Praise for the Third Edition:“This newly revised edition provides in-depth coverage of all areas of Japanese grammar, in an accessible style which will make it a valuable pedagogical and reference work for anyone interested in Japanese linguistics.”

—Peter Sells, University of York

“An impressive, judiciously revised 3rd edition with comprehensive coverage of the main areas within Japanese linguistics. Superbly suited as a textbook at both undergraduate and graduate level. Highly recommended.”

—Bjarke Frellesvig, University of Oxford

“This book is an excellent invitation to Japanese Linguistics for anyone interested in the language. It covers a wide range of topics from phonetics to dialectal vari-ation and presents a comprehensive picture of what the language is like. The insightful discussion in each chapter provides students of Japanese and linguistics with many good research topics to investigate further.”

—Mamoru Saito, Nanzan University

Praise for the Second Edition:“While maintaining the well-balanced coverage of Japanese linguistics of the earlier edition, Tsujimura manages to explore a variety of new issues in the experimental and applied areas. The well-chosen additional problem sets guide students towards important topics for future research.”

—Junko Ito, University of California, Santa Cruz

Blackwell Textbooks in Linguistics

The books included in this series provide comprehensive accounts of some of the most central and most rapidly developing areas of research in linguistics. Intended primarily for introductory and post-introductory students, they include exercises, discussion points and suggestions for further reading.

1. Liliane Haegeman, Introduction to Government and Binding Theory (Second Edition)

2. Andrew Spencer, Morphological Theory3. Helen Goodluck, Language Acquisition4. Ronald Wardhaugh, An Introduction to Sociolinguistics (Sixth Edition)5. Martin Atkinson, Children’s Syntax6. Diane Blakemore, Understanding Utterances7. Michael Kenstowicz, Phonology in Generative Grammar8. Deborah Schiffrin, Approaches to Discourse9. John Clark, Colin Yallop, and Janet Fletcher, An Introduction to

Phonetics and Phonology (Third Edition)10. Natsuko Tsujimura, An Introduction to Japanese Linguistics

(Third Edition)11. Robert D. Borsley, Modern Phrase Structure Grammar12. Nigel Fabb, Linguistics and Literature13. Irene Heim and Angelika Kratzer, Semantics in Generative Grammar14. Liliane Haegeman and Jacqueline Guéron, English Grammar:

A Generative Perspective15. Stephen Crain and Diane Lillo-Martin, An Introduction to Linguistic

Theory and Language Acquisition16. Joan Bresnan, Lexical-Functional Syntax17. Barbara A. Fennell, A History of English: A Sociolinguistic Approach18. Henry Rogers, Writing Systems: A Linguistic Approach19. Benjamin W. Fortson IV, Indo-European Language and Culture:

An Introduction (Second Edition)20. Liliane Haegeman, Thinking Syntactically: A Guide to Argumentation

and Analysis21. Mark Hale, Historical Linguistics: Theory and Method22. Henning Reetz and Allard Jongman, Phonetics: Transcription,

Production, Acoustics and Perception23. Bruce Hayes, Introductory Phonology24. Betty J. Birner, Introduction to Pragmatics

An Introduction to Japanese Linguistics

Third Edition

Natsuko Tsujimura

This third edition first published 2014© 2014 John Wiley & Sons, Inc.

Edition history: Blackwell Publishers Ltd (1e, 1996); Blackwell Publishers Ltd (2e, 2007)

Registered OfficeJohn Wiley & Sons Ltd, The Atrium, Southern Gate, Chichester, West Sussex, PO19 8SQ, UK

Editorial Offices350 Main Street, Malden, MA 02148-5020, USA9600 Garsington Road, Oxford, OX4 2DQ, UKThe Atrium, Southern Gate, Chichester, West Sussex, PO19 8SQ, UK

For details of our global editorial offices, for customer services, and for information about how to apply for permission to reuse the copyright material in this book please see our website at www.wiley.com/wiley-blackwell.

The right of Natsuko Tsujimura to be identified as the author of this work has been asserted in accordance 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 any means, electronic, mechanical, photocopying, recording or otherwise, except as permitted by the UK Copyright, Designs and Patents Act 1988, without the prior permission of the publisher.

Wiley also publishes its books in a variety of electronic formats. Some content that appears in print may not be available in electronic books.

Designations used by companies to distinguish their products are often claimed as trademarks. All brand names and product names used in this book are trade names, service marks, trademarks or registered trademarks of their respective owners. The publisher is not associated with any product or vendor mentioned in this book.

Limit of Liability/Disclaimer of Warranty: While the publisher and author(s) have used their best efforts in preparing this book, they make no representations or warranties with respect to the accuracy or completeness of the contents of this book and specifically disclaim any implied warranties of merchantability or fitness for a particular purpose. It is sold on the understanding that the publisher is not engaged in rendering professional services and neither the publisher nor the author shall be liable for damages arising herefrom. If professional advice or other expert assistance is required, the services of a competent professional should be sought.

