IPA Handbook Complete

The Handbook of the International Phonetic Association is a comprehensive guide to theAssociation's 'International Phonetic Alphabet'. The aim of the International PhoneticAlphabet is to provide a universally agreed system of notation for the sounds of languages,and for over a century the Alphabet has been widely used by phoneticians and othersconcerned with language. The Handbook presents the basics of phonetic analysis so thatthe principles underlying the Alphabet can be readily understood, and exemplifies the useof each of the phonetic symbols comprising the Alphabet. The application of the Alphabetis then extensively demonstrated by the inclusion of over two dozen 'Illustrations'

concise analyses of the sound systems of languages accompanied by a phonetictranscription of a passage of speech. These Illustrations cover languages from all over theworld. The Handbook also includes a range of other useful information. The 'Extensions'

to the International Phonetic Alphabet cover speech sounds beyond the sound systems oflanguages, such as those with paralinguistic functions and those encountered inpathological speech. A full listing is given of internationally agreed computer codings forphonetic symbols, including not only those of the International Phonetic Alphabet but alsothose of other traditions. And there is extensive information on the history of the'International Phonetic Association and its current activities. The Handbook is an essentialreference work for all those involved in the analysis of speech.

The International Phonetic Association exists to promote the study of the science ofphonetics and the applications of that science. The Association can trace its history back to1886, and since that time the most widely known aspect of its work has been theInternational Phonetic Alphabet. The Handbook has been produced collaboratively byleading phoneticians who have been on the Executive of the Association, and itincorporates (for instance in the case of the Illustrations) material provided by numerous

members of the Association world wide.

Sound files to accompany the book are available athttp://uk.cambridge.org/l ingu ist ics/resources/i pahand book.

Handbook of the International Phonetic Association

A guide to the use of the International Phonetic Alphabet

• CAMBRIDGE::: UNIVERSITY PRESS

I

CAMBRIDGE UNIVERSITY PRESS

Cambridge, New York, Melbourne, Madrid, C,ape Town, Singapore, Silo Paulo

Cambridge University PressThe Edinburgh Building, Cambridge CB2 8R~, UK

Published in the United States of America by Cambridge University Press, New York

www.cambridge.orgInformation on this title:www.cambridge.org/9780521652360

© The International Phonetic Association 1999

This publication is in copyright. Subject to statutory exceptionand to the provisions of relevant collective lice~sing agreements,no reproduction of any part may take place wi¢outthe written permission of Cambridge University Press.

First published 1999Eighth printing 2007

Printed in the United Kingdom at the University Press, Cambridge

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

ISBN-13 978-0-521-65236-0 hardbackISBN-IO 0-521-65236-7 hardbackISBN-13 978-0-521-63751-0 paperbackISBN-IO 0-521-63751-1 paperback

The IPA fonts and several language fonts used in the preparation and printing ofthis work are available from the Linguist's Software, Inc., PO Box 580, Edmonds,WA 98020-0580 USA. Tel: (425) 775-1130.

Cambridge University Press has no responsibility for the persistence or accuracy of URLsfor external or third-party internet websites referred to in this publication, and does notguarantee that any content on such websites is, or will remain, accurate or appropriate.

CONTENTS

Foreword vii

The IPA Chart ix

PART 1: Introduction to the IPA

1 What is the International Phonetic Alphabet? 3

2 Phonetic description and the IPA Chart 32.1 Linguistically relevant information in speech 42.2 Segments 42.3 The consonant-vowel distinction 62.4 Consonants 72.5 Non-pulmonic consonants 92.6 Vowels 102.7 Suprasegmentals 132.8 Diacritics 152.9 Other symbols 17

3 Guide to IPA notation 183.1 Exemplification of the symbols 183.2 Languages usedfor exemplification 25

4 The phonemic principle 27

5 Broad and narrow transcriptions 28

6 IPA transcriptions for a language 30

7 Working with the IPA 317.1 Symbol names 317.2 Using the 1PA in handwriting 317.3 Using the 1PA in print 317.4 Using the 1PA on computers 317.5 The /PA and braille 32

8 Going beyond the IPA 32

9 Some problematic issues 339./ Segmentation 339.2 Aligning transcriptions and speech 359.3 Transcribing the speaker or the hearer? 36

10 The IPA and phonological theory 37

-

+I

I

vi Contents

PART 2: Illustrations of the IPA

American EnglishAmharicArabicBulgarianCantoneseCatalanCroatianCzechDutchFrenchGalicianGermanHausaHebrewHindiHungarian19boIrishJapaneseKoreanPersian (Farsi)PortugueseSindhiSloveneSwedishTabaThaiTukang BesiTurkish

PART 3: Appendices

Appendix 1 The Principles of the International Phonetic Association

Appendix 2 Computer coding of IPA symbols

Appendix 3 Extensions to the IPA

Appendix 4 About the International Phonetic AssociationThe History of the AssociationStatutes and By-lawsHow tofind out more about the AssociationHow to join the Association

Appendix 5 Reference charts

4145515558616670747882869096

100104108III117120124126131135140143147151154

159

161

186

194194197199199

200

Foreword

The Handbook of the International Phonetic Association is a resource containing conciseinformation on the International Phonetic Alphabet and guidance on how to use it - a kindof 'user's manual'. It replaces the Principles of the International Phonetic Association,which has been out of print for some time and which had not been revised since 1949. Butalthough the Handbook replaces some of the functions of the old Principles, it is acompletely new work with wider objectives.

The old Principles contained a short tutorial on phonetic description, examples of theuse of phonetic symbols, and a large number of 'specimens' consisting of very briefcomments on the phonetics of a language and a transcription of the 'North Wind and the

Sun' text translated into the language. Additionally, there was some information about theAssociation, and, printed on the inside covers, a brief history of it.

The new Handbook broadly speaking retains these components. It is divided into threeparts: part I contains an introduction to phonetic description and exemplification of theuse of the symbols; part 2 consists of 'Illustrations' of the use of the International PhoneticAlphabet for different languages (these Illustrations are ones which have appeared in theJournal of the International Phonetic Association since 1989); and part 3 contains

appendices with a variety of reference material.Beyond the basic similarity of structure, the Handbook is very different from the old

Principles. Most superficially, perhaps, it reflects the changes which have been made inthat most tangible and widely known product of the Association's work, the IPA Chart.Discussion and exemplification is based on the most recent (1996) edition of the chart.

More substantively, the Handbook acknowledges the fact that over the past half centurythe advance of techniques for acoustic analysis means that many readers will be familiarwith, and quite possibly working with, speech as an acoustic signal. This means it nowseems appropriate to use an acoustic display such as a spectrogram not only as a way ofpresenting one facet of speech, but also to discuss problems which arise in the relationbetween a segmentally based system of notation and the physical speech event. TheHandbook will also contain practical information to do with the use of the IPA oncomputers, such as the computer codes for phonetic symbols.

The most fundamental differences between the old Principles and the new Handbookperhaps arise from the expectation that the readership of the new work will be much lesshomogeneous than that envisaged for the old one. The new Handbook is intended to be areference work not only for language teachers and phoneticians interested in the sounds ofdifferent languages, but also for speech technologists, speech pathologists, theoreticalphonologists, and others.

This breadth of readership is to be encouraged, given the goal that the InternationalPhonetic Alphabet (hereafter 'the IPA') should be a standard for the representation ofspeech. But it poses particular challenges for the writing of the 'tutorial' sections of theHandbook. The challenges are further increased by the vertical spread of readers fromthose who are experienced phoneticians to those who know nothing about phonetics. The

"

viii Foreword

breadth of readership has led perhaps to a more equivocal tone in the presentation of the

premises behind the IPA than in the Principles. For instance, the way in which the IPAdeveloped historically was closely boun8 up with a 'strictly segmented' phonemic view,

and in section 10 the fact that the+ are alternatives in phonological theory isacknowledged. The vertical spread of readers poses the recurring question of how much or

how little to say. The lower bound is pr~sumably what a novice needs to pick up in orderto have some idea of the principles governing the organization of the chart. The upper

bound is the practical goal of a compact booklet, readily affordable by students, andconcise enough to be easily digested by non-specialist readers.

The resulting text in part 1 is more discursive than that of the old Principles. It should,be borne in mind, however, that it does riot attempt the job either of a phonetics textbook,

or of a critique of the IPA. Nowadays th~re are many good phonetics textbooks available,

and it would be expected that students of phonetics would read one or more of these in

conjunction with the Handbook. The purpose of the Handbook is not to provide a

comprehensive or balanced education in: phonetics, but to provide a concise summary of

information needed for getting to grips with the IPA. Likewise, whilst a full-scale critique

of the assumptions on which the IPA is founded is perhaps due, the practically-oriented

Handbook is not the place for it. The IPI\ is a working tool for many, and whilst it may bepossible to improve that tool, the role of the Handbook is that of an instruction manual forthe too! which is currently available.

The creation of the Handbook has been in every sense a collaborative effort. The text

in part I is largely the responsibility of Francis Nolan, and the exemplification of the use

of sounds was provided by Peter Ladefoged and Ian Maddieson. Ian Maddieson, and

Martin Barry, as successive editors of the Journal of the International PhoneticAssociation, have been responsible for overseeing and collating the rich and ever growing

stock of lIIustrations. Martin Ball was instrumental in formulating the Extensions to theIPA (appendix 3), and Mike MacMahon wrote appendix 4 on the history of the

Association. John Esling is responsible for appendix 2 on the computer coding of symbols,and for most of the work involved in the final stages of preparing the Handbook includingthe final editing of the Illustrations. And, of course, particular thanks are due to the authors

of the Illustrations, and to the large number of members of the International Phonetic

Association who responded with suggestions and corrections when a draft of parts of the

Handbook was published in the Journal of the International Phonetic Association.

THE INTERNATIONAL PHONETIC ALPHABET (revised to 1993, updated 1996)

CONSONANTS (PULMONIC)

B.lnblOl Lnblodentnl Dent>!! IAlveolnr ]Pasl nlveol:u RctroOcx Pnl..nl Velar UVIliar Pharyngeal Glottnl

Plosivc p b t d t et c J k 9 q G [. ? F..Nasal m II] n 11. J1 I) N hg II ..~

Trill B r I R

Tap or Flap [ r ,- ';11

Fricative <I> 13 f V e oIs z II 3 ~ ~ <; J X Y X ff h l h fiLateral i E -~fricative ,,>11

Approximant U 1 { J UILmeral .. 1 1 A L l"'a IIapprOXimant ,

'"Where symbol!r. appenr In pairs. the one (Q the nght represents a vOIced consonilnl Shaded tife;)S denole nnlL:ulatlom Judged ImpossIble

CONSONANTS (NON-PULMONIC) VOWELS

oDIACRITICS Diacritics m:lY be placed above <l symbol with a dcscender, e.g. I)

o

D

uBack

Long

Half-long

Primary l;tress

Secondnry stress

,foun~'t1I~n

e:

Central

Where symbol!> appear In pOlrs, the oneto the nght represenls n rounded vowel.

SUPRASEGMENTALS

Y-I~-U1I Y U

e 0--~ 8--1(

g

E re-3\G-A

re B

a CE~a

Open-mId

Close

Close-mid

Open

Front

Alvcolo-p:l.lalnl fflCM1VCS

Alveolar Interal nap

Slmullaneous J Jnd X

A{{ncales and double aruculallonscnn be rcpresented by two ~ymbols

JOln~d by aIle bar I{ neces~ar)'

<;; 41fj

VOiced epigioltallncallvt:

VOiceless l;tbtnl-velar fnc.'lllVC

VOIced lablnl-velal <lpproximant

VOIcelesS cplglonnl fnenllve

VOiced lablal-palnlnl npproxlmanl

M.

W

Clicks VOiced Irnploslves Ejectlvcs

0 5,

BI]abl<l1 Bilabial Examples

I Dental d Oenl.tUalvcolnr p' Bllabl<ll

I (Posl)<llvcol,u f Pid;)l ...! t' DentaVnlvcolar

:f P<llalO<llvcolnr 9 Velar k' Velar

II Alveolar laternl G Uvulnr S' Alveolnr fncnllve

OTHER SYMBOLS

Voiceless lJ <} .. Drenlhy voiced R a Dental ! Q0 n

VVoiced § ! - Cre.'lky voiced g a ApiCi:tl ! 9-h th dh ! Q ~ QAspirmed - Linguoilibini

cL.'lminnl

y W tW dW - e, More rounded Labialized Nasalized

Less rounded ':)j

P<llat",lizcr1 tj dj n dn(

Nnsnl release

Advanced 1} Y Velarized tY dY I dl, L.'llcl7Il relense

e 'I t'l d~, d'- Relracted Phnryngcnhzed No ;'Iudible relense

.. e tCentralized - Vel:l.riz.ed or phnrynge:lhzedx x

Mid-centralized e . R;'Iised ~ <t = voiced olveolilr (ricali",e)

Syllnbic I).T

Lowered y (§ = voiced bilabial i1pprOXimanl)

~

Non·syl1ablc ~ ~Advnncetl Tongue ROOI y

~ Rholicily d'- a:- Retracted Tongue ROOI y.

Extrn-shon

I Minor (foot) group

II Major (illionalion) group

Syllnble brenk Ji.a:kt

Liuking (nbsence of a break)

TONES AND WORD ACCENTSLEVEL CONTOUR

eN l EXlr;)eor 11 RisingIllSh

e 1 High e \J Falling

e -1J

'1 HighMid e risinge --1 Low e "j Lowrising

e J EXIra e 1 Rising-low falling

J,Downstcp /' GlobJ.l rise

i Upstep \. Global {all

PART 1

Introduction to the IPA

1 What is the International Phonetic Alphabet?The aim of the International Phonetic Association is to promote the study of the science ofphonetics and the various practical applications of that science. For both these it isdesirable to have a consistent way of representing the sounds of language in written form.From its foundation in 1886 the Association. has been concerned to develop a set ofsymbols which would be convenient to use, but comprehensive enough to cope with thewide variety of sounds found in the languages of the world; and to encourage the use ofthis notation as widely as possible among those concerned with language. The system isgenerally known as the International Phonetic Alphabet. Both the Association and itsAlphabet are widely referred to by the abbreviation IPA, and here the Alphabet will

generally be abbreviated to 'the IPA'. The IPA is based on the Roman alphabet, which hasthe advantage of being widely familiar, but also includes letters and additional symbolsfrom a variety of other sources. These additions are necessary because the variety ofsounds in languages is much greater than the number of letters in the Roman alphabet. Theuse of sequences of phonetic symbols to represent speech is known as transcription.

The IPA can be used for many purposes. For instance, it can be used as a way to show

pronunciation in a dictionary, to record a language in linguistic fieldwork, to form thebasis of a writing system for a language, or to annotate acoustic and other displays in theanalysis of speech. For all these tasks it is necessary to have a generally agreed set ofsymbols for designating sounds unambiguously, and the IPA aims to fulfil this role. Thepurpose of this Handbook is to provide a practical guide to the IPA and to the conventionsassociated with it.

Phonetics, like any science, develops over time. New facts emerge, new theories arecreated, and new solutions to old problems are invented. The notational system of anyscience reflects facts and theories, and so it is natural that from time to time the Alphabetshould be modified to accommodate innovations. The Alphabet presented in thisHandbook is the version revised by a Convention of the International PhoneticAssociation held in Kiel in 1989, subject to a subsequent set of minor modifications and

corrections approved by the Council of the Association. Despite these and earlier changes,the Alphabet today shows striking continuity with the Association's Alphabet as it was atthe end of the nineteenth century. The development of the IPA has, throughout the historyof the Association, been guided by a set of 'Principles', and these are listed in appendix I.

2 Phonetic description and the IPA ChartBehind the system of notation known as the IPA lie a number of theoretical assumptionsabout speech and how it can best be analyzed. These include the following:- Some aspects of speech are linguistically relevant, whilst others (such as personal voice

quality) are not.- Speech can be represented partly as a sequence of discrete sounds or 'segments'.- Segments can usefully be divided into two major categories, consonants and vowels.- The phonetic description of consonants and vowels can be made with reference to how

4 Harulbook of/he [PA

they are produced and to their auditory cha~acteristics.

- In addition to the segments, a number o~ 'suprasegmental' aspects of speech, such asstress and tone, need to be represented independently of the segments.

The IPA is summarized in the 'IPA Ch~rt', which is reproduced in its entirety after theforeword, and section by section in appenqix 5; readers are encouraged to photocopy andenlarge the Chart for ease of reference. Th~ structure of the Chart reflects the assumptionsabove. The following subsections provide a brief introduction to phonetic description inthe context of these underlying assumptiops, while referring to the relevant parts of theChart. This introduction can only deal with a few important points, and readers who needa more thorough treatment of phonetic desc:ription should consult phonetics textbooks.

In introducing the IPA, it is necessary to refer to examples from languages. As far aspossible, languages are used with which ~any readers may be acquainted, but of coursethis is not possible for many sounds. Variation also provides a problem: all languages havedifferent accents and other varieties of pronunciation. When a sound is exemplified by aword in a particular language, this should, be taken to mean that the sound can often beheard in that word, not that it will always occur in pronunciations of the word.

In the case of English, reference is made here mainly to two varieties, GeneralAmerican and Standard Southern British. These varieties are widely heard, in the UnitedStates and Britain respectively, in formal areas of broadcasting such as newscasts. GeneralAmerican is regarded as a variety which transcends regional divides. Standard SouthernBritish (where 'Standard' should not be taken as implying a value judgment of'correctness') is the modern equivalent of what has been called 'Received Pronunciation'CRP'). It is an accent of the south east of 9ngland which operates as a prestige norm thereand (to varying degrees) in other parts of the British Isles and beyond. Where necessary,reference will be made specifically to one of these varieties, but normally the term'English' will be used, indicating that th: sound occurs in both General American andStandard Southern British. In most cases of course, the sound will also occur in manyother varieties of English. I

2.1 Linguistically relevant information .if! speechAlthough phonetics as a science is interested in all aspects of speech, the focus of phoneticnotation is on the linguistically relevant aspects. For instance, the IPA provides symbols totranscribe the distinct phonetic events corresponding to the English spelling refuse(['refjus] meaning 'rubbish' and [n'fjuz]' meaning 'to decline'), but the IPA does notprovide symbols to indicate information s~ch as 'spoken rapidly by a deep, hoarse, malevoice'. Whilst in practice the distinction between what is linguistically relevant and whatis not may not always be clear-cut, the principle of representing only what is linguisticallyrelevant has guided the provision of symbols in the IPA. The need to go further, however,is now recognized by the 'Extensions to the IPA' presented in appendix 3.

2.2 SegmentsObservation of the movements of the speech organs reveals that they are In almost

Introduction to the IPA 5

continuous motion. Similarly the acoustic speech signal does not switch betweensuccessive steady states, but at many points changes gradually and at others consists ofrapid transient events. Neither the movements of the speech organs nor the acoustic signaloffers a clear division of speech into successive phonetic units. This may be surprising tothose whose view of speech is influenced mainly by alphabetic writing, but it emergesclearly from (for instance) x-ray films and acoustic displays.

For example, the movements and the acoustic signal corresponding to the Englishword worry will show continuous change. Figure I presents a spectrogram of this word.Spectrograms are a way of making visible the patterns of energy in the acoustic signal.Time runs from left to right, and the dark bands reflect the changing resonances of thevocal tract as the word is pronounced. In the case of the word worry, the pattern ebbs andflows constantly, and there are no boundaries between successive sounds. Nonetheless theword can be segmented as [wl:ui] - that is, as [w] + ['B] + [1] + [i]. This segmentation isundoubtedly influenced by knowledge of where linguistically significant changes in soundcan be made. A speaker could progress through the word making changes: in a Britishpronunciation, for instance, [w'B1i] worry, [h'B1i] hurry, [hre1i] Harry, [hreti] Hatty, [hret;}]hatter. There are thus four points at which the phonetic event can be changed significantly,and this is reflected in the analysis into four segments. Languages may vary in the pointsat which they allow changes to be made, and so segmentation may have to be tentative ina first transcription of an unknown language (see section 9). Nonetheless there is a greatdeal in common between languages in the way they organize sound, and so many initial

guesses about the segmentation of an unfamiliar language are likely to be right.

8

7:- + + + + + + + +6 - + + + + + + +

N' + + + + + +g 5

>,4 - + +

!11~t~l~+(,)

cQ)::l0' 3:" + +Q)

<l::

'1\ ~~i2 + ++ + +,I " I,

00.1 0.3 0.4 0.5 0.6 0.7 time (s)

.W B 1 1

Figure 1 Spectrogram of the word worry, spoken in a Southern British accent.

6 Handbook of the IPA

Phonetic analysis is based on the crucial ~remise that it is possible to describe speechin terms of a sequence of segments, and on the further crucial assumption that eachsegment can be characterized by an articulatory target. 'Articulation' is the technical termfor the activity of the vocal organs in making ia speech sound. The description of the targetis static, but this does not imply that the articulation itself is necessarily held static. So, forexample, [J] (as in the word worry above) is ~escribed as having a narrowing made by thetongue-tip near the back of the alveolar ridge (the flattish area behind the upper frontteeth). The tongue-tip actually makes a continuous movement to and from that target, as

reflected in the dipping pattern of higher resonances on the spectrogram in figure Ibetween 0.4 and 0.5 s. In other sounds, a ta~get will be held for a fixed amount of time.The important point is that the use of segme~ts and associated 'target' descriptions allowsfor a very economical analysis of the complel< and continuously varying events of speech.

2.3 The consonant-vowel distinctionBroadly, speech involves successive narrowing and opening of the vocal tract, the passagethrough which the air flows during speech. Tohis can be seen clearly in an example such asbanana ([~'nreng] or [~'nong]), in which the vocal tract is closed three times (first by the

lips and then twice by the tongue), each closl,lre being followed by an opening of the vocaltract. The successive openings are the basis of syllables, and the word banana consiststherefore of three syllables. The open parit of the cycle is regarded as the centre, ornucleus, of the syllable.

Sounds like [b) and en] which involve a closed, or nearly closed, vocal tract, areconsonants. Sounds like [g) and [a] which involve an open vocal tract are vowels. Moreprecisely, any sounds in which the flow of air out of the mouth is impeded at least enoughto cause a disturbance of the airflow are cOI,lsonants. So a sound such as [s], in which the'hissing' that can be heard results from th~ airflow being made turbulent, is as much aconsonant as [b). Conversely any soundk in which the air flows out of the mouthunimpeded are vowels. The distinction bet~veen consonant and vowel is fundamental tothe way segments are described in the framework underpinning the IPA.

It follows from the definitions of 'consonant' and 'vowel', and from the origin of thesyllable in the repeated opening and narrowing of the vocal tract, that vowels are well

suited to playing the role of syllable nuclei,! and consonants are well suited to defining themargins of syllables. The relationship bbtween syllables and type of sound is not,however, totally straightforward. For one thing, a sound which is a consonant maynonetheless act as a syllable centre. So in a common pronunciation of the English wordbutton as [bAt!)] there are two syllables, but the nucleus of the second is a consonant, asjudged from the way it is produced. Conversely in the word Uet] yet, the first sound, ifprolonged, is very similar to the vowel of [hid] heed, and does not involve a narrowingextreme enough to produce friction. However because [j] plays the same role in thesyllable as sounds which are by definition consonants (e.g. [b) in [bet] bet), it is oftenincluded in the class of consonants and described accordingly.

On the IPA Chart, there are separate sections for vowels and for consonants, reflecting


different techniques for describing them. The different techniques arise from the moreclosed articulation of consonants and the more open articulation of vowels.

2.4 ConsonantsBecause consonants involve a narrowing or 'stricture' at an identifiable place in the vocaltract, phoneticians have traditionally classified a consonant in terms of its 'place ofarticulation'. The [tl of len, for instance, requires an airtight seal between the upper rim ofthe tongue and the upper gum or teeth. Phonetic description of place of articulation,however, concentrates on a section or 'slice' through the mid-line of the vocal tract, themid-sagittal plane as it is known, and in this plane the seal is made between the tip orblade of the tongue and the bony ridge behind the upper front teeth, the alveolar ridge. Thesound is therefore described as alveolar. Figure 2 shows a mid-sagittal section of the vocaltract, with the different places of articulation labelled. As further examples, the [pJ of penis bilabial (the closure is made by the upper and lower lips), and the [k] of Ken velar(made by the back of the tongue against the soft palate or velum). Other places ofarticulation are exemplified in section 3.

~ velari;-'I>\.7>

<1

Figure 2 Mid-sagittal section of the vocal tract with labels for place of articulation

On the IPA Chart, symbols for the majority of consonants are to be found in the largetable at the top. Place of articulation is reflected in the organization of this consonanttable. Each column represents a place of articulation, reflected in the labels across the topof the table from bilabial at the left to glottal (consonants made by the vocal cords or vocalfolds) at the right. The terms 'bilabial' and 'labiodental' indicate that the consonant is

made by the lower lip against the upper lip and the upper front teeth respectively;

I

I8 Handbook of the IPA !

Iotherwise it is normally assumed that the s9und at a named place of articulation is made

by the articulator lying opposite the place qf articulation (so alveolars are made with thetip of the tongue or the blade (which lies jUft behind the tip». The exception to this is theterm 'retroflex'. In retroflex sounds, the tiR of the tongue is curled back from its normalposition to a point behind the alveolar ridg~. Usually alveolar [l] shares some degree ofthis curling back of the tongue tip, which d'stinguishes it from other alveolars. Note thatexcept in the case of fricatives only one symbol is provided for dental/alveolar /

postal veolar; if necessary, these three plafes can be distinguished by the use of extra

marks or 'diacritics' to form composite SYlllbols, as discussed in section 2.8. For example,

the dental/alveolar / postalveolar nasals ca* be represented as [I} n 0] respectively.

The rows of the consonant table, labe'led at the left side by terms such as plosive,

nasal, trill, and so on, reflect another major descriptive dimension for consonants, namely'manner of articulation'. Manner of articulation covers a number of distinct factors to do

with the articulation of a sound. One is th~ degree of stricture (narrowing) of the vocal

tract involved. If the articulation of the pJosive it] is modified so that the tongue tip or

blade forms a narrow groove running from. front to back along the alveolar ridge, instead

of an airtight closure, air can escape. The airflow is turbulent, and this creates sound of a

hissing kind known in phonetics as fricatiort Such a sound is called a fricative. In this case

the resultant sound would be [s] as in sin. Glther fricatives include [f] (as infin) and [f] (as

in shin). If even less narrowing is made in' the vocal tract, an approximant will result, in

which the airflow is not turbulent and no frication is audible. Approximants are

exemplified by the sound OJ at the start of .yet, and the first sound in red in most varietiesof English ([l], [U, or [0] according to the variety).

'Manner of articulation' also includes important factors such as whether the velum (the

soft part of the palate at the back of the mouth) is raised or lowered. If it is lowered, as for

the sounds [m] and [n] in man, the resonances of the nasa) cavity will contribute to the

sounds. Consonants where this happens ate called nasals. Laterals (lateral approximantsI

such as English [I] in let and lateral fricaiives such as Welsh it] in llan 'church (place-

name element)' are sounds where air escapes not in the mid-line of the vocal tract but atthe side. Trills are sounds like [r] in Spanish perro 'dog' in which the air is repeatedly

interrupted by an articulator (in this case the tongue tip) vibrating in an airstream. A very

short contact, similar in duration to one cycle of the vibration of a trill, is called a tap, suchas the [f) in Spanish pero 'but'.

A further important factor in the description of consonants is not shown in the columnor row labels. This is whether the conson~nt is voiced or voiceless. In voiced consonants

the vocal cords are producing acoustic energy by vibrating as air passes between them,

and in voiceless ones they are not. A symbol on the left of a cell in the table is for a

voiceless consonant, e.g. [p] and [7], and 'one on the right is for a voiced consonant, e.g.

[b) (the voiced counterpart of [pJ) and [m]. Voicing distinctions are actually more fine

grained than implied by this two-way distinction, so it may be necessary to add to the

notation allowed by the two basic symbols. For instance, the symbolization [ba pa pha]implies consonants in which the vocal cords are, respectively, vibrating during the plosive


closure, vibrating only from the release of the closure, and vibrating only from a time wellafter the release (giving what is often known as an 'aspirated' plosive). Where a cellcontains only one symbol, it indicates (with one exception) a voiced consonant and isplaced on the right. The exception is the glottal plosive (7] (as the vocal cords are closed,they are unable simultaneously to vibrate).

It should be clear that the consonant table is more than a Jist of symbols; it embodies aclassificatory system for consonants. It allows the user to ask a question such as 'howshould I symbolize a voiced sound involving complete closure at the uvula?' (The answeris (0].) Or conversely, 'what sort of a sound is [j]?' (The answer is one which is voiced,and in which frication can be heard resulting from a narrowing between the tongue front

and the hard palate.)Not all cells or halves of cells in the consonant table contain symbols. The gaps are of

three kinds. Shaded cells occur where the intersection of a manner and a place ofarticulation define a sound which is thought not to be possible, either by definition (a nasalrequires an oral occlusion combined with lowering of the velum, and so a pharyngeal orglottal nasal is ruled out), or because the sound is impossible or too difficult to produce,such as a velar trill or a bilabial lateral fricative. Unless phoneticians are mistaken in theirview of the latter category of sound, no symbols wi]! be needed for any of the shadedcells. An unshaded gap, such as the velar lateral fricative, may indicate that the sound inquestion can be produced, but has not been found in languages. It is always possible that alanguage will be discovered which requires the gap to be filled in. A case of this kind isthe velar lateral approximant (L], which only became generally known among phoneticiansin the 1970s when it was reported in Kanite, a language of Papua New Guinea. An

unshaded gap may also occur where a sound can be represented by using an existingsymbol but giving it a slightly different value, with or without an added mark separatefrom the symbol. A symbol such as (13], shown on the chart in the position for a voicedbilabial fricative, can also be used to represent a voiced bilabial approximant if needed. Ina similar way, no symbols are provided for voiceless nasals. A voiceless alveolar nasal canbe written by adding the voiceless mark L] below the symbol (n] to form an appropriatecomposite symbol (I}]. Many of the gaps on the chart could be filled in this way by the useof diacritics (sections 2.8 and 3). The formation of this kind of composite symbol isdiscussed further in the section on diacritics below.

2.5 Non-pulmonic consonantsAll the symbols in the main consonant table imply consonants produced using air from thelungs ('pulmonic' consonants). Whilst some languages rely exclusively on air from thelungs for sound production, many languages additionally use one or both of two other'airstream mechanisms' to produce some of their consonants. Symbols for these soundsare given in a separate box below and to the left of the main consonant table. Thesesounds are exemplified in section 3.

The more common of the two non-pulmonic airstream mechanisms used in languages,the 'glottalic', involves closing the glottis, and squeezing or expanding the air trapped


between the glottis and a consonant stricture further forward in the vocal tract. If the air issqueezed, and therefore flows outwards ~ abruptly when a closure further forward is

Ireleased, or briefly but continuously through a fricative stricture - the sound is known asan 'ejective'. Ejectives are symbolized b~ the appropriate voiceless consonant symbolwith the addition of an apostrophe, e.g. [p'O, [s']. If instead the air between the glottis anda closure further forward is expanded, red6cing its pressure, air will flow into the mouthabruptly at the release of the forward closure. Usually the closure phase of such sounds isaccompanied by vocal cord vibration, giving '(voiced) implosives' such as [6]. If it is

necessary to symbolize a voiceless version of such a sound, this can be done by adding adiacritic: [9].

'Velaric' airstream sounds, usually known as 'clicks', again involve creating an

enclosed cavity in which the pressure of ~he air can be changed, but this time the backclosure is made not with the glottis but with the back of the tongue against the soft palate,such that air is sucked into the mouth when the closure further forward is released. The'tut-tut' or 'tsk-tsk' sound, used by many English speakers as an indication of disapproval,is produced in this way, but only in isolation and not as part of ordinary words. Some

other languages use clicks as consonants. A separate set of symbols such as [i) isprovided for clicks. Since any click involves a velar or uvular closure, it is possible to

symbolize factors such as voicelessness, v~icing, or nasality of th~clic~by combining theclick symbol with the appropriate velar or uvular symbol: [kt if IJiJ. [ql].

2.6 VowelsVowels are sounds which occur at syllable centres, and which, because they involve a lessextreme narrowing of the vocal tract than consonants, cannot easily be described in terms

of a 'place of articulation' as consonants fan. Instead, they are classified in terms of anabstract 'vowel space', which is representep by the four-sided figure known as the 'VowelQuadrilateral' (see the Chart, middle rigHt). This space bears a relation, though not an

exact one, to the position of the tongue in v~owel production, as explained below.Figure 3 shows a mid-sagittal section of the vocal tract with four superimposed

outlines of the tongue's shape. For the vowel labelled [i), which is rather like the vowel ofheed or French si 'if', the body of the tongue is displaced forwards and upwards in themouth, towards the hard palate. The diagram shows a more extreme version of this vowelthan normally found in English at least, m~de so that any further narrowing in the palatalregion would cause the airflow to become turbulent, resulting in a fricative. This extremevowel is taken as a fixed reference point for vowel description. Since the tongue is nearthe roof of the mouth this vowel is described as 'close', and since the highest point of the

tongue is at the front of the area where vo~el articulations are possible, it is described as

'front' .

Introduction to the IPA II

Figure 3 Mid-sagittal section of the vocal tract with the outline of the tongue shapefor each of four extreme vowels superimposed.

Conversely, for the vowel labelled [0], which is rather like the vowel of StandardSouthern British or General American English palm, the tongue body is displaceddownwards and backwards, narrowing the pharynx. The most extreme version of thisvowel, made so that any further narrowing in the pharynx would result in a fricative, istaken as a second fixed reference point. The space between the tongue and the roof of themouth is as large as possible, so this vowel is described as 'open', and the tongue is nearthe back of the mouth, so it is described as 'back'.

If the tongue body is raised as close as possible at the back of the mouth, just short ofproducing a velar consonant, and (as is common in languages) the lips are simultaneouslyrounded and protruded, the close back vowel [u] results (see figure 3), which is similar tothe vowel of French vous 'you' or German du 'you'. And if a vowel is produced in whichthe highest point of the tongue is at the front of the mouth and the mouth is as open aspossible, the result is [a]. This is rather like the quality of the vowel in cal incontemporary Standard Southern British English (other dialects may have less openqualities or less front qualities). These two extreme vowels may also be regarded as fixedreferences.

The first part of figure 4 shows that joining the circles representing the highest point ofthe tongue in these four extreme vowels gives the boundary of the space within whichvowels can be produced. For the purposes of vowel description this space can be stylizedas the quadrilateral shown in the second part of figure 4. Further reference vowels can nowbe defined as shown in the third part of figure 4. Specifically, two fully front vowels [e]and [e] are defined between [i] and [a] so that the differences between each vowel and the

next in the series are auditorily equal; and similarly, two fully back vowels [:J] and [0] aredefined to give equidistant steps between [0] and [u]. The use of auditory spacing in the


definition of these vowels means vowel de' cription is not based purely on articulation, andis one reason why the vowel quadrilateral must be regarded as an abstraction and not adirect mapping of tongue position. These vowels and those defined below are exemplifiedin section 3.

lor------.--------, U

'-------'-- 0

1.-------...,---------, U

'<---\-----1 J

'-------'--0

1

Figure 4 The vowel quadrilateral and dardinal vowels. Above, the relation between thevowel quadrilateral and the vowels s~own in figure 3; below, the primary cardinalvowels and all cardinal vowels. I

There are now four defined vowel heiJhts: [i] and [u] are close vowels, [e] and [0] areclose-mid vowels, [e] and [::>] are open-mid vowels, and [a] and [a] are open vowels (notethat in this last pair the difference in letter shape is important, signifying a front vowel anda back vowel respectively). The vowel space can be seen to be taking on the form of agrid. The eight reference vowels are known as the 'primary cardinal vowels'. 'Cardinal' inthis sense refers to points on which the system of description hinges. The description ofthe primary cardinal vowels outlined above differs slightly from that of the Englishphonetician Daniel Jones who first defined them, but is in accord with a widespreadconception of them today. The primary cardinal vowels are often referred to by numbersordered anticlockwise round the quadrilateral: I [i], 2 [e], 3 [e], 4 [aJ, 5 [a], 6 [::>J, 7 [oJ.8 [u].

So far, lip activity has been largely ignored. In the back series of cardinal vowels([a ::> 0 u]) lip-rounding progressively increases, from none on [a] to close rounding on


[u]. By convention unrounded vowels are placed to the left of the front or back line of thequadrilateral, and rounded vowels to the right. Conversely in the front series [a e: e i) thelips are neutral for [a], and become progressively more spread through the series to [i).The fact that [i e e: a 0] are unrounded, and [:> 0 u] have increasing rounding, reflects arelationship commonly found in languages between vowel height, backness, and rounding.Lip activity is, however, independent of tongue position, and many languages exploit thisin their vowel systems.

To reflect this, eight 'secondary cardinal vowels' are therefore defined which differonly in lip position from their primary counterparts. These are shown paired with theirprimary counterparts in the fourth part of figure 4. So, for example, of the close vowels [i

y w u], [i w] have spread lips and [y u] have closely rounded lips; and of the open-midvowels [e re A :>], [e A] have slightly spread lips and [re :>] have open rounding. A furthertwo secondary cardinal vowels are defined; these are the close central vowels [i] (spread)and [B] (close rounded). The secondary cardinal vowels are sometimes referred to by thenumber of the corresponding primary cardinal vowel, for example [~] is 'secondarycardinal 2', or they are numbered anticlockwise round the quadrilateral from 9 [y] to 16[w]. [i] and [101] are then numbered 17 and 18 respectively.

The complete set of IPA vowel symbols is shown in the quadrilateral on the Chart. Inaddition to the cardinal vowels already discussed, which lie on the outside edge of thequadrilateral, there are symbols for mid central vowels, and for vowels at a number of

intermediate locations. There is a pair of symbols for unrounded and rounded close-midcentral vowels, [~ 8], and a corresponding open-mid pair [30]. The vowel [d], oftenreferred to as schwa, lies at the middle of the vowel quadrilateral, and [e] lies betweenopen-mid and open. The vowels [I Y u] are mid-centralized from [i y u] respectively.

Since the vowel space is continuous, it is a matter of chance whether a vowel in a

language exactly coincides with one of the reference points symbolized on thequadrilateral. In particular, languages may use vowels which are similar to, but not asperipheral as, the reference points indicated by the cardinal vowels. If detailed phoneticdescription is required, most vowels in a language have to be placed in relation to areference vowel, for instance 'a vowel centralized and lowered from cardinal [er. Thisdescription can be symbolized by adding diacritics (see section 2.8) to the cardinal vowelsymbol: [~].

2.7 SuprasegmentalsA number of properties of speech tend to form patterns which extend over more than onesegment, and/or to vary independently of the segmental targets. This is particularly true of

pitch, loudness, and perceived timing. These properties are often referred to as'suprasegmenta]s', and part of the process of phonetic analysis is the separation of theseproperties from the rest of the speech event. The IPA provides a separate set of symbolsfor suprasegmentals, to be found on the Chart at the bottom right corner.

Pitch variation, for instance, can operate over complete utterances to convey meaningadditional to that of the words in what is generally termed 'intonation'. This is true in all

I

14 Handbook of the IPA I

languages, though the complexity of the i~tonational system varies across languages. The

symbol [it] can be used to mark the end o( the domain of an intonation pattern, and [[] todemarcate a smaller unit. The symbols [/1'] for 'global rise' and 'global fall' respectivelymay also be useful for intonation, altho~gh a complete intonational transcription willrequire symbols not provided on the IPA Ghart.

Another domain of pitch variation is 'the word or syllable, and such pitch variationserves to distinguish words in much the same way as their segmental make-up does.

Languages in which pitch has this function are called tone languages, and are thought to

form a majority of the languages of the w<jlrld. In Thai, for example, [kha:'J] with a falling

pitch (indicated by the diagonal part o~ the symbol following the segments) means

'servant' and [kha:A] with rising pitch mears 'leg'.

The IPA has two alternative sets of symbols for indicating tones. In languages inwhich lexical contrasts are predominantly dependent on the pitch movement on each

syllable, such as Thai and the various forms of Chinese, so-called tone letters are often

used. These letters, as in the Thai exampl~s, indicate the tone of the preceding syllable by

a vertical stroke with a line preceding it. the vertical stroke is assumed to represent five

possible pitch heights within the speaker's range, and the position of the line shows the

height and movement (if any) of the pitch on the preceding syllable. The tone letters are

often used to indicate general tone movements. For example, if there is only one fallingtone in a language, and no strong reason to draw attention to the particular level of its

endpoints, it can be noted as going from the highest to the lowest level. Thus a

transcription of the Standard Chinese word for 'scold' is [ma'J], although most Chinese

speakers will not produce this syllable with a fall extending through their whole pitch

range. It is also possible to use the tone letters to show more detailed transcriptions for

certain purposes. Thus, the Thai high to~e can be transcribed with the symbol [1]; but

measurements of the fundamental frequency in high tone syllables show that there isactually a rise and a fall in syllables of this! sort, so the tone could be represented as [1 ].

The other IPA system for transcribing tone has often been used for languages in which

tonal contrasts depend predominantly on the pitch height in each syllable. There are three

diacritics, corresponding to high [e], mid [e] and low [e] tones, which can be placed above

the segment bearing the tone (here exemplified by feD. Thus the three tones in the West

African language Yoruba can be transcribed as exemplified in the phrases [6ba] 'he/she

met', f6 ba] 'he/she hid', [6 ba] 'it perched'. Notice that these tone symbols must not be

interpreted as iconic; that is, although the 'acute accent' n looks rising, it in fact means'high'. To represent a rising tone it is necessary to combine a 'low' and a 'high', and

similarly for other contour tones. So a syllable such as [e] occurring on a rising tone is [e],and on a falling tone is [e]. On the other hand the 'tone letters' such as [ell (meaning

'high') and reA] (meaning 'rising') are directly iconic.

The chart shows the tone letter [1] as if it were equivalent to n, the extra-high tone

symbol in the other set of symbols, and so on down the scale. But this is done only to

simplify the layout of the chart. The two sets of symbols are not comparable in this way.The four tones of Standard Chinese are often symbolized as [mal] 'mother', [ma'J]


'scold', [ma1] 'hemp', [mavl] 'horse'. If they were transcribed in the other system theywould be [ma ma ma ma].

The symbols [t] for upstep and [J.] for downstep are used to show modifications(raising or lowering) of the pitch indicated by ordinary tone symbols. Upstep occurs, forexample, in Rausa in that the last of a series of high toned syllables before a low tone is

pronounced with a higher pitch than the others. Thus the Hausa word [turantjl] 'English'has three high tones with the same pitch when said by itself. In the phrase [turanttJl ne] 'it

is English', the raising of the high tone can be indicated by the modifier [t] as shown.

Downstep occurs in the Ghanaian language Akan, as in the word [:'ik:5J.t:5] 'crab' which

has a downstepped high tone on the last syllable. This tone is demonstrably a high in that

it has the same pitch level as an initial high tone in a following word.Symbols are also provided for indicating the relative prominence or stress of syllables,

differing segmental length, and syllable divisions. The exact nature of syllable prominence

or stress varies from language to language, but the IPA provides for up to three degrees ofprominence to be indicated; in [,prerasaI'kDlad3iJ parapsychology the highest level occurs

on the fourth syllable, and the second highest on the first syllable, while the remaining

unmarked syllables are less prominent (a further division among these may be inferred

from vowel quality, those syllables with [a] being least prominent in English). Extra

strong stress can be indicated by doubling the stress mark: [a"meIZIIJ] amazing! Segmental

length can be marked on a continuum from short to long as [e e e' e:], though the

possibility exists to show even greater length as [e::]. Syllable divisions, which it may be

useful to indicate for phonological reasons or where the syllable division determines

phonetic difference as in [naI.l,1elt] nitrate versus [nalt..Ieit] night-rate, can be symbolizedas shown. The use of suprasegmental symbols is further demonstrated in section 3.

2.8 DincriticsDiacritics are small letter-shaped symbols or other marks which can be added to a vowel

or consonant symbol to modify or refine its meaning in various ways. A symbol and any

diacritic or diacritics attached to it are regarded as a single (complex) symbol. The set of

diacritics approved by the International Phonetic Association is given in the table at the

bottom left of the Chart.A number of diacritics deal with different aspects of phonation. Two are available to

reverse the voicing value otherwise implied by any symbol. Voiceless trills or nasals, for

instance, for which there are otherwise no symbols, can be notated as [r], (1)] etc. (some

diacritics may be placed above a symbol when a descender on the symbol would interfere

with legibility). Vowels which occur without voicing can also be indicated, e.g. [~]. More

rarely employed is [.1 which indicates voicing in a symbol otherwise implying

voicelessness. It sometimes indicates the spreading of voicing from an adjacent segment('assimilation' of voicing), as in French [Ja~3ulf] chaque jour 'each day'. It is a moot

point whether [~] and [9] refer to phonetically identical sounds, and likewise [s] and [~]. Itis possible that the distinction between [k] and [9] or between [sl and [z] can involve

dimensions independent of vocal cord vibration, such as tenseness versus laxness of

16 Handbook of/he [PA

articulation, so that the possibility of not'lting voicing separately becomes important; but

in any case, it can be convenient to be aBle to preserve the lexical shape of a word (e.g.French [Jak] chaque 'each') while noting kssimilation. The diacritic [h] is used to indicate

a release of air after a consonant, most co~monly between a voiceless plosive and a vowelas in [thaI] tie. Two different phonation types which are used contrastively by some

languages, creaky voice and breathy voic6, can be indicated on vowels or consonants (seeexamples at the top of the second column of the diacritic table).

The diacritics shown in rows four to nine of the first column of the diacritic table,together with the diacritics for 'raised' and 'lowered' shown to the right in rows nine and

ten, can be used to modify the lip or tong1.1e position implied by a vowel symbol. Thus [y]indicates a vowel like cardinal [u] but witH a lip position further from the 'rounded' end of

the 'spread-rounded' continuum than implied by the cardinal symbol, and (as seen in

section 2.6 above) [~] indicates a vowel centralized and lowered from cardinal [ej. Vowel

qualities between [u] and [tI] might be symbolized [!:!], indicating retraction relative to thecentral vowel, or (if nearer back than central) [y] or [til indicating fronting relative to the

back vowel. The diacritic for 'mid-centralized' indicates a quality displaced in the

direction of the mid central vowel [;J]; thus: [e] is equivalent to [n and [0] to [9].The diacritics for 'raised' and 'lowered', when applied to a consonant symbol, change

its manner category, so that [I] could be lused to indicate an articulation like that of analveolar plosive but one in which complete closure is not achieved, yielding a fricative

like sound (but lacking the grooved tongue shape of [s)) as in some Irish English

pronunciations of the sound at the end of right. The diacritics for 'advanced' and

'retracted' are also commonly used to modify consonant place of articulation. So, for

instance, a voiceless fricative at the front of the velar region could be symbolized [Ji.], anda specifically postalveolar nasal [JJj.

The 'rhoticity' diacritic [-] indicates a vowel with a speyific auditory effect like that ofthe vowel in General American [fCl'] far and [f;l' ]fur (the combination of the 'rhoticity'

diacritic with [;J] is often written and printed [a-]). The auditory effect is probably caused

by a constriction in the pharynx combined with an expansion of the space in the mouth infront of the tongue, either by curling the tongue tip up and back, or by retracting it into the

tongue body while 'bunching' the tongue body up towards the pre-velar region. In some

languages the tongue root functions independently of other determinants of vowel quality,adjusting the width of the pharynx, and at the bottom right of the table there are two

diacritics to indicate advancement and retraction of the tongue root. The 'syllabic'diacritic is used to mark consonants which are acting as syllable nuclei, and the non

syllabic diacritic to mark vowels which are' not fulfilling their customary syllabic role.The 'dental' diacritic (third column) modifies those consonant symbols found under

'alveolar' to indicate unambiguously a dental articulation. As noted in section 2, although

only one symbol is provided in the consonant table (except in the fricative row), dental /

alveolar / postalveolar can be distinguished as [IJ n 0] (postaJveolar being marked by the

'retracted' diacritic). The 'linguolabial' diacritic, which is used to symbolize an otherwise

omitted (and very rare) consonantal type, indicates a sound made with the tip or blade of


the tongue against the upper lip. The diacritic is used to modify the relevant alveolar

consonant symbol. The diacritics for 'apical' and 'lamina!' make specific which part of thefrontmost area of the tongue is making an articulation: the tip (apical), or the blade(laminal).

Secondary articulations are narrowings of the vocal tract which are less narrow than

the main one producing a consonant. The names palatalization, velarization, and

pharyngealization, make explicit where the narrowing is. In one sense a secondary

articulation is the superimposition of a close-vowel-like articulation on a consonant - [i)for palatalization, symbolized for instance [tj ), [ill) for velarization ([tV)), and [a) for

pharyngealization ([t')). Labialization strictly means reduction of the opening of the lips.

However it has tended to be used for the commonly found combination of rounding(protrusion) of the lips accompanied by velar constriction. It is for such labially roundedvelarization that the superscript [W) is most appropriate. If it is necessary to distinguish a

secondary reduction of the lip opening accompanied by neither protrusion nor velar

constriction, a superscript [IlJ (the symbol for a labiodental approximant) might be used.

These superscript diacritics which are placed after the symbol look rather as if they imply

a sequence of events; but strictly the notation means that the secondary articulation is

simultaneous with the consonant. This is unlike the case of the aspiration diacritic (e.g.[th)) where the plosive and the aspiration are sequential. The simultaneity of the secondary

articulation is clearer from the alternative diacritic for symbolizing veJarization orpharyngealization, [-), which is placed through the consonant symbol in question (often to

the detriment of legibility). Nasalization, despite the similarity of name, is not a secondary

articulation in the same sense, but the addition of the resonances of the nasal cavities to asound. Vowels (e,g. [e)) and consonants (e.g. (1)) can be nasalized.

Finally, there are three diacritics in the third column dealing with release ('nasal

release', 'lateral release', and 'no audible release '). All three show that a stop consonant

has not been released into a vowel. Instead, the air escape is through the nose (e.g. [bAt"!)1button), round the side of the tongue (e.g. [bot't] bottle), or the air is not released until alater sound [1 reg'breg) ragbag . The use of diacritics is further exemplified in section 3.

2.9 Other symbolsThese symbols are included in their own section of the Chart for presentationalconvenience. The section contains several consonant symbols which would not fit easily

into the 'place and manner' grid of the main consonant table. In some cases, such as the

epiglottals and the aJveoJo-paJatals, no column is provided for the place of articulation

because of its rarity and the small number of types of sounds which are found there. Inother cases, such as [w), the sound involves two places of articulation simultaneously,

which makes it inconvenient to display in the table. If separate columns for all consonantswith two places of articulation were provided, the size of the grid would become

unmanageable. Most consonants that involve two simultaneous places of articulation arewritten by combining two symbols with the 'tie bar' [~), for example [kpl whichrepresents a voiceless labial-velar plosive.


3 Guide to IPA notation

3.1 Exemplification of the symbols

The general value of the symbols in the chart is listed below. In each case a symbol can beregarded as a shorthand equivalent to a ph01etic description, and a way of representing thecontrasting sounds that occur in a language. Thus [m] is equivalent to 'voiced bilabialnasal', and is also a way of representing on,e of the contrasting nasal sounds that occur inEnglish and other languages.

When a symbol is said to be suitable for the representation of sounds in two languages,it does not necessarily mean that the sounds in the two languages are identical. Thus [p] isshown as being suitable for the transcriptioh of pea in English, and also for pis in French;

similarly [b] is shown as being suitable for fhe transcription of bee in English, and also forbis in French; but the corresponding soun~s are not the same in the two languages. TheIPA has resources for denoting the differences, if it is necessary to do so, as illustratedbelow in section 4; but at a more general level of description the symbols can be used torepresent the sounds in either language.

All languages exhibit variation in their pronunciation. Sometimes an example belowwill only be valid for some varieties of a language. For instance the exemplification of [8]by the English word thief is not valid for dialects which pronounce the dh> as a labiodentalfricative [f] or a dental stop [1]. An examplb means that the symbol exemplified is, at least,

appropriate for one or more widely spoken varieties of the language. In the case of Englishexamples, where the variety is not further specified, it should be assumed that theexemplification is appropriate, at least, in General American English and Standard

Southern British English (see section 2).The symbols are exemplified in the or?er in which they appear on the chart; and they

are discussed using the terms given as hea~ings for the rows and columns. English (Eng.)and French (Fr.) examples are given when unambiguous. Where practical, an orthographic version of the exemplifying word is provided, in italics. English glosses of wordsin other languages are giv.en in quotation marks. The languages used for exemplificationare identified at the end of the list at the end of this section.

I

I

I

I

i

I

PLOSIVES

p Eng. pea [pi]; Fr. pis [pi] 'worst' bt Eng. tea [til; Fr. the [tel 'tea' d

Hindi [tal] 'postpone' etc Hungarian tyuk [cu:k] 'hen' Jk Eng. cap [krep]; Fr. quand [ko] g

'when'; K'ekchi [kala] 'grindstone'

q K'ekchi [qa] 'our' G? Hawaiian Hawai'i [hawai?i] '(place

name)', ha 'a [ha?a] 'dance'

Eng. bee fbi]; Fr. his [bis] 'encore'Eng. deep [dip]; Fr. diJ: [dis] 'ten'

Hindi fetal] 'branch'Hungarian gyur (ju:r] 'to knead'

Eng. gap [grep];Fr. gant [go] 'glove'

Farsi [Gar] 'cave'

Introduction 10 the IPA 19

The plosives in the left-hand column above are said to be voiceless, and those on the

right are said to be voiced. The extent of voicing may vary considerably. The voicelessconsonants may be not only voiceless, but also aspirated; and the voiced consonants maybe voiced throughout their duration, or may have voicing during only part of that time.

