CTLF Corpus de textes linguistiques fondamentaux • IMPRIMER • RETOUR ÉCRAN
CTLF - Menu général - Textes

Jones, Daniel. An Outline of English Phonetics – T01

“No theory is of any value unless it works
in practice. There are no such things as
pure science and applied science: there is
only science and the application of science.”

Louis Pasteur.

Chapter I
Introductory

The Nature of Speech

1. Spoken language consists of successions of sounds 11 emitted
by the organs of speech, together with certain ‘attributes’. 22

2. These successions of sounds are composed of (1) speech-sounds
proper, and (2) glides.

3. Speech-sounds are certain acoustic effects voluntarily produced
by the organs of speech; they are the result of definite
actions performed by these organs. A glide is the incidental
transitory sound produced when the organs of speech are passing
from the position for one speech-sound to that of another by the
most direct route.

4. Speech-sounds are made voluntarily; they require that the
speech-organs shall be placed in certain definite positions or moved
in certain definite ways. The speaker has to go out of his way
in order to make a speech-sound.

5. On the other hand the speaker does not have to go out of
his way in order to make a glide; glides occur as the natural and
inevitable result of pronouncing two speech-sounds one after the
other. 33

6. Most glides are inaudible or hardly audible even to the most
practised ear; most of the glides occurring in English require no
special consideration in the practical teaching of the language. 441

Difficulties of Pronunciation

7. The student of spoken English or any other spoken language
is faced at the outset with difficulties of five kinds in the matter
of pronunciation. They are as follows:

8. Difficulty No. 1. He must learn to recognize readily and
with certainty the various speech-sounds occurring in the language,
when he hears them pronounced; he must moreover learn to
remember the acoustic qualities of those sounds.

9. Difficulty No. 2. He must learn to make the foreign sounds
with his own organs of speech.

10. Difficulty No. 3. He must learn to use those sounds in
their proper places in connected speech.

11. Difficulty No. 4. He must learn the proper usage in the
matter of the ‘sound-attributes’ or ‘prosodies’ as they are often
called (especially length, stress and voice-pitch).

12. Difficulty No. 5. He must learn to catenate sounds, i.e.
to join each sound of a sequence on to the next, and to pronounce
the complete sequence rapidly and without stumbling.

13. The ultimate object of the language learner is to be able
to pronounce properly without having to pay any particular
attention to the way in which he does it. To attain this end he
must in the initial stages of his study focus his attention continually
on the above-mentioned details of the mechanism of speech. After
long practice he will gradually acquire the power of pronouncing
correctly without thinking of these details.

14. The student who wishes to become proficient in the written
as well as the spoken language, has an additional difficulty, which
we may call Difficulty No. 6. He has to learn the shapes of the
conventional letters and the relations between the conventional
orthography and the pronunciation.

15. Ability to speak a language or understand it when spoken
does not involve the ability to read or write it in the conventional
way. One may learn to speak English perfectly without ever
seeing ordinary English orthography. And conversely it is possible
to learn to read and write the language without being able to
pronounce it.2

16. As, however, those who wish to learn to speak and understand
English almost always wish to be able to read and write
it as well, a good deal is said in this book about ‘orthoepy’ or the
relation between pronunciation and conventional spelling. Every
word given as an example is accordingly shown both in phonetic
transcription and in ordinary spelling.

How to surmount the Difficulties of Pronunciation

17. We will now explain more fully the nature of the five
difficulties of pronunciation, and indicate shortly the appropriate
methods for enabling the student to surmount them.

18. Difficulty No. 1 is a matter of ‘ear-training’ or more accurately
‘cultivation of the auditory memory.’ No one can hope to be a
successful linguist unless he has a good ear. If his ear is unsensitive
by nature, it may be made more sensitive by training;
and if his ear is good by nature, it can be made still better by
training.

19. The possession of a good ear is necessary to the linguist
for two reasons. (1) If he has a good ear, he will be able to tell
whether he pronounces the foreign sounds correctly or not. (2) A
good ear helps him to understand the language readily when
spoken by natives; he recognizes words instantly, and does not
mistake one word for another.

20. The possession of a good ear involves (1) ability to
discriminate between sounds, (2) ability to remember the acoustic
qualities of foreign sounds, and (3) ability to recognize foreign
sounds with ease and certainty. In other words, the student
must be able (1) to hear the differences between the various sounds
of the foreign language, and between foreign sounds and sounds
of his mother tongue, (2) to bring into his consciousness, without
the aid of any external stimulus, memories of foreign sounds
previously heard, and (3) to compare sounds subconsciously with
the memory-images of sounds previously heard.

21. To cultivate a good linguistic ear requires systematic practice
in listening for sounds. There is only one effective exercise for
this purpose, namely, dictation of isolated sounds and meaningless
words by a teacher who can pronounce the foreign sounds
3accurately. The pupil should write down these sounds and words
phonetically. If he makes a mistake in his transcription, it shows
that he has confused one sound with another. The teacher will
in this case repeat the two sounds a number of times (both isolated
and in syllables) in order to impress on the pupil's mind the difference
of acoustic quality. 55

22. If the pupil is a beginner, the teacher may use for ear-training
exercises real words which the pupil has not yet learnt.

23. Examples of invented words suitable for training the ear to
recognize the English sounds are given in § 237 and in Appendix B.

24. A few invented words for ear-training practice should be
given at the beginning of every pronunciation lesson, until the
pupil can be fairly certain of doing the exercises without mistakes. 66

25. Difficulty No. 2 is a matter of gymnastics of the vocal organs.
In order to learn to form the speech-sounds of a foreign language,
the student has to learn to put his tongue, lips, and other parts
of the organs of speech into certain definite positions, or to perform
with them certain actions. He will learn to make such sounds
with the greatest accuracy and in the shortest time if (1) he is
told precisely what to do with his organs of speech, and (2) he is
given, as far as may be necessary, exercises which help him to
carry out the instructions.

26. In other words, the language learner should (1) study
phonetic theory, and (2) do, when necessary, exercises based on
that theory.

27. Difficulty No. 3 requires very different treatment. The
student has to learn what is the appropriate order in which to
4place the sounds so as to make intelligible words and sentences.
This is a matter of memorizing.

28. The student of spoken English has, for instance, to know
that if he wishes to communicate the idea expressed in French
by ‘armoire’ and in German by ‘Schrank,’ he must form the
English sounds kʌbəd 77 one after the other in this order. No
5other English sounds will do, nor may these sounds be placed in
any other order. However well the student may pronounce the
sounds, he will not convey the meaning, unless he uses this particular
sequence of them. He must therefore take care to remember that
this is the required sequence.

29. The task of learning to remember what is the appropriate
sequence of sounds to use in any given word or sentence is greatly
facilitated by the use of Phonetic Transcription.

30. Phonetic transcription may be defined as an unambiguous
system of representing pronunciation by means of writing, the
basic principle being to assign one and only one letter to each
phoneme of the language. (See Chap. X and Appendix A.)

31. Phonetic transcription, then, is a convenient method of
showing sound-order graphically. This graphic representation of
sound-order appeals to the visual memory and thus assists the
auditory memory.

32. Conventional English spelling is far from being phonetic;
it does not give the accurate information as to sound-order required
by the student of spoken English. In the first place English assigns
to many of the letters of the alphabet values quite different from
those which people in foreign countries are accustomed to associate
with them: e.g. the a in gate, the i in find, the u in tune. 88 Doubtless
these values may be learnt without difficulty; but as soon as the
foreign student has learnt them, he finds innumerable words in
which these letters have quite different values: compare the a's
in father, fall, any, fat, watch, 99 the i's in wind (noun), machine,
bird, 1010 the u's in rule, put, hut 1111; compare also the o's in stove, move,
love, 1212 the ea's in meat, head, great, bear, 1313 etc.

33. He also finds that many English sounds may be spelt in a
large number of different ways. Thus the words meet, meat, niece,
pique, key, quay, seize all have the same vowel-sound 1414; so also
6have the words sauce, lawn, stalk, stork, board, warn, thought, broad,
floor (in Southern English). 1515

34. Discrepancies between pronunciation and ordinary spelling
are not confined to the English language. In French -lle has
different values in ville and fille, 1616 o has different values in grosse
and gosse, 1717 portions is pronounced in two different ways according
as it is a noun or a verb 1818; on the other hand, the sound o is spelt
differently in the words mot, tôt, beau, chevaux. 1919 In German ch
has different values in rauchen and Frauchen, 2020 and u has different
values in Fuss and Nuss. 2121

35. The result of such inconsistencies is that the foreign learner
who depends solely on ordinary orthography is in innumerable
cases at a loss to know what sounds should be used, and is continually
mispronouncing words. Such mispronunciations may,
however, be avoided by the use of Phonetic Transcription.

36. The phonetic alphabet used here is that of the Association
Phonétique Internationale
. 2222 A list of the symbols used for transcribing
English in this book is given on p. xvi. Further information
regarding phonetic transcription is given in Chap. X and in
Appendix A.

37. It must be borne in mind that phonetic transcriptions are
valueless to students who have not learnt to form the sounds
which the phonetic letters represent, i.e. who have not surmounted
with tolerable success Difficulty No. 2. When, however, the
student can make the individual sounds with fair accuracy, he will
be in a position to begin learning sequences of sounds; phonetic
transcriptions will tell him what are the proper sequences to learn in
order to express the ideas he wishes to communicate.7

38. Unless the student can pronounce every one of the sounds
kʌbəd with tolerable accuracy, he will not be able to say the
English word for ‘armoire’ (‘Schrank’) in an intelligible manner,
even if he is furnished with a phonetic transcription. If on the
other hand he has learnt to make those sounds, he will be able
to say the English word provided he knows in what order the sounds
are to he placed
. He will remember this order if he sees the word
written ˈkʌbəd better than he would if he simply heard the word
pronounced and never saw it written.

39. Difficulty No. 4 concerns certain characteristics which sounds
and syllables have relative to other sounds and syllables in the
sentence. 2323 In particular the student will generally be able to
pronounce correctly in the matter of length, stress, and pitch, if
accurate information as to the foreign usage in regard to these
matters is supplied to him. Sometimes such information may be
supplied by means of rules, and sometimes (as in the case of English
stress) it is better conveyed by marks in the phonetic transcriptions.
There is as a rule nothing particularly difficult in carrying out such
instructions. The main difficulty in connexion with them is to
bear them in mind—again a question of memory.

40. Difficulty No. 5 must be carefully distinguished from all the
preceding. It sometimes happens that a student can pronounce
isolated sounds correctly, knows what sequence of sounds to use
in a given word or sentence, and knows the necessary details in
regard to length, stress, and pitch, but he stumbles over the sound-sequence.
He has not acquired facility in passing from one sound to
another, and he cannot always say sequences of sounds rapidly and
without stumbling. In other words, he does not ‘catenate’ properly.

41. Moreover he may have acquired the bad habit of stopping
between words; he has not realized the important principle that
the only places where pauses are normally made are at the ends
of certain grammatical groups called sense-groups (Chap. XXX).

42. Ability to catenate sounds, i.e. to pronounce sequences of
sounds with rapidity and without stumbling, can be cultivated by
continued repetition of such sound-sequences as present difficulty.
The required sequences must be pronounced at first slowly and
then with gradually increasing speed. Some suggestions for such
8repetition exercises will be found in Appendix C. Exercises of this
type may with advantage be followed by systematic fluency
exercises on the lines of ‘substitution tables.’ 2424

43. It must be borne in mind continually that in ordinary
conversation people say their sentences at an average rate of
some 300 syllables to the minute or five syllables per second.
This then is the ideal rate at which the student should aim, when
practising any given sentence. When practising catenation
exercises the student should frequently time himself, to see how
near he can get to this ideal rate.

44. It is worse than useless to try to say words fast or to
attempt catenation exercises until the individual sounds have been
thoroughly mastered. Above all, the student must beware of being
led astray by an idea that he can learn to make a difficult foreign
sound by merely repeating words containing that sound. That
idea is an absurd one. Repeating words with badly pronounced
sounds has the precisely opposite effect of fixing the student's bad
pronunciation. Speech-sounds are learnt by the methods referred
to in § 25.

Need for Oral Instruction

45. Some features of pronunciation can only be learnt with
the aid of a teacher; others can be learnt from books.

46. The services of a teacher are required mainly in connexion
with Difficulties Nos. 1 and 2.

47. The functions of the teacher in regard to these difficulties
are: (1) to act as a model of pronunciation, (2) to give the pupil
ear-training exercises as described above (§ 21), (3) to tell him
whether his attempts at the pronunciation of the foreign sounds
and sound-sequences are successful or not, and (4) where the
instructions in books are inadequate, to devise means which will
help the pupil to improve his pronunciation of the difficult sounds
and sound-sequences.

48. Good gramophone records can to some extent relieve the
teacher of the first of these duties.

49. In regard to the teacher's fourth duty, it should be remarked
that all students do not have the same difficulties, and a book
9on pronunciation cannot provide for the needs of every individual
student. The most that a book can do is to deal with the difficulties
of pronunciation most frequently met with. The rest must be
left to the phonetically trained teacher.

50. Sound-order can be learnt from books of phonetic texts.
Usage in regard to length, stress, and pitch can likewise be ascertained
from books. To attain ability to catenate properly requires
neither book nor teacher; it is a matter of private practice on the
part of the student.

Utility of Books on Pronunciation

51. It will be seen from what has been said, that though the
acquisition of a spoken language is essentially an oral process,
yet a book on pronunciation may be of service in several ways.

52. The directions in which such a book may assist the language
learner are illustrated by the present ‘Outline’ which contains:

(1) descriptions of the English speech-sounds together with
information as to their usage (Chaps. X1V-XXVII),

(2) information as to English usage in the matter of length,
stress, and pitch (Chaps. XXVIII-XXXI),

(3) descriptions of mistakes of pronunciation that foreign
learners frequently make (Chap. XIV, etc.),

(4) indications of methods which will help them to avoid
such mistakes (Chap. XIV, etc.),

(5) specimen catenation exercises and lists of words illustrating
the use of the various sounds (Chaps. XIV-XXVII
and Appendix C),

(6) specimen exercises to be dictated for ear-training (Chap.
XIII and Appendix B).

53. To perfect his pronunciation the learner should do a considerable
amount of reading aloud from phonetic texts. Some
books suitable for this purpose are mentioned in the Bibliography
(Appendix E). He should also have much practice in transcribing
phonetically passages of the language he is learning. His transcripts
should, as a rule, be corrected by his teacher. He may,
however, derive much benefit from transcribing passages from a
book of phonetic texts containing a key in ordinary spelling (by
transcribing from the key and correcting his transcripts by reference
to the phonetic version).10

Chapter II
Types of Pronunciation

54. The first question that confronts a person wishing to acquire
an acceptable pronunciation of a foreign language is: Which of the
various forms of pronunciation ought he to learn ?

55. No two persons of the same nationality pronounce their
own language exactly alike. The differences may arise from a
variety of causes, such as locality, social surroundings or early
influences, and there are often individual peculiarities for which
it is difficult or impossible to account.

56. Thus, the pronunciation current among people brought up
in Manchester differs from that of those from Exeter, and both
differ from the pronunciation of those brought up in Edinburgh
or in London. The French of Paris is different from that of
Marseilles or Lausanne; the pronunciation of educated Germans
from Berlin differs considerably from that used by those coming
from Dresden, Cologne or Hamburg.

57. An example of differences of English pronunciation according
to locality may be found in the treatment of the letter r in such
words as part. In Scotland the r in this word is generally pronounced
as a slightly rolled or flapped r, but in London English the
pronunciation is pɑːt. In many parts of the North and the West of
England on the other hand, the effect of the r appears as a modification
known as ‘retroflexion’ or ‘inversion’ of the preceding vowel,
thus pɑɹt or pɑːt/ɹ(see §§ 831-833). In the South of England the
vowels in boot and book are different (phonetically buːt, buk), but
in Scotland a short close u is generally used in both words.

58. The following are examples of differences between ‘educated’
and ‘uneducated’ speech. People of limited education in many
parts of England omit the sound h altogether; they say elp for
help. In London Dialect (Cockney) words like name are pronounced
with a diphthong ai or sei instead of with ei; and words like house,
about are pronounced with æu, or sometimes æə, instead of au
(æus, əˈbæut or əˈbæət). In uneducated Yorkshire speech the
vowels of put (put) and cut (kʌt) are levelled to a vowel intermediate
between these two.11

59. Differences between the pronunciation of old and young
people, and between that of women and men of the same locality
and similar position, may sometimes be observed. Thus in English
the word soft is pronounced sɔːft in the South by many educated
elderly men, but younger people, and particularly the women,
pronounce for the most part sɔft 125; and the use of hw in place of the
more usual w in such words as which, when, would seem to be more
prevalent among women than among men (in the South of England).

60. Individual peculiarities may be the result of habit, e.g.
childish mispronunciations which have never been corrected, or
they may arise from some physical defect.

61. The existence of all these differences makes it difficult for
the foreign learner to know which type of English pronunciation
to acquire. I do not consider it possible at the present time to
regard any special type as ‘Standard’ or as intrinsically ‘better’
than other types. Nevertheless, the type described in this book is
certainly a useful one. It is based on my own (Southern) speech,
and is, as far as I can ascertain, that generally used by those who
have been educated at ‘preparatory’ boarding schools and the
‘Public Schools.’ 226 This pronunciation is fairly uniform in these
schools and is independent of their locality. It has the advantage
that it is easily understood in all parts of the English-speaking
countries; it is perhaps more widely understood than any other
type. For further information about it, see the Introduction to my
English Pronouncing Dictionary, 11th ed., re-set, 1956 327; also Wyld's
History of Modern Colloquial English, 428 Chap. I.

62. The term ‘Received Pronunciation’ (abbreviation RP) is
often used to designate this type of pronunciation. This term
is adopted here for want of a better. I wish it, however, to be
understood that other types of pronunciation exist which may be
considered equally ‘good.’

63. It should be noticed here that all speakers use more than one
style of pronunciation. A person may pronounce the same word
or sequence of words quite differently under different circumstances.
Thus in ordinary conversation the word and is frequently
12pronounced n when unstressed (e.g. in bread and butter ˈbred n ˈbʌtə
but in serious recitation the word, even when unstressed, might
often be pronounced ænd rhyming with hand hænd.

64. Several different styles of pronunciation may be distinguished.
Notable among them are the rapid familiar style, the
slower colloquial style, the natural style used in addressing a fair-sized
audience, the acquired style of the stage and the acquired styles
used in singing. Of these the slower colloquial style is probably the
most suitable for the use of foreign learners, and is the style indicated
throughout this book, except where the contrary is stated. 52913

Chapter III
The Organs of Speech

65. It is necessary that the student of phonetics should have a
fairly clear idea of the structure and functions of the organs of
speech. Those who have not already done so should make a

image nasal cavity | mouth | tongue | larynx

Fig. 1. The Organs of Speech.

B Back of Tongue. Bl Blade of
Tongue. E Epiglottis. F Front
of Tongue. FP Food-passage. H
Hard Palate. LL Lips. P Pharyngal
Cavity (Pharynx), R Root
of the Tongue. S Soft Palate.
TR Teeth-ridge. TT Teeth. U
Uvula. V Position of Vocal Cords.
W Windpipe.

thorough examination of the inside
of the mouth by means of a
hand looking-glass. The best
way of doing this is to stand with
the back to the light and to hold
the looking-glass in such a position
that it reflects the light into
the mouth and at the same time
enables the observer to see in the
glass the interior thus illuminated.
It is not difficult to find
the right position for the glass.

66. Models of the organs of
speech will be found useful.
Suitable models of sections of
the head, mouth, nose, larynx,
etc., may be obtained from
dealers in medical instruments.

67. Figs. 1 and 2 show all that
is essential for the present book.

68. A detailed description of
the various parts of the organs
of speech is not necessary; the
following points should, however,
be noted.

69. The roof of the mouth is
divided, for the purposes of
phonetics, into three parts called the teeth-ridge, the hard palate,
and the soft palate. The teeth-ridge is defined as the part of the
14roof of the mouth just behind the teeth which is convex to the
tongue, the division between the teeth-ridge and the palate being
defined as the place where the roof of the mouth ceases to be
convex to the tongue and begins to be concave (see Fig. 1). The
remainder of the roof of the mouth comprises the other two parts,
the front part constituting the hard palate, and the back part the
soft palate. These two parts should be examined carefully in the
looking-glass; they may be felt with the tongue or with the finger.
The soft palate can be moved upwards from the position shown
in Fig. 1, and when raised to its fullest extent it touches the back
wall of the pharynx as shown in
Fig. 10, etc. (see also § 165).

image

Fig. 2. The Mouth.

A A Pharyngal Arch. PP Pharyngal
Cavity (Pharynx). S Soft
Palate. T Tongue. U Uvula.

70. The pharynx is the cavity
situated in the throat immediately
behind the mouth. Below it is the
larynx which forms the upper part
of the windpipe (the passage leading
to the lungs). The epiglottis is a
sort of tongue situated just above
the larynx. It is probably contracted
in such a way as to protect
the larynx during the action of
swallowing, but it does not appear
to enter into the formation of any
speech-sounds.

71. For the purposes of phonetics it is convenient to imagine
the surface of the tongue divided into three parts (see Fig. 1). The
part which normally lies opposite the soft palate is called the
back; the part which normally lies opposite the hard palate is
called the front; and the part which normally lies opposite the
teeth-ridge is called the blade. The extremity of the tongue is
called the tip or point, and is included in the blade. The definitions
of ‘back’ and ‘front’ are particularly important. It is sometimes
convenient to use the term middle of the tongue to denote a part
of the surface of the tongue including the fore part of the ‘back’
and the hinder part of the ‘front.’

72. The tongue is extremely mobile. Thus the tip can be made
to touch any part of the roof of the mouth from the teeth to the
15beginning of the soft palate. The other parts of the tongue may
likewise be made to articulate against different parts of the roof
of the mouth.

image

Fig. 3. Lateral Spreading
of the Tongue.

73. Moreover it is possible to spread
out the fore part of the tongue laterally
(after the manner shown in Fig. 3), or
to contract it laterally (after the manner
shown in Fig. 4). The presence or
absence of such lateral contraction is
probably immaterial for most sounds,
but there are a few in which lateral
contraction appears to play an essential
part (see particularly §§ 747, 831).

image

Fig. 4. Lateral Contraction
of the Tongue.

74. The vocal cords are situated in the
larynx; they resemble two lips (see
Fig. 6). They run in a horizontal
direction from back to front. The space
between them is called the glottis. The
cords may be kept apart or they may
be brought together so as to touch and
thus close the air passage completely. When they are brought
near together and air is forced between them, they vibrate, producing
a musical sound (see Chap. V).

75. In the larynx just above the vocal cords is situated another
pair of lips somewhat resembling the vocal cords and running
parallel to them. These are known as the false vocal cords (see
§ 82).16

Chapter IV
Experimental Methods

76. The analysis of speech-sounds in general and the differences
in articulation between English sounds and foreign sounds which
resemble them may, if desired, be investigated and demonstrated
by means of specially designed apparatus. Such demonstrations
belong to the branch of phonetic science known as ‘experimental’
or ‘instrumental’ phonetics.

77. It is not suggested that experimental phonetics is a necessary
study for all those who wish to pronounce a foreign language
correctly, but demonstrations by means of special apparatus are
sometimes found helpful by students as fixing in the memory facts
which they have previously learnt by the ordinary methods of
practical phonetics.

78. The apparatus used in elementary experimental phonetics
includes the artificial palate, the phonetic kymograph, the laryngoscope,
X-ray photography, sensitive flames, tape recorders and
other recording and reproducing machines. In more advanced
work use is made of apparatus for enlarging the curves on records,
cathode ray oscillographs, harmonic analysers, spectrographs and
much other apparatus. It is not necessary for the purposes of
this book to say much about experimental methods, beyond giving
palatograms of various sounds. 130

79. A palatogram is a drawing showing the parts of the roof
of the mouth with which the tongue makes contact in pronouncing
sounds. Palatograms may be made by means of a special kind of
artificial palate. Suitable palates in metal or vulcanite can be
made by any dentist. They should be of the form shown in Fig. 5,
the material must be very thin, and the palate must fit the
observer's mouth exactly. It must be so made that it will keep
in position by itself, and it should be provided with Little projecting
17pieces in the front (AA Fig. 5) so as to admit of its being removed
from the mouth easily. If the material is not black the under
side should be blackened with varnish.

80. The artificial palate is used as follows. The under side of
the palate is first covered with a little finely powdered chalk and
inserted into the mouth. A sound is then pronounced and the
palate is withdrawn. The parts of
the palate from which the chalk has
been removed show the points at
which the tongue touched it. These
marks on the artificial palate may
then be examined at leisure. They

image (I)
image (II)image (III)

Fig. 5. The Artificial Palate.

(I) Side View. (II) Seen from above. (III) Seen from below.

may also be photographed if desired, or the marks may be copied
in projection on outline diagrams of the palate. D. Abercrombie,
of the University of Edinburgh, has recently devised a means of
photographing the palate directly, thus dispensing with the use of
an artificial palate. See his article Direct Palatography in the
Zeitschrift für Phonetik (Berlin), 1955.

81. The palatograms in this book have been drawn from observations
made with vulcanite palates. These palates extend so as
to cover the whole of the front teeth. The limits of the gums
adjoining the front teeth are marked on the present diagrams
by the dotted line (Figs. 37, 39, etc.).18

Chapter V
Breath and Voice

82. The vocal cords are capable of acting in much the same
way as the lips of the mouth. Thus they may be held wide apart,
they may be closed completely, or they may be held loosely together
so that they vibrate when air passes between them. When they
are held wide apart (i.e. when the glottis is open) and air passes
between them, the sound produced is called breath. When they
are drawn near together and air is forced between them so that
they vibrate, the sound produced is called voice. If the false
vocal cords (§ 75) are drawn towards each other leaving only a
narrow space for the air to pass between them, the resulting sound
is one variety of whisper. It is believed that certain positions of
the glottis intermediate between those for breath and voice give
rise to other varieties of whisper.

83. The vocal cords may be made to touch tightly along their
whole length so that no air can escape at all. This is the position
known as closed glottis. The explosive sound heard when this
position is released is known as the ‘glottal stop’ (see §§ 552, ff.).

84. Breath is heard most clearly in the sounds represented by h.
Voice occurs as part of the articulation of numerous speech-sounds,
and particularly of the vowels.

image front | A | backimage front | A | back

Fig. 6. The Larynx as seen through the Laryngoscope. A Position for Breath.
B Position for Voice. TT Tongue. VV Vocal Cords. W Windpipe.

85. The positions of the vocal cords in the production of breath
and voice are shown in Fig. 6. These diagrams show the larynx
as seen from above through a laryngoscope.19

86. The Laryngoscope in its simplest form is a small circular
mirror, about 3/4 of an inch in diameter, which is fixed to a long
handle at an angle of 120°. When the instrument is held in the

image

Fig. 7. The Laryngoscope.

position shown in Fig. 7 and inserted into the mouth so that the
mirror is pressed against the soft palate as far back as possible,
and is adjusted so that a strong light is reflected down the throat,
the interior of the larynx is visible in the mirror.

87. Breath and voice may be illustrated artificially by the
following experiment. Take a short tube of wood or glass T,
say 4 cm. long and 1 cm. in diameter, and tie on to one end of
it a piece of thin rubber tubing R, of a somewhat larger diameter,
say 3 cm., as shown in Fig. 8. The tube is taken to represent

image

Fig. 8. Instrument to
illustrate
Breath and
Voice.

the windpipe, and the rubber part the larynx. The
space enclosed by the edge of the rubber E, E,
represents the glottis. If we leave the rubber part
in its natural position and blow through the tube,
air passes out, making a slight hissing sound. This
corresponds to breath. If we take hold of two opposite
points of the edge of the rubber tube, E, E, and
draw them apart so that two edges of the rubber
come into contact along a straight line, we have a
representation of the glottis in the position for voice,
the two edges which are in contact representing the
two vocal cords. Now, if we blow through the tube,
the air in passing out causes the edges to vibrate
and a kind of musical sound is produced. This sound corresponds
to voice.

88. Most ordinary speech-sounds contain either breath or voice.
Those which contain breath are called breathed or voiceless 131 sounds,
20and those which contain voice are called voiced sounds. Examples
of breathed sounds are f, s; examples of voiced sounds are v, z, b,
and the vowels.

89. When people speak in a whisper, whispered sounds are
substituted for all the voiced sounds, the breathed sounds remaining
unaltered.

90. It is possible to pronounce various sounds with simultaneous
closure of the glottis (§§ 83, 569-571). Sounds so pronounced are
neither breathed nor voiced. Sounds of this type do not occur in
English.

91. It does not require much practice for a person with a fairly
good ear to be able to recognize by ear the difference between
breathed and voiced sounds. Any students who have difficulty in
this should practise prolonging such sounds as s, z, f, v, ʃ, ʒ, θ, ð. 232
They may also try the following well-known tests. (1) Stop the
ears with the fingers, and pronounce the following sounds p, ɑ, 333
f, v; a loud buzzing sound will be heard during the utterance of
a and v, but not when p and f are sounded. (2) Pronounce the same
sounds while touching the outside of the larynx with the fingers; the
vibrations will be felt in the case of the voiced sounds. (3) Notice
that voiced sounds such as ɑ, v, can be sung, while breathed sounds
cannot.

92. When an assimilation (Chap. XXVI) takes place by which a
voiceless sound is substituted for a voiced sound, the voiced sound
is commonly said to become unvoiced or to be devocalized or better
devoiced.

93. The presence or absence of voice may be demonstrated
experimentally in various ways. One method is to use a metal

image

Fig. 9. Zünd-Burguet's Voice Indicator.

ring placed loosely in a tin match-box. A flat side of the box is
placed firmly against one side of the larynx; when voiced sounds
21are produced the ring rattles, though when breathed sounds are
produced it remains silent.

94. Zünd-Burguet's Voice Indicator (Fig. 9) is a convenient
instrument for demonstrating the presence of voice in a similar
way. 434

95. The presence or absence of voice may also be shown by
means of a phonetic kymograph, and voice vibrations appear very
clearly on magnifications of sound-tracks.22

Chapter VI
Vowels and Consonants

96. Every speech-sound belongs to one or other of the two
main classes known as Vowels and Consonants.

97. A vowel (in normal speech 135) is defined as a voiced sound
in forming which the air issues in a continuous stream through
the pharynx and mouth, there being no obstruction and no
narrowing such as would cause audible friction.

98. All other sounds (in normal speech 136) are called consonants.

99. Consonants therefore include (i) all sounds which are not
voiced (e.g. p, s, ʃ), (ii) all sounds in the production of which the
air has an impeded passage through the mouth (e.g. b, l, rolled r),
(iii) all sounds in the production of which the air does not pass
through the mouth (e.g. m), (iv) all sounds in which there is audible
friction (e.g. f, v, s, z, h).

100. The distinction between vowels and consonants is not an
arbitrary physiological distinction. It is in reality a distinction
based on acoustic considerations, namely on the relative sonority
or carrying power of the various sounds. Some sounds are more
sonorous than others, that is to say they carry better or can be
heard at a greater distance, when pronounced with the same
length, stress, and voice-pitch. Thus the sound ɑ pronounced in
the normal manner can be heard at a much greater distance than
23the sound p or the sound f pronounced in the normal manner.
It so happens that the sounds defined as vowels in § 97 are on the
whole more sonorous than any other speech-sounds (when pronounced
in the normal manner) 237; and that is the reason why these
sounds are considered to form one of the two fundamental classes. 338

101. The relative sonority or carrying power of sounds depends
on their inherent quality (tamber) and must be distinguished from
the relative ‘prominence’ of sounds in a sequence; prominence
depends on combinations of quality with length, stress and (in the
case of voiced sounds) intonation. When length and stress (degree
of push from the chest wall) are constant and the intonation is
level, the sounds defined as vowels are more prominent than the
sounds defined as consonants; ‘open’ vowels (§ 153) are mostly more
prominent than ‘close’ vowels (§ 153); voiced consonants are more
prominent than voiceless consonants; 1-sounds and voiced nasal
consonants are more prominent than other voiced consonants. The
voiceless consonants have very little prominence in comparison with
the voiced sounds, and the differences in prominence between the
various voiceless consonants may as a rule be considered as
negligible for practical linguistic purposes. It must always be
remembered, however, that more sonorous sounds may become
less prominent, and therefore more consonant-like, by diminishing
length or stress, and that sounds of relatively small sonority may
be made prominent by increasing length or stress.

102. It is as a consequence of this principle of relative prominence
that certain short vowel-glides must be regarded as
24consonants. Such are the English j (as in yard jɑːd) and w (as
in wait weit). In making these sounds the speech-organs start
in the position of i and u respectively and without remaining
there any appreciable time proceed very quickly to the position
of another vowel (ɑ in the case of yard and e in the case of wait).
Such vowel-glides are often called semi-vowels. It must be remembered
that such sounds have to be regarded as consonants on
account of (1) their gliding nature, (2) their shortness and (3)
their lack of stress as compared with the succeeding vowel. (See
§ 800.)

103. In the case of the word you juː the sound is actually
less sonorous than the i with which the j begins. Nevertheless,
the shortness and lack of stress of the vowel-glide suffice to render
the sound consonantal. (For further information regarding prominence,
see Chap. XII.)25

Chapter VII
Vowels

How to learn Vowels

104. Practical experience in teaching pronunciation shows that
consonants are as a rule best acquired by directing attention to
tactile and muscular sensations, whereas in learning vowels it is
necessary to direct attention more particularly to the acoustic
qualities of the sounds.

105. This does not mean that the learner is expected to acquire
vowels by ‘simple imitation.’ On the contrary, he will find a
knowledge of the organic formation of vowels of considerable
use to him. But this knowledge is not in itself sufficient. The
finer adjustments of the tongue have to be done by means of
sensory control from the ear.

106. In order to be able to use this sensory control properly,
the student must learn to estimate by ear the extent and nature
of the acoustic differences between one vowel and another. Fortunately
it is not difficult to devise a means by which he can
learn to do this, and hence to know what to do in order to make
any given foreign vowel.

How to describe Vowels

107. The method consists in explaining to the student the
relations between the foreign vowels and certain vowels already
known to him.

108. A bare description, however accurate, of the tongue-position
of a foreign vowel cannot convey much to the learner. A most
accurate diagram of the tongue-position obtained by X-ray photography
will not of itself enable the student to pronounce the
sound correctly.

109. If, however, the descriptions or diagrams are such as to
show the exact relations of the foreign vowel to certain known
vowels, they immediately become intelligible.26

110. Thus, if a certain foreign vowel is known to be formed with
a tongue position half-way between two sounds familiar to the
student, 139 that student will with practice be able to adjust his tongue
until he hears an acoustic quality which seems to have an equal
amount of resemblance to each of the two known sounds.

111. Those whose ears are naturally very keen will be able to
acquire the foreign vowels in this way with little or no trouble.
With those whose ears are duller by nature the process will take
longer; with many, a course of ‘ear-training’ of some length may
be necessary before they can learn to pronounce foreign vowels
with success.

112. An apt student whose ear has been well trained can estimate
pretty accurately not merely whether a sound is half-way between
two known sounds, but whether it is one-third or one-quarter
or even a smaller fraction of the distance between two known
vowels.

113. The process of acquiring vowels by means of estimates of
the acoustic distances between them and known vowels may be
carried out with tolerable success by a student working without a
teacher. But the results will be more successful and will be attained
in a shorter time if he has a teacher. Not only will the teacher
tell the student when he has hit upon the required intermediate
shade of sound, but if the student's attempt is unsuccessful, the
teacher will tell him in which direction to modify his sound in
order to improve it.

114. The lesson may be conducted in the following way: Teacher:
‘I want you to try and make a vowel-sound about one-third of
the way from so-and-so (which we may call X) to so-and-so (Y).
Will you please say these two sounds (X, Y)? Now try to make
the intermediate one. No, that won't do; that's hardly different
from X; make it sound more like Y. It's still too near X; make
it still more like Y. Now you've gone too far in the other direction;
it mustn't be quite so much like Y as that. Etc., etc’ By proceeding
in this way the student learns to produce the exact shade
of sound required, generally in a very short time.27

115. In some cases the directions will take the following form.
Teacher: ‘Now put your lips into this position, and while keeping
them there, try as hard as you can to say the vowel so-and-so.’

116. Now how are the principles above described to be applied
to the teaching of English vowels to foreign learners? With
reference to what vowels should the English vowels be described?

117. One answer is that the English vowels may be described
with reference to the vowels of the learner's mother tongue. A
phonetically trained teacher can easily ascertain the nature of
his pupil's vowels and use them as a basis for getting him to make
English vowels.

118. The author of a book cannot, however, base his descriptions
of foreign vowels on the vowels occurring in his readers'
pronunciation of their mother tongue, because all the readers of
a particular nationality do not have the same pronunciation.
Thus there exist several different ways of pronouncing the French
word bonne; the shade of vowel used depends on the locality from
which the speaker comes. So also the pronunciation of the first
vowel in the German Vater varies considerably according to locality.
It is therefore meaningless to speak of ‘the French vowel-sound
in bonne’ or ‘the German vowel-sound in Vater.’ Any description
of an English vowel which compared it to ‘the French vowel in
bonne,’ without further explanation, would be ambiguous; it would
be interpreted in different ways by different French readers.

119. There is only one way of making written descriptions of
vowels intelligible to a large circle of readers of different nationalities,
and that is to describe the sounds with reference to a scale of
‘Cardinal Vowels,’ i.e. a set of fixed vowel-sounds having known
acoustic qualities and known tongue and lip positions.

120. It has been found by experience that a scale of eight
cardinal vowels forms a convenient basis for describing the vowels
of any language. They are represented in the International
Phonetic Alphabet by the letters i e ɛ a ɑ ɔ o u. (See The Principles
of the International Phonetic Association
, 1949, pp. 4-6.)28

Chapter VIII
The Classification of Vowels

The Nature of Vowels

121. It follows from the definition of a vowel given in Chap. VI
that the characteristic qualities of vowels depend on the shape of
the open passage above the larynx. The passage forms a resonance-chamber
which modifies the quality (tamber) of the sound produced
by the vibration of the vocal cords. Different shapes of the

image

Fig. 10. Limit of Tongue-positions
for Vowels.

passage modify the tamber in different
ways and consequently give rise to
distinct vowel-sounds.

122. Now the shape of this passage
can be varied very greatly, even when
the organs are limited to vowel positions.
Fig. 10 illustrates the approximate limit
of vowel-positions; if the breath-force
is normal, the tongue must be in a
position below the dotted line in order
to produce a vowel. It will be seen
from this figure that the number of
possible vowels is very large. A good ear can distinguish well
over fifty distinct vowels (exclusive of nasalized vowels, vowels
pronounced with retroflex modification [§ 831], etc.). In any one
language, however, the number of distinct vowels is comparatively
small. In English it is not essential for ordinary purposes to
distinguish more than twelve pure vowel-sounds and nine
diphthongs.

123. The effect of a resonance-chamber in modifying tamber
may be illustrated experimentally by means of an instrument

image

Fig. 11. Instrument to show the effect of a resonance-chamber
in modifying quality of tone.29

such as that illustrated in Fig. 11; it is made by Messrs. Spindler
and Hoyer of Göttingen. It consists of a cylindrical resonator A,
open at one end, fitted with a piston B, the rod of which C passes
out at the other end. The piston-rod is hollow and the piston
contains a reed D, so that by blowing down the piston through
the opening E at the end of the rod, a musical sound of definite
pitch is produced by the reed. The tamber of this sound depends
on the length of the part of the cylinder projecting beyond the
piston; by varying the position of the piston a large number of
distinct tambers are obtainable, some of the sounds having
resemblance to well-known vowels.

124. The chief organs concerned in modifying the shape of the
passage are the tongue and the lips. Vowels are classified for
linguistic purposes according to the position of the tongue. (Note
that the position of the tip of the tongue has no great effect on
vowel quality, except when the tip is very much retracted or very
close to the roof of the mouth; see footnote to § 831.)

125. Some vowels have a clear and well-defined quality; others
have a more obscure sound. The latter are chiefly those which
are formed with the tongue in an intermediate vowel-position, not
raised markedly at the back or in the front, and not too low down in
the mouth. The most typical intermediate position gives rise to the
sound known as the ‘neutral vowel’ or ‘schwa’ (phonetic symbol a). 140

126. The vowels of well-defined quality are chiefly those in
which the tongue is remote from such intermediate position, that
is to say, they are those in which the tongue is markedly raised in
the front or at the back or is quite low down in the mouth. It
is from among these vowels which are as remote as possible from
‘neutral’ position that it has been found convenient to select the
eight ‘Cardinal Vowels’ referred to at the end of the last chapter.

127. The positions of the tongue in the formation of the different
vowel-sounds may, to a large extent, be felt, and in many cases
they may be seen by means of a looking-glass. They may also
be determined experimentally in various ways.

128. Palatograms are useful in this connexion (see Figs. 37,
39, etc.). It is desirable in making palatograms of vowels to
30take care that the teeth are always kept at the same distance
apart, because the diagram obtained depends not only on the
height of the tongue, but also on the degree of separation between
the jaws. The distance between the jaws may be kept constant
by holding the end of a pencil firmly between the teeth. The
pencil should not be more than 1 cm. in diameter. 241 When the
teeth are kept at a constant distance apart, the palatograms show
the correct relative positions of the tongue independently of the jaw.

129. The late Dr. E. A. Meyer of Stockholm obtained excellent
diagrams of the tongue-positions of vowels by means of a row of
fine leaden threads attached to an artificial palate along its centre
line. An account of this work was given in his Untersuchungen
über Lautbildung
in Festschrift Wilhelm Viëtor. 342

130. Valuable results have also been arrived at by means of
X-ray photography by E. A. Meyer (who invented the system of
laying a light metal chain on the tongue), Trevelyan George,
O. Russell, H. Gutzmann, and others.

Description of the Cardinal Vowels

131. Cardinal vowel No. 1 (i) is the sound
in which the raising of the tongue is as far
forward as possible and as high as possible
consistently with its being a vowel 443 (see
Figs. 10, 12), the lips being spread.

image

Fig. 12. Approximate
Tongue-positions
of the Cardinal Vowels
i and ɑ

132. Cardinal vowel No. 5 (ɑ) is a sound
in which the back of the tongue is lowered as
far as possible and retracted as far as possible
consistently with the sound being a vowel 544
(see Fig. 12), and in which the lips are not
rounded.31

133. Cardinal vowels 2, 3, and 4 (e, ɛ, a) are vowels of the
‘front’ series chosen so as to form an acoustic sequence between
the vowels 1 and 5 such that the degrees of acoustic separation
between each vowel and the next are equal, or, rather, as nearly
equal as it is possible for a person with a well trained ear to make
them. Cardinal vowels 6, 7, and 8 (ɔ, o, u) are vowels of the
‘back’ series chosen so as to continue this series of acoustically
equidistant vowels.

134. The approximate tongue-positions of these vowels are shown
in Figs. 13 and 14. The drawings of the tongue-positions of Cardinal

image

Fig. 13. Approximate
Tongue-positions of the Front
Cardinal Vowels, i, e, ɛ, a.

image

Fig. 14. Approximate
Tongue-positions of the Back
Cardinal Vowels, ɑ, ɔ, o, u.

vowels 1, 4, 5, and 8 are adapted from X-ray photographs. 645 The
drawings of the remaining vowel-positions are approximate.

135. The lip-positions of the Cardinal vowels are shown in
Figs. 15-22.

136. It will be observed that in passing from vowel No. 1 to
vowel No. 2 and then to No. 3 and then to No. 4, the tongue is
lowered through approximately equal intervals. Also that in passing
from No. 5 to No. 6 and then to No. 7 and then to No. 8, the tongue
is raised through approximately equal though smaller intervals.32

image

Fig. 15.
Lip-position of Cardinal i.

image

Fig. 16.
Lip-position of Cardinale.

image

Fig. 17.
Lip-position of Cardinal ɛ.

image

Fig. 18.
Lip-position of Cardinal a.

image

Fig. 19.
Lip-position of Cardinal ɑ.

image

Fig. 20.
Lip-position of Cardinal ɔ.

image

Fig. 21.
Lip-position of Cardinal o.

image

Fig. 22.
Lip-position of Cardinal u.33

137. The differences in tamber between Nos. 1, 2, 3, 4, and 5
are produced mainly by means of differences of tongue-position;
such differences of lip-position as there are (see Figs. 15, 16, 17,
18, 19) have but little effect on the sounds.

138. But the differences in tamber between Nos. 5, 6, 7, and 8
are produced by differences of tongue-position combined with
important differences of lip-position (see Figs. 19, 20, 21, 22).

139. It is for this reason that the distances between the tongue-positions
of Nos. 5, 6, 7, and 8 are less than the distances between
the tongue-positions of Nos. 1, 2, 3, and 4 (Figs. 13, 14).

140. The values of cardinal vowels cannot be learnt from
written descriptions; they should be learnt by oral instruction
from a teacher who knows them. The teacher will impart them
to the student by the means described in §§ 107-115, basing his
instruction on the sounds which he finds the learner to possess
in his mother tongue.

141. The cardinal vowels here referred to have been recorded on
gramophone discs. 746 The student who has not access to a teacher
familiar with the cardinal vowels may learn them with fair accuracy
by listening over and over again to the sounds as reproduced by
the gramophone, and adjusting his speech-organs till he makes a
sound indistinguishable (to his ear) from that produced by the
machine. His success will depend mainly on the sharpness of his
ear. A study of the tongue and lip positions will help him in some
measure, especially if he has been able to cultivate some control
over the movements of his tongue.34

142. Those who have access neither to a qualified teacher nor
to a gramophone cannot expect to learn the values of these or any
other cardinal vowels with accuracy. For the benefit of such
students we append some very rough indications of the values of
our cardinal vowels by means of key-words. It must be remembered,
however, that to attempt to describe cardinal vowels by
means of key-words is to put the cart before the horse. It is the
vowels of the ‘key-words’ that should be described with reference
to the cardinal vowels. Moreover, most ‘key-words’ are pronounced
in different ways by different people; accordingly descriptions of
sounds by reference to key-words convey different meanings to
different readers.

143. In the following table Eng. refers to what I believe to be
the average speech of educated Southern English people; Fr. means
French as spoken (as far as I am able to judge) by the average
educated Parisian; Ger. means German as spoken (as far as I a
able to judge) by the average educated inhabitant of Berlin.

144. Number of cardinal vowel | Phonetic sign | Nearest equivalent in Eng., Fr. or Ger.

1 | i | Fr. sound of i in si; Ger. sound of
ie in Biene

2 | e | Fr. sound of é in thé; Scottish pronunciation
of ay in day

3 | ɛ | Fr. sound of ê in même

4 | a | Fr. sound of a in la

5 | ɑ | Nearly what is obtained by taking
away the lip-rounding from Eng.
sound of o in hot; Fr. vowel in pas

6 | ɔ | Ger. sound of o in Sonne

7 | o | Fr. sound of o in rose; Scottish pronunciation
of o in rose

8 | u | Ger. sound of u in gut

Secondary cardinal vowels

145. A set of secondary cardinal vowels may be derived from
the eight primary cardinal vowels by changes of lip-position.
The most important of these are:

(1) the sounds obtained by adding close lip-rounding to Cardinal
Vowels Nos. 1 and 2 (phonetic signs y, ø);35

(2) the sounds obtained by adding open lip-rounding to Cardinal
Vowels Nos. 3 and 5 (phonetic signs œ and ʋ 847);

(3) the sounds obtained by combining the tongue-positions of
Cardinal Vowels Nos. 6, 7, and 8 with lip-spreading (phonetic
signs ʌ, ɤ, ɯ).

146. Further secondary cardinal vowels may also be selected
in ‘central’ (§ 155) positions. Thus vowels having tongue-positions
half-way between those of i and u may be considered as cardinal
vowels. The unrounded and rounded vowels with this tongue-position
are represented by the letters ɨ, . When they occur
as subsidiary members of i or u phonemes, they may according to
the usage of the I.P.A. be represented by ï, ü.

147. It is possible, though perhaps hardly necessary, to fix
other cardinal vowels in lower central positions, namely, sounds
having tongue-positions half-way between those of e and o and
between those of ɛ and ɔ. The rounded central vowel half-way
between o and ø is represented in International Phonetic notation
by ø.

Details of Vowel Classification

Tongue Positions

148. The tongue-positions of the eight primary cardinal vowels
(see Figs. 13, 14) may be represented diagrammatically as in
Fig. 23 where the relative
positions of the highest points
of th tongue are shown by
dots.

image front | i | e | ɛ | a | back | u | o | ɔ | ɑ

Fig. 23. Diagram illustrating the
Tongue-positions of the eight primary
Cardinal Vowels.

image i | e | ɛ | a | u | o | ɔ | ɑ

Fig. 24. A more
accurate Form of
Vowel Diagram.36

149. The shape of this diagram is a compromise between scientific
accuracy and the requirements of the practical language teacher.
Scientific accuracy would require a diagram with curved sides
somewhat as shown in Fig. 24. This shape, however, is inconvenient
for use in practical teaching. Practical teaching requires
a definiteness which can only be attained by means of a figure
bounded by straight lines. Fig. 23 indicates with very considerable
accuracy the relative tongue-positions of the vowels, the relative
positions of i, a, ɑ, and u having been obtained from X-ray
photographs. 948

150. The cardinal vowels have by definition (§ 126) tongue-positions
as remote as possible from ‘neutral’ position. Accordingly
if other vowels are represented by dots on the above geometrical
figure, they will be situated either on the circumference of that
figure or within it. Thus a dot placed on the circumference of
Fig. 23 half-way between the second and third points would indicate
a sound which (if unrounded) would have an acoustic quality
half-way between the cardinal sounds e and ɛ.

151. The tongue-positions of vowels can thus be classified by
means of a system similar to the latitude and longitude principle
used in geography. They are classified (1) according to the height
37to which the tongue is raised, and (2) according to the part of
the tongue which is raised highest.

152. When we classify according to the height of the tongue,
we distinguish four classes. Vowels which have their tongue-positions
on the line i—u (Fig. 23) are called close vowels; those
which have their tongue-positions on the lines e—o, ɛ—ɔ, a—ɑ,
are called half-close vowels, half-open vowels and open vowels
respectively.

153. The terms close, half-close, half-open and open may be
defined more precisely as follows:

Close vowels are those in which the tongue is raised as high as
possible consistently with the sounds remaining vowels
(see §§ 97, 122);

Open vowels are those in which the tongue is as low as possible;

Half-close vowels are those in which the tongue occupies a position
about one-third of the distance from ‘close’ to ‘open’;

Half-open vowels are those in which the tongue occupies a position
about two-thirds of the distance from ‘close’ to ‘open.’

154. When we classify vowels according to the part of the tongue
raised, we distinguish three classes. The vowels which have their
tongue-positions on or near the line i—a (Fig. 23) are called front
vowels
; in other words, front vowels are those in the formation of
which the ‘front’ of the tongue is raised in the direction of the
hard palate. Vowels which have their tongue-positions on or
near the line ɑ—u are called back vowels: in other words, back
vowels
are those in the formation of which the ‘back’ of the tongue
is raised in the direction of the soft palate.

155. Vowels in which the
highest point of the tongue is in
the centre part of the vowel
figure are called central vowels.
They lie in an area which is
triangular in shape, when refer
red to a diagram bounded by
straight lines such as Fig. 23.
The limits of this triangle are
necessarily arbitrary. Convenient
limits in the longitudinal direction

image front | i | e | ɛ | a | central | back | u | o | ɔ | ɑ

Fig. 25. The Central Vowel area.38

are two lines parallel to the lines i—a, ɑ—u and placed so that they
meet on the central line at a point intermediate between the lines
ɛ—ɔ and a—ɑ, as shown in Fig. 25.

Lip Positions

156. Vowel quality, though chiefly dependent on the position
of the tongue, is also affected to a considerable extent by the
positions of the lips. The lips may be spread, rounded or neutral.
Vowels with spread lips or neutral lips are generally grouped
together under the term unrounded.

157. Two degrees of lip-rounding may be distinguished, if
desired, viz. close lip-rounding and open lip-rounding.

158. Cardinal i is characteristic of a vowel formed with spread
lips (Fig. 15); cardinal ɑ (Fig. 19) may be considered as having
neutral lips; cardinal ɔ and u have open lip-rounding and close
lip-rounding respectively (Figs. 20, 22).

Tenseness and Laxness

159. Another element which is considered by some to be of
importance in determining vowel quality is the state of the tongue
and lips as regards muscular tension. Those who consider that
vowels may be differentiated by degrees of muscular tension
distinguish two classes, tense vowels and lax vowels. Tense vowels
are those which are supposed to require considerable muscular
tension on the part of the tongue; lax vowels are those in which
the tongue is supposed to be held loosely. The difference in
quality between the English vowels in seat siːt, and sit sit is ascribed
by some writers to a difference of tension (the vowel in seat being
considered tense and the vowel in sit lax).

160. It is not by any means certain that this mode of describing
the sounds really corresponds to the facts. A description of the
English short i as a vowel in which the tongue is lowered and retracted
from the ‘close’ position (see Fig. 34) is generally sufficiently
accurate for ordinary practical work. The term ‘lax’ may also be
used to describe the organic position of the English short u (in put
put) as compared with the long ‘tense’ (in boot buːt). Here the
39organic characteristics of short u as compared with long might
be more accurately described as a lowering and advancement of
the tongue and a wider opening of the lips.

161. Although the terms ‘tenseness’ and ‘laxness’ probably do
not describe accurately the action of the tongue in differentiating
certain vowels, nevertheless some teachers can actually get good
results by telling pupils to make their tongues tense or to keep
them lax. If a teacher can impart correct sounds by such instructions,
he should certainly do so.

162. It is generally advisable to apply the terms tense and lax
only to the case of close vowels. It is extremely difficult to
determine in the case of the opener vowels whether the sensation
of ‘tenseness’ is present or not, and there is in regard to some
vowels considerable difference of opinion on the subject. 1049

163. The ‘tenseness’ or ‘laxness’ of a vowel may be observed
mechanically in the case of some vowels by placing the finger on
the outside of the throat mid-way between the larynx and the chin.
When pronouncing for instance the English i (as in sit), this part of
the throat feels loose, but when pronouncing the corresponding tense
vowel (the in seat), the throat feels considerably tenser and is
somewhat pushed forward.

Nasalized Vowels

164. The position of the soft palate may affect vowel quality.
In the articulation of normal vowels the soft palate is raised so
that it touches the back of the pharynx as shown in Figs. 12,
13, 14. The result is that no air can pass through the nose. It
is, however, possible to lower the soft palate so that it takes up
the position shown in Fig. 1; the air can then pass out through
the nose as well as through the mouth. When vowels are pronounced
with the soft palate lowered in this way, they are said to
be nasalized. Nasalization may be expressed in phonetic writing by
the symbol ~ placed over the symbol of the sound which is nasalized.
An example of a nasalized vowel is the French ɑ̃, as in cent
sɑ̃.40

165. The movements of the soft palate may be observed by
means of a pencil about 15 cm. long inserted into the mouth.
If this is held between the finger and the upper teeth so that the
end inside the mouth rests lightly against the middle of the soft
palate, and groups of sounds such as ɑŋɑŋ…, ɛ̃ɛɛ̃ɛɛ̃… are
pronounced, the outer end of the pencil is seen to rise for the
sounds ŋ, ɛ̃ and to fall for the sounds ɑ, ɛ. Again, if we breathe
in through the nose and out through the mouth the end of the
pencil rises and falls in a similar manner.41

Chapter IX
Consonants

Cardinal consonants

166. The principle of cardinal sounds, which is so fundamental
for the proper study of vowels, is not applicable to consonants
as a whole. It is only applicable in cases where it is possible
to pass from one sound to another through a series of sounds
each of which is hardly distinguishable from the preceding, and
where it is therefore necessary to fix arbitrarily some points to
which any given sound in the series may be referred.

167. There are only a few limited families of consonants within
which these conditions prevail. As a rule it is not possible to
pass by almost imperceptible degrees from one consonant to
another. For instance, there is no continuous series of sounds
between t and m, nor is there between p and r.

168. The fact is that most consonants fall naturally into well-defined
classes, classes which are clearly separated from the neighbouring
classes by essential differences in the place or manner
of articulation.

169. The chief cases where the principle of cardinal sounds
can be applied to consonants are those in which the sounds under
consideration all belong to one class as regards manner of articulation,
and are all articulated by the tongue against the roof of the
mouth. 150 Descriptions of such sounds may in some cases be made
more intelligible to the reader by referring their tongue-positions
to certain cardinal tongue-positions.

170. It has been shown (§§ 140-142) that the nature of cardinal
sounds cannot be explained by means of key-words. It should
be noticed, however, that key-words are as a rule less open to
42objection in describing consonants than in describing vowels,
because in many cases consonants are not subject to notable
variations in the speech of different individuals. Thus the tongue-position
of the sound k in the English word back does not vary
to any marked extent with different English speakers. Consequently
a definition of cardinal k as ‘the final consonant of the
English word back’ will give to most English readers a fairly accurate
idea of what cardinal k is.

171. Some cardinal consonants can be deduced from sounds of
particular languages with fair accuracy; thus a French speaker
may deduce cardinal ç (Fig. 113) from i, and cardinal x (Fig. 110)
from the k in cas ka; a German can deduce cardinal c (Fig. 29)
from i or j. But to make quite certain of these cardinal sounds
oral instruction is required as in the case of cardinal vowels.

172. Fortunately most consonants either cannot be or do not
need to be referred to cardinal consonants. Consonants can as a
rule be learnt from plain descriptions of the actions which have
to be performed by the organs of speech.

Breathed and Voiced Consonants

173. Some consonants are breathed, others are voiced (see
Chap. V). To every breathed sound corresponds a voiced sound,
i.e. one articulated in the same place and manner, but with voice
substituted for breath, and vice versa; thus v corresponds to f,
z to s, b to p. It should be noted that voiced consonants are
usually pronounced with less force of exhalation than breathed
consonants.

174. Breathed consonants pronounced with weak force occur
in some languages. Such sounds are common in South German
and occur as occasional variants of b, d, z, etc., in English (see
§§ 567, 789). These sounds are sometimes called ‘mediae.’ They
may be represented thus , , , etc. People accustomed to strong
breathed consonants are apt to mistake them for voiced sounds.
People accustomed to hear very fully voiced consonants in initial
and final positions, e.g. the French, are liable to mistake ‘mediae’
for strong breathed consonants in these positions.

175. The distinction between breathed and voiced consonants
is a very important one. Some foreign people have difficulty
43in recognizing the difference between them, and in bringing out
the distinction clearly in their speech.

176. It is a good phonetic exercise to deduce unfamiliar breathed
consonants from familiar voiced ones, e.g. to deduce from m,
which is a voiced consonant, the corresponding breathed consonant
(phonetic symbol m), and to deduce from l the corresponding
breathed consonant . 251 This is done by practising sequences such
as vfvf… , zszs… , until the method of passing from voice to
breath is clearly felt, and then applying the same method to m,
l, etc., thus obtaining mm̥mm̥… , ll̥ll̥… , etc. (In practising
these exercises, the sounds should follow one another continuously
without break.)

177. The distinction between the breathed and voiced ‘plosives’
(p, t, k, and b, d, g) offers special difficulty to some foreign people
(particularly to Germans, Scandinavians and some Chinese). The
difficulty generally lies in the voiced sounds, for which ‘unaspirated’
breathed sounds (‘mediae’) 352 are commonly substituted. When the
attention of foreign learners is called to the nature of the fully
voiced sounds, they sometimes imitate them by prefixing a nasal
consonant, saying for instance mpa, nta, instead of ba, da. A true
voiced b may be acquired by practising the exercise pmpmpm
pronounced without opening the lips, followed by the exercise
bmbmbm… also pronounced without opening the lips, and taking
care that voice is distinctly heard during the pronunciation of
the b. The student should also practise repeating the ‘stop’ (§ 562)
of b, i.e. pronouncing bbbbwithout separating the lips. (Take
care that this exercise does not degenerate into mmmm…)
Voiced d, g may similarly be acquired by practising tntntn…,
dndndn…, dddd…, kŋkŋkŋkŋ…, gŋgŋgŋ…, gggg…, without
moving the tongue
. These exercises present considerable difficulty
to some learners, and they should be practised until thoroughly
44mastered. Besides being useful in teaching voiced sounds, they
are of value for obtaining control over the movement of the soft
palate.

Further Classification of Consonants

178. Descriptions of the manner of forming consonants should
take into account the following particulars: (i) the place (or places)
of articulation, (ii) the state of the air-passage at the place (or
places) of articulation, (iii) the position of the soft palate if not
already mentioned under (i) or (ii), (iv) the state of the larynx
if not already mentioned under (i) or (ii).

179. The chief classes which have to be distinguished under
heading (i) for the purposes of English may be termed bilabial,
labio-dental, dental, alveolar, post-alveolar, palato-alveolar, palatal,
velar and glottal. Of these the classes dental, palatal and velar
may be considered as cardinal classes of tongue consonants. 453
It has been suggested also that alveolar (with tongue-tip articulation)
should be considered as a cardinal class of tongue consonants.

180. The above terms are defined as follows:

Bilabial: articulated by the two lips. Examples p, m, w.

Labio-dental: articulated by the lower lip against the upper
teeth. Example f.

Dental: articulated by the tip of the tongue against the upper
teeth. Examples θ, ð, Spanish t.

Alveolar: articulated by the tip or blade of the tongue against
the teeth-ridge. Example English t.

Post-alveolar: articulated by the tip of the tongue against the
back part of the teeth-ridge. Example English r.

Palato-alveolar: articulated by the blade of the tongue against
the teeth-ridge with raising of the main body of the tongue
towards the palate. Examples ʃ, ʒ (Fig. 99).

Palatal: articulated by the ‘front’ of the tongue against the
hard palate. Example ç (one variety of German ich-sound).45

Velar: articulated by the ‘back’ of the tongue against the central
and forward part of the soft palate. Example the k in pack.

Glottal or laryngal: articulated in the glottis. Example the
‘glottal stop’ (§ 553).

181. Typical tongue-positions of the chief classes of tongue-consonants
are shown in Figs. 26-31. They illustrate the following
sounds: cardinal dental t (Fig. 26), cardinal alveolar t (Fig. 27),
a retracted variety of t (Fig. 28), the cardinal palatal sound represented
in international phonetic notation by the letter c (Fig. 29),
an advanced variety of k (Fig. 30), and cardinal k (Fig. 31).

image

Fig. 26.
Tongue-position of
Cardinal Dental t.

image

Fig. 27.
Tongue-position of
Cardinal Alveolar t.

image

Fig. 28.
A retracted variety
of t.

image

Fig. 29.
Tongue-position of
Cardinal c.

image

Fig. 30.
An advanced variety
of k.

image

Fig. 31.
Tongue-position of
Cardinal k.

182. The chief classes of consonants which can be distinguished
under heading (ii) (§ 178) are called plosive, affricate, nasal, lateral,
rolled, flapped, fricative, frictionless continuant and semi-vowel.46

183. These terms are defined as follows:

Plosive: formed by complete closure of the air-passage during
an appreciable time; the air is compressed (generally by
action of the lungs) and on release of the closure issues suddenly,
making an explosive sound or ‘plosion.’ Examples p, d.

Affricate: formed as plosive consonants, but with slower separation
of the articulating organs, so that the corresponding fricative
is audible as the separation takes place. Examples ʧ (as in
choose ʧuːz), German ts (as in zehn tseːn).

Nasal: formed by a complete closure in the mouth, the soft
palate being, however, lowered so that the air is free to pass
out through the nose. Examples m, n.

Lateral: formed by placing an obstacle in the centre of the air-channel,
but leaving a free passage for the air on one or both
sides of the obstacle. Example l.

Rolled: formed by a rapid succession of taps of some elastic
organ. Example rolled r (§ 752).

Flapped: formed by a single tap of some elastic organ; the
position of contact is not maintained for any appreciable time.
Example flapped r (§ 753).

Fricative: formed by narrowing the air-passage to such an
extent that the air in escaping produces audible friction (i.e.
some kind of hissing sound). Examples f, z.

Frictionless Continuant: made with the organic position of a
fricative consonant, but pronounced with weak breath-force
so that no friction is heard. 554 Ex-ample the principal English r.

Semi-Vowel: a voiced gliding sound in which the speech organs
start by producing a weakly articulated vowel of comparatively
small inherent sonority and immediately change to another
sound of equal or greater prominence (§§ 101, 102). Examples:
English j (as in yard), w.47

184. Two classes of consonants are distinguishable under heading
(iii) (§ 178), according as the soft palate is raised, as shown in
Figs. 26, 27, etc., or lowered, as shown in Figs. 74, 75. Different
names are given to these classes according as the mouth-passage
is completely closed or not.

185. The combination of complete closure in the mouth with
raised soft palate gives rise to plosive consonants, as already
mentioned (§ 183). The combination of complete closure in the
mouth with lowered soft palate gives rise to nasal consonants,
as already mentioned (§ 183).

186. When there is not complete closure in the mouth, sounds
are distinguished as oral (or buccal) and nasalized according as
the soft palate is raised or lowered. (See § 164 and Chap. XXIV.)

187. There are several classes of consonants distinguishable
under heading (iv) (§ 178). The four principal classes have been
described in §§ 82, 83. Only two of them, the breathed and voiced
classes, occur in normal English; the ‘glottal stop’ occurs, however,
as an occasional sound.

188. Consonants which can be held on continuously without
change of quality are sometimes classed together as continuatives
or continuants; they include nasal, lateral, rolled and fricative
consonants and the frictionless sounds described in § 183. Nasal,
lateral and rolled consonants are sometimes classed together under
the not very satisfactory name liquids. (Some writers do not
include nasal consonants among ‘liquids.’)48

Chapter X
Phonemes, Principles of Transcription

Phonemes

189. In describing the sound-system of any language it is
necessary to distinguish between speech-sounds and what are called
phonemes.

190. A speech-sound is a sound of definite organic formation and
definite acoustic quality which is incapable of variation.

191. A phoneme may be described roughly as a family of sounds
consisting of an important sound of the language (generally the
most frequently used member of that family) together with other
related sounds which ‘take its place’ in particular sound-sequences
or under particular conditions of length or stress or intonation.

192. For detailed information regarding the theory of phonemes
readers are referred to my book, The Phoneme, its Nature and Use. 155
A few examples must suffice here to give an idea of what is meant
by the term ‘phoneme.’

193. The k's in the English words keep, cool, call, are three
distinct sounds articulated at different parts of the palate; but
they are regarded as belonging to the same phoneme, since the
use of these different varieties of k is dependent solely upon the
nature of the adjoining vowel. The j's in the French yeux
and pied pje are distinct sounds, the j in yeux being fully voiced
and the j in pied being breathed (or with some speakers partially
voiced); the sounds must, however, be regarded as members of a
single phoneme in French, since the breathed (or partially voiced)
j occurs only after p or t or k, while the fully voiced j never
occurs in this position. The l-sound in the French loup and boucle
(in final position) are different; the words may be transcribed
allophonically (§ 200) as lu, bukl̥, representing a voiceless l. They
are members of the same phoneme, because never occurs initially
in French, while voiced l does not occur finally when k precedes. It
does not occur to ordinary French people that the sounds are different.49

194. The German x's in Buch and ach are different sounds,
but as their use is dependent upon the nature of the adjoining
vowel, they must be considered as members of the same phoneme. 256
The initial consonants in the Japanese words hito (man), hata
(flag), huzi (fuji) (wistaria) are very different to the ears of a
European, the first resembling a German ich-sound (ç), the second
being an ordinary h, and the third being a ‘bi-labial f’ (ϕ); but
in the Japanese language the three sounds are merely members
of a single phoneme, their use being determined by the vowel
following. In the Japanese ‘Kunreisiki’ Romazi spelling introduced
in 1937 they are represented by a single letter (h).

195. On the other hand, sounds of the n and ŋ types belong to
separate phonemes in English, because the use of the two sounds
is not dependent upon neighbouring sounds in words, n can occur
in positions which ŋ can also occupy, e.g. in the terminations -in,
-iŋ. The h-sounds and ç (the ich-sound) belong to separate
phonemes in German, since ç and the appropriate varieties of h
may occur in identical phonetic contexts.

196. But though n occurs occasionally in French, it does not
constitute a separate phoneme in that language. It is only found
as a member of the g-phoneme, as when langue maternelle is pronounced
lɑ̃ŋ matɛrnɛl. Most French people are unaware of the
existence of ŋ in their language, and they have difficulty in making
the sound properly in English words (see §§ 655, 656). In Italian
and Spanish ŋ also exists, but only before velar consonants (e.g.
Italian banca ˈbaŋka, lungo ˈluŋgo, Spanish cinco ˈθiŋko, venga
ˈβeŋga). As n never occurs in such positions in Italian or Spanish,
ŋ is to be regarded as a member of the n-phoneme in these languages.

197. The most frequent sound of a phoneme may be called its
principal member or norm. It is usually the sound which would be
given if a person with unstudied pronunciation were asked ‘to say
the sound by itself.’ The other sounds belonging to the phoneme
are called subsidiary members. The term allophone is used to denote
a particular member (principal or subsidiary) of a phoneme.50

198. Phonemes are capable of distinguishing one word of a
language from other words of the same language. There is an
English word sin and another English word siŋ. There is a German
word ˈçɔrdə 357 (Chorde) and another German word ˈhɔrdə (Horde).
The existence of such words is a proof that the n and ŋ sounds
belong to separate phonemes in English, and that the ç and h
sounds belong to separate phonemes in German.

199. The distinctive elements of language, i.e. the elements
which serve to distinguish one word from another are the phonemes
(not the sounds). The distinction between two phonemes is
significant, i.e. capable of distinguishing one word from another;
the distinction between two sounds is not necessarily significant.
Different sounds which belong to one phoneme do not distinguish
one word of a language from another; failure on the part of the
foreigner to distinguish such sounds may cause him to speak with
a foreign accent, but it will probably not make his words
unintelligible.

Principles of Transcription

200. As a general rule it is only necessary in practical phonetic
writing to have symbols for the phonemes. The use of allophones
(special members of the phonemes) is, in most languages, determined
by simple principles which can be stated once for all, and
which can be taken for granted in reading phonetic texts. A transcription
based on the principle ‘one symbol per phoneme’ is called a
phonemic’ or ‘linguistically broad’ transcription. A transcription which
provides special signs for allophones (special members of phonemes)
is called an ‘allophonic’ or ‘linguistically narrow’ transcription.

200a. There are also ‘comparative’ or ‘typographically narrow
forms of transcription, in which special symbols are introduced in
order to show that certain sounds of the language transcribed differ
from sounds of another language. For a full discussion of the
subject of types of transcription, see Appendix A. 45851

The Phonetic Representation of Vowels

201. In transcribing particular languages the following mode
of representing vowels is recommended 559:

(1) When the principal member of a vowel-phoneme is identical
with a Cardinal Vowel, the appropriate cardinal vowel symbol
should as a rule be used to represent it.

(2) In cases where the principal members of vowel-phonemes
are not cardinal vowels, the cardinal vowel letters i e ɛ a ɑ
ɔ o u y ø œ
, etc., should, as far as possible, be used to
represent vowels lying within certain areas in the vowel
figure, as shown in Figs. 32 and 33.

image front | i | e | ɛ | a | central | back | u | o | ɔ | ɑ

Fig. 32. Areas served by
the eight primary Cardinal
Vowel letters.

image front | y | ø | œ | central | back | ɯ | ɤ | ʌ

Fig. 33. Areas served by
the secondary Cardinal Vowel
letters y, ø, œ, ʌ, ɤ, ɯ.

(3) ə is the letter normally used to represent an unrounded
vowel lying within the inner central triangle.

(4) When a vowel lies near the boundary of a vowel area, and
is in consequence acoustically remote from any cardinal
vowel, it is sometimes necessary to represent it by a special
(non-cardinal) symbol. Thus in some types of transcription
of English the symbols æ, ɩ, ɷ are introduced to represent
the vowels in hat, bit, put. (In broad transcription 660 these
sounds may be written with a, i, u.)

The phonetic representation of consonants

202. Principles similar to the above must be adopted in transcribing
those consonants which have to be described in relation
to cardinal consonants. Other consonants are represented by
arbitrarily chosen symbols.52

Chapter XI
Diaphones

203. It often happens that a certain sound used by one speaker
of a language is consistently replaced by another sound by another
speaker of the same language. Thus different speakers of Southern
English pronounce the word get with vowels of different degrees of
openness (§ 271). The diphthongal sound ou in such a word as
home is not pronounced by all English people exactly in the manner
described in § 394; with some the initial element is more retracted,
with others it is opener, with others it is more advanced; the degree
of lip-rounding is not the same with all speakers, and with some
(most Scotsmen and many in the North of England) the sound in
home is not diphthongal but is a pure long .

204. The term diaphone is suggested to denote a sound used by
one group of speakers together with other sounds which replace
it consistently in the pronunciation of other speakers. Thus the
various kinds of on mentioned in § 203 and the Scottish and
Northern English may be said to be members of the same
diaphone.

205. It has been mentioned in § 63 that everyone has different
styles of pronunciation. Such different styles are merely different
ways of pronouncing the language. When a person consistently
uses one sound in one style of speech but substitutes another for
it in another style, it is as if two different people were speaking,
and the two sounds must be regarded as two members of the
same diaphone.

206. Examples are when an actor pronounces such a word as
possible with a ‘dark’ l 161 in ordinary conversation, but uses a ‘clear’
l 162 when reciting on the stage, 263 or when a ‘forward’ a is substituted
for æ in singing, or where a variety of a is substituted for ai in
very rapid pronunciation of such an expression as I'm going
(normally aim ˈgouiŋ, very rapidly am ˈgoiŋ).53

207. Care must be taken to distinguish diaphones from phonemes.
The different members of one phoneme are sounds used by one
single person
speaking in one particular style; their use is conditioned
by the nature of the surrounding sounds in the sequence and
on the degree of stress, sometimes also on length and intonation.
The different members of one diaphone are found in comparing
the speech of one person with that of another
, or in comparing two
styles of speech of the same person.

207a. The theory of diaphones is discussed at greater length in
my book The Phoneme, Chap. XXVII.54

Chapter XII
Prominence, Syllables, Diphthongs

208. In every spoken word or phrase there is at least one sound
which is heard to stand out more distinctly than sounds next
to it. Thus in the English word letter ˈletə the sounds e and ə
are heard more distinctly than the l or the t. If the speaker
is at a certain distance, or if the word is spoken on the telephone
or on a gramophone, the e and ə may be heard clearly, while l
and t are often indistinct. The e and ə are in fact the sounds
of the word letter which are ‘prominent’ in the sense explained
in §§ 100, 101.

209. The prominence of sounds may be due to inherent sonority
(carrying power, § 101), to length or to stress or to special intonation,
or to combinations of these.

210. Thus in every sentence there is a kind of undulation of
prominence which is easily perceived by the hearer. This undulation
may be visualized as a wavy line with ‘peaks’ (denoting
maxima of prominence) and ‘troughs’ (denoting minima of prominence).
It is generally quite easy to count the number of peaks
of prominence in a word or phrase.

211. Each sound which constitutes a peak of prominence is
said to be syllabic, and the word or phrase is said to contain as
many syllables as there are peaks of prominence. In the word
ˈletə the e and ə constitute peaks of prominence (by reason of
their inherent sonority) and are therefore called ‘syllabic,’ and the
word is said to contain two syllables. In button-hook ˈbʌtnhuk
there are three peaks of prominence and therefore three syllables,
the syllabic sounds being ʌ, n and u.

212. In theory a syllable consists of a sequence of sounds containing
one peak of prominence. In practice it is often impossible
to define the limits of a syllable because there is no means of
fixing any exact points of minimum prominence. In many cases
the bottoms of the ‘troughs’ must be considered as flat, that is to
say there is no one point which can be regarded as the point of
55syllable separation. 164 Fortunately the exact determination of points
of syllable separation is a matter of no practical importance to the
language learner. When it is desirable to divide words into syllables
for the purpose of practising pronunciation or for finding a convenient
place to put a stress-mark in phonetic transcription, the
separation often has to be made in some conventional way. For
instance, it is customary to divide the spoken word extremity thus
iks-ˈtre-mi-ti, though the m (or part of it) may well be considered
to belong to the same syllable as the e. Many teachers would
divide astray thus ə-ˈstrei on account of the derivation of the word,
although from the point of view of prominence əs-ˈtrei is a better
division. The actual minimum of prominence continues through
the whole of the ‘stop’ of the t. 265

213. The syllabic sound of a syllable is generally a vowel, but
consonants may also be syllabic. The more sonorous consonants
such as n, l often are so, as in the English words people ˈpiːpl, little
ˈlitl, button ˈbʌtn. 366

214. When it is necessary to show in phonetic transcription
that a consonant is syllabic, the symbol, is placed under the letter.