Library of Congress Cataloging-in-Publication Data

Tsujimura, Natsuko. An introduction to Japanese linguistics / Natsuko Tsujimura. – Third Edition. pages cm Includes bibliographical references and index. ISBN 978-1-4443-3773-0 (pbk.) 1. Japanese language. PL523.T74 2013 495.6–dc23 2013014763

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

Cover image: Robert Delaunay, Rythme, 1934. Paris, Musée National d’Art Moderne – Centre Pompidou. Photo © 2013 Gaspart/Scala, FlorenceCover design by Nicki Averill Design

Set in 10/13 pt Sabon by Toppan Best-set Premedia Limited

1 2014

To my mother and my sister

Contents

Preface to the Third Edition xiAcknowledgments xiv

1 Introduction 1Suggested Readings 4

2 Phonetics 51 Phonetic Inventory 7

1.1 Consonants 81.1.1 Stops 81.1.2 Fricatives 101.1.3 Affricates 121.1.4 Liquids 121.1.5 Glides 131.1.6 Nasals 141.1.7 Further Notes on Consonants 18

1.2 Vowels 222 Suprasegmental Feature – Accent 273 Acquisition Issues 33Notes 35Suggested Readings 36Exercises 37

3 Phonology 391 Phonological Rules in Japanese 40

1.1 Devoicing of High Vowels 401.2 Nasal Assimilation 441.3 Alveolar Alternations 471.4 [h]/[ɸ]/[ç] Alternations 511.5 Digression on the Phoneme Status of [ts, č, š, ǰ, ɸ, ç] 521.6 Rule Ordering and the Nature of Rule Statement 54

2 Sequential Voicing – “Rendaku” 56

viii Contents

3 Mora vs. Syllable 653.1 Speech Errors 68

3.1.1 English 683.1.2 Japanese 70

3.2 Language Games: “Babibu” Language 723.3 Syllable-Based Phenomena 74

4 Length Requirements 755 Loanwords 816 Accentuation in Japanese 85

6.1 Compound Accentuation 866.1.1 Accentuation of Long Nominal Compounds 876.1.2 Accentuation of Short Nominal Compounds 916.1.3 Accentuation of Superlong Nominal Compounds 95

6.2 Accentual Variation among Endings 966.3 Accentuation of Loanwords 102

7 Mimetics – Palatalization 1048 Acquisition Issues 109Notes 114Suggested Readings 118Exercises 119

4 Morphology 1251 Parts of Speech Categories 126

1.1 Nouns 1261.2 Verbs 1281.3 Adjectives 1311.4 Adverbs 1321.5 Postpositions 1331.6 Case Particles 1341.7 Adjectival Nouns 1371.8 Verbal Nouns 140

2 Morpheme Types 1423 Word Formation 150

3.1 Affixation 1503.2 Compounding 1523.3 Reduplication 1543.4 Clipping 1543.5 Borrowing 155

4 Issues in Japanese Morphology (1): Transitive and Intransitive Verb Pairs 157

5 Issues in Japanese Morphology (2): Nominalization 1686 Issues in Japanese Morphology (3): Compounding 174

6.1 Background 1746.2 N–V Compounds 179

Contents ix

6.3 V–V Compounds 1846.3.1 Lexical vs. Syntactic V–V Compounds 1846.3.2 Lexical V–V Compounds: Semantic Relations 1956.3.3 Lexical V–V Compounds: Transitivity and

Argument Structure 2006.3.4 Transitive and Intransitive Compound

Verb Pairs 2056.3.5 Compound Verbs and Nominalization 213

7 Acquisition Issues 214Notes 221Suggested Readings 223Exercises 224

5 Syntax 2291 Syntactic Structures 229

1.1 Syntactic Constituency 2291.2 Phrase Structures 2311.3 Phrase Structure Rules 234

2 Word Order and Scrambling 2382.1 Basic Word Order 2382.2 Scrambling Phenomenon 2412.3 Noun Ellipsis 251

3 Reflexives 2543.1 Zibun 2553.2 Zibun-Zisin 263

4 Passives 2654.1 Direct Passives 2664.2 Indirect Passives (Adversative Passives) 269

5 Causatives 2725.1 O-Causatives and Ni-Causatives 2725.2 The Double-O Constraint 2745.3 Causative Passives 2795.4 Adversative Causatives 2805.5 Lexical Causatives 281

6 Relative Clauses (Sentence Modifiers) 2846.1 The Ga/No Conversion 2886.2 Relative Clauses without Gaps 2906.3 Internally Headed Relative Clauses 291

7 The Light Verb Construction 2938 Acquisition Issues 295Notes 300Suggested Readings 302Exercises 304

x Contents

6 Semantics 3091 Word Meaning and Sentence Meaning 310

1.1 Word/Phrase Meaning and Types of Relationships 3101.2 Sentence Meaning 3131.3 Metaphors and Idioms 3171.4 Deixis 3191.5 Mimetics 331

2 Tense and Aspect 3352.1 Tense 3352.2 Aspect 340

2.2.1 Grammatical Aspect 3412.2.2 Lexical Aspect 3472.2.3 Aspectual Verbs 361

3 Verb Semantics 3703.1 Semantic Classes of Verbs and their Syntactic Patterns 3713.2 Lexicalization 373

4 Pragmatics 3754.1 Speaker’s Meaning 3754.2 The Nature of Information 377

4.2.1 Wa vs. Ga 3774.2.2 Interaction with Syntax and Morphology 380

4.3 Relevance of Contextual Information 3845 Acquisition Issues 387Notes 396Suggested Readings 397Exercises 399

7 LanguageVariation 4031 Dialectal Variation 4032 Styles and Levels of Speech 4103 Gender Differences 4154 Acquisition Issues 425Notes 429Suggested Readings 429Exercises 430

Bibliography 431Index 455

Preface to the Third Edition

Throughout my education I have been blessed with a number of great teachers from whom I received unfathomable influence. Adrian Akmajian is high on that list. Adrian used to say that writing a textbook is more challenging than one might imagine because you get criticized no matter what. Of course I did not realize what exactly he meant until I wrote one. While I may not agree with some of the criticisms on the earlier editions, I am fortunate that many of the criticisms were constructive and informative so as to help me produce yet another edition which I hope can be considered an improvement. I think I have achieved that goal – at least to some degree – in this third edition.

