Introduction to Phonetics for

Students of English, French,

German and Spanish

Dr. Rodney Ball

University of Southampton

Introduction to Phonetics for Students of English,

French, German and Spanish

`
 This Introduction to Phonetics was originally a booklet produced in the School of
Modern Languages at the University of Southampton, to serve as a background
and further reading text for the Articulatory Phonetics component of our first-year
Linguistics unit. It focuses on the structure and linguistic function of the vocal
tract, the classification of vowels and consonants, the International Phonetic
Alphabet and its use in phonetic transcription. Though phonology/phonemics is
not explicitly covered, the references to broad and narrow transcription in the
final section will point the user in that direction.

It is primarily addressed to native anglophones, drawing on their knowledge and

experience of English. However, it also contains extensive illustration from
standard French, German and (Castilian) Spanish, with plenty of emphasis on the
phonetic resemblances and differences between these four languages. There are
around a hundred exercises (answers are supplied), in which, again, English,
French, German and Spanish all figure. Though the course is not intended to
provide a systematic indepth analysis of the sound system of any individual
language, there is enough basic material here to serve as the starting-point for
subsequent language-specific Phonetics or Linguistics units.

CONTENTS
1. Introduction
1.01 What is Phonetics? 1
1.02 Why study Phonetics? 2
1.03 Working through this course 2
2. The Vocal Tract
2.01 Speaking and breathing 4
2.02 The source of air for speech sounds 4
2.03 The larynx 5
2.04 Voicing 6
2.05 The upper vocal tract 8
2.07 The oral tract from lips to uvula 9
2.07 The tongue 11
2.08 The pharynx 11
2.09 Pronunciation: an acquired skill 12
3. Vowels
3.01 Tongue position for vowels 14
3.02 The cardinal vowels 18
3.03 Lip rounding 23
3.04 Reversing the lip position 24
3.05 The secondary cardinal vowels 25
3.06 Focus on English 28
3.07 Diphthongs 32
3.08 Length and nasalization; Diacritics 36
3.09 Semi-vowels 38
4. Consonants
4.01 Classifying consonants 40
4.02 Place of articulation: bilabials, dentals, alveolars and velars 40
4.03 Manner of articulation 43
4.04 Stops (or plosives) 44
4.05 Fricatives 45
4.06 More places: palatal, uvular, pharyngeal, glottal 49
4.07 Aspirates 52
4.08 Affricates 53
4.09 Nasals 55
4.10 Laterals (the l sounds) 57
4.11 The r sounds 59
4.12 The IPA consonant chart 63
5. Phonetic Transcription and General Revision
5.01 Guidelines for transcription 64
5.02 Transcribing English 67
5.03 Transcribing French, German and Spanish 68
5.04 General Revision 71
6. Answers to Exercises
7. Further Reading
 introduction
[1]
1.01 What is Phonetics?

Languages can basically be thought of as systems — highly complicated ones —

which enable us to express our thoughts by means of “vocal noises”, and to
extract meaning from the “noises” (speech sounds from now on) that are made by
other people. Linguistics is the study of the nature and properties of these
systems, and its various branches focus on different aspects of the
communication process.

Phonetics is the branch concerned with human speech sounds, and itself has three
different aspects:

• Articulatory Phonetics (the most anatomical and physiological division)

describes how vowels and consonants are produced or “articulated” in
various parts of the mouth and throat.

• Acoustic Phonetics (the branch that has the closest affinities with
physics) studies the sound waves that transmit the vowels and consonants
through the air from the speaker to the hearer.

• Auditory Phonetics (the branch of most interest to psychologists) looks at

the way in which the hearer’s brain decodes the sound waves back into
the vowels and consonants originally intended by the speaker.

Closely associated with Phonetics is another branch of linguistics known as

Phonology. This focuses on the way languages use differences between sounds in
order to convey differences of meaning between words, each language having its
own unique sound pattern. Phonology is really the link between Phonetics and
the rest of Linguistics. This course focuses on the first of these aspects:
Articulatory Phonetics.

Warning. The word phonetics is often incorrectly used to refer to the

symbols of the International Phonetic Alphabet (the IPA). So people say:
“How is this written in phonetics?”, “It was all in phonetics, so I couldn’t
understand it”, or “Dictionaries use phonetics to show pronunciation”.
This isn’t how the term should be used. As has just been explained,
Phonetics is a branch of Linguistics, not an alphabet. So it would be more
appropriate to say: “How is this written in phonetic script?”, “It was all in
1
 phonetic transcription...”, or “Dictionaries show pronunciation by using
the phonetic alphabet”.

Introduction

You will be introduced to the IPA as you work through this course. Its symbols
are identified by square brackets: [p], [u], [ð], etc. Ordinary letters and spellings,
on the other hand, will always be given in italics. As you can see, some of the
phonetic symbols are the same as ordinary letters, but others will be new to you.

1.02 Why Study Phonetics?

Obviously it’s a fundamental part of Linguistics, so no-one studying this subject

can ignore it. But for students of languages, there are also practical advantages to
be gained from knowing some basic Phonetics.

Firstly, you should be able to improve your pronunciation of foreign languages if

you have a clearer idea of how the sounds are actually produced. Troublesome
sounds like French r, German ü or Spanish j lose their mystery and become less
daunting once you know how they relate to other more familiar sounds. And
there are various general features of the “British accent” which can be
characterized by phonetic analysis: when you know what it is that makes British
accents so British, you’ll be well on the way to getting rid of yours (if you have
one: most people do to some extent at least). What’s more, you’ll be able to look
up the pronunciation of words in the dictionary once you’re familiar with the
phonetic alphabet.

Secondly, many of you will at some stage or other find yourselves teaching a
language to other people: either French, German, Spanish, etc. if you make a
career of teaching, or English if you are involved in ESOL (English as a Second
Language, also known as EFL: English as a Foreign Language). ESOL is not
just a useful source of vac jobs: it is a serious career in itself. And many Modern
Languages students spend a year of their degree course working abroad as
English language teachers. In all such cases, you are likely to have to help
learners to improve their accents. If someone is having difficulty with English th,
it’s not much help just to tell them “don’t say it like that, say it like I do”.
(Unless they’re natural mimics, in which case they won’t need instruction from
you anyway.) Much better if you can guide them to make the appropriate tongue
movements, on a basis of your knowledge of phonetics.

In short, Phonetics always looks good on a language teacher’s cv.

2
 1.03 Working Through This Course

It contains a section describing the organs of speech, a section on vowels, a

section on consonants, and a concluding section on phonetic transcription,
together with a few suggestions for optional further reading.

There are also a large number of exercises, answers to all of which can be
accessed. Some of the exercises are to enable you to check that you’ve absorbed

Introduction

and understood the material covered, others encourage you to think more about
the languages you are studying and more particularly to draw on your experience
and knowledge of English.

1
 Introduction

2
 the vocal tract
[2]
2.01 Speaking and Breathing

All speech sounds in all languages are produced by modifying ordinary

respiration. In quiet breathing, air enters and leaves the lungs without any
obstruction, passing freely through the throat and mouth (or nose). If, however,
the tongue or some other organ is placed in the path of the airstream, this free
passage of air is disturbed; the air from the lungs may be set into vibration or the
flow momentarily interrupted. For example, the lips close and briefly cut off the
airstream for [p] and [b]. Any such disturbance generates a sound wave — a
ripple effect that travels through the air between speaker and hearer(s) and is then
interpreted as a particular speech sound. Articulatory phonetics studies the
various ways in which airstreams can be “interfered with”.

2.02 The Source of Air for Speech Sounds

The LUNGS (Fig. 1) are basically sponge-like in design, except that they hold air
(in a myriad of tiny airsacs), not water. When we breathe in, we enlarge the chest
cavity (in part by lowering the diaphragm). This in turn expands the lungs, and
air rushes in to fill the vacuum. Breathing out involves the opposite procedure.
The chest is contracted and air is squeezed out of the lungs, passing through the
two BRONCHI (or bronchial tubes), then through the windpipe (more
technically the TRACHEA), and finally emerging in the throat.

The Vocal Tract 5

 Larynx

Trachea

Right Bronchus Left Bronchus

Fig. 1
One or two refinements on this simple picture might be noted in passing.

First, we normally speak only while breathing out. It’s also quite possible to speak
while breathing in (for example when counting and not wishing to pause to draw
breath), but this is an inefficient way of making sounds and therefore not a regular
feature of any language. In some speech-communities, though, people use “ingressive
air” as a conventional means of disguising their voices.

Second, there are various ways of making speech sounds with air that doesn’t originate
in the lungs. The disapproving noise conventionally represented as tut tut! is an
example. Some languages make regular use of “click” sounds like this one, as well as
other “non-pulmonic” sounds that from a European point of view seem even more
exotic.

Third, if we used the same breathing rhythm for talking as for just breathing quietly,
we’d have to pause for breath every couple of words. (Try it and see.) In speech, quite
complex adjustments of the chest muscles and diaphragm are constantly being made in
order to slow down the airstream and hold it back as it leaves the lungs.

2.03 The Larynx

The statement above that the airstream “emerges from the trachea (windpipe)
into the throat” is actually an over-simplification. Before the air reaches the
“throat”, it has to pass through one of the most important speech organs, the
LARYNX. It’s at this point that the first possibilities occur of modifying the
airstream and generating sound.

The larynx can conveniently be thought of as an irregularly-shaped, hollow box

made of cartilage, which sits on top of the trachea. (This is reflected in the non-
technical name for it: the “voice-box”.) The front of the larynx can easily be seen

4 The Vocal Tract

and touched: it forms the projection an inch or two below the chin, known as the
“Adam’s apple” (more prominent in males than in females — hence presumably
the name).

Across the interior of the larynx are stretched two horizontal sheets of muscle
tissue. When these are relaxed and wide apart, then the air is free to pass between
them. This is how they are held for normal respiration (Fig. 2). But if they are
brought together with their inner edges in close contact, then air is prevented
from entering or leaving the lungs: the only way in or out is through the larynx
cavity, which is now sealed off (Fig. 3). This is the configuration for swallowing:
it prevents not only air but, more importantly, foreign bodies from getting into
the lungs.
Front of larynx (Adam’s apple)

Vocal Folds

Fig. 2 Fig. 3

A third possibility is shown in Fig. 4. The sheets of muscle are again in contact, but very
loosely this time, instead of being pressed firmly together as they were in Fig. 3. As a
result, air is able to pass through, but not freely: it has to force its way, so to speak. This
sets the inner edges of the muscles into vibration, and this vibration causes a disturbance
in the airstream — i.e. a sound wave. The sound is greatly amplified by the resonance of
the mouth and throat cavities, and the result is: the human voice. As a consequence, the
inner edges of the muscles stretched across the larynx are known as the vocal folds
(alternatively vocal cords or, occasionally, vocal lips). Say aaah, for instance: the sound
you’re producing is amplified vocal fold vibration. In essence, the vibration is similar to
the effect which you get by folding over a piece of thin paper and blowing between the
edges.

The space between the vocal folds is known as the GLOTTIS. So Fig. 2 shows an open
glottis, Fig. 3 a closed glottis, and Fig. 4 a vibrating glottis.

The vocal folds also control the pitch of the voice. As with the strings of a musical
instrument, the greater the tension, the higher the pitch. The larynx is provided with a
number of muscles which, together with the vocal fold muscles themselves, carry out
the complex adjustments of vocal fold tension that take place continually during speech.

The Vocal Tract 5

 Fig. 4

The larynx and vocal folds of women and children are smaller than those of
adult males: hence the difference between soprano and bass voices. When a
boy’s voice “breaks” at puberty, this is due to a rapid increase in the size of the
larynx.

Subtle and complex adjustments of the glottis give rise not just to “normal”
voice at a range of pitches, but also to such varied vocal effects as stage
whisper, falsetto and so-called “breathy voice”. But it’s worth remembering
that voice is a only secondary adaptation of the “vocal” folds, despite the name.
(Other mammals and even reptiles have a larynx too.) Biologically the primary
function of the larynx in general and the vocal folds in particular is to serve as a
valve for the lungs. As has been mentioned already, it’s advisable to close the
glottis firmly when swallowing — we all do so instinctively in fact. A second
important reason for having a larynx is that the closed vocal folds, by holding
back the airstream, can create a firm column of air in the chest, against which
we can push during various kind of physical exertion. Weightlifting, defecation
and childbirth all involve a tightly closed glottis!

2.04 Voicing

Although it’s merely a biological by-product, the importance of voice for

languages can’t be overestimated. All vowel sounds are normally uttered with the
vocal folds in vibration (i.e. they are voiced) and so are around half the
consonants. Take [s] and [z], for instance. As we’ll see in more detail later, for
both these sounds there is a constriction of the airflow just behind the upper front
teeth. The difference between them is that [s] is voiceless (vocal folds held apart
in the Fig. 2 position and the airstream able to pass between them unhindered)
whereas [z] is voiced, with the vocal folds in the Fig. 4 position and consequently
in vibration. You should be able to spot the absence or presence of voicing easily
enough if you say [s] ... [z] ... [s] ... [z] loudly several times in alternation. The
difference becomes even more obvious if you place your fingers firmly over your
ears while doing so.

The same relationship exists between the th in thin and the th in this. This time,
unfortunately, the spelling doesn’t show any difference. But by repeating these
two words in alternation you should be able to tell that in the case of thin we
have a voiceless th and in the case of this a voiced one. The phonetic alphabet
uses a separate symbol for each: [ T] (read “theta”) for the th of thin and [ð] (read

4 The Vocal Tract

“eth”) for the th of this. [ð] is also sometimes called “thorn” (from the name of a
letter of the Old English alphabet).

Yet another voiceless/voiced pair is [p] and [b]. Try to say [apa] and [aba] in
alternation. In both cases there’s a momentary blockage of the airstream between
the lips when the consonant is made. What makes the difference between them is
voicing (present for [b], absent for [p]).

In fact almost all consonants come in voiced and voiceless pairs — an very
efficient use of a single feature (voicing) in order to double, at a stroke, the
number of available sounds.

Exercises

1. Choose an appropriate term from the list, and insert it into one (or more)
of the gaps in the paragraph that follows:

bronchi lungs
glottis trachea
larynx vocal
folds
The _______________ supply the air for almost all speech sounds. Air passes from
them into the _______________, one from each of the two ________________, and
these two airstreams merge in the _______________, a short tube situated in the
lower part of the neck. On top of this is a valve known as the _______________.
Here the supply of air to the throat and mouth is controlled by opening or closing the
_______________ the gap between the two ________________. In ordinary quiet
breathing the _______________ is open; for swallowing it is closed in order to
protect the _______________. A noteworthy evolutionary adaptation in humans
allows voice to be produced by positioning the _______________ in such a way that
passage of air between them causes them to vibrate.

2. Say whether the following consonants are voiced or voiceless. The first
group have already been mentioned — see if you can answer without looking at
the text. For the second group the decision is up to you.

(a) [z], [T], [b], [p], [D], [s].

(b) [f], [v], [t], [d], [k], [g], [S] (this is the symbol for the sh of ship)

3. The following pairs of words are distinguished in pronunciation partly or

solely according to whether they contain a voiced (vd) or a voiceless (vl)
consonant. Say which is the crucial consonant in each case, and specify its
voicing status.

The Vocal Tract 5

 pin, bin zoo, sue
fail, veil wreath, wreathe (“in
toll, dole smiles”) either, ether (a
gin, chin kind of gas) Aleutian,
allusion.
4. Changing the voicing of a single consonant in each of the following
results in a different word. Which consonant and which word? (Sometimes
there’s more than one possibility. Concentrate on the pronunciation, not on the
spelling.)

seal bicker
razor lunge
ice Jews
scarce choke
ankle thigh
dug Confucia
n
2.05 The Upper Vocal Tract

The next thing you need to get a clear idea about is the configuration of the
cavities through which the air passes once it has left the larynx. These are
referred to collectively as the UPPER VOCAL TRACT — “upper” because the
vocal tract as a whole includes the larynx as well. You can call it the supra-
glottal tract if you prefer (supra is a Latin word meaning “above”). Fig. 5 gives
the overall picture. You can see from this that the upper vocal tract consists of
the mouth and throat cavities (together referred to as the oral tract) and the nasal
cavity (or nasal tract). The oral tract has a part to play in all speech sounds.
Sounds like [n] or [m] or the nasal vowels of French or Portuguese involve the
nasal tract as well, as will be seen later.

nasal tract

oral tract

Fig. 5 larynx

4 The Vocal Tract

Let’s go through the various parts of the oral tract, drawing attention to those that
are directly involved in the production (or articulation) of speech sounds, and are
therefore known as articulators. A number of consonant sounds will be
mentioned in passing in order to illustrate the part played by different
articulators, but there’s no need to concentrate too much on individual sounds at
this stage: they will be presented in more detail in later chapters. For the moment
you should aim at understanding the location and names of the various “organs
of speech” labelled in Fig. 6.

2.06 The Oral Tract from Lips to Uvula

a. THE LIPS. These are too familiar to need further comment, and the
involvement of the upper and lower lip in sounds like [p] and [b] is also very
obvious. (Details about exactly what happens will be provided later.)

nasal cavity soft palate / velum

hard palate
teethridge / alveolum
oral tract
front uvula
blade
tip
back

pharynx
root

food passage / oesophagus

trachea / windpipe
vocal folds
Fig. 6 larynx
b. THE UPPER FRONT TEETH. These are involved for example in the
production of [T] and [ð] (as in thin and this), for which the tongue comes into
contact with the back of the teeth. As the tongue is the moveable organ which
initiates the contact, it is said to be an active articulator, and the teeth, which
don’t move, are a passive articulator. The lower teeth and the remaining upper
teeth don’t appear to have any role in language.

c. THE ALVEOLAR RIDGE. Place the tip of your tongue against the rear
of your upper front teeth. Then draw it slowly backwards along the roof of the
mouth. You’ll notice that there is a bulge or ridge just behind the teeth, after
which the roof of the mouth rises in quite a steep, domelike way. This is the
teethridge — in phonetics more commonly called the alveolar ridge or alveolum.
It’s an important passive articulator for sounds like [t], [d], [s] or [z]. Again the
tongue is the active articulator.

The Vocal Tract 5

 d. THE HARD PALATE. This is the steeply rising section of the roof of the
mouth behind the alveolar ridge. It serves as a passive articulator in sounds like
the h of huge.

e. THE SOFT PALATE or VELUM. If you continue to run your tongue

backward along the roof of the mouth (as far back as it can go) you will come to
a point where the hard bone of the palate gives way to soft tissue. This section of
the roof of the mouth is accordingly known as the soft palate, or, more
commonly, the velum. The back of the tongue comes into contact with the velum
for consonants like [k] and [g].

The velum is an important organ of speech because it’s moveable and its
movement controls the entrance to the nasal cavity. (That’s why it’s soft not
hard: it consists of muscle tissue.) Raising the velum so that it’s pressed against
the rear wall of the throat has the effect of closing off the nasal tract, so that air is
diverted into the mouth (dashed line in Fig. 6). If you want to breathe through
your nose, you have to lower the velum (solid line in Fig. 6).

