Practical

PHONETICS
J CVfells
&
Greta Golsorv
Practical Phonetics
Practical Phonetics
J. C. Wells PhD
Lecturer in Phonetics, University College, London

and Greta Colson

Principal Lecturer in charge of Voice, New College of Speech and Drama

Pitman
PITMAN BOOKS LIMITED
128 Long Acre, London WC2E 9AN

Associated Companies
Pitman Publishing Pty Ltd, Melbourne
Pitman Publishing New Zealand Ltd, Wellington

© J C Wells and Greta Colson 1971

First published in Great Britain 1971

Reprinted 1975, 1978, 1979, 1980, 1982

All rights reserved. No part of this publication may be reproduced,

stored in a retrieval system, or transmitted, in any form or by any
means, electronic, mechanical, photocopying, recording and/or
otherwise without the prior written permission of the publishers.
This book may not be lent, resold, hired out or otherwise disposed of
by way of trade in any form of binding or cover other than that in which
it is published, without the prior consent of the publishers.

Text set in 10/12 pt Monotype Times, printed and bound

in Great Britain at The Pitman Press, Bath

ISBN 0 273 01681 4

Preface

This book has been written as an easy introduction to phonetics, the study
of pronunciation. It is intended for those whose mother tongue is English—
and particularly for those who may wrongly have imagined phonetics to be
a dull, dry, or difficult subject. It aims to hold the student’s interest, while
leading him steadily onward. Plenty of practical examples and exercises are
provided, although teachers may of course wish to supplement these with
further material.
The book covers the basic elements of general phonetics and English
phonetics. A student who has mastered its contents will be well on the way
to success in the International Phonetic Association’s Certificate examination
or the phonetics paper of such examinations as that for the Licentiate
Diploma of the College of Speech Therapists. We hope the book will prove
of value to many students of English language, speech therapy, linguistics,
speech and drama, and allied subjects.

v
Acknowledgements

The authors’ thanks are due for permission to reproduce the following
passages:

Page 33. From Three Men in a Boat by Jerome K. Jerome, published by

J. M. Dent & Sons Ltd.
Pages 46 and 88. “Cri Zok,” “Bombast” and “Tan tandinanan” by Bob
Cobbing.
Page 56. “The Rise and Fall of the River Mersey” from A Plea for Mersey by
Peter Molony, published by Gallery Press.
Pages 58-9. From the book The House at Pooh Corner by A. A. Milne,
published by Methuen & Co. Ltd. (Copyright 1928 in the United States and
Canada by E. P. Dutton & Co. Inc. Renewal 1956 by A. A. Milne.)
Page 65. From The Crane Bag by Robert Graves. Published by Cassell &
Co. Ltd.
Pages 69, 90, 96. From the Sunday Express.
Pages, 70, 72, 75, 79, 102. From the Observer.
Page 74. From Under Milk Wood by Dylan Thomas, by permission of
J. M. Dent & Sons Ltd. and the Trustees for the copyright of the late Dylan
Thomas. (Copyright in the United States 1954 by New Directions Publishing
Corporation.)
The chart on pages viii-ix is reproduced by permission of the International
Phonetic Association.

vi
Contents

Preface v
Acknowledgements v*
The International Phonetic Alphabet viii
1 Air-streams
2 Transcription 3
3 The Vocal Folds 8
4 Place of Articulation 12
5 More Transcription 16
6 Transcribing Connected Speech 21
7 Manner of Articulation 25
8 The Soft Palate 28
9 The Lips 31
10 Marginal Sounds of English 34
11 Voicing 38
12 Voicing (continued); Intonation 43
13 Vowel Sounds 47
14 Assimilation 33
15 Elision 57
16 Plosive Theory 60
17 Lateral Approach and Release 66
18 Overlapping Plosive Consonants 72
19 Aspiration; 1-Sounds 76
20 Phonemes 81
21 More about Articulation and Intonation 87
22 English /r/ 91
23 More about Vowels 97
24 Other Sound-types 103

Appendix 1 Some English Phonemes and their Allophones 108

Appendix 2 Books for Further Reading 112

Index
Index of Phonetic Symbols 113

vii
 THE INTERNATIONAL PHONETIC ALPHABET.
(Revised to 1851.)

Labio¬ Dental and Alveolo-

Bi-labial Retroflex Palatal Velar Uvular Pharyngal Glottal

o a

^ 'a
II
dental Alveolar palalal

rQ
o
M

P<
Oh

Plosive . t d

*-*%
a*

CD

•XJ*

rt

a
Nasal
z

S’
G3
CT*
Lateral Fricative.

rCp
-
Vj
Lateral Non-fricative .

u
Rolled
C2

&INVNOSNOD
Flapped .
to

CO
N

«CH
>
<x>
XO
A x
A

Fricative . .
to
4G

CSk
OH

co,
O’

cn>
to
X

Frictionless Continuants

&
to

&*
and Semi vowels
^
r*>
°

g
TS 33
2 *
■3 p
«j g

CTose ....

S
1*
<D

Half-dose .
°

o'
3
8
8
o
<

Half-open .

STJMOA
£8
C4

Open ....
vs
o

^
-5*
(Secondary articulations are shown by symbols in brackets.)
1 Air-streams

Phonetics is the study and description of pronunciation. It is concerned with

what we pronounce and how we pronounce it. In order to learn how we
make speech sounds, we shall study the workings of the various organs of
speech. In order to get a better idea of what we pronounce (which may be
by no means what we think we pronounce), we shall make use of special
phonetic symbols, which enable us to refer to sounds independently of
spelling.

For the practical work in this course, a teacher is needed. The

teacher should give dictation for ear-training purposes and check
the student’s production of unfamiliar sounds and sound sequences.

For us to make any sound with our organs of speech, air has to be set in
motion. If no air moves, no sound results. In ordinary speech, air is set in
motion from the lungs. It then passes through the throat and the mouth
and/or nose, and so into the outer air. Breath enters and leaves the body in
accordance with the alternate increase and decrease in the size of the chest,
and hence in the size of the lungs (since the lungs expand and contract
according to the size of the chest). These changes in the size of the chest are
brought about by alternate contraction and relaxation of the muscles. The
size of the chest is increased laterally by the contraction of the external
intercostal muscles and vertically by the contraction of the diaphragm.
After the air-stream has left the lungs it may be modified in various ways
before passing out through the mouth or nose.

Make a good long f sound. Feel the movement of the air involved.

The sound f is articulated by means of contact between the lower lip and
the upper teeth. Distinguish between this articulation (the nature of the
barrier to the air-stream) and the air-stream mechanism (the way the air-
stream is set in motion—also called the initiation).

A sort of f sound can also be made while sucking breath into the
lungs. Try this.
Air stream

The air-stream mechanism in ordinary speech is termed pulmonic egres-

sive. It is pulmonic, because the air is set in motion in the lungs; it is egressive,
because the direction of the movement is outwards. But a pulmonic air-
stream can move in either direction, outwards or inwards. Inflowing, it is
called ingressive; outflowing, it is egressive. Although ingressive sounds are
easy to make, they rarely occur in speech.

Make s first egressively, then ingressively. Alternate them, first one

then the other. Do the same for a (i.e. an aA-sound), for t and for b.
Try talking with an ingressive air-stream—try phrases like Good
morning, What do you think? Would you excuse me, please? Try
counting, from one to ten and further, as far as you can get on one
ingressive breath. Compare counting on an egressive breath.

We can only talk for very short stretches with an ingressive air-stream.
With an ordinary egressive air-stream, on the other hand, we can talk for
twenty seconds or more in one breath. This is because we have had much
more practice at fine control of the egressive breath.
An ingressive air-stream is sometimes used to make speech sounds. Yes
may be said ingressively sometimes; there is also an ingressive ejaculation
rather like sh! used as a response to pain.
The air-streams we have dealt with so far have been pulmonic. A familiar
but non-pulmonic air-stream is that used to make clicks. These are sounds
made entirely within the mouth, not using the lungs. Their air-stream
mechanism is oral.

Make the sound usually written as tut-tut. (Its phonetic symbol is j.)
Can you make it between vowel sounds ?

The air flows inwards in the tut-tut click, so we say it has an oral ingres¬
sive air-stream. The air is sucked into the mouth by the muscular action of
the tongue against the roof of the mouth—the tongue acts rather like a
rubber sucker pulling against a ceiling.

If the tongue is released at the side instead of at the tip we get

another click, the gee-up! click used to encourage horses (phonetic
symbol 5). It too has an oral ingressive air-stream.

It is also possible to make sounds with an oral egressive air-stream: that

is, an air-stream initiated solely inside the mouth, not using the lungs or
throat—but flowing outwards.

Experiment with oral egressive air-streams. One sound of this type

is associated with removing food from between the teeth. Another
is known colloquially as a “raspberry.”
2
 Air-streams

It is possible to make a pulmonic egressive air-stream sound, namely an

ng-sound as in sing (phonetic symbol g) at the same time as any click. The
result is called a nasalized click.

Try the nasalized clicks corresponding to the j and s already

practised. The nasalized j is sometimes used in English as a kind of
interjection to show contempt or disparagement.
Can you make a nasalized “raspberry”?

In fact all sounds made with the oral air-stream mechanism require a
closure between the back of the tongue and the soft palate, so clicks can be
made in the middle of a g or k as well as an ij.
A third pair of air-stream mechanisms involve holding the vocal folds
together and using the air above them in the throat (pharynx). If a pharyn¬
geal air-stream made like this is egressive, the result is an ejective consonant.
The air is compressed behind its articulation by the raising of the larynx,
which acts like a pump.

Take a breath and hold it. While still holding the breath, articulate |
p, t, k, or any other consonant.

Some people use ejectives in English when words ending in p, t, k, or tj

(a c/z-sound) come at the ends of sentences.

Say the following words with ejective consonants at the end: pick,
plate, stamp, bench.

Sounds made with an ingressive pharyngeal air-stream are called implosive.

The necessary suction is made by holding the vocal folds closed and lowering
the larynx sharply.

Try making implosive p, t, k, tj, etc.

If, during an implosive, a little lung air trickles through the vocal folds
with a “straining” noise, the result is a voiced implosive. Sounds of this type
occur in several African languages.

Try and make implosive sounds corresponding to b, d, g (phonetic

symbols: 6, cf, g-).

We have now covered six different air-stream mechanisms. We can tabulate

them as follows—
3
Air-streams

(egressive ordinary speech

PULMONIC
(ingressive in-breathing speech

(egressive ejective
PHARYNGEAL
(ingressive implosive

(egressive reverse click

ORAL
(ingressive click

Note: some phoneticians use the terms glottalic and velaric instead of
pharyngeal and oral respectively.

It is also possible to use a belch (burp) as an air-stream for speech. People

who have had their larynxes removed for surgical reasons learn to do this.
Since the air involved is expelled from the oesophagus, we call this the
oesophageal egressive air-mechanism.

If you care to, try talking with an oesophageal air-stream. To begin

with, limit yourself to one syllable per burp.
There is no need to worry if you find it difficult to produce sounds
made with non-pulmonic air-streams

4
2 Transcription

By systematic practice in listening to sounds one can greatly improve one’s

ability to recognize and distinguish different speech sounds. Such practice
is known as ear-training. It is best done by taking dictation from someone
familiar with the phonetic material to be studied.
Another form of practice very useful in increasing one’s phonetic awareness
is making transcriptions—writing down in phonetic symbols the way a
passage of English would be pronounced. Doing this, one has to concentrate
on sounds and free one’s ideas from the influence (often misleading) of
ordinary spelling.
For both these purposes we represent sounds by phonetic symbols, some
of which are ordinary letters used in a consistent and specified way, and
some of which are new letter shapes. They help us to note down and refer
to sounds independently of spelling. We can begin with five vowel-sounds
and three consonant-sounds, as follows—

SYMBOLS KEYWORD
i beat
a father
u moon
3 bought
3 bird
P pip
t raughr
k cook

We can use these symbols to transcribe various English words—

pit peat
kak cork
tu too, two

Looking at the last example, note how English spelling obscures the fact
that two and too are pronounced identically. (Sometimes to is pronounced
tu as well.)

Transcribe the following words—

tea, paw, purr, park, coo
5
Transcription
Write down the following words (or similar ones) from dictation.
Put down both ordinary spelling and phonetic symbols—

key ki
tar, ta ta
curt k3t
talk tak
peat pit

Obviously, not everyone has exactly the same pronunciation of English.

For some accents we should need to use a partly different set of symbols, or
use them in different ways. The pronunciation used in this book is an educated
pronunciation of England—a kind which does not have any regional charac¬
teristics. But the transcription used for this southern British standard
(also called received pronunciation or RP) will serve with very little
modification for any educated accent of England and Wales, Australia, New
Zealand, or South Africa. For the English heard in America, Scotland,
Ireland, and the West Indies we might need a rather different type of
transcription.
Notice that in Southern British Standard pronunciation there is no r-sound
corresponding to the r in words such as park, curt, cork. Accordingly we
transcribe them pak, k3t, kak.

Say these words over to yourself, noticing whether you actually

pronounce an r-sound. If you don’t, try them for contrast with an
r, the way they are pronounced in the West Country, Scotland,
Ireland, or America—
park k3rt kork
If on the other hand you usually pronounce these words with an
r-sound, try them without—
pak k3t kak
The following nonsense words (or similar ones) should be
written down in phonetic symbols only, from dictation. They give
practice in recognizing sounds when no help can be had from the
meaning—
k3p tap t3t put
kik pap kit tuk

EXERCISES
1. Say aloud the words represented by the following phonetic tran¬
scription—
pak t3k kup kat it
ak pat tik kat pat

6
 Transcription

2. Transcribe the following words—

pa, paw, peak, kirk, Pooh, purr, talk, pert, ought, car.

3. How many differently spelt words are pronounced kat? pa? pa?
In how many different ways can the sound k be represented in
ordinary spelling? The sound i? The sound a?

7
3 The Vocal Folds

EAR-TRAINING
p3tuk
topit
kakakt
pikput

The vocal folds (or vocal cords, or vocal lips) lie in the larynx, behind the
adam’s apple. They are like elastic lips, running fore-and-aft across the
windpipe. If they come firmly together, they block the passage of air: this
is what we call “holding the breath,” and is how we make the sound known
as the glottal plosive or glottal stop. If they lie open and apart, they allow
the outgoing air a free passage. But one of their most important functions
is producing voice.

Say fff (a long f-sound), and then m (a long v-sound). Feel the
difference between them as you repeat them and alternate them.
These two sounds, f and v, have the same air-stream mechanism
and the same articulation. The difference between them lies in the
voicing of v.

Voicing is due to the rapid vibration of the vocal folds: that is, they open
and shut repeatedly. For a man’s middle note the vibration may be 130 to
140 times a second; for higher notes the rate of vibration is greater, and
for lower notes less.

Experiment with further pairs of voiceless and voiced sounds. Say

sss, then zzz. Say an m-sound (which is voiced), and then try its
voiceless counterpart ipijup.
Similarly say aaa and then voiceless 999. Alternate quickly
fvfvfv, szszsz, mipmipmip, 090909.
There are several ways of testing whether a sound is voiced or
voiceless.
1. Place the thumb and forefinger gently on the outside of the
larynx and feel the vibration present during voiced sounds.
2. Put your hands over your ears and feel the loud buzzing which
fills the head during voiced sounds.
8
 The Vocal Folds

Other ways involve the use of apparatus (stethoscope, oscilloscope,

speech spectrograph).
Decide whether the following sounds are voiced or voiceless:
h w n.
Practise control over voicing, until you can switch it on or off at
will.
Say n 9 n p. (The symbol»underneath a letter shows voicelessness.)
Then try 5 on its own and between vowel-sounds.
Try mipimjnmp and ljljlj. Try devoicing any other voiced sounds
you can make—e.g. a rolled r-sound, uvular or lingual. (If you can
make a uvular roll and devoice it, try whistling at the same time—
like a referee’s whistle.)
In doing these exercises you are extending the range of sounds
you can make.

The glottis is the space between the vocal folds. The term “voiced” and
the term “voiceless” thus refer to different states of the glottis. If the glottis
is vibrating, a voiced sound results: if the glottis is wide open, a voiceless
sound results. If the glottis is open but not very wide open, the air can only
pass through with a certain amount of difficulty, and whisper is heard.

Make the voiced sound a and then the corresponding whispered and
voiceless sounds.

By closing the glottis we hold the breath and make a glottal plosive
(phonetic symbol ?).

Alternate vowel-sound and glottal plosive, thus a?a?a?.

We also close the glottis for other, physiological, reasons—when lifting

something heavy, for example.

Try this way of testing whether the glottis is open or closed. 1. Make
an h-sound, and meanwhile flick the neck just above the larynx with
the forefinger. A dullish sound is heard. 2. Do the same thing while
holding your breath, i.e. making a glottal plosive. A louder and
more hollow sound is now produced by flicking, and its pitch can
be varied by changing the position of the tongue, rounding the lips,
etc.

Note three new phonetic symbols for vowel-sounds—

1 bit
e bet
ae bat
9
The Vocal Folds

And some symbols for consonants, all familiar—

b baby
d dandy
9 giggte
f Nty
V vivid
s cease (note that the sound is the same,
though the spelling is different)
z zones (the same remark applies)
r red
lead
witch
b /zot/iouse
m murmur
n nanny

Transcribe, or write in phonetic symbols from the teacher’s dictation,

the following words—
bee, gawp, do, bird, guard, kit, tan, ten, knit, ban, fend, web.

What words are represented by the following transcriptions ?

bag men gas pis fit kas van f3st Inn wot huf kaf
rad V3b vest

Now try some two-syllable words—

mimt p3imt kapit aneks bradkast

Where the spelling is c the sound may be either k or s. Transcribe

the following, being careful never to write c, only k or s as appro¬
priate.
cat, call, calm, cent, cigar, panic, pencil, access.
Note how misleading the ordinary spelling may be: seas, sees and
seize are all pronounced siz.

Words pronounced the same but spelt differently are called homophones.
So troupe and troop are homophones, as are meat and meet. The phonetic
transcriptions of a word and its homophone are identical: trap, mit. (Such
pairs can also be termed, less specifically, homonyms.)

How are sent, scent, and cent transcribed ?

Different ways of pronouncing English lead to the existence of different

sets of homophones. In Southern British Standard carve and calve are
10
 The Vocal Folds

homophones, since both are pronounced kav; in other accents the two words
may be distinguished. Similarly cause and cores, SBS kaz: these are distin¬
guished not only by r-pronouncers but also by many Londoners.

Test the following pairs by saying them aloud. Some are homo¬
phones for everybody, some only in certain accents.

feet—feat kernel—colonel
ant—aunt tear (rip)—tear (weeping)
pair—pear—pare faint—feint
rows—roes—rose sure—shore
plaice—place slow—sloe
hare—hair which—witch

Transcribe from dictation—

ist end—izd end

let mi—led mi
haet pin—haed bin
stavd—staff

11
4 Place of Articulation

EAR-TRAINING
Transcribe from dictation—
big ben pik ten
brad bm sit bait
tu wiks liv fa waeks livz

We must now consider not so much how a sound is articulated as where

it is articulated. In all sounds we shall be dealing with here, the air-stream
passes from the lungs through the trachea (windpipe) and larynx into the
mouth and/or nose.

Study the diagram of the organs of speech (page 13).

To make p, the lips are pressed together so as to prevent the air-stream

from escaping from the mouth. (As we shall see later, the air cannot pass
out through the nose, because the soft palate is raised.) The air comes into
the mouth but is blocked behind the bilabial barrier, b and m are also
bilabial sounds, since they too are articulated with a barrier formed by the
two closed lips.

Say p, b, and m several times each, getting the feeling of their

articulation. For variation, try making them in an f position—with
the barrier formed by the lower lip and the upper teeth instead. Say
happy, baby, Mummy while laughing!

f and v are made with the lower lip and upper teeth and are accordingly
termed labio-dental.

You can also make rather similar sounds to p, b, etc., by placing

the tongue against the upper lip. Try saying happy in this way. But
as far as we know no language uses these linguo-labial sounds.

We have two sounds in English made by placing the tongue tip against
the backs of the upper teeth (sometimes by putting it further forward,
between the teeth). They are the th-sounds: the voiceless one, 0, as in thick,
12
 Place of Articulation

Diagrammatic view of a section through nose, mouth and neck, showing the
organs of speech

thought, thirteenth, and the voiced one, 6, as in this, though, other. They are
termed dental.

Read the words transcribed as follows: 0at, 6set, Sen, ha0. Tran¬
scribe: wreath, thirst, these, earth. How do different members of the
class pronounce with ?

t, d, n, I, s, and z are made by placing the tip or blade of the tongue against
the gum just behind the upper teeth. This ridge of gum is called the alveolar
ridge, and the sounds articulated there are termed alveolar.
13
Place of Articulation

Place the tip of the tongue against the backs of the upper teeth.
Then draw it back slightly until you can feel the alveolar ridge. Say
tar, den, seize, loot, checking whether the consonants are alveolar.

In some languages t, d, and n are made not as alveolars but as dentals.

This is true of French, and also of certain accents of English.

Try saying ten with dental t and n instead of alveolar. Can you hear
the difference between dental and alveolar when dictated ?

Even in Southern British Standard t, d, n, and 1 are usually dental before

0 and d.

Say wealth, width, eighth, tenth, add this. Are your t, d, n, I dental ?
If you have access to native speakers of other languages, find out
whether they use dental or alveolar sounds (or both).

k and g are made by raising the back of the tongue to touch the soft
palate (also called the velum). Their place of articulation is velar.

Make k and g alone and in words such as car, guess, dog, baker,
bigger. Try and feel the velar articulation.

The ng-sound of sing is also velar. Its symbol is g.

Make g and feel how it is made in the same place as k and g.

Words with g include rung, hang, banger, longing. Note that g is also used
in words such as think 6igk, bank baegk, anger segga.

Why is it wrong to transcribe thank as 0aenk? What should it be?

In some accents of the Midlands and North, g never occurs by itself but
only with a following k or g. So singing may be said as siggigg rather than
the standard sigig.

Try out those two ways of saying singing. Other possibilities are
sigin, siggin. Find out which each member of the class commonly
uses.
Transcribe your pronunciation of the word dinghy (a small boat).
Some say it with gg after the first vowel, others with just g. Similar
variations occur with English and language: poll the class or your
friends.
14
 Place of Articulation

In the word morning manir), the m is bilabial, the n alveolar, and the g
velar.

Take the words packet paekit, topic tDpik, lacking laekirj, band baend.
Write down the place of articulation of each sound.

Here are three new symbols for vowel sounds—

a cup,love
u good, pwt
o hot (this symbol is a upside down).

Transcribe: look, suck, dog, guard, vast, knot, could, cud, cod, card,
cad.
Read: mak, mDk, mAk, bok, maed, nud, W3d, wad, stok, stak,
staek, stak.

