1 OXFORD TEXTBOOKS IN LINGUISTICS ‘Gone edtors: Keth Brown, Univesity of Cambadoe: Eve V. Clark, Sinford University "Aor MeMabwn, Universy of Sheed: Fim Mille, University of Edintuesh: Lesley Milrey, Univesity of Michizan ‘This series provides aczessble and authoritative textbooks onthe approaches, methods, and theories asncited with the main sues of nguisies PuBuis4eo ‘A PnctealInrocucion to Phonics ‘Second ection by. Catford Meming mn Lanmage {An Inrodvcion to Semantics and Prasrates ‘oy Alan Cre Principles and Parameter ‘an Intrdetion @ Syntactic Thesry by Peter W. Colieorer Lingistie Reconsesetion ‘An tatrodnction to Theory and Medthod by Anthony Fox Semantic Analysis ‘A Practica Inioducion by Cliff Godéaré IN PREPARATION Cognitive Grammar ‘Nn Introduction plea R, Taylor The Gammar of Words An introduction 9 Linguitie Morptolory by Peofesior G. Bool) Pragmatics ay ean Fung Linguine Categorization “Thin edition dy fae R. Taylor A Practical Introduction to Phonetics Second Edition J.C. Catford OXFORDSeries editore ith Brows, Eve V. Clark, April MeMabon, Jim Millerané Lesley Mlitoy A Practical Introduction to Phonetics Second EditionOXFORD (creat Garendon Steet Orford 942 600 Onond University Pret ic adeparamentof che Universi of Oxford Ie farther the Universigys objective of excellence in research, scholarship, nd education oy publishing wordwide in ‘Oxford New Yor ‘Auckland Bangkok Buenos Aes Cape'Town Cienaai Dares Silat Dela Hong Kong Ieansul Karachi Reba eal: Lampor Madsid Melbourne: Mexico Cry Murabal Nairobi ‘Slo Paalo Stanghat Tape Toyo Torento Oxford i a registered ide mark of Oxford University Press inthe UX and inceraia other countries, ‘Published in the United States by Oxford Univerigy Press inc, New York: ©}. €. Catford 1988, 2001 ‘The moral rights of the authorhave beea asserted Database right Oxford Univesiy Press maker) estou by Oxford University es 1958 Reprinted with corecions 1994 Secon eduon 2001, Reprinted 2002, 2003 tpperemewver quan us woedye | ! Preface It may be werth drawing attention to the fact that the tite of this book is, designedly, ‘A Practical Introduction to Phonetics’ and not ‘An Introdue- tien to Practical Phonetics’ ori is, indeed, an introduction to general. or theoretical, phonetics, though it proceeds towards that goal in a highly practical way." i» Readers are introduced to the phonetic classification of the sounds of speech by mieane ofa series of simple introspective experiments carried out inside their owa vocal tacts their own throais and mouths By actually ‘making sounds (very often silently) ard attending to the muscular seasa- tions that accompany their production one can ciscover how they-are produced and leara how to deceribe and clasify them. ‘At fst sight ‘making sounds silently” may appear eontradiciony, but, as Abercrombie (1967) has aptly pointed out, speech is ‘audible gesture’ and [Ad sights reserved. No partof tis publication may be reproduced, ote ina etieval ters. or ronimitted, in ary foes ov by any mee, "shout the por permission in waiting of Oxdord University Pres, cas expressly permitted oy iw, or under terms agreed with the sppropeiaie Teprogaphie Aght onganiation enquiries concerning reproduction coutaide the scape ofthe above should be sent tothe Rights Deparonent, ‘Oxford University Pes. 2 the adress above ‘Yow must nt cincolats this boc in any aher binding ot cover the'principal aim of this book is to enable the reader to discover and to ‘analyse the gestural aspéct'of speech (upon which most phonetic classifica- tion is based) and to bringit under conscious control. This must be done, 10 ‘a large extent, in lence, since the auditory sensations of loud speech tend tomask the motor sensitions. which are the perceptual accompaniment of : ‘andvyou must impore tis same condition on any cui Beith Library Cataloguing in Publication Dara Dan availible "any nf Congract Catalngingin Bobiction Dots Carford J. (ohe Cunsison)19:7- [A praciaa intodvction to phonetics Bbbogmapay:. Tacludes ince. 1. Phonetice. 2: Grumman, Comparasivesnd ‘general—Phonology. 1 ie PLCS 1988 414 85-1256 ISBN 0-19-924698-1 357910864 “Typeset in Times and Stone Sane by RefigeCaten Limited, Banay SuTolk Panted in Great Brean “Ty. tncernational Ld, Padstow, Comwall the gestural aspect of speech. ‘That this Kind of experimentation is-aa effective means of acquiring a knowledge of the categories and principles of general phonetics I know from personal experience, for this was precisely how I learned phonetics 2s a boy, without a teacher, eagerly reading Sweet's Primer of Phoreties and constantly experimenting in my own vocal tract. Although. as this reference to boyaood experience sugsests, phonetics is «a fascinating hebly for young peopl, itis primarily an indispensable tool for all those adults who have to work with language: students of linguistics, teachers and students of languages, teachers of the deaf, the hearing- impaired themselves who may be striving to acquire intelligible speech, actors, and many others. Armed with the understanding of the basic pria- ciples of phonetics which this bock seeks to inculcate, they should be'abie to read and fully understand any specialist work on.whatever aspect of pphoneties is of special interest to them. - Much of the material of the book has been used for some years past atvi PREFACE the University of Michigan, in teaching phonetics to Targe groups of stu- dents of linguistics, spe ch pathology, anthropology, languages, education, druma, and many other elds. 1 am grateful to all those students who contributed comments nd I sould also like to thank Dr Harriet Mills who read most of the text and made aumerous valuable and suggestions, Some readers have Suggested that it would be useful to have a set of recordings of the experiments, or at least of the sounds that should result ‘when cach experiment is curried out. This might be helpful, but it also carries the risk that some readers might be Induce¢ to try and learn sounds ‘merely by imitation of heard examples. This would tend to deft the pur- pose of the book, which isto inculeate an anderstanding of the mechanism of speech and of phonetic taxonomy by experimentation in one’s own vosal act. However it may be useful for some to be able to hear exaruples of many of the unfamiliae sounds that they have been led to produce in these experiments ‘A cecording is now available of all of the Sounds represented on the latest chart of the International Phonetic Alpbabet, which can be found on pages, 114-15 of this book. This demonstration, The Sound of che IPA, spoken by John Wells, Professor of Phonetics, and Jil House, Senior Lecturer in Phonetics, at University College, London, is aailsble on audio cassette of CD. For further information weite to Listening Centce, Department of Phonctics and Linguists, UCL, 4 Stephenson Wary, Londoo NW1 2HE, UK (Fax: +4420 7383 0752, Email: [email protected]), new edition contains additions and corrections, and most impor- tantly, presents an expanded and updated list of items for further reading, fon pages 217 to 219. 1am particalarly grateful 10 Jimmy G. Harris and Professor John Esling for the many helpful suggestions thet they provided, aand 19 Jobn Davey and Jo Stanbridge of Oxford University Press for their dance and careful editorial work, Ce. Ann Arbor, August 2001 Acknowledgements The International Phonetic Association has kindly given permission for the reproduction of the (PA chart and the Intemational Phonetic Alphabet. For further information, please write to: The Intemational Phonetic Association, elo Department of Linguistics, University of Vitoria, Victoria, British Columbia, Canada. Permission to reproduce figs. 1, 19, 21, 22, 23, 24, 25, and 35, which originally appeared in Futdamental Problems in Phonetics, has kindly been granted by Edinburgh University Press,i Contents Listof Figures 1 Introduction 1, Theuses of phonetics 2, The phases of speach _ 3, Thevocal tract, 4, Thefunctional components of speech 2 Basie Components of Speech Initiation and articulation 1 2, Fricative and stop articulation 3. Pulmonic pressure and suction initiation 4, Glottatc initiation 5, Velaricinitation 6, Review of initiation types 7, initiator velecity, and initiator power (stress) Phonation: A Third Basic Component \oiceless and voiced frcatives 5. Volced ard wuivehas 101 . Voiced and voiceless vowels °° , 2 3 4, The three components Iritiation, articulation, and phonation 5. The production of voiced stops 6. Voiced implosives 7, Principal phonation types 2, Relationship of phonation to initiation 8. Aspiration 4 Articulation: Stricture Types. 1, Articulation: The firal shaping’ of sounds 2, Maintainable stricture types: Stop, frcative, approximant, resonant, til u u 16 18 2 26 2» ar 38 35 28 40 0 a a 0 3 3s 39 98 0x coNTENtS 3. Essentially momentary stricture iypes: Tap, flap, semnivowel 4, The transverse dimension: Median and lateral articulation 5 Articulation: Locatins 1. The nasal area 2, The ord area: Upper and lower aticulators 3. Lobial eticulations: Bilabial and ibiodentat 4. Dentaheolarartculations 5, Retroflex and palatal articulations 6. Velaeard uvular articulations 7. Pharyegal and glotta articulations 6 Conarticulation and Sequences 1. Co-ordinate, or double, articulation 2. Primaryand secondary articulation 3. Homorganie sound sequences: Cominotes and afficates 4, Lateral plosion, nasal plosion, and prenasalzed stops 5, Heterorganic sequences and contiguous sequences, 6. Dighthengs 7. Close and open transition 7 Vowels: Invoduction 1. Vowels end consonanis: importance of silent study of vowels, 2, Lip- andtonque-positions for vowels 3. Invoduction to Cardinal Vowels 8 The Cardinal Vowels (CVs) 1. General characteristies of CVs: Errars to avoid _ The front Cvs The back Cvs |. Types oflip-rounding Seconday CVs 5. Central vowels and other additions to the Cvs ‘The acoustics of vowels: Vowel formants Making the formants aucible [Additional vowel modifications a B 2 % » 8 8s 2 95 % 100 ol 106 108 09 133 13 oT 0 3. ie us 152 186 161 CONTENTS xi 9 Prosodic Features 1. Intiatory prosadies: Initiator power (~ stress) 2. The sylable 3. The foot 4, Phonatory prosedies: Pitch variation (~ intonation and tone) S. Aniculatory prosody: Duration of aniculation (= length) 10 Sound-systems of Languages 1 1, Phonology: The utilization of speech-scunes 2. Acontinuum of vowel-sounds 3. The vowel continuum diferently dissected by English ded Spanish . Voice-onset-times differently exploited by cilferent languages Phonemes 5. Distinctive features . English consonant phonemes . Allophones| Consonant clusters ‘The Engish vowels 11 Review For Further Reading References Index 163 163 74 7 a ns 0 181 ws 189 91 we 195 we 2 a7 za 2ai ust or rguRes: xii a 31. The vowel limit a List of Figures $ 38, Tongue-congurations for Cardinal Vowels 18 BD. Eerorto avoid in pronouncing front CVs 130 ! : ‘9. The Cardinal Yowe's 31 i 41, Schematic representation of lip-positions of Cardinal Vowels 140 1 Tevet : a eee ¢ A 2. Basie components of speech-sound production B Bi Addeealowes ie - 2 ee ee eee Me 44, First and second formants of Cardinal Vowels 154 4, Experiment 4: Remove articulation a 45, Cardinal Vowels with frequencies of Fl and F2 (Hz) : 5. Experiment 5: Add initiation Is Seca iss 6 Epes Aus econ i Pe idm ipa ts is 7. Experiment 7: Superimpose stop on air-sream 18 ee caw and wate ie 8 Pressure and suction varieties of [1] 2 Ee rece or wilioem gina g i 9. Pressure and suction varieties of 5} 2 Suse goupe of fe Ay i 10, ‘Three stage inthe production of glotalic pressure] 2 Ee cain is 11: Experiment 17: Three phases ia the production of a velanic ‘Thea vowel coatinuum: English ané Spanish divisions 181 suction stop (tae click (1) + 2 The VOT congneum 12 Experiment 23: Voicelaas and voiced frcatives, ” deietcon wnat unt Snatd te Aerodynamics ofa volved stop mn Voicetess and voiced stop and fricatve" “6 Mechanism of voiced glottal suction stop (voiced implosive) 7 16, Tnitation of voiced plosivesand voiced implosives 2 States ofthe gots, 32 3. Aspiration and voicing of tors 55 | Major stricture types 6 Major stricture types in the transverse dimension n The major articulatory areas ‘ Some features ofthe cral cavity Upper articulatory locations Subdivisions of the tongue j Lower artculstory locations £ 5. Two extreme types of alveolar ridge Apiso-dentalveolar stop articulations Lamino-dentalveolar stop articulations . Apro-dentaleoiartreatve acculations 1. Lamino-dentalveolar fricative articulations @ l Retrolex articulations: (a) stop (b) Bap : fl Docso-palatal articulation a . Some pestalveolar and (pre)palatal articulations Timing selations in co-aticulation 103 ii Named homorganis sequences wr i ‘The Intemational Phonetic Alphabet (revised to 1999, c™ corrected 1996) us \ W1 Introduction 1. The uses of phonetics Phonetics is the systematic study of human speech-sounds, It provides means of describing and classifying virtually all the sounds that ean be Produced by human yocal tacts, How this is done is the principal subject ‘matte of this book, But before we begin to investigate the sounds of specch it may be useful to say something about why it is interesting and useful to €o this: in other words, to review some of the uses of phonetics, Over a century ago, the great English philologist linguist, and phoneti- cian, Henry Sweet (who, as Shaw tells us, was in part the prototype of Profestor Higsins in Pygmalion—perhaps better known nowadays as the musical, My Fair Lory) described phonetics as... che indizpensable Joun- dation of all study of language—whether thar study be purely theoretical, ‘or practical as well .. ’ (Sweet (1877), . v). This is as true today as it was in the time of Sweet, Any person wao works with language would do well 6 have a basic knowledge of phonetios, The teacher of languages for example, including the wacher of English as a second language, must be able to diagnose the pronunciation errors made by students, and to devise mears of correcting them—this is impossible without both theoretical and practical knowledge of phoneticx mes is alu usefel wy Wiese Concerned with various aspects of te mother tongue: the phonetically trained teacher of reading will have a bet- ter understanding of orthographic problems and the relationship of spell- ing to the spoken language in the teaching of speeeh- production phonetios is obviously essentia!—sctors, particularly these who wish to master ‘numerous dialects and foreign acvents, certainly ought to have a thorough knowledge of phonetics, which, alas they usually lack. ‘Speech pathologists have aa obvious need for phoveties, whieh they tae2 mraopueTion readily acknowledge, both for a general understanding of how the vocal apparatus works and for the diagnosis and treatment of minor articulatory ~ defects ‘Communication and computer engineers and other ‘speech-scientist” ‘working on the improvement of speech transmission systems, on speech synthesis, and on automatic speech recognition, also need to have @ considerable knowledge of phonetics ‘Another important application of phonetics is to what Sweet cals ‘scien- Life philology’—or. what we would now call ‘comparativeshistorical line guistcs. In his words: “Without a knowledge of the laws of sound-changs, scientific philology... is impossible, and without phonetics their stucy degenerates into a mere mechanical enumeration of letter changes’ (9.1). ‘And of courte phonetics is absolutely essential to the student of line guistcs Ieis virwally impossibie to do serious workin linguists without a thorough knowledge of phonetics. Clearly, without phonetics, field-work, the most important source of linguistic data, is impossible, and phono- logical rules become (like the sound-lavs referred to above) meaningless and unmotivated rules of letter-substirution. Even in the suey of syntax and morphology questions of phonetics frequently arise: Now itis perfectly possible to acquire a good theoretical knowledge of phonetics by reading, and even more so by working ia a phonetics laboratory where aspects of the physiology and acoustics of speech are investigated instrumentally, But the kind of superficial, purely intelectual, knowledge of phonetics that is acquired in this way is quite inadequate asa basis for carying out many of the activities referred to above. What the) competent phonetican must acquire 1s 2 deep, internally experienced, | awareness of what is going on within the vocal tract—an ability to analyse, | and hence describe and ulimately control, the postures and movements of corgaas that produce the sounds of spesch. Its fairly obvious that tis kind | of practical ability is essential for those, like language learners and teachers, or actors who have to identify and produce exotic or ‘unaceustomed counds. What is not s0 ebvious, bur is undoubtedly the case, Js that the acquisition of these “practical” skils is by far the best way of acquiring a deep understaiding of phonetic theory—of the principles underlying the description and classification of the sounds of speech—and is consequently of the greatest importance also for more ‘theoretical’ uses, of phonetics. Thus, the nature of a historical sound-change can usually best be under mxTRODUCTION 3 stood by those who cen actually carry through the change in their own ‘vocal tracts and internally, introupentively, experience its mechanism: and the interpretation of physiological or acoustic instrumental records of speech is most eficienlly carried out by specch-scientists who possess the same kinds of skillinvestigators who cannot themselves pronounce, and internally experience, most of the pheaomena they are investigating sometimes misinterpret their data. Its because of the great importance of this kind of introspective aware ‘ness of the phenomena of speech that in this book we introduce the reader to the principles and categories of phonetic classifeation not only by meazs || of descriptions, which can produce a merely intelectual comprchension of phonetic theory, but by means of experiments waich readers are asked to | carry outin their own vocal tracts. In this way. they will acquire that deep vaderstandiag of phonctis theory which isthe indispensable stock-in-trade of the competent phonetician. 2. The phases of speech Before we begin the experimental approach to the subject, in Chapter 2, ie wall be useful to consider the nature of the speech-event, and what particular aspects of tars, oF may be, the domain of phonetics ‘When someone speaks to semeore, the sequence of evens is in outline, as fellows Ta response 0 the need to communicate about some event (which may either be in the world at large or within his own consciousness). the speaker-concepruaites the event in « panvulat way aul dieu eres chat conceptualizaton ina form laid down by the grammar of his language. The linguistically encoded utterance is exteralised and apprehended by the heaser through the agcncy of a series of events that we term the phases of speech. These phases star in the speaker, and culminate in the hearer decod- ing the’ uteraree and arriving at a conceptualzation which, assuming he is- familiar with the speaker's language; closely matches the speake:’s cconceptualization, which was the start ofthe process ‘The processes of conceptualization and coding/decoding are outside the domain: of phonetics. Tae purely phonetic part of the speech process begins, we assume, with the execution of a short-term neural programmein the central nervous system, which is triggered by che lexico-grammatical structure ofthe utterance ard determines the nature and the sequencing of4 iNtRODUCTION everything that follows. We may call this the newrolingsistie prograening pluase of the utterance. ‘Thereafter in a sequence no doubt determined during the stage of neuro linguistic programming, Specific “motor commands’ flow out through ‘motor nerves to muscles in the chest, throat, mouth, ete. As a result, these muscles contract—in whole or in part, successively or Simultaneously, more or less strone’y ‘We call this whole prozess of motor commands (the outflow of neural impulses from the central nervous system), together with the indissolubly related muscle contractions, the neuromuscular phase. ‘As aesul of the muscular contractions cecurting in this neuromuscular phase, the organs to which these muscles are attached adopt particular postures or make particular movernents—the ribcage may contract, the veal folds inthe larynx may be brought close together, the tongue adopta particular configuration, and so on. In short, the sequel to the neuro ‘muscular phase is a posturing or movement of whole organs in the vocal tract. We thes recall this the organie phase, The movements of organs during dhe organic phase aet upon the air contained within the vocal tract. They compress the air, or dilute it, and they set it meving in various ways—in rapid puffs in sudden bursts, in a turbulent stream, and 30 on. All of this smooth flow, n a rough, eddyi ‘constitutes the aerodynamie phase of speech, As the air flows through the vocal iract during the aerodynamic phase the things that happen to it set the air molecules o pececived by our sense of hearing, In other words the aerodynamic events generate sound-waves, and these constitute the acoustic phase of speech. In the acoustic phase, an airborne sound-wave radiates from the speaker's rmiouth and reaches the ear of anyone within hearing distance, including the lating im ways that ean be speaker himseit The sound-wave, impinging on the hearer’s ear-dram, sets it vibrating in step with the wave-form, and these vibrations are transmitted, by the little bones of the middle car, to the inner ear, or cochlea, where they stimulate sensory endings of the auditory aerve Neural impulses from the nerve endings travel up che auditory nerve to the brain, where they give rise 10 sensations of sound. We call this whole process of peripheral stimulation sand afferent neural transmission the meuroreceptive pase. Finally, an interpretative process occurs in which the incoming neuroreceptive signals are identified as this or that particular vocal sound INTRODUCTION. 5 or sound-sequence. This is the phase of newolinguistie identification, which we can regard as more or less the obverse of the newrolinguistie program- ming phase with which dhe phonetic event began. Taough there may ahvays be some awareness of sound in this phase, the identification as purticular specch-sounds is usually below the threshold of conssiousness, In the actual exchange of conversation, attention is dirveted more to the meaning Of wha'is said than to the sounds by which that meaning is manifested The final steps in the process—the hetrer’s decoding and ultimare e the concepzualization—are outside the domain of phonetics, just as ‘matching conceptualization and encoding in the speaker, We can now summarize the phases of speech as follows: (1) Newrotinguistc programming: the selection, sequencing, and tining of what follows. Q) Mewrormscular phase. eansmision of outbound (moter) neural {impulses and the contraction of incividual muscies ) Organic phase: postures and movements of whole ongans @) Aerodynamic phase: dilation, compression, and flow of air in and through the vooal tract, (5) Acoustic phase: propagation of sound-waves from speaker’s vocal tract. (©) Nesroreceptive phase: peripheral auditory stimulation and transmis- sion of inbound neural impulses, (D Neurolinguistic identifcation: potential or actual identification of incoming signals as specific speech-covunds In addition to all this we must take note of two other phases, or aspects, Of the speech process. These are the twe Kinds of feedback: kinacstherie feedback and auditory feedback. As the organs of spesch posture and move about in the performanes of speech, sensory nerve-eadings within the muscles and on the surfaces of the omgaas are stimulated by muscle contraction and by contact, and pressure, We may be, but often are aot, conscious of this feedback as proprioceptive sensations (feelings of muscular contraction and tensioa) and tactile sense- ‘ions. As @ genera} name for these proprieceptive and tactile sensations we use the cerm “Kinaesthesis", hence kinaesthetie feedback. ‘The second type of feedback consists of the stimulation of the speaker's peripheral hearing organs by the sound-wave issuing from his own mouthL 6 aw7ropuerion “which reaches his ears both externally, by air conduction, and internally, by ‘bone conduction, This is auditory feedback. ‘These feedback systems monitor and control speech by inserting nto the ‘motor system information concerning the continting muscular. organic, aerodynamic, and acoustic events Much of phonetic training involves ‘making these feedbacks, especially kinaesthetic or proprioceptive fesdback, ‘conscious. Analysis, and conscious contro, ofthe activities of speech must be based upon awareness of waat the vocal organs are doing and this, aaceacss it derived from the feedback systems, ‘Of tae seven phases of speech described above only thee lend themselves convenieatly to categorization for general phonetic purposes: these ar the ‘organic phase, the aerodynamic phase, and the acoustic phase. Tradition- ally, phonetic classification has teen based on the organic phase. This was the basis of classification of the earliest phoncticians—the Indian gram- rarians of 2,500 years ago—and also of the ancient Greek and Roman grammarians, the medieval Arab grammarians, and the English phoneti- cians from Elizabethan tinied onwards The seoustié phase has ouly been fuly-accessible since the development iit twentieth century of electtoaic devices for acoustic analysis, such 2s the cathiod=-ray oscilloscope and the sound spectrograph in the 1930s and 194(s. Nevertheless, since such instruments became wailable an cnormous amount has teca learned about the acoustic phase, aad the study of this phase of speech is known as acoustic phonetics. The aeroéyfiamic phase is also accessible to insttu- ‘mental iavestigation and aerodynamic data have been used sisce the nineteenth century, chiefly as a means of acquiring information about the preceding, organic, ohase: ty looking at variations in the fate of airflow out of the mouth, measuring intra-oral pressure: and so on, one'can make ‘many useful inferences about the organic activities that give rise to these acrodynamic effects It is only recently that the suggestion has been made that there should be a more or less independent aerodynamic phonetics, parallel to acoustic phoneties General phonetic taxonomy, however-—that is, the general or basic clas- sification of speech sounds—is still based on the organic phase, with some contributions from aerodynamis and acoustic phonetics where helpful, “This type of phonetics is often called crcieulatory phonetice,a term which is somewhat inaccurate, since, as we shall see, articulation is only one (though a very important one) of the componeats of speech sound production. INTRODUCTION. 7 3. The vocal tract All the sounds of speech are produced in the vocal tract. For the purpose of thie book we take the vocal tract to consist of the entire respiratory tract, rom lungs to nose, plus the mouth. This is a somewhat wide: application of the term than is found in most other works, where ‘vocal tract’ means only the tract.from the larynx up through the mouth and nose. It is. howeves, more useful for phonetic purposes to use the termi in the wide sense of all those tracts within the suman body that normally participate in the production of vocal sounds. ‘The student of practical phonetics does not require a very detailed know- ledge of the vocal tract aad vocal organs, and jn this book we stall intro- duce suct detail a is necessary asand when it is oteded. However, itmay be useful in this introductory chapter to outline the structur and phonetic fonctions ofthe principal orzans in the vocal tract. Figure I andthe follow ing description are primarily for refercnce, and at this stage need not be deeply studied, but should be apprehended in a general, impressionistic, ‘As we have seen, the function of the organic phase of speech ir to create certain acrodysamic conditions—to set the air in the vocal tract ia motion, and to control the fow of air in ways that ultimately generate sounds. The veal tract can thus beregarted as a pneumatic device—a device consisting ofa bellows and various tubes ard valves and chambers whose function ig toset air in motion ant: to control is flow. +. Figure’ is-2!sketch of this ‘pneumatic device’ alongside a somewhat more naturalistic sketch of the Vocal tract with lines connecting the two to show the relationships betwoen their parts The brief account of the vocal tract that follows should be read in close conjurction with a study of the figure. The bellows (ungs).can expand to draw in half‘ gallon or so of ir, andjean contract t0 blow out like quantity: io spesch they contract quite slawiy. There are two tubes leading from the bellows (the bronchi) which ‘unite in a larger tube (the trachea, or windpipe). ‘Near the upperend of the windpipe there is a piston (the larynx) that can slide up and down for an inch or 80, You can feel the front of the laryn- pistom as a projection’in the front of your acck (the“Adam’s appie’)-and, you can also fec! that it can slide up and dowa—this is specially noticeable when you swallow. The larynx is usually more prominent in mea then in8 irkopueTioy [oes . = Lechambee 8 Satie anireanig 2C _SCieate rem, HES ectube aN. Fig. 1 The vocal tect ‘womer but the swallowing movement can casily be felt by both. Within the piston there is a valve (the glottis—that is, the space between the vocal cords, or weal folds, as we shall call them). The glottisevalve can be tightly sloced ce opened to warying Segecr or olce rapsily and chy thimicallyopesed and shut in the course of spexch, ‘Above the larynx theve are three chambers, A (pharynx), 3 (oral cavity, ie. mouth), and C (nasal cavity), whied can be put into communication ‘with each other, or separated ofl from each other by the valves v(velum, or soft palate) and ¢ (tongue). The tongue-valve is highly mobile and cain control aitllow through chamber & (mouth) at a numberof different places and ia 2 oumber of different ways. Fivally, the outer end of chamber B {mouths provided witha double valve, numely the upper and lower lip, Study of this brief description of the ‘paeumatic device’ and the corres- ponding sketch‘of the vocal tract in Fig. I should make clear the main parts ofthe vocal trast and their major phonetic Functions. [As we said earlier, phonetic taxonomy is primarily based on the organic mvtRoDuETION 9 phase; but it requires contributions frem the aerodynamic phase. This i} inevitable, since the production of spesct-sounds is an acrodynamie pro- | cess. The organic postures and movements do not themselves generate | sounds; they merely set the aia the vou tract in motion, ard itis the flow of air through the vocal tract that generates sounds initiation on ticuletion 4, The func nal components of speech From the organie-aerodynamic point of view the produetion of specch- sounds involves two estential fanetionc! components: (\) some method of initiating a Row of air in and through the weal tract—that is, some form of initiation, a5 we calli; and (2) some method of shaping or articulating the sirestream so as to generate a specific type of sound—that is, some form of articulation. There is a third functional component of speech-prodvection, erbresent in most, but n0% all, sounds, and consisting of certain types of / modulation of the sirstream as it passes through the larymx—that is phonation. In order to understand the mechanism of speech itis essential to hhave a clear conception of these three furctional comaponents—of the forms (hat they can take, and of how they interact to produce speech-sounds, The following chapters introduce the reader to these three basic Func- tional components of speech, and t0 ine various parameters, of ranges of variation, of the sub-componeats or features that characterize cach of them, Because of its concem with these two aspects of phonetics, the phonetic theory expouaded in this beok may be called componential parametric phoneties, ‘These compouentsand parameters and other features of speech produe- tion are introduced in a series of experiments to be caftied out in the reader's own vocal tract, These experi personally experienced insight iato phoactic theory, but will also develop the motor and auditory skills that are esential for the competent phoneti- cian, and which cannot be developed by the mere reading of descriptions and explanations is will not only provide a deep, Phonettion (09 met Saltsii i 2 Basic Components of Speech 1, Initiation and articulation The fst stop in the study of phonetic i to discover experimentally, the basie components that go into the production of aay sfeech-sound. ‘We begin by’ producing a prolonged [f]-sound—the frst sound in the words ‘first’, four’, and last sound in ‘huff’, ‘puff’, and so on. While making the ‘oind ond sust thiak about what is going on—that is, one must analyse the production of the'sound introspectively, by taking vote of the Kinaesthetic sensations, the feelings of contact and movement associated vith 1 Bevin by comfortably filling the lungs—that is take ina deep, but not exces le sively deep, breath—then start up a prolonged (f-type sound: [F ff ff. On one full intake of breath you will no doubs easily be able to keep the sound going for 5 co 10 secouds. Repeat this three or four dimes—breathe i, say ‘prolonged (FFF FF 1 then relax—concentrating your attention upon waat is “happening, upon what you are doing in order to produce that sound. Thwill quickly become clear from introspective observation that the produc tion of this prolonged {f-cound requires the simultaneous cocurtence of two events; in other words, the production of the sound has fo com ponents. These components ace (i) deflation of the lungs, which, sends a stream of ait up the windpipe and out through the mouth, and (if) contect ‘between the lower lip and the upper teeth, forming an obstacle to the lowing ‘ir, which forces its way through in a turbulent stream that generates the speci hissing sound of f). 22° Now carry ovt the same experiment with 2 prolonged [s}—ie first sovad in i" see the ast sound in Ass. Breathe in, startup a prolonged [s}4ype sound fs $ 2 b's 583 lsting for Sto 10 seconés or $0, Repeat several times—breatbe is, ay12 Basic components oF seeecn [bs $5 5}, eeiax—coneentrating your at doing to produce that sound, jon all the time upon what you Once again, it will be clear,that there are (wo simultaneous components (9 detition of the lungs, initiating an upward and outward flow of air, end G2) an obstacte to the Now, past which the air forces iself in a turbulent stream that genecates the specific i Any speech-sound that we care 10 examine will always exhibit these two components in some form or other namely (9 An activity (lung-defation in the examples just studied) that intintes a flow of air, and is consequently called initiation (or ‘ait-siream sing sound of [} mechanism’) —the organ(s) usec! for this purpose being called an initiator (de lungs in the present case), (@ Anactivity that modulates or articulures the sir-stream, thus generat- ing a spevific type of sound, The process itself is called articulation and the organs utilized io articulation are artictlators Por [f] the articwation was formed by the juxtaposition of the lower lip and the upper teeth, which thus were the articulators. For (s] the articula- med a littl Further bacl tongue, just behind its Gp, und the sidge that you can fee! behing the upper tion wa by the juxtaposition ofa part of the teeth Now repeat Experiments 1 and ? saying both [FFF {aad [s ss $3) vga, several times, snd consciously inirospectiag about what the evo basic components initiation and articulation, fea! ke Figure 2 isa simplified diagram of the vocal tract, showing the locations of the two basic components of speecli-production for the two sounds we have studied sa far, (and {6}. For many other types af sound both articulation ‘and initiation would be located elsewhere, as we shall see. ‘The next two experiments demonstrate that itis essential for both basic components to be present in order to produce a spetific sound. In thesé experiments we first remove the initiation (i. stop the air-steeaan) but keep the articulation intact (Experiment 3) then remove the arieulation, keeping, the initistory air-stream intact (Experiment 4). Figures 3 and 4 show these processes diagrammatically. The double line labelled ‘articulation’ sepresenss the juxtaposed articulating organs being held together through time. In Fig. 4 the removal of the articulation is indicated by the separation of the juxtaposed lines. representing BASIC COMPONENTS OF SPEECH 13 ). 2 Basic components of speceb-sound production acticultion, disintegriting into broken line: when the articulation has thus been abolsten 1 lowes line labeled “initiation” symbolizes deflation ofthe lungs and resultant inflow, and in Fig, 3 the cessation of initiation is indicated by a short vertical line ‘The irregular scraichy-locking line lubelled ‘sound represents the hiss noise of f] or (6): itis, in fact, taken from an actual instrumental recording, Now continue with Experiments 3 and 4, 3 Removing initiation, Breathe in, and start saying [PCE 2] asin Exp 1, only this de, while carefully maintaining we lip-teeth articulstory contact, sound _phentephihenarht aiewation Agsasssa » Fig. 3 Experiment 3: Remove initiation1 BASIC COMPONENTS OF SPEECH Bere hn OTC TF established, suddenty puling the ip away from the teth, taking care to keep say nop beating oo-.e woh testo estan Fe ines ‘ the air-stream going as long as possible. 3) Take eaze to reiain the articulation fo at east 5 sevonds afer you have abolished the initiation. Now cepeat the experiment with aa (s-articuition Breathe in, start up fe 16s 53} then, while carefully maintaining the (5 ‘ardgalation, suddenly abotish the initiation ie. stop breathing out, Retain the [spartcwlation for another 5 seconds or so after you have abotshed the initiation: Both of these experiments, 2 and 4, demonstrate that the presence of both initiation and articulation is essential forthe production of the specific sounds of [Pf FF and [ss s ¢s 9. In Experiment 3, the abolition of the {nidation (complete cessation of airflow) simply resulted in silence—sileat maintenance of the articulatory pesture. In 4, the abolition of articulation resulted, not in absolute silence, but in the immediate loss of the specific sound of (ff ff Ff] or[s ss], For the shor period of time during which sirfow continued, the unmodified sound of exhalation, the non-specific hushing sound of exhaled breath, could be heard,’ We can represent this non-specific exhalation sound by [tJ thus the process of holding ané then abolishing the articulation of s] may be transcribed {ss 85 b h In both casesitis quite clear that the moment the iitiatory ar-stream stops } the sound of (fof [s] ceases: there is only sleace—silently held erticvlation. ‘This experiment demonstrates that Doth of the components infsasion ana artiouation, must be present to preducea sound like (f] or{s}. Remove I the initiation, ané, even though the articuladon is still maintained, the evult is silence. Incidentally, this is « first example of a proveduse of very grvat ue in the study of phonetic: the formation of silent ertinwation. AS twe shall see repeatedly in what follows, the production of ‘sient sounds that i, the formation of Silent articulation with introspective analysis of the kkinaesthetic sensations, is by far the most powerful way of learning about articulation, and we shall use it frequently “The next experiment, 4, demonstrates a second way of breaking downa sound into its components, namely, by abolishing aricalation. “The last two experiments demonstrated the necessity of toth initiation and articulation by-méars of what may de called a proses of analysis. resolving the production ofa sound into its two component paris. “The nent two experiments, § an 6, involve the opposite process, namely pmthesis, Wecan synthesize a sound by starting with onc component, ané thea adding the second one, and it will be clear that no specific sound is produced ntl we have both components in place, 5 Adding initiation, Synthesizz a f}-type sound by starting with the articulation (which, by itself, produces no sound) and then adding initiation fat which time the specific sound of [becomes audible). Bresthe in, form, and bold, the artewation—that is put che lover lip firmly against the upper teeth— then, after slently helding that articulation for a few seconds, stat up the “4 Removing erticulation. Breathe in and thea sart up s prolonged (FFF EA tu. this time, suddenly abolish the articulation while taking care o keep the Jnitatory airflow going as long as you can. This will probably be only = very few seconds, because, when the retarding! effect, of the articulation is removed, te yevahual initiatory ais Sawe ent very rapid, Do the same with [855 9- Gee Fis. 4) In che eae of[f ££], "abolishing the arealation’ means removing the , rower ip ro the upger teeth This car be done forebly. by aking fe over ‘ad fo tip bereen Bager and thumb and then whea the sound of (FFF) wel gots eeeereres Fa ee Fig. $ Experiment SAdd initiation " This statement wilds fe the prevent but ar we shal eater sity speaking thi set the ‘eanatifed 1omd of extlation, but more aecuntey, exalaon modulated by pastre throush the inva, whee te rbuen! fw of sie generates lryngeal or got, eine sound (See Chap. 5) Fig. 4 Experiment 4: Remove articulation16 MASiC COMPONENTS OF SPEECH Initiation, ie. release the breath ad let the intiatory aivsteeam Bowe pas the articulation generating the wsusl specific having sound of {PCF 1, Do the same experiment with (3) Breathe in, bolt breaw, form, and silently hold, the [sjarteulation, release the breath and note how the addition of the initiation generates the specific sound of [¢8 $s] {In Experiment 6 we start withthe initiation, and then, after the intiatory aie-siream is going, we add the articulation, 6 Adding aricuetion. Breathe in. Start up the initiation, i, begin to exhale, and then, befor You have lst too shuch air, sudden place the lowe li firmly against the upper teeth: the sound of exhalation i immodiotchy replaced by the specie sound off]. thus: hb h EF, (See Big. 6) Repeat the same experiment with): hi $8 88 SU a as Fig. 6 Experiment 6 Add acticuttion These initial experiments, 1-6, were designed to demoxstiate that the Production of speech-sounds requires the simultaneous presence of two base componests: initiation, or the production of a flow of ait, and articue ation, oF the formation of some kind of obstacie that modulates the airflow to generate « specific type of sound So far, we have experimented with only two sounds {f] and [s); but the findings ofthese eaperiments can, infact, be generalized, All speech-souncs ‘fequire the simultaneous occurrence ofthese two components, initiation (or air-stream mechanism) and articulation. 2, Fricative and stop articulation ‘There are different varieties of both components: an enormous aumber of cilferent varieties of articulation-and several different varieties of initi- ation: and we will make a detailed study of these in duc course. $o fat, SIC COMPONENTS OF SPEECH 17 however, we have discovered only one variety of initution (iavolving the exhalation of air from: the lugs) and two examples of articulation, [0] and Is}, Both of these articulations involve turbulent airflow through a narrow chanael, resulting in @ hissing noise, This type of articulation is eae ‘frieutive, The sounds represented by the symbols (i) and {5} are both Fricatives, but they are articulated at different places. As we shall soz Inter, ariieulations can be described and classed ia terms of their mannec of ardculation, o type of stricture (in this ease “ircative), and thei place of aricvlation within the vocal tract. Another manner of articilation— another artiuletory complete momentary blockage of the air-sirsam, with the conscquence that pressure builds up behind the closure as the aie tries (0 continue owing, When the blockage is resnoved, the somewhat compressed initiatory air bursts out in a small explosion. This type of articulation is known as stp, and examples of stop sounds, or stops, are the sounds of (p, [6 [k], repre seoted by the initial letters of per, tar, cor. The final sounds of lbp, lot, bce are ako stops, butin the pronunciation of many Enalish soeakers, particu ture type—involves complete closure, and hence larly in America, they differ quite noticeably trom the inital stops. The Mifference is that, while the initial stops of par, rar, sar are explosively released, the final stops of “op, fot, lock may {in some persons" speech) not be noisily or immediately released at al. We now experiment with the production of stops, aad, in order to yet stUre, we start by superimposing them upon an sn insight into their airstream, 7 First, silently make and release the stops [p}[, ] several times, and chen proceed to the following experiment, Breathe ie, ad then start up the ieiti- ation; that is, begin to exhale dirough the open moun, Very soon after the ‘exhalation has started, and while trying to keep theexnaledaicstream going, close the ips and suddenly open them again after a very brie? period of closure, and continue to exhale until the air in your ings is exhausted, ‘Throughout this operation you must strive to keep the iniating airstream _g0ing all he time, Think of it 2s a continuous streum of initiaiory air, with Stop arenstinn (| mamentanysuprnpel o9 i thei done peop erty, the eet should bea sight peesurebuldap behind the ctescd lps, so that when they ve suddenly opened again a moment Iter thee & an explo sive relesceof the high-pressure ir tht had been pent up behind thestop. We, sent theprocessas[h hh pH hb where the capital H presets the, ‘momentary explosive release of air Fig. 7 indicates whats going on.| [ SS | 18. BASIC COMPONENTS OF SPEECH BASIC COMPONENTS OF SPEECH 19 pressure buidup behind sop Fig, 7. Experiment 7: Superimpose scop oa airstream Now carry out the same experiment with the stops (1 (tip of tongue mck og moreentary sonia withthe ridge behind the upper teth), and (x (oak of toogve making. momentary contact with the velum or soft palato—the ~sott back partof the rdofof te moun) han tea A} DAA KA DD). In each case, strive to keep up a continous initiatory effort ard consequent exhalation, in order to get the feeling of the stop’ articulation as being a somentary closure siperimposed on the ar-stream. Repeat the experiment Several times, and nete, of course, thit'you cal supermpose a stop articulation on the initatory airstream at any point in time, Tr, for instanee, starting the iatation with the closuce already in place. Breathe i form the stop articulation, start the intiaion then, a8 aresure ule up behind the top suddenly lest (pT BAI[E BAA] wiruay. 3. Pulmonic pressure and suction initiation We have now experimented with two types of articulati (a) ‘and stops (pl. (0. [kD. but still with only one type of initiation: namely, one that involves exhalation—a steeam of air forced out of the lungs under pressure, and propelled up the windoipe and out through the mouth. Since itmakes use of the lungs, this type of initiation iscalled pulmonic—we shall see shortly that speech-sounds can also be initiated at.other places in the ‘Yocal tract. And since it involves defation of the lungs and consequent ‘compression of the air contained in them, itis pulmonic pressure initiation. It i notat all difficult to use inkalation in place of exhalation as a form of initition. To ¢o this, we dilate the Tungs, thus lowering the pressure wwithia them and so sucking air down the wiadpipe and into the lungs: act surptisingly this type of initiation is called pulmonte suction. 5 In thisenpesiment we inate te airstream by pulmonie suction, Begin this time by empryiag the iungs. Now form the articulation for (f] and begia 0 inhale. The articulation is exact the same as fora ‘normal’ but this time "the hisine sound off generated by air owing from outside. through the articulatory chanrel ito the mouth (ond down the windpine} Repeat wih J Empty the lungs form the [s}articulaion andhold it while Ihling, draving air into me mouth and through the articulatory chante Now experiment with altereating pulmonic pressure ané pulmonic sucioa initio, There are no special phonetic. symbols for puimonis sveion sounds, so we wil present them: by mass ofan arrow pointing downward nxt to the regular symbol. Thus pulmeuie presur {f]=(9, pulmonis suction ]= f). Breathe in, startup an ordisary pulmonic presurs [f] and when the lungs are nearly empty, rovere the aistveam so that you eed up making a pul moric etion (1) (FFF... [ft tL tL. J, Continue aerating pressure and suction (9 (9p (Carry cut the same experinent with pulmonie prssue(6] ard puimonie suction SU} {8 sb] {se (6J st] ete ‘These experiments demonstrated that more than one type of initiation can be used in the production of fricatives [tis easy to discover that the same is rue of stops. 9° Tust as you altemated (9f4) and (s (U]. now start altemating pulmonie pressure and suction stops Make pulmonic pressure [pM], then close the lips again and generate pulmonic succdon (Le. attempt to inkale while tbe lips are closed), so that on the release of the stop air rushes into the mouth [pHLy ‘Then continue to alternate: [pH (pHUI (pH) (pH [oH IDHU ete. ‘Nowrdo the same with {].and fe. i.e aerate (5) (41) (E41 IE ve return: for moment to Experiment 8 and produce the pulmonie pressuré and suction fricatives [f] [FL] and (s] (ol) we get an interesting insight into the serodynamics of fricative production. 