Nasal consonants like [m] or [n] and nasalized vowels are articulated with the
velum lowered. For non-nasal sounds (that’s the vast majority), the velum must
be in the raised position, so that the airstream passes into the mouth. Note that
the velum can’t block the entrance to the oral cavity, even when it’s lowered. So
even for nasal sounds, some air enters the mouth. More about this point in 3.08
and 4.09.

f. THE UVULA. This is the extreme tip of the velum, and isn’t directly
involved in the closure of the nasal cavity: you can see from Fig. 6 and several of
the other figures how it dangles down instead of being pressed against the rear
wall. Some r sounds in French and German involve the uvula. More about these
in 4.11.

2.07 The Tongue.

The tongue has long been thought of the speech organ par excellence, even
though its biological role lies in tasting and swallowing, not in vocalizing. In
many languages the word for “tongue” and the word for “language” are one and
the same (French langue, Spanish lengua, Russian iazyk for instance, or tongue
in Biblical and Shakespearean English). In actual fact the larynx is also
important, as we have seen — but as people are much less conscious of it, it
seems to have attracted less attention.

Anyway, the tongue is certainly involved in the articulation of a large number of

sounds, just a few of which have been mentioned above. Its versatility is due to
the fact that it consists entirely of nerve and muscle tissue, so it is highly flexible

4 The Vocal Tract

and mobile. You can see from Fig. 6 that the tongue is not thin and flat (even
though it may feel that way), but has a considerable amount of depth or body.

It’s convenient to consider the tongue as consisting of a number of different

sections (see Fig. 6). As there are no clear cut-off points on the tongue itself,
these division are somewhat arbitrary, and can vary from one authority to
another. But most phoneticians distinguish between the TIP, the BLADE, the
FRONT (not a good name, as it’s more like the middle than the front!), the
BACK and the ROOT. These articulate against different parts of the roof of the
mouth, giving sounds like the s of so (with the blade), the sh of shall (with the
front), and so on.

2.08 The Pharynx

Even more so than roof of the mouth and tongue, the term throat is somewhat
vague and general. (Should it be taken as including the larynx, for example?)
Consequently throat isn’t a word that’s used much by phoneticians, who prefer
more specific terms. Larynx is one which you already know, and another — not
to be confused with it — is PHARYNX. This designates the tubular cavity
bounded by the larynx, the root of the tongue and the soft palate, shown in Fig. 6.

You can see from the figure that the pharynx is a kind of crossroads: air passes
through it from the lungs to the nasal cavity; food passes through it from the
mouth to the oesophagus or food-pipe.

On the face of it, this mingling of food passage and airway sounds a rather
unsatisfactory arrangement. And in fact in animals the larynx is situated higher
up, so that it’s linked directly with the nasal cavity: no danger of choking for
our dumb friends. But a high larynx is much less efficient for the articulation
of speech sounds. It looks as though the “low-slung” human larynx has been
favoured by evolution, as it allows better vocal communication. That the
occasional unfortunate individual should choke to death is presumably a price
well worth paying! Incidentally human babies have a high larynx, which
“migrates” downwards during the first months of life: a nice example of
“ontogeny recapitulating phylogeny”.

In some languages, the rear wall of pharynx serves as a passive articulator. The
root of the tongue is pulled back towards it, causing a constriction used for
certain characteristically “guttural” sounds in Arabic or Hebrew (see 4.06).

Remember the difference: the larynx is a cartilaginous box immediately above

the trachea; the pharynx is the cavity or “crossroads” above the larynx.

The Vocal Tract 5

 2.09 Pronunciation: an Acquired Skill

Two points by way of conclusion.

Firstly, the structures described above are universal in humans: whatever the
race or speech-community, the speech organs are the same. There are of course
differences in the size and even precise shape of particular parts of the vocal
tract, but differences between individuals in a given speech-community are
greater than the average differences between one community and another.

It follows from this that the pronunciation difficulties experienced by people

learning a foreign language are due to psychological factors, not to anatomical
ones. After the age of 10 or so, most people find it difficult to break away from
their native sound system. So someone who finds French r, German ü or Spanish
j troublesome doesn’t have anything wrong with their uvula, tongue or velum (or
at least this is highly unlikely): it’s just that their brain is finding it hard to adapt
to new articulatory habits. A person’s knowledge of his or her mother tongue is
in no way congenital. Had you been kidnapped at the age of six months and
brought up in France, Germany or Spain — or China or the Amazon Basin for
that matter — you’d be speaking the local language in a totally “native-like”
manner ... and no doubt having problems with English.

Exercises

5. Show where the following are situated on this diagram of the upper vocal
tract:
1. blade of tongue, 2. front of tongue, 3. velum, 4. pharynx, 5.
larynx, 6. uvula, 7. alveolum, 8. root of tongue.

6. Identify these close-ups of sections of the upper vocal tract.

4 The Vocal Tract

(a) (b)

(c) (d)

7. Which of the following parts of the vocal tract are moveable?

lips, tongue, alveolum, pharynx, velum, uvula, nasal cavity.

8. Say whether the following are TRUE or FALSE. If false, say what the
correct statement should be.

(a) The pharynx is a cartilaginous structure attached to the trachea.

(b) The alveolum serves to open and close the entrance to the nasal
cavity.

(c) The lower front teeth are not involved in speech.

(d) The uvula is the extreme tip of the soft palate.

(e) Larynx is an alternative name for the teethridge.

9. Give the technical names for the following:

wind-pipe, voice-box, space between the vocal cords, food-pipe, soft

palate, teethridge, cavity at the back of the throat.

10. Distinguish (when appropriate) between the following:

the vocal tract the

upper vocal tract
the supra-glottal
tract the oral tract
the nasal tract.

The Vocal Tract 5

 11. Draw from memory a labelled diagram of the upper vocal tract.

4 The Vocal Tract

 Vowels
[3]
The story so far:

Speech sounds are produced (articulated) by placing an obstacle of some

kind in the path of air passing through the laryngeal, pharyngeal and/or oral
cavities.

3.01 Tongue Position for Vowels

There are a huge number of different vowel sounds in the languages of the world,
and, as will be seen in a moment, one of the jobs of phonetics is impose order on
chaos by providing a way of describing them coherently and concisely. But
whatever their differences, all vowels involve free passage of lung air through the
upper vocal tract. (This is what distinguishes them from consonants, for which, as
you began to see in the last chapter, there is always some obstruction above the
level of the larynx). The impediment to the airstream for vowels is located at the
glottis, not in the supra-glottal tract: the vocal folds are in close enough contact
for vibration to occur. The sound wave that this generates is amplified by the
resonance of the cavities above the larynx.

This means that vowels are normally voiced: “normally” because voiceless
vowels are a possibility — they can best be thought of as whispered vowels.
Although the vocal folds aren’t actually vibrating for these, there’s still sufficient
constriction at the glottis for the airstream to be impeded and for turbulence (i.e.
a sound wave) to be generated — you can feel this constriction if you say a few
vowels in as loud a whisper as you can.

Quite a few languages (Portuguese is a well-known example) have voiceless or

whispered vowels in unstressed syllables. (They aren’t sonorous enough to be
much use in stressed syllables). Voiceless vowels are even possible in English
in words like police or polite: here the stress is on the last syllable, so the o in
the first syllable, instead of being fully voiced, may, with some speakers, be
devoiced as a preliminary to disappearing altogether: p’lice, p’lite.

Now, although the air passes freely through the vocal tract, the shape of the tract
varies from one vowel to another according to the position taken up by the
tongue. (But it never gets so constricted that the airstream is slowed down or
blocked.)

Vowels 15
 Figures 7, 8, 9 and 11 below are from tracings of X-ray photographs taken while
the vowel sound of the word in question was being uttered, so they give a true
representation of the shape of the tongue — something which we’re usually only
very dimly aware of.

You can see how the position of the tongue differs from one vowel to another.
This is the crucial factor that distinguishes, say, an i from an a, or a u from an o.
For each vowel the differently shaped interior of the oral cavity generates a sound
wave with unique properties, and this gives the hearer an auditory impression of
vowels that are different in “quality”. (The shape of the lips can also have an
effect, though it doesn’t show up on these pictures, as they are taken from
sideways on: the role of the lips will be considered in 3.03.)

Vowel quality (determined by tongue — and lip — position) is independent of

loudness (determined by the force with which the air is expelled from the lungs)
and pitch (controlled by the tension of the vocal folds — see 2.03). Even when
pronounced at the same pitch and with the same degree of loudness, an i will
always be different in quality from an a, or an e from an o. In much the same
way, an identical note will differ in quality according to whether it’s played on a
violin or a trumpet: violins and trumpets being different in shape (and indeed in
material).

Classifying vowels is in large part a matter of specifying tongue positions. Here

is the diagram for heed (i-type vowel). The surface of the tongue is high in the
mouth — close to the hard palate, but not close enough for the airflow to be
impeded, which would turn the vowel into a consonant.

Fig. 7 heed
[i] is said to be a high vowel.

For had, the tongue is low in the mouth — almost flat; the mouth is much more
open inside than for heed. Fig. 8

14 Vowels
 had

[a] is said to be a low vowel.

Having seen these diagrams, you should be able to sense the high/low difference
simply by saying [i] and [a] to yourself, and feeling the difference in tongue
posture.

In some books you’ll find the terms close and open instead of high and low
respectively. But there’s no difference in meaning: a high vowel is the same
thing as a close vowel; a low vowel is the same as an open vowel. High and
low are the commonest terms, and they will be used in most of what follows.

Now the [u] of who’d is also a high vowel. What makes it different in quality
from
[i]?

who’d

Fig. 9

It’s not quite so easy to sense the answer just by saying [i] and [u] and trying to
compare the effects. But Fig. 9 should be clear enough: for [u] the tongue is
pulled backwards (“retracted”), in such a way that it’s arched towards the soft

Vowels 15
 rather than the hard palate. So [u] is a back vowel as well as a high one. Vowels
like [i] and [a], by contrast, are said to be front vowels:
Fig. 10
[i] high + front [u] high + back

[a] low + front

The [A] of spa differs from each of these three. For [A] the tongue is low (unlike
[i] and [u]), but retracted (unlike [i] and [a]).

spa

Fig. 11
So the gap in Figure 10 can now be filled in:

[i] high + front [u] high + back

[a] low + front [A] low + back

The high-low and front-back “parameters” thus operate independently of one

another, intersecting to give four basic vowel sounds which occur in a great many
languages.

Notice how the IPA makes a distinction between the [a] symbol (known as “front
a”) and the [A] symbol (“back a”). These represent different sounds and are not
interchangeable, as they would be in the ordinary alphabet. In other words, [ kat]
cat is one word, [kAt] cart is quite another.

Exercises

12. Without looking at the above text, say which of the vowels [i, u, A] is
represented in each of the following diagrams.

14 Vowels
 (a) (b)

(c)
13. Again without looking at the text, say whether the following statements are
true or false. Correct any that are false.

(a) [a] is a high back vowel.

(b) [i] is a high front vowel.

(c) [u] is a low back vowel.

(d) [a] is a low front vowel.

14. Practise using the alternative terminology:

(a) Which is [i], open or close?

(b) Which is [u], open or close?

(c) Which is [a], open or close?

15. Match the articulatory terms in the first column with the descriptions in the
second column

Vowels 15
 (a) low “surface of tongue raised
towards hard palate”

(b) open “tongue advanced”

(c) front “tongue retracted”

(d) high “narrow oral cavity”

(e) back “tongue flat”

(f) close “wide oral cavity”

3.02 The Cardinal Vowels

[i], [a], [u], and [A] are the basis of the standard “international” chart of vowels —
devised in the 1920s by the London University phonetician Daniel Jones, and a
particularly useful invention for anyone studying Modern Languages. This chart
is independent of any individual language: it provides a general grid or matrix in
terms of which vowels in particular languages can be specified.

Jones’s starting-point was to make a recording of the highest and “frontest” vowel
which the human vocal apparatus can produce (move the tongue any higher or
further forward, and the vowel will turn into a consonant because the airstream
will become constricted). This is a kind of archetypal [i] vowel. For the English
vowel of heed, the tongue is somewhat lower and less fronted than this; Jones’s
“Cardinal Vowel No. 1” is much more like the [i] vowels of French, German or
Spanish.

The other three “cardinal points” of the “vowel compass” were provided by the
lowest possible front and back vowels (physically impossible for the tongue to be
lowered any further), and by the highest and “backest” possible vowel (again, any
further tongue raising or backing results in a consonant).

Here are these four vowels displayed around a quadrilateral. For anatomical
reasons low front vowels can’t be as far forward as high front vowels, so the
resulting figure isn’t quite symmetrical.

Fig. 13 i u

14 Vowels
 a A

The English vowel of spa is a fairly good approximation to [a], but, in Southern
British pronunciation at any rate, the vowels in hat and who’d are respectively
higher and lower than cardinal [a] and [u]. French patte and sous, German hat
and Fuß, Spanish más and su have vowels that are much closer to these two
cardinals in quality.

Here is the chart again, with marking the position of three “non-cardinal” English
vowels in relation to the “basic four”.
Fig. 14 i u
heed who’d

hat

a A

Such diagrams are schematic representations of the “vowel area” of the mouth:
just like the vowels of heed, hat and who’d in Fig. 14, each of the hundreds of
different vowels that the the human vocal apparatus can produce is locatable at
some point within the space bordered by [i, a, u, a]. Not that there is any need
actually to include hundreds of vowels on the chart: another dozen cardinals are
enough to allow all the others to be specified.

So in between the highest and lowest vowels we have “mid vowels” at equidistant
intervals:

[e] (mid-high front) as in Fr bébé, Ger Weh, Sp de

[E] (mid-low front) as in Eng head, Fr terre, Ger Hemd, Sp estar

[O] (mid-low back) as in Fr donne, Ger toll, Sp donde

[o] (mid-high back) as in Fr rose, Ger Kohl, Sp boca.

Vowels 15
 Figure 15 shows how the four mid-vowels fit into the quadrilateral alongside the
previous four. We now have a set of eight cardinal vowels that are numbered
anticlockwise, starting with [i] (no. 1). Often they are referred to just by their
numbers.
(8)
(1)
i u

(2) e o (7)

(3) E O(6)

(4)
Fig. 15 a A (5)

As before, Southern British English isn’t much of a guide. The vowel of on is a

bit lower than Cardinal 6, though not very different from it. But the vowels of
day and own really are very unlike Cardinals 2 and 7 (they’re actually diphthongs
— see 3.07).

Note the name of the special phonetic symbol for Cardinal 3 [ E]: “open e”
(recall that a low vowel is also known as an open vowel, and that [ E] is open
compared with [e], or “close e”). Similarly “open o” for No. 6 [O] — this
resembles an o which has been “opened up” on one side (printers use a
backwards-facing c for this symbol.)

One more vowel to complete this section. This is actually the commonest vowel
in English, although it doesn’t have its own letter in the spelling (probably
because it only occurs in unstressed syllables). It’s the vowel spelt a in about, e
in patent, i in virginity, o in convince, u in pursue. It also occurs in German (e.g.
in the last syllable of haben) and in French words like le, de. As the IPA operates
on a strict principle of each sound having its own symbol, with no overlapping of
the sort that occurs in English spelling, this vowel is represented as [ @]. The
name of the symbol is schwa (it was adapted from the Hebrew alphabet). So
header would be transcribed [hEd@] and you would read this out as “aitch, open e,
dee, schwa”.

Make sure you write the schwa symbol as an inverted (upside-down) e, not as a
reversed (back-to-front) e. And certainly not as an inverted and reversed e!

14 Vowels
For [@] the tongue is held in a neutral position: neither particularly high nor
particularly low (but “mid”); neither particularly front nor particularly back (but
“central”). You can see this if [@] is added to the vowel chart:
i u

e o
@

E O

Fig. 16 i A

This mid-central position explains why in English (and many other languages)
[e] serves as a “hesitation noise” — a sound uttered when you’re not sure what
you’re going to say next, but want to signal that you’ve not finished yet. The
tongue is, so to speak, poised to move in any direction once you’ve made up
your mind. So the phonetic transcription of um, er would be [@m, @], or better
still [@:] with a length mark (see 3.08).

Exercises

16. Multiple-choice revision test (don’t look at the text!)

a. Daniel Jones was Professor of Phonetics at the University of:

 London
 Oxford

 Southampton

 Derby

 Edinburgh

b. Cardinal vowels are so called because:

 the symbols used for them look like cardinal’s hats

 they are more important than the other vowels

 they guide you like the cardinal points of a compass

 they occur in all languages

c. The cardinal vowel chart is important because:

it provides a framework for the description of the vowel sounds of

any language

Vowels 15
  it makes the difference between vowels and consonants clearer
 the first eight cardinal vowels are exactly the same as those of

Southern British English.

 it shows the most important vowels of any language

17. Without referring back to the text, complete this chart by inserting the
symbols for cardinal vowels 1-8.

18. Write out the schwa symbol a few times, then check that you’ve done it
properly by turning the page upside down. What you’ve written should now
look like an e.

19. For whichever of the languages you know, indicate the cardinal vowels that
correspond to the vowels underlined in the spelling:

French: bête, foule, premier, patte, été, rose, prêt, pré,

dehors, sortie, petit, sous

German: sehen, Hände, wenden, ruhen, Sonne, Sohn,

hatte, Vater, Tier, See, wahr

Spanish: ser, amigo, hombre, hoja, cabo, verde, clase,

hemos, rojo, grande, salir.

20. Read out the following nonsense words (i.e. words not actually occurring in
any of the languages). Don’t just pronounce them as though they were

14 Vowels
 English words: the more “foreign” you can make them sound, the more
likely your cardinal vowels are to be authentic!

pEm, fep, lak, lAk, tOb, tob, fuba, esiv melAg, wOp@s,

sEtAf, akamAz, AkAmaz, fusEki

kOmus@, ponOsE, ohelEg, tim@nA, @pezu, elOkEso

3.03 Lip Rounding

The two high cardinal vowels 1 and 8 aren’t only differentiated by the fact that [i]
is front and [u] is back. For [u] the lips take on a rounded position, much as for
whistling. But for [i] the lips aren’t rounded in this way — in fact the most
authentic cardinal [i] is obtained if the lips are spread energetically sideways (as if
the teeth were being bared). This difference between “rounded” and “unrounded”
applies to other vowels as well. The front vowels [e, E, and a] are unrounded just
like [i]; the back vowels [o and O] are rounded like [u].

As before, English isn’t the best language to illustrate this point. The lips are indeed
somewhat rounded for the vowels of words like food or on (try these for yourself), and
rounding is unquestionably absent in the case of heed or head (try these too). But the
rounded vowels aren’t very strongly rounded, nor are the lips noticeably spread for the
unrounded ones. In French, German or Spanish, the rounding/unrounding effect is more
marked — this is something to be imitated if you want to make your pronunciation more
“authentic”.

So we can add a further parameter to the two we met before. Vowels are
distinguished by tongue position (the high/low axis and the front/back axis) AND
by lip position (the rounded/unrounded axis). In practice there are significant
intermediate positions between high and low (the mid-vowels described in 3.3)
and sometimes between front and back: on the other hand, the
rounded/unrounded axis operates in a binary way, with vowels counting as either
one or the other.

Exercises

21. Which of the following words contain rounded vowels and which unrounded
vowels?

French: tête, fou, patte, cru, rose, prêt, pré German:

Reh, Hand, hoch, Sohn, früh, Tier, wahr

Vowels 15
 Spanish: ser, loco, verde, cara, tu, lo, si.