Transcribe the words wasp, clasp, asp.

In Southern British Standard wasp is wosp, clasp is klasp and asp is aesp.
Notice how the same spelling with a corresponds to three different vowel-
sounds. In the Midlands and North of England, however, clasp may rhyme
with asp.
There are many words where social or personal differences of pronuncia¬
tion are found. Thus plastic may be plaestik or plastik—both pronunciations
are acceptable in Southern British Standard. So also drastic as draestik or
drastik, and elastic as ilaestik or ilastik.

Conduct a poll along the class or your friends to find out which
pronunciation is commoner. What about gymnasticl And spastic?
And is graph more commonly graef or graf ?

Often such differences are regional. So past is past in SBS and in the
South of England, but paest in the North (and in America). The number
one, usually WAn, is often won in the Midlands and North.

15
5 More Transcription

REVISION EAR-TRAINING
bigmig
fokshAnt
Agkmnan
kukbuk
kAm on Ap
bi Apstaendig and dngk
onwad and Apwad

Refer now to the table of phonetic symbols for transcribing English

(page 17). We have already dealt with most of the consonant symbols; note
the following new ones—

J sheep
3 pleasure, vision
tj church
d3 judge
j yet

Be careful over the last two—the phonetic symbol j has a different meaning
from the j of English spelling.

Transcribe—
sheet, ash, rouge, chest, catch, jet, edge, charge, young, you.

But the sound j is not always spelt y. Study the following—

few fju
cute kjut
tune tjun
new nju
amuse amjuz

Compare moo mu with mew mju.

Transcribe the following—
due, duke, rebuke, argue, accuse, loose, rude, view, blue, shoe.
16
 More Transcription

PHONETIC SYMBOLS FOR TRANSCRIBING ENGLISH

•
1 l bead /bid/ P pip /PIP/
2 i bid /bid/ b barb /bob/
3 E bed /bed/ t tact, tacked /takt/
4 X bad /bad/ d dared /dead/
5 a card /kod/ k cake /keik/
6 0 cod /kDd/ 9 gargle Imil
7 0 caught, court /kot/ tj church /tM/
8 u good /g«d/ d3 judge /d3Ad3/
9 u food /ftid/ f fearful /fiaf}/
10 A bud /bAd/ V vivid /vmd/
11 3 bird /bad/ 0 thirteenth /03tin0/
announcer /anaunsa/ 5 the brother /6a brA 6a/
12 9 laboured /leibad/ s cease /sis/
vanilla /vanila/ z zones /zaunz/
13 ei may /mei / J sheepish /JipiJ/
14 3U mow /mau/ 3 treasure /trs3a/
15 ai high /hai/ h hothouse /hDthaus/
16 au how /hau/ m murmur /m3ma/
17 oi boy /boi/ n nanny Inxm/
18 13 rear /n3/ 0 fingering /figgani)/
19 89 rare /rsa/ 1 lull /w/
20 U3 Ruhr /rua/ r reared /nad/
—
1 battle /bat}/ j use, yews Iwl
— ni fatten /fatn/ w witch /witj/

Note: Some people prefer to use the letter v rather than i and o rather
than u. This affects the symbolization of vowels 2, 8, 13-18, and 20.
The symbol ? (glottal stop) should not be used in ordinary (phonemic)
transcription of English.
The following extra symbols which may be needed are discussed in
Chapter 6—

Da pores /pDaz/ (if different from pause or /?uw^)

aa fire /fa a/
aa power /paa/
hw whine /hwam/ (if different from w/ne)

17
More Transcription

Conduct a poll for suit—do more people say sjut or sut? Does it
depend on where they come from? Do the same for lewd, ljud or
lud, and assume, osjum or asum.
Where do people come from who say nu for newl List other
words where such people have u against the SBS ju.

We have already dealt with the vowels numbered 1 to 11. Number 12, a,
is very common in unstressed syllables and can be spelt in a large number
of different ways.

Read—

beta salad agri biga kalskt kamandmant

Note how the prefix con- is usually pronounced kon when stressed (kDnsat,
kDntxkt), but kan when unstressed (kansida, kantinju). This does not apply
in Northern accents, where it may always be kon—observe in this connection
that the highest-prestige pronunciation is not always the one that follows
the spelling most closely.

The prefix ex- may be iks, eks, or aks when unstressed as in except,
explain, extend. Which pronunciation is commonest? Are all three
SBS, or is one of them regional ? Are except and accept homonyms ?

Distinguish carefully between number 11, 3, and number 12, a. Both

occur in the words further f3da, occur ak3 (rarely auk3, Dk3), surplus S3plas,
murderers m3daraz. The word commerce is usually pronounced kmn3s, so
that it does not rhyme with Thomas tomas.

Transcribe—
urban, murmurs, Persia, versus, disturb, purpose, astir, refer, sermon.

It may be more difficult for some students to grasp the difference between
12, a, and number 10, a. Indeed, in transcribing Midlands, Northern, or
many American accents we should not be justified in distinguishing them at
all. But speakers of SBS and those from the South usually make quite a clear
difference between the two vowel sounds. Study the following examples—

summer SAma
other Ada
thorough 0AT3

above abAV

The word income may have either a or a in its second syllable, and these
are alternative pronunciations on a par with drastik versus drsestik and the
18
 More Transcription

like. Compare also humdrum lumdrAin versus conundrum kanAndram, unending

AncndiT) versus an ending an endnj.

Transcribe—
butter, supper, another, coloured, among, brothers, suggest, collect,
succeed, pick-up.

The closing diphthongs, numbers 13 to 17, should offer no difficulty.

The two letters making the symbol in each case suggest, but do not precisely
specify, the beginning and end of the diphthongal glide used. So the starting-
point of an, for example, may be just like a; but it may also sound rather
different, depending on regional, social, and personal factors.

Transcribe, or write from dictation—

cry, crow, die, day, dough, now, know, nigh, neigh, toy, late, light,
lout, voice, vice, mouse, mace, coat, kite, bake.

The centring diphthongs, numbers 18 to 20, may present more difficulty.

In words like beard, fear, speakers of SBS use diphthong number 18, ia,
thus biad, fia. In other accents of English, however, the sequences ia, ir, or
nr may be used. Note the following for SBS—

nearer mara
hero hiarau
really riali
idea aidia

Does nearly rhyme with really ? Poll several people, and draw con¬
clusions about how their pronunciation of really and nearly should
be transcribed.
Many Americans rhyme mirror and nearer. How does this come
about? Would you expect such speakers to distinguish in pronuncia¬
tion between serious and Sirius ?
Transcribe the following in SBS—
beer, clearer, dearest, Vera, fearing.

Somewhat similar considerations apply to sa. Note that ca and 3 are not
distinguished in the west of the North of England (e.g. Liverpool). So stir
(SBS st3) and stair, stare (SBS stsa) are homonyms for some people.
19
More Transcription

Note the following—

fair fea
sharing Jeariij
Mary mean
Sarah seara

Transcribe—
mare, mayor, glare, bearing, heiress, squarest, dairy, fairing, bareback,
millionaire.

Number 20, ua, is difficult to find a satisfactory English keyword for.

Some people use it in tour tua; but others pronounce this word like tore, ta
or tDa. Some use it in some or all of sure, poor, moor, cure, during (Jua, pua,
mua, kjua, djuarnj); but others pronounce these as homophones of shore,
pore, more, or rhymes of four, boring, with a or Da. Then again, many people
have ua in words such as fluent fluant, steward stjuad; rather more frequent,
though, are the pronunciations fluant, stjuad.
In fact sure, shore, and Shaw may all be homonyms (all being pronounced
Ja); but they may equally well all be distinct (being Jua, Joa, Ja respectively).
Likewise moor, more, maw.

Test these out among the class. What about poor, pour, pore, paw—
are any of them homonyms ? All of them ?
Investigate and transcribe your own and other people’s pronuncia¬
tion of the following words—
jury, curious, manure, Europe, brewer, lure, bureau, secure, influence,
valuer.
Note that most of these have j before the possible ua.
Transcribe from dictation—
it mei W3k out.
WDts 5a deit?
du J\t 5a da!
hav ju got da ki dea?
da pausts leit: its infjuarieitn)!
lets tua da taun.

20
6 Transcribing Connected
Speech

EAR-TRAINING
Transcribe from dictation—
kwik matj. left, rait.
derndja, raud W3ks ahed.

English is a language in which one distinguishes stressed and unstressed

syllables. Stressed syllables are symbolized by the mark ' placed before the
syllable concerned; unstressed syllables are left unmarked. Thus 'msait
insight, in'sait incite.

Say insight and incite over several times to grasp the idea of stressed
and unstressed syllables.

All words when said in isolation, and most words when in a connected
passage, have at least one stressed syllable. Study the following examples—

'said? cider
a'lau allow
pa'teitau potato
'm£33rn) measuring
'kjumjulativ cumulative

Decide which is the stressed syllable in each of the following words—

collide, bacon, calendar, judgment, another, average, enjoy, bigger,
happily, particularly.

Here is a connected text in transcription, with the stresses marked—

ail 'gau an irj'kwaia wsda '6is iz 6a 'wei ta 61 'empaia '0iata.

I'll go and inquire whether this is the way to the Empire Theatre.

Say the sentence over, banging out the stress on a desk or chair.

Stress results from a combination of several factors, the most important

being extra loudness, extra duration, and change in the pitch of the voice.
The recurrence of stresse dsyllables at regular intervals gives speech its
rhythmical qualities.
21
 Transcribing Connected Speech

Note the tendency of “function words” (fully grammatical items, such as

pronouns, conjunctions, articles, auxiliary verbs, and prepositions) to be
unstressed in connected speech. Notice also how many of them are pro¬
nounced in connected speech with a vowel (usually a or i) different from
that used when the word is used in isolation. Thus the by itself is 3i; in
connected speech it is usually 3a before consonants and often too 3i before
vowels. 3a 'bm, 3a ’g3l, 3i 'end. (Some people, though, never use the form
3i.)

Find out how you pronounce the in each of the following phrases—
the chair, the answer, the wheels, the year, the helmet.

8a and 3i are known as the weak forms of the\ 3i is the strong form.
Similarly, at has a weak form at (at 'kntj at lunch) and a strong form aet.
Discover and transcribe the weak and strong forms of the following
words—

of, was, can, she, them, an, that, for, but, are.

You are now equipped to transcribe (or write from dictation) a connected
English text. Pay particular attention to vowels in unstressed syllables
(including weak forms): mostly they are a or i, whatever the spelling may
suggest. Mark all stresses.

EAR-TRAINING
'wots 3a taim?
'du get a 'muv on.
aim 'gaunj 'out far a 'mmit.
'kip 'Df da 'gras,
a 'stitj m 'taim 'servz 'nam.
3a 'pen iz 'maitia 3an 3a 'sad.

There are one or two more points about transcription of English to be

explained. The symbols j and n denote syllabic consonants—1 and n
sounds which are rather longer than usual and have syllable-making function
like vowels. Compare 'kodlnj codling (small cod) with 'knd}n) coddling (from
the verb coddle). The word gluttony may be 'gUtam or 'gktpi; the latter
pronunciation contains no vowel between the t and the n, yet does not
rhyme with chutney 'tjAtm or Putney 'pAtm.
Study the following examples of words with syllabic consonants—
'gadn garden (childish pronunciation 'gadan)
'mid| middle (“improved Cockney” pronunciation 'midal)
'bitn bitten
'bDtj bottle
22
 Transcribing Connected Speech

There is sometimes a choice between syllabic and non-syllabic consonants:

threatening, for example, can be pronounced either 'GrEtmg (three syllables)
or 'GrEtmg (two syllables).

Transcribe the following—

bundle, hidden, burden, sandals, cotton, cattle, curdled, maddening,
bottling, handled.

Write from dictation—

'niDtJ 'padn 'saednd 'mcdlsam maen’haetn 'g3dl Totn 'dwmdt
'potl 'switn 'haitn 'pudl 'tjatl

In many words pronunciation fluctuates between n and an, J and 1: ssvn

or 'sevan seven, ’spejl or 'spejal special.
Occasionally other consonants are syllabic, notably m g r as in 'aupm (or
'aupn or ’aupan) open, 'steik g 'kidm (or other possibilities) steak and kidney,
’neibrig (or 'neibarig or ’neibng) neighbouring. But do not attach the syllabic
mark to any consonant symbols other than m n g 1 and r.
The symbols na, aa, and aa could be dispensed with altogether for some
people’s speech. For others, though, it is useful to have them available.
Most people do not distinguish saw from sore, board from bored, or cause
from cores. They should transcribe the vowel simply as a, thus sa bad kaz.
Others make a difference, having an opener and/or diphthongal sound in the
second word of each pair: we shall transcribe it na and number it la. They
should write sa saw but sna sore, bad board but bnad bored, kaz cause but
koaz cores. (Other people again use Da where there is an r in the spelling,
but a where there isn’t.)

Investigate which members of the class use Da. Can anything be

observed about their regional background ?

Now consider words such as fire, tyre, science, pious, liar. Is the vowel
part of these words best analysed as ai followed by a? For some people it is.
For others, though, it sounds rather different, being collapsed into a simpler
diphthong (or even monophthong) which we shall write as aa.

Decide which transcription is appropriate for yourself and other

members of the class—
fire 'faia or 'faa
hiring 'haiarig or 'haarig
quiet 'kwaiat or 'kwaat

Words like hour, power, towering are rather similar. They may have au
followed by a; or they may have a reduced diphthong aa; or they may be
23
Transcribing Connected Speech

pronounced just like are, par, tarring, in which case of course we simply
transcribe a.

Similarly investigate power 'paua or 'paa or 'pa, flowering 'flauarnj

or 'flaanij or 'flariij, sour 'saua or 'saa or 'sa.

In words with the spelling wh, most English people pronounce a simple w,
so that whine 'wain is a homophone of wine, and which 'witj is pronounced
identically with witch. Scots, Irish, and many Americans, however, use hw,
thus 'hwain whine, 'hwitj which, and this pronunciation is often considered
abstractly correct in England, even though its use is virtually restricted to
those who have undergone a certain type of speech training.

Discuss the question of w or hw in words such as when, where, why,

whirl, whisk, whistle. What about who ?

Words like happy, valley, coffee end with i in Southern British Standard,
thus 'haepi, 'vaeli, 'kDfi. This is also their pronunciation in many regional
accents; but in the local speech of the South of England, and on Merseyside,
in the North-east and in Wales, they end in i, thus 'haepi, 'vaeli, 'kDfi. So
when making transcriptions of one’s own speech one should write i or i
according to one’s own pronunciation; but when transcribing SBS one
should put i.

Investigate your own and other people’s pronunciation of these

words: happy, committee, donkey, city. Does it make any difference
if a vowel sound follows immediately ?

Similar considerations apply to words such as various 'vEarias or 'vEarias,

carrying 'kaerng or 'kaeriirj, create kn'eit or kri'eit.

Discuss the above words and also maniac, prettier, react.

When writing phonetic symbols, be careful to distinguish between ae and

ae, a and a, i and i, u and u, £ and 3 and finally z and 3.

Revise any symbols you are unsure of. Write from dictation various
alternative pronunciations of often.

24
7 Manner of Articulation

EAR-TRAINING
Transcribe from dictation—-
a 'kwaat 'kana
its 'as 'haos
'midjig t3 'gud
its 3 'weI 'ritn 'bok

To identify consonants there are three principal things we have to know.

The first two are the state of the glottis (voiced or voiceless) and the place
of articulation (bilabial, alveolar, velar, etc.). The third concerns what sort
of articulation is used.
Let us look at alveolar consonants of various kinds,
t and d are termed plosives. They involve the complete blocking of the
air-stream.

Make several d sounds and feel the articulation. Note how the
tongue tip contacts the alveolar ridge, while the side rims of the
tongue complete the closure by contacting the side teeth. The air is
thus completely prevented from escaping out of the mouth for as
long as the plosive is held.

n is called a nasal.

Make several n sounds. Feel how the alveolar articulation is identical

with that for d (and t).

The difference between n and d depends on the soft palate, the valve
that controls the entry of air from the throat (pharynx) into the nose. For
t, d, and other non-nasal sounds the soft palate is in a raised position to
prevent air passing into the nose; but for n and other nasal sounds it is in
a lowered position to allow air through.

Thus in n air continues to escape (via the nose) even while the tongue
blocks any escape through the mouth. Test this by pinching the end
of your nose while saying first n and then d. Note the build-up of
pressure during n but not during d. Why is this ?
25
Manner of Articulation

Make a long m sound, and see if you can feel the soft palate with
your tongue. Curl it back along the roof of the mouth until you
feel the soft flesh at the back: this is your soft palate. Then change
your m into a b: you should feel your palate move sharply up and
away. Make sure you understand what is happening and why.

A rolled r is called a roll or trill. Rolls are rapid series of closures and
openings. So a rolled tongue-tip r is a sort of intermittent d.

Make a rolled r if you can. Compare it with d. Contrast ara, ada.

If you cannot make a tongue-tip roll, try a bilabial or uvular roll.

Both plosives and rolls involve closure; but for rolls it is rapid and repeated,
while for plosives it is longer and single.
Another manner of articulation is lateral. In laterals the closure is only
partial: air escapes round the sides of a blockage. Our only English lateral
is 1.

Make an 1-sound as in leap lip. Try and feel how the sides of the
tongue are down, allowing the air out on either side of the blockage
formed by the tongue tip on the alveolar ridge. You will perhaps
feel this more clearly by making an 1 articulation while breathing in
and out voicelessly.

Yet another, very common, manner of articulation involves not a complete

closure but merely a narrowing. Sounds of this kind are called fricatives:
in English we have alveolar fricatives s and z.

Make long s and z sounds. Feel how the air is hindered and made
turbulent but not completely blocked as it passes between the tongue
and the alveolar ridge. Make an s sound with a pulmonic ingressive
air-stream (breathing in): feel the cold air along the groove in the
tongue beneath the alveolar ridge. Compare this with a pulmonic
ingressive voiceless 1-sound, where the cold air is felt at either side
of the mouth only.

These different manners of articulation apply at other places of articulation,

too. So p and b are bilabial plosives, while m is a bilabial nasal.

What organs make the closure for bilabial plosives and nasals?

The closure of the articulators in plosives makes the breath pressure

increase behind the closure, so that a slight explosion (plosion) is heard
when the articulators part again.
26
 Manner of Articulation

It is also possible to make bilabial fricatives, though they are not used in
English. They sound rather like f and v, but are made by narrowing between
the two lips instead of between the lips and teeth. (Their symbols are <f»,
voiceless, and p, voiced.)

Try and make $ and p. How would you blow out a candle?

f and v themselves are fricatives: labiodental fricatives.

We have already seen that it is possible to make a bilabial roll. It is also
possible to make a bilabial lateral.

To do so, get the lips together in the centre but open at either side.
Then blow, with or without voicing as desired.

At the velar place of articulation, k and g are plosives, while g is a nasal.

Velar fricatives occur in several languages, though not in English. The
voiceless velar fricative (symbol x) is like the ch in loch.

Make voiceless and voiced velar fricatives, x and y.

Sounds such as tj and d3 consist of a plosive followed immediately by a

fricative in the same place of articulation. Such sounds are known as
AFFRICATES.

Say tjtjtjtj and observe the alternation of plosive and fricative

segments.
Make the bilabial affricates p$ and bp; then the alveolar affricates,
ts and dz; then the velar affricates, kx and gy. Put them between
vowel sounds.
Why are sequences such as ps, k0, dv not considered affricates?
Make a table with columns labelled bilabial, alveolar, and
velar, and rows labelled plosive, nasal, lateral, fricative,
affricate. Enter as many as possible of the sounds so far dealt
with.

EAR-TRAINING
The following are examples of nonsense words for dictation. They
may be altered or added to by the teacher as desired.
'mApt 'skb
'dsgz 'tjaup

EXERCISE
Choose a sentence with at least eight different vowel or diphthong
sounds in it and put it in phonetic transcription.
27
8 The Soft Palate

EAR-TRAINING
Transcribe from dictation-
li'ncid ’Jt aem
'nvmod ’3ops
’kmgig ’froz

We have already seen (page 25) that the difference between n and d, or
m and b, lies in whether the soft palate is up (raised) or down (lowered).
When raised, it shuts off the nasal cavity from the mouth and pharynx; when
lowered it allows air to pass through into the nose.

We have little or no direct feeling for the soft palate. Repeat the
exercise mentioned at the top of page 26.
Say ama. Then try to repeat it holding the nostrils. Compare the
lack of effect doing the same while saying aba. (The symbol a
denotes a vowel between ae and a. But in this exercise the precise
quality is not important.)
What articulatory difference is there between g and g ?
Say ama—aba—ana—ada—aga—aga to study the interaction of
place of articulation and soft palate position.

The soft palate, then, affects the manner of articulation by causing what
would otherwise be a plosive to be a nasal instead.
Consider the word amber aemba. At the end of the ae the soft palate comes
down (and the end of the vowel is nasalized). Then the lips close, and the
air-stream is directed through the nose only. This gives m. Then the soft
palate rises, while the lips remain together: this turns the m into a b. Then
the lips part and the b is released; the vowel follows. Note that the m and
the b in this word share the same bilabial closure: as we go from one to the
other the only change is the rising of the soft palate.

Describe the movements of the tongue tip and the soft palate in
pronouncing the word hinder.
Describe the movements of the back of the tongue and the soft
palate in pronouncing the word anger.
28
 The Soft Palate

In a word such as kidney, where a plosive is followed by its corresponding

nasal, an analogous process occurs. In 'kidm, after the first i the tongue tip
comes up to form a contact with the alveolar ridge, while the sides of the
tongue rise to prevent the air escaping laterally: closure for d is thus made.
Then the soft palate comes down, releasing the blocked air through the nose
and turning d into an n. After the n the tongue tip and sides drop again and
the soft palate rises as the final i is pronounced.

Describe the articulation of the bm sequence in submerge.

When someone has a cold he does not, as is often said, talk through his
nose: rather, he does the exact opposite, since the nose is blocked, and
replaces nasal sounds by the corresponding non-nasal sounds.

Test this out with good morning gud'mamg: it becomes gud'badig.

Why?
To gain control over the movements of the soft palate, try saying
dndndndn. Try mispronouncing Hendon, Camden German fashion:
'hendn, ‘kaemdn (usual pronunciation 'hendan, 'kaemdan). Between
d and n in these sequences the alveolar closure is retained while the
soft palate comes down, giving nasal release to the plosive.
Practise non-English sequences mba, nda, gga. Try these words
from the African language Mende (Sierra Leone): ndovo, ggulu,
mbawa.
As b is to m, so p is to voiceless m. Make it; then alternate apa,
ama. Practise the same thing at other places of articulation: ata,
apa; aka, aga.
Say the following English words, releasing the relevant plosives
nasally and avoiding the common glottalizing pronunciation—
chutney 'tjAtm, Clapham ‘klaepin, atmosphere ' aetmasfia, topmost
'tDpmaust, mutton 'nutn, bacon ‘beikg.