30 As in Expl 8 alternate pulmonic pressure and suction [1] C1] (9 (FL) several times, Notice that the sound generated ty the suction [fl] is very much the same as‘that. generated by the pressure (9). Now alternate pressure20 BASIC COMPONENTS OF SPEECH ‘and suction {5} Sb] [9 {sl}... several times. Notice that the sound gener. lated by the Suction [sb] is cousiverably eifferent rom that generated by the pressure (s Experiment 1D showed thst the ciecton of dhe aiflowout of the mouth (epressive) for pressure (f] and [i], into the mouth (ingressive) for suction (FJ and (s1]—has noticeable effect on the -sound in the case Of fi), but nor in he case of [i], Indeed, while the pulmonie suction [FL] sounds fairly normal, the pulmonic suction [st] hardly sounds like an fs] at all, Why? “The reison for this diference is easly explained, with the help of Figs & aad 9. Fi § shows how the tevam, whether egressive as for pressure [f] Siren f Seb Siow Press (Girenivetow) Anges tow) Fig. 8 Pressure and suetion varieties off] (Ga) or ingressive as for suction [F4] (8b), squeezes dhrough between the teoth and lower lip and through the interstices between the teeth, bezoming turbulent and noisy as it does so. But the articulatory channel through which the air squeezes is of much the same shape whichever way the sir is Aowing—so the turbulent airflow and the resultant biss-sound are much the ‘same for each direction of flow, Lr welook now at Fig. 9 we see tht the aerodynamic conditions are quive turbileat er turtutentet “irecton ofsiow cane! direction wie Rebulence — olurfon 2 oy Presrels Sucttn|s| (Exess tow) (ngeontve omy Fig, 9 Presure and suction vaietes of 3} fo Dastc COMPONENTS OF SrEECH 21 different Inthe pulsaonie pressure [3] (Fig. 9a) the exressive airstream frst squeezes through a narrow channel between the tongue and the ridge bbchind the upper teeth (alveolar ridge). As it passes through this channel it is accelerated, becomes turbulent (channel turbulence’), and issues from the channel in a high-relocity curbulent jet. This jet strikes the wech, and this imparts further turbulence to it: beyond the teeth, Yéownstream’ from the teeth, there isa turbulent wake, fullof swirling movements or ‘vortices’ ‘The turbulence of the wake (‘wake turbulence’) sontsibutes an éditional rather high-frequency component to the hiss-noise of [) In the pulmonic suction [sL] (Fig. 95) however, the ingressive si coming from outside of the mouth fows rather slowly and non-turbulently past the teeth and inwards through the 3 Once again itis accelerated as it passes through the channel, and emerges as « turbulent high-velocity jet. But this inflowing, ingressive, et, as it shoots ico the interior of the mouth, meets with no obstacle comparable to the stream ow [s}type articulatory channdl. teeth, and so acquires no additional wake curbuleree. Because of the absence of wake turbulence the ingressive, pulmonic suction, (sl), sounds noticeably different from the egressive, pulmonic pressure. [s). (See Fis. 9a,b) Experiment 10 and Figs. § and 9 provide a simple and easily observed example of the importance of what we called in Chapter 1 the “aerodynamie phase" in the production of speech, We have now discovered twe different forms of initiation. Both involve the use ofthe lungs and are thus called pelmoaic—puimonic pressure and pulmonic suction, But the lungs are not the only organs that can act 3s initiators. The pressure-change and consequent airflow ncoessary for speech be inated by other organs than the kings One of the most important other initiatory organs is the larynx. The laryne, oF voice-bow’, willbe referred to again in thene«t chapter, where we study phonation, Hers itwill suffice to point out that the laryax is located ia the throat--the front of it can be felt, uné often seen, as a forward projec- tion in the neck—the ‘Adam's appie’. Within the larynx are lorated the vvozal cords of, more descripiively, the vocal folds-twe shelves of cartilage and muscle that run from back to front, and which can be pat into various positions. For the moment we need only note thet’ the vocal folds can De.quite widely separated. a5 in breathing. and as in the various sounds wwe have experimented with so far. They may also be tightly closed. Now, the space, of pastage-way between the vocal folds is called the glottis22 BASIC COMPONENTS OF SPEECH ‘Therefore, another way of putting what has just been said is to say that the slettis can be wide open, asin bresthing, or it can be tightly closed, ‘We close the glottis tightly when we "hold the breath’ or when we cough. Tna cough, we build upa strong pulmonic pressure below the tightly closed glottis, then release the closure explosively If you cough, rather gently, Wo lor three times, introspesting about what you are doing, it will be clear that a cough is a strong stop-type articulation—what is called a gloctal stop (phonetic symbol [2). The glottis is functioning as articulator, just as the lips do in fp] 4. Glottalic initiation ‘Wo must, however, aow tura to the larynx met as the locus of a. ype of amicalation, but a an initiator. Ifthe giousis dghtly closed, and there at the same time a closure in the mouth, say between the back ofthe tongue and the soft palate (a [k}-closur), 2 small quantity of air will be trapped between the closed glotis and the oral closure. If, now, the laryaxis slightly raised, the air trapped between the closed glotis and the oral closure wil be compressed. Then, if the oral clesure is suddenly released, the entrapped high presoure air wil momentarily bust forth ina short starp explosion, Here the air-compression, and eventual airflow whea the arscalatory clos- ture is released, ae initiated by the larynx. The leryax is thus the initiator, and because of the importance of the glottal closure within the larynx this type of initiation is called glotiatic: and since the larynx rises, in the initi= AUiON of thls sound, aad euupacases the ir capped above. it, thi ion exampl of alotaiie pressure initiation, Prom sonis eyo wth i005 dose | Yo]. Fig. 10 Three stages in the production of glottalic pressure (k] BASIC COMPONENTS OF SPEECH 23 i: Figure 10 shows, somewhat schematicaly, three stages in the production of a glottalic pressure stop, (k7. Nove that souns with glottalie pressure initiation are indicated by an apostrophe, thus: [k,["] [ph ete: ‘The Biststep towards learning to make glottal pressure stops is to make éerain that you can tightiy close your glottis and hold it closed for some fime If you are net sure if you can do this, develop it from a cough, asin Experiment 11. 11. Make 4 series of quiet coughs; that ic, series of glottal stors (2,7, ‘Now, make the closure for [7 but don’t release it; hold it for some time. In ‘oder to make the beginning anc end of the period of glotal closure clearly audible, begia with a momentary exhalation, fh], whichis abruptly cutoff by the dosing ofthe slottis, thus {h7], Then release the alotal closure again into ‘an exhalation, [Ph Now preduce'the whole sequence again'2 number of ‘mes, [AP] [APA] <. . progressively increasing the duration of the P}of [s}1ype arteulation rmly in place throughout, and keep ‘he pulmonic pressure initiation going stexdily chroushout, but ‘Switch on” voice, then switeh it off it. and 0.08, Ta doing thi, vigor ously, but dliberately and introspectivey, you will get the feeling hat phone ‘tion (rotce and woevtessness) constitutes a ter, independently controllable component of sound-production. It is essential to sequite precise constious control of this component—the ability to add or subsract vote t0 or from virwally any sound ‘What you are doing when you alternate between voiselass and voiced phorstion is changing the ser of the glottis, The glottis, you will reall from Chapter 2, is dhe space between the voeal folds, In producing voiceless friatives like [f] or [3] the glottis is wide open, so that the initiatory si-stream can flow through freely: when you ‘switch on voice’, you bring the vocal folds together. narrewing the glotis so shat the airstream sets the vocal folds rhythmically vibrating rapidly opening and shutting and thus letting the aie go through in a series of regularly recurting litle pufls that wehearas the buzzing sound, or cone, of voice. Fig. 12 represents what was huppeningsin Experienent 23—which i wood be well to repeat while look ing atthe Ggure. Fig 12 is similar to easier figuces (Pigs. 3-6) in that it represents initiation, articulation, and the resultant sound—but it also includes phonation—indicating the open and vibrating states ofthe glotis, “The noiry hise-sound ofthe voiceless Frcatve is represented by the ines var scratchy line (as in Figs. 3-6) while the buzz or tone geverated by the vibrating vocal folds is represented by a regularly recurring (‘serioie’ is he technical term) wavesform, with the irregular Capeciodic’) fricative hist-avise supsrimposed wpon it per'edic seen 7 vonfoide wea fos phonation otis pen <>< ODO {water ! 17 ibating| | Fig. 12 Experiment 23: Voiceless and vaived frieatives Both cepresentations of the sound—the scrately irregular hiss and the more regular wave with frivative hiss superimposed on it—are taken from instrumental recordings of the sound of actus! voicele fricaives, ‘To test your cotitrol of phonation—your control of the ability to produce voiceless or voiced sounds at willexperiment with various otter articulations, 24 solute the oh sound atthe beginning o end of sacl words shop and push ‘Tue phonetic symbol for this voiceless feieative is {f]-Say a prolonged [ff Jang, while igily maintaining the articulation and keeping te iniiiaion going, suddenly switch on voice. [n other words say {f [53 23} the long (3) being the symbol for the voived counterpart off38_BHONATION: A THIRD sast¢ COMPONENT Now alternate, canging nothing but the phonation: [1331335 ---] Next produce a voiceless fricative by bringing the back of your tongue lose to the veluen (the soft palte), at about che place of articulation of kJ, Dut without actual closure. Thisis the sound of ck in Scots Lach, German ach (phonetic symbo! (x). If you are net familiar vith these words the strong noisy voiceless fticaive at the end of the expletive of disgust yeech! will do forthe present ‘Now say a prolonged (x}—aed, while rigidly maintaining the arieulacory posture of the tongue, suddenly switch on voice (the symbol for the voioed version of x] is aD: [xx xxx YY YY. i Now alterate [xx yy xx ¥X3Y¥.. J Remimber-in going fom (x]:0 {J and back again, make absolutely no changes exceptin the phosation, In Experiment 24 we derived voiced fricatives from voiceless ones. In 25 ‘we discover some new or unfamiliar sounds by devicing—teplacing voice by voicelessness. The voiceless sounds we are about to discover have 10 specific symbols: they are represeated by the symbols for the voiced sounds witha small circle beneath then: [,], reminiscent of the open glottis, 2. Voiced and voiceless mal r 25. Saya prolonged [m] like te intial and final sounds of muon: [m mam... ‘Notice that in the articulation of [ml] there is complete closure atthe lips, but the seft palateis lowered so that airs flowing out through the nose, Note also that a cormal [m} is voiced, Youcan detect this very clealy by closing your ears—as m Exp. £é—waule saying mmm... J. Whale carefully main ‘ining the articulation for [a (the closed lips and lowered soft palate) and. carefuly and vigorously maintaining the initiation. suddenly ‘switch off voice'—ise opea the glocts:(m mon mm qj). When they retry this expeciment some people simply stop the initiation (i. sop breathing) when they change ‘rom voiced (m) to voiceless [-Itis important net to do this.You must keep the inkiator vigorously going throughout, so that when you.switch.from voiced {ml to voiceless fm] you can hear the sirttrsam Sowing noisy, but voiceless, out of the nose fora second of two: {om m mp ‘Try the same experiment witha: fp 2g 9.9) Voiccless nasal consonants like (ip]and{g|donotregularly occurin Eagiist, ‘though they are perfectly aormal sounds in some other languages, such as Burmese. There are, however, some exclamations or occasional utterances PHONATION: A THIRD BASIC COMPONENT 39 containing (m] used by English speakers. One of these is che exclamation (muni), with strongstress (marked by preceding’'|on the lastpart—x kindof indignant hm, Anotheristhe sequence | map], stressed oa the firs or cond (inl. sometimes written as mip/im, meaning ‘yes 26 Now experiment with voicing (l} and [sounds Begin by isolating the Gof Lee o* leaf Incidentally. we have up to now taken it for granted that the ‘reader can isolate a sound from a word in whichit is normally produce, but itmay de helpful to say something about this, Toisolats the [of these words, say the word normally several times; then several mare times long/hening the ‘nikal sound, but aking careto keep it atherwise exectly dhe sare as t was in the wore as proncuneeé normally: Next, say this somewhat lengthened [] but ‘merely nagine the test of the wore, Now say a prolonged (J, which you will clearly heae to be voiced, and, walle tigily maintaining the (aricuation and keeping the intitory airstream going suddenly switch off voice: 111] ))- Do this several times, and then pronounce voiceles [] by itsell. A voiceless Glovers in many iypes of English, after [p}and {Xin such words as playand clay Experiment with these words and you will probably observe tha ths = 50 Ifyou say the words pay and Kay you will otis that there isa alight putt Of breath, a brief moment of woicelessoesy, after the release of tbe stop before the voicing of the vowel begins: this is called aspiration and sounds with this pul of breath after them are aspirated. When an (follows the [5] (G8) the aspiration is caried over into the [), which thus becomes partially ‘oiceless {1}. Note that voiceless stops in the English of Seotland and the North of England are usvaly anaspirated: so the I] of play and elay may be Filly unined in thece varieties of Luisi, The best-known example of a language. regularly using voiceless {] is ‘Welsh. This is the pronunciation of {in Welsh, a8 im the place-name ‘Llangollen, In fact, the Welsh Wis not strictly the voiseless counterpart of English 1. Tae Welsh sound hes a narrower articulatory chennel, usually at aly one side of the tongue This unilateral Welsh [lis a ‘voiceless lateral ‘icatve’, for. which there is a special symbol (8. See ifby squeezing your fongueout sidewaysinto thecheeks, you can turn therather weak] youhave ‘been making into noisy fricative[t), and then say [Hengoten} for Llangollen 27 Experiment with one or moré Varieties of raioind—a tsiléd {4 such as is ‘sofmoniy ‘éed in Taian aod in the populac (but ercoucous)- conception of * Scettish Englishy-or an unwilled (fricative or approximant) r for whizh "we may ase the symbol [J], such as is used ia most types of English, For40. PIONATION: A THIRD asic COMPONENT convenience I will here use the symbo! [x for either or both of these Say a prolonged (cf (tr re. and note tha itis voized. Now do the same thing, but switeh off woies (open the glotts) in the miiddle—earefully maintaining Nowalternate wiced and voiceless (ff ery er re reerer [Note that in most ¢ypes of English an (¢] following aspirated [p] [0 fk] is partly woiveless; thus, pray ray, and cray begin with pr} (C4 eek 2 e te articulation and initationsfe ere ¢ rete, 3. Voiced and voiceless vowels It is not only rather obviously consonanial sounds like (), [] (dete. that ‘can be made *oived oF voiceless the same distinction also applies to novels, as Experiment 28 shows. 28 Suy the word see, then isolate and prolong the vowel (, thus(s i] iti] it ii}. Note tbat this sound is voiced. Now start up 3 prolonged fil ii... Siop the initiation—stop breathing —but deliberately hold the articulation of the {Fi]and intospect about it. You should eel that your tongues dusched up in the front of your mouth, the sides of the tongue being pressed aginst tie “upper teeth from the molars atleast as far forward asthe canines, Now that you are clear about what fi] feels Tike, start again: take a deep breath, stat up a prolonged ii Jand while taking eare to keep the ination going and rigidly keeping the ariculation in place, suddenly open the slots switch off voice, Keeping everything else eorstant: fifi} This sound [jis very similar 10 the “yoeeless palatal (iative’ [g] used in such German words as ih fi] ‘Whit are ten called ‘voiceless vowels’ (although they may actvally be ‘whispered) oceurin such languages as Cheyenne. Comanche, Malagasy and Portuguese, 4, The three components: Initiation, articulation, and phonation tis now clear shat there are theoe important components in the production of specch-sounds:initistion, articulation, and phonation, Just as we eOukL break down, ot analyse, and syntiesize sounds by removing or adding the two basic components in Experiments 3-6, so we can do similar things with | HONATION! A'FHIKD 9ASIC COMPONENT 41 these three components. Thus we can start by silendy forming an articulation—then add initiation, then add voice: or we ean set the glottis for voice then add initistion: or we can start with a fully formed voiced sound, eg, {2} thea remove first voies then initiation or first articulation ‘thea voice .. . ete. Experiment 29 carries out several such operations tuoval of articulation, keepin 29. Startby repeating Exp. 4 (the the jnitatory airstream going)—thas [Ff hb h bf [s ssshh bbl, Having reminded yourself of waat it feels Bke to abolish srtevlation while keeping initiation [going strongly, try the same experiment with a volved frizative—[v v This lime, wen you suddenly semove the articulation (I necessnry, gripping the lower lip between finger an¢ thumb and forcibly renoving ic foo dhe upper teeth) carefully keep both initiaion and phonation gomg, Whea the arteula tion is removed, but both the other comronentsare stil ative, the result will rot be the [hh bi] that resulted in Exp. 4, but a vowellike sound that we ean represent ss [2 23) thus ly vv ¥ 22 ‘Now camry out some further experments, adding and suburveting comporents in various orders, e. ieathe in, then hold the breath, silently forming an {f-arteulation: hold Ue articulation for @ fow seconds then start up pulmonic pressure inétiation—immediately the fricatve sound of [fs heard, Keep initation going, thea suddenly switch on voice (puonation), Maintain voies (y) or 1 few seconds, then remove the articulation, leaving simply voice (al. finaly, celax the voea! folds switching to simple voiceless and urarticulated brea fb). “The whole sequence may be represented as (Gilent FF) EEL wy 99h tah}, Repeat the sequence, and then try omeothers eg. [vv vi ftnh h999 vvw[bhhsaavvvEil, (Corry out ihe sameexperiments with s}and [2] and with I] and of ane (1 inplace If you cartied out the sequence in Experiment 29 carefully and intro~ spectively you should now have cca ea, tse on actual experience, of aihat it means to produce unarticulated voiceless initaton fh] to add ardevlation {fj to add voice [vf to remove articulation leaving only ination and phonation (and so on fs shor, you should have a solid \ understanding of the base components of speect-produetion-—intaton, | aricalaion, and phonation-—ani how they may be combined in oréer to ) produce verious type: of sound.42. PHONATION: A THIRD BASIC COMPONENT 5. The production of voiced stops Upto this point we have experimented with voiceless and voiced ficatives, {fl and (v4, [s] and (2, ane [x] and [y, te, and a few other voiesless and ‘iced pairs, suca 2s cn} and mi (Jane (), i) and (J. Wehave net, nowerer, experimented with voiced stops, that is, sounds like [b], [4], [g]. In order to do this we shall begin by superimposing the stops (pl. [eon a voiceless pulmonis pressure airstream. 30 (9 Saperimpoce [p] upon pulmonic pressure initiation, as in Exp. 7. Fill he Jungs, siart up she ination then, while che air is lowing out [hhh], suddenly close the lips for a moment, [p], thea release them. In‘doing this try to keep tiation going throughout—i.c keep up the pulmonic pressure without ay relaxation cight through the moment when the lips aré closed. If you do this, there will, of course, bea buildup of pressure behind the lipe during the short period of their closure, and a consequent explosive release of air when the lip-closur: is relecsed—represented here by H: thus [bh 9 Hh hh}. Fig: 7 indicates what is happening; Do the same thing with [ah a th bh) (hh kh} (i The nest step isto eee what happens when you! cay out exactly the same operation, except that you substitute wice for the vecelessness Of (9. ‘This time, breathe in, startup inidation with che glottis set for voize, produce 4 prolonged [o 2 92 3), but momentarily'superimpose lip-closureipon the ‘voiced airstream, taking care to keep the vice going throughout [> > > ata} Hae sain, here will bea momentary buildup of presire'li the moun ‘behind the fo}-closure,and an explosive release into voice when the closure is broken. The important thing isto contrive to keep the voice soing throughout: do not let it cease to sound during the period of lip-closure. Lirtening to yourself (ty somal ais-coaduction) while you repeatedly say (2222 2] you will observe that the sound of voice which is quite loud during [p 2 3, is somewhat muted during the moment when the lips are closed—but it mast not cxase during that time, ne Now say [39.6 23] several times with the ears eloced, Lieteningin this wayt bby bone- conduction, you will obser that the sound of voice suddenly gets louder during'the moment of lip-closure, If this docs not happen you are doing something wrong, probably allowing the voicing to stop during the articulatory closure of (bl. thus tuming it into a voiceless’ stop, « variety of tpl PHONATION! A THIRD RASIC COMPONENT 43 De the same thing with [p22 d 2 4] [22 g2 4, striving in every ease to ‘keep the voice going tareugnout—paticulary during the stop. If Experiment 30H was carried out correctly, the sound produced in the riddic ofthe stretch [9 99 ...9 2.3] should have been a fully voiced stop [b] [Bl Ie]. In cach case, however. the stop was a very short one-—that is, the closure was maintained for a very short period of time. In the next experi- rent, the duration of the stop will be increased, but voicing must continue throughout, 31 (9 Sty fp 29 2 95), as before, with voicing going on continuously, right through the stop Now try and prolong the peciod of closure while koeing the vice going throughout {9913 3] 3b b aa] fo 2b b ba al (where [o][b 5] > Dj represent short and then longer and longer (b}artculations). Check by closing your ears and listening by boneconduction to see that voicing ges on srouthoue the stop Do the same with 34 95] (93dd-99] 00d édoajandoseo3) 00882 sbsseea- (ip Now produce initia fully voiced (b). As a guide to what to do, fist say {9 a] with stongly-voieed [v], That i ake in a breath, form the [v} articulation, get the gitis set for voice, then star up the plotaic pressure fnitiation, Voizing shoold start immediately and continue through the [7] 20d oninto the [22 3). Now try it wita[p]. Take a breath, form the (b-articuation, get the glottis set for voiee, and start up the initiation, saying [b 2 2 =), making absolutely sure that the voicing starts while the lips are sil closed for the (b). Do the enine with fd 9 aL You may find it dificult to make fully voiced stops that is stops like (b) {G1 (gl, particulanly the lengthened ones introduced in Experiment 314, If you experiment assiduously with voiced stops, you will find that if you ‘maintain a tight stop closvce, as for [b], then no matter how hard you try you canizot keep the voicing going duriag the closure ofthe stop for more thaw 2 fraction of a second. You will find that the initiation, and tbe voicing, come toa stop very quickly: The reason for this becomes obvious if wwe look at the aerodynamics of a voiced stop. Figure 13 shows a diagrammatic representation of the vocal tract, as a kind of bent tube, closed atthe bottom by the lungs, which are contracting to force air up the windpipe, Some distance above the Iungs is the laryax, ‘within which are the vocal folds. When they are viteating to produse voice44 PHONATION: A TIURD DASIC COMPONENT rc 1 sega cee? == a { sae Pere Foe ini ‘Transglotal flow al 6 Lowering of laryax Sy = Presue‘aquiat fe presureciference Femtlao he owees = recur foe owoesses Fig. 13 Acrodynamics ofa voiced stop thisis indicated by a wavy lin. Above that, the tube bends into a horizontal Position. to simulate the mouth, which can be slooed bythe lips, wish are shown in Fig, 13a as separaced 10 allow sir fo dow out ofthe mouth Figure 13a represents [p 2 2, the lungs ace contracting, rising the Pressure (+) below the glottis ‘The pressure above is pretty mush atmos- Phere, represented by (0), 80 the aie ean easily low from the high-pressure ~one (+) below the glottis, up through the glottis, where it sels the folds vibrating ait passes through, intoths ow-pressize zone (0) above ihexlotia, In Fig. 13b the lips have closed (for |b), aad as more and more ait pours upwards into the mouth but eannot get out (because 0” the elosed lips), the Dressure above the glottis rises (0 - +), untl, in Fig. 15e. it comes to ‘qual —or nearly equal—pressurebelow the glottis. At :hispoiat, ofcourse, flow through the glottis ceases, the voc ols ease Co vibrate ad voising stops, 7 PHONATION: & THIRD BASIC CONFONENT 45 ‘All ths happens in a moment, ang this is why it is difteult co ma the voicing in a stop for more than a fraction ofa second: dificult, but not impossible. Our study in Fig. 13 of the aerodynamics of a voiced stop suggests how voicing cart be maintained for a longer period. We have to maintain the pressurecirop across ihe glottis for a longer time, and we can Go this in one or both of two ways. One, we ean increase the pulmonic iniviatory activity, é increase the sub-glottal prescure; two, we ean increase the volume ofthe cavity above the glottis se that ths pressure wll ot build up there so quickly. The second. infact, is the strotegy that appears to be ‘genzrelly adopted in languages that demand full and somewkat prolonged voicing of stops It is done by slightly lowering che larynx and possibly also cnlarging the cavities above the glottisin other ways such as expanding the haryax (the space behind the mouth immediately above the larynx), and puffing out the cheeks (an expedient that is not available for (gl, where the articulatory strictures toe for back, and searely availble for (4D). Figure 13“shows the larynx being lowered, so that the pressure above the larynx is gain reduced t0 about zero, bat gradualy increasing (0 —> +, ‘The supcagiottal pressure is low enough for ow through the glottis (and consequent voicing) te conmtinve fora short time. In the nextexperiment, 32, we put this into practice. 32, While looking at Fig. 13 uy once again to produce a (b] with lengthened losure--in face [b b]-in [2 b b 9}. Proceed exaetly asin Exp. 31/ but this time, the moment the [b] elosuce is made, start co lower che laryns, and see how long youcan keep voice going, while keeping the articulatory closure or [bjtight ia place. Ifyou have forgotten howtolower your larynx go back to Exp, 12in Chap.2, 4 Now experiment with produsing long [8 Jona {¢) We must remember that when we speak of & voiced stop (in a general phonetiecontext such as the present one} we meun a stop that has the vocal oles vibrating during the entire closed phase of the stop, just as a voiced fricative is one that has the vocal folds vibrating during the entire duration of the fricative stricture, Fig, (4 indicates, diagrammatically, the difference between voiceless and voiced fricatives and stops, Hanked by a vowel-like [29] sound 1 was aecesseity to emphasize that a voiced stop is one with voicing continuiag throlghout the closed phase especially for the benefit of readers in whose native language what are often called “voiced stops’ are partially or46 PHONATION: A THIRD BASIC COMPONENT Fig, 14 Voiceless and voiced stop and fricative compietely voiceless, This is true, for example, of English, where in a word like by the short medial [5] may be fully voice, but inthe initial (b] there ray be no voicing at all during the closed phase, or the voicing may start only about ont hundredth of a second (one centisecond, orl -I[ed.]leg-, where the “voieeless diacrtic’ ,]is offieca litte to indicate that ic is only tbe final part of thestop that is voiceless, Readers can convince themselves of the near-voiceesiness of Eaglish 6, PHONATION: A THIRD BASIC COMPONENT 47 4, and e by saying such words as bay day gay, Bore door gore, buy die uy and ebb Ed epg, lab lad log, dub dd dug, with the ears covered in the manner described in Experiment 22 so that aay voicing that is present during the closed phase of the stops is heard as a loud buzzing inside the head. Readers whose native language is Chinese (Mandarin and most other ‘major dialects) will find, by carrying out the same experiment, that the sounds represented by b d g in the Pinyin spelling system of their language are totally voiceless. They differ from p tk only in terms of aspiration, which we discuss below (p. 5), Speakers of languages such as English and Chinese, which have no fally voiced stops initially and finally, should try to produce them as in Experiment 32, 6. Voiced implosives If the larynxiowering manguvce. carried out:in Experiment 32 is per- formed in 2 particularly energetic manner, and if, at the same time, the pulmonic pressure is much reduced or entirely suppressed, then’ what started olt as in attempt at a'pulmionie pressure voiced stop may end up as a glotualic suction Voiced stop, a voiced impiosive. Figure 15 stows an ideal Voiced implosive, The initiatory movement that ‘manipalates the air pressare in the vocal trac isthe downward thrust of the sirscepswpwatds, Nopuimenic ‘henge rating peeture lows s oo 15 Mechanism cf voiced glottalic suction stop (voiced implosive)48 PHONATION: A THIRD RasiC COMPONENT larynx, with glotis set for voice. The downward larynx movement enlarges the space above the glottis, suddenly lowering the pressure there © below the atmospheric pressure. Sinc: the lungs are inactive, thereis little pressure below the larynx (0), but the pressure above the larynx is very low (~), and becuse the voosl folds are not tightly closed, but set for voiee, a small amount of air seeps upward through the glottis into the vacuum above it, setting the voeal cords vibrating in the process. But this upward seepage of dir, though enough to produce voice, is not sullidient to fll up the enlarged space above the glottis Consequently, when the ora closure is suddenly released, the air pressure in the mouth is lower than atmospheric pressure, and asa result thereis a momentary fale of sir into the mouth, This type of sound is called 2 glovtatic suction stop (symbols 6, «, ) Decause the essentia! alr-pressure change that initiates the sound is caused. by the sudden downward movement of the larynx, and the effect of this is te negative pressure (suction) in the air tapped between the glottis, toga and the oval closure. Now that you understand the mechanism of voiced glotialic suction stops (‘voived inmplosives') you can experiment with producing them, 33. Close the lips and while keeaing thein closed, jerk the laeyax downwards trying to produce voice at the same time, Remember that the luaes do not Bricipate i this type of initiation, so try to keep the chest fixated —the Jungs inaetive.[f you suddenly open the lips just before the larynx hss com pleted its downward movement you should observe 4 sight infix of ai into the mouth, tis somerimes helpful wy think of "sucking with the laryns” and to exper ment with the production of unreleased vied gotalic suction stops— ie, uneéleased (6], In doing this you make a sucking. gulping sound chat generates a momentary buzz of voice in the larynx, but in which the lips are newer actually open, A common error of peuple trying to produce (5) [d] [gis to turn these sounds into pulmonic suction sounds, that is to Icy to produce them with inhalstion—deawing air downwards through the laryox, instead of letting it seep upwuids through the glouis as Remember that, even though there is an infux of air from outside into the mouth on the release of [6] [a [, there is no inkaation-—no in-breathing — the lungs are not iovolved at all and the vacuum in the mouth is crested solely by the dowiiward ihovement of the larynx. ‘ynX itself moves downwarc, A pHONATION: A THIRD BASIC COMPONENT 49 A good way (0 think of the difierence between voiced pulmonic pressure stops [b][d][g] and voiced glowalic suetion stops [6] (6 [is as follows: in voiced pulmonic pressure stops we have a moving column of air passing upwards theough a storie eDT# while ip voiced glottalic suction stops we have a moving glouis siding downwards over a static colunn of air. As we know, this isnot stritly accurate, since the laryox may move down alitilein producing a (lengthened) voiced pulmonic pressure stop—but itis a very helpful way of visualizing the dillerence between the two types of sound: Fig, 16 ilustrates the difference diagrammatically. Voiced glottalie suction stops (voiced implosives) like [6] [] [¢] eccar in| quite a number of languages in the Americas, in Africa, and in South and j South-east Asis. But chat is not why we ask the reader to experiment with | ‘hem; we do this because experiments like 32 ané 33 belp the reader to get an experisneed, internalized, understanding of some aspeets of the aero- dynamics of spesch, rather then a purely theoretical and intellectual Cece ; i . Owasso a oe noe ine = = a twit Fig. 16 Initistion of voiced plosives and voiced implosives 7. Principal phonation types So fain this chapter, we have considered oaly two types of phonation: vwiceless (gts. wide opea so that air passes through with minimal50. PHONATION: A THIRD BASIC COMPONENT obstruction) ard voiced (vocal folds broueht together and thrown into vibration by the passage of sir through the gloitis). But these are only 40 of several diferent states of the glots that are responsible for phonation. In exploring states of the glottis we stat with the voedl folds as widely separated as possible. This is the position of deep breathing. Even when the folds are as wide apart as possible the larynx stll represents an orice only about halfas wide asthe windpipe, or trachea, thar extends below it to the Jungs Consequently, even the open glotis offers some resistance to air passing through. and in breething in and out one can normally hear some sound generated by turbulent flow through the open glottis Only if we slow down our breathing to a considerable extent the sound of exhalation and inhalation abolished, From this wide-open position (the state of the glottis for breathing or for voiceless sounds) ve can narrow the glotis, so that the orice presents a much greater obstacle tothe passage of ar through it, and the flow thus becomes quite strongly turoulert, generating the ‘husting’ sound that we call whisper 34 Breathe in and oat seweral times through the wide open slots As you will observe the Row of air through the gots is turbulent and heace noisy, unless you slow it dows considerably ‘Now produce the sound of whiper—if necessary, begin by saying a few ‘words to yoursel’ in a whisper—thex ty to produce a prolonged uamodi= fied whisper, kind of [993 where the litte Book under the vowel symbol incicates whisper. This hook (a subscript comma) isnot a currentiy ued symbol ofthe IPA, butit was wed in the early year ofthe Association for example in Aim ond Principles f the (PA (1904206 1812) ‘Now alternatebreathand whisper, a frstinseparatelile urstsofbreat and whisper [bh] ] [ga]. and thea in a continuum fh79 99]... -ete ‘When you are quite sure that you can produce either breath or whisper at will, you can easily demonstrate that waisper invelves a narrower channel ‘hrough the glottis than breath docs by means of Experiment 34A, 354A. Havea watch that shows seconds available, asin Exp. 19 Breathe in tll the lungs are comfortably full, then allow the lungs to collapse by themselves, expelling the air as in normal exhaling, (This breath, oF voicelessness). You will probably find that the lungs empty quite quicky—io no more than 2 seconds, ‘Now breath in again till the lungs are comfortably full aid this ime when yor allow them to collapse, doa’t just breathe out, but produce the sound PHONATION: A THIRD BASIC COMPONENTS] SS Ee of whisper. You will find that the whispered breath sakes considerably onger to ow out—probaby 5 to 7 seconds or more. ‘You used approximately the same quantity of air each time, but it obvi- ‘usly Bowed out more siowly for whisper than for breath, This éemonstrates ‘thatthe glottal orifice is much narrower for whisper than for reat, So, then, for breath (voicelesness) the glottis is wide open: for whisper it 's considerably narrowed. A further narrowing of the glottis brings it into position for the production of voice, with which we have already exper. ‘mented. Incicentally if you camry out the same timing test with veice as You did with breath and whisper in Experiment 34A you will find that it ‘akes stil longer to empty the hings: perhaps twice 25 long as it does for ‘whisper, This is because the rapid opening and shutting of the vibrating alottis lets the air pass through only.in a series of tiny puflyy and this ‘naturally increases the time it takes for the full complemeat of air ta flow out. ‘There is ene more type of phonation to be considered. This is oraak, Iris a very low frequency reaking’ or ‘crackling’ sound which—pethaps because it reminds.some people of the sputtering noise of fiying eggs or ‘bacon—is sometimes called ‘glottal fry’. Creak is produced with the glottis ‘completely (but not very tightly) closed, except for a small segment near the front end of the vocal folds, which is vibrating rather slowly. 35 Produce a sostined vowel souad of the type (293) of fa aa] on avery low piteh—about as ow as you can reach, While keeping the sound going, try to lower tbe pitch stil further till the tone of the voice Gegenecatss into a care ‘9ptonous series of taps. At this point the frequency ofthe glottal vibration is 50 dow that you can perceive the individval litle explosive bursts of air. By tial and error try to achieve a clear crackling noise with no admixture of the tone of voice. To do ths, you may have to conseiouuly tighten up your woeal foies, though not so much asto produce a glottal stop. This is creo. ‘We have now discovered four different types of phonation Woieelesiness (or treat): withthe plots wie spen, as in tating oc the production of voiceless sounds, such as [fs Bt Xl. A Whisper: with a considerably narrowed glottis 0 thatthe hihiy turboe lent airflow through the glottis generates astrong “hushing’ noise: this is the phoration type of whisnered speech, in which whioper is substituted for voiced sounds (while voiceless sounds remain voiceless)52. PHONATIDN: A THIRD BASIC COMPONENT Voice:the tone produced by the vibrating weal folds: the phonation type of all wiced sounds, sveb as [¥ 2.111 isa a Creak: the crackling sound produced by keeping the voeal folds closed along most of their length, but allowing the air to escape in a low fre- {quency series of bursts through a small vibrating segment near the front end of the gous Figure 17 shows, purely diagrammatically, the state of the glottis as it might appear if you were looking dowa on it, for these four phonation types, and for complete glottal closure. The conventions used in Fig, 17 are mostly obvious, but note that the difference in the disgonal hatchiag in (¢) aad (d) is intended to suggest a dilerence in the frequency of voeal fold vibration—medium to high frequency for voice (¢), and very low frequency for ereak (d). Note that complete glottal closure is not a type of phonation, Phonatioa always involves a flow of air through the glottis, Waen the stls i held in the closed position there i, of course, n0 flow. The] Functions of glottal closure are thus purely articulatory (for glotal stop 7) | and initiatory (for glottalic pressure sounds and unphonsted glottal | suction sounds). In the ease of volved impioives the slots funetions! simultaneously as initiator ane phonator. ‘There are various pessible combinations ofthese types of phonation, the chief ones being breathy woe, whispery wie (also known as murmur), and creaky wice. Experiment 38 intoduces these Uuee combined phonation types. 36 Greatty voice: the glouis is rather widely open, but the rate oF airflow is 0 1 that the vocal ‘Olds are set “Dapping in the breeze" asthe air reshes by, Exhale very strongly, costriving 10 produce some voice ai the same time, Imagine chet you are tying to blurt out « mesage when extremely out of breath, oF speaking while sighing deeply Wuspery vice or murmur: the voral Tolds are vibrating to produce voice a © rT © Open otis Whspee Voice Crk Closed glue (vorclenness) Gowers) Fig. 17 Se es of the glottis PHONATION: A THIRD BASIC COMPONENT bbut atthe same time thors iga continaous eseape of air generating the sound ‘of whisper. Produce an energete and prolonged stiong whisper, and waite it Js going on acd voice, to produce the mixture called ‘waispery voise’, Prox ducing voice with very relaxed vocal olds may help (Creaky voice: proceed 2s 1m 35 start with low-pitehed voice, then go still ower tl some creak is heaed mixed with voise—but in thiscase make 20 \WCempt 10 produce ihe pure creak that was the end eesult ia Exp. 33. Breathy voice, as we imply in 36, is the type of phonation we use when out of breath and panting. It is not otherwise used in English, althou high-speed cinephotography of the larynx shows the glottis passing momentarily through a “breathy voice’ phase during the transition from voicelessness (breath) t9 voice in an inital fh, as in happy gh Whispery voice or ‘murmur’ is sometimes used in quiet, initimate, con- versation in English, and is often assoviated with tenderness: itis, for ‘example, the type of phonation used in saying ‘I love you’. If you experi- ‘ment with ssying those words in a normal or tense voice, you will perecive Ue inappropristeness of this type of phonation! Whispery voice is used in seme lunguages to cistinguish words: thus in Hindi che sounds known as “voiced aspirated stops’ [bt] [d¥ [4] [a5] are procuced with whispery voice which persists into the following vowel. In fact ia some varieties of Hindi the most noticeable differenee between such pairs of words as [b4a*t] and [ba+t] 8 simply that the vowel in (b*a-t] is pronounced with relaxed, ‘whispery, voice, while the vowel of [ba -t] hae rathar tense voice, (Creak or creaky voiee is a not uncommon phonation type of very low- pitched syllables in English, particularly in the British variety of English known as RP (received pronunciation’) We can now provide a formal definition of phonation. Phonation isany of those phonetic activites of the larynx which have neither inisiatory nor articulatory function, 8. Relationship of phonation to initiation Since phonatios (the various types ef vieelesenes, whisper, and voiee) ean ‘occur only when there isan ai-stream passing through the kirynx, it follows that it con Gombine only with certain types of initiation. ‘Velaric initiation (¢2, velarie suction stops, or ‘licks’ utilizes exchusive- licks wuantity of air trapped in the mouth. Consequentl ly a small have no phonation. They may, however, be performed against a phonatedSd PHONATION: a THIRD BASIC COMPONEN’ —S background. Thus, as we saw in Chapter 2, itis possible to produce a series of clicks, say (], [i], [], while uninterruptedly humming through the nose. Tae inom is a pulmonic pressure voiced sound. But the pulmonically ini ated air-stream used for Une hum passes cut through the nose and is not at all used for the click. The later is a velaric suction sound, and itis unpho- nated, because the sir used in its production does not pass through the larynx. and thus has nothing to do with the production of voice. “Voiced clisks’ are sometimes referred to, but these can only be unpkonated velaric suction sounds, performed against a background of pulmonicelly (or pos- sibly glottalically) initiated voicing. : Glottalically initiated sounds, such os slottalie pressure (p']{°} [¢] and their inverse, glotcalic suevion counterparts, ate unphonated, since in their production the glottis is tightly closed—ro air is passing through to ‘encrate voiceless, whispered. or voiced phonation, As we have seen, however, siotalic suction sounds cun be phonated, In voised glottalic suction steps (voiced implosives) the intiatory movement, 'S indeed gotialic. However, the glottis is not tightly closed, but is disposed for the production of voice, and as the laryax moves suddenly downwards sonall quantity of air seeps upwards through the larynx, generating veice. Pulronic sounds, are, in principle, all phonated, i.e voiced, voiceless, ‘whispered, etc., with the exception of glo:ta! stop [F]. Glottal stop i of course, pultoni, since the pressure builéup below the glotis is initiated by the lungs But since the gloitis is tighdy closed, forming the articulation of the sound, it cannot simultaneously be opea to produce voicelessness, whise er, ot voice. All other pulmonic sounds are phonated: Bat, it should be foted that there are constraints on the phonation of pulniontc suction sounds IT you try to produce @ voiced vowel sound {9 2. ..J on inhalation, fe. with pulmonic suction initiation, you will find that its vittually impos. sible What is produced is a raucous eroaking type of inverse voice. The share of the alottis is such that it vibrates well under pressure from below, but aot at all well under pressure from above. Consequently, a normal-sounding voice is net possible with pulmonic suction initiation, ‘When voice is produced, the vocal folds vibrate a a frequency determined. by the tension of the vocal folds andlor the subglottal airpressure. It is possible to vary the frequency of vocal fold vibration over a wide range. ard consequently to produce the auditory effect ofa wide range of pitches, All languages make some use of voice pitch variation, and we shall study that in more detail in Chapter 9, ‘Prosodic Features’; PHONATION: ATHIRD BASIC COMPONENT 55. 