22. Characterize the following vowels in terms of the three parameters. For
example: [i] = high, front, unrounded.

[u]
[e]
[E]
[a]
[O]
[@]

3.04 Reversing the Lip Position

With the vowels considered so far, if they are FRONT they’re UNROUNDED and if
they are BACK they’re ROUNDED. But things don’t have to be that way: these
parameters are independent of one another, and both French and German
contain examples of FRONT vowels that are ROUNDED. Thus the vowel of
French lune or German für resembles [i] in tongue position (high, front), but
is like [u] in lip position (rounded). Anyone studying these languages knows
only too well that this is an entirely distinct vowel: rue is not the same as
roue or ri; für is different from fuhr and vier. Therefore it needs its own
phonetic symbol, and [y] is used for it. (In the IPA y has the same value as in
ordinary German spelling where it’s identical to ü: Psychologie, etc.).

Be sure that you’re clear about the difference between:

[i] high, front, unrounded

[u] high, back, rounded
[y] high, front, rounded

And clear about how the vowels in the various words just mentioned are
transcribed:

14 Vowels
 [ri] = ri, [ru] = roue, [ry] = rue [fir] = vier,
[fur] = fuhr, [fyr] = für

(See 4.11 for other ways of transcribing r)

English speakers tend to mispronounce [y] as [u], i.e. they keep it rounded, but
use a high back instead of a high front tongue position. So in French they
confuse rue and roue. This is because all the rounded vowels of English happen
to be back, so the reflex of English speakers is to retract the tongue for any
rounded vowel they come across.

An alternative anglophone subterfuge is to make French pu sound like English

pew. This time the single French front rounded [y] vowel is being split into a
sequence of two vowels: a short [i] (front) then a [u] (rounded). Ingenious, but
it won’t do either.

Curiously, speakers of Spanish or Italian, whose languages don’t have front

rounded vowels either, often mispronounce [y] in the “opposite” way — as [i].
That is, they keep the front tongue position, but make the vowel unrounded.

So the only way to get it right is to try as hard as you can to keep the vowel front
(tongue position for [i]) while strongly rounding the lips (lip position for [u]).
At all costs avoid retracting the tongue — this turns front [y] into back [u],
which is precisely what is not wanted.

Exercises

23. Give the IPA symbols for the vowels underlined in each of the following
words, in either or both languages:

French: pu, pou, pis, voulu, lugubre, fourrure

German: Ufer, über, Mühe, Fuß, Füße, Zypresse, Statue.

Now read the words aloud, using the appropriate vowel.

24. Rewrite, using ordinary spelling:

French: [ful] [lyn] [bude] [fyme] [vu] [vy] [fu] [fy] [u] [y]

German: [hyg@l] [mut] [ku] [zys] [fyl@n] [hun] [nud@l]

Vowels 15
 3.05 The secondary Cardinal Vowels

French and German have more front rounded vowels than just [y]. If the mid-
high front [e] is pronounced with rounded lips, the the result is the vowel of peu
or the first syllable of mögen. In other words this is a vowel that has the same
relationship to [e] as [y] has to [i]. It’s a distinct and important item, and
therefore it too has its own symbol: [ø] (“slashed o”).

Try to keep the front tongue position of [ e] while rounding the lips. A common
anglophone mispronunciation is to pronounce [ø] like a long version of [@].
But this is too far back, and it isn’t rounded, so it won’t do.

As Spanish and Italian don’t have [@], the usual hispanophone or italophone
mispronunciation is simply to unround [ø] and pronounce it as [e], so that deux
and des become identical.

Yet another front rounded vowel in both French and German is the counterpart of
[E]. This occurs in peur or Götter, and is transcribed [œ], following the spelling
of French words like soeur. Get the distinction between [e] and [ E] fixed in
mind, then say [e] with rounded lips, and the result should be a perfect [œ].
Again, a long [e] isn’t acceptable: purr is a very poor substitute for peur!

The correct term for a composite letter like [œ] is digraph, not diphthong:
diphthongs are something different, as will be seen in 3.07. Read [œ] as
“O-E digraph” or “O-E ligature” (the latter is the official IPA name, the
term ligature referring to the tie-bar sometimes used as an alternative
way of 7 linking two elements of a digraph: ts).

One problem with [ø] and [œ] is that French and German spelling don’t
distinguish clearly between them: the vowels of peu and peur are not in fact the
same. As a rule of thumb, in French the letter sequence eu is pronounced [ø] at
the ends of words, but [œ] before a consonant. Accordingly heureux is [œrø].

To summarize, we’ve looked at the following examples of reversal of lip

rounding:

[y] high, front, rounded [i] high, front, unrounded

[ø] mid-high, front, rounded [e] mid-high, front, unrounded

[œ] mid-low, front, rounded [E] mid-low, front, unrounded

14 Vowels
It should come as no surprise if it’s now revealed that the lip position of any of
cardinal vowels 1 to 8 can be reversed in this way. Many languages of the Far
East, for example, have BACK UNROUNDED vowels. Try saying the Japanese name
Suzuki with lips resolutely spread throughout (not rounded), but no less resolutely
keeping a back tongue position for the two u’s, and this should give you an idea
of what [Ì] — the unrounded equivalent of [u] — sounds like. Again, schwa
won’t do — this time because it’s too far forward.
Japanese lacks a back rounded vowel of a [u] type altogether — in fact the only
one of this language’s five vowels that’s to any extent rounded is [ O]: all the
others ([i, e, a, Ì]) are unrounded. This relative lack of rounded vowels naturally
has a marked effect on Japanese pronunciation of European languages,
especially French, which, with more rounded than unrounded vowels, is the
“opposite” of Japanese in this respect.

So alongside the “primary” series of cardinal vowels (1 to 8), there is a

“secondary” series (9 to 16) — identical in tongue position, but opposite in
respect of lip rounding. This too figures on Daniel Jones’s historic
recording. Here is the chart with both primary and secondary cardinals
included:
(9) y (1) i
u (8)
Ì (16)

(2) e
(10) Ô o (7)
Ê (15)

(11) ˚ (3) E O (6) ˆ (14)

Fig. 17 (12) œ (4) a A (5) ¡ (13)

At this stage, you should be able to reproduce the chart of the eight primary
vowels from memory, and also locate on it [@] and cardinals 9, 10 and 11. You
needn’t worry about 12-16 — apart from noting their existence. You should
also have a clear auditory image of what each of these vowels sounds like, and a
clear understanding of why each of them occupies its particular place on the
chart.

Exercises

25. Match the vowels in the first column with the descriptions in the second.

Vowels 15
 [y] high, front, unrounded
[e] mid-low, front, unrounded
[E] high, front, rounded
[ø] mid-high, front, unrounded
[i] mid-low, front, rounded
[œ] mid-high, front, rounded

26. Say whether the following statements are True or False:

 front vowels are always unrounded in English

 any vowel can be rounded or unrounded

 intermediate lip positions between rounded and unrounded are important

in distinguishing vowels
 intermediate tongue positions between high and low are important in

distinguishing vowels.

27. Which of the following languages make(s) use of lip rounding to distinguish
between vowels?

English, French, German, Italian, Japanese, Spanish.

28. Give the IPA symbols corresponding to the underlined vowels in one or
more of the languages indicated:

French: bonheur, soeur, deux, milieu, jeune, stupeur, peut-être.

German: Söhne, völlig, Körper, mögen, köstlich, Goethe, töten.

29. What words do you get if you reverse the lip rounding in the following
(concentrate on pronunciation, not on spelling)?

French: père, mère, ces, des, si, cri, heure, noeud

German: Heere, Hölle, Tier, Kölner, lügen, kennen,

fühle

30. Rewrite in ordinary spelling:

French: [sœl], [famÔ], [idÔ]

German: [fÔg@l], [gœn@n], [hÔl@].

14 Vowels
3.06 Focus on English

In this section, you will see how the cardinal vowel system can be used to identify
and characterize a few of the pronunciations of English vowels heard in the
British Isles and across the world.

One term you should note at this point is Received Pronunciation or RP (another
Daniel Jones invention). This slightly odd expression refers to the “standard
British English accent”, the one used for example by most BBC television and
radio newsreaders and presenters — or, in a somewhat more conservative form,
by the Queen herself. (Hence the expression Queen’s English, which is roughly
equivalent; BBC English is another term sometimes encountered). Middle-class
pronunciation in England approximates to RP. This is especially true in the
South, as Received Pronunciation derives from southern — particularly London
— varieties, not from Midland, northern, Scots, Welsh, Irish, etc.

As has been hinted already, the vowels of RP are not particularly close to the
cardinal values. The pronunciation is less tense than that of, say, standard French:
the tongue takes up positions that are less “extreme”, and the lips are less
energetically spread or rounded. So RP vowels tend to be located towards the
centre of the vowel area, not around its edge. For the more obviously “non-
cardinal” RP vowels, the IPA uses additional symbols; for the others the cardinal
symbols are retained — with the proviso that the pronunciation may deviate from
the strict cardinal value (compare the [i] in Eng need with the one in Fr midi).

Here are a few extra symbols commonly used in the transcription of English. The
sounds they represent aren’t unique to English of course, and the symbols can be
used for other languages if appropriate. None of them are required for French or
Spanish, but two of them are needed for German.

[{] This symbol (“a-e ligature”) is for the vowel of pat, which is somewhat higher
than Cardinal 4, though still front — it’s midway between [a] and [ E] in
fact:
(3) E

Fig. 18 (4)
a

This is why the symbol consists of an a and an e linked in a digraph.

Vowels 15
 As for the pronunciation, think of the difference between an authentically French
pronunciation of patte and the usual S. British pronunciation of this word that
makes it sound too much like pat. But many speakers from other parts of the
British Isles pronounce pat with a vowel that is close to, or identical with Cardinal
4 — and accordingly close to or identical with the French, German or Spanish [a].

[I] This is for the vowel of hit, which is lower and backer than the [i] of heed —
itself a little lower and backer than Cardinal No. 1:
(1) i

(2) e

Fig. 19

The [i]/[I] distinction is a crucial one in English, as hundreds of pairs of words

depend on it: feet, fit; meal, mill; relieve, relive, etc. It is no less important in
German (ihn, in; Miete, Mitte). But standard French and Spanish have only an [i]
and no [I]. However, [I] is a feature of the French of Belgium and Canada.

The symbol is referred to as “small capital i”. For languages like English and
German which need [I] as well as [i], the latter can be referred to as “lower-case i”
when clarity is required.

[U] Read “upsilon”. It’s the mirror-image of [ I] in the opposite corner of the
vowel chart, being further forward and lower than back [u], which itself is
slightly further forward and lower than Cardinal No. 8. Full [U] and fool
[u] illustrate the difference in English; Fluß [U] and Fuß [u] in German.

u (8)

o (7)

Fig. 20

Germanists should note a further symbol! English has two high vowels [i] and
[u], each with a more central equivalent [ I] and [U], as you’ve just seen. But
German has three high vowels [i], [u] and [y], and, symmetrically enough, each
of the three has a “centralized” equivalent [ I], [U] AND [Y]. This last vowel

14 Vowels
 (read “small capital y”) is found in many common words (fünf, müssen) and
sometimes serves to distinguish pairs, e.g. hüte, with [y], and Hütte, with [Y].
Standard French, on the other hand, which lacks [ I] and [U], also — not
surprisingly — lacks [Y]. (But the latter, like [I] and [U], is used in Canada and
Belgium.)
[¡]
This is sometimes used for the vowel of on, hot, lost (read “reversed italic
a”). It’s actually the symbol for Cardinal 13. The English vowel in
question is located between this and Cardinal 6 [ O], so some transcribers
prefer to use [O] for on, hot, lost, etc.
[O:]
This is Cardinal 6 with a length-mark (see 3.08) and the normal way of
representing the vowel of horse, caught, lawn: [hO:s], [kO:t], [lO:n].
“Turned v” is actually the symbol for Cardinal 14 (one of the back
[ˆ] unrounded
vowels you were told not to worry about in 3.05). The RP vowel found in
words like up, butter, come is close enough to this for [u] to be used to
represent it in transcriptions of English. Note it carefully: it’s a very
common sound in RP, where put, with [U], is not at all the same thing as
the golfing term putt, with [ˆ].

These symbols are in general use for transcribing English. Less widely adopted
E
is [ ] (“reversed open e”), which some phoneticians use for the vowel in her,
girl, first, etc. But this sound is really just a long schwa, so it’s OK to represent
it as [@:] and economize on a symbol (see 3.08 for the length-mark :).

Exercises

31. Can you remember?

a. what the standard British accent is called?

b. the term used for composite symbols like [æ] or [œ]?
E
c. the names of the symbols [i],[u] and [ ]?
d. the characteristic general difference between Southern British
vowels and those of standard French, German or Spanish?

32. Southern pronunciation is the basis for the British standard because:

a. London is the capital of the UK

b. It is a more attractive way of pronouncing words
c. Southern vowels are less like the vowels of continental languages
d. It is an easier, more natural way of speaking.

Vowels 15
 33. Say what IPA symbols would best represent the vowels underlined in the
following words:

hand, car, alphabet, foot, butter, further, roar, swan, wish,

monkey, some, cough, enough, yawn, pull, knowledge.

34. The following are all identical in pronunciation: hallo, hello, hullo.
What’s the transcription?

35. Here are some common English words in IPA transcription. Rewrite them
in ordinary spelling and/or read them out. Make sure you are clear why the
symbols are used in the way they are.

[lif] [lIv] [bˆt] [bæt] [but] [lUk] [kOd] [kO:d] [gUd]

[kˆntrI] [sEk@nd] [stˆm@k] [EksEpt] [bUz@m] [hæbIt].

36. The pronunciation of the vowels in the following words differs from RP in
the localities indicated. Insert the vowel symbol appropriate to the local
pronunciation.

foot (Scotland) [ ], fast (N. England) [ ], fast (USA) [ ] , up (N.

England) [ ], cap (S. Africa) [ ], off (Buckingham Palace) [ ], cross
(USA) [ ].

37. Complete this vowel chart by inserting [I, {, U, ¡, O:, ˆ].

y i u Ì

Ô e o Ê
@
˚ E

œ a A

3.07 Diphthongs

In spite of their various crucial differences, all the vowels studied up to now have
one thing in common: the tongue stays in the same position all the time any of
them is being articulated.

14 Vowels
But there are also vowels like the one in how, during which the tongue moves
from one position to another (passing through a series of intermediate points as it
does so). In the case of the ow of how, the starting-point is an [a]-type low front
position, and the end-point is an [u]-type high back position. If you say how very
slowly to yourself, the difference between the beginning and end of the vowel
should be clear, as should the intermediate stages.

A gliding vowel like this is known as a diphthong. The term monophthong

designates vowels in which the tongue position doesn’t change. (Sometimes the
expression “pure vowel” is used for these, but is perhaps best avoided, given the
implications of superiority which it conveys.) The IPA represents diphthongs by
means of a sequence of two symbols, the first indicating the starting-point and the
second the finishingpoint. So how would be transcribed [hau]. [a] and [u] are
referred to as the first and second elements of this diphthong. (Note that the two
symbols are not joined in a digraph, and remember that digraphs and diphthongs
are different things.) The movement of the tongue can be indicated on the vowel
quadrilateral:
English abounds in diphthongs. Here are some more. In each case repeat the
example
y i u Ì

Ô e o Ê
@

˚ E O ˆ

A
Fig. 21 œ a A
word to yourself so as to get a clear feel for the starting and finishing points.

In the following cases, the second element is high and front:

[ai] as in find, right

[Oi] as in toy
[ei] as in day

As the tongue doesn’t actually move all the way to the Cardinal No. 1 point, these
diphthongs are often transcribed more realistically with the retracted [ I] symbol:
[aI], [OI], [eI].

The second element of the diphthong found in so, below, home or though is high
and back. But the first element, for most RP speakers, is similar to schwa.
(Check whether this applies to your own pronunciation.) So the usual

Vowels 15
 transcription is [@u] or [@U]. [U] is sometimes used instead of [u], for the same
reason that [I] is sometimes used instead of [i] (see previous paragraph).

Many languages — standard French and Spanish for example — have no

diphthongs. A typical anglophone mispronunciation of the vowels in words like
bébé or rose in French or de or solo in Spanish is to “diphthongize” them, i.e.
supply them with a glide leading to a second element: the result in the former
case is a diphthong [ei] instead of a monophthong [e] and in the latter [ @u]
instead of [o]. There is no surer give-away of an English accent — so this is a
point of pronunciation which is worth paying a lot of attention to.

The same also applies to German. Even though this language does have some
diphthongs of its own (e.g. in mein, neu and Haus), the vowels of so or Weh are
monophthongs, and English [@u] or [ei] are not satisfactory substitutes for anyone
who aims at passing for a native speaker.

Anyone who has problems with words like bébé can console themselves with the
thought

Be careful not to confuse spelling and pronunciation. The term

diphthong relates only to pronunciation. A sequence of two or more
vowels in the spelling isn’t necessarily evidence for a diphthong. (And,
as was pointed out in 3.04, if two vowels are joined together as in [æ], the
result is known as a digraph.) Thus, in French, ou isn’t a diphthong —
it’s just a sequence of vowel letters corresponding to the monophthong
[u]. And eau isn’t a triphthong: it too corresponds to a monophthong —
this time [o]. In Spanish, beware of books that talk about the
“diphthongs” in bueno, tiene, etc. These words don’t contain diphthongs
— any more than English wet or yet do. (See 3.08 for more discussion.)
Again, ue, ie are just sequences of vowel letters.
that learners of English as a foreign language have the opposite difficulty, and
may well have even greater problems: learning to imitate the exact starting and
finishing points of English diphthongs is a very difficult task. It’s all the more
difficult as quite slight variations in the precise tongue movement for some of the
English diphthongs can change their nature greatly: anglophones are well attuned
to such differences, as these can be informative about a person’s regional or social
origin.

Thus if the o of so is pronounced [Eu] rather than [@u] (i.e. with the first element
fronted), the effect is perceived as “refined” or “posh”. By contrast, if the first
element is again fronted but also lowered slightly (to [æ]), the resulting [æu] is
felt to be “vulgar” or “uneducated”. Not that there is anything inherently posh or
otherwise about the sounds themselves: it’s a matter of who uses them and in

14 Vowels
 what context. If the “refined” [eu] just referred to is used instead of the RP [au]
in how, it instantly takes on “sub-standard” associations itself.

English has monophthongs as well of course, as we saw earlier (in words like
head, had, hid, for example). But so prone are RP-users (and Southern British
speakers in general) to “diphthongize” vowels that the long monophthongs in feed
or food are preceded by a distinct “onglide” — i.e. a movement of the tongue
giving a hint of the presence of another element besides the vowel itself. So not
only is the vowel of RP see more retracted than Cardinal No.1 (illustrated by
French si), it’s also slightly diphthongized: [Ii].

Similarly for English route compared with Cardinal 8 (or French route): [Uu].
You can see from Fig. 22 that the tongue moves through only a short distance. So
these
Fig. 22 i u

I
U

e o

sounds are considered as “diphthongized vowels” rather than fully-fledged

diphthongs, and are normally transcribed just [i] and [u] for convenience. At
least, that applies to RP: for some London speakers, this onglide is of much the
same duration as a true diphthongal element, and the distance travelled by the
tongue is greater: therefore the transcription [@i] is justified for diphthongs like
the one heard in the Cockney pronunciation of feed.
Exercises
i u

e o
@

E O

Fig. 23 a A
38. Explain the difference between a vowel, a pure vowel, a monophthong, a
diphthong and a digraph. Then say why the phrase “vowels and
diphthongs” isn’t quite logical.

