1994, �Stekauer 1998, 2000, Dressler 2000).

b. problems (largely) related to the meaning of the derivative

• Truncatory processes do not create new words in the lexicon. (cf., e.g., Dressler& Merlini Barbaresi 1994, �Stekauer 1998, 2000, Dressler 2000).

c. other problems

• Truncation is a conscious process, whereas word formation in general is uncon-scious (cf., e.g., Dressler & Merlini Barbaresi 1994, Dressler 2000).

• Truncations are a phenomenon of usage, not of grammar (cf., e.g. Jespersen1965repr., Marchand 1960).

• Truncation is a diachronic, not a synchronic process (cf., e.g., Mel�uk 2000).

In what follows I will discuss the two factors listed under (10.a) and (10.b). I willargue that – under closer inspection – none of the points listed provides sufficientjustification for excluding English truncatory patterns from the realm of (grammati-cal) word formation. This is particularly true for the claim that outputs are struc-turally unpredictable, but also for the claim that truncation does not produce newwords in the sense that other word formation processes do. The factors listed in(10.c), by contrast, contain assumptions for which, to my knowledge, no large-scaleempirical evidence is available in the literature on English.

1.3.1 The Form of the Output

The alleged contrast between unpredictable processes like truncation and predictablemorphological processes such as derivational affixation has led many scholars topostulate a systematic distinction among word-forming processes between regular(i.e. rule-governed) word formation processes and so-called creative processes (Lyons1977: 76ff., 549, Bauer 1983: 63f., van Marle 1985: 43, cf. also Zwicky & Pullum 1987,who advocate a distinction between plain morphology and expressive morphology).

However, the unpredictability claim has generally come under attack with the riseof output-oriented analyses of truncatory and reduplicative processes in ProsodicMorphology. Using a templatic framework, Weeda (1992: 102ff., 176ff.) has alreadyshown that many allegedly idiosyncratic properties of English truncated names can bepredicted on the basis of a syllabic template. The empirical analysis in chapters 4–6of this book will be devoted to showing that the output form of the most productive

MORPHOLOGY 17

truncatory processes in the language is predictable to the extent to which the outputof other, uncontroversially morphological processes is predictable, too. The key to theanalysis lies in the adoption of output-oriented, constraint-based framework, and amechanism that incorporates systematic variability into the model (cf. below fordiscussion of the latter).

The claim that English truncated forms are formally irregular seems to be theproduct of mainly three factors: First of all, previous, rule-based frameworks of theSPE-type have not been able to account for the structure of truncatory forms whichare defined through their output structure (cf., e.g., the argumentation in Bauer 1983:232). Secondly, very little attention has been devoted in the literature to the questionof which of the attested truncated forms actually correspond to a pattern that syn-chronically produces truncated forms in the language. Most descriptive accountsfocus on showing the variety of different kinds of truncatory forms (Jespersen1965repr., Marchand 1960, Adams 1973, Bauer 1983; but cf. Plag 2003 and, with afocus wider than English, Booij 2005 for a different view). Few attempts have beenmade (apart from occasional remarks) to systematise the inventory and, in particu-lar, to distinguish between productive and marginal cases. Furthermore, accountstypically do not distinguish between different processes, name truncation and wordclipping. The main means of classification employed is either which part of the baseword is retained (the beginning, the middle, or the end, e.g. Bauer 1983), or whichpart of the base word is deleted (the end or the beginning, e.g. Marchand 1960).Examples from Bauer (1983) are provided in (11).

(11) traditional formal classification of truncations (Bauer 1983: 233f.)

a. The beginning of the base lexeme is retainedmike (� microphone)binocs (� binoculars)jumbo (� jumbo jet)mimeo (� mimeograph)

b. The final part of the base lexeme is retainedCong (� Viet Cong)loid (� celluloid)

c. Middle of the base word is retainedjams (� pyjamas)shrink (� head-shrinker)

In (11) we find cases in which the clipped output is identical to a morpheme inthe base word (mimeo � mimeograph) as well as cases in which this is not the case(e.g. binocs � binoculars, jams � pyjamas). Secondly, the data in (11) differaccording to which material of the base is preserved in the truncated form (cf.,e.g., mike � microphone vs. jams � pyjamas). Whereas it is frequently noted thattruncated forms preserving the initial part of the base are more frequent than those


that retain other parts of the base, accounts have nevertheless always preserved thetraditional classification according to the part of the base retained or deleted, andhave provided examples of all classes. No attempt has been made to quantify therelation between the different patterns. Instead, the existence of forms which do notpreserve the initial part of the base is generally taken as evidence that the processproduces unpredictable outputs.

The third factor that plays a role in assessing whether or not outputs of truncationare unpredictable (and, thus, irregular) is the definition of unpredictability itself, andits relation to variability. Grammatical models generally agree that there existsa clearcut distinction between what is regular (i.e. grammatical) and what is irregular(i.e. ungrammatical) in a language. However, they differ in terms of how muchvariability they incorporate into their notion of 'regularity'. Most traditional models ofmorphological grammar put a strong emphasis on the idea that productive morpho-logical processes produce invariable outputs (e.g. Aronoff 1976: 22, Scalise 1984:137, 1988: 232, Szymanek 1985: 95; cf. Plag 1999: 49 for a critical discussion).However, there has also been a long-standing debate about how cases of systematicvariability are to be incorporated into this system. Much of this debate has centeredaround semantic variability. For example, differences in the meanings of English -erderivatives lead Stekauer (2000: 111) to postulate five independent, homonymoussuffixes -er. A very different view is taken in Ryder (1999), who proposes a unifiedtreatment of the semantics of English -er formations in a cognitive framework.

In contrast to semantic variability, structural variability has not received muchattention. Nevertheless, cases of structural variability exist, and they are not confinedto truncation. We also find them in processes whose status as morphological processis less contested. The variability of linking elements in compounds in Germanic lan-guages is a case in point. For example, the Dutch translation of the word 'spellingchange' can be spellingverandering (spelling � verandering) or spellingsverander-ing (spelling � s � verandering). 's' is the linking element. The observed variationcan occur within a single speaker (cf. Krott et al. 2001: 56). Further examplesinclude Tagalog -um- infixation in loanwords from English and Spanish (Orgun &Sprouse 1999, Zuraw 2005) and Finnish genitive allomorphy (Anttila 1995). Theyare illustrated in (12).

(12) variability of output forms in morphological processes

a. Tagalog -um- infixation (Orgun & Sprouse 1999: 204)gr-um-adwet or g-um-radwet (� gradwet) 'to graduate'pl-um-antsa or p-um-lantsa (� plantsa) 'to iron'pr-um-eno or p-um-reno (� preno) 'to break'

b. Finnish genitives (Anttila 1995: 3)naapur-eid-en or naapur-i-en (� naapuri) 'neighbour'Reagan-eid-en or Reagan-i-en (� Reagani) 'Reagan'moskeij-oid-en or moskeij-oj-en (� moskeija) 'mosque'

english prosodic morphology...nar on prosodic morphology at the university of marburg, taught by...

Documents