Usually the use of a pair of symbols such as [p] and [b] in a given language signifies onlythat there is a contrast in the degree of voicing within that pair of sounds. Either of the

variant letter shapes [g] and [g] may be used to represent the voiced velar plosive.

NASALS

m Eng. me [mil; Fr. mis [mil 'put'

IT] Eng. emphasis [£IT]f:lSlS]

n Eng. knee [nil; Fr. nid [nil 'nest'

11. Malayalam [kel1.l1.i] 'link in a chain']1 Fr. agneau [a]1o] 'Iamb'; Malayalam [kB]1]1i] 'boiled rice and water'

I) Eng. hang [hreI)]N Inuit [saaNNi] 'his bones'

Note that the symbols [t, d, n] listed above, and the symbols [I', f, I, ~, J, I] which willbe exemplified below, all represent sounds that can be either dental, or alveolar, or

postalveolar. If there is a need to represent specifically one of these places of articulation,

there are IPA resources for doing so, which will be exemplified later.

TRILLS

B Kele [mbBuen] 'fruit'

Spanish perro [pero] 'dog'; Finnish ranta [ranto] 'shore'

R Fr. rat [Ra] 'rat'; Southern Swedish ras [Ras] 'breed'

Note: most forms of English, French, German, Swedish do not have trills ex.cept inover-articulated speech, for instance when trying to be clear over a poor telephone line.

TAPS OR FLAPS

Spanish pero [pew] 'but'; Am. Eng. atom ('ref:lm]Hausa shaara Ua:ra] or [Ja:-I.a] 'sweeping'

(Some speakers of Hausa have [r] and others have [D.)

FRICATIVES

The fricatives in the left-hand column below are voiceless, and those on the right arevoiced. To a somewhat lesser degree than in the case of the plosives, the extent of voicingmay vary.

<p Ewe e fa [e <Pa] 'he polished' 13 Ewe ejJt; [e13e] 'Ewe'

20 Handbook oJthe IPA

f Eng. fee [fi]; Fr.fixe [fiks]; v Eng. vat [vret]; Fr. vie [vi) 'life';

Ewe efa [e fa) 'he was cold' Ewe ~v~ [eve) 'two'

8 Eng. thief [8it] 5 Eng. thee [oils Eng. see lsi]; Fr. si lsi] 'if' z Eng. zeal [zil]; Fr. zero [zelS'o] 'zero'

J Eng. she Ui); Fr. chic Uik] 'chic' 3 Eng. vision [v13n]; Fr.joue [3u) 'cheek'

~ Standard Chinese sha [~a) 'to kill' ~ Standard Chinese rang [~al)] 'to assist'

y German ich [Iy] T j Eng. variant of UJ in yeast [jist]

x German hoch [hox] 'high' V Greek ya;ta ['vala] 'milk'

X Hebrew [maxar] 'he sold' IS' Fr. riz [Il'i] 'rice'

h Hebrew [hor] 'hole' 'I Hebrew ['lor) 'skin'

Although it is traditional to pair Hebrew and Arabic [h], ['I] as fricatives, the voicedsound ['I] is usually perceived as an approximant.

h Eng. he [hi] Ii Eng. ahead [~lied]

[Ii] represents a breathy voiced sound, rather than an ordinary voiced sound.

LATERAL FRICATIVES

t Zulu hlanza [Hl.nza) 'vomit';Welsh lIan [tan] 'church'

!I Zulu dlala lliala] 'play'

APPROXIMANTS

lJ Hindi [n~lJB] 'ninth'

Eng. read [lid] I{ Hausa shaara Ua:ra] or Ua:{a] 'sweFping'

(Some speakers of Hausa have [r] and others have [{].)j Eng. yes [jes]; Fr. yeux [j¢] 'eyes'Uj Turkish aga [aUja] '(a title)'; Korear [Ujisa) 'doctor'

LATERAL APPROXIMANTS

I Eng. leaf[lif]; Fr. lit [Ii] 'bed'l Tamil [vaU 'sword'; Swedish parla ![pre:la] 'pearl'i. Italianfiglio [fiU::>] 'son'; Spanish llegar [i.e'var] 'to arrive'L Mid-Waghi aglagle [aLaLe] 'dizzy'

NON-PULMONIC CONSONANTS

CLICKS

o 'X66 [1<066] 'dream'I Xhosa ukucola [ukuk!ola] 'to grind finely'

Xhosa ukuqoba [uku005a] 'to break stones'


:j: 'X66 [k9=aa] 'bone'II Xhosa ukuxhoba [ukuki[h05a] 'to arm oneself'

VOICED IMPLOSIVES

5 Sindhi [5dlli] 'field'cf Sindhi [cfmu] 'festival'

f Sindhi [fdtU] 'illiterate'9 Sindhi [gdnu] 'handle'd' Mam [d'a] 'fire'

EJECTIVES

p' Amharic [p'ap'as] 'bishop (loan word)'t' Amharic [t'il] 'fight'k' Amharic [k'alat] 'word'

s' Amharic [s'ahaI] 'sun'

VOWELS

The symbols on the vowel chart can be regarded as providing reference points in thevowel space. They can also be used to represent vowel qualities generally in the area ofthe corresponding reference points. With the vowel symbols it is especially important tonote that they may represent slightly different sounds in different languages. For example,[i] may be used for the vowel in the English word heed or in the French word lit (bed),despite the fact that the English vowel may be slightly diphthongal and less close than theFrench vowel.

Because of their status as reference points, it is difficult to illustrate some of the vowelsymbols appropriately in terms of particular languages; this is particularly true of the midcentral vowels [5, e, 3, 0]. The symbols [d] and [Il] are available for representing vowels

in the mid central and lower central regions. [a] is often used for an open central vowel.The open front rounded reference quality [CE] is rarely found in languages, though reportedfor Austrian German.

The symbols in the right-hand column below specify vowels with more rounded lipsthan the corresponding symbols in the left-hand column.

Eng. heed [hid]; Fr.lit [Ii] 'bed' y Fr. lu fly] 'read';

German Fufie [fysd] 'feet'Eng. hid [hid] v German FLUsse [flvsd] 'rivers';

Swedish nytta [nvtta] 'use (noun)'e Scottish Eng. hay [he]; '/J Fr. peu [P'/J] 'few'

Fr. les [Ie] 'the (pl.)'e Eng. head [hed]; Fr. lait [Ie] 'milk' re Fr. peur [prer;] 'fear'oe Eng. had [hoed]a Fr. patte [pat] 'paw' CE Austrian German Seil [SCE:] 'rope'


a Eng.jather [f003(1») 0

A Vietnamese [AI]) 'favour' (This symbol :>is sometimes used for a different vowel,the central vowel in Eng. hut [hEt).)

y Vietnamese [ty] 'silk' 0

lU Vietnamese [tlU] 'fourth'

Korean [gim] 'gold'

u

u

British Eng. bother [bo/Y.})

British Eng. caught [btl;Gel. GOll [g:>t) 'god';Vietnamese [t:>] 'large'Fr. lot (10) 'share';Vietnamese [to) 'soup bowl'Eng. book [buk]Eng. school [skut);Fr. loup [lu] 'wolf';Vietnamese [tu) 'to drink'

Norwegian bUll root] 'blunt'

OTHER SYMBOLS

M Scottish Eng. whether [Mdhl)w Eng. weather [W£Og(l»); Fr. oui [wi] 'yes'

q Fr. huit [qit] 'eight'H Avar [maH] 'odour''t Avar [ma't] 'nail'-; Agul [jai'ar) 'centres'

/; Polish Rasia [baca) 'Barbara (dim.)'j KiChaka [iJaa) 'to dress oneself'

o Some dialects of Swedish schal [oal) 'scarf' (Note: for some speakers there is littleor no [f] friction in this sound.)

Affricates and double articulationskp, (f etc. Eng. chief [(fif]; Yoruba apa [akpa) 'arm'; Tswana tsetse I!§e!.§e) 'tsetse fly'

Note: the tie bar can be placed above or below the symbols to be linked.

SUPRASEGMENTALS

In general, only one or two degrees of stress are marked:

may be used to indicate extra strong stress.Eng. phonetics [fg'netlks)Eng. phonetician [ ,fOUng'tIJgn]

Length may be contrastive for vowels andlar consonants:Finnish matlo [mot::» 'carpet'; maalOn [ma:t:>n) 'landless'; maatlo [mo:t::»

'electrical earth/ground'Finnish mato [mat::>] 'worm'

Note: as in Finnish orthography, length can also be indicated in phonetic transcriptionby double letters: e.g. Finnish maallo [moalt::>] 'electrical earth/ground'.


Estonian has a three-way length or quantity contrast:Estonian saada [sa:da] 'to get'Estonian saada [sa'da] 'send (imperative)'Estonian sada [sada] 'hundred'

Length is not contrastive (at least, without concomitant changes in quality) in English,but allophonic differences exemplify the use of the length diacritics:

Eng. bead [bi:d]Eng beal [bi'!]

Eng. police [p~li's]

White spaces can be used to indicate word boundaries. Syllable breaks can be markedwhen required. The other two boundary symbols are used to mark the domain of largerprosodic units. There is also a linking symbol that can be used for explicitly indicating thelack of a boundary.

Eng. lamb prepared ['!a:m.pl;l.'pe;ld], lamp repaired [Ia:mp.l;l.'pe;ld]

Eng. Jack, preparing the way, wenl on [d3a:k I pl;l'pe;l.III) O;l 'weI I went 'on" ]Fr. Jacques, preparant le sol, tomba [3ak I pl!epal!Q 1;l s:Jl I t5ba II ]'Jack, preparing the soil, fell down'

Fr. petit ami [p;ltigJ.mi] 'boyfriend'

As explained in the previous section, there are two alternative systems of tonetranscription. The chart shows these two systems as if there were direct equivalenciesbetween them. However, they are usually used in different ways.

Bariba [ne na na ko] 'I am the one who came'Yoruba 0 bci [6 ba] 'he/she met'Yoruba 0 ba [6 ba] 'he/she hid'Yoruba 0 biz [6 ba] 'it perched'Trique [ell 'bitter'

It is also possible to combine these symbols so that, for example, [e] represents a hightone followed by a low tone on the vowel [e], i.e. a falling tone. Similarly [e] represents arising tone, and [e] and [e] represent high-rising and low-rising tones.

There are two symbols for showing that subsequent tones may be a step lower orhigher. The introduction of a downstep is phonologically contrastive in the Igbo examplebelow, but the Hausa upstep indicates only a predictable allophone.

Igbo ¥lQ any,i [6.11:5 .lap! ] 'our house'Hausa [turanTt[i ne] 'it is English'


The use of the other set of symbols is illustrated below.

CHINESE (STANDARD)

l [mal] 'mother'1

-I

~

J

'l [ma'l] 'scold''I

,.I

1 [ma1] 'hemp'-A [ma-A] 'horse''I

-.I

CANTONESE

[~ikl] 'to,know'

[~i-I] 'to try'; [~it-l] 'to reveal'[~i~] 'matter'; [~ik-l] 'to eat'

[~iJ] 'tim~'

[~i'l] 'poem'[~i,.l] 'city'[~i1] 'to cause'

THAI

[na:'l] 'face'

[na:1] 'aunt/uncle'[na:-A) 'thick'[na:'1) 'paddy field'[na:-.I) '(nickname)'

The symbols for global rise and global fall are appropriate for use in many languagesto mark intonation.

/' Eng. No? I/' nou)

" Eng. No. b nou], How did you ever escape? [/'hau dId ju £v~r I\oskeIp)

DIACRITICS

The diacritics allow symbols to be created to represent many additional types of sounds. Inthe representation of many languages (i;ncluding English) diacritics are necessary onlywhen making detailed transcriptions.

Burmese [Qa) 'nose'The voiceless diacritic can also be used to show that a symbol that usuallyrepresents a voiced sound in a particular language on some occasions represents avoiceless sound, as in a detailed transcription of conversational English Please

say ... as [pI i~ se .. .).The voiced diacritic can be used to show that a symbol that usually represents avoiceless sound in a particular laryguage on some occasions represents a voicedsound, as in a detailed transcription :of conversational English back of as [bre15 ;}v).Hindi [khan) 'mine'Detailed transcription of English pea, tea, key [phi, thi, khi)

Assamese [Nt) 'to bury'In some forms of English, e.g. Standard Southern British, over-rounded [J) is found,

e.g. caught [kyt).In many forms of English, e.g. Californian, under-rounded [u) is found, e.g. good

[gyd).

Eng. [~) in key [I$:i)Eng. ttl in tree [1Ji]


Eng. [e] in well [wet]Eng. [0] in November [novemlY.>(l)]Eng. LI] in fiddle [fId!]Spanish poeta [pQ'eta] 'poet'Am. Eng. [~] in bird [~d] This sound can also be written [!].Hindi [kumar] 'potter'Mazatec nda' [nd~] 'buttocks'

Tangoa [!ese] 'butterfly'W Eng. [t] in twin [tWm]; Cantonese [kw:>k] '(family name)'

Russian [matj] 'mother'Y Russian (lYisij] 'bald'

Arabic [s'ad] (letter name)Eng. [I] in hill [hIt]Some forms of South African Eng. [l] in dry [d~al]

Danish [9] in lade [1<e9d] 'barn'Igbo 6bI [¢b)] 'heart'Igbo 9b} [Vb}] 'poverty of ability'Eng. [!l in width (WI!S]Ewe e da [e ga] 'he throws'

Ewe e qa [e \la] 'he cooks'Fr.fin [fe] 'end'

n Russian [dnno] 'bottom'

Navajo [diM?] 'prairie dog'Eng. [k'] in act [<ek't]

3.2 Languages usedfor exemplificationThe principal country in which a language is spoken is given only when it is not apparentfrom the name.

Agul, Caucasian, spoken in the N.E. Caucasus.Amharic, Afro-Asiatic, spoken in Ethiopia.Arabic, Afro-Asiatic, spoken in many North African and Middle Eastern countries.Assamese, Indo-European, spoken in India.Avar, Caucasian, spoken in the N.E. Caucasus.Bariba, Niger-Congo, spoken in Nigeria.Burmese, Sino-Tibetan, spoken in Myanmar.Cantonese, Sino-Tibetan, spoken in China.Chinese (Standard), Sino-Tibetan.Danish, Indo-European.English, Indo-European.Ewe, Niger-Congo, spoken in Ghana and Togo.Farsi, Indo-European, spoken in Iran.

26 Handbook of/he [PA

Finnish, Finno-Ugric.

French, Indo-European.German, Indo-European.Greek, Indo-European.Hausa, Afro-Asiatic, spoken in Nigeria.Hawaiian, Austronesian.Hebrew, Afro-Asiatic, spoken in Israel.Hindi, Indo-European, spoken in India.Hungarian, Finno-Ugric.Igbo, Niger-Congo, spoken in Nigeria.

Inuit, Eskimo-Aleut.Italian, Indo-European.K'ekchi, Mayan, spoken in Guatemala.Kele, Austronesian, spoken in Papua New Guinea.KiChaka, Niger-Congo, spoken in Tanzania.Korean, Altaic.

Malayalam, Dravidian, spoken in India.Mam, Mayan, spoken in Guatemala.

Mazatec, Oto-Manguean, spoken in Mexico.Mid-Waghi, Papuan, spoken in Papua New Guinea.Navajo, Na-Dene, spoken in United States.Norwegian, Indo-European.Polish, Indo-European.

Russian, Indo-European.Sindhi, Indo-European, spoken in Pakistan.Spanish, Indo-European.Swedish, Indo-European.Tamil, Dravidian, spoken in India.Tangoa, Austronesian, spoken in Vanuatu.Thai, Tai-Kadai.Trique, Oto-Manguean, spoken in Mexico.Tswana, Niger-Congo, spoken in Botswan.a.Vietnamese, Austro-Asiatic.Welsh, Indo-European.

Xhosa, Niger-Congo, spoken in South Africa.!X66, Khoisan, spoken in Botswana.Yoruba, Niger-Congo, spoken in Nigeria.Zulu, Niger-Congo, spoken in South Africa.


4 The phonemic principleFrom its earliest days (see appendix 4) the International Phonetic Association has aimed toprovide 'a separate sign for each distinctive sound; that is, for each sound which, beingused instead of another, in the same language, can change the meaning of a word'. Thisnotion of a 'distinctive sound' is what became widely known in the twentieth century asthe phoneme. Its history is far longer, though. For instance, the phonemic principle isimplicit in the invention of alphabetic writing. However a lot of languages, such asEnglish, have spelling systems in which the relation between phonemes and letters of thealphabet has become obscured. This very fact was a motivation for the creation of a

universally agreed system of phonetic notation. So, in English, the IPA provides a symbolIkJ which stands unambiguously for the phoneme which is variously written as <c> (car),<k> (kettle), <ck> (back), <ch> (monarch), <q> (quick), and in other ways.

Each language can be analyzed as having an inventory of phonemes. This inventorymay range in size from around a dozen phonemes to nearer a hundred depending on thelanguage. Conventionally, as in the English example above, symbols for the phonemes ofa language are placed within oblique lines: I I.

In general, the symbol for a phoneme will be an unmodified letter of the IPA, butletters may also be combined to make a phoneme symbol (for instance ItJ/, as at thebeginning and end of English church; if necessary the phonological unity of the two

segments can be shown by a tie bar: It} I). Diacritics may also be employed to createsymbols for phonemes, thus reducing the need to create new letter shapes. This may beconvenient in particular when a subset of the phonemic system of a language shares aphonetic property, as in the case of the nasalized vowel phonemes of French If'. ce a 51,

which when they stand alone represent French hein 'huh', un 'a, one', an 'year', and on'one (impersonal pronoun)'.

The use of the phrase 'distinctive sound' above implies that there are other soundswhich do not change the identity of a word, sounds which are not 'distinctive' in thistechnical sense. Central to the notion of the phoneme is the recognition that many finelydistinct sounds can be phonetically identified which do not have the word-distinguishingrole of, say, English Ikl and It! (as in Ikil key vs. Itil tea). For instance, the English IkJphoneme is made with a tongue closure further forward in the mouth before a front vowel(such as the iii of key) than before a back vowel (such as the I'JI of caw). But crucially it isnot possible, in English, to exchange these two varieties of IkJ to make two new words, sothe two varieties of IkJ are not 'distinctive' in English.

A phoneme can be regarded as an element in an abstract linguistic system, an elementwhich has to be realized in the physical world by an acoustic signal produced by vocalactivity. Variation arises in the process of realization. Some of this variation can beattributed to the influence of adjacent sounds affecting the articulation, so for instance theIkl of key may be thought of as being further forward to facilitate integration with thefollowing iii, while in other cases the variation seems to be merely a language-specific butphonetically unmotivated habit. Variant realizations of a phoneme are known as itsallophones.

28 Handbook of the {PA

The IPA aims not only to provide symbols which can unambiguously represent

phonemic inventories, but also to be able!to represent details of phonetic realization. Theabove example could be represented as [l$:hi:] key and [k h:):] caw, where the 'Subscript

plus' and 'Under-bar' indicate advanced and retracted articulation respectively (seeappendix 2 for diacritic names). A further detail of realization is also indicated here - the

'Superscript H' indicates aspiration, a dflay in the onset of voicing after the voicelessplosive, characteristic of such plosives ilt the beginning of stressed syllables in many

varieties of English. Square brackets are ?sed conventionally to make clear that a symbolor sequence of symbols represents phonetic realizations rather than phonemes.

In providing the means to show the d~tail of phonetic realization in a given language,

the IPA also achieves the delicacy of notation needed to compare the phonetic detail of

different languages. For instance, although a phonemic representation Itrul might be

suitable for the English word true or the French word trau, the difference in pronunciationof the two words is reflected in phonetically more detailed representations such as l!;H](true) and [lliU] (frau). These show the dental realization of It! in French compared to the

alveolar realization in English, here retracted under the influence of the following

postalveolar; the uvular realization of Irl in French compared to the postalveolar

realization in English, both realizations devoiced after the voiceless plosive; and the fullyback realization of lui in French compared to the central realization in (many varieties of)

English.

5 Broad and narrow transcriptionsA connected text represented in terms of phonemes is known as a 'phonemictranscription', or, almost equivalently, a 'qroad transcription'. The term 'broad' sometimes

carries the extra implication that, as far as possible, unmodified letters of the roman

alphabet have been used. This restriction may facilitate printing, and might be considered

particularly if a phonemic transcription is to form the basis of a writing system. Under this

definition a transcription of English hide6ut as Ihaidautl would be broad, while /haldaut/would not be because it introduces letter shapes to the symbol for the phoneme laii and the

phoneme laul which are not absolutely necessary for the unambiguous representation of

the phonemes of English, but which may be desirable to remind the reader of the phoneticrealization of these phonemes. Frequentl,y, though, 'broad' is used merely as a way of

referring to transcriptions which are phonemic, regardless of the letter shapes used torepresent the phonemes. Phonemic transcriptions are one type of 'systematic'

transcription, meaning they require the phonological patterns or 'system' of a language to

be known before they can be made.The term narrow transcription most commonly implies a transcription which contains

details of the realization of phonemes. There are two ways in which such a transcriptionmay come about. If a transcription is made in circumstances where nothing can beassumed about the phonological system, it is necessary to include all phonetic details

because it is not clear which phonetic properties will turn out to be important. The

transcription would be made taking into account only the phonetic properties of the


speech. This type of narrow transcription, as might be made in the first stages offieldwork, or when transcribing disordered speech, is sometimes called an impressionistictranscription or a general phonetic transcription. If an impressionistic transcription weremade of an utterance of the English phrase check the lens well it might be[tJ~?k~dl£nzw~tj. This includes a glottalized velar stop, a dental approximant (thelowering diacritic indicating that the stricture was not close enough to cause frication), avelarized or pharyngealized lateral (probably involving, for many English speakers, asecondary articulation of 'uvularization' intermediate between velarized [IYj andpharyngealized [l~]), and three different vowel qualities in the stressed syllables, eventhough these vowels are the same in phonemic terms.

The other kind of narrow transcription containing realizational information is termedallophonic. If the relevant phonological system is known, a transcription can be devisedwhich includes any number of additional symbols to indicate the phonetic realizations ofthe phonemes, i.e. their allophones. An allophonic transcription is also known as asystematic narrow transcription. In the knowledge that a possible phonemic analysis ofcheck the lens well is ItJekodlenzwelJ, one allophonic or systematic narrow transcriptionwould, perhaps surprisingly, be [tJ~?k~dl£nzw~tj, that is, one which is identical to theimpressionistic transcription in the previous paragraph, incorporating all the phoneticdetail which can be heard. The difference is that now it would be possible to express, inconventions, the relation between the allophones transcribed and the phonemes which theyrealize. Alternatively, it is possible (and customary) to be selective about the informationwhich is explicitly incorporated into the allophonic transcription. The choice might bemade, for instance, to leave out the information about vowel height (the closer vowel incheck is triggered by the high tongue body position of the following velar, and the moreopen and retracted quality in well caused mainly by the secondary articulation of thefollowing lateral), and about vowel nasalization (which is very general before a followingnasal), giving a transcription which focuses on consonant realization: [tJe?k~dlenzwet].

Minimally, if the focus of interest were glottalization of plosives, the allophonictranscription could be [tJe?kodlenzwelj, or if the focus were the 'dark' lateral,[tJekodlenzwdj. These last two transcriptions look superficially very like a phonemictranscription, but they are in principle different because information has been included(albeit sparingly) which is not required for the unambiguous representation of the words.Narrowness is regarded as a continuum, so that [tJekodlenzwdj might be regarded as aslightly narrow (or 'narrowed') transcription, and [tJ~?k~dl£nzw~t) as very narrow. (In allthese transcriptions, no spaces between words have been included. This is inevitable in animpressionistic transcription where it is not yet known how the utterance divides intowords. In phonemic and allophonic transcriptions it is common to include spaces to aidlegibility, but their theoretical validity is problematic.)

Any transcription is connected to a speech event by a set of conventions. In the case ofan impressionistic ('general phonetic') transcription, the conventions are precisely thoselying behind the IPA Chart, indicating for instance that the phonetic value of [?k] is asimultaneous velar and glottal closure. In the case of a phonemic transcription, the

30 Handbook of the lPA

conventions also include the 'phonological rules' of the particular language which

determine the realization of its phonerrles, such as the fact that for some varieties ofEnglish the lateral phoneme III is realiz~d with an accompanying secondary articulation(rtD when not followed directly by a ~owel or Ijl in the same word. Likewise, therealizational information which is not eXBlicit in a particular allophonic transcription is, inprinciple, provided by conventions.

6 IPA transcriptions for a languageThere can be many systems of phonemic!transcription for the same variety of a language,all of which conform fully to the principles of the IPA. Sometimes the differences betweenthe systems result from the fact that mo~e than one phonetic symbol may be appropriatefor a phoneme. For instance, the vowel phoneme of get in Standard Southern British

English has allophones, according to phonetic environment, which mostly lie between thecardinal vowels [e] and [e], some realizations being closer to one and some to the other. Itis therefore permissible to choose either symbol as the one to represent the phoneme.

In other cases the differences between competing transcriptions result from alternativeways of representing the phonological contrast between sounds. In English, for example,the contrast between the words bead and bid has phonetic correlates in both vowel qualityand vowel duration. A phonemic represeptation which explicitly notes this might use thesymbols li:1 and III, where the difference in letter shape reflects the difference in vowelquality, and the length mark on the first letter reflects the difference in duration. But it isequally possible unambiguously to represent these phonemes as Ii:! and iii (where thephonemic symbol only explicitly shows the length difference), or as iii and II! (where only

quality is shown explicitly). All three pairs of symbols are in accord with the principles ofthe IPA (as long as the principle chosen for this pair of vowels is applied consistentlythroughout the vowels of the language).

Other differences may stem not from 'alternative representations of what is essentiallythe same phonemic analysis, as above, but from alternative phonemic analyses. Forinstance, English long vowels and diphthongs are often analyzed as unitary phonemessuch as Ii:! (as in heed) and laol (as in how). In this view, the fact that the phoneme symbolis made up of two phonetic symbols or a symbol and diacritic does not affect the status ofa sound as a single phoneme in the analysis. Alternatively it is possible to analyze them as

combinations of a short vowel and an approximant: iii + Ijl and lal + Iwl, or even (in thecase of long vowels) as combinations of short vowel phonemes (represented iii, lui etc.)and a 'chroneme' 1:/: iii + 1:1 and lui + 1:/. It may not be possible to infer the particularanalysis being used from the phonemic transcription. However the point here is that therepresentation resulting from any of these analyses is in keeping with the principles behindthe IPA. The IPA does not provide a phonological analysis for a particular language, letalone a single 'correct' transcription, but rather the resources to express any analysis so

that it is widely understood.

Introduction to the IPA 3\

7 Working with the IPAThere are a number of practical issues that may arise when using the IPA. Some of theseinvolve problems of how to refer to symbols. In what follows, reference will be made toappendix 2. This contains a comprehensive listing of symbols used in phonetics, includingthose of the IPA, but also many which are not recommended by the International PhoneticAssociation but which may be encountered. The listing indicates which symbols are notIPA usage, or which were once recommended but are no longer recommended. The listingwas produced by the IPA Workgroup on Computer Coding, set up at the 1989International Phonetic Association Convention in Kiel.

7.1 SymbollUlmesIt is often useful to be able to refer to symbols by an agreed name. If it is a question ofreplacing [B] by [oj, it is easier to say 'not "Turned A" but "Turned script Am than toattempt a verbal description of the relevant symbols. Although the International PhoneticAssociation has never officially approved a set of names, many symbols have informallyone or more names, and a greater degree of consensus has arisen as the result of the use ofnames in Pullum and Ladusaw's Phonetic Symbol Guide (2nd edition, 1996. University ofChicago Press). Appendix 2 therefore includes with each symbol a convenient and

systematic name, most of which are those used by Pullum and Ladusaw.

7.2 Using the 1PA in handwritingThere are cursive forms of IPA symbols, but it is doubtful if these are much in use today.They may have been of greater use when transcription by hand was the only way ofrecording speech, and so speed was essential. The cursive forms are harder for mostpeople to decipher, and it is preferable to use handwritten versions which closely copy theprinted form of the symbols.

7.3 Using the 1PA in printPrinters should normally have a font including IPA symbols. Even if they do, however,

there will be a danger of superficially similar symbols being mixed up (for instance [aJwith [aJ or [yJ with [YJ). Some publishers have tables in which unusual symbols can beidentified by index numbers and letters, but practice is variable. It should therefore benoted that the Chart in appendix 2 provides for each symbol a unique identifying number,its 'IPA Number'. It may therefore be helpful to identify symbols which might provedifficult by that number, and to supply to printers and publishers a copy of the table.

7.4 Using the lPA on computersCharacter sets including most or all of the IPA are available for several computing

environments. Most straightforwardly a number of commercial and free fonts are availablefor Macintosh and Microsoft Windows. The situation in other computing environmentsmay be less straightforward. One problem for those devising IPA character sets which hashindered the interchangeability of data containing phonetic symbols was the lack of an


agreed standard coding for the symbols. The International Phonetic Association, throughits Workgroup on Computer Coding, ~as worked with the International StandardsOrganization in its project to set up a universal character set (UCS) for all alphabets. An

I

agreed set of UCS 16-bit codes is included in the list in appendix 2.

7.5 The IPA and brailleOver the years, a braille version of the I~ternational Phonetic Alphabet has evolved. Abook by W. Percy Merrick and W. Potthoff, A Braille Notation of the InternationalPhonetic Alphabet with Keywords and Sp~cimen Texts, published in London by the RoyalNational Institute for the Blind in 1932 I-Vas reviewed in the Association's journal, LeMaitre Phonetique, in 1936 by E. E. Quick (p. 51). This book is archived in the Library ofCongress in the US and elsewhere, and has served as the basis for the development of abraille standard for the rendering of phonetics. The code book used for many years by theCanadian National Institute for the Blind and by many other institutions was a 1977volume entitled the Code of Braille Textbook Formats and Techniques, published by theAmerican Printing House for the Blind (1839 Frankfort Avenue, P.O. Box 6085,Louisville, Kentucky 40206-0085, USA. Tel. 1-800-223-1839). Rule XIX, section 45, ofthis book refers to phonetic notation, 'principles, and the phonetic alphabet, withindications of braille equivalents and illus~rations of braille usage in print (dot) form. Thebook of codings was available at US$50 'in print and US$235 in braille in 1996. A newedition of the book, Braille Formats: Principles of Print to Braille Transcription 1997,was released by the American Printing House for the Blind in September 1998. It is a 300page volume with print and braille editions, each costing US$30. The phonetic symbolsrepresented in Braille Formats: Principles of Print to Braille Transcription 1997 are forthe most part still valid and in use in the IPA. Some may no longer be in use; and somenewer symbols added to the IPA more recently may not appear.

8 Going beyond the IPAAs noted in section 2, the descriptive resources of the IPA were developed principally forthe linguistically relevant aspects of speech. This was because the whole tradition ofphonetic description was concerned with the properties which realize the phonologicalsystems of languages. A phonological system can be seen as the conventions whichspeakers of a language share about its sounds. Many aspects of individual utterances suchas personal voice quality, emotive modifications of speech, accidental mis-articulations,dysfluencies, and speech pathologies are not relevant to the phonological system, and sophoneticians have tended to ignore such aspects when working on the phonology of alanguage. The IPA reflects this orientation, being, in essence, a system for describing thelinguistic-phonetic properties of error-free utterances not specific to a particularindividual. There are, however, many circumstances in which it is essential to be able to

transcribe other properties of speech.One important set of such properties constitutes a conventionalized system of

communication beyond the verbal component of speech, and which is often referred to as


paralanguage. This includes the use of phenomena such as voice quality, pitch range, andrate of utterance variously to convey aspects of the speaker's emotional state and attitudeto other conversational participants, to indicate the status (e.g. confidentiality) of theinformation being conveyed, and to regulate the course of a conversation by encouragingor discouraging others from speaking. Researchers involved in the analysis of spokeninteraction, for instance, clearly need resources for the description of such speechphenomena.

In other situations the phonetic properties of interest may be ones which realize thephonological system, but specifically in speakers who for one reason or another do notachieve normal realizations of the system. Most obviously, clinical practice and researchin the field of speech pathology require a system of phonetic notation which will copewith sounds and combinations of sounds which lie outside the usual range. Research, too,on children's utterances during language acquisition, which contain many sounds that donot occur systematically in the languages of the world, also requires notational devicesoutside those provided by the IPA.

Researchers in these fields have, of course, developed their own notational devices asrequired. But clearly it would be preferable to have a widely agreed standard set ofconventions for these additional applications, comparable to the standard provided by theIPA. With this aim the International Phonetic Association's Clinical Phonetics andLinguistics Group has proposed a set of 'Extensions to the IPA' for transcribing nonlinguistic speech events, and other aspects of speech such as deviant or pathologicalspeech. These are listed and explained in appendix 3.

9 Some problematic issues

9.1 SegmentationIn making an impressionistic transcription of a language whose phonological system is notknown, uncertainties over the division of an utterance into segments may arise. Somearticulatory sequences produce a speech signal which different languages may interpret asmade up of a different number of segments. This is sometimes the case, for instance, whensecondary articulations are added to primary articulations. An articulatory sequence suchas would be represented by [b] is relatively unproblematic. Here the syllable begins witha lateral, which involves a closure by the tongue tip or blade against the alveolar ridge,with the tongue body left free to anticipate the position required for the following vowel.The acoustic signal, as shown in the left half of the spectrogram in figure 5, clearly fallsinto two distinct parts corresponding to the lateral and the vowel. But if the tongue body ishigh during the lateral, there will be an [i]-like transition or palatal approximant betweenthe consonant and the vowel- see the right half of figure 5. Given only the phonetic event,it is not clear whether to transcribe this sequence as three segments ([Ijo]) or two segments([ljo], where the superscript 'j' indicates a modification of the lateral by palatalization, andnot an additional segment).

There may be some evidence in the phonetic signal to help resolve the issue. For

34 Handbook of/he IPA

instance, if there is a noticeable [i)-like transition into the lateral from a preceding vowel

as well as into the following vowel, it is more likely that the high tongue position isI

associated with the lateral. But ultimately;the answer will lie in the phonological patternsof the language. If the language has seql\ences like [jo), where the palatal approximantappears independently of another consonant articulation, this points in the direction oftreating it as a separate segment elsewhere (e.g. [ljo)). Pointing in the other directionwould be the discovery that the language contrasted consonants extensively in terms ofsecondary articulation, even where an approximant would be unlikely to occur, forinstance after a word-final voiceless fricative. Sequences such as [asj] and [asw] are notlikely, since approximants normally occur adjacent to vowel nuclei, and such a contrastwould normally be attributed to the consonant and phonemicized as /asi / and /asw/,

indicating contrastive secondary articulations of palatalization and labial-velarization.With this knowledge, then, [Ii;)] would be seen as the segmentation more appropriate tothe phonological patterning of the language.

time (s)1.00.80.40.2

1

8'

7 ~ + + +i

6 + + +---.

ALiiN 11,1 + +:r:: 5c-

itlll.~~1 +>.

4uc<) I ',." ,I I:::l<:T 3 1'1 H- +<)

<t:

'Iqf~'" '" '''\~ll 1+2

111111 -I +

Figure 5 Spectrogram of [10] and an utterance whose segmentation is ambiguously

[li]+[;)] or [l]+[j]+[;)).

The point of considering such examples is to underline the fact that the segmentationwhich phonetic description requires is not always transparently available in the phoneticevent, and impressionistic transcriptions may have to contain unresolved ambiguities untilsufficient is known about the structure of the particular language. Moreover, suchuncertainties of segmentation will often form the basis of alternative proposals for

phonemic interpretation,


9.2 Aligning transcriptions and speechEven where the structure of the language is known, the alignment of a phonetictranscription with records of the physical speech event will sometimes be problematic.This is because the effects of a particular segment overlap with those of others, or, to put itanother way, because the changes in the various parameters which make up speech

(voicing, nasality, and so on) are neither instantaneous nor aligned simultaneously. Themore closely the physical speech event is observed, the greater the tension between the

segmentation derivable from the phonological structure of a language and that suggestedby the structure of the physical signal.

The English word sleeting, for instance, is phonemically /slitJI)/, a sequence of sixphonemes. The acoustic signal of this word as spoken by a speaker of a variety of Englishwithout voicing of word-medial /U is represented by the spectrogram in figure 6.

+ +ii' fI

8.

7 -

6

N'g 5 -

>-.<.) 4 -c::<l)::lcr< 3~

2

+0 " " I

0.1

S 1 1o

1

++

t s h j

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

+";,1,,,,1,

time (s)

1)

Figure 6 Spectrogram of the word sleeting, illustrating the complex relationship betweenacoustic patterns and phonemic segmentation.

This shows considerably more than six identifiably different successive aspects. Apossible acoustic segmentation is indicated by the vertical lines drawn below the

spectrogram. A narrow transcription can to an extent capture this acoustic segmentation,and might be suitable for some applications, for instance in speech technology research,where a close annotation of the acoustic signal is required. The representation [sllitSITjI)]given below the spectrogram reflects the fact that the voicelessness of the [5] persists intothe first part of the lateral articulation, so there is no single acoustic phase correspondingto the 'voiced lateral approximant' phoneme; that the [t] is released first with a phase of


affrication ([s]-like friction locatable to the alveolar region) and then aspiration (nonlocalizable [h]-like friction); and that the Ilasality of the final nasal stop is anticipated inthe preceding vowel. In the case of the nas:.tlity, the narrow transcription captures the wayin which cues to a particular phoneme are distributed beyond what would normally beconsidered the boundaries of the sound; t~e nasality on the vowel is an early indicationthat a nasal consonant is imminent. But other instances of this distribution of cues to asound cannot be captured in an IPA transcription; for example, the changing acousticpattern corresponding to [i] is caused by the movement of the tongue body, during thevowel, from the alveolar [t] towards the velar closure of [I)]. This changing pattern is animportant cue to the place of articulation of both consonants.

Problems of segmentation and alignm~nt provide a challenge to one of the theoreticalassumptions behind the IPA mentioned at" the start of section 2, namely that 'speech canbe represented partly as a sequence of discrete sounds or "segments'''. The word 'partly'acknowledges the fact that (section 2) 'in addition to the segments, a number of"suprasegmental" aspects of speech, such as stress and tone, need to be representedindependently of the segments'. But it turns out that even the 'segmental' aspects ofspeech can prove harder to allocate unambiguously to a sequence of discrete segmentsthan might be anticipated. This does not jnean that the segmental assumption should berejected. It is the foundation of phonetic ~escription, and hence of the IPA. What it doesmean is that users of the IPA should be aware that the analysis of speech in terms ofsegments does involve an analytic assumption, and that tensions between the analysis andthe data will arise from time to time.

9.3 Transcribing the speaker or the hearerThe relation between a sequence of words and its phonetic realization, far from beingunique, is highly variable. A speaker may choose to pronounce carefully, that is with ahigh degree of 'phonetic explicitness', or to take short cuts. Articulatory short cuts aresometimes known as phonetic reduction. There are tendencies, by no means absolute, formore phonetic reduction to happen the faster someone speaks and the more predictable thecontent of the speech is.

Many of the differences between explicit and reduced forms can be captured in IPAtranscription. For instance, a careful utterance of the word educated in Standard SouthernBritish English might be transcribed narrowly as [Edju~heltl(;I], and a phonetically reducedutterance as [E~<J;~Ji.e!~9], where unstressed vowels are mid-centralized, the first [d] andthe following palatal are assimilated to tbe alveolo-palatal place of articulation, and thevelar and alveolar voiceless plosives of the careful form are instead fricatives.

In other cases the transcriber is faced with theoretically problematic forms as a resultof reduction. For instance in the phrase mad cow, a careful utterance of which would be[mred'k"au], the alveolar at the end of moo is susceptible in less careful pronunciation toassimilation to the place of articulation of the following velar. Traditionally, this might betranscribed as [mreg'khau), indicating complete loss of the alveolar. Instrumental recordsof articulation, however, show that sometimes in forms where the alveolar sound cannot


be heard the speaker is nonetheless making a reduced tongue movement towards thealveolar ridge. There is then a discrepancy for such an utterance between the transcriptionwhich is 'right' for the hearer ([mreg'khau) as above) and one which would better reflectthe speaker's behaviour - perhaps [mrecfkhau), showing an incomplete articulation of thealveolar stop at the end of mad. Such a discrepancy violates an assumption implicit inphonetic description, namely that the form to be transcribed is common to speaker andhearer.

10 The IPA and phonological theoryThe IPA is intended as a commonly agreed tool for analyzing and representing thephonetic properties of any language. Often, such phonetic analysis will be done in tandemwith phonological analysis, that is, the discovery of ways in which sounds pattern in alanguage and interact with other levels of linguistic structure, particularly morphology(word-building). Views on how best to carry out phonological analysis are constantlyevolving as new theories and their associated representational devices are developed.

Although it might be thought ideal if the IPA provided a means of representingphonetic facts independent of theoretical premises, it is inevitable that any means ofrepresentation which goes beyond simple replication (as by a tape recorder) must beshaped by hypotheses about the object being analyzed. Historically, the IPA has its rootsin a tradition of phonology in which the notions of the phoneme, as a contrastive soundunit, and of allophones, as its variant phonetic realizations, are primary; and in whichutterances are seen as the concatenation of the realizations of phonemes. The use of analphabetic notation underlines the conceptualization of speech as a sequence of sounds.

That conceptualization was shown in the previous section to be sometimes at oddswith the physical speech event. It has also been departed from several times in thephonological theories of the last hundred years. Distinctive Feature Theory stressed theimportance not of the 'sound' or 'segment'. but of the phonetic properties which co-occurin different combinations in sounds. Autosegmental Phonology, and before it FirthianProsodic Analysis, broke free from the 'slicing' of speech into a single linear sequence ofphoneme-sized slots, and allowed some phonetic properties to have larger domains (suchas the syllable or word) where this seems in accord with the patterns of a language. Otherdevelopments have emphasized the importance of structures such as the mora, the syllable,the foot, and the phonological word in the organization of the phonetic properties ofutterances.

These developments in theoretical phonology have had relatively little effect on theIPA. Distinctive Feature Theory has been indirectly acknowledged in the 1989reformulation of the Principles of the IPA (see appendix I); Principle 2 now includes thefollowing:

The representation of [...) sounds uses a set of phonetic categories which describe howeach sound is made. These categories define a number of natural classes of sounds thatoperate in phonological rules and historical sound changes. The symbols of the IPA areshorthand ways of indicating certain intersections of these categories. Thus [p) is a


shorthand way of designating the intersection of the categories voiceless, bilabial, andplosive; [m] is the intersection of the cate~ories voiced, bilabial, and nasal; and so on.

But there has been no loosening of the segmental 'slicing' of a traditional phonemic view.The IPA Chart, in its fundamental conception, remains much as it has been for over acentury. Only in the case of those properties explicitly recognized as suprasegmental andin the 'Extensions to the IPA' (appendix 3) are devices provided for properties extendingover domains larger than a segment.

The conservatism inherent in the IPA tradition has advantages. Phonemic analysis isstill the most widely understood and practised form of phonological analysis, at leastoutside the ranks of theoretical phonologists, and its principles are fairly accessible to all

those familiar with alphabetic writing syst~ms. This favours a system of general phoneticI

description such as the IPA which is closely compatible with a phonemic view. Secondly,the inertia of the IPA protects it from the shorter-lived of the winds of phonologicalchange, and provides an element of continuity which is particularly important to thosewho use the IPA as a tool for practical purposes. Nonetheless, the IPA should not be

regarded as immutable, even in its fundamental assumptions, and there needs to be acontinuing reappraisal of their appropriateness.

PART 2

Illustrations of the IPA

Part 2 of the Handbook contains the twenty-nine 'Illustrations' which have appeared in

the Journal of the International Phonetic Association from 1989 to 1997. These arephonetic analyses of a language, showing how the IPA can be used in the description of its

phonological inventory, and in the transcription of a continuous text.

The Illustrations include a transcription of a spoken text, traditionally a translation ofthe fable 'The North Wind and the Sun'. Of the Illustrations presented here, only that ofTaba uses a different text. The British English text of the fable is given here for reference:

The North Wind and the Sun were disputing which was the stronger, when a travellercame along wrapped in a warm cloak. They agreed that the one who first succeeded inmaking the traveller take his cloak off should be considered stronger than the other.

Then the North Wind blew as hard as he could, but the more he blew the more closelydid the traveller fold his cloak around him; and at last the North Wind gave up the

attempt. Then the Sun shone out warmly, and immediately the traveller took off hiscloak. And so the North Wind was obliged to confess that the Sun was the stronger ofthe two.

Recordings of the words and text contained in most of the Illustrations are availableto accompany the Handbook. They may be accessed at hltp://uk.cambridge.org/Ii ngu isti cs/resou rces/ ipahand book.

American English

PETER LADEFOGED

Department of Linguistics, UCLA, 405 Hilgard Avenue,

Los Angeles, CA 90095-/543, USA

There are many different dialects of English spoken in North America, so it is somewhatimproper to refer to anyone of them simply as 'American English'. The style of speechillustrated here is that of younger educated Americans in the Far-Western and some of theMid-Western parts of the United States. The speech in the recordjng on which thetranscription is based is that of a 21-year-old speaker who has lived all her life in SouthernCalifornia. Speakers from other parts of the United States, such as the East coast and theNorthern cities of the Mid-West have different dialects, nearly all of them being moreconservative, with a greater number of vowels.

Consonants

Bilabial Labio- Dental Alveolar Post- Palatal Velar Glottaldental alveolar

Plosive p b t d k 9Affricate tJ d3

Nasal m n 1)

Fricative f V 8 0 s z J 3 h

Approximant 1 J wLateral

1Approximant

p 'pie' t 'tie' k 'kite'b 'buy' d 'die' 9 'guy'm 'my' n 'nigh' I) 'hang'f 'fie' S 'thigh' h 'high'v 'vie' (5 'thy' tJ 'chin'

s 'sigh' d3 'gin'z 4 ZQO ' J 'shy'

w 'why' J 'rye' 3 'azure''lie' j 'you'

42 HandbookoflhelPA

Vowels

Monophthongs Diphthongs

The qualities of the vowels shown in the charts are based on observations of nine speakersof the dialect being described. The unstressed vowel [a] is not shown on the chart as itsquality varies considerably. Four different forms of transcription of the vowels are givenin the list of key words. In (I), which is the style of transcription used in the firstillustrative passage that follows, the differences in quality are explicit, the length and other

, differences among vowels being regard6d as a matter of the conventions required forI

interpreting these particular symbols; in (2) the length differences are made explicit, withthe quality differences being regarded as amatter of interpretation conventions; in (3) bothlength and quality differences are shown; and in (4) diphthongs are treated as consisting ofa nucleus and an offglide. If only a single style of transcription had been given it wouldhave been necessary to make all these aspects of vowel quality clear by means ofadditional conventions for interpreting th9 symbols. All of these (and several other) stylesof transcription are properly regarded ps IPA transcriptions of Californian English,provided that they are accompanied by suitable conventions. The vowel symbols in (I) are

Isimilar to those used in the 1949 Principles, the differences being that the 1949 versionused [a] in words such as 'bad', [a] in 'bud' and [1] in 'bird'.

(I) (2) (3) (4)1: 1: ij 'bead'

'bid'

e e: e: ei 'bayed'

e e e e 'bed'

re re re re 'bad'

a a a a 'pod'

0 0: 0: OW 'bode'

u u u ua 'good'

u u: u: UW 'booed'

A A A A 'bud'

a-- a--: a--: a-- 'bird'

~C-Illustrations of the IPA 43

(I) (2) (3) (4)al al al aj 'buy'au au au aW 'bough'

::>1 ::>1 ::>1 ::>1 'boy'

d d d d 'a(bove)'Olf

StressEnglish has a very strong distinction between stressed and unstressed syllables, withstressed syllables being longer, louder, and often marked by a pitch excursion. Althoughstress placement is partially predictable, there are many instances where it is not, such asin noun/verb pairs like ['eksporUek'sportJ in contrast to [siportls;)'port]. Longer wordsfrequently have one or more syllables with a secondary stress. Stress is transcribed usingthe marks' (primary stress) and, (secondary stress), as in [,fon;)'tIS;)n] 'phonetician'.

Conventions[ p, t, k] are aspirated in word-initial position, and elsewhere when initial in a stressedsyllable, but they are always unaspirated when following [s] in the same syllable, as in'spy, sty, sky'. [b, d, '9] have little or no voicing during the stop closure, except whenbetween voiced sounds. When intervocalic and before an unstressed vowel, as in 'city,vicinity', [t] is a voiced flap, resembling [e). Cd, n] are also flaps in similar circumstances.

[I] is velarized except before [j].The vowel symbols in column (I) have the qualities shown in the accompanying

charts when pronounced in the key words. Vowels are raised before [I)] in the samesyllable, so that the vowel in 'sing' is nearer that in 'seen' than that in 'sin', the vowel in'sang' is close to that in 'sane', and the vowel in 'length' is intermediate between that in

'sing' and 'sang'. Vowels are lowered and centralized before [1], and many contrasts arelost, so that 'merry; Mary, marry' and 'Murray' are often all pronounced 6na'li]. [e] and[0] are usually slightly diphthongized. [u] and [u] are unrounded, [u] often being

pronounced with spread lips. [u] is considerably fronted after [t, d, n, I], all of which arefollowed by a mid-high front glide when preceding [u], as in 'two, new', which arepronounced [tiH, niH].

Transcription of recorded passageTwo transcriptions are given, the first being a broad phonemic transcription using thesymbols in the charts above. This transcription should be interpreted with the aid of the

conventions. The second transcription is a narrow phonetic transcription in which theconventions and other details have been incorporated. The' is often pronounced as [0;)]before words beginning with a vowel, but not on this recording. This speaker also has [h]in some words (e.g. 'he') where others might have omitted it.

44 Handbook of the [PA

Broad transcriptiona~ 'no18 ,wmd ~n (ap 'sAn W~ dIS'pjutII) 'WltJ W~Z a~ 'stJal)g~, wen ~

'urev~I~ ,kern ~'Ial) 'lrept In ~ 'WOlm 'kiok. oe ~'glid oot o~ 'WAn hu 'fer-sts~k'sid~d In 'mekII) o~ 'tJrev~I~ 'tek IZ 'kIok ,of Jud bi k~n'sId~d 'stJal)g~,oon 01 '~o~. oen o~ 'n0l8 ,wmd 'blu ~z 'hOJd ~z i 'kud, b~t 00 'mal hi 'blu 00'mOl 'kIosii dId a~ 'tJrevI~ 'fold hIZ 'kiok ~'mund 1m; ,ren ;}t 'Irest o~ 'no18,WInd ,gev 'Ap ai ~'tempt. 'oen o~ 'SAn 'Jamd ,aut 'wolmIi, ~nd I'midi~tIi a~

'tJrevI~ 'tuk ,of IZ 'kiok. ~n 'so o~ 'n0l8 ,wInd w~z ~'blaI3 tI bn'fes adt ad'SAn W~Z 00 'stJal)g~ ~v 00 'tu. I

Narrow transcription00 'n:)]8 ,wmd dn ~ 'SAn W~ dIS'pjUfIl) 'WltJ W~Z ad 'stJOl)g~, wen ~ 'tJrevl~

,kern d'lol) 'lrept In ~ 'w:)]m 'klok. ae ;}'glid oot ad 'WAn hu 'f~st s~k'sid;}d m'mekII) 00 'tJrevl~ 'tek IZ 'kIok ,of Jud bi kdn'sId~d 'stJOI)g~ o~n 01 'AO~. 'oenO~ 'n::)18 ,WInd 'blu ~z 'hOJd dZ hi 'kud, bdt Od 'm:)] hi 'bIu 00 'mOl 'klosli dIda~ 'tJrevl~ 'fold luz 'klok d'mund lum; ,ren ~t 'Irest 00 'no18 ,wmd ,gev 'Ap ai~'tempt. '5en 5d 'SAn 'Jamd ,aut 'wOlmli, dn I'mididtli ad 'tJrevl~ ,tuk 'of IZ

I

'kiok. dn 'so 00 'no18 ,wmd W;}Z d'bHu3 tI bn'fes oot 00 'SAn WdZ 00 'stJOl)g~

~v 00 'tu.

Orthographic versionThe North Wind and the Sun were disputing which was the stronger, when a traveler camealong wrapped in a warm cloak. They ag~eed that the one who first succeeded in makingthe traveler take his cloak off should be considered stronger than the other. Then the NorthWind blew as hard as he could, but the more he blew the more closely did the traveler foldhis cloak around him; and at last the North Wind gave up the attempt. Then the Sunshined out warmly, and immediately th~ traveler took off his cloak. And so the NorthWind was obliged to confess that the Sun !'Vas the stronger of the two.

Amharic

KATRINA HAYWARD* AND RrCHARD J. HAYWARD**

*Depanment ofSouth East Asia, **Department ofAfrica. School ofOriental and African Studies.

Thornhaugh Street. Russell Square, London weIH aXG, UK

Amharic, the national language of Ethiopia is the Semitic language with the greatest numberof speakers after Arabic. However, while there are large numbers of people throughoutEthiopia who speak Amharic as a second language, mother-tongue speakers areconcentrated in the highland plateau extending from somewhat south of Addis Ababa, thecapital, northwards to a line running approximately WNW from Korem. This territory isbounded to the east and west by lowland areas where other languages are spoken.