39. Making use of the IPA, try to characterize:


Northern Irish pronunciation of the vowel of face (RP [feis]).

Vowels 15
 
Northern English pronunciation of the same vowel.


American pronunciation of the vowel of so.


Australian pronunciation of the vowel of bean.

40. Transcribe the diphthongs heard in the RP pronunciation of the following

words (i.e. by giving the IPA symbol for each of the two elements):

bough [ ] aisle [ ] toe [ ] buy [ ] weigh [ ]

though [ ] buoy [ ] soap [ ] town [ ] same [ ]

41. Rewrite the following English words in ordinary spelling:

[taid@l] [klaim@] [nOiz] [fEl@U] [kwai@] [flau@]

[prOdˆkt] [aid@daun] [kæt@lOg] [mIsail] [tEl@f@Un].

42. (a) For Spanish specialists, Does the ei in deinde represent a diphthong
like the one in RP day, or a sequence of two separate vowels (as found in
RP pay in)?

(b) For German specialists. Transcribe the diphthongs in mein, neu and
Haus. (N.B. each of these is subtly different from the RP diphthongs in
mine, boy and house — can you capture any of these differences in your
transcription?)

(c) For French specialists. Why is it wrong to describe the vowel of Fr

peu as a diphthong (or the vowel of soeur as a triphthong)? And can you
specify the difference between the vowels of English high and French haï ?

43. The [i@] of RP fear is sometimes described as a centering diphthong, the

[ai] of my and the [au] of house as closing diphthongs. Can you explain
why?

3.08 Length and Nasalization; Diacritics

Changing the height, fronting, or rounding of vowels (i.e. their “quality”) isn’t the
only way of modifying them. Consider the words forward and foreword (the
latter meaning “preface in a book”). The main difference between them is that the
[@] in the second syllable of foreword is longer than the corresponding [ @] in
forward. Vowel length isn’t often distinctive in this way in English — but there

14 Vowels
are languages in which it’s important, and in any case the IPA needs a way of
indicating it. This is done by placing the mark [ :] after a long vowel: [@:] in this
case. As you know, the vowel of the first syllable of forward/foreword is itself a
long version of [O], so the two words are transcribed [f O:w@d] and [fO:w@:d]
respectively.

As an alternative to the rather elaborate [:], a simple colon [:] is sometimes used.
And when writing phonetic symbols by hand, the colon is the obvious way to
represent the length mark.

Note that for cases where a vowel is intermediate between “short” and “long”, the
IPA provides a “half-length” mark [ ;] (more simply a raised dot .). Exercise 46
will give you an opportunity to use this sign.

German also sometimes distinguishes words by means of vowel length: the first
syllables of trennen and Tränen are a case in point.

In both English and German, the difference of quality between [ I] (fit, im) and [i]
(feet, ihm) is accompanied by a length difference, with [i] usually slightly longer
than [I]. Indeed older studies of English phonetics treat the length difference as
the primary feature: they refer to the vowel of fit as a “short vowel” and the
vowel of feet as a “long vowel”, and transcribe them as [fit] and [f I:t]
respectively, making no use of [ I] at all. Nowadays, however, it is believed that
the quality difference is the one that hearers mainly listen out for, so the [i]/[ i:]
style of transcription is less common than it used to be.

A second way of modifying a vowel is to lower the velum while it’s being
articulated. Some of the air will continue to pass through the mouth, and the
positions of tongue and lips will have their usual effect on vowel quality. But
some of the air will now also pass through the nasal cavity, so that its distinctive
resonances will be superimposed on those of the mouth and pharynx. Hence the
nasalized vowels found in French — and in a range of other languages including
Portuguese, Polish, and many languages of Africa and the Indian sub-continent.
Note that the term nasalized vowel is preferred to nasal vowel by phoneticians:
they are really just “oral” vowels with nasality superimposed.

In order to indicate vowel nasalization, the IPA places a tilde [~] over the relevant
vowel symbol, e.g. [ẽ].

Any vowel can be nasalized. Standard French has just four nasalized vowels:
[]̃ (un, parfum), [] (bon, dont), [] (vin, pain), [] (blanc, lent).
French had more in the Middle Ages, and in this respect resembled modern
Portuguese which still

Vowels 15
 has, among other things, a nasalized [ i] (sim, “yes”) and nasalized
diphthongs like [au] (não, “no”). Notice how in Portuguese (unlike
Spanish where it occurs only over n), the tilde is used in ordinary spelling, as
well as in phonetic transcription, to identify certain nasalized vowels or
diphthongs (it’s placed over the first element of a diphthong).

At this point you might like to note a useful term covering all the various
lengthmarks, tildes, accents, and other items which are placed above, below,
before or after phonetic symbols in order to indicate a modification of some sort.
Collectively they are known as diacritics, or diacritical marks. An individual
accent, length-mark or whatever is a diacritic.

Further examples of diacritics are the superscript [ h] denoting aspiration (see

4.07), the subscript circle [8] used to indicate certain rare kinds of voiceless
consonant (see 4.09), or the “tooth mark” [9] placed under dental [t ] or [d] (see
4.02).
9 9

Exercises

44. Rewrite the following in ordinary spelling:

[kO:d] [kA:d] [kE:d] [rEp@twA:] [sO:lzbrI] [k@:t@sI]

45. [i] and [i] aren’t the only pair of English vowels that differ in length as well
as in quality. For each of the following pairs, say which word has the longer
and which the shorter vowel:

good, food
cad, card
pet, paired.

46. A characteristic feature of English pronunciation is the way in which the

length of some individual vowels can vary according to whether they are
followed by a voiceless consonant, a voiced consonant, or no consonant at
all. Arrange the words in each of the following groups in terms of the length
of their vowel, and see whether you can produce a general rule for what is
going on. (N.B. you’ll need the half-length mark as well as the length mark
— and of course no mark at all for the shortest vowels).

bee, bead,
beat food, boot,
do cart, car,
card caught,

14 Vowels
 cord, caw her,
heard, hurt.
47. A properly written tilde starts low, then rises, then falls, and then ends up on
a rise. Some people, quite wrongly, start high and end on a fall. Check that
you are writing your tildes correctly.

48. (For French specialists in particular.) Though standard French has only four
nasalized vowels, there are considerably more than four ways of spelling
them. How many differently spelt French words can you find that are all
pronounced [s]? And how many that are all pronounced [s]?

3.09 Semi-Vowels

In English, a short version of [i], spelt y, occurs in yet and a short version of [u],
spelt w, occurs in west. If you say yet giving the y the length of an ordinary
vowel, you’ll notice that it is in fact the same as [i] in quality (high, front,
unrounded). Similarly with the w of west (high, back, rounded.)

These truncated high vowels are, for obvious reasons, known as semi-vowels.
They always occur at the beginnings or ends of syllables, just as consonants do,
never as the centre of a syllable: thus met, pet, set (consonants) and yet (semi-
vowel). In short, semi-vowels are sounds which are articulated like vowels but
positioned in words like consonants. (Indeed the term semi-consonant is
occasionally applied to them.)

The symbol [w] has been adopted unchanged by the IPA to represent the high
back semi-vowel of west ([wEst]). However, the [y] symbol isn’t available for
yet, as it’s already in use for Cardinal 9 (the vowel in rue and über). Instead [j] is
used: so yet is transcribed [jEt]. Like [y], [j] in the IPA has the same value as in
German spelling (cf. Jahr, etc.). But it’s not called “jay”, as this would be too
suggestive of the sound it has in English. Instead you should read it as “yod”.
[w] on the other hand is still read “double-you”.

Note that [j] isn’t always spelt with the letter y in English. Often u serves to
indicate the combination [j + u], as in unique [juni:k], and occasionally other
spellings indicate the presence of [j] (e.g. i in senior [si:nj@], or ew in few [fju:]).
A point to note about [w] relates to words like when, which, whether. In some
accents (notably American, Scots, and conservative RP), [w] in such cases
becomes voiceless: the wh in the spelling is an attempt to represent this, in the
absence of any distinctive letter. A special symbol has been devised for the IPA,
however — an inverted w: [˜]. So for some English speakers there is a
difference between [wain] (wine) and [˜ain] (whine), or [weilz] (Wales) and [˜eilz]
(whales).

Vowels 15
 French specialists should know that as well as [j] and [w], French (unlike
English, German or Spanish) has a third semi-vowel — a shortened version of
[y]. The phonetic symbol is [ ] (“turned h”). This is the sound that is usually
represented in ordinary French spelling by u when followed by another vowel,
e.g. in puis or nuage. It should be distinguished from [w], which corresponds to
ou in the spelling. So there’s a difference between Louis [lwi] and lui [li], and
between “bury oneself” — s’enfouir, with [w]: [sfwir] — and “run away” —
s’enfuir, with [ ]: [sfi]. The best way to practise a word like puis is to
start by pronouncing it with two full vowels [py] + [i] and gradually shorten the
[y] so that you end up with a word of one syllable, containing a semi-vowel and
just one vowel: [pi]. But make sure you’re pronouncing the [y] as [y] and not
as [u]: otherwise you’ll be saying [pwi]!

Notice also that [j] can occur at the end of words in French: travail [travaj],
pareil [parj], grenouille [grnuj]. [aj, j, uj] aren’t diphthongs: in [aj] for
example there is a rapid transition from low [a] to high [j], with minimal time
spent on the intervening stages. This is quite unlike the much more “drawn out”
diphthong of English high, with its gradual transition. (The case of haï —
mentioned in Exercise 42c — is different again: this word has two syllables each
consisting of a full-length vowel [a + i], and the same amount of time is spent on
each.)

Exercises

49. Which of the following English words contain the semi-vowel [j]?

Europe, ensure, new, pursue, revolution, beauty, behaviour,

failure, union, duty, salute, suit, onion, piano, absolute.

50. Rewrite these English words in ordinary spelling:

[jOt] [fju:d] [m@nju@] [p@sju:] [j@:] [kwait] [kweint] [wu:m]

[wOt] [kwi:n] [kju:].

51. Some English accents distinguish between [wEt] and [˜Et]. What do these
two transcriptions correspond to in ordinary spelling?

52. Do (a) German and (b) Spanish have both [j] and [w], just one of them, or
neither of them? Give example words when appropriate.

14 Vowels
 consonants
[4]
The story so far:

Vowels are classified in terms of three parameters affecting the shape of the
oral cavity: tongue height, tongue fronting, lip rounding.

Diphthongs are vowels involving a change of tongue position.

4.01 Classifying Consonants

You’ll recall that the basic way in which vowels and consonants differ is that,
whether it’s voiced or voiceless, producing a consonant involves some sort of
constriction above the level of the glottis, with ensuing airstream turbulence. The
obstruction may be partial (as for s), intermittent (as for a trilled r) or complete
(as — momentarily — for p). But consonants, unlike vowels, always involve a
“supra-glottal” constriction of some kind.

Consequently, although consonants are also classified according to three

parameters, these are different from the ones you have been studying for vowels.
One consonant parameter has already been described in 2.04: voicing.

There are two others: place of articulation and manner of articulation.

Respectively they specify where the airstream is constricted and how it’s
constricted.

4.02 Place of Articulation: Bilabials, Dentals, Alveolars and Velars

Where, for each consonant, is the point of narrowest constriction along the vocal
tract?
Compare three pairs of consonants which occur in all European languages:

[p, b], [t, d], [k, g].

Consonants 41
 [p] and [b] are both produced by means of a constriction involving the lips, as is
obvious if you just say [apa], [aba] slowly to yourself. The vocal folds continue
to vibrate in the case of [b], but not in the case of [p]: otherwise there’s no
difference between them, and the following diagram, which doesn’t show the
vocal folds, applies equally well to both.

Fig 24

[p,
b]

Consonants like [p] and [b] are BILABIAL (this word is simply the Latin for
“both lips”).

For [t] and [d] the lips aren’t involved in the obstruction of the airstream. This
time (in English at least) the constriction is produced by bringing the tip of the
tongue into contact with the teethridge (alveolum). Say [ata], [ada] to confirm
this.

Fig 25

[t,
d]

So English [t] and [d] are ALVEOLAR. (Strictly speaking, as the tongue is
involved as well as the teethridge, the term ought to be lingual-alveolar: however,
most consonants involve the tongue in some way or other, so there’s no point in
specifying lingual- each time.)

Finally, for [k] and [g], the constriction involves the back of the tongue, which is
brought into contact with the soft palate (velum). Therefore these are VELAR
consonants. Although this part of the oral cavity is less easy to monitor than the
area around the lips and teeth, you can get a rough impression of the tongue
position by saying [aka] and [aga] slowly to yourself. But only a diagram can
show the surprising extent to which the back of the tongue is raised for velars.

Fig 26

40 Consonants
 [k, g]

There’s a slight but significant difference between [t, d] in English as compared

with the corresponding consonants in French, Spanish and Italian. In the last
three languages the tongue is always placed against the back of the upper front
teeth for [t, d], not against the teethridge, and this results in a slightly “sharper”
sound. The term DENTAL is used in this case. (German [t] and [d] are alveolar,
just as in English.) Fig. 27 shows the difference between the places of
articulation for dental and alveolar consonants.

Fig 27

Dental [t, d] 9 9 Alveolar [t, d]

As you can see from Fig. 27, the IPA specifies the dental/alveolar difference by
placing a “tooth mark” below the dentals: [ t9 d9].This is obviously useful when
languages are being compared. However, in everyday transcription of French,
Spanish or Italian, the mark can be omitted: [t, d] are always dental in these
languages, so it’s unnecessary to specify the fact every time one of them comes
up.

Though not exactly one of the most crucial pronunciation differences between
English or German on the one hand and French, Spanish or Italian on the other,
the [t d]/[t9 d9] distinction isn’t difficult to achieve, and it’s worth the effort to
make your pronunciation that little bit more authentic. Just remember to place
your tongue fractionally further forward when you make the consonant.

There’s no difference in place of articulation between English and the other

languages in the case of [p, b] or [k, g].

Now we can start building up the IPA consonant chart. Here are the four places
of articulation mentioned so far. As with the vowel chart, the left-hand side
corresponds to the front of the mouth, the right-hand side to the back.

FIG. 28

bilabial dental alveolar velar

Consonants 41
 Exercises

53. Resisting the temptation to look at the text again, place the following
consonants in the appropriate square in Fig. 28 above (say the sound to
yourself if in doubt). Voiceless and voiced pairs go next to one another in the
same square (voiceless first).

[k, d, p, g, t, b, d, t9 9 ]

54. Complete the following characterizations by inserting, in each case, two of

the following terms: voiced, voiceless, bilabial, dental, alveolar, velar. N.B.
voicing comes before place of articulation in such characterizations.

[b] is a ____________________ ____________________ consonant

[k] is a ____________________ ____________________ consonant

[d] is a ____________________ ____________________ consonant

[t] is a ____________________ ____________________ consonant.

55. Give the IPA consonant symbol corresponding to each of the following
characterizations:

voiceless alveolar _____________________________

voiced velar _________________________________ voiced

dental ________________________________ voiceless

bilabial ______________________________

4.03 Manner of articulation.

If we now take two further consonants — [s] and [z] — and specify them in
terms of voicing and place of articulation, we get the following characterization:
[s]: voiceless alveolar

[z]: voiced alveolar.

40 Consonants
Just like [t] and [d]: airstream obstruction at the alveolar ridge. So what’s the
difference between [s, z] on the one hand, and [t, d] on the other?

The answer is that different kinds of obstruction are involved. Or, as the
phoneticians put it, [s] has the same place of articulation as [t], but a different
manner of articulation.

Let’s consider in more detail how consonants like [p, b, t, d, k, g] are produced
(now that we’ve seen where they’re produced). Then we’ll consider how the
“mechanism” differs for consonants like [s, z].

4.04 Stops (or Plosives)

[p, b, t, d, k, g] all involve a complete blockage of the airstream, albeit one of

very short duration. Taking the nonsense word [apa] again, here are your
instructions for articulating the [p], in three “slow motion” stages.

Say the first [a] vowel. Then:

1. Close the lips firmly.

2. Keep them closed while continuing to allow air to come up from
the lungs: air pressure builds up behind the closed lips.
3. Now part the lips. The air under pressure will suddenly be released,
generating a disturbance (a sound-wave) in the surrounding air.

Then say the second [a] vowel.

These three stages are known respectively as (1) closure, (2) hold, (3) release.
Fig. 28 shows them in diagram form.

hold
close
FIG. 28

The X-ray tracings that were given in 3.01 represent the “hold” stage. But it is of
course not until the moment of “release” in stage 3 that the actual consonant
sound is perceived, thanks to the sound wave that reaches the ear a fraction of a
second later. In stages 1 and 2, there’s no sound. Say [apa] to yourself extremely
slowly, and you should become aware of this period of silence, and also, during
it, of the build-up of air-pressure behind the point of closure. But in normal
speech we aren’t conscious of any of these processes: they occur far too quickly

Consonants 41
 and automatically for us to notice them, and in any case we’re too busy thinking
about what we want to say next.

The same three stages are involved for [b], [t, d], [ t d] and [k, g]. For the
dentals the airstream is blocked at the teeth (by the tip of the tongue), for the
alveolars at the teethridge (by the blade of the tongue) and for the velars at the
soft palate (by the back of the tongue). But the “close-hold-release” mechanism
is identical in all cases.

Consonants like these are known either as STOPS (this term relates to the
blockage of the airflow in stage 1) or as PLOSIVES (this relates to the explosive
release of air in stage 3). But both terms refer to the same category of consonant:
nowadays, stop is used more often than plosive.

4.05 Fricatives

[s] and [z] don’t involve a complete obstruction of the airstream. The blade of the
tongue is held against the teethridge as for [t, d], but sufficiently loosely for the
air to be able to force its way through. Nonetheless, there is enough resistance at
the point of constriction for turbulence to be caused and a sound wave generated.
So [s] and [z] are examples of consonants with partial obstruction of the air-
stream, and are known as FRICATIVES (this term refers to the friction caused
by the air as it passes through the narrow gap).

The following diagram should make the difference between [t] and [s] clear. (The
voiced pair [d] and [z] would look the same, as voicing isn’t shown here.)

FIG. 29 Stop [t] Fricative [s]

So now we can give a complete characterization of the sounds considered so far.

This time we specify (a) voicing, (b) place of articulation, AND (c) manner of
articulation. Thus:

[s] is a voiceless alveolar fricative

40 Consonants
 [t] is a voiceless alveolar stop[b] is a voiced

bilabial stop.

Many other fricatives can be produced at various points along the vocal tract. In
all cases, the airstream is forced through a narrow channel or aperture, and
there’s always the same hissing or scraping effect.

Take for example [T] and [D]. The point of articulation is dental, but the manner
of articulation is fricative, not stop. You should easily be able to feel the air
passing between the tongue-tip and the back of the teeth, and the loose contact
between them.
(With voicing added in the case of [ð] of course.)

It follows from the difference between these two manners of articulation that
you can prolong a fricative until you run out of breath (try it with [s]), whereas
a stop is an instantaneous sound (an explosion can’t be prolonged). Try to
prolong a [t]: you won’t be able to (and saying [t@:] is cheating!).