215. When a consonant is immediately followed by a vowel,
it is usually not syllabic, since the vowel has the greater inherent
sonority. However, a consonant in this position is sometimes
given extra prominence by increasing its length, and it may thus
become syllabic. Examples are found in such English words as
gluttony ˈglʌtn̩i, 467 muttony 568 ˈmʌtn̩i, lightening ˈlaitn̩iŋ, bubbling ˈbʌbl̩iŋ,
flannelly 669 ˈflænl̩i. The n and l in these words are quite distinct
from those in Putney ˈpʌtni, lightning ˈlaitniŋ, publish ˈpʌbliʃ, manly
ˈmænli. In the latter cases the n and l are very short; in the
56former they are lengthened so that their prominence is sufficient
to make them syllabic in spite of the greater sonority of the adjacent
vowel. (Further information concerning the use of syllabic consonants
in English is given in my article The Use of Syllabic and
Non-syllabic
1 and n in Derivatives of English Words ending in
Syllabic l and n
in Zeitschrift für Phonetik und allgemeine
Sprachwissenschaft
Berlin, Vol. 12, No. 1-4, 1959 (Calzia-Festgabe).

216. In the comparatively rare cases when two consecutive
vowels form two syllables, the necessary diminution of prominence
between them is generally supplied by the glide (§ 3) which connects
them. This glide is a transitory sound of very short duration
and consequently of little prominence. Thus the word create
kriˈeit consists of two syllables because the first i is clearly pronounced,
and in passing from it to the e a very short j-glide is
present. (The -eit counts as one syllable because the ei is a
diphthong, see §§ 219-221.) Other examples of two consecutive
vowels forming two syllables are react riːˈækt, he ought hiː ˈɔːt,
screwing ˈskruːiŋ, freer ˈfriːə. (When the first syllable has the
stronger stress, there is a tendency in some sequences to reduce
the two vowels to a single diphthong; thus screwing, freer are
sometimes pronounced skruĭŋ, friə̆. See further my article Falling
and Rising Diphthongs in Southern English
in Miscellanea Phonetica
II, 1954, published by the I.P.A.)

217. In cases like chaos ˈkeiɔs, 770 co-operate kouˈɔpəreit, high up
ˈhai ˈʌp, coward ˈkauəd, buoyant ˈbɔiənt, the syllable separation is
marked by the ends of the diphthongs ei, ou, ai, au, ɔi. 871

218. When a vowel is immediately followed by the same vowel,
as in bee-eater ˈbiːˌiːtə, we saw all of it ˈwiː ˈsɔː ˈɔːl əv it, the syllables
are generally separated by a slight diminution of loudness of the
vowel due to a diminution in the force of exhalation. 97257

219. A diphthong is defined as an independent vowel-glide not
containing within itself either a ‘peak’ or a ‘trough’ of prominence.
By a vowel-glide we mean that the speech-organs start in the
position of one vowel and move in the direction of another vowel.
By ‘independent’ we mean that the glide is expressly made, and is
not merely an unavoidable concomitant of sounds preceding and
following.

220. During a diphthong the prominence may fall continuously
or it may rise continuously, 1073 but by definition it may not contain
a fall of prominence followed by a rise nor a rise of prominence
followed by a fall.

221. A diphthong must necessarily consist of one syllable. In
order that a vowel-glide should constitute two syllables, it would
be necessary that it should contain a ‘trough of prominence,’ i.e.
a fall of prominence followed by a rise.

222. Diphthongs may be long or short, according as the glide
is performed slowly or quickly. They may also be ‘wide’ or
‘narrow’ according as there is a large or a small movement from
the initial position.

223. One end of a diphthong is generally more prominent than
the other. The greater prominence may be due either to greater
inherent sonority (§ 100) or to stronger stress or to a combination
of the two. When the beginning of a diphthong is more prominent
than the end, the diphthong is said to be falling. When the
beginning is less prominent than the end, the diphthong is said
to be rising. Most of the English diphthongs (Chap. XV) are
falling diphthongs, but there are two important rising diphthongs,
ĭə and ŭə (§ 378) and four unimportant ones, ŏi, ŭi, ĕə and ŏə
(§§ 466v, 466x).

224. When a diphthong is ‘falling’ as the result of a gradual
diminution of inherent sonority, the correct effect will generally
be given if the speech-organs perform the greater part of the
movement towards the second vowel; it is not necessary that
the limit of the movement should be actually reached. Thus
the English diphthong ai is one which begins at a and moves in the
58direction of i. To give the right effect it is not necessary that i
should be quite reached; the diphthong may and generally does
end at an opener vowel than this, such as a fairly open variety
of e. i merely represents the furthest limit of movement; if the
movement were to extend beyond this point, the diphthong would
not sound correct.

225. When the vowels beginning and ending a diphthong are
of approximately equal inherent sonority, one end of the diphthong
is generally rendered less prominent than the other by reducing
the force of exhalation. Thus the sounds ɛ and ʌ when isolated
and pronounced with equal stress (push from the chest wall) have
approximately equal sonority, but in the English diphthong ɛʌ (a
variant of ɛə, § 449) the beginning of the glide is pronounced with
greater stress than the end and therefore has greater prominence.
The diphthong is in consequence a falling diphthong. Again, the
English sound i is normally less sonorous than ə; nevertheless in the
English (§ 440) the first part of the diphthong is pronounced with
so much more stress than the latter part that its prominence is
greater, and the diphthong is a falling one (iə̆). 1174

226. Rising diphthongs are sometimes difficult to distinguish
from sequences consisting of a semi-vowel followed by a vowel.
Some authorities consider the English juː, as in music ˈmjuːzik,
and ju, as in monument ˈmɔnjumənt, to be rising diphthongs (ĭuː
and ĭu rather than juː, ju).

227. Another kind of diphthong, called an imperfect diphthong,
is produced (1) when the initial vowel of a falling diphthong is
appreciably lengthened before the glide begins, or (2) when the
final vowel of a diphthong (falling or rising) is lengthened after the
glide ends. Thus if the initial element of the English diphthong ai
is prolonged, as is done in singing, the result is the first type of
imperfect diphthong. The second type of imperfect diphthong
is heard if the final element of the English diphthong ou is prolonged,
as is sometimes done when saying the interjection Oh.

228. An imperfect diphthong forms only a single syllable.59

229. It is convenient to represent diphthongs in phonetic transcription
by digraphs (sequences of two letters), the first letter
representing the commencement of the glide and the second representing
its termination or its direction of movement. In the case
of many falling diphthongs the point of termination is somewhat
variable (§ 224); when this is so, the second letter is selected so
as to show the termination most remote from the initial element
of the diphthong. Thus the transcription ai is used to represent
the English diphthong in fly flai; the glide begins at a and proceeds
in the direction of i, but it seldom reaches i (see § 224).

230. When it is desired to show in phonetic transcription which
part of a diphthong is the least prominent, the mark ˅ is placed
over the letter indicating the less prominent part. Thus the
English falling diphthongs ou, ai, au, ɛə, etc., may, if desired, be
written , , , ɛə̆, etc., and the rising diphthongs are denoted
in this book by ĭə, ŭə, ŭi etc. (see §§ 466a-466x).

231. The term consonantal vowel is sometimes used to denote
the less prominent part of a diphthong. Thus it is sometimes
said that the English diphthong ai consists of ‘an a followed by a
consonantal i.’ This manner of regarding a diphthong, though
not quite accurate, is sometimes convenient in practical teaching.

232. When a vowel glide contains a peak of prominence (i.e. a
rise followed by a fall), it is called a triphthong. ŏaĕ, a careless way
of pronouncing why (normally wai) is a triphthong.

233. The English sequences commonly written aiə, auə in phonetic
texts are not triphthongs; they are disyllables, since the i and u
are less prominent than the a and ə. These i and u are often
lowered towards ɛ and ɔ (§§ 414, 430); the groups then approach
nearer to triphthongs, but even then they are not actually true
triphthongs. In their extreme forms these sequences are reduced
to diphthongs of the type or to a single long vowel of the type
(§§ 414, 430). It is, however, sometimes convenient to call aiə
and auə ‘triphthongs’ for want of a better name and in view of
the fact that they are often treated in poetry as forming single
syllables.60

Chapter XIII
Ear-Training for the English Vowels
and Diphthongs

234. Those learning to speak a foreign language should begin
their study by ear-training, which will enable them to recognize
the sounds of the language. We therefore give in this chapter
some exercises for learning to recognize by ear the English vowels and
diphthongs, before proceeding to explain how the sounds are formed.

235. The general principles of ear-training have been indicated
in §§ 18-24. It must be added here that, at any rate as far as vowels
and diphthongs are concerned, it is convenient to assign numbers
to the sounds. The teacher should begin by dictating isolated
sounds, or single syllables containing easy consonants combined
with the various vowels and diphthongs. In the earlier lessons
the teacher should ask the pupil to name the numbers of the
vowels and diphthongs dictated. Later he should ask the pupil
to write with phonetic symbols the sounds or words he dictates.

236. The following system of numbering the English vowels
and falling diphthongs is a convenient one. The first line contains
the ‘pure’ vowels, the second line contains the diphthongs ending in
i and u, and the third line contains the diphthongs ending in 9.
The pupil should have a copy of this table always ready at hand
for reference.

iː i e æ ɑː ɔ ɔː u uː ʌ əː ə
1 2 3 4 5 6 7 8 9 10 11 12

ei ou ai au ɔi
13 14 15 16 17

iə ɛə ɔə uə
18 19 20 21

The rising diphthongs ĭə and ŭə (Nos. 22 and 23), important
though they are, need not be introduced until a later stage. Nor
need the unimportant rising diphthong ŭi, or the non-essential
diphthongs oi, ui, , ŏi, ŏə (§§ 466v-466x).61

237. The following are some examples to illustrate the most
elementary type of exercise to be dictated. The teacher should
vary the lengths of the isolated sounds; pronouncing some of
them fully long and some quite short and others of medium length.
This is in order to familiarize the student with the differences
in tamber apart from the differences of length. (It is, however,
unnecessary to dictate ə (No. 12) in isolation, this sound being
always very short in English; if lengthened, it is difficult to
distinguish from əː (No. 11).) In the syllables it is better to say
the vowels with the lengths they would usually have in English
words. Their lengths depend (1) on their nature (, ɑː, ɔː, , əː
being longer than the other pure vowels in similar positions), (2)
to a large extent on the nature of the following consonant (§§ 863 ff.).

(1) Isolated vowels and diphthongs and single syllables:
, æ, ɔ, i, u, e, ɑː, i, ɔː, əː, ʌ, e, , æː, , əː, , ʌ, ɑː, ɔː, u,
ɔ; ei, ɔə, ɔi, au, ou, ai, ɛə, ou, , ei, ɜə, , ɔə, ou, ɔː, ɑː,
ɔ, ɔə, ʌ, ai, , e, ɔ, ɛə, ou, , ɔː, i, əː, ʌ, , e, i, æ, ei, ɔə,
u, ɔ, ɛə, ɔː, əː, ou, e, ɔ, æ; pəːt, pʌt, pet, pæt, pɔt, pæt, puːt,
put, pɔːt, pɛət, pait, piːt, pit, pɔit, den, dɔːn, dɔn, doun, dɔən,
dən, dəːn, dʌn, duən, diən, dɑːn, diːn, dɛən, doun, etc.

(2) Sequences of more than one syllable:
tinɑːlɔd, sumiːdef, bækɑːzug, pesiːvɔː, gʌmuːbik, fægɔʒis,
brigetæ, nɑːkɔːndu, lesæʃkʌl, θɔːŋeznəːf, kʌrmædiguː, trəːsimɑːfgæk,
ðuːgʌʒtelezæ, plerkjɑːfkɔː, muːləːvɛəs.

238. When the teacher finds that the pupil has difficulty in
distinguishing certain vowels, he should repeat them a number
of times in various orders and with various surrounding consonants.
Thus if the pupil has difficulty in distinguishing æ from e and ʌ,
the teacher should dictate e, e, æ, æ, ʌ, ʌ, æ, ʌ, æ, e, e, ʌ, ʌ, æ,
e, ʌ, e, ʌ, ʌ, æ, etc., tæm, tem, tæm, tem, tʌm, tem, net, nʌt, næt,
nʌt, net, lʌp, læp, etc., etc.

239. A complete course of ear-training includes much more
difficult combinations of sounds than those given above; it includes
also sounds other than those of the language studied. Teachers
should, when possible, pay some attention to the sounds of the
pupils' mother tongue, and give some ear-training exercises containing
those sounds which resemble but are not identical with
sounds of the language studied. Some graduated exercises are
given in Appendix B.62

Chapter XIV
The English Vowels

240. The term ‘pure’ vowel is used in this book to designate
a vowel (during which the organs of speech remain approximately
stationary) in contradistinction to a ‘diphthong’ (during which
the organs of speech perform a clearly perceptible movement).

241. There exist many shades of pure vowel-sound in Southern
English. Of these twelve are of special importance for the foreign
learner of English. They are represented in this book by the
notation , i, e, æ, ɑː, ɔ, ɔː, u, , ʌ, əː, ə. It is convenient to
number them 1 to 12 as shown in the first line of the table on
p. 61.

242. Four pairs of these vowels may be considered as belonging
to single phonemes in one type of Southern English, viz. long
and short i, long ɔː and short ɔ, long and short u, and long əː
and short ə. The tamber of the English short i differs considerably
from that of the English long , but in this kind of English the
difference in tamber always coincides with a difference of length;
that is to say is always longer than i when surrounded by the
same sounds and pronounced with the same degree of stress.
Similarly with the pairs ɔː, ɔ, and , u. There is not much
difference in tamber between the long əː and the most frequently
used short ə (ə1, § 356). There are thus eight pure vowel phonemes
in Southern English (represented in this book by the letters i, e,
æ, ɑ, ɔ, u, ʌ, ə). For further particulars concerning the phonemic
classification of the Southern English vowels, see my book on
The Phoneme (Heffer, Cambridge), especially §§ 510 ff.

243. Of the above-mentioned important vowel-sounds eight
(, i, e, æ, ɑː, ʌ, əː, ə) have spread or neutral lips, while four
(ɔ, ɔː, u, ) have various degrees of lip-rounding.

244. The approximate tongue-positions of these vowel-sounds
are shown in Fig. 34 (p. 64). In this diagram the vowels are placed
in relation to the Cardinal Vowels (for which see Chap. VIII). The
tongue-positions of the Cardinal Vowels are represented by the
63small dots in Fig. 34, while the tongue-positions of the English
vowels are shown by the large dots. The nature of the tongue-positions
will be realized by comparing these diagrams with
Figs. 13 and 14.

245. Fig. 35 is a simplified chart of the chief English vowels.
It is less accurate than Fig. 34, but is a convenient form for use
in practical teaching.

image front | central | back | close | half-close | half-open | open

Fig. 34. Diagram showing the relations of
the chief English Vowels to the Cardinal
Vowels. (Small dots represent Cardinal Vowels.
Large dots represent English Vowels.)

image front | central | back

Fig. 35. Simplified Chart
of English Vowels for use
in practical teaching of the
language.

The English Vowels in Detail

English Vowel No. 1:

246. is the member of the English i-phoneme used when the
vowel is relatively long.

247. Its tongue-position is shown by the position of the dot in

image

Fig. 36. Lip-position of
English long

Fig. 34. The following is a formal
description of the manner of forming
the vowel:

(i) height of tongue: nearly ‘close’
(§ 152);

(ii) Part of tongue which is highest:
centre of ‘front’;64

(iii) position of lips: spread or neutral (see Fig. 36);

(iv) opening between the jaws: narrow to medium. 175

The sound is considered by many to be pronounced with considerable
muscular tension of the tongue (see §§ 159-163). In normal speech
the tip of the tongue touches the lower teeth, but small variations
in the position of the tongue-tip do not materially affect the
acoustic effect of the sound.

248. A palatogram of the vowel as pronounced
by me is shown in Fig. 37.

image

Fig. 37. Palatogram
of English long .

249. i: is the so-called ‘long’ sound of the
letter e; examples: tree triː, see siː, even ˈiːvn,
complete kəmˈpliːt, immediate iˈmiːdjət. is
also the sound of ea, ie, ei and i in many
words, examples: sea siː, east iːst, field fiːld,
seize siːz, machine məˈʃiːn. Note the exceptionally
spelt words key kiː, quay kiː, people ˈpiːpl.

250. The English is similar in tamber (quality) to the French
sound of i, as in ici isi, and to the German ‘long’ as in mir miːr,
sie ziː. It is, however, less close than these sounds. The average
continental close i does not, however, sound wrong (in quality) 276
when used in English words such as sea, even. But those foreign
people (mainly French and German) who use a particularly close
i should endeavour to hold their tongue a little more loosely in
pronouncing the English .

251. Many English people use a diphthong in place of a
pure . The diphthong begins with an open variety of i and
moves to a closer position; it may be represented by įị 377 or ɩi 478 or
65ij. 579 It is not necessary for foreign learners to attempt this diphthongal
pronunciation. An exaggerated diphthongal pronunciation
sounds dialectal, an extreme form of the diphthong being used in
the local dialect of London (Cockney), where see is pronounced səi.

252. Words for practice: peak piːk, beak biːk, team, tiːm, dean
diːn, keen kiːn, geese giːs, chief ʧiːf, Jean ʤiːn, meat, meet miːt,
need niːd, leaf liːf, wreath riːθ, feel fiːl, veal viːl, these ðiːz, siege
siːʤ, zeal ziːl, .shield ʃiːld, heap hiːp, yield jiːld, queen kwiːn.

English Vowel No. 2: i

253. The letter i without the length-mark stands for the members
of the English i-phoneme used when the sound is relatively short
(§ 863). The distribution of these members in words is determined
by the nature of the surrounding sounds in the sequence and on
the degree of stress (see §§ 260-263). For the purposes of practical
teaching it is sufficiently accurate to use the commonest of them
in all cases.

254. In pronouncing this common sound, the general position of
the tongue and lips resembles that of the long (§ 247), but the
tongue is lower and retracted. Its nature is shown by the position
of the dot in Fig. 34. Some writers express the difference by
saying that for the short i the speech organs are ‘lax’ or held
loosely, while for the long they are more ‘tense.’

255. The following is a formal description of the manner of
forming this English short i:

(i) height of tongue: nearly ‘half-close’ (Fig. 34);

(ii) part of tongue which is highest: the hinder part of the ‘front’
(Fig. 34);

(iii) position of lips: spread or neutral (Fig. 38);

(iv) opening between the jaws: narrow to medium.66

In normal speech the tip of the tongue touches the lower teeth,
but small variations in its position do not materially affect the
tamber. As with all normal vowels, the soft palate is in its raised
position and the vocal cords are in vibration.

image

Fig. 38. Lip-position of
English short i.

image

Fig. 39. Palatogram
of English short i.

256. A palatogram of the vowel as pronounced by me is shown
in Fig. 39. It will be observed that the air-passage is considerably
wider than in the case of the long (Fig. 37).

257. i is the ‘short’ sound of the vowel letters i and y; examples:
fit fit, rich riʧ, king kiŋ, symbol ˈsimbl. It is also the sound of e
and a in various prefixes and suffixes when unstressed; examples:
become biˈkʌm, descend diˈsend, remain riˈmein, engage inˈgeiʤ,
except ikˈsept, examine igˈzæmin, 680 horses ˈhɔːsiz, useless ˈjuːslis,
goodness ˈgudnis, village ˈviliʤ, private ˈpraivit 781; it is also the
sound of unstressed -ies, -ied, as in varieties vəˈraiətiz, carried
ˈkærid 882. Note also the miscellaneous words minute ˈminit, threepence
ˈθripəns or ˈθrepəns, women ˈwimin, Sunday ˈsʌndi, Monday
67ˈmʌndi, etc., 983 pretty ˈpriti, England ˈiŋglənd, English ˈiŋgliʃ, busy
ˈbizi, business ˈbiznis, lettuce ˈletis. 9a84

258. Many foreign people, and especially speakers of Romance
languages, use a sound which is too ‘tense’; in fact they do not
make the necessary difference of tamber between the English short
i and long . They pronounce rich too much like reach rːlʧ and
sit too much like seat silt, etc. The correct vowel may be acquired
by trying to pronounce the sound in a slack sort of way, or by
making it more like e. French learners should notice that the
English short i resembles the French sound of é.

259. The English short i is slightly opener than the corresponding
German vowel as in bitte ˈbitə, Sinn zin, etc., but it is
less open than the Dutch sound of i, as in ik ɩk (I), dit dɩt (this).

260. A notable subsidiary short i is a ‘lower’ variety, i.e. a
vowel having a tongue-position lower than that of the i just
described. It resembles in quality a not very close e. It may be
written with the letter e when it is desired in transcription to
distinguish the two members of the phoneme; this transcription,
however, involves writing vowel No. 3 with ɛ in place of e.

261. This subsidiary i is used in final positions, for instance
in such words as heavy ˈhevi, city ˈsiti (second i), many ˈmeni,
when a pause follows. 1085 If another word follows in the same sense-group,
the ordinary short i is used; thus the ordinary short i is
used in both syllables of city in the expression the City of London
ðə ˈsiti əv ˈlʌndən. 1186

262. A minute analysis of the pronunciation of words containing
short i reveals the existence of a number of shades of
68i-sound ranging mainly between the common i and the subsidiary
i described in § 260. The use of these intermediate shades of
sound varies with different speakers, and the shade used depends
on the nature of the surrounding sounds in the sequence, and on
the degree of stress. For instance, it is not uncommon to hear the
termination -ity (as in ability əˈbiliti or əˈbilete) pronounced with
the penultimate i lower than the ultimate; some use ə in this
position, pronouncing əˈbiləti. It is instructive to observe the
different pronunciations of visibility used by radio news readers
announcing the weather reports.

263. There is a tendency with some English speakers to use
lowered varieties of i in unstressed positions generally, as well as
when final. Their pronunciation might be represented thus (using
e to represent the lowered i, and ɛ to represent vowel No. 3): waited
ˈweited, ladies ˈleidez, goodness ˈgudnes, become beˈkʌm, except
ekˈsɛpt, village ˈvileʤ, limit ˈlimet, Cambridge ˈkeimbreʤ, profit
ˈprɔfet (= prophet), indeed enˈdiːd, bringing ˈbriŋeŋ, solid ˈsɔled.
These speakers also have distinct weak forms for words like it,
this, in, if, when unstressed, thus: I'll get it in the morning ai l ˈgɛt
et en ðə ˈmɔːneŋ
, I wonder if it is ai ˈwʌndər ef e ˈtiz.

264. Although it is desirable that the foreign learner should
be aware of the existence of a number of shades of i, yet it is
in my view not necessary that he should make any special effort
to use them in his speech. If he ignores the differences altogether,
it does not matter; with many English speakers the differences
are so small as to be negligible.

265. Words for practice: pin pin, bill bil, tip tip, dish diʃ,
kitten ˈkitn, give giv, chin ʧin, Jim ʤim, milk milk, knit nit, lip
lip, risk risk, fit fit, village ˈviliʤ, thin θin, this ðis, sing siŋ, zip zip,
ship ʃip, hill hil, winter ˈwintə.

266. A sound of approximately the quality of short i also occurs
in English at the beginning of the diphthongs (§ 440) and ĭə
(§ 466a) and at the end of the diphthongs ei (§ 386), ai (§ 406),
ɔi (§ 437), oi (§§ 403, 466x, 869), ui (§§ 327a, 869), ŭi (§ 466v).
Foreign learners should be careful not to use a close i in these
diphthongs.

English vowel No. 3: e

267. The English phoneme e, in my speech, has several allophones,
i.e. it comprises several shades of sound, the use of which
69is determined by the nature of the surrounding sounds in the
sequence (see § 274). The differences are, however, slight and of
no importance for the foreign learner. If the foreign learner always
uses the principal member of the phoneme, his pronunciation will
always sound correct.

268. The tongue-position for the principal English e is shown
by the position of the dot in Fig. 34. The following is a formal
description of the manner of forming the sound:

image

Fig. 40. Lip-position of
English e.

(i) height of tongue: intermediate
between half-close and half-open;

(ii) part of tongue raised: the ‘front’;

(iii) position of lips: spread or neutral
(Fig. 40);

(iv) opening between the jaws: medium.

In normal speech the tip of the tongue touches the lower teeth,
but small variations in its position do not materially affect the
tamber. As with all normal vowels, the soft palate is in its
raised position and the vocal cords are in vibration.

269. A palatogram of the sound as pronounced by me is shown
in Fig. 41. It will be observed that the air-passage is wider than
in the case of short i (Fig. 39).

image

Fig. 41. Palatogram
of English e

270. e is the so-called ‘short’ sound of the
letter e 1287; examples: pen pen, red red, seven
ˈsevn. e is also the sound of ea in many words;
examples: head hed, breath breθ. Note the
exceptional words any ˈeni, many ˈmeni,
Thames temz, ate et, Pall Mall ˈpelˈmel. 1388

271. The vowel in these words varies a good
deal with different English speakers. Some
Londoners use a closer sound than that described above; other
speakers use an opener sound nearer to cardinal ɛ. The symbol
e may in fact be taken to represent a diaphone (Chap. XI) with
several members. The intermediate or ‘average’ sound described
70in § 268 is recommended for foreign students of English. Slight
divergences from this will, however, not cause the student's pronunciation
to sound un-English.

272. Many foreign people, and especially the French, replace the
English e by a very open ɛ. This is especially the case when r
follows, as in very ˈveri. This mispronunciation may be rectified
by remembering that the sound to aim at for English is not identical
with the French sound in même mɛːm, père pɛːr, belle bɛl, etc., but
is intermediate in quality between this and the sound of French é.

273. Words for practice: pen pen, bed bed, text tekst, deaf def,
kept kept, get get, check ʧek, gem ʤem, men men, neck nek, lend
lend, red red, fed fed, very ˈveri, then ðen, seven ˈsevn, zest zest,
shed ʃed, head hed, yes jes, well wel.

274. The only subsidiary member of the English e-phoneme
worthy of note is an opener and retracted variety, which is used
when ‘dark’ l follows, as in tell tel, 1489 shell ʃel, 1490 felt felt, else els,
elder ˈeldə. It is not necessary for the foreign learner to make
any special effort to use an opener e in these cases. The use of
the same e as in other words does not sound un-English.

275. An opener variety of e (ɛ) occurs in English as the first
element of the diphthong ɛə (see § 446).

English vowel No. 4: æ

276. The English phoneme represented in this book by the
symbol æ may, from the point of view of the foreign learner, be
regarded as comprising only one sound. There is one member of
the phoneme which differs from this sound, namely a rather opener
variety used before ‘dark’ l (as in alphabet ˈælfəbit); this is a variety
which may, however, be ignored by the foreign learner.

277. It will be seen from Fig. 34 that in forming æ the tongue
is low in the mouth, and occupies a position which is roughly
71intermediate between the positions for cardinal ɛ and cardinal a. 1591
The following is a formal description of the manner of forming æ;

image

Fig. 42. Lip-position of
English æ.

(i) height of tongue: between half-open
and open;

(ii) part of tongue which is highest: the
‘front’;

(iii) position of lips: spread or neutral
(Fig. 42);

(iv) opening between the jaws: medium
to wide.

In normal speech the tongue-tip touches the lower teeth, but
small variations in its position do not materially
affect the tamber. As with all normal vowels,
the soft palate is in its raised position and the
vocal cords are in vibration.

image

Fig. 43. Palatogram
of English æ.

278. A palatogram of the sound as pronounced
by me is shown in Fig. 43.

279. æ is the so-called ‘short’ sound of the
letter a 1692; examples: glad glæd or glæːd, bag
bæg or bæːg, pad pæd, cat kæt, lamp læmp.
The sound is regularly represented by the letter
a, the only exceptions being plait plæt, plaid plæd, Plaistow ˈplæstou.
Note that have is hæv (strong form), 1793 and that bade is often pronounced
bæd but has an alternative form beid.72

280. Many foreign people, and especially the French, replace the
vowel æ by an opener sound of the a-type (§ 409), which is the
sound in the French patte pat, cave kaːv, and the initial element
in the variety of English diphthong ai described in this book.
Germans, on the other hand, commonly replace æ by some variety
of ɛ, thus making no difference between man mæn and men men,
pat pæt and pet pet.

281. The correct sound of æ can generally be obtained by
remembering that se must have a sound intermediate in quality
between an ɛ and an a. It is useful in practising the sound to
keep the mouth very wide open.

282. The sound may also be obtained by trying to imitate the
baaing of a sheep, which resembles ˈbæːˈbæː. Those who are unable
to obtain the exact quality by practising such exercises should note
that it is better to err on the side of a rather than on the side of ɛ.
a is actually used for æ in many parts of the North of England.

283. Words for practice: pat pæt, bad bæd or bæːd, tax tæks,
damp dæmp, cat kæt, gas gæs, chat ʧæt, jam ʤæm or ʤæːm,
man mæn or mæːn, nap næp, lamb læm or læːm, rash ræʃ, fat fæt,
van væn, thank θæŋk, that (demonstrative pronoun) ðæt, 1894 sand
sænd or sæːnd, exact igˈzækt, shall ʃæl, 1995 hang hæŋ, wag wæg.

English vowel No. 5: ɑː

284. The English phoneme represented in this book by the
symbol ɑ may be regarded as comprising only one sound 2096; there
73are no members of the phoneme which differ to any marked extent
from this sound. It is always relatively long and is therefore
generally written with a length-mark.

285. It will be seen from the position of the dot in Fig. 34
that in forming the English ɑː the tongue is held very low down
in the mouth, and that the vowel is nearer to cardinal ɑ than to
cardinal a. The following is a formal description of the manner
of forming the sound:

image

Fig. 44. Lip-position of
English ɑː.

(i) height of tongue: fully open;

(ii) part of tongue which is highest: a
point somewhat in advance of the
centre of the ‘back’;

(iii) position of lips: neutral (Fig. 44);

(iv) ,opening between the jaws: medium
to wide.

The tip of the tongue is generally, though not necessarily,
somewhat retracted from the lower teeth. As with all normal
vowels, the soft palate is in its raised position and the vocal cords
are in vibration.

286. The sound ɑː gives no palatogram.

287. ɑː is the usual Southern English sound of the sequence of
letters ar when at the end of a word or when followed by a
consonant; examples: far fɑː, part pɑːt, garden ˈgɑːdn. A has
the sound ɑː in half hɑːf, calm kɑːm, and several other words in
which the l is silent (see § 662); also in numerous words when
followed by ff, ss, or by f, s, or n followed by another consonant,
e.g. staff stɑːf, class klɑːs, pass pɑːs, after ˈɑːftə, fast fɑːst, castle
ˈkɑːsl, ask ɑːsk, command kəˈmɑːnd, grant grɑːnt, can't kɑːnt, also
in most words ending in ath, e.g. bath bɑːθ; also in some words
of recent foreign origin, e.g. moustache məsˈtɑːʃ, drama ˈdrɑːmə,
tomato təˈmɑːtou, vase vɑːz. Note also the words ah ɑː, are ɑː, 2197
aunt ɑːnt, draught drɑːft, laugh lɑːf, clerk klɑːk, Berkeley ˈbɑːkli, 2298
Berkshire ˈbɑːkʃiə or ˈbɑːkʃə, Derby ˈdɑːbi, Hertford ˈhɑːfəd, sergeant
ˈsɑːʤənt, example igˈzɑːmpl, heart hɑːt, hearth hɑːθ, father ˈfɑːðə,
74rather ˈrɑːðə, and words borrowed from modern French, such as
memoir ˈmemwɑː, reservoir ˈrezəvwɑː, 2399 barrage ˈbærɑːʒ. (See
further §§ 294, 295.)

288. The English vowel ɑː is about the same as the vowel used
by many Parisians in pâte pɑːt.

289. Most Germans 24100 and people from many other foreign
countries (e.g. Scandinavians, Hungarians, Portuguese) have a
tendency to use a forward a approaching Cardinal Vowel No. 4
(§§ 133, 144) in place of the English ɑː. By practising a fully
back variety of ɑ with the tongue held as low down and as far
back as possible, they will realize better the nature of the English
ɑ. It should also be noticed that the English ɑː is somewhat,
similar in quality (though not in quantity) to the English short ɔ.
thus card kɑːd is rather like cod kɔd with the vowel lengthened.
It is helpful to practise the sound with the tip of the tongue
touching the lower teeth.

290. When ɑː is followed by a nasal consonant, the Portuguese
often replace it by a vowel resembling əː (§ 343), pronouncing
for instance answer (Southern English ˈɑːnsə) almost ˈəːnsər (or
ˈə̃ːsər with a nasalized əː).

291. Foreign learners wishing to acquire the pronunciation
described in this book must be careful not to add a r-sound of any
sort after the sound ɑː unless a vowel follows. Thus the English
word marsh mɑːʃ is practically identical with the French mâche;
many Germans (from Saxony, Hamburg, etc.) pronounce Bahn
exactly like the English barn bɑːn; far is pronounced fɑː, though
far away is ˈfɑːr əˈwei (§ 756).

292. Some English speakers diphthongize slightly the sound ɑː
especially when final, saying for instance fɑə for fɑː. This pronunciation
is not, however, the most usual in educated Southern
English.

293. Words for practice: palm pɑːm, bath bɑːθ, task tɑːsk, dark
dɑːk, carve, calve kɑːv, guard gɑːd, charm ʧɑːm, jar ʤɑː, marsh
75mɑːʃ, nasty ˈnɑːsti, clerk klɑːk, rather ˈrɑːðə, far fɑː, vase vɑːz,
psalm sɑːm, sharp ʃɑːp, hard hɑːd, yard jɑːd.

294. A number of words written with a are pronounced with
æ by some Southern English people but with ɑː by others. Such
are the words ending in -aph, like photograph ˈfoutəgræf or
ˈfoutəgrɑːf, telegraph, cenotaph, various words where the vowel is
followed by s, z, θð, d, ns, nt, such as ass æs or ɑːs, 25101 mass (catholic
service) mæs or maːs, masque, masquerade, contrast (noun) ˈkɔntræst
or ˈkɔntrɑːst, contrast (verb) kən'træst or kən'trɑːst, blasphemy,
askance əsˈkæns or əsˈkɑːns, ant, lather, catholic, and all the words
beginning with trans-, such as translate trænsˈleit or
trɑːnsˈleit, transfusion, transparent, transatlantic ˈtrænzətˈlæntik
or ˈtrɑːnzatˈlæntik, transact trænˈzækt or trɑːnˈzækt (also trænˈsækt,
trɑːnˈsækt), transmit trænsˈmit or trɑːnsˈmit (also trænzˈmit,
trɑːnzˈmit). 26102

295. Many similarly spelt words are not subject to this variation
of sound. Some are pronounced with æ and others with ɑː. There
are no rules governing the use of one vowel or the other, so that
the foreign student is obliged to learn the words individually.
Examples of words pronounced with æ are: photographic
foutəˈgræfik, and (strong form) ænd, band, hand, land, sand,
romance, finance fiˈnæns or faiˈnæns, manse, substantial səbsˈtænʃl,
bass (fish), 27103 Bass (name), crass, lass, mass (quantity), morass
məˈræs, molasses, Passe (surname), Passfield, crevasse, gas, Ascot,
aster, bast, cant, Levant, rant, scanty. Examples of words pronounced
with ɑː are: after ˈɑːftə, calf, half, giraffe, laugh lɑːf, craft,
draft, draught, raft, shaft, waft, 28104 demand, remand, command, reprimand,
advance, chance, dance, glance, lance, trance, advantage
ədˈvɑːntiʤ, chant, grant, plant, slant, 29105 an't, 30106 can't (cannot), ask
ɑːsk, bask, basket, cask, flask, mask, təsk, brass, class, glass, grass,
76pass, blast, fast, aghast, last, mast, past, repast, vast, castle, caster,
Castor ˈkɑːstə, master, plaster, bath bɑːθ, lath, path, 31107 father, rather.

English vowel No. 6: ɔ

296. The English vowel represented in this book by the symbol
ɔ without length-mark is the member of the English ɔ-phoneme
used when the vowel is relatively short.

297. It will be seen from the position of the dot in Fig. 34
that in forming short ɔ the tongue is held in the lowest and most
backward position possible. Any further retraction of the tongue
would give rise to a fricative consonant of the ʁ-type (§ 763).
The vowel has the tongue-position of Cardinal Vowel No. 5 (a)
combined with open lip-rounding.

298. The following is a formal description of the manner of
forming the sound:

image

Fig. 45. Lip-position of
English short ɔ.

(i) height of tongue: fully open;

(ii) part of the tongue which is highest:
the back;

(iii) position of lips: open lip-rounding
(see Fig. 45);

(iv) opening between the jaws: medium
to wide.

The tip of the tongue is generally, though not necessarily, somewhat
retracted from the lower teeth. As in the case of all normal
vowels, the soft palate is in its raised position and the vocal cords
are in vibration.

299. English short ɔ gives no palatogram.

300. ɔ is the short sound of the letter o; examples: not nɔt,
pond pɔnd, dog dɔg, sorry ˈsɔri, solid ˈsɔlid. 0 is also pronounced ɔ
with a variant ɔː in many words where f, s or θ follows; examples:
off ɔf (or ɔːf), often ˈɔfn (or ˈɔːfn), loss lɔs (or lɔːs), cost kɔst (or kɔːst),
cloth klɔθ (or klɔːθ). Ou is similarly pronounced in cough kɔf (or
kɔːf) and trough trɔf (or trɔːf). A often has the sound ɔ when the
vowel is preceded by w and not followed by k, g or ŋ; examples:
77want wɔnt, what wɔt, squash skwɔʃ, quality ˈkwɔliti 32108 (but wax wæks,
wag wæg, twang twæŋ). Many Southern English people use ɔ
instead of the older ɔː before l or s followed by a consonant, e.g.
false fɔːls or fɔls, fault fɔːlt or fɔlt, halt hɔːlt or hɔlt. Austria and
Australia are now generally pronounced ˈɔstrĭə, ɔsˈtreiljə (less
commonly ˈɔːistrĭə, ɔːsˈtreiljə). Note the exceptional words gone
gɔn (rarely gɔːn), shone ʃɔn, because biˈkɔz, cauliflower ˈkɔliilauə,
laurel ˈlɔrəl, (ac)knowledge (ək)ˈnɔliʤ, Gloucester ˈglɔstə, yacht jɔt.

301. Foreign people generally do not make the English sound ɔ
open enough. The French use their vowel in note nɔt, bonne bɔn;
Germans use their vowel in Gott gɔt; and so on. The usual German
vowel in Gott is about Cardinal Vowel No. 6. This word is very

image

Fig. 46. Relation between English short
ɔ and French and German ɔ-sounds.
(The small dots represent Cardinal Vowels.)

distinct from the English
word got gɔt; the tongue-position
of the German ɔ is
notably higher than that of
the English ɔ. The French
(Parisian) vowel in note nɔt
is not merely higher than
the English ɔ but also more
advanced. The relations
between the English short
ɔ and these French and
German vowels are shown
in Fig. 46.

302. Foreign learners must remember that in pronouncing the
English short ɔ the tongue must be held as low down and as far back
as possible. Usually the best way of acquiring the vowel is for
them to aim at a sound intermediate between ɑ and their variety
of ɔ.

303. Cases in which the sound ɔ occurs in unstressed syllables
often seem particularly difficult to foreign learners and require
special practice. Examples: cannot ˈkænɔt, 32a109 a day on the river ə
ˈdei ɔn ðə ˈrivə
, What are you thinking of? ˈwɔt ə juː ˈθiŋkiŋ ɔv?78

304. Words for practice: spot spɔt, bother ˈbɔðə, top tɔp, cotton
kɔtn, got gɔt, chop ʧɔp, John ʤɔn, moss mɔs, not nɔt, long lɔŋ,
rock rɔk, foreign ˈfɔrin, involve inˈvɔlv, 33110 methodical miˈθɔdikl, sorry
ˈsɔri, shop ʃɔp, hop hɔp, yacht jɔt, squash skwɔʃ, watch wɔʧ. 34111

English vowel No. 7: ɔː

305. ɔː is the member of the English ɔ-phoneme which is used
when the vowel is relatively long. Its tongue-position is low,
though not quite so low as for the short ɔ. The lips are rounded
so as to leave an opening which is much smaller than in the case
of the short ɔ (see Fig. 47). The vowel differs from Cardinal ɔ
in two respects: (i) it is formed with the tongue a little lower than
for Cardinal ɔ, (ii) the lips are more closely rounded than for Cardinal
ɔ (see Fig. 20).

306. The following is a formal description of the manner of
forming the English long ɔː:

image

Fig. 47. Lip-position of
English long ɔː.

(i) height of tongue: between half-open
and open;

(ii) part of tongue which is highest:
the back;

(iii) position of lips: between open and
close lip-rounding (Fig. 47);

(iv) opening between the jaws: medium
to fairly wide.

The tip of the tongue is generally, though not necessarily, slightly
retracted from the lower teeth. As with all normal vowels, the
soft palate is in its raised position and the vocal cords are in
vibration.

307. The sound ɔː gives no palatogram.

308. ɔː is the regular sound of aw and au; examples: saw sɔː,
lawn lɔːn, author ˈɔːθə. 35112 It is also the regular sound of or when
final or followed by a consonant; examples: nor nɔː (like gnaw),
79short ʃɔːt, form fɔːm. The groups ore, oar are commonly pronounced
ɔː, though a diphthong ɔə is also frequently used in such cases
(see § 458); examples: more mɔː or mɔə, roar rɔː or rɔə, board bɔːd
or bɔəd. ɔː with the variant ɔə is also heard in some words spelt
with our; examples: pour pɔː or pɔə, couræ kɔːs or kɔəs. 36113 A
frequently has the value ɔː when followed by l final or followed
by a consonant; examples: appal əˈpɔːl, all ɔːl, halt hɔːlt. 37114 Ar
frequently has the value ɔː when the vowel is preceded by w and
followed by a consonant, examples: swarm swɔːm, quart kwɔːt. 0
is pronounced ɔː by some in words like off, loss, cost, cloth, cough, as
mentioned in § 300. 38115 Ough has the value ɔː when followed by t,
as in bought bɔːt, thought θɔːt. 39116 Note the exceptional words, broad
brɔːd, door dɔː or dɔə, floor flɔː or flɔə, water ˈwɔːtə, wrath rɔːθ.

309. The sound ɔː is best acquired by imitation, while observing
carefully the position of the lips. A very near approach to the
correct quality is obtained by trying to produce the tamber of the
English short ɔ with lips in the position for the continental close o
(as in the French côte koːt, German wohl voːl). Many foreign
people do not use sufficient lip-rounding in pronouncing the English
ɔː, especially when there is no r in the spelling (as in all, saw,
thought). When there is an r in the spelling (as in sore, soar, four,
nor), Germans generally replace the vowel by the close and
say soːr, foːr, etc.80

310. Foreign learners wishing to acquire the pronunciation
described in this book must be particularly careful not to add a
r-sound of any sort after the vowel or, unless a vowel follows.
Nor said by itself is pronounced exactly like gnaw nɔː, stork is
identical with stalk stɔːk. Note, however, cases like more easily
mɔːr ˈiːzili where a ‘linking’ r is inserted on account of the following
vowel.

311. Some foreign people (and especially the French and Italians)
have difficulty in distinguishing the sound ɔː from the diphthong
ou. Those who have this difficulty should study carefully the
differences between the two sounds (§§ 305, 394).

312. Words for practising the sound ɔː; paw, pour, pore pɔː, 40117
bought bɔːt, talk tɔːk, door dɔː, 40118 caught kɔːt, Gordon ˈgɔːdn, chalk
ʧɔːk, George ʤɔːʤ, more mɔː, 40119 gnaw, nor nɔː, law lɔː, raw, roar rɔː, 40120
drawer (sliding box in a table or cupboard) drɔː, 41121 for, four, fore fɔː, 42122
Vaughan vɔːn, thought θɔːt, sauce, source sɔːs, 43123 short ʃɔːt, horn hɔːn,
your jɔː, 44124 warn, worn wɔːn.

313. A sound near in quality to ɔː occurs as the first element
of the diphthong ɔə (see § 455) and a very similar sound occurs
as the first element of the diphthong ɔi (see § 437).

English vowel No. 8: u

314. The sound represented in this book by u without lengthmark
is the member of the English u-phoneme used when the
vowel is relatively short.

315. It will be seen from the position of the dot in Fig. 34
that the English short u has a tongue-position considerably higher
than that of the English long ɔː, and somewhat advanced. The
tongue is not so high as for long . The lips are rounded fairly
81closely, but not so closely as for the long (see Figs. 48, 49).
The distance between the jaws is less than for ɔ and ɔː. Some
writers call this sound a ‘lax’ vowel (see §§ 159-163).

316. The following is a formal description of the manner of
forming the English short

image

Fig. 48. Lip-position of
English short u.

(i) height of tongue: just above half-close;

(ii) part of tongue which is highest: the
fore part of the back;

(iii) position of lips: fairly close lip-rounding
(Fig. 48);

(iv) opening between the jaws: medium.

The tip of the tongue is generally, though not necessarily, somewhat
retracted from the lower teeth. As in the case of all normal
vowels, the soft palate is in its raised position and the vocal cords
are in vibration.

317. The English short u gives no palatogram.

318. u is one,of the two so-called ‘short’ sounds of the letter
u; examples: put put, full ful, bush buʃ, cushion ˈkuʃin. 45125 Oo has
the sound u when followed by k, as in book buk, look luk, 46126 and
in the following miscellaneous words: foot fut, good gud, hood (and
the suffix -hood) hud, stood stud, wood wud, wool wul. In broom
(for sweeping), 47127 groom, room, and soot both and u are heard,
the u-forms brum, grum, rum, sut being perhaps the more usual
in Received English. 48128 Soon is generally suːn, but some English
people pronounce sun. Note the miscellaneous words bosom ˈbuzəm,
bouquet ˈbukei, could kud, 49129 courier ˈkurĭə, should ʃud, 49130 wolf wulf,
Wolverhampton ˈ wulvəhæmptən (and a few other similar names),
woman ˈwumən, Worcester ˈwustə, worsted (woollen material)
ˈwustid, 50131 would wud. 5113282

319. Many foreign people, and especially speakers of Romance
languages, use a sound which is too ‘tense’; in fact they do not
make the necessary difference of tamber between the English
short u and long . Thus they will pronounce pull too much
like pool puːl, and full too much like fool fuːl. The correct sound
of the short u may be generally acquired by trying to pronounce
the vowel in a slack sort of way, using only the amount of liprounding
shown in the photograph, Fig. 48.

320. Words for practising short u; push puʃ, butcher ˈbuʧə,
took tuk, could kud, good gud, nook nuk, look luk, room rum, full
ful, soot sut, hook huk.

321. A sound of approximately the quality of short u also
occurs in English at the beginning of the diphthongs (§ 460),
ŭə (§ 466m), ui (§ 327a) and ŭi (§ 466v), and at the end of the
diphthongs ou (§ 394) and au (§ 420).

English vowel No. 9:

322. The notation is employed to denote those members of
the English u-phoneme which are used when the vowel is relatively
long. Two of these members require notice here, the common
long and an ‘advanced’ variety (see § 326).

323. The tongue-position of the common long is shown by
the position of the dot in Fig. 34. It will be seen that the sound
is noticeably different from Cardinal Vowel No. 8, its tongueposition
being rather lower and more forward than the cardinal
sound. The lips are fairly closely

image

Fig. 49. Lip-position of
English long .

rounded as shown in Fig. 49; the lip-rounding
is normally a little less close
than that of cardinal u, but when
pronounced with exaggerated distinctness
the close lip-rounding of cardinal u
may be used. The distance between
the jaws is less than for the short u.
Some writers call the English long a ‘tense’ vowel.

324. The following is a formal description of the manner
forming this English long :

(i) height of tongue: nearly close;

(ii) part of tongue which is highest: the back;83

(iii) position of lips: close lip-rounding (Fig. 49);

(iv) opening between the jaws: narrow to medium.

The tip of the tongue is generally, though not necessarily, somewhat
retracted from the lower teeth. As in the case of all normal
vowels, the soft palate is in its raised position and the vocal cords
are in vibration.

325. This gives no palatogram.

326. The most important subsidiary long is an ‘advanced’
variety. It is used when j precedes, as in music ˈmjuːzik, tube
tjuːb, deluge ˈdeljuːʤ. By calling it ‘advanced’ we mean that
the part of the tongue which is highest is the central part-a part
more forward than the ‘back’. The use of this advanced variety
is not essential for foreign learners. 52133

327. is the so-called ‘long’ sound of the letter u (the sound
j being inserted before it in many cases, see rules in § 817); examples:
rule ruːl, June ʤuːn, blue bluː, music ˈmjuːzik, future ˈfjuːʧə,
tube tjuːb. Oo has the sound in most words in which the oo
is not followed by r or k; examples: too tuː, food fuːd, spoon spuːn
(for exceptions see § 318). 0 has the sound in ado əˈduː, do
duː, 53134 to tuː, 54135 who huː, whom huːm, lose luːz, move muːv, prove
pruːv, tomb tuːm. Ou has the sound in some words, the principal
being routine ruːˈtiːn, soup suːp, croup kruːp, douche duːʃ, group
gruːp, rouge ruːʒ, route ruːt, 55136 through θruː, uncouth ʌnˈkuːθ, wound
(injury, injure) wuːnd, 56137 you juː, youth juːθ. 57138 (with or without
a preceding j, see rules in § 817) is also the usual sound of eu, ew
and ui; examples: feud fjuːd, new njuː, crew kruː, suːt sjuːt, 58139 fruit
fruːt. Note the exceptional words beauty ˈbjuːti (and its derivatives)
and shoe ʃuː, canoe kəˈnuː, manoeuvre məˈnuːvə.84

327a. When stressed is followed by i, the sequence is sometimes
reduced to a falling diphthong ui. Thus ruin, bluish are
pronounced ˈruːin or ruin, ˈbluːiʃ or bluiʃ, and doing ˈduːiŋ is often
reduced to ˈduiŋ. When is followed by ə there is generally an
alternative pronunciation with the diphthong (No. 21). For
instance, fewer, doer are pronounced either ˈfjuːə, ˈduːə or fjuə, duə.
See § 461.

328. The common English long has nearly the same quality
as the normal French vowel in rouge ruːʒ. It differs slightly from
the usual North German vowel in gut guːt, which is cardinal .
The result is that the of Germans speaking English generally
sounds somewhat too ‘deepe’ in quality. This deep quality of
is often very noticeable when Germans pronounce the phrase How
do you do?
The correct pronunciation is ˈhau dju ˈduː with the
English variety of ; Germans often say hau ˈduː juː ˈduː with
the deeper German variety of .

329. This ‘deep’ variety of sounds particularly unnatural to
English ears in the words requiring the advanced (§ 326), e.g. in
new njuː, music ˈmjuːzik, tube tjuːb, produce (verb) prəˈdjuːs, few
fjuː. The use of a ‘deep’ is less objectionable in other words,
such as food fuːd, lose luːz, soup suːp.

330. Many English people diphthongize slightly the sound ,
especially when final. This diphthongization takes the form of
a gradual increase of the lip-rounding; it may be symbolized
phonetically by u̜ụ 59140 or ɷu 60141 or uw 61142: thus, shoe, few are pronounced
ʃuː, fjuː, or ʃuw, fjuw.

331. It is better for foreign learners not to attempt to diphthongize
the English , because an exaggeration of the diphthong
sounds incorrect.

332. Words for practising the sound ; pool puːl, boot buːt,
tomb tuːm, doom duːm, cool kuːl, goose guːs, chew ʧuː, June ʤuːn
85move muːv, noon nuːn, loose luːs, lose luːz, blue bluː, rule ruːl, root
ruːt, food fuːd, soup suːp, Zoo zuː, shoe ʃuː, who huː, you, yew juː,
woo wuː, pew pjuː, beauty ˈbjuːti, tune tjuːn, dew djuː, cue, Kew
kjuː, music ˈmjuːzik, new njuː, lute luːt (or ljuːt), few fjuː, view vjuː,
sue sjuː, presume priˈzjuːm, 62143 hew, hue, Hugh hjuː.

English Vowel No. 10: ʌ

333. The English phoneme represented by ʌmay be regarded
as comprising only one sound; there are no members of the phoneme
differing to any marked extent from this sound. It is heard in
such words as cup kʌp, lump lʌmp.

334. It will be seen from the position of the dot in Fig. 34
that the tongue-position of my variety of ʌis that of an advanced ɔ.
The vowel is, however, pronounced with lip-spreading (see Fig. 50).
The distance between the jaws is wide; the sound cannot be pronounced
properly with a narrow opening between the jaws.

335. The following is a formal description of the manner of
forming my English ʌ

image

Fig. 50. Lip-position of
my English ʌ.

(i) height of tongue: half-open;

(ii) part of tongue which is highest: the
fore part of the back;

(iii) position of lips: spread (Fig. 50);

(iv) opening between the jaws: wide.

The tip of the tongue generally touches
the base of the lower teeth, but its precise
position does not appreciably affect the
tamber. As in the case of all normal vowels, the soft palate is in
its raised position and the vocal cords are in vibration.

335a. The vowel in the above words varies to some extent with
different Southern English speakers. In particular there are many
who use a more ‘advanced’ and less ə-like vowel than mine. Their
sound tends towards Cardinal a. In the North of England a raised
and retracted variety resembling a is very commonly heard. So
the letter ʌmay be taken to denote a diaphone comprising several
members.

336. My ʌgives no palatogram; nor do the other varieties.86

337. ʌ is one of the two ‘short’ sounds of the letter u; examples:
cut kʌt, mutton ˈmʌtn, hurry ˈhʌri. 0 has the sound ʌin a good
many words; the principal are: among əˈmʌŋ, come kʌm, comfort
ˈkʌmfət, company ˈkʌmpəni, compass ˈkʌmpəs, conjure (to do things
as if by magic) ˈkʌnʤə, 63144 constable ˈkʌnstəbl, done dʌn, front frʌnt,
frontier ˈfrʌntjə, 64145 honey ˈhʌni, London ˈLʌndən, Monday ˈmʌndi,
money ˈmʌni, -monger -mʌŋgə, mongrel ˈmʌŋgrəl, monk mʌŋk,
monkey ˈmʌŋki, month mʌnθ, none nʌn, one wʌn (same pronunciation
as won), once wʌns, onion ˈʌnjən, pommel ˈpʌml, some sʌm, 65146
Somerset sʌməsit, son sʌn (same pronunciation as sun), sponge
spʌnʤ, stomach ˈstʌmək, ton tʌn, Tonbridge ˈtʌnbriʤ, tongue tʌŋ,
won wʌn, wonder ˈwʌndə, above əˈbʌv, cover ˈkʌvə, covet ˈkʌvit,
covey ˈkʌvi, dove dʌv, glove glʌv, govern ˈgʌvən, love lʌv, oven ˈʌvn,
shove ʃʌv, shovel ˈʃʌvl, slovenly ˈslʌvnli, borough ˈbʌrə, thorough ˈθʌrə,
worry ˈwʌri, other ˈʌðə, brother ˈbrʌðə, mother ˈmʌðə, smother ˈsmʌðə,
nothing ˈnʌθiŋ, dozen ˈdʌzn, colour ˈkʌlə, twopence ˈtʌpəns. 66147 Ou
has the value ʌ in a few words; the principal are: courage ˈkʌriʤ,
country ˈkʌntri, cousin ˈkʌzn, couple ˈkʌpl, double ˈdʌbl, enough
iˈnʌf, flourish ˈflʌriʃ, hiccough ˈhikʌp, nourish ˈnʌriʃ, rough rʌf,
southern ˈsʌðən, southerly ˈsʌðəli, Southwark (London borough)
ˈsʌðək, 67148 tough tʌf, trouble ˈtrʌbl, young jʌŋ. Note also the exceptional
words does dʌz, 68149 blood blʌd, flood flʌd.

338. Foreign people generally replace this vowel by some variety
of a (§§ 133, 409) or ɑ (§ 285), or by some variety of front rounded
vowel, for instance the half-open front vowel (phonetic symbol œ)
heard in the French oeuf œf, German zwölf tsvœlf. 69150 Thus they
commonly pronounce up as ap or ɑp or œp.

339. ʌ as I pronounce it can often be acquired without much
difficulty by imitation, provided that care is taken not to add any
87trace of lip-rounding. Some foreign people are able to obtain the
sound by unrounding continental varieties of ɔ, such as those heard
in the French bonne bɔn, German Kopf kɔpf, etc.; it is also sometimes
useful to start by unrounding the German close in wohl
voːl, and then to lower the tongue. There is, however, no objection
to using a more a-like sound, as long as the ‘fronting’ is not overdone;
but it is essential to keep the vowel well separated from æ. Much
mʌʧ and struggle ˈstrʌgl must be distinguished clearly from match
mæʧ and straggle ˈstrægl.

340. It is a good plan to learn Vowel No. 11, əː (§§ 343 ff.),
before learning ʌ. It will be seen from Fig. 34 that ʌ, as I pronounce
it, is intermediate between əː and ɑː, and in practice it is found that
ʌ may often be taught by directing the learner to make a sound
about half-way between əː and ɑː.

341. Words for practising ʌ: sponge spʌnʤ, butter ˈbʌtə, tug
tʌg, dull dʌl, come kʌm, gun gʌn, chuckle ˈʧʌkl, judge ʤʌʤ,
money ˈmʌni, nothing ˈnʌθin, luck lʌk, trouble ˈtrʌbl, fuss fʌs,
vulture ˈvʌlʧə, thumb θʌm, thus ðas, such sʌʧ, result riˈzʌlt, shut
ʃʌt, hurry ˈhʌri, young jʌŋ, won, one wʌn.

English vowel No. 11: əː

342. The English sound represented in this book by əː is the
member of the ə-phoneme used when the vowel is relatively long.
(Reasons for regarding əː and ə as belonging to the same phoneme
are given in my book, The Phoneme, §§ 197 ff.) The sound əː
varies to some extent with different speakers of Southern English;
the vowel described in the next paragraph is the one used by
myself, and I believe it to be the most frequent variety. 70151

343. It will be seen from the position of the dot in Fig. 34
that əː is a central vowel; in other words the central part of the
88tongue is raised in order to make it. The tongue is raised to
about mid-way between the ‘half-close’ and ‘half-open’ positions,
or perhaps a shade higher than this. The lips are spread almost
as for (compare Figs. 51 and 36). The opening between the
jaws is narrow; it is impossible to make the sound properly with
a wide open mouth; the sound is in this respect very different
from ʌ(see § 334).

344. The following is a formal description of the manner of
forming my variety of English əː:

(i) height of tongue: about half-way
between ‘open’ and ‘close’;

(ii) part of tongue which is highest: the
central part, culminating at the
junction between ‘front’ and ‘back’;

(iii) position of lips: spread (Fig. 51);

(iv) opening between the jaws: narrow.

image

Fig. 51. Lip-position of
English long əː.

The tip of the tongue generally touches the base of the lower
teeth, but as long as it is near the lower teeth, its precise position
does not appreciably affect the quality of the sound. As in the
case of all normal vowels, the soft palate is in its raised position,
and the vocal cords are in vibration.

345. The vowel əː, as I pronounce it, gives no palatogram.

346. əː is the usual sound of stressed er, ir, ur, and yr when
final or followed by a consonant; examples: her həː, 71152 fern fəːn,
fir fəː, bird bəːd, fur fəː, turn təːn, myrtle ˈməːtl. Ear followed
by a consonant is generally pronounced əː; examples: earn əːn,
earth əːθ, heard həːd. 72153 Or is generally pronounced əː when preceded
by w; examples: work wəːk, world wəːld; it is also pronounced so
in attorney əˈtəːni. Our is pronounced əː in adjourn əʤəːrn. courteous
ˈkəːtjəs, 73154 courtesy ˈkəːtisi, 74155 journal ˈʤəːnl, journey ˈʤəːni, scourge
skəːʤ. Note the exceptional words colonel ˈkəːnl, amateur ˈæmətəː, 75156
89connoisseur kɔniˈsəː, chauffeur ʃouˈfəː 76157 and a number of other words
ending in -eur. Year is pronounced jjəː or jiə. (I pronounce jjəː.)
The word were has two pronunciations, wəː and wɛə (besides a weak
form ). The word girl is usually pronounced gəːl; gɛal and
giəl are also not infrequent. Foreign learners are recommended
to use the forms wəː, gəːl.

347. The English əː is a very difficult sound for most foreign
people. They often replace it by some variety of front rounded
vowel such as œ or ø, 77158 and in addition to this, they usually add
some kind of r-sound at the end. The word word wəːd will generally
betray a foreigner. Germans usually pronounce it as vœrd or
βœʀt.

348. The most important point to be borne in mind is that
there is no lip-rounding in pronouncing a normal əː; the lips are
spread as for (Figs. 51, 36). Foreign learners who wish to pronounce
in the manner described in this book must take care that
the quality of the sound remains absolutely unchanged while it is
being pronounced, and that no trace of a r-sound is added after
the vowel (unless another vowel follows, as in stirring ˈstəːriŋ, § 756).

349. Many foreign people have a tendencjr to curl back or ‘invert’
the tip of the tongue (§ 831) when trying to pronounce the English
sound əː. This is especially the case with Norwegians and Swedes.
Such a pronunciation is common in American and various forms
of dialectal English, but it is not used by Londoners. The usual
sound of əː may be acquired by keeping the tip of the tongue
firmly pressed against the lower teeth, holding it there if necessary
with the finger, or with the end of a pencil. It is useful to practise
the exercises kəːkəːkəː…, gəːgəːgəː… keeping the tip of the
tongue against the lower teeth.

350. Some foreign learners use a vowel which is too open and
ʌ-like or ɑ-like. Such a fault can generally be remedied by taking
care not to open the mouth too wide; in fact it is often advisable
to practise the sound əː with the teeth kept actually in contact.90

351. Other foreign people, Spaniards and Greeks, for example,
use a vowel which is more front than central, which has too much
resemblance to e or æ. For them it is useful to practise the sound
arrived at by unrounding an , or in other words, to do their best
to say through spread lips. (Unrounded u is represented in
phonetic transcription by ɯ.) The English əː is between this ɯː
and e, and may therefore be learnt by the process described in
§§ 110-114.

352. Germans should note that the English sound əː is very
similar in quality to the variety of ə heard in the second syllable
of the German word bitte ˈbitə (stage pronunciation). This fact
may be utilised in learning to pronounce the English əː.

353. It is helpful for all foreign learners, and particularly for
Germans and Scandinavians, to practise energetically the exercise
uːəːuːəːuːəː… with the teeth in contact, taking care that the
corners of the mouth move horizontally and that there is no vertical
opening of the mouth. Another effective exercise is to practise
iːəːiːəːiːəː… keeping the lips stationary in the position shown
in Fig. 36.

354. Words for practising the vowel əː; pearl pəːl, bird bəːd,
turn təːn, dearth dəːθ, curb kəːb, kernel, colonel ˈkəːnl, 78159 girl gəːl
(see § 346), church ʧəːʧ, germ ʤəːm, myrrh məː, nurse nəːs, learn
ləːn, fur, fir fəː, verse vəːs, thirst θəːst, sir səː, 79160 deserve diˈzəːv,
shirt ʃəːt, hurt həːt, yearn jəːn, work wəːk.

English vowel No. 12: ə

355. The letter ə without length-mark is employed to denote
those members of the English ə-phoneme which are used when
the vowel is relatively short. It is sufficient for practical purposes
to distinguish three of these members, which we may indicate,
when necessary, by the notation ə1 (the principal member), ə2, ə3.
An ə of intermediate quality is often called ‘the neutral vowel’ or
‘schwa.’91

356. ə, is similar to the German sound of e in bitte ˈbitə. It

image front | central | back

Fig. 52. Relation between the English
sounds ə1, ə2, ə3. (The small dots
represent Cardinal Vowels.)

is very near to əː in tamber, but
it is always extremelv short in
English, so that its exact value
is difficult to observe or describe;
the vowel is subject to slight
variations depending on the individual
speaker 80161 and on the
nature of the adjoining sounds. 81162
The approximate tongue-position
of ə1, is shown in Fig. 52. Its lip-position
is similar to that of ə3
shown in Fig. 53-

357. ə1 is the sound of a in along əˈloŋ, attempt əˈtempt, admit
ədˈmit, gentleman ˈʤentlmən, Thomas ˈtɔməs, salad ˈsæləd, breakfast
ˈbrekfəst, malady ˈmælədi; of ar in particularly pəˈtikjuləli,
forward ˈfɔːwəd, standard ˈstændəd; of e in pavement ˈpeivmənt;
of er in modern ˈmɔdən, 82163 concert ˈkɔnsət, manners ˈmænəz, Underground
(railway) ˈʌndəgraund; of i in horrible ˈhɔrəbl; of o in method
meθəd, protect prəˈtekt, melody ˈmelədi, lemon ˈlemən; of or in
effort ˈefət; of oar in cupboard ˈkʌbəd; of u in chorus ˈkɔːrəs, minimum
ˈminiməm; of ou in famous ˈfeiməs. ə1, is also the usual vowel of
the articles the (before consonants) and a; examples: the table ðə
teibl
, a window ə ˈwindou.

358. ə2 is a vowel which has a higher and more retracted tongue-position
than ə1, a vowel which is therefore a kind of ɯ (see
Fig. 52 and § 351). It is a member of the English ə-phoneme
frequently heard when the adjoining consonant is k or g, as in
condemn kənˈdem, to go tə ˈgou, back again ˈbæk əgein, the ground
ðə ˈgraund, hypocrite ˈhipəkrit, 83164 suffocate ˈsʌfəkeit; the ə in these
words is almost equivalent to a very short ɯ (kɯnˈdem, tɯ ˈgou,
ˈbæk ɯgein, ðɯ ˈgraund, etc.).92

359. Sounds intermediate between ə1 and ə2 are also common,
but it is difficult to say precisely in what cases they are used.
Moreover the pronunciation varies considerably from speaker to
speaker. In the words hammock ˈhæmək, Jacob ˈʤeikəb, I think
I generally use a sound nearer to ə1, than to ə2 in spite of the fact
that there is an adjacent k. On the other hand there are many
speakers who use a sound approximating to ə2 in breakfast, pavement,
method and many other words in which the majority would use ə1.

360. Owing to these divergences of pronunciation it is unnecessary
for the foreign learner to distinguish between ə1 and ə2.
His pronunciation will sound quite English if he uses ə1 in all
the above-mentioned cases.

361. ə3 is an opener and more ʌ-like sound than ə1; it is also
pronounced less short than ə1. It may be placed on the Vowel
Figure as shown in Fig. 52. It is used in final position, whereas
ə1 and ə2 never occur in final position.

image

Fig. 53. Lip-position of the
English ‘neutral’ vowel ə (ə2).

The following are examples of words
pronounced with ə3; it is to be understood
that they only have this vowel
when a pause follows: china ˈʧainə,
villa ˈvilə, collar ˈkolə, over ˈouvə,
manner ˈmænə, bitter ˈbitə, father ˈfɑːðə,
actor ˈæktə, honour ˈɔnə, borough ˈbʌrə,
thorough ˈθʌrə, picture ˈpikʧə, centre ˈsentə.

362. Many English speakers actually use ʌ in such words,
pronouncing ˈʧainʌ, ˈvilʌ, ˈkolʌ, etc.

363. When such words are immediately followed by another
word in the same sense-group, the vowel is replaced by ə1. 84165
Thus ə1 is used in china tea ˈʧainə ˈtiː, over there ˈouvə ˈðɛə, a
picture we like
ə ˈpikʧə wiː ˈlaik, my father and I mai ˈfɑːðər
ənd ˈai
.93

364. Sounds intermediate between ə1 and ə3 exist, but are for
the most part of no importance as they may always be replaced
by ə1 or ə3. One, which we may denote by ə4, is sometimes heard
in place of ə1 when words ending in -rə are immediately followed
by another word. Example: borough council ˈbʌrə ˈkaunsl, an
error of judgment
ən ˈerə əv ˈʤʌʤmənt, Dora is here ˈdɔːrə z hiə.

365. In such a word as honoured (which I should pronounce
ˈɔnəd with ə1) some speakers use a vowel almost identical in quality
with ə4 but somewhat lengthened. Being relatively rather long,
yet quite distinct in quality from əː, it has to be regarded as
belonging to a separate phoneme unconnected with the above
described members of the ə-phoneme. It is therefore necessary
in transcribing the pronunciation of these speakers to use a special
phonetic symbol for this vowel. ɐ is the appropriate international
phonetic letter for it.

366. ɐ is then a sound intermediate in tamber between ə1 and
ə3 (see Fig. 52). It is always distinctly longer than ə1 and ə2,
but, unlike ə3, it is never replaced by ʌ. ɐ is found chiefly in
derivatives formed by adding d or z to words ending in ə3, e.g.
honoured ˈɔnəd or ˈɔnɐd, delivered diˈlivəd or diˈlivɐd, manners
mænəz or ˈmænɐz, fatherˈs ˈfɑːðəz or ˈfaːðɐz. Some words not
derived in this way are also pronounced with ɐ by some English
people; they all appear to be literary or rather uncommon words.
Examples are: laggard ˈlægəd or ˈlægɐd, rampart ˈræmpət or
ˈræmpɐt, hazard ˈhæzəd or ˈhæzɐd and the adjective divers ˈdaivəz
or ˈdaivɐz. 85166

367. ɐ is only used in words where there is an r in the spelling
It cannot be used in such words as breakfast, salad, method. The
use of ɐ is probably a spelling pronunciation. It is mentioned
here because it is not uncommon, but it must be understood that
the use of this vowel is in no way necessary for an acceptable
pronunciation of English. Foreign learners are recommended for
94the sake of simplicity to use ə1 in all cases where ɐ is a possible
variant.

368. The foreign learner therefore need only learn two of the
numerous varieties of ə occurring in English, namely ə1 and ə3.
ə3 is to be used in final position (see § 361), and ə1 in all other
cases in which the ə-phoneme occurs short.

369. ə1 and ə3 are both easy sounds for most foreign people. In
the case of ə1 hardly any difficulty arises, owing to the fact that
it is extremely short and that slight deviations from the normal
value pass unnoticed by English hearers. ə1 is moreover almost
identical with the North German sound of e in bitte ˈbitə. French
people generally make the mistake of using the French variety
of ə (‘e mute’) which is said with rounded lips; they must remember
that all the English varieties of ə► are made with spread lips.

370. ə3 lies between əː and ʌ, and is therefore easy to learn
when once the two latter vowels have been acquired. The method
described in §§ 110-114 may be used. Foreign learners who have
difficulty in distinguishing between the ə-sounds and ʌ (Spaniards,
Greeks and the Japanese, for instance) may always use ʌ in place
of ə3 (see § 362).

371. The chief difficulty for foreign people in regard to short ə
lies not in making the sound, but in knowing when to use it.
Ordinary English spelling gives no indication as to when ə is to
be used, and consequently foreigners continually replace it by
some other vowel which the spelling suggests to them. Misled
by the spelling, they say doctor, consider, particularly, amusement
with some such pronunciations as ˈdɔktɔr, kɔnˈsider, parˈtikjularli,
eˈmjuːzment instead of ˈdɔktə, kənˈsidə, pəˈtikjuləli, əˈmjuːzmənt.
Moreover they are not generally aware of the differences between
such words as experiment (noun) iksˈperimənt and experiment
(verb) iksˈperiment, workman ˈwəːkmən and coal-man ˈkoulmæn.

372. Some guidance is to be found in the fact that ə only occurs
in unstressed syllables. The following comparisons illustrate this:

present (noun, adj.) ˈprez(ə)nt | present (verb) priˈzent

company ˈkʌmpəni | companion kəmˈpænjən

history ˈhist(ə)ri | historical hisˈtərikl

august (month) ˈɔːgəst | august (adj.) ɔːˈgʌst95

photograph ˈfoutəgræf 86167 | photography fəˈtɔgrəfi 87168 | photographic ˌfoutəˈgræfik

chronology krəˈnɔləʤi | chronological ˌkronəˈlɔʤikl

illustration iləsˈtreiʃn | illustrious iˈlʌstrĭəs

labour ˈleibə | laborious ləˈbɔːrĭəs

magic ˈmæʤik | magician məˈʤiʃn

japan ʤəˈpæn | japanese ˌʤæpəˈniːz

373. Too much reliance must not, however, be placed on the
fact that a syllable is unstressed. All other vowels occur quite
frequently with weak stress (see, however, footnote 3 to § 920).
Examples are: insect ˈinsekt, torment (noun) ˈtɔːment, ferment (noun)
ˈfəːment, comment ˈkɔment, contract (noun) ˈkɔntrækt, asphalt
ˈæsfælt, knapsack ˈnæpsæk, Afghan ˈæfgæn, Zodiac ˈzoudiæk,
cannot ˈkænɔt, epoch ˈiːpɔk, chaos ˈkeiɔs, record (noun) ˈrekɔːd,
statute ˈstætjuːt, hubbub ˈhʌbʌb, convert (noun) ˈkɔnvəːt, Exmouth
ˈeksmauθ, 88169 Greek proper names such as Logos ˈlɔgɔs, Pythagoras
paiˈθægəræs, Thucydides θjuːˈsididiːz; exotic egˈzɔtik, anticipate
æn'tisipeit, carnation kɑːˈneiʃn, Norwegian nɔːˈwiːʤən, mercurial
məːˈkjuərĭəl. Foreign learners who have accustomed themselves
to the frequent occurrence of ə and i in unstressed syllables often
have difficulty in pronouncing words such as the above. In their
anxiety to use a properly they will sometimes produce non-existent
pronunciations such as nəˈwiːʤən, 89170 ˈnæpsək, 89171 ˈstæʧət (for
stætjuːt), ˈhændikəp (for -kæp), ˈeiprikət (for -kɔt), kəmpənˈsəiʃn
(for kɔmpenˈseiʃn or kɔmpənˈseiʃn), ˈgrænsən (for ˈgrænsʌn). Such
mistakes are just as un-English as the failure to use ə in words
which ought to have it.

374. The use of some other vowel instead of ə is particularly
un-English in terminations like -əbl, -əns. Miserable, consequence
are pronounced ˈmizərəbl, ˈkɔnsikwəns and not, as many foreign
people say, ˈmizərabl, ˈkɔnsekwens. Foreign learners can often
improve their pronunciation very much by omitting altogether the
ə's in such words, and practising ˈmizrbl, ˈkɔnskwns. Similarly
96preferable, afterwards, solicitor, successful, sufficient, comfortable may
with advantage be practised as ˈprefrbl, ˈɑːftwdz, ˈslistə, skˈsesfl,
ˈsfiʃnt, ˈkʌmftbl. The word difficult ˈdifikəlt is nearly always
pronounced badly by foreign people; it should be practised as
ˈdifiklt or ˈdifklt.

375. Attention should be given to two special cases in which
the sound ə may not be omitted, namely

(i) when followed by a nasal consonant and preceded by another
nasal consonant, as in woman ˈwumən, German ˈʤəːmən,

(ii) when preceded by a nasal consonant + plosive and followed
by another nasal consonant, as in incumbent inˈkʌmbənt,
London ˈLʌndən, Hampton ˈhæmptən, Islington ˈizlŋtən.

Germans are apt to drop out the ə in such words and to pronounce
ˈwumn (or ˈβumn), ˈʤəːmn (or ˈʤœʀmn), ˈlʌndn (or ˈlœndn or
ˈlœnn), etc.

376. Especially noteworthy is the fact that foreign people
continually use ‘strong forms’ in the case of words which have
‘strong’ and ‘weak’ forms. They generally fail altogether to use
the weak forms of such words as them, have, and, of, from, for
(ðəm, həv or əv or v, ənd or ən or n, əv, frəm, ). The subject
of strong and weak forms is discussed at length in Chap. XVI.
All that need be said here is that English people use weak forms
such as those just mentioned much more often than the strong
forms, though there are a certain number of cases in which the
strong forms must be used (see for instance §§ 996, 997).