Some good descriptions of Amharic phonetics and phonology are to be found inArmbruster (1908: 4-50), Cohen (1970: 29-68), Ullendorff (1955), and Podolsky (1991).As regards its dialect situation, Amharic is (n great need of systematic research. The onlypublished work on the subject (Habte Mariam Marcos 1973) is both useful and suggestivefor future work, but it is a brief pioneering effort. The speech of Addis Ababa has emergedas the standard dialect and has wide currency across all Amharic-speaking communities. Themost divergent dialect is that of Gojjam province, though the Manz and Wallo varieties alsoshow their own marked features, especially in phonology.

Consonants

Bilabial Labio- Alveolar Post- Palatal Velar Glottal Labializeddental alveolar Velar

Plosive (p) b t d k 9 kW gW

Affricate tJ d3

Nasal m n J1Fricative [ S Z J 3 h

TapITrill r

Approximant W JLateral

1Approximant

Ejecti ve Stop (p') t' k' kW'Ejective tJ'Affricate

Ejective ,Fricative S

Additional labialized consonants: [W, bW, m W, pW', tW', hW

46 Handbook of the !PA

p posta 'post (mail)'

b b;:lkk';:ll;:l 'it sprouted' d

p p'app'as 'church patriarch' t'

m m;:lkk;:lr;:l 'he advised' nf;:lll;:lk';:l 'it gushed up'

Z

s'

tJd:)

tI'

t;:lkbl;:l 'he planted'

d;:lIT;:ls;:l 'he arrived'

t' ;:lrrdg;:l 'he swept'

n;:ldd;:lf;:l 'it Slung'

s;:lbb;:lr;:l 'he broke s.l.'

z;:lff;:ln;:l 'he sang'

s'af;:l 'he wrote'

tJ;:lllam;:l 'il got dark'

d:)arnm;:lr;:l 'he began'

tI';:lITa~;:l 'he finished'

k bbb;:lb;:l 'he encircled'

9 gaITdm;:lJl 'il surprised me'k' k';:lddad;:l 'he tore S.I.'

h hakim 'doctor'

rdZZ;:lll]a 'il became long'

J J;:lrr;:lb~ 'he plaited'

:) :);:lmb;:lr 'sun'

1 l;:lmm;:ln;:l 'he begged'

Jl taJlJla 'he lay down'

j j;:lll;:lm 'there is not'

w wat';:l 'he swallowed s.l.'

pW' pW'agWume '13th month' tW' tW'af

bW bWambwa 'pipe (conduit)'

mWmWarnmwa 'il dissolved'

fW fWaffWa 'falling water of

waterfall'

'wax taper'

kW kWas 'ball'

kW' kW'akW'ate 'whooping

cough'gW gWaggWa 'he became

full of suspense''after'

Ato Yalew Kebede is a 29-year-old male Amhara whose speech was recorded andtranscribed for this illustration. He was also responsible for the translation of 'The NorthWind and the Sun'. He grew up in Gondar, an urban centre noted for its 'good' Amharic.Gondar Amharic is extremely close to the Addis Ababa standard dialect in all its features.However, one or two things emerge in the passage which identify the origin of the speaker;for example, [b;}g;}nza f;}k' adu] instead of standard [~g;}zza f;}k' adu] 'by his own will'.

The voiceless bilabial stops Ip, p', pW'1 are extremely rare, and are confined to words of

foreign origin. Phonologically, the postalveolar affricates pattern with the stops. Because ofits affinity with the labialized consonants, we have placed Iwl in the 'labial' column. Allconsonants with the exception of Ip, pW', tW', h, hWI have geminate counterparts. In the case

of IJl/, single and geminate do not contrast phonologically, and it is usually claimed that thegeminate variant occurs intervocalically, while the single variant occurs elsewhere.Consonants may be geminated (strengthened) following nasals; an interesting example inour text is ~t'inkarel 'in strength' (see further below).

Illustrations of the IPA 47

Vowelskis 'pocket'min 'what?'jlh - jih 'this'

e k'es 'priest'

e jemmil 'he who says'

a bal 'husband';;l k;;lbt 'cattle'

::> gw::>rf 'flood'

0 s'om 'fast' (n)

u kW'ulf 'lock'

u t'ut 'breast'

u

o

(~).a

Phonetic diphthongs [al], [au], [;;ll], and [;;lU) occur, but phonotactic patterns suggest thatthese should be analysed as sequences of lal or l;;ll followed by Ijl or Iw/. The latterinterpretation is the one adopted in the transcription of the specimen passage below.

The vowels written iii and l;;ll here are often represented by the symbols l;;ll and lalrespectively. It is possible to argue that Iii is not present in underlying representations, but isalways epenthetic (see Hetzron 1964; Hayward 1986). In general in the Ethiopic script, eachsymbol represents a consonant plus a vowel; for example aD =Im;;l/, IJ'l =Imel and IJ'l =Ima/. However, the symbol for a consonant plus Iii also does duty for a consonant standingon its own, i.e. go = Imil or Im/.

The vowels represented by the symbols [I, U, e, ::» are not independent phonemes, butallophones of the central vowels Iii and l;;l/. Following posta!veolar and palatal consonants,(which, from a phonological point of view, form a 'palatal' series), Iii and l;;ll are oftenfronted to !II and lei. Following labialized consonants and Iw/, Iii and l;;ll typically haveretracted and rounded pronunciations ([u) and [::>)). Spelling also needs to be taken into

account here, since a literate speaker, whose spelling of a particular form indicates iii or l;;l/,may pronounce [i) or [;;l), especially in careful speech (thus, the variation jlh - jih notedabove). In the specimen passage below, we have written [I, U, e, ::>], rather than morestrictly phonemic Iii and l;;l/, in such cases.

The question also arises of whether [I, U, e, ::» (as allophones of central vowels) are

really distinct from the independent phonemes /ii, lui, lei, and 101 respectively, as our chartindicates. With regard to the back rounded vowels, the issue is further complicated by thefact that consonants preceding them have anticipatory lip rounding. For example, the Ikl inIkul is phonetically [kW), and for this reason it is often difficult to choose between Ikwul(= Ikwil in a more strictly phonemic interpretation) and /kul on strictly phonetic grounds. In

non-final closed syllables and final syllables closed by two consonants, the two series ofvowels do not contrast. We have interpreted all vowels in these environments as phonemiccentral vowels. For example, we have written IgWul~tl rather than Igul~tl 'strength', andIwuddl rather than Iwuddl 'dear, expensive'. This contrasts with our informant's spellingsof the same words, which indicate 'gu-' and 'wu-' respectively.

48 Handbook of the [PA,

In non-final open syllables and final syllables closed by a single consonant, both series

of vowels may occur. In such environments, there is potentially a length contrast, Ii, u, e, 01being susceptible to prolongation. Prototypically, these long vowels contrast with shortII, u, C, ',)1 in both duration and quality, butlthere is a less certain middle ground. In the caseof verbs and their derivatives, morphological considerations will often lead to a decision infavour of phonemic central vowels. For example, in the case of /kW' ut' irl 'number', thedecision to write /kw'u-I (/kw'if in a more strictly phonemic representation) was influencedby the cognate verb Ikw

' ',)tt';m~1 'he counted'. This belongs to the same verb class asIk'~tt'~r~1 'he hired' and should, therefore, have the same underlying stem vowels; thus, aphonemicization IkW'~tt'~rnl is to be preferred to Ik'ott'~r~/. Our IkW'ut'irl agrees with our

informant's spelling, which also indicates :an initial labialized velar. However, we found itvery difficult to decide between /b:lkkwul! and /b~kkul! 'side, direction', where no suchcognate forms are available for comparison. Our final choice of /b~kkWuUcontrasts with the'-kkul' of our informant's spelling. (The complexity of the situation is well-described inUllendorff 1951; cf. especially pp. 82-3.)

StressStress is weak, and its position is variable,. Further investigation is needed concerning therelationship between stress and intonation on the one hand and between stress andgemination on the other. I

ConventionsIp, t, tf, k, kWI are all moderately aspirated. Voiced obstruents are devoiced pre-pausally andwhen a voiceless obstruent follows, for example in /libsl 'clothes', [li~s]./bl is realized asan approximant [~] medially between sonorants (for example, in IgWulb~U [gWulf}~t]). III is

always clear, never dark. Single Irl is a tap, geminate Irrl a tril!. When they precede thevowels iii and lei, but more especially the latter, consonants may be strongly palatalized, forexample in Igizel [giizje] 'time'.

In rural speech, Ipl may be replaced by /fl. In initial position, 131 is now found only inthe speech of Manz. In Addis Ababa, it has been replaced by the affricate Id3/. IJlI occursinitially in only one or two rare words. Is'l is most commonly pronounced as an affricate,[ts'). It is tending to merge with It'l, especially in initial position. However, complexregional and sociolinguistic factors affect the occurrence of initial Is'l, which may beretained in certain lexical items in educated speech. Our informant consistently pronouncedinitial [[s'] in the word Is'~hajl 'sun'.

Ihl is voiced between vowels, and may be realized simply as breathy voice on apreceding and/or following vowel (for example, in Ibnzihiml 'and on these terms',phonetically [bnzA:m)). Following iii or Ii/" /hi is typically pronounced as a palatal fricative[9]. Thus, the form [bnzi] 'at this' (lb~zzih/) would be pronounced [bnzi9] in isolation or

in slow speech.A prothetic [i] is often inserted before word-initial Ir/, for example [ir~33im] 'tall' (no

examples occur in the text). [i) may also be inserted after word-final consonants, when the

l/luslralions of Ihe IPA 49

following word begins with a consonant. In our speaker's rendering of the text below, thereis variation, for example [libsi I;)bso] with epenthesis, but [Jibs k' adimo] without it. Suchcases of epenthetic [i] are not noted in the transcription. Another case of epenthetic [i] in ourinformant's rendering of the text which is not noted in the transcription, but which needs tobe mentioned, occurs in the form Ib;)t'inkarel [b;)t'inikkare]. Here, the Ik/ undergoes postnasal strengthening and this conditions [i] insertion. In other cases, [i] may be devoiced andso short that it is barely audible. This occurs in two forms in our text, when it follows thefeminine prefix It-I (in these morphological contexts geminated to tt-): [sittid3emmir] 'whenshe was beginning' and [jemmittibalt'] 'she who is greatest'. In the second of these, thepronunciation of fbi as W] confirms the presence of the vowel.

Transcription of recorded passage

s';:lhajmna k;:ls;:lmen b;:lkkWu] jemmin;:lfs::lw n::lfas ine nej1 t';:lnkarra ine nej1t';:lnkarra b;:lmmil jrkbrakk~ru n;:lb~r. b;:lzzi gize and m;:ll)g;:ldepj1a jebirdm;:lkk;:l]abja libs ];:lbso jrggWaz n;:lbb;:lr. bzjam m;:ll)g;:ldej1j1;:lw l;:lbirdm;:lkblak;:lja l;:lbsot jemmihed;:lw Jibs k';:ldimo jasw:)l!;:lk';:l b;:lt'inkarejlb;:llt' a) b;:lmmi] t;:lsmammu. b;:lzzihim m;:lS;:lr;:lt bS;:lmen b;:lkkwuljemmin;:lfs;:lw n;:lfas ball;:l ~)ell;:lw gWulb;:lt ~haj) mff;:ls;:l. honom gin ~haj)

b;:ln;:lff;:ls;:l kW'ut'ir m;:ll)g;:ldej1j1;:lw jebas;:lwnu b;:l);:lbb;:ls;:lw libs j1dd3;:lbb:mjemm;:lr. b;:lm;:ltJ';:lrr;:lfam bS;:lmen b;:lkkwul jemmin;:lfs;:lw n;:lfas b;:lg;:lnzaf;:lk'adu akw'om;:l. s'ahajm b;:lt;:lrawa w;:ltt'atfmna muk';:ltwan mawr;:ldsittid3;:lmmir m;:ll)g;:ldej1j1;:lw minimm saj kW':)j w:)dijawnu jd;:lbb::ls;:lwl libsaw:)ll;:lk';:l. b;:lm~tJ';:lrr;:lfam bS;:lmen b;:lkkWu) jemmin;:lfs;:lw n;:lfas s'::lhajb;:lt'inkare bssu jemmitti~lt' m;:lhonwan jal;:lwudd b;:lgidd amm;:ln;:l.

Orthographic version

6th£f hl'llJll"l ntH\. VIl1J..~~l'l(])- ~4-l'I >.\. ~"i ffi"ll)l. >.\. ~"i ffi"ll)l. nlJll..l\£hl.h'- ~nc:: nlt\J 1.tl ')'''l£" """l1f~ Vf)C£" ""hl\hf l\f)/'I "f)f'\ £~11 ~nc::

htUlJ" """l1f?(])- t'tf)C£" ""hl\hf t't{)('T VIl1J..'H(])-"l l\f)l'I t£"'fD fl'lIDt'tt n'Y"ll)l. £nl\fTll\ nll1J..l\ tl'llJl}au-:: nH.\J"" ""LPLT hl'llJll"l nh-l\ VIl1J..~~l'l(])- ~4-l'l

IMnt\,t't(])- ?-l\nT nU£l\ " .. l'l:: lfllJ" "'''l nU£l\ mL.l'l "",'YC lJU"l1f.rQlVI1l'l(])-'J. nt'tnl'l(J)- l\f)l'I £f(ll f""c:: n""a\.L7ilJ" hl'llJll"l nh-l\ VIl1J.. ~~l'lQl- ~4-l'I

n1"lt1 L.~£" ,),4>"":: 6th£lJ" nt l. ep IDfTl,.f au- t t"l lJI}(J)- L£" /'IT flJ" C """l1 f. rQllJ"1lJ" fl£4>£ IDSlf(J)-'J. vt'tnl'l(J)-"l l\f)l'I ')'IDt'tt:: n""ffiL7ilJ" hl'llJll"l nh-l\VIl1J..~~l'l(J)- ~4-/'I 6th£ n'Y"ll)l. hI'\. VlJ"'l-nl\'Y ""lfl"l ft't(J)-£" n..,£" ,),""~::


ReferencesARMBRUSTER, C. H. (1908). Initia Amharica: An Introduction to Spoken Amharic.

Cambridge: The University Press.

COHEN, M. (1970). Traite de langue amharique (Abyssinie). Paris: Institut d'ethnologie.HABTE MARIAM MARcos (1973). Regional variations in Amharic. Journal ofEthiopian

Studies II, 113-29.HAYWARD, R. J. (1986). The high central vowel in Amharic: New approaches to an old

problem. In Fishman, 1. A. et al. (editors), The Fergusonian Impact. Vol. I: FromPhonology to Society. Berlin, New York, Amsterdam: Mouton de Gruyter.

HETZRON, R. (1964). La voyelle du sixieme ordre en amharique. Journal ofAfricanLanguages 3, 179-90.

PODOLSKY, B. (1991). Historical Phonetics ofAmharic. University of Tel-Aviv.

ULLENDORFF, E. (1951). The labiovelars in the Ethiopian languages. Rassegna di StudiEtiopici 10,71-84.

ULLENDORFF, E. (1955). The Semitic Languages ofEthiopia: A Comparative Phonology.London: Taylor's (Foreign) Press.

Arabic

ROBIN THELWALL* AND M. AKRAM SA'ADEDDlN**

'2121 1st Avenue NW, Calgary, AB T2N OB6, Canada

"University of Kuwait, Kuwait City

There are many tendentious issues in proposing a text in any form of Standard or ClassicalArabic in spoken form. We will not justify the present text more than to say that thespeaker was born in Safad, North Palestine, lived and was educated in Beirut from age 8to 15, subsequently studied and taught in Damascus, studied phonetics in Scotland andsince then has resided in Scotland and Kuwait.

It seems widely accepted that there are two dominant (prestigeful) dialect centres ofgravity for Spoken Arabic: first, that of AI-Shaam (greater Syria, from the Mediterraneancoast to the eastern edge of the Fertile Crescent) and, second, Egypt (focused on AI-AzharMosque and University in Cairo). This is not to disregard the status of the ArabianPeninsula, Iraq or other areas as the sources of models, but the matter remains to beadequately investigated.

Consonants

Bilabial Labio- Dental IAlveolarl Post- Palatal Velar Uvular Pharyn- Glottaldental alveolar geal

Plosive b t d k q 7

Nasal m n

Fricative f e 51s z IJ X Y h h

Affricate d3

Trill r

Approximant J W

Lateral1Approximant

Pharyngealized consonants: t" d" s", 5" 1", 7", , .Note: 17'1 is represented as I'll in the transcriptions below.

t tijn 'figs' k kalb 'dog'

b balla 'recovered' d dijn 'religion' q qalb 'heart'

m maHa 'got bored' n nadda 'released' ? sa?ala 'asked'

f dafara 'stank' saara 'walked' x xilaaf 'dispute'

8 da8ara 'covered' z zaara 'visited' y yilaaf 'cover' (n)


0 haakaoa 'thus' J Jadda 'doubling mark' 'i' sa'i'ala 'coughed'

h huruwb 'wars'

d3 d3adda i 'grandmother'raw?a

I'splendour' h huruwb 'escape' (n)r

law?a ~ 'sorrow' j jaraqaan 'jaundice'I

was'l 'receipt'w

s' s'arf 'exchange' t' xit'aab 'letter'

0' o'arf 'envelope' d' xid'aab 'henna'

I' al'I'aa: 'god'

Vowels'i'idd 'promise l ' u 'i'udd 'come back!'

ij 'i'ijd 'feast' a 'i'add 'counted' uw 'i'uwd 'lute'

aj 'i'ajn 'eye' aa 'i'aadd 'came back' aw 'i'awd 'return'

U IJ UW

a aaStressStress within Arabic words falls on the fi l1 al heavy syllable of the morpheme, disregardingsuffixes. A heavy syllable is defined as having either a long vowel or a (C)VCC structure.In words with no heavy syllable the matter is still to be investigated for spoken ModemStandard Arabic, and it is assumed that the stress patterns operating in the backgroundColloquial Arabic of the speaker will operate.

Syllabification and the determiner [al·]The assimilation of the consonant of the definite article to following [+ coronal] segmentsis shown in the transcription. The vowel is assimilated to the final vowel of the precedingword, if any. Since word division is shown according to orthographic conventions, wehave left out the vowel of the determiner in such cases. The actual realization is, of course,continuous.

ConventionsIt, k / are aspirated, /t'/ is unaspirated, /t" d'i, s" 0'/ are Retracted Tongue Root, involvingsimultaneous pharyngealization and greater or lesser degrees of velarization. Theconvenient phonological term within Arabic phonetic studies is 'emphasis/emphatic'.


These consonants exert a strong co-articulatory effect on all sounds in preceding andfollowing syllables in given lexemes. This is not blocked by labials, the labialized III, and

Ir I. Only Ij I blocks the process. I'll is a Retracted Tongue Root glottal stop. Thisrealization is supported by Gairdner (1925), AI-Ani (1970) and Kastner (198 I) as well as

extensive observation of a range of speakers from different regional origins residing in

Kuwait at present (1990). Nowhere have we observed a pharyngeal fricative. Ixl is

accompanied by uvular trill. Geminate consonants in word-final position in pre-pausal

realization have an incipient syllable onset with [~], e.g. I 'Iaddl ~ ['Iad.d°], where [.]represents a syllable boundary. In context they are realized without this feature.

The primary vowel allophones are conditioned by the presence or absence of

neighbouring emphatics in the same word and by word-final position for non-emphatic

contexts. lal has three main allophones: [0] before Retracted Tongue RoOl consonants, [B]

before word boundary, and [a] elsewhere, taking phonological length into account.

Length is neutralized in open syllables. fil and fiji become [i] before Retracted TongueRoot and pharyngeal consonants. This may be more retracted than a high central

unrounded vowel, reaching towards [V]. In most contemporary colloquial dialects there is

no phonological opposition between [i] and [u]. lui and luwl become [u] and [v] before

Retracted Tongue Root and pharyngeal consonants. In word final position the length

opposition is neutralized.

Points of particular interest in the sound system include:1. The 'd3im', which is here pronounced as a voiced lenis palato-alveolar affricate

(Gairdner 1925: 23; see however Kastner 1981: 65, where it is described as a fricative).

2. The 'O~aa?', which is here pronounced IO~1 'a voiced dental fricative with tongue root

retraction and concomitant pharyngealization and velarization'. See Kastner (198 I: 62ff),

and in contrast, Gairdner (1925: 21).

3. The ''Iajn', which is realized as a pharyngealized glottal stop (see also Kastner 1981:

49; AI-Ani 1970: 62-71; and Gairdner 1925: 28-9).

There are various different reflexes of these sounds in the present day spoken dialects ofArabic, and in the phonetic descriptions of past dialects, which it is irrelevant to pursuehere.

Transcription of recorded passagekaanat rijhu IJamaali tatad3aadalu wa IJamsa fij ?ajjin minhumaa kaanat?aqwaa min al?uxraa, wa ?io bi-musaafirin jatQu1'u mutalaffi1'an bi1'abaa?atin samijka. fa tafaqataa 1'alaa 1'tibaari ssaabiqi fij ?id3baariImusaafiri 1'alaa xal1'i 1'abaa?atihi l?aqwaa. 1'as~afati rijhu IJamaali bi?aqs~aa ma stat~aa1'at min quwwa. wa laakin kullumaa zdaada 11'as~f

izdaada Imusaafiru tadaqquran bi1'abaa?atih, ?ilaa ?an ?usqit~a fij jadirrijh fa taxallat 1'an muhaawalatihaa. ba1'da?ioin sat~a1'ati IJamsu bidif?ihaa, fa maakaana min almusaafiri ?illaa ?an xala1'a 1'abaa?atahu1'alaa ttauw. wa haakaoa id~t~urat rijhu IJamaali ?ilaa li1'tiraafi bi?annaIJamsa kaanat hija l?aqwaa.



ReferencesAL-ANl, S. H. (1970). Arabic Phonology: An Acoustical and Physiological Investigation.

The Hague: Mouton.

GAIRDNER, W. H. T. (1925). The Phonetics ofArabic. London: Oxford University Press.

KASTNER, H. (1981). Phonetik und Phonologie des modernen Hocharabisch. Leipzig:Verlag Enzykloplidie.

Bulgarian

ELMAR TERNES * AND TATJANA VLADIMIROVA-BUHTZ**

*llIstitut!iir Pholletik, Allgemeille Sprachwissenschaft //lui Indogermallistik, Universittit Hamburg,Bogellallee II, D-20144 Hamburg, Germany

**Aumallellweg 2, D·79117 Freiburg im Breisgau, Germany

The style of speech illustrated is Standard Bulgarian as used by people with an academicbackground. Historically, the Standard contains elements of western and eastern Bulgarian,but in its present form, it cannot be localized in any particular region.

The speech on the recording is that of a 33-year-old speaker holding a post at theUniversity of Sofia. Bulgarian examples are given in a transliteration from Cyrillic script.

Consonants

Bilabial Labio- Alveolar Post- Palatal Velardental alveolar

Plosive p b t d k 9Nasal m n

Fricative f V S z J 3 X

Affricate ts dz tJ d3

Trill r

Approximant JLateral

IApproximant

p pija 'I drink' t tom 'volume' k kol 'pole'

b bija '1 beat' d dam 'home' 9 gal 'naked'

IS tsar 'tsar' tJ t'Sar 'charm'

dz dzift 'tar' d:) dzob 'pocket'm most 'bridge' n nos 'nose' x halka 'ring'f far 'lighthouse' s sarna 'roe' J 'Sal 'scarf'v var 'limestone' z zarna 'corn' (pi) :) wi 'pity'

row 'rose' lale 'tulip' j jak 'strong'

The phonemic analysis underlying the present transcription does not assume the existence ofpalatalized consonants. An alternative analysis postulates the following palatalizedconsonants: / pj, bi , ti , d i , ki, gi, ts i, dzi , mi , nj, ri, 0, vj, si, zi, xi, Ii /.

Chinese (Hong Kong Cantonese)

ERICZEE

Phonetics Laboratory, Departmeilt ofChinese, Translation and Linguistics,

City University of Hong Kong, 83 Tal ehee Avenue, Kowloon, Hong Kong

The style of speech illustrated is that typical of the educated younger generation in HongKong. The recording is that of a 22-year-old male university student who has lived all hislife in Hong Kong.

Consonants

Bilabial Labio- Dental Alveolar Post- Palatal Velar Labial- Glottaldenial alveolar Velar

Plosive p ph t th k kh kW kwh

Affricate ts tsh

Nasal m n I)

Fricative f s h

Approximant J W

Lateral1Approximant

p pal 'father' t tal 'dozen' k kal 'to add'ph phal 'to lie prone' th thal 'he/she' kh khal 'truck'

m mal 'mother' n phanl 'to climb' fJ phafJl 'to cook'

f fal 'flower' s sal 'sand'ts tsal 'to hold'

tsh tshal 'fork'

w wal 'frog' jBUl 'worry'

IBUl 'angry'

kW kWal 'melon'kwh kwhal 'to boast'

h hal 'shrimp'


VowelsMonophthongs(I) or (2)

i: si1 'silk' hip-l 'to assist' him1 'moderate'y y: syl 'book' hyt-l 'blood' synl 'sour'

E E: SE1 'to lend' hEk-l 'to eat' SEl)l 'sound'

0:: 0::: ho::l 'boot' so::k-l 'to whittle' SO::l)l 'hurt'

a a: sal 'sand' hak, 'guest' saml 'three'

::> ::>: s::>l 'comb' h::>k-l 'shell' bl)l 'square'

u u: fUl 'husband' fut-l 'wide' funl 'cheerful'

slkl 'colour' Sll)l 'star'

e e sen 'shirt' senl 'to inquire'

e e sepl 'wet' seml 'heart'

U U sukl 'uncle' SUl)l 'loose'

• luu.

Diphthongs \~ &\e •J(I) or (2)

ai a:i sail 'to waste'

ei ei seil 'west'

au a:u saul 'basket'

eu eu seul 'to receive'

el ei heil 'rare'

EU E:U teu-l 'to throw'

ey ey seyl 'bad'

::>1 ::>:i s::>il 'gill'

Ul u:i fuil 'ash'

iu i:u siUl 'to bum'ou ou soul 'beard'

Tones

1 (high, level) sil 'silk' Slkl 'colour'-I (mid, level) si-l 'to try' sit-l 'to reveal'-I (low-mid, level) si-l 'matter' slk-l 'to eat'J (low-mid to low, falling) siJ 'time'1 (low-mid to high, rising) Si1 'history'A (low-mid to mid, rising) siA 'city'


color [ku'lo] 'colour' but coloraina [kul1J'fajn;l] 'bright colours'. There are however a

number of exceptions, likefutur [fu'tur] 'fJture'.In emphatic speech and in certain dialects, initial postalveolar fricatives are affricated:

ximple ['(t)fimpl;l] 'foolish', germa [(d)3;lrima] 'brother'. Underlying Ivl is realized as Ibl in

Central Catalan; [b] or [~] occur before 'a vowel and [w] occurs finally; this results inalternations between [b] or [~l and [w]. rhus blava ['bla~;ll 'blue' (fern) vs. blau [blaw]

'blue' (masc).

Transcription of recorded passage1~ t[~mun'tan~ j ~1 s:>1 ~z oispu'ta~~n I sust~'niI) 'kao u k ef.. 'e[~ 1 mes br Ikwan d~ 'sop~ I 'bew~n urn bi~'d3e k~ s ~'bst~ mbuli'kat ~n un~ 'yraI) 'ka~ [Ibal) kum~'ni b 1 ki pri'me f~'ri~ b 1~i~'d3e s t[~'yes 1~ 'ka~ I s~'[i~ til)'gup~1 mes br II l~ t[~mun'tan~ s 'p:>z ~ ~u'fa m 'tot~ 1~ 'se.l3~ 'fu[i~ I ~'[:> bmmez .I3u'fa~~ I mez ~1 ~i~'d3e s ~~[i'ya.l3~ m 1~ 'ka~ II ~ 1~ fi I ba d~'Ja 'kor~

ferli 't[ew[~ II ~1~'z:>[~Z ~1 s:>1 kum~n'sa ~ ~[i'f..a I j ~1 kab d urn mu'men I ~1.I3i~'d3e I ben ~sbl'fat I ~s trew 1~ 'kap~ II ~'Ji I 1~ tr~mun'tan~ ~a '.I3e o~

kUIl)f~'sa k~ 1 s:>1 'e[~ 1 mes br II

Orthographic versionThis text was published in the Principles oj the International Phonetic Association edited in

1912 but was omitted from the specimens in the subsequent versions of the Principles. Ithad appeared for the first time in Le MaItre Phonitique vol. 26 ns. 7-8 (1911) p. I 19 and

was prepared by losep M. Arteaga (Barcelona, 1846-1913), a Catalan phonetician who

became a member of the Council of the IPA in 1907.

La tramuntana i el sol es disputaven, sostenint cada u que ell era el mes fort, quan de sobte,

veuen un viatger que s'acosta embolicat en una gran capa. Van convenir que el qui primer

faria que el viatger es tragues la capa seria tingut pel mes fort. La tramuntana es posa a bufaramb tota la seva furia; pero com mes bufava, mes el viatger s'abrigava amb la capa; a la fi,

va deixar correr fer-Ii treure. Aleshores el sol comen<;:a de briBar, i al cap d'un moment, el

viatger, ben escalfat, es treu la capa. Aixf, la tramuntana va haver de confessar que el sol era

el mes fort.

AcknowledgementsWe would like 10 thank Canne de la Mota and Antonio Rfos for their useful comments and suggestions.

ReferencesALARCOS, E. (1983). Estudis de linguistica catalana. Barcelona: Ariel.BADIA I MARGARlT, A. M. (1951). Granuitica historica catalana. Barcelona: Noguer.

Catalan translation: Valencia, Tres i Quatre, 1981.BADIA I MARGARlT, A. M. (1988). Sons i 'jonemes de la llengua catalana. Barcelona:

Publicacions de la Universitat de Barcelona.


BONET, E.-LLORET, M.-R. (1998). Fonologia catalana. Barcelona: Ariel.GILl, J. (1974). Introductory Catalan Grammar. Oxford: Dolphin Books.HUALDE, J. 1. (1992). Catalan. London: Routledge.INsTITUT D'EsTIJDIS CATALANS. (1990). Proposta per a un estimdard oral de la llengua

catalana I: Fonetica. Barcelona: Institut d'Estudis Catalans.RECASENS, D. (1991). Fonetica descriptiva del catala. Barcelona: Institut d'Estudis

Catalans.RECASENS, D. (1993). Fonetica ifonologia. Barcelona: Enciclopectia Catalana.VENY,1. (1985). Introducci6 a la dialectologia calalana. Barcelona: Enciclopedia Catalana.WHEELER, M. W. (1979). Phonology of Catalan. Oxford: Basil Blackwell.YATES, A. (1975). Catalan. London: Hodder and Stoughton (Teach Yourself Books).

Ctoatian

ERNESTINA LANDAU

Vuceticev prilaz 5, 10020 Zagreb, Croatia

MilO LONCARIC

Institute ofCroatian Language, Strossmayerov trg 2, 10000 Zagreb, Croatia

DAMIR HORGA AND Iva SKARIC

Department ofPhonetics, Faculty of Philosophy, University ofZagreb, 10000 Zagreb, Croatia

Croatian is one of the Slavic languages. As the national language of Croats, it consists ofthree main dialects, Stokavian, Kajkavian and Cakavian, named after the different formsof the interrogative pronoun meaning 'what' IItol, Ikaj/ and ItIa!, orthographically §to,

ka}, ca. Standard Croatian was established in the eighteenth and nineteenth centuries. It isbased on the New Stokavian Jekavian dialect.

The speech on the accompanying reclFding is that of a 57-year-old female announcerat the Croatian Television Network reading in a colloquial style. The style of speechillustrated is that of many educated speakers of Standard Croatian as spoken in theRepublic of Croatia.

Consonants

Bilabial Labiodental Denial Alveolar Postalveolar Palatal Velar

Plosive p b t d k 9Affricate ts tJ d3 tc d*Nasal m n J1Fricative f S Z J 3 X

Trill r

Approximant 0 JLateralApproximant 1 A

p pi:tce pIce 'drink' (n)

b bi:tce bice 'creature't tt1:gad dt1:ga

tl.1gadl.1ga

'sorrow''rainbow'

k ko:st kosi 'bone'9 90:st gost 'guest'

/

lllustrations of the IPA 67

ts tsar car 'czar'

tI tIep cep 'cork' te le:tea Mea 'lentils'

d3 d3ep d'tep 'pocket' dz le:dza IMa 'back' (0)

m m6:j rnoj 'my' n n6:s nos 'nose' J1 J16:j njoj 'to her'fa:za taza 'phase' s selo seJo 'village' x xl:r hJr 'caprice'

u va:za vaza 'vase' z za:jam zajam 'loan'

I Ia:1 Mil 'scarf'

3 3a:l 'tal 'beach'ra:d rad 'work' (0) j jug jiJg 'South'l6:v lov 'chase' I. f.eti !jeli 'in summer'

VowelsThe system of Croatian vowels consists of five monophthongal vowels, Ii, e, a, 0, ul,which occur distinctively short and long, and a diphthong liel occurring only in a longsyllable. These vowels have more or less constant qualities in stressed and unstressedpositions. There is also a syllabic trill IfI which can either be long (4 to 5 contacts) or short(I to 2 contacts). When occurring between two consonants, it is sometimes pronouncedtogether with nonphonemic [~l, i.e. as [~r), e.g. [wrt) V!l 'garden'. A short [~) also occursin certain other contexts, as for example in pronouncing the names of some letters.

The diphthong liel begins at the position of the Iii monophthong and ends at theposition of the monophthong lei. It can also be pronounced [ijeJ, but thio e.:11 e

a single syllable.

Short

vlle vUe 'hayfork'

e tek (13k 'only'

a pas pas 'dog'

0 k6d kod 'by, at'

u duga diJga 'stave'

Longi: v'i:le vile 'fairies'e: te:k lek 'appetite'a: pa:s pas 'belt'0: k6:d kOd 'code' [~) [~rstJ P!"5l 'finger'

u: di1:ga duga 'rainbow' ie bielo bijelo 'white' (adj)

ConventionsWhen If, ts, xl precede a voiced obstruent they occur as [v, dz, V): [gr6v~bi) grOf bi 'theearl would', [otadz~bil alae bi 'father would', [ouiy~bjl avih bi 'of these ... would'.

r

I

IIII,

Handbook a/the IPA

ere are similar alternations affecting other voiceless obstruents, but these may berded as the replacement of one phoneme with another; similarly, voiced obstruents

It mate with their voiceless counterparts when followed by a voiceless obstruent.)I The post-alveolar affricates II, 31 appear as the palatal sibilants [c,~] when they;pr cede ltc, d~/: [mlc_tee] mB ce 'the mouse will', [pG:~_tce] puz ce 'the snail will'.I/u occurs as [w] before the vowel lui: [wu:k] vuk 'wolf'. 1m! occurs as [ll]] before If, u/:[t aTl]uaj] lIamvaj'tram'. Inl occurs as [I)] before Ik, g/: [sta:l)ka] stfinka 'pause'.I I occurs as [h] when it is initial in a consonant cluster: [hmeA] hmelj 'hops'.

A sequence of vowels across a word boundary may be separated by a glottal stop:[':_?6nda] i anda 'and then'.

tress and accentn addition to the contrast of length mentioned in connection with vowels, Croatian alsoas rising ([~]) and falling ([ -]) pitch accents. It is traditional to regard these two

istinctions as forming a four-way contrast of word accent, namely, long rising ('), short

ising ('), long falling C) and short falling C); n denotes length on an unstressed vowel.he parenthesized signs above are not part of the usual orthography, but are used to mark

the accents in specialized linguistic works. The syllable on which the distinctive pitchmovement begins and in which it may be completed is the stressed syllable. Monosyllabic(full) words always have a falling accent: ImetII mec '(sporting) match', Ime:d/ med'honey'. Falling accents may also occur on the first syllable of a disyllabic or longerword, e.g. Imamal mama 'mama', If6rmal forma 'form (n)', but they do not fall on latersyllables apart from a few exclamations, e.g. lahal aha 'aha!', and loan words, e.g. 13ele:1zele 'jelly'. Rising accents may occur on any syllable but the last of a word, e.g. /zT:ma/zima 'winter', /terasal lerasa 'terrace', ITzuiJle:Jlel izvinjenje 'excuse (n)' (with a

secondary accent on the initial iiI).There are two groups of words, proclitics and enclitics, which form a unit with the

following or preceding stressed word respectively. Enclitics are always unstressed:

16:n_tije to: rekaol on tije to d3kaa 'he told you that'. Proclitics can be either unstressed

or stressed. They are unstressed when the following word has a rising tone, e.g. lu~u6di/

u vodi 'in water', Ipri~ra:dulpri radu 'at work', but they can be stressed when thefollowing word has a falling tone, e.g. lokol aka 'eye' but IG_okol iJ aka 'in the eye';Igra:dl grad 'town' but IG~gra:dl iJ grad 'to town'; IIGmal siJma 'wood' but 1i1~Iumil

u:Sumi 'in the wood'.The stress position in Croatian is relatively free. Unstressed syllables preceding the

stressed one are always short while the unstressed syllables following the stressed one caneither be short or long, e.g. IkGtcal kiJca 'a house' vs. I kG:tca:l kl1ca 'houses (gen pI)'.

The unstressed syllables are shorter than the corresponding stressed ones, by about 50 percent in the case of long vowels and 30 per cent in the case of short ones.


Transcription of recorded passage\I sjeue:mi: ledeni: ujetar i~sG:ntse~su~se prepirali o~suojoj sna:zi II st6gaadlutfe: da~anome 6d~J1i:x prlpadne p6bjeda kaji: suu:tfe: I tf6ujekapG:tnika II ujetar zapotfe Sna:3nO pu:xati I a~budu:tci da~je tf6ujek tfUf:stodf3aO 6djetcu I nauali 6:n j6f jatJe: \I tJ6ujek pa:k j6I jatJe: ocst11denipr'itisnu:t I nauu:tJe: na~sebe j6J ulJe: 6djetce: I d6~se ujetar ne~umori:

i~prepusti:~ga tada sG:ntsu \I ana: u~potJe:tku zasija umjereno II kad~je

tJ6ujek skinuo suuI:Jak 6djetce: I patJi:si ana: j6J jatJe: 3egu I d6~setJ6ujek I u~nemogu:t<;;nosti da~odoli suntJeuo:j toplI ni ne~stJu:tJ e: Ii~ne~pod~e: na~ku:paJ1e u~rije:ku tekutcitsu II prf:tJa pokazuje:da~je~tJe:stouspjeJnije: uujera:ua:J1e I negoli na:si:A:e \I

Orthographic version with diacritics

Sjeverni ledeni vjetar i Sunce su se prepirali 0 svojoj snazi. Stoga odluce da onome odnjih pripadne pobjeda koji svuce CQvjeka pfitnika. Vjetar zapoce snazno puhati, a buducida je covjek cv?sto drzao odjecu, navali on jos jace. Covjek pak, jos jace od studeni

prltisnut, navuce na sebe jos viSe odjece, dok se vjetar ne umori i prepust! ga tada Suncu.Ona u pocetku zasija umjereno. Kad je covjek sklnuo suYisak odjece. povisi ona jos jacezegu dok se covjek. u nemogucnosti da odoli suncev6j toplini. ne svuce i ne pode nakupanje u rijeku tekUcicu. Prlca pokazuje daje cesto uspjesnije uvjeravanje negoli nasilje.


Sjeverni Iedeni vjetar i Sunce su se prepirali 0 svojoj snazi. Stoga odluce da onome odnjih pripadne pobjeda koji svuce covjeka putnika. Vjetar zapoce snazno puhati, a buducida je covjek cvrsto drz.ao odjecu. navali on jos jace. Covjek pak, jos jace od studenipritisnut. navuce na sebe jos vise odjece. dok se vjetar ne umori i prepusti ga tada Suncu.Ono u pocetku zasija umjereno. Kad je covjek skinuo suvisak odjece, povisi ono jos jacezegu dok se covjek, u nemogucnosti da odoli suncevoj toplini, ne svu ce i ne pode nakupanje u rijeku tekucicu. Prica pokazuje da je cesto uspjeSnije uvjeravanje negoli nasilje.

Czech

lANA DA1"KOVlcovA

Departmenr of Phonetics and Linguistics, Universiry College London,I

4 Stephenson Way, London NWl 2HE, UK

Czech belongs to the western group of Slav'ic languages and is spoken as a mother tonguemainly in the Czech Republic, In its two provinces of Bohemia and Moravia it is spoken byabout 10 million people, There are also some relatively large Czech-speaking communitiesin North America and smaller, isolated ones in neighbouring European countries. Theclosest related languages are Slovak and Polish.

A continuous tradition of Czech writing begins in the late thirteenth century. Arecognisably near-modern spoken language and written standard had evolved by thesixteenth century. The spoken language continued to develop after the establishment of arelatively stable written standard. Thus when the written form underwent a modernizingrevival from the late eighteenth century, the result was a noticeable divergence betweenwritten and spoken usage. This state has persisted to the present day.

The transcription of the sample text is :based on the speech of a native speaker fromPrague speaking standard Czech. I

Consonants

Bilabial Labiodental Alveolar Postalveolar Palatal Velar Glottal

Plosive p b t d c j k 9

Nasal ill n p

Fricative f V S z f 3 X fi~

ijAffricate tsr

Trill rApproximant JLateralApproximant 1

All the plosives occur in voiced/voiceless pairs and are normally not aspirated. The glottalstop, while not phonemic, is typically used before vowels at the beginning of words orbetween vowels within the word after a prefix. Apart from the glottal and velar fricatives, allthe fricatives occur in voiced/voiceless pairs. The glottal fricative is always voiced. The

velar fricative is normally voiceless, although its voiced counterpart [V] may occur


allophonically. For more detail about the interaction between the glottal and velar fricativeswith respect to assimilation, see the section on Assimilation below. Two affricate phonemes

occur, a voiceless alveolar Itsl and a voiceless post-alveolar Itj/. They both have an

allophonic voiced counterpart arising from voicing assimilation (see below).

There are two trills, /rl, as in ruka lrukal 'hand', and Irl, as in reka Ireka/ 'river'. The

first one is an alveolar apical trill with 1-3 periods of vibration. It is immune from voicing

assimilation and occurs voiced in all positions in the word. In the case of Irl, the place of

articulation is normally similar to that for /r/. Although it may be produced with the blade ofthe tongue, the main differentiating characteristic from II'I is the number of vibrations, whichmay be 1-2 greater than in 11'/, and their lesser amplitude than for the vibrations in 11'/. Also,

the constriction is narrower and the velocity of air greater. This sound often starts as a trillbut continues as a fricative and thus probably the best term for it is 'alveolar trill fricative'

with the symbol Ir/. (The laminal diacritic IfI used in Ladefoged and Maddieson (1996) does

not capture the sound's defining property.) A voiceless allophone of hI occurs in places

where voicing assimilation applies.

There are three nasal phonemes, Iml, Inl and Ipl. A labiodental nasal may occur

allophonically, as the result of assimilation of the place of articulation with a followinglabiodental fricative. The velar nasal occurs allophonically before a velar stop, and in this

position it is obligatory. The lateral IV is always voiced and, in standard Czech, it has a clearor neutral quality in all positions in the word. However, some Czech dialects have a

phonemic contrast between III (which is often palatalized) and dark It/. The bilabial nasal

1m!, lateral III, and sonorant trill II'I can be syllabic, e.g. sedm Isedll)l 'seven'; vlk Iv\kJ

'wolf'; vrbalvrba/ 'willow'.

p pEro pero 'pen' f fakulta fakulta 'faculty'

b bota bOla 'shoe' v va:za vdza 'vase'

t lEoto tento 'this' s si:la sEla 'strength'

d du:m dum 'house' z ZIma zima 'winter'

c cdo telo 'body' S SESt fest 'six'

j jdo delo 'gun' 3 3cna lena 'woman'

k kolo kolo 'wheel' x xlEba ehleba 'bread'

9 galmjE galerie 'gallery' fi fiora hom 'mountain'

m matka matka 'mother' r ruka ruka 'hand'

n nofia nooo 'leg' r rEka reka 'river'

p ]1Etso neeD 'something' j joocn jeden 'one'

tS tscna eena 'price' 1 IEs les 'forest'

tJ tSIsti: Cistj 'clean'


Vowels

The vowel system consists of a set of five short vowels II e a 0 uJ, their long counterparts

Ii: e: a: 0: u:/, and three diphthongs, all falhng: lou!, lau! and leu! (the last two can be found

only in Joan words). The short and long vO,wels are in opposition in all positions.The quality cf short non-low vowels and their corresponding long counterparts differs,

in general, only to a small extent. However, in the pair IrI and 1i:/, the short vowel is

noticeably less close and more central than the long one. The greater difference in qualitybetween these two vowels than in the case of the other vowel pairs is reflected by the usage

here of a different vowel symbol for each of the two. The vowel 10:/ has a phonemic status

only in loan words (and therefore in the table below it is shown in brackets). The short lal is

pronounced usually as marginally fronter than the long la:/. Vowel distinctions are

preserved in unstressed positions.

I ITlIle mile 'nicely' i: mi:le mile 'mile'

e let let 'flight' e: le:t let 'summer (gen pi)'

a rat fad 'row (gen pi)' a: ra:t hid 'order'

0 voda voda 'water' (0:) go:1 gol 'goal'

u domu domu 'house (gen sg)' u: domu: domu 'home (adv)'

ou mouxa moucha 'fly (n)'

(au) auto auto 'car'

(eu) lleutra:l]1.i: neutrdln£ 'neutral'

AssimilationCzech has assimilations of place and manner of articulation, and also voicing assimilation.Both assimilation of place of articulation and of voicing are normally anticipatory. Someexamples of assimilation of place were mentioned in the discussion of consonants.Assimilation of manner of articulation is relatively rare; if it occurs, it usually concerns

combinations of It! and lsI, It! and IS!, Id! and IzI, or /d! and /:), which may become affricates

Its/,NI, /d-i/ and /cl3/ respectively (e.g. detskY can be pronounced either as [jetski:) or


(jetski:]).Voicing assimilation affects most consonants. Apart from matching voicing to the

following consonant, it also causes devoicing at the end of a word before a pause. Whensubject to voicing assimilation, the voiced glottal fricative lfiJ and voiceless velar fricative IxJ

behave like a voiced/voiceless pair. For instance, in beh Prahou /fiJ in bih is realized as [xl.

SuprasegmentalsPrimary lexical stress falls on the first syllable of a word; thus it does not have a contrastivephonological role and functions rather as an indication of the word boundary. The stressmarks in the transcription below indicate which syllables were accented by the speaker inthe recorded passage. II represents the major and I the minor intonation phrase boundary.


'severa:k a 'sluntse se 'fia:dah I gdo 'z j1rx je 'srlj1ejSi: II f tom 'spat~rlr

'pOtSestne:fio "kteri: 'kra:tSe! 'zafialen 'pla:Scem Il'ujednah tedr I 3e 'ten se rna:'pova30vat 'za srlj1ejSi:fio "gdo 'prvj1i: 'doka:3e I 'abrsr 'potsestni: 'svle:kl'pla:Sc lI'tu zatSal 'severa:k 'foukat 'ze fJi: 'sUr II ale 'tSi:m vi:tS 'foukal II ci:m'vi:tS se 'potsestni: 'zafialoval 'do sve:fio 'pla:sce II 'konetSj1e se 'severa:k 'vzdal'marne:fio 'u:srIi: II 'pak zatSalo 'sluntse 'svi:crt a 'fira:t II a 'za j1ejaki: 'okam3rk'potSestni: I 'ktere:mu brio 'fiorko " 'SXOjrl 'pla:Sc lI'tak muse! 'severa:k 'uznat I3e 'sluntse je 'srlj1ejSi: II

Orthographic versionSeven'ik a Slunce se hadali, kdo z nich je silnejsf. V tom spatFili pocestneho, ktery kraeelzahalen plastem. Ujednali tedy, ze ten se rna povazovat za silnejsfho, kdo prvni dokaze, abysi pocestny svlekl plasi. Tu zaeal Seven'ik foukat ze vsi sHy, ale eim vic foukal, tim vic sepocestny zahaloval do sveho plaste. Koneene se Severak vzdal marneho !lsilf. Pak zacaloSJunce svitit a hl'at a za nejaky okamzik pocestny, kteremu bylo horko, shodil plasi. Takmusel Seven'ik uznat, ze Siunce je silnejsf.

AcknowledgementsI would like to thank Martin Barry, Francis Nolan, Zdena Palkova, Miroslav Ptacek, Pfemysl Janota, MarieSvobodova and James Naughton for their comments and suggeslions.

ReferencesKUCERA, H. (1961). The Phonology ofCzech. The Hague: Mouton & Co.LADEFOGED, P. AND MADDIESON, I. (1996). The Sounds of the World's Languages.

Oxford: Blackwell.PALKOvA, Z. (1994). Fonetika afonologie ceStiny. Prague: Karolinum.ROMPORTL, M. (1973). Zdklady fonetiky. Prague: Statni Pedagogicke Nakladatelstvf.

I?utch

CARLOS GUSSENHOVEN

Vakgroep Engels·Amerikaan~. Katholieke Universiteit Nijmegell.Erasmusplein J. 6525 HT Nijmegen. The Netherlands

Apart from the Republic of Surinam and the Leeward islands of the Dutch Antilles, Aruba,Bonaire and Curattao, where Dutch is an official language, Dutch is spoken in the

Netherlands and the northwestern half of Belgium by about 20 million speakers. Variationin the pronunciation of educated European Dutch is substantial. For instance, the Belgianvarieties (also collectively known as Flemish) tend to have monophthongal realizations of[e:, 0:, 0:] as opposed to realizations as closing diphthongs in the Netherlands. Moregenerally, the southern varieties tend to have a full set of voiced fricatives Iv, z, vi by theside of If, s, xi - which system of voiced fricatives is reduced to just Iv, zl or just hi as onegets closer to the prestigious western part of the Netherlands ('Randstad'). (There is greatvariability in the voicing of fricatives. Low-prestige urban varieties in the west may also

lack hI.) Roughly south of a line Rotterdam-Nijmegen, which is marked by the riversRhine, Meuse and Waal, lx, vi are velar, while to the north the corresponding voicelessfricative is post-velar or uvular. The p~oneme Irl tends to be alveolar in Belgium, inAmsterdam and in the north-east of the Netherlands, but uvular elsewhere. The varietyillustrated represents Western, educated, middle-generation speech, and a careful colloquial

style. More information is given by Collins and Mees (1982), Mees and Collins (1983), andBooij (1995).

Consonants

Bilabial Labio- Alveolar Post- Palatal Velar Uvular Glottaldental alveolar

Plosive p b t d (C) k (7)

Nasal m n (J1) !]

Fricative f V S z (J) (3) x. fi

Tap r

Approximant U JLateral

1Approximant

p penb ben

'pen'

'(1) am'

t takd dak

([c] keljap

'bough''roof''soy sauce')

k kat

X gal

'cat''hole'


m mens 'human being' n nek 'neck' I) eng 'narrow'([P] oranje 'orange' adj)

fiets 'bicycle' s sok 'sock' ([JJ chef 'section head')v oven 'oven' z zeep 'soap' ([3J jury 'jury')

v wang 'cheek' jas 'coat' ([?] beamen 'confinn')Lente 'springtime' rat 'rat' fi hoed 'hat'

Ip, t, k/ are voiceless unaspirated, Ib, dl are fully voiced. Alveolars (except leI) are laminal

and Is, z/ may have only mid-to-low pitched friction. [c, p, J] are the variants of It, n, sl that

occur before Ijl. In fact, all occurrences of these sounds can be analyzed as sequences of

alveolar-plus-Ijl; by analogy, [3] could be seen as /zjl and is therefore parenthesized in the

chart above. [c, p, J, 3J are prepalatal, the tip of the tongue being held in the lower jaw.

Ivl tends to be only weakly voiced. Ivl is [v] in the onset, and [Q] in the coda. lei is [e] in

the onset, and [lJ in the coda; in careful speech, a trill may occur word-initially. [?] precedes

vowel-initial syllables within words after la:, 'JI; because of its predictable distribution it is

also shown in parentheses.Voiced obstruents and lfil do not occur in the coda; Ivl is restricted to positions between

voiced segments within the word. Marginal [g] (not listed) occurs in a small number of

loans.Sentence phonology is characterized by a number of consonantal adjustments.

Sequences of identical consonants are reduced to single consonants by a process ofdegemination. Progressive devoicing will affect fricatives after obstruents, while obstruentsmay be voiced before Ib, d/. Moreover, post-sonorant word-final fricatives, particularly lsi,

may be voiced before vowels.

VowelsDutch has a set of lax vowels, a set of tense vowels, and a reduced vowel. The first columngives the lax vowels and the reduced vowel I'JI, the second column the tense vowels.Vowels in the third column are marginal in the language, and only appear in recent loans.The nasalized vowels [e:, 0:, 5:] (not listed) also have this status. There are also three widediphthongs.

bit 'bit' biet 'beetroot' i: anaLyse 'analysis'y hut 'cabin' y fuut 'grebe' y: centrifuge 'spindryer'e bed 'bed' e: beet 'bite' e: serre 'conservatory''J 't 'the' ¢: neus 'nose' re: oeuvre 'works'


a bad 'bath' a: zaad 'seed'::> bot 'bone' 0: boot ;boat' ::>: zone 'zone'

u hoed 'hat' u: cruise 'cruise'

ei ei 'egg'aey ui 'onion'AU lOut 'salt'

Ii, y, ul are long before Ir/ in the same stress foot. Ie:, 0:, 0:1 are narrow closing diphthongs([ei, 0y, au]), except before Irl in the same word, when they are pronounced [eOl, 0Ol, Ool] .

•u(:)

:)(:)

•

o.