The second edition included significant revisions in the chapters on morphol-ogy and semantics with an additional chapter on first language acquisition. For this current edition, I have made extensive revisions in the chapters on phonetics (Chapter 2) and syntax (Chapter 5), some in the chapter on phonology (Chapter 3), and minor modifications in the rest of the chapters. The chapter on first lan-guage acquisition in the second edition has been revised and incorporated as sections in each of the chapters in the third edition with additional discussion of second language acquisition. For the phonetics chapter, unlike in the previous editions, which separated the discussion of the Japanese and English sound inventories, the presentation of various phonetic issues in this third edition keeps its primary focus on the Japanese sound system and pertinent phenomena with occasional comparative reference to English. Responding to the frequent criticism that the syntax chapter in the previous editions is too technical, I have decided to eliminate most of the tree representations as well as the theoretical discussion, but instead to keep the chapter as descriptive as possible and to include some typological comparisons. Putting together all the revisions made in the second and third editions, the current edition is quite a departure from the original book that was published in 1996. The theoretical approach underlying the book is still primarily that of generative grammar, but where possible and appropriate, I have tried to shy away from theoretical matters and to include occasional crosslinguis-tic comparison regardless of theoretical orientation so that the Japanese-internal as well as typological descriptions are more enriched than in earlier editions.

As I stated in the preface to the second edition, the primary goal of this book is to examine spoken Japanese, presenting linguistic description and analyses of

xii Preface to the Third Edition

a wide range of phenomena. It is intended to serve as a descriptive source and a theoretical foundation for an audience that includes students and scholars in linguistics as well as in Japanese language pedagogy and education, and more generally for those interested in the Japanese language. It is hoped that students will find the book a useful pedagogical tool to learn basic notions and terminol-ogy in linguistics and to develop their skills in linguistic analysis and argumenta-tion. Depending on the chapter and on the reader’s background, the user may feel that the extent of the details and the level of complexity are either too advanced or not adequate. Especially in a classroom setting, the instructor is advised to make pedagogical judgments appropriate for the students’ background and needs, and additional materials may be used as supplements; or this book can be used as a supplement to other pedagogical tools. For the novice student, this book is more effectively used as a textbook in a classroom, rather than as a self-guide to a general introduction to Japanese linguistics. Professional linguists who are not proficient in Japanese have found this book useful as a reference in providing examples and descriptive analyses of linguistic phenomena in Japanese. A suggestion was made to include examples written in the Japanese script, but I have kept the same representation in Romanization with such an audience in mind.

The publication of the third edition is somewhat of an irony since I was origi-nally not interested in writing it, still feeling a residual sense of burnout. I am grateful to Wiley Blackwell’s linguistics acquisition editor, Danielle Descoteaux, for encouraging me to work on a new edition and for once again planting the seed of excitement in me back in 2009. Danielle quickly solicited four reviews for the second edition, which served as a springboard for the current edition. Without these reviews I would have been close to clueless as to how I might improve upon the previous edition. Two additional reviewers examined the pre-final version of this edition, and thanks to their careful reading, I was fortunate to make the last round of significant revisions. I thank Danielle and the six reviewers for leading me to the path I needed to take. In the process of revising the third edition, I further benefited from comments on previous editions. Atsushi Fukada at Purdue University, who adopted the book for his linguistics course, sent me e-mail messages to inform me of a number of counterexamples to the generalizations pertinent to accentuation. His comments gave me a chance to revisit the topic in the phonology chapter. The revisions of the phonetics chapter would not have been possible without the technical assistance on spectrograms and wave forms that Kyoko Okamura kindly and patiently offered to me. I par-ticularly appreciate her generosity in having spent hours of recording (and re-recording) with me during the hot and humid summer of 2010 in Bloomington. Shigeru Miyagawa and Peter Sells, my long-time colleagues, gave me very useful suggestions as to how I might re-work the syntax chapter. Although the chapter ended up being one that is not consistent with their recommendations, I would not have been able to make my decision without their initial input. Sara Sowers did an incredibly thorough job in going over the references and the text, which

Preface to the Third Edition xiii

saved me countless hours. The frequent appearance of Stuart Davis in my acknowledgments faces the risk of obscuring their sincerity, but I do not know any other way than simply thanking him profusely for the professional and per-sonal guidance he has endlessly offered me during this project and far beyond. Julia Kirk of Wiley Blackwell patiently responded to my numerous questions on copyright permissions. It was to my pleasure and relief that Fiona Sewell once again took care of fine-tuning of the book, for the third time. The finesse and professionalism she provided in editing this edition have made the very last stage of the production much less stressful than it would have been otherwise. I am greatly indebted to her. Thank you all for helping me survive another round of revising the book through your professional expertise. Finally, I am indebted to my mother and my sister for my professional career, of which this book forms a significant part, as well as for my personal life; their sacrifice is something for which I can never repay them. This last edition is dedicated to them.

Acknowledgments

The author and publisher gratefully acknowledge the following for permission to reproduce the copyright material in this book:

Figure 2.2, vocal folds states, and Figure 2.3, the vocal tract, Figure 2.3 and Figure 2.4 from Contemporary Linguistics by William O’Grady, Michael Dobro-volsky, and Mark Aronoff. Copyright © 1989 by Bedford/St. Martin’s. Used by permission of the publisher. Contemporary Linguistics is now available in its sixth edition, copyright © 2010.

Figure 4.1, distribution of phonological constraints, adapted from Junko Itô and R. Armin Mester, “Phonological Constraints”, in John A. Goldsmith (ed.), The Handbook of Phonological Theory, p. 823, figure 8. Oxford: Blackwell, 1995. Copyright © 1995 by Blackwell Publishing Ltd. Reprinted by permission of Wiley-Blackwell.