377. The proper use of ə in words which have only one pronunciation
may be learnt from a pronouncing dictionary. The
proper use of ə in words which are said sometimes with this vowel
and sometimes with another is acquired by extensive reading of
phonetically transcribed texts.97

Chapter XV
The English Diphthongs

General Remarks

378. A common form of Received Southern English contains
twelve essential diphthong phonemes. Nine of these are included
in the vowel table on p. 61, where they are represented by the
symbols ei, ou, ai, au, ɔi, , ɛə, ɔə, , and are numbered 13 to 21.
To these must be added three ‘rising’ diphthongs, ĭə, ŭə and ŭi,
which may be identified by the numbers 22, 23 and 24. These
latter are dealt with in §§ 466a-466v. If it is desired to show in
writing that the diphthongs 13 to 21 are of a ‘falling’ type, this
can be done by placing the I.P.A. mark ˘ over the second letter
of each digraph thus , , , etc.

378a. Of the above diphthongs two, ɔə and ŭi, may be ignored
by the foreign learner, the first because many Southern English
people (including myself) never use it but replace it by ɔː (§ 458),
and the second because it can always be replaced by disyllabic
u-i (§ 466v).

378b. There exist nine further unessential diphthongs in Southern
English, namely oi, ui, , , a̲ə, , ŏi, ĕə and ŏə. They are
reductions of oui (stressed), uːi, eiə (stressed), aiə, auə, ouə (stressed),
oui (unstressed), eiə (unstressed) and ouə (unstressed) respectively;
they are unessential because they may always be replaced by
these fuller forms. Reference is made to them in §§ 403, 327a,
392a, 414, 430, 403, 466a;. It is necessary to mention the existence
of these diphthongs, since they may often be heard from English
people, but it is not needful for foreign learners to use them.

379. The diphthongs , ɛə, ɔə, have been aptly termed
‘centring’ diphthongs. 1172

380. In some of the English diphthong-phonemes we may
distinguish more than one member; for instance my ei in gate geit
is not quite the same as that in pay pei.98

381. The mode of forming the principal members of the English
diphthong-phonemes is shown in Figs. 54, 61. The dots show the
starting-points, and the arrows show the direction in which the
diphthongs proceed.

382. The positions of the ends of the arrows in Fig. 54 show
the limits of movement of the
diphthongs ei, ou, ai, au, ɔi.
The sound is not heard to be
essentially different if the movement
falls somewhat short of the
limit; in fact, in the pronunciation
of most English speakers,
the limit is not nearly reached.

image front | central | back

Fig. 54. Diagram showing the
nature of the English Diphthongs
ei, ou, ai, au, ɔi.

383. The limit of movement
of the centring diphthongs (,
ɛə, ɔə, ) is ə3 (§ 361). This
limit is usually reached when
the diphthongs are final, as for
instance when the words near niə,
fair fɛə, door dɔə, tour tuə are said by themselves. In non-final
position ə3 is generally not quite reached; thus many speakers
make a slight difference between the diphthongs in near niə (said
alone) and nearly ˈniəli or near together ˈniə təˈgeðə. The difference
is, however, negligible from the point of view of the foreign learner
of English.

384. Some speakers continue the diphthongs and ɛə as far
as ʌ especially in final position. Thus it is not uncommon to
hear near, fair pronounced as niʌ, fɛʌ; the pronunciation ˈniʌli
for ˈniəli is less common and would be considered by many English
people to be an ‘affected’ way of speaking.

385. For the purpose of practical language teaching it is convenient
to regard a diphthong as a succession of two vowels, in
spite of the fact that, strictly speaking, it is a gliding sound.
When diphthongs are described in this rough way, the less prominent
part of a diphthong is commonly said to be ‘consonantal.’
Thus in practical teaching it is convenient to regard the diphthong
ei as a succession of two vowels, e and i, of which the i is
99consonantal; and the diphthong may be easily taught by telling
the pupil to say a certain variety of e followed by i.

385a. Sometimes it is difficult to distinguish a true diphthong
from a sequence of two separately pronounced vowels. For a
discussion of this question see my article Falling and Rising
Diphthongs in Southern English
in Miscellanea Phonetica II, 1954,
published by the I.P.A.

The English Diphthongs in detail

English Diphthong No. 13: ei

386. ei is the so-called ‘long’ sound of the letter a, as in came
keim, make meik. It is also the usual sound of ai and ay; examples:
plain plein, daisy ˈdeizi, day dei, play plei. Ei and ea have the
sound ei in a few words, e.g. weigh wei, veil veil, great greit, break
breik. Note the exceptional words bass (in music) beis, 2173 gauge
geiʤ.

387. The diphthong ei starts at about the English e (Vowel
No. 3) and moves in the direction of i (see Fig. 54).

388. Speakers of Received Southern English do not, however,
all use the same variety of diphthong. In other words the notation
ei stands for a diaphone with several members. Many English
people use a diphthong which begins with a lower variety of e
than this; their varieties of pronunciation may be represented (in
‘comparative’ notation) by ɛi. But if the initial sound of ɛi is
lower than cardinal ɛ, the pronunciation must be considered dialectal
(London and Eastern Counties). There are others who use a variety
of ei beginning with a somewhat closer variety of e than that
shown in Fig. 54. Others again use a very slightly diphthongal
sound which may be symbolized by ɛe; this is a comparatively
recent development, and is becoming very common among RP
speakers.

389. In consequence of the existence of all these variants, the
teaching of ei to foreigners presents no particular difficulty. If
the learner is told to pronounce the English e (Vowel No. 3) with
i (No. 2) immediately after it, the result will be a sufficiently
near approximation to ei.100

390. But the foreign learner must not forget that the Southern
English sound is a diphthong. Foreign people commonly replace
ei by a long ‘pure’ e:, like the vowel in the German See zeː. The
diphthongal nature of the English ei may be well seen by asking
any Southern English person to repeat the sound a number of
times in rapid succession, thus ei-ei-ei… It will be observed
that the lower jaw keeps moving up and down.

391. It should be noticed that the German sound of ee and the
French é are closer than the beginning of the English ei. Their
tongue-positions are in fact higher even than that of cardinal e.

392. Words for practice: pay pei, bathe beid, table ˈteibl, day
dei, scale skeil, game geim, change ʧeinʤ, James ʤeimz, maid,
made meid, neighbour ˈneibə, late leit, railway ˈreilwei, face feis,
veil, vale veil, they ðei, same seim, haste heist, Yale jeil, wake weik,
player ˈpleiə, theyˈre (= they are) ˈðeiə.

392a. The sequence eiə is often replaced by a diphthong ,
thus pleə, ðeə. This diphthong has to be distinguished from ɛə
(§ 446).

English Diphthong No. 14: ou

393. ou is the so-called ‘long’ sound of the letter o; examples:
so sou, home houm, noble ˈnoubl, roll roul, 3174 bolt boult, post poust,
both bouθ, only ˈounli, don't dount. ou is the regular sound of
oa when not followed by r; examples: road roud, toast toust. 4175 Ow
is pronounced ou in many words; examples: know nou, sow (verb)
sou, 5176 growth grouθ. Ou is pronounced ou in the following words:
dough dou, mould mould, moult moult, poultice ˈpoultis, poultry
ˈpoultri, shoulder ˈʃouldə, smoulder ˈsmouldə, soul soul, though ðou.
Note the exceptional words oh ou, brooch brouʧ, sew sou, and words
recently introduced from French such as bureau bjuəˈrou or
ˈbjuərou.

394. The English diphthong ou, as I pronounce it, starts with a
tongue-position in advance of and somewhat lower than that of
101oardinal o (Fig. 54), and a lip-position of medium rounding
(Fig. 55); the speech-organs then move in the direction of u
(Figs. 54, 56).

image

Fig. 55. Lip-position of the
beginning of the English
diphthong ou.

image

Fig. 56. Lip-position of the
end of the English diphthong
ou.

395. The formation of the beginning of the diphthong ou may
be described formally as follows:

(i) height of tongue: a little nearer to ‘half-close’ than to
‘half-open’ (see Fig. 54);

(ii) part of tongue raised: the fore part of the back;

(iii) position of lips: slightly rounded (Fig. 55);

(iv) opening between the jaws: medium.

The tip of the tongue is touching or nearly touching the lower
front teeth, and, as in the case of all ordinary vowels, the soft
palate is in its raised position and the vocal cords are in vibration.
In normal speech the opening between the jaws is not so wide as
for ɔ, ɔː, and ʌ.

396. People often do not realize that the vowel in so, home,
etc., is diphthongal. The fact may be demonstrated by asking
any Southern English person to say Oh! Oh! Oh! … rapidly. It
will be observed that the lips do not remain in one position, but
keep closing and opening.

397. Foreign people generally replace the English diphthong ou
by a pure vowel , such as that heard in the French, cote koːt,
German wohl voːl. This is another sound of the half-close type,
but it has the tongue further back and higher than the English o,
and the lips are very much more rounded than for the English
sound. The differences between it and the English o are summed
up by some writers by describing this foreign sound as ‘tense.’

398. It is important that foreign learners, and particularly
Germans, should remember that in the English o the tongue is not
102in the standard back position, but is advanced towards the central
position (‘narrowly’ ö). This gives to the English o a trace of
œ-quality (§ 338). Many foreigners who recognize the diphthongal
character of the English ou, fail to advance the tongue sufficiently
and so to make the first element enough like œ; the result is that
the diphthong which they produce sounds too much like ɔu. 6177

399. In such cases it is well to start by practising the diphthong
œu. When this diphthong œu is mastered, learners usually do
not have much difficulty in modifying its quality until the true
sound of the English ou is arrived at. French people may obtain
a near approximation to the English diphthong ou by pronouncing
their so-called ‘e mute’ (the usual vowel in le ) followed by the
English u, thus əu. Those of other nationalities can often learn
to make the diphthong by noticing that it is not far removed from
əːu, a sequence of the English vowels Nos. 11 and 8. Most foreign
learners find it helpful to keep the tip of the tongue firmly pressed
against the lower teeth when practising this diphthong.

400. The diphthong ou is particularly difficult for foreign people
when followed by the ‘dark’ l (§ 668) as in old ould, whole houl,
rolls roulz. In practising such words a break should at first be
made, thus ou-ld, hou-l, rou-lz, and then the sounds should be
gradually joined together.

401. Foreign learners should avoid overdoing the diphthongal
character of ou or replacing it by forms like ɑu, au, ʌu, all of which
may be heard in London and other dialects. It is better to use
the continental than these exaggerated forms, is actually
used in Scottish pronunciation.

402. Many foreign people (especially the French, Spaniards,
Italians and Japanese) have extreme difficulty in distinguishing
ou from ɔː. Those who have this difficulty should study carefully
the differences between the two sounds (§§ 305, 394).

403. The English vowel o occasionally appears without a
following u, but only in unstressed syllables or before another
vowel. Such cases are comparatively rare, and there are always
alternative forms with ou or ə. Thus November, obey, molest
are often pronounced noˈvembə, oˈbei, moˈlest, but the forms
103nouˈvembə, nəˈvembə, ouˈbei, əˈbei, mouˈlest, məˈlest are also
common. Again going ˈgouiŋ, slower ˈslouə may be pronounced
goiŋ, sloə with diphthongs oi, (see § 466;r).

404. Some Southern English people use a subsidiary member of
the ou-phoneme when dark l follows. This subsidiary member
starts with a more retracted tongue-position than the ordinary
ou; it is consequently a kind of ɔu. Some English speakers, then,
use a different diphthong in bowl boul, bolt boult from that which
they use in bowling ˈbouliŋ, roll it ˈroul it. It is well for foreign
learners to know of the existence of this distinction, but it is not
advisable for them to try to make it, since an exaggeration of the
retracted pronunciation would become a mispronunciation.

405. Words for practising the diphthong ou: post poust, both
bouθ, tone toun, don't dount, cold kould, go gou, choke ʧouk,
Joseph ˈʤouzif, motion ˈmouʃn, no, know nou, loaf louf, roll roul,
foe fou, vote vout, though ðou, sole, soul soul, zone zoun, show ʃou,
hope houp, yoke, yolk jouk, won't wount.

English Diphthong No. 15: ai

406. The English diphthong denoted here by ai is the so-called
‘long’ sound of the letters i and y; examples: time taim, idle ˈaidl,
night nait, child ʧaild, find faind, fly flai. Ie has the value ai
when final, as in pie pai, and in inflected forms such as tried traid,
cries kraiz. Ei is pronounced ai in the words height hait, sleight
slait, either ˈaiðə, 7178 neither ˈnaiðə, 7179 eider ˈaidə. Exceptionally spelt
words are buy bai, eye ai, choir ˈkwaiə, aisle ail.

407. We have here the case of a diaphone with several members.
That is to say different speakers of Received English do not all
pronounce ai in the same way. In the pronunciation of many
the diphthong begins at Cardinal Vowel No. 4 (a) and immediately
proceeds in the direction of i (Fig. 54). With some speakers the
diphthong begins at a point between Cardinal a and Cardinal ɑ,
while with others the beginning is at a point between a and æ.
Others again, especially in the North of England, use a sound
nearer to ʌi or əi.

408. With each individual speaker ai represents a phoneme
comprising one of the above-mentioned forms as the principal
104member and at least one notable subsidiary member, namely the
variety used when ə follows (see §§ 414, 415).

409. In the practical teaching of English pronunciation to
foreign learners it is generally convenient to take as the normal
value of ai the variety which begins at Cardinal a. The formation
of the beginning of this diphthong may be described formally as
follows:

(i) height of tongue: low;

(ii) part of tongue raised: the front;

(iii) position of lips: spread to neutral (Fig. 57);

(iv) opening between the jaws: rather wide.

image

Fig. 57. Lip-position of the
beginning of the English
diphthong ai

image

Fig. 58. Lip-position of the
end of the English diphthong
ai

The tongue-tip is touching or nearly touching the lower front teeth,
and, as in the case of all ordinary vowels, the soft palate is in its
raised position and the vocal cords are in vibration. The nature
of the diphthong is seen in Fig. 54, and by comparing Figs. 57
and 58.

410. To pronounce the English diphthong ai correctly it is not
necessary that i should actually be reached. A certain portion of
the movement towards i is sufficient to give the proper effect.
In other words, a diphthong of the type ae will suffice. Even as
would not strike an Englishman as wrong, as long as the ɛ-element
is not too open.

411. The English diphthong ai does not present difficulty to
most foreign learners, owing to the fact that the pronunciation
of English people varies (§ 407). Some foreigners, however, start
the diphthong too near to ɑ (Vowel No. 5). The pronunciation
then becomes dialectal. To correct this mistake, the learner should
make the diphthong sound more like æi; he must learn to make
105a sound intermediate between ɑi and æi, by the process described
in §§ 110-114.

412. It should be noted that the sound a occurs in Southern
English only as the first element of the diphthongs ai and au.
In some languages the sound occurs as a pure vowel, e.g. French
la patte la pat. In many types of Northern English a is used
where Southern English has æ.

413. Words for practice: pile pail, bite bait, tie tai, dine dain,
kind kaind, quite kwait, guide gaid, child ʧaild, mine main, nice
nais, like laik, right, rite, wright, write rait, five faiv, vine vain,
thy ðai, side said, resign riˈzain, height hait, while wail.

414. ai sometimes forms a so-called triphthong 8180 with a following
ə, as in fire which is generally transcribed ˈfaiə. In pronouncing
this triphthong, the tongue never really goes near to the i-position;
aɛə more nearly represents the actual pronunciation. The ‘levelling’
of the triphthong is often carried so far that it is replaced by a
diphthong or even simply a lengthened ; thus fire is often
heard as faə or faː (which is distinct from far fɑː), and empire
is often pronounced ˈempaə or ˈempɑː. 9181 This levelling of the
triphthong is particularly common when a consonant follows, as
in fiery ˈfaːri, society səˈsaːti, entirely inˈtaːli, violin vaːˈlin, giant
ʤɑːnt, hire it ˈhaːr it, 10182 etc. (instead of ˈfaiəri, etc.). The English
word wires, usually transcribed ˈwaiəz, is often pronounced in a
manner indistinguishable from the French word Oise waːz.

415. English people do not as a rule use these reduced forms
when the ə is a suffix with a definite meaning. They employ in
this case a clearly disyllabic pronunciation which may be symbolized
phonetically by ai-ə. These speakers distinguish higher ˈhai-ə from
hire ˈhaiə, dyer ˈdai-ə from dire ˈdaiə, etc. It is not essential for
the foreign learner to make these distinctions.

416. There is another special case in which the reduction of
aiə to or does not take place, namely when ‘dark’ l follows, as
trial ˈtraiəl. The tendency here is rather to drop the ə and pronounce
trail. It must be noted, however, that if the word trial is
106immediately followed by a word beginning with a vowel, the ə
must be inserted and the aiə may then be reduced to or ,
‘dark’ l not being used in that case. Thus in the trial ended ðə
ˈtraiəl ˈendid
, trial could not be reduced to trail, but might be
reduced to traəl or traːl.

417. Foreign learners often pronounce aiə without making
allowance for the diminution of movement referred to in § 414.
The result is that in their pronunciation fire, society sound like
ˈfajə, soˈsajeti. Those who have this tendency should aim at
reducing aiə almost to a single long vowel (in words other than
those referred to in §§ 415, 416).

418. Words for practising aiə: piety paiəti, ˈpaəti or ˈpaːti
(distinct from party ˈpaːti), Byron ˈbaiərən, ˈbaərən or ˈbaːrən,
tyrant ˈtaiərənt, ˈtaərənt or ˈtaːrənt, diaphragm ˈdaiəfræm, ˈdaəfræm
or ˈdaːfræm, liable ˈlaiəbl, ˈlaəbl or ˈlaːbl, fiery ˈfaiəri, ˈfaəri or
faːri, violent ˈvaiələnt, ˈvaələnt or ˈvaːlənt, scientific saiən'tifik,
saən'tifik or saːn'tifik, desirable diˈzaiərəbl, diˈzaərəbl or diˈzaːrəbl,
iron ˈaiən, aən or aːn.

English Diphthong No. 16: au

419. The English diphthong written phonetically au is the usual
sound of ou; examples: loud laud, house haus, out aut, bough bau.
It is also a frequent sound of ow; examples: cow kau, town taun,
flower ˈflauə. Eo has the value au in the name MacLeod məˈklaud.

420. au represents a diaphone with several members. In the
pronunciation of some speakers of Received English the diphthong
begins at Cardinal Vowel No. 4 (a) and immediately proceeds in
the direction of u. With many others the diphthong begins at a
more retracted point as shown in Fig. 54, 11183 or even further back
than this. Some, on the other hand, begin the diphthong at a
point slightly higher than a, i.e. at a more æ-like sound. When
the starting-point is near to æ, the pronunciation is dialectal;
æu is one of the London dialectal variants of au.

421. With each individual speaker au represents a phoneme
comprising one of the above-mentioned forms as the principal
member and at least one notable subsidiary member, namely the
variety used when ə follows (§ 430).107

422. Taking the variety of diphthong shown in Fig. 54 as the
normal, we may describe as follows the formation of the vowel
with which the diphthong begins:

(i) height of tongue: low;

(ii) part of tongue raised: the hinder part of the ‘front’;

(iii) position of lips: neutral (Fig. 59);

(iv) opening between the jaws: rather wide.

The tongue-tip is touching or nearly touching the lower front teeth,
and, as in the case of all ordinary vowels, the soft palate is in its
raised position and the vocal cords are in vibration. The nature
of the diphthong is seen in Fig. 54, and by comparing Figs. 59
and 60.

image

Fig. 59. Lip-position of the
beginning of my English
diphthong au.

image

Fig. 60. Lip-position of the
end of my English diphthong
au.

423. To pronounce the English diphthong au correctly it is
not necessary that u should actually be reached. The proper
effect will be given as long as a considerable portion of the movement
towards u is performed. In other words a diphthong of
the type ao will suffice.

424. In the practical teaching of English pronunciation to
foreign people the most suitable variety of au to learn depends
to some extent on the nationality of the learner. Thus, if the
learner is French, it is well to teach an au with a somewhat retracted
a, for the reason given in § 425. If the learner is German, the
variety beginning with Cardinal a should be taught (see § 427).

425. French people generally make the beginning of au too
near to æ. 12184 They may improve the sound by starting from the
French ɑ (the sound of â as in pâte). This may or may not give
108the correct pronunciation; it depends on the variety of ɑ used by
the French speaker. If his ɑ is too retracted to produce a diphthong
near enough to the English au, he must aim at a diphthong
intermediate between ɑu and æu, using the method described in
§§ 110-114.

426. French learners sometimes extend the movement of the
diphthong as far as a very close u. To correct this they should
aim at pronouncing ao.

427. Many foreign people, and especially Germans, mispronounce
the English au by using a diphthong which begins at or near
Cardinal Vowel No. 5 (ɑ). In other words, the initial element
of their diphthong is too retracted. To correct this they should
aim at making a sound intermediate between ɑu and æu, after
the manner indicated in §§ 110-114.

428. Many Germans also extend the movement of the diphthong
as far as a close u. Other Germans, on the other hand, do not
make sufficient movement in the diphthong and pronounce ɑɔ.
The diphthong, as I pronounce it, lies between these two extremes.
Germans may learn an adequate variety of the English diphthong
best by pronouncing in sequence an advanced a and English
short u. 13185

429. Words for practice: pound paund, bough, bow (bend the
body) bau, 14186 town taun, doubt daut, cow kau, gown gaun, mouth
mauθ, now nau, loud laud, row (noise) rau, 15187 wound (past tense
of the verb wind waind) waund, 16188 fowl, foul faul, vow vau, thousand
ˈθauznd, thou ðau, sow (pig) sau, 17189 resound riˈzaund, shout ʃaut,
how hau.

430. In the so-called triphthong 18190 auə the tongue never reaches
the u position, aoə or aɔə more nearly represents the pronunciation
usually heard. The levelling is often carried so far that a
diphthong of the type results. This diphthong is distinguished,
109in the pronunciation of many English people, from the other variety
of which replaces aiə (§ 414). To show this distinction in writing
we may write the reduced auə as a̲ə, putting under the a the mark ˍ
denoting retraction. Thus power which would usually be transcribed
ˈpauə is often pronounced ˈpa̲oə or ˈpa̲ɔə or pa̲ə. The levelling is
sometimes carried so far that the triphthong becomes reduced to a
single long a̲ː. Thus power may often be heard as pa̲ː. This
levelling of the triphthong is especially frequent when a consonant
follows, as in powerful (transcribed ˈpauəfl, but usually pronounced
ˈpaoəfl or ˈpa̲əfl or ˈpa̲ːfl), our own (transcribed auər ˈoun, but
usually pronounced aoər ˈoun or a̲ər ˈoun or a̲ːr ˈoun).

431. The differences between a and and the English ɑ are
so slight that many English people are unable either to distinguish
by ear or to make the intermediate ; either they replace it by
ɑ or they make the reduced forms of aiə and auə identical (saying
them with a variety of a). But with those who do distinguish,
and make the three vowels, the sounds belong to separate phonemes;
in my pronunciation ˈtaːriŋ with the fully front a is the reduced
form of tiring ˈtaiəriŋ, ˈta̲ːriŋ with a retracted is the reduced form
of towering ˈtauəriŋ, and both these words are distinct from tarring
ˈtaːriŋ.

432. Further examples of the retracted a are: dowry ˈdauəri,
ˈda̲əri or ˈda̲ːri (distinct from ˈdaːri, the reduced form of diary
ˈdaiəri), Gower Street ˈgauə striːt, ˈga̲ə striːt, or ˈga̲ː striːt, now-a-days
ˈnauədeiz, ˈna̲ədeiz, or ˈna̲ːdeiz, flowerpot ˈflauəpot, ˈfla̲əpot or
ˈfla̲ːpot, devouring diˈvauəriŋ, diˈva̲əriŋ or diˈva̲ːriŋ.

433. It is not necessary for foreign students to learn to make the
distinction between and a̲ə.

434. Some words which may be said with auə have alternative
pronunciations with a very definite disyllabic au-ə. This way of
pronouncing is usual when the ə is a suffix or belongs to a suffix,
as in plougher ˈplau-ə, allowable əˈlau-əbl, allowance əˈlau-əns. The
same pronunciation may also be heard in a few other words usually
transcribed with auə, e.g. coward ˈkau-əd and the surname Cowan
ˈkau-ən. These words may be said with ordinary auə, but they
do not as a rule have a̲ə and a̲ː as alternatives.

435. There is another special case in which the reduction of auə to
a̲ə or a̲ː does not take place, namely when the triphthong is followed
110by the ‘dark’ l, as in towel ˈtauəl, vowels ˈvauəlz. The tendency
here is to drop the ə and pronounce taul, vaulz. A word such
as towel is never pronounced ta̲ːl, since the l in words where u
precedes is rather ‘dark’ even when the next word in the sentence
begins with a vowel. For instance, English people may often be
heard to pronounce such an expression as the towel isn't in its place
as ðə ˈtaul ˈiznt in its ˈpleis. (Compare the case of words ending
-aiəl, where the form -aːl may be heard when the following word
begins with a vowel, § 416.)

English Diphthong No. 17: ɔi

436. The English diphthong written phonetically ɔi is the
regular sound of oi and oy. Examples: oil ɔil, noise nɔiz, boy bɔi,
employs imˈplɔiz, employer imˈplɔiə, royal ˈrɔiəl or rɔil.

437. The chief member of the ɔi-phoneme is a diphthong
beginning about half-way between the English Vowels Nos. 6
and 7 (ɔ and ɔː) and terminating near to i (see Fig. 54). The
phoneme has subsidiary members, but they do not differ greatly
from the chief member. Thus there certainly is a slight difference
between the diphthong in choice ʧɔis and that in oil ɔil; in choice
the movement of the diphthong is continued nearly as far as the
position of the English short i, while in ɔil the movement does
not reach such a high position. In employer, too, i is not reached;
the actual pronunciation might be written imˈplɔeə. 19191 The so-called
‘triphthong’ oia is, however, never reduced to the same
extent as aiə and auə.

438. Foreign learners may ignore the differences mentioned in
the last paragraph, and may learn the diphthong with sufficient
exactitude by starting with the quality of the English long ɔː
(No. 7) and proceeding immediately to i (No. 2). The diphthong does
not as a rule present any difficulty to them. Some Dutch people
and Germans are apt to finish the diphthong with rounded lips 20192;
111it is easy to correct this as soon as the learner realizes that the
English diphthong ends with the lips spread; if necessary, he must
practise the exercise ɔiɔiɔi… with energetic motion of the lips.

439. Words for practice: point pɔint, boy bɔi, toy tɔi, Doyle
dɔil, coin kɔin, choice ʧɔis, joint ʤɔint, noise nɔiz, loyal ˈlɔiəl,
royal ˈrɔiəl, foil fɔil, voice vɔis, soil sɔil, hoist hɔist.

English Diphthong No. 18:

440. is a ‘falling’ diphthong (§ 223) which starts at about the
position of the English short i and terminates at about ə3 (§ 361).
Its formation is shown in Fig. 61. This diphthong-phoneme may

image central | front | back

Fig. 61. Diagram showing the
nature of the English ‘centring’
diphthongs, , ɛə, ɔə, .

be regarded as consisting of a
single member; there are no
phonemic variants differing to
any marked extent from the
above value. The slight modification
mentioned in § 383 is
negligible from the point of view
of the foreign learner of English.

440a. The ‘falling’ character
of is effected by the use of
‘diminuendo stress.’ This means
that the beginning part of the
diphthong is uttered with stronger
stress than the end part. This stress is felt subjectively by the
speaker; it is not always apparent objectively to the hearer on
account of the greater inherent sonority (carrying power) of ə as
compared with i (§§ 100, 101).

441. There exist diaphonic variants of ; that is to say, there
are English speakers who use a somewhat different diphthong.
The most notable of these variants is one (already noted in § 384) in
which the movement of the diphthong proceeds to a point distinctly
lower than ə3; this form of diphthong may be represented by . A
less common variant begins nearer to long than to short i; it is
heard from those whose pronunciation has been influenced by some
dialect (generally Northern, but also London and Australian).

442. is the usual sound of eer; examples: deer diə, peering
ˈpiəriŋ, steerage ˈstiəriʤ. Ear, ere, eir, ier, ea, ie also have the
112sound in some words; examples: ear , beard biəd, here hiə,
weird wiəd, pierce piəs, fierce fiəs, idea aiˈdiə, Ian iən.

442a. Some of these words have alternative pronunciations with
jəː. For instance, some English people pronounce here and fierce
as hjəː, fjəːs. This does not apply as a rule to words spelt without r,
such as idea, Ian, nor to words in which the is preceded by r
or w, such as career kəˈriə, queer kwiə. But there are some exceptions,
e.g. theatre ˈθiətə or ˈθjəːtə. See Falling and Rising Diphthongs
in Southern English
in Miscellanea Phonetica II.

443. Foreign learners are recommended to ignore the variants
mentioned in §§ 441, 442a and to use always the diphthong described
in § 440. It is easily learnt by treating it in the first instance
as a succession of the two sounds i and ə3.

444. Their most usual fault is to begin the diphthong with
long tense instead of with lax i, besides which they often finish
the diphthong with some variety of r-sound. It is very common
to hear foreign people pronouncing here as hiːər or hiːʀ instead of
hiə. Even if they are able to make lax i, they are generally not
aware that this is the sound to use at the beginning of this diphthong,
if they wish to acquire normal Southern English pronunciation. It
is quite easy for the foreign learner to correct the above mispronunciations
as soon as he has learnt to make lax i and ə3 correctly.

445. Further words for practice: pier piə, beer biə, tier, tear
(of the eyes) tiə, 21193 dear diə, Keir kiə, gear giə, cheer ʧiə, jeer ʤiə,
mere miə, near niə, leer liə, real riəl, fear fiə, veer viə, theatre ˈθiətə,
seer siə, sheer, shear ʃiə, here, hear hiə, year jiə, 22194 weir, we're 23195 wiə,
weary ˈwiəri.

English Diphthong No. 19: ɛə

446. ɛə, as I pronounce it, is a diphthong which starts about
half-way between the English Vowels Nos. 3 and 4 (e and æ)
and terminates at about ə3 (§ 361). The formation is shown in
Fig. 61.113

447. It will be noticed that the starting-point of the diphthong
ɛə is about Cardinal ɛ. The mode of forming this initial part
may therefore be summarized as follows:

image

Fig. 62. Lip-position of the
beginning of my English
diphthong ɛə.

(i) height of tongue: half-open;

(ii) part of tongue raised: the front;

(iii) position of lips: spread to neutral
(Fig. 62);

(iv) opening between the jaws: rather
wide.

The tip of the tongue is touching or nearly touching the lower
front teeth, and, as in the case of all normal vowels, the soft
palate is in its raised position and the vocal cords are in
vibration.

448. The phoneme ɛə may be regarded as consisting of a single
member; there are no phonemic variants differing to any marked
extent from the above value.

449. There exist diaphonic variants among speakers of Received
English. The most noteworthy is æə, which is by no means
uncommon. On the other hand there are some English people
who begin the diphthong with a sound nearer to English e (English
Vowel No. 3) than to æ. Variations in the termination of the
diphthong may also be observed. Thus many speakers finish the
diphthong at a point lower than ə3; this pronunciation may be
represented by the notation ɛʌ.

450. ɛə is the regular sound of the group of letters air; examples:
pair pɛə, fair fɛə, cairn kɛən. It is also the sound of ear and are
in many words; examples: bear bɛə, spare spɛə. Note the exceptional
words there and their which are both pronounced ðɛə, 24196
scarce skɛəs and aeroplane ˈɛərəplein.

451. The manner of teaching ɛə to foreign students depends
considerably on their nationality. Most foreigners make the
mistake of sounding some kind of r at the end of it. This must
not be done if they aim at pronouncing like normal Southern
English people. When this fault has been corrected, others
generally remain. French learners usually begin the diphthong
114correctly but mispronounce the termination by rounding their
lips; and if they have been taught to keep the lips spread at the
end of the diphthong, the final sound is often still incorrect owing
to the tongue being in too high a position.

452. Mispronunciations of a similar kind are often to be observed
with people of other foreign countries — sometimes in cases where
theory would lead us not to expect them, e.g. with some Germans.
In such cases the pronunciation can generally be greatly improved
if the learner will increase the distance between the jaws as the
diphthong proceeds
; in other words, if he will take care that the
jaws shall be wider apart at the end of the diphthong than they
are at the beginning. To pronounce like this often requires considerable
practice on the part of the learner. 25197

453. Another mispronunciation frequently heard from Germans
is to begin the diphthong with a very close e; this e is often
lengthened by them, giving a pronunciation eːə. The best way of
correcting this is to teach the variant pronunciation æə; to do this
he should practise saying æ (English Vowel No. 4) with ə3 or ʌ
immediately after it.

454. Words for practice: air ɛə, pair, pear pɛə, bear bɛə, tear
(verb) tɛə, dare dɛə, care kɛə, chair ʧɛə, fair fɛə, vary ˈvɛəri, there
ðɛə, Sarah ˈsɛəra, share ʃɛə, hare, hair hɛə, wear wɛə.

English Diphthong No. 20: ɔə

455. The diphthong ɔə starts very near to English Vowel No. 7
(ɔː) and proceeds in the direction of ə3 as shown in Fig. 61. The
lip-rounding of the initial part is less close than for ɔː (No. 7).
With many speakers the movement of the diphthong does not
reach as far as ə3; this is especially the case when a voiceless
consonant follows, as in coarse, course kɔəs.

456. The initial part of the diphthong may be described shortly
as follows:

(i) height of tongue: somewhat below half-open;

(ii) part of tongue raised: the back;115

(iii) position of lips: open Up-rounding;

(iv) distance between jaws: medium to wide.

The tip of the tongue is generally, though not necessarily, somewhat
retracted from the lower teeth. As in the case of all normal
vowels, the soft palate is in its raised position and the vocal cords
are in vibration.

457. The diphthong ɔə may be heard in the pronunciation of
many speakers in words written with oar, ore, and in some words
written with our. Examples: coarse kɔəs, score skɔə, four fɔə,
course kɔəs. It may also be heard in the words door dɔə and
floor fɔə.

458. It must be noticed on the other hand that many speakers
of Received English, myself among them, do not use the diphthong
ɔə at all, but replace it always by ɔː (English Vowel No. 7); we
pronounce the above words kɔːs, skɔː, fɔː, kɔːs, dɔː, flɔː. Those who
wish to learn to make the sound ɔə may make a sufficiently near
approximation to it by pronouncing Vowel No. 7 (ɔː) followed by ə3.

459. Further examples of ɔə: pour, pore pɔə or pɔː, 26198 boar, bore
bɔə or bɔː, tore tɔə or tɔː, 27199 core kɔə or kɔː, 28200 more mɔə or mɔː,
Nore nɔə or nɔː, 29201 lore lɔə or lɔː, 30202 roar rɔə or rɔː, 31203 fore fɔə or
fɔː, 32204 soar, sore sɔə or sɔː, 33205 Azores əˈzɔəz or əˈzɔːz, shore ʃɔə or
ʃɔː, 34206 your jɔə or jɔː, 35207 wore wɔə or wɔː. 36208

English Diphthong No. 21:

460. is a diphthong which starts at u (English Vowel No. 8)
and terminates at a sound of the ə-type. The formation of its
usual value is shown in Fig. 61. It is a ‘falling’ diphthong, its
116falling character being effected by means of ‘diminuendo stress.’
This means that the beginning part of the diphthong is uttered
with stronger stress than the end part. This stress is felt subjectively
by the speaker; it is not always apparent objectively to
the hearer on account of the greater sonority (carrying power) of
ə as compared with u (§§ 100, 101).

461. A diphthong of the type is used in two categories of
words, which we may call (a) and (b). Category (a) comprises most
words written with ure 37209 and oor and their derivatives; examples:
sure ʃuə, cure kjuə, endure inˈdjuə, poor puə, moor muə, surely ˈʃuəli,
cured kjuəd, poorer ˈpuərə. It also comprises many words spelt
with ur followed by a vowel; examples: curious ˈkjuərĭəs, duration
djuəˈreiʃn, security siˈkjuəriti. It comprises further some words
spelt with our, such as tour tuə, gourd guəd and bourse buəs and
French names like Lourdes luəd, when pronounced in English
fashion.

461a. Category (b) comprises words spelt with ua, ue or ewe
followed by a consonant letter, the syllable being stressed. Such
are truant truənt, fluency ˈfluənsi, jewel ʤuəl. They all have
variant pronunciations with uːə, e.g. ˈtruːənt, ˈfluːənsi. The
diphthong occurs too as a variant of uːə when the termination
-er is added to words ending in -uː, as in doer duə or ˈduːə, fewer
fjuə or ˈfjuːə, which are formed from do duː, few fjuː.

462. The words of category (a) do not have variant pronunciations
with uːə, but they nearly all have variants with a diphthong
ɔə. Thus sure, curious, tour are pronounced by some as ʃoə,
ˈkjoərĭəs, toə.

463. Many of the words of category (a), and especially the
commoner ones, also have alternative pronunciations with ɔə or
ɔː. For instance, a great many Southern English people pronounce
poor, sure, cure, pure, endure, curious, secure as pɔə, ʃɔə,
kjɔə, pjɔə, etc., or pɔː (like paw), ʃɔː (like Shaw), kjɔː, pjɔː, etc.
Less common words, such as tour, tourist, moor, are generally
said with or (tuə or toə, ˈtuarist or ˈtoərist, muə or moə);
the forms tɔə, ˈtɔərist, mɔə and tɔː, ˈtɔːrist, mɔː also exist, but
are not so frequent as the ɔə and ɔː variants of the commoner
117words poor, sure, etc. The same applies to steward, which may
be heard as stjɔəd and stjɔːd as well as stjuəd and stjoəd. Rarer
words such as gourd and bourse are pronounced with or ;
I do not recall ever hearing these words said with ɔə or ɔː. The
proper names Stewart and Stuart are generally pronounced stjuət
or stjoət. 38210

464. The forms with are generally taught as being the most
‘correct’ for words of category (a), and on the whole they seem
to be the best for foreign pupils to learn. It is, however, necessary
that foreign learners should know of the existence of the forms
with , ɔə and ɔː, since they are used by large numbers of people
whose speech must be regarded as Received English. 39211

465. The most common mispronunciation of by foreign people
is to begin it with an (similar to English Vowel No. 9) instead of
with lax u. They often also add a r-sound at the end of the diphthong
when there is an r in the spelling. Thus it is very common
to hear foreign people pronounce poor as puːər or puːʀ instead of
puə. This should not be done, if they wish to speak with normal
Southern pronunciation. It is quite easy for the foreign learner
to correct such errors as soon as he has learnt to make u (English
Vowel No. 8). All he has to do is to pronounce this u with ə3
immediately after it.

466. The following are some words for practice; those marked *
have variants with , ɔə and ɔː, those marked ‡ appear to have
variants with only, those marked † have variants with uːə; poor
puə,* boor buə,* tour tuə,* doer duə,† gourd guəd,‡ jewel ʤuəl,†
adjure əˈʤuə,‡ moor muə,* bluer bluə,† truer truə,† sure ʃuə,*
your juə* (less common than jɔə and jɔː), pure pjuə,* endure
inˈdjuə,* cure kjuə,* skewer skjuə,‡ Muir mjuə,* newer njuə,†
lure ljuə* or luə,‡ fewer fjuə,† sewer (drain) sjuə.‡

English Diphthong No. 22: ĭə

466a. Resembling the diphthong (No. 18) but yet differing
from it in some respects is another sound which may be represented
by the notation ĭə. It is heard in such words as hideous ˈhidĭəs,
118glorious ˈglɔːrĭəs, happier ˈhæpĭə, easier ˈiːzĭə, area ˈɛərĭə, aquarium
əˈkwɛərĭəm, radius ˈreidĭəs, theoretical θĭəˈretikl, axiomatic æksĭəˈmætik,
realistic rĭəˈlistik (see § 466i), archæological ɑːkĭəˈlɔʤikl,
Neapolitan nĭəˈpɔlitn, Leonora lĭəˈnɔːrə, Cleopatra klĭəˈpɑːtrə.

466b. ĭə (No. 22) is always unstressed, (No. 18) on the other
hand is probably to be considered as always either fully or partially
stressed; it is in its nature that it should be so (see § 440a).

466c. The exact nature of ĭə is difficult to establish. Many
English people ‘feel’ it as a sequence of i and ə forming two syllables.
They may even at times, especially in precise speaking, use a
disyllabic sequence of i and ə in the words here written with ĭə.
This way of pronouncing is shown in some dictionaries; it may be
denoted phonetically by i-ə, thus ˈhidi-əs, ˈiːzi-ə, ˈɛəri-ə, θi-əˈretikl,
etc. 40212 To me, however, the sound generally heard in the words
such as those in § 466a appears to be a ‘rising’ diphthong, i.e. a
diphthong in which the end part has greater prominence than the
beginning part (§ 223). I do not feel this prominence at the end
to be attributable to stress. I find the whole diphthong to be
pronounced evenly with weak stress, and that the latter part owes
its prominence to the greater sonority (carrying power) of a as
compared with i.

466d. In spite of the absence of stress, ĭə taken as a whole has
a noteworthy degree of prominence on account of its length. This
prominence is liable to give the impression of ‘secondary stress’
(i.e. moderately strong force of utterance) to those who have
difficulty in distinguishing prominence by stress from prominence
due to other factors. This ‘secondary stress’ is probably less in
degree than that attributable to (No. 18) in comparable situations.
See next paragraph; also § 440a and Chap. XII.119

466e. The fact that (No. 18) probably always has a primary
or a secondary stress in English, while ĭə (No. 22) is always unstressed,
suggests that the distinction between the two may not
be phonemic. There are, however, objections which render it
difficult to establish this definitely. It may be said, in particular,
that although (No. 18) by its nature must have a certain degree
of stress (§ 440a), yet in such words as reindeer ˈreindiə, nadir
ˈneidiə, emir ˈemiə, sclerosis skliəˈrousis, it has the weakest degree
of stress that it is capable of having
; some people might therefore
feel that it should be considered as unstressed. Again, if it can be
shown that the analysis of ĭə as a diphthong is of doubtful accuracy,
its phonemic status must necessarily be obscure. On the other
hand, allowance must be made for the fact that in some environments
the difference between ĭə and a weakly stressed is difficult
to perceive; this suggests that the two sounds are not likely to
be phonemically different, since as a general rule distinctions
between phonemically different sounds are very easy to hear in
the same phonetic context. The whole question of the nature of
the sounds , ĭə and i-ə and their phonemic status in English
deserves further investigation. 41213

466f. ĭə has a certain resemblance to the sequence , but it
is distinct from this. ĭə is a gradual glide, whereas begins with
a sudden rapid glide. It is noteworthy that many, though not all,
of the English words containing ĭə have alternative pronunciations
with . For instance, some English people pronounce hideous as
ˈhidjəs and easier as ˈiːzjə. Some even use after w, as in colloquial
kəˈloukwjəl (more usually kəˈloukwĭəl). Words like area, in which
ĭə is preceded by r, do not have a variant pronunciation with .
Nor do some of the words in which ĭə immediately precedes a stress,
e.g. Neapolitan. 42214

466g. Examples of weakly stressed (No. 18) are not common.
They are found mainly in compound words, such as reindeer
ˈreindiə, the verb dog-ear ˈdɔgiə, lop-eared ˈlɔpiəd, Shakespeare
120ˈʃeikspiə, Ellesmere ˈelzmiə, Bluebeard ˈbluːbiəd. There are, however,
a few simple words (mostly of foreign origin) in which a weak
occurs. Such are nadir, emir and sclerosis, quoted in § 466e;
also vizier which is pronounced ˈviziə or viˈziə, and some proper
names such as Angier ˈænʤiə, Gambier ˈgæmbiə.

466h. Sometimes the difference between and ĭə is accompanied
by a difference of ‘juncture’ of syllables or a difference in place
of syllable division. For instance, the distinction between corn-ear
(ear of corn) ˈkɔːniə and cornea (of the eye) ˈkɔːnĭə involves not
only the difference iə/ĭə but also a difference in the point of syllable
division: this point is after the n in corn-ear but in the middle of
the n in cornea. This in turn involves a slight difference in the
length of the ɔː-sounds in the two words, the ɔː in corn-ear being
somewhat longer than that in cornea. This does not mean that
the differences in juncture can always be held to account for the
difference between and ĭə, since cases occur where and ĭə are
differentiated in similar environments showing no differences of
juncture.

466i. Most words containing and ĭə have alternative pronunciations.
Many of those containing strongly stressed have
variants with jəː, as already mentioned in § 442a. For instance,
clear is kliə or kljəː, serious is ˈsiərĭəs or ˈsjəːrĭəs. 43215 There exists
a pronunciation (now old-fashioned) of ear as jəː (like year). Some
words with the weakly stressed have other types of variant:
thus nadir has a variant ˈneidə, emir and vizier have variants
eˈmiə, viˈziə. And, as already mentioned in § 466f, many words
containing ĭə have variants with . A few such words have
variants with ; such are frontier ˈfrʌntĭə which is also pronounced
ˈfrʌntiə and ˈfrʌntjə (also ˈfrɔntĭə, ˈfrɔntiə, ˈfrɔntjə) and realistic
which may be heard as rĭəˈlistik and riəˈlistik.

466j. It is probable that many English people do not distinguish
from ĭə at all, at any rate in quick speech. For this reason and
for the reason that with some the distinction between and ĭə is
probably not phonemic (§ 466e), the distinction between the two is
121often ignored in phonetic transcriptions. I now think it on the
whole preferable that the foreign learner should observe the distinction
if he can. 43a216

466k. The following are some further words illustrating the use
of ĭə heavier ˈhevĭə, morphia ˈmɔːfĭə, Victoria vikˈtɔːrĭə, nuclear
ˈnjuːklĭə, audience ˈɔːdĭəns, historian hisˈtɔːrĭən, material məˈtiərĭəl,
variable ˈvɛərĭəbl, Hilliard ˈhilĭəd, requiem ˈrekwĭəm, alienation
eilĭəˈneiʃn, amiability eimĭəˈbiliti, theological θĭəˈlɔʤikl, ideological
aidĭəˈlɔʤikl, rearrange rĭəˈreinʤ. Many of these words have
alternative pronunciations with . Rearrange has an alternative
pronunciation riːəˈreinʤ.

466l. The following are some examples contrasting ĭə with weakly
stressed : reindeer ˈreindiə, windier (more windy) ˈwindĭə; nadir
ˈneidiə, shadier ˈʃeidĭə, Canadian kəˈneidĭən; wheatear (name of a
bird) ˈwiːtiə, meatier (more meaty) and meteor ˈmiːtĭə; vizier ˈviziə,
busier ˈbizĭə; Gambier (surname) ˈgæmbiə, Gambia ˈgæmbĭə. The
words windier, meteor, Gambia, Canadian have alternative pronunciations
with ; so do shadier, meatier and busier, though
perhaps less commonly. The words ending in (No. 18) do not
have alternatives with .

English Diphthong No. 23: ŭə

466m. Resembling the diphthong (No. 21) but yet differing
from it in some respects is a sound which may be represented by
the notation ŭə. It is heard in such words as influence ˈinflŭəns,
incongruous inˈkɔŋgrŭəs, arduous ˈɑːdjŭəs, vacuum ˈvækjŭəm,
valuer ˈvæljŭə, valuable ˈvæljŭəbl, arguer ˈɑːgŭə, Papua ˈpæpjŭə,
usual ˈjuːʒŭəl, puerility pjŭəˈriliti, Juanita ʤŭəˈniːta, 44217 fluorescent
flŭəˈresnt.

466n. ŭə (No. 23) is always unstressed, (No. 21) on the
other hand is probably to be considered as always either fully or
partially stressed; it is in its nature that it should be so (see § 460).

466o. As with ĭə, the nature of ŭə is difficult to establish with
precision. Some English people ‘feel’ it as a sequence of u and ə
forming two syllables, and may even pronounce a disyllabic u-ə
122in precise speaking: thus ˈinflu-əns, ˈaidju-əs, pju-əˈriliti. This
kind of pronunciation is indicated in some dictionaries. Personally,
I regard the ŭə of my pronunciation as a ‘rising’ diphthong (§ 223).
The ə-element is more prominent (§ 223) than the u-element on
account of the greater sonority (carrying power) of ə as compared
with u, and not on account of any increase in stress. 45218

466p. In spite of the absence of stress of ŭə as used in English,
the sound taken as a whole has a considerable degree of prominence
by reason of its length. This prominence may give the impression
of ‘secondary stress’ to those who have difficulty in distinguishing
stress-prominence from prominence caused by other factors. See
Chap. XII.

466q. The fact that (No. 21) probably always has a primary
or a secondary stress, while ŭə (No. 23) is always unstressed,
suggests that the distinction between them may not be phonemic,
though objections similar to those adauced in the case of and ĭə
(§ 466e) render it difficult to prove this as a fact. In particular,
although (No. 21) cannot be unstressed in the ordinary sense
of the term (§ 460), it may nevertheless sometimes have the minimum
stress that it is capable of having; and it is a question whether
this degree of stress should not count as complete lack of stress.
Likewise if there is any doubt concerning the correctness of analysing
ŭə as a diphthong, its phonemic status must necessarily be obscure.
On the other hand, in some phonetic contexts and ŭə are difficult
to distinguish, and this fact supplies evidence in favour of treating
them as members of a single phoneme, since as a rule phonemic
differences are easily heard in one particular context. The available
evidence leads me to think that the distinction between and
ŭə is phonemic with some English people, though not with others.
(The analysis is rendered difficult by the fact that words containing
weakly stressed are few in number and for the most part uncommon
and subject to variant pronunciations. The case is,
123however, illustrated by such words as tenure ˈtenjuə, contour ˈkontuə,
penury ˈpenjuəri, manicure ˈmænikjuə, bureaucracy bjuəˈrɔkrəsi,
neurologist njuəˈrɔləʤist, and the names Amoore ˈeimuə and
Tweedsmuir ˈtwiːdzmjuə, when pronounced as shown here.)

466r. ŭə has a certain resemblance to the sequence , though
it is distinct from it. ŭə is a gradual glide, whereas begins
with a sudden rapid glide. It is to be observed that some words
normally said with ŭə have alternative pronunciations with .
For instance, the name Joshua is pronounced ˈʤɔʃwə at least as
frequently as ˈʤɔʃŭə, and some English people pronounce influence
as ˈinflwəns and usual as ˈjuːʒwəl. A preceding r does not preclude
the use of the variant ; for instance, incongruous may be pronounced
inˈkɔŋgrwəs. The variant may even be heard
occasionally after j, e.g. ˈɑːdjwəs for the more usual ˈɑːdjŭəs.

466s. Most words containing and ŭə have alternative pronunciations
of one sort or another. The words of category (a)
containing (§ 461) generally have variants with or ɔə or ɔː;
this applies not only to stressed syllables as shown in §§ 462-464,
but also to such words as manicure, bureaucracy. Most words
containing ŭə followed by a consonant have other types of variant.
Thus valuable and vacuum are often pronounced ˈvæljubl, ˈvækjum,
and usual has variants ˈjuːʒul and ˈjuːʒl as well as the form ˈjuːʒwəl
mentioned in the last paragraph. A few (very few) of the words
normally said with ŭə have variant pronunciations with uːə, like
the words of category (b) (§ 461a); for instance, fluorescent is pronounced
fluːəˈresnt as well as flŭəˈresnt.

466t. It is probable that many English people do not distinguish
from ŭə at all, at any rate in quick utterance. For this reason,
and for the reason that with some the distinction between and
ŭə is probably not phonemic with many speakers (§§ 466e, 466g),
the distinction between the two is often ignored in phonetic
transcripts. My present view is that the foreign learner had
better observe the distinction if he can.

466u. Examples showing clearly the contrast between and
ŭə are not numerous. The distinction is, however, exemplified
by tenure ˈtenjuə and valuer ˈvæljŭə, contour ˈkɔntuə and Mantua
ˈmæntjŭə or ˈmæntŭə, rest-cure ˈreskjuə (when pronounced without t)
and rescuer ˈreskjŭə.124

English Diphthong No. 24: ŭi

466v. A rising diphthong ŭi is not uncommon. It is always
unstressed, and it occurs in one pronunciation of such words as
valuing ˈvæljŭiŋ), issuing ˈisŭiŋ or ˈiʃŭiŋ), casuist ˈkæzjŭist, ruination
rŭiˈneiʃn. It is always replaceable by the disyllabic sequence u-i,
which is difficult to distinguish from it. It is distinguishable
(subjectively at any rate) from the stressed diphthong ui referred
to in § 327a, but to the hearer the distinction is hardly perceptible.
For these reasons the diphthong ŭi may be ignored by the foreign
learner; it is sufficient that he should use the disyllabic pronunciation,
namely u followed by i.

466w. ŭi is often replaced by uːi when immediately preceding a
stress. Thus ruːiˈneiʃn is an alternative to rŭiˈneiʃn (ruination).

Other Diphthongs

466x. The following diphthongs occur in English in addition to
those described above: oi, ŏi, , §ə, , ŏə. oi, and are
referred to in §§ 403, 392a, 403; examples of them are going goiŋ,
player pleə, slower sloə. ŏi, ŏə are heard in such words as
yellowish ˈjelŏiʃ, follower ˈfɔlŏə. ĕa is very uncommon; it may be
heard in the rare word forayer (person who forays) ˈforĕə. As
already noted in § 378b, these diphthongs are all colloquial
reductions of fuller forms, and are not necessary to an acceptable
pronunciation of English, oi and ŏi are stressed and unstressed
reductions of oui; and ĕə are stressed and unstressed reductions
of eiə; and and ŏə are stressed and unstressed reductions of
ouə. These full forms are quite commonly employed by English
people, and are well suited for use by the foreign learner. Readers
interested in the theory of all these diphthongs are referred for
further particulars and examples to Falling and Rising Diphthongs
in Southern English
in Miscellanea Phonetica II (I.P.A.).125

Chapter XVI
Strong and Weak Forms

467. One of the most striking features of English pronunciation
is the phenomenon known as gradation. By gradation is meant
the existence in many common English words of two or more
pronunciations, a strong form and one or more weak forms. Weak
forms occur only in unstressed positions; strong forms are used
chiefly when the word is stressed, but they also occur in unstressed
positions.

468. A weak form of a word is generally distinguished from
a strong form either by a difference of vowel-sound, or by the
absence of a sound (vowel or consonant), or by the difference in
the length of a vowel.

469. When the forms differ in vowel quality, it is generally
found that the weak form has ə where the strong form has some
other vowel. Examples:

tableau strong form | weak form (or one of the weak forms) | illustration of weak form | the | the man | them | take them | at | at once | are | the boys are ready | of | out of date | for | itˈs for me | should | I should think so | do | nor do you | but | all but one | sir | sir James | a | a house | by | sold by the pound | there | there's no time | and | you and I

1219126

470. In a few cases the vowel of a weak form is i. Examples:

tableau strong form | weak form (or one of the weak forms) | illustration of weak form | the | the order | be | I shan't be long | been | I've been out | me | show me the way | she | she said so | my | I must do my work

2220 3221 4222

471. Usage in regard to these weak forms with i varies considerably
both with individuals and according to the situation of
the word in the sentence. Many speakers pronounce ðiː (with iː
somewhat shortened 5223) instead of ði before vowels, pronouncing the
order
, the end as ðiː ˈɔːdə, ðiː ˈend rather than ði ˈɔːdə, ði ˈend;
this is particularly common when the following vowel is i, as in
the ink ðiː ˈiŋk. Some English people use bin as a strong form of
been; others (apparently an increasing number) use the strong
form biːn in unstressed positions. Thus some would pronounce
I've been there as ai v ˈbin ðɛə instead of the more frequent ai v
ˈbiːn ðɛə
; others would pronounce I've been out as ai v biːn ˈaut
instead of ai v bin ˈaut. Be is generally pronounced bi when
unstressed, except in final position; examples: you'll be late juː l
bi ˈleit
, it ought to be finished it ˈɔːt tə bi ˈfiniʃt, Be sure to come
bi ˈʃuə tə ˈkʌm, but it ought to be it ˈɔːt tə biː. It is also usual to
pronounce biː instead of bi when the following word begins with i
as in I shall be in ai ʃl biː ˈin; ai ʃl bi ˈin is also possible, but
would, I think, be less frequent. Be often occurs in a position
where it may be either unstressed or pronounced with secondary
stress; in such cases the pronunciation with is possible even
when the stress is too slight to mark in a phonetic transcript.
Thus he wants to be a teacher may be pronounced either hiː ˈwɔnts
tə bi ə ˈtiːʧə
or hiː ˈwɔnts tə biː ə ˈtiːʧə. The most convenient
rule for foreign learners is to use bi whenever the word is unstressed
and not final.127

472. The treatment of me, he, she, we is different from that of
be. Weak forms with short lax i are fairly common in the case
of me and she, but they are not regularly used like bi; the use
of the strong forms miː, ʃiː (with vowel shortened owing to lack
of stress) is also quite frequent in unstressed positions. Thus
show me the way may be pronounced either ˈʃou mi ðə ˈwei or
ˈʃou miː ðə ˈwei; she said so, though usually said with ʃi, may
also be said with ʃiː, even when the word is quite unstressed. In
the case of he and we, the strong forms hiː, wiː are commonly
used in unstressed positions; weak forms hi, wi exist but are not
often used. Thus the usual pronunciation of he said so, we thought
so
is hiː ˈsed sou, wiː ˈθɔːt sou. The most convenient rule for
foreign learners is to pronounce unstressed me, he, she, we always
with .

473. The use of a weak form (mi) for my is not by any means
universal; many English people, always use mai in unstressed
as well as in stressed positions. 6224 On the stage in serious drama
it is customary to use mi in all unstressed positions. 7225

474. The short lax u occurs in certain weak forms of to, do,
who and you.

475. The principal weak forms of to are tu and (with ə1, or
ə2, see §§ 356, 358), and the general usage (which incidentally
gives the most practical rule for foreign learners) is to pronounce
tu before vowels and before w and finally, but (with ə1, or ə2)
before consonants other than w. Examples: to eat tu ˈiːt, close
to us
ˈklous tu əs and ˈklous tu ˈʌs, to win tu ˈwin, Where are you
going to?
ˈwɛər ə juː ˈgouiŋ tu, to give tə ˈgiv, going to London ˈgouiŋ
tə ˈlʌndən
.

476. It must be noted, however, that English people do not
always observe this rule consistently. In the first place the unstressed
to before a consonant is sometimes said with a vowel that
128is neither ə1 nor u, but is a sound between them; with some speakers
the sound is ə2, but it is ə2 used as separate phoneme and not
an ə2 dependent upon the adjacent sounds as described in § 358.
This usage may be illustrated by the words amend, the men, to
mend
. Most English people would probably pronounce əˈmend,
ðə ˈmen, tə ˈmend, using ə1 in all three cases; but some would
distinguish tə ˈmend from the other cases by using ə2 in it. (With
such speakers ə2 is a separate phoneme and should be written
with a special sign, say ə̣, in phonetic transcription. 8226)

477. Again, in slow speaking tu may often be heard instead
of as a weak form of to before consonants, but there is no
consistent rule in regard to its use.

478. In ordinary talking, when to is unstressed and followed
by a consonant, the vowel is extremely short and sometimes
almost disappears altogether. It may sometimes be a help to the
foreign learner if this special shortness is marked in transcriptions,
thus ˈgouiŋ tə̆ ˈlʌndən, ˈwel-tə̆-ˈduː (well-to-do). It seems possible
that this extreme shortness may be distinctive, though I have not
met with two sentences distinguished by it. Compare, however,
a later day ə ˈleitə ˈdei with from day to day frəm ˈdei tə̆ ˈdei; the ə
in the first example is distinctly longer than the ə̆ in the second. 9227

479. Do has a strong form duː (and occasionally du before another
vowel) and weak forms du, and d (d being a variant of ). The
weak form du is used before vowels and before w; is used before
consonants, the variant d being commonly used before the word you.
Examples: nor do I ˈnɔː du ˈai, so do we ˈsou du ˈwiː, What do others
think?
ˈwɔt du ˈʌðəz θink; How do they do it? ˈhau də ðei ˈduː it
(or… ˈdu it), Why do people go there? ˈwai də piːpl ˈgou ðɛə, What
do you want?
ˈwɔt də juː ˈwɔnt (more commonly ˈwɔt d juː ˈwɔnt).

480. As with to, the use of ə in the weak form of do before
consonants is not invariable. The weak form may be said
with either ə1 or ə2, and sometimes a sound between ə and u may
be heard. These shades of sound appear to be used by some
129speakers without any apparent consistency. Foreign learners are
recommended to use the following consistent rule: Pronounce
unstressed do as du before vowels, du before w, d before the word
you, and before consonants in all other cases.

481. The strong form duː is used in final unstressed position.
Example: I do (with emphasis on I) ˈai duː.

482. Who has weak forms , hu, u, which may be used in place
of huː in unstressed position. The four forms are interchangeable,
except that , u do not occur initially. See example on p. 135.
Foreign learners are recommended to use always the strong form
huː, which can never be incorrect.

483. You has a weak form ju which may be used in place of
juː in unstressed position. Thus I'll see you to-morrow may be
pronounced either ai l ˈsiː juː təˈmɔrou or ai l ˈsiː ju təˈmorou.

484. The following are the chief words in which one of the
weak forms differs from the strong form by the absence of a vowel:
am, an, and, can, do, for, from, had, has, have, is, ma'am, many,
not, of, Saint, shall, should, some, than, the, them, till, us, will,
would. The forms will be found in the list in § 487, and illustrations
of their use are given in § 488 (p. 133).

485. The following are the chief words in which weak forms
lack a consonant which is present in the strong forms: and, had,
has, have, he, her, him, his, must, of, Saint, shall, them, who, will,
would. See §§ 487, 488.

486. The following are words of which the strong forms are
frequently used in unstressed position (with some shortening of
the vowel owing to the lack of stress): be, been, do, he, her, me,
or, she, we, who, you. All these words also have weak forms with
a different vowel, see §§ 471-483, 487, 488.

487. We now give a list of all the English words which have
weak forms differing notably from the strong forms. Some people
do not use weak forms of the words marked *.

tableau strong forms | weak forms | a | am | an | and130

tableau strong forms | weak forms | are | as | at | be | been | but | *by | can (be able) | could | do (auxiliary) | does (auxiliary) | for | from | had (auxiliary) | has (auxiliary) | have (auxiliary) | he | her | him | his | is | ma'am | *many | me | must | *my | *nor | not | of

10228 11229131

tableau strong forms | weak forms | or | per | Saint | shall | she | should | Sir | *so | some | *such | than | that (conj., relative pron.) | the | them | there | till | *time(s) | to | upon | us | was | we | were | who | will | would

12230 13231 14232132

tableau strong forms | weak forms | you | your

15233

488. The following are examples illustrating the use of important
weak forms, to supplement those already given in preceding
paragraphs.

tableau am | I'm coming | an | an egg, I have got an egg | and | three and eight, two and six, for good and all, bread and butter, cut and dried | are | the boys are here, the boys are over there, the shops are all shut | as | it is as well | at | at home, look at that | but | tired but successful | can | George can stay, we can do it, we can get it | could | he could have told me | does | What does it matter? | for | I did it for fun, out for a walk | from | a long way from London | had | the train had gone, before we had finished, I had got to | has | George has come, he has come today, the fire has gone, it has been very good, the book has disappeared133

tableau have | What have you done?, I have bought a book | her | she had her hat in her hand, he paid her what he owed her, take her out, give her her hat | him, his | Give him his coat | is | he's never there, Alfred is not well, it's all right | *many | How many more? | must | I must go now | *nor | neither he nor I know anything about it | not | are not, does not, it does not matter, he used not to | of | cup of tea, first of all | or | two or three more, one or two others | per | five per cent per annum | Saint | St. Alban, St. Paul's, St. John's Wood | shall | Shall I come with you?, Where shall we go? | should | I should have thought so | sir | Yes, sir, Sir John, Sir Edward

16234 17235 18236 19237 20238134

tableau some | some paper | *such | I never heard of such a thing | than | more than that | that (conj., rel. pron.) | I found that I was wrong, the book that was on the table | them | Take them away | there | there wasn't one, there is no one, if there isn't there ought to be | till | I'm staying there till Tuesday | *time(s) | the first time I went there, three times four are twelve | upon | piled up one upon another | us | Give us one, let us go | was | he was right | were | they were very kind, the books were on the table | who | the man who did it | will | that will do, the church will be full | would | it would be a pity, you would like to | your | Make up your mind

21239 22240135

489. It is to be noticed that the words on, when, and then have
no weak forms; they are pronounced ɔn, wen (or hwen) and ðen
even in the weakest positions. Not has a common weak form nt
after certain verbs (as in can't kɑːnt, did not ˈdidnt). Some English
people use a weak form nət in the single word cannot ˈkænət (more
usually ˈkænɔt). These words are mentioned here because foreign
learners sometimes make mistakes in regard to them.

490. There are other words which have weak forms when they
occur as the second element of certain compound words. Such are:

tableau strong form | weak form | example | berry | gooseberry | board | cupboard | ford | Oxford | land | Scotland | man | gentleman | men | gentlemen | most | topmost | mouth | Plymouth | pan | saucepan | pence | twopence, fivepence | penny | halfpenny | sense | nonsense | shire | Devonshire | where | anywhere near, anywhere else | yard | vineyard

23241 24242 25243

491. The proper use of weak forms is essential for a correct
pronunciation of English, and is one of the most difficult features
136of English pronunciation for foreigners to acquire. Foreign people
generally have an almost irresistible tendency to use strong forms
in their place. Such a pronunciation gives to the English ear the
impression that unimportant words and syllables are receiving
undue prominence, and in consequence the important words and
syllables lose some of the prominence they ought to have.

492. When a weak form contains an ə which may be omitted,
foreign learners will generally do well to use the form without ə.
Thus though away from the city, I should have thought so are
commonly pronounced əˈwei frəm ðə ˈsiti, ai ʃəd əv ˈθɔːt sou, yet
foreigners generally pronounce better by practising əˈwei frm ð
ˈsiti
, ai ʃd v ˈθɔːt sou. This latter pronunciation will strike an
English person as much more natural than the common foreign
pronunciation with strong forms instead of weak ones; the long
successions of consonants arising in such exercises are not so difficult
as they look.

493. The correct use of weak forms is best acquired by continual
reading of phonetic transcriptions. In a few cases there are rules
which help the learner, e.g. the rule as to the use of strong forms
of prepositions given in §§ 996, 997.137

Chapter XVII
The English Plosive Consonants

Detailed Descriptions

494. Plosive consonants are formed by completely closing the
air passage, then compressing the air and suddenly opening the
passage, so that the air escapes making an explosive sound.

495. There are six plosive consonant phonemes in English.
They are represented in phonetic transcription by the letters p,
b, t, d, k, g. The ‘glottal stop’ (ʔ) also occurs, but it is not a
significant sound of the language.

p

496. In pronouncing the principal member of the English
p-phoneme the air passage is completely blocked by closing the
lips and raising the soft palate; the air is compressed by pressure
from the lungs, and when the lips are opened the air suddenly
escapes from the mouth, and in doing so makes an explosive
sound; the vocal cords are not made to vibrate. The formation
of the sound may be expressed shortly by defining it as a voiceless
bilabial plosive
consonant.

497. In Southern English when p is followed by a stressed
vowel as in pardon ˈpɑːdn, payment ˈpeimənt, it is pronounced
with considerable force, and a noticeable puff of breath or ‘aspiration’
(i.e. a slight h) is heard after the explosion of the p and
before the beginning of the vowel. The pronunciation might be
shown thus: phɑːdn, pheimənt. This aspiration is less strong when
the p is preceded by s (e.g. in spider ˈspaidə). Also the aspiration
is less strong when a very short vowel follows, as in picked pikt.
When p is followed by an unstressed vowel, as in upper ˈʌpə, the
aspiration is weak — with some speakers almost imperceptible.
It is not usually necessary to indicate the aspiration of p in practical
phonetic transcription, since the less aspirated varieties are determined
by their situations; they are ‘members of the same phoneme.’
138(For further discussion of aspirated plosives see Theory of Plosive
Consonants, §§ 561-568.)

498. Many Northern English people pronounce p (also t and
k) with little or no aspiration.

499. A subsidiary member of the p-phoneme with nasal plosion
is heard when m or n follows as in topmost ˈtɔpmoust, hypnotize
ˈhipnətaiz. Another subsidiary member with hardly any plosion is
heard when t or k follows, as in wrapped ræpt, napkin ˈnæpkin,
and yet another with no plosion at all before p and b (§§ 579, 582).

500. p is the usual sound of the letter p; example: pipe paip.
P is silent in the initial groups pt, pn, generally also in initial ps;
examples: ptarmigan ˈtɑːmigan, pneumatic njuˈmætik, psalm sɑːm, 1244
also in the single words raspberry ˈrɑːzbri and cupboard ˈkʌbəd.
Note the exceptionally spelt word hiccough ˈhikʌp. 2245

501. Scandinavians and some Germans are apt to aspirate
initial p too strongly, pronouncing ˈpɑːdn as ˈphɑːdn or ˈphɑːdn.
Other Germans, on the contrary, especially South Germans, replace
p by a feebly articulated sound not followed by any h, a consonant
which sounds to an English ear rather like b (phonetic symbol 3246).
Scandinavians also have a tendency to replace p by when it
occurs at the beginning of an unstressed syllable as in upper ˈʌpə,
apple ˈæpl and after s as in spend spend. They should practise
aspirating the p in these cases, and should not take notice of the
diminution of aspiration referred to in § 497.

502. French people, on the other hand, pronounce the consonant
p strongly as in English, but they usually do not insert the aspiration
as Southern English people do (§ 497). They should rather
aim at saying ˈphɑːdn, etc.: they are never likely to exaggerate h
like the Scandinavians and Germans.

503. Words for practising p with ordinary plosion: peel piːl,
pill pil, pencil ˈpensl, patch pæʧ, pass pɑːs, pocket ˈpɔkit, paw pɔː,
pull pul, pool puːl, public ˈpʌblik, purse pəːs, pail peil, post poust,
139pie pai, power ˈpauə, point pɔint, pier piə, pair pɛə, pour pɔə, 4247 poor
puə 5248; capable ˈkeipəbl, happy ˈhæpi, pepper ˈpepə, people ˈpiːpl;
lip lip, map mæp, top tɔp, help help; spin spin, spend spend,
spot spɔt, sport spɔːt, spoon spuːn.

b

504. The principal English b is formed like the principal English
p (§ 496) except that the force of exhalation is weaker and the
vocal cords are made to vibrate (§ 82) so that ‘voice’ is produced
during the articulation of the sound. The formation of the principal
English b may therefore be expressed shortly by defining it as a
voiced bilabial plosive consonant.

505. A subsidiary member of the English b-phoneme with nasal
plosion is used when m or n follows, as in submit səbˈmit, abnormal
æbˈnɔːmal. Other subsidiary members with no plosion occur before
p and b (§ 578), and with hardly any plosion before other plosives,
as for instance in obtain əbˈtein. Yet other subsidiary bˈs with
partial voice are used in initial and final positions, and when a
voiceless consonant precedes, as in Whitby ˈwitbi. Many English
people use a completely devoiced b () in these positions (see
§§ 573,576).

506. As the chief members of the b-phoneme are wholly or
partially voiced, they cannot have ‘aspiration’ in the ordinary
sense of the term.

507. b is the usual sound of the letter b; example: baby ˈbeibi.
B is silent when final and preceded by m as in lamb læm, comb
koum, Coombe, Combe kuːm 6249; also before t in a few words such
as debt det, doubt daut, subtle ˈsʌtl.

508. Many foreign people, and especially Germans, do not voice
b properly, but replace it by (§ 501). This sounds wrong to
English people when voiced sounds precede and follow, as in above
əˈbʌv, table ˈteibl, the boat ðə ˈbout. But the use of does not
matter when a word like bring briŋ is said by itself or is initial in
a sentence, nor when a word like rub rʌb is said by itself or is
140final in a sentence (see § 505). For exercises for acquiring a fully
voiced b see § 177.

509. Spaniards and Portuguese people do not always make the
full contact which is necessary for the proper pronunciation of the
sound b. This is especially the case when the b comes between
two vowels in an unstressed position, as in labour ˈleibə. The
result is that the b is replaced by the bilabial fricative β (§ 692).
Some Germans, especially Bavarians, have a similar tendency.

510. Words for practice: bee biː, bid bid, bed bed, bad bæd or
bæːd, bark bɑːk, box bɔks, bought bɔːt, bull bul, boot buːt, bud bʌd,
burn bəːn, bay bei, boat bout, buy bai, bough bau, boy bɔi, beer biə,
bare bɛə, boar bɔa or bɔː, boor buə; October ɔkˈtoubə, robin ˈrɔbin,
bubble ˈbʌbl; web web, bulb bʌlb, hubbub ˈhʌbʌb, tribe traib.

t

511. In pronouncing the principal member of the English
t-phoneme, the air passage is completely blocked by raising the
soft palate and raising the tip of the tongue to touch the teethridge,
as shown in Fig. 27; the air is compressed by pressure from
the lungs, and when the tongue is removed from the teeth-ridge,
the air suddenly escapes through the mouth, and in doing so makes
an explosive sound; the vocal cords are not made to vibrate. The
formation of the sound may be expressed shortly by defining it
as a voiceless alveolar plosive consonant.

512. In English when t is followed by a vowel in a stressed
syllable, as in taken ˈteikn, it is ‘aspirated’ in the same way as p,
that is to say, a slight h is inserted between the explosion and
the beginning of the following vowel. This articulation may be
regarded as an essential element of the principal member of the
English t-phoneme. A subsidiary member with less aspiration is
used in unstressed positions, as in letter ˈletə, quantity ˈkwɔntiti;
also after s, as in step step, stood stud.

513. Other subsidiary members are (1) a dental t which is used
before θ and ð, as in eighth eitθ, look at this ˈluk ət ˈðis, (2) a postalveolar
t which is used before r, as in rest-room ˈrest-rum, at Rome
ət ˈroum, (3) a nasally exploded t which is used before nasal consonants,
as in mutton ˈmʌtn, that man ˈðæt ˈmæn, (4) a laterally
exploded t which is used before l, as in bottle ˈbɔtl, at last ət ˈlaːst,
141(5) a t without plosion which is used when t, d, ʧ or ʤ follows,
as in that time ˈðæt ˈtaim, not done ˈnɔt dʌn, that church ˈðæt
ˈʧəːʧ
, that gentleman ˈðæt ˈʤentlmən, and with many speakers
before other plosive consonants, as in at present ət ˈpreznt, at
Brighton
ət ˈbraitn, Atkinson ˈætkinsn, it goes it ˈgouz.

514. t is the usual sound of the letter t; example: tent tent.
It is, however, represented by -ed in the past tenses and past
participles of verbs ending in voiceless consonants (other than t);
examples: packed pækt, missed mist, rushed rʌʃt (but compare
waited ˈweitid). 7250 Note also the exceptionally spelt words eighth
eitθ, thyme taim, Thames temz, Thomas ˈtɔməs, Mathilda məˈtildə,
Esther ˈestə. T is silent in words ending in -stle, -sten: castle
ˈkɑːsl, thistle ˈθisl, fasten ˈfɑːsn, hasten ˈheisn, listen ˈlisn 8251; also
in Christmas ˈkrisməs, chestnut ˈʧesnʌt and
many similar words.

image

Fig. 63. Tongue-position
of dental t (variety
with tip of tongue
against lower teeth).

515. Many foreign people, e.g. the French,
Italians, Hungarians, and some Germans,
use a t articulated by the tip of the tongue
against the upper teeth, somewhat as shown
in Fig. 26 (less commonly against the lower
teeth, as shown in Fig. 63). They pronounce
a dental consonant instead of an alveolar
consonant. This articulation produces a
very unnatural effect when used in English,
especially when t is final, as in what wɔt.