0:

AU

Stress and accentuationMain stress falls on the antepenult, the penult, or the final syllable of the word if the penultis open, and on the penult or the final syllable if the penult is closed. The long marginalvowels only occur in stressed syllables. In unstressed syllables, the otherwise long [e:, 0:,0:, a:] are short. Minimal pairs are rare, examples being canon ['ka:n::>n] 'canon', kanon[ka'n::>n] 'cannon'; Servisch ['servis] 'SerQian'. servies [ser'vis] 'dinner service'.

Largely depending on focus, intonational pitch accents will occur on the syllable withmain stress of some words, marked with [ , ] in the transcription. The realization of thesepitch accents will extend as far as the next pitch accent or the single bar I, which marks off aminor intonational phrase. Utterance-final boundary tones may be added before the double

bar II.

Transcription of recorded passaged;} 'no:rd;}omt en d;} 'z:)n I fiod;} ;}n dIS'kYsi o:v;}r d;} 'fra:x I 'oi fon fiyn 'rue:;}d;} 'st£r;}kst;} oos I tun ;}r 'jceyst imont fo:r'bei koom I di ;}n 'dIk;} 'oorm;} 'josa:nfiot II Z;} spra:k;} 'of I dot oi d;} fo:rbeiXOlpr d;}rtu ZAU 'kreiX;} Z;}Jl 'jos cey t;}trek;} I d;} 'ster;}kst;} ZAU zein II d;} 'no:rd;}omt b;}x:)n ceyt 'al;} mox t;} 'bla:Z;} IIma:r u 'fiord;}r i 'blis Ides t;} 'dIxt;}r d;} fo:rbeixol);}r Z;}p joz :)m ZIX 'fie:n tr:)kII t;}n'sbt;} xof t;} no:rd;}omt ;}t ma:r ':)p II f;}r'blx;}nz ~x:)n d;} 'z:)n I 'kroXt;}Xt;} 'stra:l;} I en :)'mId;}l;}k da:'r:)p I tr:)k t;} fo:rbeixol);}r Z;}P 'joz ceyt II d;}'no:rd;}omt bn tun slexs ~'?a:m;} Idot;} 'z:)n d;} 'ster~kst~ oos.

I/lustrations of the IPA 77

Orthographic versionDe noordenwind en de zon hadden een discussie over de vraag wie van hun tweeen desterkste was, toen er juist iemand voorbij kwam die een dikke, warme jas aanhad. Zespraken af dat wie de voorbijganger ertoe zou krijgen zijn jas uit te trekken de sterkste zouzijn. De noordenwind begon uit aile macht te blazen, maar hoe harder hij blies, des tedichter de voorbijganger zijn jas om zich heen trok. Tenslotte gaf de noordenwind het maarop. Vervolgens begon de zon krachtig te stralen, en onmiddellijk daarop trok de voorbijganger zijn jas uit. De noordenwind kon toen slechts beamen dat de zan de sterkste was.

ReferencesBOOu, E. E. (1995). The Phonology of Dutch. Oxford University Press.COLLINS, B. AND MEES, I. (1982). A phonetic description of the consonant system of

Standard Dutch. Journal of the International Phonetic Association 12,2-12.

MEES, I. AND COLLINS, B. (1983). A phonetic description of the vowel system ofStandard Dutch. Journal of the International Phonetic Association 13,64-75.

French

CECILE FOUGERON*,AND CAROLINE L. SMITH**

Phonetics Laboratory, UCLA, 405 Hilgard Avenue, Las Angeles, CA 90095·1543, USA

and

*Institut de Phonbique, CNRS URA 1027, Universite Paris 11/, France

"Eloquent Technology Inc, Ithaca, NY, USA

The following description of French is based on the speech of a young Parisian femalespeaker. Varieties of French have almost identical inventories; the main differences are tobe found in the maintenance or loss of certain contrasts.

VowelsOral vowels. French vowels, which are all monophthongs, are normally described usingfour values for height and two or possibly three (front, central, back) for backness. Allback vowels are rounded; there are both rounded and unrounded sets of front vowels.

The mid-high and mid-low vowels have limited overlap in their distributions, butcannot be regarded as allophones. The contrast between [e] and [e] is found in opensyllables in final position (e.g. [se] ses (3rd person possessive, plural) vs. [se] sait 'knows'and [pwajle] poignet 'wrist' vs. [pwajle] poignee 'handful'). Elsewhere, [e] occurs in

open syllables and [e] in closed syllables: compare [se.vir] sevir 'to rage' vs. [sell'.vill']servir 'to serve'. For the other pairs of mid vowels, the contrast between mid-high andmid-low is limited for the most part to closed monosyllables. In other environments, the

mid-low vowels roe] and [::>] occur in closed syllables and the mid-high vowels [0] and [0]in open syllables. However, in addition to this general rule, the nature of the syllable-finalconsonant also has an effect: the mid-high vowels [0] and [0] are found in syllables ending

si si 'if' u su sous 'under'

e se ses 'his, hers' (pi) 0 so sot 'silly'

e se sait 'knows' ::> S::>ll' sort 'fate'

sel! serre 'greenhouse'

y sy su 'known'

0 S0 ceux 'these'

oe soel! soeur 'sister'

;) s;) ce 'this'

a sa sa 'his, hers' (I)


in [z], and [0] is not found even in monosyllables ending in [ff], lfl], and [9] (Delattre1966; Leon 1992). It is noticea.ble that our speaker often produces a markedly centralizedallophone of [:>], especially before [ff]. This is transcribed [5] in the passage below. Somespeakers have a contrast between two low vowels [a] and [0]; our speaker has only onelow vowel, [a], which is central. Schwa, [~], is also a central vowel, with some rounding.

Some speakers retain a contrast between long and short vowels in a very few words,but most speakers no longer maintain any length contrast.

Nasalized vowels. Traditionally, French is described as having four distinctivelynasalized vowels, [e), [0], [5], and [eel. However, our speaker, like many others, lacks

[ee], which is replaced by [£].The vowel [e] is produced with tongue and lip position very similar to its oral

counterpart [el; however, several articulatory studies (e.g. Zerling 1984; Lonchamp 1988)have suggested that [0] and [5] differ substantially from their supposed oral counterparts[a] and [::>]. The position of the tongue is similar in [0] and [5], and the main articulatorydifference is that [5] has a greater degree of lip rounding. These authors suggest that othersymbols might be more appropriate for transcribing these two vowels, but we have

retained the traditional usage as none of the alternative proposals has gained widespread

acceptance.

0 so sans 'without'

5 s5 son 'his, hers' (m sg)

8 S8 saint 'saint'

Consonants

Bilabial Labio- Delllal Palato- Palatal Velar Uvulardental alveolar

Plosive p b t d k 9Nasal m n J1 (I))

Fricative f V S z I 3 ff

LateralIApproximant

Palatal Labial- Palatal Labial- Velar

Central Approximanls J q w

p pu pou 'louse' fu fou 'crazy'b bu boue 'mud' v vu vous 'you' (pI)

t tu tout 'all' su sous 'under'd du dOtlX 'sweet' z zo zoo 'zoo'


k ku cou 'neck' J Ju chou 'cabbage'

g gu gout 'taste' 3 3u joue 'cheek'

Il' Il'U roue 'wheel'

m mu mou 'soft' lu loup 'wolf'

n nu nous 'we, us'

J1 aJ10 agneau 'lamb' w sw£ soin 'care'

l) pall'kil) parking 'parking lot' j sj£ sien 'his, hers'

4 s4£ suint 'grease on

sheep's wool'

French voiced stops are typically fully voiced throughout; voiceless ones are described as

unaspirated. When preceding high vowels, they are often followed by a short period of

aspiration and/or frication. The velar nasal occurs only in final position in borrowed(usually English) words.

French has one rhotic, whose production varies considerably among speakers and

phonetic contexts. The speaker presented here uses a uvular fricative [Il'] that is sometimes

reduced to an approximant [~], particularly in final position; it may also be devoiced (for

examples see the transcribed text), and can be reduced to zero in some word-final

positions. For other speakers, a uvular trill [R] is also fairly common, and an apical trill [r]

occurs in some dialects. Vowels are often'lengthened before this segment.

The approximants [w], [4], and [j] each correspond to a high vowel, [u], [y). and [i)

respectively. There are a few minimal pairs where the approximant and corresponding

vowel contrast, but there are also many cases where they are in free variation. Contrasts

between fj) and [i) occur chiefly in final position, as in [abej) abeille 'bee' vs. [abei)

abbaye 'monastery'.

Word-final consonants are always released, and in keeping with a general favouring of

open syllables, they are usually resyllabified as onsets when followed by a vowel-initial

word (enchafnement). Underlying word-final consonants that are not pronounced before a

consonant, are pronounced only when preceding a vowel in the same rhythmic group.

This process, known as liaison, also contributes to this canonical open-syllable pattern.

ProsodyAlthough French is often described as having stress on word-final syllables, in connected

speech this is pre-empted by the accent on the final syllable of a group of words (sense

group or accentual group, see Vaissiere 1992 for discussion).

Transcription of recorded passageThe transcriptional style adopted in this illustration is a relatively narrow one, which

reflects the particular pronunciation used in the recording of the passage made for the

illustration.


1a biz e 1;:) s::>lei s;:) dispyte II Jake aSyffa kilete 1;:) ply f5~ II kat ilz5 vy evwaja3ce ki savose II avlope do s5 mate Iii: s:5 t:5be dab~ b s;:)lqi kiaffivffe l;:)p~;:)mje a 1;:) lqi feffote II S;:)ff;:) ff;:)gaffde bm 1;:) ply f5~ II a15~ labiz s~ miz a sufle d;:) tut se f5'1S II me ply d sufle ply 1;:) vwaja3ce~ seffe s:5mototllIf d;:) 1qi II fina1mo d ff;:)nosa 1;:) lqi feffote II a15~ 1;:) s::>lei bmosab~ije II e 0 bu d~ m::>mo 1;:) vwaja3ce ffeJofe ota s5 mota II esi la biz dyff;:)bnet b 1;:) s::>l£i ete 1;:) ply f5~.

Orthographic versionLa bise et Ie soleiI se disputaient, chacun assurant qu'iI etait Ie plus fort. Quand ils ont vuun voyageur qui s'avan9ait, enveloppe dans son manteau, ils sont tombes d'accord quecelui qui arriverait Ie premier aIe lui faire ater serait regarde comme Ie plus fort. Alors, Iabise s'est mise a souffler de toutes ses forces, mais plus elle soufflait, plus Ie voyageurserrait son manteau autour de lui. Finalement, elle renon9a a Ie lui faire otero Alors, Iesolei! commen9a a briller et au bout d'un moment Ie voyageur, rechauffe, ata sonmanteau. Ainsi, la bise dut reconnaitre que Ie solei! etait Ie plus fort.

AcknowledgementThe first author was supported by an Allocation de recherche M.R. T. awarded to the D.E.A. de phonelique

de Paris.

ReferencesDELATIRE, P. (1966). Studies in French and Comparative Phonetics. The Hague:

Mouton.LEON, P. (1992). Phonetisme et prononciations dufranfais. Paris: Nathan.LONCHAMP, F. (1988). Etude sur la production ella perception de la parole, les indices

acoustiques de la nasalite vocalique, la modification du timbre par lafrequencefondamenlale. These de Doctorat d'Etat, Universite de Nancy II.

VAISSIERE, J. (1992). Rhythm, accentuation and final lengthening in French. In Sundberg,J., Nord, L. and Carlson, R. (editors), Music, Language, Speech and Brain(Stockholm: Wenner-Gren International Symposium Series 59), 108-20.

ZERLING, J-P. (1984). Phenomenes de nasalite et de nasalisation vocaliques: etudecineradiographique pour deux locuteurs. Travaux de l'lnstitut de Phonetique deStrasbourg 16, 241-66.

GalicianI

XOSE L. REGUElRA

lnstituto da Lingua Galega and Dept. de Filoloxfa Galega

University ofSantiago de' Compostela, J5703 Santiago, Spain

Galician is a Romance language closely related to Portuguese. It is spoken in the farnorthwestern part of the Iberian Peninsula. As a result of historical circumstances, Spanishhas exerted a strong political and cultural influence over this region since the end of theMiddle Ages, and there is in consequence in Galicia today a socially ascendant Spanishspeaking urban minority, while the rest of the population maintains Galician as its mothertongue. As a result, Galician is in a situation of language shift and the spoken language nowincludes numerous Joan-words from Spanish. However, in recent decades Galician hasbeen recognized as an official language in Galicia along with Spanish, and the language hasgone through a process of standardization and social recovery.

The language variety described here is that of colloquial Galician as spoken by a middleaged male speaker whose speech may be considered representative of an informal varietyused by educated urban speakers.

Consonants

Bilabial Labio- Dental Alveolar Post- Palatal Velardental alveolar

Plosive p b t d j k 9Affricate tJNasal m n J1 l)

Trill r

Tap f

Fricative f e § JApproximant j W

Lateral Approximant 1

Some varieties of the language, particularly among elderly speakers, display two contrastinglateral consonants, one alveolar [I] and the other palatal [,(]. However, in the language ofmost speakers, especially in urban and younger speakers, the palatal lateral consonant hasbeen replaced by a voiced palatal plosive [j], as can be seen in the table above.

In the dialects of western Galician, instead of 181 and I§I there is only one fricativeconsonant, usually realized as a voiceless lamino-alveolar fricative [s]. This phenomenon is

/llustrations of the IPA 83

known as seseo (Fernandez Rei 1991: 189-215). Another dialectal characteristic which isencountered chiefly in the western half of the Galician-speaking region is known as gheada;this dialectal feature consists of the absence of the voiceless velar plosive /g/, which isreplaced by a voiceless fricative with several possible realizations, including pharyngeal [h],uvular [X], velar [x] and glottal [h]. Standard Galician, including the variety ofpronunciation described here, does not show these variants.

p 'papo papo 'jowl' I '1a xa 'already'

b 'bimbjo vimbio 'wicker' m 'mimo mimo 'cuddle' (n)

'tiI].ta tinta 'ink' n 'neno neno 'child'

d 'doI].do dondo 'soft' ]1 'a]1o aii.o 'Iamb'

] 'a]o allo 'garlic' l) 'ul)a unha 'a', 'one'

k 'kuko cueo 'cuckoo' ko'ree correr 'to run'

9 'gOl)go ganga 'bargain' 'oweo Duro 'gold'

tJ 'tJutJo chucho 'kiss' 'lil)gwo lingua 'language'

f 'bfo fofo 'flabby' 'baja vaia 'goes' (subj)

8 '8iI].8a cinza 'ash' w 'bow vou 'r go'

§ '§i§o siso 'good sense'

The approximants fj) and [w] may appear after a vowel, giving rise to sequences such as[ew], [ow], [ej], [oj], etc. While in most phonetic descriptions of Galician these are labelleddiphthongs, in the present description they are viewed as sequences of vowel plusconsonant, rather than as sequences of two vowels.

Vowels

•

Nasalization is not distinctive for vowels, but any vowel can become fully or partiallynasalized in contact with a nasal. The effect is strongest when a vowel occurs between twonasal consonants, and nasalization is systematically indicated for this sequence of sounds inthe examples in this article. (In the recorded passage, nasalization is indicated wherever thespeaker produced a nasalized vowel, i.e. in more than just the NVN context.)

84 Handbook of the {PA

i 'bir vir 'to come~ :l 'n:l§ nos lwe'

e 'ter ter 'to have' 0 'koro corro 'I run'

e 'bero berro 'shout' (il> u 'duro duro 'hard'a 'pa [Xl 'shovel' '

StressStress in Galician may fall on any of the last three syllables of a word. If a word has aclosed syllable in final position, stress usually falls on that syllable, e.g. final [fi'nal] 'end',roncon [rol]'kol]] 'pipe in a bagpipe', armar [ar'mar] 'to arm'. In certain loans and learnedwords stress may be on the penultimate syllable, as in revolver [re'~ol~er] 'revolver',dolmen ['d:llmel]] 'dolmen', uti! ['uti I] 'useful'. If there is a final open syllable in a word,stress will normally fall on the penultimate syllable, e.g. casa ['ko§a] 'house', home [':lme]'man', carballo [kQ['~ajo] 'oak', edificio [eQi'fi6jo] 'building'. However, some words endin a stressed vowel, e.g. avo [a'~o] 'grandfather', ala [a'la] 'over there'. This is true of thethird person singular of the future tense in all verbs: cantara [kOl}ta'ra] 'will sing', lera[Ie'ra] 'will read', vira [bi'ra] 'will come'. There are in consequence some minimal pairswhere stress placement is contrastive, e.g. rev6lver [re'~ol~er] 'revolver' vs. revolver[re~ol'~er] 'to mix'; cantara [kal].'tara] 's~ng' (subj) vs. cantara [kaI)ta'ra] 'will sing'.

Some words ending in two open syllables bear the stress on the antepenultimatesyllable, e.g. cobado ['ko~aQo] 'elbow', 16strego ['lo§treUjo] 'lightning', medico ['mel}iko]'doctor', 16xico ['bfiko] 'logical' , calculo ['kotkulo] 'calculation'.

ConventionsThe consonants represented in the table above as Ib d gl are only realized as a plosive after apause or after a nasal consonant (and also after a lateral in the case of (d]), whereas in otherpositions they have approximant variants [~ I} Uj] respectively. Thus in vecUio [be'6ipo]'neighbour', [urn be'6ijlo] 'a neighbour' a plosive is pronounced, while an approximant is

found in 0 vecifi.o [0 ~e'6iJlo] 'the neighbour', mal vecino [mal ~e'6iJlo] 'bad neighbour',ter vecinos ['ter ~e'6iJlo§] 'to have neighbours'.

Place of articulation in syllable-final nasals is non-contrastive, and assimilates to that ofa following syllable-initial consonant, e.g. inferno [ill)'ferno] 'hell', enredar (enre'Qar] 'toentangle', ancho ['aniij'0] 'broad, wide' , nunca ['niil]ko] 'never'; before a pause or before aword beginning with a vowel the realization is always velar [1]]: non [no!)] 'no, not', son unhome ['§Ol] iiI] '5me] 'I am a man'. The velar nasal also occurs intervocalically in a smallnumber of words, for instance unha ['ul]o] 'a, one', algunha [at'Ujul]o] 'some' (fern),ningunha (nl!)'gu!)a] 'no, none', where the place of articulation is distinctive: unha ['ul]a] 'a,one' vs. una ['una] 'joins' (subj), una ['uJla] 'nail' (Alvarez, Regueira and Monteagudo1995: 33). Nevertheless, some authors do not include [1]] among the phonemes of Galician,suggesting that in cases such as those above the velar place of articulation is conditioned bythe location of the syllable break: ['ul].a] vs. ['u.na] (Veiga 1976: J05-7; Castro 1989: 14455). The present description prefers to regard II]I as a distinct phoneme.


As with the nasals, the place of articulation of the lateral approximant in syllable-finalpositions is also affected by a following dental, palatal or velar consonant: alto ('alto] 'tall',colcha ['koljija] 'bedspread', algo ['ahljo] 'something'. Elsewhere the lateral approximant isarticulated as an alveolar [I).

Fricatives 181 and I§I may become partially or fully voiced in syllable-final positionbefore a voiced consonant: dez meses ['de~ 'me§e§] - ['deo 'me§e§] 'ten months', tresmeses ['tre?- 'me§e§] 'three months'. Elsewhere they remain voiceless. The open centralvowel Ial is realized as an open back [a] in contact with velar consonants.

Transcription of recorded passageo '~el)to ~o 'n::lrte e majlo '§oll poffi'a~al) §o~re ka] 'dele§ 'era 0 maj§ '£::JrteI kOt)do ka'Orow ~e pa'§ar urn bia'Iejro m'b::l]to nUl)a 'ampla 'kapa IIkombi'j1efol) elJ ke 0 ke 'al)te§ kOIJ§e'tqi§e fa'8erJe ki'tala 'kapa ::l ~ia'Iejro

§e'ria kOIJ§i~era~o: maj§ 'f:)rte II 0 '~eQto ~o 'n::lrte §o'pcow kOIJ graIl) 'furja Ie 'kaQto maj§ §o'pra~a maj§ §e mbol'~i ::l ~ia'Iejco na '§ua 'kapa IIfinal'meQte I 0 '~el)to ~o 'n::lrte a1}aQdo'now 0 '§ew em'pej1o II eQ'tolJ 0 '§olket)'tow kOIl) 'for8a I e IlJme~jata'met)te 0 1}ia'Iejro §a'kow a 'kapa II e~a'kela I 0 '1}et)to ~o 'mrte 'ti~o ke rekoj1e'8ela §uperjori'~a~e ~o '§ol II

Orthographic versiono vento do norte e mailo sol porfiaban sobre cal deles era 0 mais forte, cando cadrou de

pasar un viaxeiro envolto nunha ampla capa. Conviiieron en que 0 que antes conseguise

facerlle quita-Ia capa 6 viaxeiro serfa considerado 0 mais forte. 0 vento do norte soprou congran furia, e canto mais sopraba mais se envolvfa 0 viaxeiro na sua capa; finalmente 0 vento

do norte abandonou 0 seu empeiio. Ent6n 0 sol quentou con forza e inmediatamente 0

viaxeiro sacou a capa. E daquela 0 vento do norte tivo que recofiece-Ia superioridade do sol.

AcknowledgementsI should like to thank Xose Ram6n Varela for his initiative in this description and for his help and

comments. I am also grateful to him and to John Barlow for the translation of this text.

ReferencesALVAREZ, R., REGUElRA, X. L. AND MONTEAGUDO, H. (1995). Gramcitica galega, 6th

edition. Vigo: Galaxia.

CASTRO, O. (1989). Aproximacion a la fonolog{a y morfolog{a gallegas. Dissertation,Georgetown University. Ann Arbor: University Microfilms International, 1991.

FERNANDEZ REI, F. (1991). Dialectolox{a da lingua galega. 2nd edition. Vigo: Xerais.VEIGA, A. (1976). Fonolog{a gallega. Valencia: Bello.

German

KLAUS KOHLER

Institul flir Pholletik, Olshausenstraj3e 40, D-2300 Kiel, Germany

The style of speech illustrated is that of many educated Germans in the North. Theaccompanying recording is of a 62-year-old speaker reading in a colloquial style.

Consonants

Bilabial Labio- Dental IlAlveolarl Post- Palatal Velar Uvular Glottaldental alveolar

Plosive p b t d k 9 ?

Nasal m n 1)

Fricative f V I s zls 3 ~ X 'ff h


1Approximant

The table of consonants lists phonemes with the exception of [c;:) and [X] as well as [7].Their distributions can be predicted from context, as stated below under conventions,provided morpheme boundaries are marked. Frauchen 'little woman' and rauchen 'tosmoke' are differentiated as ['flJ'au~n] and ['lJ'aux~n] because in the former [c;:) is initial in

the diminutive suffix -chen, but in the latter [X) is final after a back vowel in the stemmorpheme rauch-. So [c;:) and [X) can be said to be allophones of a phoneme Ixl with

reference to different positions and contexts within morphemes. In the absence of suchmorphemic information, as is usually the case in IPA transcription, the difference betweenthese consonant segments has to be symbolized. Similarly [fe'7alz~n) vereisen 'to freeze

over' and [fe'lJ'alz~n) verreisen 'to travel' show a paradigmatic opposition of [7) and [IJ') inthe same segmental context. If the morpheme structure of the stems -eis- vs. -reis- after theprefix ver- is taken into consideration the occurrence of [7) can be automatically inferred; ifnot, [7] needs to be transcribed.

p passe 'skip' (I sg) t Tasse 'cup' k Kasse 'cash desk'

b Raf3 'bass' d das 'that' g Gasse 'lane'

m Masse 'mass' n nasse 'wet' I) lange 'long'

fasse 'catch' (I sg) s reif3e 'rip' (I sg) f rasche 'quick'

v Wasser 'water' z reise 'travel' (I sg) 3 Garage 'garage'

!LLustrations of the IPA 87

If dichj ja

Vowels

'you''yes'

X DachK Rasse

'roof''race'

h hasseI lasse

'hate' (J 5g)

'let' (I 5g)

Monophthongs

a, a:

Diphthongs

'had' (I pi)

'requested' (I pI)

'booty' (5g)

bietenbitten

e beten

e Betten

e: Mten

'to offer''to request''to pray''beds''if they requested'

y hiUeny Hiitten

~ Goethe

re Gotter

a hatten

a: baten

~ Beuti!.

'to guard''huts'(name)'gods'

u sputenu Butter

o boten;) Botten

'to hurry''butter''offered' (I pi)

'clogs'

ill beiden 'both' (\ pI) ;)1 Beuten 'booty' (pi) au bauten 'built'

Stress, (primary stress) and, (secondary stress), as in compounds, e.g. ['JoBnJtain,fegB]Schornsteinfeger 'chimney sweep'.

Conventions/p, t, k/ are aspirated when not preceded by a fricative within the same word (e.g. ['ftat]Stadt 'town'), nor followed by a syllabic nasal (e.g. ['Ialt':l] leiten 'to guide'); the aspirationis strongest before a stressed vowel, weakest in unstressed function words.

/15/ can be an approximant intervocalically (e.g. Herren 'gentlemen'); after voicelessplosives and fricatives, especially those within the same word, it is devoiced (in e.g. trat

'kicked' it is completely voiceless [xl); postvocalically before a consonant or word-finally itis vocalized to [B), which results in diphthongs (e.g. ('haBt] hart 'hard', ['OB] Ohr 'ear', see


the vowel chart below); the ending -er is realized as [BJ (e.g. [,butBJ Butter 'butter'); the

place of articulation of the consonant va,ries from uvular in e.g. rot 'red' to velar in e.g.treten 'kick', depending on back or front vowel contexts.

in

a(:)n

[B] Diphthongs

Except for unstressed function words, word and stem initial vowels are prefixed by glottalstops (e.g. [?eB'?aBbaltn] erarbeiten 'achieve through work').

[c;:J occurs after front vowels and consonants within the same morpheme, as well asmorpheme initially (e.g. [ye'mi] Chemie 'chemistry', [<?nJ -chen = diminutive suffix). [xJoccurs after low back vowels (e.g. ['baxJ Bach 'brook', ['d:)XJ doch 'yet'); [x] is usedinstead of [X] after high and mid back tense vowels (e.g. ['buxJ Buch 'book', ['hox] hoch'high').

The closer vowels in a close/open pair are always longer under the same conditions ofstress and environment; in unstressed position the close vowels are shortened and may keeptheir close vowel quality, as in [mo'l!a:IJ Moral 'moral', [fy'zikJ Physik 'physics', but thelatter constraint does not apply to unstressed function words (e.g. [YbBJ aber 'over').

Transcription of recorded passageams 'Juntn zrr; 'n:)ntv1Ot un 'z:)n;}. ven f;}n im 'bardn vol dn 'Jtenk;}l{;} vel{;},als am 'vand;}l{n, deB m am 'vaBm 'mantI g;},hYlt vaB, d;}s 'veg;}sda'heBka:m. zr vundn 'arnrr;, das 'deBjenrg;} fYB d;}n 'JteBk;}B';}n ,geltn z:)lt;},deB d;}n 'Vand;}B'B 'tsvrIJI) vynd;}, zarm 'manti 'aptsu,nemm. deB 'n:)Btvrm 'blismrt 'alB 'maxt, abB je 'meB eB 'blis, desto 'festn 'hylt;} zrr; dn 'vand;}B'n 10

zarm 'mantI am. 'enthr; ga:p dB 'n:)Btvrn d;}I) 'kampf 'auf. nun eB'veBmt;} dr'z:)n;} dr 'Iufp mrt inn 'fB':)rnthr;n 'ftrm:ln, un fonax 'venrgIJ 'augIJ,bIrkI] tsok dn'Vand;}B'B zarm 'manti aus. da must;} dB 'n:)Btvrn 'tsugebm, das dr 'z:)n;} f;}nim 'bardn dB 'JteBk;}l{;} vaB.


Orthographic versionEinst stritten sich Nordwind und Sonne, wer von ihnen beiden wohl der Stlirkere ware, alsein Wanderer, der in einen warmen Mantel gehUJlt war, des Weges daherkam. Sie wurdeneinig, da~ derjenige fUr den Stlirkeren gelten soJlte, der den Wanderer zwingen wUrde,seinen Mantel abzunehmen. Der Nordwind blies mit aller Macht, aber je mehr er blies, destofester hUJlte sich der Wanderer in seinen Mantel ein. Endlich gab der Nordwind den Kampfauf. Nun erwlirmte die Sonne die Luft mit ihren freundlichen Strahlen, und schon nachwenigen Augenblicken zog der Wanderer seinen Mantel aus. Da mu~te der Nordwindzugeben, da~ die Sonne von ihnen beiden der Stlirkere war.

HausaI

RUSSELLG. SCHUH AND LAWAND. YALWA

Department of Lillguistics, UCLA, 405 Hilgard Avellue, Los Allgeles, CA 90095-1543, USAI

The following description of Hausa is based on the variety of the language spoken in Kano,Nigeria. The sample text is transcribed from a recording of a male native of Kano in his latethirties. This variety of Hausa is considered 'standard'. Though Kano is a large urbancenter with some internal variation in speech, the sound inventory is relativelyhomogeneous within the city and surrounding area. Kano Hausa is the variety mostcommonly heard on national and regional radio and television broadcasts in Nigeria as wellas most international broadcasting, such 'as the BBC, Deutsche Welle, Radio Moscow, andVoice of America. Kano Hausa is therefore familiar throughout the Hausa speaking areas ofNigeria as well as Hausa speaking communities in Niger, Ghana, and other areas outsidenorthern Nigeria. Hausa has a standard orthography, in use since the 1930' s and also basedon the Kano variety. It is familiar to all Hausa speakers literate in the Romanizedorthography. (Many Hausas are also literate in Arabic orthography, a variety of which hasbeen used to write Hausa, probably for several centuries. The Arabic orthography forHausa is less standardized than the Roman orthography and has little formally publishedIiterature.)

VowelsPhonologically, Hausa has a IO-vowel system comprising five vowels, each with a longand a short counterpart. The five long vowels occupy roughly the five canonical vowelpositions [i, e, a, 0, u]. Their pronunciation is relatively stable regardless of environment,and they are consistently longer in duration than their short counterparts. Before pause, thefive short vowels also fall roughly into the respective canonical positions, but medially(both word-medial and word-final when not followed by pause), they are stronglyinfluenced by environment, including both preceding and following consonants and vowelsin contiguous syllables. Short iii may thus range across [i - I - i] and short lui may rangeacross [u - U - H]. In normal conversational speech, medial short high vowels are probablyfrequently neutralized to a high, centralized vowel, with rounding or lack of roundingdetermined by environment. The placement of the short high vowels in the vowel chartbelow is a compromise for their variant pronunciations. Short Ia! is a fairly low,centra!(ized) vowel, with fronting, backing, and rounding determined by environment. Inthe sample text below, short Ia! has been consistently transcribed as [:l], but it should beunderstood that this represents a range of pronunciation in the low to mid area. Medial shortlei and 101 are neutralized with short Ia!. Thus, the words [zo:be:] 'ring' and [re:Je:]'branch' in their plural forms (formed by lengthening the second consonant and changingthe final vowel to [a:]) are, respectively, [z~bba:] and [r~ssa:]. See Parsons (1970) for a

IlLustrations of the lPA 91

discussion of vowel variability in Hausa. See notes introducing the illustrative text belowfor the marking of tone. Orthographic forms are in italics:

I: ki'i:ra : Kira 'forging' u:•e: ki'e:ra: Kera 'to forge'• ua: k'a:ra: Ram 'to increase'

0: kWo:ra: kora 'to chase' •u: kW'u:ra : Kura 'dust' 0:, 0

kiira: kira 'calling'

e ta:re tare 'together'

a kara : kara 'stalk'

0 gWo:ro goro 'kola nut'

u kW'ura: Kura 'to stare'

Consonants

Bilabial Alveolar Post- Palatal Palatalized Velar Labialized Glottalalveolar Velar Velar

Plosive & b t d tJ d3 ki gi k 9 kW gW ?AffricateImplosive &

ki'Ejective Stop 6 ts' cf (tJ') j' k' kW'

& Affricate

Nasal m n

Fricative <I> s z J h

Taprrrillr[

Approximant W JLateral

1Approximant

Plosives and affricates: Word-initial voiceless plosives and affricates are moderatelyaspirated. We have no information on voicing onset time for the voiced counterparts.Among the velars, the plain/palatalized/labialized distinction is neutralized to labializedbefore rounded vowels, e.g. [kwo:ra:] kora 'ringworm' but no contrasting *[ko:ra:] or*[kio:ra:]. The plain/palatalized distinction is neutralized to palatalized before front vowels,e.g. [kii:fJ.:] kishi 'jealousy' but no contrastive *[ki:fl:]' Labialized velars contrast with theplain/palatalized counterparts before front vowels, e.g. [kwi:51:] 'side of body'. The glottalstop contrasts with other consonants medially. It also occurs predictably before wordswritten in the standard orthography with initial vowels, e.g. [?a:ja:] aya 'verse of the


Koran', and as a terminator of short vowels before pause, e.g. [gWo:ro?] goro 'kola nut'.See Carnochan (1952) for a discussion of:these and other glottal phenomena.

t ta:ra : 'to gather' k ka:ra: 'to screen off'b ba:ra 'last year' d da:ra : 'to laugh' 9 ga:ra: 'wedding gifts'

ts' ts'a:ra : 'to arrange' k' k'a:ra : 'to increase'5 5a:[<1: 'to peel' cf cfa:na: 'to measure'm ma:ra: 'belly' n na:ma: 'meat'<jl <jla:ra : 'to begin' s sa:ra: 'to excavate'

z za:na: 'to draw'ra:6a: 'to flog'

r ra :na: lsun'

1 la:ka: 'mud'

tI tIa:ra : 'crowing' kj kja:wu: 'goodness'

d3 d3a:r): 'assets' gi gja:ra: 'to repair'

(tf tf a:da: 'expensive') kj' kj'a:ra 'grass' (sp.)

I Ia:ra: 'sweeping'w wa:ra: 'to select out' j ja.:ra : 'children'

j' j'a:j'a: 'offspring'kW kwa:ra: 'to pour much'gW gWa:no: 'stink-ant'kW' kW'a:ra: 'shea-nuts'? sa:?a: 'time'h ha:d3a: 'goods'

Implosive and ejective stops and affricates: The sounds [5, 6) have glottal vibration.Ladefoged (1968: 16) transcribes these sounds as [?b, d?] respectively and notes that 'thesesounds may be incidentally implosive on some occasions; but they are always distinguishedfrom their voiced counterparts by being laryngealized'. Unpublished instrumental studieswhich we have done showed consistent and strong implosion with these consonants. Onecan therefore only conclude that the production of these consonants varies depending onspeaker and probably rate of speech. The sound represented here as [ts'] ranges from anejective alveolar affricate with clear plosive component to an ejective fricative [s'] - in theKano dialect, it tends to be realized as the affricate. Some dialects, though not that of Kano,have a post-alveolar ejective affricate [tf'] in contrast with the alveolar affricate or fricative.The velar ejectives [k', k j ', kW'] show the same distribution with respect to followingvowels as do the pulmonic velars discussed in the preceding paragraph. Finally, a broadrange of dialects, including that of Kano, have a sound described here as a palatalized glottalstop, [?j). Historically, this represents a contraction of the sequence [6ijJ, as can be seen indialects which preserve the original pronunciation, e.g. Sokoto dialect [cflja:] 'daughter',Kano dialect [j'a:] (orthographic d'iya and 'ya respectively).


Nasals: Hausa has only two contrastive nasals, 1m, n/. Phonetically there are also[]1, 1]]. The palatal nasal []1] is orthographically ny, which represents the correct phonemicanalysis, shown, for example, by the fact that in a word like hanya 'road', the phonemicsyllable boundary falls between nand y (cf. plural hanyoyi, where only the y isreduplicated, not ny as would be expected in this plural type if ny represented a unitphoneme). The velar nasal [1]] is the predictable form of Inl before velar and glottalconsonants and the labial-velar approximant [w], e.g. [s;;ll]k'o:] 'baldness', [s;;ll]ho:) a typeof basket, [bl]wa:] 'potash' (orthographic sanKo, sanho, kanwa respectively). The velarnasal is also a frequent variant (probably universal for some speakers) of any nasal beforepause, e.g. [w;;lnn;}l)] 'this one', [mutul]] 'person' (orthographic wannan, mutumrespectively).

Fricatives: In the Kano and many other dialects, the only voiceless labial consonantphoneme is phonetically a bilabial fricative [4>], represented as f in the orthography.However, there is much dialectal and individual variation in the pronunciation of thisphoneme. Some speakers tend to pronounce it as a bilabial plosive [p]. A labiodental variantof the fricative, [f]. as well as bilabial variants of varying degrees of constriction are alsoheard. Some dialects, particularly in the north and west, have no voiceless labial consonantat all, instead having plain glottal fricative before front vowels, e.g. [hi) 'to exceed' (cf.Kano [4>i, fi, pi]) and a labialized glottal [hW

) elsewhere, e.g. [hwa:ra:] 'to begin' (cf. theKano example in the list above).

Tap/trill: Most Hausa speakers distinguish two tap sounds in the alveolar region, [r, r].The first is the 'canonical trilled r' sound, found in many of the world's languages. It can berealized minimally as a single tap, but often has multiple taps, especially word-initially orfinally, and obligatorily when geminated for morphological purposes. The [r] is a retroflexflap, realized by flipping the tongue forward across the alveolar ridge. Newman (1980)provides the most extensive discussion of the status, distribution, and history of thesesounds in Hausa. Ladefoged (1968: 30) was the first to investigate the soundsinstrumentally, and though he notes the distinct articulations, he states, 'Indeed I have notbeen able to find any consistent acoustic difference between the two sounds'. The two areclearly contrastive as shown by the oft-cited though far from unique minimal pair, [b;;>ra:]'begging' vs. [b;;>ra:] 'servant' (both orthographic bara), but in the one tap mode illustratedby a pair such as this, they are often difficult to distinguish. However, when they arelengthened through one of several morphological processes, the long variants share littlearticulatory or acoustic similarity. In pairs such as [ja: r;;lrrobke: Ji] 'he flogged him' vs.[Ja:[;;>[[e:] 'swept' (orthographic ya rarrabke shi and shararre respectively), the long [rr] isan alveolar trill whereas the long [rr] is a prolonged retroflex approximant (see Ladefogedand Maddieson (1996) for instrumental data and discussion). It should be noted that in thesample text below, the distinction between the two r's is not indicated. The speaker for thistext is among the minority of Hausa speakers who have only the single r sound, [r].


Transcription of recorded passageThe transcriptional style adopted in this illustration is a relatively narrow one, which reflectsthe particular pronunciation used in the r~cording of the passage made for the illustration.Note the following features of transcrip~ion: (1) Tone: Hausa has two distinctive tones.high. which is unmarked here. and low, which is marked with a grave accent (') over thevowel of the syllable bearing the tone. There is also a phonetic falling tone, occurring onlyon heavy (CY: or CYC) syllables and analyzed by most Hausa specialists as high followedby low on a single syllable. This falling tone is marked with a circumflex accent (A) over thevowel of the syllable bearing the tone. Aside from the normal high and low tones, certainclasses of words, particularly interjections and ideophones, are pronounced with an extrahigh tone, which falls above the normal pitch register. These are marked with an acuteaccent C). Hausa has a downdrift intonation pattern, i.e. in a sequence high-low-high-low,each subsequent high is somewhat lower in pitch than the preced.ing high, and likewise forsucceeding lows, but with less pitch declination than for the highs. (2) Punctuation: In thetranscription here, I =a significant pause at a fairly large syntactic boundary but with thedowndrift intonation carried on after the pause; , = a significant pause at a fairly largesyntactic boundary with the intonation level reset to a higher register after the pause; . =sentence level boundary, after which a new intonation register always begins. (3) Glottalstop: As noted in the section on C0!1sonants, [7] contrasts with other consonantsintervocalically, it occurs predictably at the beginning of words written in the standardorthography with initial vowels, and before pause. it terminates words with final shortvowels. In the transcription here, word-initial and medial glottal stops are all written. Wordfinal glottal stops are indicated in those cases where the speaker paused and pronounced aclearly audible [7]. (4) Short vowels: As noted above, the pronunciation of short vowels isheavily influenced by consonantal context as well as vowels in contiguous syllables. Thetranscription of the short vowels in the text attempts to roughly approximate the phoneticvariability.

7Iska: d~ ra:na:

W~I~ ra:na:, d~ 7Isk~r hunlu:ru: I~ 7~re:w~7 I d~ ra:na: suk~ ji g~rd~ma.: 7~

k~n ko:wa.:IJe:IJe: d~g~ IJikmsu I~ <j>i k'~r<j>i:. 10 suna.: IJikm w~nn~l)

g~rd~ma.: I S;)j ga.: w;)ni m;)IMlji:, ja: zo: s~J1£7 d~ n:g;)r s;)J1i:. 10 Ji:ke:n;)n, S;)j7Isk~r hunlu:ru: d~ ra:na: suk~ ji j~rd3e:d3e:n)ja:, 7~ k~n j~w, za:7~ jikW'ure7, d:)n 7~ g;) ko: wa.: z~ 7ij~ sa: m;)l~fIjln j~ lu:5e: n:g;)rs~7 7~ k~n

li:l~S. 10 Ji:ke:n;)n, s~j 7Isk~r hUntu:ru: I~ bu:so: s;)J1ml~7, d~ k';)r<j>i: I 7ij~ jml~

I ~mma: )na:. S;)j I;) ka:s~ sa: Ji: w;)nn~m m~IMIji: I j~ kW;)6e: r):g;)rs~7, d:)nkWuwa, ja: d~nd~nne: r):g~r g~mg~m 7~ J,)kms~7. 10 d~g~ k'~rJe: d;)j, S;)j7ISka.: I~ s;)ll~ma:, d~g~ n;)n ne: kWuma. 711;) ra:na: I I;) flg~ na:l~ 7ajk'rl). 7ajko: ra:na: I S;)j I;) bu:eo: h;)skj;)nl~ d~ za:fml~ w~rw~r. h~~: I k~n k~ IJe: me:,s~j ga.: Ji w;)nn~m mUlum I m;)l~fIji: I ja: kW~5e: r):g;)rsa.:, ba: gilrma: d~

?~rz)ki:. g;)nil) h~k~ kWt'Jwa I S;)j 7ISka.: d6:1e I;) ?~m):nIJe: 7~ k~n IJe:wa:, l;)lle:I ra:na: la: fi: I~ k' ~r<j>i:.

Illustrations of the [PA 95

Orthographic versionIska da Rana

Wata rana, da iskar hunturu ta arewa da rana suka yi gardama a kan kowace ce daga cikinsuta fi Rarfi. To, suna cikin wannan gardama, sai ga wani matafiyi ya zo sanye da rigar sanyi.To, shi ke nan, sai iskar hunturu da rana suka yi yarjejeniya a kan yau, za'a yi Rure don a ga

ko wa zai iya sa matafiyin ya tu5e rigarsa a kan tilas. To, shi ke nan, sai iskar hunturu ta

buso sanyinta da Rarfi iya yinta, amma ina?! Sai ta kasa sa shi wannan matafiyi ya kwa5e

rigarsa, don kuwa ya dandanne rigar gamgam a jikinsa. To daga Karshe dai, sai iska ta

sallama, daga nan ne kuma, ita rana ta shiga nata aikin. Ai ko, rana sai ta budo haskenta da

zafinta warwar. Haba! Kan ka ce me, sai ga shi wannan mutum, matafiyi, ya kwa5e

rigarsa, ba girma da arziki. Ganin haka kuwa, sai iska dole ta amince a kan cewa, lalle rana

ta fi ta Karfi.

ReferencesCARNOCHAN, J. (1952). Glottalization in Hausa. Transactions of the Philological Society

78-109. London.LADEFOGED, P. (1968). A Phonetic Study of West African Languages, 2nd edition.

Cambridge University Press.LADEFOGED, P. AND MADDIESON, 1. (1996). The Sounds of the World's Languages.

Oxford: Blackwells.NEWMAN, P. (1980). The two R's in Hausa. African Language Studies 17,77-87.

PARSONS, F. W. (1970). Is Hausa really a Chadic language? Some problems of

comparative phonology. African Language Studies 11,272-88.

Hebrew

ASHER LAUFER

The Phonetics Laboratory, Hebrew L!Jnguage Department, The Hebrew University,

MOLlnt SCOpLlS, Jerusalem 91905, Israel

There are two main pronunciations in Modern native Israeli Hebrew: Oriental and NonOriental. The Oriental pronunciation -is usually spoken by people with a Near Easternorigin, who have some sort of Arabic or Aramaic in their own or their parents'background. These speakers may have been born in Israel, and by now most of them donot know any Arabic or Aramaic. The Non-Oriental pronunciation is spoken by the rest ofthe population. It should be noted that, for various reasons, there are 'Oriental Israelis'who use the Non-Oriental pronunciation, and 'Non-Orientals' who use the Orientalpronunciation.

The Oriental pronunciation was chosen by the 'Va'ad Hallashon' (the committeepreceding the Academy of the Hebrew Language) to be the representative and theprestigious speech in Israel. This pronunciation was selected to be the preferred one forthe official broadcasting services. However, this decision is not always followednowadays.

The recordings for this illustration were taken from two Israeli-born informants,whose education was in Hebrew. The Oriental informant was a 61-year-old woman whoseparents were also born in Israel. The Non-Oriental informant was a 49-year-old man bornto parents of East European origin; Hebrew and Yiddish were spoken at their home.

The main difference between these pronunciations is in the consonants; the Orientalpronunciation has two pharyngeal phonemes, which the Non-Oriental lacks. In the NonOriental pronunciation, le;I merges with /?I and Ih/ with IX/.

Consonants

Bilabial Lab-dent. Alveolar IPoslalv. Palatal Velar Uvular Pharyng. GloHal


Nasal m n

Trill r XFricative f V S z If 3 h h

Approximant J 1

Lateral I

The phonemes Ih/ and le;I occur only in the Oriental pronunciation.


P par 'bull' t tar 'tours' k kar 'cold'b bar 'wild' d dar 'dwelt' g gar 'lives, lived'

? lor 'light'm gam 'also' n gan 'garden'

f tsaf 'floats' s sar 'minister' f Jar 'sings, sang'

X ma'xar 'sold' h har 'mountain'

v tsav 'turtle' z zar 'stranger' 3 3a'ket 'jacket'

ram 'high' gal 'wave' j jam 'sea'

Oriental pronunciation only

I h ma'har 'tomorrow' I 'I 'lor 'skin'

Affricates might be regarded as phonological units, but they can be treated as sequences ofstops followed by homorganic fricatives: Itsl as in Itsarl 'narrow'; ItJI as in ItJips/ 'chips'

and Id31 as in Id3ipl 'jeep' .

Vowels

eaou

hilhelhalhoIhul

'fear''army of''occurred''sand, workaday''abroad'

a

• U

• 0

Diphthongs can be treated as sequences of vowels. If one of the vowels in such a sequenceis an Iii it can be interpreted as Ij/. Thus, lei] can be interpreted as lej/.

StressStress is distinctive /'berexl 'knee'

Ibe'rexl 'he blessed'.

Differences between the Oriental and the Non-Oriental pronunciationsThe Oriental pronunciation has the two pharyngeals Ih, 'II. Some of the Oriental speakers,especially in liturgical reading, have also pharyngealized sounds [s", t". k"]. Most of thespeakers of the Oriental pronunciation have gemination, especially in careful and slowpronunciation (I'danu/ 'they discussed' vs. /'dannul 'we discussed'). They also preservethe traditional 'mobile schwa' more than the Non-Oriental speakers. Usually the OrientalHebrew speakers pronounce Irl as an alveolar trill (or tap [rl), and in the Non-Oriental


pronunciation it is usually a uvular approximant [Il'] (still, some Non-Oriental speakerspronounce it as a trill). Some occurrences Iof the front-mid vowel lei are diphthongized byNon-Orientals as [ei], but most of the Oriental speakers retain a monophthong [e]. Yearsago this distinction marked a distinct diffbrence between these two pronunciations: whilethe Non-Orientals differentiated between pairs like [mo're 'derex] 'guide' vs. [mo'rei'derex] 'guides', the Orientals produced both as [mo're 'derex,J. It seems that thisdifference is gradually disappearing, as more and more Oriental speakers acquire thediphthong rei] in specific words.

ConventionsIk/ is aspirated and Ip, t/ are slightly aspirated. Ib, d, gl are voiced throughout. In a normalstyle of speech I<;.! is a pharyngeal approximant. 17/, especially in unstressed syllables, canbe elided. Irl in the Oriental dialect is usually an alveolar trill [r], but it is usually a uvularapproximant [Il'] for Non-Orientals. IXI is usually a voiceless uvular fricative trill. Inl isusually [I)] before a velar plosive, and [J1] before Ij/. Especially in unstressed syllables, theglottal stop, the glottal fricative and the pharyngeal approximant 17, h, <;.! may be omitted(though their frequencies seem to depend on the personality, on the style and on therapidity of speech). When intervocalic Ihl is pronounced, it is voiced [fi]. Vowels andconsonants are long in stressed syllables. 'fhey are even longer at the end of sense groups.Vowels are centralized and shorter if they are unstressed.

Phonemic transcriptions of recorded passages

] Oriental Hebrew'ruah hattsa'fon, vehaJ'JemeJ, hitvake'hu bene'hem, 'mi me'hem ha'zakjO'ter. game'ru, ki ?et hannitsa'hon, jin'hal, 'mi Jejjats'lijah lif'Jot me'<Jal<Jo'ver '?orah ?et bega'dav. pa'tah 'ruah hattsa'fon vena'Jav behoz'ka. hid'dekha?a'dam ?et bega'dav ?el gu'fo. '?az, hista'<Jer <Ja'lav ha'ruah, be'jeter '<Joz,'?ax ha?a'dam, miffeho'sif hak'kor le<Janno'to, la'vaJ me'<Jil <Jel'jon <Jalbega'dav. no'?aJ mim'mennu ha'ruah, umesa'ro bi'de haJ'JemeJ. tehil'la,za'rah <Ja'lav haJ'JemeJ berak'kut. veha?a'dam he'sir ?et big'do ha<Jel'jonme<Ja'lav. hig'bir haJ'JemeJ ?et hum'mo, 'lad Jd'lo ja'xol ha?a'dam la<Ja'modbif'ne haffa'rav, ufa'Jat '?et bega'dav, venix'nas le'tox hanna'har, Jeha'jabekir'vat ma'kom, ke'dei lir'hots bemei'mav.

2 Non-Oriental Hebrew'ruax hatsa'fon, veha'JemeJ, hitvak'xu beine'hem, 'mi me'hem xa'zak jo'ter.gam'ru, ki ?et hanitsa'xon, jin'xal, 'mi Jejats'liaX lif'Jot me'?al ?o'ver '?orax?et bega'dav. pa'tax 'ruax hatsa'fon, vena'Jav bexoz'ka. hi'dek ha?a'dam ?etbega'dav ?el gu'fo. '?az, hista'?er ?a'iav ha'ruax, be'jeter '?oz, '?ax ha?a'dam,miJeho'sif ha'kor le?ano'to, la'vaJ me'?il ?el'jon ?al bega'dav. nO'?aJmi'menu ha'ruax, umsa'ro bij'dei ha'JemeJ. texi'la, za'rax ?a'lav ha'JemeJ

Illustrations of the lPA 99

bera'kut, veha?a'dam, he'sir ?et big'do ha?el'jon me?a'lav. hig'bir ha'JemeJ?et XU'mo, 'fad Je'lo ja'xol ha?a'dam la?a'mod bit'nei haJa'rav, uta'Jat '?etbega'dav, veniX'nas le'tox hana'har, Jeha'ja bekir'vat ma'kom, kedei lir'xotsbemei'mav.


'~ "I]~~ Tin¥~iJ M~ ':;l ,~,~~ •'1)i' P!O c~~ '~ c~'J.'~ ~nflJ;1il lZ'??~iJ1 Ti!:)~iJ·TJ1'

P·7.i:J .ili?10~:l~ ~1 Ti!:)~iJ -1]1' n1J~ ."1H M~ n1~" '~i~ ':l~ tjlZ'~? I]"¥~o/.l:l'oi;,~~ 'C1t'Q l~ ,Ti~ '~T.~ l]nQ "?¥ '¥f:)9i} T~ .i!:)u"~ "1~~ M~ C1~Q.lZ'??~iJ '7.'~ i'9~1 l]1'Q 1J~~ lZ'~iJ ."1H ':l Ti'7¥ "'¥~ lZ'~? ,iniJ:l7 'ipi]lZ'??~iJ "~~i} •"?¥~ li'7¥Q i'~~ n~ "!?iJ C1~Q1 ,m::l'p lZ'??~iJ "?¥ n"n il?r:tf;1"Q~i] liM? oJ.~~1 "1H M~ tl~~~ ,:ll~iJ 'J.~~ ,~~? c1~Q ','J: ~'o/. ':l ,i~~ M~

."1?'~~ yn:? '7.~ ,ciP1? n~:p~ ;':Qo/.

J:lindi

MANJARI OHALA

Department of Linguistics and Language Development and Department ofEnglish,

San Jose State University, San Jose, CA 95192. USA

The variety described here is Standard Hindi used in everyday casual speech by educatedspeakers in cities such as Varanasi, Lucknow, Delhi, etc. Although there are a few

differences in pronunciation among speakers of these cities, the differences are minimal.