9. Aspiration One further important phenomenon connected with phonation isthe tne relation betweca phonation changes (voiced-to-voieles, and voiceless-to- voiced) and articulatory events For example, if we pronounce an initial, stop consonant followed by a vowel, the oicing, which is ceeential-for the vowel, may start simul tancously with the release of the stop-clesure, or it may be delayed for a fraction ofa second, or, again the voicing may start before the stop-closure is released. Fig. 18 shows, diagrammatically four diffeent time-relations between the release of a stop and the onset of wong. + {In Fig. 18, he upper line represents the closee pace of an initial stop, for example, the lip-losure ofa p}, followed by the sudden explosive opening ‘of the lips. The lower lines represent the glottis either slightly open (as for whisper), wide open (as for breath), cr vibrating. In 18a the vocal folds begin vibrating at the moment when the lips separ- ate In 18b the wide-open siottis takes some time to close sufficiently for the vocal folds to start vibrating. Consequenty, there isa period of voiceless: acc, 3 kind of [h}-sound, before the vocal folds begin to vibrate: In 8c the vocal folds begin to vibrate a moment before the ips separate, and in 18d wie ee) or ewe bro ed @ ray weed Fig. 18: Aspiration and voicing of stops56 PHONATION: A THLIKD BASIC COMPONENT the vocal (old vibrations begin much earlier—as soon as the articulatory closures formed. Its clear that both 1Sa.and ISbare voiceless stops: there isabsolutely no ‘vocal fold vibration during the articulatory closure, No. 18 is a slightly voiced stop~mostly voiceless. but with voicing just beginning, before the release of the closure: initial (b] [d] (s] of English are often of this type, Finally, 18d is a fly votced stop, with the vocal folds vibrating throughout the entire duration ofthe articulatory closure ‘The difference between [Ss and 18b is what chiefly concems us at pres ent, A stop ia the prodactioa of which the vocal folds start vibratiag to produce voice more ot less simultaneously with the release of the articula- tory closure (ike 18a) is waspireted. A stop in the production of which the vocal folds start vibrating only after a delay is axpiraied, the voiveless put of breath that intervenes between the release of the stop and the start of voicing is aspiration, Jn most types of English the voiceless stops (pl {tl (kl are aspirated when they occur before a stnesued vowel asin par, tar, car Ipha-@] (@a-( [Kra-(o], where (e) implies that in some didlects the x} is pronounced, in others not. In varieties of English spoken in Scotland and the North of England, |p} {t] [k] ane unaspirated So, t00, are the [p] [1] [3] after (s] in spar, star, sear, ia viewually all types of English, {Cis important to be quite clear about the distinction between unaspi- rated and aspirated stops, end the frst thing to Uo, if you are a speaker of English, is to experiment informally with paics of words such 3s pietapy, pankspon, plspi, tele, tonelstone, elise, Keuski, kinlskin, Kill. 1b you say these energetically, but very slowly. you will probably obsecve the puff of breath that occurs hefore the vowel in the first member of cach of these pais, {In earning t0 control aspication it is helpful co note that the delay in voicing, which is so evident in aspirated stops, is. 2s it wire, only a symptom of a more fundamental feature This fundamental feature is the width of the glowal opening during the period of closure of the stop articulation, Ifyou look again at Fig, 18a you will see that the lines indicating the state of the vlottisare quite close :ogether, symbolizing a glottis that 's only open a little during the closed phase of the articulation. Since the vocal folds are already close together, they can spring into the vibrating action needed for voice the moment the articulatory closure is released. This stare of the glottis has been appropristely named prephenction by Hlartis (1999). The PHONATION: A THIRD BASIC COMPONENT 57 area of glottal openiag for prephonation is similar to that for whisper, though the location and shape of the narrow channel thus formed is differ- ‘ent from that indicated in Fig. 17b, which represents only one type of ‘whisper. In prephonation the channel extends forwards slong the greater part of the glottis, as ir may do in other forms of whisper ‘The knowledge that uspirated stops are produzed w-th a wide-open| (breath-type) glottis, while unaspiraied stops are produced with «narrowed glottis (of prephonation type, somewhat resembling whisper) suzgests an | experiment that will help to develop conttol of aspiration. \ 37 (9 Breathe in then begin a prolonged wiceles exhalation—s pulnionie pres sure initiated airstream flowing up through a wide-open glotis, Momentar= lly superimpose 2 fp] on the air-sueam as in Exp 7, thus [ka p Hh b]. Note that in doing this your alotis is wide open chrouzhout Repent the experiment, but this time “switch oa’ voice immediately after reksave of the stop, crus (ht hp H99 9]. This should produce anaspicated [p]. Do the sume with [1] and fx). (a Breathe ines before but now besin a prolonged whisperedexialation— ‘4 pulmonic pressure initiated airstream flowing up thioush te alos, generating the soued of whisper, Momentasly cuperimpose 2 [p] on this Whispered ai-seeam, [ap @ al (whew @ represents & whispered 2) Now repeat the experimeat, but this time ‘switch on’ voice just as you release the stop, If you have indeed kept the glostis in thie whisper configur- ation up to that pint, voice should star: immediately—with no delay, In other words, the result should be an unaspiretel fp}. Do the same with (0) ‘and (i) Youshould aow kave a clear, experiential (not merely intellectual) know- ledge of the difference beoween unaspirated [p] [1] {k] and aspirated [p(t] [bk], Itis possible to produce exactly the same difference between two types of fricatives: unaspirated (} (s] [f] ee, and aspirated [PY [s¥] J]. In the unaspirated acquiing this distinction i is hep at Gs, to whip ones, but make the aspirated ones very breathy, Aspirated fricatives are ‘much rarer than aspirated stops in the world’s languages, but they do exist. In some cise forexample Burmese [the eave itaelF as amach wider articulatory shane! than hat of the unasprstd (5) The aspirated stops we have cealt with here are all veeelss. There is another type, known as voiced aspirated stops’ of ‘aspirated voied stops particularly ommoa inthe languages of aon Tea. ta the production oF thes siops thé phonation typeis whispery vos onsicerabl pressure'ssareace'obthe ‘sop there isa vibevocab Ibidssewtle Uowa to normal voicing ‘Setepresenieeas (b%] [2] [g"), ete.,or fb] [2] ” sistxhe-IPAsymbol for ‘voiced fh]’, which is, in Gia morentarsburss OF whispery! voice. As we mentioned above, in EINORH Indias thygtapes the otire vowel following (b" [2] ec, may be Fphoiiarcll withiwbispety voice. ‘isteenjsebapter we have investigated, ard experienced, the principal types Of phonation, samely voiceless (‘breath’), whispered, voiced, creak, and the compound types breathy voice, whisper vice murmu), and creaky vice, We have investigated the special problems of voicing stop consonants, and she aerodynamic differences between wiced pulmonic presse stops [b (Gl lah, and voiced slottalie suction stops (void implocives) (8 [f(g We noted that the piteh phenomena of language depend upon the free quency of vibration of the vocal folds, and thus constitute a phonatory phenomenon, i ‘And, finally, we have investigated the difference ‘between unaspirated and aspirated sop. 4 Ss Articulation: Stricture Types 1, Articulation: The ‘final shaping’ of sounds We tum now to the third: besic component of speech production— = ‘Theresre three other stricture types to be considered: rill, ap (and flap, and semivowel, Trill involves neither the complete and maiatained closure of a stop, nor the maintained openness ofthe three stricture typesin the table we have just seen, but, rather, an alternation betweea them. In till one dexible organ repeatedly taps against another under the infuence of a powerful air-stsam. Approuching this stricture-type experimentally we start in Experiment 44 with the easiest type of till, a bilabial ((wo-lip) one, 48 Pluce the upper and lower lips lightly together, in a somewhat ‘pouting’ pposture—thatis, kept perfetly fat (not rounded) but pushed forward so chat it is the imer parts of lips that make contact. Now start up a powerful pulmonic pressure initiation (‘low had’) and, perhaps after a lithe ‘adjustment of lip-tension and iniatory pressure, the ps will start lapping segularly, periodically, against each other, This isa voiceless pulmonic pressure bilabial trill, and it resembles a sound wide by horses (and ro horses by horse-lrivers in the days when such people ‘were common), While doing a prolonged voiceless bilabial will, switeh on ‘A voiced bilabial tril is sometimes used as part of a gesture expressing, coldness, written brrr! Veiced bilabial til is a realar sound of few lan- ruayes, such as Nias, the Language ofanisland off the mutt coast of Sumatra, Bilabiat cil i represented by a small capital (0), : Experiments with bilabial will will have shown you the yeneral pria- ciple of vill-pmduction There are, however, two other typer of trill that are more commonly used in languages, namely the apico-elveolar trill [4]. produced by allowing the tip (apex) of the tongue to flap against the ridge bbehiad the upper teeth (alveolar ridge), and he wular rill [x], produced by letting the uvula (the pendulous rear extremity of the roof of the mouth) vibrate ina longitudinal groove formed in the buck of the tongue66 ARTICULATION: STRICTURE TYPES 445 Raise the tip ofthe tongue stighly and let rest noselyagninet he exteme ‘ck of the rdge behind the epperteth, Now srt wpa powesal pulmonic res intatryairsteam (veces) ie. Dow hard By thal and enor agjusmens of the precise postion ané tension ofthe tongue and ofthe intinory presure, you may beable to star the tonaue-tip resvlay periodicals, Banping agains the aveolr rie. This i). Having acquired the voiceless (r] produce a prolonged [j 7 ¢. . .] and switca on voios in the middle offer rye. Incidentally, inexpesimening with lst sofa bso start witt eels ones snc it iseasir tose: organsiilng withthe ror: poverfl voices ai-steam than with the voied or. Uvuar{n] ean te developed from garting: most people are fie with the technique of garaing with ditintctan, alt water oo. To develop uvalac {ny take led wazer into the mouth teow the hea backard garsle, Thea ‘eduoetheamount of water til Sally youn do it merely with ite salva atthe back ofthe mouth For some reason mos peonle practise garging witha wived acstream, bot this i quite unnesesany. (i probably better, in making the waasiten froma gaetoa wvaar ilo usea cele xtra Ast, prodacing a} ‘The apico-alveolar tril [tis a fairly common sound in languages: itis 2 typical Italian 7 or’the Castilian Spanish + (aote that in mazy Latin American forms of Spanish rr represents quite different sound): itis also 2 type of r traditionally used by stage Scotsmen (real Scots rarely use 2 strongly trlled 1) ‘The uvular [x] is one form of N. German, itt also. type of roften used in modem Isracl Hebrew, and occasionally in French—though it must be noted that in tese languages r is quite often a uvulatfricative, ot even a vvular approximant, This is particulary trve of French All the articulatory stristure types dealt with so fer—siop, /ricative, epproxtrant, resonant, and trlt—have ore thing in cominon. This common feavure isthe fact thatall of these stricture types are maincainable articula- tions. The closure of a stop (which is Of course, the one essential phase) can ‘be held for an indefinite period of time. The articulatory strictures of fica tives, approximants, and resonants can also be maintained indefinitely. A tril, too, is an articulatory posture which can be maintained for an indefine ite time: the vibratory movement characterizing a tall isnot an active per~ formance of the articulator, but simply a passive response to the airstream, ‘which sets the articulator ‘lapping in the breeze’So then, all these sristure ‘types are essentially maintainable—a stop or a fricative, for instance ean be HIER GERD ARTICULATION: STRICTURE TYPES. 67 guite short, but they can also be very long, i.¢.can be maintained for a long ume, 3, Essentially momentary stricture types: flap, semivowel ap, ‘The remaining two asticslatory strctuce types are, by contrast, character ized by being essentially momentary. Tap and flap both invelve aa essen- tially momentary contact between articulators. In 2 tap one articulator starts out towards another, makes momentary contact and immediately withdraws: the whole cient is 2 momentary Aicking movement. Oze of the best-known taps is the flicking movement of the tip of the tongue against the teeth-ridge [r, which often reoresents af or d between vowels in Ameri- can English, asia latter or ladder (keea, or an r between vowels in some types of British English, as in very [veri This ype of apical (ongue-tp) tapis also the single r of Spanish pero [pero} ‘A flap typically differs from a tap in that the momentary contact occurs, ot with a ‘Bicking’ motion, but as one articulator momentarily strikes nother in passing. A typical example isthe ‘retroflex fap’ [Ja sound chat cecars in Hindi in such words as [gota] “horse’. To produce this retrofex fap curl the tip of the tongue up and back (this is ‘reirofexion”) then lt it shoot forward so that it momentarily flaps against the extreme back ef the ‘teeth ridge before it flaps down on 10 the floot of the mouth, Thus the cap/ fap sttictuse type ‘involves an essentially momeatary contact between articulators’itis iipossible tomaintain the contact inthe way that on= can for stops “The other essentially momentary articulatory stricture is semiowel. A sernivowel isa momentary approximant. Tus tesemivoweloccurring at the beginning of the English words word’ wallorin the middleof the words away, aware (phonetic symbol {w) is essentially a momentary or vltra-short (u] ‘Vowel. The semivowel odcurring at'the beginaing of yard, youorin the middle cf Maya (written y in English, but Gin TPA) is an ultrashort [i] vowel. In Experiment 46 we develog semsivowels from the corresponding vowels: 46 The vowel {you vill recall s'a typical approximant, and the same is tne of ful, The semivowels ae similar approximants, except that they are ‘tra short68 ARTICULATION: sreieruRs TwPes “Take in a deep breath, and then start saying « very prolonged vowel, prolonged [a] {like the vowelin cain), then suddenly changing to a prolonged [i (Ike the ee of sce), then changing back 10 a prolonged fa, thus: fasaaaiiiliaaaaal, Produce thissequence several times, then start systematiealy shorting the [thas starting.as [aasaaiiiiiaaayal vaaaiitiaaaal [asaaaiiiaaag] faasasiiaas] faasasiaag sfaaaasjana) ‘As you ari at the last sequence, with 4 sery short (i, you will observe that the [] no longer sounds ikea sea! vowel, but has turned into something that sounds ike English y (IPA Q). ‘This demonstrates that the semivowel {jig in eect, an ultrashort fi. Carry ovtthe same experiment with {u] and note how i finally urns into the semivowe fl foaaaguunuusaagl fuaaasaueaas) sannauusaa] faaaaauaaal [Experiment 45 shows how a semivowel is the same as av approximant vowel, except that itis momentary. As soon as you prolong the approximant stricture of lo fw}, it turns back into a vowel. Vowels are maintainable sounds, Semivowels are momentary. Another way of describing the difer ence is to say that a vowel has three phases: sn upproack (or “on-glide’), when the articulators are going into position, a hold when the articulators remain in position for an appreciable time, and a release (or ‘off glide’) ‘when the articwittors move away from the position they took up during the hold’, A semivowel, on the other hand,'bas only one or two phases—it completely lacks the hold phase, An initial zemivowel, like the y {ia yes haz no audible approach. no hold, but only a rapid release, or olllide om the [i] position. A medial semivowel, like the [w] in away, tas an approach (onglide) followed immediately by a release (of-gide) with no intervening hold. The ‘aiticustors go from a neutral or relaxed position into the position for [}, which they immediately leave again, ‘We can now sum up the various types of articulatory stricture 2s shown ARTICULATION: STRICTURE TYPES in Fig. 19.19 this figure the movement of one articulator up t9 snother is epicted in a selfevident way, and, where cpproprate, the type of airflow through the articulatory channel is conventionslly represented as follows: voiceless, nonsturbulent voicel voiced, no: voiced, with turbulence (hence hiss superimposed on thesound of voice). So far we have coverad the “vertical” dimension of articulation—the with turbulence (hence hiss-naise) turbulent I degree or type of articulatory stricture, and the ‘time’ dimension, in ‘which we must distinguish between those articulation types which aremain- tsinghle for an indefinite period of time and those which are essentially, obligatorily, momentary, nos-prolongable. wird ~ s 7 \ sinay spprouimant po semi-vowel x Romnt Fig. 19 Major stricta types 4, The transverse dimension: Median and lateral articulation We now tur to the éransrerse dimension 0” articulstion—the location of the ora} air-path, that ig the distinction between 2 medium-or central and a aera! articulation charnel. ‘The vast majority of speech-sounds have 1 median articulation channel but a fow have a complete obstruction in the median or eentral part of the mouth lesving.open a chanael, or channels, along one or both sides of the tongue We have, in fact, already referred to the characteristic lateralpurve si} ts, vers wasdde) 34 deveg uous Bad te y a canon [248 | wossoounp asst ap uy dks axons soley, 02 ‘614 oe sypodeoro ye “(epumos ad [jo qred-are om —on0y aun Jo (aps axa Buope Avo mo swog ame wey os anus 2p U! PayDOI sf aBessed-umnow om oom w!) jar) poe “(qom om Jo uN [pou 2m 00R mo soy are ay: yon Ur “yNed asovoUMOD amt 7ey Aq) TenI9 10 woppows ‘syzod-ate [310 1EOINTP Ons 7A TE PEOOT sAPy se GONIPP “Qs Te) uonarsqo-uow poe (nen pure aneaty doxs Aauren ‘aogsTe 0: dononnsa0 sofew v ayoum ysrys ssctp) 2uaKu2Wgo sexm 29400 9m vopan sedi am “us asaxp Jo awos dno on rypsn seuamewos sy yomonejues pur doyle —ojqeBao(oxd-uow) Keenawou pure ‘runwoses ‘jumunxoudde ‘oeeotif “ptt ‘dois—(oarSeorexd) sraouoinmu ‘emtsans dvotomonan 30 sadks SupsoI amp pazesoosip sacy ayy “punos 30 adh aynods © jo Suidoys jeuy, a 20} 2jgusuodsar st ret) Cononposd-yosads Jo wauoduros ayy “wonejonre {Jo sicedse amos 32 payor aney om 1oideyo snmp wt ‘vam in ws 0} esse, Jon wren xo4730 5] yay Sunsoys arajdu9 Se ¥ st umogs 81 dois ‘pauredxa imnf wostax ayy 30g “ypedesre sein, 40 woypaLe AUK sodha amnsouns sofeu amos Jo uorreMquien aap susoBHns gz aunt ‘aBessed-ae exo am Jo aBe9016 mica © sheaye st gy wuowuodag wosy 20s om sv *do3e amp Jo ured jenussse SiWL dois m1 paryag da spring aunssaid yous Buunp amsope Jo pousd Tenuse ata —pjoy aun ‘dsmnoo Jo ‘s dons v Jo wed onsua.DETeGD OW aU] “fait 7 jo aseqas mnezoyeeiew oa tai Kops 1 peda.) ‘peop resi um.doas v1 sa (908) 1119] Sus) oug aun Buorasd nang 1 saoqs dun Suyeay [=] 3p go mo vay > 2) sp ways Fusc08 “a op 2g yo 19 eat, puw aR a Keg Hnuegs 27099 pom amp Aes (4) {B11/p.2 = Jorwanye Aways oy vyowossee erp feaoU 283 ML APUONS, “ay amdanoD cyowasdée jee, Ye (“dows v st SHEL “a aes 29j sanbop atyétson Arsszoog aa axew 01 Meas sno Suns Jqwoppns wor“ aps 20; vogsod wr dp-aneon am 9) UO nok eR op no sy Geom) [4111 Annes ned ie xp exqon dem ua) £11 Lye did ie Saved ans wee 6 pene 797 pi0m 4S |) ey ‘eso 30 Aspe 2 patoau x0 poypvesdtn 29 ueD dons v YaXaMoK “sdo's UeIpeM, 2 12]51 1,vop am :xanom om Jo tpuat‘amiua etn dn syo019 ausopo am anus “exoret son werpoct pepo 99 ues deyev 0 areyd ammsoqe 9— HP eNOAGO #13 {9 aancous ein Sspso108 ap 01 [8] sao00 uatp AeED!OAAP {8 ‘SARWOL Teo] paoo. oruT[f] 1arUMKardde peor" wou wes su ‘Hs019m ‘ooag anor om jo saps ogy Suzaonbs Shemaps oxBu01 9 mo Seas 4a awerpmadde essere 9 reso om “passaoostp Kpeaiteanzy ane SY acres “Cray reser sep da wads oy yen yoog 20 aps dn oun wos Ame Suryeaig or3u01 a Jo OPS 9m (!) SEPL-NED, am MUS NUD aye oF én SurUOD andes si jo dr 2m [pay woo ns (]2u23s 0% [s} sate wos; sied nok Sy ‘uonejnonse RW pre wepam e uanatea SOUAHAYID a Jay Tes nok Ud pur [20 samasod'aonmnanse aya useniog stewiine 41. wok 5 juueyp uonejrome iano) ¥ sey (i “AeOID “Wem aps addin mp awed FumoG pye anuao op yeep gs saps ayn ve ypnoat 24 Baza Be poo Yo} we MOF °L1M-p sso} uonons arustitA su msn sy "apse an sane CTT Sexkos nny, EIT tr Dat stouep pax f=" ld peBeojad des mon nun Jo auyanuso 2 Buoy aps wey ut soy ate poo wat oj Axe 8032 3q Te? pry JouNBYD UONEIMoTE ura T sey [x] “spIoM Io UT MVEA: apys szddn sup pce apes sn snp pm yoreoa ayeur aa8u0? 2—2 JO 2S =p oyrya‘onSuoy oqn jo ued penuso pur dr avy aosoymowstp ow Fueed B89 psa pros nas oprosy {459} vse Fompaid sy —weansare om oust mon 58 59] anon 2S Le ‘eraousadio uorematse PaerPeE a pom aopas—G vonocnsip ay qeordde am s+ wourLadxg atang ‘spunes-[i]JoMOnEALONTe IL S8dkt suntoruss :NovIndIEY STEKLTEIONUS ‘NOUVINOIEY OL| 5 Articulation: Locations In the last chapter we looked at various articulatory stricture types, representing what we led the ‘vertical’ and ‘time’ (prolonguble/aon- prolongable) dimensions of articulation, and at the location of the oral 4ir-path, representing the ‘transverse’ dimeasion of articulation (median/ lateral) Now we turn to the very important quasiloagitudinal dimension of tures within the vocal tract. For the purpose of desczibing articulatory locations we divide the vozal tract into thres areas: nasal, oral, and phuryngeo-laryngeal. These articulation, namely, the /ocation of articulatory sti areas ars indicated in Fig, 21 There is a clear natural division between the nasal area and the others, constituted by the orifice a raising the velum or soft palate, as shown by a broken line in Fig, 21 phonetic parposes, the oral urea consists of the mouth eavity, bounded by the whole of the under surface of the roof of the mouth, hack ta the vila, and by the whole of the surfuce of the tongue buck to the tip of the sryngeal area consists of the pharynx, the space behind the mouth and down to the larynx, and the laryax itself back of the nose, which can be closed by «piglotts. The pharyngeo: 1, The nasal aréa ‘The nasel area consists of the nasal cavity, which is for the most part a complex but immobile chamber coated with mucous membrane which may swell pathologically, as when we have a cold, buts not capable of voluntery movement. Seme voluntary control, and hence some variety of acticular ‘Gon, is possibly only at the iwo ends of the nasal cavity, the nostrils and the pharyngeo-nasal orfico—the ‘nasal port, as it is sometimes called. &TA amricuLATion: LOCATIONS ARTICULATION: LOCATIONS. 75 Fig. 21. The major articulatory areas The nostrils can be narrowed, or widely opened (‘fared’), and can thus modulate airflow out of the nose, but this potentiality is not known to de exploited for articulation in any language It may be noted in passing, how- ever, that when you devoice a nasal sound suck as fra] or fa] you can hear a slight hiss-noise of turbulznt airflow through the nostrils Since these nasal scunds are quite free of turbulence when voiced, the airfiow becoming turbaleat when voiceless, they are typical approximants, and might well be called ‘nestail’ (0%, better, using the latinate term) ‘narial’ approximants. However, we do not, infect, use the term ‘narial approximent’. All sounds articalated with the velum lowered (the ‘nasal port’ open) so that air flows through the nose are simply called nasal or nasalized.» Ya masale,cush as (er) fa} fg) ae im mre, mn, and the velum or soft palate is'lowered, but there isa complete closure im the ‘mouth (at the lips for (mi, between tongue-tip and teeth or teeth-ridge for {a}, Between tongue-back and soft palate (velum) for (3). Consequentiy all the air used in their production is sbunted through the nose. Tn nasalized sounds, the nasal port is open (exactly as for nasals), but at the same time the passageway through the mouth is also open, so that the air fows out through both mouth and nose: Typical aasalized sound: are the.nasilized vowels of French, asin wn bon vn blanc [éeb0 v8 Dlg}. These are sometimes called; simply, ‘nasal’ vowels—but iis clear that they differ from the nasal conscnants [ml] [n] ete. as indicated above. Experiment 49 explores the diffsrences between nasal consonants, nasalized vowels, and purely oral Final soured of lang, 49 Say a prolonged fm mm. J and ncte how zr is fowing oat of the nose. TF you told your hang just below the mosis you cas faintly fel the warm air gently Bowing ont. Ifyou suddenly devows [rr] the nasal aiclow vecomes such more obvious: mm 7 1 ‘To ge the fs of taxing and lowering the velum—closing and opening the nasal port—say a prolonged {m] punctuated by izerted (6 stops. Keep the lips tight closed toroughout the cate experiment merely lipping the nasal ‘port momentarily shut for cach fo] then opening t again for the nasal (mj fombmbmdmdmbm..}. ‘De the same with fo] and f9}(the nasal hen at theend of logic éa da dadnde..Jf9898989 5989-4) Youmay be cle to fr! the nasal port opening and closing more cleacly if you repest the experiment with voiceless masls: [m1 PAP TE PAE- «pete Finally, do the same sequonce of experiments completely silenty—that is, with no initistory airstream, In this cleat experiment you can feel, even note cleauy, the profoceptive scmsaton of opering and closing the nasal port im pm pp pm... (Sent), et, ¢ ‘As a result of Experiment 49 you should find it possitte to lower ang raise the velum—to open and close the nasal port—at will, The next experiment concems naszlized vowels and the difference between thera and purely oral vowels, 50. Produce a prolonged vowel ofthe -type something ike the vowel oF Eng Tish bed or bad, Wile this prolonged [¢€¢..]is going on lax tho velar, Jet itdzop down to open the nasal port while carefuly Keeping the [¢} going: or simply think nasal while keping thee] going, think about the sexsa- tian of making smal sound. In other words. contrite to gear low out of oh meuth and nose converting [et the corespondingrasilized vowe! {G thus fecesesegg) Havngdonethsoncrortwice aernat:[etetcéee) ‘Now do the same with some other vowels, e.g. [i] as in see (7 sev! '9395..J.[] asin pa [ad a.) Ifyou old the bick of your hand, fora Sages very close fo your nostls you shold fel warm ir coming ou: otyoar nose for] GIG} te, but ne rE] Some people nase very open vowels, ik (oa th tii saying 8] [ka(s)], ete. for pa spa car. Test for this as just deseribed. If you do nasalize these vowels experien with ‘thinking oral’ Ge. izing the sensation of closed nasal pets as fora purly oral sound) while eying [o} ao back to thea uy co getthe same eastiosinto (28.08. ete16 ARTICULATION: LOCATIONS 2. The oral area: Upper and lower articulator We tura now to the vral ares. In Purely oral sounds (that is, in the majority ‘of all speeeh-sounds) the velum is raised, closing the entrance to the nose, ‘and che air Hows solely through the mouth, Articulations in the oral area are carried out by the juctaposition of lower nel upper articulators, The ower arciculacors are those attached to the lower jaw—the lower lip, lower teeth, and tongue. The upper articularors are the upper lip, the upper teeth, and the’ whole of the soof of the mouth, We will examine all of these in some detail, and get to know them by feel, tactilely and proprioeeptively, Dut first the reader should examine what be can see of bis mouth in the mirror Open the mouth very wide and compare what the mirror shows with Fig, 22, The lips and the upper and lower teeth are obvious. The floor of the ‘mouth is mostly occupied by the tongue, which ean be put into many differ tent shapes and positions. Some people, but apparently not all, are capable of forming such a deep groove in the centre of the tongue that they can see the tip of the epigloitis like a litle yellow and rad-streaked second tongue at the very back of the mouth. The tip of the epigiotts is shown in Fig, 22 but don't be surprised if you can't see it Looking at the roof of the mouth, one can see the back of the soft palate, or velum, terminating ia the uvula, and, at dhe extreme back of the mouth, te back wall of the pharynx. We now consider the upper and lower articulstors and the zones where articulation can take place within the mouth. It will be useful to heve a Fig. 22 Soms features of the oral ewity ARTICULATION: LOCATIONS 77 ‘quick look at Figs 23, 24, und 25 before we carefully work through various artioutatio Figure 23 shows how the upper articulatory area is subdivided, first into the naniral distinction between a labial and a tectal division, the latter embracing the entire roof ol the mouth (from Latin tectum “coo!”) from the upper teeth back fo tne wvul, ‘The labial division includes an outer (exo) and aa ioner (ende-) part of the lips ‘The testal division breaks down naturally into two resions: a den- ‘alveolar region, which includes the upper tecth and the testheridge or veolar ridge, and a domal region, which covers the whole zemaining “domed” part of the roof of the mouth. Each of these two regions (dentalveolar and domal) breaks down natur- aly into two zenes, The dentalveotar region includes the dental zone, con- sisting of the upper teeth, and an alwolar zone, consisting of the whole alveolar ridge, The domal region breaks down imio the hard palate (palaiat zone) and the soft palate, of velum (velar zone). You can feel the division between these zones if you run your finger back over the roof of your ‘mouth, You will observe that the front part is quite hard and unyielding, ‘but when the finger reaches the end of the palatal and the beginning of the ‘velar zone, the roof of the mouth feels quite sof Svisions | regions LABIAL 28 prepa78 ARTICULATION: LOCATIONS axricuraTion: LocaTions 79 ie For the present itis aot necessary to study Fig. 23 in detail, or to memor~ ize the terminology introduced ia it. Use it for reference when we system- atically work backwards through the mouth feeliag out every zone and = subzone and becoming completely familiar with them through the tactile and proprioceptive sensations. ~ Figures 24 ané 25 summarize the lower articulatory area. Fig. 2$ shows Ee subdivisions of the tongue The reader should examine the tonguc in a mirror, Note is position of res, look at ts shaoe when protruded, when the tip is raised, and turned back, witen it is placed in position for (1, for (), is andso on. There are no vsibledivisions on the tongue surface, butt should. not be difficult to identify the tip or apex, and the rim, The Blade (Latin emine) is thax part of the upper surface of the tongue, exteading about 1.0 to 1.Sem. back from the apex, that usvally lies just under the alveolar ridge wwhea the tongue's at rest, andits rim touching the backs ofthe lower teeth. “The remainder of the upper surface of the tongue is the dorsunt, The front part of the dorsum (anterodorsum) practically always articulates against the roof of the mouth in the paletal zone, while the posterodorsum articu> Tater in the velar zone. It 8 thus seldom necessary to’specify:anterd= or pastero-dorsum in describing sounds: y ‘Returning to the front end of the tongue, if you turn the apex up aad somewhat back 2 certain amount of the underside of the tongue becomes visible. This underside of the tongue. lying largely beneath the blade, is called the “underblade’ or sublaminel part of the tongue. 80209) Yade-lorina = age apex “ete dot ste hr ti stecdoroum > bp yr TthDade ~ sublominef wor ig. 24 Sabdivisionsof the Tongue Seecherndle Fig. 25 Lower articulatory locations In naming the lower arculators we us latiate prefixes, labionapicon, et. tached to the-names of the upper articulatory zones or subzoves. Thus, juxtaposition of lower lip and upper teeth is labiodenta, juxtaposition of tongue surface apd soft palate is dorsovelar.and so on Figure 25 shows the profixes used in naming the lower articulators. Those in parentheses (exolabie-, endolabio~, anterodorso-, posteredorso-) are not very commonly used, but, the first two are sometimes useful and will be referred to further below. ‘We arg aow ip position to start working experimentally through the various articulatory locations : Z Longe abialandlobiodental § MEE 2. Labial articulations: = ade Jive iw SI. labinlabial Ge bilabial The bilabial stops {p] and fb] need no insrodiition, bit now close the lige ab fr (ph startup pulmonic pretsure and alow ike pe {3 A one to separate very slighty 16 that a turbulent sirsteca escapes Oroueh this i rarrow.chann, generating voiceless bilabial fncaive (9). Produce alonsl_- 7) Wl GPhisg 48. htheayoiceit(® 9 6 8 30 6 taking care te maiotain turbulent ow. Thsga wierd blgpil frearnelBl. — (Telnph) in dus gs lable be ‘on mice a Jong voiced bilabial frcatita ic fore « prolonged woited stew Susan somewhat noisy through a very aarow aperture between the lps (BBB...) Do this agai, but this time, while carefully keeping’the voiced ‘vansteam going and maintaining essentialy the sume bilbiel stricture, slowly and cazefally open up the channel between the lips very very slightly. “The fricative noise shosld disappeac: ctive (8) has beea comeried into theARTICULATION: LOEKTIONS 81 approximant [i] there betmy ao special symbol or a bilabial spprosimant, we represent itby the symbol forthe bilabial ricaive plus the diueritie[, | meaa= ing ‘more open’ ‘Now experiment with siying the veiedess and voieed bilabial fricatives (9), and (J and the veice bilabial approximant [fl between vowel: fa alle eh | . 80. ARTICULATION: LOCATIONS | [a Balle Be) {alba) [eB El.ete. Some textdookssay that, in Spanish, between vowels asin haber and beter is a bilabial freatve More commonly, however, this Spanish medial (ike Spanish medial d snd g)is sn approximant, rota frictiv, thas: fa'er] (bef. The voiceless bilabial icaive [sa common pronunciation of Japanese before, as in Met Fuji = fui rade (the exact values of the symbots [us] and [wil be vive late) Other bilsbial sounds sre the nasal [ml, and the bilabial till, We might alsa mention the semivowel [w] as in English we and wit, The articulation of ] obviously involves the lips, but itis not a pure bilabial. In place, it requires some degree of rounclng of the lips Secondly, in addition to the bilabial articulation [w] also bas a dorso-velar component: the back of the congue is raised up « Sound and will be referred to again below, under co-articulation, In Chapter 2 we discovered experimentally that the articulstion of the fictives [f] and (] requires the juxtaposition of the lower lip and upper teeth, and in Chapter 4 we discovered the corresponding labiodental approximant ‘Now we must carefully contrast bilabia/(] [J and [8] with fabicdentat (9 (end [o}. wards the velum, Ie ie thus a co-articulatod 52 bilubiol ané lubindental: Akerate aloud and then, more importantly, sileatly, between bilabial |p] ane labiedental (0: (FFG FS $ f.--and row between tilapia (P) and labiodenta [¥. [B ¥ By Pv By ...Jeand aow between bilabial [f] and labiodental fe), (B eff o Bw ‘You must by now be very clear sbout the distinction between bilabial and labiodental artiowation, Before leaving bilabial and lebiodental articala tions (For both of which the general cover term labial can often be used) we must take note of the distinction between outer (exo*) and inner (endo-) labial articulations. ‘53 Observe that it is possible to make the labial closure for [p] ard [b] in 10 ferent ways, (i) Tease th lips somewhat, adopting a kind of severe, lpped, posture so that the parts of the lips that come together aze near this ‘outer edges, and what you seein the mirror is avery thin line of lip. This type of bilabial srticulation, bringing together the outer surfices of the lips is exolabial(bi-exolabisl, to be precise). (8 Let the lips relax and push them forward somewbat, while keeping them fat (got rounded) and let the sof inner surfaces of the lips come together. Now you cin see reatively thick lips in the misros. This type of bilabial articulation, juxtaposing the inner surfaces of the lips i endolabial (pi-endolabial, 10 be pres) % In those few languages that have a bilabial will itis of a hux endolabial type. There is also a linguistically relevant contrast between bi-exolabial and bi-endolabal {p] an (b] in at east some varieties of Irish Gaelic. The [f] and [¥] of English (and of French, Russian, ete) are uaually endolabo-dental, ang tis is an mportant point to not in teaching these sounds to speakers of languages (Buch as Japanese) with no labiodentals, Learners must be explicitly cole to place the inner part of the lower lip against the edges and ower surface of the upper teeth (otherwise they are liable to place the outer sucface of che lower lip against the inner surface of the upper teeth, with bizarre restls). (eis dificult to produce an aitight closure between the lip and the teeth, since the air tends to escape throug the interstices beiween the (etl Prob- ably for this reason labjodental stops apparently do not occur in languages and the IPA provides no symbols for them, though it does provide & sym- bol, iy}, for labiodental nasal. This occurs as a variant, oF alleplone (see Chapter 19) of {mm} in such English words as criumph and nymph, It is probably reulized most frequently asa nasalized approximant sather than the usual type of nasal, which requires un sirtight oral closure Passing further into the moutt we must cake note brielly of articulations that involve the lower teeth, chat is dentélobial and deattdentel (or bidental) 54° Silently bring the wer edt ino contact with the upper fipand then start up voiceless pulmonic pressure intistion. The result isa kind of [like tieative, bata denclabial onc, not a lakiodental one, There is a0 phonetie symbol for this dentilabal fricative, whieh is not known to oveur regularly in any lane guage. Note, however, that ia the phonets literatue (particularly ie French) fone sometimes finds the term “dentilabia” meuning what is propery called lebiodental, used by persons who do not adhere 10 the convention that we striily folow, namely, thatthe prefix (ee. labio-) always refers to the lower82. ARTICULATION: LocaTroNS ARTICULATION: LocATIONS 83 articular, while the rest of the term (e8. dental always refers to the upper artisulater, Bidertol articulation hardly warrans a special experiméatal approach you simply bring theupperandlowertxthtogethec leah te eth) an bow nosiy through them, This isa biderta! ficarive—a sound thst practically unknown in languages though it does occur (as a variant of the corso-velaefcctive [x] in one sub-cialeer of the Slapoug dialect of ‘Adyghe (Cieassian) ofthe north-west Caucasus. 4. Dentalveolar articulations We come now to @ part of the mouth where we must spend.a good deal of time, since a considerable variety of articulations can be produced thers This is the dentalveoiar region. Both the tip, or apex, and the Alade of the tongue can articulate in vari- ous ways against the upper teeth, and ogainst the front and back subzones of the teeth ridge—the alveolar ridge. We thus have the possiblity of both apico- and lomino- articwations against the dental zone, anc against two parts (front and back) of the alveolar zone. These are what we must now explore, ss atly plac the apex aad rim ofthe tongue against the backs ofthe upper teeth, Slowly and intrespectvely drow the tongue buckwards, feling the alveolar sds, jus: behind the upper ee. As tz tongue sides very slowly ‘accwares over the suriace ot the aleola Acge you snoutc get an impres- sion of the shape ofthe ridge. Immediately behind the teeth ti relatively fiat then, asthe tongue slomly sides farther back, you can fel thatthe ridge is no longer fat and more oF less Roszontel, but ie beginning to eae upwards Ifyou kep on sliding te congue-tip slowly backwards you will el itpassing the most ‘ridge-like'—the most convex—partof the avenla:sdge, aad thea moving on tothe more concive arching front part of the bard palate. You have now gone beyond the alveolar ridge andl hive entered the fiont-paltal, ot prepalatal bane ofthe hac palate: Repeat this investigavon several mes unl you ar gt! fair withthe Shape of the alveolar ridge, as et by your tongue Some people have a more prominent alveolar ridge than others. Fig. 25 shows, schematically, two extremes of this kind. If you run your tongue Pee Fig. 26 Two extreme types of alveolar ridge ‘over your alveolar ridge, as you look at Fig. 25 you will be able to estimate the degree of prominence of your own alveolar ridge. The front edge of the slvoolar ridge sa the place wher the upper teeth recadeinto the gums, tat most convex part, at the point where the convexity of the riége ives way, to the concavity of the hard palate, Thus, as you can feel with ‘yous, tongue, the alveolar ridge may be considered to have two parts—a ‘rather flat front part, and a curved, convex, back part. These two parts of ‘he alveolar ridge are what we call the alveolar subzone (the front part), and ‘eack part, waish might more appropriately be called the ‘posterior alveolar” subzone). ‘We can now begin, in Experiment 56, to experiment with articulations in the deatalveolar zone, articulations made with the tip (apex) of the tongue as the lower anicalator. Look at Fig. 27 while carrying out Experiment $6. the portalveolar subzone (t ‘36. Silently place the tip (and sim) ofthe tongue against the backs of the upper teethand make astep inthis position, Tass apico-dewtal[j], the small toozh- like elacritis sack andec the J means thet i is dowtel rather theo, alvolar, (Fig. 274.) In fact Fig. 272 represents a rather uncommon variety of dental stop. articulation. More often than not in languages that have dental stops, the bad: of the tongue simultaneously makes contact with the alveolar ridge behind the teeth eee L Cz) mpice - pest alviolar Fig. 27 soe entrun anticalationsS4 ARTICULATION: LOCATIONS Now, very slowly, carefully, and introspectively draw the tip of the 1ongve backwards, As soon as the rongue-tip is completely Iree of the teeth, but still in contact with the reatively Bat part of the alveolar ridge, bold chst re. This i apiconaieotar( postion, anc then makea voissless stop from # (Fig 270) Alter producing two oF three apico-alvolar (js, and being quite clear bout the tactile and proprioceptive dlferences between alveolar) and den tal) slide the tongue-tip very slowly hack, Keeping contact with the ridge tl you can feel it touching the extreme back of the ridge, at is most conver point, just before it begins 1o merge with the eoneave palate Make a voiceless stop ftom this point. This is at apico-postalveotar (2, the subscript line (minus sige) means that it i retracted from the alveolar position. (Pig, 272) 1 dena, alveo- You should now have acquired a clear understanding of d lar aval postalveolar places of articulation. The three stops pronounced in Experiment 56 were all apical. But itis possible (o articulate stops at these same locations usiag the blade of the tongue, that is, the part of the upper immediately beiind the tig, and extending back surface of the tongue lyin stom the tip along the centre-line about { 19 1.Sem. Articulations made with the blade are called fantinal, or, in the prefixed form, lamina. (Sez Fig. 28,) 57 Place the tip of the tongue lightly azsinst the backs of the lower teeth, or better the lower gums, Keep it anchored ther, out of the way, while you ieuly bring the bade ofthe tongue into contact with the backs of the upper teeth, This sa lamino-denta contaet, und you can makea famino-dental fg st this location (Fig, 28). [Now wile keeping the tongue-tip anchored to the lower teeth, and thus oat of the way, silently bring the blide into contact with the alveolar ridge erbaps the best way to do this isto start rom the lamino-deatal position Zz» we ch Fig. 28 Lamino-dentalveolar stop articulations ASTICULATION: LOCATIONS $5 (blade against backs of upper teeth) and slide the blade back very slightly tl iisjust clear of the upper teeth. This is laino-alveolar, and you can make:a Jmino-alveolar stop [t] at this lozation, If you carelully compare apico- aveolur (with Jonino alveolar [t] you may notive shat the release sound— the little burst of noise—heard as the tongue breaks away from the ridge, tends to be ess clean-cut, litle more ‘sloppy’ in the cise ofthe lamin [1 In ict, it may sound a litle lke [ts] rather than simple (2) (Fig. 288) [Now in order to shift bask to makea lemino.postalwalar stop articulation ‘youl probably have to remove the tonsue-tip from the lower teeth. Never= theless, you must contrive (© make contact Yetween the blade (not the apex) and the postalveolir subzoae—the most convex part of the ridge, (Fig. 280) Male a lamino-postalveolar stop from ths place of srtivulation [t, You may find that the articulatory location of this amino-postalvealir stop is abowt that of the starting point (the stop sxgment) ofthe Baglish ffveace [1] 55 ja church. (Or aftricates see Chap. 6) ‘You will have observed that the IPA supplies no special symbols, iacrtics for the laminal [Js This is not 2 serious disadvantage, since) distinction between apical and laminal [sounds is rarein languages, Now we must experiment with a series of fricativee in the dentalveolar 58 Sletly se the pnt of the ong nt spe and vi jum toch he cuting cages ofthe upper tech, Helding this rteulton sat ep pulmoris esr itkiston, and esltntegesive air-sea and the rst sould Be an opicodera rari (8, exacy or very neatly the Enlist of sin ‘Note that typically the English [0] isa rather wide channel fricative, the blade a the tongue i eased and eather fot aod the cin ofthe tongue either touches or ust behind the ees of tae upper tcth—the tongue does aot svemallyproirade betveen the eth for this sound Fi. 3) ‘Now silenty meme the onguotip avery tl ad tar i lites that the eds ofthe tongues sake: contac with the alwolr td, leaing 2 sry nacrow cent channel. This ste potion fran aplce-abeolr fia tbe it yp uo piven gma dca yo villa he tis Jing’ kind of [s}-sound, (Fig. 29.) Retracting the tongue-tip still further, til it Jnatiha mania ooarex earcan back Gf Psion ge Jou cta ealaa Ace pestalolar rcv ype of areulston. An egress airsteam wil now geverate a [Jsoand, Ske a kiad of sh a in Englah sop, bat by ne tmeuns the commoner vaity ofthis (wich is lamina the so sound of Norh Geman, sné the Rassan sh mare rater typically of this | apico-postalveolar type. (Fig. 29¢.) «86 ARTICULATION: LOCATIONS a > . Fig, 29 Apico-dentalveolacfricative arteulations “The hain things achieved in 58 wil bea farther fariiaity with the three major dentalvcolar zoxes of articulation: dental, alveolar and poitalveolar. [Experiment $9 investigates laminal articulation a two of these locations. 59 Let the ape and rim of the tongue le Hghtly against the backs of the lower teeth, Press the sides ofthe ‘ongue-blace up against the alveolar ridge, leav- ing a very narrow channel in the crates, Aa egressive air-stream through this ‘narrow chanel generates a typical laminoabveolar [s-type- jricative. (Fig, 300;) ‘Now retrict the tongue a tte detaching the dp from the lower tect, aid form a narrow erticulatory channel between the blade and the most con- vyer back part of the ridze—a lominoportatvenlar frcative, a kind of (J) (Fig. 308) Sileaty, and slowly, alternate (s/f till you are sure you can feel the éiffer cence between them. | Ne have now covered the major types of stop ond fricative artioulation ia | de derzalveolar regon. You can use the knowiedge acquired from Expat ‘ments SE-D7o tiake a silent, introspective, analysis of some sounds of your own language: for example, are your [4 {n) [] apical, or laininal: dental, of alveolar? And if your language has a tiled or capped) or [where is it artioulated? If you have (3}andlor [Jpsounds—in your languags, are they Fig. 30 Lamino-dentalveclar fricaive articulations urea eset eae aera ARTICULATION: LOCATIONS ST apices or iamino-, dental, alveolar, ot posialveolar, etc.? If your native lan~ guage & Polish, what are the articulatory differences berween 2,32, 22 IF your language is Russian, now does the primary articulation of x] and [4] in {tot] ‘that’, [da] ‘yes’ compare with that of palatalized [v] in [Vows] ‘aunt’ and [dads] ‘uncle? (On palatalization see Chapter 6) If your languaee is “Arabic compare the dentalveolar articulations of plein (t] and [asin [sn] ‘fg’ [sin ‘the letter sane of “emphatic [t] and fs] asin [tia] ‘mud and fia ‘China’. You may find the dentalveolar articulations much the same, the ierence between the plain and emphatic consonants depending chiefly on the back ofthe tongue (See Chap. 6 Sect. 2). Ifyour language is English, lent compare the (ane [¢}sounds in (@) | ‘eight wide, (®) eighth width, (6) ry dry. Can you observe differences between | them? And if ou use a tap in the middle of better (American), or in the | riddle of very (British) sit gpico- oF famino-, dental, alveolar, or postalveo~ Jar? How do you articulate the r in red? In British types of English it is likely tobe a variety of apico-postalvolar approximant (or, mote rarely, rictive) {g). Ta American types of English you may find that che tongue-tp is rather far behind the postalveolar location and that the whole body of the toayue jis bunched up, coming rather near the velar articulatoty zone. There is a considerable amount of variation—both regional and: personal—in the promunciation’of English rs, se your own r may aot exactly correspond to aay described bere. By silently isolating your r anc introspecting about it sce ityod ean discover how itis arieulate. “We can now review the principal types of dentalveolar articulation and tie IPA Symbols lied to represent them: (Hope Gos se’ dental f(g) alvolarltd. a) ey cece ebanne) (0}(3) - ateclar(aommaly mmo). (SICI pies puspolatae oat stctatolarapico-or kino), IU l_.f “PC? ~Postonela fostines exproximante: 20 special S760) "which is most gominonly approximant, but may ‘not much diference between the fricative apice-postalveolar [3] as in + Pols sig] Russian he) lived” and the dpico-postalveolarfricatve(3] (es often ia British Englisn dry (dsl): but there it some difference. Ta (3, but no in [gj there is a slight spooa-shaped hollow in the centre of the tongue just behind the blade, which is absent in (5) Approsimants.of other dentalveolar types can be represented ty using AG)88 aricutaTion: Locan | the ‘opening? diacritics thus (Q [el represent apico-dental and lamino- .