Map 7.1, map of Japan, from Masayoshi Shibatani, The Languages of Japan, map 3, p. 188. Cambridge: Cambridge University Press, 1990. Copyright © 1990 by Cambridge University Press. Reprinted with the permission of Cambridge University Press.

Every effort has been made to trace copyright holders and to obtain their permis-sion for the use of copyright material. The publisher apologizes for any errors or omissions in the above list and would be grateful if notified of any corrections that should be incorporated in future reprints or editions of this book.

An Introduction to Japanese Linguistics, Third Edition. Natsuko Tsujimura.© 2014 John Wiley & Sons, Inc. Published 2014 by John Wiley & Sons, Inc.

1 Introduction

We use language every day to communicate with each other. Even young children use language. Children naturally acquire the language (or languages) spoken in the community around them: it could be English, Japanese, Russian, Tagalog, Zulu, or, in the case of the hearing impaired, American Sign Language (ASL), for example. While language consists of sounds (or signs as in ASL), words, and sentences, it is not simply a random sequence of sounds, words, or signs. For instance, a native speaker of English knows that a sequence of sounds like “abpmeshnsch” or the string of words “the walked yellow a yesterday pet three where quickly” do not represent utterances of English. Language is a more systematic, rule-governed mechanism. Sounds pattern in certain regular ways in forming words, and words combine to form sentences in a consistent manner as well.

The field of study where language is investigated in a systematic way is called linguistics. A primary goal of much linguistic research then is to discover the patterns that underlie languages. When linguists find such patterns in a particular language, they posit that there are rules or constraints in the language, which produce these patterns. Linguists hypothesize that when children acquire their native language during early childhood, they subconsciously learn the rules and constraints of their language that enable them to speak and understand the lan-guage in a fluent manner without hesitation. Thus, linguists are particularly interested in uncovering rules or constraints that speakers must subconsciously know when they speak a language – rules that speakers themselves are completely unaware of.

Some people have thought that children learn their first language by imitating what their parents say. That is, children were thought to learn their first language gradually, by listening to what their parents say to them and by imitating it. However, this assumption has been questioned for various reasons. For example, children have the ability to create sentences they have never heard before. If they

2 Introduction

learned their first language simply by imitating their parents, it would be impos-sible for them to create sentences that are completely novel to them. The fact that children are capable of constantly creating new sentences immediately casts doubt on the assumption that children learn their first language by imitation.

Another piece of evidence in conflict with the assumption that children learn their first language by imitating their parents comes from the nature of the mis-takes they make. In learning their native language, children make mistakes, but the mistakes often reflect their ability to make sophisticated generalizations con-cerning their first language on the basis of their observations. There are plenty of examples of this sort. One instance can be seen in English-speaking children’s mistakes in the irregular past tense verb forms. In English, regular past tense verbs are formed simply by adding -ed to the present tense verb. So, we get laughed from laugh and smiled from smile. English also has a large class of irregular verbs including go–went, bring–brought, and break–broke. For the irregular verbs, many children go through a stage where they use goed instead of went and bringed instead of brought. Adult speakers of English know irregular past tense verb forms, and they would not utter goed, bringed, and breaked; instead, they would say the correct forms, went, brought, and broke. If children learn their first language simply by imitating what their parents say, they should never produce incorrect past tense verb forms such as goed, bringed, and breaked because their parents do not say them. Rather than imitating their parents’ speech, children subconsciously make an observation concerning the formation of past tense verb forms on the basis of regular forms such as laughed and smiled, and then make a generalization that past tense verbs are formed by adding -ed to present tense verbs. The incorrect outputs like goed, bringed, and breaked result from an overgeneralization of such a “rule” to irregular verbs. This sug-gests that children do have the ability to generalize on the basis of their observations.

Children also make mistakes in their use of verbs. Some English verbs have causative counterparts. Consider the pairs in (1)–(2), which are taken from Pinker (1989).

(1) a. The horse walked/galloped/trotted/raced/ran/jumped past the barn.b. I walked/galloped/trotted/raced/ran/jumped the horse past the barn.

(Pinker 1989: 131)

(2) a. The log slid/skidded/floated/rolled/bounced.b. Brian slid/skidded/floated/rolled/bounced the log.

(Pinker 1989: 130)

The verbs of the (b) sentences in (1)–(2) display causative use. For example, in (1a) the horse voluntarily performed the action of walking, galloping, trotting, etc. In (1b), on the other hand, I instigated some action that led to the horse’s walking, galloping, trotting, etc. That is, I caused the horse to walk, gallop, trot,

Introduction 3

etc. Similarly, in (2b), Brian caused the log to slide, skid, float, etc., while such a causative interpretation is not available in the (a) sentence. So, the same verb can be used to induce the causative interpretation. Children observe this phe-nomenon, but their output is not necessarily grammatical; (3)–(6), taken from Pinker (1989: 305–306), are actual utterances by children (cf. Bowerman 1982; Pinker 1989).

(3) You can drink me the milk.

(4) Will you have me a lesson? [Request to adult friend in swimming pool]

(5) Andrea, I want to watch you this book.

(6) Remember me what I came in for.

The verb drink in (3) is used with a causative interpretation, although such a reading is not allowed with this verb in adult English. So, the child used the verb to mean “to feed” or “to help to drink”. Similarly, have in (4) is used as “give”, the causative counterpart of have (“let me have”); in (5) watch you is used to mean “have you watch”; and in (6) remember is meant to be “remind” (“let me remember”). Another example that is often observed in children’s speech is a sentence like “He learned me real good”, in which learn is intended to be “make me learn”. In these errors which actually occur in children’s speech, we can see that they are analyzing drink/have/watch/remember/learn as causative verbs just like walk, run, and roll in (1)–(2), for instance. Notice that children would not make these sorts of mistakes if they spoke the language just by imitating their parents, because adult speakers would not make such mistakes. Instead, children are making a generalization and applying it to new words and sentences based on what they hear. What is important here is that children are actively making generalizations, trying to figure out the language, although this task is largely subconscious.