Here is an expanded version of our embryonic IPA consonant chart (still far from
complete though), with all the consonants discussed so far — and a few more.
This time there are two extra points of articulation (underlined), and fricative has
been brought in as a second manner of articulation.

FIG. 30

labio post-
bilabial dental alveolar velar
dental alveolar

STOP p b t d t d k g

FRICATIVE f v  D  

Note the two additional pairs of fricatives here:

For [f] and [v] the air is forced between the upper lip and the lower front teeth
(hence labio-dental at the top of the column).

[f, v]

FIG. 31

[S] is the voiceless consonant commonly written sh in English, ch in French, and

sch in German. The IPA symbol is read “esh”, or “long s”. Its voiced partner [ Z]

Consonants 41
 (read “ezh” or “long z”) is of frequent occurrence in French (spelt j, as in jour),
but is relatively rare in English, where it’s the second consonant of leisure, or the
last one of camouflage. In German [Z] is only used in words taken from French
(Passage). Neither [S] nor [Z] occur in standard (Castilian) Spanish, but [ Z] is a
common pronunciation of -ll- (e.g. in calle) and of -y- (e.g. in yo) both in Latin
America and in Spain (where, however, it is frowned upon as “substandard”).

[S,
Z]

FIG. 32

As you can see from Fig. 32, the constriction for [ S] and [Z] involves the front,
rather than the blade of the tongue, and is located just behind the teethridge.
Hence the term post-alveolar at the top of the column (post- meaning “behind”).

So [s] is alveolar and [S] is post-alveolar. Say them in succession and you should
feel the difference.

There are still some empty squares in Fig. 30. They can be filled up with
pronounceable consonants — though not ones that are encountered in standard
English (or French). The following are worth knowing about — particularly if
you are studying the language in question.

The voiced bilabial fricative (IPA symbol ¬, read “beta”). In Spanish, this often
replaces the voiced bilabial stop [b], in particular between vowels (haber [a¬Er],
se baja [se¬axa]). The closure made by speakers for the b isn’t quite complete,
and this results in a fricative [¬] rather than a stop [b].

The voiceless velar fricative. This is spelt ch in German and j in Spanish —

and occurs in words like Buch and baja. This consonant has the same point of
articulation as k, but, again, partial not complete closure of the vocal tract. The
IPA uses the symbol [x] to represent it: so Buch is transcribed [bux] and baja
[baxa].

A voiced velar fricative, transcribed [γ] — read “gamma” — replaces the

voiced velar stop [g] between vowels in Spanish words like luego [lweγo] or
pagar [paγar].This parallels the replacement of the bilabial stop [b] by the
fricative [¬].

Non-specialist accounts of pronunciation try to describe sounds like [g] or [γ]

by means of expressions such as “hard vs. soft g”. Hopefully you’ll agree that
stop and fricative are a lot more precise and informative, at least when used in
conjunction with the name of a place of articulation. “Soft g” could as easily

40 Consonants
 refer to the g of germ as to [γ], but a term like “voiced velar fricative” is
unambiguous.

Exercises

56. Complete the blank diagram as follows:

a. Insert the names of the points of articulation studied so far, involving

(listed here in random order):

(1) tip of tongue + teethridge

(2) lower lip + upper lip

(3) back of tongue + soft palate

(4) lower lip + upper teeth

(5) front of tongue + hard palate

(6) tip of tongue + back of upper front teeth.

Use the appropriate technical terms, making sure that you insert them in the
right squares and arrange them in the correct order.

b. Insert the two manners of articulation studied so far: (1) with

complete closure, (2) with partial closure.

c. Insert the sounds represented by the following IPA symbols: [x]

[z]
[S] [¬] [γ] [t] [g].

57. The consonants in each of the following pairs are alike in some respects,
but different in others. Specify the resemblances and differences.
Example: [p] and [b]. Both are bilabial. Both are stops. [p] is voiceless,
[b] is voiced.

Consonants 41
 [z] and [d]

[s] and [d]

[t] and [k]

[v] and [g]

a. and [¬]

[v] and [¬]

[b] and [v]

[s] and [S]

58. Though the voiceless velar fricative [x] doesn’t occur in RP (or most other
varieties of English) it is a feature of Scots English and Liverpool English,
and occurs in some Irish place and personal names. Think of examples of
words containing it.

59. If the IPA symbol [x] represents a velar fricative (unknown in most
varieties of English), how would a word containing x in the spelling be
transcribed? Extra, for instance. (Think carefully about what this letter
corresponds to in the pronunciation.)

60. Non-technical books sometimes try to describe Spanish [ ¬] by talking

about “a cross between b and v”. What are they getting at?

61. Rewrite in ordinary (English) spelling:

[lA:fs] [kl@UDz] [hi:Dnz] [peISns] [@UT] [w@:Tl@s] [Tˆm]

[f{Sen] [kl@UZe] [lˆS@s] [beIZ] [ju:ZU@l] [S@rA:d] [lˆkS@rI].

62. Transcribe, using IPA symbols:

cake, bouquet, bagpipes, oaths, gnaw, sash, pleasure, Asia, breath,

breathe, photo, physics, warmth, theatre, though.

40 Consonants
4.06 More Places: Palatal, Uvular, Pharyngeal, Glottal

So far six places of articulation have been introduced, from bilabial to velar. In
actual fact, a constriction giving rise to a consonant can be made at any point
along the vocal tract by bringing two “articulators” into contact. The tongue in
particular being highly mobile, different parts of it can close against different
parts of the roof of the mouth, giving a range of subtly varied consonants, all of
which occur in some language or another. However, the IPA contents itself with
ten places of articulation, so here is the chart again with the remaining four added
(underlined). Finer intergradations can be indicated if required by use of pre- (“in
front of”) or post-
(pre-velar, post-palatal, etc.)

FIG. 33
bilabia labiodenta alveola postalveola pharyngea
l dental r palatal velar uvular l glottal
l r

STOP ?
FRICATIV C
E h

Most of these additional places of articulation are relevant to at least one of the
wellknown European languages, and you should note the existence of the others.
So a few further symbols have been inserted. Here is some information about the
sounds that they represent.

[h] (read as “aitch” — it’s incorrect to call it “haitch”, whether you’re using the
ordinary alphabet or the IPA!). This is for the first consonant of English or
German hat. It’s classified as a voiceless glottal fricative: the vocal folds
constrict the airflow, but aren’t close enough together to vibrate. Effectively [h]
is a whispered version of the (voiced) vowel that follows it: compare the [h] of
hard with the [h] of he: during the articulation of the [h] the tongue takes up the
appropropriate position for the vowel that is to follow.

[ç] (“c cedilla”) is for a voiceless palatal fricative, with a constriction

between the centre part of the tongue and the hard palate. It occurs in
English as a variant of [h] in words like hue (i.e. [çju:] — with the tongue
high and front in anticipation of [j]). But the [ç] symbol is used only in very
detailed or “narrow” transcriptions of English: normally hue would be
written with [h].

[ç] is an important sound in standard German, where it alternates with its

neighbour on the consonant chart, the velar fricative [x]. Both correspond to
ch in the spelling: the [ç] variant occurs after high or mid front vowels (ich,
Bücher, echt), the [x] variant after other vowels (Buch, Bach, hoch). Note
that the diminutive suffix -chen is pronounced with [ç] not [x]: tauchen (“to
dive”) is [taux@n], but Tauchen (“little rope”) is [tauç@n].

Consonants 41
 [?] is for the glottal stop. (The symbol — read “glottal stop” — is adapted from
the Arabic alphabet: in Arabic glottal stops are important enough to have a letter
to themselves.) As the name suggests, this consonant is produced at the glottis
(i.e. between the vocal folds), using the same three-stage manner of articulation
as all the other stops. One example of its use in English is the exclamation
commonly spelt uh-uh (an expression of wary agreement, slight surprise, or
refusal, depending on the intonation). Phonetically this might be transcribed [ @?
@]: after the first schwa the vocal folds are (1) closed tightly, then (2) held
together while air-pressure builds up beneath them, and finally (3) released
suddenly. Though there’s no perceptible “explosion”, the sudden release gives a
distinctively sharp onset to the second schwa, and the silent pause between the
two vowels is very noticeable. Another situation where a glottal stop can used in
English is to separate adjacent vowels in phrases like India [?] and Pakistan
(though many speakers insert an r instead: India [r] and Pakistan).

Separating adjacent syllables (the second of which begins with a vowel) is a

common function of the glottal stop in German. Thus in der [?] Apfel, it would be
positively incorrect to run the first word on to the second by means of a “linking
r”. In French, on the other hand, it would be just as incorrect to insert a glottal stop
in equivalent phrases (e.g. after the r of cher ami). Similarly a more authentic
effect is obtained in le homard if a glottal stop is avoided between the two words,
and the [@] of le run directly on to the [O] of homard.

That leaves the places of articulation labelled UVULAR (involving the tip of the
soft palate) and PHARYNGEAL (involving the wall of the pharynx). The r
sounds used in standard French and German are uvular, and will be discussed
separately in 4.11. Meanwhile, though the following uvulars and pharyngeals
don’t occur in English or any of the European languages you are studying, you
may like to note them for interest’s sake (and remember that they are important
in the languages in which they do occur).

In some languages (Arabic for example), stops occur which are produced by
closing the back of the tongue against the uvula. The effect is that of a retracted,
rather hollow-sounding [k] or [g]: the IPA uses the symbol [q] for the voiceless
uvular stop — as does the ordinary Western alphabet when Arabic names are
written in it: the Al Qaida network, for instance. [G] is the symbol for the
corresponding voiced uvular stop.

It is also possible to pull back the root of the tongue so as to create a

constriction between it and the rear wall of the pharynx, giving rise to the
pharyngeal fricatives which are a feature of Arabic and Hebrew (midway
between [x] and [h]). These can
4
be voiceless [h] (“crossed h”) or voiced [] (“reversed glottal stop”). It’s not
possible for most people to to retract the tongue root far enough to close the
pharynx completely, so pharyngeal stops don’t occur.

40 Consonants
Exercises

63. You are now in a position to enter on the IPA chart all the stops and
fricatives that have been presented, so do so — together with the names of
the ten points of articulation and the two manners. You might as well
include the “exotic” uvulars and pharyngeals, making 11 stops and 15
fricatives altogether.

64. Glottal stops are a conspicuous (and often frowned-on) feature of some
nonstandard varieties of English, in words like daughter or butter — or
glottal itself. Say how they would be used, and in which accents.

65. You may have noticed in the above text the two spellings uvula and uvular.
This isn’t a misprint, and they are not interchangeable: work out when each
should be used.

66. With reference to as many as possible of the following languages (in their
standard version), say whether these sounds are part of the repertoire or not:
English, French, German, Spanish. [x], [ç], [?], [q], [h].

4.07 Aspirates

Here’s an final point about stops which has been kept from you so far. But it
should be noted, as it’s another important difference between English and
German on the one hand, French, Spanish and Italian on the other.

Say the English word pa very slowly, clearly and deliberately. You may notice
that the [A] vowel doesn’t begin immediately after stage (3) of the stop (the
“release” stage). Instead there’s a short pause before the vocal folds begin to
vibrate, during which air passes freely through the glottis. Hence the effect of a
puff of breath, or brief [h] sound, between the [p] and the vowel. The same
applies to the other voiceless stops, e.g. in tar or car.

Consonants 41
 If the same syllables are pronounced in the French, Spanish or Italian way, the
vowels begin immediately after the release of the [p], [t] or [k] — no delay, so no
suggestion of a puff of breath or an [h]. It is incorrect in these languages to delay
the onset of the vowel.

Voiceless stops followed by a puff of breath (more technically “followed by

delayed vowel onset”) are said to be aspirated. Those without the “aspiration”
are — obviously — said to be unaspirated. The IPA uses a superscript h to
indicate aspiration: [ph], [th], [kh], but there’s no need to include this detail in an
ordinary transcription of English. Diagramatically the process can be represented
as follows (aspirated stop in Fig. 33, unaspirated stop in Fig. 34):
release
Fig. 33

p h A:

aspiration voicing

Fig. 34
release

p A:

voicing
Remember: In English and German voiceless stops are aspirated at the beginning
of stressed syllables. In French, Spanish and Italian they are never aspirated. If
you disregard this point, your pronunciation won’t sound authentic, however
good it may be in other respects, and native speakers will notice: when French
people, for example, try to imitate a “British accent”, they go to great lengths in
an attempt to produce aspirated stops.

Occasionally it’s necessary to specify that a stop is unaspirated. The symbols

[p=], [t=], [k=], are used for this purpose.

Note that in English and German, voiceless stops are unaspirated if preceded by
a fricative. Thus: [st=] star, [sp=] spare, [sk=a] sky in English; [t=at]
Staat, [p=its] Spitze in German. The presence of [s] or [ ] at the beginning of
the syllable allows more time for the voicing of the vowel to be initiated, so
there is no reason for any delay in onset.

The traditional way of checking that you’re making voiceless stops without
aspiration is to hold a thin piece of paper in front of your mouth while saying
[p=a], [t=a], [k=a]. If you aspirate the consonant, the puff of breath will make the
paper jump; if you’re saying it correctly, the paper will remain motionless,
however loudly you speak. A time-honoured practice sentence for French is:

40 Consonants
 Tintin, ton thé t’a-t-il ôté ta toux? (Make the French [t] dental as well as
unaspirated.)

4.08 Affricates

You may recall from 4.05 that [Z] (as in camouflage) isn’t a very common sound
in English and that [S] and [Z] don’t occur at all in standard Castilian Spanish.
But what about the ch of Spanish mucho, the second g of most people’s
pronunciation of garage or the dg of everyone’s pronunciation of porridge? Or
the two ch’s of church? These sounds do contain [S] or [z], but the fricatives are
combined with a [t] or a [d], so that ch or dg actually represent composite sounds
known as AFFRICATES.

These consist of a stop immediately followed by a fricative with the same or

similar point of articulation: [t + S] or [d + Z]. Compare hash (fricative) with
hatch (affricate) or leisure (fricative) with ledger (affricate), and you will
appreciate the difference. Say the tch or the dg very slowly, and you’ll see how a
[t] or [d] is fused with the following [S] or [Z]. To indicate the close fusion of the
stop and the fricative, the IPA symbols for them are digraphs: [ tS ] and [dZ]. The
official names are “T-Esh ligature” and “TEzh ligature” respectively. These two
affricates are classified as post-alveolar (this is the point of articulation of the
fricative element): [tS] is a voiceless post-alveolar affricate; [dZ] is a voiced
post-alveolar affricate.

There’s a fairly obvious difference between an affricate like [ tS], which counts
as a single consonant, and a sequence of stop + fricative (two consonants).
Compare he cheats (affricate: [hitSits]) and heat sheets (stop + fricative:
[hitSits]). In heat sheets, the separate identity of the [t] and the [ S] should be
quite perceptible.

[tS] and [dZ] are the two commonest affricates found in the world’s languages,
but stops and fricatives can be merged at other points of articulation. Thus a
voiceless alveolar affricate [ ts] is common in German, and a voiced alveolar
affricate [dz] in Italian. Both are represented by the letter z in ordinary spelling
(often zz in Italian).

Affricates are a manner of articulation in their own right, so we have now

encountered three manners: stop, fricative and affricate.

Consonants 41
 Exercises

67. What two features make [t] in French, Spanish and Italian doubly different
from [t] in English and German?

68. Show the difference between the following phrases by rewriting them in
ordinary spelling and restoring the spaces between the words:

[waItSu:z] [waitSu:z].

69. What day of the week is [tSu:zdeI]? What kind of dish is [stSu:]? In what city
can you travel by [tSu:b]? What religion is practised by [krIstS@nz]? At what
kind of institution can you be [EdZukeItId]? How can the sixth month of the
year and a mountain of sand have the same pronunciation? What can you
conclude about [tju, dju] in British English? And in American English?

70. Insert these affricates in the appropriate places on the following chart: [ tS],
[dZ], [ts], [dz].
bilabia labiodenta denta alveola postalveola palata vela uvula pharyngea glotta
l l l r r l r r l l

71. Are there post-alveolar affricates in (a) French, (b) German? If “yes”, say
whether voiced or voiceless or both, and give example words.

72. How would a voiceless velar affricate be transcribed, and what would it
sound like? (They are found in some southern dialects of German.)

73. Underline the correct features for the following sounds, and give the correct
IPA symbol.

(a) English t in tar: stop, fricative, affricate, aspirated, unaspirated,

alveolar, dental, palatal, velar, voiceless, voiced.
(b) French qu in quand: stop, fricative, affricate, aspirated,
unaspirated, alveolar, post-alveolar, dental, palatal, velar, voiceless,
voiced.

(c) French d in donne, Spanish d in dar. stop, fricative, affricate,

aspirated, unaspirated, alveolar, post-alveolar, dental, palatal, velar,
voiceless, voiced.

(d) Spanish ch in muchacho. stop, fricative, affricate, aspirated,

unaspirated, alveolar, post-alveolar, dental, palatal, velar, voiceless,
voiced
.

40 Consonants
 (e) Spanish j in trabajo, German ch in Koch. stop, fricative, affricate,
aspirated, unaspirated, alveolar, post-alveolar, dental, palatal, velar,
voiceless, voiced.

(f) German z in Zeit. stop, fricative, affricate, aspirated, unaspirated,

alveolar, postalveolar, dental, palatal, velar, voiceless, voiced.

74. Rewrite in ordinary (English) spelling:

[tSu:] [mIdZIt] [lektSe] [dZ@:k] [dZi@] [neitS@] [s@dZEstSn] [w¡tS]

75. Transcribe:

adventure, righteous, lounge, issue, luncheon, question, wretched,

Norwich, Jewish, culture, butcher, merchant, cheer.

4.09 Nasals

Nasals are our fourth manner of articulation, after stops, fricatives and affricates.

For all the consonants so far, the soft palate (velum) has been raised, closing off
the nasal cavity, so that all the air from the lungs has had to pass through the
mouth (getting partly or completely blocked en route).

The uniqueness of nasals is the fact that, during their articulation, the velum is
lowered, allowing air to exit through the nose. Nasality is such a usefully
distinctive feature that few, if any, languages in the world lack these
characteristic n or m type sounds. Three conditions are normally present for a
nasal:

(1)The vocal folds are vibrating (nasals are typically voiced).

(2)Thoughout the duration of the sound, the oral cavity is closed at some
point, so that the air set into vibration by the vocal folds enters the mouth,
but can’t leave through it.

(3)The velum is lowered, so that the vibrating air escapes via the nose.

Check this by saying [mmm] to yourself. In the case of [m] the oral cavity is
closed at the lips (same place of articulation as [p], [b]). So [m] is a voiced
bilabial nasal. Fig. 35 shows that [m] differs from [p, b] only in respect of the
position of the velum: the entrance to the nasal cavity remains open throughout
the duration of [m], so that it can be prolonged in spite of the stop-like closure at
the lips.

Consonants 41
 Fig. 35

[m ] [p, b ]

If the oral closure is alveolar instead of bilabial, [n] is the result. (The slightly
different shape of the oral cavity gives rise to a slightly different sound wave). As
you might expect, the [n] of French, Spanish or Italian — like the corresponding
stops — is dental not alveolar: [n]. 9

Also to be noted is the velar nasal, which is of frequent occurrence in English (ng
in the spelling, e.g. singer, singing). This has the same place of articulation as [k]
and [g], but the same manner of articulation as [m] and [n], specified above. The
IPA uses the symbol [Œ] for this sound: [sIŒ@], [sIŒIŒ]. Its official name is
“eng”.