> | alveolar approximants. } srill'and tap: [1] and (4). | trsetaesinen Get aon orpoebwos k | deter fricativer: (dental alveoli, or postalveola) (lj (__ rave cecal avn oF poster vebalemingpripatal articulation ™ 5, Retroflex and palatal articulations Having explored the dentalveolar region of articulation we now proceed fursher buck into the mouth. First, immediately behind the postilvweolar subzone we come to sublamino-prepaletal asticulation, oti wise known as retrofiex 60 Silently place the rip of he tongue aguist the postaveolar part of the alveolar ridge. Now slide it bck beyoad the postalveolar subzane, to whers the hard pate ests to be ennvex. As yous tongue enters this concave zone it ispointing almost straight up, an the underblade,orsubamina, bests to spake contact withthe prepaltal arch. This is seflamuno-prepalatlaticla. os, of sine the apex ofthe tongusis vituily curled backwards, reopen Produce a series of reroiex sounds from this pice of aticultion ps [erates fl, apron 9 tera apprxinene ane) ( nasa. See Fie 30a) / - re Then i as 8 7Paraculatcd in the reteonex zone, symbol [To produce his start withthe conguo-tip cried wel! back, thea leit shoot forwards nd downwards, lightly striking the prepalstal arch, just behind the alveclar Tide, cn the way down, (See Fig. 315) Retrofer consonants are particularly charucteristie of languages of Incis, They are often quite strongly retrofex in the Dravidian languages, but (cad to be somewhat less so in the Indic langvages of northern India, In Hindi, for example, [1] and [das in [in] ‘tin’ and (ord “buck ‘be mich further back than apico-postalveo "may not j (tl, However, they contrast with apico-dental stops (Ql [4] a8 in [fen] ‘three’ [do:] ‘two’, and the more retraced stops are normally culled ‘rere’ and writen [tJ and [4]. Hindi {1 however, eceurting in such words as [gioga) horse flstka] ‘boy’ is a truly retrofex flap. ‘Next, we must examine dorsal articulations in the prepalatal zone—that ARTICULATION: LOCATIONS 89 Fig. 31, Retrflec artiewlations (3) stop (6) Map, is npn-retrotlex articulations invo.ving juxtaposition of the dorsal surface of the tongue and the bard palate, The hard galate is divided into a font half—the upward-backwsrds sloping prepalatal arch—and a rear half-—the hhigh vault of the hard palate back to the line of division between the hard and soft palate: this is the palate! subzone proper (see Fig, 22). We start swith dorso-palecal (not prepalatal) articulation (sec Fig. 32). G61 Silently form and hold the articulation for 2 vowel [] asin see. Introspesting bout the tactile and proprioceptive sensations of sent [i] note bow the tongue is bunched up ia the froat of the mouth. The vowel [J, of course, sa orso-palael approximant [Now push the central part of che tongue upwards narrowing the artiula- tion channel of seat i] until it disappears altogether as the centre of the tongue makes contact with the highest part of the hard palate. (See Fig. 31) Fro this position make « voiceless dorso-pslatl stop [e] and a vsiced dorso-palatal stop Foon and hold the stop of {eli hold the tongue dorsan: in firm contact With the hard palate, but nothing else: there must be absolutely n0 contact ‘between the anterodossum: aad che prepalatal subzone, or between the blade and the postalveolar subsone,,Now, zlease the centre of the tongue very slighty, so that you forsn a ver} narrow dorso-palatal articulation channel: oreo 32 Dorso-palatal articulation90 anmcvterion: LOCATIONS {ovate an agressive airstream which ought to become strongly turbulent a¢ is Rows through this narrow channel, forming adocso-palatal voces rae tie i}. Wealready eached a sound very much ike(¢] by devocing (jin Exp. 29, This [gis the sound of oh in German ich Now add voice, but be sure that you have a really narrow frcative ee channel, o that when you voice (6 itbecomes a voioaé dorsapalztal Fcative Gand aot the approximant {The symbol j] wseful in order to gash BED He reaive (and the epprosiman! or semevowe! = / cxempliied by the yin Engish yer. ‘SURE Wis normaly We GORA fan of the tongue (and, specifically, the anterodorsal. part) that articulates against the hard palate, we commoaly érop the prefix and talk simply of palatal articulation, The palatal stops (¢] and j] are not very common in languages, but they ars sometimes exempl fed by the 1y and gy of Hungarian ('a:col] ‘veil’ and [mayor] “Huagarian’ ey are pethaps more often pronounced as prepalatal age) 2 addition tothe palatal stops [¢ a [J the fisatives @) and 6} approximant (], and the semivowe:{] we Gan havea palatal nasal [aid lateral ff. ——— a “The palatal nasal) is tracitionally said to be the pronunciation of the French gr in campagne, the Halian gr in aga, the Spanish 7 in matana, However, in these languages ‘he [ns not always pronounced as.a geauine palatal nasal, It may, instead, be pronounced as an apico-alveolar,lamino- postalveolas or amino-prepalatal nasal followed bya palatal emivewsel (2) Palatal (fis traditionally said 1 be the pronunciation of Italian gin el, Castilian Spanish iin pollo. Again, however, it isnot always pronounced as a genuine palatal [in these languages, but ather as I]. In Latin-American varieties of Spanish ils commonly pronounced as (6, but sometimes at a semivowel [] ora frcatve (5) ‘As a-matter of principle, the student of phonetics should ekperiment with producing genuine darso-palatal in] and (0), with contact only between the dorsal surfce of the tongue andthe Righ vault ofthe hard palate Bat [HOGI To Contact Betweed ihe apex OF VU Of thé Tongue ane the aveolar nde of prepalatal arch. Tt may be useful to compare geauize [ni] and {<] with the sequences {nj} and {i} that ovour in English. Thus compare English onion (aojen] with 2, ARTICULATION: LOCATIONS. 91 ‘Before leaving the palatal zone we must mention the possibility of articu- lation in the front part of the zone—the prepalatal subzone. Experiment 62 introduces this subzone. 62 Produce a voiceless pulaal fricative ()—vif necessary develop it from [i] by Lamino-postsveote L} prlete-alecstor eo fy. ce € Lamine-pfepalati é Sa aie ni pict ¥ iz) orn a i vice) 1m Fig. 23 Some pestelveolar and (pre)palatal articulations agpeet French agneau [apo], Talian agnello (aneilo| ‘lamb’, Spanish aro [apo] “year's English biliards (bilj(e}dz, Italian bigletto [bideto] ‘ticker’, Spanish Diller fifa} Billard’, te ”92 ARTICULATION: LecarioNs (33a, and a dorro-palacal fg] ($32). Ik may be user for the reader (0 produce these four types of fricative, silently and sloud, while looking at the figare It should be noied that in the older terminology of the IPA ico- oF lamino-postalveolar frieatives of the type [] and [5] are called “palato-alveolar’, while laminc- or dorso-prepalatal tricatives of the type {[c] and [2] are still called ‘alveolo-palatal” This terminology is not \ recommended, since itis consistent with the stiet principle of using the | prised term co designate the luwer smiculstor (as in labiowentl, ‘ eolat) 6. Velar and uvular articulations We go on now to dorso-velar articulation—that is, articulation between the back of the tongue and the soft palite This isthe place of articulation of typical [k] and (3] sounds, Sq in 63 we begin with those, ish car; hold it silently foes momen 63. Foon theartevlaion for a[kJasin Beg 3 thea silently release it, Repeat this Several times, introspecting about weat it feels like. Contest this darso-velar top {k] with a dorso-palatl stop Te), Note how the body of the tongue is thrust wel forward for [e, so that contact can be made with the highest part of the hard palate. For [sl however, the body of die congue, though clearly ferther tuck than for (@), doesn't feet particularly strongly drawn Back, (Once you are satsted you caa feel the midvelar (mid-soft pala) contact for fe] boll Hae tomgue in that the tongue either oeward ox back, open up a very small chanel between the tongue and the sot palste. If you propel an earessive air-stream tive, blow) through that narrow dorse-velar channel you should bear the sound of the vwieeless dorso-relarjricaive (x). Make sure it is velar. Some people tend to Jet the tongue slp back and make u wuulorfeicatve fz, We will come co that x isrequired isa pussly dorso-veta (x, J, then do it again, switehing on voice, but J, where [yl isthe symbol be a voiced sition and, while king ener motto shift in Exp. 65, bat for the mom Make a prolonged (x x xxx dng no other change: [Xx ¥¥ dorso-velsefrcative Starting fiom the voizee (uh Say a prolonged [y yy. .], noting that its tuly fricative, Le there is a ficative hiss-noise superimposed upon the smooth sound of voice Now while saying prolonged fy yy. very slow. and very slighly, open up the 1 Hicetive [y] develop a veler spprosismant ARTICULATION: LOCATIONS 93 articulatory channel, just to the point where the hiis-noise of turbulent aire flow ceases:(y ¥ yu wil. This isthe velar approximant. t] Now that you are clear about dorso-velar (or simply velur, as we often calli) articulation you can note that there is also a velar nasel (9). This is the nasal that occurs at the end of lung (lag) in Gnglish. Note that [9] if you isolate [p] you will nd it quite easy to put a vowel after it, and say [pa] [oil ete In English, and many other languages (ic is very noticeable in French, for example), (kJ and [g]are purely velar only before such vowels as {a} and fu, as in English car, ewan, and cool, goo, Before front, or palatal, vows ike never cocurs at the beginning of a syllable in English. Howeve [i] as in key, geese, however, the articulation of fk] {g] is shifted forward a liude, If you artieulate these words silently, then isolate the (kya) of each and compare the ones that occur before [i] withthe others, the forward shift ‘of tongne-position befor [i] will be obvious. But note that even though the tongue is somewhat advanced in key and geese, it does not go nearly as Car forward as the palatal position of fe] and [}}. Before fi] the English velar stops are somewhat ‘Ironted’ (ie. slightly advanced from the fully velar location that they have in such a werd as car}, but they are not fully palatalized, and certainly Go aot become palatals, ‘The next, and the last, ariealatory location in the oral area is dorso- wvular, or simply wvular. We investigate souncs made at this location in Experiment 64. 64 Make » (K.closure and then, silenty, or almost silently, make a prolonged caries of faint felype sounds [RR RY J ete. while slowly slidiug dhe tongue back and cov 3s far as you can. You wil ead up making a stop at the very furthest buck part ofthe soft palate. The extreme back of the tongue isin contact with the uvela and the exireme back of the vehum (oft pate) yeu let a litle pressure build up behind this extreme back closure, thea felease the closure, you will hear a wiular stop (gh. If you repent this cexperiment—a series of faint ops of the [k* kk. Iype steadily moving ‘back from the velar to the unuar positions—you will observe thst the sound of the litle burst of noise occurring on the release of ench’stop goes dows in pitch by about an ective over the whole range. ‘Anotier observation you may make is that the release of velar (K] rela: tively ‘clean’, while that of [q] is more ‘sloppy’. This is because tbe convex toague-surfuce can break away from the whole contact-area of the concive ‘war surface almost instantansously but separation from the more Henible94 aRricutaTion: LocaTIONS and frvegular surface of the extreme back of the velum, including the wvulay is less inscantancous less clean-cut Hioving produced the voiceless uvular siop {5] (not a dificult sound) and having copeatec it several times, you might try to produce the cocespond+ ing voiced sourd—the voiced uvular stop {¢). This is much more‘dffcul, because the tongue is so far back in the articulation of a uvulay, chat the space between the oral closure and the glottis is very smal: consequent as the air used in generating voice flows upwards through the glottis the essen- tial pressure-difference across the slots is abolished almost immediately andvoisingceates ‘Sir wralay arcilations a te Hisatves and [uy] the approxima ‘te nasal and he til [), They areinvetigated in Experiment 65.) eens ne 65 Fora Be desure fora aval stop (a and hold i. Now, while bolding that, uwular ariculaton posture, open up a small central channel. aid propel an egresive airstream through it. The result should be the veizelo ular Sica ‘Another way of approaching 8 san witha vel {x] and thea move progressnely backwards as you did with & .. ¢] in Exp. 64 In this case rake a var iative (x. and while keeping the fictive rote going siw!y slde yourtongue Sack and down, til youhave arrived as far bag ae youcaa 89 3. Once ass, notice os you do this thatthe pte of the tia ise: gecedown by neatly an octave Having produced [3], prolong it and then switch on voice [2 44 « # w x} The results the voiced wvularfricaive (ef you now make a prolonged (x) ten. while Keening the veion ening and the me genaral tongue posture vary lightly widen the artiulstony share and you wil convert the wvslar aie [oto the war approximant) io ‘The uvular nasal [fy] should give 20 trouble, since it can easly be veached, | lke (@] and [a by siding beck from! the corespondine velar, (r]. There | remains the urular tril fe. [Fou can gergle, and most people ena, then ou can produce & valor teil It is oaly neoery to reducs the amount of water used in gargling, finally using only Séliva, to pass from gargle 16 [k In adcition, as we pointed out in the las chapte its casie 10 predce al tis with a pewerful voiceless ai-ream, so you might try that ‘300 espsrneat wih wruar sounds you may note tat the ra oem iSO ate hain ARTICULATION: LOCATIONS 95 simply because the uvula, being small sné flexste, tends to be thrown into vibration by the airstream cf the fricative, You may be able to avoid it by trying to got a strong feeling of tenseness into the rear of your soft palate and the back of the tongue as you produce [i] and [v] An imiporant | feweane between wvulac friative [u] ad@ WT a] & that whereas, the | ‘tongue-back is relatively flat or coavex for {s).a decp longitudinal grdtve is |” formed in the back ofthe tongue for the wl x} and the uvula wibrates By © the groove, "WE hive now covered all the major articulatory locations within the oral area. You might find it useful to run through them again. In particular, it would be useful to compare the three major tectal articulatory locations: palatal, velay, wrular, By contrasting series of sounds such 2s [ek] (a. () [8] (2b Fl [a] [s] one becomes more clearly aware of the differences between ‘palacal, velar, and urular articulations, Finally, we must turn our attention to articulations in the pharyngzo-laryngeal area: these are articulations in the pharynx and larynx 7.. Pharyngal and glottal articulations ‘Articulations ia the pharynx are ealled pharyngeal, or pharyngak those performed ia the larynx are gloital (not to be confused with ‘glottalic, which is the name of a type of initiation). ‘The pharynx is the cavity behind the mouth, rumning from the back of the nose and the ‘nasal port’ down to the laryns, Phargngal aztievlasions ccan be made both ia the part of the pharynx just detind me mouth! (ihe oropharynx) avd in the lower part of the pharynx, immediately above the Jaryax and rherging with it (he laryngepharyr). ‘Here we will deal with only two types of erticulatiom in the pheryax: one a rather generalized sphincterie compression of the oropharymx—waich we shall call pharynga: the other involving the epislottis, which we skal cal epigiotial, Taese.are dealt with in Experiments 66 and 67 respectively. (66 The bist way to duce the pharyrigea! compression that we want to achieve Js to activate what is called the ‘gag" reflex. Unless the reader is encegticaally insensitive he can do this by sticking a linger Into" his mouth so that it ‘ouehes, or merely approackes, the uvula. The extreme convulsive costf2c~ tion of the pharynx that this inducrs is the starting point from which 10 develop'a milder, ‘less intense, contraction of the pharynx. If you send a wy ond96 ARTICULATION: LOCATIONS voiceless and then a voiced airstream through the eonlracted pharynx you will produce the pharyngal souncs, voiceless h|and voiced |S, TGS ive 4 vey Gotition variety of the sounds oF the Arabic letters © {e) and & Cain). They are ten dascrbed inthe lterture as ‘pharyngal | realty Uiey are more ellen approximants. Note wat ne | lk, hss although « noticeable _/ fricatives’, but i voiced sound [S] hss a0 turbulent, ii tiem tia The pharyngal approximants [h] and [5] are very characteristic of most varictes of Abie and of a few other languages, including Somali, Berber, and some varieties of Oriental Hebrew (in modern Israzli Hebrew, the ‘ancient Hebrew [hs replaced bya velar or uvula [x] or fzh-and she {5} by a lotta stop [). In a few varies of Arabic and Oriental Hebrew the “aia (2 appears to be pronounced, notasa simple pha:yngal constriction, but as a complete closure, formed by folding the epigloltis back. as in the act of swallowing, CERT epigorl seop pa i the Jnest (1996) IPA chact by (7) 6F Start by swallowing several times, and introspesting about what is going on. In the middle of the process of swallowing there isa feeling of complete closure when the epiglottis folds dovn over the larynx to prevent food from eenteringit, Hold that stop position for a moment, Do that several times, then {ty to lank that moment of elosuse by a vowel, oe example a]. thus saying fa # 9] (a 7 af Tals is an epiglowef cop. Once you can say [a? a] Lally easly, compare t with simpleglottal sop: [a a} [a 23], Notice how glottal stopla a) is usta simple momentary ‘holding of breath’ simple hiatus Between the Rankine vous that hardly ffees their quality ‘The epigiotial stop, however, in [a ? a] ig not only a very steong stop, b also affects the quality of the faning vowels, which acquire a peculiar ‘squevzal” quality (pharyngalization) Enislonal suop [@\ seems to ozcur not only in varieties of Arabic and Hebcew es indicated above, but also in several languages of the Caucasus, for example in Chechen, where epiglotal (2 eonteats with glut! (7). The cxistence of epiglotaimwrep Ti Arabicind Hebrew was demonstrated instrumentally by Laufer and Condax (1979). Gotalasticulation occurs, of cours, ia the laryax, by the justaposition ‘of the voral cords, We have already seen many examples of gioral stop FP, articulated by makinga complete closure between the oral cords—closing the glottis Glottal stop is sometimes used in English befoce a strongly ARTICULATION: LOCATIONS. 97 stressed initial vowel, as in “AR! (2a). It may accompany final voiceess stops, producing co-articulated glottal + oral stops in many types of Eng lish, hough more frequently ia the USA than in Bsitain, perhaps, a8 i exp, fat, hack, ew pronounced (ka) faz], [ee, where the ligarue. | indicates that glottal stop and oral stop are simultzneous. In some Eaalish and Scottish dialects glottal stop may totally replace an intervocalie[f], thus Cockney [bu?s], Glasgow [ba?ar] for butter’ “The Sound [h, as in hot, is often described asa wieeless glottal jriative, singe ike other consonants it occupies the marsinal (initial) position in the syllable, rather than the central position in the syllable appropriate to 9 ‘vowel. Ip cerms of is articulation, however, i¢ might also be regarded as & voiceless vowel of about the same quality asthe voiced vowe! that follows, “The corresponding voiced loutal fricati, (8) is essentially a brief span of brccthy voice oF whispery voice functioning as a censcnant, Voiced (8] may English in such words as Ala! and perhaps. (Re the end of Chapiers 2 and 3 we mentioned the prosodic features that) jrelae to initiation and t@ phonstion respectively, oamely sess and pitch phenomena, The prosodic feature that is related to articulation isthe durae ‘ion, of length (also known as quantity) of sounds. Clearly all maintainable articulations can be held for a shorter or longer time. We will deal with the guration of sounds in Chapter 9, ‘Prosocie Features’, {-—————canhery octicobel'a ons Col: 'WG GP st? 6 Co-articulation and Sequences We completed ovr survey of articulatory stricture types (stop, til, icative, ete.) and articulatory locations (bilabial, Inbiodental, apico-dental, etc) in ‘Chapter 5. However, there ae stl somethings that remain to be said about amtcalation,and we discuss them bere under the headings of covartialation and sequences. Co-articulation, All the sounds we have dealt with so far (with one excep~ tion, fw) have a single place of articulation: thus, [p] is bilabial, fois dome sopdlata, [is pharynga, ete, But its pectecty possible for articulation 10 ‘occur at two different pléces simultaneously. As we sew in Chapter 5, the semivowel [w] involves an approximation and rounding of the lips, and , consequently is bilabial, but, atthe seme time, the buck of the tongue is raised towards the velum, so that there is @ simultaneous dorzo-velar! articulation, ‘We call such sinmltaneous articulation at two different locations co- articulation, Covarticulaion isan essential feature of sora sounds such a8 {iw}, but it aso occurs ‘accie xt were inthe elose transition trom ‘one sonsonant 1 another. In the English word play for example, as usvally prononced alii intzospection will show that the tongue-tip makes con~ tact with the alveotat'nége for the apico-alveolarlatefl approximant (1] while the lips are still closed for the bilabial stop [p]. There is thus 2 short period of overlapping artieulation—and this is a period of transitory caktcuiation, (On the other hand, some consonants, ike the (w]already mentioned, are covartiguated in their owm right, asit were, and these are the subject of the present section. “There are two types of co-articulation: (9 coordinate, or double} articulation, and (if) secondary articulation.( Co-ordinate ) 100 CO-AKHICULATION AND srQUENCIS Co orbrealatlon (double) ote (ump uey s (deuble) (we, LPR] 1, Co-ordinate, or double, articulation In double articulation there are to simultancous articulations of the sume sank-—that is (0 say, of the same degree of stricture, hence: stop + sto ricative + frieative, o approximant + approxennt ~ 68 slyp + stop, Simultaneously plase the back of the tongue against the soft palate (form a velar stop, (kD, and elose the lips (Torr a bikibial step, [p) Hold this double acticulation [a] or [kp] in whish the ligatues |_| snows that the erticulations are simultancous, build up litle pulmonic pressure behing 1H,and thea release i into a Vowe, sty a), ths ipka}. This bdabiad + velar stop. Nowputa vowelbeforeardafter[rklandsay ppka],makingsurethatiberwo articulotions are as nearly as possible coterminous, ie begin and end at the same time Experiment withothereo-articulatedstops (by), [gh tp (bd) We label co-ordinate, or double, articulations by means of either a. plus alyphen: thus {pk] or [kp] = bilabial + velar or bilabiat-velar. Double articulated bilabial + velar sounds are sometimes referted to 2s labio-velar’ of ‘labiovelar’. This usage should be avoided, since the latinate prefixes labion, apico- ete, refer sirietly to lower articulators and should not be used to fer to the whole conjunet, lower ~ upper, articulation © frouive+frzatve Sinalancoudy form teariltor stewes Orff and [sardehcn ty [Now FE ah You wil End iio Ke eet to generate cave trtulenc:simlineouy a both sriturelocations Tis i pola way ceordnse facine + ae ep-aieuaton ie vitally Dobe Waa enrexinant + agprocinont Tistype f co-ordinate (ote) artctation inquiteeay. You te ate cen one expe (I yousxperimentaly lengthen [¥] (whic tr nto the vowel fu, you aa then Sesompooe it ini ts two composca Fist while erpingte [ow] going Sowty and deberatelyuarouod telips bat make wo ofr cnge, Wats tafe ou dla atalino lar have removed the bilabial articulation, is approximant, Thus the vowel [a] isa double articulation (@ s0-ordinate co: aniculatioa) consisting of a rounded bial and 2 velar approximant, And te ultrashort (uaa is the semivawel (w] is. bilebial-vlar semivowel, Now, produce & palatal approximant {] and wile earefuly maintaining slowly and carefully adda labial result will be a bilabiel-patera! the dorso-pulatal articulation, ariculation—round the lips andey: See om Choe it) OTe CO-ARTICULATION AND SEQUENCES 101 Joproximant, the vowel ly} asin French kane, Make it ultrashort, and you ave the iibia-palatalsemivowel[] Tosh ofthese double-articulatedsemivewnds geosr in French, for example in oii} "yes! and had (yt) "eight, Lous Iwi} "Lois ancl “i 2. Primary and secondary articulation “The second type of soaiculation involves two simultaneous artiealations 1 iene seer Basia roa oe Sih jicatne—preSimon—feoha:Secbalary rata a abut the rank of appiSafmant. Fo example, an (with a simultaneous site ton Of eff o€ Bo pl cia we mune Vasey ccelaiganby siicture types is srteulations a the ‘aa fi] wih secondary palatal articulation ‘means of adjectives in The only departures from the basic ranking concern nayalized and pha rmgalized sounds, and lateral approximants, Any nasal articulation is regarded as secondary to any accompanying oral articulation. For a we eal this a palatalized 0 crampk, even though the oro-aasal orifice (the “nasal port) may be smaller (ané therefore apparently of higher rank) than the wide oral articulatory channel of an open vowel like [%] or 3], nevertheless the oral antcalstion is regarded as primary, We call these (and all vowels With simultaneous alow through the nose) aazalized wwels—and never “oralized nasal Again all pharyngal articulations are regarded as secondary to any simul- taneous oral vowel seticulation, ane, fnslly, any approximant-type artiea- Jacioa accompanying a lateral approximant & always reyarded as secondary tothe lateral, Thus, an alveolar lateral appreximant {1} witha smaltaneous palatal approximent (\iJ-type) co-articulation would always be called a ppelaialzed alveolar lateral approximant (and never an ‘alvealor-hteralized laa appivrinau’ The principal types of secondary articulation ore (labislization), polaialized {palatulization), velarized (velarization), and pharyngalized (pharyngalization). ‘Secondary articulations are transeribed by diacritic marks: (¥) for lab ized (labialization, ) for polarized (palatalization),[-] trough the letter, eg. i. forvelarized (velarization), and either the same [-] or a small raised ] for pharyngalized (pbaryagalization). labialized[Now closely round the lips, simultaneously form a kj-closure, build up a slight oressure bekind the stop, and thes say fk aa] simultaneously releas- ving the labial and velar strictures. Now experiment with other labialized sounds: [1] gr) [er] s"] fer}, ate —initally, medially, and final, eg, (5° aa] (a 5¥ 3] [oo] et. In every case note that 2 uuly labialized consonant must ‘have simultaneous labialization. Thus the transcription [s¥ aa] represents an intl] pronounced though cocely rounded ips. nt a sequence 1» Um mtich woud be tansribd sw a] oe Je is sometimes difticult to decide whether we are. ‘dealing with a sequence of eonsonants followed (or preceded) by [w], oF witha single cou oasonant, with scondary Bbilzaoa, The reason, of couse Fat though in theory secondary articulation Fequires total simultaneity of arvculadon, in fact the timing-lations between two artaulatons are lnGinitely variable We indicat this state of afainssuhly in Fp. 2 In Fig. 3 the upper line(6) epreseat the primary arcalatory scare, sy 0 being made, eld, and thea released (eg, into a vowel). Tae loner line represents the degree of stricture of the: ‘secondary articulation, say, lip- approximation and rounding for labialization. The solid labial articulation lines represent unequivocal simuitoncous(a) and sucessve(o)arieutations Thersisne doabs that () iepreenis co-ertcuiaton secondiy atealaion, ‘whereas (b) represents a labial semivowel or approximant (w] following the [k]. The broken lines represcatswo of the innit numberof posibe states aMfirsbetwoca these extremes, unerining the fact that the status ofa Possible secondary articulation is often ambiguous, In a seal ambiguous case of this iyre we may Sally decide whether we art dealing vith» soe ondary aniculation, say lBialized (kr), or a sequence of distinct sounds [kw], on the basis 0 rom the general phonological stricture of “the language rater than on purely phorete grounds But the aadeor a Dhonetics must beable to dstnguth between the unambiguous extreme cases Te would be a good idea, while looking at Fig. 34, to experiment with git SS. 10 tatemnicoatiini CO-ARTICULATION AND SEQUENCES 103 produciig series of secondary articulations with diferent time-relations. = to the primary articulation, like those indicated in the figure. yw ‘We now coatinue working through the remaining types of secondary articulation, staring with polatalisetion in Experiment 71, 71 Pinlnton wren eb he gba fr hepinany “A articulation. To produce a palatalized sound onc must hold the tooyue in a [ ] (J ‘erso-pusia! position (anproximatey as fr {) while mating the primacy arteulation, Experiment Set with psi of voiced ficatives suchas pain y}, oe palataliced i. V1 wih scultancous 6) lan. netatalves Er. ie Go 1 Smaleaneoes Gh [ Now plac thse corsénaatsberwcen vowes in et sayfa xia. (or(a.a jaa)ard(@avaal {2az aa} simultaneousy: ths (vaya gaz a, Nett iperinent wt plain [plain nk peated, pallid if that [1 is, simultaneous (or [a] and ff. A} Bo Vole fccatvs can also be palatal, btn tis can ofeobng EPS 6.9) ‘Secondary articulationis not{i, but vaeefes thus [0 [] -EBI+ aD. Finally, experiment with simultaneous secondary palétalization added to \ stops Say fea ia aa} (42 3 paaa)then combinethem: [a2 plas a} and sowith faablas][aat calfaad aalfaa¥ aaliaagaa). % Since palatalization consists in raising the anterodorsim of the tongue towards’ the hard. palate, it is not surprising that sometimes, when theprimary aiticulation also involves the tongue, the secondary pulatalization slightly changes the primary articulation. Russian, which isa language with 2 whole serie: of palatalized consonants contrasting with single or none palatalizee ones, provides examples The Russian [| and fd, asin (tot that? and (Jal ‘yes’, are normally apwo-dental. However, theit palatalized counterparts [8] or (2!) as in (Yora] ‘aunt’, (Dadis] Yoncle’ are usually Lamino-alwolar or even Ismino-postalveolat. The upward thrust of the anterodorsumn of the tongue tefids (0 retract and depress the apex, with the result indicated. Again, in Russian and many other languages, including Faglish, pelatalization of dorso-velar (kc, as in English kay (oi merely lilt the anterodorsum of the tongue, but it also shifts the front edge relum contact slightly forwards, We can now turn 10 velarization. Velarized sounds have a secondary does not of the area of wong ‘r-upprcximant-type articulation i.e. in addition to the primary articulation he back of the tongue is raised up towards the soft palate 72 One-way to achiovea velar approsizun sto produce a fy} Gitmecesacy, go sight back to experiment (63)and start fiom [x or from [thee [x then [yD From fécative[y| open up the ariulaion channel and you have the velar approxinast( Another approach isto stat from a vowel ofthe type (ul (ore oF Issa a Engl who hus) aud, while protoning the Vowel without ckanging it inary way, slowly and caceflly waraend the ips. The result should be 2 velar approximant [i]. Now say « prolonged IL), then prolonged (ai. }. Now both seultancously (H. you compare this velarized [], and the palutalized (B) you produeed in Experiment 71, with the J of English fee! und leaf, you will probably find ‘that these two English & are somewhat similar 1p the Velarized and palatal ied (s you have been experimenting with, thus [fi] and [itl]. Note, howe ever that these 1wo fs—clear P occurring before vowels and often before [) in million {muljon], and “dark # occucring after vowels and finally, as in ‘middie no very strongly moditied. Note, also, that there are consider- able difterenaes in the quality of sounds in different English accents. Most Scots. for example, use rather ‘dark’ 5 in all positions, seying Pill, for example for ica Trish speakers may use very clear in all positions, saying [fc] feet, ete, Americans tend to have darker (s than speakers of British RP (‘received pronunciation’), and s0 on 73 Apply wearizaton to some other sounds, for example (e}—say (and (et) wey simltameously thus 2. .], Then combine velurization with voiceless fra tives, fOr example [s}, thus [¥ 3 ...J. Wea the primary articulation is a voiceless tricutive, as here, you may be able to develop the velarized form fom simultaneous pronunciation of [s] and Ix] Having achieved a co- articulated [st], iryto relax the velar artieuation a ite, so that ite no long Fricative Thi should give fe). Now try te combioe vlarization with stops, eg. (¢}. Say i aguin, several ‘mes, aloud and silestly, concentrating ypoa what it feels like to have the beck ofthe tongue bunched up in the velarizing position. Now, transfer thae feeling to silent {a} or [da], and finally sey [4] aloud, Do likewise with [¢ _-Note that the 1996 chart of the International Phonetic Alphabet shows the disenitic [Jasan alternative mark of vesevation. Velurized [2 and s"]are one variety ofthe Arabic ‘smphatic"corson= ants > b >, which conirast with plain 2.5 fl], (0, and fs] as in [dem] “hurt vs [dim] ‘perpetuate, [sn] mud’ os fin] gs’ and (se Summer vs [bel] Sword’, Moreover, note that velarized (jis the I-sound of Abie . The mare [] over the ft] indicates the wesk or unstressed element of the diphthong. A decrescendo dipithong like this is ‘often called a ‘falling’ diphthong because of the fact that the stress falls, away ftom a peak near the beginaing of the diphthong. Crescendo, or ‘rs ing’, diphthongs, which start weakly stressed and build up to a peak near the end, repcesented by <, are also possible, Some people might describe such English words as yes [jes], yawn [on], you (uj war [wou], ete., as containing rising diphthongs [Je] (Jol [fu] [Vo], but it is more usual to eserite them (and transcribe them) as we have done, that isas sequences of the semivowels Gand [W] +a vowel. 7. Close and open transition Before leaving the subject of sequence ghenomena we must consider one ore topic, namely two different ways of making the transition from one consonantal articulation to another: what we call dose trarsition and open In cose ‘ransiion the successive consonants are articulated as closely together as possible What shis means diffess according to whether we are talking of a homoreanic. heteroreanic. or contiguous sequence, In a hhomorgaaie sequence, close transition simply means thatthe articulation is ‘maintained, unbroken, through both of the successive sounds: thus the [pp] in top part, is artioulated with the lips closed forthe duration of the two (ps the (f] of helf fare holds the labio-tental fiicative articulation throughout both ffs, and 20 on. In open transition between homorganic consonants there is a momentary and minimal relaxation of the articulation which is reformed immediately. “Tus ia “ake the top apart’ in Eaglsh there is normally no vowel between the (ph just an open transition which we can represent as [-]: thus top « ‘part (as opposed to teppar’). Siailarly in half a fare there is a mere open transition between the [fls, thus, half + fare. 75 Experiment with making close and open transitions: between Komorganic2 CO-ARTICULATION AND SEQUENCES consonants. Thus, sty fa f Fa} [a ss a] f¥ val [av a} in cach case articulating the two coasoaants as a prolonged! articulation with a syllable division in the middle oft, Do the same with pairs of homorganic stops: p pa}fabbalfattalfad ¢ 8] [arda]fad 1} ete, Notice that in each case {even when there is change of phonation, 2s in [4 td] [2 d tal) what you ‘articulation, maintained unchanged for the duration of two have is bg. successive sounds These are all lose transitions. Now say the aime sequences with an absolutely minimal release and remake of tbe artisulation, Sey [sf Fa, for example with a sight selxation of the lover lip between the [f}s The relaxation nest be ne move than sul ficient toinserta momentary voiceless approximant between the [DS hus; [a fy fa] Notice that this is what happens. i rapid speech at east, in saying! a fare, cough affair, ete.in English, Carty out similar experiments with the other examples just given, [8°] fave vs) ete. and then with the stops fa pp a] [a d+ da, exe, Note thee the minimal open transition between [p] and [p] may take the form of a momentary bilabialapproximant orwide fricative [9], thus[a p} pal “These were examples of close aud open transitions between homorganic ‘consonants, In the case of hererorgamic sequences, the difference between close and open transition is somewhat more marked. In close transition Detween hcterorganic consonants thereis a moment of articulatory ovedap: in other words, the articulatory striciure for the second consonant is formed before the stricture for the fist is released, 77 Say fa kl and hold the final [k], Now. while stil bolding that final (k] form a bilabial stop articulation. Silently reoase the [K] and only after that release the [p}. The whole sequence might be transeri reproser's the moment of co-artieulation—of articulatory overlap—in the iniddle of the sequence, Now say the English words back part, slowly and silently, giving particular attention to the articulation of the sequence (& F) In virtually al varieties of English, this will be an example of los, sd fan k k p pa}, where kp] Carry out similar experiments with [Ukin that car (p ] in top ten K] 0 staf car (sia enphali | Nosie tow, in every eate (in normal English), thereis close transition—in / overy cose the second stricture is formed before the Gist i released, If you are not a rative speaker of English you may have 10 carry out this series of experiments with panieular eare, to ensure that you ar, in fact using close transitions CO-ARTICULATION AND SEQUENCES 113 Next experiment with oper gansition between heterorganic articulations Use the same examples, but this time, rake sure that the frst stricture is released a moment before the second stricture is formed. The actual transi- tion between the consonants must be absolutely minimal: there must be no | vow inserted between the consonants, The azproach of the artieutators to the second articulation must have startad by the moment when the dist sicture is released Observe that, in rapid specch at least. English otlizes open transition betwee consonants in such plrases a Boot apart (ep thar a car] sop acer {p 1}, al a cam (CK), this affar's 1, es Finally we must consider the cransitions (iat oocur between consonants with contiguous articulation, In close transition between contiguous ariula- “tions there is ccommodat phonation) oT S36 SOREGHEAt aconmodates 10 the aitiulation (o: phoaa- sion) of the other. In many languages, including English, the place of articulation of a velar is shifted forward slighily belore a palatal semivowel Gor a paletal vowel ff, In other words, the velar articulation is accommodated to the contiguous palatal articulation: dhe ve, tha isis somewhat palatlized In. open transition between contiguous a 5 there Fite or a0 accommodation Thesincture forthe rs onsonuat made at Teoria ‘oeation, and the movement of eecommodation to the location of the see- ond consonant is actually made during the momentary period of \n: the place of articulation (and sometimes the culati transition, 18 Soy the Boplish word backyard, contzinieg [kj] several times, aloud and silently, Note the ascommodation of (k] 10 Ui}. Now say eighth, containing [0] several times, aloud snd silently, Note the aovorumadstion of (\ 10 [2]. ‘The articulation [ys apico-dental, Compare its alveolar articulationia eigie Soy (aia, containing (aloud and silent, and note how the apico-alveolar [1] accommodates 19 the post-alveolae[r]. The (¢ either articulated com- pletely posi-aiveolarly (1), or the tonggue-tip slides rapicly backwards from the apico-alveotar to the apico-postalvectar position, Now compare al! these examples of else transition between contiguous articulations, with open transitions between the sime consonants In English atleast in apid speech, there ie open transition (eather than the actual prs {nce & vowel) in the sequences +] im bask a yard, [8 in ae a tin [2] in remain. Say these aloud and silently, and note how in this second, ‘open:transition series there is no articulatory accommodationII co-aRTICULATION AND SEQUENCES CO*ARFICULATION AND SEQUENCES 115 CONSONANTS (PULMONTO) jie vi gos. ic (Tinta chia [es Pei im] wee [losin eel Se [ewe [pb td idle si | mm a a = fe 3D et ae zl SEV eSEz SA Ee ee © diss (es Biss CL cenaatecar £ Piece [ACh ve 1 anaveste FPN asada |G. at : ae cosoxasqoseawona Pabre! sb [2] at Cicts | Voiced implosives Ejectives «7 tp’ C7 dentaieote kK S?_nbeotrficaive Bramstes Bilbist Ver mine pat OTHER SYMBOLS S AN Voices wois-venrineawwe” — GB Aiveolopaats thessves. larmain2 W Voiced bbia-vctaragpraimane vost svar tee top i we SAU vires pponimat — f}simatincen fant X, SUD), BL vote ett eve mi i 4147 Erigowal sive Seen ees ps Pa Fig. 36 The International Phonetic Alphabet (revised to 1993, corrected 1996) — aun \ ou —ske— +10 5 ons ge she at a 8 ds ava (aly Senate Toa ann cence a es Sar cial En EE Be A, tie Sapte Gt Nes 1 Maltfong * é Ma e 4 or & & me ey copes © eS [menor § er 7 Syllble break Zi.cekt Downstep 7 Gobal rise T Upmep Na. Globat fatt angen DUAGRETES Dene may eed hove symbian A Yostion sreanyvoiee QA | oom td ioe, SESE eee td anes” COE upton CO [ut £O| = Now tie OY eS ta FT rasmiine tt GI | neainine hen Yo vemos CGY] centene ee Yeteie oesanraemies oe a vowed, B= ieee itil appronian) samme Tege tee & eset Torgitow ©L1G co-ARTICULATION AND SEQUENCES. Table 3 NS jequence type ‘open iansiion (Close transition Homorganic Arfculatory Aniculatey nor-eontinlty ‘continuity Nooverlsp Ariicunory overlay Contiguows Noaecommodation Ariculawory caccommedation ee We can sum up the differences between close and open transition as shown in Tele We have devoted some time to the (wo dillerent ways of making the transition between successive consonants becaus: this isa feature that dif fees somewbat from ove language to another—thus in (ct) in English evar open there close triton, but ia French acteur there is, usual transition, ‘With this survey of e-ariculaion snd soqucnse phenomena we come to the end of ou stay of what are usvally called consonants. It wil be noted thee we have given no format definition of consonaats though it wll be tear thai by consenants we mean ckicly sounds Functioning as elable vans, in the wort eat [et [End (form the margins of te syle, ihe the vowel fe] is centre or auleus” Again, the dovopelatal Spano Gj incleany vowel when iis he conte of a sliabl asin seat Tit, but when essentially the Same sou is reduce 9 a mere slide aay from the dorsopaltal approximant postion, nthe palatal semwvowei as in yu js it yale margin, and consequently consonant, We stall have ile more to say about this at he beginning of Chapter 7. th order to rrapituate most of what has been discussed up to now, Fook Fig. 36. This is a chart of the phonetic symbols of the International Phonetic Association, and it wil be adviable ro ead through i several times, both silently und aloud, both horizontally along the rows of stops fricatives, and so forth, and vertically down the columas. To assist in this we provide come notes here Terminology. Reading across the top (location labels) of the main chart wwe find the headings Dental, Alveolar and Postalveotar. The symbols [2] [a] * (5) (4 (4G GS) Ll Mare listed under Alwolar, bat with no vertical lines cm from Dental and Postalveolor. This means that these separat CoraRticuLaTion AND SEQUENC! 47 symbols can be used te represented sounds articulated at any ofthe three locations, It is necessary to speci that such sounés are in fact dental the appropriate dental daerits ean beused 06 wodr ihe lit of DIACRITICS); if they are retraced alveolars, oF postalveolars, the underline meaning ‘retrgoted’ can be used (jllustzated on the diacritics list only fora vowel). Note that special symbols are provided for fricaives at all thre lagations. But nots also that dental 6} and (2 can also oocur, and can be symbolized bby means of he dental dizeritic, Dental x] and [2] differ from [0] and [0] in terms ofthe size and shape of theit articulatory channels: fr [s]and (7 this isa narrow, grooved central channel for [0] and [8] a wide, fat channel Note that under OTHER SYMBOLS [6] and {| are labelled atveolo- palatal, This older term violates the strict rule of systematic phonetic terminology in which the prefixed term refers 10 the lower articulator. We should prefer to eall them lomino-prepalatal or palialized leming- postalveoer. For the present the display of VOWELS, SUPRASEGMENTALS and TONES AND WORD ACCENTS ean be passed over, since we are going to deal with these matters in subsequent chaptersDs Spats ALC Yn te: 7 Vowels: Introduction 1. Vowels and consonant: study of vowels Importance of silent In Chapter-6 (p. 116) we partially distinguished besween consonants and ‘vowels oin the bass of their different functions, merginal and oenteal, in the structure of syllabies For general phonetis purposes, however, that distine- tion is inadequite There are, for instance, certainly syllabic téntral units that we would prefer not'to'deserite as vowels: for example the syllabic tailed {¢] that forms the syllable ezotre in Czech words like (krk] ‘neck’ or [pit] finger’, dot to mention syllabic (] and [r] in English middle and button." hs . "In‘ordér to maiko fa clase of aricalations that comespodds closely to what are traditionally called vowels, we have to be somewhat arbitrary. So by vowels, we understand a class of pulmonic pressure sounds normally voiced. with 2 maintainable central oral arproximant or resonaat dorso-domal, or pharyngal, articulatory channel. f ‘We have already seen, more than once, thatthe articulation of one vowel at least, namely fil, can be pertecily well deseribed in accordance with the same principles that we use in deseribing consonants’ that is as a dorso- palatal approximant, Ta fact, all vowels. can, ia, principle, be described as approximan(s, or resonants. articulated at various oral ané pharyogal locations... . However, it hat long teen the custom to cefine vowels ia terms quite ‘iffezet from those used in defining consonants. This traditional éeserip- tion if vowels, instead of specifying stricture type amd location, as for consonants, seeks in effect, to define the shape and size-of the resonance chimbers'of the mouth and pharynx by specifying the position of the tongue and lips. The traditional way of classifying vowels Works well in120 vowels: INTRODUCTION practice, and, indeed, is the only basis for the successful acquisition of practical skill in producing, identifying, and classifying vowels. Veirls ot more precisely, the month-shopes foe vowels aro specifid in terms of three variableY—thee garumelers: vertical fonguegositon (high-low), horizontal tongue-position (front-back], and lip-position (unrounded-rounded). Iu is essentisd for the student of phoneles 10 become: intimately acquainted withthe tactile and proprioceptive sensations essocited with in bis own vocal tact, Itis absolutely useless merely to he awa of thet ntelleeealy. Consent, one must earry outa pret dal these paramet Of stient introspection concerning these sensations, and silent practice of vowels. This was the recommendation of the great English phonetician Henry Sweet, over a century ago: “The first und indigpeasable qualification of the phonetician is a thorough practical knowledge of the formation of the vowels ‘Those who try to leam mew sounds by ear alone, without ary systematic training in the use of their vocal organs, generally succeed only partially. (Sweet (1877), p. 21.) Considerable experience (and at least one small experiment—see Catford and Pisoni (1970)) confirms Swezt’s view, In all experimentation on vowels it is helpful to use a band mirror (and a I Rashlight where necessary) so that one ean correlate the visible movements and positions of the tongue and lips, with the proprioceptive sensations, and also with the auditory sensations when they are whispered or voiced. Z. Lip: and tongue-positions for vowels ‘The most obvious and most easily controlled of the vowel parameters is lip-position, $0 we stact with that. 79. Sienly alernace a vowel of me (J4ype (a8 is [ubtype (as in English eco), Concenzrave attention ou the postion of the lips sis see) witha vowe! of the in articulating these vowels, Speakers of many varieties of English aay notice that in saying s word like roo there ie no fixed degree af lip-rounding, ‘At the start of the word the lips may be more oF less spread, but as the luterance of the word proceeds, they bexome more and_more closely rounded. Weare, for the moment, interested only in the most xtreme round- ing ofthe ips as atthe end of too. Say along (ae] vowel with the lips fixed in this closely rounded position. VowsLs: INTRODUCTION 12] Haring slently produced this counded fu), switch to [i] (as in see). Sidenely alternate (aii) A¢ this point one must detach the lip movemen's from the ‘vowel uounds In other words i necessary to expesment with slowly and Geliberately coundigg anc spreading the lips as an activity by itself, with 10 particular vowel sounds in mind. Having discovered experimentally that lip-rounding is an independently controllable parameter, one can make further experiments to discover the effect ef rounding oF its absence on various types of vowel 80 Say a prolonged {1 . J aloud and silently and introspect about the lipe position, which you will probably find to be moderately spread. Exaggerate the ip-spreading a lit, 50 that the ]-vowel is being said through a broxd senile, Now wile slenty sayinga prolongesf iii. ..]lowly and deliberately round the lips. taking care to think fi? all the time, rigidly maintaining the tongus-postion of [i. Having reached maximum lip-rounding, with the ‘ongue sll in the f] posion, produce whisper, and then voice: The audible results an [i-ype vowel with rounded lips, which we symbolize by ly}. This i approximately the French w of faweand the German d of Bile. Say a prolonged fu w w...J-roughly the vowel of tao, only take care to say ie through filly rounded lige, While holding a prolonged silent fu wu 1 v.Jpslowly and deliberately unround the lips hing care co “think fa)" all the time, rigidly maintaining the tongue-position of [u), with tongue bunched up at the back of the mouth, The unronding process begins ‘chiefly with a eelavation of the closely rounded lips but this is followed by su increasing tension at the corners of the mouth, as they are drawn bask, ino a wide smile Having reached a state of completely unrounded, widely spread lips, with the tongue stil firmly in the [uppostion, produce whspe, 12 voice, The audible result isa totally unrounded {ul (a vowel that ie [up fhe which the phonetic symbol is fw). This is doesn't sound at all somevht similar to the Russian vowel sf of suis [muuf] “mouse” or the Scots Gaelic a9 of Hogh, fruy] “call”. Experiment further, silently and aloud, with adding and subtracting lip-rounding, saying fu wu us am. Jete Now experiment silently in a whisper and with voice, adding lip-rounding, to fe]. s in Read, and (akiag lp-zounding away from (as in sow, and 30 on. By now you should have good control of the parameter of lip-position, and the ability to pronounce any vowel at will with the ips unrounded or rounded. If this is not so, repest Experiment 8) as often as necessary, until you fed that your lip-control is perfecta 122 YOWELS: INTRODUCTION oe ‘The second parameter of vowel description that we shall examine is that of ‘tongue-beight’, as itis often called, 81 Say aseries of vowels, jl fe] la} roughly asin English deat bat bet bar—at this stage the presse quality ofthe vowels is rot important. Say this seces of voniis several times. silently. (el [Ca Oe) Cefa, Kewl te obvios than progressing through this set the mouth becomes more an mor open Experiment with a comparable srics of vowels articulate atthe Bask of he mouth YET ETA), very rosy as in too toe paw pat. Say this sees silently and make the same observations as you did on Gi [6 fi]. Once agiin, it shouldbe clear tha the mouth gets prowresively moze open 2s your ge TETRIS te. Repeat thesis tigain several times. In carping cut these experiments in varying vowel in’ die versal measion—the parameter of torigue-beight'—you wil probably become -rvar thatthe opeting ofthe movth ia going from, sy, i]t (aan be done in tvte diffrent ways: (by keep the lower jaw stl ané peéaressively lower and Baten the tongue, (by Snatng the tongue With respect to the lower jaw, and siroply lowering the jaw, (i) by a Bi of bot: From the phonetic point of view the impertaat thing is the varying istance detween the surface of the tongue and the roof of the mouth, however achieved, since that determines in large part the size of the oral resonance chamber. For: the student of phonetics however, the’ most important thing is to develop awareness of the positiow and shape of the tongue, and this can be done most readily by fxating tke jaWand concen- trating attention oa tbe proprioceptive and tactile sensations associated with loweridg and raising the tongue $2 Bite theend of a pen of pereil, ot any other convenient object that will old the teeth about | ino { em., apart, Of course itis importart to make sure tut the toth-separator docs net project into the mouth to interfere with -moverneats ofthe tongue, ‘Now, holding the jaw rigid in this manner, experiment repeatedly with silent vowels ofthe types fil fee] [a] and {ul fo] fol moving through the series ‘upwards and dowawaede slowly and quick'y, many times Try to become clearly aware of the diferent degrees of tongue-raising through the proprio ‘eptive sensations from the tongue-muscles, ané (or the front series of |vowe) the tactile sensations of contact between tongue sides and molar teath, at least for Gand (¢). Finally, aé far a5 you can by looking. with asblight anc misror through the narrow space between the teeth iry to see lhe different positions of the tongue. Ta order to see the tongue-positions VOWELS: INTRODUCTION 123 well you will have to experiment with going through as neatly as possible the same series of tongue postures with the mouth rather wide open. In Experiment 82 the vowe's were all articulated silently, Experiment 83 adds whisper and voice to them. 82. While carmying out the same experiments asin Exp 82, first whisper and then voice the vowels You miy Gnd at first that the auditory sensations of the ‘vowels completely mask the proprioceptive seasutions. Whisper is @ good compromise, however. In whisper you cau bear, mote of lest, the auditory qualities of the vowels, but you can stil fee} the tongue-position pretty clearly. IF necessany, cepeat the last few experiments agsin until you are confident that you can clearly fee! the difereiees of ‘tongue-height’ ‘The third parameter used in the description of vowels is what we have ‘called “horizontal tongue-position’—the relative advancement oF retraction of the body of the tongue. 