Noam Chomsky, the most influential American linguist in the second half of the twentieth century, is a strong proponent of the hypothesis that there must be something innate in the human cognitive system which enables children to create sentences they have never heard before, and enables them to figure out and learn their language. Under this view, one of the main tasks of a linguist is to figure out the exact nature of the innateness. In undertaking this sort of task, linguists first observe some language-related phenomenon and describe it. Second, they try to figure out whether the phenomenon is of an arbitrary nature, or whether there is some systematic pattern associated with it. When the latter is found to be the case, they formulate a hypothesis on the basis of this pattern. Often the hypothesis makes further predictions about patterns in the language. Third, the hypothesis is tested against a new set of data. If the new data are inconsistent with the predictions, the hypothesis is falsified, and hence needs to be discarded or modified in order to account for the patterns found in the language.

4 Introduction

We shall see that this notion and strategy of hypothesis testing will play an important role in the presentation of the language data in many of the chapters in this book. To this end, we will primarily examine the Japanese language, but on occasion we will also compare Japanese with English as well as other lan-guages of the world.

Suggested Readings

A general introduction to the field of linguistics can be found in Akmajian et al. (2010), Fromkin et al. (2011), Finegan (2012), Parker and Riley (2010), Napoli (1996, 2003), Pinker (1999), and Jackendoff (2002), among many other textbooks and introductory books. Parker and Riley (2010) is particularly accessible to those who have no prior knowledge of linguistics. Chomsky’s program of linguistic research can be found in Chomsky (1986), Cook and Newson (2007), and Pinker (1994).

An Introduction to Japanese Linguistics, Third Edition. Natsuko Tsujimura.© 2014 John Wiley & Sons, Inc. Published 2014 by John Wiley & Sons, Inc.

2 Phonetics

When we try to list all the sounds in a language, it is important not to confuse orthography, i.e. the writing system, with the actual sounds. For example, the sound [k] in English can be exemplified by various spellings, as in kiss, sick, choir, quit, cow, Iraq, and unique. (Note: The brackets “[ ]” are used when representing the pronunciation of each sound. This symbol will be further explained in chapter 3.) Even though we are looking at seven different ways of spelling here, we are nonetheless referring to the single sound [k].

Confusing spelling and actual sounds is particularly problematic in Japanese because, as figure 2.1 (p. 6) shows, the majority of the Hiragana syllabary – one of the writing systems used for Japanese – consists of the combination of two independent sounds, i.e. a consonant and a vowel.

Take さ and て for instance.

(1)

s a t e

The さ stands for the consonant [s] and the vowel [a] in that order; and て stands for the consonant [t] and the vowel [e] in that sequence. Thus, the Hira-gana syllabary is unable to isolate the individual sounds of Japanese and does not reflect the phonetic inventory (i.e. the list of sounds available) of the language.

Another option for listing Japanese sounds would be to consider utilizing the specific Romanization system that has been developed for Japanese in order to isolate the individual sounds. However, even this system does not neces-sarily reflect the actual sounds. Consider the series in (2), which is part of a commonly used Romanization system for Japanese (i.e. the “Kunrei”-style Romanization).1

6 Phonetics

Figure 2.1 Hiragana syllabary2

a i u e o

ka ki ku ke ko

sa si su se so

ta ti tu te to

na ni nu ne no

ha hi hu he ho

ma mi mu me mo

ra

wa

o

n

ri ru re ro

ya yu yo

(2) a. tab. tic. tud. tee. to

Under this Romanization system, the same consonant is shared by each of the five in (2a–e), and the only difference is supposed to be the vowel that accompanies it. However, the consonant that is represented by t in (2a–e) is not pronounced

Phonetics 7

in the same manner. The consonant t in (2a, d, e) is a sound similar to the t sound in the English words top, ten, and tone; the consonant in (2b) is similar to the sound represented by ch as in the English word cheese; and the consonant in (2c) is pronounced very much like the sequence ts as in the English word cats.

This illustration of the Hiragana syllabary and the Romanization system clearly shows that writing systems do not correlate with the phonetic inventory of the language, and we need a system that enables us to describe a sound as it is pronounced. This is why we focus on spoken language rather than written language when we investigate the phonetic system of a language. To avoid the confusion that we have demonstrated above, we use phonetic symbols to tran-scribe the sounds that exist in a language.

1 Phonetic Inventory

Before describing the sounds of Japanese, it is helpful to understand the general mechanism of how speech sounds are produced. While there exist some sounds in the world’s languages that are produced as we inhale, most sounds are pro-duced as air flows from our lungs. The stream of air, which originates in the lungs, travels through the trachea – commonly called the windpipe – and then reaches the larynx, less technically known as the Adam’s apple. The larynx is an important speech organ because this is where the vocal cords (or vocal folds) are found: depending on the nature of the vocal cords’ activities, different types of sounds are produced. For instance, place your fingers at the Adam’s apple, and sound out each of [s] and [z] continuously. The presence and absence of vibration you feel at your throat is attributed to the activities of the vocal cords. The vocal cords are made of two muscular folds, and when they are pulled together leaving a slight opening, the air from the lungs tries to escape through the narrow opening of the vocal folds, which leads to their vibration. This is what you feel when [z] is pronounced. In contrast, when the vocal cords are open, creating a space referred to as the glottis, air flows freely without causing their vibration. This is the situation with [s]. The presence or absence of the vibration of the vocal cords, as is reflected by the difference between [z] and [s] in our example, is called voicing: voiced sounds like [z] are produced with vocal cord vibration while voiceless speech sounds like [s] are produced without such vibration. Figure 2.2 illustrates the contrastive states of the vocal cords.