At least that’s the RP pronunciation of singer, singing. Many speakers in the UK always
have a [g] after their [Œ] and therefore pronounce these words [ sIŒg@], [sIŒgIŒg]. In
RP the [g] is present only in the pronunciation of some words: finger for example
[fIŒg@] (where er is not a suffix).

Fig. 36 shows the resemblance between [ Œ] and [k] — and the crucial difference:
the lowered velum in the case of [Œ].

Fig. 36
[Œ] [k]

Here is the nasal row of the consonant chart, with an impressive array of symbols
included. The palatal nasal m (“left-tail N”) occurs in French (signe) and Spanish
(señor). Note that a few languages outside Europe even make use of a uvular
nasal.

Fig. 37
bilabia labiodenta alveola postalveola palata uvula pharyngea
dental velar glottal
l l r r l r l

40 Consonants
 m Õ n9 n Æ Œ N
Voiceless nasals are also a possibility. French has a voiceless [m] in a few
words like rythme, where it follows a voiceless [t]. As there’s little point in
devising a special symbol for such a relatively rare sound, the IPA uses a
diacritic (see 3.08) and simply puts a small circle under the ordinary symbol:
[m8 ] or [n8]. (A subscript circle always signifies “voiceless”, whatever
consonant it’s placed under.)

Exercises

76. As well as being common in English, [Œ] occurs in French (but only in a
rather special category of words), in Spanish (though in Spanish it’s not
separately represented in the spelling) and in German. Give examples of
words with [Œ] from at least one of these languages.

77. Those very old-fashioned RP speakers who go “huntin’, shootin’ and

fishin’” are accused of “dropping their g’s”. Why is this an odd way of
describing their pronunciation (from a phonetic point of view at least)?
How might it be improved on?

78. The above chart includes a symbol for a labio-dental nasal. This sound
actually occurs in English — though it’s not recognized separately in the
spelling. Which of the following contains a labio-dental nasal, and why?
impossible, invalid, inglorious, indiscrete.

79. How, in rapid speech, is the letter n pronounced in: (a) ten people, (b)
unclear, (c) unveil?

80. Rewrite in ordinary (English) spelling:

[tˆŒ] [l¡ŒIŒ] [k{Œg@ru:] [{Œk@] [tSuŒk] [streŒkT] [{ŒkS@s]

81. Transcribe, using IPA symbols:

language, nightingale, anger, anxiety, hung, anything, dinghy.

4.10 Laterals (the l sounds).

This chapter and the one that follows are particularly important for modern
linguists, since they include many of the most conspicuous consonant differences
between English, French, German and Spanish.

Consonants 41
 For an l in any language (e.g. the one at the beginning of like), the following
conditions usually have to be met. Check them by saying the [l] of [laik] as you
read through this.

(1) the velum is raised (as it is for all consonants except nasals)
(2) the vocal folds are vibrating
(3) the blade of the tongue is closed tightly against the alveolum (or
occasionally the front of the tongue against the hard palate)
(4) the side of the tongue is grooved, so that the air leaves through the corner
of the mouth.

Condition 3 means that the point of articulation is alveolar (or palatal). But,
uniquely among consonants, the air escapes “along the side” (condition 4), so the
manner of articulation is said to be LATERAL.

Whether the air flows out through the right-hand corner of the lips or the left-
hand corner, or even through both corners (i.e. with a groove on either side of
the tongue) is something that varies from one individual to another, irrespective
of the language being spoken. So you can carry on using whichever side you’re
used to. (Indeed it’s quite difficult to produce a lateral using the opposite side!)

Laterals are usually alveolar, but Hispanists will have met a palatal lateral in
words like llama or calle. The IPA symbol is [˘], and, as the tongue position is
similar to that for the high front semi-vowel [j], the effect is not unlike that of
[lj] in English value. In fact there is a strong tendency in Spanish to replace [ ˘]
by [j] altogether, so that pollo (“chicken”) and poyo (“bench”) fall together as
[pojo] — or even as [poZo] (see 4.05).

Few speakers of English are aware of it, but there are actually two variants of [l]
in most English accents (including RP and American and Australian varieties).
At the end of a word or before a consonant (well, child), the [l], in addition to
having the features listed above, is pronounced with the back of the tongue raised
slightly towards the soft palate. But this doesn’t apply at the beginning of a word
or after a consonant, e.g. in like or play. It’s not easy to sense these tongue
positions, but the difference between the “dark” [ ] of well and the “clear” [l] of
play shouldn’t be to difficult to hear, particularly if you try to interchange them,
i.e. to say well with the [l] of play, and vice versa.

The technical term for “dark” [ ] is “velarized [l]”. When it’s important to

register the distinction, the IPA uses [l] for the clear variety and [ ~ ] can be

used for both clear and ] (“l tilde”) for the dark

variety. But in ordinary transcription of English, [ l

dark versions.

40 Consonants
This alternation between two types of [l] is a feature of English which is not
shared by standard French, German, Spanish or Italian. In each of these
languages, [l] is always clear and dark [ ] is unknown. For example the French
word belle doesn’t sound like the English bell as far as its final consonant is
concerned. Similarly for German (voll, etc.) and Spanish (arból, etc.). Using dark
[] in these languages is a typical and all too conspicuous feature of a British
accent. (However, in languages like Dutch, Portuguese or Russian, both types of
[l] do occur, as in English, though they may be distributed differently.)
Exercises

82. Say which of the following varieties of English has/have only clear (and no
dark) [l]:

Scots, Irish, Welsh, Geordie, Cockney, Yorkshire.

83. Can you transcribe with IPA symbols the Cockney pronunciation of words
like meal or milk ?

84. Rewrite in ordinary (English) spelling:

[aɪlənd] [jɛləʊ] [iːgl] [miːzlz] [feɪljə] [ljuːd] [ɔːfl] [wɛlʃ] [ɒzwəld]

85. Transcribe, distinguishing if you wish between [l] and []:

bill, foolish, possible, lady, handled, splice, Holborn, blow, cold,

fly, wealthy, should, Ethel, leer, gimlet, ugly.

4.11 The r Sounds

These are more varied even than the laterals: there are at least half-a-dozen
different types. Taken together, English, French, German and Spanish provide
examples of all of them, but of course they differ from one another as regards
which particular ones they use.

Indeed it’s quite surprising that the various kinds of r should be perceived as
having something in common: significantly perhaps there isn’t any widely-used
collective name for them, apart from r sounds and the technical term rhotics (rho
being the Greek letter corresponding to r). Oddly enough though, there is a well-
known term covering the l and r sounds taken together: LIQUIDS.

Fortunately, the r sounds can be classified quite easily.

There are two possible areas of the vocal tract in which an r can be made: dental/
alveolar and uvular. Typically the r’s used in English and Spanish are dental or

Consonants 41
 alveolar; those of standard French and northern varieties of German are uvular.
(There are plenty of regional exceptions to this general rule of course.) As only
two main areas of articulation are involved, people sometimes simply talk about
“front” vs. “back” r.

In addition to this, there are three different manners of articulation for each type.
The r’s used in English and Spanish illustrate the different “front” varieties.

First the TRILL, for which the tip of the tongue vibrates rapidly against the
alveolar ridge. Unknown in RP, but the commonest kind of r in Spanish. It’s the
normal r in Italian and Russian, as well as in the German of Austria, Switzerland
and southern Germany. And it’s found in many regional varieties of French. It
may be voiced or voiceless, depending on its position in the word. So common is
it in the world’s languages generally that the letter r, when used as an IPA
symbol [r], designates the alveolar trill, and not any of the other varieties.

Second the FLAP, which, as its name implies, involves a single tap of the
tonguetip against the rear of the upper front teeth or the alveolar ridge. Indeed,
TAP is an alternative term for this sort of r. A flap can be thought of as a
truncated trill. The IPA symbol for both the dental flap and the alveolar flap is
[ɾ] (“fish-hook r”), i.e. [r] minus its ornamental bits and pieces. Flaps (often
dental) are found in Spanish as well as alveolar trills, but the two aren’t
interchangeable: sometimes the difference between them serves to distinguish
one word from another: [karo] (carro) means “coach” (with a trill), but [ka ɾo]
(caro) means “dear” (with a flap). Note how the spelling uses single versus
double r to convey the difference. In American English an alveolar flapped [ ɾ]
commonly replaces [t, d] between vowels ([siɾi] city, [leiɾi] lady, etc.

Thirdly the APPROXIMANT — unknown in Spanish, but the normal variety of

r in most accents of English, and most likely the kind that you use yourself.
Prolong the first consonant of red so as to get the feel of it. You’ll notice that the
tip of your tongue is close to the teethridge, but probably not quite touching, and
certainly not close enough for a fricative to be produced. (And obviously there’s
no question of trilling or flapping.) The sound is actually rather like a vowel
produced with the tongue turned towards the palate (contrast the position for [i],
where the surface of the tongue is much flatter). The term approximant will
remind you that the sound isn’t quite a fricative, but not a proper vowel either.
The IPA symbol for the alveolar approximant is [Ú] (read “turned r”). But in
ordinary transcription of English it’s acceptable to use [r] instead (on the
understanding that it’s not being used to represent a trill). Note how alveolar
flaps and approximants can be adjacent to one another in American English:
[la:ɾɹ] ladder, [d@ɹɾi] dirty.

40 Consonants
 Occasionally in English the [Ú] is devoiced to [Ú8] — e.g. when it follows a
voiceless stop in a word like train. The loss of voicing has the effect of turning
the approximant into a fricative, as you can see if you say train slowly to
yourself. What’s more, in words like this, the fricative combines with the t to
form an affricate, so that in many people’s pronunciation there’s little
difference between train and chain, or chip and trip.

r serves to demarcate two major accent types of English worldwide. In RP (and

some other varieties), r is heard only before vowels (red, rich); after vowels,
any r in the spelling is silent in the pronunciation (car, bird), UNLESS a word
begining with a vowel immediately follows it, in which case a “linking r” is
present (drive the car [r] away). But in many accents of English (notably
American), r is always pronounced — after vowels (post-vocalically) as well as
before them (pre-vocalically). Accents of the latter type are known as rhotic,
accents of the RP type as non-rhotic (after the Greek letter rho).

Now let’s see how three manners of articulation (trill, fricative and approximant)
are used with the uvular place of articulation, this being particularly relevant to
French and German. Recall that the uvula is the extreme tip of the soft palate
(velum).

For a TRILL, the uvula itself vibrates against the back of the tongue (this is
the converse of the alveolar trill, for which it’s the tongue, or at least the tip of
it, that vibrates). The IPA symbol for the uvular trill is [R] (read “small
capital r”). This symbol is the one customarily used in transcribing French,
since the uvula trill is the traditional way of pronouncing standard French r.
Thus [RaR] for rare, [ReEl] for réel, etc. Nowadays though, French uvular [ R]
tends to be pronounced with little or no vibration or trilling, being
“weakened” to a FRICATIVE, with the back of the tongue simply in loose
contact with the uvula. This sound is represented by [] (“inverted capital
r”): it can either be voiced [ ] (as in [ab] arbre), or else voiceless []
(as in [tt] traître), depending on whether neighbouring consonants
are voiced or voiceless.

Parallel to the alveolar approximant characteristic of English is the uvular

APPROXIMANT: a vowel-like sound with “back r colouring”. The IPA
represents this as []. The diacritic [4] indicates that the tongue is lowered
slightly relative to the uvular fricative [] — enough to rule out any friction,
but remaining sufficiently close to the uvula for an approximant to be
possible. This sound is to be heard increasingly often in contemporary French,
particularly at the end of a word ([œ] heure), or between vowels
([pai] Paris). In the latter case, though, it is still considered a somewhat
“vulgar” pronunciation.

In northern pronunciations of German, a trilled [R] is possible at the beginning

of words, after consonants and between vowels (rot, treffen, waren); but, as in

Consonants 41
 French, the usual tendency is to weaken the trill, and to use a fricative [] or [
 ] instead. Between vowels an approximant [] may also occur (Ehre).
After vowels (as in wird, Uhr) any rquality is lost altogether and
[]/[]/[]/[] is replaced by a lowish central vowel, not unlike RP [ ],
for which the IPA and many dictionaries use the symbol [ ] (“turned a”).
Thus: [vit] wird, [u] Uhr. Here it is on the vowel chart:
e o

E O

Fig. 38 a A

Unstressed er in the spelling (hundert, höher) corresponds to [·] not to [@],

the erstwhile r having absorbed the e altogether: [hund·t], [hø:·]. The
difference between [·] (low central) as in bitter and [@] (mid central) as in
bitte, is quite an important one: think of the [ˆ] of S. English up, rather than
the [@] of rather and you’ll be on the right track.

To resume the complexities of r:

English, Spanish and southern German use dental or alveolar r’s of one sort or
another (the IPA symbols are always variants on lower-case r). Northern
German and standard French use various kinds of uvular r (the IPA symbols
are variants on upper-case R).

English (RP) has the approximant, but doesn’t normally use trills or flaps.
German uses voiced or voiceless fricatives (sometimes approximants) and
Spanish has trills and flaps, but no approximant. In standard French trills
(occasionally), fricatives and approximants occur. In northern German and in
RP there is no “post-vocalic” r: in German this becomes a vowel, in RP it
disappears altogether.

Exercises

86. What does the term liquid refer to?

40 Consonants
87. State the various places and manners of articulation associated with r .

88. Say whether the English accents typical of the following places are “rhotic”
or “non-rhotic”:

Edinburgh, Cardiff, Dublin, Manchester, Preston, Bristol, Southampton,

Brighton, Sydney, Chicago, Basildon.

89. The r in RP “take the car away” is known as a “linking r”. The following
contain examples of so-called “intrusive r “. Why “intrusive”?

Russia [r] and China, India [r] and Pakistan, drama [r] and music,
law [r] and order.

And what makes some people insert an [r] into drawing, sawing and awe-
inspiring?

90. Can you transcribe and/or comment on the rather “posh” or “stagey” variety
of r heard from some RP speakers in words like very ? And how might the
same speakers pronounce the r of rather ?

91. Rewrite in ordinary (English) spelling:

[rˆt] [rut] [ri@] [rO:] [f@:rI] [hˆrI] [i@rI] [dE:rI] [daI@rI] [mIr@]

92. Transcribe:

royal, rude, rag, arrow, fury, dowry, gregarious, treasury,

umbrella, birthright.

4.12 The IPA consonant chart.

Here is the chart for reference, including all the consonants mentioned in this
chapter. It’s still incomplete compared with the full official chart, which you will
find reproduced in many of the books listed in the Further Reading section (e.g.
Gimson’s Pronunciation of English).

The exercises following the chart are intended to help you to revise the material
on consonants (including one or two finer points).
Labio- Post-
Bilabial Dental Alveolar Palatal Velar Uvular Pharyngeal Glottal
dental alveolar

t
STOP p b t d k g q G ?
d
¬ f v T D s z S Z C x γ  h  h
FRICATIVE


Consonants 41
 ts tS
AFFRICATE
dz dZ

NASAL m Õ n n Æ Œ N

LATERAL
  ˘

TRILL
r R

FLAP
Q

APPROXIMANT
ɹ 
Exercises

93. Why is there no voiced/voiceless pairing in the case of the glottal stop?
94. Why aren’t voiced/voiceless pairs given in the nasal row?
95. Why are two alveolar lateral symbols given?
96. Why is no uvular flap symbol given?
97. Which symbols are used in ordinary transcriptions of RP?
98. Which additional symbols might be used in more detailed transcriptions of
RP?
99. Which symbols represent consonants occurring in the standard
pronunciation of: (a) French, (b) German (northern), and (c) Spanish?

40 Consonants
phonetic transcription and
general revision
[5]
5.01 Guidelines for Transcription

If you’ve worked through the various exercises given so far, you will already
have a good idea of the basic principles of phonetic transcription. But for
convenience, they will now be stated explicitly, and some further practice
material will be given from English.

Phonetic transcription of other languages follows the same principles as it does

for English (see below), and, as you know, the same phonetic alphabet is used —
plus or minus various symbols. So anyone who can transcribe English can adapt
without much difficulty to French, German, Spanish, etc. You can see this for
yourself by looking at the phonetic spellings in an up-to-date bilingual dictionary
— by now you should be able to follow them.

English is a good language to train with: it contains a larger than average number
of vowels and diphthongs, uses several of the less common IPA symbols, its
spelling is often highly idiosyncratic and unrelated to modern pronunciation, and
its vowels are prone to appear and disappear mysteriously according to where the
stress falls (see below). All this makes the transcription of many other languages
seem easy in comparison.

There are two “golden rules” for transcription:

 Think of the pronunciation, not the spelling. Instead of letting

yourself be mesmerized by the written form of the word, look
away, or close your eyes, and ask yourself how you actually
pronounce it. Among a thousand other pitfalls, this will help you to
avoid inserting an [l] into your transcription of salmon, it will
ensure that you transcribe the s of his with [z] not with [s], or the f
of of with [v] not with [f], and it will prevent you from including
the silent b of doubt in your transcription.

 Use one symbol for each identifiable sound. In many cases, the
orthography (i.e. the spelling system) uses more than one letter in
cases where a single symbol is required in a transcription: ps, sh,

Phonetic Transcription and General Revision 71

 ch, tch are cases in point. And a transcription, though it may be
shorter, can convey more precise information than the orthographic
form (the spelling): th corresponds to either [T] or [D]) and the
notorious -ough to [ʌf], [uː], [ə], [aʊ], [əʊ] or [ɒf] (enough,
through, borough, bough, although, cough).

Here are some punctuation conventions which you should note:

Capital letters and apostrophes are used only if some special IPA
value is intended (as with [I] and [R] for example). So Tony Blair’s
Britain would be transcribed [ten blz brtn], with lower-case
throughout, and no apostrophe. Otherwise you can use the same
punctuation and word spacing as the original. But it’s usually best
to transcribe numbers and dates in full, and also initials like U.S.A.
[ju.s.e]

Broad and narrow transcription

Some transcriptions provide less detail than others. In so-called “broad”

transcriptions, differences between sounds are shown only if they can serve to
distinguish words from one another in a particular language. For example, in
English the difference between the [l] of look and the [r] of rook obviously has to
be shown. However, “clear” and “dark” [l] (see 4.10) cannot distinguish words,
as in English they don’t occur in the same place. An [l] at the beginning of a
word is always clear (like), but at the end
(well) only the dark variety [ɫ] occurs, so a pair of words [wɛl] and [wɛɫ] would
not be possible. Accordingly in a broad transcription we write [lʊk] and [rʊk],
differentiating
[l] and [r], but [laik] and [wɛl], not differentiating between the two types of [l].
Similarly with the longer and shorter variants of vowels (see Exercise 46). The
vowels of bee and beat would be transcribed with [i] in both cases, even though
in bee the vowel is appreciably longer than in beat (this is because beat ends in a
consonant and bee doesn’t). A pair of words like [bi] and [biː] would not be
possible in English. Nor, in a broad transcription, need it be specified that RP r is
an approximant: replacing approximant [ɹ] by some other kind of r wouldn’t
change the identity of the word rook, so [rʊk] is a sufficiently distinctive
transcription.