EAR-TRAINING
Transcribe from dictation—-
'bmirnp 'ksggng 'mtn 'guki
ka'lnsj 'lDriz 'kaerng 'kreits ag 'kreits av 'bia 'bsekfaiarig am
'bAmpig alDg 'knkjwud 'lem.

In nasalized vowels, such as occur in French, the air-stream passes out

both through the nose and through the mouth, simultaneously. Example:
un bon vin blanc a b5 ve bla.
(The mark ~ over a vowel symbol denotes nasalization.)
29
The Soft Palate
Try the phrase, not worrying too much about the vowel qualities,
but rather concentrating on making sure that the vowels are properly
nasalized. Do not insert any n or g after the vowels.
If you find nasalized vowels difficult to make, proceed as follows.
Say a long m; keeping everything else the same, open the lips. The
result will probably be some kind of nasalized vowel such as s.
Vary the vowel quality to get the vowel desired.
What is the position of the soft palate during a nasalized vowel ?

30
9 The Lips

EAR-TRAINING
Transcribe from dictation—
bmau'hAijk
'krindi'dnav
pa'bumaif

There are some sounds, of course, where the lips are the primary articu¬
lators : for instance p, b, and m, made by the closure of the two lips, and f
and v, made by an articulation of the lower lip against the upper teeth. But
there are also many other sounds where the lips play an important role.
These include certain vowels and other sounds which have lip-rounding.
Positions and movements of the lips are easily observed. The lips are very
flexible, and can vary in several ways, including: (1) different degrees of
lip-rounding, (2) lip protrusion (usually with rounding), (3) lip-spreading,
giving a wide shape. The extent of vertical separation between the lips tends
to depend on the jaw opening, though the lips can also move independently
in this dimension.
Among English sounds usually given some degree of lip-rounding are the
vowels d, o, u, and u, the diphthongs au, au and the semi-vowel w; often
also the consonants r, I, 3, tj, and d3.

Try these sounds, observing your neighbour or using a small mirror

to investigate lip positions. As keywords use those shown on the
Table, page 17.

A word such as moon ‘mun, for example, usually has some lip-rounding,
but not necessarily very much.

See if you can say moon with (a) close lip-rounding and (b) no lip-
rounding. Do the same for good 'god and law 'la. Are these three
vowels the same as far as lip-rounding is concerned ?

Comparing words such as cot 'kDt and caught 'kat, we see that most
English people use considerably more lip action for a than for d. (Scots, on
the other hand, don’t usually distinguish these two words, using the same
vowel in both and making them homonyms.)
31
The Lips
Most people make the diphthong au with lip movement as well as tongue
movement.

Try the word toe 'tau and observe possible lip-rounding at the end
of the diphthong. Try (a) increasing and (b) decreasing the amount
of lip-rounding you use: does this affect the sound ? If so, in what
way?
For practice in control over the lips, recite some passage (such as
Jack and Jill went up the hill. . .) with lip-rounding throughout, and
then with none at all—smile broadly as you say it! Can one com¬
pensate in some other articulatory way for absence of lip action?

The w sound, on the other hand, seems to need definite lip action. Indeed,
the lip-rounding associated with w often extends over other consonants
which immediately precede it.

Say win 'win, twin 'twin, quin 'kwm, swim 'swim. Do the t, k, and s
have any lip-rounding in these sequences ?

Such lip-rounding occurring at the same time as some other more important
articulation is known as labialization. Thus t does not usually have any
lip action; in the word twin 'twm, though, it is labialized, that is, it has
lip-rounding. Labialization is one kind of secondary articulation. (All
labels for secondary articulations end in -ization: others are palatalization,
velarization, etc. In each case the secondary articulation accompanies some
other, more important, primary articulation.)
As mentioned above, many people have considerable lip action for the r
in red Tad. Some people, indeed, have little or no tongue articulation, though
such pronunciation is often considered defective. Note how initial r is often
spelled wr (e.g. write, wrong, wrap), which suggests that this lip-rounding has
a certain historical justification.

Say red and other words beginning with r, using varying degrees of
lip rounding and protrusion. Try it with a weak labiodental sound
(u, discussed in Chapter 22 below) instead of the usual English r.
Then do the same with a word where the r occurs in the middle
rather than at the beginning, such as carry 'kaeri or story 'store
Then say all these words with any other kinds of r-sound you are
familiar with.

In one or two local accents the vowel 3 has lip-rounding—in New

Zealand, for example.

Say bird 'b3d with and without lip-rounding for the 3.

32
 The Lips

EAR-TRAINING
Transcribe from dictation—
aa 'fain} ‘matj 'daun 5e 'midj av 9a 'haistrit 'niAst av bin az
im'pDsabJ a 'spektakj az 'malau ad 'sin fa 'mem a 'log 'dei.
di 'adar av 5a pra'sejn waz az 'fDlauz:
’mmnpma'rensi 'kaerug a 'stik.
'tu dis'repjutabl lukig 'k3z, 'frendz av 'mmitma'rensiz.
'd3ad3 'smaukig a 'Jat 'paip.
'haeris 'kaerug a 'bAld3d 'out 'glaedstan 'baeg im 'wAn 'haend and e
'bDtj av 'laim ’d3us in Si 'Ada.
'grin'graosaz 'bm am 'beikaz ‘bm wi5 ‘baskits.
'buts frm au'tel, kaerug 'haempa.
kan'fekjnaz 'bm wiS 'baskit.
'grausaz 'bm wiS 'baskit.
'tJizmAggaz 'bm wi5 'baskit.
'Dd 'maen 'kaerug 'baeg.
'buzm kam'paenjan av 'Dd 'maen wi5 iz 'haendz in iz 'pokits, 'smaukig
a 'Jat 'klei.
'frutaraz 'bm wi5 'baskit.
mai’self, 'kaerug '0ri 'haets and a 'pear av 'buts.
'siks 'smal 'bmz and 'fa 'strei 'dngz.
'wem wi got ta 5a 'laendig 'steid3, 5a 'bautman 'sed: ‘let mi 'si,
S3, waz jaz a 'stim 'lantj ar a 'hausbaut?
'Dn aar in'famig him it waz a 'dAbl 'skAlig 'skif, i 'simd sa'praizd.

33
10 Marginal Sounds of
English

EAR-TRAINING
Revise the following sounds, not part of the ordinary English sound
system—

? ip 5 J (Chapter 3)
$ P x \ (Chapter 7)

Transcribe from dictation some nonsense words including them,

e-g-
‘auip '$3X 'ya| 'Pnnn 'Ju?x3$

In some kinds of English, e.g. Scots, the sound x is quite common and has
a proper linguistic status (in technical language, constitutes a phoneme of
the dialect concerned—see Chapter 20). But most educated speakers of
English use x, or attempt to do so, in at least a few words, including some
proper names. The composer Bach, for example, is commonly called box—
with an English-type vowel but a German-type final consonant. And everyone
has heard the Scots word loch with its pronunciation Idx.

Think of other words with x.

Does any member of the class use x in words such as technical,
monarch, patriarch, architect ? Say the following surnames and place-
names with x—

MacLachlan ma 'kloxlan
Auchindachie D'xinaxi
Auchtermuchty 'Dxta'nuxti
Bwlch ‘bulx
Amlwch 'aemlux

A similarly marginal sound of English is the nasalized vowel we may

write 5, used mainly in words borrowed from French, e.g. grand prix ’grb
'pri, salon ’sseId, clientele 'klib'tal.

See how members of the class pronounce the words just mentioned.
Do some put in an n sound after the 5 ? Does anyone distinguish in
English between the two French vowel sounds 5 (as in son s5) and
fi (as in cent sfi) ?
34
 Marginal Sounds of English

Investigate the pronunciation of the following words, where some

people use 5 (or something like it) but others do not—

restaurant
baton
entente
pension (in the sense of “boarding-house”)

Transcribe all pronunciations you encounter.

What other words can you think of in which 6 may be used?

SUBSTITUTIONS
It is a useful exercise, particularly for budding speech therapists and speech
specialists, to do a form of ear-training in which English words are mis¬
pronounced in some respect: the student has to identify the wrong sound
which has been used to replace a correct one. So, for example, see might be
dictated as pi instead of si; the student could either write down the phonetic
symbol for the wrong sound, noting also what it has replaced, or could
answer verbally, thus “the voiceless alveolar fricative has been replaced by
a voiceless alveolar nasal.”

Examples of Substitution Ear-training Exercises

For s in see: x, J, hw, p;
for f in rough: <J>, ip, x, 0;
for w in wet: v, p, y.

Note that most sounds can be simply and quickly identified by using a
three-part label referring to their voicing, place of articulation, and manner
of articulation. Thus v, as in river, is a voiced labio-dental fricative;
t, as in tip, is a voiceless alveolar plosive; m, as in come, is a voiced
BILABIAL NASAL.

Work out the appropriate three-part, phonetic label for the following
sounds—
n as in thin
p as in peck
rj as in long
d as in order
g as in baggage
0 as in think

But we have not yet studied all the labels needed to describe the places of
articulation used in English. Note the following additions to what was learnt
in Chapter 4.
35
Marginal Sounds of English

J as in ship is a voiceless palato-alveolar fricative;

3 as in measure is a voiced palato-alveolar fricative;
tj as in chip is a voiceless palato-alveolar affricate;
d3 as in agile is a voiced palato-alveolar affricate;
r as in red is (for most people) a voiced post-alveolar frictionless
CONTINUANT.

Practise saying these names until you know them off by heart. (The
precise articulations involved are described in Chapters 21-2.)

TEST ON CHAPTERS 1-10

It is useful to do this test (or a similar one devised by the teacher) in order
to check up on what you have learnt so far.

A. Here are ten words with s in their spelling. Decide whether the appro¬
priate phonetic transcription in each case is s or z—

missing, base, phase, choose, goose, position, consent, rags, bits, edges.

B. Tick the phonetic term which correctly describes the sound indicated
by italics in the spelling in each of the following—

Example begin voiceless ( ) voiced (vO

1 happy voiceless ( ) voiced ( )

2 busy voiceless ( ) voiced ( )
3 remove voiceless ( ) voiced ( )
4 choose voiceless ( ) voiced ( )
5 reduce voiceless ( ) voiced ( )
6 willow voiceless ( ) voiced ( )
7 crisis voiceless ( ) voiced ( )
8 worry voiceless ( ) voiced ( )
9 skin voiceless ( ) voiced ( )
10 lie voiceless ( ) voiced ( )
11 rough plosive ( ) fricative ( )
12 stab plosive ( ) fricative ( )
13 kiss plosive ( ) fricative ( )
14 bath plosive ( ) fricative ( )
15 vote plosive ( ) fricative ( )
16 rubber bilabial ( ) alveolar ( ) velar ( )
17 make bilabial ( ) alveolar ( ) velar ( )
18 cog bilabial ( ) alveolar ( ) velar ( )
19 stew bilabial ( ) alveolar ( ) velar ( )
20 print bilabial ( ) alveolar ( ) velar ( )

C. Underline a suitable phonetic transcription for each word in the

36
 Marginal Sounds of English

following connected passage. (Sometimes there is more than one possible

correct answer: choose the one that fits the passage best.)

Jim and Mary decided they would move into a new house being tired
■jim an 'mem di'saidid 5ei wud 'mauv mtu a 'nu 'haus bug 'taiad
'djim send 'mcri di'said3d 0ei wuld 'muv mtu ae ‘nju 'haus biii] 'tai3d
'd3im en 'mean di'saidvd thei wud 'muv inta ei 'nju 'haus bin) 'taad

of their flat and wanting more space. They needed a garage for Jim’s
Df 6cr 'flat an ’wantig 'ma 'space. 6ei 'nidid a 'gaerid3 foa 'jnns
av Qca 'flat and 'wantig 'moa 'speis. 0ei 'nidid ei 'gaera3 fa 'd3imz
dv 6a 'flaet n 'wmitig 'mua 'spais. thei 'nided 3 ga'ra3 fa 'd3ims
car, too.
'ca, 'tu.
'kar, 'tau.
'ka, 'tu.

The following nonsense-words and English dictation material should

be taken down from the teacher’s dictation.

'dlepm 'wodnAj '<JmdD? tj£'YDi3 gae'x3j

ta 'plei 'd3aeks, a ta 'grv 6a 'geim its ful 'taitl, 'd3aekstaunz, 6a

'tfildran 'nidid 'faiv 'smal 'round 'pebjz. '6iz 6ei 'tDS, ag 'kaetj am
'pik 'Ap in 'vsarias 'weiz. bat 6ei 'pleid 'd3aek 'stra wi6 '0m 'naerau
'strips av 'wud witj 6ei ’0ru mtu a 'd3Ambld 'hip an 'dsn 'traid ta
ri'muv 6am, 'wvn at a 'taim, wid'aut 'tAtJig 'em av 6i 'Adaz.

The English passage may also be used for practice in reading

aloud from a phonetic text.

Transcribe this passage—

At a meeting of the Committee of Management of the Royal

National Lifeboat Institution for the preservation of life from ship¬
wreck held at their offices, London, on the eleventh day of March 1937
the following minute was ordered to be recorded on the books of the
Institution:
That the Institution gratefully recognizes the services of the St
Mary's Life-boat Station, established in 1837 in the cause of life¬
saving from ship-wreck and on the occasion of the centenary of the
station desires to record appreciation of the voluntary work of the
officers and committee and the devotion and courage of the life-
boatmen of St Mary's who have never failed to maintain the high
traditions of the life-boat service.
Total number of lives saved by the St Mary's life-boat by September
5th 1969: 579.
37
11 Voicing

EAR-TRAINING
'ho? 'pmnt
'Idx 'nss
mmm

The test you have just done may have revealed some weaknesses in the
detection of voicing. If so, revise what was studied in Chapter 3. Remember
that we can observe voicing directly in two ways—

(a) by placing the hand on the throat and feeling the vibrations present
in voiced sounds;
(b) by completely covering the ears and experiencing the buzz of voicing
heard in the head during voiced sounds.

Carry out these procedures for any of the sounds you got wrong in
Question B. 1-10. Remember that for the procedures to work you
must say the word out aloud: whispering it won’t do.

It’s now time to back up the observations we have made by some theory.
Look at the following chart of voiced and voiceless sounds in English.

I. Always voiceless P t k tj
f 0 s I h
II. Usually voiced b d 9 d3 ^
V 6 z 3 all referred to loosely
III. Usually voiced m n 0 as “voiced sounds.”
w j r 1
IV. Always voiced vowels and diphthongs.

(There are occasional exceptions to these generalizations. Be on the look-out

for them and prefer the results of observations to the predictions of theory.)
38
 Voicing

GROUP I
Test out the sounds in Group I.
Think out two words for each of the sounds in the group, and
check whether the sounds are indeed voiceless.

Some people voice t when it comes between vowels (when it is inter¬

vocalic).

Say butter and water in your ordinary pronunciation. Is the t voiced?

Then say the same word(s) imitating an American accent, with a
voiced t.
(If your ordinary t in these words is voiced, try to say them instead
with a voiceless t. What sort of accent does that remind you of?)
Then say the words several times, alternating between a voiceless
t and a voiced one.
Is a voiced t the same as a d ?

Some people voice h intervocalically, too.

Decide whether your h is voiceless or voiced in the words behind,

ahead, rehearse.

GROUP II
Group II, marked “usually voiced,” are in fact always voiced between
vowels. But they can be voiceless at the beginning and end of words.

Check on voicing of Group II sounds in intervocalic position, using

the words abbey, order, again, lodging, cover, either, fuzzy, pleasure.
(It might be helpful to transcribe them first, so as to be sure what
sounds are being investigated.)

A word such as b3d bird, said in isolation, will tend to have its b and d
partially or wholly voiceless. In such cases we speak of these consonants
being devoiced. Note, though, that when the same word is surrounded by
voiced sounds, as in the phrase 5a 'b3d a'laits the bird alights, the consonants
b and d are fully voiced. When the word is surrounded by voiceless sounds,
as in the phrase bat '6is 'b3d 'siijz 'switli but this bird sings sweetly, we again
find devoicing of the initial b and the final d.

Test these statements: are they true for your pronunciation ?

We can summarize by saying that Group II sounds tend to be partially

or completely devoiced whenever they are not between voiced sounds.
39
Voicing
We can show the voicing pattern of a word or phrase diagrammatically,
by drawing a voicing diagram like this—

A straight line denotes no voicing

seat s t (vocal folds not vibrating): a
jagged line denotes voicing (vocal
folds vibrating).

oasis au ei s i s

bird (in isolation) > b 3 d

(in the phrase this bird sings) -\ i/WW -

(in the phrase the bird appears) wvw /wwv WWW

Make similar diagrams for the words face, safe, ticket, assist, faster.
(These should all prove easy.)
Then make voicing diagrams for the word cab. (N.B. The final
sound, b, is a Group II sound, and so subject to devoicing.) Show
it (i) as said in isolation; (ii) in the phrase the cab stopped-, (iii) in
the phrase the cab arrived.
Next make voicing diagrams for the word gap. (Here it is the
initial consonant, g, which is a Group II sound.) Show it as pro¬
nounced (i) in isolation; (ii) when a voiceless sound preceded, as in
the phrase this gap; (iii) when a voiced sound precedes, as in the
phrase the gap.

When two or more Group II sounds occur consecutively, they behave just
like a single Group II sound. So the final consonants gd in the word tagged
'taegd are (partly or completely) devoiced before a pause or a voiceless
sound—

tagged t g d

(The initial t here is aspirated (see Chapter 19)—a fact we ignore for the
time being.)
40
 Voicing

But they are fully voiced if a voiced sound follows, as in the phrase tagged
it: 'taegd it—

tagged it t X g d I t

Similarly, the gb sequence in big boy ’big 'bni is voiced throughout (except
in the unlikely event of the speaker pausing between the two words).

FORTIS AND LENIS

If the difference between Group I sounds and Group II sounds were merely
that Group I were voiceless and Group II voiced, then we should find that
this devoicing of Group II sounds would make them indistinguishable from
the Group I sounds. Yet b3d, even when said in isolation with devoicing of
b and d, does not sound like p3t. Nor does this gap sound like this cab.
Evidently something more than voicing distinguishes p from b, t from d,
f from v, etc.
Group I sounds are more energetically articulated and have stronger
breath force than Group II sounds. On this basis we can divide English
plosives, affricates and fricatives into the fortis (strong) Group I—p, t, k,
tj, f, 0, s, 1, h —and the lenis (weak) Group II—b, d, g, d3, v, 6, z, 3.

Say etch atj and edge ed3 in isolation. Notice the difference in force
of articulation between the final tj (fortis) and the final d3 (lenis).
Do the same with rope, robe; calf, calve; place, plays.

This fortis-lenis difference is reinforced by two further features of the

fortis Group I sounds: aspiration (Chapter 19) and shortening of preceding
vowels (Chapter 23).

PREPARATION
Transcribe the following passage. (Can you transcribe without
referring to the list of symbols and keywords yet?)

The weather today will be warm for the time of year and fine on
the whole. There will be showers here and there though some places
will miss out completely. The good spell should hold over the next two
days but there may be fog over low ground in the early mornings.
That is the end of the general forecast.
41
Voicing

EAR-TRAINING
(i) Nonsense words—

'yitjov 'Buthas ' Psawaem au'pa3

(ii) English—
ta 'nul 'wain, 'kwin vik'tariaz 'resipi.
'bml sm 'spais in a litl 'wata til 6a 'fleiva bi 'gemd. 6sn 1 aed an
'ikwal 'kwnntati av 'gud 'pat 'warn, sm 'Jugar an 'iutmsg. ’bml, an
's3v wi0 'knsp An'switnd 'biskits.

42
12 Voicing (continued);
Intonation

EAR-TRAINING
S'3Df0d
'zrsagkt

GROUP III
Group III, also marked “usually voiced,” are fully voiced initially and
finally as well as intervocalically.

Check this statement by saying the words real, wine, young, male,
lemon. (Note that no English words begin with g or end with w or j.
In SBS no words end in r, either, when said in isolation.)

However, most sounds in Group III can occur after other consonants and
so form clusters. In such cases, if the preceding consonant is voiceless, the
whole cluster is usually voiceless.

Say the word 'prei pray and investigate the voicing of the r. In most
people’s pronunciation it is devoiced.
Similarly investigate the 1 in play 'plei, and the w in twin 'twin
and the j in pure ‘pjua.

We can again show this devoicing graphically by drawing a voicing

diagram, like this—

pray P r ei play P 1 ei pure P j U3

rWW> rWWV

GROUP IV
Sounds in Group IV, i.e. vowels and diphthongs, are in principle always
voiced. Occasional exceptions may arise when a vowel is entirely unstressed
and between two voiceless sounds.

43
Voicing {continued); Intonation

Say the word success sak'ses. The first vowel is sometimes devoiced:
try saying it in that way, and compare it with a pronunciation
having a voiced a.
Do the same with support sa'pat. If the a is devoiced, does the word
become identical with sport ?

We are now in a position to make voicing diagrams for any word or

phrase. Here is the diagram for the word common—

common k D m 3 n

Voicing starts just after the k is released. It is switched off again at the end
of the word: notice that the final n does not get devoiced. (Why ?)
Here are some more words with their voicing diagrams. Study them care¬
fully and check that they agree with both observation and theory.

sticky s t I k I

music m j u 2 i k

What would these words sound like if all the sounds were voiced?
If all the sounds were voiceless ?

Exercise
Make voicing diagrams for the following words or phrases—
crocodile
renewing
grab it
watch them

EAR-TRAINING
These nonsense words aren’t easy so be pleased when you identify
sounds correctly and don’t be discouraged if you get them wrong.
44
 Voicing {continued); Intonation

Your teacher will repeat the correct sound and your version of it so
that you can hear the difference. Make them yourself and feel as
well as hear the difference.

'd3AtJYprias kujl'puakts 'xaePtn

INTONATION
By intonation we mean the linguistically significant variation in the pitch of
the voice during speech. Changes of pitch, together with features of stress,
length, and rhythm, enable us to convey meanings sometimes just as impor¬
tant, though rather different in kind, as those we express through the use
of particular words and grammatical patterns. In learning to analyse intona¬
tion patterns, ear training is even more important than in other branches of
phonetics.
We can start by listening to high and low level pitches.

Listen to the word two said with high and low pitches. Then imitate
them.

But we rarely use level tones on one-word sentences in English. To see the
kind of intonation we might get if two was said as a one-word answer in
conversation, we have to look at changing pitches.
A falling pitch (usually just called a fall) is one that goes from a higher
pitch to a lower one, thus for example—

\ Two.

t u

Listen to some falls. Then imitate them. Say the words now, fine,
oh, where, with falls.
In phonetic notation we can show a fall thus—

'tu, 'nau, 'fam, 'au, 'wsa.