516. The difference between the articulation of t in French
and English may be shown by palatograms. Figs. 64, 66 show
palatograms of the English two tuː and the French tout tu. Figs.
65, 67 show palatograms of the English tea tiː and the French type
tip. 9252142

517. Indians generally use a ‘retroflex’ t (phonetic symbol ʈ) in
place of the English t. In pronouncing the retroflex sound the
tongue-tip touches the roof of the mouth further back than for the
English t (Fig, 114, p. 214). Norwegians and Swedes also use ʈ
in some cases (see § 829).

image

Fig. 64. Palatogram
of the English word
two.

image

Fig. 65. Palatogram
of the English word
tea.

image

Fig. 66. Palatogram
of the French
word tout.

image

Fig. 67. Palatogram
of the French
word type.

518. Scandinavians and some Germans are apt to exaggerate
the aspiration of t and to pronounce taken as ˈtheikn (or ˈtheːkn).
There are, however, other Germans who pronounce t very feebly,
and do not insert any h after it; the consonant then sounds to
an English ear like a weak d (phonetic symbol ). These latter
must be careful to pronounce the English t with considerable
force of the breath. Scandinavians have a tendency to replace
t by d when it occurs at the beginning of an unstressed syllable,
as in matter ˈmætə, bottle ˈbɔtl; also after s, as in storm stɔːm. They
should practise aspirating the t in all such cases, and should not
take notice of the diminution of aspiration referred to in § 512.

519. French people on the other hand pronounce the consonant
strongly as in English, but they usually do not insert the aspiration
143properly. The sound they produce is known as ‘unaspirated’ t.
They should therefore rather aim at pronouncing ˈtheikn, thiː
(tea), thuːl (tool), etc.

520. The difficulties experienced by foreign learners in connexion
with the nasally and laterally exploded t's and with the
unexploded t are dealt with in §§ 578-590.

521. Words for practising t with ordinary plosion: tea tiː, tin
tin, tell tel, attack əˈtæk, task tɑːsk, top tɔp, talk tɔːk, took tuk,
two tuː, tumble ˈtʌmbl, turn təːn, take teik, toast toust, time taim,
town taun, toy tɔi, tear (of the eye) tiə, tear (to rend, a rent) tɛə,
tore tɔə or tɔː, tour tuə; writing ˈraitiŋ, water ˈwɔːtə, native ˈneitiv,
theatre ˈθiətə, constitute ˈkɔnstitjuːt, potato pəˈteitou; print print,
profit ˈprɔfit (= prophet), doubt daut.

d

522. The principal member of the English d-phoneme is formed
like the principal English t (§ 511) except that the force of exhalation
is weaker and the vocal cords are made to vibrate so that
‘voice’ is heard. The formation of the principal English d may
therefore be expressed shortly by defining it as a voiced alveolar
plosive
consonant.

523. The chief subsidiary members of the English d-phoneme
are (1) a dental d which is used when θ or ð follows, as in width
widθ, add them ˈæd ðəm, (2) a post-alveolar d which is used before
r, as in he would write hiː d ˈrait, (3) a d with nasal plosion which
is used when m or n follows, as in admire ədˈmaiə, road-mender
ˈroudˌmendə, sudden ˈsʌdn (see § 586), (4) a laterally exploded d,
as in middle ˈmidl (see § 590), (5) a d without plosion which is used
when t, d, ʧ or ʤ follows, as in bed-time ˈbedtaim, good jam ˈgud
ˈʤæm
, (6) partially voiced varieties of d, which are often used in
initial and final positions and when a voiceless consonant precedes,
as in birthday ˈbəːθdei. Many English people use a completely
devoiced d in these positions (see §§ 573, 576).

524. As the chief members of the d-phoneme are wholly or
partially voiced, they cannot have ‘aspiration’ in the ordinary
sense of the term.144

525. d is the regular sound of the letter d; example: deed dild.
Note that final -ed is pronounced -d in the past tenses and past
participles of all verbs ending in vowels or in voiced consonants
(other than d); examples: played pleid, seized siːzd, begged begd. 10253

526. Like t, the English sound d is articulated by the tip of
the tongue against the teeth-ridge (Fig. 27); but many foreign
people, and especially those speaking romance languages, replace
it by a sound made with the tip or blade of the tongue against the
teeth (Figs. 26, 63). This articulation produces an unnatural
effect in English, especially when the d is final as in good gud.

527. The palatograms for d are practically identical with those
for t (Figs. 64, 65, 66, 67).

528. Many foreign people, and especially Germans, do not voice
the sound d properly, but replace it by (§ 518). This sounds
wrong to English people when voiced sounds precede and follow,
as in addition əˈdiʃn, hide it ˈhaid it. But the use of d does not
matter when such a word as do duː is said by itself or is initial
in a sentence, or when such a word as card kɑːd is said by itself
or is final in a sentence. For exercises for acquiring a fully voiced
d, see § 177.

529. Spaniards and Portuguese people are apt to reduce d to
a weak form of the corresponding fricative ð (§ 702), especially
when intervocalic and unstressed, as in ladder lædə.

530. Words for practice: deal diːl, did did, debt det, dash dæʃ,
dark dɑːk, dog dɔg, door dɔː, doom duːm, dust dʌst, dirt dəːt, date
deit, dome doum, dine dain, down daun, dear diə, dare dɛə; hiding
ˈhaidiŋ, louder ˈlaudə, garden ˈgɑːdn, middle ˈmidl 11254; lead (to conduct)
liːd, lead (metal) led, hard hɑːd, load loud, wood wud.145

531. As regards the variety of d known as retroflex d (phonetic
symbol ɖ, see Chap. XXV.

k

532. In pronouncing the common varieties of k the air passage
is completely blocked by raising the back of the tongue to touch
the fore part of the soft palate, the soft palate being at the same
time raised so as to shut off the nose passage (see Fig. 31); the air
is compressed by pressure from the lungs and when the contact of
the tongue with the palate is released by lowering the tongue, the
air suddenly escapes through the mouth and in doing so makes an
explosive sound; the vocal cords are not made to vibrate. The
formation of the principal English k may be expressed shortly by
defining it as a voiceless velar plosive consonant.

533. The English k-phoneme contains several easily distinguishable
members. Firstly, there are variations in the place of tongue-articulation
dependent upon the nature of a following vowel.
Taking the k in come kʌm as the principal member of the phoneme,
we find that a more forward k is used before (as in keep kiːp)
and a more backward k before ɔ (as in cottage ˈkɔtiʤ); other
intermediate sounds are used before other vowels according to
their nature. Secondly, there exist varieties of k with different
degrees of lip-rounding, the most notable being a strongly lip-rounded
k used before w (as in queen kwiːn). The precise sound
used also depends to some extent on preceding vowels. The
k-sound used finally when a consonant precedes (as in ask ɑːsk)
is about the same as that in kʌm, that is to say the principal
member.

534. Thirdly, the amount of ‘aspiration’ of k before a vowel
varies like that of p and t (see §§ 497, 512). Thus the principal
k (as in kʌm) has considerable aspiration, while an unstressed
k with the same tongue-position (as in baker ˈbeikə) and a k
following s (as in sky skai) have less aspiration. The same applies
to the fronter and backer members of the phoneme; compare
kingdom ˈkiŋdəm in which the k has fairly strong aspiration with
talking ˈtɔːkiŋ and skin skin in which the k has less aspiration.

535. Fourthly, k has nasal plosion before nasal consonants (as
in acme ˈækmi, Faulkner ˈfɔːknə, bacon ˈbeikn or ˈbeikŋ). And
146fifthly, a k with little or no plosion is used before other plosives
(see §§ 578, 579, 581, 582).

536. The principal English k pronounced by itself gives no
palatogram on my artificial palate, but forward members of the
phoneme give palatograms. A palatogram
of the word key kiː pronounced by me is
shown in Fig. 68.

image

Fig. 68. Palatogram
of the English word
key.

537. Members of the k-phoneme are
regularly used where the ordinary spelling
has k, and where it has c followed by one of
the letters a, o or u or a consonant letter
or finally; examples: king kiŋ, cat kæt, coat
kout, cut kʌt, fact fækt, electric iˈlektrik. Ch
is pronounced k in some words, e.g. character
ˈkæriktə, chemist ˈkemist, Christmas ˈkrisməs,
ache eik. Qu is generally pronounced kw (e.g. queen kwiːn, quarter
kwɔːtə), but there are a few words in which it is pronounced k
(e.g. conquer ˈkɔŋkə, liquor ˈlikə, antique ænˈtiːk). X is generally
pronounced ks (e.g. box bɔks); for the exceptional cases in which
it is pronounced gz, see § 547.

538. In French the subsidiary k's preceding front vowels are
more forward than in English; in fact with many French people
the contact is so far forward that the sound is the true palatal
consonant c (Fig. 29). The use of such a sound gives to the
ordinary English ear the effect of kj; thus when a Frenchman
pronounces kept as cɛpt, it sounds to English people like kjept.
French people are also liable to use a fronted k before ʌ, ai, and
au, as in cut kʌt, kind kaind, count kaunt. The nature of this
mistake will be realized by comparing Figs. 29, 30, and 31.

539. Some Scandinavians and Germans exaggerate the aspiration
of k, and say ˈkhiŋdəm, ˈkhɔtiʤ, etc. Other Germans on
the contrary, and especially those from Central Germany, are apt to
pronounce the sound very feebly, and not to insert any aspiration
after it; the consonant then sounds to an Englishman like a weak g
(phonetic symbol ). Those who have a tendency to pronounce in
this way must therefore be careful to pronounce the initial k with
considerable force of breath. Scandinavians are also liable to
147replace k by when the sound occurs at the beginning of an
unstressed syllable, as in thicker ˈθikə, pocket ˈpɔkit; also when
preceded by s as in school skuːl. They should practise aspirating
the k in such cases, and should not take notice of the diminution
of aspiration referred to in § 534.

540. French people on the other hand pronounce the consonant
k strongly as in French, but they usually do not insert the aspiration.
They should therefore rather aim at pronouncing ˈkhiŋdəm,
khaind, ˈkhɔtiʤ, etc., with exaggerated aspiration.

541. Words for practice: key kiː, kill kil, kettle ˈketl, cat kæt,
cart kɑːt, collar ˈkɔlə, cushion ˈkuʃin, 12255 cool kuːl, cut kʋt, curl kəːl,
cave keiv, cold kould, kind kaind, cow kau, coil kɔil, Keir kiə, care
kɛə, course kɔəs or koːs; acre ˈeikə, cooking ˈkukiŋ, rocky ˈrɔki; leak
liːk, cake keik, pack pæk, duke djuːk.

g

542. The principal English g is formed exactly like the principal
English k (§ 532) except that the force of exhalation is weaker
and the vocal cords are made to vibrate so that ‘voice’ is heard.
The formation of the principal English g may therefore be expressed
shortly by defining the sound as a voiced velar plosive consonant.

543. The English g-phoneme, like the k-phoneme, has subsidiary
members with places of articulation different from that of the
principal member. Thus if we take the g in govern ˈgʌvən as
the principal member of the English g-phoneme, we find that the
sound in geese giːs has a fronter articulation and the sound in
gɔt got has a backer articulation. There are other varieties of g
with intermediate places of articulation, and their use depends
upon the nature of adjacent vowels. There exist also varieties of
g with different degrees of lip-rounding, the most notable being
a strongly lip-rounded g used before w, as in language ˈlæŋgwiʤ.

544. Other subsidiary members of the g-phoneme with partial
voice or occasionally without voice are used by many speakers
148in initial and final positions and when a voiceless consonant precedes
(see §§ 573, 576).

545. As the chief members of the g-phoneme are wholly or
partially voiced, there cannot be any ‘aspiration’ in the ordinary
sense of the term.

546. A further subsidiary member of the g-phoneme with nasal
plosion is used when m or n follows, as in dogmatic dɔgˈmætik,
Agnes ˈægnis. Other gˈs with little or no plosion occur before
other plosive consonants (see §§ 578, 581, 583).

547. Members of the g-phoneme are the regular sounds of the
letter g when followed by one of the letters a, o or u or a consonant
or when final (as in game geim, go gou, good gud, gum gʌm, green
griːn, big big). The g-phoneme is also used in some words spelt
with ge and gi, for instance get get, give giv, girl gəːil, 13256 finger
fiŋgə. 14257 The x in the prefix ex- is generally pronounced gz when
immediately followed by a stressed or semi-stressed vowel, except
in words beginning with exc-; examples: exact igˈzækt, examine
igˈzæmin, examination igˌzæmiˈneiʃn, exhaust igˈzɔːst, exhibit
igˈzibit (but except ikˈsept, excite ikˈsait); compare exhibition
eksiˈbiʃn, exercise ˈeksəsaiz, in which the vowel following the prefix
is quite unstressed.

548. As in the case of k some French people articulate g too
far forward for English, and sometimes even replace it by the
149voiced palatal plosive ɟ, when a front vowel follows, as in gay gei,
guest gest, gallop ˈgæləp. Some French people do this also before
ʌ, ai and au, as in gun gʌn, guide gaid, gown gaun. (ɟ has the
same tongue-position as c, see Fig. 29.)

549. Many foreign people, and especially Germans, do not voice
the sound g properly, but replace it by (§ 539). This sounds
wrong in English when voiced sounds precede and follow, as in
regard riˈgɑːd, eager ˈiːgə. But the use of g does not matter when
such a word as go gou is said by itself or is initial in a sentence,
or when such a word as jug ʤʌg is said by itself or is final in a
sentence. For exercises for acquiring a fully voiced g, see § 177.

550. Spaniards and Portuguese people often reduce g to the
corresponding fricative sound (phonetic symbol ɤ), especially when
intervocalic and unstressed, as in sugar ˈʃugə. Danes and some
Germans have a similar tendency.

551. Words for practice: geese giːs, give giv, guess ges, gas gæs,
guard gɑːd, got gɔt, gauze gɔːz, good gud, goose guːs, gum gʌm, girl
gəːl, gate geit, goat gout, guide gaid, gown gaun, gear giə; eager
ˈiːgə, tiger ˈtaigə, organ ˈɔːgən, sugar ˈʃugə; big big, egg eg, log lɔg,
mug mʌg.

ʔ

552. The sound commonly known as the ‘glottal stop’ or ‘glottal
catch,’ but more accurately termed the glottal plosive consonant,
is not an essential sound of the English language, but it is necessary
to say a few words about it here.

553. In forming the sound ʔ the glottis is closed completely
by bringing the vocal cords into contact, the air is compressed
by pressure from the lungs, and then the glottis is opened (by
separating the vocal cords) so that the air escapes suddenly. It
is neither breathed nor voiced.

554. An exaggerated form of this consonant constitutes the
explosive sound heard in coughing. Some coughs can be
represented in phonetic transcription. A common kind of cough is
ʔʌhəʔʌh.150

555. The consonant ʔ occurs as an essential sound of many
languages; but in Received English it is not an essential sound,
though it may often be heard incidentally. It sometimes occurs
in Received English when a word which normally begins with a
stressed vowel is specially emphasized. Thus it is absolutely false
spoken with emphasis on absolutely would often be pronounced it s
ˈʔæbsəluːtli ˈfɔls
. The sound ʔ is also often prefixed to initial
vowels when people speak with hesitation.

556. Most foreign people, and more particularly Germans, have
a tendency to insert the sound ʔ at the beginning of all words
which ought to begin with vowels. Thus they will pronounce it was
all our own fault
as it wəz ˈʔɔl ʔauə ˈʔoun ˈfɔːlt. Sometimes they
insert the sound in the middle of a word before a stressed vowel,
saying for instance, wɛəˈʔæz, kriˈʔəit, fiziˈʔɔləʤi instead of wɛəˈræz,
kriˈeit, fiziˈɔləʤi (whereas, create, physiology).

557. It is important that the foreign learner should remedy this
error. The mistake is one which will effectually spoil what is
otherwise a good pronunciation, and it is one which often necessitates
a great deal of practice on the part of the learner. It must be
remembered that in normal English there is no break between
consecutive words which are closely connected by the sense.
The normal English way of pronouncing may often be acquired
by dividing up the sounds into syllables, thus: it wə ˈzɔː lauə
ˈroun ˈfɔːlt
.

558. In phonetic transcriptions the absence of the glottal stop
may be marked, if desired, by ‿ : thus, it wəz‿ˈɔːl‿auər‿ˈroun ˈfɔːlt.

559. Further examples for practice: far away ˈfɑːr‿əˈwei, anywhere
else
ˈeniwɛər‿ˈels, the ends of the earth ði‿ˈendz‿əv ði‿ˈəːθ,
to eat an apple tu‿ˈiːt‿ən‿ˈæpl, all over again ˈɔːl‿ˈouvər‿əˈgein,
not at all ˈnɔt‿əˈtɔːl, to live on an island tə ˈliv‿ɔn‿an‿ˈailənd, he
put on an overcoat
hiː ˈput‿ɔn‿ən‿ˈouvəkout.

560. Very frequently in dialectal English ʔ is used as a
substitute for t in unstressed positions. Some speakers of received
English pronounce like this, especially when m, n, r, j, or w follows,
as in fortnight ˈfɔːʔnait, Tottenham ˈtɔʔnəm, quite right ˈkwaiʔ ˈrait,
not yet ˈnɔʔ ˈjet, that one ˈðæʔ wʌn. It is not necessary for foreign
151learners to adopt this pronunciation, but they should know of its
existence.

Theory of Plosive Consonants

561. To pronounce a complete plosive consonant (§ 183) two
things are necessary: (i) contact must be made by the articulating
organs, (ii) the articulating organs must subsequently be separated.
Thus, in pronouncing an ordinary p the lips must be first closed
and then opened.

562. While the organs articulating a plosive consonant are
actually in contact they form what is termed the stop. In the
case of voiceless consonants, e.g. p, no sound is heard during the
stop; in the case of voiced consonants, e.g. b, some voice (a greater
or less amount according to circumstances, §§ 572 ff.) is heard
during the stop.

563. The explosion of a plosive consonant is formed by the air
as it suddenly escapes at the instant when the stop is released.
The rush of air, however, necessarily continues for an appreciable
time after the contact is released. A plosive consonant therefore
cannot be fully pronounced without being followed by another
independent sound, namely the sound produced by this rush of
air. This independent sound may be either breathed or voiced.

564. When we pronounce a voiceless plosive, e.g. p, “by itself,”
it is generally followed by a short breathed sound which may be
represented by h, thus ph. When we pronounce a voiced plosive,
e.g. b, by itself, it is generally followed by a short vowel, which
may be represented by ə, thus bə.

565. When a voiced plosive consonant is followed by a vowel,
as in the group bɑː, the vowel itself constitutes the necessary
independent sound.

Voiceless plosives

566. It is possible to pronounce a voiceless plosive consonant
followed by a vowel, e.g. the group pɑː, in such a way that the
vowel constitutes the additional sound necessary for the full
pronunciation of the consonant; the effect of this manner of
152pronouncing the sequence is that the vowel-sound begins at the
instant of the explosion of the consonant. It is also possible to
pronounce a voiceless plosive consonant followed by a voiced consonant,
e.g. the group pl, in such a way that the voice begins
at the instant of the plosion. Voiceless plosive consonants pronounced
in such a way that voice begins at the instant of the
plosion are said to be unaspirated.

567. Unaspirated voiceless plosives fall into two classes, viz. those
uttered with considerable force of exhalation, and those in which
the force of exhalation is weak. 16258 The former strike the English
ear as belonging to the p, t, k class; the latter, though voiceless,
strike the English ear as belonging to the b, d, g class. Examples
of the first kind are the French initial p, t, k, as in père pɛːr, tard
tɑːr, cas ka (see §§ 502, 519, 540); examples of the second are the
sounds b, d, g, referred to in §§ 501, 518, 539, which are heard in
many parts of Germany instead of the distinctly voiced b, d, g
of normal North German.

568. In English, initial voiceless plosives are not generally
pronounced in this way, but as already remarked in §§ 497, 512,
534, breath is heard immediately after the plosion. The sounds
are then said to be aspirated. Thus part, pair could be written
‘narrowly’ as phɑːt, phɛə; praise might be written ‘narrowly’ as
transcription preiz. In Denmark and some parts of Germany
aspiration of this kind is so strong that there is practically a
full independent h inserted between p, t, k, and following vowels
(§§ 501, 518, 539).

Ejective sounds

569. It is possible to pronounce consonants of plosive nature
in which the necessary air pressure is produced by some other
means than by the lungs. Sounds in which the air is forced
outwards by these means are called ejective consonants.

570. The most important ejective sounds are those formed by
a closure in the mouth (as for p, t, or k, for instance), keeping
the soft palate raised and closing the glottis. The air in the
153completely enclosed cavity thus formed is slightly compressed,
chiefly through muscular action in the throat causing the larynx
to rise somewhat; when the closure in the mouth is released, the
air therefore escapes with a plosive noise, although the glottis
remains closed. When exaggerated these ejective sounds have a
peculiar hollow quality resembling the sound made in drawing a
cork out of a bottle.

571. These sounds are mentioned here because French people
occasionally use them instead of ordinary voiceless plosives when
final. Such a pronunciation may be corrected by pronouncing a
distinct h after the explosion, e.g. practising the words up ʌp, get
get, look luk, as ʌph, geth, lukh.

Voiced plosives

572. In voiced plosive consonants the amount of voice heard
during the stop may vary. In English and French when a voiced
plosive, e.g. b, occurs between two vowels (as in about əˈbaut),
voice sounds throughout the whole of the stop. Many French
people also pronounce initial voiced plosives in this way, e.g. the
b in bas , the d in doute dut.

573. In English when b, d, and g occur initially, as when bee biː,
day dei, go gou are said by themselves, they are partially devoiced
in the pronunciation of most people, that is to say, voice is not
heard during the whole of the stop but only during part of it,
generally the latter part. With some speakers the voice disappears
altogether, so that the sounds are replaced by , , .

574. In the cases mentioned in the two preceding sections, the
voice of the following vowel begins at the instant of the plosion.

575. Another variety of plosive consonant may be made, in
which the stop is voiced but breath is heard when the contact is
released. Final voiced plosives are often pronounced in this way
in English. This is especially the case when another consonant
precedes, as in bulb bʌlb (narrowly bʌlbh).

576. With many speakers the stop itself is partially or even
completely devoiced in these circumstances. In the latter case
the consonants are very weak voiceless plosive consonants, or sometimes
weak ‘ejective’ sounds (§ 570). These weakened forms of
154final voiced consonants may be represented by , , , without
inconvenience, being very similar in acoustic effect to the sounds
, , , previously described. Thus when bulb is said by itself, it
is generally pronounced bʌlb.

577. In French, final voiced plosives are generally completed
by the addition of a weak central vowel ə, herbe for instance being
pronounced ɛrbə. French people should be careful not to make
this final ə at all strong in saying such English words as globe
gloub, knob nɔb.

Incomplete plosive consonants

578. Sometimes plosive consonants are not fully pronounced.
This happens in normal English when a plosive consonant is
immediately followed by another plosive consonant or by an
affricate. Thus in the usual pronunciation of such words as act
ækt, picture ˈpikʧə, the tongue does not leave the roof of the
mouth in passing from the k to the t or ʧ. There is therefore no
explosion of the k; only the stop of it is pronounced. In Act 2
ˈæktˈtuː there is in normal pronunciation no explosion to the k or
to the first t; the first t is in fact only indicated by a silence.
Similarly in empty ˈempti there is no explosion to the p; its
presence is indicated by a silence. 17259 It is a case of a sound and
a silence belonging to the same phoneme. Similarly in cub-pack
ˈkʌbpæk, begged begd there are no explosions to the b and g; only
the stops of the sounds are pronounced.

579. In that time ˈðæt ˈtaim, red deer ˈred ˈdiə, the first t and
d are not exploded; in fact, the only difference between the tt,
dd here and the t, d in satire ˈsætaiə, red ear ˈred ˈiə, readier ˈredĭə,
is that in the former cases the stop is very much longer than in
the latter. Further instances of the same kind are Iamp-po3t
ˈlæmppoust, book-case ˈbukkeis, cockcrow ˈkɔkkrou.

580. In apt æpt, ebbed ebd the t, d are formed while the lips
are still closed for the p, b. The result is that the p and b do
not have normal plosion, that is to say no h or ə is heard when
the lips are separated. Similarly in suit-case ˈsjuːtkeis the k is
formed during the stop of the t, with the result that little or no
plosion is heard when the t is released.155

581. In iŋk-pot ˈinkpɔt, big boy ˈbig ˈbɔi, the lips are closed for
the p and b during the stop of the k and g. The result is that
no explosion of the k or g is heard.

582. The sequence td in that day ˈðæt ˈdei differs from the d in
muddy ˈmʌdi in having a longer stop, the first part of which is
voiceless. 18260 In ˈðæt ˈdei, midday ˈmiddei (or ˈmidˈdei) the stops
are of the same length, but in the former the first part of the
stop is voiceless and the second part voiced, 18261 while in the latter
the stop is voiced throughout. Further instances of the same kind
are scrap-book ˈskræpbuk, black gown ˈblækˈgaun, soap-bubble
ˈsoupˌbʌbl.

583. The sequence dt in bedtime ˈbedtaim only differs from the
t in better ˈbetə, in having a longer stop, the first part of which
is voiced. In ˈbedtaim, ˈðætˈtaim, the stops are of the same length,
but in the former the first part of the stop is voiced and the second
part voiceless, while in the latter the stop is voiceless throughout.
A further instance of the same kind is egg-cup ˈegkʌp.

584. Many foreign people pronounce all the above sequences of
consonants in an unusual manner, by inserting h or ə between the
consonants. The mistake is particularly noticeable in the sequences
kt, gd. Foreign people usually pronounce act as ækht, picture as
ˈpikhʧə, begged as begəd. The foregoing explanations (§§ 578-583)
should enable them without much difficulty to pronounce such
words as most English people do.

585. Additional examples for practice: picked pikt, wrecked rekt,
locked lɔkt, cooked kukt, worked wəːkt, fogged fɔgd, tugged tʌgd,
exactly igˈzæktli, 19262 expectation ˌkspekˈteiʃn, big dog ˈbig ˈdɔg.

Nasal plosion

586. In sequences consisting of a plosive immediately followed
by a nasal, e.g. the sequence tn in mutton ˈmʌtn, the plosive is not
pronounced in the normal way. The explosion heard in pronouncing
such sequences is not formed by the air escaping through the mouth,
but the mouth closure is retained and the explosion is produced
by the air suddenly escaping through the nose at the instant when
156the soft palate is lowered for forming the nasal consonant. This
kind of plosion is known as nasal plosion.

587. Many foreign people pronounce such English sequences as
tn incorrectly. Thus they often pronounce mutton, topmost, etc., as
ˈmʌtən, ˈtɔphmoust, etc., instead of ˈmʌtn, ˈtɔpmoust, etc.

588. Those who have difficulty may acquire the correct pronunciation
by practising (i) pmpm… and bmbm… without
opening the lips, (ii) tntn… and dndn… without moving the
tip of the tongue, (iii) kŋkŋ… and gŋgŋ… without moving
the back of the tongue.

589. Additional examples for practice: shopman ˈʃɔpmən,
written ˈritn, certain ˈsəːtn, sudden ˈsʌdn, hidden ˈhidn, bacon ˈbeikŋ
(alternative form of ˈbeikən), oatmeal ˈoutmiːl, sharpness ˈʃɑːpnis.

Lateral plosion

590. In the sequences tl, dl, as in little ˈlitl, middle ˈmidl, the
explosion of the t is lateral, that is to say the tip of the tongue
does not leave the teeth-ridge in pronouncing the sequence. Many
foreign people have difficulty in doing this, and consequently
replace tl by təl or something similar (thus ˈlitel, ˈmidəl). The
correct pronunciation of the tl in little may be acquired by practising
the exercises tltltl…, dldldl… with the tip of the tongue kept
firmly pressed against the upper teeth, where it can be seen. In
pronouncing these exercises the tip of the tongue should not move
at all.157

Chapter XVIII
The English Affricate Consonants

591. An ‘affricate’ consonant is a kind of plosive in which the
articulating organs are separated more slowly than usual. In
pronouncing ordinary plosives the separation is made with great
rapidity, and the acoustic effect of the consonant is what might
be called ‘clean-cut’; the plosion itself may be regarded as an
instantaneous noise; if a vowel or an aspiration (h) follows, the
ear cannot detect any intermediate glide between the plosion and
that vowel or aspiration.

592. When, however, the separation of the articulating organs
is performed less rapidly, the ear perceives distinctly the glide
between the plosion and a following sound. The effect of this
glide is essentially the sound of the homorganic fricative consonant,
through the position for which the articulating organs necessarily
pass.

593. There exist degrees of affrication corresponding to the
degrees of rapidity with which the separation of the articulating
organs is performed. If affrication is perceptible but only very
slight, the sound is classified as a plosive. But if there is strong
affrication, such that the effect of the homorganic fricative is
distinctly perceived by the hearer, then the sound is classified as
an affricate.

594. As there exist fricative consonants corresponding to every
plosive, so also there are affricates corresponding to every plosive
(with the exception of ? which has no corresponding fricative).

595. The nature of an affricate may be well seen by articulating
the affricate corresponding to the plosive b. Pronounce the
syllable firstly in the normal manner and then performing the
separation of the lips slowly in such a way that the bi-labial
fricative β1 1263 is heard before the vowel begins; the consonant then
becomes an affricated b. A similar exercise should be tried with
158a dental d. First pronounce the syllable (with a dental d),
then make a similar movement but withdrawing the tongue slowly
from the teeth in such a way that the dental fricative ð 2264 is heard
as a transitory sound before the vowel begins; the consonant then
becomes an affricated dental d.

596. It is convenient as a rule to represent affricates in phonetic
transcription by digraphs consisting of the symbol for the normal
plosive followed by the symbol for the homorganic fricative. Thus
the affricate corresponding to b may be written , and the affricate
corresponding to dental d may be written . Affricates corresponding
to other sounds of the d-type may be written ʣ, ʤ,
, etc. The breathed affricates may be represented in a similar
way thus pϕ, 3265 , ts, ʧ, tɹ̥, 4266 kx, 5267 corresponding to p, dental t,
other varieties of t, k, etc.

597. The fricative glide which finishes an affricate is an essential
part of its pronunciation; this glide gives to each affricate its
distinctive character, and it is never suppressed. 6268 In this respect
affricates differ from plosives (see §§ 578-5æ). Thus the affricate
is pronounced with plosion and off-glide in the past tenses of
verbs ending in ʧ or ʤ, as in attached əˈtæʧt, pledged pleʤd.
But in the past tenses of verbs ending with the plosives p, b, k
or g, e.g. in stopped stɔpt, rubbed rʌbd, cracked krækt, dragged
drægd, these consonants have no plosion. The same thing is
well seen in compound words and in juxtapositions of words.
The affricates ʧ and ʤ are pronounced with plosion and off-glide
in latch-key ˈlæʧkiː, which place ˈwiʧ ˈpleis, lodge-keeper ˈlɔdʒˌkiːpə,
large town ˈlɑːʤ ˈtaun, Bridgetown ˈbriʥtaun, orange juice ˈɔrinʤ
ʤuːs
. (The plosive consonants have no plosion in similar combinations,
e.g. in hat-box ˈhætbɔks, hat-pin ˈhætpin, that place
ˈðæt ˈpleis, neck-tie ˈnek-tai, back door ˈbæk ˈdɔː, cardboard ˈkɑːdbɔːd,
egg-plant ˈegplɑːnt, pug dog ˈpʌg dɔg, Egton ˈegtən, Bridport ˈbridpɔːt.)159

598. For the above reason the representation of affricates by
digraphs is particularly appropriate.

599. At the same time it must be realized that it is sometimes
more convenient to represent affricates by single letters. Such
representation is feasible in writing the many languages which
contain relatively few affricates; it could hardly be applied to a
language containing a number of affricates, owing to the difficulty
of devising a sufficient number of good symbols. When a language
contains only two affricates of the ʧ and ʤ type, the letters c
and ɟ are appropriate for representing them, provided that these
letters are not required for representing palatal plosives in the same
language. 7269

600. In Received English there are six affricates which may be
represented phonetically by the digraphs ʧ, ʤ, ts, ʣ, tr, dr. (tr,
dr are written in place of tɹ̥, , in order to avoid the introduction
of the additional symbols ɹ, ɹ̥).

ʧ

601. In pronouncing the principal member of the English ʧ-phoneme,
the air-passage is completely blocked by raising the
soft palate and raising the tip and blade of the tongue into the

image

Fig. 69. Tongue-position
of the ‘stop’ of the
affricate ʧ.

position shown in Fig. 69, that is to say
a closed position in which the main part
of the tongue is shaped nearly as for ʃ
(Fig. 99); while the ‘stop’ is being held,
air is compressed by pressure from the
lungs; when the tongue is removed from
the teeth-ridge, the air escapes through the
mouth: the removal of the tongue is
performed in such a way that the effect of
the homorganic fricative J is audible before
any following sound is reached (see § 592);
the vocal cords are not made to vibrate.
160The formation of ʧ may be expressed shortly by defining it as a
voiceless palato-alveolar affricate consonant.

602. Those whose languages contain aspirated and unaspirated
plosives regard the English ʧ as aspirated in stressed position,
as in chair ʧɛə, enchant inˈʧɑːnt. This aspiration is, however,
combined with the ʃ-element and is not heard clearly following it.
Nevertheless, the notation ʧʰɛə, inˈʧʰɑːnt may be used when it is
desired to show the aspiration. In unstressed
position there is little or none of this aspiration
in English (e.g. in kitchen ˈkiʧin, lecture ˈlekʧə).

image

Fig. 70. Palatogram
of the English ʧ in
the syllable ʧɑː.

603. ʧ really stands for a diaphone; that is
to say the sound varies to some extent with
different speakers. In particular there is variation
in lip-articulation. With some (probably
the majority) the tongue-articulation is accompanied
by protrusion of the lips as for ʃ
(Fig. 101), while with others the lips are spread.
Slight variations may also be observed in the
position of the tip of the tongue.

604. Fig. 70 shows a palatogram of the English ʧ. It should
be compared with the palatograms of English t (Figs. 64, 65),
ts (Fig. 72) and tr (Fig. 73).

605. ʧ is the usual English sound of ch and tch, as in chain
ʧein, choose ʧuːz, orchard ˈɔːʧəd, watch wɔʧ, wretched ˈreʧid.
It is also the usual sound of t in unstressed -ture, as in furniture
fəːniʧə, nature ˈneiʧə. 8270 Ti has the value of ʧ when the termination
-tion is preceded by s, as in question ˈkwesʧən, combustion
kəmˈbʌsʧən. Te is pronounced ʧ in righteous ˈraiʧəs, but not in
other words ending in -teous.

606. Most foreign learners do not experience difficulty in pronouncing
an adequate ʧ. Danes, however, are apt to substitute
tj for it, and make choose sound too much like tjuːz; to remedy
this error it is first necessary to learn to make a good ʃ by the
method given in § 735; then a good ʧ may be acquired by prefixing
to ʃ the appropriate variety of t. The sound must, if necessary,
161be somewhat exaggerated by articulating with the tip of the tongue
a little too far back, and care must be taken to round and protrude
the lips well as shown in Fig. 101.

607. Words for practice: cheap ʧiːp, chin ʧin, check ʧek, chap
ʧæp, charm ʧɑːm, chop ʧɔp, chalk ʧɔːk, choose ʧuːz, chum ʧʌm,
church ʧəːʧ, picture ˈpikʧə, chain ʧein, choke ʧouk, child ʧaild,
choice ʧɔis, cheer ʧiə, chair ʧɛə; each iːʧ, ditch diʧ, sketch skeʧ,
porch poːʧ, much mʌʧ, birch bəːʧ, H eiʧ, broach, brooch brouʧ,
couch kauʧ.

608. The affricate ʧ must be distinguished from the sequence
t + ʃ which also occurs in English. When it is desired in phonetic
transcription to show that this sequence is used, a hyphen must
be placed between the t and the ʃ. Examples of the group t + ʃ
are courtship ˈkɔːt-ʃip, nutshell ˈnʌt-ʃel, light-ship ˈlait-ʃip, Dorsetshire
ˈdɔːsit-ʃə.

ʤ

609. The principal member of the English ʤ-phoneme is formed
like ʧ except that the vocal cords are made to vibrate so that
‘voice’ is produced during the articulation of the sound. The
formation of the sound may therefore be expressed shortly by
defining it as a voiced palato-alveolar affricate consonant.

610. Being a voiced sound it cannot have ‘aspiration’ in the
ordinary sense of the term.

611. The ʤ-phoneme has subsidiary members with partial voice
which are used in initial and final positions (as when generally
ˈʤenrəli is initial or when bridge briʤ is final), and when a breathed
consonant precedes (as in gas-jet ˈgæsʤet). Many English people
use a completely voiceless ʤ in these situations.

612. ʤ is subject to diaphonic variations similar to those
mentioned in § 603.

613. ʤ is the usual English sound of j, and the usual sound
of g before e, i, and y 9271; examples: jump ʤʌmp, jaw ʤɔː, jet ʤet,
gem ʤem, giant ˈʤaiənt, page peiʤ, pigeon ˈpiʤin, religion riˈliʤən,
162gymnastic ʤimˈnæstik; dg has this sound in edge , judgment
ˈʤʌʤmənt, etc. Note also the miscellaneous words grandeur
ˈgrænʤə, soldier ˈsoulʤə, Greenwich ˈgriniʤ, Norwich ˈnɔriʤ
sandwich ˈsænwiʤ. 10272

614. Most foreign people, except Danes and South Germans,
pronounce ʤ sufficiently well without difficulty. Danes are apt
to replace it by dj and make June ʤuːn sound too much like dune
djuːn; to correct this error, learn ʒ first (§ 742) and then prefix
the appropriate variety of d, taking care to articulate with the
tip of the tongue against the teeth-ridge (not too far forward)
and to protrude the lips as shown in Fig. 101.

615. South Germans are liable to use ʤ, a voiceless sound
resembling ʧ, in place of ʤ. This sounds wrong to English people
when voiced sounds precede and follow, as in engaging inˈgeiʤiŋ),
adjoin əˈʤɔin. But the use of ʤ does not matter when such a
word as join ʤɔin is said by itself or is initial in a sentence, or
when such a word as edge is said by itself or is final in a sentence.
Foreign learners who have difficulty in giving sufficient voice to
ʤ should practise exercises of the types recommended in §§ 177,
792.

616. Words for practice: Jean ʤiːn, jig ʤig, gem ʤem, Jack
ʤæk, jar ʤɑː, job ʤɔb, jaw ʤɔː, June ʤuːn, just ʤʌst, journey
ˈʤəːni, injure ˈinʤə, James ʤeimz, joke ʤouk, gibe ʤaib, joy
ʤɔi, jeer ʤiə; bridge briʤ, large lɑːʤ, George ʤɔːʤ, age eiʤ.

ts

617. The sequence of letters ts is used in more than one sense in
transcribing English. There exists an affricate ʦ, but it is a rare
sound and only occurs in words and names borrowed from foreign
languages such as tsetse ˈʦetsi (first ts), 11273 Tsana ˈʦɑːnə, Tsushima
ˈʦuːʃimə. There exist also sequences consisting of t followed by s;
163in some cases, e.g. in outside ˈautˈsaid, outset ˈauʦet, the two sounds
are clearly separated, while in other cases, e.g. in cats kæʦ, curtsey
ˈkəːtsi, they are pronounced together in more or less intimate
combination. Thus though the ts of curtsey is not a true affricate,
it is more like an affricate than the ʦ of outside is. In ˈʦetsi the
first ʦ is a true affricate, but the second is like the ts in curtsey.

image

Fig. 71. Tongue-position
of the ˈstopˈ of the
affricate ʦ.

image

Fig. 72. Palatogram
of the affricate ʦ in
the syllable ʦɑː.

618. The affricate ʦ is formed by placing the main part of the
tongue as for s (§ 709) and bringing the blade to touch the teethridge
as shown in Fig. 71; air is compressed by pressure from the
lungs, and then the tongue is removed not too rapidly from the
teeth-ridge; at the beginning of the separation there is a plosion,
and as the separation proceeds, the effect of a short s is audible;
the vocal cords are held apart during the production of the sound, so
that no ‘voice’ is present. The formation of the sound may be
expressed shortly by defining it as a voiceless blade-alveolar affricate
consonant.

619. Fig. 72 shows a palatogram of the affricate ʦ. It should
be compared with the palatograms of English t (Figs. 64, 65), ʧ
(Fig. 70) and tr (Fig. 73).

620. As the true affricate ʦ is so uncommon in English, and as
it appears to occur exclusively in initial position, it is not necessary
for practical purposes to have a special symbol to distinguish it
from the group t + s. When it is desired to show specially that t
and s are separately pronounced, a hyphen may be introduced,
thus ouʦet ˈaut-set.164

ʣ

621. The affricate ʣ is formed like the affricate ʦ except that
its articulation is accompanied by vibration of the vocal cords so
that ‘voice’ is heard. The formation of the sound may therefore
be expressed shortly by describing it as a voiced blade-alveolar
affricate
consonant.

622. The sound is very uncommon in English. It only occurs in
foreign proper names beginning with ʣ, e.g. Dzungaria ʣʌŋˈgɛərĭə.

623. There also occurs in English the sequence d + z, in which
the two sounds are in fairly intimate combination, though not
sufficiently intimate as to constitute a true affricate. This sequence
ʣ is common in final position, as in reads riːʣ, fields fiːlʣ, woods
wuʣ; in this position it is often partially or completely devoiced
(see §§ 722, 788, 789). In medial position ʣ is rare, and occurs
mainly in compound words, such as bird's-eye ˈbəːʣai, and in
borrowed foreign words, such as piazza piˈæʣə.

tr

624. As already mentioned in § 617, it is often difficult to draw
a line of demarcation between an affricate and an intimate combination
of two sounds. It is, however, probably correct to class the
Southern English tr (as in tree triː) as an affricate. The sound tr is
formed by placing the main part of the tongue as for r (§ 747) and
bringing the tip of the tongue to touch the back part of the teethridge
very much as shown in Fig. 28; air is compressed by pressure
from the lungs, and then the tongue is removed not too rapidly from
the teeth-ridge; at the beginning of the separation there is a plosion,
and as the separation proceeds the effect of ɹ̥ (breathed fricative r)
is audible; the vocal cords are held apart during the production of
the sound, so that no ‘voice’ is present. The formation of the sound
may be expressed shortly by defining it as a voiceless post-alveolar
affricate
consonant.

625. If the separation of the tongue were performed very rapidly,
the corresponding plosive would be produced. This plosive is one
variety of Indian retroflex (cerebral) ʈ ().165

626. Fig. 73 shows a palatogram of the affricate tr. It should
be compared with the palatograms of English t (Figs. 64, 65), ʧ
(Fig. 70), and ʦ (Fig. 72).

image

Fig. 73. Palatogram
of the affricate tr in
the syllable trɑː.

627. The affricate tr is the usual sound of
tr in English; examples: tree triː, straight streit,
entrance ˈentrəns.

628. Foreign learners who have difficulty in
pronouncing the English r have also difficulty
with tr. There are two methods of acquiring
tr. One is to learn r first, by the methods
suggested in § 766, and then prefix to it the
appropriate variety of t. The other is to start
from ʧ and try to pronounce it with the jaws
widely separated. Place two fingers one above
the other between the teeth, and try to say ʧ; the resulting sound
approximates very nearly to the English tr. This exercise may be
done still better with a cork about an inch in diameter; the
endeavour to pronounce chain tʃein with the cork held between
the teeth produces a syllable almost identical with train trein.

629. Words for practice: tree triː, trick trik, treasure ˈtreʒə,
travel ˈtrævl, trance trɑːns, trot trɔt, trawler ˈtrɔːlə, true truː, trust
trʌst, tradition trəˈdiʃn, train trein, trophy ˈtroufi, try trai, trout
traut, Troy trɔi, matron ˈmeitrən, poultry ˈpoultri, symmetry ˈsimitri,
actress ˈæktris.

630. The affricate tr must be distinguished from the sequence
t + r which also occurs in English. When it is desired in phonetic
transcription to show that this sequence is used, a hyphen must
be placed between the t and the r, unless there is a stress-mark
separating the letters. Examples of the sequence t + r are restroom
ˈrest-rum, outrageous autˈreiʤəs.

dr

631. The affricate dr is formed like tr except that the vocal
cords are made to vibrate so that ‘voice’ is produced during the
articulation of the sound. The formation of the sound may therefore
be expressed shortly by defining it as a voiced post-alveolar
affricate
consonant.166

632. If the separation of the tongue-tip from the roof of the
mouth is performed very rapidly, the sound produced is no longer
an affricate but is the corresponding plosive. This plosive is one
variety of Indian retroflex (cerebral) ɖ ().

633. The affricate dr is the usual sound of dr in English; examples:
dream driːm, draw drɔː, hundred ˈhʌndred. It occasionally occurs
finally in words borrowed from French, such as cadre kɑːdr; in
these cases the dr is often partially or completely devoiced.

634. Foreign learners who have difficulty in pronouncing the
English r have also difficulty with dr. dr may be acquired by
methods similar to those recommended for tr (§ 628). Jaw ʤɔː,
jug ʤʌg pronounced with a large cork held between the teeth
become practically draw drɔː, drug drʌg.

635. Words for practice: dream driːm, drip drip, dread dred,
drag dræg, draft, draught draːft, drop drɔp, draw drɔː, drew druː,
drum drʌm, dramatic drəˈmætik, draper ˈdreipə, drove drouv, dry
drai, drought draut, Droitwich ˈdrɔit-wiʧ, dreary ˈdriəri, drawer 12274
drɔə or (more commonly) drɔː, Drury ˈdruəri; laundry ˈlɔːndri,
Andrew ˈændruː, bedroom ˈbedrum, 13275 kindred ˈkindrid.

636. The affricate dr must be distinguished from the sequence
d + r which also occurs in English. When it is desired in phonetic
transcription to show that this sequence is used, a hyphen must
be placed between the d and the r, unless there is a stress-mark
separating the letters. Examples of the sequence d + r are headrest
ˈhed-rest, head-room (room for one's head) ˈhed-rum, 13276 bloodred
ˈblʌdˈred or ˈblʌd-red, hand-writing ˈhænd-raitiŋ). 13277167

Chapter XIX
The English Nasal Consonants

637. Nasal consonants are formed by closing the mouth-passage
completely at some point, the soft palate being held in its
lowered position so that the air is free to pass out through the
nose.

638. There are three nasal consonant phonemes in English.
They are represented phonetically by the letters m, n, I).

m

639. The principal member of the English m-phoneme is formed
as follows. The mouth-passage is completely blocked by closing
the lips; the soft palate is lowered so that, when air is emitted by
pressure from the lungs, it passes out through the nose; the tongue
is held in a neutral position; the vocal cords are made to vibrate
so that ‘voice’ is produced. The formation of the sound may be
expressed shortly by defining it as a voiced bi-labial nasal consonant.

640. When a vowel follows, the position of the tongue during
the production of m approximates to the position required for
that vowel. To this extent, therefore, it may be said that there
are subsidiary members of the phoneme. With most speakers,
however, these differences of tongue-position are slight and their
effects on the acoustic quality of the sound are negligible.

641. A labio-dental nasal (ɱ) is used by some speakers as a subsidiary
member of the m-phoneme when f or v follows, as in
triumph ˈtraiəmf, comfort kʌmfət, Dumville ˈdʌmvil, information
imfəˈmeiʃn (a variant of infəˈmeiʃn). This subsidiary sound is
used chiefly by those whose upper teeth project considerably beyond
the lower teeth.

642. A partially devoiced m (phonetically ) sometimes occurs
as a subsidiary member of the m-phoneme when s precedes in the
same syllable, as in small smɔːl (see § 845 (i) a).168

643. m is the regular sound of the letter m; examples: make
meik, come kʌm. M is, however, silent in initial mn-, as in
mnemonic niːˈmɔnik.

n

644. The principal member of the English n-phoneme is formed
as follows. The mouth-passage is completely blocked by raising
the tip of the tongue to touch the teeth-ridge as shown in Fig. 74;
the soft palate is lowered so that, when air is emitted by pressure
from the lungs, it passes out through the nose; the vocal cords
are made to vibrate so that ‘voice’ is produced. This formation
may be expressed shortly by defining the sound as a voiced alveolar
nasal
consonant.

645. Subsidiary members of the English n-phoneme exist, and
notably an advanced (dental) variety which is used when θ or ð
follows (as in enthusiasm inˈθjuːziæzm, in there ˈin ˈðɛə) and a
retracted variety used before r (as in enrol inˈroul). Practically
these varieties are of no importance, since they are acoustically
almost indistinguishable from the principal n. A partially devoiced
n also occurs as a subsidiary member of
the phoneme. It is used when s precedes
in the same syllable (as in sneeze sniːz,
see § 845 (i) a).

image

Fig. 74. Position of
Tongue and Soft Palate
for English n.

646. Some foreign people, chiefly
speakers of Romance languages, regularly
use a dental n, i.e. a n articulated by the
tip of the tongue against the upper teeth.
The difference of sound is unimportant
except in final position, where the use of
dental n gives an unnatural effect to English
ears. These foreign people have no difficulty
in pronouncing such words as own
oun, done dʌn with an alveolar n, when once the formation has
been explained to them.

647. Some Germans use a slightly palatalized n differing from
the usual English n in the same way as the German l does from
the English final l (see Chap. XX). The correct English n has a
169duller quality than this German variety of n. The “clear” quality
of this palatalized variety is often strengthened by lip-spreading.
The effect of the sound is strange to English ears when final or
followed by a consonant, and especially when preceded by a back
vowel, e.g. in pond pɔnd, soon suːn. The correct English n presents
no great difficulty after the English final l has been acquired
(§§ 670-672). Note that lip-spreading should be avoided in
pronouncing the English n, and that if a back vowel precedes,
as in pɔnd, suːn, it is well to maintain the lip position of the back
vowel until the completion of the n.

648. The palatograms of the various kinds of n are similar to
those of t (Figs. 64, 65, 66, 67).

649. French learners have to be told that -gn- is pronounced
with g followed by n in English. Those who do not know this
follow French usage, and use a palatal nasal consonant (ɲ) in
such words as ignorance ˈignərəns.

649a. n is the usual sound of the letter n; examples: nine nain,
linen ˈlinin.

ŋ

650. The principal member of the English ŋ-phoneme is formed
as follows. The mouth-passage is completely blocked by raising
the back of the tongue to touch the fore part of the soft palate as
shown in Fig. 75; the soft palate is in its lowered position, so that
when air is emitted by pressure from the lungs it issues through
the nose; the vocal cords are made to vibrate, so that Voiceˈ is
produced. The formation of this q may be expressed shortly by
defining it as a voiced velar nasal consonant.

651. Varieties of ŋ with fronter and backer tongue-articulation
occur as subsidiary members of the phoneme. Their use is determined
by the nature of the adjacent vowels. Thus the principal ŋ
is used after ʌ, as in young jʌŋ trunk trʌŋk, and when ɔ precedes
and i follows, as in belonging biˈlɔŋiŋ (first ŋ); a backer variety
of ŋ is used after ɔ finally, as in long lɔŋ; and varieties of different
degrees of advancement are used after the front vowels, the frontest
occurring after i, as in sing siŋ, bringing ˈbriŋiŋ. These differences
of articulation can easily be felt, but they are of no practical
importance, because the acoustic differences are hardly appreciable.170

652. The ŋ-phoneme is represented in spelling by final ng, as
in king kiŋ, and very often by n before letters representing k and
g sounds, as in ink iŋk, anchor ˈæŋkə, finger ˈfiŋgə, strongest ˈstrɔŋgist.

653. In regard to the pronunciation of the sequence of letters
ng when medial, it is to be noted that (i) ŋ alone is used in words
formed from verbs by the addition of the suffixes -er and -ing,
e.g. singer ˈsiŋə, hanging ˈhæŋiŋ; (ii) the prefix con- when followed
by the sounds k or g, is pronounced by most people with ŋ when
the following syllable is quite unstressed, but with n when the
following syllable has stress (primary or secondary); thus, congress
ˈkɔŋgres, congregation ˌkɔŋgriˈgeiʃn have ŋ, while concur kənˈkəː,
congratulation kənˌgrætjuˈleiʃn have n; (iii) the prefixes en-, in-,
un- are pronounced with n by most speakers of Received English:
thus engage inˈgeiʤ, ingredient inˈgriːdĭənt, ungrateful ˈʌnˈgreitfl
have n. These latter prefixes are also generally pronounced with
n when k follows, as in encourage inˈkʌridʒ, increase (noun) ˈinkriːs,
increase (verb) inˈkriːs, uncomfortable ʌnˈkʌmfətəbl. There is, however,
a tendency at the present day to use ŋ in place of n in cases
(ii) and (iii).

image

Fig. 75. Tongue-position
of Cardinal ŋ.

image

Fig. 76. Tongue-position
of Cardinal ɲ.

654. The principal English ŋ gives no palatogram on an ordinary
artificial palate, since no part of the contact is against the hard
palate. The subsidiary ŋ used after i gives the palatogram shown
in Fig. 77.

655. The sound ŋ is often pronounced incorrectly by French
people. They have a tendency to replace it by the palatal nasal
171ɲ, especially when a front vowel precedes. The difference between
ŋ and ɲ will be seen from Figs. 75, 76.

656. ɲ is the ordinary French ‘n mouillé,’ as in montagne
mɔ̃taɲ. French people have to remember that for the English ŋ
the contact of the tongue with the palate is much further back
than for the French ɲ. It is often useful for them to practise the
sound ŋ with the mouth very wide open. 1278

image

Fig. 77. Palatogram
of the ‘advanced’ ŋ in
the English sequence
-iŋ (my pronunciation).

image

Fig. 78. Palatogram
of French ɲ in the
sequence aɲa.

657. Some Germans have a tendency to replace final ŋ by the
sequence ŋk, thus confusing for instance sing siŋ and sink siŋk.
This defect may be remedied by pronouncing final ŋ very long, thus
siŋː. It should be observed that the substitution of ŋk for ŋ in
nothing, something, anything is found in London dialect (Cockney)
but is not considered a desirable pronunciation.

658. Words for practice: bring briŋ, sang sæŋ, long lɔŋ, rung
rʌŋ; longing ˈlɔŋiŋ, singer ˈsiŋə; longest ˈlɔŋgist, anger ˈæŋgə, anchor
ˈæŋkə, younger ˈjʌŋgə, handkerchief ˈhæŋkətʃif.172

Chapter XX
The English Lateral Consonants

659. The lateral consonants occurring in English are represented
phonetically by the letter l. Several varieties occur in Received
Southern English, but for practical purposes it is sufficient to
distinguish two. These are known as ‘clear’ l and ‘dark’ l. They
are members of the same phoneme, the principle governing their
use being that clear l occurs only before vowels and before j, while
dark l is only used before all other consonants and finally. Thus
‘clear’ l is used in leave liːv, lake leik, along əˈlɔŋ, million ˈmiljən,
while ‘dark’ l is used in feel fiːl, field fiːld, people ˈpiːpl. 1279

660. Both these consonants are primarily articulated by the
tip of the tongue touching the teeth-ridge in such a way that
though there is complete closure in the middle of the mouth, yet
a passage for the air is left on one or both sides of the tongue;
the soft palate is in its raised position; the vocal cords are made
to vibrate so that ‘voice’ is produced. This formation may be
expressed shortly by defining the sounds as voiced alveolar lateral
consonants. In order to give a complete definition of any particular
variety of l-sound it is, however, necessary to specify in addition
the position of the main body of the tongue (see §§ 665-669).

661. In narrow (allophonic) transcription the clear l and dark l
are distinguished as l and ɫ respectively. Thus the word little, which
is usually transcribed simply ˈlitl, might be written in narrow transcription
ˈlitɫ.

662. The English l-phoneme is always represented in current
spelling by the letter l. Examples: let let, look luk, collar ˈkɔlə,
bell bel (narrow transcription beɫ), belt belt (beɫt), people ˈpiːpl
(ˈpiːpɫ). L is silent in calf kɑːf, half haːf, behalf biˈhaːf, chalk
ʧɔːk, walk wɔːk, Fa(u)lkner ˈfɔːknə; balm baːm, calm kɑːm, palm
pɑːm, psalm saːm, qualm kwɔːm, 2280 Malmesbury ˈmɑːmzbəri, salmon
ˈsæmən; could kud and kəd, should ʃud and ʃəd, would wud and
173wəd and əd, Holborn ˈhoubən 3281; folk fouk, yolk jouk, Folkestone
ˈfoukstən; holm houm; Lincoln ˈliŋkən; calve kɑːv, halve hɑːv, salve
(soothe) sɑːv 4282; colonel ˈkəːnl.

663. Many foreign people articulate their l-sound with the tip or
blade of the tongue against the teeth. It should be noticed,
however, that such variations in the position of the tip of the
tongue do not appreciably affect the quality of l-sounds. Variations
in the quality of l-sounds are due chiefly to the position of
the main part of the tongue (see § 665, also footnote 6 on p. 176).

664. l-sounds are pronounced unilaterally by many. In this
pronunciation the tongue obstructs the air passage in the middle
of the mouth and on one side, the air being free to pass out on
the other side. The sounds thus produced are not appreciably
different from the normal lateral sounds, in which both sides are
open.

665. Many varieties of l-sounds may be formed with the tip
of the tongue in the lateral position against the teeth-ridge or
teeth. These varieties depend on the position of the main part
of the tongue
and not on the position of the tip; this is a point
of considerable importance. While the tip is touching the teethridge
or teeth, the main part is free to take up any position, and
in particular it may take up any vowel-position. The l-sound
produced with a given vowel-position of the main part of the
tongue always has a noticeable acoustic resemblance to that vowel;
it may be said to have the ‘resonance’ of that vowel. It is not
difficult to pronounce a whole series of l-sounds having the resonance
of all the principal vowels, i, e, ɑ, ɔ, u, ə, etc. These varieties
of l may be represented, when necessary, by the notation li, le,
lɑ, lɔ, lu, lə, etc. 5283

666. Figs. 79, 80 and 81 show the approximate positions of
the tongue in pronouncing li, lu, and lə with the tip of the tongue
against the teeth-ridge. Similar diagrams may be drawn to show
the formation of li, lu, and lə pronounced with the tip of the tongue
against the teeth.174

image

Fig. 79. Tongue-position
of ‘clear’ l (li).

image

Fig. 80. Tongue-position
of ‘dark’ l (lu).

image

Fig. 81. Tongue-position
of intermediate
l (lə).

image

Fig. 82. Palatogram
of li with tip of
tongue placed as in
English.

image

Fig. 83. Palatogram
of le with tip of
tongue placed as in
English.

image

Fig. 84. Palatogram
of lɑ with tip of
tongue placed as in
English. The palatogram
of ləː (l with
resonance of English
long ɔː) is very
similar to this.

image

Fig. 85. Palatogram
of lɔ (l with resonance
of English short ɔ)
with tip of tongue
placed as in English.

image

Fig. 86. Palatogram
of l with tip of
tongue placed as in
English.

image

Fig. 87. Palatogram
of lə with tip of
tongue placed as in
English.175

667. Figs. 82 to 87 are palatograms showing the differences
between some of the chief varieties of l pronounced with the tip
of the tongue placed as in English. A similar set of diagrams
may be obtained showing the differences between the same varieties
of l pronounced with the tip of the tongue placed further forward.

668. The difference between ‘clear’ varieties of l and ‘dark’
varieties of l is thus simply a difference of vowel resonance. In
clear varieties of l there is a raising of the front of the tongue in
the direction of the hard palate (in addition to the tongue-tip
articulation), while in dark varieties of l there is a raising of the
back of the tongue in the direction of the soft palate. In other
words, clear l-sounds have the resonance of front vowels, whereas
dark l-sounds have the resonance of back vowels. 6284

669. The English ‘dark’ l, which is used finally and before
consonants, 7285 generally has the resonance of a back vowel approaching
u. The Southern English ‘clear’ l, which is used before vowels,
generally has the resonance of a front vowel approaching i. 8286176

670. Most foreign people use a clear l in English in all situations,
instead of using a dark l when final or followed by a consonant.
It is often a matter of considerable difficulty to them to acquire the
pronunciation of dark l. The best way of obtaining it is to place
the tip of the tongue between the teeth 9287 in the lateral position, and,
while the tip of the tongue is pressed firmly against the upper
teeth, to try to pronounce the vowel u without rounding the lips.

671. Many foreign learners find it easier to acquire lɔ first, by
pressing the tip of the tongue firmly against the upper teeth in the
lateral position and trying to pronounce simultaneously the vowel
ɔ. When lɔ is obtained, the quality of the sound has then to be
gradually modified until the correct lu is arrived at. It should be
remarked, however, that the sound lɔ should only be used as an
exercise and should not be used instead of lu in speaking. (The
Portuguese have a tendency to do this.)

672. Other foreign learners find it more helpful to press the tip
of the tongue firmly against the upper teeth in the lateral position
and try to pronounce a series of vowels, beginning with i, e.g. i, e,
ɑ, ɔ, u. With a little practice they are generally able to produce
readily the various varieties of l, namely li, le, lɑ, lɔ, lu, and can
therefore in particular pronounce the lu of Received English.

673. The easiest words for practising the dark l are those in
which the sound is syllabic (§ 211) and not preceded by t or d
(§ 590), e.g. people ˈpiːpl, table ˈteibl, knuckle ˈnʌkl, struggle ˈstrʌgl;
the most difficult words for most foreign people are those in which
the preceding vowel is ɔː or ou, e.g. all ɔːl, old ould.

674. The Japanese are generally unable to make any kind of
l with certainty. They confuse it with r, and use varieties of l
and r indiscriminately for both l and r when speaking English.
It is not difficult to teach a Japanese to make a l by explaining
177the manner in which the sound is produced. In the first instance
the best results are obtained if the Japanese learner practises l
with the tip of the tongue pressed firmly against the upper teeth.
When he has mastered the sound pronounced in this way, he
may proceed to form a l in the normal English position against
the teeth-ridge.

675. The chief difficulty for the Japanese is not to learn to
make l, but to remember to use it in the proper places in connected
speech. It is a help to practise reading from phonetic texts in
which every l is underlined and every r is marked in some distinctive
way (e.g. by drawing a circle round it).

676. Russians have difficulty in making the English clear l.
Before sounds of the i and e type they substitute a ‘palatalized’ l
which is followed by a distinct j-glide; their pronunciation of
live, let sounds like ljiv, ljet. Before other vowels they use a
dark l; thus they pronounce like, lock as ɫaik, ɫɔk. To improve
their pronunciation they should learn to use a l of intermediate
resonance, such as lə or le, by the method indicated in § 672.

677. Words for practising clear l 10288: leave liːv, lick lik, let let,
lamb læm, large lɑːʤ, long loŋ, law lɔː, look luk, lose luːz, love
lʌv, learn ləːn; lake leik, loaf louf, line lain, loud laud, employ
imˈplɔi; clear kliə, flare flɛə, floor flɔə or flɔː; cellar ˈselə, calling
ˈkɔːliŋ, jelly ˈʤeli.

678. Words for practising dark l: double ˈdʌbl, noble ˈnoubl,
possible ˈpɔsəbl, struggle ˈstrʌgl, eagle ˈiːgl, angle ˈæŋgl, vessel ˈvesl,
partial ˈpɑːʃl, little ˈlitl, settle ˈsetl, middle ˈmidl, candle ˈkændl;
feel fiːl, fill fil, fell fel, shall ʃæl, snarl snɑːl, doll dɔl, fall fɔːl,
full ful, fool fuːl, dull dʌl, curl kəːl, fail feil, foal foul, file fail, fowl
faul, foil fɔil; field fiːld, milk milk, health helθ, Alps ælps, scald
skɔːld, bulk bʌlk, pulpit ˈpulpit, ruled ruːld; nails neilz, cold kould,
child ʧaild, owls aulz, coils kɔilz.178

Chapter XXI
The English Fricative Consonants

Detailed Descriptions

679. Fricative consonants are formed by a narrowing of the
air-passage at some point so that, when air is expelled by pressure
from the lungs, it escapes with a kind of hissing sound.

680. All fricative consonants may be pronounced with a varying
amount of audible friction. In the case of voiced fricative consonants,
when the friction is so reduced as to become practically
imperceptible, the sounds become ‘frictionless continuants’ (Chap.
XXII).

681. There exist in English ten fricative consonant phonemes.
They are represented in phonetic transcription by the letters f,
v, θ, ð, s, z, ʃ, ʒ, r, h.

f

682. The sound f is formed by pressing the lower lip against
the upper teeth and allowing the air to force its way between
them and through the interstices of the
teeth; the soft palate is in its raised
position and the glottis is left open.
This formation may be expressed shortly
by defining the sound as a breathed
labio-dental fricative
consonant. The lipposition
is Shown in Fig. 88.

image

Fig. 88. Formation of f.

683. The positions of the tongue and lips during the articulation
of f approximate to those required for adjacent vowels. To this
extent therefore it may be said that there exist subsidiary members
of the phoneme (considering the f with neutral tongue-position
to be the principal member). These differences of tongue and
lip positions are, however, slight, and their effects on the acoustic
179quality of the sound are negligible for the ordinary linguist. 1289
For practical linguistic purposes it may therefore be said that the
English f-phoneme consists of a single sound and has no subsidiary
members differing appreciably from the principal member.

684. f is the regular sound of f and ph; examples: far fɑː,
faithful ˈfeiθful, philosophy fiˈlɔsəfi. Gh is pronounced f in the
following common words: enough iˈnʌf, rough rʌf, tough tʌf, cough
kɔf, 2290 trough trɔf, 2291 laugh lɑːf, draught draːft; also in the less common
words chough ʧʌf, slough (skin of a snake) slʌf. ʒ292 Note the pronunciation
of lieutenant lefˈtenənt. 4293

685. The Japanese generally replace f by a breathed bilabial
fricative ϕ (Fig. 89). (One form of ϕ is the sound made in
blowing out a candle; ϕ is the breathed
consonant corresponding to the voiced
sound β described in §§ 692, 806.) The
error may be remedied by holding the
upper lip out of the way, and practising
the sound with the lower lip firmly
pressed against the upper teeth.

image

Fig. 89. Formation of ϕ
(‘bilabial f’)

686. The same error is occasionally
met with from Germans and Norwegians, especially when the sound
is preceded by a consonant, e.g. in useful ˈjuːsfl.

687. Words for practice: feed fiːd, fit fit, fence fens, fat fæt,
farm faːm, fond fɔnd, force fɔːs, foot fut, food fuːd, fun fʌn, fir
fəː; fail feil, fold fould, fine fain, found faund, foil fɔil; fear fiə,
fair fɛə, four fɔə or fɔː, Balfour ˈbælfuə 4a294; safe seif, loaf louf, half hɑːf.

v

688. The principal English v is formed like the principal English
f (§ 682 and Fig. 88) except that the vocal cords are made to
180vibrate so that ‘voice’ is produced during the articulation of the
sound. The formation of v may therefore be expressed shortly
by denning it as a voiced labio-dental fricative consonant.

689. There exist unimportant subsidiary members of the English
v-phoneme formed by approximating the positions of the tongue,
and to some extent the lips, to the positions required for adjacent
vowels in connected speech.

690. Partially voiceless varieties of v () occur in the speech of
some as subsidiary members of their v-phoneme in initial and final
positions (see §§ 788-794).

691. v is the usual sound of v; example: voice vɔis, wave weiv.
Ph is generally pronounced v in nephew ˈnevjuː, though some English
people pronounce ˈnefjuː.

692. Many Germans have a tendency to replace v by the
bilabial fricative β. β has the same lip-position as ϕ (Fig. 89).
The English sound v has the same lip-position as f (Fig. 88),
and is acquired by simply pressing the lower lip firmly against
the upper teeth (taking care to keep the upper lip out of the way)
and producing voice, forcing the air through the narrow passage
thus formed. In practising the sound the upper lip may, if
necessary, be held out of the way with the finger. The German
tendency to use β is particularly strong when the sound occurs
in the neighbourhood of the sound w,
as in equivalent iˈkwivələnt.

image

Fig. 90. Formation of the
Frictionless Continuant ʋ.

693. Indians generally replace v by
a frictionless continuant ʋ in which the
lower lip touches the centre front teeth
lightly and is so held as to allow the
air to escape chiefly at the sides (see
Fig. 90). I am informed that Austrians
of the Tirol also use this sound. To pronounce v correctly, they
must observe carefully the difference between Figs. 88 and 90.

694. Words for practice: veal viːl, vicar ˈvikə, vest vest, van
væn, vase vɑːz, volume ˈvɔljum, vault vɔːlt, vulgar ˈvʌlgə, verse
vəːs; vain vein, vote vout, vine vain, vow vau, voice vɔis; veer viə,
various ˈvɛərĭəs; give giv, glove glʌv, prove pruːv, wives waivz,
very well ˈveri ˈwel; a very vivacious and vain villain visited various
181villages of the valley
ə ˈveri viˈveiʃəs ənd ˈvein ˈvilən ˈvizitid ˈvɛərĭəs
vilidʒiz əv ðə ˈvæli
.

ɵ

695. The English phoneme represented by θ may be regarded
as comprising only one sound. There are no members of the
phoneme differing to any marked extent
from this sound.

image

Fig. 91. Tongue-position
of θ.

696. The sound θ is articulated by the
tip of the tongue against the upper teeth,
the main part of the tongue being fairly
flat (see Figs. 91, 92); the air passage
between the tip of the tongue and the
upper teeth is narrow; the soft palate is
in its raised position and the vocal cords
are not made to vibrate. The formation
of θ may be expressed shortly by defining
it as a breathed dental fricative consonant.

697. Fig. 93 is a palatogram of the English θ.

image

Fig. 92. Front view of mouth
in pronouncing θ.

698. θ is one of the sounds of th.
Th is pronounced in this way (i) initially
except in the words mentioned in
§ 704, e.g. thin θin, thank θæŋk, (ii)
medially in non-Germanic words, e.g.
method ˈmeθəd, author ˈɔːθə, sympathy
ˈsimpəθi, (iii) finally in all words except
those mentioned in § 704, e.g. mouth mauθ, month mʌnθ.

image

Fig. 93. Palatogram
of θ.

699. Plurals of words ending in th take the
pronunciation θs in the following cases, (i)
If one of the short vowels precedes, e.g. smiths
smiθs, breaths breθs, moths mɔθs, mammoths
ˈmæməθs; (ii) if a consonant precedes, e.g.
lengths leŋθs, healths helθs, months 0iʌnθs;
(iii) if the letter r precedes in the spelling,
e.g. births bəːθs, hearths hɑːθs (compare baths
bɑːðz), an exception being berths which is
182bəːθs or bəːðz; (iv) in the exceptional words heaths hiːθs, faiths
feiθs, growths grouθs, sloths slouθs. In other cases ðz is used, e.g.
baths bɑːðz. mouths mauðz, youths juːðz (compare the singular bɑːθ,
mauθ, juːθ). In wreaths, sheaths, the pronunciation varies; some
say riːðz, ʃiːðz and others say riːθs, ʃiːθs. My pronunciation is
with ðz. 5295 In cloths and broths the pronunciation varies according
to the vowel used. These words are now generally said with short
ɔ, and the plurals are klɔθs, brɔθs; those who, like me, use the
more old-fashioned pronunciation with long ɔː generally make the
plurals klɔːðz, brɔːðz. Some of the latter, however, distinguish two
plurals of cloth, klɔːðz meaning ‘pieces of cloth’ and kiɔːθs meaning
‘kinds of cloth.’

700. Many foreign people replace θ by f or by some variety of s.
They may learn to acquire θ by starting with an exaggerated
form of it, placing the tip of the tongue so that it projects out
between the upper and lower teeth. When the tongue is in this
position, they must blow so that a stream of air passes out between
the tongue-tip and the edge of the upper teeth. The lower lip
must be kept out of the way when practising this exercise. The
quality of sound produced in this manner is about the same as
that of the ordinary English θ. When the learner has become
familiar with the sound formed in this exaggerated way, he can
soon learn to modify the articulation and articulate with the tongue
in the normal English position shown in Fig. 91.

700a. It should be observed that in making θ the teeth are
separated more widely than in the articulation of s.

701. Words for practice: theme θiːm, thin θin, theft θeft, thank
θæŋk, thong θɔŋ, thought θɔːt, thumb θʌm, third θəːd, thermometer
θəˈmɔmitə; Thane θein, 6296 three θriː, thwart θwɔːt; heath hiːθ, smith
smiθ, breath breθ, bath bɑːθ, north nɔːθ, truth truːθ, birth bəːθ;
both bouθ, mouth mauθ; method ˈmeθəd, author ˈɔːθə, sympathy
ˈsimpəθi, ether ˈiːθə; thirty-three things ˈθəːti ˈθriː ˈθiŋz.183

ð

702. The principal member of the English phoneme represented
by ð is the voiced consonant corresponding to the breathed θ. Its
formation may be expressed shortly by defining it as a voiced
dental fricative
consonant. (See Figs. 91, 92.)

703. Some English people use partially voiceless varieties of ð as
subsidiary members of their ð-phoneme in initial and final positions
(see §§ 788-794).

704. ð is one of the sounds of th. Th is pronounced in this
way (i) initially in pronouns such as this ðis, they ðei, and in than,
that, the, then, thence, there, thither, though, thus, and their derivatives
such as themselves ðəmˈselvz, thenceforth ˈðensˈfɔːθ, therefore ˈðɛəfɔː,
(ii) medially in words of Germanic origin, e.g. father ˈfɑːðə, northern
ˈnɔːðən, (iii) in plurals of nouns ending in -th not preceded by r
containing a long vowel or a diphthong, e.g. paths paːðz, youths
juːðz, oaths ouðz, mouths mauðz (exceptions are faiths, heaths,
growths, sloths and with some speakers sheaths, wreaths, cloths,
see § 699), (iv) finally when there is a mute -e in the spelling (e.g.
bathe beið), and in the single words with wið, 7297 bequeath biˈkwiːð,
booth buːð, smooth smuːð and the rare verbs mouth mauð and
south sauð. 8298

705. Foreign people have the same difficulty with ð as with θ,
and the correct sound may be acquired as directed in § 700.

706. Some foreign people, especially Scandinavians and Germans,
do not always voice the sound ð properly. They will find it useful
to practise singing the sound, sustaining it on various notes.

707. Words for practice: these ðiːz, this ðis, then ðen, that ðæt,
thus ðʌs; they ðei, though ðou, thy ðai, thou ðau, there ðɛə; breathe
briːð, with wið, soothe suːð, bathe beið, loathe louð, scythe saið;
gather ˈgæðə, worthy ˈwəːði, hither and thither ˈhiðər ən ˈðiðə.

708. θ and ð are particularly difficult for foreign learners when
they occur near the sounds s and z. Students are recommended to
practise carefully such phrases as this is the thing ˈðis iz ðə ˈθin,
the sixth street ðə ˈsiksθ ˈstriːt, the hyacinths and the chrysanthemums
ðə ˈhaiəsinθs ən ðə kriˈsænθəməmz.184

s

709. The English s-phoneme may be considered for practical
language teaching as comprising only one sound. 9299 This sound is
articulated by the blade (or tip and blade) of the tongue against
the teeth-ridge, the ‘front’ of the tongue being at the same time
somewhat raised in the direction of the hard palate (Figs. 94, 95).

image

Fig. 94. Tongue-position
of s pronounced with tip
of tongue raised.

image

Fig. 95. Tongue-position
of s pronounced with tip
of tongue lowered.

The teeth are close together; the sound cannot be pronounced
with the mouth wide open (see Fig. 96). The space between the
blade of the tongue and the teeth-ridge is extremely narrow. The
soft palate is in its raised position, and the vocal cords are not
made to vibrate. The formation of s may be expressed shortly by
defining the sound as a breathed blade-alveolar fricative consonant.

710. The tip of the tongue is with
some speakers raised towards the teethridge
(as shown in Fig. 94), and with
others kept against the lower teeth (as
shown in Fig. 95). The first formation
seems the more usual in English.

image

Fig. 96. Front view of mouth
in pronouncing s.185

711. Fig. 97 is a palatogram of the sound s, as pronounced by
me (tip of tongue raised). Fig. 98 is a palatogram of the sound
s as pronounced by a Frenchwoman (tip of tongue lowered). The
two sounds though formed slightly differently strike the ear as
being very similar.

image

Fig. 97. Palatogram
of English s pronounced
with tip of
tongue raised.

image

Fig. 98. Palatogram
of French s pronounced
with tip of
tongue lowered.