“Narrow” transcriptions, on the other hand, do include these less crucial

differences. So well might indeed be transcribed as [w], bee as [bi], rook as
[k], and so on. The more detail is included, the narrower the transcription
becomes. When it is desired to specify which kind of transcription is involved —

72 Phonetic Transcription and General Revision

broad or narrow — the convention is to place broad transcriptions in slashes / /,
reserving square brackets for narrow transcriptions. So rule could either be /rul/
(broad) or [u] (narrow). The following transcriptions of English are of the
broad variety (although, as is customary, the reduction of unstressed vowels to []
is indicated — see below). Further examples of narrow transcription are given in
5.03 in connection with French, German and Spanish.

A few other points now follow, relating particularly to features of English.

Regional accents.

Most regional variations are far too subtle to be picked up by the broad system of
transcription being introduced here. However, there are one or two conspicuous
differences which do have an effect on transcription: [fast] versus [f Ast] for fast;
[hAɹd/ hArd] versus [hA:d] for hard; [ʊp] versus [ʌp] for up. If you’re not an RP
speaker, you should represent your own regional pronunciation, unless
specifically asked to base your transcription on RP. The most important thing is
(a) to avoid mistranscriptions that have nothing to do with regional variation
([fəst] for fast, or [hɛːd] for hard) and (b) to be consistent: don’t put [fast] for fast,
then use [kAsl] for castle a few lines later.

Word stress

In English, as in other European languages (apart from French), certain syllables

in words and phrases are pronounced with greater force than others. Thus in
conversation, the main stress, or “primary stress”, is on the third syllable, there’s
a “secondary stress” on the first, and the vowels of the second and fourth
syllables are “unstressed” — in rapid speech these may disappear altogether.
Primary stress is indicated in the IPA by [ˈ] and secondary stress by [ˌ]. Both
these diacritics are placed before the syllable in question. Syllables with no mark
preceding them are unstressed. So in a transcription indicating stress,
conversation would be [ˌkɒnvəˈseɪʃən].

Vowel reduction.

A big problem for foreign learners of English is the way vowels change in quality
according to whether they are stressed or not. Compare (a) photograph with (b)
photographic and (c) photographer. The o of the first syllable is pronounced
[@U] in (a) and (b) but [@] in (c). The second o is [@] in (a) and (b) but [¡] in
(c). As for the a it changes from [ɑ] (in RP at least) to [æ] to [ @]! If this sounds
confusing, repeat the three words to yourself to check: native speakers of English
know instinctively where to place the stresses and what quality to give the
vowels. None of this variation is revealed in the orthography, but it’s very

Phonetic Transcription and General Revision 73

 important for acceptable pronunciation, and should therefore be shown in the
transcription.

So once again the basic principle applies of not paying too much attention to the
spelling. Here’s the transcription of (a) photograph, (b) photographic, (c)
photographer, with this intricate interplay of stresses and “reduced vowels”
indicated:

(a) [ˈfəʊtəˌgrɑf] (b) [ˌfəʊtəˈgræfɪk] (c) [ˌfəˈtɒgrəfə].

Reduced vowels are particularly common in the case of prepositions (to and of
for example) and common verbs like was, were, has, have — especially in rapid
speech. In some cases the reduction has gone so far that it’s acknowledged in the
spelling: I’d, he’s, etc. But often the spelling doesn’t reflect it: you’ll find plenty
of examples in the passages given below.

The best thing is to base your transcription on a style that isn’t too stilted and
unnatural — one that sounds more like informal conversation than formal
reading. But you probably won’t want to go too far in the other direction. So
shouldn’t have would be [ʃʊdnt həv] or [ʃʊdnt əv] (registering the dropping of
the h in rapid speech — even in RP), but maybe not [ʃʊnv], even though this is
what it is sometimes reduced to. Unless you’re making a specific study of
degrees of English vowel reduction, that is. Basically the moral is that there’s
often not just one single “correct” way of transcribing words and phrases: there
are many different ways, depending on who is speaking, under what
circumstances, and what the purpose of the transcription is exactly.

5.02 Transcribing English

Here are the transcriptions (broad — see above) of some short items for further
illustration. In the first one, the stresses have been marked, and in all of them
vowels have been reduced to an extent appropriate for informal conversation.

a. Dozens of holidaymakers travelling on Leicestershire’s only remaining

steam railway escaped injury yesterday after a collision with a lorry on an
unguarded level crossing.

ˈdʌzənz əv ˈhɒlɪdɪˌmeɪkəz ˈtrævlɪŋ ɒn ˈlɛstəʃəz ˈəʊnlɪ rɪˈmeɪnɪŋ

ˈstiːm ˈreɪlˌweɪ ɛsˈkeɪpt ˈɪnʤərɪ ˈjɛstəˌdeɪ ˈɑftə r ə kəˈlɪzən wɪð
ə ˈlɒrɪ ɒn ən ʌnˈgɑːdɪd ˈlɛvəl ˈkrɒsɪŋ.

74 Phonetic Transcription and General Revision

b. Ireland’s Jimmy Logan won the Eurovision Song Contest on Saturday
night, the second time in a decade that his country has taken the prize.

aɪələndz ʤɪmɪ ləʊgən wʌn ðə jurəʊvɪʒən sɔŋ kɒntɛst ɒn sætədɪ

naɪt, ðə sɛkənd taɪm ɪn ə dɛkeɪd ðət hɪz kʌntrɪ həz teɪkən ðə
praɪz.

c. More than 330,000 of Britain’s poorest children will lose their right to
free school meals when social security changes come into effect tomorrow.

mɔː ðən Triː hʌndrəd ənd T@:tɪ Taʊzənd əv brɪtənz puəːrɪst

ʧɪldrən wɪl luːz ðɛ raɪt tu friː skuːl miːlz wɛn səʊʃəl sɪkjuːrɪtɪ

ʧeɪnʤɪz kʌm ɪntu ɪfɛkt təmɔrəʊ. Now see if you can read the following
unaided.

d. weɪlz rɪvəːst lɑst ɔːgəsts stjuːdənt wəːld kʌp dɪfiːt baɪ ɪŋglənd, bʌt ðə
wʌndə r əv lɑst naɪts ʌtəlɪ fəgɛtəbl geɪm ɪn kɑːdɪf wəz ðət ɪŋglənd ɛvə
r əraivd ɑːftə ðə sevən brɪʤ həd bin kləʊzd wɛn ə lɒrɪ təːnd əʊvə.
ðɛː bʌs wəz snɑːld ʌp ɪn ə træfɪk kjuː ənd ðeɪ meɪd ðɪ ɑːmz pɑːk
əʊnlɪ ɑftə rɪtəːnɪŋ tə pɑːkweɪ steɪʃn ənd kæʧɪŋ ə treɪn.

5.03 Transcribing French, German and Spanish

A broad and a narrow transcription is given for each language. Various other
narrow transcriptions are possible, depending on how much detail is included.

Note that the transcriptions given in bilingual or monolingual dictionaries are

essentially broad, though departures from a strictly broad representation are
sometimes made in order to give a clearer idea of the actual pronunciation: thus
length marks may be included for vowels in English or German; French r is often
indicated as [R] (see below), and so on. Dictionaries vary in their practice in this
respect — hence some of the discrepancies between their transcriptions.

Exercise: What features of pronunciation in each language do the narrow

transcriptions reveal?

a. Broad transcription of French (educated Parisian variety).

Le territoire français est le plus étendu d’Europe, hors Russie. La densité

de population est plus faible que celle de tous ses voisins, à l’exception de
l’Espagne. La France dispose de réserves territoriales comme aucun de

Phonetic Transcription and General Revision 75

 ses partenaires: sa position d’isthme (situé à l’ouest du continent) donne
accès à tous les types de milieux existant en Europe: littoraux,
méditerranéens, continentaux, atlantiques, montagnards.

l tritwar fr~s e l ply z et~dy dørp, r rysi. la d~site de

ppylasj~ e ply fbl k sl d tu se vwaz~, a lkspsj~ d lspa. la
fr~s dispoz d rezrv tritrjal km okœ̃ d se partnr: sa pozisj~
dism (sitye a lwst dy k~tin~) dn aks a tu le tip de miljø gzist~
~ n ørp: litro, medit~rane~, k~tin~to, atl~ ~tik, m~taar.

N.B. Many otherwise broad transcriptions represent standard French r as [R]. This is not
strictly necessary, as the difference between different types of r isn’t used to distinguish
between words in French. In any case, the actual pronunciation is generally a uvular
fricative [] rather than a uvular trill [R] (see 4.11).

`
Many speakers of standard French never use the nasal vowel` [ ˚ ], as in
aucun, brun, parfum, replacing it by the [E] vowel of main, voisin, dessein.
So, for
`
example, they pronounce brun and brin in the same way: [brE].

b. A narrow transcription of the same French text.

l t  itwa  f ~s e lply zet ~dy d ø p, 

ysi. la d ~site t p pylasj~ e ply fbl k sl d tu se
vwaz ~, a lkspsj~ d l spa. la f~s dispoz d  ezv
t it jal km okœ̃ t se pa t  n: sa pozisj` ~

dism (sit  e a lwz dy k~tin ~) d n aks a tu le t ` ib

d  miljø ki gzist  nøp: lit o, medit ane~, k~tin
~to, at l ~tik, m ta a.

N.B. The plus sign [ ] beneath a vowel symbol indicates that the tongue position is more
fronted than would be the case if it had the strict cardinal value.

c. Broad transcription of German (educated northern variety).

Kein zweites Land in Europa grenzt an so viele Nachbarstaaten wie

Deutschland — größere und kleinere, ärmere und reichere, romanische,
germanische und slavische. Sie alle werden — ob sie es gewollt haben

76 Phonetic Transcription and General Revision

 oder nicht — mehr oder weniger von Deutschland beeinflußt:
wirtschaftlich sowieso, politisch wieder nachhaltiger als früher, aber auch
kulturell (obschon das traditionell meist eine Straße in zwei Richtungen
war).

kain tsvaites lant n yropa grntst an zo file naxbataten vi

dylant — grøsere unt kleinere, rmere unt raiçere, romane,
grmane unt slave. zi ale verden — p zi s gvolt haben
oder nçt — mer oder venger fn dylant beainflst: vrtaftliç
zovizo, plit vider naxhaltiger als fryer, aber aux kulturl
(opon das tradtsjnl aine trase n tsvai rçtuŒen var).

N.B. As [] in German is simply an unstressed variant of [e], it is not identified separately in
this transcription.

d. A narrow transcription of the same German text.

khan tsvats lant n opa gntst an zo fil

naxbatatn vi dlant — gøs un klan, m
unt aç, oman, gman un slav. zi al
vedn — p zi s gvolt ham od nçt — me od veng fn
dlant banflst: vtaftliç zovizo, plit vid
naxhaltig als fy, ab ax kltul (opon das
tadinl an tas n tsva çtuŒŒ va).

e. Broad transcription of Spanish (educated Madrid variety).

La unidad de España le viene de su condición peninsular, de ese cerco de

mar que la cierra en sí misma. Vista en el mapa, España no puede ser más
que unidad. Sin embargo, todos los paisajes posibles se dan en la
Península Ibérica: variedad cuya grandeza llegó a ser desconocida. Hay
igualmente una España de belleza fácil que salta a los ojos, y otra que
solo se sabe ver en todo su riqueza después de un aprendizaje
prolongado.

la unidad de espaa le bjene de su kondiTjon peninsula, de ese

Teko de ma ke la Tiera en si misma. bista en el mapa, espaa
no pwede se mas ke unidad. sin embago, todos los paisaxes
posibles se dan en la peninsula ibeica: baiedad kuja gandea
ego a se deskonoTida. ai igwalmente una espaa de beea
fail ke salta a los oxos, i ota ke solo se sabe beQ en todo su
rikea despwes de un apendiTaxe polongado.

Phonetic Transcription and General Revision 77

 f. A narrow transcription of the same Spanish text.

la unia de spaa le jene e su kndiTjn pninsula,

dese k de ma ke la Tjra n si mizma. bista n l
mapa, spaa n pwee s mas ke unia. sin mbag,
ts ls paisaxs psibls se an  n la pninsula ieica:
bajea kuja gan dea e a s dsknia. ai
iwalmnte un a spaa e ea fail ke salta a ls xs,
i ta ke sl se sae  n t su rikea dspws de un
 apndiaxe plŒga.

N.B. [j] is an increasingly frequent alternative to [ ] (see 4.05 and 4.10).

Sources: A. Frémont, Portrait de la France, Flammarion 2001; M. Gorski, Gebrauchsanweisung für Deutschland, Piper 1996; G. Torrente-Ballaster, España,
Pueblos y paises 1986.

Now make your own transcriptions of the following:

g. French. Give a broad transcription.

Entretien général du magnétophone. Placer l’appareil sur une

surface dure et plate pour qu’il soit bien droit. Ne pas laisser des
CD, des piles ou des cassettes en contact avec l’humidité, la pluie,
le sable, à la lumière directe du soleil ou dans des endroits
susceptibles de connaître des températures élevées comme à
proximité d’appareils de chauffage ou dans des voitures garées au
soleil. Manipuler toujours le CD en le tenant par les bords et le
ranger dans sa boîte après l’usage, la partie imprimée tournée vers
le haut.

h. German. Give a relatively broad transcription, but distinguish between [e]

and [@] and between [x] and [ç]. Also indicate vowel lengthening.

Allgemeine Pflege des Kassettendecks. Die Anlage auf eine harte,

ebene Fläche stellen, damit sie nicht umkippen kann. CDs,
Batterien und Kassetten vor Feuchtigkeit, Regen, Sand und direkter
Sonnenstrahlung schützen und nicht an Orten aufbewahren, an
denen es zu hohen Temperaturen kommen kann (z.B. in der Nähe
von Heizungen oder in einem in der Sonne geparkten Auto) . Die
CD stets am Rand halten und nach Gebrauch wieder in die
Schachtel legen, um sie vor Zerkratzen und Staub zu schützen.

78 Phonetic Transcription and General Revision

i. Spanish. Give a relatively broad transcription, but distinguish between the
stops [b, d, g] and the fricatives [, ð, ]; also between [n] and [Œ].

Mantenimiento general del magnetófono. Ubique el aparato en una

superficie dura y plana de tal manera que no se incline. No exponga el
aparato, los CD, las pilas ni los cassettes a la humedad, la lluvia, la
arena, o a la luz directa del sol o dejarlos en lugares donde ocurren altas
temperaturas, tales como en las cercanías de aparatos de calefacción o en
automóviles estacionados en el sol. Siempre aguante el CD por el borde y
guárdelo en su estuche después del uso, para evitar que se raye y se
ensucie.

5.04 General Revision

Here are ten quotations from more advanced books on phonetics, which, in terms
of style and terminology, are typical of much specialist writing on the subject. All
the topics referred to have been covered in this book, so there’s nothing here that
you don’t know about. However, the language used may seem somewhat
forbidding and technical at first sight. It should yield quite readily to a little
thought, so your exercise is to explain in your own words what each statement
means (preferably with one or two examples for illustration when appropriate).

a. Consonants are made with a stricture involving contact of relatively large

areas of both active and passive articulators.

b. The distinctive quality of sound of any vowel results from the general shape
given to the oral cavity during its production.

c. The stricture which produces the type of consonant called a fricative is one
of close approximation of the articulators, with central passage of the
airstream.

d. We need more than two parameters for adequate description of a vowel.

e. A cardinal vowel is a fixed and unchanging reference point, established

within the total range of vowel quality, to which any other vowel sound can
be directly related.

Phonetic Transcription and General Revision 79

 f. Stops are produced with a supra-glottal articulatory closure, unless they
are glottal stops.

g. Aspiration involves adjustments of timing between laryngeal and oral

articulations.

h. Affricates are an intermediate category between simple stops and a

sequence of a stop and a fricative.

j. Nasals have an articulatory similarity to stops by virtue of their oral

closure, but in other respects they are similar to approximants.

k. The r-sounds form a heterogenous group from the phonetic point of view,
exhibiting a wide variety of manners and places of articulation.

80 Phonetic Transcription and General Revision

 answers to exercises
[6]
Section 2 (The Vocal Tract)
1. The lungs supply the air for almost all speech sounds. Air passes from them into the
bronchi, one from each of the two lungs, and these two airstreams merge in the trachea, a
short tube situated in the lower part of the neck. On top of this is a valve known as the
larynx. Here the supply of air to the throat and mouth is controlled by opening or closing
the glottis — the gap between the two vocal folds. In ordinary quiet breathing the glottis is
open; for swallowing it is closed in order to protect the lungs. A noteworthy evolutionary
adaptation in humans allows voice to be produced by positioning the vocal folds in such a
way that passage of air between them causes them to vibrate.

2. (a) [z] voiced; [T] voiceless; [b] voiced; [p] voiceless; [ð] voiced; [s] voiceless.

(b) [f] voiceless; [v] voiced; [t] voiceless; [d] voiced; [k] voiceless; [g] voiced; [ ]
voiceless.

3. pin (vl), bin (vd)

fail (vl), veil (vd) toll (vl),

dole (vd) gin (vd), chin (vl)

zoo (vd), sue (vl) wreath (vl),

wreathe (vd) either (vd),

ether (vl)

Aleutian (vl), allusion (vd)

4. zeal, racer, eyes, scares, angle, duck/tug, bigger/picker, lunch, juice/choose, joke,
thy, confusion.

5. See Fig. 6.

6. (a) the velum: here it is pressed against the wall of the pharynx and is closing off the nasal
cavity. The uvula “dangles down” below this point of closure.

(b) the pharynx: the projection on the left is the epiglottis, an appendage at the top
of the larynx, which rarely if ever has any function in speech.

(c) the tip of the tongue, closing against the alveolar ridge (upper front teeth shown
at far left).

(d) the tongue, with root to the right and tip to the left.

7. lips, tongue, velum, uvula.

Answers to Exercises 81
 8. (a) FALSE (correct answer: the larynx)

(b) FALSE (correct answer: the velum)

(c) TRUE

(d) TRUE

(e) FALSE (correct answer: the alveolum).

9. trachea, larynx, glottis, oesophagus, velum, alveolum/alveolar ridge, pharynx.

10. the vocal tract runs from the lungs to the lips; the upper vocal tract (also known as the
supra-glottal tract) is the part situated above the larynx (including the nasal cavity). Oral
tract is an alternative term for the mouth cavity (from the top of the pharynx to the lips).
The nasal tract is the same as the nasal cavity.

11. See Fig. 6.

Section 3 (Vowels)

N.B. [a ] [e] [], [a] [e] may, in broad transcription, also be written [ai] [ei] [ i], [au] [eu].

12. (a) [A] (b) [i] (c) [u].

13. (a) FALSE (it is low front)

(b) TRUE

(c) FALSE (it is high back)

(d) FALSE (it is low back)

14. (a) [i] is close

(b) [u] is close

(c) [a] is open

Answers to Exercises

15. (a) low = “tongue flat”

(b) open = “wide oral cavity”

(c) front = “tongue advanced”

2
 (d) high = “surface of tongue raised towards hard palate”

(e) back = “tongue retracted”

(f) close = “narrow oral cavity”.