A rising pitch (usually just called a rise), is one that goes from a lower
pitch to a higher one, thus for example—

Two ?

t u

Listen to some rises. Then imitate them. Say the same words now,
fine, oh, where, this time with rises.
45
Voicing (continued)/ Intonation

In phonetic notation we can show a rise thus: 'tu, 'nau, 'fam, 'au, 'wea.
Is it true that statements always have a falling intonation, while
questions always have a rise ? (How come we can say ' Where ? with
a fall ? Does 'Fine have to be a question ?)

EAR-TRAINING
Listen to the following words dictated sometimes with a fall, some¬
times with a rise. Identify the intonation used.

Right
Ten
Soon
Run

Comment on the speaker’s attitude implied by the use of these

intonations.

TRANSCRIPTION
Transcribe these lines by Bob Cobbing and indicate some possible
falls and rises.

Example: 'kri 'znk 'kri 'zDk 'kri 'zDk

1. Cri zok cri zok cri zok

Rinkle stammen rinlcel stammen
Tak tak tak tak
Gros temps gros temps gros
Temps temps temps tempe
To two too door
A door adore
Toe toe toe toe
Zzzzz

2. Bombast bombast
Bomb bomb bomb bast
Bombast
Emphase
Em- em- em- phase
Bombast emphase
Bombast
Phebus.

46
13 Vowel Sounds

EAR-TRAINING
Nonsense words—
'pop pDpakaeta'petl
'Dp a'pDsm
'pDm pDma'lDdgikl

Analysing vowel sounds is a more difficult task than analysing consonants.

This is because a consonant usually has an obstruction at some point in the
mouth, and we can easily locate and identify the obstruction. But a vowel
sound involves no obstruction in the mouth, although its quality (or tamber)
does depend on the attitude and position of the tongue. The mouth is really
a sort of tube, ending at the lips and with the tongue for a floor; as the
tongue changes position it changes the shape of the tube through which the
air passes. As the shape of the tube changes, so the resultant vowel tamber
alters.
It is usual to classify vowels in terms of three primary variables—

(1) the height of the tongue—how close is it to the roof of the mouth?
(2) the part of the tongue which is highest, i.e. nearest the roof of the
mouth;
(3) the position of the lips—rounded or spread ?

We can often judge the height of the tongue by investigating the degree of
jaw opening. Other things being equal, an open jaw implies an open tongue
position, while a narrow jaw opening implies a close tongue position.

Take the words

bead bid
bid bid
bed bad

What do you observe about the degree of jaw opening when you
say these words? Which has most, which has least? What do you
conclude about the height of the tongue in the vowels i, i, z ?

The following chart shows five English vowels arranged according to the
height of the tongue in their formation.
47
Vowel Sounds

i (Pete) close tongue position JK

i (pit)
t (pet)
x (pat)
a (part) open tongue position ,,

Notice that the term close here is an adjective (it rhymes with gross and
dose). Don’t confuse it with closed : for a vowel such as i, the tongue is not
closed but raised close to the hard palate.
Because of their tongue positions, we call vowels such as i and i close
vowels, and vowels such as x and a open vowels.

Now try the vowel u in shoe |u. Does it have an open or close
tongue position? What about the sound in hot!

Here is a simple diagram of the English (SBS) vowel sounds arranged by

tongue position.

i u CLOSE
I U
£ 3 D
X A

a 0 OPEN

(a can be regarded as having the same tongue position as 3 from which it

differs mainly in length.)

What can you say about the height of the tongue during the diph¬
thong ai? (The tongue starts open and goes to a nearly close
position.)
Now find out about au. (Again open to close.)
The diphthong ia is rather different—how? (The tongue starts
rather close and then moves opener to a more central position.)

The second dimension for vowel classification is implicit in the diagram

just given. According to the part of the tongue which is highest in the mouth
for their formation, we can classify vowels as front, central, or back.

Say the vowel i, but with an ingressive air-stream (sucking in). The
air feels cold at the point where the tongue is closest to the roof of
the mouth. Feel how for this vowel sound the tongue is relatively
forward in the mouth.
48
 Vowel Sounds

Then do the same thing with the vowel d. Notice how the cold
part is now at the back.

It is convenient to divide the continuous surface of the tongue into several

areas—

the tip (used in articulating t)

the blade (used in articulating s)
the front (part highest in the mouth for i)
the centre (part highest in the mouth for 3)
the back (part highest in the mouth for u)

(See diagram of organs of speech, page 13.)

Beware of the term front. Don’t confuse it with the blade or tip.
Look at your mouth in a mirror while pronouncing i. Can you see
what the front, blade and tip respectively of your tongue are doing?

We can now add to our simple vowel chart labels showing the part of the
tongue highest in the mouth for the formation of each vowel.

FRONT CENTRE BACK

i u
I U
£ 3 D
St A

a D

Look again at the organs of speech (page 13). Note that the tongue is
the shape of a clenched fist (not long and thin like a dog’s tongue). The tip
can be pointed, though.

Experiment with the tip of the tongue, pointing and unpointing it.
Is it usually pointed while we speak ?

The tip and blade of the tongue are easy to see. The front lies under the
hard palate when at rest. The back lies under the soft palate.
We could draw a cross-section showing the tongue position for any vowel-
sound, e.g.

' n
t

(highest part of the tongue arrowed)

49
Vowel Sounds
But usually we abstract from such diagrams a schematic chart representing
the locus of the highest points of the tongue. Marking a point on this chart
implies an auditory vowel tamber and a corresponding tongue position

The last main variable we have to consider in vowel classification is the

LIPS.

Say u. What do the lips do? (Use a small mirror to find out if you
are not sure.)
It is quite possible to say a sort of u-sound without any lip¬
rounding. Try this. Then say u exaggerating the lip-rounding. Is
there much effect on the resulting sound ?

For most English vowels the lips are unrounded (spread or neutral).
Lip rounding is usually greatest and most consistent with the vowel o.
We are now all set to identify vowels in terms of our three main variables.

Height of Part of tongue Position of

tongue highest lips

i close front unrounded

D open back rounded

(Some people have little or no lip rounding with d.)

Now try and classify the French vowel y (as in lune lyn, “moon”).

y close front rounded

So y differs from English i principally by having rounded rather than

spread lips. It is difficult for English people to learn since it involves
an unfamiliar combination of lips and tongue position. Now
analyse the vowel ra, which occurs in Japanese.

ui close back unrounded

50
 Vowel Sounds

To make ra, do u but spread the lips. This, too, is unfamiliar to

English ears and mouths.
Analyse the English vowels i and 3.

i fairly close fairly front unrounded

3 midway central unrounded
between
close and
open

Some vowels vary a good deal in different parts of the country and in
different accents. Thus a, as in love Iav, is usually unrounded and fairly
open; but it may be fairly back, central, or fairly front (the last being typical
of Cockney). In Ireland it is often rounded; in the Midlands and North it
is often unrounded but midway between close and open.

Try out as many varieties of a as you can. Do all the members of

the class pronounce love in the same way? (A “broader” Midlands
or Northern accent has u rather than a, thus luv.)

As the tongue moves during the pronunciation of a diphthong, all three

variables may change. Thus—

di from
fairly open back rounded
to
fairly close front unrounded

EAR-TRAINING
Try the opening of Richard III. It’s rich in vowel sounds, especially
au.

['ritjad] 'nau iz 5a 'wintar av aua 'diskan'tsnt

meid 'glorias 'sAina bai 6is 'sAn av 'jak,
and 'al 6a 'klaudz 6at 'lauad apDn aua 'haus
in 6a 'dip 'buzm av 6i 'aufn 'bsnid.
'nau ar aa 'brauz 'bound wi6 vik'tarias 'ri6z,
aua 'bruzid 'amz hAi) 'Ap fa 'mDnjamants,
aa 'st3n a'laeramz 'tfeindjd ta 'mcri 'mitigz,
aa 'drEdfl 'matfiz ta di'laitfl 'me3az.
'grim 'vizidjd 'wa ha0 'smu6d hiz ‘rigkld 'frAnt,
and 'nau, in'stsd av ‘mauntig 'babid 'stidz,
ta 'frait 6a 'saulz av 'fiafj ' aedvasariz,
hi 'keipaz 'nimbli in a 'leidiz 'tfeimba,
ta 6a la'sivias 'plizig av a 'lut.
51
Vowel Sounds

Transcription
There is a police message for motorists in the Barnet area of London.
A lorry has shed its load at the Apex Corner roundabout on the Al.
You are asked to avoid the area as much as possible. South-bound
traffic will be diverted for the next two hours. That is the end of the
message.

52
14 Assimilation

EAR-TRAINING
Here are some nonsense words. (Remem¬
ber that nonsense words are good
practice, since they make us concentrate
on sounds and forget about meaning.)

tij fns pum sank gDS deig nAtJ

bAV d33ug kmd3 f3b mauf tfin
haip vaeg

Doing dictations in phonetic transcription should by now have made it

obvious that many words can have varying pronunciations according to the
surroundings they are in—of, for example, may sometimes be av, v, or just a.
But there are other kinds of variation in connected speech. One of them
is called assimilation.
The word good by itself is pronounced gud. But in phrases such as good
boy, a good man, good people, it may become gub instead.

Try these phrases out, saying them as 'gub 'bni, a 'gub 'maen, 'gub

Pip?
In these instances the final d of gud has been replaced by a b. This has
the effect of giving it the same place of articulation as the following con¬
sonant (p, b, or m—bilabial). The articulation of the phrase is thereby
simplified, since no tongue-tip movement is now needed at the end of good.

Work out just what movements of the organs of speech are needed
(i) for the sequence dm in the unassimilated pronunciation a 'gud
'maen, and (ii) for the sequence bm in the assimilated version a 'gub
'maen.

Assimilations of this kind are in fact very common in rapid, colloquial

speech, though many people find this difficult to believe. They are fewer in
slow or formal speech.
Now let us consider the word good when it occurs in a phrase such as
a good cook or a good girl. Here it may be said as gug.
53
Assimilation
Try out the unassimilated and assimilated pronunciations. How is
the assimilated one simpler to say ?
(Notice that in the assimilated 'gug 'g3l the final consonant of
good is not left out altogether. For the gg sequence, the back of the
tongue stays pressed against the soft palate for a longer time than
it would for a single velar consonant. Compare 3 ’big 'g3l a big girl.)

Assimilations like these may occur whenever a final d is followed by a

bilabial (p, b, m) or velar (k, g). They are part of people’s ordinary speech
behaviour.

What assimilation is possible in the phrases salad cream and broad

beans ? Transcribe the unassimilated and assimilated pronunciations.
Can assimilation occur within compound words? (Consider words
such as mudguard, broadcast, headman.)

In fact the alveolars t and n are just as subject to possible assimilation as

d is. (Since it is the alveolar consonants that are affected, this commonest
kind of assimilation in English is known as de-alveolar assimilation.)
So final n may become m before a following bilabial, or g before a following
velar.

Explore what can happen to the word ten in phrases such as ten
minutes, ten kings. (It may become trm or teg respectively.)
The prefix un- is often assimilated, thus Ain’pkznt, Ag'kaind. But
remember that the unassimilated pronunciations occur too.
Construct other examples of assimilation of final n, using the
words one and in.

Final t may become p before a following bilabial, or k before a following

velar. Thus daet 'bm may become Qaep ‘bm; and Saet 'g3l may become daek
'g3l. However many people tend to pronounce the t in such cases as a glottal
plosive [?], rather than as an alveolar, and a glottal plosive is not usually
assimilated.

EAR-TRAINING
Practise until you can hear the difference between unassimilated and
assimilated versions, e.g.
Dn 'kas—Dg 'kos
'staend 'bai—'staemb 'bai

Can you distinguish between that with a final alveolar [t], bilabial
[p], and glottal [?], in the phrase that boyl With practice you should
not find it too difficult.
54
 Assimilation

Assimilation may occasionally bring about an ambiguity of meaning—

ju 'nid sm 'hop ma'njua.
dear a 'raip 'pear av 'fulz.
'kAm m an 'sip bai da 'faia.

But such cases are rare, and the context usually makes things quite clear.

See if you can think of other examples of assimilation leading to

ambiguity. (It proves rather difficult to think up plausible examples.
What do you conclude about some people’s assertion that assimila¬
tion is to be avoided because it is lazy and leads to ambiguity ?)

In all these cases we saw an alveolar sound replaced by one which is

identical in manner of articulation and voicing but different in place of
articulation. Thus t becomes p or k; d becomes b or g; and n becomes m
or g. It is a characteristic of English that assimilation most typically affects
just place of articulation; in some languages—French, for example—it is
most usually voicing that is affected. English speakers, though, don’t assimi¬
late voicing (except some Scots): ’blaekbad blackboard has no tendency to
become ‘blaegbod.
There are other alveolar consonants in English which we have not yet
considered. The alveolar fricatives, s and z, are also subject to assimilation,
but only when they are followed by J, 3, or j. So 'dis 'fop may become 'dij
'jDp, while 'diz 'fops may become 'di3 'fops and spaceship tends to be
'speijfip.
Some people do, and some people don’t, assimilate s and z before j. For
those who do, 'siks 'jadz six yards may become 'sikj 'jadz.

Try out the following phrases, all potentially assimilable:

yes, you can
these shoes
in all these years
What assimilations are possible? Do you think you yourself com¬
monly make them in colloquial speech?

In such cases j may even disappear, and we may speak of the coalescence
of sj into J or of zj into 3, thus in 'keifu fa'get in case you forget. Another
common kind of coalescent assimilation is the change of tj or dj into the
affricates tf or d3, for example 'wud3u would you.

PREPARATION: WRITTEN WORK

Show how the final consonants of the words bad, fine, and nice may
be assimilated in connected speech.
Remember that the aim of phonetics is to increase your awareness
of what you do (or may do) when you talk.
55
Assimilation

READING OR DICTATION PRACTICE

'leidiz an ^entjman, in a 'fju ’mmits wi JI bi a 'rami) ak 'glazgau.
'pliz 'fasn ja 'sipbdts and ik'stigwij fa sigaTets. pliz 'stei in ja 'sits
antil di 'cakraft az 'kAm tu a kam'plit 'staenstil. bifa 'livirj di 'aakraft
'meik 'Ja ju haev ja 'haembaegidj 'wid ju. wi 'haup juv m'dgmd da
'flait.

Transcribe this passage—

The Rise and Fall of the River Mersey
“Let us consider Liverpool as a whole" (Sir James Mountford).
The history of this great city is spread around it in the panorama of the
place names of the neighbouring towns and villages. Within the city
boundaries is the great cemetery of the early pioneers, once known as
Boot Hill, now shortened to Bootle. To the north, on the Irish Sea
coast, is the town that kept alive the pig trade with Ireland during the
time of “the troubles,'1' still known to many Liverpudlians by its old
name of Sowport, refined by its inhabitants to Southport. Further north
still is a monument to Liverpool's connection with the slave trade,
once known as New Liverpool, but because of the number of escaped
slaves who made their way there, now known as Blackpool. To the
south lies Runcorn, whose name comes from the corn runners of
eighteenth century corn prohibition.
Even over the Mersey, in the peninsula known as Wirral (from the
oft-repeated Lancashire expression “It's luvly today, wirral we go?'')
the place names are of great historic interest. Liverpool first had a
mayor in 1352. Up to that date the man who messed everything up
was a lord's bailiff.

56
15 Elision

EAR-TRAINING
Nonsense words to help locate difficulties
of aural discrimination—■

rub 0ak sau0 rag fm6 0ij m3

tad J3V tjus

Once upon a time the words listen and Christmas were pronounced with a t
after the s. Nowadays there is no pronounced t corresponding to the t of the
spelling: the t which was formerly pronounced has been elided, historically
speaking.

Try saying 'listp, 'kristmas instead of 'lisn, 'krismas. Which pro¬

nunciation is correct? Why? Find out how members of the class
pronounce pestle, then look up its pronunciation in a dictionary.
Discuss how people pronounce postman: which is more usual,
'paustman or ’pausman? Is either of these pronunciations wrong?

But elisions are not just a matter of historical development, as with listen
and Christmas. Like assimilation, elision affects the pronunciation of words
in running speech.
Consider the phrase last month. This can be pronounced as ‘last 'mAn0.
But in ordinary colloquial speech it is more usual to elide t, giving 'las ‘mAn0.
Similarly with the phrase round the corner: pronounced carefully it is
‘round 6a 'kana but in faster or less formal speech 'raun 6a 'kana. Here we
say that the d has been elided.

Try out the phrases just given, saying them first without and then
with elision of the t or d.
Distinguish carefully between historical elision (e.g. listen, where
there is no question of pronouncing the elided consonant today)
and contextual elision (where the elided and unelided forms are
both to be heard).

Just like assimilation, elision makes words and sentences easier to say.
Since it does not usually lead to any confusion, it is accordingly very common.
57
Elision
But its nature and incidence differ from language to language, so it is evi¬
dently institutionalized and part of our cultural behaviour.
Thus in English assimilation mainly affects the consonants t and d and
the vowel a. It occurs mainly in the following contexts—

(i) When t follows a fortis consonant and precedes any consonant,

e.g. 'neks 'wik, at 'f3s ‘salt.
(ii) When d follows any consonant and precedes any consonant, e.g.
‘staen 'fast, 'aul 'maen, 'kainnas.
(iii) When a is between consonants in an unstressed and non-final syllable,
e.g. 'm3dra, 'naejnl, 'trifik, 'klosl. It is particularly common where the
strings of a and r occur: both a and r may be elided, e.g. 'litaran
becomes 'litrari or ‘litri.
Discuss the pronunciation of the words contemporary, believe,
secretary, police, probably, in the light of possible elisions.

Exercise
Take the present-tense and past-tense verb forms send and sent.
Under what conditions do their final consonants tend to be elided?
Can this ever cause confusion between present and past? Do the
same considerations apply to fetch and fetchedl (Try sentences such
as I'dfetch(ed) them.)

Notice that the final d of and can be elided before vowels as well as con¬
sonants: ‘beikan an 'agz.
There are many cases where assimilation and elision are both possible,
singly or together. Thus soft cloth may be 'soft 'klD0 (full form), 'sDfk'klD0
(t assimilated to k before k), or 'sDf 'klD0 (t elided). And handbag may be
'haendbaeg (full form), 'haenbaeg (elision), ‘haembbaeg (assimilation), or
'hxmbxg (assimilation and elision).

Try out all variants just mentioned. Make sure you can (a) distinguish
them when dictated, and (b) pronounce them at will. Which is your
normal pronunciation?
Show how the pronunciation of the phrase stained glass may be
affected by (a) assimilation, (b) elision, (c) both together.

DICTATION
Either this passage or a similar one with plenty of assimilations and
elisions should be written down phonetically from dictation.

'bai am 'bai, ’pu am 'piglak 'keim a'log.

Note the assimilations! If you failed to notice them, practise hearing

the difference between the unassimilated and the assimilated forms
58
 Elision

—e.g. the teacher should alternate 'bai an 'bai with 'bai am 'bai
until everyone can clearly hear the difference.
'pu waz 'tclig 'piglit m a 'sigig 'vrns dat it 'didnt 'sim ta 'maeta, if i
'diggk get cm 'faeta, 'wDt i 'did.
i

Did you get the difference between the alternative pronunciations

of Piglet ? If not, listen carefully while they are compared and
contrasted for you.
N.B. 'sigig, not 'smgmg or 'siggigg. (Why not?) The pronunciation
with g is of course heard in parts of the Midlands and North of
England, and also in New York.
Note the pronunciation of 'didnt. If necessary, compare and
contrast it with 'didant. The second time didn’t occurs, it is assimi¬
lated to the following velar.
h is often elided in pronouns when they are unstressed and not
initial, thus here i for hi he.
How are you getting along with marking the stress?
'luk pu, seb 'piglit 'svdpli, daz 's\m0ig im 'wah av 6a 'pain triz.
Did you get seb (assimilated before the p in Piglet)? Note the
weak forms daz (there is, there's, also deaz, dar iz), a da (of the, also
av da without elision of the v).
'sau dar 'iz, seb 'pu, 'lukig 'Ap 'wAndrigli, daz an ' aenimj.
Assimilation again in said; elision of a in wonderingly.
Note animal. Compare the forms ' aenimal, ' aenimul, ' aemmu often
heard in the South-East of England. Discuss final -1 and ul in awful,
careful, spoonful.
Finally, try saying the whole passage with strong forms instead of
weak forms throughout ('bai send 'bai, 'pu send 'piglit 'keim a'log.
'pu wdz . . . ). Note the resultant effect: like a badly-read script
spoken by someone not used to reading aloud. Weak forms, and a
modicum at least of elisions and assimilations, give more natural
speech.

Preparation
Find a short passage of English. Transcribe it, marking the stress,
and hand it in for correction.

PASSAGE FOR READING OR DICTATION

'0£egk 'gudms wi 'sim tu av 'stDpt '0igkig av 'maerauz 'aunli az
'huvist 'festavl ig'zibits am 'weitig antil dea laik 'baera3 balunz bi'fa
wi 'it dam. da 'smal 'imerauz haev 'fa 'ma 'fleiva ag kam bi 'kukt m
a 'lumbar av 'weiz, 'fraid, 'stAft, ar in an An'ju3l ‘sup, a 'gud 'wei av
'juzig Ap da 'biga wAnz.
59
16 Plosive Theory

EAR-TRAINING
'pli3 'jAt 3a 'da.
ai ‘daump ba'liv it.
Compare—
daump, daunt, daum, daun.
bi’liv, ba'liv.

Quite often in English we get the same consonant twice running, where
one word (or grammatically separate part of a word) ends in a consonant
and the next word (or part) begins with the same consonant. Thus penknife
is pronounced 'pcnnaif, with two n sounds.

Say the word penknife over. Notice how we usually run the two n’s
together, making a single long n sound.

Similarly, coolly is 'kulli, with two l’s—realized as a single long 1-sound.

Compare coolie 'kuli, with a single 1.
(Note: the paradox of whether we have one n or two in penknife, one 1 or
two in coolly, will be resolved when we deal with phoneme theory, Chapter
20. In these words we have two occurrences of the phoneme concerned, in
each case realized phonetically as a single long articulation.)
We get the same thing in a phrase such as stafffund ‘staf ’fAnd. When the
double consonant is a fricative, a nasal, or 1, it is pronounced just as a single
consonant, but longer.

Try and think up more examples of this and check on their pro¬
nunciation. Investigate doubled m, v, 0, 3, s, z, J. (Why are doubled
g, h, and 3 unlikely to occur in English?)
Remember we are talking about doubled sounds, not doubled
letters in the spelling—the two m letters in commit, for example,
correspond to only a single m sound.

When two identical plosives come together in this way, the results are
similar. Thus book-case is pronounced 'bukkeis, where kk denotes a voiceless
plosive that lasts perceptibly longer than a simple k.