712. There exist many diaphonic variants of s differing in the
quality of the hiss or the degree of its penetrating power. Some
very slight changes in the adjustment of the tongue produce
considerable alterations in the acoustic quality of s. Moreover,
the kinds of s which an individual is capable of pronouncing depend
to some extent on the formation of his teeth. The occurrence
of the various types of s in speech is thus partly a matter of language
or dialect and partly individual. 10300

713. It may be said that as a general rule the s of French people
has a more penetrating hiss than that of English people, and the
s of Germans has a still more penetrating hiss than that of the
French. The use of a particularly penetrating s is a characteristic
of the pronunciation of many Germans, and it sounds incorrect
if used in English. An English s may be acquired by those who
naturally use a sound of more penetrating quality either by
diminishing the force of the breath or by articulating with the
tip of the tongue raised and held somewhat further back than
for the speaker's habitual s.186

714. s is the normal sound of the letter s in English, as in so
sou, sets sets. S is always pronounced s at the beginnings of
words, but in other positions it is very frequently pronounced z.
Compare absurd əbˈsərd, absolve əbˈzɔlv; cease siːs, please pliːz,
base beis, phrase freiz; close (adj.) klous, close (v.) klouz; use (noun)
juːs, use (v.) juːz, used to (in the senses of ‘accustomed to,’ ‘was
in the habit of’) juist, used (‘made use of’) juizd; this 5is, is iz.
Most of the rules regarding the use of s and z are so complicated
and subject to such numerous exceptions, that the foreign learner
will find the easiest way of acquiring the correct pronunciation is
to learn the pronunciation of each word individually as he comes
across it.

715. The following points should, however, be noted, (i) The
s denoting the plural of nouns or third person singular of verbs
is pronounced s when the preceding sound is a voiceless consonant,
e.g. cats kæts, takes teiks, laughs lɑːfs. (ii) The s in the terminations
-sive, -sily is nearly always pronounced s, e.g. conclusive kənˈkluːsiv,
curiosity kjuəriˈɔsiti. (iii) Final s preceded by one of the letters
ɑ, i, o, u or y is pronounced s (when not mute 11301), e.g. gas gæs,
atlas ˈætləs, this ðis, basis ˈbeisis, chaos ˈkeiɔs, us ʌs or əs, 12302 genius
ˈdʒiːnjəs, precious ˈpreʃəs, 13303 Gladys ˈglædis. The only exceptions
are the inflected forms of nouns and verbs (e.g. plays pleiz, was
wɔz or wəz), and the single words his hiz (weak form iz), as æz
(weak form əz), whereas wɛəˈræz, avoirdupois ˌævədəˈpɔiz.

716. The following is a list of the chief words ending in -se
in which the final consonant is s: abase əˈbeis, base beis, case keis
(and compounds, e.g. encase inˈkeis, staircase ˈstɛəkeis), chase ʧeis,
purchase ˈpəːʧəs or ˈpəːʧis; cease siːs, crease kriːs, decease diˈsiːs,
decrease (noun) ˈdiːkriːs, decrease (v.) diːˈkriːs, grease (noun) griːs, 14304
increase (noun) ˈinkriːs, increase (v.) inˈkriːs, lease liːs, release (noun
and v.) riˈliːs 15305; Chersonese ˈkəːsəniːs, geese giːs, obese ouˈbiːs; anise
ˈænis, concise kənˈsais, paradise ˈpærədais, practise ˈpræktis, precise
187priˈsais, premise (noun) ˈpremis, 16306 promise ˈprɔmis; tortoise ˈtɔːtəs;
bellicose ˈbelikous, close (noun meaning ‘enclosed place,’ and adj.)
klous, 17307 dose dous, jocose ʤəˈkous, morose məˈrous, purpose ˈpəːpəs,
verbose vəːˈbous; goose guːs, loose luːs, noose nuːs (also pronounced
nuːz); obtuse əbˈtjuːs, profuse prəˈfjuːs, recluse riˈkluːs, refuse (noun)
ˈrefjuːs, 18308 use (noun) juːs, grouse graus, house haus, louse laus,
mouse (noun) maus, 19309 souse saus; also all words ending in -lse, -nse,
-pse, -rse (with the single exceptions of cleanse klenz and parse pɑːz),
e.g. else els, dense dens, lapse læps, course kɔːs.

717. The sound s is also the usual sound of c before e, i, and
y, as in cell sel, face feis, cinder ˈsində, mercy ˈməːsi.

718. S is silent in isle ail, island ˈailənd, aisle ail, corps (sing.)
kɔː, 20310 chamois ˈʃæmwɑː, 21311 rendezvous (sing.) ˈrɔndivuː, 22312 debris
ˈdebriː, demesne diˈmein, viscount ˈvaikaunt.

719. Some foreign learners tend to voice the sound s, especially
when it occurs between two vowels, thus replacing it by z; others
will use a partially voiced z or an ‘unvoiced z,’ a sound which
has an effect intermediate between s and z (phonetic symbol ).
Those who have this tendency should practise words like necessary
ˈnesisəri or ˈnesisri, ceaseless ˈsiːslis.

720. Further words for practice: see siː, sit sit, set set, sat sæt,
psalm sɑːm, song sɔŋ, saw sɔː, soot sut, soon suːn, son sʌn, certain
ˈsəːtn, say sei, so sou, sigh sai, sound saund, soil sɔil; serious
ˈsiərĭəs, Sarah ˈsɛərə, soar, sore sɔə 23313; this ðis, less les, pass pɑːs,
gross grous, course kɔəs, 24314 scarce skɛəs, places ˈpleisiz, ceases ˈsiːsiz,
exercises ˈeksəsaiziz.

z

721. The principal English z is the voiced consonant corresponding
to the breathed s. The formation of the sound may
188therefore be expressed shortly by defining it as a voiced bladealveolar
fricative
consonant. It is articulated by the blade (or
tip and blade) of the tongue against the teeth-ridge, the front
of the tongue being at the same time slightly raised in the direction
of the hard palate (see Figs. 94, 95). The teeth are brought close
together, and the passage between the blade of the tongue and
the teeth-ridge is extremely narrow. The soft palate is in its
raised position, and the vocal cords are made to vibrate so that
‘voice’ is produced. Some English people use some lip articulation
in addition (see footnote 10 to § 712).

722. Partially voiceless varieties of z occur as subsidiary members
of the English z-phoneme in initial and final positions (see § 788).
Completely devoiced may be heard from some English speakers
in these positions. 25315

723. z is the sound of the letter z; examples: zone zoun, razor
reizə. It is also very frequently represented by the letter s, when
not initial; examples: raise reiz, easy ˈiːzi, observe əbˈzəːv, his hiz.
Final s denoting the plural of nouns or third person singular of verbs
is pronounced z when preceded by a vowel or by a voiced consonant;
examples: trees triːz, plays pleiz, rushes ˈrʌʃiz, dogs dɔgz, ideas
əiˈdiəz, falls fɔːlz, gives givz; also does dʌz, has hæz, is iz, was wɔz.
Final s is pronounced z in other words whenever it is preceded
by a pronounced e, e.g. species ˈspiːʃiːz, Hades ˈheidiːz, aborigines
æbəˈriʤiniːz. 26316 Note the exceptional words with final z mentioned
at the end of § 715; also Mrs. ˈmisiz. Note that ss is pronounced
z in the words dessert diˈzəːt, dissolve diˈzɔlv, hussar huˈzɑː, possess
pəˈzes, scissors ˈsizəz, and that house haus has the irregular plural
ˈhauziz.

724. Some foreign people, and especially Scandinavians and
Germans, do not voice the sound z properly, but replace it habitually
by a consonant which sounds like a weak s (phonetic symbol ).
They go beyond the permissible devoicing referred to in § 722.
Those who have this tendency will find it useful to practise singing
the sound z, sustaining it on various notes.189

725. Words for practice: zeal ziːl, zest zest, Zoo zuː, zones zounz;
scissors ˈsizəz, reserves riˈzəːvz, diseases diˈziːziz.

ʃ

726. The English phoneme represented by the letter ʃ may be
considered as comprising only one sound. There are no subsidiary
members of the phoneme differing to any marked extent from the
principal member.

image

Fig. 99. Tongue-position
of English ʃ pronounced
with tip of
tongue raised.

image

Fig. 100. Tongue-position
of English ʃ pronounced
with tip of
tongue lowered.

727. The normal English ʃ is articulated by the tip and blade of
the tongue against the hinder part of the teeth-ridge, the whole of
the main body of the tongue being simultaneously held in a raised
position after the manner shown in Fig. 99. The teeth are close
or fairly close together; the sound cannot be properly pronounced
with the mouth wide open. The space between the blade of the
tongue and the teeth-ridge is narrow, though wider than for s; on the
other hand the air channel in the region of the palate is narrower
than in the case of s. There is protrusion of the lips as shown in
Fig. 101. The soft palate is in its
raised position, and the vocal cords
are not made to vibrate. The formation
of ʃ may be expressed shortly by
defining the sound as a breathed palatoalveoktr
fricative
consonant.

image

Fig. 101. Lip-position of
English ʃ.

728. With most speakers the tonguetip
is raised and articulates against the teeth-ridge. Some people,
190however, make the sound with the blade only, keeping the tip
lowered as shown in Fig. 100. This formation does not entail
any perceptible difference in acoustic effect.

729. Some English people use a variety of / made with spread
lips. It has a ‘clearer’ acoustic quality than the normal ʃ.

730. Figs. 102 and 103 are palatograms of ʃ, the first being
my own and the second being that of a Frenchwoman. Not withstanding
the considerable differences of tongue position shown by
these palatograms, there is not much acoustic difference between
the sounds.

image

Fig. 102. Palatogram
of English ʃ pronounced
with tip of
tongue raised.

image

Fig. 103. Palatogram
of French ʃ pronounced
with tip of
tongue lowered.

731. The chief differences between the articulation of ʃ and s
are well seen by comparing the sectional diagrams (Figs. 99 and 94)
which are adapted from X-ray photographs, the palatograms
(Figs. 102 and 97), and the photographs of lip-positions (Fig. 101
and 96).

732. ʃ is the usual sound of sh in English; examples: shoe
ʃuː, wish wiʃ. It is also often used where the spelling has -si-,
-ci-, -sci-, -ti-, etc., followed by an unstressed vowel or syllabic
consonant; examples: mansion ˈmænʃn, Persia ˈpəːʃə, special
ˈspeʃl, provincial prəˈvinʃl, musician mju(ː)ˈziʃn, precious ˈpreʃəs,
ancient ˈeinʃənt, ocean ˈouʃn, permission pəˈmiʃn, conscious ˈkonʃəs,
nation ˈneiʃn, vexatious vekˈseiʃəs, partial ˈpɑːʃl, 27317partiality pɑːʃiˈæliti,
191associate (verb) əˈsouʃieit, (noun) əˈsouʃiːt 28318; so also in words
like censure ˈsenʃə, pressure ˈpreʃə. S is pronounced ʃ in sure
ʃuə, assure əˈʃuə, etc., and in sugar ˈʃugə. Ch is pronounced
ʃ in various recently borrowed French words, such as champagne
ʃæmˈpein, chandelier ʃændəˈliə, machine məˈʃiːn, moustache məsˈtɑːʃ.
Chivalry used to be pronounced with ʧ, but is now usually pronounced
with ʃ (ˈʃivlri).

733. ʃ may be considered as an element of the affricate ʧ
For details see §§ 601 ff.

734. Some Central and South Germans replace ʃ by the corresponding
voiced sound ʒ, especially in intervocalic position. They
must master the difference between the voiced and breathed
sounds, if necessary making use of the tests mentioned in § 91.
They must give special attention to the pronunciation of such
words as nation ˈneiʃn, marshes ˈmaːʃiz, social ˈsouʃl.

735. Danes generally make ʃ too palatal, with the result that
it sounds to an English ear like ʃj; thus they pronounce shine,
which should be ʃain, in such a way that it sounds very like ʃjain.
A correct English ʃ may be acquired by trying to keep the tongue
very loose, and by retracting the tip of the tongue and exaggeratmg
the lip-protrusion.

736. Words for practising ʃ: sheaf ʃiːf, ship ʃip, shell ʃel, shadow
ˈʃædou, sharp ʃaːp, shock ʃok, Shaw ʃɔː, shoes ʃuːz, shut ʃʌt, shirt
ʃəːt; shake ʃeik, show ʃou, shy ʃai, shout ʃaut; shear ʃiə, share ʃɛə,
shore ʃɔə, 29319 sure ʃuə 30320; fish fiʃ, ash æʃ, marsh mɑːʃ, squash skwɔʃ,
bush buʃ.

ʒ

737. The principal English ʒ is formed like ʃ (§ 727) except
that the air-pressure is weaker and the vocal cords are made to
vibrate so that ‘voice’ is produced during the articulation of the
sound. It may be described as a voiced palato-alveolar fricative
consonant.192

738. Partially voiceless varieties of ʒ occur in initial and final
positions (see § 788); they are subsidiary members of the English
ʒ-phoneme. Completely voiceless ʒ̊ may be heard from some
English speakers in these positions. 31321

739. ʒ is the sound of s in words like measure ˈmeʒə, pleasure
ˈpleʒə, -si- in occasion əˈkeiʒn, hosier houʒə and numerous other
words in which -si- is immediately preceded by a stressed vowel. 32322
ʒ is also heard in the miscellaneous words usual ˈjuːʒul or ˈjuːʒŭəl,
azure ˈæʒə, seizure ˈsiːʒə, transition trænˈsiʒn, 33323 and words recently
borrowed from French such as rouge ruːʒ, garage (noun) ˈgærɑːʒ. 34324

740. ʒ may be considered as an element of the affricate ʤ.
For details see §§ 601, 609.

741. Some foreign people, and especially Scandinavians and
Germans, do not voice ʒ properly, but replace it by a weak ʃ.
Those who have this tendency will find it useful to practise singing
the sound ʒ, sustaining it on various notes.

742. Danes generally use a variety of ʒ which is too palatal.
The sound which they use sounds to an English ear like ʒj when
a vowel follows; thus their pronunciation of measure sounds too
much like ˈmeʒjə. To correct this the tongue should be held
loosely and the articulation should be made with the tip of the
tongue rather retracted; the lips should be rounded and protruded
as shown in Fig. 101.

743. Words for practice: seizure ˈsiːʒə, pleasure ˈpleʒə, treasure
ˈtreʒə, leisure ˈleʒə, enclosure inˈklouʒə, composure kəmˈpouʒə;
prestige presˈtiːʒ, barrage ˈbærɑːʒ, massage ˈmæsɑːʒ, camouflage
ˈkæmuflɑːʒ, espionage espĭəˈnɑːʒ, 35325 rouge ruːʒ, gamboge gæmˈbuːʒ,
cortege kɔːˈteiʒ, beige beiʒ, Vosges vouʒ.

r

744. There exist a number of sounds which fall under the
general heading of r-sounds. The one with which we are chiefly
193concerned here is the most usual English r-sound which is a fricative
consonant.

745. The International Phonetic Alphabet provides special letters
for representing various kinds of r-sound in narrow (‘allophonic’ or
‘comparative’) transcriptions. It is, however, not as a rule necessary
to employ these, since it does not often happen that two of these
varieties occur as separate phonemes in a language. The particular
variety or varieties used in each language can generally be specified
once for all.

746. The chief varieties of r-sound are as follows:

tableau special symbol | rolled lingual r | r | flapped lingual r | ɾ | fricative lingual r | ɹ | rolled uvular r | ʀ | fricative uvular r | ʁ

There exists also a retroflex flap (ɽ), which though easily confused
by Europeans with flapped lingual r, should not be regarded
as a ‘variety’ of r; in most languages in which it occurs it is a
separate phoneme from one of the above-mentioned r-sounds. 36326

image

Fig. 104. Tongueposition
of English
fricative r.

747. The most usual English r is a ‘fricative
lingual’ sound. It is articulated by the
tip of the tongue against the back part of the
teeth-ridge, the main body of the tongue being
kept low and the ‘front’ being held concave
to the palate, as shown in Fig. 104, and the
whole tongue being laterally contracted (§ 73).
The distance between the jaws is immaterial;
the sound can be pronounced with a wide
separation between the upper and lower
teeth. The soft palate is in its raised position,
and the vocal cords are made to vibrate so
that voice is produced during the articulation of the sound. The
formation of this r may be expressed shortly by defining the sound
as a voiced post-alveolar fricative consonant.194

748. A partially voiceless variety of r occurs as a subsidiary
member of the English r-phoneme when a voiceless consonant
precedes in the same syllable, as in prove pruːv, crowd kraud. See
§ 845 (i).

749. Many English people pronounce r with a certain amount
of lip-protrusion, especially in stressed position. Others regularly
use a frictionless continuant r (§ 796), which is likewise generally
accompanied by lip-protrusion in stressed position (§ 797).

750. Many speakers of Received English use a ‘flapped’ r
(§ 753) as a subsidiary member of the r-phoneme; it occurs chiefly
in unstressed intervocalic position, as in very ˈveri, period ˈpiərĭəd,
and when inserted at the end of a word (§ 756). The use of this
subsidiary r is not essential; the fricative r is also quite commonly
used in such cases.

751. In the North of England and in Scotland rolled r is generally
used in initial position in place of the Southern fricative r. The
use of rolled r is also generally recommended
by teachers of singing and elocution
in all parts of the country. The
sound is also often used in making telephone
calls by those who can pronounce it.

image

Fig. 105. Action of
tongue in pronouncing
Rolled Lingual r.

752. Rolled lingual r is formed by a
rapid succession of taps of the tip of the
tongue against the teeth-ridge, as shown
in Fig. 105. The taps are not made by
any conscious muscular movement of the
tip of the tongue; the tongue is held loosely
in the appropriate position, and the airstream
causes the tip to vibrate. The action is similar to that of
a musical reed.

753. Flapped r is formed like rolled r but consists of only one
single tap of the tip of the tongue against the teeth-ridge.

754. Figs. 106 and 107 are palatograms of fricative r and flapped
r (my pronunciation).

755. In non-dialectal Southern English a r-sound is the usual
pronunciation of the letter r when a vowel-sound follows, as in
red red, round raund, write rait, grow grou, arrange əˈreinʤ, story
195ˈstɔːri, for instance fər ˈinstəns or ˈfrinstəns. In this type of English
no r-sound is ever used finally or before a consonant, except
occasionally when a is elided. Thus far, fir, err, fear, fair, four
are pronounced fɑː, fəː, əː, fiə, fɛə, fɔə or fɔː, and farm, cord, first,
erred, fierce, scarce, fours are pronounced fɑːm, kɔːd, fəːst, əːd, fiəs,
skɛəs, fɔəz or fɔːz; nearly ˈniəli rhymes exactly with really ˈriəli.
Exceptionally r occurs before n and l in one pronunciation of words
like barren ˈbærn, quarrel ˈkwɔrl (more usually ˈbærən, ˈkwɔrəl).

image

Fig. 106. Palatogram
of English Fricative
r (ɹ).

image

Fig. 107. Palatogram
of Flapped r (ɾ) in the
sequence ɑːrɑː.

756. But when a word ending with the letter r is immediately
followed by a word beginning with a vowel, then a r-sound (generally
the flapped variety, § 750) is usually inserted in the pronunciation.
Thus though pair by itself is pronounced pɛə, yet a pair of shoes
is usually pronounced ə ˈpɛər əv ˈʃuːz. Similarly your by itself is
pronounced jɔː, 37327 your book is pronounced ˈjɔː ˈbuk, but your own
is pronounced jɔːr ˈoun; similarly our by itself is auə but our own
is auər ˈoun 38328; far by itself is fɑː, but far away is ˈfɑːr əˈwei; other
by itself is ˈʌðə, but the other end is ði ˈʌðər ˈend. 39329 r is also generally
inserted in compound words, such as over-eat ˈouvərˈiːt, razor-edge
ˈreizərˈeʤ. r inserted in this way is called ‘linking r.’196

757. There are, however, special circumstances in which a
final r has no consonantal value even when the following word
begins with a vowel. The principal cases are: (i) when the vowel
of the syllable in question is preceded by r, e.g. the emperor of
Japan
ði ˈempərə əv ʤəˈpæn, a roar of laughter ə ˈrɔː əv ˈlɑːftə,
a rare animal a ˈrɛə ˈæniml, nearer and nearer ˈniərə ən ˈniərə,
there are at least four of them ðɛər ə ət ˈliːst ˈfɔːr ɔv dəm, (ii) when
a pause is permissible between the two words (even though no
pause is actually made), e.g. he opened the door and walked in
hiː ˈoupnd ðə ˈdɔː ənd ˈwɔːkt ˈin.

758. Cases may also be found which do not seem to admit of
any satisfactory explanation. Thus very many speakers say
ˈmɔː an ˈmɔː for ˈmɔːr an ˈmɔː (more and more), biˈfɔː it s tuː ˈleit
for biˈfɔːr it s tuː ˈleit (before it's too late). Some people say
a ˈpɛə əv ˈbuːts instead of ə ˈpɛər əv ˈbuːts and ai ˈdounou ˈwɛə iˈtiz
for -ˈwɛər iˈtiz (I don't know where it is). There appears to be
an increasing tendency, especially among younger people, not
to use linking r at all, particularly when the vowel following the
word ending in r is unstressed. Sometimes even compound words
such as fire-engine, hair-oil may now be heard without the r:
ˈfaiəˌenʤin, ˈhɛəɔil instead of the more normal ˈfaiərˌenʤin,
ˈhɛərɔil.

759. Many English people add r to words ending in a when
the following word in the sentence begins with a vowel, even if
there is no r in the spelling. Thus the idea of it is very often
pronounced ði aiˈdiər əv it instead of ði aiˈdiə əv it. Other examples
are china and glass ˈʧainər ən ˈglɑːs, the sofa over there ðə ˈsoufə
ouvə ðɛə
, a vanilla ice ə vəˈnilər ˈais, Asia and Africa ˈeiʃər and
ˈæfrikə
, a diploma of honour ə diˈploumər əv ˈɔnə, a banana or an
apple
ə bəˈnɑːnər ɔː ən ˈæpl, Lena Ashwell ˈliːnər ˈæʃwel, the sonata
in F
ðə səˈnɑːtər in ˈef. r inserted in such cases is called ‘intrusive
r.’ Most teachers discourage its use, but it cannot be denied
that a very large number of people, educated as well as uneducated
pronounce in this way.

760. An intrusive r may also sometimes be heard after ɑː and
ɔː. Thus the Shah of Persia, the law of England are sometimes
pronounced ðə ˈʃɑːr əv ˈpəːʃə, ðə ˈlɔːr əv ˈiŋglənd instead of ðə ˈʃɑː
əv ˈpəːʃv
, ðə ˈlɔː vv ˈiŋglənd. The use of intrusive r after these
197vowels is less frequent than its use after ə. For further information
concerning linking r and intrusive r, see my Pronunciation of
English
(1950 and subsequent editions), §§ 357-366, and Explanations
XV in the 11th (1956) edition of my English Pronouncing Dictionary.

761. It is not necessary for foreign people to learn to use intrusive
r. They should, however, know of its existence; otherwise they
may sometimes fail to understand what is said to them by English
people who insert it.

image

Fig. 108. Position of
tongue and action of
uvula in pronouncing
Uvular Rolled r (ʀ).

image

Fig. 109. Tongue-position
of Uvular Fricative
r(ʁ).

762. Many European foreigners, including most French people
and most Germans, replace the English r-sound by a uvular rolled
consonant (special phonetic symbol ʀ). This sound is formed by a
vibration of the uvula against the back of the tongue, somewhat
as shown in Fig. 108. 40330 This vibration may be seen in a lookingglass,
when the sound is pronounced with the mouth wide open.

763. Some European foreigners use the corresponding fricative
(narrow phonetic symbol ʁ), Fig. 109.

764. The sounds ʀ and ʁ give no palatograms.

765. The use of ʀ or ʁ is one of the commonest mistakes made
by French, German and Danish people in pronouncing English.
It may be added that foreign people often make their pronunciation
198still more un-English by pronouncing or giving some indication of
the sound where the letter r is final or followed by a consonant —
positions in which r-sounds do not exist in non-dialectal Southern
English (§ 755); thus European foreigners often pronounce part,
bird as pɑʀt, bœʀd, instead of pɑːt, bəːd.

766. Experience has shown that the foreign learner who uses a
uvular r must first dismiss from his mind his foreign sound of r,
and persuade himself that the English sound to be learnt is something
entirely different and related only to z and ʒ. Keeping
this idea in mind, and remembering also the manner of forming
the English r described in § 747, many foreign learners are able
to acquire the English sound without much difficulty. For those
for whom these directions are insufficient the following exercise
is generally effective. Keep the mouth very wide open by placing
the bent knuckle of the thumb, or a cork about an inch in diameter,
between the teeth and try as hard as possible to articulate a ʒ
with the tip of the tongue against the teeth-ridge. The resulting
sound is very nearly the English fricative r. Some foreign people
obtain the sound more easily by trying the same exercise with, z
or a retracted ð instead of ʒ. The sound may often be improved
by pushing the tip of the tongue backwards with the end of a
pencil (the end of the pencil being placed underneath the tongue).

767. Thus foreign people may learn to pronounce r correctly in
rock, rat by inserting the cork between the teeth and trying to
produce the syllables ʒɔk, ʒæt, or zɔk, zæt, or ðɔk, ðæt (with a
retracted ð).

768. Foreign learners who are still unable to pronounce a
satisfactory fricative r may use a rolled or flapped lingual r.

769. Rolled lingual r is best acquired by imitation. If simple
imitation is not successful, the following well-known method may
be tried. Pronounce təˈdɑː… təˈdɑː… təˈdɑː… with dental
t's and alveolar d's, at first slowly and then with gradually increasing
speed. By keeping the tongue loose, and pronouncing this exercise
very fast, the d tends to become a kind of flapped r (§ 753), thus
ˈtrɑː… ˈtrɑː… ˈtrɑː…. When the flapped r has been thus
acquired, after a little practice the action can generally be extended
to the fully rolled sound.199

770. If this exercise is not successful, the best thing to do is
to practise all kinds of voiced alveolar fricative sounds (ʒ, z, ð
and other similar sounds), using considerable force of the breath
and trying to keep the tongue loose. It is useful to practise with
sudden jerks of the breath. After a little practice students usually
manage to hit upon the position in which the tongue will begin
to vibrate slightly. To attain a clear sustained r: often requires
considerable practice, say five or ten minutes a day for several
weeks.

771. Some learners can acquire rolled r more easily when d
is prefixed. Others again find it quite easy to make a breathed
rolled r (either with or without a prefixed t), and can by practice
acquire the voiced sound from it.

772. Words for practice: reason ˈriːzn, rid rid, red red, rash
ræʃ, raft rɑːft, wrong rɔŋ, raw rɔː, room rum, 41331 rule ruːl, run rʌn;
race reis, rope roup, right rait, round raund, royal ˈrɔiəl or rɔil;
real riəl, rarer ˈrɛərə, roar rɔə or rɔː, brewery ˈbruəri; recruit riˈkruːt,
retrograde ˈretrougreid, literary ˈlitərəri or ˈlitrəri.

773. Some foreign people, when they have learnt to realize the
fact that in London English the letter r is never sounded when
final or followed by a consonant, nevertheless still persist in trying
to give the effect of a r-sound by curling back, or ‘inverting’ as
it is technically called, the tip of the tongue while pronouncing
the vowel (see §§ 831-834).

774. Words for practising the omission of r: car kɑː, tar tɑː,
war wɔː, fur fəː, stir stəː, over ˈouvə, later ˈleitə, beer biə, care kɛə,
more mɔə or mɔː, tour tuə; part pɑːt, mark mɑːk, short ʃɔːt, corn
kɔːn, warm wɔːm, port pɔːt, 42332 force fɔːs, 42333 court kɔːt or kɔət, 42334 source
sɔːs or sɔəs, 42335 earn əːn, fern fəːn, girl gəːl, world wəːld, church
ʧəːʧ, hurt həːt, concert ˈkɔnsəːt, lizard ˈlizəd; pierce piəs, beard
biəd, weird wiəd, scarce skɛəs, stairs stɛəz, dared dɛəd, pears pɛəz
(= pairs and pares), gourd guəd, assured əˈʃuəd.

775. Fricative r may be considered as an element of the affricate
dr, and unvoiced fricative as an element of the affricate tr. For
details see §§ 624 ff.200

h

776. The letter h denotes the sound of pure breath having a
free passage through the mouth. This letter is used in transcribing
English and many other languages to represent any one of the
sounds produced when the mouth is held in a vowel-position and
air is emitted through the wide open glottis. The different varieties
of h are known as breathed glottal fricative consonants, since the
friction produced by the air passing through the glottis is the
feature common to all of them.

777. There are as many varieties of h as there are vowels. In
fact, h-sounds may be regarded as breathed (devoiced) vowels,
and they might in very narrow notation be represented by , ɑ̥,
, ə̥, etc.

778. The English h-phoneme comprises a great many members,
the variety used in any particular case being that which corresponds
to the vowel immediately following. Thus the h in hit
hit is a breathed i, the h in hard hɑːd is a breathed ɑ, the h in
hook is a breathed u, and so on. 43336 The variety of h which has
the mouth in the neutral vowel position (as in hurt həːt) may be
considered as the principal member of the phoneme.

779. There exists also a ‘voiced h,’ which is represented in
narrow transcription by the letter ɦ. It has the mouth-position
of a vowel but is pronounced with such strong exhaling-force that
the air produces considerable friction in the glottis besides causing
the vocal cords to vibrate.

780. Voiced h occurs in the speech of many English people
as a subsidiary member of the h-phoneme, when voiced sounds
both precede and follow. Thus voiced h may often be heard in
such words as behind biˈhaind, boyhood ˈbɔihud, perhaps pəˈhæps, 44337
inhabit inˈhæbit, the hedge ðə ˈheʤ, two hundred ˈtuː ˈhʌndrəd.
Some speakers would use the ordinary breathed h in such cases,
especially when speaking slowly. There is therefore no need for
the foreign learner to make any special effort to use voiced h.201

781. The h-phoneme is represented in ordinary English spelling
by the letter h; examples: heap hiːp, heavy ˈhevi, hard hɑːd, home
houm, inhabit inˈhæbit, dishearten disˈhɑːtn. It is represented by
wh in who huː, whole houl and their derivatives. The letter h
is silent in hour ˈauə, heir ɛə, honour ˈɔnə, honest ˈɔnist and their
derivatives; it is also often silent in unstressed syllables, and
notably in names ending in -ham, such as Balham ˈbælam, Wykeham
ˈwikəm (= Wickham and Wycombe) and in the words hedgehog
ˈheʤɔg or ˈheʤhɔg, vehicle ˈviːikl, annihilate əˈnaiəleit. 45338

image

Fig. 110. Tongue-position
of the Velar
Fricative x.

782. Speakers of French and other
Romance languages usually have considerable
difficulty in pronouncing h. They
generally leave it out altogether. Spaniards
usually replace it by the breathed velar
fricative (phonetic symbol x), that is, the
sound heard for instance in the Scottish loch
lɔx, Spanish jabon xaˈβon (Fig. 110). Those
who have this difficulty should bear in mind
that the h-sounds are simply vowels pronounced
with strong breath instead of with
voice. A near approach to the h-sounds in
hard hɑːd, he hiː, hook huk, etc., may be
obtained by whispering the vowels ɑː, , u, etc.

783. There is a peculiarity of French pronunciation which may
be made use of for acquiring the English h-sounds. In French,
final vowels are often devoiced, e.g. tant pis is often pronounced
ˈtɑ̃ ˈpi̥ with devoiced i, c'est tout is often sɛ ˈtu̥ with devoiced u.
In such cases the final sounds are simply varieties of h, being
the same as the h in hiː and being the same as the h in huː.

784. Most foreign people do not pronounce the h nearly strongly
enough in words beginning with hj, e.g. huge hjuːʤ, human
ˈhjuːmən, hue hjuː (= hew, Hugh). Those who have difficulty in
acquiring the correct pronunciation should notice that the h in
the group hj is very similar to the sound ç (the German ich-sound,
202§§ 820, 821). Many English people, in fact, actually use
the sound ç, pronouncing çuːʤ, etc. It is, therefore, often
advisable for foreign learners to adopt the forms with ç rather
than those with hj.

785. Words for practice: heat hiːt, hill hil, help help, hat hæt,
hard hɑːd, hot hɔt, hall hɔːl, hook huk, hoof huːf, hut hʌt, hurt
həːt; hay hei, hold hould, high hai, how hau, hoist hɔist; here hiə,
hair hɛə (=hare); behave biˈheiv, childhood ˈʧaildhud, buttonhook
ˈbʌtnhuk.

786. In educated English h is often dropped in unimportant
words such as him, her, have, when unstressed: thus I should have
seen him
is generally pronounced ai ʃəd əv ˈsiːn im. This omission
of the h of unstressed words is especially frequent when stressed
words beginning with h occur in the same sentence; thus in such
a sentence as she had her hat in her hand ʃiː hæd əː ˈhæt in əː ˈhænd
it would sound pedantic to sound the h in the words her. See
further, §§ 485, 487, 488.

787. h is occasionally dropped in initial unstressed syllables
of longer words, such as horizon, historical, hotel. Thus it would
be quite usual to pronounce on the horizon, from the historical
point of view
as ɔn ði əˈraizn, frəm ði isˈtɔrikl pɔint ə(v) vjuː.
Those who pronounce the h in hotel when said by itself would
often drop it in a good hotel ə ˈgud ouˈtel.

Initial and Final Voiced Fricatives

788. When one of the phonemes z or ʒ occurs initially or finally,
partially voiceless sounds are generally used. When initial, as
the z in zeal ziːl, the sound usually begins without voice and ends
with voice; when final, as the z in please pliːz, the sound usually
begins with voice and ends without voice. Similarly the ʒ in
garage ˈgærɑːʒ usually begins with voice and ends without voice.

789. When the phoneme is final and preceded by another consonant,
a completely voiceless sound is generally used. Thus with
most English speakers the z of heads hedz or sounds saundz (when
those words are said by themselves) is completely voiceless and
resembles a weak s. This voicelessness may be indicated in narrow
transcription by the mark,, and the pronunciation of the above
203words may be shown thus: heʣ̥, saunʣ̥ (also pronounced saund̥z̥
and saunz̥).

790. With some English speakers initial z and all final z and ʒ,
whether preceded by consonants or not, are completely voiceless.

791. French people often pronounce initial and final z and ʒ
with very full voice; the effect is unnatural to English ears. The
ordinary English pronunciation may be acquired by pronouncing
the sounds with very weak force.

792. Many other foreign people, and notably Germans and
Scandinavians, have difficulty in giving any voice to these sounds.
It is necessary for them to learn to make fully voiced z and ʒ
in isolation. This may be done as follows. Pronounce a long
əː, and, while this sound is going on, gradually bring the lower
lip against the upper teeth in the v-position; the əː-sound (voice)
must be continued while the lip is pressed close enough to the
upper teeth to give rise to friction as the air emerges. By this
means it is easy to acquire a fully voiced v. Then a similar
exercise will give the other fully voiced fricatives ð, v, and ʒ.

793. Very often foreign learners who have taken pains to acquire
fully voiced z and ʒ use them without modification in initial and
final positions. In fact they do purposely what so many French
people do unconsciously (§ 791). They can generally attain a
good pronunciation of the z-sounds in zeal, please by trying to
use a sound intermediate between z and s. Similarly they can
acquire the partially devoiced varieties of ʒ by aiming at a sound
intermediate between ʒ and ʃ. The sounds should be pronounced
with weak force, and should have greater resemblance to voiced
z and ʒ than to s and ʃ.

794. v and ð are subject to devoicing in similar circumstances,
though not to the same degree as z and ʒ. It does not sound wrong
to an English ear to hear such words as veal viːl, wave weiv, then
ðen, smooth smuːð pronounced with fully voiced v and ð.204

Chapter XXII
Frictionless Continuants

795. There exist voiced consonants which have the same or very
nearly the same articulatory positions as fricatives, but in which no
friction is audible; the absence of audible friction is due either
to the fact that less exhaling-force is used than for the corresponding
fricative, or to the fact that the aperture at the place of articulation
is somewhat wider, or to a combination of both these features.

796. Many English people pronounce r as a frictionless continuant
instead of as a fricative. The tongue-position required
for this variety of r is almost identical with that of fricative r
described in § 747, but the aperture between the tip of the tongue
and the teeth-ridge is slightly wider and the sound is produced
with less exhaling-force than fricative r. The sound is equivalent
to a weakly pronounced ‘retroflexed’ ə (§ 831). It may be denoted,
when necessary, by the special symbol ɹ.

797. Many of those whose r is a frictionless continuant use a
variety which has a certain protrusion of the lips when it occurs
at the beginning of a stressed syllable, as in red red, arrange
əˈreinʤ. In unstressed position, as in very ˈveri, miracle ˈmirəkl,
a r without lip-protrusion is generally used.

798. Examples of other frictionless continuants are a frictionless
B (corresponding to the fricative ʁ mentioned in § 763) and the
labio-dental frictionless continuant represented phonetically by ʋ
(§ 693). Frictionless ʁ is used by many Germans in final position;
consequently they often introduce it in such English words as
more, better, pronouncing moːʁ, ˈbetʁ instead of mɔː, ˈbetə, u is
used by many Indians in place of both w and v.

799. Frictionless continuant variants of w and j are also sometimes
heard in English in place of the ordinary semi-vowels
(§§ 802, 813). These variants may be indicated phonetically by
writing a length-mark after the letters w and j. Thus the interjections
well and yes are occasionally pronounced wːel, jːes. These
are single syllables, and are distinct from the sequences uːel, iːes
which would be disyllabic. It is weakness of exhaling-force which
causes the continuant and to be consonantal.205

Chapter XXIII
Semi-Vowels

800. Semi-vowels are defined as independent vowel-glides in
which the speech-organs start by forming a weakly articulated
close or fairly close vowel and immediately move to another sound
of equal or greater prominence; the initial vowel-position is not
held on for any appreciable time. It is the rapid gliding nature
of these sounds, combined with the use of rather weak force of
exhalation, that renders them consonantal. (See § 102.)

801. In English there exist two semi-vowels; they are represented
phonetically by the letters w and j. 1339

w

802. In pronouncing w the speech-organs start in position for
a variety of u and immediately leave this for some other vowel
position, or occasionally for one of the consonants l or j. The starting
point varies slightly with different speakers and according to the
vowel following, but for the purposes of practical teaching it may
be considered to be a variety of u with the lips pursed up to about
the same degree as for the English ‘long’ (Fig. 49). The position
of this starting point may therefore be described as follows. The
lips are closely rounded; there is considerable raising of the back of
the tongue in the direction of the soft palate; the soft palate is in its
raised position; the vocal cords are made to vibrate so that voice
is heard. The formation of w may be expressed shortly by defining
the sound as a labio-vehr semi-vowel.

803. The glide away from the above-mentioned starting point
may be considered as the principal member of the English w-phoneme.
The subsidiary members are not important. It may,
however, be noted that the lip-rounding is closer when long
follows (as in woo wuː), and may be less close before vowels remote
from u (as in wide waid).206

804. On the other hand, if w is pronounced emphatically before any
vowel the lip-rounding may be closer
than that of (compare Figs. 111 and
49). For practical purposes therefore
the existence of subsidiary members of
the w-phoneme may be ignored.

image

Fig. 111. Lip-position of the
beginning of w in emphatic
pronunciation.

805. w is the consonantal SOUnd
of the letter w. It is used when w
occurs at the beginning of a syllable
(except in the group wr, in which the w is silent) or is preceded
by a consonant, e.g. wait weit, away əˈwei, twelve twelv. U is
generally pronounced in this way when preceded by q, e.g. quite
kwait, 2340 and often when preceded by g in unstressed syllables, e.g.
language ˈlæŋgwiʤ. Note the exceptional words one wʌn, once
wʌns, choir ˈkwaiə, suite swiːt (= sweet). 3341 Examples of w followed
by l and j are equal (ˈiːkwl), colloquial (kəˈloukwjəl) (alternative forms
of ˈiːkwəl, kəˈloukwĭəl).

806. The English sound w causes difficulty to many foreign
people, and especially to Germans. They generally replace it by
a different kind of bi-labial fricative, namely one in which the lips
are kept flat instead of being rounded and pushed forward, and
in which the tongue is in a neutral position instead of being raised
at the back. The phonetic symbol for this consonant is β. Its
lip-position is the same as that of ϕ (Fig. 89). It is a sound intermediate
in acoustic effect between w and v; it is very frequently
heard in German words like Quelle ˈkβɛlə or ˈkvɛlə, zwei tsβai or
tsvai. Sometimes foreign learners replace w by v.

807. The best way of acquiring w is to substitute the vowel
for it, and gradually to shorten this . Germans should begin
by practising win win, well wel, for instance, as uiːn, uːel, etc.
It is also very useful to practise the exercise uːəːuːəː… with
energetic motion of the lips. The motion of the lips in this exercise
should be entirely horizontal (exactly as for uːiːuːiː…); most
foreign learners have an almost irresistible tendency to pass from
207the to the əː by a vertical motion of the lower jaw. It will be
found helpful to practise this exercise with the teeth kept tightly
together.

808. The English sound w must be distinguished from the
French consonantal sound which is heard at the beginning of huit
(phonetically ɥit) and which French people often
substitute for w in some English words, such as
persuade pəˈsweid, Swiss swis. ɥ is the semi-vowel
corresponding to the French sound of u
(y); in forming its starting-point the lips have
a position similar to that of w, but there is
a simultaneous raising of the ‘front’ of the
tongue towards the hard palate. Fig. 112 is
a palatogram of ɥ. The English w gives no
palatogram.

image

Fig. 112. Palatogram
of French ɥ
in the group ɥə.

809. Words for practising w: we wiː, with wið, wet wet, wag wæg,
quaff kwɑːf, want wɔnt, warm wɔːm, wool wul, wound (injury) wuːd,
won, one wʌn, word wəːd; wake weik, won't wount, wife waif, wound
(past of verb wind) waund; weird wiəd, wear wɛə, wore wɔə, 4342 wooer
wuə 5343; waver ˈweivə, equivalent iˈkwivələnt. The following sentence
affords good practice for foreign learners who have difficulty with
w: we would work if we were wise wiː ˈwud ˈwəːk if wiː wə ˈwaiz.

810. The breathed consonant corresponding to w (phonetic
symbol ʍ) is used by many English people in words spelt with
wh. Thus what, which are often pronounced ʍɔt, ʍiʧ. This
pronunciation, with a variant hw which is difficult to distinguish
from it, is regularly used in Scotland, Ireland, the North of England
and in America. In the South the more usual pronunciation of
these words is wɔt, wiʧ, etc., though the use of ʍ or hw is sometimes
taught as being more ‘correct’.

811. ʍ, being a breathed sound, is a fricative consonant and
not a semi-vowel. The friction is always clearly audible. (Semi-vowels
have no audible friction, see definition of a vowel, § 97.)

812. ʍ may be defined shortly as a breathed labio-velar fricative
consonant.208

j

813. In pronouncing the most usual English j the speech-organs
start at or near the position for the English ‘short’ i (§§ 254, 255)
and immediately leave this for some other sound of equal or greater
prominence. The sound following j is generally a vowel, but it
may on occasion be one of the consonants l or w. It is the glide
away from i that constitutes the consonant j. The starting-point
of the principal English j may be described as follows. The front
of the tongue is raised rather high in the direction of the hard
palate (as for i, Fig. 34); the lips are spread; the soft palate is in
its raised position; the vocal cords are made to vibrate, so that
voice is heard. The formation of j may be expressed shortly by
denning the sound as an unrounded palatal semi-vowel.

814. The above is a description of the principal member of the
English j-phoneme. The actual sound used in particular words
depends to some extent on the nature of the following vowel.
The starting-point of j is generally closer than the following vowel.
Thus it is very close indeed before , as in yield jiːld, but much
less close before such sounds as ɑ or ɔ, as in yard jɑːd, yacht jɔt.
In such a case as four yards ˈfɔː ˈjɑːʣ the j hardly rises above
the position of English Vowel No. 3 (e). It is thus possible to
distinguish several subsidiary members of the English j-phoneme.
The distinctions, however, are unimportant, and may be ignored
in the practical teaching of English.

815. The palatogram of the j in the group jɑː is practically
identical with the palatogram of lax i (Fig. 39).

816. j is the consonantal sound of the letter y; examples: yes
jes, vineyard ˈvinjəd. I and e often have the value j when the
following sound is e; examples: onion ˈʌnjen, familiar feˈmilje,
simultaneous simelˈteinjes. 6344 Examples of j followed by l and w are
labial (ˈleibjl), arduous (ˈaːdjwes) (alternative forms of ˈleibjel or
ˈleibĭel, ˈaːdjŭes).

817. In words spelt with u, ue, ui, ew, and eu, representing
long , j is sometimes inserted before the (as in uniform ˈjuːnifɔːm,
209few fjuː) and sometimes not (as in rule ruːl, chew tʃuː). The
rules with regard to this are as follows. (i) The j is never inserted
after ʧ, ʤ, or r, or after l preceded in turn by a consonant; examples:
chew ʧuː, June ʤuːn, rule ruːl, blue bluː. (ii) The j is regularly
inserted after p, b, t, d, k, g, in, n, f, v, h; examples: pew pjuː,
beauty ˈbjuːti, tune tjuːn, due djuː, queue kjuː, argue ˈɑːgjuː, music
ˈmjuːzik, new njuː, few fjuː, fugue fjuːg, view vjuː, huge hjuːʤ.
(iii) The j is regularly inserted after l preceded by a vowel, when
that preceding vowel is stressed (examples: deluge ˈdeljuːʤ, value
ˈvæljuː 7345) or semi-stressed (example: aluminium ˌæljuːˈminjəm 8346).
(iv) Usage varies in words in which l is initial or preceded by an
unstressed vowel; thus lute, absolute are pronounced ljuːt, ˈæbsəljuːt
by some, and luːt (like loot), ˈæbsəluːt by others; the forms with
j are recommended by some teachers, but the forms without j
appear to be the more usual in ordinary speech, at any rate in the
commoner words, (v) After s, z, and θ usage also varies; thus
suit, presume, enthusiasm are pronounced sjuːt, priˈzjuːm,
inˈθjuːziæzm by some and suːt, priˈzuːm, inˈθuːziæzm by others;
I use the forms with j.

818. Some foreign people, and especially North Germans, use a
fricative j; the effect is somewhat strange to English ears. Fricative
j is made by holding the tongue in position for a rather close i and
producing voice with considerable exhaling-force. The English
semi-vowel j may easily be acquired by observing its gliding nature
and by diminishing the force of exhalation.

819. Words for practising j: yield jiːld, Yiddish ˈjidiʃ, yet jet,
yak jæk, yard jɑːd, yacht jɔt, yawn jɔːn, yew juː (= you), young
jʌŋ, yearn jəːn; Yale jeil, yolk jouk; year jiə or jəː, your jɔə, 9347
youˈre juə; beyond biˈjɔnd, 10348 million ˈmiljən, India ˈindjə.

820. The breathed consonant corresponding to j (phonetic
symbol ç) is used by some English people in place of hj in such
words as huge hjuːʤ or çuːʤ, human ˈhjuːmən or ˈçuːmən. A
gliding ç (corresponding to the semi-vowel j) must be considered
as a fricative consonant, since the friction is clearly audible; in
210fact it is almost impossible in ordinary connected speech to
distinguish by ear between a gliding ç and a continuous fricative
ç (corresponding to fricative j).

image

Fig. 113. Tongue-position
of the Palatal
Fricative ç.

821. The continuant g may be termed the breathed palatal
fricative
. It is one variety of the German ‘ich-sound.’ Its tongue-position
is that shown in Fig. 113 or somewhat opener than this. 11349211

Chapter XXIV
Nasalization

822. When sounds (other than plosive and nasal consonants) are
pronounced with simultaneous lowering of the soft palate, so that
the air passes through the nose as well as through the mouth, they
are said to be nasalized. Nasalized sounds are generally represented
in phonetic transcription by the mark ̃ placed above the symbol
of the normal sound. The best known cases of nasalized sounds
are the French vowels ɛ̃, ɑ̃, õ (or ɔ̃), œ̃ heard in vin vɛ̃, sans sɑ̃,
bon (or bɔ̃), un œ̃. Such sounds do not occur in Received
English.

823. Some foreign people are apt to nasalize vowels whenever a
nasal consonant follows: thus French people often pronounce jam,
hand, won't as ʤæ̃m, hæ̃nd, wõnt, instead of ʤæm, hænd, wount;
the Portuguese regularly pronounce the English word tense (which
should be tens) as tɛ̃ns or even tɛ̃s. The Dutch and many South
Germans have a similar tendency; with these the nasalization is
especially noticeable in the diphthongs, e.g. wãĩn or βãĩn instead
of wain (wine). Some foreign people nasalize all vowels or at any
rate all the more open vowels independently of any nasal consonant.
Such nasalization is abnormal when introduced into English.

824. Those who habitually nasalize their vowels 1350 often have
difficulty in getting rid of the fault. It can be cured by constant
practice of isolated vowel sounds. It is better to start practising
with close vowels, such as , , there being always less tendency
to nasalize these. It is also a good plan to pronounce z before
each vowel, because z is a sound which cannot be nasalized without
losing most of its characteristic quality. When by means of
exercises such as ziːziː… zuːzuː… the student is enabled to
pronounce a pure and , which should not require much practice,
the opener vowels may be rendered pure by exercises such as
212ieie…, uouo…, iaia…, uɔuɔ… pronounced without break
between the i and e, u and o, etc. When all the isolated vowels
can be pronounced without nasalization, easy words should be
practised. The greatest difficulty will probably be found in words
in which the vowel is followed by a nasal consonant, e.g. wine
wain, want want; such words should therefore be reserved till the
last. In practising a word such as wain a complete break should
at first be made between the ai and the n, thus wai-n; this interval
may afterwards be gradually reduced until the normal pronunciation
is attained.

825. It can be shown by experimental methods that slight
nasalization of vowels occurs in English when nasal consonants
follow. Such nasalization is, however, not sufficient to give to
the vowels the characteristic nasal tamber. For the purposes of
practical teaching it is therefore necessary to state definitely that
vowels are not nasalized in normal British English.

826. Words for practice: stream striːm, limb lim (= Lympne),
stem stem, jam ʤæm or ʤæːm, calm kɑːm, Tom tɔm, form fɔːm,
room rum, 2351 boom buːm, come kʌm, germ dʒəːm; game geim, home
houm, time taim; seen, scene siːn, tin tin, then ðen, ran ræn, man
mæn or mæːn, barn bɑːn, can't kɑːnt, on ɔn, corn kɔːn, spoon spuːn,
one wʌn, learn ləːn, rain rein, alone əˈloun, wine wain, town taun,
coin kɔin; end end, hand hænd, pond pɔnd, warned wɔːnd, under
ˈʌndə; owned ound, find faind, found faund, joined ʣɔind.213

Chapter XXV
Retroflex Sounds

827. Retroflex sounds (also called ‘cerebral,’ ‘cacuminal’ or
‘inverted’ sounds) are those in the formation of which the tip of
the tongue is curled upwards towards the hard palate. Thus
‘retroflex t’ and ‘retroflex d’ are plosive consonants made by
articulating with the tip of the tongue
against the hard palate as shown in
Fig. 114.

image

Fig. 114. Tongue-position
of the retroflex consonant ʈ.

828. The principal retroflex consonants
are represented phonetically by the letters
ʈ ɖ ɳ ɭ ʂ ʐ. A retroflex rolled consonant
is not known to occur in any language,
so the letter ɽ is used to represent the
‘retroflex flap’ formed by placing
the tongue in a retroflex position with the
tip near the hard palate and then
shooting it forwards and downwards
in such a way that the under side of the tongue strikes the teethridge
producing a flapped sound of very characteristic quality.
The retroflex frictionless continuant of this series is near in quality
to the alveolar ɹ and does not as a rule require a separate symbol.

829. Retroflex consonants do not occur in Received English;
they may be heard in some foreign pronunciations. Indians
generally use ʈ and ɖ in place of the English alveolar t and d;
they pronounce ten as ʈen, did as ɖiɖ, to-day as ʈuɖeː (instead of
təˈdei). Norwegians and Swedes often use retroflex consonants in
words which have the spelling r + alveolar consonant-letter; thus
they pronounce part, hard, barn, pearl, first (which are in South-Eastern
English pɑːt, hɑːd, bɑːn, pəːl, fəːst) as paːʈ or paɹʈ, haːɖ or
haɹɖ, baːɳ or baɹɳ, pəɹɭ or pəɹɭ, fəːʂʈ or fəɹʂʈ.The correct English
alveolar sounds are easily acquired by feeling with the tip of the
tongue the appropriate place of articulation on the teeth-ridge.214

830. ɽ is sometimes heard from English people in place of
the ordinary flapped r after short vowels; thus the interjections
hurry up!, sorry! may sometimes be heard as ˈhʌɽi ˈʌp, ˈsɔɽi.
This pronunciation appears to be merely a peculiarity of individual
speakers, and does not belong to any recognized dialect.

831. Vowels may be ‘retroflexed,’ i.e.
pronounced with retroflex modification. In
making retroflexed vowels the main body
of the tongue is held as for an ordinary
vowel, but the tip of the tongue is simultaneously
curled up towards the hard
palate with lateral contraction (§ 73). The
resulting sounds have a peculiar hollow
quality. 1352 Fig. 115 shows the approximate
tongue-position of a retroflexed ɑ.

image

Fig. 115. Tongue-position
of a retroflexed ɑ.

832. Retroflexed vowels may be represented in phonetic transcription
by superposing ɹ on the vowel letter, thus ɑɹ, ɔɹ, əɹ, etc., or
by means of digraphs, ɑɹ, ɔɹ, əɹ, etc., with the convention that
each of these sequences of two letters represents only a single
sound. Retroflexed ə (əɹ) may be represented more simply by ɹ
alone (see § 796). Some writers prefer to represent these sounds
by letters with a ‘retroflex modifier,’ thus ɑ, ɔ, ɚ. 2353

833. Retroflexed vowels do not occur in the type of English
taken as a model for the purposes of this book. They are, however,
found in many other types of English in words written with r final
or r + consonant, such as far, garden, door, sort, verse. The use
of retroflexed vowels in such words is not confined to local dialects,
but may be heard in the speech of many educated English people,
and particularly of those who come from the South-West of England;
retroflexed vowels are also a characteristic feature of Irish and
American English. Some English people use retroflexed ə(ː), but
not any other retroflexed vowels; such speakers pronounce bird,
verse, murmur, as bəɹːd, vəɹːs, ˈməɹːməɹː, but garden, door as ˈgɑːdn,
dɔː or dɔə, etc.215

834. Some foreign people, and especially Norwegians and Swedes,
use retroflexed vowels in speaking English in much the same
way as the South-Western Englishman does. If they wish to
acquire the ordinary pronunciation of educated Londoners, they
must avoid this retroflexion; it may be easily avoided by keeping
the tongue-tip firmly pressed against the lower teeth while pronouncing
the vowels in such words as garden, door, verse, murmur,
bark, curve. Words for practice are given in § 774.216

Chapter XXVI
Similitude. Assimilation

Similitude

835. It often happens that a particular sequence of two phonemes
involves the use of a certain subsidiary member of one of them
which has a greater resemblance to a neighbouring sound than
the principal member has. In this case there is said to be similitude
between that subsidiary member and the neighbouring sound.
Thus a partially breathed l () is used in English in such words
as please pliːz, play plei (see § 845 (i) a), and we say there is
similitude between this l and the p. 1354

836. Examples of similitude may be stated by means of a
formula of the following type: the subsidiary sound B belonging to the
phoneme whose principal member is the sound A is used when the sound
C is adjacent to it or near to it
. Thus the example in § 835 may
be stated thus: the subsidiary sound (partially breathed l) belonging
to the English phoneme whose principal member is a fully voiced l
is used when p precedes in a stressed syllable.

Assimilation

837. Assimilation is defined as the process of replacing a sound
by another sound under the influence of a third sound which is
217near to it in the word or sentence. The term may also be extended
to include cases where a sequence of two sounds coalesces and
gives place to a single new sound different from either of the
original sounds; this type of change may be termed ‘coalescent
assimilation.’

838. Assimilations are of two chief kinds, historical and contextual.
By a ‘historical assimilation’ we mean an assimilation
which has taken place in the course of development of a language,
and by which a word which was once pronounced in a certain way
came to be pronounced subsequently in another way. By a ‘contextual
assimilation’ we mean one which is occasioned when words
are juxtaposed in a sentence, or in the formation of compounds, and
by which a word comes to have a pronunciation different from that
which it has when said by itself.

839. An example of historical assimilation is the change of m
to n which has taken place in the word ant ænt. In the thirteenth
and fourteenth centuries this word was written amete and amte,
and was no doubt pronounced ˈæmətə and (later) ˈæmtə and æmt;
spellings with n in place of m first appeared in the fifteenth century,
clearly indicating the change to the modern pronunciation ænt. An
example of historical coalescent assimilation is the reduction of the
sequence tj to the affricate ʧ in such a word as picture ˈpiktʃə which
some hundreds of years ago was doubtless pronounced ˈpiktjur.

839a. An example of contextual assimilation is the change of
s to ʃ when horse hɔːs and shoe are put together and form horse-shoe
ˈhɔːʃʃuː. An example of contextual coalescent assimilation is when
don't dount and you juː are put together and pronounced ˈdounʧu,
as is frequently done.

840. Changes of the kind mentioned in § 839 which occur when
a word is borrowed from one language into another may be considered
as particular cases of historical assimilation. Such a case
is the change of s to z when the English roast beef is borrowed into
French, where it is pronounced rɔzbif.

841. Historical assimilations are conveniently described by means
of the following formulae: (i) (for ordinary assimilations) the sound
A has been replaced by the sound B under the influence of the sound C
,
(ii) (for coalescent assimilations) the sounds A and C have influenced
218each other and coalesced into the single sound B
. Thus in the two
examples quoted in § 839 the assimilations may be described thus:
(1) m has been replaced by n under the influence of t, (2) t and j
have influenced each other and have coalesced into the single
sound ʧ.

841a. Contextual assimilations may be described by the following
formulae: (i) (for ordinary assimilations) the sound A is replaced by
the sound B under the influence of the sound C
, (ii) (for coalescent
assimilations) the sounds A and C influence each other and coalesce
into the single sound B
. Thus in the two examples quoted in
§ 839a the assimilations may be described thus: (1) s is replaced
by ʃ under the influence of ʃ, (2) t and j influence each other and
coalesce into the single sound ʧ.

842. The distinction between similitude and assimilation should
be carefully observed. 2355 The term ‘similitude’ is used to describe
an existing fact; ‘assimilation’ is a process by which certain pronunciations
are evolved. (Thus it would not be accurate to say
that the use of a partially breathed l in please is a case of
‘assimilation.’ Such a statement would imply that the l of please
had at one time been fully voiced and had subsequently lost part
of its voice owing to the presence of the p; there is, on the contrary,
every reason to believe that the l in this word has had its present
value ever since the word first appeared in the language.)

843. It is likely that many similitudes have been arrived at
by a process of assimilation, but it is generally not possible to
tell this with any degree of certainty.

Types of Similitude

844. Similitudes are of various kinds. The most important are
(i) resemblances in the matter of voice or breath, (ii) resemblances
in tongue-position in the case of a consonant, (iii) resemblances in
lip-position in the case of a consonant, (iv) vowel harmony, (v)
resemblance of a vowel to an adjacent consonant, (vi) nasality in
phonemes of which the principal member is not nasal.219

Examples of Similitude

845. The following are some noteworthy examples of similitudes
occurring in English. The types are numbered as in § 844.

(i) a. When the phonemes mnlrwj are immediately preceded
by a voiceless censonant in a stressed syllable, partially breathed
varieties of m n l r w j are used. Examples: small smɔːl, sneeze
sniːz, place pleis, cream kriːm, quite kwait, pew pjuː. 3356 The pronunciation
may be represented in narrow (allophonic) transcription
thus: sm̥ɔːl, sn̥iːz, etc.

b. When the h-phoneme occurs medially between voiced sounds,
a voiced h (ɦ) is used by most English people (§§ 779, 780).
Examples: behind biˈhaind, adhere ədˈhiə, inhabit inˈhæbit (narrow
transcription biˈɦaind, ədˈɦiə, inˈɦæbit).

(ii) Different varieties of k and g are used before different vowels
(§§ 533, 543). Different varieties of ŋ are used after different
vowels (§ 651). Different varieties of h are used before different
vowels (§ 778). Special varieties of t are used before θ and r,
as in eighth eitθ, at rest ət ˈrest (§ 513). A labio-dental m is used
by some speakers before f and v (§ 641). A dental n is used before
θ and ð, as in one thing ˈwʌn ˈθiŋ, on the ground ɔn ðə ˈgraund.

(iii) Lip-rounded varieties of k, g and other consonants are
used before w, as in queen kwiːn, language ˈlæŋgwiʤ (§§ 533, 543,
and footnote 9 on p. 185).

(iv) No example in modern English. 4357

(v) When the long is preceded by j, an advanced variety of
vowel (üː) is used (§ 326); examples: music ˈmjuːzik, deluge
ˈdeljuːʤ. When a front vowel phoneme is followed by dark l,
220a somewhat lowered and centralized sound is employed. Thus the
e in well is an opener and more retracted vowel 5358 than that in
get, and the æ in alphabet ˈælfəbit is more a-like than the æ in
alley ˈæli (which has a clear l). See also the dialectal examples
bowl, rule referred to on p. 224.

(vi) Vowels have slight nasality before nasal consonants (§ 825)

Types of Assimilation

846. There exist various types of assimilation parallel to the
various types of similitude. The most important are (i) assimilations
of breath to voice and voice to breath, (ii) assimilations
affecting the position of the tongue in pronouncing consonants,
(iii) assimilations affecting the position of the lips in pronouncing
consonants, (iv) assimilations by which a vowel is affected by
another vowel, (v) assimilations by which a vowel is modified by
an adjacent consonant, (vi) assimilations affecting the position of
the soft palate.

Examples of Historical Assimilation

847. The following are examples of historical assimilation. The
types are numbered as in § 846.

(i) For the words width, breadth, amidst, many people use the
pronunciation witθ, bretθ, əˈmitst instead of the commoner and
presumably older pronunciation widθ, bredθ, əˈmidst. Here d has
been replaced by t under the influence of θ or s.

In absorption əbˈsɔːpjn, description disˈkripʃn, action ˈsekʃn,
election iˈlekʃn, direction diˈrekʃn, affliction əˈflikʃn, etc., it may
be presumed that in Latin times an assimilation took place by
which b (or g) was replaced by p (or k) under the influence of a
following t. 6359 (The ti has undergone a further change to ʃ in
Modern English — an assimilation of type (ii).)

If, as is likely, the final s of plurals such as dogges, wordes was
pronounced in Early English as s (ˈdɔgəs, ˈwɔrdəs), then the
resemblance seen in the modern dogs dɔgz, words wəːʣ has been
221arrived at by an assimilation: s has been replaced by z under
the influence of g, d, etc. 7360 (More modern words, such as globes
gloubz, jugs ʤʌgz, tunnels ˈtʌnlz, schemes skiːmz are formed by
analogy, but no assimilation has taken place, since these plurals
have never been pronounced otherwise than with z.)

image

Fig. 116. Tongue-positions of t,
j, and ʧ (characteristic features
somewhat exaggerated).
— t ––– j … … ʧ.

(ii) Words like picture ˈpiktʃə, question ˈkwestʃən were no doubt
at one time pronounced ˈpiktjur, ˈkwestjən, etc. In such cases tj
has undergone a eoalescent assimilation resulting in ʧ. The nature
of this assimilation is shown in Fig. 116. Similarly in grandeur
ˈgrænʤə, which was doubtless formerly pronounced ˈgrændjur, the
sounds d and j have influenced each other and have coalesced into
the affricate ʤ.

The words sure and sugar, which are now pronounced ʃuə,
ˈʃugə, were probably pronounced in former times sjuːr, ˈsjugər.
There has been a coalescence (§ 837) of sj to ʃ, the s and j having
influenced each other. The way in which this assimilation has
worked may be seen by comparing the tongue-position of s
(Fig. 94) with those of j and ʃ (Fig. 99). The same assimilation
has taken place in unstressed position in a large number of words,
e.g. in most of those ending in -tion, -tial, -cial, -cian, and consonant
letter + -sion or -sure, such as nation ˈneiʃn (formerly ˈnɛːsjən),
222position pəˈziʃn (formerly poˈzisjən), portion ˈpoːʃn (formerly
porsjən), essential iˈsenʃl (formerly eˈsensjəl), musician mjuːˈziʃn
(formerly miuˈzisjən), mansion ˈmænʃən (formerly ˈmænsjən),
permission pəˈmiʃn (formerly pərˈmisjən), pressure ˈpreʃə (formerly
presjur).

The corresponding assimilation of the voiced zj to ʒ is also
fairly common in unstressed position. Examples are seen in vision
ˈviʒn (formerly ˈvizjən), measure ˈmeʒə (formerly ˈmezjur), azure
ˈæʒə (formerly ˈæzjur or ˈɛːzjur).

The majority of the ʒ's occurring in Modern English have been
arrived at by this assimilation.

The word handkerchief was doubtless at one time pronounced
ˈhæn(d)kəːʧif; it is now ˈhæŋkətʃif. This development illustrates
a historical assimilation by which n has been replaced by ŋ under
the influence of k.

The pronunciation ˈbeikŋ which is often heard for bacon instead
of the more usual ˈbeikən or ˈbeikn illustrates a similar historical
assimilation: here again n has been replaced by ŋ under the
influence of k. 8361

The ‘fronting’ of k in the development of such a word as chin
is also a historical assimilation of type (ii). This word is believed
to be derived from an early form kinn which subsequently became
cin 9362 and eventually ʧin. A comparison of Figs. 29, 30 with
the tongue-position of an i (Fig. 12) shows the nature of the
assimilation: k was replaced by c under the influence of i. 10363

(iii) The words happen, open are pronounced by some people
hæpm, ˈoupm, instead of in the more usual and presumably older
way ˈhæpən or ˈhæpn, ˈoupən or ˈoupn. In these cases n has been
replaced by m under the influence of p.223

(iv) The old English mutations are good examples of vowel
harmony. It is believed that men men is derived from an early
form ˈmaniz, and that myːs, the old English form of mice, 11364 was
derived from an earlier ˈmuːsiz. In the first case a was replaced
by e under the influence of i, and in the second case u was replaced
by y under the influence of i.

(v) The common pronunciation of children as ˈʧuldrən furnishes
an example of historical assimilation of type (v). Here i has
been replaced by u under the influence of the following dark l.
The occasional pronunciation of pretty (normally ˈpriti) as ˈpruti
is probably also due to assimilation: i has become u under the
influence of pr.

In the pronunciation of those whose speech shows marked
traces of London dialect many examples may be found of modifications
of other vowels under the influence of a following dark l.
Thus with many Londoners bowl and rule have vowel-sounds
quite distinct from those in bowling and ruling (where the l is
clear); an extreme form of these differences may be shown thus:
bɔul, ˈböüliŋ, roil, ˈrïüliŋ. (For ö see § 398.) In such pronunciations
of bowl and rule, the normal English ou is represented by an ɔu-tike
sound and the normal English is represented by an -like sound
before dark l; the similitudes shown in these words have presumably
been arrived at by historical assimilation.

There exists a not uncommon pronunciation of the adverb just
as ʤest. This form is probably the result of an assimilation: A
has been replaced by e under the influence of the adjacent ʤ
and s.

(vi) Historical assimilations affecting the position of the soft
palate are not common in English. An example is seen in the
pronunciation of some people who substitute nn for nd before
unstressed vowels in some words, who pronounce for instance
individual as inniˈvidjŭəl. Here d has been replaced by n under
the influence of n. 12365224

848. It will be seen that historical assimilations account for
most of what are known as ‘combinative’ sound-changes.

Examples of Contextual Assimilation

849. The following are some examples of contextual assimilation.
The types are numbered as in § 846.

(i) The full pronunciation of is is iz. In connected speech the
word is generally reduced to z when preceded by a voiced
sound other than z or ʒ; e.g. Who is there? ˈhuː z ˈðɛə, dinner
is ready
ˈdinə z ˈredi, When is he coming? ˈwen z (h)iː ˈkʌmiŋ. 13366
And it is reduced to s when preceded by a breathed consonant
other than s or ʃ; e.g. it is ready itsˈredi, that is all ˈðæt s ˈɔːl,
What is the time? ˈwɔt s ðə ˈtaim, the shop is open ðə ˈʃɔp s ˈoupn,
Mr. Smith is coming mistə ˈsmiθ s kʌmiŋ. 14367 The latter case
illustrates contextual assimilation: the z of the full pronunciation
is replaced by s under the influence of a preceding voiceless
consonant.

Has is treated similarly. Its strong form is hæz, and its weak
forms are həz, əz, z and s, the latter occurring as the result of
contextual assimilation after voiceless consonants (other than
s and ʃ): Who has been here? ˈhuː z bin ˈhiə, John has finished
ˈʤɔn z ˈfiniʃt, but Jack has been here ˈʤæk s bin hiə, What has
he done?
ˈwɔt s (h)iː ˈdʌn.

The expression used to, meaning ‘accustomed to’ or ‘was in the
habit of,’ is now generally pronounced ˈjuːst tu or ˈjuːs tu. 15368 Contextual
assimilations from voice to breath thus take place when to is
added to used: d is replaced by t under the influence of the following
t, and z is replaced by s under the influence of t. 16369225

In newspaper ˈnjuːspeipə, fivepence ˈfaifpəns we see the effect
of contextual assimilation in compound words. News, five by
themselves are pronounced njuːz, faiv; in the compounds the z
and v are replaced by s and f under the influence of the following p.

I should have thought so is sometimes pronounced in rapid
colloquial speech ai ʃt f ˈθɔːt sou, instead of the more usual ai
ʃəd əv ˈθɔːt sou
. Here d and v are replaced by t and f under the
influence of the following θ.

(ii) A common instance of contextual assimilation of type (ii)
is the replacement of s by ʃ under the influence of a following
ʃ. Examples are horse-shoe ˈhɔːʃʃuː, and such expressions as
this shop, of course she does, just shut the door, which are very
commonly pronounced ˈðiʃˈʃɔp, əv ˈkɔːʃ ʃi dʌz, ˈʤʌʃ ˈʃʌt ðə ˈdɔː.
The replacement of z by ʒ under the influence of a following ʃ is
also common: Does she?, butcher's shop are generally pronounced
dʌʒ ʃiː, ˈbutʃəʒ ʃɔp.

Another contextual assimilation made by many English people
is the replacement of a terminal s or z by ʃ or ʒ under the influence
of an initial j in the next word: ˈðiʃ ˈjəː for ˈðis ˈjəː (this year),
ˈtelʒ ju for ˈtelz ju (tells you), ˈmisiʒ ˈjʌŋ for ˈmisiz ˈjʌŋ (Mrs.
Young
). (It is not necessary for foreign learners to make this
assimilation.)

Many English people make the contextual assimilation of replacing
initial unstressed j by ʃ or ʒ when the preceding word in the sentence
ends in t or d, or of making tj, dj coalesce into the affricates ʧ, ʤ
in such sequences: ˈʃʌt ʃr ˈəiz (or ˈʃʌʧˈrəiz) for ˈʃʌt jər ˈəiz 17370 (shut
your eyes
), ˈʃʌt ʃ ˈmauθ (or ˈʃʌʧˈmauθ) for ˈʃʌt jə ˈmauθ 18371 (shut your
mouth
), ˈdiʤuː for ˈdid juː (Did you?). (It is likewise not necessary
for foreign learners to make this assimilation.)

In careless speech other contextual assimilations of type (ii) are
continually made. One further example must suffice. I am going
to buy some
is normally pronounced ai m goiŋ tə ˈbai sʌm, but
this is often reduced in careless speech to aiŋənə ˈbai sʌm. Here
m is replaced by ŋ under the influence of g, and ŋ is replaced by
n under the influence of t, and then the g and t are elided. Assimilations
which are confined to careless speech may be termed ‘negligent’
226assimilations (see my Pronunciation of English, 1050 and subsequent
editions, §§ 404, 405).

(iii) Contextual assimilations of type (iii) are not unfrequently
made, especially by careless speakers. The following are some
examples: ˈtem ˈminits for ˈten ˈminits (ten minutes), where n is
replaced by m under the influence of the following m; sm ˈpɔːlz,
sm ˈpæŋkrəs for sn(t) ˈpɔːlz, sn(t) ˈpæŋkrəs (St. Paul's, St. Pancras),
ˈstæm-pɔint for ˈstæn(d)-pɔint (standpoint), ˈlʌdəm ˈbridʒ for
ˈlʌndən ˈbriʤ (London Bridge), ai l ˈsuːm ˈbriŋ ðəm for ai l ˈsuːn
briŋ ðəm
(I'll soon bring them), in which cases n is replaced by
m under the influence of p or b. Tadpole, which is normally
ˈtædpoul, is sometimes pronounced ˈtæbpoul, d being replaced by
b under the influence of p. I don't believe it, which is normally
ai ˈdount biˈliːv it, is sometimes reduced to ai ˈdoump ˈbliːv it or
ai ˈdoum ˈbliːv it, where t is replaced by p under the influence of
b, and n is replaced by m under the influence of p or b. Similarly,
it can't be done (normally it ˈkɑːnt bi ˈdʌn) is often reduced to
it ˈkɑːmp bi ˈdʌn.

Foreign learners are not recommended to make any of these
assimilations of type (iii), but they should observe to what extent
they occur in the speech of English people.

(iv) Contextual assimilations of type (iv) are rare in English.
An example is seen in the use of i in the ordinary pronunciation
of we are (wiə); here is replaced by i under the influence of ə.

(v) Contextual assimilations of type (v) are likewise not common.
The most noteworthy are the replacement of ə by i or u under the
influence of a following j or w. Thus it is not uncommon to hear
What are you doing?, Which way are you going?, Go away pronounced
wɔt i ju ˈduiŋ, ˈwitʃ ˈwei i ju ˈgoiŋ, ˈgou uˈwei instead of the
normal ˈwɔt ə ju ˈduiŋ, ˈwitʃ ˈwei ə ju ˈgoiŋ, ˈgou əˈ wei.

(vi) I have not come across any English example of contextual
assimilation affecting the position of the soft palate. 19372227

Progressive and Regressive Assimilation

850. Assimilations are termed progressive and regressive according
as the assimilated sound is influenced by a preceding or by a
following sound. Thus the assimilations which have taken place
in dɔgz (s > z), ˈbeikŋ (n > ŋ), or which take place contextually
in it s ˈredi (z > s), ˈʃʌt ʃr ˈaiz (j > ʃ) are progressive, while those
which have taken place in ænt (m > n), witθ (d > t), ˈtʃuldrən
(i > u), ˈnjuːspeipə (z > s), ˈhɔːʃʃuː (s > ʃ) are regressive.

Difficulties of Foreign Learners

851. French people speaking English often make assimilations
of voice to breath and breath to voice where they are not required.
When there are two consecutive consonants, one of which is breathed
and the other voiced (neither, however, being a liquid), they have
a tendency to assimilate the first to the second as regards presence
or absence of voice: thus, they are apt to pronounce

medicine (normal English ˈmedsin) as metˈsin,

anecdote (normal English ˈænikdout) rather like anɛgˈdɔt,

absurd (normal English əbˈsəːd) as apˈsœʀd (compare the French
absurde apˈsyʀd),

absolute (normal English ˈæbsəluːt or æbsəljuːt) as apˈsəlyt,

plenty of time (normal English ˈplenti əv ˈtaim) as plenˈti ɔf
ˈtaim
,

this book (normal English ˈðis ˈbuk) as ðiz ˈbuk,

like that (normal English laik ˈðæt) as laig ˈðat.

The Dutch have a similar tendency. Phonetic transcriptions of
the correct and incorrect pronunciation will help foreigners to
avoid such errors. French people should also note the English
word observe əbˈzəːv which they generally pronounce ɔpˈsɛrv as
in French.