84° Soy. siledt (thea earetony could it to a stent {y). Be careful to preserve exactly the tongue-position of fi] ven after you have closely rounded the Eps. ‘Now rigidly maintain the lip-postion, but slently shift fom rounéed [i thai is back to (ul Sileaty aleenate [yw y i] Several times slowly and elberately As you iotrespect wile ding this you will obéerve that in going from silent [j] to silent [u] the tongue moves backwards Gecper into the ‘mouth, As you seal alternate fy vy vlete.youcan fel the tongue dicing backwards and forwards Ifyou ate a speaker of magy types of Eneliso, and ityoa are using a (Clocely Suaded) version ofthe oo in fo0 as your {u}. you ‘wil probably notice tat ie congue dossn't move back very far in going frome [pj to ful. Ths w because m most varsties of English me 90 voWe Is m0. very back vowel, To indicate this ‘non-back’ quality of oo we can write two dots over theletteru i]. So, in carrying out the present experiment, see what happens if yousilenty go from (y] to [x] as before, and then strive your hardest to pull the tongue evea further back Keep the tongue fully bunched up, elote fo the reof of the mouth, as it was for [ano (but seve «fe prolong the backward side of te rongue teyoné [6 ea flly back [ut bEEH * The reason we started Experiment 8¢ by closely rounding [i] to [y], and then sliding back. from [y} to {i} (and subsequently {v]) was this: the proprioceptive sensations of lip-movernent tend to mask those of tongue movement. So if you hed started from unrounded fi] and then slid silently back to rounded [3] and (u] the strongly percsived feeling of change of4 VOWELS. INTRODUCTION lip-poston would have tended to blot out your perception of the tongue movement. Consequently, it was necessary to keep the lip-position constant FE ota Sopa li a atiod tv daa seed We can, ofcourse eliminate shane oflipposion in another vay: by szating with unrounéed[t} and sing forward to (and in Experiment 85 wed that S85 Says silent fu) and try co make ita eealy back one, ie, a back [a rather than ack (Gi. Now slowly and deliberately anround and speead inglish non the lips, while “thinking [up ll the time, and carefully maintaining quite Fisidly the tongue-postion of lu} ‘The result is silent ts). Now trom this ( position siently slide forward to [i]. Allthis time the lips must be kept spread ina broad smile, so that there is no distracting change of lip-positon. Silently slide back and forth, slowly then mere quickly, between (wand [i fui wi ee. ‘As you do this repeatedly you should begin to identity the propriogeprive sensationsof aving the body of the tonguepushed as far forward in the mouth tus possible as lor {or [yan pull far back in the mouth a for fu] oF fu) Carry out similar expetiments with lower or more open vowels, for example, slently produce an [1], soughiy as in fed: while carefully retaining the tongue-position of {e] add lip-rounding, changing [e] to (a), which is roughly the vowel of French neuf. Now, without changing the lip-position, silendy slide back to (2, 08 in sane snd altemare slowly, quickly, at fst silenly, then Biter Whispered and voce” fu >«e 9] ‘This should give you once again the feeling of moving the tongue back fand forth in the mouth: varying, that is, along tbe parameter of forizontal tongue-position If you still have some uncertainty sling of moving the tongue to the front and back of the mouth, do Experiments 84 and 85 again ané ‘again, until you are absolutely confident about the feelings of horizontal tongue-moverent, Now you have # clear idea, intimately experienced in your own voeal tact, no! merely intellectually known, of how vowsls can be deseribed and classified: in terms of vertical and horizontal tongue-position, and position cof the lips We have seen, and verified experimentally, that vowels differ in terms of three parameters, and can thes be classied as () rounded or unrouded (i) high o€ low, oF to use the terms we shall use benceforth close or oper, it) rant or bate VOWELS: INTHODUCTION 125 Tables Front ack Unrounded Rounded | Unrounded Rounded Close: (beet): yilne) — | en (ogi) u (oo) (Gat) « | Opens (buds ere | 2 am) We can lay out this scheme of classification as shown in Table 4, with 2 few examples, The general outline of vowel-description chat we have presented is aot sufficiently detailed or specific to enable us to unambiguously specify fine istinedions beeween the vowels of different languages and dialecis, For that purpose we have to learn a rather previsely defined set of universal reference-vowels, known as'the Cardinal Vowals\. This system was invented by the great English phonetician Daniel Jones 3. Introduction to Cardinal Vowels The Jones system of Cardinal Vowels is based upon the idea of a “vowel limit’, and of a ‘vowel space’ delimited by it. IL is easy to discover experi- mentally what is meant by the vowel limit, 1S Produce a silent close front vows, a vowel ike that heacd ia English se, French i et ff, Note, as you have frequcaly doue befor, how the tongue +s bunched up very bigh inthe (ont of the mouth, lewing only 2 narmow chanacl between the dorsal surface ofthe tongue and the hard palate and prepslatal arch. Voice this vowel ff. AS we have observed before. (is a docse-pulaia approximant. Now, while pronouscig a prolonged tii. J ‘onl, make the Longue very tense aad push its doisal surface up closer and closer co the bard palate unt audible turbulence ean be heard, At this point the approximant} bas obviously cursed into the friatve fj] By making very fin adjustments inthe tenseness and convexity ofthe tongueserface, pass back and forth several ies beween the approximant i} the fricative fasigiat. Experiment 86 has shown that there is a limit to the ‘closeness’ of a vowel: if you raise the tongue any further, the vowel will turn into a fricative126 vowexs: mvrropverion 87 Produce a prolonged voiesd [i] oree again, roting particulary iss extreme “Froncness, te tut that the anterodorsum of the toague is pressed very fir forwards in the mouth, Now strive to push the anterodorsum even further forwards, coming closer and eloser to the prepalatal arch (immediately above and behind tHe alvedlar ridge). Once again you will find tha’ as the channel between the tongue-surface and the palate-curface narrow, the hse of turbu= Jenes becomes audible and the approximant [iJhas turned into a fricatve: not exactly the Gj] we reached by pushing upwarcs from (jj, but rather an 4, fj anterodorso-prepalatalfricative which we caa represent by (2), and which _ an sounds lke 6] modified in he direction oflminc-postalveolr i]. Cerefaliy altemate zi zi}. A similar expeiinont with fu] (of betty unsounded counteract i) ves raul of thesame kind. 88 Produce an [u}type vowel—domewhat like a tense, monophthongal (F! scompléely unchanging throughout is duraor) vesion of the English | ‘vowel in ‘oo, of better. the German vowel in du—which is generally more | Bock. From this very back {u] slently, and while rigidly retaining the close! back tongue-position of [y, slowly and deliberately unround. the lips: be careful to retain exaenly the close back tongue-position walle doing this Silently caiatain the. [oy] position, noting how you caa feel the iongue bunched up towards the back of the mouth with part ofits dorsal surface ‘very near the soft palate Carefully maintaining the close back position, initiate and voice a prolonged (wi this lke (9, isan approximant Now, while continuing to produce prolonged (a), tense the tongue-tack and strive 0 push it apwards coming stil closer :o the soft palate. As you do this, the hiss of turbulent aidtow will besin io be hearé: the aparoximant [ea] is being ‘converted into the frieative[y). Carefully alternate wy 2 0]. A similar experiment with a very open very beck vowel will help to.define the vowel limit further, The kind of, vowel to aim at is roughly ofthe type of (a] in far or the [0] of cot in British RP or the [o] of caught in some varieties of American English, But it must be muck deeper—more retracted —than any vowel ou are farniliar with 89 Draw the tongue down and back, particularly back. Get the sensation of pulling zhe tongue as far dowa and as far back into the pharynx as possible. Retain this very low back position, and voice it, producing a very low back [ol-type vowel. Prolong this vowel and while doingso force your tongue even further back. As the space between the tongue-toct aad the back wall of the vowns: mtmopuction 127 pharynx narrows the quality of the fa} vowel will become somesthat strangue lated, and you may hear it turning into a pharyagel fictive or possibly a sind of pharyngalized uvularfrieative,Itis clear, at east, that you ultimately reach a point where the clear vowel-like quality of [lis lost. Experiments 86,87, &8, and §9 have served to demonstrate that, atleast in the upward and backward direction, there isa kind of boundary in the mouth and pharynx beyond which agproximant-type vowel souads become {ricatives: this boundary delimits the ‘vowel space’ in these two dimensions, as indicated by the broken line in Fig. 37. It is more diffi to specify, in'd precise manner, the forward and down? vied lis of toague position. Fig. 37, which is based on X-tay photo- srepbs of the author’ articulations, shows the ‘vowel limit’, represented by a broken line, continuing down the prepalatal arch to just behind the post- alkeolar zene This is quite in ancordance with the findings of Experiment 70, in which the convex anterodorval surface of the tongue was brought as ‘lose as possible t0 the concave’ prépalstal arth. But itis not so easy to incicate thé vowel limit for completely open vowels in an analogous man- ace The’ problem is that the tongue is a mobile and polymorphous mas itt Is € say, a mass chat'cin take up different position dad somewhat different shapes. Since itis difficult to describe and compare the positions of the tongue-mass as a Whole it is convenient to select some easily identi- ‘ied reference point for purposes of comparison. For over a century it has been traditional to define relative tongue-positions in terms of the location Fig. 37 The vowel limit128 voweLs: iNTxoDUCTION et ici ial of the summit of the convex tongue-mass—Lhe highest point of the tongue. In Fig, 37 the highest point of the tongue for each of the vowels [i tw} and [al is indicated by the tip of g small black tviangle, ft must always be remembered that the highest point of the tongue is no more than a coaveni- nt reference-point;it has no significance other than that, Indeed, as we have already seen, another Location on the tongue-surface is more important in defining the vowel limit, and indeed, in other ways: this is the location of ‘narrowest linguo-fectal or linguo-pharyngal articulatory channel. ‘Anticipating a litt lock at Fig. 38. This shows the tongue-configurations for the Cardinal Yowels that we are about to study: The black dots here Show the highest point of the tongue for each vowel, while broken lines round the periphery show the location of the narrowest articulatory chan- ighest ne for some of the approximant-type vowels. This shows that the bi point (a mere refesence-point) does not always eoineide with the location of cticulatory stricture. ‘As we have already pointed ont, the underlying idea of the Cardinal Vowels is that of a ‘vowel limit’ within the mouth, aad, consequently, a “vowel spuce’ inside that limit. In theory, any vowel of any language must have its tongue-position either on the vowel limit itself, or within the vowel spucs, The problem, then, isto define chat spice, and to specify a sot of sference-points on the vowel limit, in @ way that is independent of say articular language. — Frost } oo Rack i 12m tighescpont of tote Q)ete = nartowest eiculstory char! Fig. 38 “Tongus-configuralioas for Czrdinal Vowels Vowsts: INTRODUCTION 129 ‘The key reference-points for the Cardinal Vowels are two that we have already experimented with, namely the closest and most front vowel pos sible fil, and the openest and most back vowel possible [a]. The vowels] and [u] represent relatively fixed points that anyone can locate for himself by following the directions given in Experiments 85, 87, and 88, without ever having heard the sounds, Between [which we will call Cardinal Vowel 1, and [a] waieh is Car! inal Vowel 5, are three intermediate Cardinal Vowels: 2(e}, 3 [c}, and 4 a, ‘which are sometimes said to represent equidistant points between ! [i] and 5 {c]. The systems completed by continuing the series of ‘equidistant’ vowels (on past fo] up to [, with Cardinal Vowels [6] 2)? fo} 8 (ul. Damvel Jones, the inventor of the system, described the Carcioal Vowels ofthe sets sé and 5-8 s bein orgonical articulstorily) equidistant, ie, separated by equal steps of tongue-lowering/raising, and also acoustically (by which he meant aeattorily) equidistant, On the whole, he emphasized the articulatory equidistance more, though the question whether it is better to regard the Cardinal Vowels as an articulatory or an auditory scales Stil disputed. The present writer's view (ike that of Henry Sweet) is that for the purpose of learning the Cardinal Vowels itis essential to concentrate on their articuletory characteristics: make correct tongue-configurations and the correct souads will avromaticlly follow. Daniel Jones emphasized the articulatory (and auditory) equidistane: of the separate senes{i-e~c-a] and [a-o-o-u], not the equicistanze of the entire set. This is important because, 2s many users of the system have observed, the distance berwuen [a] as the cpenest possible extremely front vowel, and [, as the most retracted possible open vowel, clearly seems to be greater than the distance between the successive ‘equidistant’ vowels within the separate front and back sets 4 Partly because of this, itis desirable to learn the front series and the back scties as two distinct, thouga risted, sets. This helps to avoid a problem. that arises if one tries to produce [-e-c-a-a] as a single set of equidistant vowels. This problem is the tendency to let the more open freat vowels slip buck from the fully fronted position as one begins, as it were, aim for [ol Fig. 39 shows schematically what I mean, The soli line through [fe] [el [a} indicates the progression by equal steps through a purely front series oF vowels: the broken Hine suggest the kiad of dwvistion from frontness that ofien occurs when «1 student ties to say i-2-2-2-0] as a single series of equidistant vowels,Fig. 29 Bros te avoid in pronouncing front CVs. Typical error of leamer trying to say F-e-e-8-¢ in one sores (coe text) Before finally beginning to study the Cardinal Vowels we must introduce a. ‘ype of digram thatis commonly veedo represent the Cardinal Vowslsand the universal vowel spare that they sunoundand dene This dagram Gig. 40) is a simplified version of earlier diagrams that purported to correspond more closely othe actual tongue-positions ofthe Cardinal Vowels Iti easy to draw th Cardiol Voveldigram if you observe the length proportions 2:3:4fo the Dotom, right, and top sides ofthe clagram, The slope of heeft sid, representing front vowe’s is supposed to correspond more ores te the fact thatthe location ofthe highest point ofthe tongue retreats backwards {nto the mouth the more the tongue is lowered, but sis showa by come argon with Fig. 3, this retreats exaggerated, The fact thatthe top tng connecting [i] and| is twice as long as the bottom-line, connecting [a] and [oh reflect the achat the ezoure (end advo) diferene between and {uJ is more thaa twice the dfence between [a] and fo} The very great acoustic difference between [i] and [u] results largely from the fact that [u) is’ rounded while [i] is unrounded: Asccan be seen from the symbols used to label (ximary) Cactinal Vowel thee ont of four back vowels niney jm are rounded only te mostopec bck vowel, isuarounded. ‘The reason for selecting rouaded vowels for.the back:series is that, more often than not, in natural languages all: but the most open of back vowels are rounded. Theresa second sere of scondasy’ Cardinal Vowels waich have the opposite lpspstions that the font vowels and the most ope bac vowels are rounded, while the closer back vowels are uarounded $0, then, the coounenly used Cardinal Vowel diagram isa Bybri, to S:INTRODUCTION 131 FRONT (CENTRAL back Fig. 40 The Cardisal Vowels largeextent reflecting the relative tongue-postions ofthe vowels, with front vvowe's down the left side, back vowels on the right side, and with its vertical spacing redecting equal articulatory steps of tongue-lowering in going from 6] to [a] or from (ul to fo}, but at the same time reflecting the different acoustic distances between {i] and {u), and between [a] ane (a] ia the differ- cent lengths of the upper and lower horizontal lines. In spite of its curious hybrid mature, the diagram works very well in practic: Since the tongue-position of any vowel mest lic on the periphery of the figure or somewhere within it, dots placed om its periphery or within it ean indicate the position of any vowel in relation to the Cardinal Yowe's, which serve as universal reference-poiats, This is way the Cardinal Vowel system is an essential too! of all practical phoneticians, who must leara ‘these universal reference-vowels as well as they possibly ca. Tt has often been stated that the Cardinal Vowels can only te learned orally, from a teacher who already knows them. This is only partly sue, Perhaps a final polish can best be applied to one’s Cardinal Yowels by appealing to the judgement end the corrective hints of @ competent teacher. But all students can go a very long way towards accuracy in the Cardinal ‘Vowels—and some can achieve virtual mastery of them—withouta teacher, providing they carry out experiments and exercises on them with meticu~ Iouscare. In Chapter 8 we introduce the system of Cardinal Vowels in detail through a series of expesiments8 The Ca rdinal Vowels (CVs) 1. General characteristics of cardinal vowels: Errors to avoid There is some dives system of Cardinal Vowels (which, from now on, we shall refer to as CVs) should be regarded primarily as an articulatory or as an auditory scale of referencewonels. From the practical point of view, however, i is best to proceed as if the system is basically an articulatory ong and (o familiae- ize oneself thoroughly with the ‘Teel'—the proprioceptive and tactile seasations—of the CVs ‘As indicated in Chapter 7, the CVs are a set of eight tongue- configurations, five of them with unrounded lips the remaining theee with rounded lips There is also a set of eight secondary CVs with exactly he © sine tonguo-configurations, but opposite lip-positons, Finally, there area few additional vowel articulations, which, though not usually considered part of the CV system, are exsntial for the student of phonetics to get 10 know. “The CVs are not the vowels of any particular language, but are, as we have sen, a set of systematically established, language-indepondent, uni- .nce of opinion among phonstisians as to whether the versal reference vowels. Naturally, some of the vowels of sonie languages hnappen to coincide with, o¢ to approximate closely 0, sowie Uf he CV, Thus typical French vowel, of si [si ces (se, sat (se, are rather close to} the CVs [J els} Practically no English vowels are a al close to CVs. This | means that the English-speaking reader must start from the assumption that not singl one ofthe yowels wsed in the English ianguage is exactly 2 Cardinal Vowel. Morcover, the vowel-sounds of English, and of other languages, vary very considerably from one disleet to another, ‘All of this means that in learning o- teaching the CVs we must use the134 THe CARDINAL voweLs (Cvs) vowels of the reader's language with the greatest care, From time to time they provide useful starting-points for arriving at CVs, but readers must always assume that the CVs are not by any means the same as any vowel of their own language. ee ‘Two characteristics of the CVs must be emphasized at the outset, First, they should all be pronounced in a rather energetic, tense, mannet. They are, after all, peripheral vowels—that is, vowels produced with the tongue thrust as far forward as possible or pulled back as far as possible, and the close ones have the tongue pusked up as far as possible. It is therefore reasonable to assume that if you do not at times feel some degree of strain, or tirecness, in some of the muscles within, or attached to, the tongue, you are not exerting yourself suficiently to produce good CVs. Secondly they are all menophuhongr—that is, they are simple or ‘pure’ vowels, that can be prolonged for 2s long as possible with absolutely no change of tengue- (or lip-) position, and consequently no change of quality. This is very import- ant for English speckers to remember (and we shall remind them cepeat- ccdly), since many of the vowels of English, in, most dialects of English, are diphihongs. That is, they are gliding vowels that start with one tongue- and! oF lip-position, and end up with another. Obvious diphthongs are the sound of I [2 1] and the ow of Aow (au. If you say these words aloud, in a whisper, and silently, several times, the gliding movement of the tongue will bbe obvious in both of them, and in (au) a cadical change of thelip-position will also be obvious. ; + But some other English vowels are, less obviously diphthoagal. Thus, ‘many speakers—it is probably safe to say most speakers—af English pro- ounce the vowels of such words as day and go in a decidedly diphthongal manner. Experiment 90 deals with the vowel of day (normally a diphthong) and with the production of.a rigidly fxed monophihong.. somewhat resembling it i 90 Say the word day aloud, in a whisper'and silently. Fednounce it let, quickly, siowly, and at an intermediate rate: As’ you introspect about it, it is ‘ej probable that you will discover that your promuaciation of thea in day is diphthongal. If you conte from Scotland, Ireland, Wale, India| or come parte‘of the Nerth of England, it is possible that you wil find your ay to be ‘un unchanging monophthong, Most probably, however. you will Sad tht ay ssarts with mouth slightly open, ané then, as the sound progresses, themouth loses somewhat. The tongue rises slightly (when you are doing it ently, you may feel an ineceasing contact between the Sides of the tongue and the ‘THE CARDINAL Vowers (cvs) 135 molar teeth). The auditory effect i likely to be that ofa vowel that changes its quality throughout its duration: its exact starting and finishing points vary very widely according t dialect, and the diphthong might be represented as {cil fed, [et]. and so on: you will learn the precise values of these letters later. If you find that your ay-vowel is indeed, diphthongal, even to a very slight degree (as you probably will), it will be necessary to start training your tongue to adapt and hold rigid. fixed, position which will reduce a ‘rowel having, oughly the quality of your ay diphthong (a kind of average of mid-point between its Seginaing and end) but with an absolucely xed, urmoving, tongue-, av, and lip-positon, 91. Make the tongue very tens, and prest its sides agniast the back molar seth, Hloid the tongue silently, inthis position, anchored, as it were, to the molar teeth, for several seconds ‘After you. are confident that you are able to hold the tongue rigidly in iM 1 this position, inate an airstream and produce a prolonged vowel, fist 18) we whispered, thea voiced. : ff } Testould be something oughly ofthe type represented phonetzaly ase. Mice sue'tha-your tongue is pushed wel forsards to ensure that Jou are producing etrly front vowel a¢wellas pwardeto maintincontact withthe molar tert, Usten and fel very cieall- Ifthe sound stangesinthe sights cst, or if there isthe slightest movement of the tongue during the uiterance u* of the vowel, stop and go back to the stage of silently holding the fixed ‘7 y? postion ofthe tongue anchored to the molar teeth Ts sorta ard on tllig ht Sad tng ston and J incidentally, by stalin asic were, bea 0 Voke yKeping almostallof your aflenton of the fied tongue postion, he ticki fo Drea te Asocation with that olé dpithonsd vowel sound, an the:vay todo itis 9 concentrate rateey on the artevlatoe, You hive ostyto yourself NOT. Tm pngacing ta wntype roel soon’, 3UT, Ti bolting a coral artultory pestere— aoe Fn going to Rody where sound comes gu wie oe" "We next-carry outa similar procedure with the diphthong of go, or better, the Solated:diphthong of Oh! This, again, is almost certain to’ be a.diph- ‘thong funless: you'speak certain varieties of Scottish, aofthern English, Irish, Welth, or Indian English), involving progressive raising of the tongue and a plogressive increase in lip-rounding as the vowel proceeds, The exact starting and inishing point of the ciphthong varies very widely according to dialect: you may be sayiag an only slightly diphthongal [ou} or avery iphthongal {ov}, whichstarts with the tongue low in the movth “and the136 ram caRDaNAL vowsLS (cvs) lips in a neutral position and ends up with the tongue bunched up quite high and the lips quite closely rounde¢: or anything in between. In any ease, the task now is to produce some kind ef [o}-like vowel articulation, pro- longed for several scvonds, with absoluely no change of tongue: oF tip- position trom beginning to end. Tzking your diphthongal English oli as a starting point, try, as you did with the ay-vowel, to produce a kind of average or mid-way [ol-sound, 9% Draw the tongue well back into the phurynx and keep it raised eather high (Gere Stould be some feeling of strain at the base of the tongue and just above the larynx.as you do this), withthe ips rather closely mundad. The ips should be as close as they wereat the end of your diphthongal o (or closen if you feet that your o really involves very tle lip-rounding, as some do). Now, silently hold this pesition rigidly: tongue bunched up and drawn back, lios well rounded. Continue to concentrate hasd on holding that tonsue and lip ily ieitisting an. air configuration while st ream, ané starting up whisper, then voice Once again, the aim mast be t bold a tongue and lip-pesitios: the voicing, aad consequent audible sound, are incidental. Concentrate almost entively on the articulation, and atevely casually and incidentally notice the vowelssound that is produced: nevertheless, be cuice 0 stop the performance at once if you fee! (oF heat) that the souad is aot a igidly maintained unchanging menophidionge vowel Some students quickly get to the point where they are producing a good maintained monophthong but then they spoil it by finishing it off with a brief diphthong/like glide away from that Sxed position, Watch out for that When you produce a long, monophthongal [e¢ € ¢¢] or [0.0 9.0 ol, as in Experiments 91 and 92, concentrate on holding that fixed tongue- (and lip.) position right up to the end. Let the initiation and phonation sop BEFORE you relax the tongue or ips. Now go back over 91 and 92 once aagaia, checking that you are not spoiling your monophthongs in this way. Itis, of course, not only the vowels of day and go that are commonly diphthongal in English, There are varieties of English ia which viewally fry vowel i ciphtonga, and speakers tee vaites mut be pret lly careful to dsipine tsi tongues (and lps into absolute sity in producing the Cardinal Vowels, In tbe expucineate tht follow, vbich ae designed to Kad de reader to discover llth CVs we wil repiuely wim against dipthongization, Bu tray be useful fo English-speaking readers torun though series of Engish vowels aloud, nd then sel tying to ‘TIL CARDINAL VowsES (Cvs) 137 iscover which, if any, of them are truly monophthongal ia their proauaciation, The following weeds provide examples of ull the front and bacie vowels of English (0) fe, heats (2) i (3) hay, hayek (8) heat; (5) haut (6) hat, eat; (9) boas (8) Aa, awed, (9) hoe, hoed; (10) hoods (11) who, who'd Nos 2, 4,5,7, and 10 occur only with 2 consonant aftee thera in English, which is why oaly one example of each is riven. In experimenting with these Faglish vowels, ry to cut olf the final consonant of nos. 2,4, 5, 7, and 10, saying first id, then (carefully preserving the same vowel quality) fi... and likewise heard and then fe. .., etc. In other words, try co isolate these (end all the othe:) vowels, the better to be able to introspect about ther, and feel exactly what is going on when you produce them. Note that nos. 4, 6, 7, 10 are virtually never diphthongal. Al the otbers may be, in various varieties of English Having determined which of your Eaglish vowels are diphthongal, be particulaily careful to avoid dipbthongizing CVs that resemble them 10 ‘some extent 2, The front CVs ‘We are now ceady to start acquiring the artioulations of the CV, 93. Peeduse CV 1 [i], Make your tongue very tense and bunch it up in the front of the mouth as ligh and as far jorwurd as you can without actualy pro- ducing a fricave Maintain the articulation silently and feel the following: the tip ofthe tongue isin contact withthe backs or roots ofthe Lower teeth the anterodorsum of the tongue is pushed up very high snd very foe forsard you can feel quite strong contact between the sides of the tongue ang che ypper teeth on both sides, foom the molar to the canines ihe eaaine ‘octh isthe uid upper tooth fiom the centce on either side Now veioe the voweland note that it sa very “shurp'fi}sound, noticeably higher and teoser an ‘Sharper’ than the vowel of English me, or heed. This is CV 1 ( We now proceed to CV2 el. = 94 Produce a tense, silent CV 1 [J and then deliberately lower the tongue a very ltl, Leaping it quite tense, and oeping the tip ofthe tongue in contact withB | I 1 E 138 THE canpuvaL vowsts (cvs) THE CARDINAL VOWELS (Cvs) 139 the backs o* roots of the lower teeth. Lower the tongue only to the point | where contact. with the canine and froat molar teeth i | “eel outact Between the sides of the tone Ane WHE BASE molars. This should \ ve approximately CY 2 (el. Slowly and deliberately move back and forth between this new. vey slightly lowered position, and the (i) postion, then try to fixate the tongue in the (¢] position, The tongue should be tense, tip touching the backs of the lower tecth, thrust well forward, sides pressing against upper molars, Produc: voice with the tongue fixated in this position, and maintain the resultant fz ¢ € € € ...] vowel for several seconds, The tongue must be rigidly fixed, the lips spread, and there must be absolutely no movement of tongue of lips during this prolonged [ee ee. J. Be specially ‘careful at che end of the utterance: don’t lett fnish with the tongue sipping ‘up to the {i position (producing [ee ... i), nor with the tongue sipping back aind down to the [2] position (producing fe ¢ ¢. .. a). Let the voicing stop before jou relax or move the toaizue in any way, Concentrate on the fixed tengue position, net'on the sound that is produced: let that be merely incidental, If you think about the sound too much you may find yourself ‘equating it with your (ell or {ex in aay, which is almost certainly a éipkehong. SGulyou cans Having consolidated the tongue-postures for the CVs [ij and {e) by repeated silent positioning of the tongue for these vowels, with occasional volving to see what they sound ike we move on to the halfopen front cv 31 95 Oneway.to approack [eis thie: silsadly articulate [i] then {e, several times concentrating on the perctived distance between the cose [] and the Aa: close (e: Now start at sient [i silently trop down to |e}, then try to crop down the same distance, Keep the tongue in a very front position. Repeat {hie manceuvee silently many Smet! tongue dose, tajselace,halfopen. When you are confident that you ban relay do this, cacry out the Stmte motes in a ‘whispes and chen aloud, crying all tae time to produce equal steps of tongue- lowering, resating in eual step of sound, n going from i]to[e]an from c} toe! The third CV, [¢, is closely similar to the French [e] vowel of béte.For a {ew speakers of English, particularly in the North of England, the vowel of head is about the same as CV 3, For most others itis not, Speakers of RP usually have a vowel between CV Zand CV 3in head, that is; oné that is foo high for CV fel. Maay Americans, specially in the Middle West, have:2 vowel in head that is about the same height as CV 3 [6], but not neatly front exough. On the other hand, such people often have a vowel in ad which, starts off rather cose to CV 3, but then slides backwards: type (e°2), [he = ‘24). So although these vowels of French and English may be some kind of guide to CV 3, do not rely on them. CV 3 is simply a tense, rigid, monoph- thongal vowel, two steps down from [J], as it were. This is what we tried to achieve in Experiment 95. 96 A further equal step downwards, with tongue stil pressed well forwards, {ves us the tongue-configuration of CV 4 [al, The commonest ecror here is to let the tongue slip back, so that the vowel produced no longer belongs to the front series. Remember that in going from CV 3 [s] to CV ¢ [i] the movemeat must be only downward slight lowering of the tongue andlor of the jaw—not inwaré at all. ‘e-wall be well'nw to experiment with the four front CVs fi-e-e-a] af series The points tébearin ming afe: (9 allare very front vowels: rom tals point of View they form a'single, unified, group, all produced with the tongue thrust as fat forwaird in the mouth as possible. (if) CV 1 (Js the highest frootest possible vowel, (iD CV 4 (al is a very low vowel—itmay bs thought of as the ‘loivest possible’ front vowel, though that is not so clearly defined as ‘highest possible’. A feeling of ‘lowest possible’ can be achieved by lowering the jaw somewhat (not excessively, front teeth not more that about 1 om. of 4 in. apart) and drawing the body of the tongue down to the bottom of the mouth, always being careful to keep it thrust as far forward as possible, At the same time, note that having the tongue ‘slow as possite’ does not mean that the tongue is hollowed: it must retain essen- tially the same convexity as it has lor all the other front vowels. Americans may find it helpful to note, that CV 4 (a) is someting like the [az] vowel that maay speakers of South Midland and Southern dialects use for (atin such words as time (taxa). 3 " 97 Having repeatedly switched (in silence) (rom CY 1 [Jo CV fa] to get a clear feeling of the two extremes of front vowel positions, now go from CV I to CY 4 silenty/in equal sens: CV 1,2, 3 4, Do this repeatedly, watching in your mirror, and concsatrating on fronmere, and the Teeing of equal steps Fiaally, d6 the same thing aloud: (i [e (2 (6, many times, Combine the feeling of equal steps with extreme frontness and rigidly monophthongal pronunciation of all the vowels. Repest the series many times, quickly ané slowiy, maintaining each position for a moment, and then fora longer period of time, Listen and fel in an extremely ertical way, The tongue must not slip14). THE CARDINAL VowELS (cvs) THE CARDINAL VOWELS (C¥s) Id] back from the extreme {roat position ast goes down, and each vowdl must, as already pointed out many tintes, be absolutely monophthongal. Ifyou find yourself diphthongizing ary of che vowels (saying, for example [t) instead of (e]. {5} instoed of c), immediately stop, and silently practise the acticulation of that vowel, then ‘eotatively whisper and voice it taking care to concentrate primarily on the fasling—not the sound—beesuse itis the position that you must 6on~ proprioceptively and tactilely perceived tone trol: the sound will take care of sell Watching in your mirror compare your Lp-postions with chese showa ia Fig. 41 3. The back CVs Assuming that the front CVs fe] [el [a] have been acquired, approximately at last, we turn to the back vowels 5 fo), 60), Tfol, § (ul ‘The frst problem is to make sure that all four vowels ofthe series 5-8 are as much retracted as possible, In articulating ll of them there must be a distinct feeling of muscular strain as you pull the tongue-mass as far back = cer ©)« = © FRONT Fig. 41 Schematic representation of Sip-positions of Cardinal Vowels in the mouth as posible. Even before carrying out detailed experiments on the back CVs sileniy articulate many times a set of vowels of the type [a] [fo] ful with toogue pulled as far back as possible 98 Start with 0), with che tongue absolutely cs low and ss far back a8 possible With the lips in a more or less autre) position. Then, move the congue {silndy) upwards, by equal steas—about equal, as nearly 88 you can judge, to the steps you made forthe front vowels—to three more positions, roughly these ofa} fo}and fu. Note that(9}, fo}, and [u] have rounded lips—progressively more rounded as you move up. Fig. 41 shows the lip-positions of the four back vowels (Compore these repeatedly with your own lip positions as you articulate the scties, and as you practise individual vowels. ‘We must now take 2 more careful look at the individual vowels 99 CV5 is, a8 we have seen, ss low and as back as possible, with neutral Hips Moving the tongue andlor jaw up one step, and bringing forwacd the corners of the mouth to form the typical kind of open rounding, articulate CV 6 f Make an attempt at this postion, then whisper it, and finally voiee it, The resulting sound rust be absolutely mosophihoogal, and it must sound! 33 well as fe? about one-shied of the way between the most open [e] ané the closest ful, AS usual, we cannot use the vowels of any particular language as an ‘absolute guide, but they can provide some hints. A typical RP is not far off CV 6: Americans American dialects the typical vowel of saw is much too open, Is iso, with ‘a vowel almost a5 open as (a), rather thon (o:], On the other hand, some New Yorkers havea vowel more like 2}: but here again, there are problems. First they tend to dlpishongize their (a), to something like [a]: and the starting point in what one might call “broad NY’ is much too high, so chat the diphthong is something iike (03), specially in words ending in a conson- nt, such. as caaght (xaot] or [koot]. French typically has a vowel of about the right height, but not acarly far enough back, in words like nore (u5t), where the two dts indicate a strongly centralized vowel These negative indications may help the reader (0 avoid some of the worst emors, however, have more difficulty, In, some Leavidy[3]in'a possibly ubcertain state, to be linproved later if necessary we progress to fo]. As this suggests, the learning of the CVs is a matter of making closer and closer approximations.142. THE CARDINAL VOWELS (Cvs) 100 CV7 [o} has the tongue mised’ by two equal steps from (el, or about two thirds of the way to [u. The ips are rather clovely couaded. Astuming that you can articulate & very close very back, very rounded [u), you should ‘experiment by silentty moving from [a] to (3) by equal step: silent fo {9}[0] [u]. Having arrived, by this kind of experimentation, at a rather closely sounded, very back vowel, about two-thirds of the way from [a] to {u), whisper it, and then voice it, maintaining a prsloaged, unchanging Jo00...] forsome seconds. “Most English speakers will ealize that CV 7 [o] is somewhat reminisoont of the vowel of go. As we have already pointed out,'2 monophthongal {o} type vowel very close to CV 7 is used by many Scots, northem English speakers. and by Irish, Welsh, and Indian speakers Speakers of RP and of ott types of American English have a ¢iphthongal sound, usuelly starting considerably forward ftom the back position, anc involving a progressive raising of the tongue, and 2 progressive rounding of the lips,.eg. {oul. Clearly, this i very emots from the rigidly monophthongal, Fully back, half-ciose, sounded CV 7. That is why suck people must try 10 acquire the articulation of this vowel, as suggested, in,2 manner totally uarelated to thoughts about the vowel of go. However, if one can accurately imitate oné of the moophthongal [o}type vowels mentioned above, that may be 2 \seful guideline. The French (o] in beau is of about the right height, but usually not back enough. 101 CVS fujis meciosestpossibe very back vowel. It may bearchedatyaswe ave already mentioned, from an English (x vowel but onl with great care snd conscious effort it must be as bask as possible (which English (iis far fom being). rigidly monophthongal, and strongly rounded. Anothet starting point for ariving atthe articulation of CY 8 {us fw, asin English wall ‘Whisper that word, several tes, and do the same with a prolongation,of ‘he intial (ws lenty make an¢ hold that (w}ariculation, It will probably, te dove 1o CV $ fu} Ht may requize some tersine and seme retraction, to ‘make it as back as possible. ae Figure 40, which’ we introduced in the last chapter, shows the eight pri- ‘mary Cardinal Vowels oa the commonly used diagram. You shoulé- now ge tarough the whele series many times, waile looking at the diagram, which you will find at the end of Chapter 6. 9\\% sie Gothrough the CVs from I to 8 several times, silently Remember, all are rigidly monophthongal. Be careful toumake.1-4 very front—don’t start THE CARDINAL VOWELS (CVs) 143 Sipping away fcom the front position as you get down towards 4 fa} —thrust the tongue energetically forward all the time, and take equal steps between Land 2,2 and 3, 3 and 4, Take care to make 5-8 very back—the tongue rust be strained, tiringly, back towards the pharynx—and there should be ‘equal steps between 5 and 6, 6 aad 7, and 7and 3. ‘Next, whisper the whole series, taking exactly the same precautions, but ‘row that the articulations are made audible by whispering, try to ser that the auditory impressions age as equally spaced as possible. as you say 1.2, 3 and then 5, 67,8. Finally, say the CVs aloud, taiang the same precautions. Having acquired a close approximation to the eight primary Cardinal Vowels the next step is to acquire the eight secoadary CVs. This is not iflicult, since it merely catails changing the lip-positions of the CVs. ‘Figure 42 shows the CV diagram with both primary and secondary CVs indicated. A short arrow points out froma the symtol for eack primary CV t that fer the coresponding secondary CX . a 0. ay = : — © oo Ne © ore be “oe ome 20 Fig-42 (Primary) and sevoudary Cardinal Vowels 4, Types of lip-rounding Before we go through, all the secoadary CVs, is necessary, to make two remarks about lip-rounding. The first point to, note. is that (as,e have already seen to some extent) there is normally 2 close correlation. between. tongue-height and degree of rounding: the closer the vowel th a mae144 THE CARDINAL VOWELS (CVS) THE CARDINAL VOWELS [evs) 45 “labial aperture and viee versa, Close vowels like [j] und (x) bave very close rounding—a very small aperture barely the diameter of @ pencil; opener ewes, lke fx} and [a] have much bigger aperture. In other words, corre- sponding to the different levels of iongve-height” we have close rounding, helfclose rounding, half-open rounding, and open rounding. The secon point is thatthe form of the lip-rounding often differs according to whether itis applied to bask vowels or to front vowel. 