Compare 'bukkeis with the non-existent word 'bukeis.

60
 Plosive Theory

Usually, when two plosives with the same place of articulation come
together, they are realized with one long plosive articulation. Occasionally,
and in very careful speech, there may be two successive plosive articulations.

How would you normally say bad dog ? Are there two d sounds here
realized as one long plosive, or as two separate ones? Try the phrase
both ways, and compare the two possible ways of saying it with dd.

To understand more fully what is involved here, we must look in greater

detail at the articulation of plosives.
Plosives are sounds in which the air-stream is entirely blocked for a short
time: in English, p, b, t, d, k, g. We can distinguish three phases in the
articulation of a plosive: the approach (as the articulating organs come
together), the hold (as they stay together, preventing the air-stream from
escaping), and the release (as they separate and allow the blocked air to
escape). Some people distinguish a fourth stage, the plosion, when the
characteristic noise of the escaping air is heard. We can show this diagram-
matically—

* i

approach release

Draw similar diagrams for the plosives in echo and eddy.

It is useful to distinguish these three phases clearly, since the approach

and release phases of plosives may vary considerably according to the
phonetic context, whereas the hold phase is relatively free from variation.
In order to make any plosive, it is necessary to have both the nose and
mouth blocked off—otherwise the air would continue to escape through the
one that was not blocked. Therefore there are always in a sense two closures:
the soft palate is raised to prevent the nasal escape, and the lips or tongue
articulate in the mouth to prevent oral escape.
In the case of the b in abbey ' aebi just considered, the soft palate was
already raised for the vowel which preceded the b. It remained raised for the
vowel which followed, and indeed did not move during the articulation of
the word.
But now consider the word amber 'aemba. As we saw in Chapter 8, for the m
sound in this word, the soft palate is lowered to allow the air out through the
nose, and the lips come together to block any air escape through the mouth:
as we pass on to the b sound, the soft palate rises, stopping any further nasal
escape, while the lips merely stay together. They do not need to come
61
Plosive Theory

together for this b, since they already are together. Then, as we pass from b
to a, the lips separate in the ordinary way.

Say over the word amber, using a pocket mirror to check that there
is only a single lip articulation. Changing from the m to the b
requires no movement of the lips at all. (What does change as we
go from m to b ?)

We say that the b in amber has a nasal approach, since the approach
consists not in the coming together of the primary articulators but in the
rising of the soft palate.
The same thing applies to the d in handy. Here the tongue is the primary
articulator, forming a block in the mouth with the tip on the alveolar ridge.
For the n the soft palate is down; it rises to turn the n into d.

ae m a h * n i

nasal approach nasal approach

Show how the same thing applies to the g in anger. What are the
articulations here?

But it is not only the approach that can be nasal, made by a movement
of the soft palate only. We can have nasal release too.
Consider the b in the word crab-meat 'kraebmit. It has an ordinary ap¬
proach, with the lips coming together to complete the blocking of the passage
taken by the air escaping during the ae. (The soft palate is already raised
from the beginning of the word.) To turn the b into an m, all that happens
is that the soft palate lowers. This allows the air to escape through the nose
and converts the plosive into a nasal. The lips remain firmly closed together:
there is only a single closing and a single opening of the lips for the sequence
bin. (After the lips have separated from the m, when the i is being pronounced,
the soft palate rises again.)
The same applies to the bm in submerge. Here again the b has a nasal
release: the release is performed by a movement of the soft palate, not by
any movement of the lips.

Say the words crab-meat and submerge, observing how there is a

single bilabial closure and a single bilabial release for the sequence
bm. What would the sound be like if the b was released separately?
Would you ever say them that way?
62
 Plosive Theory

We can have nasal release in just the same way at other places of articula¬
tion. In kidney 'kidm and midnight 'midnait the d has nasal release.

Does saying that a d has nasal release mean that air escapes through
the nose during its articulation? (No, not during the hold phase—
only during the release phase. Where a sound has air escaping
through the nose throughout, e.g. n or 5, we say it has nasal escape.
Don’t confuse the two terms escape and release.)
Say whether the plosives in the following words have (a) nasal
approach, (b) nasal release.

end
longer
submit
standing
Rodney
kindness
slumbered
designed

Voiceless plosives can have nasal approach and/or release, of course.

Nasal approach for p, t, k is quite common: e.g. lamp 'laemp, rant 'rant,
thank '0aei)k.

Think up some further examples of nasal approach to a voiceless

plosive. Describe carefully the articulations involved.

Nasal release of p, t, k varies from speaker to speaker. Some people do

use it in words or phrases such as topmost, stop me, greatness, cat-nap,
button 'bAtn, steak and kidney 'steik g 'kidm; but many people use glottaliza-
tion (see Appendix 1) in some or all of these words, which means that the air
is blocked by, and released by, the glottis rather than by some organ in the
mouth.

Listen to a demonstration of the difference between button said with

a nasally released t and with a fully glottalized t. Imitate both
pronunciations and make sure you can produce either at will. Do
the social connotations of the two pronunciations differ ?

INTONATION
We can make a distinction between falls which fall a long way (from high
to low) and those which fall a short way (from mid to low).
63
Plosive Theory

HIGH FALL ; LOW FALL ^

Two! Two.

We can symbolize falls more precisely thus: high fall, 'tu; low fall, Ntu.

Do ear-training on high and low falls until you can distinguish

them with certainty.

We can also distinguish a high rise (low to high) from a low rise (low
to mid)—

HIGH RISE ; LOW RISE

Two? Two...

We can symbolize rises more precisely thus: high rise, 'tu; low rise, ,tu.

Do ear-training on high and low rises until you can tell them clearly
apart.
Then mix high and low falls. Stick for the time being to one-
syllable utterances, e.g.

'Them jOut ,Right 'All

EAR-TRAINING
(i) Nonsense words

bm'naup \u'lef0
'WEndadmp wim'mux

(ii) English
'luk. 'V1311 an 5i 'ai. 'kmnpn'hEnsiv, 'im'maekjulatli 'mauntid
Eksa'bijn kwriq 'evri 1 sespekt av 'nptiks. a'nsetami am fizi'Dlad3i av
6i 'ai, n'kadid frm '3lnst 'grik 'taimz. kl'ekjn av 'kweint 'speks
di'vElapt sins 'rndja ‘beikg, '03tin0 'sentjn, 'f3st 'hit mi 5i ai'dia.
'hDrifangli 'big 'knntaekt 'lsnziz, a'neit 'aibadz, 'vijual 'lai ditekta ta
'ksetj 'aut 'Sauz 'feinig 'blamms, 'intrakat 'daiagraemz, 'keisiz av
'tatjuas 'mstramants. 'aiz Dn 'staks, 'litrah, fr m't3nl ig'zaemi'neifnz.
aen'tik 'maikraskaups. 'siarias 'kvvrid3 ta mit 'al '1ev|z. 'kliab
iks'pleind fa 5a 'lerman. I'mjali a 'bip ba'wildrig. ab'zabnj ’wAnJ ju
■gsk 'gauig.
64
 Plosive Theory

Transcribe this passage—

Human reactions to danger are often archaic instincts, meaning
inherited memories, as when a bomb falls through the roof of a building
and explodes and the shock survivors absurdly try to scratch a hole
through the tiles with their nails—because their remote ancestors
would have acted like that in some stage of the human evolution from
the three-eyed lizard to hominoid. Nevertheless, the common flying
dream is no proof that the dreamer was ever a bird; or indeed, that
any of his ancestors were, since palaeontologists deny this link in our
evolutionary chain. It seems to be either metaphorical of a wish to
fly away from our present circumstances or else—since time is only a
convention and memory works both ways; either as reminiscence or
as prophetic anticipation—of a future age when human beings will
develop wings, as birds once did, and dispense with balloons, planes
and rockets.

65
17 Lateral Approach and
Release

EAR-TRAINING
Begin with these words, and others, and
say them several times for practice on
falls and rises.

'help! 'hei! vluk! ,hu?

In this next group notice the plosive

release.

mak'txgat mak'natn maekna'mara

After the next lines have been dic¬

tated, practise them with falls and rises.
(If you do this in threes or fours and
each do them in your own way and in
your own time, you will have the
beginnings of a tone poem.)

'da a'dad 'dgaekda 'djil

'plop 'plDmp 'plau

As we saw in Chapter 7, the difference between 1 and d depends essentially

on what the sides of the tongue do.

Make an 1-sound, as in let, but keep it on for a long time (a second

or so). Then make a d-sound, keeping the hold phase going as long
as you can. Feel how in each case the articulation is made by the
tongue tip against the alveolar ridge, while the soft palate stays up
and the vocal folds vibrate. Ask yourself what the difference is
between the two articulations. How does 1 differ from d?

For the lateral, 1, the sides are down, allowing the air to escape laterally
without any restriction. But for the plosive, d, they are up, thus preventing
the air from escaping laterally; and, since the escape of air is blocked by the
tip of the tongue in the middle of the mouth and the side rims of the tongue
at the sides of the mouth, whilst the raised soft palate prevents any escape
down the nose, the air-stream is entirely blocked and a plosive hold results.
66
 Lateral Approach and Release

Now consider what happens in a word such as badly 'baedli, where d is

immediately followed by 1. In such cases the plosive is effected by the lowering
of the side rims of the tongue. The dropping of the sides of the tongue is all
that is necessary to convert a d articulation into an 1 articulation.

- aedl - in badly 'badh.

x 1

ordinary approach lateral release

(tip rises) (sides drop)

Say the word badly over a few times, trying to feel the lateral release.
Try the further examples mudlark, oddly, good looks, Dudley.
(Compare the release of the two d sounds in the last example.)

It is very common to have a lateral approach, namely when d is immediately

preceded by 1, as in field ‘fild, seldom 'scldam. But here there are the compli¬
cating factors—

(i) simultaneously with the rising of the sides of the tongue, an adjustment
may occur at the back of the tongue, which has been raised somewhat
for the I (dark 1, see Chapter 19);
(ii) in many accents, notably those of the South-East of England, there is
often no actual lateral where SBS has 1, but just a glide resembling
the vowel u. If there is no true lateral, there can be no lateral approach.

Say the words field and seldom in your usual pronunciation and
determine whether your d has a lateral approach. If not, practise
making one.

Henceforth we assume a pronunciation which has lateral approach as

SBS.

- Id - in seldom saldam.

lateral approach ordinary release

(sides rise) (tip drops)
67
Lateral Approach and Release

In a word such as childlike 'tjaildlaik (said without elision of the d!) both
the approach and the release of the alveolar plosive are lateral—

- Idl - in childlike tjaildlaik.

1 1
d

hold
lateral approach lateral release
(sides rise) (sides drop)

Notice that the tongue tip remains motionless throughout the whole ldl
sequence.
In a word such as middle 'midi most people have a lateral release to the d,
the tongue tip remaining still through the dl sequence—

- idl - in middle 'midi.I

i 1
d

hold
ordinary approach lateral release
(tip rises) (sides drop)

Here the 1 is syllabic, but that need not affect the issue. In the South-East
of England, though, we again find a tendency to use a vowel-like sound
rather than a lateral after the d, giving a pronunciation sounding like 'midu.
Those who have been taught “not to leave their /’s out” may then add 1 on
to the end of 'midu, giving the “careful” pronunciation 'midul (or ’midal)
often to be heard from Londoners. In such a case, of course, the d does not
have a lateral release.

Say words such as middle, saddle, curdle, cradles, and see whether
you usually give d a lateral release. If not, practise doing so.

Where t is preceded or followed by 1, it too has lateral approach/release—

unless it is fully glottalized. (Of course, to go from 1 to t, or from t to 1, also
requires a change in voicing.)

Say words such as bottle, atlas, felt, filter, finding out in each case
whether your t has lateral approach or release.

We have now dealt with two special kinds of approach and release. In
nasal approach/release, the change in articulation is a movement of the soft

68
 Lateral Approach and Release

palate; in lateral approach/release, the change in articulation is a move¬

ment of the side rims of the tongue. Since the opposite of nasal is oral and
the opposite of lateral is median, the “ordinary” approach/release, charac¬
terizing for example the d in eddy, is properly termed median oral.

For each alveolar plosive in the following phrase, state whether the
approach and release respectively are median oral, lateral, or nasal—
I didn’t like getting bundled into the old lorry.

Notice that, in a long sequence of alveolar plosives, nasals, and laterals

such as ndldn in long-handled knife, the tip of the tongue does not have to
move at all. The successive plosive approaches and releases are all nasal or
lateral, depending on movements of the sides of the tongue or the soft
palate—

n d 1 d n

sides up up * down * up up

soft palate down t up up up | down

* lateral release
and approach
f nasal approach
and release

It is a marvellous thing how the human speaker can perform articulatory

movements as intricate as these entirely without conscious thought and at
very high speeds.
When 1 precedes or follows plosives at some other place of articulation
than alveolar, the approach/release is not strictly speaking lateral, though the
term is sometimes loosely used in this way. (In apple ' aepl for instance, the
bilabial closure of p is released by the separation of the lips not laterally but
in the median line; but an alveolar lateral articulation is made during the
plosive hold, so that the air which escapes on the release does have to escape
laterally round the tongue.)

EAR-TRAINING
Transcribe the following passage from dictation

'las 'taim 6a ‘bi bi 'si 'sent a ’ka ta k'lskt mi, it ’almaust 'kaptjad
6a rorj ‘man. ta pra'vsnt 6a 'draiva 'luznj imself in a 'taijgl av
'kAntri 'leinz, 6a 'rondavu waz 'nain '03ti pi 'em at 6a 'laukl. nn a
■pitj 'blak, 'stami 'wmtaz ’nait, ai bi'gan ta gat 'agfas az 6a 'kink
69
Lateral Approach and Release

'krcpt tadz 'ten. dsn 5a 'laendlad 'sAdnli 'baerjd iz 'brau an ssd da'WDz
SAmwAn 'askiij fa mista 'raian and i waz 'sent 'daun 6a 'raud ta 6a
'lDd3.
ai ‘hArid a'kms ta wea mista 'raian, 'mistifaid and aepri'hcnsrv, waz
biig 'bubd out av ‘bed. aez ai 'entad, ai ‘h3d 6a 'draiva 'sei, 'its 'nDt
fa 'ju ta 'nau 'wai. wen 6a 'bi bi 'si 'kalz fa 'ju, juv d3ASt 'gDt ta 'kAm.
it 'mei bi '6is iz ja 'laif.
'sAtJ iz 6a 'paua av 6auz 'dredid I'mjlz 'stil m 'ruarl 'iggland, if ai
had a’piad a 'mimt 'leita, mista 'raian wud av bm ‘wiskt 'Df ta 'famd
imsclf at 'midnait dis'kASig 'gDd wi6 'maelkam 'mAgrid3.

INTONATION
When a fall is followed by a string of unstressed syllables, they are low and
level in pitch—

*\ ..
Happiness
*\ ..
xHappiness

But when a rise is followed by a string of unstressed syllables, the rise

itself is spread out over the whole—

''Happiness ? yHappiness

This applies equally when the tail is quite long, as in the following—
'Never, Professor Jenkinson.
'What was it you said you wanted me to do ?

Practise hearing the difference between falls and rises with unstressed
tails following.

Transcribe this passage—

After living in three countries, Mr. Sachs has noticed that men the
world over expect most of their post to be handled by a female member
of the family, so why not, he suggests, adopt a form of letter to take
care of all usual contingencies? At first sight it’s a tempting idea.
Picture the damp winter morning, for example, when you simply
snatch the pad and mark appropriately the form that reads—
I am sorry my child cannot come to school. He is unfortunately
kept at home by measles I his grandmotherly our maths test-, there is a
bus strike I the car has broken down/his mother has broken down; I
couldn't find him when 1 got up this morning.
70
 Lateral Approach and Release

And another one we could all do in our sleep is—

Dear Mr. Lockjaw, I am sorry my husband will not be able to keep
his dental appointment this week. He is
out of town unexpectedly,
the victim of a virus infection,
working for a difficult deadline,
in a blue funk.

71
18 Overlapping Plosive
Consonants

EAR-TRAINING
Nonsense words—
'dDkbsadn 'bmiggdl
'dlaebd 'talktpm

Transcribe the following passage marking the places where two

plosives occur in sequence—
Giveaway newspapers are going great guns. Hardly a week passes
without news of another free-of-charge journal being launched. Packed
with lucrative advertising, and sometimes devoid of editorial, the ‘free
press” is catching the eye of readers and publishers all over Britain.
Post Publications, a stable of nine giveaways, was bought last week
by the Manchester Guardian and Evening News group. The price paid
is being kept secret. It was founded five years ago by Derek Meakin
“with fifty pounds, a kitchen table and a pot of glue.”

Consider the sequence bd in the phrase grabbed it 'graebd it. It is a sequence

of plosives with different places of articulation—the first bilabial and the
second alveolar—and in the usual pronunciation of such sequences the two
plosive articulations overlap. What happens is that the release of the first
plosive articulation does not occur until after the approach phase of the
second. For the bd in grabbed it the sequence of events is (i) the lips come
together (approach for b); (ii) the lips stay together, while the tongue tip
rises to the alveolar ridge (hold for b, approach for d); (iii) after a moment of
double hold, alveolar and bilabial (b and d), the lips separate while the
contact of the tongue tip on the alveolar ridge is maintained (release for b,

- aebdi - in grabbed it graebd it.

lips
tongue tip and alveolar ridge

72
 Overlapping Plosive Consonants

hold for d); (iv) after a further moment of alveolar hold (d) the tongue tip
comes away from the alveolar ridge (release for d).
The articulation of the sequence gb in Egbert is similar. The bilabial
closure is made before the velar closure is released.

Work out the detailed sequence of articulatory movements for the

gb in Egbert (as was shown for bd above in grabbed it).

Similar overlapping usually occurs in English whenever two plosives with

different places of articulation occur in sequence. It is usually true of the kt
in actor, the pt in kept, the dp in midpoint (unassimilated), the bt in obtain,
the pd in update, the kb in blackboard, etc. However these and similar
sequences are occasionally pronounced with separate, non-overlapping
articulations for the two plosives. This results in a slight h-sound or a-sound
(depending on voicing) between the plosives. It is characteristic of most
foreign accents of English and also of the careful speech of native speakers
of English.

Listen to the two ways of pronouncing kt in actor-, (i) with over¬

lapping plosive articulations; (ii) with non-overlapping articulations
and a brief escape of breath between the k and the t. Imitate the
two forms and learn to produce either at will. Do the same for the
gb sequence in big boy.

The overlapping of plosive articulations means that the release of the first
plosive in any such sequence is not audible, since it is masked by the second
closure. Hence it is often termed non-audible release. Another term some¬
times encountered, incomplete plosion, is misleading and best avoided.
The “masking” can take two forms. If the second closure is made further
forward in the mouth than the first (e.g. kt, dp, gb), then the air freed on the
release of the first closure can still not escape, since it is blocked by the
second closure. But if the second closure is further back (e.g. pt, dk, bg), then as
soon as it is formed it takes the pressure of air from behind the first closure,
and again no air under pressure can escape on the release of the first closure.
In fact it is something of a simplification to speak of two closures in some
cases—in actor, for example, the tongue never really leaves the roof of the
mouth during the kt sequence, but rolls along it from a velar closure to an
alveolar closure.

Work out what happens during the plosive sequences in the following
words or phrases—
doctor, robbed, stockpot, egg-cup, a locked door, chipped potatoes.

Things are rather similar when we have a nasal before or after a plosive
with a different place of articulation, except that now the movement of the
soft palate has to be taken into account as well. The md sequence in Camden
73
Overlapping Plosive Consonants

'kacmdan involves the following: (i) the soft palate drops during ae; (ii) the
lips close, giving m; (iii) the tongue tip contacts the alveolar ridge, as m
continues; (iv) the lips separate and soft palate rises, giving d a kind of nasal
approach; (v) after the hold phase of d, the tongue tip leaves the alveolar
ridge, giving a median oral release for d. The gm sequence in dogmas exem¬
plifies the order of articulatory events when a nasal follows a plosive at a
different place of articulation: (i) the back of the tongue rises to contact the
velum (median oral approach for g); (ii) the lips close, while the velar closure
is held; (iii) the soft palate drops and the back of the tongue drops away
from it; (iv) after the steady phase of m, the lips separate and the vowel
follows; (v) during a, the soft palate rises again.

Work out the order of articulatory movements during the bn

sequence in hobnob. In what respects does the pn sequence in Stepney
differ?

EAR-TRAINING
it 'waunt 'ksetj 'faia.
it Vaugk 'kaetj 'faia.
ip 'waug 'kaetj 'faia.
'daunt bi 'leit.
'daump bi 'leit.

Transcribe the following from dictation—

ta bi'gm at 6a bi'ginnj: its 'sprig, 'munlis 'nait in 6a 'smal 'taun,
'stabs am 'baibj 'black, 6a 'kDbl 'strits 'sailant an 6a 'hAntJt 'kataz
an 'raebits 'wud 'limpig m'vizabl daun ta 6a 'slaublaek, 'slau, 'black,
'kraublaek, 'fijigbaut 'bobig 'si. 'h\f, 6a 'beibiz a ’slipig, 6a 'famaz,
6a 'fijaz, 6a 'treidzman am 'penjanaz, 6a 'kDblaz, 'skultitja, 'paustman
am 'pAblikan, 6i 'Andateika an 6a 'faensi 'wuman, 'drAgkad, 'dresmeika,
'pritja, 'plisman, 6a 'webfut 'kDklwimin an 6a 'taidi 'waivz.

INTONATION
Falls and rises may be preceded by unstressed or stressed syllables with low
pitch.

Fm'sure he will l’m^ure he will

• • •
I think it’s disgraceful I think it’s disgraceful
74
 Overlapping Plosive Consonants

Practise recognizing falls and rises, high and low, when preceded
and/or followed by unstressed syllables. Remember to listen only to
the syllable where the fall or rise occurs (the nucleus) and any
following syllables; ignore the pitch of syllables that precede the
nucleus.
Say the following phrases with (a) a high fall and (b) a high rise
on the syllable spelt in capitals—

Have you got your BOOKS ?

I don't really MIND.
Do stop TALKing about it.
Where would you like to SIT, Veronica ?

Transcribe the following passage—

The newly formed National Sheep Association believes in going the
whole hog. No nonsense about thinking up recipes for a smart way
with half a shoulder: we've been sent a leaflet—its first—for roasting
a whole sheep. Roast one, or four, for a teen-age party or charity fete,
they suggest. But the economics seem as dicy as most farm figures.
Expenditure: up to £7 for the sheep, about £3 for the charcoal, 50p
say for the sauce, and £1.25 for the baps to make the hamburgers.
Total £11.75. Income 100 portions per sheep at 12\p a head. Maybe
there's a subsidy one can claim.

75
19 Aspiration; 1-Sounds

In words such as pay 'per, time 'taim, carve 'kav, the release of the plosive
is not immediately followed by voicing for the vowel. Between them, there
is a period of voiceless escape of breath known as aspiration : we say that
the plosives concerned are aspirated.