852. Foreign learners often have difficulty in remembering
which are the words in which assimilation from tj to ʧ or dj to ʤ
has been made (§ 847 (ii)) and which are the words where such
assimilated forms are to be avoided. There is a general rule that
this assimilation has taken place in unstressed syllables, though not
often in stressed syllables. Thus assimilation has been made in the
words picture, question, grandeur, pressure mentioned under (ii) on
228pp. 222, 223; also in ocean ˈouʃn, pension ˈpenʃn, conscience ˈkɔnʃns,
partial ˈpaːʃl, anxious ˈæiŋkʃəs, usual ˈjuːʒul or ˈjuːʒŭəl, soldier
ˈsouldʒə, righteous ˈraitʃəs, natural ˈnætʃrəl, furniture ˈfəːnitʃə. On
the other hand the assimilation has not been made in mature məˈtjuə,
endurance inˈdjuərəns, in which the syllables in question are stressed.

853. There are, however, exceptions. There is a tendency, for
instance, for less common words to be pronounced without this
assimilation; thus celestial is siˈlestjəl, plenteous, beauteous are
ˈplentjəs, ˈbjuitjəs, 20373 and overture is generally ˈouvətjuə. Christian
is generally pronounced ˈkristjən, though ˈkristʃən may also be
heard. Sure ʃuə and sugar ˈʃugə are exceptional words in which
the assimilation sj > ʃ has been made in stressed syllables.229

Chapter XXVII
Elision

854. Elision is defined as the disappearance of a sound. There
are historical elisions, where a sound which existed in an earlier
form of a word was omitted in a later form; and there are contextual
elisions, in which a sound which exists in a word said
by itself is dropped in a compound or in a connected phrase.

855. A noteworthy example of historical elision is the loss of all
r-sounds finally and before consonants in Southern English. It
cannot be doubted that up to the fifteenth century the r's of such
words as arm, horse, church, more, other were always sounded (as
indeed they still are in many types of English), and it is believed
that the elision of these sounds started in the fifteenth century
and became fairly general in court circles in the course of the
sixteenth. 1374 Similar considerations apply to l in some words,
e.g. walk (now wɔːk), half (now hɑːf).

855a. The following are a few instances of historical elisions of
other sounds. The p of cupboard (now ˈkʌbəd) was doubtless
pronounced in Early English, but the sound eventually disappeared
(probably about the fifteenth century). It is likely that ˈtɔːtəʃel,
the usual pronunciation of tortoise-shell, is a reduction of a previous
form ˈtɔːtəʃʃel or ˈtɔːtəsʃel; if this is so, this word also illustrates
historical elision. 2375 The usual pronunciation of windmill ˈwinmil,
kindness ˈkainnis shows historical elision of d, if, as is probable,
d was formerly sounded in these words.

856. Historical elisions of unstressed vowels, especially ə and i,
are common in English. Examples are seen in the words history,
university, which are now generally pronounced ˈhistri, juːniˈvəːsti,
Formerly, no doubt, the pronunciation was ˈhistəri, juːniˈvəːsiti.
230and these forms may still be heard in precise speech, though they
are not common.

857. Contextual elisions of many kinds are frequent in English,
especially in rapid speaking. The following are examples of contextual
elisions commonly made in ordinary (not rapid) speech:
blind man ˈblain ˈmæn, Strand Magazine ˈstræn mægəˈziːn, a good
deal
ə ˈguˈdiːl (elision of d), take care ˈteiˈkɛə (elision of k), last time
ˈlaːs ˈtaim. Sit down is pronounced by some people siˈdaun with
elision of t.231

Chapter XXVIII
Length, Rhythm

858. Sounds which can be held on without alteration during
a longer or a shorter period of time are called continuants. The
chief continuant sounds are the vowels, the nasal, lateral, rolled
and fricative consonants, the frictionless continuants (Chap. XXII)
and the ‘stops’ of plosive consonants.

859. The length or quantity of a sound is the length of time
during which it is held on continuously in a given word or phrase.
Vowels and continuant consonants have length. Vowel-glides (i.e.
diphthongs and semi-vowels) also have length. The non-continuants
other than vowel-glides, e.g. flapped consonants and the plosions
of plosive consonants, may be regarded for practical linguistic
purposes as having no appreciable length.

860. It is easy to distinguish many degrees of length, say five or
six, but for practical purposes it is sufficient to distinguish two
or sometimes three degrees. When two degrees are distinguished,
they are called long and short. When it is desired to distinguish
an intermediate degree, this intermediate degree is termed medium
or half-long.

861. The mark of length is ː placed immediately after the
symbol for the sound which is long; half-length is marked when
necessary by ˑ; short sounds are generally left unmarked. 1376

Length of English Vowels

862. The principles governing the length of sounds in English
are very complex, and considerable differences may be observed
in comparing the speech of one person with that of another. The
following rules of length of English vowels are, however, sufficiently
accurate and sufficiently commonly followed to serve as a guide
in the practical teaching of the language.

863. Rule I. The vowels Nos. 1, 5, 7, 9, 11 (, ɑː, ɔː, , əː) are
longer than the other English vowels in similar situations, i.e.
232when surrounded by the same sounds, and pronounced with the
same degree of stress. Thus the vowels in heed hiːd, hard hɑːd,
hoard hɔːd, 2377 food fuːd, heard həːd are longer than the vowels in
hid hid, head bed, pad pæd, rod rɔd, bud bʌd, hood hud; similarly
the vowels in heat hiːt, heart hɑːt, short ʃɔːt, shoot ʃuːt, hurt həːt
are longer than the vowels in hit hit, get get, hat hæt, hot hɔt, hut
hʌt, put put. In consequence of this rule it is customary to
designate the vowels , ɑː, ɔː, , əː as the ‘long’ vowels and the
remaining English vowels as the ‘short’ vowels. Measurements
illustrating this rule are given in my book The Phoneme, §§ 398,
403-406. (For exceptions see §§ 874-879.)

864. The diphthongs have about the same length as the ‘long’
vowels.

865. The absolute lengths of the English ‘long’ vowels and
diphthongs are very variable and depend on their situations in
words and sentences (see Rules II-V). This fact may be stated in
more technical language by saying that there are two ‘chronemes’
(‘long’ and ‘short’) applicable to the vowels of the type of English
with which we are concerned here, and that each chroneme comprises
several ‘allochrones.’ See Chap. XXIII of The Phoneme.

866. Rule II. The ‘long’ vowels (and diphthongs) are shorter
when followed by a voiceless consonant than when final or followed
by a voiced consonant. When thus shortened, the ‘long’ vowels
may if desired be written with ˑ instead of with ː. Thus, the
vowel is shorter in seat siːt, than it is in sea siː or in seed siːd;
the vowels and diphthongs in staff stɑːf, sought, sort sɔːt, use (noun)
juːs, height hait, house (noun) haus, scarce skɛəs are shorter than
those in star stɑː, saw sɔː, yew juː, high hai, now nau, scare skɛə,
starve stɑːv, sawed sɔːd, use (verb) juːz, hide haid, cows kauz, scares
skɛəz. 3378

867. Rule III. The ‘long’ vowels (and diphthongs) are also
shorter before a nasal consonant or l followed in turn by a voiceless
consonant. Thus the ɔː in fault fɔːlt is shorter than that in
fall fɔːl or that in falls fɔːlz; the əː in learnt ləːnt is shorter than
that in learn ləːn or that in learns ləːnz.233

868. Rule IV. ‘Long’ vowels (and diphthongs) in stressed
syllables are also shorter when an unstressed syllable immediately
follows in the same word. Thus the 's in leader ˈliːdə, seeing
ˈsiːiŋ are shorter than those in lead liːd, see siː or seen siːn, the
ɔː's in drawing ˈdrɔːiŋ, 4379 causes ˈkɔːziz are shorter than those in
draw drɔː, draws drɔːz, cause kɔːz, and the in immunity iˈmjuːniti
is shorter than that in immune iˈmjuːn. 5380

869. When ‘long’ is immediately followed by an unstressed
i or ə, there is generally an alternative pronunciation with short
lax u forming either disyllabic sequences u-i, u-ə or diphthongs
, . Thus ruin, truer are pronounced either ˈruiːn, ˈtruːə (with
a shortened ) or ˈru-in, ˈtru-ə or ruin, truə. The diphthongs ei
and ou are often reduced to e and o in similar situations, with the
result that disyllabic sequences e-ə, o-i, o-ə or diphthongs , oi,
are used. Thus player is pronounced either ˈpleiə (with a shortened
ei) or ˈple-ə or pleə 6381; poetry, slower are pronounced either ˈpouitri,
ˈslouə (with a shortened ou) or ˈpo-itri, ˈslo-ə or ˈpoitri, 7382 sloə.
Variations of this nature are discussed at greater length in my
article Falling and Rising Diphthongs in Southern English in
Miscellanea Phonetica II, 1954 (published by the I.P.A.).

870. Rule V. The ‘long’ vowels (and diphthongs) are shorter
in unstressed syllables than in stressed syllables. This reduced
length is particularly noticeable in syllables preceding the stress.
Thus the ɔː in audacious ɔːˈdeiʃəs and the ɑː in carnation kɑːˈneiʃn
are shorter than the same vowels in August ˈɔːgəst, scarlet ˈskɑːlit;
the ai in idea aiˈdiə, the ou in ovation ouˈveiʃn, 8383 and the in
djuəˈreiʃn are shorter than the same diphthongs in idle ˈaidl, over
ˈouvə, enduring inˈdjuəriːŋ.234

871. When the unstressed ‘long’ vowel or diphthong follows the
stress, reduction of length is still observable though less marked.
Thus the ɔː in cardboard ˈkɑːdbɔːd, and the ou in fellow ˈfelou 9384
are not so long as the same sounds in board bɔːd, below biˈlou.

872. The English ‘short’ vowels (i, e, æ, ɔ, u, ʌ, ə) are also
subject to certain variations of length. The variations are similar
to those of the ‘long’ vowels but are less in degree. Thus it will
be noticed that the short vowels are generally slightly longer
before voiced consonants than before voiceless consonants; for
instance the i in bid bid and the ʌin cub kʌb are slightly longer
than the same vowels in bit bit and cup kʌp. Measurements
illustrating this are given in The Phoneme, §§ 405, 406.

873. The variations in the length of short vowels, with the
exception of the cases noted in §§ 874-878, are, however, not
sufficiently noticeable to be of importance in practical linguistic
work. Moreover, usage varies a good deal with different speakers.

874. Some exceptional cases of lengthening of the traditionally
‘short’ vowels must be noted. The most important is a lengthening
of æ in certain words. In the South of England a fully long æː
is generally used in the adjectives ending in -ad (bad bæːd, sad
sæːd, etc.), and is quite common in some nouns, e.g. man mæːn
or mæn, bag bæːg or bæg, jam ʤæːm or ʤæm. Curiously enough
the æ appears to be more usually short in nouns ending in -ad
(lad læd, pad pæd, etc.).

875. Long æː is most frequently found before voiced consonants,
but is not confined to these situations. Thus the words back, that
(meaning ‘that thing’) at the end of a sentence are often pronounced
with long æː by some Southern English people. Foreign learners
can sometimes improve the effect of their pronunciation of words
containing æ by lengthening the vowel.

876. Some English people, and especially Londoners, make a
similar lengthening of e in some words, e.g. bed, dead (but apparently
not in fed, tread). The word yes is exceptional; it is sometimes
pronounced jes, as usually transcribed, but when said by itself
it is more often pronounced with a fully long vowel of opener
quality (jɛːs). When French people are speaking English, they
235generally fail to lengthen the vowel of this word; their pronunciation
gives an effect of abruptness which is strange to English ears.

877. Similar lengthening may occasionally be observed with
other traditionally ‘short’ vowels. Thus speakers may be found
who pronounce big, good with longer vowels than pig, hood. His
and is often have lengthened vowels when in final position.

878. The lengthened ‘short’ vowels referred to in the four
preceding paragraphs are used chiefly in syllables which are final
in the sentence, and the length is particularly marked when the
words are pronounced with the ‘fall-rise’ intonation (§§ 1054, 1055).
Thus, in saying it isn't bad in such a way as to imply ‘but at the
same time it's not very good,’ the word bad would very commonly
be pronounced with a long vowel: iˈtiznt ˈbæːd. Similarly, if yet
had the ‘fall-rise’ in I really can't go to bed yet (implying ‘it's much
too early’), many people would pronounce ai ˈriəli ˈkɑːnt gou tə
ˈbed ˈjeːt
, bed having a short vowel and yet a long one.

879. The tendency to lengthen ‘short’ vowels appears to be
on the increase. In the local dialect of London it is much more
prevalent than in normal educated speech; it may also be observed
in American English. It is, in fact, possible that a new development
of the language is beginning to take place, by which the
present distinctions of quantity combined with quality will
eventually give place to distinctions of quality only. If such a
new system of vowel-length were to become the regular usage
in educated Southern English, it would become necessary to
modify the method of vowel representation in practical phonetic
transcription by introducing special letters to distinguish the pairs
of sounds which are now sufficiently well distinguished by lengthmarks. 10385

Length of Consonants

880. The length of consonants also varies, but not to the same
extent as that of vowels. The following are the only rules of
importance for foreign learners.

881. Rule VI. Final consonants are longer when preceded by
one of the ‘short’ vowels than when preceded by one of the ‘long’
236vowels or by a diphthong. Thus the n in sin sin is longer than
the n's in seen, scene siːn and sign sain.

882. Rule VII. Liquids are longer when followed by voiced
consonants than when followed by voiceless consonants. Thus the
n in wind wind is longer than that in hint hint, the l in bald bɔːld
is longer than that in fault fɔːlt, the m in number ˈnʌmbə is longer
than that in jumper ˈʤʌmpə.

883. Plosive consonants preceded by a short stressed vowel and
followed by another consonant are rather long, e.g. the k in act
ækt, actor ˈæktə (compare the k in jacket ˈʤækit), the p in description
disˈkripʃn.

884. Liquid consonants are usually long when preceded by a
short vowel and followed by an unstressed syllable beginning with
j or w, as the l in million ˈmiljən or the m in somewhere ˈsʋmwɛə
(compare sillier ˈsilĭə, summer ˈsʌmə).

885. Consonants following stressed short vowels are sometimes
very much lengthened for the sake of emphasis, e.g. splendid
ˈsplenːdid, a little more ə ˈlitːl ˈmɔː, I never heard such a thing ai
ˈnevːə ˈhəːd sʌtʃ ə θiŋ
, numbers and numbers of things ˈnʌmːbəz
n ˈnʌmːbəz əv θiŋz
. Similar lengthening occasionally occurs after
‘long’ vowels, e.g. it was awfully good it wəz ˈɔːfːli gud.

Relation between Rhythm and Length

886. Vowel-length depends to a considerable extent on the
rhythm of the sentence. There is a strong tendency in connected
speech to make stressed syllables follow each other as nearly as
possible at equal distances. Rule IV above (§ 868) is a result of
this rhythmical tendency. The usage may be stated more fully as
follows: when a syllable containing a long vowel or a diphthong
is followed by unstressed syllables, that vowel or diphthong is
generally shorter than if the syllable were final or followed by
another stressed syllable; moreover, the greater the number of
following unstressed syllables the shorter is the stressed vowel.

887. The following are some examples to supplement those in
§ 868. In pronouncing the series of numbers eighteen, nineteen,
twenty ˈeitiːn, ˈnaintiːn, ˈtwenti, the diphthong ai in nineteen is not
so long as the ai in nine in the series eight, nine, ten ˈeit, ˈnain,
237ˈten. The ou in there's nobody there ðɛə z ˈnoubədi ˈðɛə is not
nearly so long as that in there's no time ðɛə z ˈnou ˈtaim.

888. The differences of length caused in this way may be made
evident by representing the rhythm by means of an approximate
musical notation. Thus if we take a quaver ♪ to represent the
length of time between two consecutive stresses in eight, nine, ten,
the first two of the above sequences will appear thus:

♬ | ♬ | ♬ | ♪ | ♪ | ♪
ˈeitiːn | ˈnaintiːn | ˈtwenti | ˈeit | ˈnain | ˈten.

It is clear from this that the diphthongs ei, ai are something like
twice as long in the second sequence as they are in the first.

889. In like manner the two other sequences appear thus:

♪ | ♪ ♪
ðɛə z ˈnoubədi ˈðɛə | ðɛə z ˈnou ˈtaim.

The nou in the second sequence takes up almost as much time as
the entire word ˈnoubədi in the first. The ou is therefore a good
deal longer in the second sequence than it is in the first.

890. Further examples:

♪ ♪ ♪ ♬
wiːl ˈstɑːt iˈmiːdjətli if juə ˈredi
We'll start immediately if you're ready.

Here the two syllables ˈstɑːti may be made to occupy almost as
long a time as the five syllables ˈmiːdjətliifjuə. The syllable stɑːt
accordingly occupies more time than the syllable miːd, and it can
be heard that the in miːd is short (for a ‘long’ vowel) and that
the comparative lengthening of the syllable stɑːt is distributed over
the sounds ɑː and t. If in this sentence the word start were
replaced by a longer word containing a ‘long’ vowel, this vowel
would be shortened. Thus if we were to substitute arbitrate
ˈɑːbitreit or harmonize ˈhɑːmənaiz, we should find that the whole
238of these words might be compressed into a space not very much
longer than the monosyllable stɑːt.

♪ ♪ ♬
juː kən ˈkʌm wið ˈmiː if juə ˈredi.
you can come with me if you're ready.

Here the length of miː is not much less than the total length of
the three syllables ˈmiːdjətli in the preceding example.

♪ ♪ ♪ ♬ | ♪ ♬ ♬
ðə ˈsiːn wəz ˈbjuːtəfl. | ðə ˈsiːnəri wəz ˈbjuːtəfl.
The scene was beautiful. | The scenery was beautiful.

These examples show that the in scene is considerably longer
than the in scenery.

♪ ♬ | ♪ ♬ ♬ ♬
hiː ˈmuːvz ˈveri ˈræpidli. | hiz ˈmuːvmənts ə ˈveri ˈræpid.
He moves very rapidly. | His movements are very rapid.

The group ˈmuːvməntsə in the second example is compressed into
nearly the same space of time as muːvz in the first example, and
the of ˈmuːvmənts is heard to be a good deal shorter than the
of muːvz.

♪ ♬ ♩ | ♪ ♬
ðei ˈbeið in ðə ˈsiː. | ðeiə ˈbeiðiŋ in ðə ˈsiː.
They bathe in the sea. | They're bathing in the sea.

Here ˈbeiðinðə and ˈbeiðiŋinðə occupy about the same space of
time, so that the ei in beið is longer than that in ˈbeiðiŋ.

891. The nature of spoken English rhythm, upon which the
lengths of sounds are partly dependent, is a very involved subject,
and it has not been possible for me to investigate it in any detail.
It may be remarked, however, that the rhythm is determined not
only by the number and nature of the speech-sounds in the wordgroup
and the positions of the stresses in the words of more than
239one syllable, but also by the grammatical relations between words.
Thus if an unstressed syllable occurs between two stressed syllables,
it tends to be shorter if it is grammatically closely connected with
the following stressed syllable than if its closer grammatical relationship
is with the preceding stressed syllable.

892. To illustrate how rhythm is determined by grammatical
relations
between words we may take the comparatively simple
case of such a sequence as ˈei + t + ə + another stressed syllable,
where a stressed long vowel (or diphthong) is followed by a single
consonant which is followed in turn by a short unstressed vowel
and then another stressed syllable. It will be found that this
sequence has different rhythms, and consequently different lengths
of the sounds ei, t, ə, in the three following expressions:

(He's a)way to-day | (hiː z ə)ˈwei təˈdei.
(A) later day | (ə) ˈleitə ˈdei.
(He uses) eight a day | (hiː juːziz) ˈeit ə ˈdei.

The rhythm in the first example may be expressed by the approximate
musical notation ♪♪♩, the rhythm in the second example
is approximately ♬♩, while the rhythm in the third example is
intermediate between the other two. 11386 Expressed in terms of
length of sounds the differences are as follows: in the first example
the first ei is distinctly long and the ə extremely short; in the
second example the first ei is much shorter and the ə is longer; in
the third example the first ei is also rather short, but there is a
perceptible lengthening of the stop of the t, and the ə is intermediate
in length between the ə's of the first two examples.

893. Further examples of a rhythm approximating to ♪♪♩
in similar circumstances are:

Buy the book | ˈbai ðə ˈbuk.
(From) day to day | (frəm) ˈdei tə ˈdei.
Arm in arm | ˈɑːm in ˈɑːm.
240

894. The rhythm of the type ♪♪♩ becomes clearer still when
two or more consonants intervene between a stressed long vowel
and the unstressed short vowel. Examples are seen in:

(I) quite forgot | (ai) ˈkwait fəˈgɔt.
First of all | ˈfəːst əv ˈɔːl.
.

895. Further examples of a rhythm approximating to ♫♩
in similar circumstances are:

Either book | ˈaiðə buk.
Take it out | ˈteik it ˈaut.
Is it right? | ˈiz it ˈrait? 12387
Does he like (it)? | ˈdʌz iː ˈlaik (it)?

896. The rhythms are similar if the first syllable contains a
short vowel and two or more consonants intervene between it
and the unstressed vowel.

Examples of ♪♪♩:

Come to-day! | ˈkʌm təˈdei.
(It's) not for us | (it s) ˈnɔt fər ˈʌs.
What's the time? | ˈwɔt s ðə ˈtaim?
Ring the bell! | ˈriŋ ðə ˈbel.
Well-to-do | ˈwel-tə-ˈduː.
Twelve o'clock | ˈtwelv əˈklɔk.

Examples of ♫♩:

What's it for? | ˈwɔt s it ˈfɔː?
Shelter here | ˈʃeltə ˈhiə.

897. On the other hand, the rhythms that might be expected
on syntactic grounds are often replaced by others on account of
the nature and grouping of the sounds. Thus we have a rhythm
approximating to ♫♩ in many cases where the second syllable
is in close grammatical connexion with the third; a notable case is
241where the first stressed vowel is short or fairly short and there
is no consonant or only one consonant separating it from the
following unstressed vowel, as in

How are you? | ˈhau ə ˈjuː?
Come along! | ˈkʌm əˈlɔŋ.
Get away! | ˈget əˈwei.

(The rhythms in the two latter cases are difficult
to distinguish from those of summer day ˈsʌmə ˈdei,
(a) better way (ə) ˈbetə ˈwei.)

Right away | ˈrait əˈwei.

(In this case the ai is rather short because it is
followed by a voiceless consonant in the same
word. The rhythm is almost indistinguishable
from that of (a) later day, § 892.)

One oˈclock | ˈwʌn əˈklɔk.

(Compare twelve o'clock, § 896.)

898. Again, in the following examples owing to the number of
consonants a rhythm approximating to ♪♪♩ is found although
the second syllable is in close grammatical relationship with the
first:

Take them out! | ˈteik ðəm ˈaut.

(Compare Take it out, § 895.)

(He) locked it up | (hiː) ˈlɔkt it ˈʌp.
Easter Eve | ˈiːstər ˈiːv.
Does she like (it)? | ˈdʌʒ ʃiː ˈlaik (it)P

(Compare Does he like it? § 895.)

899. The above examples show that the rules determining the
rhythm, and therefore the lengths of the sounds, in the simple
case of an unstressed syllable between two stressed syllables are
rather complex. The reader may thus form some idea of the
extreme difficulty of describing or reducing to rules the innumerable
rhythms heard in ordinary connected speech. All we can say here
is (1) that there is a general tendency to make the ‘stress-points’
of stressed syllables follow each other at equal intervals of time,
but that this general tendency is constantly interfered with by the
242variations in the number and nature of the sounds between
successive stress-points, and (2) that the rhythms heard within the
‘stress-bars’ are dependent upon the grammatical relations between
the words as well as upon the number and nature of the sounds. 13388

Mistakes in Length made by Foreign Learners

900. The most notable mistakes of length heard from foreign
learners are as follows.

901. Many foreign people make the English ‘long’ vowels and
diphthongs fully long when followed by voiceless consonants, instead
of shortening them in accordance with the rule in § 866. This is
one of the characteristic mistakes made by Germans speaking
English. They almost invariably make the vowels and diphthongs
far too long in such words as park pɑːk, use (noun) juːs, fruit fruvt,
nation ˈneiʃn, mouth mauθ, right rait, roast beef ˈroust ˈbiːf. French
people also occasionally fall into this error.

902. Again, Germans generally fail to lengthen properly final
consonants preceded by short vowels. Thus, they are apt to
pronounce thin θin, tell tel, come kʌm with very short final consonants,
instead of lengthening them in accordance with the rule
in § 881.

903. The French are inclined to shorten long vowels when final,
pronouncing, for instance, sea, too with short vowels (like the
French si, tout) instead of with long ones (siː, tuː).

904. On the other hand, when there is a final unstressed
written -er, they make the vowel too long (besides inserting some
kind of r-sound). Thus they often pronounce paper peˈpœːʀ
instead of ˈpeipə.

905. The French also have a tendency to shorten the long
vowels and when followed by b, d, g, m, n, and l, as in tube
tjuːb, food fuːd, league liːg, tomb tuːm, fifteen ˈfifˈtiːn, feel fiːl.

906. Words for practice: (for Rule I) seen siːn, sin sin, harm
haːm, ham hæm, short ʃɔːt, shot ʃɔt, call kɔːl, doll dɔl, wall wɔːl,
quality ˈkwɔliti, pool puːl, pull pul, root ruːt, foot fut; (for Rule II)
see siː, far fɑː, saw sɔː, too, two tuː, fur fəː, say sei, sow (verb)
243sou, sigh sai, sow (pig) sau, lead (conduct) liːd, lard lɑːd, lord,
laud lɔːd, lose luːz, learn ləːn, laid leid, load loud, lied laid, loud
laud, geese giːs, pass pɑːs, horse hɔːs, loose luːs, verse vəːs, lace leis,
toast toust, nice nais, house haus; (for Rule III) aunt, are'nt ɑːnt,
taunt tɔːnt, learnt ləːnt, paint peint, don't dount, pint paint, ounce
auns.

907. The rhythm of spoken English is a source of considerable
difficulty to some foreign people. The French, for instance, are
liable to make continual use of the rhythm ♪♪♪♪♪… where
it is inappropriate in English. Thus they are liable to pronounce
Ring the bell, first of all, What's the time? with the rhythm ♪♪♪
instead of ♪♪♩ they pronounce he wrote to the secretary as

♪ ♪ ♪ ♪ ♪♪♪♪
i ʀot tu zi sɛkʀɛtɛʀi

instead of the correct English

♪ ♩ ♬
hiː ˈrout tə ðə ˈsekrȅtri.

908. The greatest difficulty of all is experienced by the Japanese.
For them and all others to whom English rhythm is difficult is
will be found helpful if the teacher taps the rhythms of sentencet
with his finger on the table. The pupil should practise saying
the sentence while tapping in unison with the teacher.244

Chapter XXIX
Stress

The Nature of Stress

909. Stress may be described as the degree of force with which a
sound or syllable is uttered. It is essentially a subjective action.
A strong force of utterance means energetic action of all the
articulating organs; it is usually accompanied by a gesture with the
hand or head or other parts of the body; it involves a strong ‘push’
from the chest wall and consequently strong force of exhalation 1389;
this generally gives the objective impression of loudness. Weak
force of utterance involves weak action of the chest wall resulting
in weak force of exhalation, and giving the objective acoustic
impression of softness.245

910. It was pointed out in §§ 208-210 that one or more sounds
in a spoken word or phrase are heard to stand out more prominently
than their immediate neighbours; and in the subsequent paragraphs
it was shown that a ‘syllable’ is essentially a small sound-sequence
containing a peak of prominence. Now if a word or phrase contains
a number of peaks of prominence, it is generally found that the
degrees of prominence at the various peaks are unequal; some
of the peaks have much greater prominence than others. In
other terms, some syllables of a word or phrase are perceived
more distinctly than others.

911. The prominence of a given sound may be increased or
diminished by means of any one of the three sound-attributes,
length, stress, or intonation, or by combinations of these. A
common and effective means of increasing prominence is to increase
the stress. In English, increase of stress is generally accompanied
by a modification of intonation and sometimes by an increase of
length.

912. It is important not to confuse stress with prominence
(§§ 100, 101, 208-210). The prominence of a syllable is its degree
of general distinctness, this being the combined effect of the tamber,
length, stress, and (if voiced) intonation of the syllabic sound.
The term ‘stress,’ as here used, refers only to the degree of force
of utterance; it is independent of length and intonation, though
it may be, and often is, combined with these. 2390 (For a more detailed
discussion of the nature of prominence see my book The Phoneme,
§ 434ff.).

913. Stress without intonation may be heard in English when
a clergyman is intoning the prayers in a church service. The
relations between stress and intonation found in ordinary spoken
English are shown in Chap. XXXI (see especially §§ 1022-1027).246

914. Syllables which are pronounced with a greater degree of
stress than the neighbouring syllables in a word or sentence are
said to be stressed or (more accurately) pronounced with strong stress.
Syllables pronounced with a relatively small degree of stress are
said to be unstressed or (more accurately) pronounced with weak
stress
. In what follows I am retaining the conventional, though
inaccurate, terms ‘stressed’ and ‘unstressed.’

915. It has frequently been suggested that a hearer can distinguish
by ear and a speaker can distinguish by sensation quite
a number of degrees of stress, say four or five. For instance, I
expressed the opinion in previous editions of this book that five
degrees of stress may be perceived in the word opportunity, and
that if we denote the strongest stress by the figure l, the second
strongest by the figure 2, and so on, the stressing of this word
might be indicated thus: 2 4 1 5 3 | ɔpəˈtjuːniti. I now think that this view
needs modification, on the ground that much of what is commonly
thought of as ‘stress’ is in reality stress (as defined here in § 909)
plus ‘prominence’ effected by means other than stress, and particularly
by ‘inherent sonority’ (§§ 100, 101), by subtle degrees of vowel and
consonant length and by intonation.

916. However that may be, I do not find any need in the practical
teaching and learning of pronunciation to attempt accuracy of that
order. It is generally sufficient to distinguish two degrees only,
stressed (or strong) and unstressed (or weak). Stressed syllables are
marked in this book by placing ˈ immediately before them, thus
father ˈfɑːða, arrive əˈraiv, opportunity ɔpəˈtjuːniti, Where are you
goingˈ?
, ˈwɛər ə juː ˈgouiŋ?

917. When for any reason it is found needful to distinguish three
degrees of stress, the sign ˌ may be used to denote the intermediate
or secondary stress. Thus in examination the secondary stress is
on the second syllable, so that the word may be written, if desired,
igˌzæmiˈneiʃn. (It is useful to mark the secondary stress in this
word, because foreign people usually put the secondary stress on
the first syllable.)

918. Marking secondary stress is of particular value in transcribing
English words which have three or more syllables preceding
the principal stress (see § 941).247

Stress in English

A. Word-stress (simple words)

919. Most English words of two syllables have one strongly
stressed syllable and one weak one. The strong stress is on the
first syllable in some words and on the second in others. For
instance, in the nouns increase and insult the first syllable is strong
and the second weak (ˈinkriːs, ˈinsʌlt), but in the verbs increase
and insult the first syllable is weak and the second strong (inˈkriːs,
inˈsʌlt). There exist, however, some disyllabic English words in
which both syllables have strong stress. Such are fifteen ˈfifˈtiːn,
prepaid ˈpriːˈpeid. They are said to be ‘double-stressed.’ Their
stress is subject to rhythmical variations in the sentence (see § 932).

919a. In English words of three or more syllables there is
always one strong syllable and occasionally two. The other
syllables in the words are as a rule weak (‘unstressed’), but in
some words there is a syllable with secondary stress. In each
of the following cases, for instance, there is one stressed syllable
and several unstressed ones: excessively ikˈsesivli, portmanteau
pɔːtˈmæntou, philanthropist fiˈlænθrəpist, particularize pəˈtikjuləraiz,
symbolically simˈbɔlikəli, uncharitableness ʌnˈtʃæritəblnis. There are
secondary stresses in centralization ˌsentrəlaiˈzeiʃn, administration
ədˌminisˈtreiʃn and the other examples quoted in § 941; probably
also in the numerous single-stressed compounds like foot-passenger
ˈfutˌpæsinʤə, kettle-holder ˈketlˌhouldə (§ 946). Occasionally a very
long word may have two secondary stresses. For instance, intellectuality
may be pronounced ˌintiˌlektjuˈæliti (also ˌintilektjuˈæliti
and ˈintiˌlektjuˈæliti).

919b. Examples of double-stressed words of three or more
syllables are given in § 922. Like the disyllables their stresses
are subject to rhythmical variations in the sentence (§ 933).

920. Generally speaking there are no rules determining which
syllable or syllables of polysyllabic English words bear the main
stress. The foreign student is obliged to learn the stress of each
word individually. He has to learn, for instance, that the main
stress is on the first syllable in photograph ˈfoutəgrɑːf or ˈfoutəgræf,
on the second in photography fəˈtɔgrəfi, on the third in photographic
foutəˈgræfik and on the fourth in photogravure foutəgrəˈvjuə. When
248rules of word-stress can be formulated at all, they are generally
subject to numerous exceptions. 3391

921. In the case of double-stressed words it is, however, possible
to formulate some general principles which are of assistance to the
foreign learner. These are given in the following paragraphs.249

922. Words formed by adding to a word in common use a prefix
having a distinct meaning of its own 4392 very usually have two strong
stresses, namely a stress on the prefix and the stress of the original
word. Examples of such prefixes are: anti-, arch- (in the sense of
‘chief’), dis- (when equivalent to un- or implying separation), ex-
(in the sense of ‘former’), half-, joint-, in- (il-, im-, ir-) (in the sense
of ‘not’), inter- (in the reciprocal sense), trial-, mis- (implying ‘error’
or ‘falseness’), non-, out- (in verbs, with the sense of ‘outdoing’),
over- (in the sense of ‘too much’), pre- (meaning ‘beforehand’), re(denoting
‘repetition’), sub- (in the sense of ‘subordinate’), ultra-,
un-, under- (in the sense of ‘too little’ or in the sense of ‘subordinate’),
vice-. 4393

Examples: anticlimax ˈæntiˈklaimæks, archbishop ˈɑːtʃˈbiʃəp,
disloyal ˈdisˈlɔiəl, disconnect ˈdiskəˈnekt, discontented ˈdiskən'tentid,
disembark ˈdisimˈbɑːk, ex-president ˈeksˈprezidənt, half-finished
hɑːfˈfiniʃt, joint-tenant ˈʤɔintˈtenənt, inexperienced ˈiniksˈpiərĭənst,
250insincere ˈin-sinˈsiə, insufficient ˈin-səˈfiʃnt, illogical ˈiˈlɔʤikl, imperceptible
ˈim-pəˈseptəbl, irreligious ˈiriˈliʤəs, intermingle ˈintəˈmiŋgl,
malformation ˈmælfɔːˈmeiʃn, misquote ˈmisˈkwout, misrepresentation
ˈmisreprizen'teiʃn, non-payment ˈnonˈpeimənt, outgeneral ˈautˈʤenərəl,
overestimate (v.) ˈouvərˈestimeit, overripe ˈouvəˈraip, prepaid
ˈpriːˈpeid, rearrange ˈriːəˈreinʤ, 5394 sub-dean ˈsʌbˈdiːn, ultra-fashionable
ˈʌltrəˈfæʃnəbl, unfruitful ˈʌnˈfruːtful, unknown ˈʌnˈnoun, unpack
ˈʌnˈpæk, unobjectionable ˈʌnəbˈʤekʃnəbl, underestimate (v.)
ˈʌndərˈestimeit, under-secretary ˈʌndəˈsekrətri, vice-chancellor
ˈvaisˈtʃɑːnslə.

923. It must be observed that if the word to which the prefix
is added is not in common use or is only used in a sense different
from that attributed to it when the prefix is added, then double
stress is not generally used.

Examples: discourage disˈkʌriʤ (courage not being used as a
verb), inordinate iˈnɔːdinit (the adjective ordinate being rare),
unwieldy ʌnˈwiːldi (the word wieldy being very rare, and in fact
unknown to most people), undoubted ʌnˈdautid (doubted not being
used as an attributive adjective), underline ʌndəˈlain (the verb
line not being used in the sense of ‘to draw a line’).

924. For a similar reason some adverbs have single stress while
the corresponding adjectives have double stress. Thus unaccountably
is usually ʌnəˈkauntəbli, while unaccountable is quite
commonly ˈʌnəˈkauntəbl; so also invariably is regularly pronounced
inˈvɛərĭəbli, though the adjective invariable is pronounced either
ˈinˈvɛərĭəbl or inˈvɛərĭəbl.

925. Very common words formed from other words by the
addition of some of the above-mentioned prefixes, and particularly
words in which the stress of the simple word is on the first syllable,
are exceptions to the principle stated in § 922, and take no stress
on the prefix. Thus it is not usual to stress the prefixes of impossible
imˈpɔsəbl, unusual ʌnˈjuːʒuəl, unfortunate ʌnˈfɔːtʃnit. Some put
imperceptible (§ 922) into this category, pronouncing it im-pəˈseptəbl.

926. In many words which are not uncommon but yet not
very common, usage varies. Thus some speakers pronounce
251irregularity, overestimate (v.) with single stress (iˌregjuˈlæriti,
ouvərˈestimeit), even when not under the influence of rhythm
(§ 932); others would say ˈiregjuˈlæriti, ˈouvərˈestimeit. In cases
of doubt it is probably safer for the foreign learner to use double
stress in preference to single stress.

927. Further exceptions are archbishopric ɑːtʃˈbiʃəprik, archdeaconry
ɑːtʃˈdiːkənri, archdeaconship ɑːtʃˈdiːkənʃip, halfpenny ˈheipəni
or ˈheipni. The word archangel is usually ˈɑːk-einʤl, but is
pronounced ˈɑːkˈeinʤl by some.

928. The following miscellaneous words 6395 are commonly pronounced
with double stress (subject to rhythmical variations, see
§ 932): amen ˈɑːˈmen or ˈeiˈmen, 7396 daresay ˈdɛəˈsei, hullo ˈhʌˈlou,
inborn ˈinˈbɔːn, inbred ˈinˈbred, inlaid ˈinˈleid, conversely ˈkɔnˈvəːsli,
postdate ˈpoustˈdeit, and the numerals thirteen ˈθəːˈtiːn, fourteen
fɔːˈtiːn, fifteen ˈfifˈtiːn, sixteen ˈsiksˈtiːn, seventeen ˈsevnˈtiːn, eighteen
ˈeiˈtiːn, nineteen ˈnainˈtiːn.

929. The following words are pronounced either with stress
on the last syllable or with double stress; in any case they are
subject to the influence of rhythm: princess ˈprinˈses or prinˈses, 8397
sardine ˈsɑːˈdiːn or sɑːˈdiːn, trombone ˈtrɔmˈboun or trɔmˈboun,
bamboo ˈbæmˈbuː or bæmˈbui, masseuse ˈmæˈsəːz or mæˈsəːz.
Another variable word is indiarubber, which is pronounced
ˈindjəˈrʌbə or indjəˈrʌbə. Instances of rhythmical variations in
these words are given in § 932.

930. A number of proper names are similarly treated, e.g. Bantu
ˈbænˈtuː or bænˈtuː, Bengal ˈbeŋˈgɔːl or beŋˈgɔːl (or ˈbenˈgɔːl,
benˈgɔːl), Berlin ˈbəːˈlin or bəːˈlin, Bexhill ˈbeksˈhil or beksˈhil,
Blackheath ˈblækˈhiːθ or blækˈhiːθ, Canton (in China) ˈkænˈtɔn or
kænˈtɔn, 9398 Carlisle, 10399 Carlyle ˈkɑːˈlail or kɑːˈlail, Cheapside ˈtʃiːpˈsaid
or tʃiːpˈsaid, Cornhill ˈkɔːnˈhil or kɔːnˈhil, Panama ˈpænəˈmɑː or
252pænəˈmɑː, Dundee ˈdʌnˈdiː or dʌnˈdiː, Peiping ˈpeiˈpiŋ or peiˈpiŋ,
Piccadilly ˈpikəˈdili or pikəˈdili, Scawfell ˈskɔːˈfel or skɔːˈfel, Spithead
ˈspitˈhed or spitˈhed, Stonehenge ˈstounˈhenʤ or stounˈhenʤ,
Torquay ˈtɔːˈkiː or tɔːˈkiː, 11400 Vauxhall ˈvɔksˈhɔːl or vɔksˈhɔː!, Whitehall
ˈwaitˈhɔːl or waitˈhɔːl, and many names ending in -ness, e.g.
Skegness ˈskegˈnes or skegˈnes, Shoeburyness ˈʃuːbəriˈnes or
ʃuːbəriˈnes, also disyllabic adjectives ending in -ese formed from
proper names, e.g. Chinese ˈtʃaiˈniːz or tʃaiˈniːz, Maltese ˈmɔːlˈtiːz
or mɔːlˈtiːz. 12401 All the above are subject to rhythmical variations;
for examples see §§ 932, 933.

Rhythmical Variations

931. The stress of words normally pronounced with double
stress is very often modified in sentences. The first of the stressed
syllables is apt to lose its stress when closely preceded by another
stressed syllable; similarly the second of the stressed syllables is
apt to lose its stress when closely followed by another stressed
syllable. Thus although the word fourteen spoken by itself, or
said in answer to the question ‘How many people were there?’
has double stress (§ 928), yet in fourteen shillings it is stressed on
the first syllable only (ˈfɔːtiːn ˈʃiliŋz) and in just fourteen it is
stressed on the second syllable only (ˈʤʌst fɔːˈtiːn). Compare
similarly inlaid wood ˈinleid ˈwud with all inlaid ˈɔːl inˈleid, an
unknown land
an ˈʌnnoun ˈlænd with quite unknown ˈkwait ʌnˈnoun.

932. The words which, when pronounced by themselves, admit
of either single or double stress (§§ 929, 930) are likewise subject
to similar rhythmical variations. Compare

Princess Victoria ˈprinses vikˈtɔːrĭə with a royal princess
ə ˈrɔiəl prinˈses,

an indiarubber ball ən ˈindjərʌbə ˈboːl with a piece of indiarubber
ə ˈpiːs əv indjəˈrʌbə,

Piccadilly Circus ˈpikədili ˈsəːkəs with close to Piccadilly
ˈklous tə pikəˈdili,

Vauxhall Bridge ˈvɔkshɔːl ˈbriʤ with near Vauxhall ˈniə
vɔksˈhɔːl
,

Waterloo station ˈwɔːtəluː ˈsteiʃn with the train for Waterloo
ðə ˈtrein fə wɔːtəˈluː,253

Dundee marmalade ˈdʌndiː ˈmɑːməleid with going to Dundee
ˈgouiŋ tə dʌnˈdiː,

sardine sandwiches ˈsɑːdiːn ˈsænwiʤiz with a tin of sardines
ə ˈtin əv sɑːˈdiːnz.

933. Similar changes of stress arc sometimes found in the case of
single-stressed words. Examples are Constitution Hill, Cayenne
pepper
which are commonly pronounced ˈkɔnstitjuːʃn ˈhil, ˈkeien
ˈpepə
. In Salvation Army the stress ˈsælveiʃn ˈɑːmi seems quite
as usual as sælveiʃn ˈɑːmi. Similarly many would say ən ˈɑːtifiʃl
ˈlængwiʤ
, ə ˈdipləmætik ˈmiʃn, rather than ən ɑːtifiʃl ˈlængwiʤ,
ə dipləˈmætik ˈmiʃn (an artificial language, a diplomatic mission).
Those who pronounce finance as faiˈnæns will often speak of a
ˈfainæns ˈsʌbkəmiti (finance subcommittee).

Emphasis

934. When it is desired to emphasize 13402 words which have both
a primary and a secondary stress, and in which the secondary
stress precedes the primary (as is usually the case), the secondary
stress is often reinforced and becomes as strong as the primary
stress. Thus the words fundamental, distribution, responsibility, disappearance,
recommend, artificial (normally ˌfʌndəˈmentl, ˌdistriˈbjuːʃn,
risˌponsəˈbiliti, ˌdisəˈpiərəns, ˌrekəˈmend, ˌɑːtiˈfiʃl) would
often be pronounced ˈfʌndəˈmentl, ˈdistriˈbjuːʃn, risˈponsəˈbiliti,
ˈdisəˈpiərəns, ˈrekəˈmend, ˈɑːtiˈfiʃl for the sake of emphasis.

935. The frequent use of double stress in the words mentioned
in §§ 929, 930 is no doubt to be attributed to this tendency.
Occasionally other single-stressed words may receive double stress
for the sake of emphasis: thus unless is often pronounced ˈʌnˈles
instead of the normal ənˈles or ʌnˈles; and one may occasionally
hear such words as spectator, psychology, gesticulate, mutation said
with double stress (ˈspekˈteitə, ˈsaiˈkɔləʤi, ˈʤesˈtikjuleit, ˈmjuːˈteiʃn)
instead of the normal single stress (spekˈteitə, saiˈkɔləʤi,
ʤesˈtikjuleit, mjuːˈteiʃn).

936. In longer words, the greater the distance between the
secondary stress and the primary stress, the more readily does
this reinforcement of the secondary stress take place. Thus in
representation, characteristic, vulnerability the double-stressed forms
254ˈreprizenˈteiʃn, ˈkæriktəˈristik, ˈvʌlnərəˈbiliti seem quite as common
as the single-stressed forms ˌreprizenˈteiʃn, ˌkæriktəˈristik,
ˌvʌlnərəˈbiliti. And in very long words in which as many as three
syllables intervene between the secondary stress and the primary
stress, reinforcement of the secondary stress is so common that it is
probably to be regarded as the usual form. Thus perpendicularity,
characterization are usually ˈpəːpəndikjuˈlæriti, ˈkæriktəraiˈzeiʃn.

937. When it is desired to emphasize (for contrast) a particular
part of a word which is not normally stressed, that part may receive
a strong stress, and the normal primary stress may become a
secondary stress. Thus when reverse is contrasted with obverse, it
is commonly pronounced ˈriːvəːs. When commission is contrasted
with omission, it is commonly pronounced ˈkɔmiʃn or ˈkɔˈmiʃn.
So also with ascending and descending, offensive and defensive,
which are frequently ˈæsendiŋ ən ˈdiːsendiŋ, ˈɔfensiv ən ˈdiːfensiv 14403
(instead of əˈsendiŋ ən diˈsendiŋ, əˈfensiv ən diˈfensiv). In the
case of external, there is practically always a contrast, expressed
or implied, with internal; consequently the natural stress of the
word (eksˈtainl) is seldom heard, the usual pronunciation being
ˈeksˈtəːnl (less commonly ˈekstəːnl). Similarly demerits is almost
always pronounced ˈdiːˌmerits.

Special Difficulties of Foreign Learners

938. Most foreign people have a tendency to stress the last
syllable of words ending in -ule, -ude, -ise, -ize when the stress
should be on some other syllable. They also generally stress the
last syllable of reconcile, which is in Southern English ˈrekənsail. 15404
Examples for practice: prosecute ˈprɔsikjuːt, substitute ˈsʌbstitjuːt,
gratitude ˈgrætitjuːd, multitude ˈmʌltitjuːd, criticize ˈkritisaiz, 15405
exercise ˈeksəsaiz, recognize ˈrekəgnaiz. 15406 Foreign people are
particularly liable to stress the syllables -juːt, -juːd, -aiz in the
inflected forms such as prosecuted ˈprɔsikjuːtid, criticizes ˈkritisaiziz.

939. The French are apt to stress the final syllable wrongly
in many other words. Examples for practice: language ˈlæŋgwiʤ,
paper ˈpeipə, collar ˈkɔlə, distance ˈdistəns, circumstance ˈsəːkəmstəns,
255universe ˈjuːnivəːs, ridicule ˈridikjuːl, goodness ˈgudnis, vexation
vekˈseiʃn, disgraceful disˈgreisfl.

940. French people should pay special attention to the stress
of English words of more than two syllables. They often have a
tendency to stress the first syllable of any long word beginning
with a consonant, and the second syllable of any long word
beginning with a vowel. They should thus be careful to stress the
second syllable in such words as remarkable riˈmɑːkabl, sufficient
səˈfiʃnt, tremendous triˈmendəs, reluctance riˈlʌktəns, successful
səkˈsesfl, and to stress the first syllable in such words as absolutely
ˈæbsəluːtli or ˈæbsəˈluːtli, execute ˈeksikjuːt, excellent ˈekslənt.

941. Foreign learners should give special attention to the position
of secondary stress in long words where it precedes the primary
stress. They often mispronounce such words by putting the
secondary stress on the wrong syllable; and particularly on the
first when it ought to be on the second. The following are some
examples of words of this type 16407:

tableau Secondary stress on first syllable | Secondary stress on second syllable | centralization | modification | ornamentation | perigrination | qualification | representation | solemnization | circumlocution | archæological | temperamental | aristocratic | mathematician | disciplinarian | caricature | administration | affiliation | anticipation | assimilation | consideration | examination | interrogation | pronunciation | ecclesiastical | antagonistic | materialistic | academician | bacteriology | Iphigenia

17408 18409 19410 20411256

tableau Secondary stress on first syllable | Secondary stress on second syllable | penetrability | potentiality | instrumentality | accessibility | individuality | familiarity | artificiality | peculiarity | heterogeneous | superiority | paraphernalia | encyclopedia | peritonitis | tuberculosis

21412

942. It is noteworthy that a number of words having i in the
first syllable and stress on the third syllable have no perceptible
stress on the first syllable. Such are electricity ilekˈtrisiti, electrician
ilekˈtriʃn, electrolysis ilekˈtrɔlisis, etc., elasticity ilæsˈtisiti, detestation
ditesˈteiʃn. When pronounced as here shown, neither of the first
two syllables of these words can be said to have a stronger stress
than the other. 22413 These words have alternative pronunciations
with or e; when so pronounced, there is generally secondary
stress on the first syllable which can be shown thus, if desired:
ˌiːlekˈtrisiti or ˌelekˈtrisiti, etc., ˌditesˈteiʃn.

B. Word-stress (Compound Words)

943. By a compound word is meant here a word made up of two
words written in conventional spelling as one, with or without a
hyphen.

944. Some compound words have single stress on the first
element, others have double stress. 23414

945. Single-stressed compounds are by far the most common.
Examples are: appletree ˈæpltriː, bookbinding ˈbukbaindiŋ, bystander
bai-stændə, Buckinghamshire ˈbʌkiŋəmʃiə, 24415 daybreak ˈdei-breik,
257dining-room ˈdainiŋrum, door-handle ˈdɔːhændl, figurehead ˈfigəhed,
fireplace ˈfaiə-pleis, flowerpot ˈflauəpot, footpassenger ˈfutpæsinʤə,
flute-player ˈfluːtpleiə, grasshopper ˈgrɑːshopə, green-fly ˈgriːn-flai,
hairbrush ˈheə-brʌʃ, housekeeper ˈhaus-kiːpə, jellyfish ˈʤelifiʃ, kettleholder
ˈketlhouldə, key-hole ˈkiːhoul, lightning-conductor ˈlaitniŋkəndʌktə,
midnight ˈmidnait, orange-blossom ˈɔrinʤblɔsəm, painstaking
ˈpeinzteikiŋ, pickpocket ˈpikpɔkit, schoolmaster ˈskuːlmɑːstə,
shirt-sleeves ˈʃəːt-sliːvz, sitting-room ˈsitiŋrum, smoking-compartment
ˈsmoukiŋ-kəmpɑːtmənt, snowball ˈsnoubɔːl, tea-party ˈtiː-pɑːti, thunderstorm
ˈθʌndə-stɔːm, washingstand ˈwɔʃiŋ-stænd, waterproof
ˈwɔːtə-pruːf, weatherbeaten ˈweðəbiːtn, wind-screen ˈwindskriːn.

946. Special attention is called to the following cases of compound
nouns in which single stress is used.

(i) Where the compound noun denotes a single new idea rather
than the combination of two ideas suggested by the original words.
Examples: blacksmith ˈblæk-smiθ, bluebottle ˈbluːbɔtl, Newcastle
ˈnjuːkɑːsl, 25416 greenhouse ˈgriːnhaus, greengrocer ˈgriːngrousə, kingfisher
ˈkiŋfiʃə, walking-stick ˈwɔːkiŋstik. (Exceptions, in my pronunciation,
are great-coat ˈgreitˈkout, greengage ˈgrinːˈgeiʤ. 26417)

(ii) Where the meaning of the whole compound noun is the
meaning of the second element restricted in some important way
by the first element. Examples: birthday ˈbəːθdei (a special day),
cart-horse ˈkɑːthɔːs (a particular kind of horse), darning-needle
ˈdɑːniŋniːdl (a special type of needle), dinner-table ˈdinəteibl (a
particular kind of table), gas-engine gæsenʤin (a particular kind
of engine), cattle-show ˈkætlʃou, sheepdog ˈʃiːpdɔg. Exceptions are
words in which the second element is felt to be of special importance
(see § 949).

(iii) Where the first element is either expressly or by implication
contrasted with something. Example: flute-player ˈfluːtpleiə (where
flute is naturally felt to be contrasted with other musical instruments).

947. Double stress is used in compound adjectives of which
the first element is an adjective. Examples: redhot ˈredˈhɔt,
258good-looking ˈgudˈlukiŋ, old-fashioned ˈouldˈfæʃnd, bad-tempered ˈbædˈtempəd,
absent-minded ˈæbsntˈmaindid, first-class ˈfəːstˈklɑːs, secondhand
ˈsekəndˈhænd, bare-headed ˈbɛəˈhedid, 27418 dead-beat ˈdedˈbiːt.
Note also home-made ˈhoumˈmeid, well-bred ˈwelˈbred. 28419

948. There is an exceptional case in which single stress is used,
namely when the compound adjective is practically synonymous
with its first element. Examples: oval-shaped ˈouvlʃeipt, yellowish-looking
ˈjelouiʃlukiŋ (which are practically equivalent to ‘oval,’
‘yellowish’). 29420

949. When the second element of a compound is felt to be of
special importance, double stress is used. Thus gas-stove is commonly
ˈgsesˈstouv, the importance of the second element stove being no
doubt due to the implied contrast with ‘fire,’ the traditional method
of heating in England. (On the other hand, gas-engine ˈgæsenʤiŋ
has only a single strong stress, there being no particular contrast
between ‘engine’ and anything else, but rather a contrast between
an engine worked by gas and engines worked by other means.)
Further examples are: indiarubber ˈindjəˈrʌbə (the important part
being rubber), eye-witness ˈaiˈwitnis (‘witness’ being contrasted with
persons who had only heard of the occurrence, etc.), bow-window
ˈbouˈwindou. Armchair ˈɑːmˈʧɛə would apparently also belong to
this category. Churchyard ˈʧəːʧˈjɑːd is another example in my
pronunciation, the ‘yard’ being implicitly contrasted with the
church itself; it seems, however, that ˈʧəːʧ-jɑːd with single stress
is now the commoner form. (Graveyard ˈgreiv-jɑːd is always said
with single stress, there being no such contrast in the case of this
word.)

950. But when a compound noun of the kind referred to in
§ 950 is commonly or very frequently used attributively, it may
have single stress. Examples are midsummer ˈmidsʌmə, midnight
ˈmidnait. These words are frequently used attributively (e.g. Midsummer
Day
, midnight sun). When so used they have single
stress on the first element by the principle of rhythm (§ 931),
and this pronunciation has become permanently attached to them.
259Compare mid-winter ˈmidˈwintə, which is not used attributively
and which has double stress.

951. It may be added that it is often difficult to give satisfactory
reasons for assigning a word to the classes mentioned in
§ 947 (ii) and (iii) or to the class described in § 950. In numerous
cases both elements of the word are felt to be important for
reasons of contrast or otherwise, and the treatment of the stress
may depend simply on a very small balance of importance which
it is not easy to estimate.

952. The following are some miscellaneous compounds having
double stress (subject to rhythmical variations and to emphasis,
§§ 932, 938), although not coming under the principles in §§ 948, 950:
downhill ˈdaunˈhil, uphill ˈʌpˈhil, downstairs ˈdaunˈstɛəz, upstairs
ˈʌpˈstɛəz; hereby ˈhiəˈbai, herein ˈhiərˈin, hereinafter ˈhiərinˈɑːftə, 30421
heretofore ˈhiətuˈfoi, hereupon ˈhiərəˈpon, whereabouts (interrogative
adverb) ˈwɛərəˈbauts, thereabouts ˈðɛərəˈbauts, 31422 thereby ˈðɛəˈbai,
therein ˈðɛərˈin, thereupon ˈðɛərəˈpon, whereupon ˈwɛərəˈpon; henceforth
ˈhensˈfɔːθ, henceforward ˈhensˈfɔːwəd, thenceforth ˈðensˈfɔːθ,
thenceforward ˈðensˈfɔːwəd, elsewhere ˈelsˈwɛə (also ˈels-wɛə); inside
ˈinˈsaid, outside ˈautˈsaid, alongside əˈlɔŋˈsaid, seaside ˈsiːˈsaid; indoors
ˈinˈdɔːz, outdoors ˈautˈdɔːz; upturn ˈʌpˈtəːn, meantime ˈmiːnˈtaim,
meanwhile ˈmiːnˈwail, 32423 passer-by ˈpɑːsəˈbai, point-blank ˈpointˈblæŋk.

953. Nouns compounded of a verb and an adverb, such as
make-up, setback, set-up, get-away, are generally said with single
stress (ˈmeikʌp, ˈsetbæk, etc.), but some pronounce them with
double stress.

954. The stress of double-stressed compounds is subject to
rhythmical variations like double-stressed simple words (§ 932).
The following are examples of rhythmical variations in doublestressed
compounds:

a red-hot poker ə ˈredhɔt ˈpoukə | just red-hot ˈʤʌst redˈhɔt
second-hand books ˈsekəndhænd | all second-hand ˈɔːl səkəndˈhænd ˈbuks
260inside out ˈinsaid ˈaut | right inside ˈrait inˈsaid
the upstairs rooms ði ˈʌpstɛəz rumz | on going upstairs ɔn ˈgouiŋ ʌpˈstɛəz
greengage jam ˈgriːngeiʤ ˈʤæːm | stewed greengages ˈstjuːd griːn-ˈgeiʤiz 33424
inland revenue ˈinlænd ˈrevinjuː 34425 | further inland ˈfəːðər inˈlænd
the overhead railway ði ˈouvəhed ˈreilwei | right overhead ˈrait ouvəˈhed
an uphill task ən ˈʌphil ˈtɑːsk | a light great-coat ə ˈlait greit-ˈkout 33426
seaside sports ˈsiː-said ˈspɔːts | cold plum-pudding ˈkould plʌm-ˈpudiŋ

955. Compound words consisting of three elements generally
take single stress on the second element if the first two elements
taken alone form a double-stressed compound. Examples: gingerbeer-bottle
ʤinʤəˈbiəbɔtl, hot-water-bottle hɔtˈwɔːtəbɔtl, waste-paperbasket
weisˈpeipəbɑːskit. (These words may also be said with
stress on the first element as well as on the second: ˈʤinʤəˈbiəbɔtl,
etc.) Otherwise three word compounds have main stress on the
first element. Examples: teapothandle ˈtiːpɔthændl, teaspoonful
ˈtiːspuːnful, lodginghousekeeper ˈlɔʤiŋhausˌkiːpə, sodawaterbottle
ˈsoudəwɔːtəˌbɔtl, watercressbed ˈwɔːtəkresˌbed.

Difficulties of Foreign Learners

956. Some foreign people, and especially Germans, are liable to
pronounce double-stressed compounds with single stress on the
first element; others (especially the French) are liable to pronounce
single-stressed compounds with double stress. These errors may
be rectified by observing the relations between stress and intonation
described in Chap. XXXI (particularly §§ 1019-1027). The correct
intonations of arm-chair, plum-pudding, pronounced with a falling
intonation (Tune 1) may be represented thus:

image ˈɑːmˈʧɛə | ˈplʌmˈpudiŋ261

Germans generally say

image ˈɑːmʧɛə | ˈplʌmpudiŋ

while the correct pronunciations of dinner-table, lightning-conductor,
pronounced with a falling intonation (Tune 1), may be represented
thus:

image ˈdinəteibl | ˈlaitiŋkəndʌktə

French people often say

image dinəˈteibl | ˈlaitiŋkənˈdʌktə

956a. Readers of this book should supplement what is said here
on stress by a study of two recently published specialized books on
the subject: R. Kingdon's Groundwork of English Stress
(Longmans, 1958) and G. F. Arnold's Stress in English Words
(North Holland Publishing Co., Amsterdam, 1957).

C. Sentence-stress

1. General Principle

957. As a general rule it may be said that the relative stress of
the words in a sequence depends on their relative importance.
The more important a word is, the stronger is its stress. The
most important words are usually (in the absence of special
emphasis) the nouns, adjectives, demonstrative and interrogative
pronouns, principal verbs, 35427 and adverbs. Such words are therefore
generally strongly stressed (subject to exceptions, see §§ 962 ff.).
Thus the first sentence of this paragraph is stressed as follows:
əz ə ˈʤənrəl ˈruːl it mei bi ˈsed ðət ðə ˈrelətiv ˈstres əv ðə ˈwəːʤ
in ə ˈsiːkwəns diˈpenʣ ɔn ðɛə ˈrelətiv imˈpɔːtns
. Similarty What do
you think of the weather?
is usually stressed thus: ˈwɔt djuː ˈθiŋk
əv ðə ˈweðə
; this train generally arrives late is normally stressed as
follows: ˈðis ˈtrein ˈʤenrəli əˈraivz ˈleit.262

958. When all the important words in a sentence are equally
important they all have strong stress. In this way it frequently
happens that a number of strong syllables occur consecutively.
Thus in the sentence John has just bought two large brown dogs
every word would be stressed except has, thus: ˈʤɔon əz ˈʤʌst
ˈbɔːt ˈtuː ˈlɑːʤ ˈbraun ˈdɔgz
.

959. Foreign learners should note particularly the case of one
word qualifying another. Both the words have as a rule strong
stress. 36428 Examples: it's very important it s ˈveri imˈpɔːtnt, a useful
book
ə ˈjuːsful ˈbuk, the first prize ðə fəːst ˈpraiz, roast beef ˈroust
ˈbiːf
, a deck chair ə ˈdek ˈʧɛə, the boy's book ðə ˈbɔiz ˈbuk, Wednesday
evening
ˈwe(d)nzdi ˈiːvniŋ, George's dog ˈʤɔːʤiz ˈdɔg, North Western
nɔːθ ˈwestən, the castle wall ðə ˈkɑːsl ˈwɔːl, an orphan boy ən ˈɔːfən
ˈbɔi
, all right ˈɔːl ˈrait, so far so good ˈsou ˈfɑː ˈsou ˈgud, it was too
much
it wəz ˈtuː ˈmʌtf, Buckingham Palace ˈbʌkiŋəm ˈpælis, Hyde
Park
ˈhaid ˈpɑːk, York Road ˈjɔːk ˈroud, Chancery Lane ˈʧamsri
ˈlein
, Gloucester Terrace ˈglɔstə ˈterəs, Kentish Town ˈkentiʃ ˈtaun,
Kamden Town ˈkæmdən ˈtaun, 37429 Ladbroke Grove ˈlædbruk ˈgrouv,
Shaftesbury Avenue ˈʃɑːftsbri ˈævinjuː, Herne Hill ˈhəːn ˈhil,
Hampton Court ˈhæmptən ˈkɔːt, Bell Yard ˈbel ˈjɑːd, Hampstead
Heath
ˈhæmpstid ˈhiːθ, Hampstead Way ˈhæmpstid ˈwei. Many
foreign people, and especially Germans, omit the stress on the
second word in many expressions of this kind; they say for instance
263ˈroust biːf, ˈɔːl rait. (They also often use an incorrect intonation, as
in the examples in § 950.) Where, however, the qualifying word
is no, so or too the tendency on the part of foreign people is rather
to omit the stress on the first word and to shorten unnecessarily
its vowel (e.g. to say it ˈwɔz tu ˈmʌʧ instead of it wəz ˈtuː ˈmʌʧ).

960. The case of a verb followed by an adverb, the two words
together forming what is practically a new verb, should be also
specially noted. 38430 Thus in go away, give up, put down, leave out,
turn round, come on, etc., both words are normally stressed.
Examples: Tie got up and went away hiː ˈgɔt ˈʌp ən ˈwent əˈwei,
Put down that parcel! ˈput ˈdaun ðæt ˈpɑːsl, Take it off! ˈteik it
ˈɔf
. Phrases like get ready, make haste which are equivalent to
single verbs are treated in like manner (ˈget ˈredi, ˈmeik ˈheist).

2. Exceptions to General Rule

961. Exceptions to the general rule that nouns, adjectives,
demonstrative and interrogative pronouns, principal verbs, and
adverbs have strong stress (§ 957) are as follows:

1st exceptional case

962. 1st exceptional case. When it is desired to emphasize a
word for contrast, its stress is increased, while the stress of the
surrounding words may be diminished. Thus in the absence of
special emphasis the stress of I never gave you that book is ai ˈnevə
ˈgeiv juː ˈðæt ˈbuk
; but if it were desired to emphasize the word
I or the word you or the word that, we should have three different
ways of stressing the sentence, namely: ˈai nevə geiv juː ðæt ˈbuk
(stress on I and no stress on never), ai ˈnevə geiv ˈjuː ðæt ˈbuk
(stress on you and no stress on gave or that), ai ˈnevə ˈgeiv juː
ˈðæt buk
(stress on that and no stress on book). In I don't object,
if I is stressed, don't is unstressed, thus ˈai dount əbˈʤekt. In
that's your look-out ˈðæt s ˈjɔː lukaut, look-out is not stressed, in
order to give greater force to your. Similarly with concerned in
so far as heˈs concerned sou ˈfɑːr əz ˈhiː z kənsəːnd.

963. In the expression to make sure tə meik ˈʃuə the make is
usually not stressed, presumably in order to give greater force to
264sure; similarly with gave in he gave a final touch… hiː geiv ə
ˈfainl ˈtʌʧ…
.

964. In some people think so ˈsʌm piːpl ˈθink sou there is an
implied contrast with ‘other people,’ therefore people is not stressed. 39431
So also in the latter case ðə ˈlætə keis there is a contrast (expressed
or implied) with some other case. Similarly with way in this
way or that
ˈðis wei ɔːˈðæt, and with instant in it was light one
instant and dark the next
it wəz ˈlait ˈwʌn instant ən ˈdɑːk ðə ˈnekst.
The absence of stress on rate in the expression at any rate ət ˈeni
reit
appears to be due to a similar cause.

965. For the same reason when a sentence contains a word
which has been used just before, that word is generally not stressed.
Examples: How many times have you been there? Three times
hau meni ˈtaimz əv juː ˈbiːn ðɛə? ˈθriː taimz (no stress on the
second times), those who have read about everything are commonly
supposed to understand everything
ˈðouz hu əv ˈred əbaut ˈevriθiŋ
ə ˈkɔmənli səˈpouzd tu ʌndəˈstænd evriθiŋ
(no stress on the second
everything), we think of that as a child thinks wiː ˈθiŋk əv ˈðæt əz
ə ˈʧaild θiŋks
(no stress on thinks), the boys shouted to the other
boys
ðə ˈbɔiz ˈʃautid tə ði ˈʌðə bɔiz (no stress on the second boys),
the house called ‘The Brambles’ was chiefly conspicuous for its lack
of brambles
ðə ˈhaus kɔːld ðə ˈbræmblz wəz ˈʧiːfli kənˈspikjŭəs
fər its ˈlæk əv bræmblz
(no stress on the second brambles).

966. So also when one word in a sequence of two words is naturally
or habitually contrasted with some other word, that word alone
receives the stress. Thus acute angle would generally be pronounced
əˈkjuːt ængl (without stress on angle) even when no contrast with
‘obtuse’ is intended; similarly with railway journey ˈreilwei ʤəːni,
pleasure trip ˈpleʒə trip, lighthouse keeper ˈlaithaus kiːpə, steamship
company
ˈstiːmʃip kʌmpəni, life-saving apparatus ˈlaifseiviŋ æpəreitəs,
high-school teacher ˈhai-skuːl tiːʧə, safety razor ˈseifti reizə. Lady's
265maid
has single stress (ˈleidiz meid), like house-maid ˈhausmeid,
parlour-maid ˈpɑːləmeid, etc. 40432 These cases are sometimes difficult
to distinguish from those mentioned in § 976.

967. Sequences of three words which are equivalent to compound
words are stressed like the compound words mentioned in § 955.
Examples: public school man (ˈ)pʌblik ˈskuːl mæn, high water mark
(ˈ)hai ˈwɔːtə mɑːk.

968. The stressing of this, these, that (demonstrative), those
depends upon the amount of ‘demonstrativeness’ it is desired to
suggest. Sometimes they are equivalent to little more than the
definite article the, and in such cases they are unstressed. This is
especially the case when the noun they qualify refers to something
previously mentioned. Examples of lack of stress on these
words: he managed this matter admirably hiː ˈmæniʤd ðis ˈmætər
ˈædmərəbli
, it was necessary to take these precautions it wəz ˈnesisri
tə ˈteik ðiːz priˈkɔːʃnz
, I don't care for that other one ai ˈdount ˈkɛə
fə ðæt ˈʌðə wʌn
, he couldn't bear the gaze of those eyes hiː ˈkudnt
ˈbɛə ðə ˈgeiz əv ðouz ˈaiz
. This is not stressed in this morning,
this afternoon, this evening (ðis ˈmɔːniŋ, ðis ˈɑːftəˈnuːn, ðis ˈiːvniŋ).

969. When which is used as a demonstrative pronoun (= ‘and
this,’ ‘and these,’ etc.), it is stressed according to the general rule.
Example: which diamond was eventually lost ˈwitʃ ˈdaiəmənd wəz
iˈventjŭəli ˈlost
(meaning ‘and this diamond was…’).

970. The exclamatory what in such expressions as What a
dreadful thing!
, What beautiful weather!, What crowds of people!
is not stressed, presumably in order to give greater emphasis to
dreadful, beautiful, crowds (wɔt ə ˈdredful ˈθiŋ, wɔt ˈbjuːtəfl ˈweðə,
wɔt ˈkrauʣ əv ˈpiːpl).

971. When such is followed by an emphatic word, it is generally
unstressed. Examples: such a curious shape sʌʧ ə ˈkjuərĭəʃ ʃeip,
such quantities of sand sʌʧ ˈkwɔntitiz əv ˈsænd. But when such
is followed by a word incapable of receiving emphasis, e.g. thing,
it is usually stressed; example: such a thing ought to be impossible
ˈsʌʧ ə ˈθiŋ ˈɔːt tə biː imˈpɔsəbl. It may, however, be unstressed
266if the noun it qualifies is also unstressed; example: I don't know
anything about such matters
ai ˈdount nou ˈeniθiŋ əˈbaut sʌʧ
mætəz
, I never heard of such a thing ai ˈnevə ˈhəːd əv sʌʧ ə θiŋ.

972. When the expressions sort of, kind of are used indefinitely,
i.e. not with reference to particular varieties of things, the words
sort, kind are usually not stressed. Examples: there was a sort of
seriousness in his face
ðɛə wəz ə sɔːt əv ˈsiərĭəsnis in iz ˈfeis, they
made a kind of agreement
ðei ˈmeid ə kaind əv əˈgriːmənt. When
these expressions are followed by words which cannot be emphasized,
both are unstressed; example: I don't like that kind of thing ai
dount ˈlaik (ˈ)ðæt kaind əv θiŋ
. Also when sort of is used in
colloquial speech as an adverb meaning ‘in some kind of way,’
it is not stressed; examples: he sort of slipped hiː sɔːt əv ˈslipt,
he slipped sort of hiː ˈslipt sɔːt ɔv.

2nd exceptional case

973. 2nd exceptional case. The double stress in groups of words
such as those mentioned in §§ 959, 960 is often subject to rhythmical
variations
. The following are examples of variations in stress
due to this cause: hot roast beef ˈhɔt roust ˈbiːf, John went away
ˈʤɔn went əˈwei (cp. he went away normally pronounced hiː ˈwent
əˈwei
), a very good thing ə ˈveri gud ˈθiŋ (cp. ə ˈgud ˈθiŋ), not very
good
ˈnɔt veri ˈgud, very much better ˈveri mʌʧ ˈbetə, we can't
get out
wiː ˈkɑːnt get ˈaut 41433 (cp. Get out! ˈget ˈaut), he put on his
hat
hiː ˈput ɔn iz ˈhæt (cp. hiː ˈput it ˈɔn), Go and get ready!
ˈgou ən get ˈredi (cp. Get ready at once! ˈget ˈredi ət ˈwʌns), we
didn't see anything at all
wiː ˈdidnt siː ˈeni8in əˈtɔːl (no stress on
see) (cp. we did not see the exhibition wiː ˈdidnt ˈsiː di eksiˈbiʃn),
the disaster claimed many victims ðə diˈzɑːstə kleimd ˈmeni ˈviktimz
(no stress on claimed), there was nothing going on ðɛə wəz ˈnʌθiŋ
gɔuin ˈɔn
(no stress on going), no one went near it ˈnou wʌn
went ˈniər it
(no stress on went), it seems so funny it ˈsiːmz sou
ˈfʌni
(no stress on so), 42434 we all got home without difficulty wiː
ˈɔil gɔt ˈhoum wiˈðaut ˈdifiklti
(cp. we got home… wiː ˈgɔt
ˈhoum…
).

974. In some cases of this kind two ways of stressing are
possible. Thus he's so much kinder than he used to be may be
267pronounced hiː z ˈsou mʌʧ ˈkaində ðən iː ˈjuːs(t) tə biː or hiː z
sou ˈmʌʧ ˈkaində…
. In so many years, so much more, etc., in
the sense of ‘such a great many years,’ ‘such a great deal more,’
etc., the stress is more usually on many or much (sou ˈmeni ˈjəːz,
sou ˈmʌʧ ˈmɔː). The other stressing, ˈsou meni ˈjəːz, ˈsou mʌʧ
ˈmɔː
, is also possible but seems to be generally avoided, presumably
because when so many, so much are pronounced with stressed so,
they usually have the special meanings ‘such and such a number,’
‘such and such a quantity.’

975. Loss of stress for rhythmical reasons is not always essential
for correct pronunciation. Thus it would not be incorrect to say
ˈhɔt ˈroust ˈbiːf, it ˈsiːmz ˈsou ˈfʌni. When the foreign learner
is in doubt as to whether a stress should be suppressed on account
of rhythm or not, it is safer for him to retain the stress.

3rd exceptional case

976. 3rd exceptional case. When two nouns in sequence are
felt as being very closely connected by the sense, so that they
form practically one word, the second is generally unstressed.
(These groups may really be considered as compound words,
and many of them may be written in ordinary spelling with
hyphens.) Examples: door handle ˈdɔː hændl, gooseberry bush
ˈguzbri buʃ, camping ground ˈkæmpiŋ graund, tennis ball ˈtenis
bɔːl
, golf club ˈgɔlf klʌb, 43435 cricket bat ˈkrikit bæt, diamond merchant
ˈdaiəmən məːʧənt (even when no contrast between dealers in
diamonds and dealers in other goods is intended), violin string
vaiəˈlin striŋ, the Law Courts ðə ˈlɔː kɔːts, chimney corner ˈʧimni
kɔːnə
, barrel organ ˈbærəl ɔːgən, bank note ˈbæŋk nout, 44436 examination
paper
igzæmiˈneiʃn peipə, lemon squeezer ˈlemən skwiːzə, hair-dressing
saloon
ˈhɛədresiŋ səluːn, television set teliˈviʒn set. (It is often
difficult to distinguish this case from that mentioned in § 966).

977. There are some exceptions, namely cases in which the
second element expresses or implies a contrast, e.g. gooseberry
tart
ˈguzbri ˈtɑːt, rice pudding ˈrais ˈpudiŋ, plum cake ˈplʌm ˈkeik,
port wine ˈpɔːt ˈwain (‘tart’ being commonly contrasted with ‘pie,’
‘pudding,’ etc., and ‘pudding’ with ‘meat,’ ‘wine’ with other
beverages, etc.). Saucepan lid would usually be ˈsɔːspən ˈlid, no
doubt owing to an implied contrast between the lid and the
268saucepan itself (cp. churchyard, § 950). Birthday present and
Christmas present have double stress in my pronunciation (ˈbəːθdei
ˈpreznt
, ˈkrisməs ˈpreznt) presumably because present is felt to be
the important word. 45437

4th exceptional case

978. 4th exceptional case. The word street in names of streets
is never stressed, e.g. Oxford Street ˈɔksfəd striːt, Downing Street
ˈdauniŋ striːt (cp. York Road, etc., § 959).