102 We can easily discover these to types of rouncing experimentally Slenty ound the lips as if for a very round [o] vowel. Nove that the corners of the lips are pushed in iowards the centre so that both lips axe pushed forwares ( ‘pouted’. They form 2 kind of small tunnel i front of the mouth. In this type of rounding you can feel that the cheeks are pulleé inwards, and also thatthe charne! Batween te lip is formed by the inner surface ofthe lins rather than their outer surface: ic is distioetly Yendolabial” (6. pp.'80“1), | Rouriding of this type is called "nner counding’, and iis characteristic of rounded back vowels like (u] and (0) . For the second type of rounding do aot ‘pout’ the lips, but vertically compeess the corners of the meuth leaving, 2 small central channel betwea the lips, of 2 slitlike fit ellipical shape racher than actually round. This type of rounding is called ‘outer rouncing’,wn¢ unlike the ‘inner rounding’ that we hive just seen, it involves the outer surface of the lips, and se may be called “exolabial’ (F pp 80-1). Outer rouiding is move characteristic of front rounded vowels like [y] and [2], though not all vowels of this type in all languiges tave cuter rounding. The specife type of rounding t© be ‘applied to the secondary Cs [y] [0] [| [2] 8, not usually speciied, but front rounding woulé be rather natural. In any ease, you should experiment with botk rypes 5. Secondary CVs ‘We can now proceed to go through all the CVs, primary and secondary, silently, whispered, and voiced. We abbreviate sevondary CV #8 SCY": 103 CV 15) + SCV 1 fy. Silently articulate [i], add close rounding (preferably loie outer younding) to forer{y]. Siléntly switeh back an¢ forth between (i) tad ()) maiy mcs taking cae to leep the tongue absolutely fixéd in the “lose froat position of [i all the time ote same with whisper and: then With voice Protong the SCV [y}) =i casein ty beri MHF NTE AN AS we have already poinsed out, vowels eather close to SCY fy), though, not so close and so front, inelude French w in lure, German « in Bike, and we might add Chinese d in yu meaning ‘mu to meet” secording to tone. 10 CV2 el > SCV 2 fal. Silently articulate half-close front fd (one stp down from [] in the series of equal steps that go from fi] to (al), remembering to kvep it very tease and very front: add liperoundling, a very Bt less close thaa that of{y] (balPclose outer rounding) to form [a]. Silentyavizch beck ‘nd forth between (e] and [o), taking care to keep the tongue all the time Tigidly fixed in the half-close front position of fe. Do the same with whisper ‘and then with voice, . SCV'2{0) iscloser aiid more front than the French eu i fe, the German, Gin schdn, 10S CV3 {6} SCV 3, Silently articulate halopes Front [i] (one step down from (6) taking grat care 10 make it very front and rigidly: mozoph- {rongal; add jip-rounding somewhat more open than that foe (10 form {0}. Sieny switch back and forth between (] and fe, taking ere toksep the tongurall the ime eisily xed nthe hall-onen fll Front postion of). CV 4 [a] > SCV < [a] Silenly produce the most open fo9! vowe [a Remember it must beyery ope, the tongve yng ight Gowa.on heoor of the mouth, but stl convex (the centre ofthe tongue aust ac be polled own), and as front as possible. Run silently or in 1 whisper tarcugh the whole froat series (ile) faland take care that you keep the sume feling tarolghoutpartoulary the Feeling that (9 belongs to exactly the same frort serie, wil the toaguesmats thrist forward in the out ‘way pulled bade, ag elo}. ‘When you are sure you lave gota good [a] position, hal the frig in that positon, and sradusly bring tae comers of your mou inwards somewhat to the open rounded position, ia reality a kind of verslly placed oval shape Whisper aad then vole this (el vowel, an yvieh te onenewabovest fe} not ia aay mpare it Go through the! whole frost series again, silently, with waispei aud with ‘voice, as often as'you nedd in order to feel that you really Amve control of all the front C¥s, both primary and secondaty. Now we ean turn to the back CVs, and derive the secondaries fiom the primafies ai we eid with the front ones 106:. Throughout all ofthese experiments on the back CVs make sur that they146 TRE CARDINAL Vows (os) really are back, You mast. in other words, maintain throughout that feeling of straining the whole mee of the tongue backwards into the pharya, CVS fo] + SCV'5 fo}. Having silently achieved a good maximally opea ‘maximally back vowel a}, slowly ‘round’ the lips as you dic for (|: that is, move the comers’ of the mouth inwards to form a kind of oval ‘open rounding’. Do this with whisper, then with voice, You will probably cbserve that thewe it really not much difference between the neutral position of the lips for [a] and the open rounding of {o]. Concomitantly with this there is also not a great dliference in sound. Nevertheless, there isa sligntdiference and you must become aware ofthis Speakers of British RP ane various South Enghih dieleets mey note that, SCY 5 [0] is not unlike the vowel of hot, cot, ete. This may also be true for some New England Americans, bit for most Americans the vowel of hot 0,0, 8 unrounded, and may noteven be very back: for instance, one Ms} hear such words pronounced witha vowel zearer to CV 4 than CV 5, as tha: ka, or nearly so For some Americens the neerest vowel to-SCV 5 {o] may well be the Yowel of azuphy, which is often pronounced as Something approsching {Kost with a rather open back rounded vowel (though not acually as o7en as SCV 5). For the remaining three back vowels the change from primary to sec: oncary i tected ty unrounding, rot rounding the lis. Some people ind it alittle more difficult to remove rounding from a given vowel than to add it, Consequently, you may find it useful to re-practise the deiberite, slow seat cowiding 2nd unrowding ofthe Ups before going on 0 experiment via convectng pnamary CVs [i [Of [yl 1 ther secondary counterparts (al. (¥], and (w). 107 CV 6p} SCV 6 {a}, Mace sure that your [o]is really back. and one whole step above a] (silently running throuigh the series(a}-[5}-{o]-[u] may help to ensure that itis about the right distance from fol), Silently articulate {o) ané slowly and deliberately unround it: relax the lips, then pull thé comers of the mouth back in the direction of a, smile, but not too far; the open¢ unrounded vowels (both back and front) do not have the extreme retraction of the comers of the mouth that is characteristic of the close vowels ji] and {ss}, Now unround (9} to {n), with whisper aad with voiee, zlking care all the time to retain the zongne-position of {3}. 7 ‘The SCV 6 [a] basa sound that is noticeably different from that of fa}, but while sayingit, with the lips consciously spread, not rounded. it may help to ‘THE CARDINAL VOWELS (CVs) 187 think of the sound of fo] as wel sits articulation. Say to yourself, in effect, “Lam now going to say (5), but keep the lips widely spread: ie, deliberately say p] through wide-spread lips The symbol [1] bas often been used to roprosert the English vowel in ‘words like cut an¢ bud Thisis a convention that goes back to the nineteenth ‘century, but it must be noted that, apart from a few Scottish dialects, 20 variety of English has 2 fully back vowel in such words ‘Some ‘varieties of American English have 2 eu vowel that is half-opea and almost completely ceatral, more appropriately represented by the sy- bol [a] (see below). In RP the cut vowel is usvally more open than [a], perhaps about half-way between open and halfopen, and very, much further forward. . Bearing this in mind, one must learn SCV 6 (4) purely by deriving it from [6], without guidance ftom vowels of English or any other language, 108 °CV:7 [0] = SCV 7 fy). Makea silent fol. Make sure itis really beok, and stcongly rounded. Siectly remove the rounding, “aking care to rtsin caactly the tongue-position of [o]. Do this with whisper and voice. Note thatthe place of articulation (the aarrowest stricture) of {o] ~ {v] is prety much ia the uvular zone, You can tet our artivlation of fo] ~ x] by producing voiced {0}. unrourding it to [v) ther sishtly closing up the articulatory channel, This should produce africative (1, oF thereabouts. IF you believe you have a geod (o], you can also aim at f+] by saying fo] through wide-spread lips Once you have achieved a good thoroughly. back [x] you might experiment by allowing it to slip forward a little, becoming a somewhat centralized (F]-Now place a [g| before that [#] and o [6] after it: [e¥el. This isa rather widespread American pronunciation of the word good. 109 CY 8 [a] > SCY 8 (us, Silently articulate a good close, back, strongly rourded {o]: remember, this cardinal rowel is very 6act (unlike most ver~ Jeties of English [u), and very close, and tee! Taking care to maintai that close back tongue postion, slowly and silenty remove the ipsrouncing. Do the same thing with whispered, and then voiced, [o]— i. Having reached {os}, mainain it, then experimentally clse up the articulatory channel a little The resalt should be the voiced velar ficative fr). ‘We have now conipleted our survey of all the CVs. it is important to consolidate these by repeatedly going through the primary CYs, silent, { AX es Gn ©148 rae carpivat vowsts(Cvs) whispered and voiced, always taking care to make oll of them rigidly monophthongal, to make 00s I~ very, very front, and nos. 5-8 very, very back, and co aim at approximately equal steps in going from close [i t0 open [a], and from op2n {a} to close [u]. I: takcs a great deal of practice to seach the stage where secondary CVs can be indepencently controlled. Consequently it is best to practise.thera always as derivates of the primary CVs, i. by saying {i] = [y], [0] > [W], te. This has the advantage, too, of firmly linking them to the primary CVs, with which, of course they arc identical in tongue-position, 6. Central vowels and other additions to the CVs ‘We now have to experiment with a number of otner important, or besi vouel-sounds which are not usually regarded as part of the CV system, though they might wal be. These are the exnteal vowels, fa 2 [0 [2] (] an¢ (§. (See Fig. 3.) ‘After one has accuired virwal mastery over the-extreme froat and exareme back vouels ofthe Cardinal system, itis useful to experiment with vowels exactly half-way between Front and back, Those central vowels can be arrived at by sliding back and forth between front and buck vowels and trying to stop half-way. Addisoral 2) ‘WHE CARDINAL VOWELS (Cs) 149 110 Close contcal i} and (u). Make asilent, very back [u remember, the tongue ‘ust be sirsined back tovards the pharyns. Now, while retaicing exactly the same lip-pesition slide the tongue forward to the very front vowel [y]. (Amatempt to switeh ftom inner rounding with 1] 0 outer rounding with [5] would merely be distracting in this experiment.) Side back aad forth betwees silent and whispered (ul and [y] many times, taking care to main- tain the strongly bunched wp, close, tongue-height throughout. Now, while silently siding back and forth, try 10 stop about mid-way between (u] sad OO). This should be the position of the clese central sounded vowel (3. Having stopped tbe tongue at this mid-point, produce voise end the sound should be fe, auditory (as well as anieulatoily) about bale-way derwveen (0) and (y), Repeat this procedure many times, and see if you can consist- cently arrive at approximately the same sound of [u. When you hive achieved this unround (u] to the corespotding unrounded close central vowel fi-Now eross-chesk your {by arsving at it by silently aiding back ‘and forth bewicen close unrounded front (J and back (wl), Stop halfway between thest two vowelpositions voice the sound and see ifit isthe same as the fi] arrived at by unrounding fv]. It ought to be. 111 HalGiose ceatal [3] and (0). Cary out the same type of experiment at the half-lose level. Repeatedly slide between hallclose back rounded fo] and hal-} FIS00 F2 700 fe] F1460 21310" fo] FL360 F2 640 Ir (u] F130 F71390° "ju 1250 F2 595 | ee | Figure 4 shows the frequenciés ofthe frst and second formants (F] and ij F2) ofthe primary CVs plotted against a feequency scale. \ Fig. 44 First and second formants of Cardinal Vowels Figure 45, on the other hand, shows the CYs on two Cardinal Vowel formants charts'(CVF charts) that is, regular CV disgrams (one for ‘unrounded vowels, the other for rounded vowels) with theaddition of lines indicating the frequenties of FI aad F2. For this purpose the unrounded ané rounded vowels must be charted separatsly, because of the differences between the fequencies of thie formants, pariculaly F2 ‘These CVF charts show how the formants relate to the articulatory con fgurations for the vowels, Fl has the lowest value forthe closer vowels, increasing as the vowels get more and more open. F2 has the lowest vale for back vowels, ineeacing as the vowels become more and more front. Comparing the CVF chars for unrounded and rounded vowels one can “rie CARDINAL VOWELS (Cvs) 155 Fig. 4S Cardinal Vowels with frequencies of Fl and F2 (Hz) superimposed alse! see-the very noticeable effect of lip-rounding upon the formants, icularly F2, and more especially for the back vows, Pjur ye pronounce a eed owe, somal you edaat ras he separate formants: the fardamental and the audible harmonics are all sounding together, so tbat what.you hear is simply a vowel spoken on a partcalae pitch (derermined by the frequencyof the fundamental) and of a panicular quality (determined by the frequencies of the formants, espe: ally F and F2). Asa rule, the individual formants that characterize the156 THE CARDINAL VOWELS (Cvs) VES AL vowsts (cvs) 157 vowel are not perceived as distinct sounds. The accurate determination of the formant frequencies of a vowel is normally cartied out by instramental analysis in a phonetics laboratory. 8. Making the formants audible ‘There ere, however, non-instrumental ways of getting to hear the formants, and itis interesting and useful to experiment with making the formants of one’s own vowel sounds audible, and this is what we do in Experiments 113 and U1. Enperiment 113 is directed to making F1 audible. 113 Youcun hear F1ifyou close yous glottis and tap your throst, in the manner cescribed in Exp. 11, Close the glotis tightly and hold chat ight closure for some time, While carefully main‘aining the glotal closure, and taking care that there is no other closure in the mouth, tap the chrost with 2 pencil, or bby seapping s finger against it and note the relatively clear note. of definite pitch that this produces if you do it propery: Having acquired the tech- nique with a tte practice, now slectly produce the front C¥s [i.e] [a}— ‘hats, silently go tarough the motions of articulating these vowels, tapping your throat, with your mouth in position for esch of the vowels in turn. IF your CV 1 {iss really cose one (asit ought to be) you may not hear avery clear note when you tap your throat while mouthing [i But the subsequent ‘vowel postions, e}{¢}{s} Sught o yield clear notes of aquite distinct pitch when the throats tapped. You will observe thatthe pitch of that note goes ‘up as you pass ftom fj through fe] and [e] to [al. What you are fearing is a tapeiteh corresponding to the frequency of FI, eising a¢ you pace From closer to opener vowels. Now carry out tie same experiment with the-back vowels (a (0 (1 [uk land note how the tapepitch goes down, indicating the lowering of the fre- quency of Fl as you go up from open (and unrounded) fa} to close (ard. _ rounded) (a). Experiment farther with roueding and unrounding vowe's, such as [e}[e), [e}{e, lul-{e), ste, noting how lip-ronading affcets, the formant frequency (ard, ence the tap-piteh. , Experiment 113 shows thet F corresponds to the parameter of tongue- height. The frequency of Fis lowest with close vowels, highest with open vowels. In addition, it is somewhst affected by lip-rounding. Rounding Jowers the frequency of all’ formants, the formants of: back vowels somewhat more than those of front vowels; and F2 move than'Fl. Hoving made Fi cucible by throat-tapping we now carry out an experi= ‘ment that will enable us (0 perozive F2. This can be done by whispering vowels. Waisper is a somewhat complex noise, but within the rather rich hhushing sound of whisper we can generally pick out the sound of F2 11 Go through the CVs ia a strong whisper, listening carefully tothe pitch of the dominant note of the whisper. As you go through the primary CVs—] (6116) @ ©) &) fo} (u}you can hear the whisper-pitch (representing the frequency of F2) steadily going down as you procsed from CV | to CV 8. ‘The most striking changes in the frequency of F2.can be heard (2) when you alternate between unrounded and rounded versions of whispered vowels Which are otherwise the same, eg. ((}e), fee) [NH] [OH{e] ete. and (Wf When you go ina whisper through a horizonual series of vowels. ‘You should carefully carry out the following experiment. Form a very back, stoonaly rounded, whisrered (ul. Now, takiag care to maintain the Lip-roundiag throughout, slowly slide the buncked-up tongue forwards (carefully etaining the close toague-position) from fully back [u] though a somewhat centralized (0, fully central fu, and retracted [f} toa fully frost [y} coing all of this in a whisper. Repeat this manceuvre several times, noting how the whisper-pitch starts quite low for (u] then gets higher and higher as the tongue slides further and fuer forwasd. Slide back and [orth several tines between the extreme back aad frent vowels [3] and {) [Now try the sume experiment with the close wnyounded vowels (wf) dt may be easiest 10 start by forming (u] then unrounding it to [uo] before starting the whispered forward slide toi). ‘Note once agsin how the feegiseney of F2(heatd as the whisper-pitch of the successive vowels) goes up aad up as you go forward from (a to, {I Now edery out simiiar experiments at cuerent vongue-elynts—with the halal Scres [o} fool, fr [3} (the halPopen series [oH{eH{) [a (BL, and the open series [a}-l. Experiment-series 14 demonstrates the relatioaship between F2 and the parameter of horizontal tongue-position. As you move from back to front the frequency of F2 gets higher ‘Take another look at Fig, 45. There yousee the actual frequencies (in Hz) of F and F2, The charts in this figure will be more mecningful now that ‘you have actually discovered the selationships between formant frequencies ‘and tongue- and lip-positiors by experimentation im your own vocal. tract. You will also’ understand way it is aecessary for unrounded and rounded vowels to be plotted om separate CV formant charts. Liperounding,:as158 THE CARDINAL vowELS (cvs) Experiments 113 and 114 demonstrated, has considerable effect oa formant frequencies, particularly on F2. Note how the frequency of F2 for the close back uarounded vowel [a] is almost 1400 Hz, whereas the frequency of F2 for the rounded vowel {u), with exacty the same tongue-position is only about 600 Hz Because of such great differences in values of F2, clearly the lines cepresenting [requency Rave to be cifferenty oriented in the two charts, Ttshoil be noted, by the way, that formant Frequencies are not absolute: the specif values forthe frequencies of the formants depend upea the size fof the vocal tract that 5S produciag the sound. The values given here (approximately my own) are correc for a moderately large man, A small woman's formant would all hve somewhat higher Frequencies (ineveased bby a factor of about 1.2), a child's would be higher still, and those of a man witha very large vocal tract would be lower. However, the values given here are areasonably useful average, Reader: with come musical knowledge will ebserve that itis quite easy to ‘cquate the series of laryma-tapping pitches (FI), and the series of wisper- pitches (F2), with musical notes, ether on the diatonic scale, e=ntred on ridéle Cor, perhaps more usefully on the so-fa(doh-ray-me) seale Using tho latter scale itis interesting to go once agsin through the series of whispered vowels, rom [0] to fy] (asin Experiment 119). - ‘Call the whispec-ptch fu] 608 and then move up the scale by shifting the Dunched-up tongue forwards, exactly ab in Experiment 114, Assuming that the reader has aot yet quite got the knack of prodcing ately Fully back (u] @s is higily probable) we may take it that his starting-point—his doh— vl have a frequency of about 700 Hz. With the starting doh at 700 He, the ‘Bigh doh (a SctvaUighr] wl Kit nats tie ts engeasey ati 1400 He, which, as Fig. 45 shows, is about the frequency of F2 for te close central vowel (a. By the time (y] is reached, the frequency of F? is about that of aor i of thesesond octave. ’ Figure 46 shows the CVF charts for unfounded and rounded cardinal vowels with approximate sobfe scales added. The bold itaic levers (4, r te.) indicate the dob-rayeme values of throat-taps representing FL-The plain eters, ete) indieaté the doh-ray-me values of dominant whisper pitch, representing F2: ‘You should experiment with these scales trying in each ease to make Jour scales start at about the poiot indicated as doh on the charts, emem- bering, of course, that if you shift the stating point (the do) all the other ‘TRE CARDINAL VOWELS (Cvs) 159 Fg.46 Cardinal Vowel charts with dobeayeme superimposed notes will be shifted accordingly. You will probably find it easier to work ‘with F2 (whisper) haa with FI (throat-taps) and in particular you may discbver that whisper-pitch is a useful guide to locations slog the ‘hori- zontal’ dimension: front, central, back. Starting with dob’ as about the furthest back whispered vowel you can produce on each level (close, half- close, half-epea, open) side slowly forward on that level aad observe that ‘youeda estimate roughly when you have reached the central vowel and then ‘the front one from the pitch of the whisper. Ths, in fac, is a good way of learning the central vowels.160 THE CARDINAL VOWELS (CVS) ‘THE CARDINAL VowELS (CVs) 161 Readers with a more tectnical knowiedge of music may want to equate the frequencies of Fl and F? with the notes of the absolute, diatonic (A, B, ©) scale, Table $ shows the ratio of the frequency of each note to the frequency of C, snd also gives approximate values for notes of the octave starting at middle C (261.63 Hz), thats, C, in the aoiation recommended by the USA Standards Association. ‘Sharps ate about 1.0595 times the corresponding natural note, Thus Cathe frequency of C, multiplied by 1.0595, that is 251.63 x 1.0395 = 277, D," = 293.5 X 0595=311, and so on. Using this table, one can caleulete the frequency of ary aote, and hence ciscover approximately how the formant frequencies of the Cardinal Vowels corres- pond to musical notes For example, Cardinal [i] has FI at about 240 Fz, which is about half the frequency of B, in the scale shown ia Table 5. That means that it is about the frequency of B of the octave below the one shown row (that iB), since the Frequencis of notes double within each octave. F2 of Cardinal [] has a frequency of about 240) Hz; that is alte more than Dy, In our table, the frequency of D, is 293.5. Double this to get the frequency of De (ofthe octave next above), that is, $87. Double it agsia to 1174, the frequency of D,; Double it again, and you have 2348, the fre- quency of D inthe tied octave above middle C, xamely D;,and thisis quite rear the frequency of £2 offi]. Taus we can say that the formants of fi heve these values: Fl about B, (the B below middle C), F2 about D, (the D of the third octave above middle C) ‘Other examples arf}, FI about 619, that is nearly D,* (622), F2 about 1900, that i, benween A" (1865) dnd B, (1975), [0] FI about 500, that is about B, (493.95), and F2 700, that is, about F, (698). Tables Bags Se USA Standards Aswciation Notation piantriniseiacisoameni ae eee CD Be Fe Ge Ave BGs Ratio of note frequency toC™ 1.000. 1.122 71.260 1.335 1495 1.682 1.888: 2000 porn. values for Totes parting at ladle Cin Haz 261,65 293.5 329.65 349 39192.440 - 49395 52326 9. Additional vowel modifications We have now almost finished with vowels, but a few poincs still have to be made, Although the system of Cardiaal Vowels is an indispensable wol for the practcal phovetician, it does not take account of all possible features that may characterize the vowels of particular languages. ‘Three mocifications of vowels not covered by the Cardinal Vowel system by itself are nasalization, retroflexion, and rhotacization, ‘Nesalization has already'been mentioned, so here we need only remind the reader that any vowel, whether Cardinal or other, can be proncunced with the soft palate raised or lowered. The raised soit palate closes the entrance to the nose and characterizes oral vowels, (} (] [e] [al exe. Most vowels of most languages are basically of this type, though ‘they often become nasalized in the neighbourhood of nasal consonants. Thus English (2) in bad is normally purely oral. In the word man, however, where itis Aanked by nasal consonants, it will commonly be nasalizéd, hence {mén}: In some languages, of which'French is the best-known exemple, there is series of obligatorily nasalized vowels that contrast with purely oral vowels ‘Thus, in French we have such word.pairs as paix ‘peace'|pain “bread! = (pe! [a beau ‘beautiful’bon “good" = bo bd} Retrofiexion 2s 4 modification of vowels most commonly affects cpener vowels. Thus, in dialects of South-west England and some parts of Midwest America, the vowel [6] in a word like Jar may be pronounced with retro- flexion throughout, as [fd (or in some rural dialects of South-west England (vag) where the raised symbol for a retrollex approximant indicates retro- flexion throughout the vowel. ‘To produce a retroflexed {6 simply say a long (a]-type vowel, and while tying to maintain it, deliberately curl the tip of the tongue up and back: wards a littl, and note how this modification aifects the quality of the vowel. Carry out the same modification with other vowels, e.g. fed fod [e4] fog. Rhotecization: The term ‘rhotacized’, which means, in fact, “-coloured’, thas been appiied to the very peculiar sound represented by ir in the word bird in 2 common American pronunciation of this vowel. This vowel was formerly described as ‘retrofiexed! but thit ie not a correct description. Tt does not usually have the upward curling of the tongue that is characteristic of retroexion. Instead, the main body of the tongue is bunched up into? 162. THE CARDINAL VowEES (Cvs) Kind of balf-lose-central position, but with two peculiar modifications: ‘one modification is a moderate degree of deep pharyngalization: theroot of the tongue is drawn back into the pharynx just above the laryax, The ‘second modification isa fairly deep depression in the surface of the tongue opposite the uvular zons, This sub-uvular concavity can be acquired as fellows, Produce a uvular tril, Noie that in ofcer to do this you have to form a longitudinal furrow in the tongue within which the uvula vibrates, Now move the whole body of the tongue dightly forward, while retaining puccisely that deeply furrowed configuration, The result should be a.close approximation to the typical American ‘bird vowel’, for which the phonetic symbols f=] and (»] have been used—both representing a ceatral vowel with ie modification. ‘As we saw, this very strange American vowel imolves not only” a cconcavity—or ‘sulcalization’ (Grom the Latin sulcus ‘a furrow, or treaca’}— of the tongue in the aeighbourhood of the uvula, but also some slight degree of pharyagelization. It is because of this that a series of vowel sounds with. modifications of this totacized. type in some languages. spoken in the Caucasus area of Russiz, notably Tsakhur and Udi, are enown as ‘phacyngalized’ vowels + ‘These additions to the basic parameters of vowel classification complete our survey of vowels. cotdete 9 Prosodic Features au 1. Initiatory prosodies: Initiator power (= stress) Up to now we have been looking at speech-sounds as isolated phencmens, In reality, of cours, spoken sounds occur strung together, one after the ‘other. More precisely, speech is a continu; a continuous fx of initatory, ‘phonatory. and articulatory states and movements, constantly changing, often overlapping ad interpenetrating aad influencing. each otter. When ‘we look at isolated sounds we arc artificially cutting up that flowing chain of events: into 2 series of segments or segmental sounds, as we sometimes call thespeech-sounds that we isolate out of the continuum, ‘Although the segmentation of speech is an artificial procedure, we are obliged to do itto arrest the Now, as it nere—in order to pix down in vidual sounés for detailed study. We must, however, also give attention to those phonetic phenomena that are characteristic not so much of the inci- vidual segments as of their relations to cach other, of of stretches of the speech-coatiquam, that are greater than ous sepusout in length. Because such phenomena take account of mor: than just segments they are some- timer called snprasegmental features. Another name for them is prosodic features, ot prosodies, and these are the terms we shall use. ‘We have already noted in passing that the prosodic features of speech can bbe seated to the tree basic components of speech—initition, phonation, and articulation, We thersfore study ther under these three headings, The prosodic features siress-and the division into syllables arc features of initiation; prosodic phecomena:of pitch are features of phonation; and the duration of sounds is a feature of articulation, “The iniiccory prosodies, that is stress aad syllabifcation; relaie to.the way in which the bursts of muscular energy that drive the initiator (the ‘organ that initiates airdow for speech) art organized and distributed,164 pRosopic sRATURES PROSODIC FEATURES 165 We saw in Chapter 2 that stress een be defined as initiator power, and, moreover, that this definition is applicable to all ypes of initiation. 11 is possible :o initiate 201 only pulmonic sounds but also glottalic and velaric sounds with grester or lesser initiator power, that is, grester or lesser stress. Nevertheless, stress as a prosodic feature of langvages is virtually always manifested in terms of pulmonic pressure intition. ‘As we saw in Chapter 2, itis desirable to begin learning about stress by experimenting with voiceless fricatives. Tai isthe best way of “getting the feet’ of the difference between stressed aad unstressed syllables, sO we start from there, 145. Saya prolonged (5335. J. Now deliberately superimpose soiné variations of stress upon that prolonged {s 5 5 ...). That is, while continuously producing this voices friative sound, vary the amount of enerey you use in initating the airstream. Deliberately drive the air ot in an alternating series of stronger and weaker Bucs of initiator power: Represesting moments of increased power by bol! capital] we can indicste the alteraa- tion of initiator power as (ss SssSSésS Sete, Experiment with vared pattems of sress:[SSssSSssS5si, 6SssSsSSh[ses5ssSS] cre, 20d with different voiceless fecatives: (PR ITP 9, [fF FER LEE By ete ‘Now try the same experiment with sequences of voizeles steps, [> PM Ph ph PM PH ete, Note that, whereas ibe sequences of voiceles iticatives could be producedwhileretaining the ariculation unchanged throughoutthe Utlerance ofeach Set in the ease of the scors tis necessary to release each cone into 2 momentary pul of breath (symbolized here by superscript (H)- The IPA wiacritic for stress is a’short vertical line ['] placed before the symbol forthe steassad sound ar yilahle Phu the example in Feperiment 113aare properly transcribed as [sss 85 'ssss's] ['ss9'ss ss, etc: Note chat this convention is different from that used in many dictionaries and other publications where the stress-mark, oftenn accent ['] is placed on, or after, the stressed syllable The IPA practice of placing the siress-mark before the sound or syllable to be stressed: is of some practical advantage. It means that as you read along/a phonetic transcription (which students. at first ‘canaot do very fluently) you are being wemediin advance: “the next syllable is to be stressed’, The IPA supplics diaéritcs for various’ degrees of stress, But for the present we shal se only [], meaning ‘stressed’ as opposed to “unstressed! (which is unmarked): "As you otperinatad ith [va 619": ta: fou probably aon that scressed [5] sounds distinetly:louder than unstressed [s]. This is because the velocity of sirflow through the aniculatory channel is higher for the stressed sound, and consequently the turbulence is greater and the resultant hiss is stronger, that is, has greater acoustic intensity. Any particular sourd- type appears louder when stressed than when unstressed: thus a strongly stressed ['s] souads louder than a weakly siressed (unstressed) [j], and 2 strongly stressed ['] sounds louder than a weakly stressed (f]- However, the range of loudness varies from one sound-type to another thus a strongly stressed 'f] niay sound no louder than a weakly stressed). Try this, saying [] and then (3) with, as nearly as you can judge, the same initiator power: the 'g] will probably sound a little louder than the ('f]. Then experiment further with strogly and weakly steessed ('), (5). CIE and compere the apparent loudness of these sounds. You will probaoly discover that, fOr the same degree of initiator power (Le. stress), [J] sounds louder than {and [5] sounds louder than (f. This shows that, while all speech-sounds exhibit varistions of loudness under different conditions of stress, nevertheless some Souds aie always louder than others for a given dearer of stress. Speeck-sounds differ from each other in terms of their great sonoriy than fi the seme initator velocity (same stress) fa} sounds louder than (i, and agzin, a strongly stressed [i] may sound a0 louder than a rather weakly stressed [a] z ‘This is why it.is unwise to talk of stress (as some people have done) in terms of degrees of ouiness,-since loudness is in part a. product of the inherent sonority of sounds. It is much more reiable to think of-strss entirely in terms:nf degrees af initintar prnden—the amonnt of nergy expeaded in pumpingair out ofthe lungs Incidentally, we often talk as if there were.two cal distinct degrees of stress—ealled ‘stressed’ and ‘unstressed’. This, of course, is not the case, Initiator power is infinitely variable, frdm zero (when theinitiatorisinactve and, consequently, there is no airflow and.no sound) to an indefinite maxi- mum (depending oo the size and muscular strength of the speaker) when the initiator is operating at full power, foreing the air out at the high possible velocity against the resistance imposed upon the airflow by phox tory and antculatory strictures, However, many languages make meaning ful, inguistic, use of only two distinct degrees of stress, namely a relative! strong stress, characterizing what are called “stressed” syllables, and a rl tively weaker stress, characterizing what are called “unstressed? syllabks j166. erosonic FEATURES PROSODIC FEATURES 167 chere were a clear-cut This is one reason why itis often useful to talk as if distinction between stressed and unstressed sounds So far we have experimented chiely with voiceless sounds. There is ¢ ‘good reason for this, namely, that with voiceless sounds it is particularly easy to get the fesling of strese—that is, to become aware of the kinaesthetic sensations associated with the production of different degrees of scress ‘There's also enother reason that we will refer to agaia in a moment, Having acquired a basic feeling for stress as initiator power, we must now experiment with producing voiced sounds with varying degrees of stress, 6. Saya prolonged (e2¢2222] and tenimpoe diferent sueswpattrns upoe i justasyou dd wit 6s) ia Exp. 115, tmas:('22°22'22] ec. You may find it easier to separate the successive [z]-segments by a glottal stop, thus: (202 220] ete. but yoa eet not doth Continue wih varying pattems sucha: (e'22 22°2, @'222 222, (Ve'v vveu Net, ckperimet With vowel prefiably separated by coxsonant i suet sequences a6 la ("a a aba baba bb} “HU "Til, [bi di Bi bi bi DiJete. Tis a litle more dificult to be aware of the feeling of stress-variations with voiced sounds, and you may find it helpful to try whispering the cxamples of Experiment 116 as well as voicing chem, ‘As Thinted above, there is another reason why itis simplest to begin the study of stress with voiceless frcstives. This is the fact that with voiced sounds, particularly vowels, attention’ may be distracted from stress by she presence of concomitant differences of pitch,” and. in -real-language ‘examples—for example, in English—there are often cifferences in duration 5 well, These concomitants of seress tometimes confuse people since they ‘cad to conceal the true natucr of sicess, as simply initiator power Lf you experiment briefly with stressed and unstressed voiced syllabies, such as fla la la Ia}, la“la-la “lal '4a da da}, [ds ‘da dal. ete. you will probably find that you pronounce the stressed syllables —partculady if you siress them stiongly—on a higher pits: than the unstfesed syllables, and “You miy ako make them logger. There's a satural, physical, Comnection-between stress and pitch, s we shall sce below. Consequently stresieg syllables are often pronounced on & iffercat pitch from neighbouring unstressed sylables, I you say the Eag- lish words ‘abstract (noun), ab'straci (verb), ot ‘pervert (noun); per'vert (verb), you will notice that in the nouns the frst syllable is stressed, whereas high pitch: the pitch falls and the second syllable is quite low-pitched. In ‘represent these pitch-schemes as shown in Fig, 47, notice, incidentally, that the unstressed syllabies tend to be of shorter dura tion than the stressed syllables. You can, incidentally, get an even clearer saying: noun [ss], 96rd [s'5}. ee Nie ener pewver Fig. 47 Pitch.schemes of nouns and verbs Tt is indeed, quite easy to manouvre th: English words into a position ‘hire, in normal Eaglish-specch, the pitch diferenoes are eliminated. This ‘bappeas for instance, in such sentences as "fsaid-abstract ot "I sald pervert, ‘wheré the Zis spoken very emphatically (the double sres-mark ("] inate ing extra strong stress) meaning Fsaid it... not you. Here there is strong stress and » highyElling tone on 'T. Tn such a sentence the quoted word at the end is pronounced on a uniform low pitch, but its stress patter is preserved, We'can-iadvae che pitch pattern aid the tresses of suck snteabes 3 showin Fig 48, “Ty saying these seatences, and note how the stress differences are poe: sereg ia abrtract'vs:ab'sroer etc, even though those words are spoken on a.monotone-Thus, though pitch phenomena and stress are often reloted to each other inthe pronvnciation of a language. they are in fact distinct and isolatable features which are independently controllable, The fact that in many languages, of which English is one, stress is associated with pitch and curation has led to some confusioa. ‘Is stress realy pitch, nonde ne168 PRosoDIC FEATURES = yLvoe ooo cn ar rs Ny wo tainty gC “1 sée sara (er) ee “16 emer (nus) 2 “isd zerver (er, Fig, 48 Stress differences on uniform pitch gn been aled. The answer is that all ofthese fatuee pan uration sents of mterace is wall as che inherent sonoiy ofa seund may seas tte the Peeeptal promtience ofa sound or slable marc we degree to which it appears to stand out ftom ts Aeighbours However, —figm a general phonetic point of view: to prevent setfusion, it's best to Sonfine the application of the term sires 1p initiator power as we have defined it here 2. The syllable Wie have frequently used the trm syllable without defining it, and we have sreena, 2 implied, that stress is feature ofsylables eather than of seg rectal sounds This indeed, is why stress is ecated as a prosodic rather than a segmental phenomenon: FROSODIC FEATURES 169 sn dmitihing Somewhat, then peaking gain and soon tn thie uterance aaa eute coatinuouy unbroken [}-sound,butthercis alse ofhot or eons, besause the 1ouduest of the his ree and with the 2 ements ofthe iitiator power. Tn oher words we get e fecling of separa piesa serie of poner peaks, or peaks of iitinne velocity UT’ Sey a prolonged usbroten (2 4, gering te feeling of a singe, very Jone ilaie. NeW ay be sme sound oa a sores ofp sts, oF surges, of wine aa S2 Mat YOU get the fcing ofa series of syebie (but Mabou ee" interrupting the continuous is of 1} only ing its loud- Be) Ess... Now do the same wit thei pronounce [22 Jwitha pulsating {Satis of syabio fe}. Finally co the same thing wth 8 vowel, say [a3 sacepritis stots be no break here between the vowel fo na, insert a TAR Flsst,say'a series of syllabic (s-sounds, as in Exp. 116, separated only by Cinder Titttations of the infiator i. a contimwous unbrolen [ihtise vided by pulmonic pte into aseresofsylaties a vee 8.1 Now Produeea.prolonged (9, this time broken up by.a sees Off bests Fr ce set UWH Bettie impreston of sees of somat syllabic {tsounds but his tne the division into sabes is nponne Dy the succes. ie AP&2 apa. Jlasasara.Jespassasyi dive micartve then, ata definition of the syllable 36 ¢ mivina) pulse of HOD activity bounded by a momeniary rtarcaton oF the initiator, f iter stimposed, or, more, usually imposed by a consonantal type of / articulatory stricture, Frery syllable mas a central pare, or nucleus this is most commonly a170 paosoDIc FeaTURES ‘vowel, as in English cat, stop, slept and so on, but occasionally its asound that we would usually regard as a consonant, such 28 the ] that forms the nucleus of Czech (krk] ‘neck’, or the syllabic [a] that forms the nucleus of the final ayllable of English seldom (sadn). Since the syllable-nvcleus is not- rally a vowel we shall represent it here as V. A syllable may consist of nothing but the aueleus, as, for instance, the English words awe, ah, the name of the lever £, and so on, which we can represent as simply V. On the other hand, there may be a consonantal beginning or end to the syllable, which We can represent as C. We can thus indicate various Kinds of sylable structures that occur ia English as: V, eg, awe att; CV saw go; YC eat up: CVC cat mad, ete. TF there is more thaa one consonant at the beginning or end ofa syllabi, we ean shew this by simply adding at many Ce as may be necessary, thus: CCV play stay, VEC apr end, CCVE siog brows, COVES stopped {stopt] plant, CCCV stray serew,, CCCVCC strand, ané so oa, Lan- ‘guages differ grectly as.to the types of syllable srvcture that they permit, and we. will refer to this agein whea we discuss the sound.systems of languages We have now seen that, in the course of speaking, the activity of the initiator —most commonly, indeed nearly always, the pulmonic initistor—is. parceled out, as it were, iato short chunks, o: small peaks of inidator power separated from each ether by slight retardations of the iniviator. “These chunks are syllables, and the momentary retardatiots that mark the boundaries between syllcbles are mos: commoaly imposed by consonants, but they can abo.beselt-imposed. Moreover, cach syllable may be produced vwith greater or lesser iniator power, or stress asit is called, than its neigh- tours, The word’ phonetics, for instance, clearly has’ three syllables—is delivered in three chunks of initatory aétivity,"sach bousded by conson- ants, tus: CYCVCVCC. Moreover! it is‘clear that the second syllable is Gelivered with greater initiator power than the others, and consequently we eseribe i as stressed, and so we can represent the syllabic structure of that word as CV'CVCVCC where 'CV marks the stressed syllable, and Saal, wwe can transcribe it a5 [fe'aetuks) “There is onc further matter to discuss coriseming prosodic fectures associated with initiation. This is the question of what is-often called the rayih of speech. This is parly a maiter ofthe alternation or distribution of short ard long syllables, but itis also a matter of the timitig’of the initiator pulses prosonicrearunes 171 3. The foot In many languages, of which English 8 « good example, in addition te syllables, such as we have discussed, intiatory activity seems to be pareelled out into relatively equal chunks that are often longer than the length of a syllable, Tiere are thres English seztences with 4 siressemark [ inserted before cach stressed syllable: 1. That's what John bought ‘yesterday ‘morning. 2. That's the ‘book John bought 'yesterday ‘morning. 3. That's the ‘model Joon bought ‘yesterday ‘moming. SWe'tend-to deliver each of these sentences in four bursts of initiator activity, Each burst, or ‘chunk’, bas the initiator power rising to 2 peak in the syllable that is marked 2s stressed, then the power falls off to a low value then eises again to a peak at the next stressed syllable, and so on. Representing the rise and falls. of initiator power by a rising-falling ling and thedivision into chunks by vertical line, we can cepresert the initiatory pattern of these sentences asin Fig. 49. Ifyou read these sentences aloud in a properly English way, with what we describe as-an English shythm, you will probably get the feeling thatall the “chunks are in some respect, equal, oF nearly so. Consider the second one in particular, In seatexce (1) i contains two syllables, in (2) it bas three syllables, and in (3) ites foue. As you say each sentence you will notice that there is some attempt wo pach al tiuse syllables, two three, four into aha the same space of time Not exactly the same of course, but there fs obvi- cously’a tendency in that direction—a tendency, that is, to give about the. sim Aghount of tne to @ach chunk. At the ame time, you may fel. as you 1 ct) Joba boat vowing. 12] “ctharsihe” | boskohaboupht | yeserday | _moming 3° Tacs mvdeiJohntought || yeterday | “mering, I I Fig. 49 Stress-groups, or feet172 pRosopic reaTURES Paosonie reavuRes 173 repeat the three sentences, voiced and whispered, tht you expend about the same eneegy a each chunk ich ofthese chunks is what we cll a‘stresgroup or “oot” and you will notice that each contains one major stressed syllable, because the dis- it follows approximately the curve shown, English spect tnd to be delivered in a seria of fc, each’ containing from one to several sylables, @0a wih a power cine of the ‘type shown, Within short stretches of speech,.feet in. English tend to be } equal dartdog) al tolincras (oehmvig asak tribution of initiator power within each f lsookrencnas the same output of initiator power) : Notall languages are like English In Freneh and Japanese for éxample, the parcelling out of initiator aower is done differeaty. In these languages, if there is any sozhrony (equal timing) or isodyramism (equal powering) it applies eather to syllables, then to those longer stretches called “fet that ‘may be several syllables in leng 4, Phonatory prosodies: Pitch variation (=intonation and tone) ‘We'vun now to those prosodic features that relate to phonaiion, These inclide features of voice-guality, that may, in some lsnguages, affect stretches of speech longer thai the segment, and features: of pitch. We concentrate here on the second type of phonatory prosody, pitch Icis obvidus that the vocal folds caa vibrate at different frequencies, and ‘that voice can thus be produced at different pitches 119 Experiment with deliberately varying the pitch of jour volte. Keeping preity much within the normal compass of your speaking voice (unlike Exp. 12, where you had to go as high andas ow as possible), say the syllable [la] om @ low level piteh, on a sting pitch, and so on. This is just to verify that you can deliberately control the pitch ef your voice ‘Change of giteh is produced in two ways: (a) by stretching and teasing the vocal folds—the tenser the folds the higher the pitcl. (6) By changing the pressure below the vocal folds, the subglottal pressure. The higher the subglottal pressure, the higher the piteh. “You probabiy won't be able to fee! the tension of the vocal folds, but if you say, or sing, a series of noves, starting low and going higher and higher with, as far at possible the seme initiator power throughout, then you can |e be preity sure you are raising the pte by tensing te vocal elds 3 ‘The second method of changing ptch—charging subgiotal pressure— can be demonstrated this vay, Say. or sing. a very prolonged vowel on one unchanging pitch (any pitch that is comfortabe for you). Arrange with fiend tht, while you are producing that long unchanging note, he should suddenly give your chest (at about the level of the diaphragm) a short sharp squeeze. The result will probably be a short sharp rise in pitch. This demon- strates the ces0nd method of pitch chang: ‘The short shacp seuecze rmomentarily increased the pressure of the air below tae vocal folds, and this sudden increase in subglotta air pressure caused the vocal folds momentarily :0 vibrate faster and produce a higher pitched note. This demonstrates in simple way that relationship betweva pitch and pressure that we referred to above ‘This is why strongly stressed syllables have a tendency to be pronounced on a higher pitch than unstressed syllables, But this tendency may be overrled by spesite ruler of a’ particular language. There are languages, including dialects of English, in which most stressed sylables are actully Pronounced on a Jower pitch than the neighbouring unstressed syllables. This is true, for example, of the English-of Aberdeenshire in Scotland, and some varieties of Welch English, Jumaican English, ard lnciaa Enaish Pitch and pitch changes are uelized in languages in two distinct ways, On the one hand, variations of pitch may be related to relatively long stretches of speech, which may be many syllables in length, and which correspond to relatively large gramtmatial unite such ac the centence Pitch variation used in this way i called inonation. On the other hand, the pitch variations of a Jenguage may be-related to short stretches of speech, typically of syllable length, and to small grammatical units’such as words and morphemes. Pitch variation used in thie way is called sone. Exainples of izonations are the pitch patterns ofan English sentence like the folowing, for which the intonations are indicated by means of a rather obvious notation in which lines represent the pitch levels of the stressed syllables chat oocus atthe Hart of each foot, and the dots yeprewent the unstressed syllables that follow them, high Fotting $+ 2ee+ piteh Sais was bere yesterday fs : XE t 2. “ane was here ‘yesterday? (__ ..=.7] 3. ane eae Re on174 PROSODIC FEATURES ee ‘inthe three versions of ths sentence the changes inthe pitch pattern or intonation contour in no way change the mening ofthe inéivicual words in the sentence—not eves of the word yesterday, which is the carcier.of the major pitch differences, What it does do, is change the function of the sentence as a whole, from 2 statement (1) 10 a question (2), to a question ‘with incredulous emphasis on ‘yesterday’ (3). Using terms thatare currently 4 rather popular in linguistics we might say that intonation, as kere, has a 27 | pragmatie rather than a semantic fonction. 5 Examples of tones might be the following. words in Mandarin Chinese. Here the pitch movement of the tones is indicated impressionisticaly as the intonational piteh movements were in the last example, ¥ bal — J'eight! 