Dangle a strip of paper in front of the mouth and pronounce pay.

The puff of breath constituting aspiration will blow the paper aside
on the release of the p.

Aspiration is characteristic of English voiceless plosives (p, t, and k),

particularly when initial in a stressed syllable. Weaker aspiration, or none,
is used with intervocalic p, t, and k ending a stressed syllable, as in upper
'Apa, outer 'auta, looking 'luknj. In final position, the amount of aspiration
is very variable: reap 'rip, fought 'fat, shock 1 jDk. After s, no aspiration is
usual: span 'spaen, esteem I’stim, scarf 'skaf.

Do the strip-of-paper test again, this time with the words pin and
spin. The paper is blown aside with pm, but not with spm. Why?

When desired we can symbolize aspiration by a small raised h, thus 'phm.

It is not, however, necessary to show aspiration in ordinary transcription.
The voicing diagram for pm looks like this—

pin P h i n

Compare spin s P I n

In a sense, the devoicing of r, 1, w, j after voiceless plosives (Chapter 12)

is a form of aspiration: compare pay and play—
76
 Aspiration; l-Sounds

pay P h ei play P 1 ei

In English it is only voiceless plosives that are aspirated. Languages differ

strikingly in whether aspiration occurs in them: in French, other Romance
languages, and Dutch, p, t, and k are not aspirated, whereas in other
Germanic languages they are.

Learn to hear the difference between aspirated and unaspirated

voiceless plosives. A convenient way of symbolizing them when
desired is ph and p=, etc., respectively.
Try pronouncing phrases such as the following without aspiration:
the party of progress
take cover
pass the potatoes, please
carry on, corporal
Do any English speakers talk like that ?

1-SOUNDS
There is a rather clearly perceptible difference in SBS and most other accents
of English between the kind of 1 used initially, e.g. in lif, 1st, lut, and the
kind of 1 used finally, e.g. in fil, ful, sel. They are both made with the tip of
the tongue touching the alveolar ridge, the sides of the tongue being down
to allow air to escape laterally (and the soft palate being up to prevent nasal
escape). But they differ in quality (or resonance or colour or tamber)
because of the differing positions of the body of the tongue. The first one,
which has a quality similar to the vowel a, is known as clear 1; the second,
which has a quality similar to the vowel a, is known as dark 1. (More
scholarly phonetic terms are non-velarized and velarized respectively.)

You can make an 1 with any kind of vowel quality simultaneously,

by trying to pronounce the vowel at the same time as the 1, keeping
the tip of the tongue firmly in contact with the alveolar ridge. Try
this out, pronouncing 1-sounds with the qualities of i, x, a, d, a,
u, 3, y, etc. Then analyse carefully the quality of your 1 in less and
law, then the quality of your 1 in sell and all.

In SBS clear 1 is used before vowels and j; dark 1 is used before consonants,
before w, and finally before a pause.

Decide whether the 1-sounds in the following words are clear or

dark: luck, call, milk, shelf, allow, pulse, blame, bilge, silver, silly.
77
Aspiration; l-Sounds
Words ending in 1 usually have clear 1 if a word beginning with a vowel
follows closely, but otherwise a dark 1.

Compare sell it! with sell! Compare feel, feel it, feeling, feel them.
What happens in the case of I feel ill?

American English commonly differs from English English by having dark 1

rather than clear between vowels in words such as valley, jelly, yellow,
pillow. People from Wales, Ireland, and Tyneside may have clear 1 in all
positions; people from some parts of Scotland may have dark 1 in all
positions.

Go through all the words with 1 mentioned in this chapter and try
to say them (a) all with clear 1, and (b) all with dark 1.
Does your speech, or that of anyone you know, differ from SBS
in the distribution of clear and dark 1 ?

German 1 is very clear in all positions. So English learners of German,

and German learners of English, often have pronunciation problems with 1.
For example, the German for milk is Milch. If we use the symbol I to denote
dark 1, and the symbol J to denote a specially clear 1, we can say that English
people learning German tend to pronounce milx instead of mij?, whereas
German people learning English tend to say mijk instead of milk.

Try the pronunciations and mispronunciations just mentioned.

Both clear 1 and dark I are usually made with the tongue tip on the alveolar
ridge. The difference in the body of the tongue can be shown in a cross-
sectional picture of the organs of speech—

As we saw above (Chapter 12), 1 is sometimes devoiced in English. And

it was mentioned in Chapter 4 that it is usually made as a dental, not an
alveolar, before 0 and 6.

English dental 1 is always dark (1)—why ?

78
 Aspiration; l-Sounds

INTONATION
Another very common tone used in English is a fall-rise. The fall and rise
can occur on a single syllable—

V
v Now v Three

—or they can be spread over several syllables—

v Sometimes v Now, Professor Jenkinson

And they can be preceded by other syllables—

But it was v silly of you

Try and analyse the difference in meaning between 'Two and ''Two.

EAR-TRAINING
Transcribe the following verse from dictation. Decide whether the
various 1-sounds in it are clear or dark.

9saz a 'blaek 'bal 'bak

'kAmii] daun 9a 'riva,
'blau, 'bubz, 'blau.
9az a 'blaek 'bal 'bak,
■pulir) daun 9a 'pul,
'blau, mai 'bub 'bmz, 'blau.

Transcribe the following passage.

Dylan Thomas, say the locals, who claim almost to a man to have been
daily drinking companions of the poet, would have enjoyed all the
fuss. At any rate, next Wednesday's auction sale of his last home.
The Boat House, at Laugharne, has been the biggest talking point in
this tiny South Wales township since the party which took place at
Brown's Hotel after his funeral. The Boat House which is a damp,
unremarkable building built about 140 years ago at the foot of a cliff,
has a breathtaking view of Laugharne Estuary with the silence broken
Aspiration; l-Sounds
only by the seagulls and the Army’s guns, which have been shelling
off-shore targets a few miles down the coast.
At one time Caitlin wanted Dylan buried in the garden, but instead
he rests in St. Martin's churchyard, where a plain wooden cross
shrieks for attention among the marble memorials put up to the local
dead.

80
 ■ I

20 Phonemes

Up to now we have been speaking rather loosely in terms of “different

kinds of 1-sound,” “d-sounds with various releases,” and so on. But it is
now time to look more closely at some of the theoretical concepts and
assumptions involved in grouping sounds together in this way—sounds
whose articulations sometimes differ quite considerably. We must try and
stand back from the material we are dealing with—pronunciations and
articulations—and look at them with an unprejudiced eye.
Why, as speakers of English, do we feel that clear 1 and dark 1 are obviously
varieties of “one sound” ? Or, again, what is it that the very different t-sounds
of top, stop, pot, eighth, bottle, button, twin, etc., have in common which
leads us to think of them all as being in some sense the “same” sound ?

List the articulatory differences between the t-sounds in the words

mentioned.

It is not just that they are articulated in similar ways. After all, there are
many other sounds whose articulation is extremely similar, yet we do not
think of them as being varieties of the same sound. For instance, the t in
twin has more in common articulatorily with the d in dwell than it has with
the t in button; so why don’t we feel the initial sounds of twin and dwell to
be in some sense the same sound, and different from the sound before the n
in buttonl

Analyse the articulations of the sounds just mentioned. Tabulate

similarities and differences between the three segments in question.
Don’t forget the activities of the lips and of the body of the tongue.

The answer seems to be that sounds belonging together in this way are
never in direct contrast with one another: we can’t distinguish pairs of words
just by using one of them rather than another. In most cases the choice of
the particular variety of sound used in a given phonetic context depends on
that context in a quite direct way. And sounds belonging together do have
considerable articulatory similarity.
Putting the last paragraph into proper technical language, we say that
particular allophones go together into classes known as phonemes. Pho¬
nemes are contrastive with respect to one another. Allophones belonging
81
Phonemes
to a given phoneme are arranged in complementary distribution (or in
some cases free variation); they also show phonetic similarity. In the
course of this chapter we shall examine these various notions individually.
It is, in fact, the phonemes of English that are represented by the symbols
we have been using for transcription of English (page 17)—we have been
using a broad or phonemic transcription. Other symbols that have been
introduced from time to time, e.g. I, denote not phonemes but either allo-
phones or general-phonetic sound-types considered independently of any
language. A transcription using such symbols is termed narrow.
So 1 and d, for example, are different phonemes in English. Clear 1 and
dark I, though, are allophones of a single phoneme: in our phonemic tran¬
scription we accordingly write them identically, both as 1.
It is usual in linguistic work to enclose phonemic symbols in slant lines,
thus HI, /d/, /fri/, but to use square brackets to enclose phonetic symbols
denoting allophones or general-phonetic sound-types, thus [I], [M], [m].
(A further commonly accepted convention is that words quoted in ordinary
spelling are italicized, or in handwriting underlined, thus fell.) We shall
follow these conventions from now on.

CONTRASTIVENESS
We can show that two phonemes of a given language or dialect are con¬
trastive by listing minimal pairs of words distinguished by the contrast
(or opposition) being illustrated. So for English /l/ and /d/ we can list

/lip/ leap —/dip/ deep

/h>t/ lot —/dot/ dot
/fell fell —/fed/ fed
/'pulnj/ pulling—/'podn)/ pudding
and so on.

But clear and dark [1,1] are not contrastive in English—being members of
the same phoneme—and there are accordingly no minimal pairs distinguished
by the choice between them. We say they are “not in opposition.”

List some minimal pairs which are evidence for the contrastiveness
of the following English oppositions—
/t/ and /d/
/au/ and /o/
111 and /w/
Are there any minimal pairs for aspirated [th] versus unaspirated
[t=] ? If not, why not ?

Notice that the contrastiveness of a given pair of sounds in something that

varies from language to language—and indeed from dialect to dialect and
82
 Phonemes

from accent to accent within a language. This is why transcription symbols

given for SBS are not necessarily appropriate for transcribing other accents,
whose phoneme systems may be different.

In Scots accents u and u are not contrastive: good and mood, for
example, rhyme perfectly for the Scots but not for the English. Can
you think of a minimal pair to use in order to test whether [u] and
[u] are contrastive in a given person’s speech ?

People learning foreign languages have difficulty when sounds are con¬
trastive in the language being learnt but not in their own mother tongue.
This is one reason why English people find it hard to acquire the French
opposition /y/ versus /u/, which distinguishes pairs such as rue /ry/, meaning
“street,” versus roue /ru/, meaning “wheel.” And Germans, for example,
find it difficult to get the English opposition /e/ versus /se/, since they have
only one vowel phoneme in the [e—se] area.

List some words which become homophonous if the opposition

between /e/ and /se/ is not made.

The same thing applies to people trying to acquire a different accent of

their own language. Many Midlanders and Northerners, for instance, start
off with no opposition corresponding to that between SBS /a/ and /«/. From
the point of view of acquiring an SBS accent, their problem is one of under¬
differentiation. Not only do they have to learn the different phonetic
qualities appropriate for /a/ and /u/, but they also have to learn which of the
newly-distinguished phonemes to use in which words—to learn the incidence
of these phonemes in SBS.

Discuss why some Midlanders and Northerners say, or seem to say,

/'kAjn/ and /'bAtJa/ instead of /'kujp/ and / 'butja/.

(Notice that sometimes accents differ in the incidence of certain phonemes

without having any difference in their phoneme systems at the point con¬
cerned. Contrastive /se/ and /a/, for example, are found in Northern accents
as well as in the South and in SBS; but their incidence differs, in that words
like path, grass, dance have /se/ in the North but /a/ in the South and in SBS.
In many Scots accents, on the other hand, there is only one phoneme we
can write it /a/—corresponding to the two found in England.)

COMPLEMENTARY DISTRIBUTION
The kind of /k/ we use before front vowels, in words such as key /ki/, is
made with the closure rather far forward on the velum—it is almost a
83
Phonemes

palatal sound, [k], Before back vowels, we make the closure in a retracted,
almost uvular, position [k], as in cool /kul/.

Say key and cool over several times and isolate the initial sounds.
Notice the difference between them.

But there is no question of the front [k] and the back [k] being contrastive
in English—the difference between them is not one that can potentially
distinguish words. (The difference between, say, keep and coop depends on
the vowel contrast between [i] and [u], not on the consonantal contrast
between [k] and [k].) In literary Arabic, on the other hand, sounds like [k]
and [k] belong to different phonemes, usually transcribed /k/ and /q/ respec¬
tively, and words may be distinguished by this consonantal difference alone.
In English, the choice between [k] and [k] is predictable from their
phonetic environment—when /k/ occurs before a front vowel, it is realized
as [It], but when it occurs before a back vowel it is realized as [k]. Before a
central vowel, and finally after any vowel, we get a middle quality, [k]. We
say that these allophones of /k/ are in complementary distribution—where
one allophone occurs, another can’t.

Check that you understand the notion of complementary distribu¬

tion. Discuss whether the following pairs of sounds are comple-
mentarily distributed in your accent of English—

[d] and [d]

[n] and [q]
[0] and [6]

The existence of relevant minimal pairs is proof, of course, that two sounds
are not complementarily distributed.
Where the use of one or another allophone depends upon the phonetic
context in which it occurs, we say that the allophones are conditioned by
the phonetic context. Thus the allophone [k] is conditioned by a following
front vowel, while for the allophone [k] the conditioning factor is a following
back vowel. Among the allophones realizing the phoneme /l/, clear [1] is
conditioned by the phonetic context “preceding a vowel or /j/”, and dark [I]
by the context “preceding a consonant or pause”. In such cases we speak of
conditioned variation between the allophones concerned.

FREE VARIATION
Sometimes, though, the choice between allophones is free in certain contexts
—one or another may occur randomly, and a speaker repeating the same
word may use sometimes one, sometimes another without any apparent
system. This is called free variation and the allophones concerned are
termed free variants.
84
 Phonemes

There are occasional instances of free variation in English. Consider the

/t/ at the end of a word—wait /weit/, for example. If the word is repeated
several times in isolation (to keep the phonetic context unchanged), most
people vary freely between aspirated [th], unaspirated [t=], and unreleased
[t(Many use glottalized [?t] and even ejective [t’] sometimes, too, but
this does not affect the point at issue.) So these /t/ allophones are in free
variation in this particular phonetic context. It is just a matter of luck which
of them is used on a given occasion.

Say Wait! several times and find out whether you have more than
one allophone in free variation for the final /1/.

Note that in other phonetic contexts the variation between [th] and [t=]
is not free but conditioned—e.g. initially in a stressed syllable we get [th],
as in /teik/ take, but after /s/ we get [t=], as in /steik/ stake, steak.
Some people also have free variation between [r] and [j] for intervocalic
/r/ (see Chapter 22).

PHONETIC SIMILARITY
In English, [g] occurs only finally and before consonants, whereas [h] occurs
only initially and before vowels. The two sounds are therefore in comple¬
mentary distribution. So ought we to regard them as allophones of a
single phoneme, the variation between them being conditioned by phonetic
context ?
As native speakers of English we feel intuitively that to call [g] and [h]
“the same sound” would be absurd. The theoretical justification for keeping
them in separate phonemes is that they do not satisfy the requirement of
phonetic similarity. If a given pair of sounds are in complementary distri¬
bution or free variation, we do not identify them as members of the same
phoneme unless they have the majority of phonetic features in common.
As [g] is voiced, velar, and nasal, while [h] is voiceless, glottal, and fricative,
they lack the phonetic similarity which would be necessary for them to belong
to the same phoneme. The sounds [1] and [I], on the other hand, share the
features voiced, alveolar, lateral, and non-fricative. They are phonetically
similar, and we regard them as allophones of the same phoneme. Similarly
[th], [t=], [f ], and [t] all share the features voiceless, fortis, coronal (made
with tongue tip or blade) and plosive; ft], [k], and [k] are all voiceless, fortis,
dorsal, and plosive.
Forgetting for the moment about the final free variation mentioned above,
we can observe that [t11] is in complementary distribution not only with [t=]
but also with [p=] and [k=]. The justification for classing it phonemically
with [t=] (as we obviously must) rather than with one of the others is that
only [t=] has the necessary phonetic similarity with [th].
85
Phonemes

Go back to the “t-sounds” mentioned in the second paragraph of

this chapter. Explain why we regard these articulatorily different
sounds as allophones of a single phoneme /t/.
Analyse and discuss the following definition of a phoneme, which
is that given by Daniel Jones in his work The Phoneme: its Nature
and Use (§. 31). “A family of sounds in a given language which are
related in character and are used in such a way that no one member
ever occurs in a word in the same phonetic context as any other
member.”
If time permits, study the treatment of phonemic analysis in some
of the books mentioned in Appendix 2.

REVISION
1. Check that you understand the significance of enclosing phonetic
symbols between (i) diagonals / / and (ii) square brackets [ ]. (Remember
that this notation is merely a convention. The same distinction could
conceivably be shown by writing one set of symbols in pink and another
in yellow.)
2. Check your understanding of the terms in small capitals: allophones
of a given phoneme show phonetic similarity. They are arranged in
complementary distribution—that is, are conditioned by their phonetic
context—or, in some cases, free variation.

TEST
1. Which of the following pairs are contrastive in English ?

'rein—'bein 'hAfi— 'hAji 'kaedi—'kaeji 'path—'pat=

2. Is [ij] a member of the English /n/ phoneme ? Give reasons for your
answer.
3. What is the phonemic status of [?] in English?
4. Describe the various lateral articulations likely to be used in pro¬
nouncing the phrase Lil climbed stealthily. What reasons are there for
classing them together as one phoneme and transcribing them by a single
symbol ?

TRANSCRIPTION
Frustration is a burst hot-water bottle, or loathing every moment of a holiday
you’re paying a fortune for. It's using the wrong side of the Sellotape,forgetting
what you were going to say, or locking yourself out. Frustration is other people
parking in front of your garage, or a stranger reading a riveting letter on the
bus and turning over before you get to the bottom of the page.

86
21 More about Articulation
and Intonation

We have already seen that [0] differs from [s] in its place of articulation: it
is dental as against alveolar. But there is another important difference
between these two sounds: the shape of the tongue. For [0] the tongue is
relatively flat, but for [s] it is grooved fore-and-aft along the median line.

Have a look at the shape of your tongue (or your neighbour’s) in

the production of first [s] and then [0]. Experiment with different
kinds of tongue shape.

In the case of [s] the air-stream escapes along a groove, but for [0] it
escapes through a slit. So we can call [0] a voiceless dental slit fricative
and [s] a voiceless alveolar groove fricative.

What is the phonetic symbol for a voiced dental slit fricative? a

voiced alveolar groove fricative ? Now try and combine the grooved
tongue shape characteristic of [s] with the dental place of articulation
characteristic of [0]. This should give a voiceless dental groove
fricative, symbol [§]. One kind of lisp involves the use of this
sound and its voiced counterpart [z].

EAR-TRAINING and Sound Production

Recognize and make words such as six, sack, bus, snake, with each
of the following sounds substituted for [s]: [0, §, x, ?, 9, J, $].
Discuss similar substitutions for [z] in words such as rose, zoo,
busy.

The sound [J], like [s], is a groove fricative; but the groove for [J] is not
so narrow as that for [s]. It also extends further back along the surface of
the tongue. The place of articulation for [J] is in fact classified as palato-
alveolar, since it involves narrowing both between the front of the tongue
and the hard palate and also between the blade of the tongue and the
alveolar ridge.

Make [J] and observe its articulation. Alternate [J] with [s] and feel
the difference in tongue shape and tongue position between them.
Are there any other differences? Is it possible to pronounce either
of the two sounds with the jaw lowered? If not, why not?
87
 More about Articulation and Intonation
Learn off by heart the following labels: [s] and [z] are alveolar
fricatives; [J] and [3] are palato-alveolar fricatives; [tj] and
[d3] are palato-alveolar affricates.

DENTAL, ALVEOLAR, RETROFLEX

We have already seen (Chapter 4) that English /t/ /d/ and /n/ are usually
realized as alveolar, [t, d, n], although before the dentals /©/ and /6/ they
are often dental—narrow symbols ft, d, n].

You can think of the diacritic mark underneath as a small tooth to

show dentality.
Make ft], fti], and ft}] and alternate them with the corresponding
alveolars. Learn to recognize them in ear-training—note how
English sounds slightly “foreign” if one uses dentals instead of
alveolars.

Going in the opposite direction from an alveolar starting-point, we can

make retroflex sounds by curling the tip of the tongue backwards and
articulating with it in this shape against the alveolar ridge or anterior part
of the hard palate. Plosives, nasals, and laterals can be produced in this
way: ft, 4,n, l].
Alternate retroflexes and alveolars.
EAR-TRAINING or Sound Production Practice
'taen 'taendi'naenaen taendi'nam 'ftenaen taeqcjma tsendi'rjam tae'narei
tseii'ijita taendi'rjam 'taenta'rata tsen’tjina taen'dita
'taenaep tae'narei taeq'krirja taen'rautu taeo'rita taifta'rapi taen'rautu
taenta'rota taenta'ram taenta'rani taen'rita taen'rfta

INTONATION: PLACE OF NUCLEUS

We have studied various intonations—pitch patterns of falls and rises—but
we have not yet really seen how they fit into sentences in English or how they
are used to convey meanings.
Let us take the sentence It was a remarkably silly idea. One way of saying
it is with a falling tone on the last word—

• • •
• • •
9 0 0 0

it waz a ri 'makabli 'sill ai 'dia

This is quite a natural way of saying it, the way you might reply if someone
said to you “What did you think of. . . . ?”
88
 More about Articulation and Intonation

But if, on the other hand, you were replying to the question Was it a
sensible idea ?, you would be more likely to say it with a fall on silly—

• • •
• • • •*\ • • m
it W3Z a ri'makabh 'sill ai'dia

This seems to put more emphasis on silly, reducing the emphasis on idea,
which is a word simply repeated from the question.
And then again if you were replying to the question Do you think it was
a silly idea ? you might well say

. . .. N.
it waz 3 ri'makabli 1 sili ai 'di3

—putting the fall on to mak, the stressed syllable of remarkably. This puts
all the emphasis on remarkably, taking emphasis away from the words silly
and idea, which are both now simply repeated from the question. (Note we
have kept stress constant while altering the intonation.)
Now we can do precisely the same kind of thing with rises, fall-rises, or
any other of the tunes we have dealt with so far.

Try saying the same sentence with a high rise on each in turn of
idea, silly, and remarkably.

The word on whose stressed syllable the fall, rise, etc., occurs is thereby
thrown into emphatic relief. We say that it (or its stressed syllable) carries
the nucleus of the intonation group. The tone which occurs at the nucleus
is called nuclear tone. The place of the nucleus in an intonation group is
sometimes called its tonicity (from the term tonic, which is a synonym of
nucleus).
Look back at the intonation patterns discussed above. Identify the
nucleus in each case. (The nuclear tone was always a fall.)
Take some English sentences such as the following and vary their
tonicity in as many ways as possible. Do not worry about the actual
tones used, but just about the varying the place of the nucleus.