The transcription is based on a recording of a female third-generation speaker of Standard

Hindi who grew up mostly in Uttar Prades~ before moving to Delhi. For a detailed analysisof Hindi segments, see Dixit (1963) and Ohala (1983),

Consonants

Bilabial Labio- Dental Alveolar Post- Retroflex Palatal Velar Glottaldental alveolar

Plosivep b 1 Q t cl k 9ph bfi !h Qfi th clfi kh 9fi

Affricate tJ d3tJh d3fi

Nasal m n I)

Tap or Flapf r

r fi

Fricative f S Z J h

Approximant 0 JLateral IApproximant

p pal 'nurture' 1 1al 'beat' (n) k kal 'span of time'

b hal 'hair' g gal 'lentil' g gal 'cheek'ph phal 'knife blade' t

h thai 'platter' kh khal 'skin' (n)

bfi bfial 'brow' gfi gfiar 'knife edge' gfi gfian 'bundle'

m mal 'goods' n nola 'drain' (n) l] OOlJm~j 'literature'

farsi 'Persian' s sal 'year' h hal 'condition'

u uala 'pertaining to' z ~min 'ground' j jar 'buddy'

ral (tree species)

lal 'red'


tJ tJal 'gait' t tal 'postpone'd3 d3al 'net' <t <tal 'branch'tJh tJhal 'tree bark' th thai 'lumber shop'd3 fi d3~1 'glimmer' <tfi <tfial 'shield'

J Jal (tree species)

r oora 'big'

rfi rorfia 'increase' (imp)

[f, z, J] only occur in loans (from Perso-Arabic, English, or Sanskrit). They are, however,well established in Modem Standard Hindi. Some authors also include [f] which occursonly in a few Sanskrit loans; in casual speech it is usually replaced by [n]. Sounds such as[x, y, q], etc., do not occur in the variety of Hindi described here although they would befound in some varieties of Urdu, the form of the language used as the official language ofPakistan and by many Indian Muslims. [Ifi, r fi , mfi , nfi ] are considered clusters of a liquid ornasal and /hI in the analysis adopted here.

GeminatesAll consonants can also occur with distinctive length (i.e. as 'geminates') except thefollowing: [b fi , r. r fi , h]. Geminates occur only medially and are always preceded by thenon-peripheral vowels [I, ~, u]. Although the orthography preserves a few geminates infinal position, in all but the most formal speech they are pronounced as singletons. All thegeminates occur monomorphemically except U:] which occurs only in a few Sanskrit loanswhere a morpheme boundary could be posited in between: [mJ:il] 'without shame' < ImJ +Jill. Geminates are longer than singletons by a ratio of about 2: I. Some illustrative examplesof the geminate contrasts are given below.

t ~!a 'address' (n) tJ b~tJa 'save'

t: ~!:a 'leaf tJ: rotJ:a 'child'th k~!ha 'narrative' t ~ta 'to make someone agree't h: k~!h:a 'red powdered bark' t: p~t:a 'a fold' (in cloth)cj g~cja 'mace' (weapon) k ~ka 'cook' (v)

cj: g~cj:a 'mattress' k: ~k:a 'finn'

VowelsThere are eleven oral vowels in Hindi, as shown on the vowel chart. The vowel [ee] onlyoccurs in English loans. All of these vowels except [ee] also have distinctively nasalcounterparts. The sequences [;Ji] and [;Ju] also occur but are not listed separately becausethey are analyzed as vowel clusters and not as diphthongs.


u·

o·:)

• a

mil 'mile' u leu1 'shore'mIl 'meet' u leul 'lineage'

e mel 'harmony' ;} rrnl 'rub' (imper) 0 bol 'speak'e mel 'dirt' :J bl (a name)<e b<et 'cricket bat' 0 mal 'goods'

sikh 'lesson' 0 sos 'mother-in-law' u bfiukh 'hunger'slk 'twig' 0 sos 'b~eath' u bfiuk 'dog's bark'sHf 'be irrigated' 0 kOwf 'prince'

e me 'in' g hgs 'laugh' (imp) 0 jo 'in this manner'e me (l sg pn) :; bfi 5 'eyebrow'

StressStress is not distinctive in Hindi. It is also controversial as to whether there is even phoneticword stress, other than for emphatic contrast. For details, see Ohala (1991).

ConventionsThe velar nasal by and large only occurs before homorganic consonants except in a fewSanskrit loans such as the word for 'literature', cited above. Although palatal and retroflexnasals followed by homorganic consonants also occur phonetically and have been includedin the list of phonemes by some authors, they have not been treated as phonemes in theanalysis presented here (and thus are not listed in the consonant chart). [u] is in freevariation with [v) (and can also be pronounced as [w)). [r] can vary with [r]. The geminate[r] is always an alveolar trill [r). Stops in final position are not released.

Vowels are nasalized before nasal consonants. The schwa is often pronounced with thesame value as the vowel in the English word 'cut', i.e. the vowel that has traditionally beentranscribed with [A). However, this vowel is central and not back, and thus perhaps a moreappropriate IPA symbolization would be [u). As in many other languages, vowel lengthvaries as a function of the voicing of the following consonant with vowels being longerbefore voiced ones, at least before stops (Ohala and Ohala 1992).


Transcription of recorded passageThis translation of 'The North Wind and the Sun' is a modified version of that presented inthe 1949 version of the Principles of the International Plwnetic Association.

u!:gci hguo JC Sucgd3 IS bot pgC d3fi gggr cghe the ki hgm gono me zjogo\)gIuon bn he. I!ne me ggcgm tfogo pehne ek musoflC ugfigC a mk.Io. hguoJC surgd3 gono IS bot pgC cozi ho ggje ki gono me se d30 pehle musofIf kotfogo U!gruo dego ughi zjogo bgluon Sgmd3fiO d30jega. IS ~c u!:gri hguagpna puro zoc Iggobc tfgIne Iggi lebn uo d3£se d3£se gpna zoc \)grfio!i ggjiuese u£se musofIf gpne \)gggn pgc tfoge ko JC bfii zjogo bs bc !gpef1o ggjo.gn! me hguo ne gpni kOflf bgng bc gi. phIf sucgd3 !ezi ke SO!h orklo JC

musaflc ne !ucgn! gpno tfoga U!OC gIjo. IS hje hguo ko manna pgra ki ungon6 me surgd3 hi zjoga \)gluan he.


Nn ~ m ~~ ~ 0fRf CJT ~rr:s \% 21 fc1:; ~ ~ B ,,~ ~

m %I~ B 1W1 ~pn~ ~q> TfTfcf;{ 3"trr 3fT RWr I ~ 3ftt~

~ffi ~~ GfRl CJT U"fT ~T rriT fc1:; cfRr B B- "TT~ 'r~rfr:n{ q>T ~Trrr «f0'fT~rrT <%t ~r GfmR ~~T ~rrT I~ CJT 3D"ft ~ 3fCf1r ~ "Tn- (1l'fTql{

~~ <1lfT c1fc1:;~ ~ ~B- ~B- 3TCRT '-'iR~ rr% ~B- ~~{ 3fcR~ CJT

m q;T m ~T~ q>{ ~Gilll"f1fr I 3f~ B~ ~ 3fCf1T CflIfww R ~~ I f':n{~~ ~"TT ~ m~ RWT 3Ttt J:Rf~ ~ (f~ 3fCRT '9Trrr 3"(fIT fc0rr I ~m~ . ~. ~

~ q>T 11T-l;n q:sr fcF~ cfRT B~~ ~ ~ "1c1"C!"fol %

AcknowledgementsI thank John Ohala for his help with the software implementation of the IPA transcription. I also thank

Usha Jain for her comments on the translation and for her assistance with the software for the Devanagari.

ReferencesDIXIT, R. P. (1963). The Segmental Phonemes of Contemporary Hindi. M. A. thesis,

University of Texas, Austin.

OHALA, M. (1983). Aspects of Hindi Phonology. Delhi: Motilal Banarsidass.OHALA, M. (1991). Phonological areal features of some Indo-Aryan languages. Language

Science 13, 107-24.OHALA, M. AND OHALA, 1. J. (1993). Phonetic universals and Hindi segment durations.

Proceedings of the Second International Conference on Spoken Language Processing,831--4. Edmonton: University of Alberta.

Hungarian

TAMf,.S SZENDE

Institute of Linguistics, Ftungarian Academy ofSciences,

P.O. Box 19, H-h50 Budapest, Hungary

The style of speech illustrated is that of Educated Colloquial Hungarian of the 1990s asspoken in Budapest. Historically, Standard Hungarian is based on the Eastern dialect,which became a supra-dialectal high prestige variety during the eighteenth century. Thespeech on the recording is that of a male s~eaker in his fifties with an academic background,using a somewhat advanced style of speed) as spoken in a formal communicative situation.

Consonants

Bilabial Labiodental Dental Post- Palatal Velar Glottalalveolar

Plosive p b t d k 9Affricate ts· dz tJ d3 C9 jj

Nasal ill n J1Fricative f V S z J 3 h

Trill r

Approximant jLateral

1Approximant

p pipa pipa 'pipe' t tol: toll 'feather' k ke:p kip 'picture'

b bot bot 'stock' d dob dob 'throw' 9 ge:p gep 'machine'

m ma rna 'today' n ne:m nem 4 no' p Jla:r nyar 'summer'

f fa fa 'tree' s so: sz6 'word' S So: s6 'salt'

v va:g vag 'cut' z Z01d zold 'green' 3 3e:b zseb 'pocket'

ts tse:l eel 'goal' tS tSak csak 'only' Cy cyu:k (yuk 'hen'

dz dze: ta dzeta 'zeta' d3 d3e:S: dzsessz 'jazz' jj jja:r gyar 'factory'I

1 10: 16 'horse' r ro: r6 'carve'

h ho: h6 'snow' i jo: )6 'good'


Consonant length is distinctive in Hungarian, cf. hal '(s)he/it is dying' vs. hall '(s)he/ithears' and hal 'fish' vs. hall 'hall'. Short/long opposition extends over the whole system ofconsonants functioning in (at least a limited number of) word forms. Accordingly, the fullsystem consists of short C : long C pairs at all positions of the pattern. However, theselong consonants are commonly analyzed as clusters of identical consonants. The phonemicanalysis underlying the present chart reflects this second view.

Vowels

1: • u:u·

• 0:

o •

.0• a:

vis visz 'carry' i: vi:z viz 'water'

y yt Ut 'hit' y: ty:z tUz 'fire'

u u ujj 'finger' u: u:t Ut 'road'

e ez ez 'this' e: e:l iL 'live'

¢ ¢l al 'kill' ¢: l¢: L6 'shoot'

0 ok ok 'cause' 0: to: t6 'lake'

a hat hat 'six' a: va:r var 'castle'

The seven basic vowel qualities occur in distinctively long and short quantities. Except forthe low pair, the short vowels are a little lower and less peripheral than their longcounterparts and the same symbol can be used for long and short members of each pair.However, the short vowel corresponding to long la:1 is higher and backer than its longcounterpart. To show its departure from the usual pattern it is transcribed as [a]; in fact, thisvowel has some rounding and might be transcribed [0]. As for the [e:/e] pair, mid frontunrounded long [e:] is considerably higher and more peripheral than its lower mid frontunrounded short counterpart [e].

In addition to the seven basic vowel pairs three additional length/quality combinations,shown below, are phonemically distinctive in a limited number of minimal pairs only, forexample 10:1 A 'the letter A' as opposed to 101 a (definite article). The marginal existence ofboth a long 10:1 and a short Ia! in the sound pattern of Hungarian provide further reason fortranscribing 101 and la:1 differently. The third marginal vowel, here written le:1 is similar toa lengthened version of the short vowel leI.


(e:) e:

(a) pas:

(0:) 0:

E (name of the lett~r E)

passz 'no bid' (in cards)

A (name of the letter A)

Hungarian has two co-existent vowel systems, one of which does not distinguish twodifferent heights among its mid front un rounded short vowels. However, a major segment(some 50 per cent) of the Hungarian-speaking population uses a vowel system with thedistinction of lei and lei. In this Regional Standard version of Hungarian. orthographicmenlek - i.e. Budapest Standard [mentek] - represents four different words: [mentek] 'go(2 pl.. pres.)', [mentek] 'go (3 pI., past)'. [mentek] 'save (I sg., pres.)', or [mentek] 'beexempt from (3 pI., pres.)'.

StressWord-level stress is nondistinctive, and is fixed on the first syllable of the morpheme.

ConventionsIp, t, k/ are unaspirated, Ib, d, gl are fully voiced. It, d, n. I, rl are laminal dental, Is, z. ts,

dzi are laminal dentialveolar, If, 31 are .Iaminal postalveolar. Itf. d31 are predorsal andpostalveolar, and ltc, dzl are mediodorsal and palatal. In formal style loy, jjl are realizedmostly as palatal stops, i.e. [c) and [j).

Irl is a trill. /hi is voiced [fi] in intervocalic position, [y] in syllable-final position afterfront vowels, and [x] word-finally after back rounded vowels; word-final Ihl is oftendeleted.

Short vowels are to some extent reduced (lax) in unstressed position; their longcounterparts are realized as full (tense) vowels. Long vowels, especially high ones, shortenin unstressed syllables with a consonant in the coda. The resulting vowel can be half-longor as short as a short vowel. A postvocalic Inl usually nasalizes the preceding vowel anddeletes before a following consonant, ,especially a continuant. (When the followingconsonant is a nasal. this process can be viewed as a coalescence of the two nasals.)


,ets:er az 'e:soki 'se:l e'J a ,nap 'vetdket:ek hOjj 'mejikyk az 'eret>:feb: II'e'p:a:ro jet>t: £jj 'va:dor ,voStok ,ket>pet>j1egbe burko1o'dzva II az ,e:saki ,sell e'J a,nap pomba- 'meg£jj:estek hOjj 'az l£s: a 'jjet>:stef oki 'fiamorab ,ra:biTja a,va:dort hOjj 'lev£jje a 'ket>pet>j1ege't II ak:;)r az ,e:soki ,se'l ,e:kezdet 'Jyvet>lteni,afiotf:ak 'bi:rt II de a ,va:dor 'an:a'l 'sorofab:aTl) vota ,maga ket>re' a ,ket>pej1t,mind 'wtl'Jeb:eTl) ,fujt II ,i:jj asta'n az ,e'soki sd ,£1 is 'vesitet:e a ,verfej1t II a,nap meg ,e:keste 'otani 'ty:zC/>' ,fugaroit ,mire a ,va:d;)r 'ets:eribe 'kibujt a,ket>pet>j1ege'bet>1 II az ,e:saki ,se:l 'ke'j1tele vot 'megadni hOjj ,bizoj1 a 'nap az,eret>:feb:

lIlustrations of the IPA 107

Orthographic versionEgyszer az eszaki szel es a nap vetelkedtek, hogy melyikiik az erosebb. Epp arra jott egyvandor, vastag koponyegbe burko16dzva. Az eszaki szel es a nap nyomban megegyeztek,hogy az lesz a gyoztes, aki hamarabb rabirja a vandort, hogy levegye a koponyeget. Akkoraz eszaki szel elkezdett siivolteni, ahogy csak birt. De a vandor annal szorosabban vontamaga kare a kopenyt, minel erosebben fujt. fgy aztan az eszaki szel el is veszitette aversenyt. A nap meg elkezdte ontani taza sugarait, mire a vandor egyszeriben kibujt akopanyegebol. Az eszaki szel kenytelen volt megadni, hogy bizony a nap az erosebb.

Igbo

CLARA I. IKEKEONWU

Department ofUnguistics and Nigerian Languages, University ofNigeria,

Nsukka, Enugu State, Nigeria

The style of speech illustrated is that of many educated speakers of Standard Igbo.Standard Igbo can be seen as a fusion of aspects of Central Igbo and Onitsha Igbo. In itspresent form, Standard Igbo cannot be localized in any particular region or area ofIgboland (Ikekeonwu 1985).

Consonants

Bilabial Labio- Dental Alveolar Post- Palatal Velar Labialized Glottal Labial-dental alveolar Velar Velar

Plosive p b t d k 9 kW gW kP 91Nasal ill n J1 l) l)w

Fricative f S Z IJ Y fi

Affricate tJ d3Approx.

1 J wLateral

1Approx.

p apa apa 'scar' J1 J1~ ny~ 'defecate'

b (jba I)M 'wealth' l) l)~ JW 'drink'

ta tel 'chew' l)w l)w~ nw~ 'die'

d da da 'fall' f fe fe 'fly'

tJ tJa eM 'ripen' s sa sa 'wash'

d3 ;i.ld3a aja 'sand' z za za 'swell'

k ka ka 'older, more' J aJa asha 'weaver bird'

9 ga ga 'go' y aya agM ~wart

kW akwa akwa 'bed' fi afia aha 'name'gW agWa agwa 'behaviour' w wa wa 'split'

kP akPa akpa 'bag' 1 11 ri 'eat'

gb agba 8gM 'fame' j ji yl 'resemble'

m mv m~ 'me' I n Ii 'bury'

n n~ mJ 'hear'


Vowels

isi lSI 'head'

i }b¢ }b¢ 'dissect'

1,1 l,Imv Ijml) 'children'

u emu emu 'derision'

0 eko ekb 'gizzard'

e uk6 ilk6 'rafter, roof'

Q ¢ml ¢mJ 'marrow'

a aklJ ak~ 'kernel'

u

\I •._-------\-----.. 0

Q

ConventionsBoth dental and alveolar plosives [1, \!] and [t, d] are found in the language. The alveolarplosive allophones occur mainly in the environment of lui and 11,1/. The voiced postalveolar approximant hi is realized as a flap [r) intervocalically, e.g. ara lara] 'madness'.

The labial-velar plosives Ikp, gbl involve double articulation; the velar and labialclosures are made and released simultaneously. The airstream used is pulmonic egressive.However, in some dialects, e.g. Enu-Onitsha Igbo, the voiced bilabial-velar plosive Igblmay be realized as a voiced bilabial implosive [6] and in the Owerri dialect IkPl may berealized as a voiceless bilabial implosive [9]. Aspirated and nasalized consonants are not

used in Standard Igbo. They are, however, used distinctively in some Igbo dialects, e.g.llmlJahia and Owerri, within Inland East Igbo (Inland East Igbo incorporates what wasknown as Central Igbo in earlier writings). A voiced labiodental fricative [v] occurs insome Igbo dialects but not in Standard Igbo; the llm1.!ahia dialect has [njvQ] for'fingernail' where Standard Igbo has [mb(i].

There is vowel harmony in Igbo. The eight vowels fall into two sets distinguished bypharyngeal cavity size (Lindau 1975), with 'expanded' vs. 'unexpanded' pharynx. The

sets are shown below. Vowels from different sets do not normally co-occur in a word.This distinction is often discussed in terms of an ATR (Advanced Tongue Root)parameter, since fronting the tongue root contributes to pharyngeal expansion. One of thediacritics [.1 or [.J can be used to distinguish one set of vowels from the other. In thisillustration, the unexpanded (-ATR) set are marked, except that since separate letters areused for lei and lal it is unnecessary to use a diacritic to distinguish this pair. The oldertradition of using symbols that suggest the auditory height differences of the vowels, asindicated by their locations on the accompanying vowel chart, can be used as analternative. The suggested symbols would be:

[+Expj I [+ATR)

II

[-Exp] I [-ATR]

ue o a


Two distinctive tones are often recognized in Igbo; in addition, there is a third tone, a

downstepped high. The tones may be marked as follows:

['] = high ['] =low [~'] =downstep.

In our text illustration, we have left the low tone unmarked. In the orthography, however,

the downstep is usually indicated with a macron [-] on the affected segment. In the

illustration of the passage in the 9nwl,l orthography (the current orthography for Igbo), we

have marked all the tones.

There is the tonal phenomenon of downdrift in Igbo: each succeeding high tone is

lower than the preceding one, especially when there is an intervening low tone. In a

sequence H - L - H, the second high is on a lower pitch than the first. Low tones are also

affected by downdrift. This phenomenon, of course, is of no phonemic significance.

Transcription of recorded passageikuku yglJ.II,J na a~l)wy naaJtPJta y~ka OJ1e ~ka Ibe ja ike rngbe fia fi1,1.II,J otuoJ1e Id3e ka 0 ji uwe yglJl'.1 ja naabJa. fia kwek<;ujtaJa na OJ1e ~bwu ~zQ

rnee ka OJ1e Id3e afilJ jIplJ uwe ~j'i1 ka a ga ewe dJ ka oJ1e ka Ibe ja Ike.ikuku yglJ1lJ wee rnallte fee, fee, fee, otu ike ~ja fia; rna ka Qna efe kaOJ1e id3e afi'.1 na ed3idesl ~uwe ~ja ~ike na afiy ~ja. ja fekata fiap'.1. rngbeafi1,1 a~l)wy wee tJaplJta, tJasi~ke rnee ka ebe n11~le kP010 Q~k1,1 naJ.Mufuyioge OJ1e Id3e afilJ jiP'.11'.1 uwe J.ja l)ke a rne1e ikuku lJg'.111,1 kwele na aJ.l)w1,.1ka ja Ike.

Orthographic versionlkuku ~gi;!n) na AnwQ na-a~(ll<i ~ka 6nye ka "ibe ya Ike mgbe ha hun) 6tu 6nye \je ka

6 y'l uwe ~gi;!n) ya na-ab.ia. Wi kwekQr!tara na 6nye biirU ~zQ mee ka 6nye \je ahi;!

y\pi;! uwe ya ka a ga-ewe d)ka 6nye k'l \be ya Ike. lkuku ~gi;!ri;! wee malhe fee, fee,

btu Ike ya ha; ma ka Q na-efe kfl 6nye \je ahi;! na-ej\des\ iiwe ya Ike na ahu ya. Ya

fekata hapi;!. Mgbe ahi;! AnwQ wee cMpi;!ta, chaslke, mee ka ebe nllle kpbro QkQ; na

atUfiigh'l 6ge 6nye ije ahi;! y'lpi;!n) uwe ya. Nke a mere 'Ikuku ~gi;!n) kwere na AnwQ ka

ya ike.

ReferenceslKEKEONWU, C. 1. (1985). Aspects ofIgbo dialectology. Journal afWest African

Languages 15,93-109.

LINDAU, M. (1978). Vowel features. Language 54, 541-63.

Irish

AILBHE Nf CHASAIDE

Centre for lAnguage and Communication Studies, Trinity College, Dublin 2, Ireland

Irish, or Gaeilge, is a Celtic language which is spoken as a mother tongue in certain parts

of Ireland, known as Gaeltacht regions. It is closely related to Scottish Gaelic and Manx,and more distantly to Breton, Welsh and Cornish. There are three main dialects of Irish,

and although there is a written standard form there is no spoken standard accent. Written

records in Gaeilge date back to the eighth century, with a literary tradition continuing in

manuscript until the mid-nineteenth century, when published books became

commonplace. The spelling system remained relatively unchanged from around thetwelfth century until a government-sponsored reform in the mid twentieth century, buteven so, orthographic forms can be rather archaic. There is on the whole a poor

correspondence of letter to sound, as will be evident from the orthographic version of the

text below. This phonetic opacity, however, confers the advantages of morpho

phonological transparency, and helps the written form bridge divergences between the

modern-day dialects, being relatively equidistant from the spoken forms. The description

and sample text given here is based on the speech of a native of Na Doirf Beaga, Gaoth

Dobhair, one of the Gaeltacht areas in County Donegal, situated in the north west of the

country. She is a teacher in her thirties, working in Dublin.

ConsonantsConsonant quality: The most striking feature of the consonantal system is the

phonological opposition of a series of palatalized and velarized segments. This is

accommodated in the table by subdividing the rows and placing the velarized series above

the palatalized. Consonants which do not enter into the opposition, such as [h], are placed

in the middle of a cell. The palatalized-velarized opposition serves not only for lexicaldifferentiation, but may also express certain grammatical functions, such as case and

number marking on nouns, e.g. [:Jl Y] 6/ 'drink' (nom. sg.) [:Jl i ] 6il 'drink' (gen. sg.);

[bYaeqY], Md 'boat' (nom. sg.) [bYaeg j ] Mid 'boat' (gen. sg., and in some dialects, nom.

plur.). Word internally, consonants in a cluster typically agree in terms of quality. The

customary term velarization is used although the secondary articulation may be further

back (in the uvular or upper pharyngeal regions) than strictly implied by the termvelarization.

The phonetic distinctions between palatalized and velarized pairs of consonants may

involve more than secondary palatalization and velarization. In the case of the labial

consonants, the velarized segments are additionally labialized. For pairs such as Iii, lYI,both primary and secondary features are involved. The former is a palatalized lamino

posta[veolar, which will here be referred to as alveolopalatal, and the latter is a velarized


apico-dental with contact extending onto the alveolar ridge. In the case of palatals andvelars, for example Ic, k/, where the pt;imary place of articulation coincides with thesecondary articulation characterizing the phonological series, the realizations aretranscribed in terms of primary articulation differences alone.

Labial Dental Alveolar Alveolo- Palatal Velar Glottalpalatal

pY bY F gY k 9Plosive pi hi !i gi c j

Fricative/ fY W sY X yApproximant fi vi C c; J h

mY I}Y I)Nasal

mi n Oi J1rY

Tapri

Lateral FApproximant I Ii

pY pYirYah Paorach (family name) pi piini pighin 'penny'bY bYi bui 'yellow' hi hii bt 'be' (imp)

l Y !Yigj~ taoide 'tide' !i !ji (ar) tt 'about to'gY gYinji daoine 'people' gi gii dt 'drink' (gen)

k kiP caol 'thin' (masc) C cia}Y ciall 'sense'g gi}Y gaol 'relative' j jia}Y giall 'hostage'fY fYi faoi 'under' fi fji fi 'weaving'w WI rnhaoigh 'boasted' vi vji bhl 'was'sY sYi sut 'sit' e ei 'si 'she is'x xiP chaol 'thin' (fern) <; <;ia}Y (rno) chiall (my) 'sense'y yi}Y (rno) ghaol (my) 'relative' j jia} Y (rno) ghiall (my) 'hostage'h hiP (le) haol (with) 'lime'mY mYi rnaoigh 'boast' (imp) mj mii rni 'month'lJY lJYi naoi 'nine' lJj niih nigh 'wash' (imp)

l jeQY leann 'education' l ien j leinn 'education' (gen)

n en ean 'bird'I) I)iP (a) ngaol (their) 'relative' p pia} Y (a) ngiall (their) 'hostage'

P pe lae 'day' (gen) Ij l jej leigh 'read' (imp)

gap Gall 'foreigner' kaI j caill 'lose' (imp)

lej let 'with her'gil (ag) gail 'boiling'

rY mYrerYah 'rndrach 'tomorrow' rj kreriah cdireach 'dirty'


Since both velarization and palatalization are generally indicated, the symbolization in thechart is over-specified, except for palatals and velars (as mentioned above) where primaryand secondary features coincide, and leI and Iwl where the secondary articulation is alsoimplicit in the symbol. In phonological terms, it would be sufficient to mark only onemember of the opposition (traditionally the palatalized one), but this over-specificationdraws attention to the auditorily striking presence of phonetic velarization in the nonpalatalized series. Strong onloff glides are generally found between velarized consonants

and adjacent front vowels.Alveolar laterals and nasals, and glottal fricatives, are a potentially problematic area in

this dialect. Here it appears that an earlier opposition of palatalized/velarized segments hasbeen lost, or at least is in the process of being lost. As produced by the present informant,these segments have a rather neutral quality and take their colouring from the adjacentsegments. The case of the nasals and laterals is discussed more fully below.

In some phonological treatments. a separate series of voiceless nasals and liquids hasbeen posited. As such sounds occur only in certain grammatical morphemes. they areprobably best treated as sequences of nasal (or liquid) plus /hI, an analysis suggested by

Sommerfelt (1964).Plosives: The voiceless series is strongly postaspirated and devoicing of the voiced

series is fairly widespread. especially in non-intervocalic contexts. The voiceless series areslightly preaspirated. Nasals and liquids before and after voiceless plosives are devoiced.The alveolopalatal plosives I!j, ~PI are affricated, as are the palatal plosives Ic, j/. though

to a lesser degree.Fricatives: lvi, w, j, yl may be realized as approximants or as fricatives. The fricatives

and approximants are grouped together because of their close relationship with their stopcongeners. An important morphophonological alternation between word-initial stops andfricativeslapproximants in Irish expresses certain grammatical functions such as presentvs. past tense. gender marking on adjectives. and case marking in certain noun phrases. Inthe Gaoth Dobhair dialect, the voiceless velar fricative Ix! tends to occur only in initialposition. Where it occurs non-initially in other dialects, Gaoth Dobhair usually has /hi.

Laterals and Nasals: Historically. there was a four-way opposition of laterals and

nasals in the dental to alveolopalatal region. This system is widely reported in the dialectliterature for Ulster (e.g. Quiggin 1906; Sommerfelt 1922; 6 Searcaigh 1925) and seemsto have involved an opposition of palatalized and velarized alveolar segments, contrastingwith a velarized dental and a palatalized alveolopalatal segment. However the present-daysituation, particularly among the younger generation in Gaoth Dobhair. is that one of thealveolar segments has more or less disappeared, leaving a robust three way contrast, withthe possibility for some speakers of a marginal fourth phoneme. The latter was not reliablyproduced by this informant. In the case of the laterals. the velarized alveolar has tended to

merge with the velarized dental !pI (see Nf Chasaide 1977, 1979). In the case of thenasals, the situation is more fluid: the palatalized alveolar has tended to merge with thepalatalized alveolopalatal nasal IUi/. In lexical items which have retained the (original)alveolar pronunciation, there does not appear to be a contrast with the (originally)


velarized alveolar.

The remaining single alveolar lateral and nasal consonants have a fairly neutral orslightly clear quality, and tend to coarticJlate freely with the quality of adjacent segments(see Nf Chasaide 1977). In prepausal position, the alveolar lateral and nasal segments tendto be devoiced, and to clip a preceding short vowel making it over-short. As the alveolarnasal does not occur in initial position, illustrations of the dental, alveolar, andalveolopalatal nasals and laterals in word-final position are included in the consonantword list.

Rhotics: An older four-way distinction among rhotics has been reduced to a two-waydistinction in Gaoth Dobhair. Initially, the IcY, eil distinction has been neutralized to avoiced post-alveolar approximant [1], which takes its colouring from the following vowel.

In non-initial position, the veJarized member can be realized as either leY] or [l Y]. Thehistorical palatalized /ci/, which in most other dialects is a palatalized tap or apicopostalveolar fricative, may also appear in Gaoth Dobhair as a voiced palatal fricative [j] innon-initial position. For both palatalized and velarized taps, there is often incompleteclosure with considerable frication.

Vowels

Vowels fall into long and short sets, with long vowels (Ii, e, re, :>, 0, uI) typically shown

in the orthography with an acute accent. Since long and short pairs of vowels arequalitatively distinct, the transcription does not mark length explicitly. Long vowels have

short allophones in unstressed syllables and before Ih/. Schwa occurs only in unstressedposition. The close rounded vowel is often realized with only weak lip rounding.

u

Allophonic realizations: The locations of vowels marked on the charts represent a fairlyhigh degree of abstraction, since consonantal quality exerts major influences on thequality of adjacent vowels. For all vowels, clearly audible on/off glides appear when theirfrontness or backness conflicts with the secondary articulation of the consonant withwhich they co-occur. Thus, a velarized consonant preceding a front vowel gives rise to astrong diphthongal glide of velar origin, and conversely, a strong palatal glide is audiblebetween a palatalized consonant and a following back vowel. The wide range of steadystate realizations of short !II and lal are illustrated by large ellipses.

I/lus/rations of the [PA 115

bYi!Yah baolach (to be) 'feared' Il~ uile ' all'

e e!Yi ealaigh 'escape' (imp) e el~ eile 'another'

re relVIDi dlainn 'beautiful' a alVah ealtach 'cattle'

:> :>lV~QY alann (he) 'drinks' 1\ I\lV~ ola 'oil'

0 orYan amhrdn 'song' y YlVu Uladh 'Ulster'

u uFi ultai 'apples'

ia ia! Yah~ iallacha 'laces'

ua uaFah ualach 'burden'

au auwIoi Eamhain (place)

SuprasegmentalsPrimary lexical stress is located on the first syllable of most words. The stress marks inthe transcription indicate the syllables accented by the speaker in the reading of thepassage. II and I indicate the ends of major and minor intonational phrases respectively.

Transcription of recorded passageyji ;:)0 'Vi ;:)'ejYuaj ISV ;:)0 'jrjian ;:)] 'arv;:)gal 1;:) yd;:) 1;:) fYal ;:)'mvah 'caku <;!v;:)n,yjerj!j ;:)bv;:) 'yrjec;:) I l)v;:)rv;:) ham] 'lVac!ja}Yi ;:)0 'bial Yah I ;:)g;:)sv 'kl v;)k;:) '!jeh axayu ep II '<;!vel)vlvi c;:)<;!v g;:)rv ;:)0 QVIgj;:) by;:) 'yrjec;:) I ;:)0 <;!VIgj;:) bY;:) '1Vua~ ;:)'horVhu ;:)rV ;:)0 'lVac!ja! Vi ;:) 'x! v~k;:) WVIgj!j 'Qv;) II m'cm 'he~V ;:)0 'vi ;:)'QvuajXYh '!jrjeQv ISV ;:) 'hyku 1e~ I ah a 'yjegj ISV ;:) 'hegj ci I ca bY;:) '1vu~ ;:) 'haI}Yan 'lVac!jalYi;:) 'x}Y;)k;:) 'hafvlv fYa 'Qvu QVA II ;:)g;:)SV sv;:) 'gjefju 'gjirji;:)o 'yVi;:)'Qvuaj ;:)SV ;:)0 'iarYarV1V II ;:)o'cm 'l VyoVrvj ;:)0 'jrjian g;:) ,!jeh I ;:)g;:)SV wVIgj ;:)n'Fac!ja}Yi '<;!v;) ;:) 'x}Y;)b Frehrjah 'bvYI}v II ;:)g;:)SV ;:)rV ;:) '<;!v;)j cm I biepn <;!v;:)o'vi ;:)'<;!vuaj ;:) 'agVwa1 I gYfY;:)bi i 0 'jrjian ;:) bY;:) Fregjrj;:) <;!v;:)n 'yjerj!j II

Orthographic versionBhf an ghaoth aduaidh 's an ghrian ag aragaille cheile Ie fail amach ce acu den bheirt a bathreise nuair a thainig taistealaf an bealach agus cl6ca te a chaitheamh aige. D'aontaighsiad gur an duine 'ba threise an duine 'ba luaithe a thabhllirfeadh ar an taistealaf a chl6ca abhaint d6. Ansin sheid an ghaoth aduaidh comh trean is a thiocfadh lei the, ach da mheid a

sheid sf 'sea ba dhluithe a theann an taistealaf a chl6ca thart fa dJaobh d6, agus sa deireadhd'eirigh an ghaoth aduaidh as an iarracht. Ansin lonnraigh an ghrian go te agus bhain antaistealaf d6 a chl6ca laithreach bonn. Agus ar a' d6igh sin b'eigean don ghaoth aduaidh aadmhail gurbh f 'n ghrian a ba Jaidre den bheirt.

AcknowledgementsI am very grateful for suggestions and assistance in the preparation of this illuslration 10 Calhair 6Dochartaigh and Prionnsias 6 Nuall~in.


ReferencesNi CHASAIDE, A. (1977). The Laterals ofDonegal Irish and Hiberno-English: An

Acoustic Study. M.A. thesis, University of Wales, Bangor.Ni CHASAIDE, A. (1979). The laterals of Donegal Irish and Hiberno-EngJish. In 6 Baoill,

D. P. (editor), Papers in Celtic Phono'logy. Coleraine: New University of Ulster.6 SEARCAIGH, S. (1925). Foghraidheacht Ghaedhilge an Tuaiscirt. Beal Feirste

(Belfast): Brun & 6 Nuallain.

QUIGGIN, E. C. (1906). A Dialect ofDonegal. Cambridge University Press.SOMMERFELT, A. (1922). The Dialect ofTorr, Co. Donegal. Cristiania (Oslo): Jacob

Dybwad.

SOMMERFELT, A. (1964). Consonant clusters or single phonemes in Northern Irish? In

Abercrombie, D., Fry, D. B., MacCarthy, P. A. D., Scott, N. C. and Trim, J. L. M.

(editors), In Honour ofDaniel Jones: Papers Contributed on the Occasion ofhisEightieth Birthday 12 September 1961, 368-73. London: Longmans.

Japanese

HIDEOOKADA

6-29-22 Sakuradai, Nerima-ku, Tokyo 176-0002, Japan

The style of speech illustrated is that of many educated Japanese brought up in Tokyo orother areas with similar pitch accent systems. The transcription is based on a recording of a25-year-old student whose speech is typical of speakers of his age group with thisbackground.

Consonants

Bilabial Labio- Dental Alveolar Post- Palatal Velar Uvular Glottaldental alveolar

Plosive p b t d k 9

Affricate tS

Nasal m n N

Flap rFricative S Z h

Approximant J W

p paN 'bread' t taijo: 'the sun' k kaze 'wind'

b baN '(one's) tum' d daNdaN 'gradually' 9 gaito: 'cloak, overcoat't' t'u:t'i 'notice' igaku 'medicine'

m maZll 'first' n mini 'what' N zeNi 'goodwill'

r raN 'orchid' (See also Ib, p, d, r/.)

s su:si 'numeral' h hana 'nose'

z zat'uzi 'chores'

jama 'mountain' w wa (particle)mjakU 'pulse'

Vowels• u

fnri 'meaning'e erne 'smile" (archaic imperative)

O·a ama 'woman diver (for abalone)'0 6mo '(sur)face' (archaic)u umu 'suppurate'


VoweI lengthozisantari 'uncles'hodo 'degree, extent'

ozi: santatSjhodo:

'grandfathers''sidewalk'

Pitchhasihasfhasi

accent'chopsticks''bridge''end'

hasj 0 nu(uhasf 0 nuruhasi 0 nu(u

'paint chopsticks''paint a bridge''paint the end'

ConventionsIp, t, t', kJ are only moderately aspirated. Syllable-final (moraic) forms of these and othervoiceless obstruents occur as the first P,art of geminates, e.g. /happo:1 'firing', Ijaltol 'atlast', Igakko:1 'school', /hossal 'attack (M disease)'. The geminate form of Irsl is Itrsl, as inIkutt'ukemasital 'pressed, attached'. Ibl is nonnally [b], but in rapid speech it may become[y] or [j3]. Igl tends to become [V] between vowels. Older and otherwise conservativespeakers use [I)] in all medial positions, except mainly for the element Ig61 'five' in numberdesignations, as e.g. Infzju:-g61 Lvfzu:g6] 'twenty-five'. Ir/, which corresponds to or' inRomanization, is postalveolar in place rather than retroflex and mainly occurs medially.Initially and after IN/, it is typically an affricate with short friction, [Q1J. A postalveolar [l] isnot unusual in all positions. Approximan( [J] may occasionally occur in some environments.It'l is normally [ts] but becomes [tel before iiI. Inl before iiI is prepalatal Lv]. Before luI, itmay be alveolar with some speakers. Consonants generally are strongly palatalized beforeiii, as Imil [m j i] 'body'. INI represents a moraic nasal with very variable pronunciation.Word-finally before a pause, it is typically a uvular nasal with a loose tongue contact or aclose to close-mid nasalized vowel. Such a nasalized vowel is also the normal realizationbefore a vowel or approximant, or before Ihl or lsI, as e.g. /zeNil [dze1i] 'goodwill',IhoNjaku/ [holjaku] 'translation', /zeNhaNI [dzerhaN] 'first half'. Before other consonants,it is homorganic with that following consonant, as e.g. IsfNpol [efmpo] 'progress', ISINnil[efj)ni] 'truly', ISIN[il [el!JQIi] 'truth'. lSI and /zl are [e] and [zl before Iii. /zl tends to be[dz] initially ([dZ] before Iii) and after IN/. /hI tends towards [~] and [cj>] before iiI and luirespectively. Ihhl is realized as [~~], [xx] or [cj>cj>] depending on the (normally identical)surrounding vowels. IjI affects the prec<;ding consonant as iiI does, and is itself absorbed,thus: ImjakUI [mjakU] 'pulse', It'jal [tea] 'tea', Isj6:1 [e6:] 'prize', lkanju:I [kaj)u:] 'joining'.It,dl before li,j/, normally limited to borrowings, remain essentially un palatalized, thusIti:pa:ti:1 [ti:pa:ti:] 'tea party', Idi(ekuta:1 [direkl}ta:] 'director', Idjda[ul [djdaru] 'dual'.Iwl has very slight or no rounding (except after 101), but involves no spreading either.

lui, resembling [w] auditorily, has compressed lips. so that it is unrounded but withoutspreading; it could be transcribed narrowly as [y] or [tV]. The slit between the lips may bevery narrow vertically and is generally much shorter in the horizontal plane than for [i]. Avery advanced variety in the [i] area may occur among the younger age groups. Except inaccented or lengthened syllables, Ii, ul tend to be devoiced [j, \I] between voicelessconsonants. As often as not, preceding fricatives replace them altogether. Final unaccentedIsu/ is very often reduced to [s].

Illustrations of the [PA 119

There are two lexically relevant pitch levels: high and low. Within a word, if the firstmora is high-pitched, the second is inevitably low-pitched, and vice versa. A moratranscribed with an acute accent, a, is said to be accented and is high. If more than one moraprecedes it, the high pitch extends towards the beginning of the word up to the secondmora. A word with no accent mark begins low and continues high from the second moraonwards. All moras following the accent are low until another accented or polysyllabicword is reached, when these accentuation rules reapply. Note that word pairs such as /hasiJ'bridge' and /hasi/ 'end' are both low-high when spoken in isolation, but will bedistinguished when something else follows. What follows will be low after the accent, butotherwise high. When a long (two-mora) vowel is transcribed with an accent, only the firstmora is high, and a pitch drop occurs between the two moras.

Transcription of recorded passagearutoki kitakaze to taijo: ga tSikarakurabe 0 simasita. tabibito no gaito: 0

nugaseta he: ga katSt to ju: kote ni kimete, mazu, kitakaze karahazimemasita. kitakaze wa, nani, hitomakuri ni site miseje:, to, hagesikuhukitatemasita. suruto tabibito wa, kitakaze ga hukeba hukuhodo gaito: 0

sikkarito karada ni kuttsukemasita. kONdo wa taijo: no baN ni narimasita.taijo: wa kumo no aida kara jasasii kao 0 dasite, atatakana hikar'i 0

okurimasita. tabibito wa daNdaN joi kokoromotsi ni natte, simai n'i wagaito: 0 nugimasita. sokode kitakaze no make ni narimasita.

Romanization (Hepburn system)Arutoki Kitakaze to Taiy6 ga chikara-kurabe 0 shimashita. Tabibito no gait6 0 nugaseta h6ga kachi to yu koto ni kimete, mazu Kitakaze kara hajimemashita. Kitakaze wa, 'Nani,hitomakuri ni shite misey6', to, hageshiku fukitatemashita. Suruto tabibito wa, Kitakaze gafukeba fukuhodo gait6 0 shikkarito karada ni kuttsukemashita. Kondo wa Taiy6 no ban ninarimashita. Taiy6 wa kumo no aida kara yasashii kao 0 dashite, atatakana hikari 0

okurimashita. Tabibito wa dandan yoi kokoromochi ni natte, shimai ni wa gait6 0

nugimashita. Sokode Kitakaze no make ni narimashita.


~ ~ B;j, ~tEB. C :*~b<:tJ < .; --< ~ L j; L t: o ~AQ)~~ ~ ~iJ<it t: 1Jilq~!> C ~,

?'::'cl;::~~L, j;T~t!f!b\';ji;~j;Lt.:o ~tmli, rt,l:l;::,-j;<l')I;::LL~it

.t ? J c, ~ L < ~ ~ n: L j; L t.: 0 1" ~ c ~A Ii, ~t!f!b<~ 1:1' Ii~ < Ii c~~~L .., iJd) c f;"q;:: < .., -:> It;t L t.: 0 ~ llii :*~ Q) ii I;:: t! l') ;t L t.: 0 *~ It~ Q) ~ ~ ,

t:.· b\ .; if L ~'l;l'i ~ tlJ L L Il!l iJ' t! JI::~z l') j; L t: 0 ~A Ii %t q .t ~,,c., b !> I;:: tJ. .., L,

L;t~";::li71-~~~~';tLt.: o i-'::'c·~tmQ)~I:tl;::t!l')j; Lt: o

Korean

HYjJN BOK LEE

Phonetics Laboratory, Department of Linguistics, Seoul National University,

Seou/151-742, Korea

The variety of Korean spoken in and around Seoul, on which the following phoneticdescription is based, is widely recognized as the standard language of the Korean peninsula.It differs from the speech of Pyongyang in North Korea, however, in phonetic featuressuch as vowel and consonant qualities, vowel length, accent, rhythm and intonation.

Consonants

Bilabial Labio- Dental Alveolar Post- Palatal Velar GlottaldenLal alveolar

Plosive p ph b t th d k kh 9

Nasal m n I)

Fricative IS Z h

Affricate C Ch j

LateralIApproximant

p pal 'sucking' t tal 'daughter' k kal 'spreading'ph phal 'arm' th thai 'riding' kh khal 'knife'

b bal 'foot' d dal 'moon' g gal 'going'

m mal 'horse' n nal 'day' I) bal) 'room'

s sal 'rice' h hal 'doing'

z zal 'flesh'

c cal 'squeezing'

ch chal 'kicking'

j ja! 'well'

I balain 'wind'

JIlustrations of the IPA 121

Vowelsa) MonophthongsKorean can be considered as having nine vowel qualities, which occur distinctively long orshort. The vowel II/J(:)I is usually pronounced as (we] in Seoul speech although there arealso words with lwei not derived from II/J/. In view of the overwhelming tendency of Seoulspeakers to pronounce a diphthong instead of II/J/, it may be more appropriate to postulate asystem of only eight vowel qualities in modem standard Korean. Vowel qualities areaffected to a great extent by vowel quantity. Long vowels are pronounced with a moreperipheral quality than the corresponding short ones, which are centralized or lowered.Long and short IA(:)I show the greatest quality difference: short IAI is a mid-open backunrounded vowel but long IA:I is realized as a central vowel (;):].

Short Vowels Long Vowels

ill:· • u:

ill • •U

• 0:.0

• A

a: •

zi'jal) 'hunger' 1: 'zi:pl) 'market'

e be'ge 'pillow' e: 'be:da 'to cut'

e the'jal) ·sun' E: 'the:do 'attitude'

a 'mal 'horse' a: 'ma:1 'speech'

0 bo'li 'barley' 0: 'bo:zu 'salary'u gu'li 'bronze' u: 'zu:bag 'watermelon'A 'bAl 'punishment' A: 'bA:l 'bee'

w 'A:lwn 'seniors' w: 'w:mzig 'food'

I/J 'gjo:hl/J 'church' I/J: 'I/J:thu 'overcoat''gjo:hwe 'we:thu

b) DiphthongsI j, wi are considered to be components of diphthongs rather than separate consonants.

Iwi I dwi 'back' Iwi I 'wiza 'doctor'Ijel 'je:zan 'budget' lwei gwe 'box'Ijel 'je:gi 'story' lwei we 'why'IjaJ 'ja:gu 'baseball' Iwal gwa:'il 'fruits'


Ijol

I'u/.IljA/

'gjo:za

ju'lijA'gi

'teacher'

'glass''here' IWI>J mWA 'what'

StressStress placement is predictable in Korean. In words of more than one syllable it is initialwhen the first syllable is a heavy syllable, i.e. one which either contains a long vowel or hasa syllable-final consonant because of the presence of an intervocalic consonant sequence.All other words of more than one syllable are accented on the second syllable. An accentedsyllable is longer and louder than an unaccented one.

1) Words with accent on the first syllable:'

a) long vowel in the first syllable:

I 'ga:Jal]1 'disguise', I 'zA:lil 'acting head'

b) closed first syllable:

I 'zanzul 'landscape', I 'chulgul 'exit'

c) long vowel and closed first syllable:

I'za:nzul 'arithmetic', I'ga:mzal 'thanks'

2) Words with accent on the second syllable:

Iga'Jal]1 'most', /zA'lil 'frost', /za'dalil 'ladder'

ConventionsIb, d, gl are voiceless unaspirated (or slightly aspirated) lenis plosives [1;>, 9, g] syllableinitially, but are regularly realized as voiced sounds in intervocalic position.

IJ I shows the same voicing variation. /zl is a Jenis voiceless alveolar fricative. It issometimes realized as voiced [z] intervocalically, especially when preceded by a nasal as inIga:mzal 'thanks' and linzal 'greetings' spoken in an informal style. Ip, t, k, c, sl arevoiceless unaspirated fortis sounds syllable-initially, produced with a partially constrictedglottis and additional subglottal pressure. Iph th, kh, chi are strongly aspirated voicelesssounds syllable-initially.

In syllable-final position Ib, p, phi are realized as a voiceless bilabial stop withoutplosion, [p'], Id, t, th, j, c, chi are realized as a voiceless alveolar stop without plosion, [f],and Ig, k, khl are realized as a voiceless velar stop without plosion, [k']. The four sonorants

1m, n, I), II are the only other consonants which can occur syllable-finally.Ij, c, chi are voiceless postalveolar affricates syllable-initially. Izi and lsi are realized as

alveolopalatal fricatives [eh], [e] when followed by Iii or a diphthong beginning with [j].Ihl is [9] before iii and [j], [x] before Im/, [fi] between voiced sounds and [h] elsewhere.Inl is [J1] before iii and [j], and [n] elsewhere. II/ is [r] intervocalicalJy, but [I] or [lJsyllable-finally, and [A] before iii or [j]; /Ill is [II] or [IA], according to the vowel context.

Illustrations of the lPA 123

Broad transcription of recorded passageba'lamgwa 'heooimi ZA'lo hi'mi 'dA 'ze:dago da'thugo i'srnl te, 'haona'grnoega ta'trnthao 'we:thulrnl 'ibgo 'gAllA wa'srnmoida. grn'drnlrnooU'gudrnnji oa'grnoerni 'we:thulrnl 'mAOjA 'bAdgiorno oiga hi'mi 'dA 'ze:dagoha'gilo 'gjAljAI)hesrnmnida. bugphuI)rno 'himkAd bu'IAsrnoa 'bu:lmjAo'bu:lsulog na'grnoeorno 'we:thulrnl 'daodaohi jA'mjAsrnmoida. grn 'tee'henoimi trn'gA un 'hed bichrnl ga'manhi oe'ljA 'cwe:oi oa'grnneorno'we:thulrnl 'Allrno bA'ZAsrnmoida. i'lihajA 'bugphuI)rno 'heooimi 'du:ljuI)ehi'mi 'dA 'ze:dago 'iI)JAI)haji a'ornl zu 'A:bzAsrnmnida.

Narrow transcription of recorded passagelta'famgwa 'henpimi ?-A'fO tyi'mi 'dA '?-e:dago qa'thugo i'srnl te, 'hanna'grnoega ta'trnthan 'we:thufrnl 'iltko 'g~:fA wa'srnmpida. grn'drnfrnooU'gudrnnji na'grnnerni 'we:thufrnl 'mAOjA 'gA9kinrnn piga ~i'mi '9A '?-e:dagoha'gifo 'gjAlcAI)hesrnmpida. ltugphUI]rno 'tyimkAq QU'fAsrnna 'ltu:lmjAo'Qu:lsufog na'grnnenrnn 'we:thufrnl 'qandanfii jA'mjAsrnmpida. grn 'tee'heopimi trn'gAun heq'pichrnl ga'manfii ne'fjA 'cwe:pi oa'grnoenrnn'we:thufrnl 'AHrnn QA'ZAsrnmpida. i'fihajA 'ltugphuI)rnn 'henpimi \lu:lcuI]etyi'mi '9A '?-e:dago 'iI)JAI]haji a'nrnl su '~:QsAsrnmpida.


1l]-~.:i!J- ~'i:lo1 ....~.£. ~01 t:~ ,\]142 4.!f-2 ~~ all, ~ y..::z.1.117]- ttj-~~ ~.!f-~ ~

2 ~~ ~irY4 . .::z.%~ i=-T~A1 y..::z.l.i1GJ ~.!f-~ 'tl~ ~71~ OJ7]- ~oJ t:i ,\]142 i>]-71.£. 7J~~irY4. *%~ ~~ ~~~y. ~~ ~?~ y..::z.l.iJ~ ~.!f-~ ~~i;J

aj'Jiiryt:}. .::z. allolJ ~'i:loJ E:.7j-&- ~~~ 7]-~i>J t.ll~~y, y..::z.1.11~ ~.!f-~ ~~

~~iryt:}. ol~"&-raj *~~ ~'i:l°1 %~oll ~01 t:i ,\]1t:}2 ~~"5}A1 ~% ? ~~

iryt:}.

Korean phonetic transcription

1::1 I- c I- 0 "1"T I- "5 HL LID I A -1 c..L "5 J 0 I t: -1 A ~I : ll.: 1-"1..L t: I- E T"1..L I J.I. u E

n:H. "5I-L LI-"1uL~I"11- n:1-n:uEI-L "T~I:ETcuE Itln..L "l4':c-1"TI-J.l.UDLIll.:I-. "lull.:ucuL LT"1Tll.:uLAI LI-"1uL~Jul "T~I:ETcuE D-1LA-1 1::1

-1 en 1 L u L L I "1 I- "5 I 0 I c -1 A ~I : ll.: 1-"1..L 0 I- "1 I c..L "l t -1 EJiU 6"5 HJ.l.u 0 L

Ill.: 1-. tlT"lliT6uL "5ID1J-1c tlTc-1J.1.uLI- I::IT:EDt-1L tlT:EJ.l.Tc..L"lL I-"l u L ~I L u L "T ~I : ETc u Eel- L ll.: I- L ~ I t -1 0 t i J.I. u 0 L I ll.: I- . "l u n: H ~I

oHLLIDJ n:u"1-1TL ~Ht:tIllxuE "ll-DI-:L~l LHcl-1 »:"T~I:LI. LI"1uL~ILuL "T~I:ETcuE -1EEuL tl-1A-1J.1.uDLIll.:I-. Ic1"5l-l-1 tlT"lliT6 u L 0 H L LID I C T : E »:T6 ~I 0 1 0 I t: -1 A ~I : ll.: 1-"1..L I LA -160 I- A J I

LuE J.l.T 4': tlJ.l.-1J.1.uDL Ill.: 1-.