5th exceptional case

979. 5th exceptional case. In phrases of a parenthetical nature
the words are often unstressed. Examples: Has he gone to town
this morning?
ˈhæz iː gɔn tə ˈtaun ðis mɔːniŋ?, How do you do,
Mr. Smith?
hau dju ˈduː mistə smiθ, ‘Yes,’ he said ˈjes hiː sed,
where the phrases this morning, Mr. Smith, he said, are of a
parenthetical nature.

980. The question of stress in such cases is, however, less important
than that of intonation (§ 1071). Thus a certain amount
of stress would often be put on the words mɔːniŋ, smiθ, sed, in
the above examples, provided that the word taun has the lowest
pitch in the first sentence, and that the whole of the phrases
mistə smiθ, hiː sed are pronounced with low pitch. Thus:

image ˈhæz iː gɔn tə ˈtaun ðis mɔːniŋ? | gud ˈmɔːniŋ mistə smiθ | ˈjes hiː sed

Further examples are given in § 1071.

6th exceptional case

981. 6th exceptional case. The various parts of the verb be
are generally unstressed even when the word is a principal verb,
269except when it is final. Examples: the train was late ðə ˈtrein
wəz ˈleit
(cp. the train arrived late ðə ˈtrein əˈraivd ˈleit, in which
the verb is stressed), you are never ready juə ˈnevə ˈredi, What
is the time?
ˈwɔt s ðə ˈtaim? (But it is stressed finally in I don't
know where it is
ai ˈdount nou ˈwɛər iˈtiz, here we are ˈhiə wi ˈɑː,
the chances are… ðə ˈʧɑːnsiz ˈɑː…, the fact is ðə ˈfækt ˈiz, the
reason being…
ðə ˈriːzn ˈbiːiŋ…)

982. The verb be is also unstressed when final and immediately
preceded by its subject, if that subject is stressed. Example:
he asked what the time was hiː ˈɑːskt wɔt ðə ˈtaim wɔz.

7th exceptional case

983. 7th exceptional case. When the subject follows the verb,
the verb is generally not stressed. Examples: ‘Yes,’ said his
father
ˈjes, sed iz ˈfɑːðə (where father may be stressed but said
should not be), after a storm comes a calm ˈɑːftər ə ˈstɔːm kʌmz
ə ˈkɑːm
(no stress on kʌmz).

3. Miscellaneous Rules

984. The following are some miscellaneous facts about stress
which it is necessary for the foreign learner to know.

985. The pronoun one in a good one ə ˈgud wʌn, everyone ˈevriwʌn,
etc., is always unstressed. Foreign people are apt to stress it.
So also with other words that refer to something which has just
gone before, e.g. things in those things ˈðouz θiŋz, matters in I'll
explain matters
ˈai l iksˈplein mætəz, affair in that's my affair
ˈðæt s ˈmai əfɛə.

986. In the expression each other the pronoun each is not stressed,
and other is generally not stressed. Example: they like each other
ðei ˈlaik iːʧ ʌðə. The reflexive pronouns, myself, etc., when used
as object, are generally not stressed. Example: he hurt himself
hiː ˈhəːt imself.

987. Auxiliary verbs are normally not stressed.

988. They are, however, stressed in the following particular
cases:

(i) In affirmative statements for the sake of emphasis, e.g. it
can be done
it ˈkæn bi dʌn, it has been done it ˈhæz bi(ː)n dʌn, 46438
270I may have said so ai ˈmei əv ˈsed sou. The auxiliary do is
always emphasized in such cases, e.g. I do want to ai ˈduː
ˈwɔnt tu
; similarly in imperative sentences, e.g. Do come!
ˈduː ˈkʌm.

(ii) When immediately followed by not pronounced nt, e.g. I
shouldn't have thought so
ai ˈʃudnt əv ˈθɔːt sou, we haven't
been able to
wiː ˈhævnt bi(ː)n ˈeibl tu. 47439

(iii) When introducing a question, e.g. Have you seen them?
ˈhæv juː ˈsiːn dəm?, Did you like it? ˈdid juː ˈlaik it? (In
this case, however, the stress is not essential.)

(iv) In other questions when there is much curiosity, surprise or
anxiety on the part of the speaker and the auxiliary is
immediately preceded by the interrogative word, e.g. What
are you doing?
wɔt ˈɑː juː ˈduiŋ, What is to be done? wɔt ˈiz
tə bi ˈdʌn?
, How did they manage it? hau ˈdid ðei ˈmæniʤ
it?
48440 (But in However did they manage it? hauˈevə did ðei
ˈmæniʤ it?
the did would not be stressed because it does
not immediately follow how.)

(v) When the principal verb is suppressed, e.g. Yes, I have
ˈjes ai ˈhæv, he always does hiː ˈɔːlweiz ˈdʌz.

989. The word going in the expression to be going to… being
of an auxiliary nature is often not stressed. Example: What are
you going to do?
ˈwɔt ə juː goiŋ tə ˈduː. (It would also be possible
to stress going.)

990. The adverbs on, forth, in the expressions and so on ən
ˈsou ɔn
, and so forth ən ˈsou fɔːθ are not stressed. The adverb
again when used to emphasize a contrast is not stressed; examples:
Put it back again (after having taken it away) ˈput it ˈbæk əgein,
heˈs going out again soon (after having come in) hiː z ˈgoiŋ ˈaut
əgein ˈsuːn
, he was dead and is alive again hiː wəz ˈded ənd iz
əˈlaiv əgein
. 49441271

991. The adverbs now and then 50442 are normally stressed, e.g.
How are you now? ˈhau ə juː ˈnau?, I couldn't do it then ai ˈkudnt
ˈdu it ˈðen
. The expression now then is pronounced ˈnau ðen
with stress on now. The adverb so in do so ˈduː sou, think so
ˈθiŋk sou, etc., is not stressed.

992. Adverbs sometimes do not take stress in final position
following a stressed object. Examples: Put your things on! ˈput
jɔː ˈθiŋz ɔn
, he let the fire out hiː ˈlet ðə ˈfaiər aut.

993. Monosyllabic prepositions and the disyllabic preposition
upon əˈpɔn (or əpən) are usually unstressed. Examples may be
found in any book of phonetic texts. These prepositions may,
however, occasionally be stressed when they occur at the beginning
of a sentence; examples: On his way he had an adventure ˈɔn iz
ˈwei hiː ˈhæd n ədˈventʃə
, In the room they found a dog ˈin ðə ˈrum
ðei ˈfaund ə ˈdɔg
.

994. Monosyllabic prepositions are also occasionally stressed
when followed by a pronoun at the end of a sentence (see § 998).

995. Prepositions of two or more syllables (with the exception
of upon, § 993), such as after ˈɑːftə, into ˈintu, 51443 between biˈtwiːn,
during ˈdjuəriŋ, besides biˈsaiʤ, along əˈlɔŋ, concerning kənˈsəːniŋ,
are often stressed in non-final position. Such stress is, however,
not essential in many cases; examples: he went after it hiː ˈwent
ˈɑːftər it
, I'll do it after tea ai ˈdu it (ˈ)ɑːftə ˈtiː, he ran into them
hiː ˈræn ˈintə dəm, he put the money into the box hiː ˈput də ˈmʌni
intə də ˈbɔks
, he searched among his papers hiː ˈsəːʧt ə(ˈ)mʌŋ iz
ˈpeipəz
, he finished it during the holidays hiː ˈfiniʃt it (ˈ)djuəriŋ də
hɔlədiz
.

996. The final prepositions in sentences like What are you looking
at?
ˈwɔt ə juː ˈlukiŋ æt, Who were you talking to? ˈhuː wə juː
ˈtɔːkiŋ tu
, What's all that fuss about? ˈwɔt s ɔːl ðæt ˈfʌs əbaut, we
asked where they came from
wiː ˈɑːs(k)t wɛə ðei ˈkeim frɔm, he wants
looking after
hiː wɔnts ˈlukiŋ ɑːftə, are not stressed though they
have their strong forms. 52444272

997. In sentences ending with a preposition and a pronoun the
final pronouns are not stressed unless special emphasis is needed, 53445
e.g. it's very good for you it s ˈveri ˈgud fɔː ju (or it s ˈveri ˈgud fə
juː
), What shall we do with it? ˈwɔt ʃl wiː ˈduː wið it, Look at them!
ˈluk æt ðəm (or ˈluk ət ðəm). Foreign learners should note that
in these cases the preposition more usually has its strong form
and has noticeably stronger stress than the pronoun.

998. Sometimes it is necessary to stress the preposition in
sentences of this type in order to bring out a contrast, e.g. the bills
were not large but there were a great many of them
də ˈbilz wə ˈnot
ˈlɑːʤ bət ðɛə wər ə ˈgreit ˈmeni ˈɔv ðəm
.

999. Conjunctions introducing dependent clauses are often
stressed when initial. Examples: When he comes I'll introduce
him to you
ˈwen iː ˈkʌmz ai l intrəˈdjuːs im tuː ju, As I was saying…
ˈæz ai wəz ˈseiiŋ…, After he had left… ˈɑːftər iː əd ˈleft…,
nor do I ˈnɔː du ˈai. 54446 If the order of the clauses in the first example
were reversed, then when would not be stressed, because the whole
sentence would be pronounced in one breath-group, and the when
would no longer be initial.

1000. The copulative conjunctions and and but are not generally
stressed. These words may however be stressed, especially when
immediately followed by two or three consecutive unstressed
syllables. Thus and at the same time… may be pronounced
ˈænd ət ðə ˈseim ˈtaim… or ænd ət ðə ˈseim ˈtaim… or ənd
ət ðə ˈseim ˈtaim…
. Even in but it's of the greatest importance
it would be more usual not to stress the but, pronouncing bət it
s əv ðə ˈgreitist imˈpɔːtns
. Foreign learners are recommended to
use the weak forms ənd and bət in all such cases, except where
special emphasis of the conjunction is required.

1001. Other linking conjunctions, such as now, then, introducing
the continuation of a narrative or conversation are not stressed,
e.g. Now when he was gone… nau ˈwen iː wəz ˈgɔn…, Then you
don't believe it?
ðen juː ˈdount biˈliːv it, So he went into the garden
sou iː ˈwent intə ðə ˈgɑːdn.273

Chapter XXX
Breath-Groups, Sense-Groups

1002. Pauses are continually made in speaking. They are made
chiefly (1) for the purpose of taking breath, (2) for the purpose of
making the meaning of the words clear.

1003. It is usual to employ the term breath-group to denote a
complete sentence that can conveniently be said with a single
breath, or, in the case of very long sentences, the longest portions
that can conveniently be said with single breaths.

1004. Pauses for breath are normally made at points where
pauses are necessary or allowable from the point of view of meaning.

1005. Sentences are usually divisible into smaller sequences
between which pauses may be made, though they are not essential.
The shortest possible of such sequences (i.e. sequences which are
not capable of being further subdivided by pauses) are called
sense-groups. Each sense-group consists of a few words in close
grammatical connexion, such as would be said together in giving a
slow dictation exercise.

1006. The divisions between breath-groups are generally made
clear in writing by the punctuation marks. In phonetic transcriptions
it is sometimes useful to mark the division of breath-groups
by the sign ‖ and the division of sense-groups by the sign | .
Another method, which has, however, certain disadvantages, is
not to leave any spaces between consecutive words in breath-groups
or sense-groups. For this see Chap. XXXII and especially § 1094.274

Chapter XXXI
Intonation

The Nature of Intonation

1007. Intonation may be defined as the variations which take
place in the pitch of the voice in connected speech, i.e. the variations
in the pitch of the musical note produced by the vibration of the
vocal cords.

1008. Intonation is thus quite a different thing from stress
(§ 909). There are, however, important relations between stress
and intonation in English, as indeed in all ‘stress languages.’ The
effect of prominence (§§ 101, 208-210) is often produced by certain
combinations of the two.

1009. From the above definition it will be seen that there can
be no objective intonation when voiceless sounds are pronounced.
The number of voiceless sounds occurring in connected speech is,
however, small in comparison with the voiced sounds, 1447 so that
the intonation in any ordinary breath-group may be regarded as
practically continuous. It is certainly subjectively continuous.

1010. In ordinary speech the pitch of the voice is continually
changing. When the pitch of the voice rises we have a rising
intonation
; when it falls we have a falling intonation; when it
remains on one note for an appreciable time, we have level intonation.

1011. The range of intonation is very extensive. It is a noteworthy
fact that most people in speaking reach notes much higher
and much lower than they can sing.

1012. The extent of the range in any given case depends on
circumstances. It is as a general rule greater in the declamatory
style of speech than in conversational style, and in each case it
is greater when the speaker is excited than when he is in a serious
mood. In reciting a passage of a light or humorous character
275it is by no means unusual for a man with an average voice to have
a range of intonation of over two octaves, rising to F image
or even higher and going down so low that the voice degenerates
into a kind of growl which can hardly be regarded as a musical
sound at all. In ordinary conversational English the intonation
(in men's voices) does not often rise above D image.

1013. In the case of women's voices the range of intonation is
not quite so extensive. The average limits for English appear to
be in declamatory style about D image and G image and in
conversation about B image and G image. 2448

1014. A good way of representing intonation for practical
teaching purposes is a system of dots (denoting approximately level
pitches) and curves (denoting rising and falling intonations) placed
above each syllable of a phonetic transcription. It is convenient
to place these marks on a stave of three lines, the upper and lower
lines representing the upper and lower limits of the voice, and the
middle line representing an intermediate pitch.276

1015. It is advantageous to show the stress on the intonation-graph.
This is conveniently done by indicating the strongly stressed
syllables by large dots. If a syllable with a rising or falling intonation
is strongly stressed, this may be shown by placing a large dot
on the appropriate part of the curve (generally at the beginning);
so when a curve has no dot attached to it, it is to be understood
that the syllable is unstressed. 3449

1016. Intonations in language have meanings which are superposed
on the dictionary meanings of the words uttered. They may
convey subtle shades of meaning which could only be expressed by
words in a cumbrous manner, if at all. Compare the following:

image jes. Yes. (meaning ‘That is so’) | image jes. Yes. (meaning ‘Of course it is so’)

image jes. Yes. (meaning ‘yes, I understand that; please continue.’
This form is very frequently used when speaking
on the telephone. The same intonation would be
used in answering a question if a further question
were expected; for instance, shop assistants may be
heard to use it in answering the question ‘Do you
keep so and so?’)

image jes?. Yes?. (meaning ‘Is it really so?’) | image jes. Yes. (meaning ‘That may be so’)

image ˈwɔt ə juː ˈduiŋ?. What are you doing?. (ordinary enquiry) | image ˈwɔt ə juː ˈduiŋ?. What are you doing?. (expressing greater curiosity)277

image wɔt ˈɑː juː ˈduiŋ?. What are you doing?. (expressing still greater curiosity) | image wɔt ə juː duiŋ?. What are you doing?. (you being in contrast with someone else)

image gud ˈmɔːniŋ. Good morning. (on meeting) 4450 | image gud ˈmɔːniŋ. Good morning. (on parting)

image nou. No. (rejoinder to a statement) | image nou. No. (answer to a question)

1017. The principles governing the use of intonation in English
have been well set forth and amply illustrated in several books
and articles, and notably in Coleman's Intonation and Emphasis
in Miscellanea Phonetica I (1912), 5451 Klinghardt's Übungen im
Englischen Tonfall
, 2nd ed., 1927), 6452 Palmer's English Intonation,
with Systematic Exercises (1922), 7453 Armstrong and Ward's Handbook
of English Intonation
(1926), 7454 R. Kingdon's Tonetic Stress Marks for
English
in Le Maître Phonetique, Oct., 1939, articles on The Teaching
of Intonation
by R. Kingdon in English Language Teaching, Jan.,
Feb., March and Sept., 1948, 8455 M. Schubiger's The Role of Intonation
in Spoken English
, 9456 M. Schubiger's English Intonation, its Form and
Function
, 10457 J. D. O'Connor's English Intonation Course, 11458 H. E.
Palmer's New Classification of the English Tones, 5459 W. Jassem's
278Intonation of Colloquial English, 12460 R. Kingdon's The Groundwork of
English Intonation
, 13461 Kingdon's English Intonation Practice, 13462 and
K. L. Pike's The Intonation of American English 14463 In this chapter
only a bare outline of the subject can be given; those who wish to get
a real grasp of English intonation must work through at least the
Armstrong-Ward Handbook, and Kingdon's Groundwork, and
preferably several other of the above-mentioned works.

1018. I find the method of classifying the phenomena of English
intonation adopted by Armstrong and Ward in their Handbook to
be effective in practical teaching, and I accordingly follow their
system in this chapter. They have shown, quite correctly in my
opinion, that most sense-groups in English are said with one of two
fundamental ‘tunes’ or with other ‘tunes’ which are formed by
modifying the fundamental tunes according to definite principles.

1019. The two fundamental tunes are generally known as
‘Tune 1’ and ‘Tune 2’ respectively. 15464 Their particular features are
shown by the following graphical illustrations:

Tune 1 image
Tune 2 image

1020. These tunes may be spread over a large number of
syllables, or they may be compressed into smaller spaces. All the
essential features of the tunes are shown in the above graphical
illustrations. When the tunes are applied to small groups of
syllables or to the extreme case of monosyllables, several of these
features disappear. I find it therefore a good plan in teaching
English intonation to start with the intonation of long sentences
and proceed subsequently to the intonation of short sentences.

Tune 1 (normal form)

1021. The following are some sentences illustrating the normal
form of Tune 1.279

(1) Statements:

image
hiːwəz əbaut ði ˈounli inˈteliʤənt ˈmæn in ˈkʌntri.
He was about the only intelligent man in the country.

image
it s ðə moust iksˈtrɔːdnri ˈθiŋ ai ˈevə ˈhəːd ɔv.
It's the most extraordinary thing I ever heard of.

image
ðei wə ˈveri riˈmɑːkəbl ˈsəːkəmstənsiz.
They were very remarkable circumstances.

image
ai v ˈʤʌst ˈbɔːt ə ˈnjuː ˈpɛər əv ˈglʌvz.
I've just bought a new pair of gloves.

image | image
it s ˈʤʌst ˈfɔːr əˈklɔk. | it ˈɔːl ˈhæpnd ˈjestədi.
It's just four o'clock. | It all happened yesterday.

image
ðɛə z ˈnʌθiŋ tə bi ˈdʌn əbaut it.
There's nothing to be done about it.

image
wiː ˈdid wɔt wiː wə ˈtould.
We did what we were told.
280

image | image
it wəz ˈkwait imˈpɔsəbl. | ai ˈdidnt ˈɑːsk ju tu.
It was quite impossible. | I didn't ask you to.

image | image
it s ˈkwait ˈrɔŋ. | hiː ˈwɔnts it.
It's quite wrong. | He wants it.

image | image | image
ai m ˈgouiŋ. | hiː z ˈgɔn. | juː ˈkɑːnt.
I'm going. | He's gone. | You can't.

(2) Questions containing a special interrogative word:

image | image
ˈhuː wə juː ˈtɔːkiŋ tu?. | ˈwɔt s ðə ˈmætə?.
Who were you talking to?. | What's the matter?.

(3) Commands:

image | image
ˈgou ənd ˈoupn ðə ˈwindou. | ˈteik it əˈwei.
Go and open the window!. | Take it away!.

image | image | image | image
ˈget əˈlɔŋ wið ju. | ˈgou əˈwei. | duː. | trai.
Get along with you!. | Go away!. | Do!. | Try!.

Numerous other examples will be found in Armstrong-War
Handbook, pp. 11-17.281

1022. It will be observed that the characteristic features of an
unmodified Tune 1 are as follows:

(a) Initial unstressed syllables are rather low, and when there
are two or more they are all said on about the same pitch.

(b) The stressed syllables form a descending sequence of notes,
the first being on a rather high level pitch and the last having
a falling intonation.

(c) When there is more than one stressed syllable, the fall of
the last stressed syllable generally begins at a pitch near to
that of the initial unstressed syllables, and falls to the lower
limit of the voice-range. The precise pitch at which the fall
begins depends to some extent on the number and height
of the stressed syllables preceding.

(d) Unstressed syllables between stressed syllables have the same
pitch as the preceding stressed syllable, except in the case
of unstressed syllables immediately preceding the last of the
stressed syllables. In the latter case the last unstressed
syllable is somewhat lower than the preceding stressed
syllable. (Sometimes two unstressed syllables are lowered
in this situation.)

(e) Final unstressed syllables are said with low level pitch.

Tune 2 (normal form)

1023. The following are some illustrations of Tune 2. They
should be compared with the sentences in § 1021.

(1) Questions requiring the answer ‘yes’ or ‘no’:

image
bət ˈwɔz hiː ði ˈounli inˈteliʤənt ˈmæn in ðə ˈkʌntri?
But was he the only intelligent man in the country?

image
bət ˈiznt it ðə moust iksˈtrɔːdnri ˈθiŋ juː ˈevə ˈhəːd ɔv?
But isn't it the most extraordinary thing you ever heard of?
282

image
ˈdid it ˈɔːl ˈhæpn ˈjestədi?
Did it all happen yesterday?

image
ˈiznt ðɛər ˈeniθiŋ tə bi ˈdʌn əbaut it?
Isn't there anything to be done about it?

image | image
ˈdid juː ˈlaik it? or did juː ˈlaik it?
Did you like it?

image | image
ˈdjuː ˈlaik it? or djuː ˈlaik it?
Do you like it?

image | image
ˈiz hiː ˈgɔn? or iz hiː ˈgɔn?
Is he gone?

image | image | image
ˈduː juː? | ˈhæʒ ʃiː? | ˈkɑːnt wiː?
Do you? | Has she? | Can't we?

image | image
hiː ˈwount? | ou?
He won't? | Oh?
(= Do you mean to say that he won't?) | (= Is that really so?)
283

(2) First parts of sentences:

image
ai d ˈʤʌst ˈbɔːt ə ˈnjuː ˈpɛərəv ˈglʌvz, …
I'd just bought a new pair of gloves
(and was walking out of the shop).

image
it s ˈʤʌst ˈfɔːr əˈblɔck, …
It's just four o'clock
(so I think I'll be going).

image
əz it wəz ˈkwait imˈpɔsəbl tə ˈfiniʃ it, …
As it was quite impossible to finish it
(we didn't hurry ourselves). 14465

image | image
wiː ˈdid wɔt wiː wə ˈtould, … | ai ˈsent him əˈwei, …
We did what we were told | I sent him away
(but it wasn't any use). | (but he came back again).

image | image
wiː ˈgɔt him ˈaut əv it, … | hiː went ˈin, …
We got him out of it | He went in
(as soon as we could). | (but found nobody there).
284

image
ai ˈlaik it, …
I like it
(because it's amusing).

(3) Statements with an implication:

image
ai ʃəd biˈglæd if juːd ˈhelp him
I should be glad if you'd help him
(if you possibly can).

image
it ˈiznt ˈbæːd
It isn't bad.
(But at the same time it's none too good.)

Further examples of Tune 2 will be found in Armstrong-Ward,
Handbook, pp. 22-24.

1024. It will be observed that the characteristic features of
Tune 2 are as follows:

(a) Initial unstressed syllables are rather low, as in the case
of Tune 1.

(b) When there is more than one stressed syllable, the first has
rather a high pitch and the last has a low pitch. The intervening
syllables (both stressed and unstressed) are said on a
descending sequence of notes.

(c) Unstressed syllables following the last stressed syllable are
said on an ascending sequence of notes. When there are no
such unstressed syllables, this rising intonation is put on to
the last stressed syllable.

(d) The pitch of the last stressed syllable (or its initial pitch
if it has a rise) is generally lower than that of the initial
285unstressed syllables. It may, however, be on the same level
with them when it is the only stressed syllable in the sense-group.

Variations in the Treatment of Unstressed Syllables

1025. The form of Tune 1 described in §§ 1019-1022 appears to
me the most convenient standard form to teach to foreign students,
being distinctive and easy to learn. The following variations are,
however, permissible in all ordinary cases, but the variations are
never essential.

(a) Initial unstressed syllables may be said as a rising sequence,
ascending towards the pitch of the first stressed syllable. Thus
It's a most extraordinary thing it s ə moust iksˈtrɔdnri ˈθiŋ may
be said with the intonation

image

instead of with the intonation

image

(b) Unstressed syllables between stressed syllables may be said
on a falling sequence between the pitches of the preceding and
following stressed syllables, 16466 or they may be said on a rising
sequence ascending from the pitch of the preceding stressed syllable.
Thus it s ðə moust iksˈtrɔdnri ˈθiŋ ai ˈevə ˈhəːd ɔv may be said
with either of the two following intonations instead of with what
I have called the ‘normal’ intonation shown in § 1021:

image286

The second of these methods of treating medial unstressed syllables
appears to introduce (in the words of Armstrong and Ward) ‘an
element of surprise, cheerfulness, enthusiasm or more interest.’ 17467

1026. In Tune 2, initial unstressed syllables are likewise often
said as a rising sequence. Thus But did you ever see one? bət
did juː ˈevə ˈsiː wʌn?
would often be said with the intonation

image

instead of with the intonation

image

1027. In Tune 2 as in Tune 1 medial unstressed syllables may
be said on a rising sequence ascending from the pitch of the preceding
stressed syllable. Thus it is possible to pronounce bət ˈwɔz hiː ði
ˈounli inˈteliʤənt ˈmæn in ðə ˈkʌntri?
with the following intonation

image

instead of with the normal intonation shown in § 1023. Other
examples are:

image | (instead of image)
ˈput ɔn joː ˈkout
Put on your coat!

image
wiː ˈtraid it ˈouvər ənd ˈouvər əˈgein
We tried it over and over again.

This variation of the tune appears to imply incredulity when
applied to questions, encouragement when applied to commands,
and cheerfulness or facetiousness or protest when applied to
statements.287

Long Sentences with Tune 1

1028. When a group requiring Tune 1 is rather long, the tune
is often modified by raising the pitch of one of the stressed syllables,
as shown in the following example:

image
ai ˈsɔː ðə ˈmæn ˈkʌmiŋ əˈlɔŋ ðə ˈroud.
I saw the man coming along the road.

This sentence might also be pronounced with an unmodified
Tune 1, thus

image

Further examples of this will be found in Armstrong-Ward,
Handbook, p. 19.

Use of tune 1 (normal form)

1029. Tune 1 is the intonation of a sense-group which is a plain
statement of fact, when there is no unspoken implication and no
contrast-emphasis on any particular word. Examples of such
statements will be found in § 1021.

1030. It is also the intonation of questions containing a specific
interrogative word, such as how, when, which, why. Two examples
are given in § 1021 (2). The following are some further illustrations.
For others see Arm strong-Ward, Handbook, pp. 14, 15.

image | image
ˈhau d juː ˈmæniʤ it? | bət ˈhau meni ˈɑː ðɛə?
How do you manage it? | But how many are there?

image | image | image | image
ˈhau ˈmeni? | ˈwɛə z ˈðæt? | hau? | wai?
How many? | Where's that? | How? | Why?
288

1031. Tune 1 is also the intonation of commands and invitations
(as distinguished from requests, see §§ 1040-1041). The
following are examples to supplement those in § 1021 (3). Others
will be found in Armstrong-Ward, Handbook, pp. 16, 17.

image | image
ˈkʌm ˈhiər ət ˈwʌns. | ˈdount ˈteik əniˈnoutis ɔv ðɘm.
Come here at once! | Don't take any notice of them.

image | image
ˈduː wɔt ai ˈtel juː. | ˈkʌm ˈɔn.
Do what I tell you! | Come on!
(As said to a dog, for
instance. Compare
the intonation used
when addressing a
person, § 1041.)

image | image
ˈkʌm ən ˈdain wið əs. | ˈlet s ˈget səˈmɔː.
Come and dine with us. | Let's get some more.

1032. When a plain statement of fact is said in more than one
sense-group, Tune 1 is the normal intonation of the last group.
Preceding groups are usually said with Tune 2 (§ 1033), but sometimes
with Tune 1 (§ 1044). Some examples of the use of Tune 1
in final sense-groups are given in §§ 1034, 1035.

Use of tune 2 (normal form)

1033. Tune 2 is essentially the intonation of unfinished sentences
and of non-final portions of sentences. When a sentence is divisible
into two or more sense-groups, Tune 2 is the intonation generally
used on the non-final groups.

1034. Examples of Tune 2 in non-final portions of a sentence
are seen in the examples in § 1023 (2) and in the first parts of
the following:

image
wiː ˈkeim ˈhoum bikɔz it wəz ˈreiniŋ.
We came home because it was raining.
289

image
ˈwen wiː get ˈhoum, ai l ʃou juː ə ˈpiktʃər ɔv it.
When we get home, I'll show you a picture of it.

image
ˈif it səkˈsiːdz, ai ʃl meik ə ˈfɔːtʃn.
If it succeeds, I shall make a fortune.

image
hiː ˈspəndz iz ˈmʌni əz ˈif iː wər ə miljəˈnɛə.
He spends his money as if he were a millionaire.

image
ˈwen juː gɔt tə ˈlʌnden, ˈdid juː gou ˈstreit tə ðə houˈtel?
When you got to London, did you go straight to the hotel?.

1035. In the following examples two or more non-final groups
are said with Tune 2:

image
hiː ˈtuk iz ˈhæt, ˈræn daunˈstɛəz,
He took his hat, ran downstairs,

image
ənd ˈhʌrid ˈaut intə ðə ˈstriːt.
and hurried out into the street.

image
əz ˈsuːn əz wiː əˈraivd, wiː ˈtuk ə ˈtæksi,
As soon as we arrived, we took a taxi,
290

image
ənd went ˈstreit tə ðə ˈhaus tə ˈsiː if aue ˈfrend wəz ətˈhoum.
and went straight to the house to see if our friend was at home.

Further examples will be found in Armstrong-Ward, Handbook,
p. 34.

1036. Alternative questions are a particular case of this form
of intonation. The last alternative is said with Tune 1, and the
preceding alternatives are normally said with Tune 2. The
following are examples:

image
ˈʃæl wi ˈdraiv ɔː ʃl wi ˈgou bai ˈtrein?
Shall we drive or shall we go by train?

image
ˈdjuː laikˈtiː ɔː ˈkɔfi ɔː ˈkoukou?
Do you like tea or coffee or cocoa?

1037. Another particular case of this normal intonation is seen
in enumerations of things. Examples:

image
ðɛə wər ˈæplz ən ˈpɛəz ən ˈplʌmz.
There were apples and pears and plums.

image
ˈwʌn, ˈtuː, ˈθriː, ˈfɔː, ˈfaiv.
One, two, three, four, five.

1038. Very often a sentence which is complete in form is said
with Tune 2 because a continuation is implied though not expressed
291in words, or because the sentence requires a rejoinder from the
person addressed. The following are examples of statements
pronounced in this way. They imply some such continuations
as those shown in brackets.

image
it ˈwount ˈteik miː ˈlɔŋ.
It won't take me long.
(So you may expect to see me back soon.)

image
ai wəz ˈounli ˈwundriŋ.
I was only wondering.
(But couldn't come to any conclusion.)

image
it wəz ˈnou ˈjuːs.
It was no use.
(The thing couldn't be avoided.)

Other examples are enumerations in which the alternatives
mentioned do not exhaust the possibilities, e.g.

image
wiː mait gou tə ˈlʌnden ɔː ˈpæris ɔː bəːˈlin.
We might go to London or Paris or Berlin.
(Or some other place.)

image
ðɛə wər ˈæplz ən ˈpɛəz ən ˈplʌmz.
There were apples and pears and plums.
(And other kinds of fruit.)
292

image
ˈwʌn, ˈtuː, ˈθriː, ˈfɔː, ˈfaiv.
One, two, three, four, five.
(Six, etc.)

The intonations of the last two sentences should be compared
with those given in § 1037, where there are no further alternatives.

1039. On the whole, statements other than enumerations are
not often said with an unmodified Tune 2. They are, however,
very often said with a modified Tune 2 (§ 1051), since they frequently
contain a word requiring contrast-emphasis.

1040. The commonest kinds of sentence pronounced with an
unmodified Tune 2 in final position are ordinary requests and
questions requiring the answer ‘yes’ or ‘no.’ Requests have Tune 2
presumably because they imply that the person addressed is given
the alternative of refusing to accede to them. Questions requiring
the answer ‘yes’ or ‘no’ have this intonation because they imply
the continuation ‘or not.’ 18468

1041. The following are some examples of requests. They should
be compared with the commands in §§ 1021 (3), 1031.

image
ˈgiv mai ˈkaind riˈgɑːdz tə jɔː ˈbrʌðe.
Give my kind regards to your brother.

image
ˈdount ˈgou ən ˈmeik ə ˈfuːl əv jɔːself.
Don't go and make a fool of yourself.

image | image
ˈʤʌst ˈʃʌt ðə ˈwindou. | ˈduː ˈkʌm ən ˈsiː əs.
Just shut the window. | Do come and see us.
293

image | image
ˈteik it əˈwei. | ˈkʌm ˈɔn.
Take it away! | Come on!
(Compare the
intonation of
this sentence
as a command,
§ 1021 (3).)
(Spoken to a person;
compare the intonation
used when the
words are addressed
to a dog, § 1031.)

Further examples of requests will be found in Armstrong-Ward,
Handbook, p. 24.

1042. The following are some examples of questions requiring
the answer ‘yes’ or ‘no,’ to supplement those given in § 1023.
Further examples will be found in Armstrong-Ward, Handbook,
pp. 23, 24.

image
ˈʃæl wiː ˈget səm ˈæplz?
Shall we get some apples?

image
ˈhæv juː ˈbiːn tə ði əksiˈbiʃn?
Have you been to the exhibition?

image
ˈhæv juː ˈeva ˈbiːn ðɛə?
Have you ever been there?

image
ˈdidnt ai ˈsiː juː ət ðə ˈsteiʃn ði ʌðə dei?
Didn't I see you at the station the other day?

image
ˈdjuː ˈlaik ˈðis ˈbuk witʃ ai ˌbɔːt ði ʌðə ˌdei?
Do you like this book which I bought the other day?
294

image
ˈwil juː ˈkʌm ən ˈdain wið əs?
Will you come and dine with us?

image
ˈiz it ˈgouiŋ tə bi ˈfain tedei?
Is it going to be fine to-day?
(= Do you think it is going to be fine to-day?)

image
ˈdid ðei ˈsei sou?
Did they say so?
(= Do you know whether they said so?)

image
ˈdid juː ˈɑːsk im wɔt iː ˈθɔːt əbaut it?
Did you ask him what he thought about it?

image | image
ˈdidnt hi? | ˈwil juː?
Didn't he? | Will you?

Tune 1 in non-final Groups

1043. Tune 1 is sometimes used in non-final groups. The
commonest case is when the following group expresses a reservation
as in the following examples:

image
it teiks əbaut ˈtuː ˈauəz ˈʤenrəli.
It takes about two hours generally.
295

image
ai l ˈlʌk fə wʌn if ju ˈlaik.
I'll look for one if you like.

image
wiː l stɑːt iˈmiːdjətli if juə ˈredi.
We'll start immediately if you're ready.

image
it s ə ˈvəri gud ˈθiŋ ɔn ðə ˈhoul.
It's a very good thing on the whole.

(The non-final groups in these examples might also be said with
Tune 1 modified for contrast-emphasis; like the examples in
§ 1049.) For further examples see Armstrong-Ward, Handbook,
p. 35.

1044. The following are further examples of the use of Tune 1
in non-final sequences. The reasons for its use are not always easy
to establish.

image
ail ˈʃou it tju wen wi get ˈhoum.
I'll show it to you when we get home.

image
ai ˈspouk tu im ɔn ðə ˈtelifoun.
I spoke to him on the telephone.

image
ˈwen juː ˈgɔt tə ˈlʌnden, ˈdid juː gou ˈstreit tə ðə houˈtel?.
When you got to London, did you go straight to the hotel?
(Compare the alternative intonation in § 1034.)
296

image
wiː ʃl ˈsiː juː biˈfɔː juː ˈgou.
We shall see you before you go.

Many other examples will be found in the texts in the Armstrong-Ward,
Handbook.

Emphasis

1045. When it is desired to give emphasis to a particular word
in a sentence, that word has to be said with greater prominence
than usual. As has already been pointed out in §§ 208-210, 911,
912, special prominence may be given (1) by increasing the length
of one or more sounds, (2) by increasing the stress of one or more
syllables, (3) by using special kinds of intonation, or by combinations
of these means. It is also to be noted that when a word
can be pronounced in more than one way, a fuller or strong form
is used in emphasis. Such full or strong forms do not of themselves
give prominence 19469; the prominence is effected by means of
the sound-attributes (prosodies) length, stress and intonation.
Of the above-mentioned methods of effecting prominence intonation
is the most important; it is generally, though not necessarily,
combined with extra strong stress on the emphatic word.

1046. There are two kinds of emphasis, which may be termed
297emphasis for contrast and emphasis for intensity. 20470 The first is
emphasis intended to show that a word is contrasted with another
word (either implied or previously expressed), or that a word
introduces a new and unexpected idea. The second is an extra
emphasis to express a particularly high degree of the quality
which a word expresses; it is equivalent to the insertion of such
words as very, extremely, a great deal of. Contrast-emphasis may
be applied to almost any word, but intensity-emphasis can only
be applied to certain words expressing qualities which are measurable,
e.g. adjectives such as huge, enormous, lovely, tremendous,
wonderful, marvellous, appalling, awful, tiny, absurd, killing, brilliant,
deafening, 21471 adverbs such as particularly, extremely, hopelessly,
plural nouns such as quantities, masses, heaps, tons, hundreds, and
a certain number of verbs such as rush, squeeze, hate.

1047. Contrast-emphasis is expressed mainly by intonation. The
special intonation may be accompanied by extra stress or length,
but these are secondary.

1048. It often happens that a word has both kinds of emphasis
simultaneously, see § 1061.

Emphasis for contrast

1049. The following are typical examples of the effect of contrast-emphasis
in a sentence which ends with low pitch. The
intonation may be regarded as a modification of Tune 1.

image
ai θɔːt hiː wəz gouiŋ tə ˈlʌnden.
I thought he was going to London.

(The other speaker having just said that his friend was going to
Edinburgh.)298

image
it s ðə ˈsaiz əv it ðət s sou əsˈtɔniʃiŋ.
It's the size of it that is so astonishing.

image
bət ðɛə z ˈnou riˈzemblens bitwiːn ðɘm.
But there's no resemblance between them.

image
ai ˈdidnt ˈɑːsk ju tu.
I didn't ask you to.

(Compare other intonations of this sentence shown in §§ 1021 (1),
1051 (1), 1051 (2).)

image
ˈduː wɔt ai ˈtel ju.
Do what I tell you.

Further examples will be found in Armstrong-Ward, Handbook,
pp. 52, 53.

1050. It will be seen that in sentences of this kind the only
syllable with a really strong stress is the stressed syllable of the
emphatic word. Other syllables may have a medium or fairly
strong stress, but they have the intonation of unstressed syllables.
The intonation is therefore a particular case of that shown in
§ 1025 (a).

1051. The following are typical examples of the effect of contrast-emphasis
in a group with rising intonation. The intonation may
be regarded as a modification of Tune 2.299

(1) No stress preceding the emphatic word:

image
ˈʤɔn dʌnzt ˈkɛər əbaut it.
John doesn't care about it.
(Implying ‘though other
people may.’)

image
ai iksˈpəkt ðeiə ˈkʌmiŋ.
I expect they're coming.
(Implying ‘but I can't say
for certain.’)

image
ai ˈhoup ðeil biˈeibl tu.
I hope they'll be able to.
(Suggesting ‘but I rather
doubt if they will.’)

image
ˈai didnt ˈɑːsk ju tu.
I didn't ask you to.
(‘I’ contrasted with someone else.
Compare other intonations shown
in §§ 1021 (1), 1049, 1051 (2).)

image
ai ˈθink ˈðæt s wɔt it ˈwɔz.
I think that's what it was.
(Implying ‘though I'm
not quite sure.’)

image
ai ˈθink it s ɔn ˈsætedi.
I think it's on Saturday.
(ditto)

image
ai ˈθink sou.
I think so.
(ditto)
300

image
ˈðæt dʌznt ˈmætə.
That doesn't matter.
(‘That’ contrasted with some
thing else.)

image
it ˈdʌznt mætə. 22472
It doesnˈt matter.
(Said with this intonation the
words are intended to reassure
the person spoken to, in case
he should think that the thing
did matter.)

image
it s ˈriːznəbl inʌf.
It's reasonable enough.
(Implying ‘though perhaps
not very feasible.’)

(2) With one or more stresses preceding the emphatic word:

image
ai ˈdidnt ɑːsk ˈjuː tu.
I didn't ask you to.
(‘You’ contrasted with someone
else. Compare other intonations
shown in §§ 1021 (1), 1049, and
1051 (1).)

image | image
ai ˈdidnt ˈɑːsk ju tu. or … ˈɑːsk ju tu.
I didn't ask you to.
(Implying ‘you did it of your own accord.’ Compare other
intonations of the same words in §§ 1021 (1), 1049, and 1051 (1).)
301

image
it ˈiznt ðæt ai ˈwɔntid tə ˈsiː ju əbaut.
It isn't that I wanted to see you about.
(Implying ‘it was something else.’)

image | image
it s ə ˈgud ˈbildiŋ ɑːkiˈtəktʃərəli. or … ɑːkiˈtəkʃərəli.
It's a good building architecturally.
(Implying ‘though not quite what is wanted from other points
of view.’)

Further examples will be found in Armstrong-Ward, Handbook,
pp. 58-67.

1052. It will be seen that in sequences of this kind the emphasized
word has a high falling intonation on its stressed syllable, that the
terminal rise begins on the last stressed or semi-stressed syllable
of the group, and that intervening syllables all have low pitch.
When there is no stressed or semi-stressed syllable after the emphatic
word, the terminal rise begins at the syllable following the emphatic
fall (see § 1054).

1053. The highest pitch in the emphatic syllable is at the point
of maximum stress in the syllable. In modified Tune 2 it is led
up to by a sharp rise, which is generally extremely short and often
hardly perceptible when the syllable has no initial consonant or
begins with a voiceless consonant. This rise is thus clearly audible
in the words reasonable and you in the last example in § 1051 (1) and
the first example in § 1051 (2); but in such words as think, ask, in
the fifth example in § 1051 (1) and the second example in § 1051 (2),
it is too short to be easily heard, and may even be objectively absent.
The speaker has, however, a subjective feeling of its presence in all
such cases. (This preparatory rise is shown in the graphical
representation of the intonation by the line preceding the stress-dot;
the line is dotted in cases where the rise is not clearly audible.)

1054. If there are several syllables following the emphatic fall,
the terminal rise is spread over them. But if there is only one
302unstressed syllable following, the terminal rise is compressed into
it. If there is no following unstressed syllable, the terminal rise
is compressed into the same syllable as the emphatic fall; the
emphatic syllable is therefore said in this case with a fall-rise.
The following are some examples to illustrate these points:

image
ai l kənˈsidər it. or ai l kənˈsidər it.
I'll consider it.
(Implying ‘though I can't promise to do it.’)

image
ai ˈwɔnt tə du it.
I want to do it.
(Implying ‘but I don't know if I
shall be able to.’)

image
it s ˈʤʌst ˈpɔsəbl.
It's just possible.
(Implying ‘though not likely.’)

image
juː d ˈbəte.
You'd better.
(Implying a clause beginning with
‘or else….’)

image
ai m ˈɔːfli ˈsɔri.
I'm awfully sorry.
(Implying ‘but it couldn't be
avoided.’)

image
ai njuː iː kept ˈhɔːsiz.
I knew he kept horses.
(Implying ‘but I didn't know
he kept any other animals.’)
303

image
it ˈiznt ˈbæd.
It isn't bad.
(Implying ‘but at the same time it's
not very good’.)

image
wiː ˈkɑːnt ˈduː it təˈdei.
We can't do it to-day.
(Implying ‘though we might
perhaps to be able to tomorrow.’)

image
ˈðæt s wɔt iː ˈsed.
That's what he said.
(Implying ‘though I don't know
if it's what he meant.’)

image
ai ˈwil if ai ˈkæn.
I will if I can.
(Implying ‘but I doubt if I shall
be able to.’)

image
it ˈwɔznt ˈauəz.
It wasn't ours.
(Implying ‘it must have been
someone else's.’)

Numerous further examples will be found in Armstrong-Ward,
Handbook, pp. 66-69, 72, 73.

1055. In connexion with the fall-rise on a single syllable the
following details should be observed:

(1) When the syllable ends in m, n, ŋ or l, the lowest pitch
is reached at the beginning of this consonant, and the whole
of the rise takes place during the pronunciation of the
consonant. Thus in the example I will if I can given in
the preceding paragraph, the whole of the rise takes place
during the n.304

(2) When the syllable contains a short vowel followed by b, d
or g, the whole of the rise generally takes place during the
‘stop’ of this consonant. Thus in the example That's what
he said
given in the preceding paragraph, the whole of the
rise generally takes place during the ‘stop’ of the d. (In the
speech of those who completely devoice final b, d, g, the
rise takes place on the latter part of the vowel.)

(3) In other cases the rise begins about the middle of the vowel.
Examples are seen in the words bad and to-day in the seventh
and eighth examples in § 1054.

(4) When the syllable contains a short vowel followed by a
voiceless consonant, the intonation has to be compressed into
a particularly short space. The final rising part is then so
short as to be difficult to hear. The speaker has, however,
a subjective feeling of its presence. This would be the case,
for instance, in
image
ˈnot ˈjet.
Not yet.
(Implying ‘but perhaps
later on.’)

1056. The intonation described in §§ 1051-1055 is sometimes
used in situations where there does not appear to be any obvious
contrast, and where it is therefore difficult to specify the reason
for the use of the intonation. Notable cases are expressions of
regret
and entreaties or urgent requests.

Examples of expressions of regret:

image
wiə ˈsou ˈsɔri.
We're so sorry.
(Pronounced in this way even if no
excuse is implied; compare I'm awfully
sorry
in § 1054.)

image
wiː wə ˈsou ˈsɔri ˈnot tə bi ˈeibl tə ˈkʌm.
We were so sorry not to be able to come.
305

image
ai ˈbeg jɔː ˈpɑːdn.
I beg your pardon.
(Meaning ‘I'm sorry’; compare the
other intonation given in § 1063)

Examples of entreaties or urgent requests (to be compared
with the ordinary requests in § 1041):

image
ˈduː kʌm ˈɔn.
Do come on!.

image
ˈgiv mai ˈkaind riˈgɑːdz tə jɔː ˈbrʌðe.
Give my kind regards to your brother.

image
ˈpliːz dount ˈtrʌbl.
Please don't trouble.

1057. When a word has to be emphasized for contrast in a
question requiring the answer ‘yes’ or ‘no,’ the intonation is an
ordinary Tune 2 with the emphatic syllable at the point of lowest
pitch. All following syllables are generally unstressed, but if any
of them have a certain degree of stress, their intonation is as if
they were unstressed. The effect of contrast is often made more
marked by pronouncing preceding unstressed syllables with high
pitch. Examples:

image
ˈdjuː ˈθink ðæt s ment fə ˈmiː?
Do you think that's meant for me?
306

image
iz ˈðæt wɔt juː ment?
Is that what you meant?

image
hæv ˈjuː evə biːn dee?
Have you ever been there?
(Compare the third example in
§ 1042.)

image
did ˈjuː laik it?
Did you like it?
(Compare the intonations given in
§ 1021 (1).)

1058. Tune 1 (with or without an emphatic word) is sometimes
applied to questions requiring the answer ‘yes’ or ‘no.’ When said
in this way the questions embody the idea of some statement or
invitation. Examples:

(1) Without contrast-emphasis:

image
ˈhæv juː ˈbiːn tə ði əksiˈbiʃn?
Have you been to the exhibition?
(Suggesting ‘I don't expect
you have,’ or ‘You really
ought tp go.’)

image
ˈʃæl wi ˈget səm ˈæplz?
Shall we get some apples?
(Suggesting ‘it would be a good
idea to get some apples.’ Compare
the intonation of the same
words in § 1042.)

image
ˈwil juː kʌm ən ˈdain wið əs?
Will you come and dine with us?
(= ‘I invite you to come
and dine with us.’ Compare
the intonation of the
same words in § 1042.)
307

image
ˈiz it gouiŋ tə bi ˈfain tedei?
Is it going to be fine to-day?
(=‘I wonder if it will be fine to-day.’ Compare the intonation
of the same words in § 1042.)

(2) With contrast-emphasis:

image
ˈdid ðei ˈsei sou?
Did they say so?
(= ‘it is open to question whether they
said so.’ Compare the other intonation
of the same words in § 1042.)

image
ˈðætz s ðə ðiˈrəkʃn, ˈiznt it?
That's the direction, isn't it?
(‘Isn't it’ being an invitation to assent, and not expressing a desire
for information. 23473)

image
ʃl wiː get səm ˈæplz den?
Shall we get some apples then?
(= ‘In that case I suggest that we get some apples.’ Compare
the intonation of the first example in § 1042.)

1059. It happens not unfrequently that two words in the same
sentence have contrast-emphasis. The following are examples:

(1) Single groups:

image
ˈʌðə piːpl dount ˈlaik it.
Other people don't like it.
308

image
ai ˈhoup ðeil bi ˈeibl tu.
I hope they'll be able to.

image
juː ˈdount siːm tə ˈkeər əbaut ˈʌðə θinz.
You don't seem to care about other things.

(2) Sentences consisting of two groups:

image
ˈif it səkˈsiːdz, ai ʃl meik ə ˈfɔːtʃn.
If it succeeds, I shall make a fortune.
(Compare alternative
intonation in § 1034.)

image
if ðə ˈskiːm ˈfeilz, it ˈwount əˈfəkt ˈjuː.
If the scheme fails, it won't affect you.

Further instances will be found in Armstrong-Ward, Handbook,
pp. 54-56.

Emphasis for Intensity

1060. Intonation is often employed (in addition to length and
stress) to intensify the meaning of words expressing measurable
qualities as explained in § 1046. The modification of intonation
for this purpose generally takes the form of increasing the pitch
intervals. The pitch of the emphatic syllable is generally led up
to by a rise. For instance, it's enormous may be said thus:

image or thus image
it s iˈnɔːməs | it s iˈnɔːməs

the second intonation giving an idea of greater size.309

Further illustrations of intensity-emphasis are shown in the
following examples:

image
ðə ˈhoul ˈθiŋ wəz ə triˈmenːdəs səkˈses.
The whole thing was a tremendous success.

image
ðɛə wə ˈmæsiz ən ˈmæsiz ɔv it.
There were masses and masses of it.

In the following example the emphasized word has a low pitch:

image or image
iznt it ebˈsəːd? | iznt it əbˈsəːd?
Isn't it absurd?

This might also be said with the intonation image

In all cases there is very strong stress on the emphatic syllable.
There is generally also a lengthening of it. In enormous the length
is given mainly by an extra long ɔː, in tremendous by lengthening
the m and n, in masses by lengthening the m and s, and so on.

1061. Intensified words often have contrast-emphasis as well as
intensity-emphasis. In this case the emphatic syllable has to be
said with extra length, extra stress and special intonation.

Thus it was a tremendous success might be pronounced as follows
in reply to someone who said ‘I hear it was a great success’:

image
it wəz ə triˈmenːdəs səkˌses.
310

Similarly in

image
hiːz ə ˈwʌnːdəfl mæn.
He's a wonderful man.
(Implying ‘though not
properly appreciated.’)

Special Intonations

1062. Coleman 24474 and others have pointed out that the reasons
for the use of particular intonations are sometimes obscure. The
following are some examples of intonations which are difficult to
explain.

1063. A gradual rise of intonation is used when the speaker
desires the person addressed to repeat what he said before.

Examples:

image | image
ˈwɔt did ju sei? | wɔt?
What did you say? | What?

image
hau meni did ju sei?
How many did you say?

image
ˈhau meni?
How many?
(Meaning ‘How many did you say?’ Compare
the ordinary intonation of How many? shown
in 1030.)

image
ai ˈbeg joː ..pɑːdn.
I beg your pardon.
(Meaning ‘What did you say?’ Compare
the other intonation shown in S 1056.)
311

This intonation seems to be a special case of Tune 2, and its use
here seems analogous to the use of Tune 2 in such cases as:

image
juː ˈdidnt?
You didn't?
(Meaning ‘Do you mean to say that you
didn't?’)

image
ə ˈsəkənd taim?
A second time?
(Meaning ‘Has it really occurred a second
time?’)

image
ðei ˈhævnt ˈθɔːt əbaut it?
They haven't thought about it?
(Meaning ‘Do you mean to
say that they haven't thought
about it?’)

image
wɔt əm ai ˈduiŋ?
What am I doing?
(Meaning ‘Are you asking what I'm
doing?’)

1064. The expressions I do, it's not, he can, they have, etc., used
in replying to a question requiring the answer ‘yes’ or ‘no’ are
said with Tune 1. But the same expressions are said with Tune
2 when they are used to contradict what the previous speaker has
said. Compare:

image | image
djuː ˈlaik it? | ˈjes, ai ˈduː.
Do you like it? | Yes, I do.

image | image
juː ˈdount ˈlaik ˈðæt? | ˈjes, ai ˈduː.
You don't like that? | Yes, I do.
312

image | image
iz it ˈgriːn? | ˈnou, it s ˈnot.
Is it green? | No, it's not.

image | image
it s ˈgriːn. | ˈnou, it s not.
It's green. | No, it's not.

1065. In asking a question containing a specific interrogative
word, the effect of great curiosity on the part of the speaker is
conveyed by saying the interrogative word or the first stressed
syllable after it on a very low tone, as shown in the following
examples. Preceding unstressed syllables are high-pitched.

image
bət ˈhau djuː ˈmæniʤ it? 25475
But how do you manage it?

image
hau ˈduː juː ˈmæniʤ it?.
How do you manage it?.
313

image
hau ɔn ˈəːθ djuː ˈmæniʤ it? 26476
How on earth do you manage it?

image
wɔt ˈɑː juː ˈduiŋ?
What are you doing?

image
wɔtˈevər ə juː ˈduiŋ?
Whatever are you doing?

1066. In exclamations of astonishment a high degree of surprise
is expressed by pronouncing the non-final unstressed syllables with
high pitch and the stressed syllables with low pitch, as in

image | image
hau ˈhai it ˈluks. | wɔt n iksˈtrɔːdnri ˈθiŋ.
How high it looks! | What an extraordinary thing!

image
wɔt ə ˈvəri ˈfʌni ˈθiŋ.
What a very funny thing!

This intonation is an exaggerated form of the second variant of
Tune 1 mentioned in § 1025 (b).

1067. Thank you is sometimes pronounced with a rising intonation
(Tune 2) and sometimes with a falling intonation (Tune 1).
314When a person performs a customary service, the acknowledgement
seems to be said more usually with the rising intonation, thus:

image
ˈθæŋk ju.

But in acknowledging an unexpected favour the falling intonation
seems more usual, thus:

image or thus image
ˈθæŋk ju. | ou ˈθæŋk ju.
| Oh thank you.

1068. Thank you with rising intonation is often reduced to
ˌŋkju or kju, 27477 thus:

image or image
ˈŋkju | kju

Thank you with a falling intonation is not generally reduced in
this way.

1069. All right is generally said with Tune 2, thus: image ˈɔːl ˈrait

The use of Tune 1 image or (with emphasis) image
may have the effect of a threat.

1070. The usual intonation of Good morning as said on parting is
image gʌd ˈmɔːniŋ.
315It is not quite clear why this intonation is used, but it appears
to imply non-finality or some such continuation as ‘I shall hope
to see you again soon.’ 28478 (Cp. § 1072, last example.)

Parentheses

1071. Expressions of a parenthetical nature have no particular
intonation of their own. They share the intonation that the main
sentence would have if the parenthesis were not there. Thus a
parenthesis occurring at the end of a sequence requiring Tune 1
is said on a low level pitch, this being a continuation of the low
pitch to which the last stressed syllable falls. A parenthesis
occurring at the end of a sequence requiring Tune 2 shares in
the rise with which the tune terminates. Examples:

image | image
hiːz ˈgɔn ˈhoum, ai ˌθiŋk. | it s laik ˈðis, juː ˌsiː.
He's gone home, I think. | It's like this, you see.

image
ai ˈkɑːnt ˈhelp it, hiː ˌsed imˌpeiʃntli.
‘I can't help it,’ he said impatiently.

image
ˈhau djuː ˈlaik ðis ˈbʌk witʃ ai ˌbɔːt ði ʌðə ˌdei?
How do you like this book which I bought the other day?

image
gud ˈmɔːniŋ, miste ˌbraun.
Good morning, Mr. Brown.
316

image
ˈɑː juː ˈgouiŋ əˈwei, hiː ˌɑːskt.
‘Are you going away?’ he asked.

image
gudˈbai, oul ˌtʃæp.
Good-bye,old chap.
(Same intonation as ‘Good-bye.’)

For further information about the intonation of parentheses see
Armstrong-Ward, Handbook, pp. 27-30.

Interjections

1072. Interjections and exclamatory phrases take as a rule the
intonation of the complete sentences to which they are equivalent.

Examples:

image
gud.
Good!
(Meaning ‘I'm glad of it.’)

image or image
ou. | ou.
Oh! | oh!
(Meaning ‘That is a surprising
or piece of news.’)

image
ou?
Oh!
(Meaning ‘Do you really mean it?’)

image or image
riəli? | ˈriəli?
Really? | Really!
(Meaning ‘Do you really mean
it?’)
317

image
wel.
Well§
(Meaning ‘Yes; and what happened next?’)

image
nou.
No.
(Meaning ‘it isn't,’ ‘I haven't,’ etc.)

image
jes.
Yes.
(Meaning ‘It is so,’ ‘I agree,’ ‘I will,’ etc.)

image
jes.
Yes.
(Meaning ‘It may be so, but I can't be certain.’)

image | image or image
igˈzæktli sou. | ˈæbseluːtli. | ˈæbsəˈluːtli.
Exactly so! | Absolutely!

image
wɔt n iksˈtrɔːdnri ˈθiŋ.
What an extraordinary thing!
(Meaning ‘It's a very extraordinary
thing.’ Alternative intonation in
§ 1066.)

image
wɔt n aiˈdie.
What an idea!
(Meaning ‘That idea is extraordinary.’)

image
əz ˈif wiː ˈʃud.
As if we should!
(Meaning ‘The suggestion that we should is
absurd.’)
318

image
ˈwɔt ˈnəkst.
What next!
(Meaning ‘I wonder what impudent thing he'll
be doing next.’)

image
gud ˈmɔːniŋ.
Good morning.
(Said on meeting. 29479 Meaning ‘I greet you.’
Cp. § 1070.)

Incorrect Forms of Intonation heard from Foreign
Learners

1073. The mistakes of intonation made by foreign people when
they speak English are very varied. The following are a few
examples.

1074. French people often employ an intonation of the type
image or image when an intonation of the
type image should be used. The following are examples:

tableau Correct pronunciation | Incorrect intonation often heard from French people | Absolutely | I've got two tennis balls319

tableau Correct pronunciation | Incorrect intonation often heard from French people | I like it | What are you looking at? | I've never been there

1075. French people are likewise liable to use an intonation of
the type image in eases where one of the type
image is required. The following are examples:

tableau Correct pronunciation | Incorrect intonation often heard from French people | Don't you think so? | Shall we go and, look at it?320

tableau Correct pronunciation | Incorrect intonation often heard from French people | One moment!

1076. The above incorrect forms of intonation used by French
people give the effect of emphasis to the final unstressed syllables.

1077. Germans are liable to make mistakes of an opposite
nature, that is to say they have a tendency to use intonations
of the types image and image where the
intonations image and image are required.
These mistakes are commonly attributed to incorrect stress;
it will be found, however, that as long as the intonation is right,
the degree of stress is not of much consequence.

Example of the first case:

Correct pronunciation
image
wie ˈgouiŋ fər ə ˈwɔːk in ˈritʃmənd ˈpɑːk.
We're going for a walk in Richmond Park.

Incorrect intonation commonly heard from Germans
image

Example of the second case:

Correct pronunciation
image
ˈʃæl wi ˈgou tə ˈritʃmənd ˈpɑːk?
Shall we go to Richmond Park?
321

Incorrect intonation commonly heard from Germans
image

1078. Very often in sequences requiring Tune 2 Grermans say the
last stressed syllable with high pitch and all following unstressed
syllables on the same high pitch. Thus in such an example as:

image
ai ˈkudnt ˈfiniʃ it, bikɔz dee ˈwɔznt ˈtaim.
I couldn't finish it, because there wasn't time.

they will pronounce

image
ai ˈkundt ˈfiniʃ it, …

1079. Opportunities for these characteristic German mistakes
occur very frequently in long descriptive passages. The following
taken at random from my Phonetic Readings in English 30480 will
serve to illustrate what happens: and the sergeant major was heard
to say that it kept better time than the station gun
. The intonation
should be

image
ənd ðə ˈsɑːʤent ˈmeiʤe | wəz ˈhəːd tə ˈsei |

image
ðət it kept ˈbəte ˈtaim | den ðə ˈsteiʃn ˈgʌn.
322

Germans are liable to mispronounce the sentence by using the
following incorrect intonations:

(1) either

image or image
ənd ðə ˈsɑːʤent meiʤe | ənd ðə ˈsɑːʤent meiʤe

(2) either

image or image
wəz ˈhəːd tə ˈsei | wəz ˈhəːd tə ˈsei

(3) either

image or image
ˈbəte taim | ˈbəte taim

(4)

image
ˈsteiʃn gʌn.

1080. Many Germans also have considerable difficulty in pronouncing
stressed syllables on a high level tone as is required in
Tune 1. They are apt to say all such syllables with a low rising
pitch. Thus in pronouncing he was about the only intelligent man
in the country
they will use an intonation of the following type
instead of that shown in § 1021:

image
ˈhiː wəz əbaut ði ˈounli inˈteliʤent ˈmæn in ðə ˈkʌntri.
323

1081. Norwegians find Tune 1 difficult. They substitute very
high pitch for the low pitch of the final unstressed syllables. Thus
they will say

image instead of image
it s ˈbəte | it s ˈbəte
It's better.

image instead of image
it wəz ˈkwait imˈpɔsəbl | it wəz ˈkwait imˈpɔsəbl
It was quite impossible.

1082. Most foreign learners have great difficulty in learning to
make a fall-rise on a single syllable, as in the examples in § 1054.
The correct pronunciation may be acquired by practising very
slowly and then gradually increasing the speed, being careful to
observe the rules mentioned in § 1055. Thus the can in the
example I will if I can given in § 1054 should be practised thus:
image kæ-n-n then image kæ-n-n then image kæ-n then
image kæn and the bad in the example it isn't bad (§ 1054) thus:
image b-æ-d then image b-æ-d then image b-æːd and so on.

Methods of Recording Intonation

1083. There are various methods of recording intonation.

1084. A notation of dots and lines on a stave, such as that
used in this chapter, may be drawn free-hand by anyone with a
good musical ear. This method is sufficiently accurate for practical
324linguistic purposes. The method has the advantage that it records
intonations which are subjectively present, even if they are not
clearly audible objectively owing to the presence of voiceless sounds
or the nature of the pitches of adjacent syllables.

1085. A more accurate method of obtaining intonation-curves
is the following. If while a gramophone record is being played
the needle is lifted from the revolving disc, the ear retains
the impression of the sound heard at the instant when the
needle is lifted. If the record is of the speaking voice and the
needle is removed in the middle of a voiced sound, the ear retains
in particular the pitch of the musical note which the voice is
producing at that instant; this may be marked on some kind of
musical stave. By taking observations at a large number of
points in a sentence and joining the points by lines, a complete
intonation-curve of the sentence results. In order to ensure
accuracy it is of course necessary to take a number of observations
at every chosen point; the chosen points should likewise not be
too far apart: thus it is necessary to record the pitch of every
vowel and a considerable number of the voiced consonants, and
where sounds are long or where the intonation is rising or falling
rapidly it may be necessary to record the pitch of two or three
portions of one sound. This method was the one followed in
preparing my book of Intonation Curves. 31481

1086. Certain small inaccuracies are unavoidable with this
method, but the method has the advantage that while a considerable
degree of scientific accuracy is attained yet the resulting curves
are such as can be made use of without difficulty in practical
language teaching. The phonetic text is continuous (not irregularly
spaced as in the case of the most accurate curves), and the ordinary
musical stave being used, the values of the curves are clearly
apparent to anyone who has an elementary knowledge of music.

1087. The most accurate methods of obtaining intonation-curves
are obtained by measuring the lengths of vibrations on kymographic
tracings or enlargements of the lines on gramophone records or
records made in other ways. (For details and an example, see
325pp. 179-182 of the first and second editions of this book, 1918,
1922.)

1088. Accurate curves obtained by such means have scientific
value, but their use in practical language teaching is limited, since
they only record what is objectively present. To get good results
in practical teaching it is necessary to have regard continually
to the intonations aimed at, i.e. the intonations which are
subjectively present to the speaker. These often differ considerably
from the objective intonations actually employed. The latter are
such approximations to the subjective intonations as are compatible
with the length and nature of the sounds in each particular case.
(The differences between subjective and objective intonations are
especially notable when vowels are very short and voiceless consonants
are present.) The graphical representations in the examples
in this chapter have been drawn by ear, and they represent the
subjective intonations to be aimed at by the learner.326

Chapter XXXII
Syllable Separation

1089. It was pointed out in § 212 that it is often impossible to
specify points at which a syllable begins and ends. There do exist,
however, circumstances where points of syllable separation are
well marked in pronunciation, and must be shown in transcriptions
in order to render them unambiguous. This happens in some
instances of compound words, and in some apparently simple words
where two parts, though joined together without pause, are nevertheless
pronounced as if they were separate words.

1090. Where such circumstances are present, the points of
syllable separation are sometimes made evident in transcriptions by
the positions of stress-marks. 1482 More often, however, they have to be
shown by a special mark. Hyphens are convenient for this purpose.

1091. When two parts of a word are pronounced as if they
were separate words, the various rules relating to sound-quality
and length in single words apply to each part; the sounds on each
side of the place of separation do not affect each other in the ways
they would if there were no clear syllable separation. For instance,
when the first syllable ends in a long vowel and the second one
begins with a breathed consonant, the long vowel does not have
a shorter length as described in § 866. And when the first syllable,
ends in a breathed consonant and the second begins with l, r, etc.,
the fully voiced allophones of these phonemes are used, and not
the devoiced sounds which would represent the phonemes if the
same breathed consonant were to precede them at the beginning
of a word (§ 845 (i) a).

1092. These principles are well illustrated by compound words
containing sequences of vowel + s + t + r + vowel, such as toe-strap
ˈtou-stræp, mouse-trap ˈmaus-træp, toast-rack ˈtoust-ræk. In
ˈtou-stræp the syllable separation is between the ou and the s. This
means that, although there is no cessation of sound at the place of
separation, the two syllables are pronounced as if they were separate
327words, and consequently the ou is made fully long as explained
in § 866. In ˈmaus-træp, however, the separation being between
the s and the t, the au is a rather short diphthong on account of
the presence of the s. 2483 Moreover, in ˈtou-stræp and ˈmaus-træp
the tr's are pronounced as if stræp and træp were said in isolation,
which means that they have the sound of the voiceless affricate
described in § 624. 3484 In ˈtoust-ræk not only is the ou rather short
in accordance with the principle stated in § 866, but the r being
pronounced as if it were initial is fully voiced.

1093. It would be ambiguous to transcribe the above words without
the hyphens. If a foreign learner were to say them with syllable
separations elsewhere than at the places shown by the hyphens,
his pronunciation would be incorrect and possibly unintelligible,
since there are no such words as ˈtous-træp, ˈtoust-ræp, ˈmau-stræp,
ˈmaust-ræp, ˈtou-stræk and ˈtous-træk.

1094. The following are some further examples illustrating the
effect of syllable separation.

(1) Syllable separation between a vowel and a consonant: biplane
ˈbai-plein, 4485 eye-sight ˈai-sait, 5486 awe-struck ˈɔː-strʌk, 6487 how-string
ˈbou-striŋ, 7488 door-plate ˈdɔː-pleit, 8489 key-stone ˈkiː-stoun. 9490

(2) Syllable separation after the first of two or three consonants:
horse-truck ˈhɔːs-trʌk, 10491 Lakeland ˈleik-lænd, 11492 heat-wave
328ˈhiːt-weiv, 12493 leap-year ˈliːp-jəː, 13494 outrageous autˈreiʤəs, 14495
boat-race ˈbout-reis, 15496 Pecksniff ˈpek-snif, 16497 undomesticated
ˈʌn-dəˈmestikeitid. 17498

(3) Syllable separation after the second of three consonants:
mincemeat ˈmins-miːt, 18499 bank-rate ˈbæŋk-reit, 19500 lamp-light
ˈlæmp-lait. 20501

(4) Syllable separation between a consonant and a vowel: lynx-eyed
ˈliŋks-aid, 21502 cat's-eye ˈkæts-ai, 21503 stomach-ache ˈstʌmək-eik, 22504
hair-oil ˈhɛər-ɔil, 23505 under-masticated ˈʌndəˈmæstikeitid. 24506

1094. In transcriptions in which words are joined together,
hyphens have to be inserted wherever the absence of a hyphen
329would render the notation ambiguous in the matter of syllable
separation. The following are a few examples:

ðəˈʃipˈsæŋk-wiðˈɔːlˈhænʤ (the ship sank with all hands), 25507
hiːˈsæŋ-kwaitˈwel (he sang quite well), 26508
itsəzˈwel-təˈweit (it's as well to wait), 27509
hiːzəˈweltə-weit (he's a welter-weight), 28510
ˈliː-pleidəˈdiːp-leidˈgeim (Lee played a deep-laid game). 29511

1095. Similarly, when it is desired to transcribe without spaces
between words, the positions of stress-marks sometimes have to be
chosen so as to show syllable separations. Compare for instance
əˈblæktˈai (a blacked eye) 30512 with əˈblækˈtai (a black tie), 30513 and itˈslips
(it slips) 31514 with itsˈlips (its lips). 32515

1096. Further illustrations of phonetic phenomena associated
with syllable separation will be found in §§ 892-898 and in my articles
The ‘Word’ as a Phonetic Entity in Le Maître Phonétique, October,
1931, and The Hyphen as a Phonetic Sign in the Zeitschrift für
Phonetik
, Vol. IX, No. 2 (Berlin, 1956).

1097. The existence of many special shades of sound and degrees
of length near word junctions, as illustrated in this chapter, shows
the necessity for defining the ‘phoneme’ and the ‘chroneme’ by
reference to ‘words’ and not to longer units of connected speech.
See The Phoneme, § 34.330

11 Also called ‘phones’ or ‘linear’ or ‘segmental’ features of speech.

22 Also called ‘prosodies’ or ‘suprasegmental’ features of speech.

33 In the technical term ‘independent vowel-glide’ the word ‘glide’ is used
in a different sense. It there denotes a particular kind of speech-sound
(see §§ 219, 800).

44 But some glides occurring in foreign languages are distinctly audible
and require special mention in descriptions of pronunciation. For instance,
the glide between the French ɲ (§ 655) and a following vowel is always
clearly audible.

55 See also my book, The Pronunciation of English, 1950 and subsequent
editions, § 490 (Cambridge University Press). When a teacher is not available,
ear-training may be done with a gramophone, provided the records are good
and there is a key giving a phonetic transcription of the words of the record.

66 The ‘ear-training exercise’ was, I believe, first invented by Jean Passy,
the brother of Paul Passy. See his article La dictée, phonétique in Le Maître
Phonétique
, Feb., 1894 (particularly pp. 36, 37).

A dictation of meaningless words for testing the sharpness of candidates'
ears forma part of various examinations in Phonetics in the University
of London, and in the examinations held by the Association Phonetique
Internationale.

77 Letters in thick type are phonetic symbols. The various sounds denoted
by them are fully described further on (Chap. XIV, etc.), and lists with keywords
are given on pp. xvi-xx.

Every teacher should adopt a definite method of naming the symbols and
sounds. As to the symbols some teachers simply call them by the sounds
they represent, e.g. if they want to mention the phonetic letters p, l, ʃ, ɔ,
they call them by the isolated sounds p, l, ʃ, ɔ. There are some objections
to this system. One is that some sounds (e.g. plosive consonants) cannot be
said without another sound to accompany them (§ 563). Another is that
isolated sounds are often indistinct, especially when uttered in a large room,
and a third is that learners not yet fully familiar with foreign sounds do not
always recognize which sound is meant. On the whole I am inclined to
recommend giving names to the consonant letters—the ordinary names
for the letters of the ordinary alphabet and the following special names
for new letters:

letter | name
ŋ |
θ |
ð | ðiː or ðɑː
ʃ |
ʒ | ʒiː or ʒɑː

The naming of vowel-symbols presents some difficulty. Probably the best
plan is to say the sound with a defining adjective, and speak of ‘the close
e letter,’ ‘the open ɛ letter,’ ‘the back ɑ letter,’ ‘the neutral ə letter,’ etc.
Some teachers, however, use non-technical language and speak of ‘Greek e
(for ε), ‘broken o’ (for ɔ), ‘round a’ (for ɑ), ‘inverted e’ (for ə). Another
plan is to use key-words and speak of ‘the cup-symbol’ (for ʌ), ‘the lock-symbol’
(for ɔ), etc.

Sounds may often be named by simply uttering them. But here again
confusion may arise in practical teaching through the indistinctness of isolated
sounds (especially in a large room) or failure on the part of the pupil to
recognize which sound is meant. Greater clearness is ensured by speaking
of ‘the kei-sound’ (for k), ‘the el-sound’ (for l), ‘the -sound’ (for θ), etc.
In the case of vowels the mention of a key wTord is often helpful: ‘the cup-sound’
(for ʌ), ‘the bird-sound’ (for əː), etc. Another plan which gives
excellent results in practice is to number the vowels, and designate them
when necessary by their numbers; thus ʌ may be called ‘English No. 10,’
a may be called ‘Cardinal No. 4,’ etc. (see §§ 235, 236).

88 These words are phonetically geit, faind, tjuːn.

99 Phonetically ˈfɑːðə, fɔːl, ˈeni, fæt, wɔʧ.

1010 Phonetically wind, məˈʃiːn, baːd.

1111 Phonetically ruːl, put, hʌt.

1212 Phonetically stouv, muːv, lʌv.

1313 Phonetically miːt, hed, greit, bɛə.

1414 Phonetically miːt, miːt, niːs, piːk, kiː, kiː, siːz.

1515 Phonetically sɔːs, lɔːn, stɔːk, stɔːk, bɔːd (also bɔəd), wɔːn, θɔːt, brɔːd,
flɔː (also flɔə).

1616 Phonetically vil, fiːj.

1717 Phonetically groːs, gɔs.

1818 Phonetically pɔrˈsjɔ̃, pɔrˈtjɔ̃.

1919 Phonetically mo, to, bo, ʃəˈvo.

2020 Phonetically ˈrɑuxən, ˈfrɑuçən.

2121 Phonetically fuːs, nus.

2222 Particulars of the Association Phonétique Internationale (in English
International Phonetic Association, often abbreviated to I.P.A.; in German
Weltlautschriftverein) are obtainable from the Secretary of the Association,
Department of Phonetics, University College, London, W.C.1.

2323 I.e. the ‘prosodies’ or ‘suprasegmental’ elements of speech.

2424 See 100 English Substitution Tables, by H. E. Palmer (published by
Heffer, Cambridge).

251 The use of short ɔ in words like soft, cross, lost is much on the increase.
It would not surprise me if the pronunciation sɔːft (which I use) were to
disappear within the next fifty years.

262 ‘Public School’ in the English sense, not in the American sense.

273 Published by Dent (London).

284 Published by Blackwell (Oxford).

295 Although those who use RP have much in common in their speech, it
must not be thought that RP is absolutely uniform. Quite a number of
variations are to be found in it. For instance, the qualities of sounds used in
some words vary from speaker to speaker (see, for example, §§ 271, 330, 388).
And in the case of some words two distinct pronunciations must both be
considered as belonging to RP; examples of such words are been (§ 471), off,
cost, loss, etc. (§ 300), the words ending in -aph (§ 294), threepence (§ 257),
association (footnote 28 to § 732).