What did you say his name was ?

I flew to Paris last Thursday.
Why don’t you relax ?
I suppose it's a joke.
Don't rock the boat.
89
More about Articulation and Intonation

EAR-TRAINING
See if you can identify the place of the nucleus in each intonation
group of the following passage of English (to be dictated).

6a 'britij 'meil, wia 'tauld, iz az 'proud az a 'pikok about iz 'klaudz—

and az 'kAlafJ az wad. hiz 'tsnd iz 'wsl 't3nd 'back on 'draebms, and
iz 'avri 'bit az Tasi az iz 'waif a 'g3l frcnd about iz 'drcs. 'Jeiplis,
'nondiskript 'gamants a 'not fa 'him.
it 'al 'saundz 'vcri kan'vinsig, bat 'bAd3 an 'mtj fram 'knjz 'raud a
'kanabi 'strit an 6i i'Iu3n 'sAfakeits in 'eikaz av 'il 'Jeipan 'klD0,
'raept around 'nuljanz av 'men m 6a 'breiv 'neim av 'klaudig.
6a 'pikok, it 'simz, az 'flaun, ar az 'daid iz 'fadaz 'grei ta iuk laik
its 'neiba.

90
22 English /r/

EAR-TRAINING
Nonsense words—

fuadcin <|ajI Jsi0s nwigofl

The commonest allophone of English /r/ is a voiced post-alveolar fric¬

tionless continuant, narrowly symbolized [i]. It is the sound commonly
used at the beginning of a word such as /red/ red, and also in words such as
/brig/ bring, /a'gri/ agree.

Say /red/ over several times and isolate the first segment, [j]. Try
and feel where your tongue is for this sound.

This sound [j] is termed “post-alveolar” because it is made with the

tongue tip just behind the alveolar ridge. It is termed a “frictionless continu¬
ant” because it is made with the organs of speech in a position similar to
that which would produce a fricative, except that the articulating organs are
not quite close enough together to cause friction. In the case of [j], the air-
stream can escape between the tongue tip and the rear of the alveolar ridge
without any friction. (The term approximant is an alternative to frictionless
continuant.”)

Many people labialize initial /r/, as discussed in Chapter 9.

After initial /p/, /t/, /k/ in a stressed syllable, most people use a voiceless
post-alveolar fricative, narrow symbol [j]. This occurs in words such as
/preiz/, /kroon/. Moreover, /t/ combines with a following /r/, the two being
realized together as a voiceless post-alveolar affricate ft*]. Examples of
this are found in /trai/, /tru/, /'kxntn/.
91
English /r/

Say aloud some words beginning with /pr/ and /kr/; try and isolate
the [j] segment. Then say some words beginning with /tr/, and
isolate the initial affricate [Jj], noting that the /t/ and the /r/ have
in this case the same place of articulation.

Similarly, /d/ combines with a following /r/ to give a voiced (or partially
voiced—see Chapter 11) post-alveolar affricate, [fli], as in /drai/, /a'drss/,
/'londn/.

Say some words with /dr/, isolating the affricate [$}j] and feeling
how it is articulated.

(The marks underneath the symbols [4] and [j] denote respectively “closer
articulation,” i.e. fricative, and “opener articulation,” i.e. frictionless con¬
tinuant. The symbol [4] without any diacritic mark can be used when the
distinction between fricative and frictionless continuant is not relevant to the
point under discussion.)
Between vowels, as in /’kaeri/, /*sorau/, most people use ordinary [j], the
voiced post-alveolar frictionless continuant. But others use a voiced
alveolar tap, [r]. A tap is like a roll (Chapter 7), except that there is only
a single touch of one articulator against the other instead of a whole series
of touches. This voiced alveolar tap, [r], is sometimes called “flapped r”;
there are some speakers of English, e.g. many South Africans, who use it
for /r/ in all positions.

Pronounce [r] and alternate it with [j] until you can produce either
at will. Do you ever use [r] in your ordinary pronunciation? (Check
for its occurrence particularly between vowels and after dental
fricatives.) For practice, say the following words with the /r/ realized
first as [j] and then as [r] in each case—
right, arrange, ferry, bright, draw.

There are several other sounds which, while not properly members of the
SBS /r/ phoneme, are generally thought of as “r-sounds.”
The voiced alveolar roll, [r], was mentioned briefly in Chapter 7. It is
popularly supposed to be used in Scottish speech: most Scots, however, use
not [r] but a voiced post-alveolar fricative, [j]. The roll is, however,
occasionally used by people in Scotland and elsewhere for special effect, e.g!
for declaiming or for extra clarity. And it is used in Welsh and in many
foreign languages, e.g. Spanish, Italian, and Russian.

Make a rolled [r] if you can. Try saying various English words with
/r/ realized as [r]. Is [r] in some sense the correct way to pronounce
/r/? Would you advise a foreign learner of English to pronounce
M as [r]?
92
 English frl

If you can make a satisfactory [r], devoice it to get a voiceless

alveolar roll, ft]. This is the pronunciation of the Welsh rh.

Some English people use a rather different kind of realization for /r/—a
voiced labiodental frictionless continuant, [u]. (It is also characteristic
of the speech of New York City.) It is like [v], except that the contact of the
lower lip with the front upper teeth is not firm enough to bring about friction
as the outgoing air passes through: one can think of [o] as a very weak
[v]. Used for /r/, it is usually considered defective.

Pronounce [v>] on its own and between vowels. Then use it for /r/
in the words right, arrange, ferry, bright, draw. Discuss whether this
type of pronunciation should be regarded as defective, and if so
why. What happens to the opposition between /v/ and /r/ ? Which
well-known public figures use [o] ?

Other realizations of /r/ occasionally encountered are the voiced uvular

fricative, [k], and the corresponding frictionless continuant. These are
often heard from French and German people speaking English. They are
also used by some Northumbrians and North Welsh. The fricative [k] is
similar to the voiced velar fricative [y] mentioned above (Chapter 7), but
articulated further back, by the extreme back of the tongue against the
uvula.

Pronounce fricative and frictionless [b]. Use them for /r/ in some
English words. Does this give the effect of a foreign accent?

The same uvular place of articulation characterizes the voiced uvular

roll, [r]. This is made by arranging the tongue and uvula so that the uvula
vibrates against the extreme back of the tongue in the outflowing air-stream.

Pronounce [r] if you can. Use a mirror to check what is happening

in your mouth. Put [r] in some words. Then devoice [r], giving a
voiceless uvular roll, 0$]. Try whistling while articulating a long
[r] : this gives a reinforcing effect like a referee’s whistle with a pea
in it.

One last kind of exotic r-sound is the voiced retroflex flap, [y]. For this
sound the tongue starts in a retroflexed position, the tip curled back but not
touching anything; the tip is then thrown forward and down, flapping once
against the alveolar ridge on the way past and ending up behind the lower
teeth. This retroflex flap is used in various languages of India and Pakistan,
93
 English //•/

and sometimes also in the English spoken by those who have [j] in their
mother tongue.
Note among the symbols for r-sounds that those which are ordinary letters
the right way up denote rolls (r, r), while those which are inverted denote
fricatives or frictionless continuants (4, k). Those using the shape of the lower¬
case letter r denote (post-)alveolars (r, 1), whereas those using the shape of
the capital (or its mirror-image) denote uvulars (r, k).

POTENTIAL /r/ IN ENGLISH

In SBS and most other accents of England and Wales, no word in isolation
ends in /r/. So we have, for example, in spite of the spelling, star /sta/,
four /fa/, near /nia/. Nor is any /r/ pronounced in these words when they
precede a consonant: starlight, star performance, foursome, four people,
nearly, a near thing; but /r/ may be pronounced if a vowel follows, as in
starring, star attraction, four-ish, four apples, nearest, Near East. The jrj used
in these latter cases is called linking /r/. Alternatively, instead of a linking
M, speakers sometimes use a glottal stop, [?], or sometimes nothing—but
within a single word linking /r/ is nearly universal.

Say the phrase four apples with

(i) linking /r/
(ii) linking [?]
(iii) zero link.

Check that you can recognize these various possibilities with this
and other similar phrases.

Linking /r/ is used with words ending in any of the vowels /a, 0, 3, a, w,
S3, 03/ and, for those who use it, /m/.

Think of two examples of phrases with linking /r/ for each of the
vowels mentioned.

Historically speaking, linking /r/ is all that is left of a final /r/ that was once
pronounced whatever followed—as in present-day Western, Scots, Irish and
American accents. But accents like SBS (the non-rhotic accents of English)
lost /r/ before consonants and pause two centuries ago: when start became
/stat/ rather than /start/, star became /sta/, the older form /star/ being
retained only where a vowel followed, i.e. as a variant with linking /r/.
This had an odd consequence. When final /r/ was lost, new rhymes arose:
star—Shah, four—law, tender—Brenda and so on. Because the first in each
pair was subject to linking /r/ when a vowel followed, the second came by
analogy to be subject to it, too. A linking /r/ is usual in a phrase like the star
94
 English /r/

of Persia: hence /r/ has come to be very common in the phrase the Shah of
Persia /5a 'Jar av 'p3ja/. Similar intrusive /r/’s may be heard in the phrases
law and order, here's Brenda again. The jxj is called intrusive in these cases
since it was not present historically, is not in general used in rhotic accents,
and does not correspond to any r in the spelling.
Apart from considerations of spelling, though, linking /r/ and intrusive
/r/ are essentially the same phenomenon in present-day SBS. They can be
regarded as two forms of potential /r/, the difference between them being
not strictly speaking a phonetic one. Hence it happens that people who try
to eliminate intrusive /r/ from their speech, having been told that it is
slovenly or vulgar, can often only do so at the expense of eliminating linking
/r/’s too—between words and even sometimes within them.

Think up two further instances of intrusive /r/ after each of the

following: /a, ia, a, a/.
Discuss whether people are justified in regarding intrusive /r/ as
slovenly and vulgar.
Why is intrusive /r/ (but not linking /r/) virtually non-existent
after /ea/ and /ua/ and completely non-existent after /3/?

INTONATION
One more nuclear tone used in English is the rise-fall—

# • « * * . «

- Oh! De*lightful It "as his * mother-in-law

Practise making and recognizing rise-falls. Distinguish them care¬

fully from fall-rises.
How would you describe the distinctive meaning(s) of the rise-
fall?

If you can identify the nucleus of each intonation group and its nuclear
tone (high fall, low fall, high rise, low rise, fall-rise, rise-fall), you have
mastered the description of the most important features of English intona¬
tion.
The pitch features of pre-nuclear syllables are less important. Any un¬
stressed syllables at the beginning of an intonation-group constitute the
pre-head, which may be low or high. Apart from the pre-head, if one is
present, any syllables before the nucleus constitute the head. Pitch patterns
of heads include high level, downward “stepping,” downward “sliding,
upward “bouncing,” and low level.
95
 English /r/

TRANSCRIPTION
Use the allophonic symbols discussed in this chapter in transcribing
the following passage—
How do you recognize a bargain when you meet one ? Right now,
take the packet on my kitchen shelf, currently distracting me while I
cook. It contains a detergent that recently appeared on the market,
and squint-wise across the centre of the packet, is a label proclaiming
3 p off the recommended price. The immediate mental picture is of
panic in the factory—Save the customer and reduce the price! Some¬
body's recommended the wrong one! And the factory leaps into action
as it relabels every box. But the law of selling isn't a bit like that.
That label is an integral part of the original package; so who recom¬
mended that 3 p too much in the first place ? See the smart supermarket
operator arranging a notice: A free balloon and 3 p plus the original
3 p off. Then running off round the corner to see what can be marked
up.

96
23 More about Vowels

CARDINAL VOWELS
If we consider the vowel sounds that are found in various languages or
dialects and try to classify and analyse them, we run up against more diffi¬
culties than we do with consonant sounds. Unlike most consonants, vowels
do not have an easily perceptible place and manner of articulation. In fact,
vowel sounds (or more precisely vocoids, defined as sounds in whose produc¬
tion the air-stream escapes through the mouth over the median line of the
tongue with no obstruction in the mouth to its free flow) are most easily
recognized and classified by ear—auditorily.
But as we saw in the chapter about phonemes, everybody tends to hear
sounds in terms of the phonemic system of his own language. So phonetically
naive people tend to hear all vocoids as related to the vowel sounds phonemic
in their language. Since different languages have different vowel systems, this
leads to chaos if we attempt to classify all vocoids by reference to any given
language.
If we were to describe a given sound by saying “it is slightly closer than
the /e/ in get", this would not be very clear. After all, people vary considerably
in how they pronounce English /e/—Australians make it relatively close,
Northerners relatively open, and even people using an SBS accent do not
all make it precisely the same.
So what we need is a neutral classification system for vocoids—a system
independent of any particular language. Such a system was devised by
Daniel Jones, and is widely used by phoneticians. It is known as the cardinal
vowel system. The cardinal vowels are particular vocoid qualities selected
as reference points for the description of the vowels of different languages
and dialects, and have been recorded (e.g. on Linguaphone ENG 252-3).
But the best way to learn them is from a teacher who knows them.
Cardinal number 1 is the closest and frontest vocoid that can be made.
The raising of the tongue is as far forward as possible and as high as possible
consistent with the sound being a vocoid, i.e. with the avoidance of friction.
The lips are spread, and the teeth are close together, almost touching. The
phonetic symbol is [i]. (Many people find it convenient to distinguish cardinal
vowels symbols from those for non-cardinal vowels, e.g. English vowels, by
underlining them to show cardinal quality, thus [i]. But this usage, though
convenient, does not have the approval of the International Phonetic
Association.)
97
 More about Vowels

Learn to recognize and make Cardinal 1, [i]. Compare it with

English [i].

The opposite extreme of the vocoid area is represented by Cardinal

number 5, [a], which is the openest and backest vocoid that can be made.
The back of the tongue is lowered as far as possible and retracted as far as
possible consistently with the sound being a vocoid; the lips are not rounded
and the jaw is fully open.

Learn to recognize and make Cardinal 5, [a]. Compare it with

English [a].

Cardinals number 2, 3, and 4 (symbols [e], [s], [a]) are front vowels falling
between [i] and [a], chosen in such a way that the intervals between adjacent
cardinal vowels form a series of auditorily equal steps.

Learn to recognize and make Cardinals 2, 3, and 4—[e, e, a]. Com¬

pare them with English [i, ei, e, ea, x, a].

Cardinals 6, 7, and 8 (symbols [a], [o], [u]) are back vowels with lip¬
rounding. They are chosen so as to continue the series of auditorily equi¬
distant steps [i—e—s—a—a—a—o—u]. The lips are open-rounded for [a],
close rounded for [of, and closer rounded still for [u].

Learn to make and recognize Cardinals 6, 7, and 8—[a, o, u]. Com¬

pare them with English [o, a, au, u, u].

The eight primarycardinal vowels dealt with so far are the basis of the
VOWEL CHART diagram often used in phonetics works (Chapter 13). It repre¬
sents the presumed position of the highest point of the tongue in the mouth:
in fact it is a schematization of the locus of the highest point of the tongue
(although a point is generally placed on it by auditory rather than articulatory
identification).

Vowels
i u

98
 More about Vowels

Study the following diagrams showing English (SBS) typical vowel

qualities placed on the vowel chart.

Monophthongs
u

Closing diphthongs Centring diphthongs

Monophthongs are vocoids during the production of which the tongue

and lips stay in the same position, so that the resultant sound does not
change; diphthongs are vocoids during the production of which the tongue
and/or lips move, so that the resultant sound changes as it proceeds (Chapter
13). Among the English vowels shown on the chart of monophthongs, /i/
and /a/ are often actually rather diphthongal, [ii, uu], particularly when
stressed and final: sea, two.
Diphthongs are termed closing if their second element involves a move
to a closer tongue position; they are termed centring if their second element
involves a move to a central position. All the English diphthongs listed are
diminuendo (or falling) in that their second element is less prominent than
their first. Sometimes the /»/ sequence in happier /'haepia/, etc., and the /ua/
sequence in influence /'mfluans/, etc., are analysed as crescendo (or rising)
diphthongs, since their second elements have greater prominence than their
first.

ENGLISH SEMIVOWELS
The English palatal semivowel /j/ as in yes /jes/, and labio-velar semi¬
vowel /w/, as in wet /wat/ are usually vocoids, but very short and gliding
99
 More about Vowels

ones. The commonest allophone of /j/ is a relatively front close unrounded

vocoid, [i] or [I] (the mark “ shows that the sound concerned is non-syllabic).
In /jes/ the front of tongue glides from a close to a nearly half-open position,
[ie]. Similarly, the commonest allophone of /w/ is a relatively back close
rounded vocoid, [u] or [u]. In what /wot/ the back of the tongue glides from
a close to a nearly open position, while the lips move from close-rounded to
open-rounded, [ud]. In /w£t/ the tongue glides from having the back highest,
from close to nearly half-open; the lips move from close-rounded to spread.

WD

Occasionally, though, /j/ and /w/ have fricative allophones—notably after

/p, t, k/,and perhaps after other voiceless consonants. In words such as
pure Ipjua/, queue /kju/, /j/ is realized as a voiceless palatal fricative, [5].
(And /k/ before /j/ becomes more or less palatal too so that /kj/ is pronounced
[k?] or [c?].)

Say over pure, queue, and other words with sequences /pj, kj/. Feel
the articulation involved for /j/. Try and isolate the [9] segment.
What part of the tongue is it made with ?
Then try tune and Tuesday. Do you say these with /tj/ or with
/t]/ ? If the latter, how is the /j/ realized ? Possibly as [9] (see next
chapter).

The sequence /hj/, as in huge, may be realized by a single segment [9], or

by a less fricative sequence.

Isolate the first sound in your pronunciation of human.

In words such as twin /twin/, quite the /w/ is realized as a voiceless

/kwait/,
labio-velar fricative, [m], articulated by the back of the tongue near the
velum and with the lips closely rounded.

Say these and other words with /tw/, /kw/. Try to isolate the [m]
segment.

We have seen that /j/ and /w/ are typically vocoid—like vowels. So the
question arises, what is special about them that makes us separate them
from vowels and put them in a distinct category called “semi-vowels” ?
100
 More about Vowels

It is a matter of their place in the structure of the syllable. An English

syllable has the structural formula
C30 V C*
i.e. from zero to three initial consonants, followed by a vowel (monophthong
or diphthong), followed by zero to four final consonants.
To avoid complications, let us consider just words of one syllable, with
the structure C V, i.e. one consonant plus one vowel. Examples are—
key /ki/
saw M
now /nau/
bar /bo/
foe /fe«/
day /dei/ etc.
If we take monosyllabic words of this type containing /)/ or /w/, we find
that the /j/ or /w/ can only come in the C place, never in the V place

you /ju/
we /wi /
year /j3/ or /jia/
war /w a/
why /wai /
We cannot have words of the type */bj/, */kw/ etc. In other words, semi¬
vowels BEHAVE AS consonants, even though articulatorily they are like
vowels (vocoid). We say that consonants, including semi-vowels, have
marginal syllabic function, whereas vowels have central syllabic function
(on the basis of CVC syllables).
Taking syllables of the structure CVC (e.g. bet /bet/), show how /j/
and /w/ behave as consonants rather than vowels. Do you notice
anything about (i) the vowels that occur in (C)VC monosyllables
but not in (C)V monosyllables ? (ii) the restriction of certain conso¬
nants to either initial or final C place ?
What place in syllable structure do syllabic consonants /l, n/
occupy? How does this compare with their articulatory classifica¬
tion? Is there a case for labelling them semi-consonants?

DURATION OF ENGLISH VOWELS

The English monophthongs can be divided into the short monophthongs,
h £ se, A, d, u, a/, and the long monophthongs, /i, a, o, u, 3/. Other things
being equal,* the long vowels have greater duration than the short ones.
The diphthongs are comparable in duration to the long monophthongs.
To check this, compare bid with bead, nod with gnawed, etc.
101
 More about Vowels

But the realization of all vowels varies considerably in duration according

to phonetic context. The long vowels and diphthongs are particularly
affected by phonetic context. The most important factors are—
(1) A following fortis consonant—one of /p t k tj f 0 s J/.
(2) A following unstressed syllable, or several such syllables, within the
same word or stress group.
(3) Occurrence in an unstressed syllable immediately followed by a
strongly stressed one.

Each of these contexts has the effect of shortening the duration of a

vowel, though without changing its quality.

Compare the j\j in see, need (fully long), with those in (1) seat,
(2) cedar, (3) Seattle. Which of the three shortening contexts has the
strongest shortening effect ?
Discuss the duration of /a/ allophones in law, lawn, talk, order,
austere.
Think up sets of words to illustrate variations of duration for /u/
and /ai/.

EAR-TRAINING
Nonsense words—

'jwaOjaz 'pdi3 vi'vsde jok ‘hja^uv

Transcribe the following passage from dictation—

'ai 'zili 3'hei
'zan a'sksl
wilfn 'bavaes
'biven o'det 'bavm
‘biban 'trig 'buban
'tizuk ka’levajina 'tezik,
sti’fok ka'pruz ste'fak
ka’bcsta 'spo^el sa'bita.

When you have checked the accuracy of your phonetic transcrip¬

tion of the following passage, underline short vowels in red, long
vowels in another colour; and double the line for those whose
phonetic context gives them a long duration.

in an ‘aidial 'w3ld, dsa wud bi a 'greit 'nAmbar av ‘difrant 'sats av

skul. 'ai wud bi 'fri fa 'Evribndi, am ‘pcarants kud 'tjuz wot waz 'best
fa dsa 'tjildran widaut kansida'reifn av ‘nuni, 'distrikt, 'klas a 'graft,
wi wil ‘nau baev *tu mmits 'sailans ta 'knntempleit di An'laikli
'prnspekt.
102
24 Other Sound-types

In this chapter we shall look briefly at quite a large number of sound-types,

some of them rather exotic. You will probably find this concentrated diet
quite indigestible if you try to swallow everything at once. To assimilate the
material thoroughly, you will need to do a lot of regular practice in—

(1) ear-training—recognizing different sound-types when produced by

the teacher or another student;
(2) description—identifying the articulation used to produce a given
sound-type;
(3) sound production—making the sounds yourself.