Once the air from the lungs passes through the larynx into the mouth, a variety of sounds are produced, and they are generally divided into two groups, conso-nants and vowels. Consonants are sounds made with an obstruction in the mouth, while no such obstruction occurs with vowels.

8 Phonetics

Figure 2.2 Vocal cords (O’Grady et al. 1989: 17)

Front of larynx

Glottis

Voiceless

Vocal folds Vocal folds

close together

Whisper

Front portion of vocal folds drawn together

pulled open

Voiced

1.1  Consonants

We have already discussed that voicing is important in distinguishing sounds. There are two additional factors that help us to isolate one sound from another. They are place of articulation and manner of articulation. With reference to the anatomical structure of the mouth, place of articulation indicates the place in the mouth where the sound is made. The lips and the various regions along the roof of the mouth, as is illustrated in figure 2.3, are all places of articulation. Different sounds are made at different places of articulation by placing articulators such as the tongue at these different locations.

Manner of articulation refers to how the articulators, such as the lips and the tongue, achieve contact with the places of articulation. For instance, the sound produced when an articulator has complete contact with a place of articulation and the sound produced with partial contact are different.

In addition to these three factors, it is also important to keep in mind whether the airflow from the lungs passes through the oral cavity (the mouth) or through the nasal cavity. When the velum is raised, the passage through the nasal cavity is completely blocked, and the air from the lungs goes out through the oral cavity. Sounds made in this way are called oral sounds. When the velum is lowered, on the other hand, the airflow from the lungs travels through the nasal cavity as well as into the oral cavity. The sounds created in this manner are referred to as nasal sounds.

Thus, consonants in a language are described by place of articulation, manner of articulation, voicing, and whether they are oral or nasal. A specific combina-tion of these factors identifies a single sound to which a unique phonetic symbol is assigned.

1.1.1 Stops

The first group of sounds that we will consider is one that is characterized by its manner of articulation, namely, stops. Stops are further divided into two types,

Phonetics 9

oral and nasal. Oral stop sounds – often referred to simply as stop sounds or plosives – are produced when the airflow originating from the lungs and coming through the oral cavity, or mouth, is completely blocked. What separates oral stops from nasal stops is that the former occur when the velum is raised, prevent-ing airflow escaping through the nasal cavity. Nasal stops are articulated when a complete obstruction takes place in the oral cavity and the air goes through the nasal cavity because of the lowered velum. Nasal stops will be discussed in section 1.1.6 below.

The location of complete air blockage in the oral cavity reflects the place of articulation. When the blockage is made at the lips by placing the upper lip and lower lip together, a bilabial stop is produced. Furthermore, when the bilabial stop is accompanied by vocal cord vibration, a voiced bilabial stop is produced,

Figure 2.3 Sound-producing system (O’Grady et al. 1989: 19)

Nasal cavity

Palate: palatalsAlveolar ridge:alveolars

Alveopalatal

Lips: labials

Blade

Teeth: dentals

Tip

BodyBack

Root

Glottis: glottals

TracheaEsophagus

Pharynx: pharnygeals

Uvula: uvulars

Velum: velars

10 Phonetics

and we transcribe it as [b]. The voiced bilabial stop in Japanese is exemplified by words such as binboo “poor” and kooban “police box”. On the other hand, when the vocal cords do not vibrate, a voiceless bilabial stop is articulated. We represent this sound as [p]. Examples of words with the voiceless bilabial stop [p] are pan “bread” and denpoo “telegram”. The bilabial stops in English are found in words like bird and label for the voiced ones and pen and pop for the voiceless ones.

A stop sound made at the alveolar ridge, where the airflow would be com-pletely blocked by the blade of the tongue at the alveolar ridge, is referred to as an alveolar stop. The voiced alveolar stop is represented as [d] and its voiceless counterpart as [t]. They are exemplified by doko “where” and denwa “tele-phone” for the former and tensai “genius” and tokidoki “sometimes” for the latter. The alveolar stops in English words like sad/adapt and tiny/pet are typically produced by the tip of the tongue, rather than the tongue blade, contacting the alveolar ridge. That is, [t] and [d] in Japanese are pronounced slightly further forward than those in English with the tongue tip almost touching the back of the upper teeth (cf. Vance 1987, 2008). Despite this difference, however, we use the same set of phonetic symbols of [d] and [t] to represent the alveolar stops in both languages.

When the air obstruction takes place by raising the tongue body to the velum or soft palate, velar stops are produced, and sounds such as the first consonant of words like gakusei “student”, gojuu “fifty”, konbu “seaweed”, and kinoo “yesterday” are articulated. The first two examples contain a voiced velar stop [g], and the second a voiceless velar stop [k]. The voiced and voiceless velar stops appear in English words agree and cat, respectively.

Although the stop sounds are roughly the same in Japanese and English, a slight difference can be detected with the voiceless members [p, t, k]. When these voiceless stops occur at the beginning of a word or syllable in English, they are accompanied by aspiration, i.e. a puff of air. This can be felt by placing a piece of paper in front of your mouth and then trying to pronounce the English word pin. The initial consonant should be accompanied by a puff of air, and hence is aspirated, as is indicated by the movement of the paper. When a native speaker of Japanese pronounces the word pan “bread”, by contrast, we notice that the paper does not move as much as in English. This is because the degree of aspira-tion in the Japanese voiceless stops is not as large as in English. Even these sounds, however, can be aspirated, especially when they are pronounced with great emphasis (cf. Vance 2008).