5 2 bal }'t0 uproot’ 3. ba[ ~]‘t0 bold” “4, bal ~ J'aharrow tis clear that this is quite different from the last example, because here, fist, the pitch. differences. occur.on'a single syllable, and secondly, they ‘completely change the meaning of the word. In other words, fore, as in these examples, has a semantic rather than a pragmatic function, and is a permanent or inherent feature of the word. 5. Articulatory prosody: Duration of articulation (= length) “The third basic component of specch is articulation, and this is particularly relevant to the prosodic feature of lengch ot duration. Of course, i is clear that in speech both initiatory and phonatory eveats can be shorter of longer. You ean obviously produce a seres.of pulmonic ‘pressure pulses of longer or shorter duration. Going back for a.moment to ‘Experiment 117, you can pump out that series of syllabic (s-sounds very npidly cs <<<... ].orinamoreleisurely fashion &.3.3.3...Lete. And itis clear that, during spesch, there cam be longer or shorter stretches of ‘voicing or voicelessness. Thus, in such a sentence as They stopped splashing [Ber ‘stopt'splefin] we have a long stretch of voicelessness ist (-ptsp-} in ‘They stopped talking the stretch of voieelessnessis shorter (-ptl in They all {alk {Sex 2 it's quite short; only during (9). But differences in duration Prosovic FEaTuats 175 ae ‘of 2 phonation type are not what we normally have in ming in talking about length, cr duration, ‘When we talk about duration in phonetics we are referring to the duration of particiler articulatory postures, In Chapte: 4 we saw that there were wo | types of stricture that are essentially momentary, namel fs flap and semni- vowel, A Pap (or tap) is an approach to. immediately followed by a depart- tare from, 2 momentary contact, IF you prolong the contact the sound wil no longer be a flap, but a stop. A semivowel is likewise an approach to and/ or departure from an approximant position: if you prolong that position you have aa approximant, or approximant-type vow fj (ul. ete. ‘All other stricture types arc, as we saw in Chapter 4, mainteinable, Its these strictures that can be maintained for a shorter or longer time, ie can have diferent durations Thus we can have shorter or longer stops, {fricatives, approximants, vowels, et. ‘We already noted earlier in this chapter that there is some tendency, at least in Engish, 10 make stressed syllables (or, rather the vowels that {orm their nuclei) longer than unstressed syllables. It is, however, perfectly possible to control duration quite independently of stress 120. Soy a series of syllables sch a (Ta Ia ln Tala...) alternating stressed and tunstiesied. Now, experiment with verying the daration ofthe syllables The IPA symbol or long duration isto dois placed after the symbol for the teagthened sound: thas: [i] [e] (uJ, ete. eepreseat short vowels. (:] [6 (8) repeesent long vowels. and fs}, ete. represent short consonants, [sl] ei. represent long comoran's. ‘Now say a: la asa ln le] and then [Tala "a ae la Te] and ae les “a. J ete. In languages Uke Czech ai Maugacian, both of whisk ‘contrast long and short vowels and both of which have stress on the frst syllable of a word, such stress and duration pattems as ‘CVC “CVCYV ‘CVCV 'CVYCVY freely orcur. lala10 Sound-systems of Languages 1. Phonology: The utilization of speech-sounds Tn the preceding chapters we have talked bout specsh-sounds for the most pact-as things-inethemselves—that is, as so many isolated physiologico acoustic events without much reference to the way in which they are used in ifferen: languages. But now we mast look at how sounds are utilized—how ‘they are organized into the souad-systems of languazes, ‘The scudy of the physiological, acrodynamic, and acoustic characteristics, of speech-sounds is the central concem of phoreties. The study of how sounds are organized into systems and utilized in languages is the central concem of phonology. Neither of these two linguistic disciplines is independent of the other. A knowiedge of what featares of sound are most utilized in languages cetermines what aspects of souné production ate most worth studying in depth. Thus phonetics depends to some extent upoa phonology:to indicate areas of linguistic relevance end importance. Phos ology on che other hand, is heavily dependent on phonetics, since itis phonetics that provides the insights that enable one to discover what sound features are linguistically utilized, and it is phonetics again, that supplies the terminology for the desévipticn and clatsifcction of the linguistically relevant features of suinds ‘Apart ftom very ninor énetomical differences all human beings have the same vocal apparatus, Consequently, all human beings are capable of pro- ducing the same sounds. In other, words, the human sound-producing potential is universal, itis also capable of infinite variation. Consonaatscaa be articulated in a great varity of ways at any of the locations. within the voral tract described in Chapter 4, orat any intermediate point. Vowelscan be articulated anywhere in the entice vowel-space defined in Chapter 6. However, out of the.infinite range of possible sounds, every, language178 SOUND-SYSTEMS OF LANGUAGES utilizes oaly a Fraction of this rich phonetic potential, and, moreover, languages difer greatly as to which parts of this potential they'make use English, for example, hes the glotal[rcative h] but otherwise makes rig use at all of the whole pharyngo-laryngeal tract in the articulation of ‘consonants, By contrast, Arabie expleits that part of the vocal tract much, ‘more extensively, with the pharyngals [5] an¢ {B] and the glottals 7] and (2), and there are a few languages which make even more vse of that area, for example, the Burkihan dialect of the North Caucasian language Agul, ‘with no fewer then seven consonants articulated in the pharyngo-laryngeal (On the other hand, Arabic has only fwo labial otstruents, &) and (0, ‘where English has four, (pl. (b}. (0, (V). French has three rounded front vowels (y) (@} (@), and Mandarin Chisese has one such vowel, [y], but Engish never combines lip-ounding with the articalation of front vowels, i.e has no rounded front vowe's. Engish makes absolutely no use of pitch to istinguish one word from another, Bat Chinese, for example, make considerable use of pitch for this purpose, with 4 word-tones in Mandarin, and 6 tones in Cantonese. 2. Acontinuum of vowel-sounds Not only do languages contrast with each other in terms of the different parts or aspects of the phonetic potential that they utilize they also differ in the ways in which they divide up any’ particular range, or coniinuurs, of possible sounds. What is meant by ‘continuum’ m this context wall be made clearer by Experiments 121 and 122. 121 Produce a continuum of vovel sounds by stacting at CV 1 fi and then care- fully siding along the periphery of the CV disgram in an unbroken slow Bide. Keeping the body of the tongue bunched well forward in the mouth, 80 a§ to retain a truly /tont fostiod tiroushout, slowly and continuously lower the tongue from fi] '0 [a]. You indy find yourself ai first wanting to ‘proceed by a series of steps or jms, trom fi] to (, €] tel [e} of}. You ‘must resi: this [Fyou do find yourself jumping fromm one CV io another, 26 ‘back to [i] and start ever agaia. The todgué must move downwards, always ‘mainiaining its very front postion, at a‘uniform slow pece: Whea you reach, [2] try to continue the slow uniform-paced glide back to [c]. This whole rocedureis represented by the broken line in Fig 50, SOUND-SYSTEMS OF LANGUAGES. 179 Fig, 50 The Ha vowel continuum. Theline ...... represects continuous gliding tongue movement. For the same, represented asa straight line, see Fig. 51 ‘We siwin Chapter 7 that tongue-teight can bévared in one or both of two ways: (1) by fxaing the lover jaw’ and passing from a closer 9 2 opener vowel by lovering and fattening the tongue and 2 by fiatine the tongue in relation 10 the lower jaw and changing tongue-height slely by lowering the jav. You will probably find i eases to carry ou the present procedure by controlling only one set of muscles at a time, as suggested in Experiment 122 122. Faate.the lover jou if ancesay by eriping the end ofa pens o of a dual ook begoce your tet Now eng Set wep thn wn, Pr a {i and slowy and uniormly lower your oagur though all possible pesitgns—not sopping x at)—t0 (o, Havng competed this slow con ruoustongvelowering fom 0 (continu slow and continuly rrirsting it (0) Ty to carry ont the whole operaon agai, fem 6] {hobeh fo ein oneslow contd continuous movement | Now try to do the same thing in the second manner, Start with a very ese cos front). Keep the toague ridin tit cose oct postion, and slowly and contimosly lower te jaw I you rodute voice wile hing ie you wile how the vo! qualty continuo changes pass- ing slowly trough al posible vowel aouads teres Gand a) a you Tower your jw more and mor. When you get wa ofa withthe tongue sil fated, you wll of course, have tweet the tongue ia order to ~ continue the unbroken slide back to [a].180 SOUND-SYSTEMS OF LANGUAGES If you spend a little time on Experiments 121 and [22 you should find quite easy to produce a continuous vowel-side, passing through aa infinite umber of intermediate points withSu: pausing at any of them. This is Quite a useful skill in is own right, and it is worth while experimenting farther with siding slowly from (u]to[y] and vice versa, Itis useful because it provides a way of acquiring new vowel sounds: if, forinstanés, you want to produce avowel between [c] and {e] you can get there by starting at [o] and slowly sliding downwards, stopping when you reach the vowel you want. Ta the present chapter, however, we are using this skill in a different way. If you carry out a slow steady slide from [i] to fel, 2s in Experiments 121 and 122, you pass through an infinite series of vowels—in other words, you experience a rome! continuum, Having got the idea of @ vowel continuum, not as a mere intelectual concept, but as something you can produce and feel and hear in your own mouth, you ean more readily understand when wwe say that different languages dissect the vowel continuum in diferent ways, that & select different bits within it for linguist use. 3. The vowel continuum differently dissected by English and Spanish : English divides up the peripheral vowel continuum from [i] to (a) into six divisions—or, to put it differently, selects six bits on or near that continuum for use in words of the language, These six selected bits of the continuum are the following: {as in Beer, (asin bi, (ei) as in ade, eas in Der, fe] as in bas, ané [a] as in part. (Li British RP the sixth vowel, [0] in pot, les beyond the short vowel continuum given here) Spanish, on the otker hand, selects only threc bits out of that Gontinuur, narnely i] asin ibis, fe] as in eres, anc fa} 28 in acd, 1 would be reasonable to suppose that vowel-sounds would be more o= less evenly spaced along the eontinaum, and to some extent this isthe case ‘The more widely spaced out they are, the more room there is for variation. For example, the Spanish, vowel represented by (e] has variants ranging {com about CV [e] to about CV [e]. But these variants do not oocur at random; they are systematically related to different environments—to dif ferent locations within the syllable. Generally speaking, the closer variants of the Spanish [e] vowel occur, in opent syllubles—that is, in syllables SOUND-SYSTEMS OF LANGUAGES. [8] not closed by a consonant—while the opener variants ocour in closed syllables, ic. those that end in agonsorant. Examples, with 2 hyphen indicating che division between syllables, are pe-ro ‘out’, with [@], but perro dog’ with [2 We caa represent the vowel continuum that we have been talking about, and the different divisions imposed upon it by English and Spanish, a¢ in Fig. 51 Spanish : . x 51 The ina vowel continuum: English and Spanish divisions 4. Voice-onset-times differently exploited by different languages A different kind of phonetic continuum is represented by the ‘voie-onset- time" (VOT) of stop consonants. As we saw earie, a charactoretic feature of intial expiraced stops is that, ater the moment of release of the oral Glosure, the vocal folds take some time to come together and stat vibrating for the following vowel. This time-lapse before the onset of voicing is what wwe call the VOT. Taking the moment of release of the oral closure at reference time, aspirated stops, with a delayed voice-onser, have a postive VOT. For example ifthe voicing does not begin until 100 thousandth of @ “second—a hundred milliseconds—after the release of the stop, we ave 2 VOT of + 100 ms. If the voicing besias simultaneously with the release of the oral eosuse, he stop is wasprated, and the VOT, of course, is zero. If tae vocal folds start vibrating before the release ofthe stop (as they do in voiced stops) we have a negative VOT. For example, if the voice-onset pre- cedes (cother than follows) the stopclease by 100 ms, then. the VOT 100 ms. So it s clea that there isa continuum of possible uations of the times Gtferencé between the release of stop and the onset of voising, runaing from, say, + 100'ms. to, say, ~ 100 ms. with ary conceivable value in between, This kiad of continuum is, of course, a purely conceptual one—a182. souNo-svstas oF LANGUAGES SOUND-SYSTEMS OF LANGUAGES 183, continuum of pessbiltis, composed of all conceivable VOTE that might «vont, You cannot actualy ren unbrokenly through icin your own voral tract as you did with the vowel continuum, You can, however, pronounce, fone ater the other, a series of discrete examples illustrating different poiats along that conceptual continuum, which is what we will do in Experiment 123, Meanwhile, look at Fig. 52, In this figure the vertical lines represent the start and finish of lip-closure for a labial stop. The horizoatal lines repre- sent, conventionally, the state of the glotis, the dotted line indicating voleelessness—the glottis being ether wide open (as for aspiratee stops), ot narrowed to the whisper, or prephonation, position (as for unaspirated stops s02 p. S6)—and the wavy line indicating voicing, The broken diagonal line running from top to bottom represents the whole conceptual con- tinuam that we have been talking about—it suggests that there's, in Principle, ¢ continuum, an infinite number, of possible VOTS, ‘Now, while looking at Fig. 2, carry out Experiment 123, feoovosnosoen. pa slightly asprcted Fig. 52 The VOT continuum 123 Close your lips as for a {pl and build up palmonic-pressure behind them. ‘When the pressure is rally high, suddenly open the lips and release 2 con- siderable blast of air before the vocal folds finally come together and start vibrating. This high pressure [p] nas along VOT, Le. it strongly aspirated, and we represent the utterance as (ph. ‘Now do the same again, bat this time build up less pressure so that the aie-bast following the release nate 3 shorter time before the voicing begins, ie the VOT is shorter. Thisisa moderately aspirated [p] perhaps such os you ‘vould normally use in Englsa, an¢ the utterance is represented as [pa]. ‘Next, produce a quite unaspirated (p]. Rememter that for zn unaspirated stop the slotis is already ‘throttled down’ to about the size ofthe whisper orifice during the closed phace of the stop: consequently, the vocal flés can ‘art vibrating immediately upon release of the oral closure, ie. the VOT is (virally) zero. You can achieve this by imagining whisper during the ctoses phase of the stop, Reinemer what it fels like to whisper a [v] and get this ‘sume glottal feeling into the closed phase of the stop—i.e. whisper the stop, ‘eat there will, of course, be no audible whisper, so that voicing starts ‘with unaspirated (pl ie with Immediately the lips burst open. This [p 240 VOT. “The three [p] Sounes produced in Experiment 123 are al voiceless, Two cof them, strongly aspirated [pi] and moderately aspirately {p], have posi tive VOT Plaasible values might be something like 120 ms. for [pha] and 60 ims. or [pPal, but of course, tie actual VOTS that you achieved might difer ‘quite widely from these values, and could oaly be established by instru- rental measurement. The unaspirated stop [p in [palis still wiceles, since he vocal flde are rot vibrating. ll uring the loge phace Af the ap 1d ur next expeciment we Continus our progréss along the conceptual continuum of VOTs into increasing negative values of VOT. Whereas in the spiraled stops there was a delay in voicing, waich did not begin vat a noticesble tine after the elese ofthe stop jn the next two examples there is an ddvanee or anicipation of voicing, which stats while tte labial loshre is stil firmly in position, It is because we assigned positive values fo delayed voize-onset that we now Rave to assign negative values to the ticipated onset. 124 Now produce a slightly voiced’ (b}, Close the lips, hold them closed, an contrive to get the vocal folds wrating just before the retease of the labial closure, This kind off] starts out as a totaly voiceless stop, and thea, just atthe las: moment before th lips part, it turns into a voiced sion. We can184 sounD-sysTEMsor LANGUAGES SOUND-SYSTEME OF CANGLAGES 185 represent it ia transcriptions as fp & al, that is, voiceless {9} passing (momentarily) into voiced [b] just before the lips open up into the vowel. Finally, produc: a fully voise (b, ie. get the voicing going well before the release ofthe con. as in Exps 3! and 32 Experiments 123 aad 124 exemplied five diffrent points alony the continuum of possible voiveonset-times. Figure 52 represents the VOT con- ‘Snusum by a slanting broken line, and labels these five ifferent VOT values. It also suggests how three different languages exploit different parts of the VOT continuum in ther contrasting pairs of initial stop consonants, Some further comments on Fig, 52 are necessary. First, it will be noted that both French and Mandarin Chinese have a voiceless unaspirated «209, but they are represented differently inthe tcanscriptions—ihe French one by [p] and the Chinese one by (b}, This indicates that the voiceless unaspi- rated stops of the two languages are not identical. Frenck [p] (and likewise (@] and (kis relatively ‘strong’—the pulmonic pressure behind the stop is relatively high, and its tense aricslated, The Chinese [b] (and likewise [¢ and g))is articulated io a more lax way. The Chinese sounds, though total vvoicsless just as the Fench ones are, have 2 lower air pressure behind the stop and probably require less muscular tension ia the articulators than the French sousds This relatively low pressure and rather Iax siticilstion is reminiscent of voiced [b], [dl [2], where some ofthe energy in the air-streami hhas been absorbed ia setting the vocal folds in vibration. Consequently we jepresent the Chinese voiceless unaspiraied stop as (p] as a reminder that, though completely voizless, it has something ke the laxness ofa [b). ‘The English unaspirated stop is represented as [pb] to underline the fact that it starts ovt a « lanly articulated vojcsless stop, lk the Chinese [pl but then it momentary (2, for not more than 20 or 30 ms) turns into a voiced stop [b] just before the release of the oral closure. Itshould be aoted that Fig, $2 represents inival stops. This is particularly relevant io the English examples, for in English aspirated voiceless st6ps occ iniilly (before stressed vowels) but isually not elseviéte. Thus the medial [p] and K] of paper and baker are muck less aspitated thin the iaital stops of pay or Kay. In most typés of American {ané some types of British) English the medial consonant of better, buster, tc, is not a stop at all, but a (voiced) apice-alveolar tap (fl, though im British RP itis usually a voiceless slightly aspirated [) In many types of English syllable-fnal voieeless stops, as in hop, hot, hock, stopgap, et. areunceleased (Le. have no auibie release) [>] [C0] moderately to strongly and may be preceded ad accompanied, by gual dosure PP"MLPM PE) Moreover as we pointe out in Chapter 3, alter sin spar, star ser, the English stops are unaspirated, like the French [p] or Chinese [b}. Weve now tom how thee diferent languages ler diferent pats of the VOT continu for us ia bulding up the phonologien forms oftheir words. The VOT continuum is universal. That is to say, all human vocal tracts are capable of producing stops with woiee-onettimes at any point along that coatiauure, But different inaguages sie eiffereat pas ofthe continuum. Thas, Chinese utilzes points towards the “goitve” end of the continuum; French utilizes points nearer the negative end of the continuum, while Engl uilzs an area between those we, When we looked at the vewel contiauum, above, we saw how ewe dif ferent languages, English und Spanish, used different parts of that coatiauum—ust as the three languages we have now been looking at use different parte of the VOT continuum, Rut we also noted that the two langvages, English and Spanish, divide up the vowel continuum! inte iffereat numbers of bits, different numbers of vowel-units—English 6, Spanish 3. It so happens that Chinese, English, and French each divide up the VOT contiguum into the same number of bits, the same mumber of different VOT-ypes, namely two, But there are languages taat utilize not just two bits of the VOT continuum, bur chree Such a language is Thai. Thi utilizes three diferent types of inital stop consonant, whish can be exemplified by voiceless aspirated [pa], voiceless unaspirated {p]; and voiced [b]- Taus, with cespect 10 the nimber of types of stop consonant that they Utilize, the three languages; Chinese, Frglish, and French, are the same. They each have jus wo typesof stop, though they diller with respect (9 the precise quality of these'stops: that is t0 say, the stop-types of these languages are differen- tiated ftom each other in different ways, as utilizing different parts of the VOT centinuum:.Thai, on the other hand, differs from all three of these languages in that it-utlizes three, not just ewo; ditfereat types of stops— three different bits of the VOT continuum. We can captare this similarity and difference ty transeribiag the two types of labial stop of all of the first three languages as /pl uad /el, but the three types utilized by Thai as /ph/ /p/ and /bl. Here ve are representing facts about the phonology of these languages rather than about phonetics, ‘we are representing phonological units, or phonemes, as they are called, rather than speect-sounds This is indicated (following the normal practice186 soUND-sYSTEMS OF LANGUAGES SOUND-SYSTENS OF LANGUAGES 187 of linguistics) by placing the symbole within slantlines // rather than within square brackets [ ]. The distinction beoween phoaemes and speech sounds will be made clearer below. Meanwhile, we can display this very small frag- ment of the sound-systems of Chinese, English, and Thai showing kow the contrastive phonological units, the phonemes, correlate with their different phonetic realizations, as shown in Table 6. Table 6 Language Phonetic realization bt el om o &) Chinese ipa! pal “Tocroues" “sight English ‘pay fay ‘pie ‘by French Il food “seo ‘Tow’ Thai ‘kaa! pas! Aaa! ‘clock ‘tant ‘ony’ ‘Other languages have more than the three types of stop of Thi additional ones do aot represent simply further subdivisions of the VOT continuum since they involve other features. Thus Hindi and many other Indian languages have four types of stop, by adding a voiced aspirated step (ope (of) to the series. The sound [bf does not At on che VOT continuum wwe have been looking at, since it involves whispery voiced or “murmured” aspiration, not the parely voiceless aspiration of these otber languages. Stil other languages have five éifferent types of stop, but the Afth one is, distinguisned by something other than VOT, namely @ difference of initi= ation type. Thus Sindhi of Pakistan and NW India has voiced implosives as, well as all the types.of stop that Hindi has. The systems of labial siops of these additional languages thus are shown in Fig. 53. Hinde fo fo) fol Pe sah: fob pl Pa Fig. 53 Labial stops in Hindi and Sindhi 5. Phonemes We have seen how ¢ifferent languages exploit the universal human sound potential in diferent ways. Every language uses a particular set of phon emt for the purpose of building up the phonological forms—the audible realzations—of its words, and we have seen some examples of these, Now we must look more closely at the mature of phonemes and at the sound-systems of English ‘Phoremes are the minimal sequential contrastive units ofthe phonelosy of lenguages, Each of the thiee terms minimal, zequential, and contrestve requires explanation We will take contrastive frst. Phonemes are contrastve in the sense that they are the Dts of sound that disinguisl one word from another. Consicer the English word bi, transcrited phonemicaly as (bv. Its distinguished from the word pit pt solely by the contrast between the initial consonants Tol sod Ip). Again, pit, is distinguished from fit by the coatrast between ip! and /t/; and bit is distinguished: from ber /beW/ by the contrast between te vowels i! and /e/ and so ca. The bits of sound that carry, or manifest, these contrasts are phoneme: Phonemes are minimal sequential units, because if you take a stretch of speech and chop it up into a sequence of phonological units the shortest stretch of speech that functions as a contrastive unit in the buildup of the phonological forms of words is the phoneme, The phonological structure (of English, ike that of other languages, can be described as a hierarchy of ‘units The largest, or most inclusive, unit in English isthe intonation con tour or tonegroup. For example, each of the three versions of the Sentence ene was here yesterday, given on p. 173, exemplies a single tone group. LIT tay te Each of these utterances represents a single phonological unit of pitch—a singlt sone group, here marked off by double vertical lines. ‘We can, however, chop up each tone group into smaller units—namely into two successive rhythmic units, or feet, shown bert as separated by single lines, thus: Hmm emt. aE ll ane was hive yesterday | ‘That “these feet sre contrastive, meaning-differentiating, units is188 SDUND-SYSTENS OF LANGUAGES SOUND-SYSTEMS OF LANGUAGES 189 demonstrated by the fact that we could givide the utterance into feet differertly, and this would convey a slishtly different meaning: Imi || “Fane was | ‘here yesterday seal Next we ean divide each foot ino stil smaller chenks, namely into a sequense of syllables, sowa here as separated by hyphens within each foot, hus: UW aneewaschere | yesterday I FFinaly, we can divide up each syllable inte a sequeree of still smaller ‘units—and here it is necessary for us to go into phonetic transcription, separating each unit from the next by a dot, thus Wl ‘ésenmewoahse | ‘That representsa possible British, RP, pronunciation. The corresponding American form might WL Gerewerhas | jesterde |] A‘ this point we can do ne Further chopping: We have reached the lowest rank in the phonological hierarchy, the smallest sequential or linear units— phonemes. There are no smaller meaning-diferentiating units: we don’t ‘built up the phonological shapes of words out of only half an [3] or a small fraction of an (e}-for example. There are, howeves, some questionable ems here. Why, for example, do we treat the sequences (63) et] and (15) as minimal units? We iake [25] first: The interpretation of [4s] as a unit phoneme is sug- gested by the text of commute. if we Wy to make English words by ‘commating, or: exchanging, cach part of the afiricate [43] with other soundi, we Sind that in such a word as Jane the [5] can commute with zero, ‘giving Dene, or with (4, giving drain, or with {w] giving Dwayne. Bus the [d) canzo: commute with anything. No consonant other than [d] can ever come ‘before{s}in English, ic. there are no such words as */bseia/ or */ggeta (the aateride indicates a hypothetical oF non-existent form). Nor ean {3] be pre ceded by zero—no genuine English word begins with {3}, ie, thereis no such sword as “ein! Ie thus appears that the (3] in [d3]s incissolubly Linked with the (d]and the sequence must thus be regarde¢ as asingle, unitary, affticate phonenc ‘Applying the commutation test fo the corresponding voicéless aflicate [if] we find that both thet] and the [f] are commutable, so, from that point of view one might regard [tf] as a sequence of two phonemes, /U and If. However, the parallelism with the unit phoneme /d7/ is obvious, and this suggests that we might well regard both as unit phonemes, and this inter. pretation is supported by other considerations, such as the historical origin of the sounds and certain phoneme alternances in mocern English. Thus, most linguists regard these two ffticstes as unit phonemes, and partly because of this, often transcribe them as /8! and J). We will continue, how- ever, 10 follow the principles of the IPA by using the digraphs tf and dy. ‘The case of the diphthongs /ey and (in the British version) /ta/ is ditfer- ‘ent, Here there are genuine options in the analysis Depending. in part. on the purpose of che analysis, one may choose to regard the Enalish diph thongs either as units or as sequeness of two vowel phonemes, or (in the case of ley fad, fa! fou/ ete.) asa vowel followed by 2 semivowel, ie. as fe Jail, aw! fowl—the first two are usually transcribed by American linguists as ey! /ay/,a transcription which vielates IPA practics,sinee the symbol [y] in the IPA system represents a rounced close froat vowel, These ambiguities apart, however. phonemes, as we bave said, represent the smallest sequential contrastive phonological units into which we can divide up a text. 6. Distinctive features Ir was necessary to make the point about phonemes being minimal sequen- taal units, because a further subdivison into contrastive units of a different ‘kind io poasible, This ig the divicion, oF bot 2 the diccostion, of phonemes into distinctive features. But this, we must emphasize, is 2 dissection in a quite different dimension, ‘Ail of the phonological units we have looked at 10 far—the intonation contour che foot, the syllable, and finally the phone ceme, ate all eits that fellow each o1her in time throughout « spoken text Distinctive features, on the other hand, ace simultaneous components of 3 singlé unit the phoneme, IF we look again at the woede bie pit and fir we can see how the initial phonemes in these words, /b/p/ ff, have certain things in common, but also differ from each other. All three are consonants, as epposed to vowels; all three are, broadly speaking, labials (nevlecting the fact that /f/ is actually labiodental); the fist wo resemble aach otker, being cops, while the third is a ficative; fois (lightly) voiced, Dut (p/ and /tf are both voiceless. We=e Te 19) souwp-sysTaMs oF LANGUAGES tabulate these fects at shown in Table 7 indicating presence of a feature bya + sign and atsence of a feature by a ~ sign Table7 Phoneme Feature cea vowel labial stop __ficative voiced voiceless ® + a ie aes p > enh ey Cee Se Obviously we could tabulate this information more economically by thinking of eack feature as reprecenting a binary choice: a sound is either a ‘consonant or not a consonant, & vowel or nota vowel, voiced or not voiced, So we can reduce the oumber of features that we have to list here to cons, labial, stop, and voiced; see Table 8, We can thus say that the phonological form of a werd consists of a sseeuence of phonemes, and tat every phoneme consists of a ser of (Simul taneous) distinctive features. The phonological forms of words are differen tiated from each other by the occurrence in ther of different phonemes: the phonemes are differentiated from each other by the existence in them of different distinctive features. In the last forty years a great deal has been written about distinctive features and several scholars have sought to discover the minimum number Of distinctive features that are used to differentiate phonemes in the lan- guages of the world, Three works dealing with distmctive features are ‘tamed in For Further Reading. at the end of this book. Readers are referred to these publications for more information on features, since we will not make ase of a formal set of distinctive features in our further discussion of the English sound systems. Tables Phonene Feature cons. labial stop—_voieed b + + + + P + + + = i Since this book is a practical introduction to general phonetics werdonot intend to give a full account of the phonetics and phonology of English: Nevertheless, it wil be useful to survey some aspects of the English sound stem, and this we will do, beginning with consonants. 7. English consonant phonemes English has a total of 24 different consonant phonemes, oF, as we say, it has 4 consonant phoneme inventary of 24 items. These consist of 6 stops: /p bt d kg, afiticates: tf dg/, 9 iricatives: /T v0.0 sz 5 hi, 3 nasals:/m nay, 2 approximants: lf and 2 semivowels: fj “That i the overall inventory; but we have to note that not all ters in this inventory occur in all positions in syllables Tn initial position only 22 items ‘occurin geauine English words, since /y occurs only in a few incompletely assimilated foreign words or names, like jabot and Jecnne, and the velar nasal never occurs inthis position. In syllable final postion the sersivowels/ wi and If never oscur, ner does fi; this reduces the investory-for Binal position to 19 in those varieties of English that admit final (called ‘rho- tig’ varieties) such as most types of American English, Scottish ane Irish English, etc. In varieties with ao Gnal /s! (‘non-raotic’ varieties), such as British RP and other varieties of English in England, in the US, Australi, New Zealand, ete, only 18 consopants o¢vur at the ends of syllables. “The full inventory is found oaly ietervocalicallyfor here we have both the non-initial consonants, eg, /9/ in pleasures vision, [9 in hanger, longing, and also the non-final consonants, ¢.g, (W/ in away, iV in Deyond, A In ahead, hl ia eround, Tis important to be aware that phonemes are abstractions or generalizae tions: they are, that is to say, abstcact phonological units, each of waich is manifested, or alized, in speech in a number of different ways, You cannot preneurcea phoneme, You ean only pronounce a specie sound which may ‘be the realization of a phoneme. If you sey, for instance, the English word cat youare producing a quite specific sequence of sounds. That sequence of sounds isnot itself a sequence of phonemtes itis the outward or concrete manifestation, or realization, of the sequence of phonemes that we repr’ sent in transcription as (kev. Say cara number of timer: every time you say it you are obviously producing a different sequence of sounds—you may even hear slight differences in your pronunciation of them. But each192. SOUND-SYSTEMS OF LANGUAGES SOUND-SYSTEMS OF LANGUAGES 193 utterance of car isa realization of one end the same sequence of phonemes: eau, So, every utterance of a phonemes a diffs aware of a difference in pronunciation or not, The actual sounds you make are concrete, physical, eveats: the phoneme js, as we have seid, an abstraction, ¢ sound—whether you are 8. Allophones ‘Apact fiom the fact that obviously every single re a diflereat sound, there are particular types of realizations of phonemes that are systematically related to particular contexts, of, a8 we say, to parti- cularphonological environments. We have already noted some environmen tal variants of English stops, namely, thet initially before stressed vowels English voiceless stops are noticeably aspirated: in syllable fizal position, particularly in American English, they are often unreleased and may be preceded and'eccompanied by glottal clovure, and there ars still other noticeable variants. Such systematic enviconmental variants of phonemes are called allophones. Allophones, just like phonemes, are abstractions— ‘which are concretely. realized in speech as partictlar sounds. We might tabulate some of the allophones of the English phoneme represented by it/ asin Table 9. tis anot uncommon practice to transcribe allophones, such as we have listed here, within square brackets as (2) °°] [t] [as if they were purely phonetis entities rather than phonological ones Strictly speaking, however, since they are abstractions made from the eaviroamentally constrained systematic realizations of variants. of phonemes, they are part of the phonology of the language, and should thus be enclosed within slant tines. However in the notes on English consonants that follow we wil iranseribe allophones withia square brackets, ‘Stops, Voiceless: pl Af fel: all noticesbly aspirated when initial, especialy tion of a phonemes Tabled Initalrstress Final Before da) ‘Before /9/ or 18! WW ° e e(retracted) (dental) Example tea stare eighth Defore stressed vowels (except in English of Seotland and Northern England), Generally, in American English, and sometimes in RP, final voiceless stops are unreleased and often uccompunied by glottal stop, fe. [9°] PC] (X’], In British English they may be relessed when absolute final (before pause) =(D°] (0) (k). \ Vojsed: fe/ (al): all partially (or completely) voieless whea inital, IQUE [8a gal o (O-(- gh, oF final 1) dg] “gg 0° 1-6) “ig Jn American English, medial (t/ and /d! after'a stressed vowel. are commonly realized as alveolar tap (4. as in latter or Indder. /M ané fel: addivionalallophones asin list above for /V. ‘ky and gf: noticeably fronted (palatalized) before (and front vowels, particularly A. Africates 'f) and df: wsvaly lamino-postalveolacafficates, but with the congue-point raised (i. articulation is aot far back on blade). Often with wide kp-rounding /d3) has partially or wholly voiceless allophones, as. forthe voiced stops. i Note that we deieibe /ord g dij a8 ‘voiced’, even though they may be (party or even fully) voiceles: this is Because they are always capable of being voiced, ie are always potentially voiced, and in addition they have the rather lax: articulation. characteristic of voiced stops. In fact some suthors describe both tke voized stops and the voiced fficativer as “lax? at opposed co ‘tense’ voiesies onss. We prefer co call them “voices! and ‘voiced’, since the ‘lax’ fticatives are vistually always 2t least partially voiced, the voiced stops are always potentially voiced, and the lax vs. tense opposition is no doubt merely caused incidentally by the aerodyaaznie effects of the diferent glotal configurations for the two clases of sound, Fricatives: the: voiced fricatives have partially voiced initial and final allophones: l= [yw] -[wyh l=! = (gah teh-tez ete. Jt ivi; endolabio-denial (i.c, inner part of lower lig makes contact with edges and froats of upper teeth), 19184 apico-dental: in American English pethaps somewhat interdental (ie. with tongue-tip sighty proteudiag through bemween the upper aad dower teeth), ia Briish English perhaps more commoalyslighily postdental (Ge, with tongue-tip and rim behind or barely touching the inner edges of theupper teeth), ‘sel: most ceramenly lemino-alveolsr (i/if:-usually lamino-postalveclar, but with the tongue-tip noticeably194 SOUND-SYSTEMS OF LANGUAGES raised: may aiso be apico-postalveolar. Usuelly with some degree of tip- rounding. Th: usually described as a glottal fricative: it has litte in the way of a turbulence-generating stricture in the mouth, the fricative sound being mainly generated by turbulent flow through the giouis, bat noticeably rodulated by oral articulations iis often described as estentially a voiee- less version of whatever vowel follows: thus, he /hi, had Phe, tall bot right be transcribed elloptonically as /f] edd] (go ‘Nasals: ial /o/ /y/: bilabial, zpico-alveolas, and velar with few notable allophones, though note that ‘n/ undergoes slight shifs ike those of // and J] before postalveolass and dentals. Approximants: i! aad Mi: both are nermally voiced, but have whelly ot parly voiceless allophores after voiceless stops, asin pray [pz], play [ple ory (kcal, day (kel. ‘dT have up to now retsined the transcription i! as a reminder that the English r-sound is practically never a will (represented in phonetic tran: scription as (c) even in Scotland, where tilled [rs are popularly imagines to be common! It is, however, normal practice to represent the /:! phoneme ‘of English simply as /r, This is in accordance with a principle of typo- ‘graphical simplicity, whereby the typographically simpler of alteraative transcriptions is used whenever there can be no possible ambiguity. From row on [ will represent the Enalish It/ phoneme as /r/, but will use (a! in transcribing allophones. Ia American English, // can be pronounced in several different ways. Ik is virtually always an approximant—ie, aeither a trillnor a fricative, except az after It oF /éf in such words as:try éry-it may be a postalveolar {ricative—wholly or partly voiceless in éy: thus [tu] and (€say} (where [t] and {d+}, you remember, represent the retracted allophones of /t and /d that occur before postaiveolar (i). One pronunciation is simply that of 2 retiofiex approximant, je an approximant made with the main body of the tongue ia roughly an (a}-type posture, but with the tip curled up somewhat. A commoner variety seems 10 be a peculiar sound like the rhotacized xowel described on pp. 161~2. In this type of /t! the body of the tongue is bunched up into roughly the position of an [¥] vowel. At the same time the root of the tongue is slightly retracted into the pharynx (slight éccp pharyngalization) and there is a short longitudinal furrow in the dorsal surface of the tongue roughly opposite the uvula, ora litle forward from there, Rather surprisingly at first sight, those who do not already use this SOUND-SYSTEMS OF LANGUAGES 195 type of fl can sometimes learn it by starting with a uvular trill (the ‘rare ling’ sound) [x and thea very slightly moving the tongue-bod. forward ‘until the triling ceases: furtkersligh tril and error adjustments oftcmcad toa reasonably good Americar /t! of the type in question, The resemblance of the articulation ofthis American le! to a uvular[R) isnot realy sumpris- ing, since both involve a rather similar dorsal furrow in the tongue. 7: apico-alveolar lateral approximant, voiced, except after the voiceless stops /pl and /k/, as mentioned above, In addition to these voiceless (or partly voiseless) allophones there are two principal allophones. [a slightly Palatalized variety of {}, often called ‘clear I" ovearring before vowels as in Teaf fit} ie as} and, in RP at least, before (jf as in mtn {ms}, and C, a vearized variety of {Il often called ‘dark I’ oceurring at the ends of syllables, as in fet (Gs) field [Red], and when syllabic, as in battle (bat) smidate (ruc). * Semivowels: jw! and fj (the latter often transeribed as /y/ in American pubiicetions—but, as mentioned earlier, this ‘anglo-centric’ usage violates IPA practic). Jai: Dilabiat (founded) + velar semivovel, and (/; dorso-palatal semivowel : Both semivowels are normally voiced, but have ray (or whelly) voice- lessallophoacs after voiceless stops, as in aveve[tyely), quoen feiss), Cube [Kjurbe) The sequences Inw/ and /hf as in-wshy and Hugh, 28 spoken by persons who pronounce the//,may terealized s[nw], hj). {hy} [hj.oras[v). that ig) in fact; a voiceless Inbial + velar fricative, and as (c). voiceless dorso-palatal ricative. 9. Consonant clusters By consonant clusters we mean those sequences of consonants that occur initially or finally in syllables. For example both /pU and /nt/ are consonant clusters in the word plant, because they eccurin one and the same syllable. But we would not call the sequence /pr in stoplight a consonant cluster, because the sequence crosses a syllable boundary: /o/ isthe final consonant in the ft cyllable and /V is the inital consonant in the stcond'syllable. In Chapter 9, on Prosodies, we cited some English syllable srvctures and referred to the fact that languages differ as to the types of syllable strectare that they acmit. These differences are chiefly a matter ofthe extent to which195 scUND-sysTaMs OF LANGUAGES different languages permit consonant clusters te oxcur st the begionings of ends of sylables English permits initial clusters of either two or three consonants, as in Saas renee os prey, Hos (CE ND emis Fert (COSY), Gal sites nat, Care, or four Snal consonants as in ask, aps (VCC), asked, elks (VECC), oF four as in waltzed iwoltst, glimpsed /ghmpst! (C(C)VCCCC). Some Eagiish speakers, ir fact, simplify some of the longer final consonant clusters by eliminating some of the consonsnts—saying, for example, fast (or est) in place of /asky (or /ask¥) for asked. You might check your own usage in these and other words. : Other languages have diferent rules celating to syllable structure: Clas. sical Arabic, and many varieties of modzrn colloquial Arabic, for example, admit no inital consonant clusters at al, but allow final clusters of two or three conscnants On the other hand, the Cavcasian language Georgian permits inital consonant clusters of from two to six items! Examples are /fq'vel “to water’, /péTSkvna, ‘to peel” (the ligatures on /BY and iS) indicate that these affeicates are unit phonemes). ’ Not ony is the number of consonants that ean occur initially and Smelly in syllablesalways subject to limitations imposed by the phonological struc ure of the language, he particular sequences of consonants that can occur in clusters is always:limited. In English, for exemple there are 22 initial consonants, This mesas that, if these could be comabined quite freely, the total number of clusters of two different coasonants would be 22"~22, that 's, 462..In fact the rules of participation in consonant clusters in English pemnit only from 36 to 44, according to dialect, that is, only 7:8 per cent to 9.5 per cont of the theoretically possible number af clusters The Engish inital 2-consonant clusters are showa in Table 10 Naturally, adjacent sounds in clusters influence each other in: various ways (some of which have already been mentioned). in other words notice- able allephones of consonant phoaemes ozcur in some clusters. Thus, the aspiration of the voiceless stops carries over into following approximants and semivowels, so that the latter become partly voiceless, such words as play ané twelve being pronounced as [plev] and [twwelvy]. Sequences of Consonant) + wi and C fj exhibit some degree of labi- alization and palatalization respectively of the fist clementin the cluster. If forexample you pronounce the words seat and sweet, carefuly listening to the sound of the initial /f in both words, you will clearly hear the auditory flect of the progressive labialization of the /¥ of sweet—as vsual, you will SOUND-SYSTENS OF LANGUAGES ‘brow (Buenos) beauty dry dwell (due } ree thyack fthews) sheik slow prow Oven (eves) fy fy fev new) (nh si) ni ) 8 b > (sh ai i (oy i ws) tw kw x) dw iy ow 7 kW u a i 3 o tsk sty sp Sphere smooth snow slow sweet stay spy hy) (Hugh) (tow) ea af Engl Gras dentl pe ohGola eonbamoats are nat (hh are tere dp i 38 BE = Io this be ( sue onty inthe Posh fares widves woe pel and fi 17=a teak’ tol oe 198 SOUND-SYSTEMS OF LANGUAGES find the difference between the two /s/ sounds easier to observe if you whisper the words, and isolate the /s/. ‘Again, if you compare such words as coo and cue you may hear thet the ‘burst of noise on the release of the palatalized /k! before / in eue is notice- ably higher pitched than hat of the k/incoo, betokening a more advanced, slightly palatalized, place of articuletion, and if you articulate the two words silently, or nearly so, you will probably fee! the slight difference in the place of articulation. However, it accessary to point out thet the allo- phonic difference between the us in these two words may be minimal. This is because many speakers of English have a very advanced (cectralizeé) and somewhat dipathoagal fur! vowel in words like con; i.e. something of the type [a0] or even fis] If your /u/is ofthat type, a better comparison would ‘be with caw, for example In addition to these 2-consonant clusters, English admits of $, and oaly 9, different initial custers of 3 consonants. These are [sp spr spt {] Si) ISkU Iskx (sow! Usk, as in splay spray spew stray (stem) sclerosis screw square skua,'of which one item, stew, is pronounced with initial /st, not Jat, in most types of American English in ateordance with the American rule of non-occurrencs of ff afte alveolars ‘AS wesaid easiest is not our intention to give « full count of English phonology, but just to provide an illustration of the highly organized, sys- tematic, nature of the sound systems of languages We will therefore not ccnumerate the final consonant clusters of English, but pass directly to a brief discussion of English vowels 10. The English vowels ‘The enumeration of the vowel paonemes of English is a little less straight forward, than that of consonants, for two reasons. First, there is slightly more divergence between major dialects (¢, British and American) with respect to'vowels than with respect to consonants, and secondly, because there are different ways of making 2 phonological analysis of vowels. ‘There is one general remark to be made about English vowels and diph- ‘thongs in stressed sylables, namely that vowels are normally longest when final (e.g. (i) in see), storter in sjlzbles closed by a consonant, particularly those closed by a voiceless stop (es. {r]in see, [i] in seat). This should be bbome in mind as we enumerate the vowel phonemes of English, with brief SOUND-SYSTENS OF LANGUAGES 199 descriptions, Tx every case readers should experiment with the Enalish ‘vowel that they use, pronotncing it both in the sample words, and in isol- ation, aloud and silently, and trying to feel where aad how it s articulated in relatos to the Cardinal Vowels. Front vowels, iV as in see, seat. Unsounded rather close front vowel, semewnal less close and tense than CV |, Sometimes slighty diphthongal—ic. starting a3 ‘an opener and laxer vowel an¢ rapidly becoming slightly closer and tenser. ‘We ean represent this as fi] or (i) Win sit. A slightly retracted and lazer and more open vowel, In Ameri- can English often at least as open as a retraced CV 2 (¢], and sometimes somewhat diphthongized, gliding inwards and downwards slightly :owards [s} particularly before [das in did [dtad]- In RP often somewhat closer. Ted in day, dave. In both RP and Midwest American this ie normally'a Giphthong, but the extent of dipathongal movement is variable, Most comtmonly, pethaps: it starts between CVS 2[e]and 3 e] and glides forwards and upwards, ut seldom, if ever at far as (} {hence our choice of symbols Jeucrather than /ei/). The starting point may be lower, at about the level of CV3 fel and even somewhat retracted, tnus [ct] or [é]. In some varieties of ‘American Engliso, and in the English of Scotland and aorthera England, it may be.a pure. vowel of the CV 2 type—our sample words being pronounced [de and: (det). You should experiment with these variuat rominciations, and try t0 discover which type you normally ust. el in set, sid. Ta xmaay types of both British and American English, this is avowel rather similanto CV'3[e}, though usvally aitle closer in RP.and often a litte retracted im American Engilsh ae J ir sat, sad. Traditionally this is generally described as a more open rowel than the preceding one, usually somewhat less than half-way from CV 3 [6} to CCV 4 {a}, and this is true for many varieties of British and ‘American English, However, some speakers of American, English. com- monly. vse a-closer, fehlike vowel in. sat ett. distinguishing it from the traditional ve! by Ieagih.