Persian (Farsi)

MOHAMMAD-REZA:MAnDI AND ELMAR TERNESI,

Institut fur Phonetik, Allgemeine Sprachw~senschaft und Indogermanistik, Universitdt Hamburg,Bogenallee 1I, D-20144 Hamburg, Germany

The style of speech illustrated is that of many educated Persian speakers in the area ofTehran. It is based on a recording of a 45-year-old male speaker.

Consonants

Bilabial Labiodenl. Dental Postalv. Palatal Velar Glottal


Nasal m n

Fricative f V S z J 3 X Y hAffricate tJ d3

Trill rI


1Approximant

p prer 'feather' t tir 'arrow' k kur 'blind'

b brer 'fruit' d dir 'late' g gur 'grave'm nom 'name' n non 'bread' x xrem 'bend'f nof 'navel' sir 'garlic' y yrem 'sorrow'v nov 'warship' z zir 'below' ? tre?sir 'impression'

siri 'satiety' J fof 'urine' h kuh 'mountain'

sili 'slap' 3 303 'idle talk' j jek 'one'

tJ tfire 'victorious'

d3 d3ire 'ration'

Vowels

udir 'late'

e del 'heart'

0 re drer 'door'

0 dDT 'gallows'

0 por 'full'

u dur 'far'0

StressStress is distinctive: ['mrerdi)

[mrer'di)mrerdimrerdi

'a man''manhood'


Conventions

Ip, t, kJ are strongly aspirated in word-initial position, in other positions slightly aspirated.

Ib, d, gl and lv, z, 3, Vi are slightly devoiced in word-final position. IVI is [G] in wordinitial position, after nasals, and when geminated; otherwise it is postvelar [y). 1m, n! are

voiceless in word-final position after a voiceless consonant. 1m! is [fl)) before If, vi and In!

is [!)] before Ik, g/, and [~] before IV/. /rl varies between [c) and [J]. Ivl is [w] after 101,

otherwise it is [v].

Vowel length is not distinctive, but all vowels are rather long in stressed position. Ii, 0,

ul are, in general, somewhat longer than Ie, re, 0/. Ie, re, 01 are long or half-long beforeconsonant clusters. Before or after nasal consonants, vowels are nasalized. Word-initialvowels are preceded by [7], vowels in hiatus are separated by [7]. 101 is underrounded.Articulatory positions of the other vowels are as on the chart.

Transcription of recorded passage'jek 'ruzi 'bode Jo'mol bo xor'Jid bohrem dre?'vo 'mikrerdrend, ke 'ojoko'domjeki yrevi'trerrest. drer 'on 'hin moso'feri re'sid, ke lrebbo'deje ko'loftibe 'dovre 'xodeJ pitJi'de bud. on'ho yre'mr 'gozoJtrend, 'hrerko'domeJon, ke'revvrel 'betrevonred moso'ferm mred3'bur konred lrebbo'dereJm 'bekrenred,mre?'lum 'miJrevred, ke 'zurreJ biJ'trerrest. 'bode Jo'mol to trevonest vre'zid,'remmo 'hrertJe biJ'trer vre'zid, moso'fer lrebbo'dereJm biJ'trer 'dovre 'xodeJ'd3rem? krerd. oye'bret 'bode Jo'mol xres'te Jod vre 'drest 'brerdoJt. 'breedxOr'Jid to'bid, vre hre'vo in'yredr 'grenn Jod ke fov'n moso'fer lrebbo'dereJm'dremvrerd. pres 'bode JO'mol mred3'bur Jod ey'mr konred ke xor'Jid 'zurreJbiJ'trerrest.


~~ ~ ~I~ ~I o.S "';';..l?w-o Iy..l ~ ~ ~...u-:" ~ JL.,...:;:. ..l~ tSjJ..) ~

...ly o~ ,p..ly:,...))..l ~ u-:;.ili :..lW o.S '~..) tS~~ 0-:!-=>- (.J..)..l .~I

I..) ,plo..lW ~..)~ I..)~~ ...;.;I~ J)-I o.S u~I~..r'" ,~I~..)I"; L....,,;I

~..r'" L:I . ..l..:-.jJ ..:.:......u..;ly l.:; JL.,...:;:. ..l~ .~I~ ,p..)Jj o.S ..ly;,w-o I'y-L>-.o '~

JL.,...:;:...l~ ~l..> ...l? F ,p..ly:,...)J..l~ I..) ,plo..lW~~ '..l..:-.jJ~

tS..)~ o.S ~ f'.?..)~lly) ,~l.:; ~..)y:,. ~ ..::......:i,I..l.r. ~..lJ ~ ~

~..)y:,. o.S ~..)I";I ~..)~ JL.,...:;:...l~ ~ . ..l..))T..)..l I..) ,plo..lW~~

.w~l~ ,p..)Jj

Portuguese (European)

MADALENA CRuz-FERREIRA

Department ofEnglish lAnguage and Literature, National University ofSingapore,

10 Kent Ridge Crescent, Singapore 119260

The following illustration of European Portuguese is based on the Lisbon dialect. The textis transcribed from a recording made by a female native speaker of Lisbon in her midforties, speaking in a style that may be described as educated colloquial. All phonemes ofEuropean Portuguese are illustrated in the passage unless otherwise noted.

European Portuguese comprises st1veral regional accents, including those of thearchipelagos of Madeira and the Azore.s. The Lisbon accent is one of the two nationalstandard varieties in mainland Portugal, the second being that spoken in the area of thecity of Coimbra.

The language is characterized by a velarized resonance superimposed on both vowelsand consonants, in that the usual posture of the tongue body is retracted and raised. Inaddition, an overall rather lax articulation further affects vowel quality and voicing, aswell as the articulation and voicing of voiced consonants. These features are not, it shouldbe pointed out, exclusive to informal or careless speech and, together with specificprosodic patterns, they constitute the most striking characteristic distinguishing Europeanfrom Brazilian accents of Portuguese.

Consonants

Bilabial Labio- Dental Alveolar PaJato- Palatal Velar Uvulardental alveolar

Plosive p b t d k 9

Nasal ill n Jl

Fricalive f V S z J 3 B'

Tap [

Lateral1 ;:.Approximant

p 'patu pato 'duck' (m) 'tatu tacto 'tact' k 'katu cacto 'cactus'

b 'batu hato 'I strike' d 'datu data 'I date' g 'gatu gato 'cat' (m)

m 'malu mato 'Ikill' n 'natu nato 'innate' (m) J1 'piJ1E pinha 'pine cone'

f 'fatu Jato 'costume' s 'kasu caro 'I hunt' J 'Jatu chato 'flat' (m)

v 'ViJ1E vinha 'vine' z 'kazu caso 'I marry' 3 '3atu jaclo 'jet'

llluslralionsojlhelPA 127

rator 'picB

'lipB

pira 'pyre'

linlw 'line' IA

K 'Katu

'piAB pilha

'mouse' (m)

'battery'

VowelsEuropean Portuguese has 14 monophthongs, 9 oral and 5 nasalized. These vowels areusually described using four values for height and three for backness. Their positions areshown on the chart below according to the values widely agreed on in the literature, withthe exception of the vowel Im/. This vowel, which occurs only in unstressed syllables, isoften represented as l'dl but does not correspond to the mid central quality associated withschwa. It is a fronted and lowered high back unrounded vowel; hence the symbol chosenhere for its transcription. The chart represents target articulations of the vowels, which arenormally 'undershot' in connected speech, resulting in more centralized qualities.

Oral vowelsvi VI 'saw' (I sg)

e ve ve 'see' (3 sg)

e se se 'cathedral'a va va 'go' (3 sg)

::> s::> s6 'alone'0 so sou 'I am'

u, u.u 'mudu mudo 'mute' (m) •B pB'gar pagar 'to pay'm pm'gar pegar 'to grip' 0, o.

Nasalized vowels

I VI vim 'came' (I sg):) .

e 'etru entro 'enter' (I sg)

E 'Etru antro 'den'

a s6 som 'sound' (n)

G 'mGdu mundo 'world'

DiphthongsThere are 14 diphthongs, 10 oral and 4 nasalized, as shown below. The diphthongs lei, ::>i,ui, elll do not occur in the transcribed passage.

ei B'neif aneis 'rings' (n)

ai sai sai 'go out' (3 sg)

Bi sBi sei 'know' (I sg)

::>i m::>i m6i 'grind' (3 sg)

oi 'moitB moita 'thicket'


ui l3'nuiJ anuis 'agree' (2 sg)

iu viu viu 'saw' (3 sg)

eu meu meu 'mine' (pass m)

eu veu Vlfu 'veil'

au mau mau 'bad' (m sg)

f!i SRi cem 'hundred'

oi l3'noif anoes 'dwarves' (m)

Oi 'mOitl3 muita 'much, many' (I)

BU miiu mao 'hand' (n)

ConventionsExcept in word-initial position (and after nasalized vowels), the voiced plosivesIb, d, gl are noonally pronounced as the fricatives [~, 5, V]. If, 31 are weakly fricated insyllable-final position. Syllable-final If I occurs as [3] before a voiced consonant (exceptbefore 131 itself, where it is deleted), and as [z] before a syllable-initial vowel both within

and across word boundaries, as in [k1)al dU3 doiz '£f::>] qual dos dois era 0 'which of thetwo (m) was the (m)' in the transcribed passage. III is veJarized in all its occurrences.

If I does not occur in word-initial position, and Ifl, 1.1 only occur initially in a fewborrowed words, and in the case of II.I also in the clitic pronoun IAwl lhe 'to him/her/it'and all the foons in its paradigm.

All vowels have lower and more retracted allophones before /II, and higher and moreadvanced allophones before alveolar, palato-alveolar or palatal consonants. IUJI andunstressed 113, ul are voiceless in word-final position.

The end-points of the diphthongs, although transcribed [i] and [uL tend to be more

central and, in the case of [uL Jess strongly rounded than these symbols suggest. The

diphthong Il3il in the Lisbon accent is tending towards a retracted onset [~iJ or [AiJ. Insome words the offglide has been altogether lost, as in the pronunciation of the wordIpl3ifwl peixe 'fish (n)' as [pI3D.

In a vowel sequence in which the first vowel is an unstressed Iii or lui and the second astressed vowel, the first vowel may lose it syllabicity, e.g. l!fi'al/ real 'real' can bepronounced as a monosyllable, [!flal], and Imu'idl31 moida 'ground' (adj, f) as ['m1)idl3].

ProsodyLexical stressLexical stress is distinctive. Stress provides very productive class-changing contrasts, suchas those between nouns and verbs in pairs like ['duvidl3] duvida 'doubt (n), [du'vidl3]duvida 'doubt (v, 3 sg)'. Less frequently, words in the same class contrast in stress, e.g.the nouns ['tunel] tunel 'tunnel', [tu'nel] tonel 'wine cask'.

lllustrations of the /PA 129

Most Portuguese words are stressed on the penultimate syllable, although lexical stressmay fall on any of the last three syllables of the word, and even on the fourth from last inthe case of verbal forms with enclitic personal pronouns such as [JB'mavBmuftw],chamcivamos-te 'we called you (sg)', [BkB'bavBsw,(w) acabava-se-Ihe 'he/she/it ran outof'. Syllables with diphthongs not bearing primary stress are assigned a secondary stress,e.g. [,au'tetiku) autentico 'authentic (m sg)'.

Rhythm, vowel reduction and devoicingEuropean Portuguese is a stress-timed language with vowel reduction in unstressedsyllables. In connected speech, unstressed vowels are either centralized or altogetheromitted, and consonants and vowels in unstressed positions may be devoiced (Mateus

1975; Willis 1971). Examples in the transcribed passage include /e'voltu/, phonetically[tV'volt], envolto 'wrapped (m)' and /dwzif'tiul [cJ~ftiu) desistiu 'gave up' (3 sg), where apotentially trisyllabic word is reduced to a monosyllable with a complex onset.

IntonationPortuguese intonation can be analyzed as a set of nuclear tones associated with phraselength intonation groups. The nuclear tones fall on a stressed syllable, typically the laststressed syllable in the intonation group. If the nuclear syllable is not the last syllable inthe group, the nuclear pitch movement continues on any following weak syllables.Normally, such weak syllables are only pronounced when required for completion of an

intonational pattern.There are six main nuclear tones in Portuguese, three falling, two rising, and a rise

fall. They are transcribed here with 'tone letters' placed before the nuclear syllable. The

low fall (-J) is the neutral tone for statements and question-word questions; a high fall ('l) isgenerally associated with emphasis; an extra-low fall is used in exclamations and

commands. The low rise (-') is the typical tone for yes/no questions, and markscontinuation or incompleteness in discourse. The high rise (1) is associated with echoquestions or with questions requiring repetition of a previous utterance. The rise-fall ("l)

typically marks implication and reserve. Only the low fall and low rise occur in thetranscribed passage. Cruz-Ferreira (1998) provides a more comprehensive outline of theprosodic system.


The transcription provided below is narrow in certain respects, and includes therepresentation of sandhi phenomena (Herslund 1986), for example, in line I where ('£f~]

is the contraction of l'efB u/, and of pitch excursions. Unstressed syllables which maintain

their target vowel qualities are not transcribed with a secondary stress; secondary stress is

shown only with diphthongs. Upstepped syllables are followed by low or falling syllables,downstepped syllables by low or rising syllables. Since upstep and downstep do notnecessarily occur on a stressed syllable, all primary stresses not indicated by a tone markare shown in the transcription.


u i'vetU 'mrt ii u s::>I 9fku'ti,BU ik1}aI dU3 doiz i'eD maif Abrtw I i'k~Bdu

susw'deu p-e'sar G Vi-ei'3iHw lip'~olt num-e --Ikap-e II au Avelu I 'p6i i,iiisdi-eJ.'kordu iii 'komu -ei'kel kw pri~'m-eiru k6s'gis ::>bri'gar u vi-e'3Btw -e ti'rar-e Akap-e I sri-e k6sidw i'radu u '\laif --Ibrt II u i'vetu 'n::>rtw kumw'so -esu'prar k6 'mGit-e Afuri-e I m-ef i'k~~tu maif sillprav-e I maiz iU vi-eJ.'3Btw si-ek6fJ.'gava su-e --Ikap-e I -e'te ikiu 'vetu 'n::>rtw --I9~ftiu II iU s::>l bri'A:o e'tiiu k6'todu fPleAdor I i iimwdiat-e'met u Vi-e i'3Btw ti'ro -e --Ikap-e II u i'vetu 'n::>rtwtev -e'sl dw ~kuJ1wi'ser -e supwriuri'dad idu --Is::>I II

Orthographic versiono vento norte e 0 sol discutiam qual dos dois era 0 mais forte, quando sucedeu passar urn

viajante envolto numa capa. Ao ve-Io, p6em-se de acordo em como aquele que primeiroconseguisse obrigar 0 viajante a tirar a capa seria considerado 0 mais forte. 0 vento norte

come~ou a soprar com muita furia, mas quanto mais soprava, mais 0 viajante seaconchegava a sua capa, ate que 0 vento norte desistiu. 0 sol brilhou entao com todo 0

esplendor, e imediatamente 0 viajante tirou a capa. 0 vento norte teve assim dereconhecer a superioridade do sol.

AcknowledgementI am indebted to Professor Paroo Nihalani for his kind help in acquainting me with the implementation of

the IPA font.

ReferencesCRuz-FERREIRA, M. (1998). Intonation in European Portuguese. In Hirst, D. and di

Cristo, A. (editors), Intonation Systems: A Survey ofTwenty Languages. Cambridge

University Press.HERSLUND, M. (1986). Portuguese sandhi phenomena. In Andersen, H. (editor), Sandhi

Phenomena in the Languages of Europe, 505-18. Amsterdam: Mouton de Gruyter.MATEUS, M. H. M. (1975). Aspectos da Fonologia Portuguesa (Publica~6es do Centro de

Estudos Filol6gicos 19). Lisbon: Instituto de Alta Cultura.WILLIS, R. C. (1971). An Essential Course in Modern Portuguese, revised edition.

London: Harrap.

Sindhi

PAROO NIHALANI

Department of English Language and Literature

Oita University, 700 Dannoharu, Oila-City 870-/192, Japan

The variety of Sindhi described here is roughly representative of the Vicholi dialect,which is considered to be the 'Standard variety'. This variety has special prestige attached

to it because the most celebrated poet of Sindhi, Abdul Latif, wrote the classic 'Shah joRasalo' in this dialect. The description, based on the author's own speech, has been kept

fairly close to the colloquial style. The transcription is based on a recording of a male firstgeneration speaker of Standard Sindhi who grew up in Sindh, Pakistan, before moving toIndia in 1947. For a detailed discussion of Sindhi segments see Khubchandani (1961).

Consonants


p b t d t ct k 9Plosive

ph bfi th dfi th ctfi kh gfi

6 cf i 9Implosive

C jAffricate

Ch jfi

m n 11. p I)Nasal

mfi nfi 11.fi

f s z ? X Y hFricative

r (Tap (fi

U jApproximant

Lateral 1Approximant Ifi

Pb

p;mu 'leaf'

butu 'shoes' d

t;}ro

dunu

'bottom'

'navel'tq

'cap'

'fear'


ph phut~ 'rift' th thali 'plate' th t~ru 'be cool'

bfi bfiul;} 'mistake' dfi dfiar;J 'separate' ct.0 ct.fi~ku 'cover' (0)

5 5aru 'child' 6 6aru 'crevice'

m m~lu 'pitcher' n nalo 'name' II m;'}l1i 'diamond'mO mOe 'buffalo' nfi s;'}noi 'thin' llfi ma'l."u 'person'

f f;'}rasi 'bed cover' s sufu 'apple' ~ ~e 'thing'

z zal;'} 'wife'

rolu 'wanderer' [ phO[O 'sore' (0)

[0 porno 'old man'

\J \Jaru 'hair' I limo 'lemon'lfi thulfii 'fat' (adj, fern)

c calu 'cunning' k kano 'straw'

j juto 'shoes' 9 gano 'song'

ch chati 'breast' kh khano 'drawer'jO j°;'}tI 'immediately' gfi gOO[O 'horse'

} faro 'cobweb' 9 g~ro 'heavy'

Jl };'}Jl~ 'marriage I) C~I)O 'good'

procession' x xofu 'fear'

¥ ¥;'}mu 'sorrow'

jaw 'friend' h haru 'necklace'

VowelsSindhi has a system of ten oral vowels.

sir;'} 'midstream'SIr;'} 'brick'

(a measure of weight)u.

e ser~

• ue ser~ 'walks' (0) o.a sar;'} 'jealousy'

~ s~r~ 'funeral'

c;'}oondo 'you will say';) .

0

c~wndo 'he will say'g

0

u sur;'} 'tunes' • a

u sur;'} 'aches and pains'


The vowels lei and I'JI tend to be diphthongized, as [e;}] and [;}u]. These vowels have alimited distribution, nOl occurring before aspirated stops, implosives, nasals or Iy/.Vowels are audibly nasalized preceding a nasal consonant, but, as in many other modernIndO-Aryan languages, there are a few words in Sindhi with distinctive nasalized vowels.Compare I;}sil 'eighty', l;}s11 'we'; ladfiil 'half-rupee', ladfiil 'storm'; lo;}hil 'yoghurt',I<f;}hil'tenth'. For this reason, vowel nasalization must be regarded as phonemic, and ithas been marked in the transcription of the passage below in positions where it is notpredictable from a preceding nasal.

StressWord-level stress is nondistinctive, and is fixed on the first syllable of the morpheme.Contrastive stress may be used to give special emphasis to a word or to point to a contrastof ideas.

ConventionsThe consonants in the post-alveolar column of the chart are pronounced as apical postalveolars (Nihalani 1974b). They have been transcribed, following tradition, with symbolsfor retroflex sounds. The consonants in the palatal column are laminal post-alveolars.Ie, ch, 1, lfil have an affricated release of relatively short duration. The alveolar nasal In!occurs mostly before dental or alveolar consonants and is pronounced homorganic withthe following consonant. [v] is in free variation with [w] as the realization of Ivl.

Breathy-voiced consonants, transcribed with [fi], are produced with the vocal foldsincompletely closed at the centre and with the arytenoid cartilages apart (Nihalani 1974a).Of the rhotics Ir, [, [fi/, Irl is realized most often as a tap but may be a trill with two or

more contacts, Ie, enl are typically flaps.Vowels in open syllables tend to be longer than those in closed syllables. I;}, I, ul are

shorter than other vowels. Stops in intervocalic position preceded by these vowels tend tobe longer than those preceded by other vowels, e.g.:

[mud:o]

[~t:o]

[pIt:;}]

'period'

'leaf'

'curse'

[mado]

[pato]

[moto]

'capability'

'worn'

'fat' (adj, male)


ut~rd ji hdua e SljU paIl,d me jfi~g[O bre rdhja huja td 5mfi i me keruu~dfiib tabtuaro ahe. etre me hIku musafIru gdrdm kotu pae uta dci

IdI)gfijo. hdua e SIp 5rn fi i bbul k~jo td jeko musafIrd jO kotu lahraesdgfidndo uho i oodfiib tak~tuaro lekhjo uendo. po ut~rd ji hdua cfaqfio zor~

sa Idgi, pdrd jetro oodfiib ZOrd sa Idgi otro oodfiib zord sa musafIrd paIld


khe qfi;}kI'1.g ji kO~I~ k;}ji. aXlf me Utgfg je hgua kO~I~ chgcfe cfmi. po SIjg ji

Usg ZOf;} sa mkti E mUSaflf;} pahIJljO kotu pldi lahe p;}re phIti k;}jo. mhid

kdfe Utgfg je hgua mdJ1jo td SIjU hi 16mfii me Uddfiikd tabtuaro ahe.


l;> ~J C!/ ~k ~ !J~7

t!~J J'';------""'~ /' 3,/ I~ \.b .-5<:.>- '.) ~ ~ 0 ?

~ ~ /5"" ~ u.' jlJq ~~ ~ ~/ .w,~

/

ReferencesKHUBCHANDANI, L. M. (1961). The Phonology and Morphophonemics ofSindhi. M.A.

dissertation. University of Pennsylvania. Philadelphia.NIHALANI, P. (1974a). An aerodynamic study of stops in Sindhi. Phonetica 29. 193-224.NIHALANI, P. (1974b). Lingual articulation of stops in Sindhi. Phonetica 30. 197-212.

Slovene

RASTISLAV SUSTARSIC, SMIUANA KOMAR

Department of English, Faculty ofArts, University ofLjubljana,

AIkerleva 2, 1000 Ljubljana, Slovenia

AND

BOlAN PETEK

Faculty ofNatural Sciences and Engineering, University of Ljubljana,

Snelnifka 5, 1000 Ljubljana, Slovenia

Slovene (or Sloven ian) is the national language of the Republic of Slovenia and is alsospoken in adjoining areas of Austria, Hungary and Italy. The variety described here isStandard Slovene as spoken by educated speakers in Slovenia. The transcription is basedon a recording of two speakers, a female and a male, from Ljubljana, the capital.

Slovene has 21 consonants and 8 vowels. The orthography generally represents thesegmental pronunciation quite faithfully. The main exceptions are that [u] can berepresented by the letters 'I' and 'v' as well as 'u', and that 'e' is used to represent [e, E,

d] and '0' to represent [0, ;)]. Note also that 'Jj' and 'nj' are pronounced as [I] and En]unless followed by a vowel.

Consonants

Bilabial Labio- Dental Alveolar Palato- PalaUll Velardental alveolar

Plosive p b t d k 9Affricate ts tJ d3

Nasal m n

Tap r

Fricative f S z J 3 X

Approximant U JLateral Approximant 1

p 'pi:ti pili 'to drink' t ti:sk lisk 'print' (n) k ki:p kip 'statue'b 'bi:ti bili 'to be' d di:sk disk 'disk' 9 gi:p gib 'movement'

ts tsi:n tsin 'tin' (metal) tS tSi:n tin (anny) 'rank'

d3 d3i:n gin 'gin'


m 'mi:ti mili 'myths' n 'ni:ti

fi:n fin 'fine' (adj) s 'si:nitiz 'zi:niti

r 'ri:tiv 'ui:diJ vidis 'you see' 'li:ti

nili 'threads'

s.inili 'to shine out' f 'fi:la sila 'awls'

zinili 'to open 3 '3i:la fila 'vein'

one's mouth'I ••

'to dig' 'xi:ti hili 'rush' (imp)nil x/iIi 'to pour' 'ji:diJ jidis 'Yiddish'

Before voiceless obstruents and in word-final position (unless they are followed by aword-initial voiced obstruent) voiced obstruents are completely devoiced: [sla:t] slad'malt', sladkor ['sla:tbr) 'sugar'. Compare: [gra:d g:>'ri:) grad gori 'the castle is on fire'.Voiceless obstruents become voiced before voiced obstruents: [le:sJ les 'wood', [le:z

g:>'ri:] les gori 'wood burns'.Is, z, tsl when followed by If, 3, tf, d31 assimilate or coalesce to U, 3, tJ] respectively,

e.g. prefix Is-I in rJtfa:s:>maJ scasoma 'in the course of time'; lizl 'from' but [i'3:e:paJ izfepa 'from the pocket'.

The nasal Inl is pronounced as velar [l)) before Ik, g, xl: ['ba:l)kaJ banka 'bank',['a:l)gelJ angel 'angel', ['a:l)xovoJ Anhovo (name of a town). The bilabial and alveolarnasals, 1m, nl, are pronounced as labiodental [ll)) before If, vi: [sill)b'ni:ja) simfonija'symphony', [s;:lll)'ue:deu) sem vedel 'I knew', [ill)f:>r'ma:tsija) informacija 'information',

[ill)va'li:d) invalid' invalid (n)'.The sonorants 1m, n, 1, 1', j, vi are normally voiced, but the approximant Ivl has four

allophonic variants: preceding a vowel it is pronounced as labiodental [uJ: ['v:>:da) voda'water', in final position or preceding a consonant it is pronounced [uJ: [siuJ siv 'grey',['brautsi) brald 'readers', in syllable-initial position preceding a voiced consonant it is avoiced labial-velar approximant [w): [wnu:k) vnuk 'grandson', [:>d'wze:ti) odvzeti 'take

away' and in syllable-initial position preceding a voiceless consonant it is a voicelesslabial-velar approximant [MJ: [Msa:k) vsak 'every', [pret'Mse:mJ predvsem 'in particular'.Instead of [w) and [MJ, the pronunciation can also be [uJ.

VowelsTraditionally, the vowel system in Standard Slovene was described as including a vowel

length distinction, with Ii, a, u, e, :>1 occurring both long and short, Ie:, 0:1 only long andl;:ll short (ToporiSil:: 1984). Vowel length, however, can no longer be regarded asdistinctive for most speakers, and it is generally accepted that long vowels occur in

stressed, and short vowels in unstressed position (Srebot-Rejec 1988), giving a phonemicinventory of eight distinctive vowel qualities, Ii, e, e, ;:l, a, 0, :>, uf, as shown on the chart.

All vowels occur in stressed and unstressed syllables, but lei and 101 occur in unstressedposition in a few grammatical words only, e.g. the auxiliary bo ('will') in the recordedpassage. Because of this restriction, some authorities consider that there are only sixunstressed vowel qualities. What is meant by 'stress' will be discussed below.

Illustrations a/the IPA 137

Stressed Unstressed

i: mi:t mit 'myth' 'mi:ti miti 'myths'

e: me:t med 'honey' e 3e've: ie ve 'already knows'

e: 'pe:ta peta 'heel' e 'pe:te pete 'heel' (gen)

a: ma:t mat 'checkmate' a 'ma:ta mala 'checkmate' (gen)

::l: 'p::l:t~n poten 'sweaty' ::l p::l'te:m potem 'then'

0: po:t pot 'path' 0 bo 'J1o: bo slo 'will go'

u: pu:st pust 'carnival' u 'pu:stu pustu 'carnival' (dat)

~: p~:s pes 'dog' ~ 'do:~r dober 'good'

Phonetic diphthongs arise when the approximants Ivl and Ijl are preceded by a vowel andfollowed by a consonant or a word boundary. The labiodental Ivl in these positionsbecomes a rounded second element of a diphthong, i.e. [u], and Ij/ becomes [i]. Thesephonetic diphthongs do not occur with salient lengthening when they occur in stressedsyllables. The degree of openness in the first element in leil is between that of lei and lei,and for lou! it is between that of 101 and hi.

1U piu pi! 'drank'ei glei glej 'look' (imp) eu peu pel 'sang'

eu leu lev 'lion'ai dai daj 'give' (imp) au pau pav 'peacock'oi looi tvoj 'your' (masc sg) ou pou pol 'half'::li b::li boj 'battle'ui tui tuj 'foreign'

~u b:V<lu topel 6 warm'

Stress and accentIn the illustrative wordlists and the transcribed passage below, certain syllables have beenmarked as stressed. These stressed syllables are characterized by greater duration as wellas by amplitude and pitch prominence. Non-compound words have no more than onestressed syllable. Stress placement is not predictable but is also rarely distinctive inSlovene, although in a few instances different forms of the same noun or verb differ only


in stress placement, e.g.: ['stua:ri], [stua'ri:], stvari ('things', dual vs. plural), ('no:sim::>],[n::>'si:m::>], nosimo ('can)" 1st p. pI. indicative and imperative).

Standard Slovene distinguishes two pitch accents, traditionally referred to as the'acute' and the 'circumflex' accents, and:these accents are tied to the location of stress.The pronunciations are variable, but in typical realizations, the 'acute' accent involves astep up in pitch and the 'circumflex' a step down in pitch (ToporiW: 1984). In non wordfinal position, these patterns are distributed over a stressed syllable and the followingunstressed syllable, so that the acute has a low-pitched stressed syllable and a high peakon the unstressed, whereas the circumflex has a high peak on the stressed syllablefollowed by a lower unstressed syllable. Hence the acute is also referred to as the low orrising toneme and the circumflex as the hiSh or faJling toneme. The accents are illustratedgraphically on the words for 'mother' and "grandfather' in the following sentences:

Acute:

Circumflex:

maMa je prisla. 'Mother has come'.

Dedek je prisel. 'Grandpa has come'.

The accents are seldom lexically distinctiye, although there are examples such as ['ki:la]kila 'hernia' (acute) and ['kl:/;:).] kila 'kilo' (circumflex). Perhaps because of this, many

words vary in which accent they are given, and some speakers, although they speakStandard Slovene, do not make any distinction between the two accents. In thetranscription below, the different accents have not been marked, although both vowel

length and stress are indicated despite the considerable redundancy involved. Anintonation group boundary is marked by [I] and a pause by [Ill.

Transcription of recorded passage'se:wrni 'ue:tdr in 'so:ntse sta se pr£'pi:rala I ka'te:ri :)d 'nji:ju je m:)tJ'neiJi Ib je 'mi:m:) pri'Jeo p:)'po:tnik I za'ui:t u b:pdU 'pla:JtJ. II d:)g:)u:)'ri:la sta seI da bo ud'jau za m:)tJ'neiJega 'ti:sti I ki mu bo 'pd:ruemu u'spe:b I da bop:)'po:tnik 'sle:bu suoi 'pla:JtJ. II 'se:wrni 'ue:tdr j£ za'pi:xau z u's:): m:)'tJjo: IItoda 'bo:l b je 'pi:xau I 'bo:l td'sn:o je p:)'po:tnik I :)'ui:jau sooi 'pla:JtJ ::lkoli'se:be. II 'b:ntJn:) I je 'se:wrni 'oe:tdr p:)'pu:stiu. II na'to: je 'so:ntse t:)'plo:p::lsi'ja:b I in p:)'po:tnik je ta'koi 'sle:kdu suoi 'pla:JtJ. II in ta'ko: je 'se:odrni'oe:tdr 'mo:rau pri'zna:ti I da je 'so:ntse m::l'tJneiJe :)d 'nje:ga. II

Orthographic versionSeverni veter in sonce sta se prepirala, kateri od njiju je motnej~i, ko je mimo pri~el

popotnik, zavit v topel pla~t. Dogovorila sta se, da bo veljal za motnej~ega tisti, ki mu boprvemu uspelo, da bo popotnik slekel svoj pla~t. Severni veter je zapihal z vso motjo,toda boJj ko je pihal, bolj tesno je popotnik ovijal svoj pla~t okoli sebe. Kontno je Severni


veter popustil. Nato je Sonce toplo posijalo in popotnik je takoj slekel svoj plas~. In takoje Severni veter moral priznati, daje Sonce mo~nejSe od njega.

AcknowledgementsThe authors would like to thank Mr. Janez Oblak for technical assistance, and Dr. Beverley Collins (Leiden

University, Holland) for his valuable advice and suggestions.

ReferencesSREBOT-REJEC, T. (1988). Word Accent and Vowel Duration in Standard Slovene: An

Acoustic and Linguistic Investigation (Slavistische Beitrage, 226). Munich: Otto

Sagner.TOPORISIC, J. (1984). Slovenska slovnica. Maribor: Zalotba Obzorja.

Swedish

OLLE ENGSTRAND

Ins tilLIte of Linguistics. University of Stockholm, S-106 91 Stockholm, Sweden

The style of speech illustrated is that o~ many educated speakers of Central StandardSwedish as spoken in the Stockholm area. The actual speech on the accompanyingrecording is that of a male speaker in his forties whose speech is typical of that variety ofSwedish.

Consonants

Bilabial Labiodental Dental Alveolar Palatal Velar Glottal

Plosive P b t d k 9

Nasal m n 1)

Fricative f V S J h

Approximant 1

Lateral1Approximant

fj (Voiceless dorso-palatal/velar fricative) C (Voiceless alveolo-palatal fricative)

P pol 'pole' t tok 'fool' k kon 'cone'b bok 'book' d dop 'christening' 9 god 'good'm mod 'courage' n nod 'node' I] lang 'long'f fot 'foot' S sot 'soot' C kjol 'skirt'v vat 'wet' fj sjok 'chunk' h hot 'threat'

rov 'prey' I lov 'tack' j jord 'soil'

Vowels

-~. u:ue

e 0:

e::>

a:a

l/lustrations of the IPA 141

i: sil 'strainer' (II: nOt 'nut'sill 'herring' re nOli 'worn'

y: syl 'awl' Q: mat 'food'y syll 'sleeper' a matt 'feeble'

l;l jul 'ugly' 0: mal 'goal'

e full 'full' :) moll 'minor' (music)e: hel 'whole' u: bot 'penance'

e: hal 'heel' u bolt 'lived' (perf)

e hall 'flat rock'

Stress and accent, (primary stress), ,(secondary stress), and' (word accent) as in ['nu:ctan ,vrnd;}n]nordanvinden 'the north wind'.

ConventionsIp, t, k/ are aspirated in stressed position when not preceded by lsi within the samemorpheme. It, d, n, s, II are dental. The retroflex sounds [l, ct, 11., ~, Ucan be consideredphonetic forms of Irt, rd, rn, rs, rll; this is also reflected in the orthography (e.g.nordanvinden). They also arise at word and morpheme boundaries, e.g.jor+soket. hi canbe approximant (e.g. starkast), voiced fricative (e.g. the first r in vandrare), or trilled; thetrilled variant is restricted to emphatic stress in many speakers and does not appear in therecording. Open varieties of Ie:, e, (II:, rei are used before hi and the retroflex allophones(e.g.forst). Long vowels are often diphthongized or fricativized, particularly the high ones:[ij, Yl/, u~, l;l~]. Lip rounding differs between Iy, (III ('outrounded') and lu, l;ll ('inrounded').IQ:I is slightly rounded. Consonants are long after short vowels in stressed position, andshort elsewhere. Voiced consonants are frequently devoiced in voiceless context, e.g.[tylstad;}], [jest go]. Compound words have the so-called grave tonal word accent (accent2) with falling pitch on the primary stressed syllable and a strong secondary stress. Mostbisyllabic and polysyllabic stems also have the grave accent but with a weaker secondarystress.

Transcription of recorded passage'nu:qan,vmd~m ::> 'su:l~m 'ty'rstada a1) 'g::>1) ::>m VEm av 'd::>m s::>m va 'stAlkast.'jest 'go: bm an 'vandJala 'vE:gan 'fJam, ~n,sve:pt i an 'valm 'kapa. d::>mk::>m do: ~Va'JEnS ::>m, at dEn s::>m 'fq:~t kenda fa vandJal~m at to 'a:v SEj'kapan, han skelg 'an,se:s vaJa 'staJkaJg En dEn 'andJa. do: 'blo:stg'nu:qan,vmdgn so 'hot han 'n:')n,sm 'kanda, mEn jl;! 'ho:qalg han 'blo:sta, dEstu'te:talg 'sye:ptg 'Vandlalgn 'kapan '::>m SEj, ::> tl 'SISt gay 'nu:qan,vmdgn 'epfq:'~~:kgt. do: l£:t 'su:lan sina 'stJo:laJ 'fji:na 'he:lt 'vaJmt, ::> 'jenast tug'vandJalgn 'a:v sEj 'kapan, ::> so va 'nu:qan,vmdgn 'tverpn at 'ea,CEna, at'su:lgn va dEn 'st3.Jkasta av d::>m 'tva:.


Orthographic versionNordanvinden oeh solen tvistade en gang om vern av dom som val' starkast. Just da kom envandrare vagen fram, insvept i en varm kll-ppa. Oom kom da bverens om, att den som fbrstkunde fa vandraren aU ta av sig kappan, han skulle anses vara starkare an den andra. Oablaste nordanvinden sa hart han nansin kunde, men ju hilrdare han blaste desto tatare sveptevandraren kappan om sig, oeh till sis! gay nordanvinden upp fbrsoket. Oa lat solen sinastralar skina helt varmt oeh genast tog vandraren av sig kappan, oeh sa var nordanvindentvungen aU erkanna aU solen var den starkaste av dom tva.

AcknowledgementAppreciation is extended to Ian Maddieson for the preparation of the vowel diagram.

Taba

JOHN BOWDEN AND JOHN HAJEK

Department of Linguistics aJUi Applied Linguistics.

Universiry of Melbourne, Parkville, VIC 3052, Australia

Taba (also known as 'East Makian' or 'Makian Dalam') is an Austronesian languagespoken in northern Maluku province, Indonesia. It is spoken on Makian, Moti and Kayoaislands and also in a few villages on other nearby islands. The speech recorded here isfrom the village of Ngofakiaha on Makian island. Slight dialectal variations are to be

found in most villages where the language is spoken.

Consonants


Plosive p b t d k 9Nasal m n 1]

Fricative f S h

Affricate tf d3

Trill r

Approximant J W

Lateral1Approximant

The consonant table shows phonemic contrasts. Although ItJ I and Id31 are relativelyinfrequent and originally borrowed, both can now be considered fully native.

p pait 'moon' t top 'sugar cane' k kam 'I see'b bait 'wrestle' d d::l 'couscous' 9 gah 'theft'm mai 'then' n n::l 'there' l) l)an 'sun'f fati 'to cover' S::l 'to ascend' h han 'to go'

tJ tJ::lat 'firewood bundle' j jan 'fish'd3 d3::lU 'good' w wah 'island'

T::lrE: 'stone'bat 'to slice'


GeminatesAll of the Taba consonants except /j, w"r, d3/ can occur with distinctive length (i.e. as'geminates'). Interestingly, although geminates are found word-medially, they are mostcommon in word-initial position. Geminates are never found in word-final position. Theymay occur monomorphemically, but they' also occur across morpheme boundaries whereinitial agreement markers are prefixed t~ verbs. Phrase-initial geminates are realized incareful speech with a greater degree of tension and more articulatory force; in more casualvarieties of speech they can be realized with the same reduced tension and articulatoryforce as singletons. Some illustrative examples of geminate contrasts are given below.

t tala 'to meet' I b 'inside'tt ttala 'we (inclusive) meet' 11 Ib 'blood'

9 9:)w:) 'place' k kut 'louse'

99 99:)w:) 'neck' kk kku 'tail'm mul 'to return' h han 'to go'mm mmul 'you (sg) return' hh hhan 'you (pI) go'I) I)an 'sun'

1)1) I)l)e 'canarium nut'

VowelsTaba has a simple five vowel system, as shown on the chart. The phone [e] is relativelyrare. Notwithstanding the effects of predictable contextual conditioning (see below),vowels are normally realized as relatively short. Very rare examples of long vowels can befound, e.g. [ki:] 'vagina', but only one marginal example of a length contrast has beenuncovered: [tel 'no(t)' vs. [te:] 'if not'.

All possible sequences of unalike vowels occur in the corpus except for lea] and [e:)],but none are listed here because they are analysed as vowel clusters and not diphthongs.

im 'fishing reel'

e hen 'turtle'

a am 'to see':) :)m 'uncle'

u urn 'house'

lllustrations of the IPA 145

StressStress is indicated by a greater degree of force in the production of a syllable, a noticeablyhigher pitch than in unstressed syllables, and some degree of lengthening. In general,primary stress falls on the only syllable of monosyllabic words or on the penultimatesyllable of polysyllabic words. Examples of antepenultimate stress are extremely rare, e.g.['uJi;,tit] 'we defecate'. Final stress in polysyllables is more frequent, but is generallyrestricted to a few loan words, e.g. [a'had] 'Sunday, week' « Arabic), and somepolymorphemic verb forms with monosyllabic roots, e.g. [na'pe] 'he does' from ['pel 'todo',

ConventionsStops are not released in final position, and [t] is usually slightly palatalized before [iJ.The nasals [m] and [n] tend to assimilate towards a following labiodental [f] while [n] alsotends to assimilate towards following bilabials and velars. [r] is always strongly trilled.

In addition to its use of initial geminates, Taba is also characterized by the appearanceof a large set of unusual clusters, e.g. ['nmu] 'muddy water', ['mhonas] 'sick', ['nhik]'bat', ['khan] 'I go', ['mt;,] 'eye'. More common initially than medially, they are often theresult of prefixing simple consonants to verb roots in order to mark person and number. Innasal consonant + [h] clusters, e.g. ['mhonasJ, the nasal starts off as fully voiced, with thelatter portion devoiced without any concomitant increase in nasal airflow. The [h) offset isclear. Heterorganic clusters with initial [h) are also found, e.g. ['hkutan] 'you ask' vs.['kutan] 'to ask'. The articulation of [h) in onset is also clear.

Initial geminate contrasts are generally stable, although there is a tendency for somespeakers to reduce initial [hh] to [h], e.g. ['hhan] - ['han] 'you (pi) go'. Conversely, somenon-geminate fricative clusters may be subject to optional metathesis or even completeassimilation, e.g. expected ['hs;,paJ)] alongside ['sh;,paI)] and ['ss;,paI)] 'you (pi) descend'.

There is not much allophonic variation in the vowels, apart from the kinds ofvariations seen in a great many languages: vowels tend to be nasalized before nasalconsonants, and they tend to be lengthened slightly before voiced consonants and instressed positions. Some dialects of Taba replace many instances of Ngofakiaha hI withla I, e.g. ['w;,g] vs. ['wag] 'canoe'.


A riddle about being sick

'banda 'ni 'We 'mh:mas I 'nim 'We na'lusa 'mh:mas II 'nim pap'pub 'm£

na'lusa 'mh:mas II 'bir]:) nam:J'lam II u'bn 'nmau 'nhan II 'P:Jt:J 'P:JP£ 'nmau

'nhan 'ntJiwi II 'sum:J na'lusa 'khan II 'mt:J 'nujak I 'P:Jj:J 'mh:Jnas I bai'bij:J

nma'br):J II u'bn 'nmau 'nhan I 'mtum:J £ 'b £ II pap'pub 'mai na'lusa

'mh:Jnas II 'bir):J na'lusa nam:J'lam II u'bn 'jas£ 'nmau mpa'r£nta 'nhan II


'P::Jt::J 'P::JP£ na'lusa ta'siaki II 'mpili £ 'b 'Ii £ II 'p::Jb 'banda ta'siak I 'nhan

'tuli na'hat£s II 'mau 'nhan 'tuli s£'dal)kan 'banda na'p£ ta'siaki II 'malE 'si;,I

'malai 'han 'tuli II 'si::J 'mul 'malai :han 'tuli a'han II 'malE 'ttfi;,tit 'hu 'malaiI

'than II 'ttfi::Jtit ';,kik 'malai 'than 'tuli I 'than 'tr;,nda II

Orthographic versionBanda ni we mhonas, nim we nalusa 'mhonas'. Nim pappuko me nalusa 'mhonas'. Bingonamolam. Ulon nmau nhan. Poto pope nmau nhan ncioi. Sumo nalusa 'khan'. Mto nuyak,poyo mhonas, baibiyo nmalongo. Ulon nmau nhan, mtumo e 10 e? Pappuko mai nalusamhonas. Bingo nalusa namolam. Ulon yase nmau nparenta nhan. Poto pope nalusa tasiaki.Mpili e 10 Ii e? Polo Banda tasiak, nhan tuli nahates. Mau nhan tuli sedangkan Banda napetasiaki. Male sio malai han tuli. Sio mul malai han tuli ahan. Male tciotit hu malai than.Tciotit okik malai than tuli, than tronda.

Thai

M. R. KALAYA TINGSABADH

Department of Linguistics, Faculty of Arts, Chulalongkorn University, Bangkok 10330, ThailanJi

AND

ARTHUR S. ABRAMSON

Haskins Laboratories, New Haven, CT, and Department of Linguistics, The University of Connecticut,

Storrs, CT 06269-1145, USA

Standard Thai is spoken by educated speakers in every part of Thailand, used in newsbroadcasts on radio and television, taught in school, and described in grammar books anddictionaries. It has developed through the standardization of a socially prestigious variety ofCentral Thai, the regional dialect of Bangkok and the surrounding provinces.

The transcription of 'The North Wind and the Sun' is based on recordings made bythree cultivated speakers of the language, who were asked to read the passage in a relaxedway. In fact, we find them all to have used a fairly formal colloquial style, apparentlyequivalent to Eugenie J. A. Henderson's 'combinative style' (Henderson 1949). In a moredeliberate reading of the text, many words in the passage would be transcribed differently.The main features subject to such stylistic variation are vowel quantity, tone, and glottalstop. Thus, for example, Ite:1 'but' is likely under weak stress to be Itel with a short vowel;the modal auxiliary 1tJ;,a.?I'about to' becomes Itr;aJ, with change of tone from low to mid andloss of final glottal stop. The prosodic and syntactic factors that seem to be at work hereremain to be thoroughly explored.

Consonants

Bilabial Labio- Alveolar Post- Palatal Velar Glottaldenial alveolar

Plosive p p" b t til d k k" ?Nasal m n I)

Fricative f S hAffricate tc tc"

Trill r

Approximant J W

LateralApproximant 1

148 Handbook of the EPA

p pa:n 'birthmark' t tii.:n 'sugar palm' k ka:n 'act'ph pha:n 'belligerent' th tha:n 'alms' kh kha:n 'shaft'b ba:n 'to bloom' d da:n 'calloused' I) I)a:n 'work'

m ma:n 'demon' n na:n 'long time' w wa:n 'to sow'

f fa:n 'to slice' s sa:n 'court' j ja:n 'sagging'ra:n 'shop' ? ?a:n 'saddle'

I la:n 'million' h ha:n 'to divide'

te tea:n 'dish'

teh teha:n 'trash'

VowelsThere are nine vowels. Length is distinctive for all the vowels. (In some phonologicaltreatments, N:I is analyzed as IVV/.) Although small spectral differences between short andlong counterparts are psychoacoustically detectable and have some effect on vowelidentification (Abramson and Ren 1990), we find the differences too subtle to place withconfidence in the vowel quadrilateral. The vowel lal in unstressed position, including theendings of the diphthongs /ia, wa, ua/, is likely to be somewhat raised in quality. The finalsegments of the other two sets of phone\ic diphthongs: (I) [iu, eu, e:u, e:u, au, a:u, iau]and (2) [ai, a:i, oi, o:i, ui, Y:i, uai, wail ai;e analyzed as Iwl and Ijl respectively .

• 1 u

~, \ wa[ /l ua(

0

J

-----1a

krit 'dagger' i: kri:t 'to cut' ia nan 'to study'

e ?en 'ligament' e: ?e:n 'to recline' wa niian 'house'

e phe? 'goat' e: phe: 'to be defeated' ua roan 'to be

a fan 'to dream' a: fa:n 'to slice' provocative'

0 kbI) 'box' 0: kI5:I) 'drum'

0 khan 'thick (soup)' 0: kha:n 'to fell (a tree)'

u sut 'last, rearmost' u: su:t 'to inhale'

Y I)in 'silver' Y: di:n 'to walk'

w khwn 'to go up' w: khlw:n lwave'


TonesThere are five tones in Standard Thai: high 1'1, mid r I, low J'I, rising r I, and falling rl.

kha: 'to get stuck'kha: 'galangal'kha: 'I'

kha: 'to engage in trade'kh;!i: 'leg'

StressPrimary stress falls on the final syllable of a word. The last primary stress before the end ofa major prosodic group commonly takes extra stress.

ConventionsThe feature of aspiration is manifested in the expected fashion for the simple prevocalic oralstops Ip", t", kh/. The fairly long lag between the release of the stop and the onset of voicingis filled with turbulence, i.e. noise-excitation of the relatively unimpeded supraglottal vocaltract. In the special case of the 'aspirated' affricate Itc,"/, however, the noise during thevoicing lag excites a narrow postalveolar constriction, thus giving rise to local turbulence. Itis necessarily the case, then, that the constriction of the aspirated affricate lasts longer thanthat of the unaspirated one (Abramson 1989). Not surprisingly, it follows from theseconsiderations that the aspiration of initial stops as the first element in clusters occurs duringthe articulation of the second element, which must be a member of the set /I, r, wi.

Only Ip, t, k, 7, m, n, I), w, jl occur in syllable-final position. Final Ip, t, k, 71 have noaudible release. The final oral plosives are said to be accompanied by simultaneous glottalclosure (Henderson 1964; Harris 1992). Final I?I is omitted in unstressed positions. InitialIt, thl are often denti-alveolar. Initial It I and IfI are velarized before close front vowels.

The consonant Irl is realized most frequently as [r] but also as [r]. Perceptualexperiments (Abramson 1962: 6-9) have shown that the distinction between Irl and IV is notvery robust; nevertheless, the normative attitude among speakers of Standard Thai is thatthey are separate phonemes, as given in Thai script. This distinction is rather wellmaintained by some cultivated speakers, especially in formal speech; however, many showmuch vacillation, with a tendency to favour the lateral phone [1] in the position of a singleinitial consonant. As the second element of initial consonant clusters, both III and /rl tend tobe deleted altogether.

In plurisyllabic words, the low tone and the high tone on syllables containing the shortvowel lal followed by the glottal stop in deliberate speech, become the mid tone whenunstressed, with loss of the glottal stop.


kha'na? thi ,lom'mua Ie ,phni?a'thlt I kam'Ial) 'thial) kan 'wa: I 'khraj tca 'mT:pha'lal) 'ma:k kwa 'kan I k5 'mT: 'mik,dY:n'tha:l) 'phU: 'null) 'dY:n 'pha :n 'rna: I'saj 'swa,kan'na:w II ,lom'nwa Ie ,phra?a'thlt 'tCWl) ,tok'lOl) kan 'wa: I 'khraj thi,sa:'ma:t 'tham haj 'mlk,dY:n'tha:l) 'phU: 'nl: I 'th3:t 'swa,bin'na:w ?3k 'da:j


,sam'ret 'kb:n I tca 'thw: 'wa: I pen 'phU: thl 'ml: phei'la!) 'ma:k 'kwa: II 'Ie? 'le:w I,lom'nwa k3 kra'phw : 'phat 'ja:!) 'sut! 're:!) II te ll!) 'phat 're:!) 'ma:k 'khwn 'phI a!)

'daj I 'nak,dy:n'tha:1) k3 'jl I) 'dwl) 's~a,kan'na:w 'haj kra'tchap kap 'tua 'ma:k

'khwn 'phIal) 'nan II 'Ie? 'naj thl 'sut I ,lom'nwa k3 'ly.:k 'lorn 'khwa:m

phaja'ja:m II 'tca:k 'nan I ,phra?a'thit tCWl) 'sa:t 's£:1) ?an 'r6:n 're:1) ?bk 'rna: II'nak,dy:n'tha:1) k3 'thb:t 'swa,kan'na:w '?b:k 'than 'thl: II 'naj thl 'sut I ,lom'nwa

tCWI) 'tcam 't31) 'j5:m 'rap 'wa: I ,phra?a't11lt ml: pha'lal) 'ma:k 'kwa: 'ton II


'lIru:;~fUJ l'VIU tlll~:;'V'I 'S:;miilllU(\1 ~\ll~ tJ\I fl'W11'If'1'Fil:;iJ'V'I ~\l3Jl fl fl11 fl'W f1 iJ11'fll~'W'Yl1\I ~"

'VI~\I L~'W ~1'W3Jl 'l~l~tl fl'W'VI'W11 ~3J l'VIU mlft:;'V'Id:;miiIllU~\llIlflft\l fl'W11 'If'1d~ftl3Jld(1

r\1'1~11'm~'W'YI1\1~d C1tlll\l~tlfl'W'VI'W11flflfl1;~ll1'ilnfl'W'ild1tl11l ,j'W~~iJ'V'I~\l3Jlflfl11 LLft:;" "

Ll;l~ 3J l'VIU tl i1m:;~m~'lI\fl~l\1~lI\lLd\l LL~~\I-Wll\lL'S\l3Jl fl~'W loWm'lll\ 11'fl L~'W'Yl1\If1~\I~\I L~fl

fl'W'VI'W11'1~m:;i1J fl1J ~13Jl fl~'W loWtJ\I,T'W Ll~:;'1'W~ftll\fl3Jl 'VIU El nl~ fl;3Jf'1113J'V'IUl Ul3J 'ill fl,,T'W'V'Id:;mii lI\U~\lftll1\ Llft\ltl'W;fl'W IId\lflflfl~l 11'm~'W'l'11\In (1flll\l~flfl'W'VI'W11flElflrl'WYi '1'W~

.or d 0 (V Q ~cj Q.I

~lI\~3Jl'VI'W fl'il\l'illl1lfl\l tJ El3J'S1Jl1'V'1'S:;m'YlllltJ3J'V'I~\l3J1 flf1Jll1l'W

AcknowledgementsWe thank Dr. Chalida Rojanawalhanavulhi, Dr. Kingkarn Thepkanjana, and Miss Surangkana Kaewnamdee,whose readings of the passage underlie our transcription. Dr. Theraphan Luangthongkum and Dr. SudapornLuksaneeyanawin read the manuscript and made helpful comments. Part of the work of the second authorwas supported by Grant HDOl994 from the U.S. National Institutes of Health to Haskins Laboratories.