1.1.2 Fricatives

Fricatives are sounds produced when the airflow in the oral cavity is forced through a narrow opening in the vocal tract, so that air turbulence is generated, resulting in a friction noise. The friction noise sounds different depending on where the narrow opening occurs. To start with fricative sounds that are created at the front part of the mouth, there is a voiceless bilabial fricative sound, which

Phonetics 11

is similar to the sound that is made when one blows out a candle but without much lip protrusion. There is a narrow opening between the lips, and the air from the lungs escape through them. This sound occurs as the first sound in Japanese words hurui “old” and hukai “deep”, and is represented by the symbol [ɸ]. There is no voiced bilabial fricative in Japanese; and there is neither voiced nor voiceless bilabial fricative in English.

There are a few fricative sounds that exist in English but are missing in Japa-nese: they are labio-dental and interdental fricatives. The first consonants of English words like f n and vase are produced when the upper teeth and the lower lip achieve contact. These sounds are referred to as labio-dental fricatives. The air passes between the upper teeth and lower lip, creating friction noise. The voiceless labio-dental fricative is [f] and its voiced counterpart is [v]. A partial blockage of the airflow can also occur when the tip of the tongue is between the upper and lower teeth. Since the sound is made between the teeth, these fricative sounds are called interdental. The voiceless and voiced interdental fricatives are [θ] and [ð], respectively, and they are exemplified by three/truth and they/smooth.

The first consonants of Japanese words like zoo “elephant” and san “three” are fricative sounds. These sounds are produced when we try to create a narrow opening by keeping the tip of the tongue very close to the alveolar ridge. They are called alveolar fricatives. The voiced alveolar fricative as in zoo “elephant” is transcribed as [z], while the voiceless counterpart, as in san “three”, is repre-sented by the symbol [s]. Thus, the voiceless alveolar fricative [s] and the voiceless alveolar stop [t] share the same place of articulation, alveolar, and voicing feature, voiceless; but they differ in manner of articulation. Similarly, [z] and [d] share the same place of articulation and voicing, but the former is a fricative and the latter is a stop. That is, they have different manners of articulation. Voiced and voiceless alveolar fricatives in English are exemplified by the respective first sound of the words song and zero.

When a fricative sound is made with the blade of the tongue just behind the alveolar ridge, the sound is characterized as a palato-alveolar or alveo-palatal fricative. The voiceless alveo-palatal fricative is represented as [š], and is exempli-fied by the first sound in Japanese words like sinbun “newspaper” and sika “deer”. The same set of sounds is found in the first sounds of English words shoe and shine. The voiced alveo-palatal fricative is [ž], and is not usually found in Japanese.3 There are not many words that have this sound in English either, but it can be exemplified by the middle consonant of vision and the second consonant of measure. For many English speakers, the alveo-palatal fricatives in English are accompanied by rounding of the lips.

When the middle of the tongue behind the blade, i.e. the tongue body, approaches the middle of the roof of the mouth, called the hard palate or just palate, creating a narrow passage, friction sounds are produced. They are palatal fricatives. Japanese has only the voiceless palatal fricative [ç], as in the words hiroi “spacious” and hitori “one person”. This sound is similar to the sound in the German word ich “I” or to the first sound in the English word huge as pro-nounced by many native speakers of American English.

12 Phonetics

Finally, when the air is partially blocked at the narrow opening between the vocal cords (i.e. the glottis), it creates friction and a glottal fricative sound is produced. The glottal fricative is represented as [h], and can be found in the words hanbun “half” and hosoi “thin” for Japanese and heart and hotel for English. There is no vibration of the vocal cords in pronouncing [h], and hence the glottal fricative is voiceless.4 Careful attention should be paid to the Romani-zation representations of the glottal fricative [h] on the one hand and bilabial and palatal fricatives, [ɸ] and [ç], on the other. The Japanese words that include these sounds are uniformly written in Romanization with h, as in haru “spring”, huro “bath”, and hirune “nap”, but the actual pronunciation that corresponds to h in each of these words has a distinct sound quality that is identified by dif-ferent places of articulation. The letter h in haru is glottal, h in huro is bilabial, and h in hirune is palatal. This is why we must focus on a sound as it is actually pronounced rather than as it is written when we consider the range of sounds available in a given language.

1.1.3 Affricates

When we pronounce English words such as church and judge, we notice that the first and the last consonants of each word are produced by the combination of a brief stop followed by a fricative. At first, the air is blocked at a designated place of articulation, but then it is released with a partial blockage. A sequence of a stop immediately followed by a fricative is called an affricate. Japanese has four affricate sounds: the alveolar [ts] (voiceless) and [dz] (voiced), and the alveo-palatal [č] (voiceless) and [ǰ] (voiced). The voiceless alveolar affricate is made with a short alveolar stop that is released into an alveolar fricative. The voiceless alveolar affricate [ts] is exemplified in Japanese by the first sounds in the words tumi “sin” and turi “fishing”. The voiced alveolar affricate, [dz], does not seem to display a clear contrast with the alveolar fricative, [z], in the pronunciation of most Japanese people (cf. Vance 1987; Shibatani 1990). For example, the pro-nunciation of the second consonant of mazusii “poor” and that of the third consonant of mikazuki “crescent moon” varies between [dz] and [z], with perhaps a slight preference for the voiced alveolar affricate. In pronouncing the word tizu “map”, on the other hand, the second consonant for many speakers is invariably the alveolar fricative [z]. The alveo-palatal affricates, [č] and [ǰ], are exemplified by the first sounds in the words tikaku “near” and zikan “time”, respectively. English has these two alveo-palatal affricate sounds. The voiceless one, [č], is exemplified by the consonants in church; and its voiced counterpart, [ǰ], is found in judge. Associated with the pronunciation of these English affricate sounds is a slight rounding of the lips.