1. PALATAL, VELAR, UVULAR

Raise the back of your tongue to form a [k] closure against the
velum. Note where the tip of your tongue is—behind the lower
teeth, out of the way. Keep the tip down while feeling forward from
the velum until the front of the tongue (not the blade or tip) is
articulating against the hard palate. Produce a voiceless palatal
plosive, [c]. Say it on its own and between vowels.
Add voicing to produce a voiced palatal plosive, [j]. Alternate
[jEce], etc.
[caja],
Go back to the [k] position (voiceless velar plosive). Now feel
backwards and downwards with the back of the tongue, moving it
down along the velum until the extreme back of the tongue articu¬
lates with the extreme end of the soft palate and the uvula. Try to
make a voiceless uvular plosive, [q]. Repeat it a few times and
put it between vowels.
Add voicing to produce a voiced uvular plosive, [g]. Compare
and contrast [c, j], [k, g], [q, g].
You should have no difficulty in producing a palatal nasal [ji]
or a uvular nasal [n]. The first is articulated just like [j], except
that the soft palate is down; the second has the same relationship
to [g].
Keep the tip of your tongue out of the way for [ji]: compare it
with the sequence [nj], as in English onion. Say the French word
montagne [m5'(aji].
103
Other Sound-types
If you can make a lateral at this place of articulation, it will be a
voiced palatal lateral, [X]. Note its articulation by the front of
the tongue, as compared with the tip (or blade) articulation for pj]
as in English million.
Repeat the velar fricatives [x, \] learnt earlier. Now produce
palatal and uvular fricatives by moving the body of the tongue
forward and back, as just practised for the palatal and uvular
plosives. With the front of the tongue raised close to the hard
palate, make a voiceless palatal fricative [9] (Chapter 23). Note
its relationship to Cardinal 1, [i]—how do they differ articulatorily ?
Alternate [91919!]. The voiced palatal fricative [2] is even more
similar to [i]". (In this book we use the symbol [J] for the fricative,
to distinguish it clearly from the frictionless continuant or semi¬
vowel [j]. The IP A alphabet uses [j] in both senses.)
Sliding the tongue back from [x] gives a voiceless uvular
FRICATIVE, [/].
Glide from palatal through velar to uvular fricative, [9—x—/],
and back again. Note how the pitch changes from high (palatal) to
low (uvular).
Produce the voiced counterpart of [/], the voiced uvular frica¬
tive [h] (Chapter 22). The French r is one or other of these: try the
following French words—

route
quatre [kaft]
le roi [la 'kwo]
lettre
arriver [airi've]

Then try some German words with [9] and [x] or [/]—

nicht [ni9t]
Achtung [’axtuij]
Milch [mil9] (clear [1]!)
Buck [bux]

Children with pronunciation difficulties (dyslalia) often replace

alveolar fricatives by palatal, velar, uvular, or other fricatives.
Practise and write down some examples of this, using appropriate
phonetic symbols.

2. MORE FRICATIVES
Make a voiceless alveolar lateral (Chapter 3) before and after
vowels, e.g. [ejs], [aja], [aja]. Then add friction between the sides
104
 Other Sound-types

of the tongue and the teeth. This gives a voiceless alveolar

LATERAL FRICATIVE, [4].
Say [e4e], [ala], [o4a], etc.
[1] is the sound of Welsh //. Pronounce Llanelli [4a'ne4i], cyllell
(“knife”) ['kalsl], being careful to avoid the common English
mistake of saying [0J] or [xj] instead of [1].
Some people who lisp use [1] for [s]. Pronounce some words
lisping in this way—e.g. see, six, voice, assist.
Add voice to [4] to get the voiced alveolar lateral fricative,
[!}]. Lisp by using it for [z] in words such as easy, noises, zinc.
Using the same kind of tongue position as was used for [t, R],
produce voiceless and voiced retroflex fricatives, [§] and fc].
Practise them before and after vowels.
Compare [§] and fa] with the voiceless and voiced palato-
alveolar fricatives, [J] and [3]. Note the difference in the pitch of
resonance: [§] has a lower pitch than [J]. Slur from [§] to [J], listening
to the gradual rise in the resonance of the friction.
Now try and continue this rise, while maintaining the production
of a voiceless [J]-type fricative. This should give a voiceless alveolo-
palatal fricative, [0]. Slur back from [0] through [f] to [§].
Make the voiced counterpart of [0], symbol [*]. Slur from [z]
through [3] to fc], and back again to [z]. Practise [0] and [z] before
and after vowels.
Note that these sounds may function as allophones of English /j/
after /t/ and /d/, as in tune, educate.
A voiceless glottal fricative, [h], is produced by passing a relatively
large flow of air through the parted vocal folds. On the other hand,
passing a large flow of air through vibrating vocal folds gives a
voiced glottal fricative, [fi]. Make this sound—be passionate,
sigh as you speak.
Try the voiceless pharyngeal fricative, [h]. This is articulated
by a constriction in the pharynx, between the root of the tongue and
the wall of the pharynx. Its voiced counterpart, [q], is often accom¬
panied by “creaky” voice.

3. MULTIPLE ARTICULATIONS
It is quite easy to pronounce [f] and [s] simultaneously. The resultant [fs] is
said to have double articulation—labiodental and alveolar. Plosives and
nasals can be similarly double-articulated: e.g. [kp], [mg]. In these cases the
two places of articulation are of equal importance.
Sometimes, however, a sound-type with two places of articulation clearly
has one of them more important, one less important. In the case of dark [1],
for example (Chapter 19), the alveolar articulation is primary, the velar
105
Other Sound-types

articulation secondary (Chapter 9). A secondary articulation is defined as

one having a lesser degree of stricture than a simultaneous primary
articulation.'
Secondary articulations are given names ending in -ized and -ization. So
the proper name for [1] is a velarized voiced alveolar lateral. The /t/ in
twin, which has lip-rounding, [t], is a labialized voiceless alveolar plosive.
A consonant with secondary palatal articulation is termed palatalized,
e.g. [tj]—and so on.

4. SECONDARY CARDINAL VOWELS

Adding lip-rounding to cardinal vowels 1 to 5 gives the secondary cardinals
1 to 5, [y, 0, ,«e a:, o]. Spreading the lips while having the tongue position
for cardinals 6 to 8 gives the secondary cardinals 6 to 8, [a, », w].

Make [y] by rounding the lips while keeping the tongue position
for [i], Practise alternating [i—y—i—y], moving the lips only and
keeping the tongue steady.
Do similar exercises to derive [0] from [e] and [a] from [e].
Say French and German words with front rounded vowels of
this type, e.g.
[plym] plume “pen”
[jn] yeux “eyes”
[eef] oeuf “egg”;
['myda] miide “tired”
[fen] schon “beautiful”
['Poefnan] offnen “to open”
Make [in]—a close back unrounded vowel—by saying cardinal
[u] and then spreading the lips while keeping the same tongue
position. Alternate [u—in—u—in—u], moving the lips only and
keeping the tongue steady. Say some Japanese words: [mizui]
“water,” [suinmi] “live,” [([auity] “Fuji.”

Ordinary [w] has double articulation, since it is a voiced frictionless con¬

tinuant with both bilabial and velar articulations (for short, labio-velar
semivowel). Just as [w] corresponds in articulation to [u], so the semivowel
corresponding to [y] is [q], the voiced labio-palatal semivowel (i.e.
bilabial/palatal frictionless continuant).

Say the French words

[qij] huit “eight”
[sqi] suis “am”; and compare
[Iqi] lui “him” with [lwi] Louis.

106
 Other Sound-types

The voiceless counterpart of [w] which some people use for wh, [m], may
often have friction at the bilabial place but not at the velar, and thus be a
voiceless velarized bilabial fricative.

EAR-TRAINING
Nonsense words

’l0<iyq pseToe/ 'ciuixaj we’^esyc qa'sytaz

Try transcribing this passage of no known language from dicta¬

tion—

0 'hy 0'jo, in 'ca ^d'z«.

y'xs ce3y 'n0 sa ‘lono, e But v'^en 'zampm, y’nee 'dixs u'cuu a 'jiaix.

107
Appendix 1 Some English
Phonemes and their Allophones
This table is not exhaustive. In particular, it does not show glottalization,
which may occur in certain of the syllable-final contexts marked 1 below,
palatalization (before /j/), labialization (for /r/ etc. and before /w, u/
etc.), or different approach features of plosives. Nor is it possible to show
the accumulation of special features in certain cases: e.g. the ft/ in true is
commonly (i) post-alveolar, (ii) affricated, and (iii) labialized.

Feature Context Example

/P/ Voicing: voiceless all

Place: bilabial all, except:
labiodental before /f, v/ cupful
Manner: plosive all
Aspiration: aspirated [ph] initial in stressed
syllable park, appear
unaspirated [p=] after /s/ spark
medial, final1,2,3 happy, cap
Release oral all, except:
nasal [pm] before nasal1 topmost
non-audible [p1] before ft, d, k, g/1,2 apt
none [p~] before /p, b/1 chipboard
Fortis/lenis: fortis all

Itl Voicing: voiceless all4

Place: alveolar all, except
dental [$] before /0, 5/1,2 eighth, but then
post-alveolar [t] before /r/1 train, country
glottal [?] syllable-final2 at large
Manner: plosive all
Aspiration: aspirated [th] initial in stressed team, attack
syllable
unaspirated [t“] after /s/ steam
medial, final1,2,3 water, hot
Release: oral all, except:
nasal [t"] before nasal1 chutney, button
lateral [t*] before /l, J/1 Scotland, bottle
i,2,3.4, por notes see page m.

108
 Appendix 1

non-audible [t1] before/p, b, k, g/1* outcome

none [t~] before /t, d/1 hot dog
Fortis/lenis: fortis all

Voicing: voiced all, except:

partly or wholly next to fortis
voiceless [$] cons, or pause bedtime
Place: alveolar all, except:
dental [d] before /0, 6/2 width, said that
post-alveolar [d] before /r/ drain, sundry
Manner: plosive all
Release: oral all, except:
nasal [dn] before nasal midnight, sudden
lateral [d1] before /l, ]/ sadly, saddle
non-audible [d7] before /p, b, k, g/: broadcast
none [d"] before ft, d/ bedtime
Fortis/lenis: lenis all

Voicing: voiced all, except:

partly voiceless after fortis cons.2 snap
fe]
Place: alveolar all, except:
dental [n] before /0, d/2 tenth
post-alveolar [n] before /r, tr, dr/ sunrise, laundry
Manner: nasal all
Fortis/lenis: lenis all
{Note: /n/ is as In/, but of longer duration.)

/g/ Voicing: voiced all, except:

partly or wholly next to fortis outgrow
voiceless [g] cons, or pause
Place: velar all, except:
pre-velar [g] before front vowel
or/j/ geese
post-velar [g] before back vowel
or /w/ goose
Manner: plosive all
Release: oral all, except:
nasal [g9] before nasal ignore
non-audible [g1] before /p, b, t, d/2 lagged
none [g"] before /k, g/ big game
Fortis/lenis: lenis all
109
Appendix 1

/d3/ Voicing voiced all, except:

partly or wholly next to fortis
voiceless [d3, cons, or pause large size
$3>
Place: palato-alveolar all
Manner: affricate all
Fortis/lenis: lenis all

/z/ Voicing: voiced all, except:

partly or wholly next to fortis
voiceless [?] cons, or pause was called
Place: alveolar all
Manner: fricative all
Fortis/lenis: lenis all

/r/ Voicing: voiced all, except:

partly or wholly after fortis cons. train, cry
voiceless [j]
Place: post-alveolar5 all
Manner: frictionless
continuant [j] all, except:
fricative [j] after /p, t, d, k/ drink
Fortis/lenis: lenis all, except:
fortis after /t/ train

/!/ Voicing: voiced all, except:

partly or wholly after /p, k/ plane, clue
voiceless [|]
Place: alveolar all, except:
dental [I] before /0, 6/ wealth
post-alveolar [1] before /r, tr, dr/ already, poultry
Manner: lateral non- all (or some friction
fricative when voiceless)
Fortis/lenis: lenis all
(Note: for velarization—dark [i]—see Chapter 19.)

VOWELS (all are voiced and vocoid)

/u/ Place: nearly back [ii]
central [h] j new
back [u] 1 fool
Height: close all
Lips: lightly rounded all
110
 Appendix 1

Duration: relatively long all, except:

[u:]
shorter [u] before fortis cons.
in same syllable root
before unstressed
syllable in
same word ruder
unstressed unite

Place: central all

Height: between half-
close and half-
open all, except:
half-open final sofa
half-close before /k, g, g/ recognize
Lips: unrounded all
Duration: relatively short all

1*1 Place front all, except:

between front
and central [£] before [1] tell
Height: between half-
close and half-
open all
Lips: unrounded all
Duration: relatively short all

1 Possible glottalization.
2 Depending on speaker or style of speech.
3 Finally: free variation between aspirated and unaspirated.
4 Some speakers voice /t/ intervocalically, [t], e.g. better.
6 Some speakers use an alveolar tap [r] between vowels and/or
after /0, d/.

Ill
Appendix 2 Books for Further
Reading

Gimson, A. C. An Introduction to the Pronunciation of English. London:

Edward Arnold, 2nd edition, 1970.
Jones, Daniel. The Pronunciation of English. Cambridge University Press,
1966.

The above are the standard works on English phonetics. For general pho¬
netics, consult

Abercrombie, D. Elements of General Phonetics. Edinburgh University Press,

1967.

And for English intonation

O’Connor, J. D., and Arnold, G. F. Intonation of Colloquial English. London:

Longmans, 1961.

Halliday, M. A. K. Intonation of British English.

The standard English pronouncing dictionary (but note that it uses a different
transcription from the one used in this book) is

Jones, Daniel, revised by A. C. Gimson. Everyman’s English Pronouncing

Dictionary. London: Dent, 1967.

See also
Miller, G. M., BBC Pronouncing Dictionary of British Names. London:
Oxford University Press, 1971.

112
Index

affricate, 27, 91, 92 egressive, 2

air-stream mechanism, 1-4 ejective, 3
allophone, 81-6, 108-11 elision, 57-9
alveolar, 13, 88
alveolar ridge, 13
approach, 61, 62, 67-9 fall, falling tone, 45, 64, 70, 74
approximant, 91 falling diphthong, 99
articulation, 1, 87, 88 fall-rise, 79
aspiration, 76-8 feature, 85
assimilation, 53-9 flap, 92, 93
fortis, 41, 102, 108-10
free variation, 82, 84
fricative, 26, 27, 36, 55
back (of tongue), 13, 14, 49-51 frictionless continuant, 36, 91
back (vowel), 49-51, 106 front (of tongue), 13, 47—51
bilabial, 12, 26 front (vowel), 47-51, 106
blade (of tongue), 13, 49
broad (transcription), 82
glottal, 8
glottalic (air-stream mechanism), 4
cardinal vowels, 97, 98 glottalization, 85, 108
central (vowel), 48-51 glottis, 9
centre (of tongue), 48, 49 groove (fricative), 87
centring diphthong, 19, 99
clear (1), 77-82
hard palate, 13, 103
click 2 head, 95
close (vowel), 48, 50, 106
height of tongue, 103
closing diphthong, 19, 99 high fall, 64
cluster, 43 high rise, 64
coalescent, 55 historical elision, 57
complementary distribution, 82-4
hold (of plosive), 61, 67, 68
conditioned allophone, conditioned var¬ homonym, homophone, 10, 18-20, 31
iant, 84
contextual elision, 57
contrastive, 81, 82 implosive, 3
coronal, 85 incidence, 83
crescendo diphthong, 99 incomplete (plosion), 73
initiation, 1
ingressive, 2
dark (1), 77-82 intervocalic, 39
de-alveolar assimilation, 54 intonation, 45, 46, 63, 64, 70, 79, 88, 89,
dental, 13, 14, 21, 78, 88 95
devoiced, 39, 44 intrusive /r/, 95
diminuendo diphthong, 99
distribution, 95, 100, 101
dorsal, 85 jaw, 47
double articulation, 105
doubled consonant, 60, 61
duration, 101, 102 labialization, 32, 106
dyslalia, 104 labio-dental, 12, 35, 93
113
Index
labio-palatal, 106 primary articulation, 31, 32, 105
labio-velar, 99, 100 primary cardinal vowels, 98
larynx, 13 pulmonic (air-stream mechanism), 2, 4
lateral, 26, 77
lateral fricative, 105
lateral plosion, lateral release, 66-9 quality, 47, 77
length, 101, 102
lenis, 41, 108—10 regional accents, 6, 14, 15, 18, 19, 23, 24,
lingual roll, 26
32, 39, 83, 92, 93, 105
linguo-labial, 12
release, 61, 62, 67-9
linking /r/, 94
retroflex, 88, 93, 105
lip-rounding, 31, 32
rise, rising tone, 45, 64, 70, 74
lips, 31, 32, 50
rise-fall, 95
long (vowel), 101
rising diphthong, 99
low fall, 64, 95
roll, 26, 93
low rise, 64, 95
rounded, 31, 51, 98, 106
lungs, 1
RP, 6 (see SBS)

manner of articulation, 25-7, 35, 55 SBS, 6, 10, 11, 15, 18, 19, 24, 43, 67, 77,
marginal sounds, 34-7 83, 94, 97, 99
marginal syllabic function, 101 secondary articulation, 32, 105, 106
median, 69 secondary cardinal vowel, 106
minimal pair, 82 semi-vowel, 99, 100, 106
monophthong, 99 short (vowel), 101
sides (of tongue), 25, 66
slit fricative, 87
narrow (transcription), 82 soft palate, 13, 14, 25, 28-30, 69, 74
nasal, 25, 28, 68, 109, 110 spelling, 5
nasal approach, 62, 63 spread (lips), 50
nasal plosion, nasal release, 29, 62, 63 stress, 21, 89
nasalization, 28-30, 34 strong form, 22
non-audible release, 73 substitution, 35
nonsense words, 6, 29, 31, 34, 37, 38, 42, syllabic consonant, 22, 23
43, 47, 53, 57, 64, 66, 72, 91, 102, 107 syllable structure, 101
nuclear tone, 89, 95 symbol, 5, 17, 82
nucleus, 75, 88-90
tail, 70
oesophageal (air-stream mechanism), 4
open (vowel), 48, 50
opposition, 82 underdifferentiation, 83
oral (air-stream mechanism), 2, 4 unrounded, 51
oral (escape, release), 69 uvula, 103
organs of speech, 12-14 uvular, 93, 103
overlapping plosives, 72-5
variants (conditioned, free), 84
velar, 14, 27, 36, 103
palatal, 99, 103 velaric, 4
palatalization, 106, 108 velarization, 77, 106
palate, 13
velum, 14, 28-30, 69, 74
palato-alveolar, 36
vocal cords, vocal folds, 8
pharyng(e)al (air-stream mechanism), 3, vocoid, 97, 99
voice, voicing, 8, 35, 38-45, 55, 103
pharyng(e)al (consonant), 105 108-10
pharynx, 13, 25
voiced sounds, 8, 36, 39, 43
phoneme, 81-6
voiceless, 8, 36, 103, 108
phonetic similarity, 82, 85
voicing diagram, 38, 40-4, 76, 77
pitch, 45 (see intonation)
vowel, 43, 47-51, 97-9
place of articulation, 35, 55, 3
vowel chart, 99
plosion, 25, 61
plosive, 25, 28, 36, 60-3
post-alveolar, 36, 91, 92 weak form, 22
potential /r/, 95 whisper, 9
114
Index of Phonetic Symbols

a, 98 g, 103
a, 17, 99 h, 9, 10, 17
a, 98 h, 105
u, 17, 19, 99 fi, 105
aa, 17, 23, 24, 99
aa, 17, 23, 24, 99 i, 97, 98, 106
a:, 9, 17, 99 i, 5, 17, 50, 99
i, 9, 17, 51, 99
b, 10, 12, 17, 31 i, 17
6, 3 ia, 17, 19, 99
p, 27, 34
j, 17, 20, 99
c, 103 X, 104
5, 87, 104 J, 103
a, 105
k, 5, 17, 61, 76
d, 10, 13, 17, 58, 61, 66, 67
l, 10, 13, 17, 66-9, 76-9, 82, 110
^ 109
!, 78
<1, 88
l, 88
tf, 3
X, 104
d3, 16, 17, 31, 36, 110
i, 105
6, 13, 17
fe, 105
e, 98 !, 34, 110
ei, 17, 19, 99 |, 17, 22, 23
£, 98
m, 10, 12, 15, 17, 31
£, 9, 17, 99, 111
ig, viii, 12
£a, 17, 19, 99
ip, 34
a, 17, 18, 99, 111
m, 23
au, 17, 19, 31, 32, 99
,
3 17, 18, 51, 99
n, 9, 10, 13, 15, 17
n, 109
f, 8, 10, 17 9, 34, 109
j, 103 p, 17, 22, 23, 109
?, 2 n, 103
g, 10, 17, 109 0, 3, 14, 15, 17
g-, 3 *,23
115
Index of Phonetic Symbols

p, 103 ua, 17, 20, 99

H, 88 », 106

o, 98 Y, 27, 34, 93
з, 98 v, 8, 10, 17, 35
o, 5, 17, 31, 99 c, 93
d, 15, 17, 31, 50, 99 a, 106
D3, 17, 23, 99, 106 1, 15, 17, 18, 51, 99
DI, 19, 51, 99
w, 9, 10, 17, 24, 31, 99
5, 34, 35
m, 107
0, 106
q, 17
ce, 106
m, 50, 51, 106
e, 12, 17, 87
x, 27, 34, 87, 104
27, 34, 87
X, 104
p, 5, 12, 17, 31, 61, 76, 108
y, 50, 106
f, 6, 10, 17, 26, 31, 32, 36, 91-5, q, 106
110 X, 106
j, 92
3, 10, 13, 17, 110
I, 85, 91, 93, 110
X, 105
X, 94
2, 87
r, 85, 92
2, 105
r, r, 93
3, 16, 17, 36
k9 93
?, 9, 17, 34, 86, 94
s, 10, 13, 17, 87
q, 105
§,87
b, 2
g, 105
J, 16, 17, 31, 35, 87 9, 104
29, 34
t, 5, 13, 17, 36, 58, 61, 76, 81
103, 104, 108
J, 88
',21
t, 88 100
tj, 3, 17, 31, 36
+, 100
1,2 108
и, 98 = , 108
u, 5, 17, 31, 50, 99, 110 //, 82
u, 15, 17, 31, 99 [], 82

116
 Phonetics is the study of pronunciation. Actual pronunciation is often quite
different from what the speaker imagines it to be, and with the techniques of
phonetics described in this book it becomes possible to analyse pronunciation
objectively. Such ability is a vital prerequisite for the study of speech, linguistics
and speech therapy, and for a soundly-based interest in dialects and accents.

In addition to its coverage of techniques and theory. Practical Phonetics also

provides practical exercises and questions for discussion. A student who has
mastered its contents will be well on the way to success in the International
Phonetics Association s Certificate examination or the phonetics paper of such
an examination as that for the Licentiate Diploma of the College of Speech
Therapists. This has proved an invaluable book for students of English
language, speech therapy, linguistics, speech and drama, and allied subjects.

J C Wells, PhD, is Lecturer in Phonetics at University College, London.

Greta Colson, LRAM LGSH, is Principal Lecturer in charge of Voice, Middlesex

Polytechnic.

ISBN 0273 016814