16. (a) London

(b) they guide you like the cardinal points of a compass(c) it provides a framework for the

vowels of any language

17. See Fig. 15.

18. Correct yourself, or consult your tutor.

19. bête []; foule [u]; premier [, e]; patte [a]; été [e]; rose [o]; prêt []; pré [e]; dehors
[, ]; sortie [ , i]; petit [ , i]; sous [u]

sehen [e, ]; Hände []; wenden []; ruhen [u]; Sonne []; Sohn [o]; hatte [a, ]; Vater
[ ]; Tier [i]; See [e]; wahr [ ]

ser []; amigo [a, ]; hombre [ ]; hoja [ ]; cabo [a, ]; verde [ , e]; clase [e]; hemos
[e, ]; grande [a, e]; salir [i].

20. Monitor yourself, or consult your tutor.

21. Rounded: fou, cru, rose; hoch, Sohn, früh; loco, tu, lo

Unrounded: tête, patte, prêt, pré; Reh, Hand, Tier, wahr; ser, verde, cara, si.

22. [u] high, back, rounded

[e] mid-high, front, unrounded

[] mid-low, front, unrounded

[a] low, front, unrounded []

mid-low, back, rounded [@]

mid, central.

23. pu [y], pou [u], pis [i], voulu [u, y], lugubre [y, y], fourrure [u, y]
Ufer [u], über [y], Mühe [y], Fuß [u], Füße [y], Zypresse [y], Statue [u].

24. foule, lune, bouder, fumer, vous, vu, fou, fut, ou/où, eu Hügel, Mut, Kuh, süß,
fühlen, Huhn, Nudel.

25. [y] high, front, rounded

[e] mid-high, front, unrounded

Answers to Exercises 83
 [] mid-low, front, unrounded

[ø] mid-high, front, rounded

[i] high, front, unrounded

[œ] mid-low, front, rounded

26. (a) TRUE

(b) TRUE

(c) FALSE

(d) TRUE

27. French, German.

28. bonheur [œ], soeur [œ], deux [ø], milieu [ø], jeune [œ], stupeur [œ], peut-être [ø].

Söhne [ø], völlig [œ], Körper [œ], mögen [ø], köstlich [œ], Goethe [ø], töten [ø].

29. peur, meurt, ceux, deux, su, cru, air, né höre, helle, Tür, Kellner, liegen, können,

viele.

30. seul, fameux, hideux

Vögel, gönnen, Höhle.

31. (a) R(eceived) P(ronunciation)

(b) digraph

(c) small capital i, turned v, reversed open e.

(d) Southern British vowels are more centralized, i.e. the tongue position is less extreme.

32. Because London is the capital of the UK.

33. hand [æ], car [], alphabet [æ, ], foot [], butter [], further [, ], roar [], swan
[], wish [], monkey [], some [], cough [], enough [], yawn [], pull [],
knowledge [ ].

Answers to Exercises

34. [hl].

35. leaf, leave, but, bat, boot, look, cod, cord, good, country, second, stomach,
except, bosom, habit.

2
36. foot [fut], fast [fast], fast [fa st], up [ p], cap [kep], off [ f], cross [kras].

37.
y i u Ì

I U
Ô e o Â

@
˚ E O ˆ
ae

œ a A ¡

38. A vowel is any speech sound in which there is free passage of air above the glottis. Pure
vowel and monophthong both refer to vowels with unchanging tongue position. During the
articulation of a diphthong, the tongue moves from one position to another. A digraph is a
group of two letters representing a single monophthong.

39. (a) face: [fj s] [j] represents the first sound of yacht (see 3.09).

(b) face: [f s]

(c) so: [so]

(d) bean: [b n]

40. bough [a ], aisle [a ], toe [ ], buy [a ], weigh [e ] though [ ], buoy [ ],

soap [ ], town [a ], same [e ]

41. tidal, climber, noise, fellow, choir, flower, product, eiderdown, catalogue, missile,
telephone.

42. (a) It represents a sequence of two separate vowels.

(b) mein [a], neu [ny] or more exactly [] (lips rounded throughout), Haus [hs] (back
[ ] rather than front [a]).

(c) in peu the spelling eu represents [ø] and in soeur the oeu represents [œ] — a
monophthong in each case. English high contains a single diphthong [a ], with a
gradual transition from [a] to []; French haï (“hated”) contains a sequence of two
monophthongs [a + i], with a very rapid transition from one to the other (see also
3.09).

43. In the case of [i] the tongue movement is towards the centre of the vowel area; in the
case of [a ] and [a ] it is from a lower (more open) to a higher (more close) position.

44. cord, card, cared, repertoire, Salisbury, courtesy.

Answers to Exercises 85
45. good (shorter), food (longer) card (longer), cad (shorter) pet (shorter), paired (longer).

46. From shortest to longest:

beak, bead, bee [bik, bid, bi ]
boot, food, do cart, carve, car
caught, call, caw hearse, heard, her

Vowels are shortest before a voiceless consonant ([k, s, t], etc.); longer before a voiced
consonant ([d, l, v], etc.); longest when no consonant at all follows.

47. This one is for self-correction!

48. [s] cent, sans, sang, (tu) sens, (il) sent.

[s] cinq (sometimes: e.g. in cinq cents), saint, sein, sain, seing, (tu) ceins, (il) ceint.

55. voiceless alveolar: [t]

voiced velar: [g]
voiced dental: voiceless [d]
bilabial: [p]

78 Answers to Exercises

49. Europe, new, pursue, beauty, behaviour, failure, union, duty, onion, piano.

50. yacht, feud, manure, pursue, year, quite, quaint, womb, what, queen, queue.

51. wet, (to) whet (e.g. someone’s appetite).

52. German has [j] only, as in das Jahr. Spanish has both [j] and [w], as in tiene and bueno.

Section 4 (Consonants)
53.
bilabial dental alveolar velar
p b t d t d k g
54. [b] is a voiced bilabial consonant
[k] is a voiceless velar consonant [d] is
a voiced alveolar consonant [t] is a
voiceless dental consonant. 56.
(2) (4) lab- (6) (1) (5) post- (3)
bilabial dent dental alveolar alv velar
stop t g

fricative  z  x 
57. [z] [d] both voiced alveolar. [z] fricative, [d] stop.

[s] [d] both alveolar. [s] voiceless fricative, [d] voiced stop.

[t] [k] both voiceless stops. [t] alveolar, [k] velar.

[v] [g] both voiced. [v] labio-dental fricative, [g] velar stop.

[b] [ ] both voiced biliabial. [b] stop, [ ] fricative.

[v] [] both voiced fricative. [v] labio-dental, [ ] bilabial.

[b] [v] both voiced. [b] bilabial stop, [v] labio-dental fricative.

[s] [] both voiceless fricative. [s] alveolar, [ ] post-alveolar.

58. loch (Scots), back (Liverpool), Haughey (Irish).

59. [kstr]

60. It’s bilabial like [b], and fricative like [v].

61. laughs clothes heathens patience oath worthless thumb fashion closure luscious
beige usual charade luxury.

62. [keik] [bukei] [bægpaips] [ uðz] [n ] [sæ ]

[pl@] [ei @] (or ei @) [br] [bri ð] [f ut u] [f zks]
[wm] [ it] [ð u] 63.

bilabia labden denta alveola Postal palata vela uvula pharynge glotta
l t l r v l r r al l

t
p b t d k g q  ?
stop d

fricativ
 f v  ð s z  z ç x    h
e

64. A glottal stop would replace the [t] of daughter, butter, or glottal, notably in London
(Cockney) and Glasgow pronunciations. Glasgow has [r] at the end of the first two words of
course. So: [b(r)] [d (r)].

65. Uvula is a noun; uvular is an adjective. As in: “uvular sounds involve the uvula”. Just like
peninsula and peninsular (“the Peninsular War was fought in the Iberian Peninsula”).

Answers to Exercises 87
 66. [x] German, Spanish
[ç] German []
German
[q] none of these languages [h]
English, German.

67. They differ in place of articulation (dental in French, Spanish and Italian, alveolar in
English and German) and in aspiration (present in English and German, absent in French,
Spanish and Italian).

68. [wa  tSuz] = why choose? [wai  uz] = white shoes.

69. Tuesday stew tube Christians educated June/dune.

Before [u] the sequence [tj] is often reduced to a single-unit affricate [tS] in British
English. Sometimes in American English the [j] is omitted but the [t] retained, e.g. [tuzdi]
for Tuesday.

70.
labdent post- glot-
bilab dental alveolar pal velar uv phar
alv tal

ts dz tS 

71. In French [ ] is found only in words borrowed or adapted from other languages (match,
putschiste ). It occurs in German in deutsch, Quatsch, etc. [] doesn’t occur in either
language.

72. [kx]. Much as though imitating gunfire.

73. (a) English t in tar: stop, fricative, affricate, aspirated, unaspirated, alveolar, dental,
palatal, velar, voiceless, voiced.

(b) French qu in quand: stop, fricative, affricate, aspirated, unaspirated, alveolar,

post-alveolar, dental, palatal, velar, voiceless, voiced.

(c) French d in donne, Spanish d in dar: stop, fricative, affricate, aspirated,

unaspirated, alveolar, post-alveolar, dental, palatal, velar, voiceless, voiced.

(d) Spanish ch in muchacho: stop, fricative, affricate, aspirated, unaspirated,

alveolar, post-alveolar, dental, palatal, velar, voiceless, voiced.

(e) Spanish j in trabajo, German ch in Koch: stop, fricative, affricate, aspirated,

unaspirated, alveolar, post-alveolar, dental, palatal, velar, voiceless, voiced.

(f) German z in Zeit: stop, fricative, affricate, aspirated, unaspirated, alveolar, post-
alveolar, dental, palatal, velar, voiceless, voiced.

74. chew midget lecture jerk jeer nature suggestion watch.

75. [dvntS] [raitSs] [laun] [ ju] (or sju)

[lntSn] [kw stSn] [r tSd] [n r] (or nritS)

88 Answers to Exercises
 [u(w) ] [k ltS] [b tS] [m tSnt] [tSi]

76. French. Only in words borrowed or adapted from English, e.g. le smoking, faire du forcing,
un brushing.

Spanish. Spelt n but pronounced [ ] in cinco, inglés, naranja, etc. (i.e. velar [ ] before
another velar consonant).

German. ng in Finger, lang; n in Prunk, hinken.

77. What really happens is that they pronounce these words with an alveolar [n] instead of a
velar [ ]. “g “ relates not to the sounds, but only to the spelling — where it would indeed
have to be omitted in order to represent this pronunciation. But there are no g‘s to be
dropped in actual speech.

78. invalid. In words like these the nasal takes on the same place of articulation as the
immediately following consonant — economy of effort really — so [n] becomes labiodental
before [v]. To get the effect you should say invalid at normal conversational speed.

79. (a) ten people [m], (b) unclear [], (c) unveil []. Same phenomenon of “assimilation” as in
exercise 78.

80. tongue, longing, kangaroo, anchor, chunk, strength, anxious

81. [lægw] [nait geil] [æ g] [æ gzait] [h ] [ n] [d g]. Many speakers
have [i], not [ ], at the end of anxiety, dinghy.

82. Irish, Welsh.

83. [miw], [m wk] (or even [m w] with a glottal stop).

84. island, yellow, eagle, measles, failure, lewd, awful, Welsh, Oswald.

85. [b] [fu li] [p sb] [leid ] [hænd d] [splais] [h ubn] [bl u] [k ud] [flai]
[wel] [] [li ] [g mlt] [ gl].

Some speakers have [i], not [ ], at the end of lady, wealthy, ugly.

86. The r and l sounds.

87. Places: alveolar and uvular.

Manners: trill, fricative, flap and approximant.

88. Rhotic: Edinburgh, Dublin, Preston, Bristol, Southampton, Chicago. Non-rhotic: Cardiff,
Manchester, Brighton, Sydney, Basildon.

Answers to Exercises 89
 89. “Intrusive” because there is never an r in the spelling, and rhotic speakers don’t have an r
at the end of Russia, etc. when these words stand alone.
The r in drawing, etc. is due to the analogy with the much larger group of words like pour,
roar, bore, etc., where all speakers have an r before the following vowel.

90. Such speakers would use a flap [] between the two vowels of very. And they might even
use a trill [r] at the beginning of a word like rather.

91. rut, root (or route), rear, roar, furry, hurry, eerie, dairy, diary, mirror.

92. [i] [ ud] [ æg] [æ u] [fj ] or [fju] [da ] [g gs] [t
] [mbl] [b ait].

For convenience, [r] could be used instead of [ ], provided it’s noted that it doesn’t
represent a trill in such cases.

93. As the glottis is closed, the vocal folds can’t vibrate, so voicing is physically impossible.

94. Voiceless nasals occur only as occasional variants of voiced nasals, and therefore don’t
have special symbols.

95. [l] is for clear l, [] for dark l.

96. A uvular flap is too rare a sound to warrant a special symbol.

97. [p] [b] [t] [d] [k] [g] [f] [v] [ ] [ð] [s] [z] [ ] [] [h] [tS] [] [m] [n] [ ] [l] [r].

98. [] [ç] (see 4.06) [ ] (see 4.09) [ ] [ ].

99. (a) French: [p] [b] [t] [d] [k] [g] [f] [v] [s] [z] [ ] [ ] [m] [ ] (see 4.09) [l] [ ]

plus, in more detailed transcriptions:

[t] [d ] (dentals)
[m] (voiceless [m], see p. 57)
[] [] [] (approximant, voiceless and voiced fricative r ) and,

in words borrowed from other languages:

[tS] [ ] (see Exercises 71 & 76).

(b) German: [p] [b] [t] [d] [k] [g] [f] [v] [s] [z] [ ] [ç] [x] [h] [ts] [tS ] [m]
[n]
[] [l] [r]

plus, in words borrowed from other languages:

90 Answers to Exercises
 [] (die Passage) and, in more detailed

transcriptions:

[] [] [] [ ].

(c) Spanish: [p] [b] [t] [d] [k] [g] [f] [ ] [s] [x] [t] [m] [n] [ ] [l] [r] plus,

in more detailed transcriptions:

[] [ð] [z] [ ] [].

Section 5 (Transcription)
Features revealed by the narrow transcriptions (see the works listed in Section 6 for further
information).

French:

(1) The common tendency to pronounce [ ] and [ ] with a fronted tongue position, as
[ ] and [  ], so that they come to resemble [ø, œ].

(2) The dental articulation of [t], [d] and [n], as [t ] [d ], [n ]

(3) The various pronunciations of r (see 4.11).

(4) The devoicing of [m] to [m ] when a preceding sound is voiceless (isthme).

(5) The insertion of [ ] to break up groups of consonants. So [f blk] in plus faible que
becomes [f blk].

(6) The omission of [] in certain other circumstances, especially in rapid speech (e.g. est
l(e) plus étendu).

(7) The tendency in rapid speech to make adjacent consonants “agree” in respect of
voicing: thus type de pronounced [tibd ] rather than [tipd ]. Similarly in l’ouest du
continent. In this case the [t] at the end of ouest may merge with the [d] (oues’du),
causing the [s] to become a voiced [z], so as to match the [d] of du, which now
immediately follows it.

(8) Tendencies (6) and (7) are seen operating together in the case of de in aucun de ses
partenaires: [] is omitted, and [d] devoices to [t] as it now adjoins the voiceless [s] of
ses.

9) The way in which liaison consonants are pronounced as though they belonged to
the following word (which in fact they do, phonetically). So plus étendu is [ply zet
dy].

German:
(1) The aspiration of voiceless stops [p, t, k] before stressed vowels.

(2) The reduction of unstressed [e] to [ ].

(3) The various pronunciations of r (see 4.11).

Answers to Exercises 91
 (4) The reduction of certain very common words in rapid speech, illustrated here by und
and haben: [n], [ha m].

(5) Another feature of rapid speech is the omission of [ ] from Richtungen, and the
assimilation of the final [n] to the [ ] which now precedes it.

(6) The variation in the length of vowels according to their position.

(7) The insertion of a glottal stop [ ] before vowels at the beginning of words.

(8) The pronunciation of the eu diphthong: [ ] (with lips rounded throughout).

Spanish:

(1) The fricative articulation of [b, d, g] as [, ð, ], except at the beginning of words
(when not preceded by a vowel) and after nasals.

(2) The dental articulation of [t], [d] and [n], as [t ] [d ], [n ].

(3) The fact that in some positions in a word [e] may have a more open variant,
approximating to [ ].

(4) The relatively open articulation of [o] — often more like [ ].

(5) The voicing of [s] to [z] before a voiced consonant (misma).

(6) The elision of a vowel in rapid speech when followed by a similar vowel (de ese).

Practice transcriptions from French, German and Spanish.

French.

  tr  tj    eneral dy ma  et  f  n. plase lapar  j syr yn syrfas dyr e plat pur kil swa
bj   drwa. n  pa lese de sede, de pil u de kas  t   k   takt av  k lymidite, la plyi, l
 sabl, a la lymj  r dir  kt dy s  l  j u d   de z   drwa sys  ptibl d  k  n  tr de t  
peratyr el  ve k  m a pr  ksimite dapar  j d   ofa  u d   de vwatyr gare o s  l  j.
manipyle tu  ur l  sede   l  t  n   par le b  r e l  r    e d   sa bwat apr  lyza 
, la parti   prime turne v  r l  o.

German.

algmein  pfleg des ka stndks. di anla g auf ain  hart  ebn flç
tln, damt zi n çt mkpn kan. tsede s, bat rin nt ka stn fr fiçt
çkait, regn, zant nt d rkt r znntral ytsn nt n çt an rtn
aufbwarn, an denn s tsu hon t mpratu rn kmn kan (ts m baipil in
d r n fn haits n oder n ain m n d r zn gparkt n auto). di tsede  ste
ts am rant halten nt na x gbraux vidr n di axt l legn, m zi fr tsrkrats
n nt taup tsu ytsn.

Spanish.
manteni mjento xene al del ma etofono. uike el apa ato en una supe fije ðu
a i plana ðe tal ma ne a ke no se i kline. no es poga el apa ato, los eðe, los
bate ias, las pilas, ni los ka setes a la ume ðað, la uja, la a ena, o a la lu 
diekta ðel sol o ðexalos en lu aes donde okuren altas tempe atu as, tales
komo ekan ias de apa atos de kalefak jon o en auto moiles esta jonaðos en el
sol. sjempe a wante el eðe pa a el bode i wadelo en su es tut e ðes pwes
del uso, paa eita  ke se aje i se en suje.

92 Answers to Exercises
 further reading
[7]
The following two introductions to phonetics cover much the same ground as
this course, but with the emphasis mainly on English:

P. Ashby, Speech Sounds (Routledge, 1995).

J. C. Wells & G. Colson, Practical Phonetics (Pitman, 1971).

More advanced coverage is to be found in:

P. Ladefoged, A Course in Phonetics (Harcourt, 2001).

M.J. Ball & J. Rahilly, Phonetics: the Science of Speech (Arnold, 1999)
86 Further Reading

H. Rogers, The Sounds of Language: an Introduction to Phonetics

(Pearson, 2000)

Most general books about Linguistics include an introductory chapter on

phonetics. One good example is:

V. Fromkin & R. Rodman, An Introduction to Language (Holt Rinehart,

1998): chapter 2.

The standard work of reference for English phonetics is:

A. C. Gimson, An Introduction to the Pronunciation of English (Arnold,

1980).

For other languages see:

B. Tranel, The Sounds of French (Cambridge University Press, 1988).

A. Coveney, The Sounds of Contemporary French (Elm Bank, 2001).
C. Hall, Modern German Pronunciation (Manchester University Press,
1992).
R.M. Hammond, The Sounds of Spanish (Cascadilla, 2001).