1.1.4 Liquids

In comparing the first sounds of the English words sun and run, we notice that the two sounds are made at the same area of the mouth, namely, at the alveolar

Phonetics 13

ridge. In both pronunciations, the tip of the tongue approaches the alveolar ridge and furthermore the airflow from the lungs continuously escapes through the space between the tongue tip and the alveolar ridge. The major difference between the two sounds, however, is the degree to which the obstruction of the airflow is created: the passage between the tongue tip and the alveolar ridge is narrower in the voiceless alveolar fricative [s] in sun than in the first sound of run. The latter sound is produced with constriction in the vocal tract, but the air flows more freely from the mouth with a lesser degree of blockage of air or friction than in fricative sounds. The sounds that are articulated in such a manner are called liquids. English has two liquid sounds, [l] and [r], both alveolar, and the initial consonants of land and run exemplify each. The difference between the two is that with [l] the air channel is on the sides of the tongue while with [r] it is in the middle of the mouth. English [l] is normally pronounced with the tongue tip touching the alveolar ridge, but with the sides of the tongue lowered. The air flows freely over one or both sides of the tongue. Consequently, [l] is referred to as a lateral sound. The American English [r] sound is often made with the tongue tip curled back. Both [l] and [r] are accompanied by vibration of the vocal cords, and are voiced.

Japanese has one liquid sound, the voiced alveolar liquid, as is exemplified by the words roku “six” and ringo “apple”. Although the sound in roku and ringo generally share the place and manner of articulation with English [l] and [r], the initial sound of these Japanese words is quite different from [l] or [r] in English. The liquid sound is technically called the alveolar tap, represented by the phonetic symbol [ɾ], and is produced by placing the tongue tip at the alveolar ridge fol-lowed by an immediate release of that contact. The alveolar liquid in Japanese sounds very similar to the “d” sound in English words like tidy and steady in the American English pronunciation (technically called flap): with both sounds, the tongue achieves very rapid contact at the alveolar ridge. This similarity between English and Japanese explains why native speakers of English learning Japanese are often unable to make a distinction between [ɾ] and [d] in Japanese words. For the sake of simplicity, we will use the transcription symbol [r] for the Japanese alveolar liquid sound throughout this book.5

1.1.5 Glides

Glides are sounds that have characteristics of vowels and consonants, and are often called semivowels. Like vowels, they are produced without (much) obstruc-tion in the mouth. They, however, function like consonants in that they occur at the beginning of a syllable immediately before a vowel. There are two glides in Japanese, as is exemplified by the first sounds of wakaru “understand” and yasui “cheap”. In the articulation of the first sound of wakaru, the body of the tongue approaches the velum, so the place of articulation is velar and is represented as [w]. The first sound of yasui is made by the tongue body approaching the hard palate; that is, it is the palatal glide and is transcribed as [y]. These two glide

14 Phonetics

sounds are also accompanied by vocal cord vibration, and are, therefore, voiced. English also has two glides, [w] and [y], as in wish and yield, respectively. The English [w], however, is accompanied by lip rounding while the Japanese [w] is not generally observed with the same type of lip movement, at least in normal speech. This difference is captured by identifying the glide as the labio-velar glide for English and the velar glide for Japanese. Although there are phonetic symbols that strictly represent the difference in the lip rounding, we will use [w] for the glide in both languages.

1.1.6 Nasals

In producing a nasal consonant, the velum is lowered and a complete obstruction occurs in the oral cavity. The airflow, however, passes freely through the nasal cavity. The sounds produced in this way are nasal stops or simply nasals. The nasals are like the oral voiced stops [b, d, g] in that they are voiced and are produced with a complete blockage in the oral cavity. There are three nasal sounds in Japanese: [m], [n], and [ŋ]. The first one, [m], is a bilabial nasal sound, and appears in words like mikan “orange” and mame “beans”. The second nasal sound, [n], is alveolar, and is found in neko “cat” and naka “inside”. Finally, the occurrence of the velar nasal [ŋ] within a word depends on the speaker. For example, some speakers, especially older speakers, pronounce the “g” in the word kagaku as [ŋ], while others pronounce it as the velar stop [g]. Even when a speaker consistently uses the velar nasal sound in her or his speech, however, it never appears in word-initial position in Japanese. Thus, the initial consonant of the word gakkoo “school” can only be pronounced as [g].

Some works describe the nasal sound before a pause, in words like yon “four” and ken “ticket”, as being pronounced with the tongue body touching the uvula (cf. Vance 2008). The uvular nasal is represented as [N]: the words mentioned above are, thus, transcribed as [yoN] and [keN] before a pause. Sometimes a nasal before a pause may make no contact with the root of the mouth. Such a nasal could be described as a nasalized glide.

The same set of nasal sounds found in Japanese occurs in English. The bilabial nasal [m] is exemplified by the first sound in the word mother; the alveolar nasal [n] occurs at the first and last consonants in the word nun; and the final consonant of the word sing is an example of the velar nasal [ŋ]. The velar nasal does not occur in word-initial or syllable-initial position in English.

Nasal sounds in Japanese present an illustration of a phonetic phenomenon called coarticulation. Coarticulation occurs when place of articulation of one sound extends to a neighboring sound and as a result the pronunciation of adja-cent sounds overlaps. Consider the following examples. (The label “Nom” is the abbreviation of Nominative Case particle.)

(3) a. ken made [kem made]ticket even