ReferencesABRAMSON, A. S. (1962). The Vowels and Tones ofStandard Thai: Acoustical

Measurements and Experiments. Bloomington: Indiana University Research Center in

Anthropology, Folklore, and Linguistics, Publication 20.ABRAMSON, A. S. (1989). Laryngeal control in the plosives of Standard Thai. Pasaa 19,

85-93.ABRAMSON, A. S. AND REN, N. (1990). Distinctive vowel length: Duration vs. spectrum

in Thai. Journal of Phonetics 18, 79-92.HARRIS, J. G. (1992). The consonant sounds of 17th century Siamese. Mon-Khmer

Studies 21,1-17.HENDERSON, E. J. A. (1949). Prosodies in Siamese: A study in synthesis. Asia Major

New Series 1, 189-215.HENDERSON, E. J. A. (1964). Marginalia to Siamese phonetic studies. In Abercrombie,

D., Fry, D. B., MacCarthy, P. A. D., Scott, N. C. and Trim, J. L. M. (editors), /n

Honour of Daniel Jones: Papers Contributed on the Occasion ofhis Eightieth Birthday

/2 September 196/,415-24. London: Longmans.

Tukang Besi

MARK DONOHUE

Departmelll of Linguistics, University ofSydney, NSW 2006, Australia

Tukang Besi is an archipelago in south-east Sulawesi, Indonesia, where an Austronesianlanguage is spoken, The speech represented here is that of the northern coast of the islandof Wanci. With mainly slight lexical variation it represents the speech of the twonorthernmost islands in the archipelago, Wanci and Kaledupa.

Consonants

Bilabial Labiodental Dental Alveolar Velar Glottal

Plosive p b t (d) k 9 ?

Implosive 6 cfNasal m n IJFricative 13 S (z) hPrenasalized

mb nt nd IJkPlosive mp IJgPrenasalizedFricative ns

Trill rLateral

IApproximant

The consonant table represents phonemic contrasts, with the exception of [d] and [z]which are allophones of the loan phoneme Id3/; only borrowed words (or suspected loans)show this phoneme. Ibl is also a loan phoneme, present only in a few recently borrowedwords, but contrastive with 151, as seen in [bale] 'turn around (walking only)' « Malaybatik) vS. [5ale] 'frond of coconut tree'.

p apa 'up to' t titi 'breast' k kai 'hook'5 a5a 'previous' cf picfi 'rubbish' 9 gai 'pull out'm ama 'father' n ana 'child' lJ alJa 'gills'13 ?aJ3a 'obtain' s asa 'one'

mp kompa 'eel' nt tinti 'run'

IlJk 5alJka 'ship'

mb komba 'moon' nd pindi 'faeces' lJg ilJgaJ3i 'yesterday'ns pinse 'squeeze'

ara 'if'

I? ?acfa 'send'

ala 'fetch' h hacfa 'imminent'

152 Handbook of/he {PA

Vowels

5ali 'turn around'e 5ale 'frond of a young coconut

tree'a 5ala (classifier for soap and

small objects)0 5alo 'answer'w 5alw 'buy'

o

a

Stress/ accentStress, realized as a generally higher a~d level pitch on the whole syllable, is almostalways on the penultimate syllable of the word. The definition of the word is sometimesextended to include the following (inherently stressless) absolutive article Ite/. Thus thesentence nomanga Ie bae 'he ate the rice' can be pronounced either as [no'mal)ate'5ae] oras [,noma'1)ate'5ae]. A secondary stres~, consisting of a greater amplitude without thehigher pitch that is the characteristic of t~e primary stress, is also found on every secondsyllable away from the primary stressed s~llable.

Contrastive accent patterns have arisen from the introduction of loan words with finalnasals; the final nasal is borrowed as a sbparate syllabic nucleus, with predictable effectson the penultimate accent, and then later ~ropped. This has produced at least one case of alexical difference based solely on accent: ['kenta] 'fish' (archaic) vs. [ke'nta] 'potato'(loan, < Malay [kenta1)] ).

ConventionsFree variation is observed between W], [<I>] and [u], with [(1] being the most commonphone except amongst children, who pr~ter [<1>]; the implosives and /gl are preglottalizedfollowing a stressed syllable ([75], [70] ~nd [7g]). /hi varies with [<1>] before a high backvowel. The lateral III is sometimes realized as a retroflex lateral flap [1] after back vowels.The phoneme Irl varies between [r], [r], i[l] and [u; thus, after a back vowel, realizationsof Irl and III overlap. The loan phonefe Id31 varies idiolectally between [d3] (mostcommon), [d] (contrasting with [efJ), and; [z]. Thus: Lkata'da:] - [,kata'd3a:] karajaa 'work(for a wage)', ['dari] - ['zari] - ['d3ari]ja~i 'then, thus'.

Voiceless or sonorant consonants ar4 often geminated in the stressed or post-stressedsyllable, in a restrictive set of vocalic environments that depends on the consonant. Thus:['aka] - ['ak:a] 'root', [,motu'twrw] - [,rpotw't:uJrw ] 'sleepy'. The voiced prenasalizedstops sometimes reduce to a geminate n~sal in casual speech. Thus [5a'mbai] - [5a'm:ai]'comb', ['nda1)a] - ['n:a1)a] 'jackfruit', [i'1)ga(1i] - [i'1):ja(1i] 'yesterday'. A glottal stopdeletes or is replaced by [k] in a syllab~e adjacent to another glottal stop: thus 17wl '2ndperson singular possessive' and 11)0701 'inose', but [1)o'70w] 'your nose' « 11)070 + 7w/),I?el '3rd person object' and Itw70/ 'chop down', but [tw'7oke] 'chop it down' « 1t1ll70 +

I7e/).


The high front vowel shows little variation in quality, nor does the low back vowel.The high back unrounded vowel Iml can round to [uJ in a syllable immediately followingthe vowel 101 or following Iw/. In these and other environments it shows occasionalvariation with [u]. The front mid vowel leI tends to raise to [eJ or even [IJ when notfollowed by a consonant (i.e. finaJly or before a syllable without an onset), with thenotable exception of the core article te, which is invariably [tel. The mid back vowel ishigher than the mid front vowel, again tending to be higher when followed by a vowel,especially when the following syllable contains 101 or Im/. Following a bilabial consonantthe height of the vowel is also more pronounced, and the closer the syllable to the end ofthe word, the higher the vowel. All the vowels tend to de-syllabify if they occur before thestressed syllable adjacent to another vowel, or in a stressless word; usually the highest ofthe vowels desylJabifies. When two vowels come together across a word boundary, thefirst of these often deletes. See the text for examples of these phenomena.

Transcription of recorded passagesa'pajra sa'pajra 'ana, 'ane ke la 'Bela j3a'nde,:!sa'I)ia 'kene la 'Bela 70'10:. tea'm:ai 'ana ,n;)po,Busw'BusU{ 'ako te e'mai na me'5ukw. mbe'aka mo'leI)o'7umpa ,n;)ma'imo na 'mia <pu'mila pU'make ?e 'Bad3w mo'k;)75a.,sa7i,ta7a'bno, ,n;)Se'twd3w kU{a te ,5a:'Ba:no pU,maJw'lw7e oa 'mia j3u'mila'iso, te 'ia na,hopo'taJw i pO,ta.{w'?ano 'isiS aj. 'maka te Ja '5ela j3a'nde,:!sa'I)ia'ana o;)bp:a, O;)'t;)p:a, o;)bp:a, bk:a ,same'5ukw n;)bp:a ,same'5ukm te 'ia'iso o;)'k;)ota te '5adw ,mob'Baoo. ka'm5ea ,moni'nino. ma,kala'7amo,mbea'kamo oJ'sJ75a na la '5ela j3a,ode,:!sa't]ia 'ana. po'?oli 'iso, la 'Bela70'10: oO,pasJ'soa te 'mia me'Bad3lU iso 'kene ,mbea'kamo J'ha7cfa pa'ke teBad3lUoo. 'd3ar ,aj3a'oa:tw nJ,hJpo'taJw na la '5ela b'lo:, 'kene ,oJmo'talwoa la 'Bela j3a,ode,:!sa'I)ia i ,pota'rw7 lU ka'j3asa 'iso aj. ,sapa'ira ,sapa'ira,taI)ka'nomo.

Orthographic versionSapaira sapaira ana, ane ke La Bela Wandensangia kene La Bela '0100. Te arnai ananopobusubusu ako te emai na mebuku. Mbeaka molengo 'umpa nomaimo na mia wumilapumake te baju rnokoba. Sa'itaakono, nosetuju kua te baabaano pumalulu'e na mia wumilaiso, te ia nahopotalu i potaru'ano iso ai. Maka te La Bela Wandensangia ana notopa,notopa, notopa, toka samebuku notopa samebuku te ia iso nokonta te baju mokobano.Kambea moninino. Maka la'arno rnbeakamo nosoba na La Bela Wandensangia ana. Po'oliiso, La Bela '0100 nopasosoa te mia mebaju iso kene rnbeakamo ohada pake te bajuno.Jari awana atu nohopotalu na La Bela '0100, kene nomotalu na La Bela Wandensangia ipotaru'a u kawasa iso ai. Sapaira sapaira tangkanomo.

rurkish

KARL ZIMMER AND ORHAN ORGUN

Department of Linguistics, univer~ity of California, Berkeley, CA 94720, USA

The speech is that of a 25-year-old nat ve of Istanbul. This variety of the language isconsidered standard; most educated speak~rs of Turkish are familiar with this pronunciationas well as a regional variety.

Consonants


Plosive b t d tI d3 kand Affricate P C j 9

Nasal m Q

Fricative f V S z I 3 Y h

Tap I fI

Approximant jLateral i 1Approximant

p put 'stamp' t tel iwire' c car 'profit'

b but 'find' d del I'pierce' j jem 'bit (for a horse)'k kar 'snow'g gam 'grief

m mat 'property' n nat 'horseshoe'

rej I'vote'f far 'headlight' s sar 'wrap'v var 'exists' z zar 'membrane' y day fda:] 'mountain'

h her 'every'ItJ tJam 'pine'

d3 d3am I'glass'I jer 'place'

ta'ta 'servant'la:'le 'tulip'


Vowels

kil 'clay' •y kyl 'ashes' ue kel 'bald'ee geel 'lake' •a kat 'stay' 0

w kwt 'hair'u !cut 'slave'0 kat 'ann'

Long vowels are [i:), [e:), [u:) and [a:]. Diphthongs can be treated as sequences of voweland Ij/.

StressWord stress tends to be on the last syllable of the word. However, there are someunstressable suffixes which cause the main word stress to fall on the syllable precedingsuch a suffix, e.g. [jap-ttur-ma'lw] 's/he must have (it) done' vs. [jap-'ttur-ma-malw] 's/hemust refrain from having (it) done', where the negative suffix [-rna] is an unstressablesuffix. There are also some lexical exceptions to final stress, e.g. ['masa] 'table'.

ConventionsThe voiceless stops are usually aspirated in syllable-initial position and are always releasedin codas unless followed by a homorganic consonant. Ic,}1 do not contrast with /k, gl inthe native vocabulary, where [c) and [j] appear only in syllables with front vowels, while[k] and [g] appear only in syllables with back vowels. There are, however, some loanwordsin which there are unpredictable occurrences of [c] and [j) with back vowels, e.g. Icarl kar'profit' (cf. /karl kar 'snow'). 11/ is a palatalized postalveolar lateral, ItI a velarized dentallateral; ItI does not occur after front vowels. /hI in final position may be realized as avoiceless velar fricative. Irl is most commonly a single tap. Ir, t, II are frequently devoicedin final position or when a voiceless consonant follows. [v] is frequently pronounced as abilabial fricative or approximant when preceded by a vowel. Iyl corresponds to the 'soft g'(g) in Turkish orthography; its use finds its main justification in accounting formorphological alternations. Iyl between front vowels is pronounced as a weak front-velar orpalatal approximant. When the Iyl is word-final or followed by a consonant it is realizedphonetically as a lengthening of the preceding vowel; elsewhere when intervocalic, it isphonetically zero. All vowels except la, 01 have a lower variant in the final open syllable ofa phrase, e.g. [kel) 'bald' but [ka'i£] 'castle'.


Transcription of recorded passage I

poj'razla jy'nef birbirlerin'den da'ta kuvvet'li olduklarw'nw ile'ri sy'rereciddia:la'Jwjodardw. 'dercen ka'tw 'bir 'patto jij'miJ 'bir jOB'd3u Jcerdy'ler.'bu jold3u'ja 'pattosu'nu tJw,kart wrabile'nin da'ha kuvvet'li otduyu'nuka'bu:l etmi'je ka'rar verdi'ler. poj'raz 'var jy'd3yle esmi'je bafta'dw.'and3ak jot'd3U 'pattosu,na 'jitjide da'ha sw'kw sarw'nwjordu. 'sonundapoj'raz uyrafmak'tan 'vazjetJti. 'bu se'fer jy'neJ atJ'tw orta'twkwsw'nwnd3a jot'd3U 'pattosunu he'men tJwkar'dw. 'bcej1ed3e poj'raznne'fin ,kendisin'den da'ha kuvv'let'li olduyu'nu ka'bu:l etmi'je med3'burkat'dw.

Orthographic versionPoyrazla gtine~, birbirlerinden daha kuvvetli olduklartOl ileri stirerek iddiala~lyorlard/.

Derken, kahn bir palto giymi~ bir yolcu gordtiler. Bu yolcuya paltosunu ylkarttlfabilenindaha kuvvetli oldugunu kabul etmeye karar verdiJer. Poyraz, var gtictiyJe esmeye ba~ladl.

Ancak, yolcu paltosuna gitgide daha S~kl sartnlyordu. Sonunda poyraz ugra~maktan

vazgeyti. Bu sefer gtine~ aytl; ortah k IsiOl nca yolcu paltosunu hemen ylkardl. Boylece

poyraz, gtine~in kendisinden daha kuvvetlt oldugunu kabul etmeye mecbur kaldl.

PART 3

Appendices

Appendix 1

The Principles of the International Phonetic Association

From its earliest days (see appendix 4) the Association has tried to make explicit theprinciples which guide its work. The statement of these principles has been amended andupdated from time to time; the current formulation (below) was approved at the 1989Convention of the Association.1 The International Phonetic Association has a standard alphabet which is usuallyreferred to by the initials IPA, or, in a number of non-English-speaking countries, API. Itis designed primarily to meet practical linguistic needs, such as putting on record thephonetic or phonological structure of languages, providing learners of foreign languageswith phonetic transcriptions to assist them in acquiring the pronunciation, and working outroman orthographies for languages written in other systems or for languages previouslyunwritten. A large number of symbols and diacritics is also provided for representing finedistinctions of sound quality, making the IPA well suited for use in all disciplines in whichthe representation of speech sounds is required.2 The IPA is intended to be a set of symbols for representing all the possible sounds ofthe world's languages. The representation of these sounds uses a set of phonetic categorieswhich describe how each sound is made. These categories define a number of naturalclasses of sounds that operate in phonological rules and historical sound changes. Thesymbols of the IPA are shorthand ways of indicating certain intersections of thesecategories. Thus [p] is a shorthand way of designating the intersection of the categoriesvoiceless, bilabial, and plosive; [m] is the intersection of the categories voiced, bilabial,and nasal; and so on. The sounds that are represented by the symbols are primarily thosethat serve to distinguish one word from another in a language.3 In the construction of the IPA attention has been paid not only to the appropriatenessof each symbol from a phonetic point of view, but also to the suitability of symbols fromthe typographical point of view. The non-roman symbols of the IPA have, as far aspossible, been made to harmonize with the roman letters. For instance, the Greek lettersincluded in the IPA are roman adaptations; as the ordinary shape of the Greek letter f3 doesnot harmonize with roman type, in the IPA it has been given the form~. The Associationdoes not favour the use of italic forms of symbols as models for the design of newsymbols.

4 The construction and use of the IPA are guided by the following principles:(a) When two sounds occurring in a given language are employed for distinguishing oneword from another, they should wherever possible be represented by two distinct symbolswithout diacritics. Ordinary roman letters should be used as far as is practicable, butrecourse must be had to other symbols when the roman alphabet is inadequate.(b) When two sounds are very similar and not known to be employed in any language fordistinguishing meanings of utterances, they should, as a rule, be represented by the same


I,

II

symbol. Separate symbols or diacritics m y, however, be used to distinguish such soundswhen necessary.(c) It is not possible to dispense entirely with diacritics. The International PhoneticAssociation recommends that their use be limited as far as possible to the following cases:(i) For denoting length, stress and pitc~.

(ii) For representing minute shades of sounds.(iii) When the introduction of a single Idiacritic obviates the necessity for designing anumber of new symbols (as, for instance, ip the representation of nasalized vowels).5 The use of symbols in representing the sounds of a particular language is usuallyguided by the principles of phonological 9ontrast. All languages use a limited number ofvowels and consonants that are able to distinguish word meanings: the contrast betweenEnglish m and n is used to distinguish the ,vords met and net, and these two sounds shouldtherefore be represented by different symbols. The three k-sounds of the English wordskeep, cart, cool can be heard and felt to be different, but from the linguistic orphonological point of view the differences are not distinctive and all may be representedby the same [k] symbol. The same applieslto the French k-sounds in qui, cas, cou, thoughthese differ phonetically from the correspohding English ones.6 The Association recommends that phonetic transcription should be enclosed insquare brackets [ ]. A transcription that no es only phonological contrasts may be enclosedin slanted Jines or slashes t t.7 A transcription always consists of a set of symbols and a set of conventions for theirinterpretation. Furthermore, the IPA con~ists of symbols and diacritics whose meaningcannot be learned entirely from written descriptions of the phonetic categories involved.The Association strongly recommends th t anyone intending to use the symbols shouldreceive training in order to learn how to produce and recognize the corresponding sounds

with a reasonable degree of accuracy. I

II

Appendix 2

Computer coding of IPA symbols

This section is a revision of an article which first appeared in the Journal of theInternational Phonetic Association 23, 83-97 (1993), entitled 'Computer codes forphonetic symbols', by John H. Esling and Harry Gaylord. The chart of IPA Numbers andthe coding tables in this section have been updated to 1998.

The process of assigning computer codes to phonetic symbols began when the 1989Kiel Convention of the International Phonetic Association was called to revise theAssociation's alphabet. The Workgroup on Computer Coding formed at that time had thetask of determining how to represent the IPA alphabet numerically, and of developing aset of numbers referring to IPA symbols unambiguously. This involved assemblingphoneticians who work with computer representations of phonetic symbols, andcommunicating with specialists in computer coding to gauge the fit between thephonetician's perspective on symbol usage and the non-phonetician's understanding ofhow to identify and use phonetic symbols. Prior to the Kiel meeting, a collection ofpractical approaches to coded representations was outlined in JIPA (Esling 1988), whichdealt mainly with keyboard assignments of characters.

At Kiel, after reviewing several submissions on current practice, the Workgroupconcluded that each symbol used by the IPA should be assigned a unique, three-digitnumber known as its IPA Number. IPA Numbers were assigned in linear order followingthe new IPA Chart which resulted from the deliberations at Kiel (IPA 1989a). IPANumbers were not only created for approved consonant, vowel, diacritic andsuprasegmental symbols, but also for symbols often referred to in IPA deliberations orimplied by IPA convention but which do not appear explicitly on the IPA Chart.

The conclusions and recommendations reached by the Workgroup at Kiel appeared inJIPA (IPA 1989b), followed by a supplementary report with an initial listing of IPANumbers by symbol and by symbol name (Esling 1990). The comprehensivedocumentation of phonetic symbol usage and categorization in the Phonetic Symbol Guideby Pullum and Ladusaw (1986, [996) assisted in this process. Several recommendationswere made at Kiel, including the use of at least two levels of transcription when enteringcoded phonetic values to accompany data, but what is most important to point out aboutthe process of associating each possible character with a discrete numerical entity is itscomprehensive nature. No symbols could be ignored, and the application of diacritics hadto be made explicit. The result is that the list of IPA Numbers includes more charactersthan are specified on the IPA Chart alone. Thus. in addition to the IPA Chart, an IPANumber Chart was also established to show the equivalent Number for each symbol; and acomprehensive list was drawn up that includes symbols cited by the IPA since 1949, aswell as some non-IPA symbols, cross-referenced to their equivalent Numbers.


The reason for the comprehensive inclusion of all symbols is to anticIpate thepossibility that some symbols may bf withdrawn while other symbols may bereintroduced into current usage; and a numerical listing of character shapes and types mustbe comprehensive enough to support slig~t revisions in symbol specification or diacriticplacement as well as to be available to al wide spectrum of phonetic users of computersystems. The set of IPA Numbers also allows for the addition of new symbols, within the100 series for consonants, the 200 series for extra symbols and cross-reference to otherphonetic sets, the 300 series for vowels, thf 400 series for diacritics, and the 500 series forsuprasegmental symbols. Ligatures for affricates, for example, are included in the 200series as formerly recognized IPA symbols although they do not occupy a specific

location on the IPA Chart. IAs the 1989 IPA Chart was subjectyd to review, several modifications emerged,

resulting in the publication of the 1993 IPA Chart (IPA 1993) which was updated in 1996.No new Numbers were required to specify these symbols, even though a few symbolswere revived and reinstated. Although the order of IPA Numbers is different from the1989 order, because the original order w~ itself arbitrary, the changes on the 1993 Chartdo not impinge on the IPA Number scheme as an effective numerical interface. In somesmall measure, the Numbers that accomp+y the symbols help to document the history oftheir development. The IPA Number Chrrt corresponding to the updated IPA Chart isshown as table I.

The set of IPA symbols and their numbers were used to draw up an entity set withinSGML by the Text Encoding Initiative (TfI). The name of each entity is formed by 'IPA'preceding the number, e.g. IPA304 is the tEI entity name of lower-case A. These symbolscan be processed as IPA symbols and represented on paper and screen with the

Iappropriate local font by modifying the :entity replacement text. The advantage of theSGML entity set is that it is independent Of the character set being used.

At the same time that this work was tieing done, two organizations were drawing upcomputer character sets for all world languages: the Unicode Consortium and ISO, theInternational Standards Organization. Thehe projects were linked so that the code for eachcharacter is the same in Unicode and the ISO Universal Character Set (UCS), also knownas ISO 10646. The IPA symbols were su~mitted for inclusion in these character sets and

I

with a few exceptions have been incorporated. Character set 10646 was approved by theISO and published on I May 1993. The f~lI set, which comprises over 40,000 characters,undergoes periodic revision. A recent u~date can be found in The Unicode Standard,Version 2.0 (Unicode Consortium 1996). I

Tables 3, 4,5 and 6 below contain thelVCS Code position for each phonetic character.UnicodelUCS is a 16-bit character set, and is represented therefore as a four-characterhexadecimal number. There are also some non-IPA phonetic symbols in UnicodelUCSwhich had been submitted by other groups than the IPA. A few IPA symbols, e.g. theChao tone characters, have not yet been i~cluded in the UnicodelUCS character set. If thesymbol is not included, this is indicated by ,------, in the table. A few symbols have twopossible encodings, e.g. IPA598 is enc0ged as 0316 if placed under the symbol and as

THE INTERNATIONAL PHONETIC ALPHABET (revised to 1993, updated 1996)

CONSONANTS (PULMONIC) NUMBER CHART

Biloblal Labiodental Dental IAlvcolnr jPost:tlveOlar Retroncx Palo,"1 Vclnr Uvular Pharyngeal Glonal

Ploslve 101 102 103 104 105 106 107 108 109 110 111 112 I' ': l13!f ~

Nasal 114 115 116 117 118 119 120 ".~

Trill 121 122 123 "",:~~Tap or Flap l24 125

...0'

Fnco(lve 126 127 128 129 130 1311132 133 1134 135 136 137 138 139 140 141 142 143 144 145 146 147

Latc,"1a a 148 149 11:"1fi.1';:- .~afncatlve

Approximant 150 151 152 153 154

Lateral 155 156 157 158 r~;tpproxtmnnt

Where symbols ..ppear In pairs, (he one 10 (he nght represents a vOiced consonant Shaded areas denote arllculnllons Judged Impossible

DIACRITICS Diacritics may be placed above a symbol with a descender, e.g. 119 + 402B

402A Voiceless l} d 405 Brealhy voiced b a 408 DenIal t g403 Voiced § 1 406 Creaky ....oiced g g 409 Apical ! 9404 Aspirnled th dh 407 LinguolnlJial t Q 410 Laminal 1cj4/1 More rounded y 420 l...Jbializcd tW dW

424 Nasalized e412 Less rounded J 421 Pal:unlizcd tj dJ 425 Nasal release dn

413 Ad....nnccd \l 422 Velnriz.cd tY dY 426 l....:lleml rc:lcnsc d'414 RelrnCled e 423 Pharyngealized t~ d)' 427 No audibll: release d'415 Ccmmliu:d e 428 Velarized or pharynge.,liud 209

X416 Mid-centralized e 429 Raised t( <t = voiced ah'eolnr fricalive)

431 Syllabic 1). 430 Lowered ~ (~ = voiced bilnbinl npproximant)

432 Non-syllabic ~ 417 Advanced Tongue Rool ~419 RholicilY 327 Cl'- 418 RelraCled Tongue Rool y

Clicks 1 VOiCed implosi \les EJcclives

176 Bllabml

I160 BilabIal 401 Example.."

162JOJ

O.labHII177 Denial DcnlaValvcolar + 401

164103

Dcnl.,l/iI]vco)ar178 (Posl)alveolar POllala] + 401

Palal0alveolar I 166109

Velar179 Vel:u +401

Alveolar lateral I 168132

Alvcol:u- fncallve180 U~'Ular +401

182 183 /\Iveolo-pnlntnl fncnllVes

501 Pnmary slress

502 Secondary siress

,fOUng'tlf;m

503 Long e:504 Half·long e'505 Extra·shon e507 Minor (foot) group

508 Major (intonalion) group

506 Syllable breok li.rekt

Where symbol'!; nppcar In patrs, The one10 Ihe nghl represenls 9. rounded vowel

TONES AND WORD ACCENTSLEVEL CONTOUR

509 Linking (absence of a break)

SUPRASEGMENTALS

512 519 EXlrn 524 529 Risinghigh

513 520 High 525 530 F<lll.ing

514 521 Mid 526 531Highrising

515 522 Low 527 532Lowrising

516 523EXlffi

528 533Rising-

low falling

517 Downslep 510 Global rise

518 Upslep 511 Global fall

VOWELS

Front CentrRI Dad:.

Close 301 309

m~.--316 308

319 320 321

Close-mId 302 310 -- 397 323 315 307

322

Open-rrud 303 311 - 326\395- 314 306

325 324

Open 304 312~ 305 313

433 (509)

Alveolar l:lter:l1 flap

Slmullaoeou~ Jnod X

Affncates find double antCUlaltOnscan be rcprc.~enlcd by two "ymbols

JOined by a lie bnr If necessary

175

181

169 VOiCelesS lablilJ·velar fncauve

170 VOIced [abla]·\lc]cu approxlmanl

171 VOiced labml-pahual apprOXlmLlnl

172 VOICelesS epiglottal fncalL ....e

174 VOIced epigiollfli fncall ....e

173 Epigiollal ploslvc

CONSONANTS (NON·PULMONlC)

OTHER SYMBOLS

Table 1 The IPA Number Chart

164 HandbookofthelPA

02CE if it follows the symbol. When this 9haracter set is in wide use, it will be the nonnalway to encode IPA symbols. I

In cooperation with ISO, the Associ~tion for Font Information Interchange (AFII)maintains a registry of glyphs. The IPA SY~bolS have been registered with AFII, and theirregistered glyph numbers are indicated tth a hexadecimal number in the AFII Codecolumn of tables 3, 4, 5 and 6. The ,AFII glyph registry may be used for fontstandardization in the future.

A TEl writing system declaration (wsg) has been drawn up for the IPA symbols. Thisdocument gives information about the sympol and its IPA function, as well as its encodingin the accompanying SGML document afd in UnicodelUCS and in AFII. The writingsystem declaration can be read as a text d9cument or processed by machines in an SMGL

process. ITable 2 illustrates that the 26 roman characters within the IPA symbol set have

retained their original or 'ASCII' numbers as their UCS codes. They belong to the 00 or'base' table of Unicode/UCS. More specialized phonetic symbols have been assigned tosubsequent tables.

Table 2 IDCS codes retain original ASCII coding,for roman characters in the IPA set.

IPA DCS AFIISymbol Symbol Name Phonetic Description / Status Number Code Code

a Lower-case A Open front unrounded vowel 304 0061 E25Bb Lower-case B Voiced bilabia~ plosive 102 0062 E2A3c Lower-case C Voiceless palat~1 plosive 107 0063 E2D9d Lower-case D Voiced dental or alveolar 104 0064 E2BI

plosivee Lower-case E Close-mid front unrounded 302 0065 E256

vowelf Lower-case F Voiceless labio1dental fricative 128 0066 E2AC

g Looptail G Voiced velar plosive 210 0067 E2E3I

Equivalent to 110

h Lower-case H Voiceless glott~l fricative 146 0068 E2EE

Lower-case I Close front unrounded vowel 301 0069 E251

J Lower-case J Voiced palatal :approximant 153 006A E2DB

k Lower-case K Voiceless vela~ plosive 109 006B E2DE

1 Lower-case L Voiced dental or alveolar lateral 155 006C E2BD

approximant

m Lower-case M Voiced bilabial nasal 114 006D E2AI

n Lower-case N Voiced dental or alveolar nasal 116 006E E2AF

Computer coding of [PA symbols 165

0 Lower-case 0 Close-mid back rounded vowel 307 006F E269

P Lower-case P Voiceless bilabial plosive 101 0070 E2A2q Lower-case Q Voiceless uvular plosive III 0071 E2E6r Lower-case R Voiced dental or alveolar trill 122 0072 E2COs Lower-case S Voiceless alveolar fricative 132 0073 E2B6t Lower-case T Voiceless dental or alveolar 103 0074 E2BO

plosiveU Lower-case U Close back rounded vowel 308 0075 E265v Lower-case V Voiced labiodental fricative 129 0076 E2ADw Lower-case W Voiced labial-velar approximant 170 0077 E2A8

X Lower-case X Voiceless velar fricative 140 0078 E2EO

Y Lower-case Y Close front rounded vowel 309 0079 E252

z Lower-case Z Voiced alveolar fricative 133 007A E2B7

The Kiel Convention Workgroup on Pathological Speech and Voice Quality hasdeveloped specialized symbols and diacritics with IPA Numbers from 600 to 699. Theyare described in Duckworth, Allen, Hardcastle and Ball (1990), and listed together withtheir Number assignments in Ball (1991). They are not included in the present tablesbecause they were not considered among the original submissions made to ISO foruniversal coding. At the moment, therefore, there are no UCS codes or AFII codes thatcan be easily associated with the 'Extended IPA' characters for disordered speech andvoice quality of the 600 series.

Table 3 lists all phonetic consonant and vowel symbols that have been given a code inthe universal coded character set, cross-referencing symbol shape, symbol name,articulatory description, IPA Number, UCS code, and AFII code. The symbols arearranged in pseudo-alphabetical order. Table 4 lists all phonetic diacritic andsuprasegmental symbols that have been given a code in the universal coded character set,arranged in the order of the original 1993 IPA Chart. Symbols in the declaration which arenot IPA symbols and those which are no longer in IPA usage are specified. with anindication of the date when they were removed from IPA usage or superseded by othersymbols. Supplementary tables 5 and 6 offer a cross-referenced listing of all phoneticconsonant and vowel, diacritic and suprasegmental symbols in numerical order by IPANumber. An earlier version of these tables appeared in the Handbook of Standards andResources for Spoken Language Systems (Gibbon, Moore and Winski 1997). Thepublication of these lists of coding assignments should not be construed as anendorsement by the IPA of every character in the list, but as a convenient reference to thelocation of any potential phonetic character in the coding tables as currently constituted.

166 Handbook ofthe IPA

Table 3

IPA ,ymbo'" Pbon.'" ,"mooanU'owo! I-YmbO"""(in pseudo-alphabetical order by symbol shape)

IPA DCS AFIISymbol Symbol Name Phonetic DerriPtion / Status Number Code Code

a Lower-case A Open front urtrounded vowel 304 006[ E25BB Turned A Near-open cejltral vowel 324 0250 E263

Script AI

305 0251 E26Ca Open back unrounded vowel0 Turned script A Open back rofnded vowel 313 0252 E26Dre Ash; Lower-case Near-open frrt unrounded vowel 325 00E6 E25A

A-E ligature

b Lower-case B Voiced bilabial pJosive 102 0062 E2A35 Hooktop B Voiced bilabial implosive 160 0253 E2A9B Small capital B Voiced bilabial trill [21 0299 E2FO

J3 Beta Voiced bilabial fricative 127 03B2 E2A5II

c Lower-case C Voiceless palatal plosive [07 0063 E2D8C' Hooktop C Voiceless paital implosive 163 0188 2376

Withdrawn ( 993)

C Cwedge Voiceless postalveolar affricate 299 O[OD F[AE

C cedillaNot IPA usagf

138 00E7 E2DAY Voiceless palatal fricative

C Curly-tail C Voiceless alveolo-palatal fricative 182 0255 E2CE

C Stretched C Postalveolar click 202 0297 E2C4Superseded by 178 (1989),

dI

Lower-case D Voiced dental or alveolar plosive 104 0064 E2BIcf HooktopD Voiced dental or alveolar implosive 162 0257 E2C2

clI

Right-tail D Voiced retroflex plosive 106 0256 E2C8c( Hooktop right- Voiced retroflex implosive 219 E219

tail D Not exPlicitlyjJPA approveddz D-Z ligature Voiced alveo ar affricate 212 02A3 E2F9

Superseded b~ 104+ 133

~ D-Ezh ligature Voiced postayeolar affricate 2[4 02A4 E2FA

Superseded by 104+135

<t D-Curly-tail-Z Voiced alveojo-palatal affricate 216 02A5 E2FB

ligature Superseded by 104+183(') Eth Voiced dental fricative 13[ OOFO E2B3

e Lower-case E Close-mid fr0nt unrounded vowel 302 0065 E256

Computer coding of IPA symbols 167

d Schwa Mid central vowel 322 0259 E25F~ Superscript Mid central vowel release 218 E21A

schwaa- Right-hook R-coloured mid central vowel 327 025A E260

schwa Equivalent to 322+419:;) Reversed E Close-mid central unrounded vowel 397 0258 E26Ee Epsilon Open-mid front unrounded vowel 303 025B E258E) Closed epsilon Superseded by 395 (1996) 396 029A E2733 Reversed epsilon Open-mid central unrounded vowel 326 025C E262a Closed reversed Open-mid central rounded vowel 395 025E E270

epsilon

f Lower-case F Voiceless labiodental fricati ve 128 0066 E2AC

9 Opentail G Voiced velar plosive 110 0261 E2DFEquivalent to 210

9 Hooktop G Voiced velar implosive 166 0260 E27Eg LooptailG Voiced velar plosive 210 0067 E2E3

Equivalent to 110

G Small capital G Voiced uvular plosive 112 0262 E2E7

G Hooktop small Voiced uvular implosive 168 029B E2FIcapital G

y Gamma Voiced velar fricative 141 0263 E2EIy Superscript Velarized 422 02EO E28B

gammay Ram's horns Close-mid back unrounded vowel 315 0264 E268

h Lower-case H Voiceless glottal fricative 146 0068 E2EEh Superscript H Aspirated 404 02BO D565

h Barred H Voiceless pharyngeal fricative 144 0127 E2EBfi Hooktop H Voiced glottal fricative 147 0266 E2EFfj Hooktop heng Simultaneous voiceless 175 0267 E2D6

postalveolar and velar fricative

4 Turned H Voiced labial-palatal approximant 171 0265 E2A6H Small capital H Voiceless epiglottal fricative 172 029C E2F2

i Lower-case I Close front unrounded vowel 301 0069 E2511 Undotted I Not 1PA usage 394 0131 00F5i Barred I Close central unrounded vowel 317 0268 E25Dt Iota Near-close near-front unrounded 399 0269 E253

vowel Superseded by 319 (1989)

168 Handbook of the IPAI

Small capital I Near-close n~ar-front unrounded 319 026A E254vowel

I

J Lower-case J Voiced pala(al approximant 153 006A E2DBI

Superscript J Palatalized ! 421 02B2 D567

j Curly-tail J Voiced palaJal fricative 139 029D E2F3

J J wedge Voiced postalveolar affricate 298 OIFO E290Not [PA usake

j Barred dotless J Voiced pala6al plosive 108 025F E2D9

J Hooktop barred Voiced palaJal implosive 164 0284 E27Cdotless J

k Lower-case K Voiceless ve.lar plosive 109 006B E2DEK Hooktop K Voiceless velar implosive 165 0199 2363

Withdrawn (V993))( Turned K Withdrawn 1979) 291 029E E2F4

Lower-case L Voiced dent lor alveolar lateral 155 006C E2BDI

approximantSuperscript L Lateral rele~e 426 02EI FDA3

t L with tilde Velarized voiced dental or alveolar 209 026B E27Dlateral approximant

i Belted L Voiceless dental or alveolar lateral 148 026C E2BBfricative

1 Right-tail L Voiced retroflex lateral 156 026D E2CC

Small capital Lapproximan~

158 029F E2F5L Voiced velar lateral approximant

t L-Ezh ligature Voiced dent~1 or alveolar lateral 149 026E E2BCfricative

A Lambda Voiceless dental or alveolar lateral 295 03BB 266Efricative Not [PA usage

Ji. Barred lambda Voiceless dental or alveolar lateral 294 019B FD7Baffricate Not [PA usage

m Lower-case M Voiced bilabial nasal 114 006D E2AI

llJ Left-tail M (at Voiced labiodental nasal 115 0271 E2AB

right)I

lli Turned M Close back unrounded vowel 316 026F E264

Computer coding ofIPA symbols 169

Ul Turned M, right Voiced velar approximant 154 0270 E2E2leg

n Lower-case N Voiced dental or alveolar nasal 116 006E E2AFSuperscript N Nasal release 425 207F FDA8

11 N, right leg Syllabic nasal Withdrawn (1976) 293 019E E2E5

J1 Left-tail N (at Voiced palatal nasal 118 0272 E2D7left)

I) Eng Voiced velar nasal 119 014B E2DD

11 Right-tail N Voiced retroflex nasal 117 0273 E2C6N Small capital N Voiced uvular nasal 120 0274 E2E4

0 Lower-case 0 Close-mid back rounded vowel 307 006F E2690 Bull's eye Bilabial click 176 0298 E2AAe Barred 0 Close-mid central rounded vowel 323 0275 E261

f/J Slashed 0 Close-mid front rounded vowel 310 00F8 E257ce Lower-case O-E Open-mid front rounded vowel 311 0153 E259

ligatureCE Small capital O-E Open front rounded vowel 312 0276 E25C

ligature:) Open 0 Open-mid back rounded vowel 306 0254 E26BG> Closed omega Near-close near-back rounded 398 0277 E266


P Lower-case P Voiceless bilabial plosive !OI 0070 E2A2

P Hooktop P Voiceless bilabial implosive 159 0lA5 2378Withdrawn (1993)

<I> Phi Voiceless bilabial fricative 126 0278 E2M

q Lower-case Q Voiceless uvular plosive III 0071 E2E6

cf Hooktop Q Voiceless uvular implosive 167 02AO E2F6Withdrawn (1993)

r Lower-case R Voiced dental or alveolar trill 122 0072 E2COf Fish-hook R Voiced dental or alveolar tap 124 027E E2CIr Long-leg R Withdrawn (1989) 206 027C E2BEJ Turned long-leg Voiced alveolar lateral flap 181 027A E2BF

R

r Right-tail R Voiced retroflex flap 125 027D E2CD1 Turned R Voiced dental or alveolar 151 0279 E2BA

approximant


.:r. Turned R, right Vo;,'" ""or." ,ppw,;m," 152 027B E2CBtail

R Small capital R Voiced uvul r trill 123 0280 E2EAff Inverted small Voiced uvular fricative 143 0281 E2E9

capital RI,

S Lower-case S Voiceless alveolar fricative 132 0073 E2B6

Superscript S Withdrawn (1989) 207 02E2 FDA7

S S wedge Voiceless postalveolar fricative 297 0161 FIDCNot [PA usa}e

~ Right-tail S (at Voiceless retroflex fricative 136 0282 E2C9left) I

,J Esh Voiceless postalveolar fricative 134 0283 E2DO

£ Curly-tail esh Withdrawn (1989) 204 0286 E2D2I

t Lower-case T Voiceless dttal or alveolar plosive 103 0074 E2BOf Hooktop T Voiceless de, tal or alveolar 161 OIAD 2379

implosive Withdrawn (1993)Left-hook T Palatalized v.oiceless dental or 208 01AB E2AO

t Right-tail Talveolar plos~ve Withdrawn (1989)

Voiceless re~roflex plosive lOS 0288 E2C7ts T-S ligature Voiceless d~tal or alveolar 211 02A6 E2FC

affricate Su erseded by 103+ 132

If T-Esh ligature Voiceless po~talveolar affricate 213 02A7 E2FDSuperseded by 103+134

te T-Curly-tail-C Voiceless alveolo-palatal affricate 215 02A8 E2FEligature Superseded ~y 103+182Turned T Dental click 201 0287 E2C3

Superseded Qy 177 (1989)

e Theta Voiceless de tal fricative 130 03B8 E2B2

8 Superscript theta Voiceless dental fricative release 217 E21B

u Lower-case U Close back rounded vowel 308 0075 E265I

l:l Barred U Close central rounded vowel 318 0289 E25E

u Upsilon Near-close near-back rounded 321 028A E267vowel !

v Lower-case V Voiced labi~dental fricative 129 0076 E2AD

u Cursive V Voiced labio.dental approximant 150 028B E2AE

A Turned V Open-mid back unrounded vowel 314 028C E26A


w Lower-case W Voiced labial-velar approximant 170 0077 E2A8w Superscript W Labialized 420 02B7 D56E

M Turned W Voiceless labial-velar fricative 169 028D E2A7

X Lower-case X Voiceless velar fricative 140 0078 E2EOSuperscript X Voiceless velar fricative release 292 02E3 D56F

X Chi Voiceless uvular fricative 142 03C7 E2E8

Y Lower-case Y Close front rounded vowel 309 0079 E252!. Turned Y Voiced palatal lateral approximant 157 028E E2DCy Small capital Y Near-close near-front rounded 320 028F E255

vowel

z Lower-case Z Voiced alveolar fricative 133 007A E2B7

Z Z wedge Voiced postalveolar fricative 296 017E FIFONot IPA usage

Z Curly-tail Z Voiced alveolo-palatal fricative 183 0291 E2CF

~ Right-tail Z Voiced retroflex fricative 137 0290 E2CA

3 Ezh; Tailed Z Voiced postalveolar fricative 135 0292 E2DI

~ Curly-tail ezh Withdrawn (1989) 205 0293 E2D32 Barred two Withdrawn (1976) 290 OIBB E2B5

? Glottal stop Glottal plosive 113 0294 E2ED~ Barred glottal Epiglottal plosive 173 02AI E2F7

stop.'> Inverted glottal Alveolar lateral click 203 0296 E2C5

stop Superseded by 180 (1989))' Reversed glottal Voiced pharyngeal fricative or 145 0295 E2EC

stop approximantSuperscript Pharyngealized 423 02E4 E28Creversed glottalstop

~ Barred reversed Voiced epiglottal fricative or 174 02A2 E2F8glottal stop approximant

I Pipe Dental click 177 OICO 23A6

:f Double-barred Palatoalveolar click 179 0lC2 23A4pipe

II Double pipe Alveolar lateral click 180 OICI 23A7! Exclamation (Post)alveolar click 178 0lC3 23A5

point


T~~4 JPhonetic diacritic and suprasegmental ymbol codes(in 1993 IPA Chart order) :

IPA DCSPhon,ti, l P,.oom,n" AFIISymbol Symbol Name Descriptio Status Number Code Code

Apostrophe Ejective p' k' ts' If' 401 02BC E249Under-ring Voiceless I} 9 402A 0325 E229

Over-ring Voiceless i) 9 402B 030A OOCASubscript wedge Voiced § ~ 403 032C E22A

h Superscript H Aspirated ph th 404 02BO 0565Subscript right More rounded y y 411 0339 E23Chalf-ringSubscript left Less rounded ? 412 031C E232half-ringSubscript plus Advanced l} 413 031F E233Under-bar Retracted ~ 414 0320 E234

Umlaut Centralized e 415 0308 E221x

Over-cross Mid-centralized e 416 0330 2311

Syllabicity mark Syllabic f t:l ,1 431 0329 E22ESubscript arch Non-syllabic ~ 432 032F 23FA

Right hook Rh .. I e- 419 020E E28AotlClty 3'- 0'- a-Subscript umlaut Breathy voi~ed 9 ? 405 0324 E22BSubscript tilde Creaky voicfd 9 g 406 0330 2308Subscript seagull Linguolabia) t 9 407 033C 22E8

W Superscript W Labialized tW dW 420 02B7 056E

Superscript J Palatalized ti di 421 02B2 0567Y Superscript Velarized tY dY 422 02EO E28B

gammaSuperscript Pharyngeali~ed t' d' 423 02E4 E28C

reversed glottalstopSubscript bridge Dental I} 9 408 032A E22C

Inverted subscript Apical !J. 9 409 033A 23FDbridgeSubscript square Laminal IJ Q 410 033B 23FE

Superscript tilde Nasalized e- o 424 0303 E222


0 Superscript N Nasal release dO 425 207F FDA8Superscript L Lateral release dl 426 02EI FDA3

Corner No audible, d' 427 031A 23F9P

releaseSuperimposed Velarized or A EI 428 0334 E226tilde pharyngeal izedRaising sign Raised <r ! (~ Dol) 429 O31D/ E22F

0204Lowering sign Lowered y 13T (eT DT) 430 031E/ E231

02D5Advancing sign Advanced tongue 1} 417 0318 23F7

rootRetracting sign Retracted tongue y 418 0319 23F8

rootTop tie bar Affricate or fP gb ts d3 433 0361 E225

doublearticulation

Subscript right Rhoticity e;. I{ <\. ~ 489 0322 E228hook Superseded by

419 (1989)Open comer Release/ burst Not IPA usage 490 E218Comma Pause (comma) Not IPA usage 491 002C 002C

Reversed Weak aspiration Withdrawn 492 02BB 00A9apostrophe (1979)Over-dot Palatalization! Withdrawn 493 0307 E224

centralization (1979)Minus sign Retracted variety Use 414 or 494 02D7 E239

(backed) 418 (1989)+ Plus sign Advanced variety Use 413 or 495 02D6 E238

(fronted) 417 (1989)y Superscript Y High-front Not IPA usage 496 02B8 D570

rounding!palatalized

Under-dot Closer variety/ Use 429 497 0323 E230fricative (1989)

Subscript left Palatalized Superseded by 498 0321 E227hook 421 (1989)Subscript W Labialized Superseded by 499 032B E22D

420 (1989)

174 Handbook of Ihe IPA

Suprasegmentals

Vertical stroke P' I ,found'tIfdn 501 02C8 E23Enmary str~ss

(Superior)Vertical stroke Secondary ~tress ,found'trfdn 502 02CC E23F(Inferior)Length mark Long e: 503 02DO E23AHalf-length mark Half-long e' 504 02DI E23BBreve Extra-short e 505 0306 E223Period Syllable break li.rekt 506 002E 002E

I Vertical line Minor (foot) 507 007C 007C(thick) group

"Double vertical Major 508 2016 2142line (thick) (intonation)

groupBottom tie bar Linking (ab~ence ,fa:l~d'weI 509 203F 230F

of a break) !