(and sometimes. dipithongization), thus. pro- ‘ouncing the:pais:setisai and said'sad as something Like (sztV/[seu] oF (seat), [sedVsezd] or (sed. Indeed, in the Midwest (very commonly in the area of Detroit) what I have called the traditional relationship between these two vowels has been virtually reversed in the speech of young people: In this moder pronuac= ation, the old la isia very front, approximately Cardinal; and: usvally200. SOUND-SYSTEMS OF LANGUAGES SOUND-SYSTENS OF LANGUAGES 201 Teble 11 Table 12 Pronunciation sacksex massless ‘Midwest phoneme Pronundition ‘RP phoneme Promuaciation “Traditional [seisl/seks) ([mes}fenes] A w fa New [seaksjfscks] (neas)lss fo diphthongal 9}, waile the old [2] isa short moaophthong, lower than (e] and considerably centralized, whith we might transcribe phonetically a (3 ‘This new pronunciation is particularly noticeable before voiceless stops and Frcatives, asin such wore-pairs as scts/vex massfmess, Table 11 shows the ‘uaditional and new pronunciations, Open central to back vowels: the vowels of father, cot, caushe. TARR, the words father, cor;and éaught contain three quite cistinet vowel phonemes, which we represent as fa//o! and fl. The fist is asomewhat long rather back open unrounded vowel which can be represented phonetically as[@], meaning aa opea vowel somewhat advanced from the position of CV 5. The vowel of cor i short; very back, and very open and somewhat rounded: ie, approximately the rounded equivalent of CV 5. The vowel of axis close to CV 6, thats, (9 In phonemic transcriptions of American English one tracition (ecempli- fed, for example, ia the American Pronouncing Dictionary of Kenyon and Kaott) represents the vowel of the first 1wo words as /a/, seeming to imply ‘that ths is, phonetically a back vowel. The vowel in thete words, however, is never (I believe) very back, in the kind of Midwest American that | am representing here, and indeed, ia the proaunciation of the-young, may be Quite front, approximating to CV 4 fa}, so that there is no justification for representing it es /a/. The.vowel of ccught, onthe other hand, isa very open back vowel, with open rounding. It is commonly represeated in phonemic transcriptions as //. But itis important to remember that in the Midwest it is nowhere near CY 6 {a}, but much closer to CV 5 rounded, that i [o), the vowel of RP cot, the word caugit being pronounced as (approximately) [kozt], Much closer vowels. oftzn somewhat diphthonsized, occur in New York, where caught may be pronounced (kort, kot), ex even {kot}. Table 12 sums up the relationships between the to Midwest and the three RP phonemes with an indication of how they are pronounced. Remaining Back Vowels ‘ Jou! in coat. Most usually a diphthong starting as a somewaat rounded vowel, between CV 6 [9] and 7 fo} but centralized and gliding towards [0] (ol ‘yeaughe fol but varying a good deal both in terms of the height and the centrality of the starting poist. In modern RP, commonly starting completely unrounded and central; type [20], and sometimes virally unrovnded throughout, hhence fev]. May be retlized as a monophthong (0!) in some varieties of ‘American English, and in the English of Scotland and northern England, ol in put. A moderately rounded, highly centralized vowel at about the halfeclose level, ic. a relaxed, and centralized CV 7 [6], or a slightly rerracted and lax half-close ceatral-rounded vowel fol. In the word good. particulary, but not only, in American English, thie vowel may be totally nrounded, thus being a strongly centralized [1] ul in boot, Traditionally regarded as 2 close back sounded vowel, and hhence transcribed phonetically as /a/. However, in virtually all types of English it is not fully back and aot fully close. In American Eaglsh, in particular, it is often articulated at least halfway from back to central, sometimes even more centrally than that—which is why Americans usually hhave-some difficulty in acquiring a good. CV: 8. Moreover, as mentioned cartier, it is often somewhat diphchongized, the diphthoagization usually taking! the form of progressive ip-ounding, hence pechaps something approximating to [dt] or even (i. Readers should carefully check their.owa Pronunciation of fu. . Central Vowels. Ind in cut. Traditionally represented in phonemic transcriptions of Eng- lish by /al, which stristly speaicng indicates CY 6 {o] uarounded, that is, a fully back half-open unrourded vowel. A fully back [A] vowel oceuss in some dialects, in Scotland, for example, but in the types of English we are dealing with here the vowel represented by /l is mors central. In.the Midwest it tends to be a little below half-open and advanced ia varying degrees from back towards central, In RP it is much more open, and may be slightly in front of cent {s/ or /o/ in tird. The RP vowel is a straightforward uarounded centcal vowel, varying from hall-open to neatly half-close, In Midwest American it202. SOUND-SYSTEMS OF LANGUAGES is commonly a rhotasized vowel of approximately a halflose central te ‘back type, sometimes with slight lip-rounding, represented by /2¥, in which the little hook, reminiscent of the kook of an r, suggests the thotacization. The setual tongue-configaration of this vowel is approximately as follows, The tongue'is bunched up into an almost spherical shaps (the apex being somewhat crawn back into the body of the tongue), coming: rather close to the soft palate asf to articulate a éorso-velar approximant, Simultancously, a rather short, but deep, longitudinal furrow is formed ia the-dorsum of the tongue, somewhat lke the furrow in which the wvala vibrates in the articulation of a uvular trill [x] but a little further forward, |At the same time, the root of the tongue is slightly retracted into the pharynx, Readers who are Midwestemers should experiment, isolating the vowel of words such as bird, curt ur, her, and articulating it voiced, voice les and silently, Others may produce it by meticulously trying to replicate the tongue-configuratioa just described. Major Dipkshongs. ‘We have already described the two miner diphthongs /ey/ and foul— ‘minor’ because they do not usually involve a very extensive movement of the articulators, and may, indeed, be represented by monophthongs in some varieties of English. lav in high height. A diphthong starting at, or a litte behind, CY ¢ fa}, and gliding towards i]. The precise starting point and ending point of the diphthongal glide vary somewhat from one individual to another, and to some exteat alloptionically, i. from one phonological eaviroument to another thus, the /al/ of highs longer, than that of aeight, and the end point of the dipathongel glide may be lower. The reader can determize het! his own variants of /at/ by experiment. Jaw in how out, This diplthong starts at about the same plac: as Ja and slides upwards towares [o], the lips, starting from a neutral pesition, becoming more closely rounded as the glide proceeds. Analogous variants, as forlash, RoBi ate lou in boy, This diphthong starts at a point between hal-open and oper, with rounded lips, and glides towards fi). é American English has an additional diphthong fut in évoy, absent. in REY “ Centring, or If diphthongs. at . In RP there isa set of diphthongs ending in a ceatral vow! of {3} type, occurring in auch words as idea; beat tol bare, Baird leak poor gourd ius); SOUND SYSTEME OF LANGUAGES 203 and, for some sgeakers, pour boeroa (otter smapy have/o/inthe lst v0 snoeds, In those eases where there isan orthographic ria the word, Mig= ese American hag ia fact, aa 7, that i the typical American if described tap 161-2, 194-5. This is infact, virtually dential with he rhotacized “Jowel fal deseribed on pp. 201-2. so that we may represent such words 2s ened, Rand, gourd, board as Pera! ed (gra | ard! or possibly ox! Poardl, te “The sound system of English consists, of course of much mors than we pavelooked at here. There are, as we hinted above (pp. 187,188) units larger than the segmental phoneme, such as the syllable, the foot, and the tone troup (or intonation eantsu:) which nave to be incorporated into a Srecripton ofthe phonology of English. And phonology isconcerned aso seth rules relaing stressed to unstressed forms of syllables :0 rues relating the diferent phonological forms of word groups such as phovograph, tography, photographic and s0 08. oor go joa te scopes ss bok, All intended 10 do here was, by looking at some of the details of English phonclogy, to STestrate the point roade at the beginning of the chapter, that languages ‘pate an orderly selection from the human sound-producing potétial, ané Gevanize sounds in 2 highly systematic and rule governed way into the ‘phonological forms that are the outward and audible manifestasions ofthe words and sentences that make up language11 Review ‘We have covered 2 good deal of ground in this book, anc looked closcly at ‘many espoot of the production of speech. At times, while the reader was learning, by experiment, the various, taxonomic categories of phonetss, it may have seemed dificult to see, the wood for the trees. For that reason, it seems desirable to give here a brief review of most of the matters we have dealt with, : 'A speech actor speech event, "as we Sew in the frst' chapter, eaa be studied in any aie ot more ofa serés of stages, or phos, itarting with seuringuinid progratising “and passing on through aéuroniusculer, orgiaig, aetoaynamic, acoustic, ard neuroreeeptive phases'to the’ final stages of neurolinguistic identification of sounds and ultimate'deéddling and comprehension." odie ‘The actual production of the Sounds of speech is, as We Saiyan dere dynamic provess—all voeel sounid sire generated! by the passage of a stream of aic through the vocal tract: drivin by the movements of various ofgans, and modulated to produce specific types of sound by moveinenté and!pos- tures of others, The categories used in phénetics for the cescription’and ~ classification of sounds are, thereforé, based very largely on the organicand aerodynamic phases of speech ve . ‘From now on we camry.out our review of the pasaetic categories mainly ‘in the form of an extended list or glossery with brief explanations Productive components of speech Incetion. Aa activity inthe vocal tract which comprestesorrarees the ait in the tract, and hence initiates, or tends to initiate, an air-stream. The ‘organ or organ-group involved in initiation isan initiator. Articalation. An. activity in the:vocal tract (chiefly inthe part of the vocalj 206 REVIEW tract above the larynx) which interrupts, or modulates, the airstream in sack a way that a specific type of sound is generated. The organ involved in articulation isan articulator. Phonetion. An activity in the larynx whieh is neither intiatory nor articular tory in which the air-stream is modulated by its passage through the glottis (the space between the vocal folds) before being finaly ‘shaped’ into 2 spesific sound-type by the articulation. Types of initiation Pulmonic Préssure. The air-stream is initiated by the'Iuags (nence ‘pul- monic’), which contract to generate positive pressure and thus to initiate an outflowing, egressive, air-sream. By far the commonest initiation type, and the only ote fegulrl used in most languages including Engi. Pulmonie Suction. The aicstream is initiated by expansion of the lungs, generating negative pressure (suction) and hence aa inflowing, ingressive, air-stream. A starp, hissing, intake of breath, caused by a sudden pain, for instance, i initiated ia tis way. In speech, occasional worés may be uttered with pulmonic suction initiation, and mort extensive talking on.a pulmonic sigtion air-siréam may be done for fun, orto disguise the voice Glottalie Pressure. The initiator is the larynx, with the glotis closed, The laryx is thrust upwards, compreising the air contained between the initi- atory closure at the glouis and some closure or narrow strieture in the mouth. Glotialie pressure sounds are also called ejectives, and. mos: com- monly are stops, or affrcates, more rately fricatives. The IPA diacritc for jectives isan apostrophe, asin [p' [¢]fi=I0] ete. Glottalic pressure sounds are particularly characteristic of languages.of the Caucasus and of many African and American Indian languages. Glotialic Suction, As. with the preceding, the initiator isthe larynx, with closed glottis, Unphorated glottalic suction sounds are, in fact, “inverse ejectives’ (sometimes called ‘injectives). A downward thrust of the lary=x generates negative pressure (suction) between the initiatory closure and.an articalatory stricture in the mouth. The commonest glottalis suction sounds (and'they are not very common) are what are called voiced implo- sires. Ta thesa; the glottis is not tightly closed, but disposed as for the production of voice, so that as the'laryax suddenly jerks downwards, con! siderably lowering the air pressure above it, there is a brief leakage of air bias REVIEW 207 upwards through the glottis producing voice. Voiced implosives are repre- sented as (6) [6] (g]. They occur in Sindhi, and in a number of African’ and American Indian languages. Yelarie Suction, The iaitiator is the tongue, which forms an airtight closure at the velum (uence “Nelaric’) trapping a small quantity of air between that and aa articulatory closure further forward. A slight downward motion ‘the tongue rarefies the trapped cir so that when the articulation is released - there is a sudden, noisy, influx of air into the mouth—the mechanism of lide sounds, eg, (I, dentalveclarclick, the sound represented by ‘tsk tI’ of ‘ut tut” in Bnlish; and {0} lateral cick, Velaric suction sounds are regular speseh-sounds ia Bushman, Hotentot, and a few other languages of southern Aftica., Holetiess. Gtotis wide open, as for (9) (9 DA fe LE] ete. And voiceless Jateral, nasals vowels, ete.:( (] lJ es. Whisper (whispered sounds). Gots narrowed so thatthe at-stream pro- duces a rich hushing sound as it paises through. Glottal narrowing also occurs in prephonation during the pre-release phase of voiceless unaspirated stopsand affricates Voice (Voiced scunds). Giottis narrowed, vocal folds vibrating, asi [b] [4] WEG N Mee, | Creat, Glottis closed along most ofits length but with a very sal vibrat- sag segment near ine fron: en¢ trough wits low-feausuey bursts of air escape, producing a erackly sound, sometimes also called ‘glottal ry. Breaiky voice Glotis,rather wide open but very high velocity sirfiow so that the vocal folds are flapping in the breeze’ ‘ Whispery voice. Vocal folds vibrate to produce voice, but air is also escaping through a parrow, whisper-type, chink. also called “mura Glotal stop, tightly closed glottis. is sometimes included among phonation types, but is more properly regarded. as either a type of articulation (ia glottal stop 7) or assimply a concomitant of glottalic initiation.208 nevew Articulation: stricture types Maintainable stricture types: ie. those in which the essential articulatory posture cen beheld for a soasiderable time, ‘Stop. Articulators come together (approzch), forming and holding a com- plete closure (hold) with buildup of (positive or negative) pressure behind it, sudden release of closure resulting in a noise-burst, A stop may lack approach or release, but the intervening hold is the one absolutely essential feature, and it can be maintained for some time. This is why stops must be classified as mainzainable sounds, even though the noise of the air escape on ‘their release is only momentary. Examples: [p] 1] (P] [4] [' f'116) [ol ee. Fricative Sound formed with very aarrow articulatory channel, such that flow theough itis turbulent, generating a hiss noise, whether voicad or ‘voiceless. Examples: {f] [3] [v] [2] x] [y) (] ls) ete. ‘ Africae Sequential artcslation—stop rleased into homorgans fictive (wiekn one and the same sllable). Examples ps) (2 i] Gomedies ‘written in non-IPA (&)) [ds] (aon-IPA ff) [ex [ey] ete. Approsiniil, Sound formed ‘nth sighty wider ardcuator) channél then feisaive—urbuteat flow when voisiess, aon-rurbuleat when voiced, Exampes i 0106 (a Resonant Soe with ety wide articulatory ckanne!—no focal tartulenes even when voiceless. Examples: some vowels, such as (¢] (2). The term is not much used, because the only examples are vowels, and they are classified in 2 diferent vay y Truk One arse taps cepeately against ater (usualy af a quency of about 25 to 35 Hz). Trills are maintainable because they require a static articulatory posture to bé eld for an appreciable tian: the repeated tapping ofthe witing organi eatly caused by the ir-steam Causing it to fap in ta breeze. Examples x] and bilabial ri. Obsiruent and soi-ebsrvent. Thee axe vse cove ims fr dass of sounds: stops areetes and feats are ostruet, the rest ae generally regarded as nor-obstruents, though the position of trill may be somewhat. ambiguous, Momentary (non-maintainable) stricture types: i.¢. those whose articulation is essentially a momentary gesture, not a maintainable posture—ihat is an pproach andlo: a release with a0 intervening fold. REVIEW 209 Tap ond Flap. One ervculator makes momentary contact with another, either ina “Ricking’ movemeat, eg. the apico-alveolar tap [1]. of by ‘momentarily strkingit in passing, eg, the cetcollex ap [t] Semivowel. A momentary spproximant-iype articulation—an approach and/or release with no bold. e.g, GJ, en approach andlor release, or depart- ure from, « palstal approximant articulation (which, if maintained, would be an i] vowel) (w}, approach and/or departure {rom an [u position, ‘The transverse dimension of articulation ‘Median articulation. Most oral articulations require the formation of an obstruction along the sides of the mouth, leaving open aa rticulatory chanrel along the centrs, or median, line of the mouth—vr, in the case of stops, total, side-to-side obstrvetion usually followed by a meciaan release 8. (6 (0 (5 2) fe) fx Gf ete. Since median articulation is by far the commorest type it is normally net explicitly indicated in descriptions of sounds. Lateral articdation Thereis an obstruction somewtere along the median line of the mouth, leaving open a lateral channel or channels along one ot both, sides of the congue, Examples lateral approximants fl) (4, lateral fricatives iE. : Articulation: Locations Articulatory Areas ‘Nasal area. Involving lowering of the soft palate so that air passes dit ‘rough she nove: nase! erioulation, ex fa] fo] bl ete Nasaized, with sir possing ou: through both mouth and nose, ee. 18] “ete Orel ares. Articulation in the mouth, &g. (p] ) 8) @) 8) @0) Elle ets. Pharyngeo-laryngeal aree. Articulation ia the pharynx, [B] [5], and glottis (2) BE ‘Principal oral articulations | Bilebiat: 2-2. (p] [>] (9)(B); Labiodentai: eg. (f]{v}; (Apico) dental: e.g. (8) (9), (0); Abeolar: eg. [t)[4) [s) (2 (al: Postalveoter:(1)(3) ah: Retroflex:() TGs) EDU; Polarok eg. fe} 4) Co} EDEN Le) Gs Volar: (eC (2) [fos Urea: OCI. i 2210 aevEW Co-articulation and sequential sound-types Co-ordinate or Double Articulation: simultaneous articulations of same stricture typeat two different ocations, e.g. simultaneous [5] ~ [Kl = FR). (P] © (= OD (6) + i) =[Ga). There are special symbols Tor simultancous approximant articulations at the labial + velar locations, [w], and at the labial + palatal Yocations, (0 Secondary Articulations: an opecer articulation (usually of approximant type) superimposed on a imaltancous closer articulation. Lebialized, Sioltandous tp-rounding, ©. (1 [8°) 6°] 8 Polataiced Simultaneous, raising of tongue dorsum towards the hard palate, £8, [P] 0] [9] ec also symbolized as (3) [4] ete Velorized. Simultaneous rising of tongue dorsum towards the velum, eg. mages. Pharymgatized. Simultaneous compression of pharyax—syrabolized a [€9] 1 fo} ete. erie Sequential souiidtypes: these re’ homorganic sequences whic ai >e regarded as unite, and may function as units (unit phonemes) in languages Geminate. Sequence of ientical sounds, within onc and the same word or morpheme, eg. tl] in Italian note, bello, Arabic Hn’ Allat, ee,” Agrieate. Sop released inte homorganie feativn ©, [26] [s] (fee (0 [éslete. within one and same syllable and morpheme, Lateral africate, Stop released into bomorganie lateral fricative, «&.(t] (#] {as} Lateral plosion. Release of stop into @ hormorganic non-fricative lateral, often syllabic, eg. [ef] and ai} in English bol, mil. Nasal plosion, Release of 2 stop into a homorganic, often syllabic, nasal ‘eg. {ty] ane [én] in English button, sudden. Prenasalized stops, Homorganic sequence of oasal + stop at the begianing ‘ofa syllable and functioning as a unit phoneme, e.g. [m+] {ad+] [9g], rather common in Austronesian languages ‘Non-homorganié sequences, “Heverorganic sequences. The articulators used for the successive sounds are completely different, e.g. [Kp] and {sp) in back part and lisp. REVIEW 211 Contiguous sequences. Articulations of successive sounds made by adjscent parts of the same organs [ti] in tree and (kin beckyard Diphthong, Sequence of two vowels within one and the same syllable, eg. {odin high, (a) in how, and, in many types of English, ft] in dey and [] ox foo] in go. Diphihongs such as these have decresozado or diminishing stress and are calle falling diphthongs. We can indicate them by masking the lst partof the diphthong, the weaker element, wit "thus (ot [a6] ate. Diphthongs with oresoendo or increasing stress are called rising diphe thongs and are indicated by placing the mark {"]over the fisstelement. The Enalish words yer, yawn, you, et. might de transcribed: (i) (oa) (fe, but these sequences are usually recarded as mage up of semivowel+ vowel, and ace ranseribed {jis} Go] Gus)- Close ad open transition between consonants Close transition, Successive Consonants articuleted as, closély togetber as ‘possible, In heterorganie sequences there is articulatory overlap, ic. the seo- ond articulation is formed beforé the firsts released, e.g, (p) in play, [kt] in actor. Taleéntiqueus sequences, there is articulstory accommodation, ic. the rtisulation of one consonant accommodates to’ the articulation of the ‘other, e2. (K] + Gj] in backyard, where the [K) Is slightly palatalized to sccommodate to the palatal {J thus fe, or (+ [6 in eighth, where he [t becoiucs dental to accommodate to the dental [9], and thus {6} In homor- ‘ganic segueness the articulation is simply prolonged, e.'Tk] + (kin Black ‘cat, where there i @' single, long, velar closure Open cransition. In open transition thes is no such overlap, aocommods- tion or cobtmuty of arucwator; instead thee b a inbueata wlaase and rmartiglaton. One can hea contrat between clot and open transition in such English pairs as back part foakpa(:)t] with close (overlapping) transi- tion kp} and back Sparc oekrpo(.)], where the ulta-shory, often voiceless, (is tls more than a mere open tatsition between ché consonants. Com pare. also, {pl] in pligte {plast) with [pel] in polite, or (ka] and [ks] in cracks with fed and [ks] it Caracas, ete. Vowels Parameters of vowel classification Vowels are tasically classified in terms of lip-position, vertical tonguer position (tongue-height), and horizontal tongue-positien.Review 213 Lipsposition. We distinguish unvounded vowels, e. ( () (2) (a) fv] [x] from rounded vowels, 8. (y 2} (6 (0) fu ec. Tonguesheight. The degre= of convex ‘bunching up’ of the tongue, from very bunched up, coming close to the roof of the mouth—high of close, e.g. (1 [ [uw] 10 very low and rather fat (though still coavex)—iow or open, 2.2, [a] (0) 6}. Intermediate positions—hal-close, eg. fe} (al (6) (0) fv), halfaper, og [a] [«) la) Horizontal tongue-position: the location of the bunched-up tongue in the front-back dimension of the mouth. Vowels produced with the tongue thrust forward as faras possible aad bunched up in the front of the mouth are front vowels, [i] [y] [el [] [el (2) [ce]. Vowels produced with the tongue drawn back as far as possible and bunched up at the back of the mouth areback vowels, eg. (6) {0} [6] (a fo] fx fu [ws]. Vowels produced with she tongue about ball-way between the front and back exteemes ae central vowels, €. (fl [e] @ etc Cardinal Vowels. set of universal reference vowels, invented by the Eng- lish phonetician Daniel Jones, by comparison with which virwaly any vowel of aay language can be specified with some accuracy. The Cardinal Vowe's (CVs) are derived from the idea of 2 ‘vowel space’ within the mouth, circumscribed by a ‘vowel limit’ beyond wich vowels eapnot be produced. For example, CV 1 [is the highest and most front possible vowel: i the tongue is bunched up any higher, or pushed any further forward, the sound produced will be a palatal or prepalata) fricative. The CV 1 (i thus marks the closest and frostest point on the vowel limit, ‘Allthe CWs are on the periphery of the vowel limite they consist of the ost /ront possisle vowels, at approximately equidistant points from high- est possible to lowest] (e] [6] [a] and the most bask possible vowels, at equidistant points from lowest to highest—fa} [9] [0] [x]. There is d set of secondary Cardinal Vowels, with exactly the same tongue-positions, but the opposite lip-positions These are front: fy fo] fo} (2); beck: (0) [a fl [a] ‘Thess refereace vowels are all teaie and absolutely monophthongal. Practicaly none is exactly the same as any vowe! éf English: so English speakers have to be particularly careful in learning the Cardinal Vowels, Acoustics of vowels ‘The vibrating vocal folds in the tarynx generate 2-complex sound rich harmonics The tongue- and lip-coafigurations for vowels shape the mouth into a series of resonators. As the sound of voice passes through the mouth resonators, they pick out certain bands of frequency within the comolex sound-wave and emphasize those. For exch vouel the acoustic energy is thus concentrated in a particular set of frequency baads called formants (Fo). Mos: vowels can be pretty well charseterized in terms of the two Lowest frequency formants, Fl and F2, Thus CV 1 fi} ean be described 2s having Fl at about 240 Hi, F2 st about 2400 Hz, CV 3 [2] has Fl at about 810 and F2 at about 1900 Hz, CV 6 {a} has F1 at about 500, and F2 at about 100 Hz. Vowel sounds are usually heard as unified sounds of particular quatiies, within which the individual formants cannot be separately perceived Howeves itis possible wo become aware of Fl and F2 as fellows, ‘Tapping the throat, with the glottis closed, generally makes F1 audible, Secondly, though the hushing sound of whisper is complex, 2 frequency repreienting F2 is usuclly audible in whisper. It is casy to ideatlly che ‘whisper-pitch’ (= F2) of various vowels with ‘musical notes, and this is a ‘useful aid to the learning of some voivel sounds’ Additonal vowel modifications i Nasalization. If the soft palate is lowered while a voisel is being produced, rabalized vowelsare generated, 0g. | Rewroftexion, if the tip of the tongue is turned up while a vowel is being produced, the vowel is retrofexed, as in some varieties of Midwest Arseticaa, particularly with [o}] as in far, fo] or [ot asin bind. Rhotasization. Rhotacization, or ‘r-colouring, is sometimes confused with retoflexion, but intact it doei nat involve the wpearning of the tongae-tip characteristic of retrofenioa, Rhotacization isa rather colmmon feature of the vowel [2] of bird in Midwest American, It is produced by two simul- taneous modiications of the configuration of a tongue which is rather ‘bunched up into approximately a half close central to back position some sstraction of the root of the tongue back into the pharynx (slight deep Pharyngzlization),.and the formation of a hollow in the centre of the tongue about opposite the uvula 5 (8) [6] as in French un bon vin blane, Prosodic features Prosodic features, or prosodies, are shonetie features that characterize relax dons becween segmental sounds, or stietches of speech more than one segment in length,214 REVIEW ot Initiatory prosodies Stress. Initiator power—the energy expended by (usually) the pulmonic initiator in pusbing sir up from the lungs against the loads imposed by phonatory and articulatory strictures. Strss is marked in PA transcrip. tions by vertical lines placed before the relevant symbol: ("] means very stronshy stressed, () means stressed, (] means weakly strested, and absence of seressamarks means unstressed, or most weakly stressed Spllales. Mizor power peaks usualy separate fom each othe by latory strictures. Rhythm. Distibution of durations and timing of initiator pulses Feet. Rhythm units-—unitary bursts of initiator power. in English initietor power appears to be delivered in a series of bursts or pulses, each quickly reaching a major peak and then dying away. Those ‘ft tend to be of very rovehly equal curation, irrespective (within limits) of how many minor power peaks, or syllables, occar withia each foot Phonatory prosodies Inconation, Meaningful pitch variations (ie, variations in frequency of ‘vocal fold vibration) often characterizing long stretches of speech, which may de many syllables in length, and relative'y large grammatical units, such 2s the centence: often used to distinguish statements from (certain kinds of) questions in many languages, such as English, French, and many otter languages in Eutope and elsewhere, | Tore. Meaningful pitch variations chéracterizi short stretches of speech, fia of oily one syllable in length, and the shdeies! grammatial vii words and morpiemes: Used in many languages such as Chinese, Tas many African and American Incian languages, to distinguish one word fromanother |. | “ Anticalatory prosody Duration, The length of time that aa articulatory stricture is maintained, ‘known as length ot quantity, In IPA transcription duration is indicated as follows: long vowe or (less commonly) consonant: {} half long [2s shor {a [p,ultra-short (or ultea-weak) [3] (5, Review 215 Sound systems Phonetics. Tae stady of the physiological, serodynamic, and acoustic characteristics of speech-sounds Phonology. The study of how sounds are orgaoized into systems and utilized in languages. The human sourd-preducing’ potential. A universal, differently exploited by different languages. For example, the continuum of possible vowel-sonds is differemily dissected by different languages (c.g. English and Spanish). ‘The continuum of different possible voice-onset-times (VOTs). VOT = time iffercnce between release of articulatory oral stricture and onset of voising fora following vowel. Phoneme. Minimsl sequential contrastive unit of phonology used in the buildup of the phonelogical forms of words. Distinctive Feature. Minimal simultaneous contrastive component of a phoneme.For Further Reading CHAPTER Introduction Tluminating but quite different invroductions to phosetcs arz given ia Aber crombie (1967), Chap. 1, aad Laver (1994), Chap. (On the phases of speceh in general see Catford (1977), Chap. ‘, and Ball and Rakilly (1999), Chin 1, pp. 13-18, The aerodynamic phase is dealt with com- prehensively in Catford (1977), and in a highly technical manner ir Shadle (1997). For the acoustic phase, set Catford (1977), Chap. 4, Clark and Yallop (1999), Chap. 6, snd Fry (199), ‘Thee isa relatively simple deteription of the vacel suet, and éf the fonctions cf the organs within it, ia Catford (1977), Chap. 2. A more detailed description is ‘ven ia Clack and Yallop (1995) Chap. 6 CHAPTER 2 Basic Components of Speech Initiation types (ai-sream mechanisirs) are gone into in detail in Catford (197), ‘Chap. 5. Seealso Laver (1994), Chap. 6, and Bal ané Ranily (1998), Chap. 2. CHAPTER3 Phonation Fhiodion i disease in'Caers (197), Chap. 6, Numerous pes of Yoke, whisper ad vowe qulily ae desnbed in Caton (1968), Laver (1999, Chap 7, and most extensively in Laver (1980). CHAPTERS 4 AND S Articulation Articlory locaton and trctare pes ce deal with (genera a ‘psc and ‘manner’ of articulation) in all manuals of phonetics. Both of these parameters of srzuaton are deat with in great detail Gretoding aeodyname dat) in Catiord (1977), Chaps. 7 sind §. Live (984) has ecllent deséiptons and digas ofthe anicunicns of stops da fines Guay, Baad 9, Cash sal Yat (1995, Cas 2, cued Segrenal artaucten, combines socnet éesriptions of intisti, phonation and articulation. Ball and Rahilly (1999), in Chap 3 on Speech Articulation, tals ie ofthe category of mainidincbiliy or prolnabilty tat was introduced in Cato (1977 and used inthe present Book chatens Conarticulstibn snd sequences ‘Multiple artioulations and-sequences are dealt with in Catford (1997), Caap. Ll,218 FoR FURTHER READING ang Laver (1994), Chap. 11 om Multiple articulctions and Chap. 12 0n Intertege mental coverdination, See also Ball and Rahilly (1999), Chap. 7. The latter two contain useful diagrams ef multiple articulations CHAPTERS 7 AND 8 Vowels On vowels generally se expecially the excelent survey in Lover (1994), Chap. 10, ‘The Cardinal Vowels and the theory underiying them are well describes in Aper- crombie (1967), Chap. 10, and Abercrombie (1985), and, with some criticism, in Ladefoged (1995), Chaps. 8 and 9, where the acoustics of vowels is also dealt with. For genera! ietroduetiont to the accusties of speech sce the referezecsciteé for Chap 1. CHAPTER 9. Prosodic features Prosodic features are deseribed in Chap. 9 of Clarke ake Yallop (1995). See also Leniste (1970), and the detailed analyses in Chaps. 14 10 16 of Laver (1994). CHAPTER 10 Sound-systems of languages Some aspects of phonology are dealt with in Abercrombie (1967), Chap. 5, and Ladefoged (1995),' Chap. 2, On feature systems see Clark and Yallop (1993), (Chap 9, Jaxobson, Fant, and Kalle (1963), and Chomsla and Halle (1968), Chap. 7. For general descriptions of Enalish see Jones (1970) and Gimson (1989), on British English, ané Bronstein (1960) and Kenyon (1997). on American English. ‘The latter the 12th edition of a stundard work, inehides spectrograms of the sounds of American English. CHAPTER 11 Review [Acompresed survey ofthe fe of phonetss is provided by Catior (1954). ‘A opie tha isnot expiciy dest wit inthis books phone aoaton, Foran excelent general and historical introduction io this subject see Abersrombic (4967. Chap 7. See also Laver (1994), Chap. 19. For general information on the IPA and specimen transcriptions of 29 languages, sec the Handbook of rhe Inceretionel Phonetic Asocation Tadeioged and Madiiovon, The Sounds of she Word's Languages isn esazemly useful referenie work, which ight be rpandcd as an epplcatin of cverything tht san be lard from the present boo. cis the mos eommpreea- Site surrey of human speech that exits containing esenptions of sounds largely supported by instrumental data, Sooner or later students of phoretiss il be reading ates and books like te pececing one that make use of the Sndings of instrumental phonetics. It is hrsfore desirable that they have some knowledge of the techniques: of FOR FURTHER READING 219 jstrumental investigation of speech and how to interpret instrumental dasa, A. ‘004 introduction to such matters is Painter (1979), and Chap. 12 of Bal aad RRabilly ia short and readable survey of the topic. A comprehensive, but very technical, survey of speech eesearch methods is Lass (1996).References ‘Aaercnonare, Dawn (1967), Elements of General Phonetir,Easabargh University Press Asmcxonae, Davao (1985). ‘Daniel Jones Teaching’ in Fromkia, VA. (Ed) "Phonetic Linger: essysin Bone of Peter Ladefoped. Orlando FL, ‘Academic Pres, 15-24, BALL, Macros J and Rasauit, (1998), Phonetic, the Seionceof perch, ‘Loncon, Eawin AmoW Lid, USA, Oxlord Univessey ress BRONSTEIN ARTHUR 1 (1960), The Pronunizton of American Engl, New York, Applton-Ceatury-Croft, CCartorD, J.C. (1954), Phonation types: the classciton of some lncyageal ‘comporents of speech production’ in Abercrombie, D. eta. (Eds) In Horowr of Daniel Jones, London, Longman, 26-37. carro, J. (1977 Fundamental Problems in Phones, Eaisbureh Universty Presi ad ladiaca Univesity Pee : CaTrORD, JC. (1994), Phonetios, Articulatory’ in Asher, R.B, (Ed.) Encyclopedia of Language and Linguistics, Oxford, Pergamon Press, 3058-3070, Carron, J.C. and Psowt,D. B. (1970, ‘Auditory vs Articulatory Tiiaing in rote Sounds’, Modern Language Journal, $8/1, 877-8. CCwowsey, Nand Hate, M (196), The Sour Pater of English, New York, Hargec Row. cus, Ind Yartor, C. (1995), dn Introduction to Phonetics and Phonology. Oxford, Biackvell Publishers Fre D, B, (979), The Posies of Spec, Cambridys University res Gisos, A.C. (1589), A ntrdetin 0 the Pronunciation of English, London, Edvard Amold Lid. Handboo' ofthe International Pkonec Association (1999), Cambridge Univesity Press Haans, fon G, (1999), ‘States of the glots fr voiceless plosives’, Proceedings file ite Intemational Congras of Phonste Srimens San Francisco (Aug. 199), vo 3, 2341-2044, IAKORSON, R., FANT, G. 206 HALLE, M, (1863), Prelrnaris to Spcoh Analysis, ‘Cambridge, Mass, MIT. Joves, Das (1970), Ouilne of English Phoreties,8t edn, Cambridge Unitesity Press anno, Jon S. (1997) Amerisan Promurcavon, 12th edn, expanded, Lance, Donald M. ané Kingsbury, Stewart A (Eds), Ann Arbor George Wat222 RerereNces Kanwox, lors $. and Kworr. TA. (1944), A Pronoxncing Dictionary of American English, Springfcid, Mass, Merriam. Lanerocen, 2 (1967), Tiree dreas of Experimental Phoncics, Landon, Oxford University Press LaDEFOGED,P: (1993). 4 Course in Phonetics, New York, Harcourt Brace Lapsrcom, P. and Manoasow,[. (1996), The Sounds ofthe Worlds Languoge, Oxford, Blackwell Publishers Lass, NORMAN J. (1996), Principles of Experimental Phonetics. St. Louis et. Mosby. Lauren, A. and Connasy I.D. (1981), Te Epiglottis a an Artoulatee’, Journal Of the IPA, 92, 3-5, Laven, JouN (1880), The Phonetic Deseripicn of Yoice Quality, Cambridge University Press Laven, Joun (1894), Principles of Phonetis, Cambridge University Press Lavasre lise (1970). Suprasezrentls. Cambridge, Mass. MIT. Parwtin, Cots (1978) An Trivoduetion to Frdtromental Phonetic, Baltimare, ‘University Paci Press Suiaois, Crmisrve (1997), ‘The Aerodynamics of Specch' in Hacdeastle, W J, and Laver, J. Eds) Honsbook of Phanesie Scences, Oxford, Bickel, 33-64, ‘Swarr, Hewny (I8T7), 4 Hancbook of Phonetics, Oxioré, Clarendon Pres. fein FOL ABA AL Had an AR ARI AO MSL “Index Ft male nd ao sccommedaton 113 sent 83,67, 98 sowurse ‘enmlvclnr foeaton 5,86 . ‘ates 155 ‘entlvois tops #3 ta hase 5 postlveolr 579: Dhonetes 5 approach phase 6,70 ‘peste 153 approximant S15, 6, 9,70, scvasienly equidisan vowes 129 ovuricsof vowels 152 soon ‘88thnal, som Candia vowels 18-59 ‘Aeyghe ) 22 aerodynamic passe 45 ProseesS sctodynamic of restive 2-1 otvoiced ope 3.94 fdoeter 107,108 sent phonemes 185, pies 9 ‘agai 78 sieeouna moving 69 sito 21.165 ioptoae si, 192 seshaeons 172 of English 192 spar of phonology of alanguage 192 akeoke tara! 101 pri saben? one AD heolo-falaal, tern not ecomnended 92 ‘Ameren “English 67, 7,199,192 108-9 ape 7892-5 pel accuations 8-9, 90 apco- 79,82 aco 67 38,93 Arabic 96,105, 128,196 ‘enphate consonants, 105 auton [2-13 1 35,37, ‘6D, 6, 8,713, 767, 79.86 289 Gimensions of 9,50 Incntors 73,1, 29 secure ype 39, 71,2089 arczulto 76,77, sesoultony rene funetion 52,58 cevedap 12 Phoaets 75 postures prosody 174 sale 03 secure 64,71. tones 77 aspirated stops 57.58, 181 aspiration 55.56 auditoy Teocoaek $ Actzencian largunges 109 ack vee 1265 of Eagish 200 ‘pase comporens 1 13,35, Berber 96 beendaabial Ddiexoadi Bt biden eave 82 babi 80 ‘eatin £0 ‘nll 80,81 blade 18, 3,85,86 Drea i ret voice 52,53, 5897 Brush agli 97,193 me 198224 wwpex RP 164, 126,198, 146,129, 151,184,192, 198 RP vowel of bin 150 Buckinaa (Agu) dle 78 Burmese 38 ‘Bushman 30 Cantonese 178 Cardinal Vowel ‘ogra (30,131,163 formant frequencies 14 ormantscharts (CVF chars) 155 wich sola sale 18-5 (Cerna Vowels (CVS) 125-3, 199-6 cima v2 37 eva ee even) 38 CVs tie V5 iat evo) a2 Cvs fy 2 ‘Catford snd Pisoni 120 Coueus38 centering ore di tel vous 1-50 Chesten 96 (Chnese <7, 178, 195-6 Cocasian x 82 leks 30,534 ese uacing 1a ese wasn 111 ‘one vowel 62,125, lesur 6,68, 7-1 coartculte! sound 80 coarteulation 9, [03,219 ‘ovordinae (or Goube) 99-100 secondary 101-6 210 cozksey 7 ‘communication anc computer engineer: 2 commutator 182 competent phoostsan 2 comparatvehistoral linguistics? cossoaant clsters 195 ‘les of pertsipatos in 196-8. theoretical number fia Eaglah 196 corsoeants 116, 118 contiguous srelation 113 degeenne 16 contnsous vue 180 ‘conceptual (81 of posse sande 178,181 ot possble VOTS 181, sof English 202 of vowel unde 1780 spec aea 163 exmtastie ents 187 ‘soorinate articulation 9-100 crak 51, 52,55, creaky vice $3 (CY se Carcinal Vowe's (CYP reeCavcial Yowe formants char areauitons #25 rove T, 2,35 dexaveclae ‘sruculsens 67 ‘eon 77, 78,82, 85,86 ones 1,85, 86 detienal (Sect 81, 82 ‘easiabia Geroicng 38 dition (A, 3, C) sale 158 ‘imensions ef articulation Tongitudial 3, time 5,73. rrassvee39,69.73 vol 59, 65,73 iphone 10 ‘ing 11 Snstelement of 110 Sieg 111 secondelemett of {10-11 sisincuve esas 189-50 somal repon77 dopalarteubtions 39-99 ‘ono. pata 89, 90-2 rue ela 9,9, €2,92, 58, dosum 78 docbie (=:0-rdinat)arcuition 100 Dravidian guages 8 uration 57, 163, 168, 1745 egrsive sro 20,91 fngolabia $1 sndoabio. 78, 81 ent tt ngs 33,90, 116,172,178, Mees 35,193 Epproninaats 104 seatral vowels 20, 201 ‘nwaintchter 195-4 ‘pasoaint 191-5 ipbtnengs 189,199, 200,202 ane) ‘henives 193 Fone vowel 159 later 05,195 Mies Amero 199-201 nasal 194 of Aberdeeshice 173, ofthe North of Exgland $6, 192,201 cof Scatlsnd 56 173, 193,199,201, ay (71 semiows 195 fas vyowes 136, 81, 158-202 ‘pitlotal arialaons 95-6 eighots 78,98 {rom te aid in eaming CVS 133 cexhaation (9 olebic 7, 81 features 190 feet 71,188 ‘stehronous 172 ‘sedynaeie 172 ap 65,57, 71,58, flow inouiga gous 32 fot the 171-2 eran frguensits 1S § ‘net absoute1St feet of ip rounding on 158 formant 153 ‘making sudle 156 ‘reneh€6, 11,90, 116,13, 138, 78,384 fequcney 1 een 1657, 59-60, 65, 72-82, 85, 90-1 front wows 134 Ofna 19 fly voted 09 3, $6 fencional components of prec 9 anderen 13 . surge 93 Georgian 36,186 German 108 Gasjow97 planet 95 closure stp 2, 32,36 Sop, Englsh96-7, 185,199 boon ‘eiiation 2, ressure sounds 25,29, 54 sounds in aba © phonation $Y saruon 26,58 suston 20930, 48,58 potas 29 Botte 82, ties ousenlues 3648, 567 alone 15,14, rusose eunding 1st alpen 158,185 199,212 hand pate 783,99 harmonic 153 Burrs 3 Huu 30 Hebe 96 Hera (22) 192 heroin sequence 109, 112,16 Iara of unite 87 dngh vonete 24 Dighet pir of the tongue 127-3 Hind 53,67, 8,186 ‘stone sound-ehange 2 eld phase, 7 homerganiccorsonans 111-12 homerganicsequetees 105,109,116 Inerzonal tongue nove horizontal tongue-poction 134,123 Hottentct 30 ‘bum sound-producing potential 17 Hunginsn 0 lnoplosve 30,47 din 8,134 dian Eaglah 179 Ti angaageess ingresive 20 ‘mulation 19,48, 5 ftiaton 9-12, 13,21, 35,40, 53, $8,206. Sntiaon types 2,206 ole 2, 206 almonic 1,206 ‘iar 26,207 Feit power 31,38, 1624 Inklatoe tele 31 imitistoryprosdies 165 inser rounding Intrational Phonetic Assocation OPA) 16 pragmatefusction 174 Inland 134 Tih Gai rh speakers (of English) 108225 anpex eeheony 172 ‘oeynario 172 lnracl Hetrow 66 Talan90,10T Jammvean English 73 Japenese 9,172 [pt ick wosty 22 Toney Davi 135,129 Kenyon and Knot 200 xy reece pat 29 Lansestete fedtack 5 bial ion 77 [Baintion 101, 106 baled babi 9 abi 79 Inbiadental 6, 79, 80, 8, lamina 79 lnmalaniulaions 8, 86 Jnmno- 52.64 Tabeola 84, 86.7 parr eenaloar 8 “postveolar 85,86, 50 prcpalal 90-1 snnyagopnaryex 35 larynx 20,36, 53, 60,73,35 ‘ising ond lowering 24, 47.99 stoping pith 3,156 129 laryaeiston 7 teal ‘ait 108 aipath 71 Apronumaee 0, 8h, 101,108 Frost 6, 88 Lauter snd Cond 96 ‘eng 178 Upepositon 29 [pspositons of Cerdiaal Vowels»: “0 pounding 157 ‘ppesof is peters os 74, 79,30 exon 73 lengtucial dimension 9-40 loudess 165 low vowels 23 lowe arias 76,77,09 lewer ciety area 78 ‘bintainble aiaultion 66, 97 poetrer eunds & siete types €0, 71 major phshoogs of English 202 sarpasofsyiabe 116 [gies ff ard 200 cestures 50 rounds & stirs 67 rmonophthengal vod, production of 346 ‘moropacheng. Cara Vowes 35134 rotor commaadss Rowhshapes for vowels 20 murmurs Mo Pair Lady sara appointment 74 aerwest articulatory chanel (f vowels) 128 m7 cemcenante 77 lsion 109 ere 3, 14,35 sstaizaion 105, 161 sasnizes 74 101 sasaized vows 73,101, sas 74,80 Neveho 30,108 nevelinguitieidetifention S| seuplinwistie powromsine 4,5 ‘euromuscuae phase, § ‘eurteytive pase 5 Nase Nigest 20 otabsiuent 71 son-turblen tw 62,62 Nort German 6,88 North of Eeglond 134, 138 ote 78,75 notes of dintni sale requncies af 158 ‘eels a sllane 16 of gees or-slde s8 opensounding lat opentrasion 11-16 opeavowels 124, 125, opeang ciate st ispath 60, 9, 71,73, 0273.76.95 fronanse chamber 122 towels 5 one 7,73 fnnizaton 201,103,106 Pastand rounds 3,101,104 Babtoateoti 52 ‘era nt recommended 72 viteatons 153 23 of itor pow 170 Prsrmtator veo 169 pedpheral vor 14 piareea sproximants 9 alton 95 pbartgaztion 101, 105 proryagaie! 10L Prange rye track we of 178 Brees speech 3 ‘Pronition 9,22 35,40, 55,57, 59,207 Bhonstion pes 9-53. 207 Baonston inreaten to inition Sit Bronson prosaces 72 rosea (17-2 plone rymbols 16 Phoceton 9.177 conrastd with phenols 177 ptonslogeal Tereay 187 snot of Bngish 187 phosblogy 177 Poyisspling «7 pach $458 15, 186,172 pach chines poduction of 72 plots 3 ‘acunate devi, val tacts Bots 87 postavenlar subaene 33-5 postrocore- 9 ostrodoru 78 Powe natn 3,168 power peak 19 ‘ener 90-90 59 Ss oe noveond! eethonaton 86 Preworeiniiason 18 primary aricalaton 101 pay Cordial Vowels rotuctve= funcional components of spe 9,205 rorrenepivesensaios 5,78, 84,122,153 proves fenrre 63. 215-18 ‘relation (éuraion) 1,214 ‘nity (ehytm, ses slab) 163,165, 2d ponatery toe, intonation) 172-4216 pulmonic ation 1,33 preoure 18 29-26 48,56, $1, $1, 59, 65.85 1 sounds in elaion to phonation S34 eeton 8, 8 Pymalion | RP rose promunsaton 33 refrence-plat (27 ‘roe tone 133 ‘ieee ae 58,70 resonant 65, 66,71 rete sections” Esp 67,89 a eotenic 6 81 fheeczaton 6 eye 170 z Shei ni 19 Fi oftonave 8, 83,85, 86 tout of mouth 7,75 rounded vowels [24 rovnding 30 Ruan 1,97, 10 “nino ations diatonic 158 ie sa dobry) ee Sod dake scone Soto onlay aon TORRE Sea ome eto no. tees Soria SCV 2fo] iss é&228 mex en SCV fa) 145 SCVaye igs SCV5 fo 6 SCV 5 146 SCV rly 157 SCV fay er segments 163 semowel 65,67, 68,71, 3,20 Sequence of ahonemes 190 sequences hombagan 17-8 ot-homerganie 103-10 voaie LID segue sound types 210 tof disincive features 190 Shape 2 ‘Shaw, Berard silt srteulstion ts introspection 120 pracice (2) simultaneous srscuions 102 omponent 19 Singha t8e Slightly voiced ston soft pale 74,77 Somali96 soneny 165 sound stems 171, 25 South and Souih-tas asin Sparish67,80,90, 160-1 185, 6 vesspinted $6, 181-4 ste 31,163 36 ph in Eagish 167 1 ator power 31, 168 a ducrii or lot “nserthete sensations of 166 of roune and verse (66 stres-aroup 7 street 165 stricture pes oF 17,6, 66, 7 ‘aden’ oftingustes 2 subgloalai presee 173 sutlamina | ‘sarinal 76 ssblumiao-pepaacal 3 deecanneurclation 102 Semtra 6S Ssuprseymentl tues 163 Sean sd Sweet Henry 1, 120,129 slab 238 minimal intaory pulse 168 ‘onsorantl tegang of {70 fontonantl ead ot 170 Passi 116 suc of 168 slabs 16,165, 16,170,186 Syshens (5 symheszingslocale pressure 24 ‘aca ensaons 5,7, 12 "ap8,67,71 ‘sachet of tenguage 1 otresdisg | tecul dvison 77 ‘eet 7, 8,85, 86, teethodg 1,2 ‘eason of veal folds 172 ‘seosaciogy 116, Talla eowetepring 156-7, 139 time dimension 58, 6 73 tone 37,175,134 senaaticfunten 14 teae-group 187 tog 75,7, 20,5245 fonguesojst 123-3 tril 65,65 turbulene 20, 15 shanae! 2 wale furbulnttiow21,82,65 unaspirated 15.36, 58,81 undeetiae 7,88 Univeral ference von 139 lunphoaated unde Se unibunded vowes 1265, nsresid 165 pp artialtors 76.77 sppeesrucltoryrones 77 rae 92,99 sppeseinane 6 ‘ations 955 featvess ‘ all 66 | velar aroaation 92 tane77.78 seine miition 26,7, 59 velaiaion 101, 13,106 vwlanzed scunds 101,108,108, 106 lot of stow 168 vehun 74,75, 77, 99 versal dimension 39.60,73, vertical tongue poution 120, vorsleorse 16 worl oles 367,54, 58 soul 8 orale Sequence 119 088 36,31, 34,63 wooeten 36 oie ose ne (VOT) 15 ‘VOT continuum 82-3 YOT negative NOT pesitve 181-2 NOT Chinese 12-4 VOT, Engish 182-5, OTK French 182-5 voip vation 5 pirate stop 3 Treaties 36-8 slot five 97 btlic suction stop 47-9 plosve 30, 47-9, 33, 58 phonation 35, 59, $4, 18 Rept 36 INDEX 229 ‘wietes ‘eral 36-40 Iara ate 39,63 asl 39 us 66 Phonacon 36,0, 5,58 fost vovele40 62-4 votelestes SI. 93, $e wore 63,115,119, ontinger 17-81 it 1207 ‘nodifitons 161 vowels conti of 152-6 slasifirien of 119,128 outhshages or 130 sent study 120 ‘rational desertion of Wate 124 Wei 39 Wels Engl 172 whisper SO weisperpiteh 1573 sonesoFaation 7-9 2us30