The following are some variations within RP which require special mention,
since their occurrence has determined certain features of the system of
phonetic transcription used in this book: (1) many of those who speak with
RP lengthen the traditionally short sound of a (æ) in various words (§ 874),
(2) the sequences eiə, ouə, are often reduced to diphthongs , , distinct
from ɛə, ɔə (§§ 392a, 403), (3) a diphthong of the -type or a pure vowel
of the -type distinct from ɑː is often used in place of aiə and auə (§§ 414,
430), (4) a monophthongal o-sound often takes the place of ou in unstressed
positions (§ 403).

Although the form of transcription here adopted allows for these four
variations, it must be understood that none of these variants are essential
to acceptable pronunciation. Consequently, if greater simplicity is desired, a
simpler form of transcription may easily be devised by taking as a basis
for study a type of English pronunciation in which the above-mentioned
alternatives do not occur. Such a simplified transcription is described in
Appendix A, 44-49; it is perfectly adequate for enabling foreign students
to learn to pronounce English like English people. In a ‘broad’ transcription
(§ 200) of this particular kind of speech the special letters æ, ɛ, ɑ, and ɔ are
not needed; they may be replaced without ambiguity by the common letters
a, e, a, and o respectively, in accordance with the principle of substituting
familiar for unfamiliar letters formulated in § 20 of the Principles of the International
Phonetic Association
, 1949. Transcription simplified on these lines, in
which the number of special letters is reduced to a minimum, is very convenient
for use by the numerous foreign learners whose sole object is to learn to speak
English well, but who have no need either to become specialized phoneticians
or to concern themselves with more than one variety of English pronunciation.

For further information concerning different types of phonetic transcription
see Appendix A.

301 A number of kymographic tracings were reproduced in the first and
second editions of this book (1918 and 1922).

311 It is convenient to use the term ‘breathed’ in speaking of continuant
sounds and ‘voiceless’ in speaking of plosive consonants. It can hardly
be said that during the 'stop' of a plosive consonant there is a current of
air passing between the vocal cords.

322 ʃ is the English sound of sh; ʒ is the sound of s in measure; θ and ð are
the sounds of th in thin and then.

333 As in half, father.

344 These voice indicators respond excellently to voiced consonants and
close vowels, but do not always respond well to the opener vowels, especially
the opener front vowels such as ɛ, a.

351 Whispered speech is not considered as normal. In whispered speech
‘voice’ is replaced throughout by ‘whisper’ and every sound consists of
audible friction and nothing else (except the ‘stops’ of voiceless plosives,
which have no sound at all). The term ‘whispered vowels’ is commonly
used to designate sounds produced with the organs in the same positions
as for the sounds defined as ‘vowels’ in § 97, but with ‘whisper’ substituted
for ‘voice.’ There is no objection to this terminology, but it should be noted
that if a whispered vowel were to occur in speech next to a voiced one, the
whispered vowel would have to be regarded as a consonant. This may be
seen by pronouncing a whispered ɑ immediately followed by a voiced ɑ.
The result resembles with a strong kind of h.

36 Voir note 35.

372 With the exception apparently of ‘cardinal’ i (see the article on The
Perceptibility of Sounds
, by Stephen Jones, in Le Maître Phonétique, January,
1926).

383 The line of demarcation between vowels and consonants might have been
drawn elsewhere. Thus since speech-sounds which consist wholly or in part
of ‘noise’ (as distinguished from ‘musical sound’) are less sonorous than those
which contain no perceptible ‘noise,’ a logical classification into vowels and
consonants might be based on the presence or absence of perceptible ‘noise.’
If this classification were adopted, the voiced sounds M, n, etc., and the
voiced 1-sounds would have to be classed as vowels, because in normal
pronunciation they are not accompanied by any perceptible ‘noise’ or ‘audible
friction.’ This method of classification would, however, be less convenient
in practice than that given in § 97.

391 We assume in this simple case that the lip-positions required for forming
the two known sounds and the foreign sound are identical.

401 The English sound of a in about əˈbaut is a characteristic variety of
neutral vowel.

412 If the distance between the teeth is much greater than 1 cm., some
vowels (e.g. English and əː) cannot be pronounced quite correctly. And
if the distance is much less than 1 cm., other vowels (e.g. English ʌ and ɔː)
cannot be pronounced quite correctly.

423 Published by Elwert, Marburg a. L., Germany, 1910.

434 If the tongue were raised higher, the breath-pressure remaining constant,
the result would be a fricative j (§ 818).

445 If the tongue were retracted further, the breath-pressure remaining
constant, the result would be a variety of ʁ (§ 763).

456 These photographs were of my mouth. They were taken by Dr. H.
Trevelyan George of St. Bartholomew's Hospital by Dr. E. A. Meyer's method
of placing a very thin metal chain on the tongue. They were first published
in the Proceedings of the Royal Institution, Vol. XXII, Part 1, Oct., 1919.
The originals may be seen in the Department of Phonetics, University College,
London, W.C.1. Reproductions of them on a small scale are given in the
frontispiece of this book.

467 E.g. on double-sided record No. ENG 252-3, published by the Linguaphone
Institute, 207 Regent Street, London, W.1. There exists also a record of
these vowels made in 1917 by the H.M.V. Gramophone Co., 363 Oxford
Street, London, W.1, and numbered B 804. Although this record was made
before the invention of electric recording, the reproduction is very good.
It is no longer on sale through the ordinary channels, but the H.M.V. Gramophone
Co. have preserved the matrix, and are willing to print off copies
specially for anyone who orders a sufficient number.

Tape recordings of these cardinal vowels said on various pitches by twelve
reliable phoneticians have recently been made by Mr. P. Ladefoged, Lecturer
in Phonetics in the University of Edinburgh, as part of a research programme.
Copies of these recordings can be made available for linguistic research
institutions. (Particulars on application to the Secretary, Department of
Phonetics Minto House, Chambers Street, Edinburgh.)

478 In broad transcriptions of particular languages it is generally convenient
to use the symbol ɔ in place of ɒ.

489 Though this figure combines accuracy
with definiteness, some teachers find that its
unsymmetrical shape renders it difficult for
ordinary pupils to draw. When such a difficulty
is experienced, it is well to adopt the
modified shape shown in Fig. 23a. In this
simplified form the lines a—ɑ, ɛ—ɔ, e—o,
i—u are parallel, the angles at ɑ and u are
right angles, and the lines a—ɑ, ɑ—u,
i—u are in the proportion 2:3:4.

image

Fig. 23a. Simplified
Form of Fig. 23.

The symmetrical form shown in
Fig. 23b has been widely used. It
appears to be preferable, however, to
adopt a form which shows that the
distance i—a is longer than the distance
ɑ—u and which does not suggest
that u has a more retracted tongueposition
than ɑ.

image

Fig. 23b. Symmetrical
Arrangement of Vowels

4910 See, for instance, the remarks on æ, p. 72, footnote 16.

501 E.g., the series of so-called ‘t-sounds’ ranging from dental t through
alveolar t to the retroflex ʈ (Chap. XXV), or the series of plosive consonants
ranging from C (Fig. 29) through k (Figs. 30, 31) to the uvular plosive q,
or the series of fricatives comprising sounds of the types θ, s, ʃ, ç.

512 This sound exists in French in such words as peuple when final (pœpl̥);
a variety of it is the sound of Welsh ll, as in Llangollen (l̥anˈgɔl̥en. When
the sound occurs in a language as a separate phoneme, it is better to represent
it by the special letter ɬ as in S. Jones' Welsh Phonetic Reader (University
of London Press) and C. M. Doke's The Phonetics of Zulu (University of the
Witwatersrand Press, Johannesburg).

528 §§ 508, 528, 549, 566, 567.

534 In general phonetics at least two other classes of tongue consonants
have to be distinguished. They are retroflex consonants (articulated by the
tip of the tongue against the hard palate) and uvular consonants (articulated
by the ‘back’ of the tongue against the extremity of the soft palate).

545 The palatal and velar frictionless continuants have the organic positions
of close vowels. They are, however, uttered with very little breath-force as
compared with the normally pronounced vowels which adjoin them in connected
speech. These frictionless continuants are to be considered as
consonants on account of their consequent lack of prominence as compared
with the adjoining vowels.

551 Published by Heffer, Cambridge.

562 According to one theory, first propounded by J. L. M. Trim in Le Maître
Phonétique
, July, 1951, p. 41, the German x-sounds should be held to belong
to the h-phoneme, ç constituting a separate phoneme. (Formerly it was
thought that the German ç and x-sounds were assignable to a single phoneme.)
I believe Trim's view to be the correct one for most German speakers.

573 More usually pronounced ˈkɔrdə.

584 The form of transcription of English adopted in this book is a very
nearly ‘broad’ one allowing for the speech of those who use the variant
pronunciations mentioned in footnote 5 on p. 13. It is ‘allophonic’ in one
particular only, namely in that it provides a letter to denote the monophthongal
o-sound that sometimes takes the place of the diphthong ou in
weak positions. It is ‘comparative’ in some respects. See Appendix A.

595 See The Principles of the International Phonetic Association, 1949, §§ 17-22.

606 Of the kind of English in which these vowels are not lengthened (§§ 874-878
and Appendix A).

611 See Chap. XX.

62 Voir note 61.

632 This is often done in accordance with a tradition, the origin of which is
obscure.

641 Thus as the ‘stop’ of t has no sonority at all, it ia impossible to say at
which part of the t the syllable separation of the word ˈletə takes place.

652 The word tray trei consists of a single syllable. The word stray strei
is conventionally considered also to form a single syllable in spite of the fact
that s has some sonority while the stop of the t has none. The s is rather short,
and its prominence is ignored in conventional syllable separation.

663 Many foreign people mispronounce these words by inserting a vowel-sound,
generally ə or e, thus ˈpiːpəl, ˈlitel, etc. See § 590.

674 Alternative pronunciation of ˈglʌtəni.

685 As in This meat has a muttony taste.

696 As in This material has a flannelly feeling.

707 Also pronounced ˈkeɔs, in which case the syllable separation is marked
by the glide from e to ɔ.

718 Many foreign people, and especially Germans, mark the syllable separation
in such words as kriˈeit, kouˈɔpəreit by inserting the sound ʔ (§ 553), thus
kriˈʔeit, kouˈʔɔpəreit (see §§ 557-559).

729 Very often the effect of syllable separation in such cases is produced by
a sudden change of pitch (intonation). It must always bo remembered that
where there is a sudden change of pitch, it is extremely difficult to ascertain,
even with the aid of apparatus, whether there is any simultaneous variation
in force.

7310 It is also theoretically possible to make diphthongs in which the
prominence remains constant.

7411 It is, however, not uncommon to meet with Southern English, speakers
who in many words do not give sufficient force to the i to make it predominate
over the latter part of the diphthong. The diphthong is then a rising one
of the type la. Some people use the sequence jai in these words (see § 442a).

751 To make the description complete it is necessary to add (v) position
of the soft palate
: raised: (vi) action of vocal cords: vibrating, producing voice.
This addition is to be undorstood in the case of all the subsequent descriptions
of vowels.

762 The vowel as pronounced by foreign people is often wrong in quantity
(length); see §§ 901 ff.

773 į denotes a very open i, similar to the English short i. denotes a
very close variety of i.

784 ɩ is a symbol used in narrow transcription for the English short i.

795 The symbol j is used here in a sense somewhat different from that assigned
to it in § 813. The two values are, however, closely related. The similarity
between them lies in the fact that the tongue-position reached at the end
of the diphthong written ij, is about the same as the tongue-position assumed
at the beginning of the sequence ji. Those who would prefer not to use the
same symbol j in these two different senses, are recommended to use in
the diphthong, thus , the mark ˘ indicating that the sound is to be regarded
as a consonantal vowel (§ 231) and not as a semi-vowel (§ 183).

806 Note the difference between explain iksˈplein and explanation
ekspləˈneiʃn, exhibit igˈzibit and exhibition eksiˈbiʃn, etc. The prefix
is quite unstressed in explain, exhibit, but it has secondary stress in explanation,
exhibition.

817 Unstressed -ate is pronounced -it in most nouns and adjectives. In
verbs on the other hand the termination is pronounced -eit. Thus the
nouns estimate, associate and the adjectives appropriate, intimate, separate
are pronounced ˈestimit, əˈsouʃiit, əˈproupriit, ˈintimit, ˈseprit, while the
similarly spelt verbs are pronounced ˈestimeit, əˈsouʃieit, əˈprouprieit,
ˈintimeit, ˈsepəreit. Intermediate and immediate are exceptional words in
which the vowel of the termination is usually ə (intəˈmiːdjət, iˈmiːdjət).
The -it is often changed to -ət in derived adverbs; thus though the adjective
deliberate is normally diˈlibərit, yet the adverb deliberately is pronounced
diˈlibrətli by many.

828 Foreign people often use the long in the terminations -ies, -ied.

839 Also pronounced ˈsʌndei, ˈmʌndei, etc., especially by younger people.

849a There is a modern tendency in England to substitute ə for i in some
of the prefixes and suffixes, e.g. to pronounce bəˈkʌm, rəˈmein, ˈgudnəs.
This is not as a rule done with -es and -ed, presumably because it is felt to
be desirable to maintain the distinction between -iz and -əz, e.g. in offices
ˈɔfisiz and officers ˈɔfisəz, charted ˈʧɑːtid and chartered ˈʧɑːtəd. See the
special section on Variant Pronunciations of -less, -ness, etc., in the 11th
(1956) edition of my English Pronouncing Dictionary (Explanations XXI).

8510 These words may also be transcribed ˈhɛve, ˈsite, ˈmɛne, when they stand
in final position.

8611 In many forms of dialectal English the final vowel of heavy, city, etc.,
is closer than the common short i; often too it is lengthened or replaced by the
dialectal diphthong əi (§ 251).

8712 With some speakers the sound is sometimes long (see § 876).

8813 These are the only words in which the sound e is represented in spelling
by the letter a. Xote that The Mall is pronounced ðə ˈmæl. Some English
people pronounce Pall Mall as ˈpælˈmæl.

8914 When words ending in -el are immediately followed by a word beginning
with a vowel, they are said with ‘clear’ l (§ 668) and the vowel is then the
principal e and not the opener variety. This would be the case, for instance,
in tell it ˈtel it.

90 Voir note 89.

9115 This seems for practical purposes the most satisfactory way of regarding
the tongue-position of this vowel. It must be admitted, however, that the
exact analysis of the manner of forming this sound presents some difficulties.
Some authorities regard ɛ as a tense vowel and æ as the corresponding lax
vowel. In passing from ɛ to æ there is (at any rate in my pronunciation) a
distinct raising of the sides of the tongue; this can be felt, or it can be seen in
a looking-glass; it is also indicated by the fact that æ gives a palatogram while
cardinal ɛ does not, though the middle of the tongue seems to be lower for æ
than for ɛ. I am also conscious of a contraction in the pharyngal region in the
production of æ. Other observers have also remarked this. This pharyngal
contraction is too vague to define precisely, though it appears to be an
inherent characteristic of the sound. I have often been able to improve
foreign students' pronunciation of æ by telling them to tighten the throat.
(The existence of the contraction in the throat is no doubt the reason why
the sound æ cannot be pronounced with good voice-production. Singers
commonly use a modified æ, or substitute a for it.)

9216 The vowel is in reality often lengthened; see § 874.

9317 The weak forms of this word are həv, əv, and v.

9418 This word has no weak form. The conjunction that has a weak form
ðət; the relative pronoun that (meaning ‘which’ or ‘whom’) is always pronounced
ðət in conversational speech.

9519 This word also has weak forms ʃəl, ʃl.

9620 In simplified transcription (see Appendix A) the symbol is used to
represent this sound. The use of for this vowel and a in the diphthongs
ai, au does not cause one word to be confused with another in phonetic
transcriptions. The letter ɑ is used in this book for ‘comparative’ reasons,
namely, to demonstrate in writing that the initial elements of ai, au are
different from Vowel No. 5, and to allow for the reduced forms of aiə and
auə (§§ 414, 430). There is, however, much to be said in favour of representing
Vowel No. 5 by in practical textbooks, and explaining once for
all the special values to be attached to the symbol a in ai, au.

9721 Are has also a weak form ə.

9822 The American name is ˈbəːkli.

9923 Also pronounced ˈmemwɔː, ˈrezəvwɔː.

10024 Except Saxons, Bavarians and some from the extreme North (Hamburg,
Lübeck, etc.), who often use an ɑ. similar to the English one.

10125 Usually æs for the animal, but ɑːs (less frequently æs) when applied
to a person as a term of contempt.

10226 These variations do not exist in Northern English. In the North all
the words written with ɑː in §§ 294, 295 are pronounced with æ or with the
Northern variant of this (a).

10327 The musical term bass is beis.

10428 Also wɔft.

10529 Want is wɔnt.

10630 Colloquial abbreviation of am not, used in the expression an't I? ˈɑːnt ai.

10731 But wrath is rɔːθ. The place-name Wrath is generally pronounced rɔːθ
by English people, but in Scotland it is raθ (which is sometimes imitated by
English people as rɑːθ or ræθ).

10832 Foreign learners often make the mistake of pronouncing this word
with ɔː instead of ɔ.

10932a Pronounced ˈkænət by some English people.

11033 Often pronounced inˈvɔːlv by foreign people.

11134 Often pronounced wɔːʧ by foreign people.

11235 In the sequences aus + consonant and aul + consonant many speakers
substitute the short ɔ, see § 300.

11336 Many people from the North and West of England use a close o or a
diphthong in words spelt with ore, oar, our and in many of the words
spelt with or + consonant. Thus more, board, course, port are pronounced
by them moə (or moər or moːr), boəd (or boərd or bord), koəs (or koərs
or kors), poət (or poərt or port), while the Received Pronunciation of these
words is mɔː (or mɔə), bɔːd (or bɔəd), kɔːs (or kɔəs), pɔːt. The chief words
written with or + consonant which have such alternative pronunciations with
close o are: afford, ford, horde, sword, fort, port (and the compounds export,
important, etc.), sport, proportion, forth, divorce, force, borne, sworn, torn, worn,
forge, pork. The chief words having no alternative pronunciation with close o
are: cord (and compounds record, etc.), chord, lord, order, form (and reform,
etc.), storm, adorn, born, corn, horn, morn, scorn, shorn, cork, fork, stork, York,
sort (and compounds resort, etc.), short, snort, north, George, gorge, horse, gorse,
remorse, corpse.

11437 The sequence represented by al + consonant has short ɔ in the speech
of some English people, e.g. hɔlt for the more usual hɔːlt.

11538 This is my natural pronunciation; it is now becoming old-fashioned.

11639 Except drought draut.

11740 Pour and pore have the variant pronunciation pO9- Roar has the variant
pronunciation rɔə. Door, more have the variant pronunciations dɔə, mɔə.

118 Voir note 117.

119 Voir note 117.

120 Voir note 117.

12141 In the less common sense of a ‘person who draws,’ the word is always
pronounced ˈdrɔːə. Drawers, the article of clothing, is drɔːz (identical in
pronunciation with draws).

12242 For has also a weak form . Four and fore have the variant pronunciation
fɔə.

12343 Source has the variant pronunciation sɔəs.

12444 Less commonly juə. There are also variants jɔə, joə.

12545 Or ˈkuʃn.

12646 The only exception is the rare word spook spuːk.

12747 In broom (plant), however, bruːm seems more frequent than brum.

12848 I pronounce these words with short u. But the use of in broom,
groom, and room is quite common in London.

12949 Weak forms kəd, ʃəd.

130 Voir note 129.

13150 Worsted from the verb to worst is ˈwəːstid.

13251 Weak forms wəd, əd, and d.

13352 It is not as a rule necessary to represent this advanced by a special
phonetic symbol; ü is, however, available for those who find a need for one.

13453 This word has weak forms and d. Before vowels the word do (whether
stressed or not) is generally pronounced du.

13554 This word has weak forms tu and . Before vowels the word to (whether
stressed or not) is generally pronounced tu.

13655 Also raut in route-march (ˈrautmɑːʧ or ˈruːtmɑːʧ).

13756 Wound from the verb wind is waund.

13857 The name Brougham used to be pronounced bruːm, but is now pronounced
bruəm or ˈbruːəm. The noun brougham is also pronounced bruəm or
ˈbruːəm.

13958 Also pronounced suːt.

14059 denotes a very open u, similar to the English short u. denotes
a very close u.

14160 ɷ is a symbol used in narrow transcription for the English short u.

14261 The symbol w is used here in a sense different from that assigned to
it in § 802. The two values are related in the same way as the two values
of j; see footnote 5 on p. 66.

14362 Pronounced by some priˈzuːm.

14463 But conjure (to appeal solemnly to) is kənˈdʒuə.

14564 Also ˈfrɔntjə.

14665 This word also has weak forms səm, sm.

14766 Brompton used to be ˈbrʌmptən, but is now more usually pronounced
ˈbrɔmptən. I pronounce accomplish with ɔ (ɔˈkɔmpliʃ), but a great many
English people pronounce ɔˈkʌmpliʃ.

14867 Southwark Bridge Road appears to be, however, more usually ˈsauθwək
ˈbridʒ ˈroud
, Southwark Bridge is ˈsʌðək ˈbridʒ or (less usually)ˈsauθwək
ˈbridʒ
.

14968 This word has also a weak form dəz.

15069 œ is obtained by adding lip-rounding to ɛ (§ 145 (2)).

15170 We have here again a case of a diaphone with several members. Some
people use a sound which has a lower tongue-position and is therefore more
ʌ-like in quality than that described in § 343. (Henry Sweet was one of these;
I remember well his variety of 91, which was a distinctly opener vowel than
mine.) Other Southern people have a closer əː than mine, i.e. a sound
resembling ə1 (§ 356) but lengthened. An exaggeration of this combined with
tongue retraction, producing almost an ɯ (§ 351), may be observed in the
‘clerical’ accent.

15271 When unstressed this word is often pronounced əː or or ə.

15372 Exceptions are beard biəd, heart hɑːt and hearth hɑːθ.

15473 Also pronounced ˈkɔːtjəs.

15574 Also pronounced ˈkɔːtisi.

15675 Also pronounced ˈæmətə, æməˈtəː and (rarely) ˈæmətjuə.

15776 More usually pronounced ˈʃoufə.

15877 œ is a lip-rounded ɛ; ø is a lip-rounded e. œ is the sound of eu in the
French neuf nœf and of ö in the German zwölf tsvœlf. ø is the sound of eu
in the French peu and of ö in the German schön ʃøːn.

15978 Colonel is the only word without an r in the spelling in which the sound
əː is used.

16079 This word has also a weak form .

16180 These variants being therefore members of the same diaphone.

16281 These variants being therefore members of the same phoneme.

16382 Also pronounced ˈmɔdn when in close grammatical connexion with the
following word (as in modern languages ˈmɔdn ˈlæŋgwidʒiz).

16483 Also pronounced ˈhipukrit.

16584 Not as a rule by ə3. It seems as if the words ending in ə (except a
and the) never occur in sufficiently close grammatical connexion with a
following word to induce a pronunciation with ə2, even if the following word
begins with k or g. Thus alpaca coat ælˈpækə ˈkout is said with ə1 while
I'll pack a coat ail ˈpæk ə ˈkout would generally be said with ə2. Cases in
which this distinction is made are so uncommon that they may be for practical
purposes disregarded.

16685 Commoner words like leopard ˈlepəd, standard 'stændəd, mustard
ˈmʌstəd, awkward ˈɔːkwəd, have ə1. Convert (noun) and bulwark are generally
pronounced with 91 (ˈkɔnvəːt, ˈbulwəːk) and occasionally with ɐ (ˈkɔnvɐt
ˈbulwɐk) but not as a rule with ə1. Forward is generally pronounced with 9i
(ˈfɔːwəd) and occasionally with ɐ (ˈfɔːwɐd); the word is distinct from
foreword ˈfɔːwəːd or ˈfɔəwəːd.

16786 Also pronounced ˈfoutəgrɑːf.

16887 Also pronounced fˈtɔgrfi.

16988 Also pronounced ˈeksməθ. Compare Bournemouth which is pronounced
ˈbɔːnməθ or ˈbuənməθ or ˈbɔənməθ.

17089 Noted many years ago by Sweet.

171 Voir note 170.

1721 This term was invented by H. E. Palmer. See his First Course of English
Phonetics
(1920), p. 23.

1732 Bass (fish) is bæs.

1743 ou is used in all words ending in -oll except doll dɔl, loll lɔl and Poll
(parrot) pɔl.

1754 Broad brɔːd is an exception.

1765 Sow (pig) is sau.

1776 A diphthong of the type ɔu is used for ou in various dialectal varieties
of English, but it is hardly to be recommended for foreign learners.

1787 Also pronounced ˈiːðə, ˈniːðə.

179 Voir note 178.

1808 It is not a real triphthong, but a sequence of two syllables (see Chap. XII,
and especially §§ 232, 233).

1819 Not ˈempɑː.

18210 Not ˈfɑ:ri, səˈsɑ:ti, inˈtɑ:li, ˈhɑ:r it, etc.

18311 This is my pronunciation.

18412 This is a good example of a case where in practice the results are not
what one would expect from theoretical considerations, since ordinary French
contains a but no æ.

18513 Not a and o, on account of the full back position of German o.

18614 But a bow for shooting, etc., is bou.

18715 But a row of houses, etc., is rou, as also is the verb meaning to propel
a boat with oars, and the corresponding noun.

18816 But the verb to wound and the noun wound are wuːnd.

18917 But the verb to sow is sou.

19018 It is not a real triphthong, but a sequence of two syllables (see Chap. XII,
and especially §§ 232, 233).

19119 It is not written so in ordinary transcriptions, since it is, as a rule, more
convenient in practice to adhere to the principle of admitting only one way
of representing each phoneme. Note that the ɔiə of employer is distinct
from the sequence ɔːjə heard in one pronunciation of lawyer ˈlɔːjə. (Some
English people pronounce the word as ˈlɔiə.)

19220 The terminal sound is then the vowel in the German Hütte ˈhytə
hübsch ˈhypʃ.

19321 Tear (verb) is tɛə as also is the corresponding noun meaning a rent.

19422 Also very commonly pronounced jəː. This is my pronunciation.

19523 The usual conversational form of we are.

19624 There has also a weak form ðə (ðər before vowels). Their before vowels
has occasionally a weak form ðər.

19725 Here is another case in which practical experience gives a method which
on theoretical grounds one would not expect to be effective. English speakers
do not as a rule separate the jaws in this manner in pronouncing ɛə or even
in pronouncing the variant ʌə.

19826 pɔː is also the pronunciation of paw.

19927 tɔː is also the pronunciation of tor.

20028 kɔː is also the pronunciation of caw.

20129 nɔː is also the pronunciation of nor and gnaw.

20230 lɔː is also the pronunciation of law.

20331 rɔː is also the pronunciation of raw.

20432 fɔː is also the strong form of for.

20533 sɔːis also the pronunciation of saw.

20634 ʃɔː is also the pronunciation of Shaw; ʃɔə and ʃɔː are also pronunciations
of sure (§ 463).

20735 Also pronounced juə and joə.

20836 wɔː is also the pronunciation of war.

20937 Except when unstressed, as in furniture 'fə:niʧə.

21038 Rarely stjɔət, and apparently never stjɔːt.

21139 ɔː was used, for instance, by the late Prof. H. C. Wyld, and this pronunciation
is recorded in his Universal English Dictionary.

21240 In verse the combination of i and ə in such words is sometimes disyllabic
(i-ə), though much more often monosyllabic (ĭə or ). In Shakespeare's
verse, for instance, the terminations of such words as audience, envious, are
always treated as monosyllabic except when they occur at the end of a line.
Compare Arise, fair sun, and kill the envious moon (Romeo and Juliet, ii, 2, 4)
and Be not her maid, since she is envious (Romeo and Juliet, ii, 2, 7).
Occasionally the monosyllabic pronunciation is needed at the end of a line, as
for instance the word bounteous in: and you yourself Have of your audience
been most free and bounteous
(Hamlet, i, 3, 93).

21341 The subject is discussed at greater length in my article Falling and Rising
Diphthongs in Southern English
in Miscellanea Phonetica II, 1954, published
by the I.P.A.

21442 Pianissimo, however, is often said with jəː pĭə'nisimou or pjə'nisimou
(also pjæ'nisimou and piə'nisimou).

21543 Fiat (make of motor-car), the legal term lien and the proper name Ian
are exceptions. They are normally said with (fiət, liən, iən) and
occasionally with iːə (ˈfiːət, ˈliːən, ˈiːən), but never with jəː.

21643a A short investigation of the status of and ĭə has been published by
B. S. Anjdrésen in Le Maître Phonetique, July, 1957, pp. 35-37.

21744 Also pronounced wəˈniːtə. (I have heard of three Englishwomen with
this name who pronounce it thus.)

21845 In verse the combination of u and ə in such words as those quoted in
§ 466m sometimes counts as two syllables, but more often as one. In Shakespeare
it counts as a single syllable except at the end of a line. Compare:

Valiant and virtuous, full of haughty courage (King Henry VI, iv, 1, 35).
The flinty ribs of this contemptuous city (King John, ii, 1, 384).
O, he was gentle, mild and virtuous (King Richard III, i, 2, 104).
CalVd Katliarina fair and virtuous (Taming of the Shrew, ii, 1, 43).

2191 The weak form used before consonants.

2202 Only used before vowels.

2213 Pronounced bin by some.

2224 See § 473.

2235 See § 870.

2246 Except in the special expressions my lord, which is generally pronounced
mi ˈlɔːd, and (at Eton College) my tutor mi ˈtjuːtə and my dame mi ˈdeim.
I do not use mi except in these and one or two other common expressions
such as never in my life ˈnevər in mi ˈlaif

2257 It is likewise customary on the stage in serious drama to give a weak form
min to the unstressed mine which in older literature replaces my before words
beginning with a vowel, thus mine eyes min ˈaiz, mine own min ˈoun.

2268 The use of a dot under a vowel letter to indicate a closer variety of vowel
was recommended by the International Phonetic Association in 1927 (see
Le Maître Phondtique, April, 1927, p. 14).

2279 There is a further difference between these examples, namely that the
ei preceding the t is shorter in ə ʥleitə ˈdei than in frəm ˈdei tə̆ ˈdei (see
§§ 892, 893).

22810 The forms without h are not used initially. Thus her hair is brown is
pronounced həː ˈhɛə z ˈbraun. For ə2 see example in § 488.

22911 Only used by servants addressing mistresses. When addressing members
of the royal family, the full form mæm is used when unstressed.

23012 In reality sʌm and səm (sm) are different words. See p. 135, footnote 21.

23113 That (demonstrative pronoun) has no weak form.

23214 Less commonly wɛə (before vowels wɛər).

23315 Also jɔə (before vowels jɔər). Also less commonly ju or joə (before
vowels juər or joər).

23416 Her can hardly be reduced to ə in this case. If ə were used, the sentence
would sound like she had a hat in a hand.

23517 In final position owed her is pronounced like odour (with ə3). The her
in paid her would also be pronounced with ə3 if it were in final position.

23618 Not ˈgiv ər ə ˈhæt which would mean Give her a hat.

23719 Also pronounced hiː ˈjuːst nɔt tu.

23820 Also pronounced ˈʃæl ai kʌm ˈwið juː (or … ˈwið ju).

23921 Meaning a certain quantity of paper. If the word paper were used in
the sense of a journal, some would have the strong form sʌm; example: I saw
it in some paper
ai ˈsɔ: it in sʌm ˈpeipə, meaning ‘I saw it in a journal, but
I don't remember which journal.’ Note that sʌm is unstressed in this case.
sʌm and səm (sm) should really be considered as different words.

24022 Meaning ‘we had better go’ (addressed to a member of the party denoted
by us), ˈlet əs ˈgou would mean ‘allow us to go’ (addressed to someone other
than a member of the party denoted by us). This difference was first pointed
out by Y. R. Chao (Le Maître Phonétique, January, 1931, p. 4).

24123 Some public speakers say ˈdʒentlmen.

24224 Also pronounced ˈtɔpmoust.

24325 Also pronounced ˌeniwɛə ˈniə, ˌeniwɛər ˈels.

2441 Some people, and particularly psychologists, pronounce ps in words
beginning with psych-; thus psychology is saiˈkɔlədʒi or (less commonly)
psaiˈkɔlədʒi.

2452 Now often spelt hiccup.

2463 ˚ is the sign devoicing, so that denotes ‘voiceless b’. It is its weak
force of articulation that distinguishes it from p.

2474 Also pronounced pɔː.

2485 Also pronounced poə or pɔə or pɔː.

2496 Iamb ˈaiæmb is an exception.

2507 This only applies to verbs, not to the termination -ed generally. Thus
wicked is ˈwikid (see footnote 10 to § 525).

2518 The only exception of importance is pestle, which is pronounced ˈpestl
by many.

2529 Incidentally, these palatograms corroborate a curious point previously
ascertained by direct observation, that while the English t is articulated
further back when followed by sounds of the u type than when followed
by sounds of the i type, yet in French the opposite is the case.

25310 When the verb ends in d (or in t) the termination is pronounced -id;
examples: added ˈædid, fitted ˈfitid. When the verb ends with a voiceless
consonant (other than t), the termination is pronounced -t (§ 514).

Note that the termination -ed in adjectives is almost always pronounced
-id. Hence a difference in pronunciation is made between aged (participle)
eidʒd and aged (attributive adjective) ˈeidʒid, blessed (participle) blest and
blessed (adjective) ˈblesid, etc. Similarly, the adverbs formed from participles
take the pronunciation -idli, whatever the form of the simple participle may
be; compare unfeigned ʌnˈfeind, unfeignedly ʌnˈfeinidli, marked mɑːkt,
markedly ˈmɑːkidli, composed kəm'pouzd, composedly kəmˈpouzidli.

25411 See § 590.

25512 Also pronounced ˈkuʃən and ˈkuʃn.

25613 Pronounced gɛəl by some.

25714 The principal words in which g before e or i is pronounced g are gear giə,
geese giːs, get get, gibberish ˈgibəriʃ (also ˈdʒibəriʃ), gibbous ˈgibəs, giddy ˈgidi,
gift gift, gig gig, giggle ˈgigl, gild gild, gill (of a fish) gil (gill, liquid measure, is
dʒil), gimlet ˈgimlit, gimp gimp, begin biˈgin, gird gəːd, girder ˈgəːdə, girdle
ˈgəːdl, girl gəːl, girth gəːθ, give giv, gizzard ˈgizəd; anger ˈæŋgə, conger ˈkɔŋgə,
eager ˈiːgə, finger ˈfiŋgə, hunger ˈhʌŋgə, linger ˈliŋgə, longer ˈloŋgə, longest
ˈlɔŋgist, (fish-)monger -mʌŋgə, stronger ˈstrɔŋgə, strongest ˈstrɔŋgist, tiger
ˈtaigə, younger ˈjʌŋgə, youngest ˈjʌŋgist; all words ending with -gger, -gging,
e.g. dagger ˈdægə, digging ˈdigiŋ; also the names Gertrude ˈgəːtruːd, Gibbon(s)
ˈgibən(z), Gibbs gibz, Gibson ˈgibsn, Gilbey ˈgilbi, Gilchrist ˈgilkrist, Gillespie
giˈlespi, Gillow ˈgilou, Gilpin ˈgilpin, Girton ˈgəːtn, Gissing ˈgisiŋ and a
number of less common names. Gill in ‘Jack and Gill’ (now more usually
written Jill) is dʒil, otherwise the proper name Gill is gil; Gifford is
ˈgifədand ˈdʒifəd, Gileon is ˈdʒilsn and ˈgilsn, Gimson is ˈgimsn and ˈdʒimsn.

25816 With voiced plosive consonants the amount of force does not appreciably
influence the effect of the sound on the European ear.

25917 The word is often reduced to ˈemti; there is also a variant ˈemm̥ti.

26018 With many speakers the whole of the stop is voiceless in these cases.

261 Voir note 260.

26219 Also often pronounced igˈzækli or gˈzækli.

2631 The Middle German sound of w, Spanish sound of b.

2642 For ð see § 702.

2653 ϕ is the symbol for ‘bi-labial f.’

2664 Generally simplified to or tr.

2675 X represents the German ach-sound.

2686 If we except the case of ‘doubling’ in languages such as Italian and
Hindi.

2697 Some phoneticians recommend using c and ɟ in transcribing English.
I am doubtful if there is anything to be gained by adding these two symbols,
and it might be urged that there would be an inconsistency in not also introducing
special signs for tr and dr.

2708 Exceptions are aperture ˈæpətjuə, overture ˈouvətjuə (rarely with -ʧə).

2719 For the words in which g before e or i has the sound g see footnote 14
on p. 149.

27210 But Ipswich ˈipswiʧ, Droitwich ˈdrɔitwiʧ. Some say ˈsænwiʧ in the
singular, but ˈsænwiʤiz seems to be universal for the plural. The placename
Sandwich is more usually ˈsænwiʧ but some say ˈsænwiʤ and there
exists an old-fashioned pronunciation ˈsæniʤ. Ostrich is usually pronounced
ˈɔstriʧ, but the form ˈɔstriʤ may also be heard.

27311 Also pronounced ˈtetsi.

27412 Meaning a sliding box in a table, etc. Drawer meaning a person who
draws is ˈdrɔːə.

27513 It is noteworthy that bedroom is usually said with the affricate dr. In
other compounds the d and r are pronounced separately, as shown in § 636.

276 Voir note 275.

277 Voir note 275.

2781 The mouth may be kept open if necessary by means of a large cork,
1½ inches wide, placed between the front teeth.

2791 Many English people use a very dark l when it is syllabic (as in ˈpiːpl)
and a less dark variety in other cases (as in fiːl, fiːld).

2802 Also kwɑːm.

2813 ˈhoulbən is now sometimes heard from speakers of Received English,
and this pronunciation seems to be gaining ground.

2824 But salve in the sense of ‘to save a ship’ is sælv. Valve is vælv.

2835 It is often convenient in oral work to refer to these sounds as ‘l with
i-resonance,’ ‘l with u-resonance,’ etc.

2846 It must not be thought that the peculiar quality of the dark l as compared
with the clear l is due to the retraction of the tip of the tongue. A
dark l with u resonance pronounced with the tip of the tongue against
the back part of the teeth-ridge is indistinguishable as regards acoustic
effect from a dark l with u resonance pronounced with the tip of the tongue
right against the teeth. Similarly a clear l with i resonance pronounced
with the tip of the tongue against the back part of the teeth-ridge is
indistinguishable acoustically from a clear l with i resonance pronounced
with the tip of the tongue against the teeth. The same applies to all the
other varieties. Note that the English dark l is articulated with the tip of
the tongue against the teeth in such a word as health helθ; note also that
if a foreign learner is unable to pronounce the English dark l with the tip
of the tongue right against the teeth, he may be quite certain that he is
forming the sound incorrectly (see § 670).

2857 Other than j.

2868 Both are subject to slight variations depending on the nature of the
adjoining vowel. The only cases of note are when the adjoining vowel
is ɑ or ɔ. When the dark l is preceded by a or ɔ, its resonance tends towards
these vowels; and when the clear l is followed by ɑ or ɔ, it tends towards
a ‘neutral’ l with the resonance of ə. These varieties are other subsidiary
members of the Southern English l-phoneme.

It should be remarked here that the treatment of l-sounds is different
in other types of English. In particular, in Scottish English and in American
English dark l is commonly used in all positions. In Irish English l is clear
in all positions. There are, moreover, English people, especially in the
North, who use intermediate l-sounds in all situations.

2879 The reason for saying ‘between the teeth’ is that many foreign learners
try to obtain the peculiar resonance of the English lu by curling back or
‘inverting’ (§ 827) the tip of the tongue. The sound so formed is quite
different from a dark l. The tendency to invert the tongue is avoided if the
tip of the tongue is placed between the teeth, and when once l can be correctly
pronounced with the tip of the tongue between the teeth, there is no difficulty
in retracting it to the more usual position just behind the upper teeth. See
footnote 6 on previous page.

28810 Most foreign learners, other than the Russians and the Japanese, do not
need to practise the words with clear l. The l which they are accustomed
to use in their own language generally suffices.

2891 I have been informed by Miss Iza Thompson, a particularly able teacher
of the deaf, that it is useful to point out the existence of these subsidiary
members in teaching speech to the deaf.

2902 Also pronounced kɔːf, trɔːf.

291 Voir note 290.

292ʒ But slough meaning a ‘morass’ is slau.

2934 In the army. In the navy there existed until recently the pronunciations
luːˈtenənt and ˈluːtnənt. Now the form lefˈtenənt is usual in the navy as
well as in the army. There exists also a pronunciation ləˈtenənt.

2944a Also ˈbælfəə, ˈbælfɔə, ˈbælfɔː.

2955 There is a growing tendency to use the θs forms in many other words,
e.g. truths, baths, oaths.

2966 But the English river Thames is temz, as also are the rivers of that name
in Canada and New Zealand; the river Thames in Connecticut is θeimz.
Thame in Oxfordshire is teim.

2977 Pronounced wiθ in the North of England.

2988 The verb south is pronounced sauθ by some.

2999 Subsidiary members with varying lip-positions do exist. They are
unimportant for ordinary language teaching, but I have been informed that
the speech of deaf-mutes can be considerably improved by directing attention
to them.

There exists also a weakened form of s which is often used at the ends of
words when a breathed consonant precedes, as in box bɔks, books buks,
shuts ʃʌts. Some English people use (§ 722) in such words. See my book
The Phoneme, §§ 171-175.

30010 Many English people use a variety of s involving an articulation by
the lower lip against the upper teeth in addition to the tongue articulation.
This kind of s differs considerably in quality from the normal English s.

30111 For examples of mute final s see § 718.

30212 Reduced to s in the expression let us… lets… meaning ‘we had better’
(see footnote 21 on p. 135).

30313 Foreign people often say ʌz, ˈpreʃəz, etc.

30414 Grease (v.) is griːz.

30515 When the noun is used in the technical legal sense, opposed to lease,
it is commonly pronounced ˈriːˈliːs.

30616 Chiefly used in the plural premises ˈpremisiz.

30717 Close (noun meaning ‘end,’ and v.) are pronounced klouz.

30818 Refuse (v.) is riˈfjuːz.

30919 The rare verb meaning ‘to catch mice’ is mauz (also maus).

31020 The plural is kɔːz.

31121 Often pronounced ˈʃæmi in the expression chamois leather.

31222 The plural is ˈrɔndivuːz.

31323 Or sɔː.

31424 Or kɔːs.

31525 differs from s in being uttered with weaker air-pressure.

31626 The only exceptions are yes jes and a few proper names such as Agnes
ˈægnis, Elles ˈelis. Foreign learners should note that the letter c is never
pronounced z. Note the pronunciation of Latin plurals in -es (-iːz), e.g.
axes (plural of axis) ˈæksiːz. Compare axes (plural of axe) ˈæksiz.

31727 And all other words ending in -tial except bestial ˈbestjəl and celestial
siˈlestjəl

31828 Note associate (verb) əˈsouʃieit, appreciate əˈpriːʃieit, appreciation
əpriːʃiˈeiʃn, negotiate niˈgouʃieit, negotiation nigouʃiˈeiʃn, but association
əsousiˈeiʃn (less commonly əsouʃiˈeiʃn), pronunciation prənʌnsiˈeiʃn.

31929 Or ʃɔː.

32030 Also pronounced ʃoə, ʃɔə, and ʃɔː.

32131 ʒ̊ differs from ʃ in being uttered with weaker air-pressure.

32232 Not, however, in words like rosier ˈrouzĭə (comparative of rosy).

32333 Pronounced by some trænˈziʃn; also trɑːnˈsiʒn and trɑːnˈziʃn.

32434 Also ˈgæriʤ. The verb to garage is usually ˈgæriʤ.

32535 Also pronounced ˈespĭəniʤ and esˈpaiəniʤ.

32636 ɽ is the Swedish dialectal ‘thick l’ and the Indian sound commonly
represented romanically by .

32737 Also joə or (less commonly) juə.

32838 Or aər ˈoun or aːr ˈoun (§ 430).

32939 Note the various possible pronunciations of for him in it's very good for
him
, when the him is unstressed. They are fɔː him, fə him, fɔːr im, fɔr im
fər im, fɔː im; of these fɔr im is perhaps the best for foreign learners to use
Perhaps is pəˈhæps or præps; either form may be used in any position
pəˈhæps is fairly common parenthetically (as in you know, perhaps, …
juː ˈnou, pəˈhæps, …), and præps is more usual in other cases (e.g.
perhaps we shall ˈpræps wiː ˈʃæl).

33040 This action is made possible by holding the tongue in a ‘sulcal’ position;
that is to say, the tongue is held so that the two sides are rather high, but
there is a depression or groove down the centre. It is in this groove that the
uvula vibrates. The height of the sides of the tongue is indicated by the
dotted line in Fig. 108.

33141 Also pronounced ruːm.

33242 Pronounced with in some parts of the country (see footnote 36 on
p. 80).

333 Voir note 332.

334 Voir note 332.

335 Voir note 332.

33643 These words might be written in very narrow (allophonic) transcription
i̥it, ɑ̥ɑːd, u̥uk, but such a mode of representation would be both inconvenient
and unnecessary.

33744 Also very commonly præps. See footnote 39 on p. 196.

33845 The letter h of ordinary spelling often has no separate consonantal value
at all, but is used in conjunction with other letters to form digraphs with
special values: thus ch, th and sh generally stand for ʧ, θ or ð. and ʃ
respectively.

3391 In speaking of these sounds they must be called , , on account of
their gliding nature.

3402 Not however in conquer ˈkɔŋkə, etiquette etiˈket, exchequer iksˈʧekə,
liquor ˈlikə, and a few other words.

3413 Note also that the verb will (strong form wil) has a weak form l.

3424 Or wɔː.

3435 Or ˈwuːə.

3446 Note that i does not usually have the value of j when followed by vowels
other than 8. Thus peculiarity, pronunciation are with most speakers
p^kjuili'seriti, pr8,nAnsi'eiJn (not pikjui'ljseriti, prarun'sjeifn as sometimes
pronounced by foreign people).

3457 Also pronounced ˈvælju.

3468 More often pronounced with short u (ˌæljuˈminjəm).

3479 Or jɔː, also jɔə and (less commonly) juə.

34810 Also biˈɔnd.

34911 The opening may be as wide as for cardinal i or even slightly wider.

3501 We are here speaking of nasalization which is merely the result of habit
and not due to any physical defect.

3512 Also pronounced ruːm.

3521 The same acoustic effect may also be produced by a lateral contraction
of the tongue (§ 73) combined with a retraction of the tip without raising it.

3532 An alternative symbol for ɚ, is ɚ (much used by American writers).

3541 The term ‘similitude’ is only used in reference to subsidiary members
of phonemes. Sounds which resemble each other but belong to separate
phonemes often occur in sequence, but such resemblances between consecutive
sounds are not called similitudes. Such are the resemblances between ŋ
and k in conquest ˈkɔŋkwest (where both sounds are velar) or between z and g
in dogs dɔgz (where both sounds are voiced). Such sequences are often
the result of the historical process known as assimilation, but the resemblances
do not come within the definition of similitudes.

Similitudes are to be found both in single words and when two words are
put together in forming a sentence. In the latter case the similitudes are
brought about by contextual assimilation (§ 838). An example of it is the
use of dental n in ˈwʌn ˈθiŋ, ɔn ðə ˈgraund quoted in § 845 (ii). Similitudes
occurring in single words are not necessarily brought about by assimilation.

3552 Many writers have used the term ‘assimilation’ loosely to include
similitude. It is desirable to avoid this ambiguity by using separate terms.

3563 The consonant is more breathed after p t k than after the fricatives;
thus the l in place has a larger proportion of breath than the l in slate sleit.

The partially breathed members of these phonemes are not used if the
preceding consonant belongs to a different syllable, e.g. in at once ət'wʌns.
at rest ət ˈrest (not əˈtw̥ʌns, ə'tr̥est). See § 1094 (1)-(3), also my article on
The Word as a Phonetic Entity in Le Maître Phonétique, Oct.-Dec, 1931.

3574 An example is the use of e and ɛ in Zulu. In that language e is only
used when the following syllable contains i or u, while ɛ is used finally and
when the following syllable contains some other vowel, e.g. leli (this) but
wɛna (thou).

3585 The e in well is near to cardinal vowel No. 3 (ɛ).

3596 In such a word as absorptive no assimilation has taken place. This
word is a modern invention and has never had a b before the t. It has
the pronunciation with p by analogy.

3607 If the supposition as to the value of final s in Early English is correct,
there has been no assimilation in the case of cats kæts, books buks, etc. The
value of the s has simply remained unchanged.

3618 It does not seem possible to instance the resemblances appearing in
such words as conquer ˈkɔŋkə, congress ˈkɔŋgres, congregation kɔŋgriˈgeiʃn
as the results of assimilation. There does not appear to be any evidence
to show that these words were ever pronounced with n.

3629 For c see § 538 and Fig. 29.

36310 The further change c > ʧ is not a case of assimilation. This change
was no doubt due to an instinctive desire to make clearer the distinction
between words beginning with ci- and those beginning with ki- which
appeared in the language at a subsequent time when the tendency to complete
fronting was no longer operative.

36411 y stands for the French sound of u; it has the tongue-position of i combined
with lip-rounding. myːs no doubt subsequently underwent changes
of the following types, which are isolative and not due to assimilation:
myːs > miːs > mïis > məis > mais.

36512 This pronunciation was first noticed by H. O. Coleman: see his article
entitled ‘inːividːjul pikjuːliæritiz’ in Le Maître Phonétique, July-August, 1911.

36613 But Whose is this? ˈhuːz iz ˈðis, George is here ˈʤɔːʤ iz hiə.

36714 But the price is sixpence ðə ˈprais iz ˈsikspəns, this fish is very good
ˈðis ˈfiʃ iz ˈveri 'gud.

36815 The older form ˈjuːzd tu is still sometimes heard.

36916 By analogy, the pronunciation juːs(t) is now employed even where to
does not immediately follow. Thus we say Used he to? ˈjuːst (h)iː tu, Yes
lie used
ˈjes hiː ˈjuːst; and a negative ˈjuːsnt has been constructed on the
model of mustn't ˈmʌsnt, etc.: ˈjuːsnt (h)iː, ˈnou hi: ˈjuːsnt. ˈjuːsnt is
quite a common word in the spoken language, but I do not remember ever
seeing it written; it would presumably have to be spelt usedn't. If, as is
probable, used to was formerly pronounced ˈjuːzd tu, the change to ˈjuːst tu
illustrates historical as well as contextual assimilation.

37017 Or '/At jo:r 'aiz.

37118 Or '/At jo: 'mau8.

37219 An example from French is seen in the frequent pronunciation of les
langues modernes
as le lɑ̃ːŋ mɔdɛrn instead of le lɑ̃g mɔdɛrn; g is replaced
by ŋ under the influence of the adjoining nasal sounds. The result is a
similitude, ŋ being an allophone (subsidiary member) of the g-phoneme in
French.

37320 -tj- is used in all words spelt with -teous except righteous.

3741 For the evidence see H. C. Wyld's History of Modern Colloquial English
(Blackwell, Oxford), pp. 298-300.

3752 The elision is by no means a recent one. The Oxford Dictionary records
a spelling torter shell in 1652.

3761 In rare cases it is useful to mark extreme shortness. This may be done
by placing ˘ over the symbol of the sound.

3772 Also pronounced hɔəd.

3783 The vowel ɑː appears not to undergo as much shortening as the other
vowels. Thus the ɑː in bark bɑːk is shorter than that in barge bɑːʤ, but
is longer than the in beak biːk.

3794 Drawing-room (salon) is pronounced exceptionally with the diphthong ɔi.
Drawing-room meaning a room for drawing is pronounced ˈdrɔːiŋrum
according to the rule.

3805 Note. The length varies according to the nature of the sound following
the ‘long’ vowel. When an unstressed vowel immediately follows (as in
seeing, drawing), the long vowel is distinctly shorter than when a consonant
intervenes (as in leader, causes).

3816 Prayer (supplication) is pronounced prɛə, while prayer (one who prays)
is ˈpreiə or preə.

3827 Pronounced 'pɔitri by some.

3838 Also pronounced o'veiʃn.

3849 Also pronounced ˈfelo.

38510 E.g. by introducing into broad transcriptions some such system as that
now used in narrower transcriptions, where ‘short’ i, u, ɔ are written ɩ, ɷ, ʋ
and the vowel in bird is written with ɜ3. See Appendix A, §§ 28, 33, 36.

38611 The lengths indicated by the musical notes are not the lengths of the
syllables but the lengths separating the ‘stress-points’ or ‘peaks of prominence’
of the syllables. The note attached to the last syllable of each example
denotes the length of time which would presumably elapse between its
stress-point and that of a following stressed syllable if there were one.

38712 Is it right? may be said with the rhythm ♪♩; this is presumably
a consequence of the presence of the two consonants t, r.

38813 For a more detailed discussion of the rhythm of English, readers are
referred to A. Classe's The Rhythm of English Prose (Blackwell, Oxford).

3891 Except in rare cases where strong stress falls on a sound which has no
exhalation. It must be observed that cases do occur where a strong stress
fails to give much carrying power to a sound, and therefore fails to make
it objectively prominent. A strong stress may even occur on a silence,
e.g. on the stop of a voiceless plosive. When a strong stress is given to a
sound incapable of receiving any noticeable increase of loudness, a person
unfamiliar with the language would be unable to tell that a stress was present
except by observing the gestures. A hearer familiar with the language
would not perceive the stress objectively from the sound apart from the
gestures, but he perceives it in a subjective way; the sounds he hears call up
to his mind (through the context) the manner of making them, and by means
of immediate ‘inner speech’ he knows where the stress is. The process is
analogous to that by which the beats of the bar are felt in syncopated music
at points where no notes are played. (This type of process is known, to
psychologists as “empathy.”)

Strong stress without strong force of exhalation and consequent loudness
is not often found in English. It may, however, be observed in one pronunciation
of Thank you, viz. the abbreviated form ˈk̩kju. Here a syllabic
k without plosion is stressed although it has no sound; the stress is generally
shown by a gesture. Strong stress without accompanying loudness is a
common feature of the Tswana language of South Africa. In that language
final low-tone syllables, as for instance the second syllable of the word thata
(‘strong’), are often said without voice and (when plosives precede) with closed
glottis. They have a very strong stress, quite as strong as that of the
penultimate, but owing to lack of sonority of unvoiced vowels they have
very little loudness.

3902 Some phoneticians have expressed the view that stress is not independent
of pitch, and have shown by experiments with a dead larynx that an increase
of stress involves a raising of pitch. This is no doubt the case for a given
tension of the vocal cords. But the living speaker does not maintain a fixed
tension of his vocal cords; he has complete control of his intonation, whatever
the stress may be, and it often happens in a language that strong stresses
are found on low-pitched syllables and weak stresses on high-pitched syllables.
It appears to me therefore that linguistic stress must be regarded as
independent of pitch.

3913 In the above paragraphs (§§ 919-920) I have taken no account of a
theory advanced by some that no syllable is really ‘unstressed’ in English
unless it contains one of the vowels ə, i, ɔ (the monophthongal reduction of
ou) or u or a syllabic consonant. Those who maintain this theory appear
to think that when e, æ, ɔ, ʌ and the long vowels and the falling diphthongs
occur in weak positions, they are pronounced in reality with secondary stress.
It would seem, for instance, that they consider the second syllables of such
words as insect ˈinsekt, asphalt ˈæsfælt, teapot ˈtiːpɔt, hiccup ˈhikʌp, concrete
ˈkɔŋkriːt, schedule ˈʃedjuːl, mundane ˈmʌndein, fortnight ˈfɔːtnait to have
secondary stresses, and the first and third syllables of portmanteau
pɔːtˈmæntou to have secondary stresses.

I am not satisfied that this view is a correct one. It is to be observed in
regard to e, æ, ɔ and ʌ that these vowels are undoubtedly more prominent
(§§ 100, 101, 208) than ə, i, o and u; but this does not necessarily mean that
they always derive their prominence from stress as here defined, i.e. from a
special push of the chest wall. I submit that they have considerable prominence
by reason of their ‘inherent sonority’ (§§ 100, 101), and that if (say)
e and əˌ (§ 356) are uttered with what the speaker judges to be equal push from
the chest wall, and the conditions are in other respects comparable, e ‘carries’
better than ə does, i.e. it is clearly audible at a greater distance than ə is.

The most that can be said in favour of the theory, in so far as it concerns
e, æ, ɔ and ʌ, is that these vowels are generally uttered with greater jaw
movement than ə, i, o and u, and perhaps that it is customary to give them
slightly greater length than that which ə, i, o, u have in comparable positions.
J. W. Jeaffreson, it is true, has maintained that the extent of jaw movement
is an indication of stress, but it is not certain that this hypothesis is always
valid. His experiments, valuable as they are, did not demonstrate that
the English e, æ, ɔ and ʌ are uttered with stronger push from the chest wall
than ə, i, o and u. Readers interested in this question are recommended
to study Jeaffreson's remarkable results. They are set out in his paper
Stress and Rhythm in Speech in the Transactions of the Philological Society,
1938, and in his unpublished Mensuration of French Verse (thesis for the
London M.A., 1924) which may be consulted in the Library of the University
of London. See also some remarks in my book The Phoneme, §§ 204, 205
and footnote 14 on p. 60.

The English vowels , ɑː, ɔː, and əː have considerable prominence by
reason of their length, but I would suggest that there is nothing to prevent
them from being ‘unstressed,’ i.e. uttered with very weak push from the
chest wall. The same applies to the diphthongs ei, ou, ai, au and ɔi,
which are ‘falling’ (§§ 220, 223) by reason of the relatively small inherent
sonority of their terminal elements. The same probably applies also to
ɛə and ɔə.

It looks as if the only syllabic sounds of Southern English which can
properly be said to have inherent stress are (No. 18) and (No. 21),
since, as explained in §§ 225, 440a and 460, these diphthongs owe their ‘falling’
character to a certain degree of stress on their initial elements. Those who
maintain that they cannot be ‘unstressed,’ and that therefore such words
as reindeer ˈreindiə, contour ˈkɔntuə, are pronounced with secondary stress
on the second syllable, have thus a good case. But even here it might be
argued that when a sound is said with the minimum stress it is capable of
having, it should be considered as unstressed. The question is discussed at
some length in the sections relating to and in my article Falling and
Rising Diphthongs in Southern English
in Miscellanea Phonetica II, 1954
(published by the I.P.A.).

In phonetic transcriptions designed for foreign learners I find it adequate
to mark with a secondary stress-mark only such syllables as those exemplified
at the end of § 919a together with those which have a reduced primary stress
in the sentence. To mark weak syllables with secondary stress-marks solely
because they contain e, æ, ɔ and ʌ or long vowels or diphthongs seems to me
hardly to represent the facts correctly. In any case, I doubt if such marking
would serve any useful purpose in texts intended for the practical teaching
of English stress, since such syllables are bound to have sufficient prominence
if the vowels are correctly pronounced.

3924 The prefixes here referred to may be conveniently termed the ‘separable’
prefixes.

393 Voir note 392.

3945 Compare recover (‘get back’) riˈkʌvə with recover (‘cover again,’ said of
umbrellas, etc.) ˈriːˈkʌvə. In reproduction the re- is not felt as separable,
and the normal pronunciation is accordingly riː-prəˈdʌkʃn.

3956 For miscellaneous compounds with double stress see §§ 948 ff.

3967 This word is usually pronounced ˈɑːˈmen in Church of England churches;
elsewhere both forms are heard, ˈɑːˈmen being probably the more usual.
Amen Corner is however 'eimen ˈkɔːnə.

3978 The plural princesses is prinˈsesiz.

3989 But Canton in Wales is ˈkæntən. The heraldic term canton is ˈkæntən
Canton meaning a state in Switzerland is generally pronounced ˈkæntɔn
but some say ˈkæntən.

39910 Carlisle in Cumberland is locally ˈkɑːlail.

40011 But Newquay is ˈnjuːkiː or ˈnjuːki.

40112 Also prounounced ˈmɔlˈtiːz, mɔlˈtiːz.

40213 Either for ‘intensity’ or for contrast. See § 1046.

40314 Or ˈæˈsendiŋ ən ˈdiː(ˈ)sendiŋ, ˈɔˈfensiv ən ˈdiː(ˈ)fensiv.

40415 In the North of England and in Scotland the words reconcile, criticize,
recognize are usually stressed on the last syllable.

405 Voir note 404.

406 Voir note 404.

40716 A number of other useful words with secondary stress on the second
syllable will be found in an article on Secondary Stress by L. J. Guittart in
English Studies, Vol. XII, No. 1, Feb. 1930.

40817 Also pronounced,ˌæntisiˈpeiʃn.

40918 Also pronounced ˌæntægəˈnistik.

41019 Also pronounced,ˌbæktiəriˈɔləʤi.

41120 Now perhaps more commonly pronounced ˈkærikətjuə

41221 Also pronounced ˌænsaikləˈpiːdjə.

41322 If there is any secondary stress, it is on the first syllable i-, and is only
subjective. The second syllables are so prominent by the nature of their
sounds that it is difficult to make a stress on the first syllable objectively
audible.

41423 A few isolated compounds have single stress on the second element.
The chief are: compounds with -ever (e.g. whenever weˈnevə), -self (e.g. himself
himˈself, themselves ðəmˈselvz), and the words hereafter hiərˈɑːftə, thereafter
ðɛərˈɑːftə, throughout θruˈaut, wherein wɛərˈin, already ɔːlˈredi, look-out
lukˈaut, uphold ʌpˈhould, shortcomings ʃɔːtˈkʌmiŋz.

41524 Or ˈbʌkiŋəmʃə.

41625 Newcastle-on-Tyne is pronounced locally njuˈkæsl.

41726 Great-coat and greengage are said with single stress (ˈgreitkout, ˈgriːngeiʤ
or ˈgriːŋgeiʤ) by some English people.

41827 But muddle-headed is pronounced with single stress (ˈmʌdlhedid).

41928 Exception long-tailed (tit) ˈlɔŋ-teild, due no doubt to the fact that this
word is always attributive and therefore takes stress on the first syllable
by the principle of rhythm (§§ 931, 954).

42029 Compare good-looking ˈgudˈlukiŋ which is not equivalent to ‘good.’

42130 Note, however, hereafter ˈhiərˈɑːftə, thereafter ðɛərˈɑːftə.

42231 Note, however, hereabouts ˈhiərəbauts and the noun whereabouts ˈwɛərəbauts.
Also ðɛərəˈbauts in the expression there or thereabouts (ðɛər ɔː
ðɛərəˈbauts
). Wherein is always wɛərˈin.

42332 Also ˈmiːntaim, ˈmiːnwail.

42433 My pronunciation. See, however, footnote 26 on p. 258.

42534 Also ˈinlənd ˈrevinjuː.

426 Voir note 424.

42735 Have used as a principal verb is exceptional. It is often unstressed,
though it. generally appears in the strong form hæv.

42836 The adverb most is exceptional. In a most important thing ə moust
imˈpɔːtnt ˈθiŋ
, their most valued possessions ðɛə moust ˈvæljuːd pəˈzeʃnz
the moust would not be stressed, except for special emphasis. The substantival
and adjectival most are, however, stressed; examples: most of the
houses were empty
ˈmoust əv ðə (ˈ)hauziz wər ˈemti, most bears are brown
ˈmoust ˈbɛəz ə ˈbraun, and the special expression for the most part fə ðə
ˈmoust pɑːt
. More is treated similarly; examples: that's a more serious
matter
ˈðæt s ə mɔː ˈsiərĭəs mætə (adverbial more), there were more than
I expected
ðɛə wə ˈmɔː ðən ai iksˈpektid (substantival more), more haste less
speed
ˈmɔː ˈheist ˈles ˈspiːd (adjectival more).

The adjective little is generally not stressed. Compare they lived in a little
house near the wood
ðei ˈlivd in ə litl ˈhaus (ˈ)niə ðə ˈwud with they lived
in a small house near the wood
ðei ˈlivd in ə ˈsmɔːl ˈhaus (ˈ)niə ðə ˈwud,
a good little boy ə ˈgud litl ˈboi with a big fat boy ə ˈbig ˈfæt ˈboi. There are,
however, exceptions, e.g. little things please little minds ˈlitl ˈθiŋz ˈpliːz ˈlitl
ˈmainʣ
; stress appears to be put on little when the word is used to imply
a considerable degree of smallness.

42937 But Canning Town is sometimes said with single stress (ˈkæniŋ taun).

43038 The case of a verb with a preposition is, however, different; in this case
the verb only has stress. Examples: meet with ˈmiːt wið, enter into (an agreement)
ˈentər intu.

43139 Some is here used in the collective sense, which is distinct from the
indefinite (partitive) sense. The indefinite some is pronounced səm or sm,
and the following word is stressed, e.g. there were some books on the table
ðɛə wə səm 'buks ɔn ðə ˈteibl. Some denoting one of a class is pronounced
sʌm but has no stress, e.g. we must try and get hold of some teacher wiː məs
ˈtrai ən get ˈhould əv sʌm ˈtiːʧə
. (But some teachers meaning ‘a few teachers’
would be səm ˈtiːʧəz, or if contrasted with ‘other teachers,’ ˈsʌm tiːʧəz.

43240 The French lady's maid may mean two different things according to
the way in which it is stressed. ðə ˈfrenʃ ˈleidiz meid means ‘the lady's
maid who is French’; ða ˈfrenʃ leidiz ˈmeid means ‘the maid employed
by the French lady.’

43341 In rapid conversation often wiː ˈkɑːŋk get ˈaut.

43442 The rhythm of it ˈsiːmz sou ˈfʌni is ♪♪♬

43543 Also, in old-fashioned pronunciation, ˈgɔf klʌb

43644 Loss commonly ˈbæŋk ˈnout.

43745 But wedding present has single stress (ˈwediŋ preznt). Some people,
especially in the North of England, use single stress on birthday present and
Christmas present. Wedding breakfast has double stress (ˈwediŋ ˈbrekfəst
in my pronunciation. So also have Christmas dinner ˈkrisməs ˈdinə
Christmas pudding ˈkrisməs ˈpudiŋ, Christmas wishes ˈkrisməs ˈwiʃiz
Christmas Day ˈkrisməs ˈdei, Easter egg ˈiːstər ˈeg, birthday greetings ˈbəːθdei
ˈgriːtiŋz
. Birthday cake, wedding cake, wedding day have single stress
(ˈbəːθdei keik, ˈwediŋ keik, ˈwediŋ dei).

43846 The normal (unemphatic) pronunciation of these sentences would be
it kən bi ˈdʌn, it s bi(:)n ˈdʌn.

43947 These expressions might also be pronounced ai ˈʃəd ˈnɔt əv ˈθɔːt sou,
wiːv ˈnɔt bi(:)n ˈeibl tu.

44048 The normal (unemphatic) pronunciation of these sentences would be
ˈwɔt ə juː ˈduiŋ, ˈwɔt s tə bi ˈdʌn, hau did ðei ˈmæniʤ it.

44149 But again meaning ‘a second time’ is stressed, e.g. ˈput it ˈbæk əˈgein
(= put it back a second time), hiː z ˈgouiŋ ˈaut əˈgein ˈsuːn (= he's soon
going out a second time).

44250 But not the conjunctions now, then (see § 1001).

44351 ˈintə before consonants.

44452 Except to, which is generally said with the weak form tu in such cases.
The strong form tuː (unstressed) would also be possible.

44553 The pronoun it would not be stressed in any case. If emphasis were
required, it would be replaced by this or that.

44654 Nor introducing a sentence is almost always stressed, unless combined
with another word, as in nor yet nɔː ˈjet.

4471 About 20 per cent, of the sounds used in speaking a connected passage
of English are voiceless.

4482 I have, however, frequently heard F imageand even E image
from women whose voices did not sound abnormally low. With women
whose voices sound distinctly lower than the average, notes as low
as D image and C image may often be recognized. Speaking
generally, however, notes cannot be clearly recognized much below
G image the voice then degenerating into a kind of growl without
recognizable pitch.

4493 This system is a modification of that used by H. Klinghardt in his
Übungen im Englischen Tonfall, which was first published by Otto Schulze
in Cöthen in 1920.

4504 Also image In modern usage image is sometimes used
on meeting, as well as on parting.

4515 Published by the International Phonetic Association, and obtainable
from the Secretary of the Association, Department of Phonetics, University
College, London, W.C.1.

4526 Published by Quelle & Meyer, Leipzig.

4537 Published by Heffer, Cambridge.

454 Voir note 453.

4558 Published by the British Council, 65, Davies Street, London, W.1.

4569 Published by Fehr'sche Buchhandlung, St. Gall, 1935.

45710 Max Niemeyer Verlag, Tübingen.

45811 Radiotjänst, Stockholm.

459 Voir note 451.

46012 No. A45 in the Prace Wrocławskiego Towarzystwa Naukowego (publications
of the Wrocław Literary and Scientific Society), Wrocław, Poland, 1952.

46113 Published by Longmans, Green & Co., 1958.

462 Voir note 461.

46314 University of Michigan Press, 1949.

46415 R. Kingdon has adduced reasons for reversing these numbers. I suggest
that those who favour the Kingdon system should call the present Tunes 1
and 2 ‘Tune K2’ and ‘Tune K1.’

46514 This might also be pronounced
image
ˈæz it wəz ˈkwait imˈpɔsəbl tə ˈfiniʃ it,…

46616 Armstrong and Ward considered this treatment of medial unstressed
syllables commoner than the level pitch given here as the normal form
(Handbook, p. 5, and throughout the examples).

46717 Handbook, p. 5.

46818 This fact was first pointed out by Coleman (Intonation and Emphasis, § 60).

46919 Except when a sound of a strong form happens to have greater intrinsic
sonority than the corresponding sound of the weak form. In this connexion
it may be repeated here that some sounds are naturally more prominent than
others when said with the same degree of stress or 'push from the chest wall';
in other terms, the inherent sonority of some sounds is greater than that of
others (see §§ 100, 101). Thus if 8e and i are pronounced with equal stress,
8e is found to be the more prominent; it will be heard at a greater distance.
It is instructive to try the experiment, suggested by Prof. Lloyd James, of
pronouncing the word mechanically (normally miˈkænikəli) on a monotone
or in a whisper, and endeavouring to give artificial prominence to the first
and third syllables by means of stress, while keeping the qualities and normal
lengths of the vowels unchanged. It will be found that though the speaker
may so pronounce the sequence as to experience a subjective impression of
prominence of the first and third syllables, it is very difficult to convey this
as an objective impression to a hearer; the inherent sonority of 86 renders
the second syllable the most prominent (objectively) in spite of very strong
stresses that may be put on the adjoining syllables.

47020 This was first pointed out by Coleman (Intonation and Emphasis,
§§ 6-15).

47121 Intensity-emphasis cannot be given to all adjectives expressing measurable
qualities. The majority of adjectives can only have their meaning intensified
by prefixing adverbs like very, extremely, rather. Such are good, hot, long, thick,
frequent, spacious, troublesome, difficult. Some are capable of having their
meaning intensified by either method; such are wonderful, absurd. Students
have to learn which adjectives can be said with intensity-emphasis and
which have to be intensified by prefixing a qualifying word.

47222 This sentence might also be said with ordinary Tune 1
image, in which case it is a plain statement of fact
without any suggestion of reassuring the person addressed.

47323 Said as an enquiry, the intonation would be
image

47424 Intonation and Emvhasis, §§ 68-75.

47525 Ordinary pronunciation image ˈhau djuː ˈmæniʤ it?
The sentence may also be said with contrast-emphasis on manage, thus
image

47626 Said with a less degree of curiosity, this sentence would be pronounced
with ordinary Tune 1, thus image

47727 Also ˈḳkju, see footnote to § 909.

47828 Some English people now use this intonation on meeting. The reason
or this is obscure.

47929 See footnote 4 on p. 278 and footnote 27 on p. 316.

48030 Page 7. Published by Carl Winter, Heidelberg (new edition. 1956).

48131 Published by Teubner, Leipzig, 1909. Now out of print.

4821 As long as the system of transcription is one (like that of the I.P.A.)
in which stress ia shown by marks preceding the stressed syllable.

4832 I.e. it is said with the shortest ‘allochrone’ of the long chroneme. (For
‘chronemes’ see The Phoneme, Chap. XXIII).

4843 Or t + voiceless r, if it is held that English initial tr is a sequence of
two sounds.

4854 ˈbai-plein has a fully long ai and a voiceless l. The word pipe-line
ˈpaip-lain has a shorter ai and a fully voiced l.

4865 The first ai of ˈai-sait is fairly long. If there were a word ˈais-ait (with
the first syllable as in ice-axe ˈais-æks) the first ai would be shorter.

4876 ˈɔː-strʌk has a fully long ɔː.

4887 ˈbou-striŋ has a fully long ou, like ˈtou-stræp (§ 1092).

4898 ˈdɔː-pleit has a fully long ɔː and a voiceless l. If there were a word
ˈdɔːp-leit, it would have a shorter ɔː and a fully voiced l.

4909 ˈkiː-stoun has a fully long . If there were a word ˈkiːs-toun, it would
have a shorter .

49110 ˈhɔːs-trʌk has a rather short ɔː. Compare awe-struck (footnote 6).

49211 ˈleik-lænd has a rather short ei and a fully voiced l. If there were a
word ˈlei-klænd, it would have a fully long ei and a voiceless l after the k.

49312 ˈhiːt-weiv has a rather short and a fully voiced w. If there were a
ward ˈhiː-tweiv it would have a fully long ei and a partially voiceless w.

49413 ˈliːp-jəi has a rather short i: and a fully voiced j. If there were a word
ˈliː-pjəː, it would have a fully long and a partially voiceless j.

49514 Outrageous is generally pronounced autˈreiʤəs with a fully voiced r.

49615 ˈbout-reis has a short ou and a fully voiced r. If there were a word
ˈbou-treis, it would have a fully long ou and the voiceless affricate tr (or
t + voiceless r).

49716 ˈpek-snif has a strong s and a partially voiceless n. If there were a
word ˈpeks-nif, it would have a weak s and a fully voiced n.

49817 Undomesticated ˈʌn-dəˈmestikeitid has a rather long n, and the ə is
very short.

49918 ˈmins-miːt has a short n, and them of the syllable miːt is fully voiced.
If there were a word ˈmin-smiːt, the n would be longer and the second m
would be partially voiceless.

50019 ˈbæŋk-reit has a short ŋ and the r is fully voiced. If there were a
word ˈbæŋ-kreit, the ŋ would be long and the r voiceless.

50120 ˈlæmp-lait has a short m, and the l of the second syllable is fully voiced.
If there were a word ˈlæm-plait, the m would be long and the l following
the p would be voiceless.

50221 ˈliŋks-aid and ˈkæts-ai have a weak s (with as a variant, as explained
in footnote 9 to § 709, also The Phoneme, §§ 171-175).

503 Voir note 502.

50422 The medial k of ˈstʌmək-eik is unaspirated. Compare this word with
summer cake ˈsʌmə keik which has a longer a and an aspirated k at the
beginning of the final syllable.

50523 ˈhɛər-ɔil is said with a weak variety of r; it is the flapped r (§ 750) in
the speech of those who use this sound. If there were a word ˈhɛə-rɔil,
it would be said either with a fricative r or with a strong variety of frictionless
continuant.

50624 Under-masticated ˈʌndəˈmæstikeitid has a medium length n, and the ə
is of moderate length. Compare footnote 17.

50725 Here the ŋ is short and the w is fully voiced.

50826 Here the ŋ is long and the w of kwait is partially voiceless.

50927 Here the l is long and the ə of very short.

51028 Here the l is short and the ə of ˈweltə is of moderate length.

51129 In ˈliː-pleid the is fully long and the l in the second syllable is voiceless.
In ˈdiːp-leid the is rather short and the l is fully voiced.

51230 In əˈblæktˈai the t is unaspirated, but in əˈblækˈtai it is aspirated.
It is worthy of note that əˈblæktˈai is much nearer in sound to əˈblækˈdai
(a black dye) than to əˈblæktˈai.

513 Voir note 512.

51431 The first s in itˈslips is a strong variety, and the l is partially voiceless.

51532 The first s in itsˈlips is a weak variety, and the l is fully voiced.