Tones and word accents

Double acute Extra high level Q 512 030B OOCDaccent (over)

Acute accent High level 6 513 0301 00C2(over)

Macron Mid level 6 514 0304 00C5Grave accent Low level a 515 0300 OOCI(over)

Double grave Extra low level a 516 030F 23E2accent (over)

1 Extra-high tone Extra high I~vel mal 519 02E5 E28Dletter

1 High tone letter High level ma1 520 02E6 E28E, Mid tone letter Mid level rna, 521 02E7 E28F-1 Low tone letter Low level ma-1 522 02E8 E29FJ Extra-low tone Extra low level maJ 523 02E9 E29E

JetterWedge; hacek Rising contqur a 524 030C OOCF

Circumflex Falling contpur a 525 0302 00C3J

Macron plus High rising a 526 E296

acute accent contour~

527 E297Grave accent plus Low rising amacron contour


~

Grave plus acute Rising-falling a 528 E298plus grave accent contour

A Rising tone letter Rising contour maA 529 E299\J Falling tone letter Falling contour ma\J 530 E29A1 High-rising tone High rising ma1 531 E29B

letter contourA Low-rising tone Low rising maA 532 E29C

letter contour-I Rising-falling Rising-falling ma-J 533 E29D

tone letter contourJ. Down arrow Downstep 517 2193 EEAFt Up arrow Upstep 518 2191 EEAD

/' Upward diagonal Global rise 510 2197 EF3E

arrow\, Downward Global fall 511 2198 EF3D

diagonal arrowSuperscript arch Long falling Not IPA usage 595 0311 23F2

tone/advanced! palatal

Wedge; hacek Falling-rising Usage re- 596 02C7 E247tone defined (1989)

See 524Circumflex Rising-falling Usage re- 597 02C6 E246

tone defined (1989)See 525

Subscript grave Low falling tone Superseded 598 0316/ E245accent (1989) 02CESubscript acute Low rising tone Superseded 599 0317/ E243accent (1989) 02CF

Transcription delimitation charactersIPA DeS AFII

Symbol Symbol Name Phonetic Description / Status Number Code Code

Left square Begin phonetic transcription 901 005B 005BbracketRight square End phonetic transcription 902 005D 005Dbracket

/ Slash Begin/end phonemic transcription 903 002F 002F( Left parenthesis Indistinguishable utterance (begin) 906 0028 0028) Right parenthesis Indistinguishable utterance (end) 907 0029 0029


« Left double Sound obscured (begin) 908 0028+ 2127parenthesis 0028

)) Right double Sound obsc red (end) 909 0029+ 2128parenthesis

Begin prosJdic notation0029

Left brace 910 007B 007BRight brace End prosod}c notation 911 007D 007D

Table 5IPA symbols: Phonetic consonant/vowel symbol codes(in numerical order by IPA Number) I

IPA UCS AFIISymbol Symbol Name Phonetic Description / Status Number Code Code

p Lower-case P Voiceless bilabial plosive 101 0070 E2A2b Lower-case B Voiced bila~ial plosive 102 0062 E2A3t Lower-case T Voiceless dental or alveolar plosive 103 0074 E2BOd Lower-case D Voiced dent~l or alveolar plosive 104 0064 E2BI

t Right-tail T Voiceless re roflex plosive 105 0288 E2C7q Right-tail D Voiced retro,flex plosive 106 0256 E2C8c Lower-case C Voiceless pa,Jatal plosive 107 0063 E2D8

j Barred dotless J Voiced palatal plosive 108 02SF E2D9k Lower-case K Voiceless velar plosive 109 006B E2DE

9 Opentail G Voiced velar plosive 110 0261 E2DFEquivalent to 2/0

q Lower-case Q Voiceless uvular plosive III 0071 E2E6G Small capital G Voiced uvular plosive 112 0262 E2E7( Glottal stop Glottal plosirve 113 0294 E2EDm Lower-case M Voiced bilatJial nasal 114 006D E2Al

rI) Left-tail M (at Voiced labiodental nasal 115 0271 E2ABright)

n Lower-case N Voiced dental or alveolar nasal 116 006E E2AFI

11. Right-tail N Voiced retroflex nasa] 117 0273 E2C6

J1 Left-tail N (at Voiced palatal nasal ] 18 0272 E2D7

left)Voiced velat nasalI) Eng 119 014B E2DD

N Small capital N Voiced uvular nasal 120 0274 E2E4

B Small capital B Voiced bila~ial trill 121 0299 E2FOI

r Lower-case R Voiced dentel or alveolar trill 122 0072 E2CO

R Small capital R Voiced uvular trill 123 0280 E2EA

f Fish-hook R Voiced dental or alveolar tap 124 027E E2CI

( Right-tail R Voiced retr1f1ex flap 125 027D E2CD


<P Phi Voiceless bilabial fricative 126 0278 E2A4

~ Beta Voiced bilabial fricative 127 03B2 E2A5f Lower-case F Voiceless labiodental fricative 128 0066 E2ACv Lower-case V Voiced labiodental fricative 129 0076 E2ADe Theta Voiceless dental fricative 130 03B8 E2B2(5 Eth Voiced dental fricative 131 OOFO E2B3

s Lower-case S Voiceless alveolar fricative 132 0073 E2B6z Lower-case Z Voiced alveolar fricative 133 007A E2B7

J Esh Voiceless postalveo)ar fricative 134 0283 E2DO

3 Ezh; Tailed Z Voiced postalveolar fricative 135 0292 E2DI

~ Right-tail S (at Voiceless retroflex fricative 136 0282 E2C9

left)

~ Right-tail Z Voiced retroflex fricative 137 0290 E2CA

Y C cedilla Voiceless palatal fricative 138 00E7 E2DA

J Curly-tail J Voiced palatal fricative 139 029D E2F3

x Lower-case X Voiceless velar fricative 140 0078 E2EO

Y Gamma Voiced velar fricative 141 0263 E2EI

X Chi Voiceless uvular fricative 142 03C7 E2E8

IS' Inverted small Voiced uvular fricative 143 0281 E2E9capital R

h Barred H Voiceless pharyngeal fricative 144 0127 E2EB<;' Reversed glottal Voiced pharyngeal fricative or 145 0295 E2EC

stop approximanth Lower-case H Voiceless glottal fricative 146 0068 E2EEfi Hooktop H Voiced glottal fricative 147 0266 E2EFi Belted L Voiceless dental or alveolar lateral 148 026C E2BB

fricative

t5 L-Ezh ligature Voiced dental or alveolar lateral 149 026E E2BCfricative

u Cursive V Voiced labiodental approximant 150 028B E2AE1 Turned R Voiced dental or alveolar 151 0279 E2BA

approximant{ Turned R, right Voiced retroflex approximant 152 027B E2CB

tailj Lower-case J Voiced palatal approximant 153 DOM E2DB

Uj Turned M, right Voiced velar approximant 154 0270 E2E2legLower-case L Voiced dental or alveolar lateral 155 006C E2BD

approximant

1 Right-tail L Voiced retroflex lateral 156 026D E2CCapproximant

I. Turned Y Voiced palatal lateral approximant 157 028E E2DC


L Small capital L Voiced velar lateral approximant 158 029F E2F5

PI

Hooktop P Voiceless b~labial implosive 159 0lA5 2378

WitJulrawn f993)5 Hooktop B Voiced bila, ial implosive 160 0253 E2A9f Hooktop T Voiceless dJntal or alveolar 161 OIAD 2379

Iimplosive WitJulrawn (1993)

cf Hooktop D Voiced dentlu or alveolar implosive 162 0257 E2C2e Hooktop C Voiceless palatal implosive 163 0188 2376

f Hooktop barredWithdrawn ~1993)

Voiced palar implosive 164 0284 E27Cdotless J

Ie Hooktop K Voiceless v~lar implosive 165 0199 2363WitJulrawn (1993)

9 Hooktop G Voiced velar implosive 166 0260 E27E

cf Hooktop Q Voiceless u~ular implosive 167 02AO E2F6Withdrawn ~1993)

G Hooktop small Voiced uvulrr implosive 168 029B E2FIcapital G

IM Turned W Voiceless labial-velar fricative 169 028D E2A7w Lower-case W Voiced labial-velar approximant 170 0077 E2A8q Turned H Voiced labia)-palatal approximant 171 0265 E2A6

H Small capital H Voiceless epiglottal fricative 172 029C E2F2-; Barred glottal Epiglottal plpsive 173 02AI E2F?

stop I~ Barred reversed Voiced epiglottal fricative or 174 02A2 E2F8

glottal stop approximantfj Hooktop heng Simultaneous voiceless 175 0267 E2D6

postalveolar:and velar fricative0 BuB's eye Bilabial click 176 0298 E2AA

I Pipe Dental click 177 OICO 23A6

! Exclamation (Post)alveol<,\f click 178 0lC3 23A5

point I

:f Double-barred Palatoalveol~ click 179 0lC2 23A4

pipe

II Double pipe Alveolar lateral click 180 OICI 23A7i

1 Turned long-leg Voiced alve91ar lateral flap 181 027A E2BF

RC Curly-tail C Voiceless alveolo-palatal fricative 182 0255 E2CE

I~ Curly-tail Z Voiced alve0lo-palatal fricative 183 0291 E2CF

Turned T Dental click 201 0287 E2C3

Superseded by 177 (1989)


C Stretched C PostalveoJar click 202 0297 E2C4Superseded by 178 (1989)

!> Inverted glottal Alveolar lateral click 203 0296 E2C5stop Superseded by 180 (1989)

I Curly-tail esh Withdrawn (1989) 204 0286 E2D2

d Curly-tail ezh Withdrawn (1989) 205 0293 E2D3

r Long-leg R Withdrawn (1989) 206 027C E2BE

Superscript S Withdrawn (1989) 207 02E2 FDA7Left-hook T Palatalized voiceless dental or 208 OIAB E2AO

alveolar plosive Withdrawn (1989)t L with tilde Velarized voiced dental or alveolar 209 026B E27D

lateral approximantg Looptail G Voiced velar plosive 210 0067 E2E3

Equivalent to 110ts T-S ligature Voiceless dental or alveolar 211 02A6 E2FC

affricate Superseded by 103+132dz D-Z ligature Voiced alveolar affricate 212 02A3 E2F9

Superseded by 104+133

tf T-Esh ligature Voiceless postalveolar affricate 213 02A7 E2FDSuperseded by 103+134

d3 D-Ezh ligature Voiced postalveolar affricate 214 02A4 E2FASuperseded by 104+135

t.e T-Curly-tail-C Voiceless alveolo-palatal affricate 215 02A8 E2FEligature Superseded by 103+182

~ D-Curly-tail-Z Voiced alveolo-palatal affricate 216 02A5 E2FBligature Superseded by 104+183

e Superscript theta Voiceless dental fricative release 217 E21B~ Superscript Mid central vowel release 218 E2lA

schwac( Hooktop right- Voiced retroflex implosive 219 E219

tail D Not explicitly [PA approved

2- Barred two Withdrawn (1976) 290 OIBB E2B5){ Turned K Withdrawn (1979) 291 029E E2F4x Superscript X Voiceless velar fricative release 292 02E3 D56F

IJ. N, right leg Syllabic nasal Withdrawn (1976) 293 019E E2E5};. Barred lambda Voiceless dental or alveolar lateral 294 019B FD7B

affricate Not [PA usageA. Lambda Voiceless dental or alveolar lateral 295 03BB 266E

fricative Not [PA usage

180 Handbook of the /PA

Z Z wedge Voiced postalveolar fricati ve 296 017E FIFONot /PA usdge

~ S wedge Voiceless p~stalveolar fricative 297 0161 FIDC

J J wedgeNot /PA uSalfe

298 OIFO E290Voiced postalveolar affricateI

Not /PA usageC C wedge Voiceless p6stalveolar affricate 299 OIOD FlAB

Not /PA usage

Lower-case I Close front lmrounded vowel 301 0069 E251e Lower-case E Close-mid fiont unrounded vowel 302 0065 E256£ Epsilon Open-mid f iont unrounded vowel 303 025B E258a Lower-case A Open front Jnrounded vowel 304 0061 E25Ba Script A Open back unrounded vowel 305 0251 E26C;) Open 0 Open-mid back rounded vowel 306 0254 E26B0 Lower-case 0 Close-mid b:ack rounded vowel 307 006F E269u Lower-case U Close back wunded vowel 308 0075 E265

Y Lower-case Y Close front lounded vowel 309 0079 E252

!/> Slashed 0 Close-mid font rounded vowel 310 OOF8 E257re Lower-case O-E Open-mid fr,ont rounded vowel 311 0153 E259

ligatureCE Small capital O-E Open front r,ounded vowel 312 0276 E25C

ligatureI

0 Turned script A Open back rounded vowel 313 0252 E26DA Turned V Open-mid b~ck unrounded vowel 314 028C E26Ay Ram's horns Close-mid back unrounded vowel 315 0264 E268ill Turned M Close back ~nrounded vowel 316 026F E264i Barred I Close centr~l unrounded vowel 317 0268 E25D

l:l Barred U Close central rounded vowel 318 0289 E25E

r Small capital I Near-close 'lear-front unrounded 319 026A E254vowel

Iy Small capital Y Near-close near-front rounded 320 028F E255

vowel

u Upsilon Near-close near-back rounded 321 028A E267vowel i

cl Schwa Mid central vowel 322 0259 E25FI

0275 E261e Barred 0 Close-mid central rounded vowel 323

Turned Ai

324 0250 E263B Near-open central vowel

re Ash; Lower-case Near-open front unrounded vowel 325 00E6 E25A

A-E ligatureOpen-mid c~ntral unrounded vowel3 Reversed epsilon 326 025C E262

Computer coding of [PA symbols 181

<3' Right-hook R-coloured mid central vowel 327 025A E260schwa Equivalent to 322+419

1 Undotted I Not 1PA usage 394 0131 00F5(3 Closed reversed Open-mid central rounded vowel 395 025E E270

epsilonf) Closed epsilon Superseded by 395 (1996) 396 029A E273~ Reversed E Close-mid central unrounded vowel 397 0258 E26E(i) Closed omega Near-close near-back rounded 398 0277 E266

vowel Superseded by 321 (1989)t Iota Near-close near-front unrounded 399 0269 E253


Table 6Phonetic diacritic and suprasegmental symbol codes(in numerical order by IPA Number)

Phonetic PlacementJ IPA UCS AFIISymbol Symbol Name Description Status Number Code Code

Apostrophe Ejective p' k' ts' If' 401 02BC E249Under-ring Voiceless l} Q 402A 0325 E229

Over-ring Voiceless ~ 9 4028 030A OOCASubscript wedge Voiced § Is 403 032C E22A

Superscript H Aspirated ph th 404 02BO D565Subscript umlaut Breathy voiced R ~ 405 0324 E22BSubscript tilde Creaky voiced g g 406 0330 23D8Subscript seagull Linguolabial t cJ 407 033C 22E8Subscript bridge Dental I) 9 408 032A E22CInverted subscript Apical l} g 409 033A 23FDbridgeSubscript square Laminal IJ Q 410 033B 23FESubscript right More rounded y y 411 0339 E23Chalf-ringSubscript left Less rounded y 412 03IC E232half-ringSubscript plus Advanced If 413 031F E233Under-bar Retracted ~ 414 0320 E234


Umlaut Centralized e 415 0308 E221x

Over-cross Mid-centratzedx

416 033De 2311

Advancing sign Advanced t ngue t,I 417 0318 23F7

root IRetracting sign 418 0319 23F8Retracted tongue ~

rootRight hook Rhoticity e-- a-- 0" Cl" 419 02DE E28A

W Superscript W Labialized tW dW 420 02B7 D56E

Superscript J Palatalized tj dj 421 02B2 D567v Superscript Velarized ' tv dv 422 02EO E28B

gamma,Superscript Pharyngeal ized t' d' 423 02E4 E28Creversed glottalstopSuperscript tilde Nasalized e a 424 0303 E222Superscript N Nasal release dn 425 207F FDA8Superscript L Lateral rele'se dl 426 02E1 FDA3

Corner No audible!,

d' 427 031A 23F9Prelease

Superimposed Velarized or B 8 428 0334 E226tilde pharyngealirzedRaising sign Raised y } (ec oJ.) 429 03ID/ E22F

02D4Lowering sign Lowered ~ 13T (eT OT) 430 031E/ E231

02D5Syl1abicity mark Syllabic f '! 431 0329 E22ESubscript arch Non-syllabiF ~ 432 032F 23FA

I

Top tie bar Affricate or,

kP 9b iSct3 433 0361 E225doublearticulation

Subscript right Rhoticity ct It ~ ~ 489 0322 E228

hook Superseded by419 (1989)

Open corner Release/ burst Not IPA usage 490 E218

Comma Pause (comma) Not IPA usage 491 002C 002C

Reversed Weak aspir~tion Withdrawn 492 02BB 00A9

apostrophe (1979)

Computer coding of lPA symbols 183

Over-dot Palatalization/ Withdrawn 493 0307 E224centralization (1979)

Minus sign Retracted variety Use 414 or 494 0207 E239(backed) 418 (1989)

+ Plus sign Advanced variety Use 413 or 495 02D6 E238(fronted) 417 (1989)

y Superscript Y High-front Not [PA usage 496 02B8 0570rounding/palatalized

Under-dot Closer variety/ Use 429 497 0323 E230fricative (1989)

Subscript left Palatalized Superseded by 498 0321 E227hook 421 (1989)Subscript W Labialized Superseded by 499 032B E22D

420 (1989)

Vertical stroke Primary stress ,foun~'tII~n 501 02C8 E23E(Superior)Vertical stroke Secondary stress ,foun~'tII~n 502 02CC E23F(Inferior)Length mark Long e: 503 02DO E23AHalf-length mark Half-long e' 504 02Dl E23BBreve Extra-short e 505 0306 E223Period Syllable break li.rekt 506 002E 002E

I Vertical line Minor (foot) 507 007C 007C(thick) group

II Double vertical Major 508 2016 2142line (thick) (intonation)

groupBottom tie bar Linking (absence ,fa:l~~'weI 509 203F 230F

of a break)/" Upward diagonal Global rise 510 2197 EF3E

arrow'\, Downward Global fall 511 2198 EF3D

diagonal arrowDouble acute Extra high level a 512 030B OOCDaccent (over)Acute accent High level 6 513 0301 00C2(over)Macron Mid level a 514 0304 00C5


Grave accent Low level a 515 0300 OOCI(over) IDouble grave Extra low I· vel a 516 030F 23E2accent (over)

.j. Down arrow Downstep 5\7 2193 EEAFt Uparrow Upstep 5\8 2191 EEAD

l Extra-high tone Extra high level mal 519 02E5 E28Dletter

1 High tone letter High level ma1 520 02E6 E28E-i Mid tone letter Mid level I ma-l 521 02E7 E28F-l Low tone letter Low level ma-l 522 02E8 E29FJ Extra-low tone Extra low level maJ 523 02E9 E29E

letterWedge; hacek Rising contour Ii 524 030C OOCFCircumflex Falling contour a 525 0302 OOC3Macron plus High rising I

J

526a E296acute accent contourGrave accent plus Low rising

~

527a E297macron contourGrave plus acute Rising-falling - 528a E298plus grave accent contour

t1 Rising tone letter Rising contour maA 529 E299'J Falling tone letter Falling contour ma'J 530 E29A1 High-rising tone High rising ma1 531 E29B

letter contourA Low-rising tone Low rising rna.-! 532 E29C

letter contour1 Rising-falling Rising-falling ma-J 533 E29D

tone letter contourSuperscript arch Long falling tone/ Not IPA usage 595 0311 23F2

advanced! p~latal

Wedge;hacek Falling-rising Usage re- 596 02C7 E247tone defined (1989)

See 524Circumflex Rising-falling Usage re- 597 02C6 E246

tone I defined (1989)See 525

Subscript grave Low falling tone Superseded 598 0316/ E245accent (/989) 02CESubscript acute Low rising tone Superseded 599 0317/ E243accent (1989) 02CF


AcknowledgementsWe would like to thank Alexandra (Smith) Gaylord for initiating discussion in 1989 between the phoneticscommunity and the Text Encoding Initiative, and for interpreting phonetic symbols in the context of codingstandardization. We are also indebted to Hans G. Tillmann, William J. Barry, and H. Joachim Neuhaus fortheir contribution to the Workgroup on Computer Coding of IPA Symbols and Computer Representation ofIndividual Languages, and for their advice on developing principles for the computerized documentation ofphonetic databases. Suggestions on the technical presentation of these tables and on details of symbolcoding from Christoph Drax1er and from Michael Everson are gratefully acknowledged.

ReferencesBALL, M. 1. (1991). Computer coding of the IPA: Extensions to the IPA. Journal of the

International Phonetic Association 21, 36-41.DUCKWORTH, M., ALLEN, G., HARDCASTLE, W. AND BALL, M. J. (1990). Extensions to

the International Phonetic Alphabet for the transcription of atypical speech. ClinicalLinguistics and Phonetics 4, 273-80.

ESLING, 1. H. (1988). Computer coding of IPA symbols and Detailed phoneticrepresentation of computer databases (1989 Kiel Convention coordinator's report).Journal of the International Phonetic Association 18,99-106.

EsLING, J. H. (1990). Computer coding of the IPA: Supplementary report. Journal of theInternational Phonetic Association 20, 22-6.

GIBBON, D., MOORE, R. AND WINSKI, R. (editors) (1997). Handbook ofStandards andResources for Spoken Language Systems. Berlin: Mouton de Gruyter.

IPA (I 989a). The International Phonetic Alphabet (revised to 1989). Journal of theInternational Phonetic Association 19(2), centre pages.

IPA (I989b). The IPA Kiel Convention Workgroup 9 report: Computer coding of IPAsymbols and computer representation of individual languages. Journal of theInternational Phonetic Association 19,81-2.

IPA (1993). The International Phonetic Alphabet (revised to 1993). Journal of theInternational Phonetic Association 23( I), centre pages.

PULLUM, G. K. AND LADUSAW, W. A. (1986). Phonetic Symbol Guide. University ofChicago Press.

PULLUM, G. K. AND LADUSAW, W. A. (1996). Phonetic Symbol Guide, 2nd edition.

University of Chicago Press.UNICODE CONSORTIUM (1996). The Unicode Standard, Version 2.0. Reading, MA:

Addison-Wesley Developers Press.

Aippendix 3

Extensions to the IPA: The ExtlPA Chart

This section on extended IPA sYmbOls~was prepared by the Executive Committee of theInternational Clinical Phonetics and I inguistics Association (ICPLA), which can becontacted through the Secretary of th ICPLA, Wolfram Ziegler, Stadt. KrankenhausMiinchen-Bogenhausen, EKN, Dachauer Str. 164, D-80992 Miinchen, Germany. Theoriginal text and chart first appeared as an article in the Journal of the InternationalPhonetic Association 24, 95-8 (1994~, entitled 'The ExtIPA Chart', by the ICPLAExecutive Committee. The list of 1P1 Numbers for ExtIPA symbols which has beenadded here appeared originally in an artIcle by Martin J. Ball entitled 'Computer coding ofthe IPA: Extensions to the IPA' in the Journal of the International Phonetic Association21, 3~1 (1991).

1 Introduction IAt the 1989 Kiel Convention of the IPA, a sub-group was established to draw uprecommendations for the transcription pf disordered speech. The report produced at theConvention appeared in print in Duckworth, Allen, Hardcastle, and Ball (1990), beingmainly a list of symbols, termed 'ExtenJions to the IPA', or 'ExtIPA' for short. Examplesof the use of the ExtIPA symbols are giv,en in Ball (1991), Howard (1993) and Ball, Code,Rahilly and Hazlett (1994) among otherS, and Ball (1993) includes them at textbook level.Various changes and additions to the ori~inal set of symbols are reported in Bernhardt andBall (1993). The changes meant, however, that there was a need for a listing of the currentset of symbols, preferably in as concise a form as possible. There was also felt to be a needfor the symbol set to receive overt recognition from a relevant society, therefore the 1994publication in JIPA marked the offi¢ial adoption of the ExtIPA symbols by theInternational Clinical Phonetics and Linguistics Association (ICPLA).

2 The ChartThe chart included here is clearly based on the current format of the IPA Chart to make itdirectly comparable and equally user-friendly. There are, however, a few comments that

Iwill aid in the use of the chart. First, sha~ed squares on the main chart represent - as in theIPA Chart - articulations that are deenied impossible to make. Blank squares stand forsounds for which the IPA already pro~jdes symbols, or for sounds that are possible tomake but for which no symbols are available. Clearly, this difference is straightforward toresolve with access to both charts.

The place of articulation labels used on the chart are mostly straightforward:dentolabial is the reverse of labiodental (upper lip to lower teeth), labioalveolar is neededfor speakers with gross overbite such that bilabial and labiodental target articulations arerealized with the lower lip articulating against the alveolar ridge. The linguolabial place ofarticulation does appear (as a diacritic) on the IPA Chart, but it is also included here as it

Extensions to the lPA 187

does occur reasonably frequently in disordered speech. The velopharyngeal place ofarticulation occurs only as a fricative: the friction originates at the velopharyngeal port(and can also accompany other articulations as noted below).

The manner of articulation labels are similar to those used on the IPA Chart with threeexceptions: the lateral plus central (median) fricative manner involves a simultaneouscentral and lateral release of friction, such as is found with some misarticulations of targetalveolar fricatives. The percussive manner involves the striking together of two rigid orsemi-rigid articulators. While the most common example of a percussive is found with abidental place of articulation, bilabial percussives are occasionally found also. Finally,there is the nareal fricative category. This sound type involves audible nasal release duringthe production of a nasal segment. As noted below, a diacritic is available to mark audiblenasal release with other sound types, and is used on the chart with nasal symbols as theyare probably the most common sounds to occur of this variety. These symbols replacepreviously proposed composite symbols of the type [!Lm), the change being approved atthe 1994 ICPLA Symposium.

At the 1989 IPA meeting at Kiet, the sub-group on disordered speech, following onfrom the recommendations of the PRDS Group (1983), recommended that as analternative to [t1)J, clinical phoneticians might wish to employ [llJ], to pattern with thesymbols for labiodental plosives. This might be especially useful in a patient presentingwith labiodental as a favourite articulation, in that the use of the dental diacritic added to arange of different symbols would clearly stand out in a transcription. This usage was,however, controversial as it was the only symbol directly competing with an official IPAsymbol, and recognition was withdrawn from it at the 1994 ICPLA Symposium.

The sets of diacritics beneath the main chart are mainly self-explanatory. The backslash for reiterated articulation can be used between individual segments or betweensyllables, but presupposes rapid reiteration; reiteration with pauses between the segmentsshould be shown with the pause markings listed in the 'Connected Speech' section.Sliding articulation is described more fully in Bernhardt and Ball (1993), but we can notehere that it is to be used to mark a rapid movement from one place of articulation to aneighbouring place within the time normally allotted to a single segment. The nasal escapediacritic is used to mark audible nareal friction accompanying another sound and,similarly, velopharyngeal friction accompanying another sound is marked with the doubletilde.

The voicing diacritics allow fine discrimination of amounts of voicing with particularsounds. There is also a diacritic for unaspirated plosives. It might be argued that it isunnecessary to mark the absence of something (i.e. aspiration). However, speechpathologists have found such a diacritic useful due to the ambiguous nature of a plainfortis plosive symbol, standing in narrow transcription for an unaspirated plosive, but inbroad transcription for a plosive unspecified for aspiration, and the need some transcribersmight have to note overtly a non-normal lack of aspiration.

Finally, the symbolizations in the 'other' category can be commented on. The'indeterminate' system utilizes the 'balloon' to mark sounds about which the transcriber is


uncertain. The amount of information t~e transcriber is sure about is entered within thebaJloon in terms of categories (e.g. consonant), or features (e.g. bilabial, plosive, etc.). Thetranscriber can also enter a specific phonJtic symbol, in which case this is read as meaning'probably [f]' etc. 'Silent articulation'l or 'mouthing' is relatively common in somedisordered speakers. Here there is a visible articulatory gesture, but no sound is made.Naturally, labial gestures may be the o~ly ones easily spotted here, but the use of thismarking convention does allow the anal~st to distinguish between speakers who omit asound and those who attempt it. I

3 ConclusionI

As with the IPA Chart, we expect this crart to be subject to regular revision by ICPLAmembers, especially in the light of clinic~1 experience. The chart may be copied by thoseinterested in clinical phonetic transcripti'on, as long as acknowledgement is made to theICPLA. ExtIPA symbols are listed in th~ following table, with an indication of the IPANumber that has been assigned to each symbol. Any symbol which also appears in thetables in appendix 2 is cross-referenced here.

Extended IPA character set: Symbol n~mes and IPA Numbers

ISymbol Name Phonetic Phonetic Description IPA

Symbol NumberConsonanUvowel symbolsDouble bridge [ : ] Bidental percussive 601

F-Eng ligature [ t) J Velopharyngeal fricative 602L-S ligature [ Il ] Lateralized [s] 603L-Z ligature [ Iz ] Lateralized [z] 604

Indeterminacy symbolsBalloon 0 Unidentified segment(s) 611

Asterisk * Placeholder symbol 612

Voice symbolsUpper-case V V Voice 621 (=722)

Upper-case F F Falsetto 622 (= 706)Upper-case W W Whisper 623 (= 723)

Upper-case C C Creak 624 (= 703)

Upper-case L L Larynx 625 (= 712)

Upper-case J J Jaw 626 (= 710)

Upper-case O-E ligature CE Oesophageal speech 627

Upper-case Greek Theta e Protruded-tongue voice 628

Extensions to the IPA 189

Connected speechBracketed single period (.) Short pause 631 (use 506)Bracketed double period ( ..) Medium-length pause 632 (use 506)Bracketed triple period ( ...) Long pause 633 (use 506)Forte f Loud speech 634Fortissimo ff Louder speech 635Piano p Quiet speech 636Pianissimo pp Quieter speech 637Allegro alkg,o Fast speech 638Lento fclto Slow speech 639

Numeral one I Slight degree 640Numeral two 2 Moderate degree 641Numeral three 3 Extreme degree 642

Diacritics

Superscript bridge[ " ]

Dentolabial 651

Subscript double arrow [,J Labial spreading 652

Superscript+subscript bridge [ : ] Bidental articulation 653

Superscript slashed tilde [ + ] Denasal 654Superscript dotted tilde [ ,: ] Nasal escape 655Superscript double tilde [= ] Velopharyngeal friction 656Subscript double syllabicity [" ] Stronger articulation 657markSubscript comer [ , ] Weaker articulation 658Backslash [\ ] Reiterated articulation 659Subscript upward arrow-head [,] Whistled articulation 660Down full arrow P] Ingressive air flow 661

Up full arrow [ 1" ] Egressive air flow 662

Left-sited subscript wedge L ] Pre-voicing 663 (use 403)Right-sited subscript wedge [ J Post-voicing 664 (= 403)Bracketed subscript wedge Lv.] Partial voicing 665Left-bracketed subscript L] Initial partial voicing 666wedgeRight-bracketed subscript [v,] Final partial voicing 667wedgeBracketed under-ring C.,] Partial devoicing 668Left-bracketed under-ring L. ] Initial partial devoicing 669Right-bracketed under-ring [ ,,] Final partial devoicing 670

190 Handbook of/he lPA

Left-sited superscript H [h J Pre-aspiration 671 (use 404)Right-sited double [ !!] Ventricular 672exclamation pointSuperscript cursive V [ U] LabiodentaJized 673Subscript right arrow [ ~ ] Slurred articulation 674Subscript equals sign [ =] Alveolarized 675Superscript inverted small [ "] Uvularized 676capital RSuperscript double-barred H [ H] Faucalized 677Superscript O-E ligature [ ~] Labialized: Open-rounded 678Right-sited exclamation point [ !] Harsh 679Superscript equals sign [ =] Unaspirated 680Subscript left pointer [ , ] Laterally offset to the right 681Subscript right pointer [ ] Laterally offset to the left 682Right-sited subscript tilde [ -] Creaky 683 (= 406)

Existing IPA diacriticsymbolsUnder-ring l . ] Voiceless 402Subscript wedge [ - ] Voiced! Diplophonia (VQ) 403

(= 663, 664)Superscript H [ h] Aspirated 404 (= 671)Subscript umlaut [ .. ] Breathy 405Subscript tilde [ - ] Creaky 406Subscript seagull [ - ] Linguolabial 407

Subscript bridge L] Dental 408Inverted subscript bridge [ u ] Apical 409

Subscript square [ D ] Laminal 410

Subscript plus [. ] Advanced 413

Retracting sign [, ] Retracted 418

Right hook [ ~] Rhoticity (Retroflex) 419

Superscript W [ W] Labialized: Close-rounded 420

Superscript J [ j] Palatalized 421

Superscript gamma [ V] Velarized 422

Superscript reversed glottal l ~] Pharyngeal ized 423

stopSuperscript tilde [- ] Nasalized 424

Raising sign [. ] Raised 429

Lowering sign [. ] Lowered 430

Comma Comma 491

Extensions to the [PA 191

Under-dot [,] Whispery 497Period Pause 506

(631.632.633)Bottom tie bar ( - ] Linking 509

Upper-case lettersUpper-case A A A 701Upper-case B B B 702Upper-case C C C (Creak) 703 (use 624)Upper-case D D D 704Upper-case E E E 705Upper-case F F F (Falsetto) 706 (use 622)Upper-case G G G 707

Upper-case H H H 708Upper-case I I I 709Upper-case J J J (Jaw) 710 (use 626)Upper-case K K K 711Upper-case L L L (Larynx) 712 (use 625)Upper-case M M M 713Upper-case N N N 714Upper-case 0 0 0 715Upper-case P P P 716Upper-case Q Q Q 717Upper-case R R R 718Upper-case S S S 719Upper-case T T T 720Upper-case U U U 721Upper-case V V V (Voice) 722 (use 621)Upper-case W W W (Whisper) 723 (use 623)Upper-case X X X 724Upper-case Y Y Y 725Upper-case Z Z Z 726

Transcription delimitationcharactersLeft square bracket ( Begin phonetic transcription 901Right square bracket ] End phonetic transcription 902Slash / BeginJ end phonemic 903

transcriptionLeft parenthesis Silent articulation (Mouthing) 906Right parenthesis Silent articulation (Mouthing) 907

192 Handbook ofthe [PA

908909910911

~ound obscured~ound obscured~egin prosodic notation

IOd prosodic notation

«»{}

ReferencesBALL, M. J. (1991). Recent developments in the transcription of non-normal speech.

Journal of Communication Disorders 24,59-78.BALL, M. J. (1993). Phonetics for Speech IPathology, 2nd edition. London: Whurr

Publishers. IBALL, M. J., CODE, C., RAHilLY, J. AND HAZLETT, D. (1994). Non-segmental aspects of

disordered speech: Developments in tr~nscriPtion. Clinical Linguistics and Phonetics8,67-83.

BERNHARDT, B. AND BALL, M. J. (1993). Characteristics of atypical speech currently notincluded in the Extensions to the IPA. Journal of the International PhoneticAssociation 23, 35-8.

DUCKWORTH, M., ALLEN, G., HARDCASfE, W. AND BALL, M. J. (1990). Extensions tothe International Phonetic Alphabet for the transcription of atypical speech. ClinicalLinguistics and Phonetics 4, 273-80. I

HOWARD, S. (1993). Articulatory constraints on a phonological system: A case study ofcleft palate speech. Clinical Linguistic~ and Phonetics 7, 299-317.

PRDS GROUP (1983). The Phonetic Representation ofDisordered Speech: Final Report.London: The King's Fund.

Left double parenthesisRight double parenthesisLeft braceRight brace

ExtIPA SYMBOLS FOR DISORDERED SPEECH(Revised to 1997)

CONSONANTS (other than those on the IPA Chart)bilabial labiodental dcnrolabial Iflbioalv. lin.e:uolabial interdent.a.l bidental alveolar velar velODhar.

Plosive

Nasal

Trill

Fricalive:central

ill

f V t V

n

r

e QF~~::~: 1.<. .... . I~:'.;:'i

f-ce_n'_"_1_rc~-,,-.·_·_-.-r-......"'""+-'-""'.,~--'I'I_"'.~,•.-;"'''''-',+---+----+....'"..,.i•.;;.'.. .,:."\ .• :,I-E__lz__. -+- . --+;.:;;-.;~ -'.Fricative: ":t:" ,

r-na_re_al_+-_ffi_+-__+-__+-__+-__t-__I-'-I',·-.=·_'""""'t"li..__n-t_---=I):.......+.....-,Percussive W M ~~.--, .~.:-.-

W M !e~

DIACRITICS

labial spreading ..§ strong articulation f ... denasa] 01-.... .. m

- weak articulation v nasal escape -dentolabial v , - v,

- interdental/bidental \ reiterated articulation p\p\p: velopharyngeal friction s- (j

alveolar J whistled articulation s ! ingressive airflow pJ.= ~ ~

- linguolabial Q .....sliding articulation ~ i egressive airflow !i

CONNECTED SPEECH() short pause(.. ) medium pause(... ) long pausef loud speech [{ ( laud rl]ff louder speech [{ ff laud~ «) 1p quiet speech [In kwal;;lt n)Jpp quieter speech [{ pp kwat;;l!;)'- nn)1allegro fast speech ((alle.ro fa:st alle.ro)][enID slow speech [{ lenlo slou Jenlo)Jcrescendo. rallentando. etc. may also be used

OTHERS

VOICING

pre-voicing zpost-voicing zpartial devoicing '~I

( initial partial devoicing z, final partial devoicing z,

( , partial voicing (§)

( initial partial voicing ,§

final partial voicing §,- unaspirated p=h pre-aspiration h p

n indeterminate sound (( )) extraneous noise ((2 sylls))

(v). (fl) indeterminate vowel, plosive, etc. i sublaminal lower alveolar percussive click

(fuJ&) indeterminate voiceless plosive, etc. (. alveolar & sublaminal click ('cluck-click')'1

( ) silent articulation (J), (m) * sound with no available symbol

© 1997 ICPLA Reproduced by permission of the Intemational Clinical Phonetics & Linguistics Association.

APrendiX 4

About the Internatipnal Phonetic Association

The History of the International Phonet Association

1 The Association \The Association was founded in Paris early in 1886 under the name of Dhi Fonetik Tfcerz'As6ciecon (the FTA), which was itself a development of L'Association Phonetique desProfesseurs d'Anglais. A small group of language-teachers came together, under theleadership of Paul Passy, to press the case for phonetic notation to be used in schools as amethod of helping children to acquire a ~ealistic pronunciation of foreign languages. Afurther use of phonetic notation was s~en to be in the teaching of reading to youngchildren. In these aims, particularly the former, the Association was supported by theexamples and encouragement of language-teachers and phoneticians in other countriessuch as Johan Storm in Norway, Henry Sweet and Henry Widgery in Britain, andHermann Klinghardt and Wilhelm Vietor in Germany.

In less than a year, the membership !had grown from an initial eleven members inFrance to 58 in twelve countries, mainly in Western Europe. By the summer of 1914,which was the high point of the Associhtion's existence in terms of membership andinfluence in education circles, there were] 751 members in 40 countries. In January 1889,the name of the Association was changed to L'Association Phonetique des Professeurs deLangues Vivantes (AP), and, in 1897, to 4'Association Phonetique Intemationale (AP!)in English, the International Phonetic Association (IPA). The Association's activities wereseverely disrupted by the First World Warl and its aftermath. No journal appeared betweenOctober 1914 and December 1922, alt~ough circumstances did allow the occasionalpublication of a handful of Supplements o~ particular aspects of phonetics.

It was the Danish phonetician Otto Jespersen who first suggested, in 1886, that thereshould be what he called an 'International Phonetic Association', which would devise aninternational alphabet rather than a se~ of different alphabets sp~cific to particularlanguages (see 4 below on the developmert of the Alphabet). Passy hImself favoured thecompromise of an 'International Society' Imade up of a confederation of national bodies,whose aim would be 'the advancement of the science of phonetics and its application topractical teaching'. Jespersen's proposal, however, failed to get the support of a majorityof the FfA's membership, and the Associ~tion continued as a forum for language-teachersdedicated to the cause of phonetic scri~t rather than as an organization of academicphoneticians. Within a few years, the g~adual increase in the number of items in thejournal dealing with matters unconnecte~ with the concerns of mainstream languageteachers (for example, specimens of Man~arin Chinese and Armenian) heralded the shift

Itowards phonetic rather than applied ph<j)netic concerns. Items dealing with language-teaching matters have, nevertheless, alwaYIcontinued to be published in the journal.

AboUl the lPA 195

2 The journalDuring the first few months of 1886, the Association's members kept each other informedof developments in the application of phonetic techniques in schools with the help of an'ever-circulator', but in May of that year a journal was published, Dhi Fonetik Ticer ('ThePhonetic Teacher') - FT for short. It quickly established itself as the focus of theAssociation's activities. (Another journal of the same name, emanating from the United

States, dealt solely with questions of spelling reform and was unconnected with the FT.)In January 1889, the name of the journal was changed to Le Maitre Phonetique (mf), atthe same time as French became the Association's official language. In 1971 the mfbecame the Journal of the International Phonetic Association (J1PA), and the officiallanguage reverted to English.

The journal was, from the very start, published as far as possible in phonetic script. (Inmany issues, a variety of phonetic scripts illustrating the range of symbolizations possiblewithin the Association's conventions was the norm.) It was only with the publication ofJIPA from 1971 onwards that phonetic script gave way to traditional orthography. Theformat of each issue of the journal has tended to be one of articles on aspects of the

descriptive phonetics of languages or language-varieties, comments on phonetic notationand examples in phonetic script (spesimen, as they were known in French in the mf) ofthe pronunciation of languages. The reading passage used to illustrate the pronunciation

was, normally, from 1911 onwards, the fable of 'The North Wind and the Sun'. Inaddition, smaller items were published describing the academic and social activities ofmembers of the so-called famille phonetique. Although the character of the articles hasalways tended to reflect the descriptivist articulatory phonetic leanings of the Association,experimental work of various kinds is increasingly represented in the journal.

3 Supplements to the journalA series of Supplements and loose inserts to the journal has been published at irregular

intervals since 1888. One of the most important of these has been the Principles of theInternational Phonetic Association, which first appeared in French in 1900, in English in1904, in German in 1928, in Italian in 1933 and in Spanish in 1944. The Principles (ofwhich this IPA Handbook is the successor) included an introductory section explainingphonetic notation, which was followed by transcriptions of 'The North Wind and the Sun'in about 50 languages from around the world. Further spesimen covering about another100 languages can be found in the journal itself. In addition, three issues of MiscellaneaPhonetica, collections of articles on phonetic topics, have been published (1914 (issued1925), 1954, 1958).

4 The development of the Association's AlphabetPhonetic notation has been one of the Association's central concerns from the verybeginning. The first alphabet that was employed and promulgated was a modification ofthe' 1847 Alphabet' of Isaac Pitman and Alexander J. Ellis. Originally, the aim was to


make available a set of phonetic symbols which would be given different articulatory

values, if necessary, in different languages. The choice of symbols was dictated by theneed to keep them as simple as possible, for the benefit of both teachers and schoolchildren. Since a large proportion of the membership was from Western Europe (or waslinked in cultural and linguistic terms with Western Europe), it was inevitable that a romanbase should be used for the Alphabet. Thus, to begin with, 'c' stood for the [J] in theEnglish word sheep. but the [fJ of French chat was represented by an 'x'. Then, as now,the policy for the development of the Alphabet involved members making proposals forchanges, which were published in the journal and voted on by the Association's Council.During the first half of 1887 there were some small alterations to the list of symbols.

Twelve months later, in August/Septemqer 1888, further revisions were proposed in thelight of a set of policy statements. The latter were to exert a profound effect and determinethe content of most later versions of the A!lphabet:

In choosing between the various proposals that have been made for revising our alphabet,we've been guided by the following principles, which I think are admitted by most of ourreaders as essential to a practical system ?f phonetic spelling:

There should be a separate sign for e,ach distinctive sound; that is, for each sound which,being used instead of another, in the same language, can change the meaning of a word.2 When any sound is found in several languages, the same sign should be used in all. Thisapplies also to very similar shades of sound.3 The alphabet should consist as much as possible of the ordinary letters of the romanalphabet; as few new letters as possible being used.

4 In assigning values to the roman letters, international usage should decide.5 The new letters should be suggestive of the sounds they represent, by their resemblance tothe old ones.

6 Diacritic marks should be avoided, being trying for the eyes and troublesome to write.

Other remarks emphasized the need for the transcription to be, in the terminology oftwentieth century phonological theories, 'phonemic', for stress to be predicted asoccurring on particular syllables within a word depending on the language in question, andfor contributors to the journal to notate their own pronunciation of their language.

During the 1890s further changes were made to the Alphabet, mainly to provide a

series of diacritics for transcriptions that ,were 'allophonic' (pace the 1888 proposal thatphonetic notation should be 'phonemic'), for prosodic features, and for widening thecompass of the Alphabet to accommodate languages well removed from the originalcluster of English, French, and German - for example, Arabic. The relatively rapiddevelopment of the Alphabet can be judged from the fact that the one that was issued inAugust 1899 is in many respects similar to today's, despite certain divergences in the

choice of symbols and their allocation to sound categories.Clearly, at that time the Alphabet was already developing into a tool for general

phonetic work rather than being merely a set of notational conventions for the phonemic

transcription of certain major languages taught in schools. Even so, the 'phonemic'

About the IPA 197

emphasis in defining the range of sounds to be symbolized has continued up to thepresent, as can be seen in successive elaborations of the Principles of the IPA. The mostrecent formulation can be found in appendix I.

The development of the Alphabet has not been without controversy. Particularlyduring the 1920s, changes suggested by a specially convened Conference on PhoneticTranscription and Transliteration would have affected many of the conventions; few ofthem, however, were incorporated into the Alphabet. In addition, other Alphabets haveexisted alongside the IPA, especially during the early period of the Association'sexistence, for example Alexander Melville Bell's Visible Speech alphabet, later modifiedand renamed the Organic Alphabet by Henry Sweet; Johan Lundell's Swedish DialectAlphabet; and Otto Jespersen's Danish Dialect Alphabet. None of them has, in the longerterm, offered serious competition to the pre-eminence of the Alphabet of the InternationalPhonetic Association as a widely, even if not universally, accepted standard. The Alphabethas always been subject to critical review, and consequent change, and this healthy cycleof renewal is certain to continue as new knowledge is brought to bear on it. It will be ofconsiderable interest to see how it develops in the future, under the influence ofdevelopments in phonological theory, the needs of new practical applications, and theknowledge about speech production and perception which experimental methods willcontinue to provide.

5 Examinations in PhoneticsA further aspect of the Association's acti vities since 1908 has been the organization ofexaminations in phonetics, leading to the award of a Certificate of Proficiency in thephonetics of English, French, or German. Examinations are still held for English, andfurther details and entry forms are available from the following address: IPA Examination,Department of Phonetics and Linguistics, University College London, Gower Street,London WCIE 6BT, UK, or via the IPA World Wide Web page (see 'How to find outmore.. .' below).

Statutes and By-laws of the Association

Statutes of the International Phonetic Association (1995)I The name of the Association is 'International Phonetic Association'.

2 The aim of the Association is to promote the scientific study of phonetics and thevarious practical applications of that science. In furtherance of this aim the Associationpublishes a journal entitled the Journal of the International Phonetic Association. TheAssociation considers that in pursuing its aim it makes a contribution to friendly relationsbetween peoples of different countries.

3 To become a member of the Association it is necessary to pay dues as fixed in the bylaws, and to be elected by vote of the Secretary acting on behalf of the Council.


4 The Association is administered by a Council of 30 members, 20 of whom will beelected by direct vote of the members. T~e elected Council has the power to co-opt up to10 further members, so as to ensure a suitable representation of phoneticians throughoutthe world.

5 Elections for the Council will be held every four years, in a manner as prescribed inthe by-laws.

6 The Council elects from its members a President, a Vice-President, a Secretary, aTreasurer, and an Editor; these officers 1;'orm an Executive Committee, reporting to theCouncil, and running the current affairs of. the Association.

7 The President and Vice-President are each elected for a single term of four years only.The Secretary, the Treasurer, and the Editor are each elected for a four-year term in thefirst place, and are eligible for re-election for not more than one further term of four years.

8 The Association holds a general business meeting at least once every four years,normally in conjunction with the four-yearly International Congress of Phonetic Sciences.At the request of any member present, any decisions taken at such a meeting are subject toratification by mail ballot of the Association.

9 Modification of these statutes requires a mail ballot of the membership, with a twothirds majority of those responding to the ballot.

By-laws ofthe International Phonetic Association (1997)I The annual dues are £13 or US $25 payable to the Treasurer on the first of Januaryeach year.

2 There are the following types of members:

Members, who pay annual dues.Life members, who pay no annual dues. Any member who has paid full annual duesfor 30 years will automatically become a life member. Alternatively, life membershipmay be obtained by paying a one-time sum of 15 times the annual dues.Student members, who pay half the annual dues. Student members must include withtheir dues a letter from their Head of Department or supervisor, certifying that they arecurrently full-time students.All of the above types of members are entitled to all the privileges of membership,including receipt of the Journal.Institutional members, who pay twice the current annual dues, and who receive theJournal, but are not otherwise entitled to the privileges of membership.

About the IPA 199

3 Elections to the Council will be by mail ballot. The time schedule for the ejection ofthe Council of the IPA will be as follows: (a) 12 months before an International Congressof Phonetic Sciences - the Secretary to request nominations for membership of theCouncil; (b) 10 months before - closing date for receipt of nominations, and for membersnominated to have notified the Secretary if they are willing for their names to appear on aballot; (c) 9 months before - ballot paper to be mailed to members; (d) 8 months beforecompleted ballot paper to be received by the Secretary; (e) 7 months before - theSecretary to request the new Council to make nominations for lOco-opted members; (f) 6

months before - closing date for the receipt of nominations for co-opted members; (g) 5months before - ballot paper for co-opted members to be mailed to Council; (h) 4 monthsbefore - completed ballot paper to be received by the Secretary; (i) 3 months before - theSecretary to request the new Council to make nominations for the posts of President, VicePresident, Secretary, Treasurer, and Editor; U) 2 months before - closing date for thereceipt of nominations for executive posts; (k) 2 months before - ballot paper forexecutive posts to be mailed to Council; (I) 1 month before - completed ballot paper to be

received by the Secretary.

4 Modifications of these by-laws require a two-thirds majority of the Council voting in a

mail ballot.

How to find out more about the AssociationMembers of the Association receive the twice-yearly Journal oj the InternationalPhonetic Association, which is also available in many libraries in academic institutions.This contains research articles, discussions, and news of the Association's activities.

A monthly electronic newsletter for the phonetic sciences, joNETiks, carries regularannouncements of IPA news and activities. This can be joined by sending an e-mail

message to: [email protected]. The message should consist of one line as follows:Join fonetiks firstname(s) lastname. There is also an IPA page on the World Wide Web;this gives up-to-date information about the IPA, including useful addresses and currentsubscription rates. The URL is: hUp:llwww.arls.gla.ac.uk/lPAlipa.hlml

How to join the AssociationAn application form is included in each number of the Journal oj the InternationalPhonetic Association and on the IPA Web Site. Completed application forms should bereturned to the Treasurer.

Appendix 5

Reference charts,

This appendix contains larger-scale versions of the IPA Chart together with thecorresponding IPA Number Chart, dividCd into sections for ease of reference. The chartsmay be copied and used while consulting parts I and 2 of the Handbook, or enlarged forteaching purposes. I


Bil,bi,1 IL.biodent.I1 Denl.1

LUJ

I\.J

l{

Poslalveolar Retronex P,I.t.1 Vel,,- UvuJar

t q C jlk glqIl J1

r~Iv J

1

tII)

B

mp b

Nos.1

Approximant

Plosive

ute",1fricative

Loter,]approximant

Fricative

Trill

T,p or FI.p

Where symbols appeilI in pairs. the one to the right represents a voiced consonant. Shaded orcas denote aniculations judged impossible.

IVo

158

154153

157

152

156

151

155

NUMBER CHART

Po,I,lveol,,- Reuotlex Vel,,-

103 104 105 106 109 110 I III

116 117-122

124 125

136 137

Bil,bi,1 IL,biodentnii DenlOl

Approximant

ute",1approximant

Where symbols appear in pairs. the one to the right represenLS a voiced consonant. Shaded areas denote anicul'llions judged impossible.

Plosive 101 102

Nos,1 114 115

Trill 121

T,p or FI,p

Fricative

Lllcralfricative


Y-i~-UI uI Y U

e (/)--5 8--Y 0

g

E ce-3\G-1\. J

re B

a CE~Q D

CONSONANTS (NON-PULMONIC)

Clicl<.s Voiced implosives EjecLives

0 5,

Bilabial Bilabial Examples:

I Denlal d Dental/alveolar p' Bilabial,(Posl)alveolnr f PaJala] t' DentaUa!veolar

:j: Palalo<Jlveolar 9 Velar k' Velnr

II Alveolar l<:llernl G Uvular S' Alveolar fricmive

VOWELS

Front

Close

Close-mid

Open-mid

Open

Central Back

NoN

,1!

CONSONANTS (NON-PULMONIC) VOWELS

Where symbols appear in pairs, the oneto the right represents a rounded vowel.

m~"316 308

319 320 321

310 -- 397 323-- 315 307

322

303 311 - 326\395- 314 306

325 324

304 312 ----.L- 305 313

Where symbols appear in pairs, the one10 the right represents a rounded vowel.

Clicks Voiced implosives Ejectives

176 BHabi..1 160 Bilnbial 401 Exnmples:

177 162101

BilabialDenml DenmlJalveolur +401103

178 (Posl)alveol<Jf 164 Palatal + 401 DentaUalveolar

179109

VelarPalatoJlveolnr 166 Velnr + 401132

Alveolar fric<llivc180 Alveolar lllleraJ 168 Uvular + 401

Fronl

Cia" 301 309

Close-mid 302

Open-mid

Open

Centra] Back

203

OTHER SYMBOLS

Affricillts and double articulationsenn be represented by two symbolsjoined by n lie bar if nccessnry.

Alveolnr Intern] fl:lp

SimullttneOU5 Jand X

Alvco)o-palntal (nctllivesc ~

1fj

Voiced epiglollal friclllive

Voiceless labial-velar fricative

Epiglollnl plo:;ive

Voiceless epiglollal fric."llive

Voicctllablill-pnlillal npproximnnt

VOIced labial-velar approximant

Nt..

W

OTHER SYMBOLS

169 Voiceless labinl-velar fric.uive 182 183 Alvcolo-palatal mCallyeS

170 Voiced lnbial-vclari1pproxirmlnl

17) VOIced labinl-pnlillnl npproxim:mr

181

175

Alveolar Ialcr;ll nap

Simull:lncous Jand X

172 Voiceless epigloual fricallve

174 Voiced epiglollnl fricntive

173 Epiglolfnl plosivc

Affricates and double ar1iculillionscan be represented by two symbols

joinc...d hy n lie bM If necessary.433 (509)

SUPRASEGMENTALS SUPRASEGMENTALS

Primary stress 501 Primary stress

Secondary Slress 502 SCl:ondary stress

,foun;;l'tlf~m ,fOUn;;l't1f;;lnLong e: 503 Long e:Half-long e' 504 Half-long e'EXlra~shon e 505 EXlra-shorl e

I Minor (foot) group 507 Minor (foot) group

"Mi1jor (inlonmion) group 508 Major (inlonaLion) group

Syllable break li.rekt 506 SyUable break li.rekt

Linking (absence of a breok) 509 Linking (absence of i\ brej]k)

TONCS AND WORD ACCENTS TONCS AND WORD ACCENTSLEVEL CONTOUR LEVEL CONTOUR

ear l EXlrlleor ;1 Rising 512 519 Extra 524 529 RiSinghigh highe , High e \J Falling 513 520 High 525 530 Fnlling

e -1 Mit! e -1 High514 521 Mid 526 531

Highrising rising

e ...j Low e A Low515 522

Lowrising Low 527 532 rising

e J Extra e 1 Rising- EXIra Rising-tow foiling 516 523 low 528 533 falling

.j,Downslep /' Glob::!1 rise 517 Downslep 510 Global ns~

't Upslep '\. Global (nil 518 UpSICp 511 Glob'll (.. II

l

204

oDIACRITICS Diacritics may be placed above a symbol with a desceoder. e,g.I]

Voiceless l) d .. Brc.1lhy voiced R ~ "Denlul ! <}

vVoiced § ! - Crenky voiced g g

uArk;)! ! g

hASp,raled th dh

Linguolabial 1 g Lamin<ll ! Q- c

9W tW dW - eMore rounded L.,bialized Nnsllli2.cd

:? J tJ dJ n dn(

Less rounded P:llnl:llizcd Nasi'll release

U Y tY dY I dl+

Advanced+

Vclarized LllcrnlrcJeasc:

~ t~ d~,

d'- Relrocled ~ Phnrynge<lliz.ed No <l.udible relense.. e tCenlmllled - Velllriu:d or ph:lryngcnliz.c.dX X

Mid·ccnlrnlilCd e ... R:used tt It = vOIced nlveolar fric;uive)

, Syllnbic t:J. TLowered e (~ ;:; voiced bilOlbinl approltimnnl)

~Non·syl1..blC ~ . Advanced Tongue Rool y

'" RholicilY ~ cr- Relrncled Tongue RoO! y.DIACRITICS Diacritics may be placed above a symbol with a deseeoder. e.g. 119 + 402B

402A Voiceless I) Q 405 Breathy voiced R ~ 408 Denial ! <}403 Voiced § ! 406 Creaky VOIced g g 409 Apical ! g404 Aspirated th dh 407 Linguolabio.l 1 g 410 Lamin<ll t Qc

411 More rounded 0 420 L'lbiolizcd tW dW424 Nllsnhzed e

412 Less rounded J 421 Pllhll;llizcd tj dJ 425 Nasnl rclCllSC dn(

413 Advonced U 422 Velnrized tY dY 426 Lnleral release d l+

414 Relr3.ctcd e 423 Pharyngcalized t'i d\' 427 No audible release d'415 Ccntralized e 428 Velnrized or phnryngealized 209

X

416 Mid-ccnlralized e 429 Raised tt It = voiced Dlveolar fric::llive)

431 Syllabic t:J. 430 Lowered y (.f? = voiced bilabial approximanl)

432 Non-syllabic ~ 417 Advanced Tongue ROOI e.419 Rholicity 327 cr- 418 RclnlCled Tongue Rool y

Documents

IPA Handbook Complete