The Organs of Speech.
§ 1. The current of air issuing from the lungs is the
essential element of all speech. It is the rough
material from which we obtain an endless variety
of sounds, the current of air being acted upon by
the articulations, which produce or modify sound.
We shall consider first the organs of breathing :
diaphragm, lungs and windpipe ; and then the organs of
articulation and of resonance — the ‘organs of speech’
in a narrower sense — larynx, mouth and nose.
Organs of Breathing.
§ 2. In ordinary breathing, the diaphragm — which
is convex if viewed from above (like the top of a
balloon, roughly) — is flattened by muscular action
and the ribs are expanded : the lungs have more
room and air is drawn in ; this is inhalation. Then
the diaphragm reverts to its natural form ; the space
which the lungs can occupy becomes smaller and the
air is forced out, after undergoing certain chemical
changes which do not concern us here ; this is
exhalation or expiration.
As a rule the breath is drawn in and expelled
through the nose, windpipe and the bronchial tubes1
The Organs of Speech
Plate III. in A. v. Luschka's ‘Der Schlundkopf des Menschen’ reproduced by permission of the Publishers, Messrs H. Laupp & Co., Tübingen.insert
which communicate with the left and right lungs
respectively ; the two processes take about the same
time in breathing.
By means of the laryngoscope we can look right down
the throat, and even see the branching off of the bronchial
tubes ; see the diagram on p. 6.
§ 3. In speaking the conditions are somewhat different.
As the exhaling of breath requires no
muscular action, but merely implies the return of
the diaphragm to what we may call its natural position,
it is better suited for the production of speech
than inhalation. Consequently certain muscles are
employed to hasten the process of drawing in the
breath and to prolong that of exhaling, as also to
regulate the pressure of air in each case, according as
the speech sounds to be uttered require it. The
breath passes almost entirely through the mouth.
It is not impossible to speak by means of inhaled breath,
and short words are sometimes produced in this way ; e.g.
no, or the first sound in oh dear. But the expiratory
stream of air is better adapted to speech, because speech
is meant to carry sound to a distance from us.
Organs of Articulation and Resonance.
§ 4. The upper part of the windpipe is called the
larynx. As it is of particular importance for the
production of speech, but cannot generally be examined
as easily as the organs of the mouth, a
somewhat detailed description will be necessary.2
§ 5. The relative position of the larynx is clear
from the plate opposite p. 1.
Its shape is due mainly to
the thyroid cartilage (or ‘shield cartilage’), known
also as the ‘Adam's Apple,’ No. 1 on the
diagrams (on pp. 3 and 5).
It consists of a shield, the two halves of
which are joined in front, forming an edge,
which can be felt through the skin, and which
is particularly noticeable in the case of men.
At the back these ‘halves’ do not meet, but
leave a space for
the cricoid cartilage (or ‘ring cartilage’), No. 4 on
the diagrams. This has roughly the shape of
3a signet ring, with the signet behind ; the front
half of it can easily be felt underneath the lower
edge of the ‘shield’ cartilage.
The two arytenoid cartilages (‘adjusting cartilages’)
have a very irregular shape ; 1 they are numbered
2 on the diagrams. With their base they rest
on the ‘signet’ of the ring cartilage ; and one
side of each is attached to one of
the two vocal chords. (The left diagram on p. 3 gives
the cartilages only ; the chords, numbered 3, appear
in the right diagram, and in that on p. 5).
These are horizontal membranes stretching from
the ‘signet’ of the ring cartilage (where each
is attached to one of the adjusting cartilages),
to the inner edge where the halves of the shield
cartilage meet. The vocal chords are connected
with the left and right walls of the larynx respectively.
The interval between the vocal chords is called
the glottis ; we may further distinguish that part
which lies between the adjusting cartilages as
the whispering (or cartilaginous) glottis, the rest
as the voice glottis.
§ 6. The following diagram is a transverse section
of the larynx, the line of section being directly in
front of that part of the adjusting cartilages which
is connected with the vocal chords.4
This clearly shows the narrowing of the larynx through
which the breath has to pass. 1,1 are the ‘half shields’ of
the thyroid, 4,4, sections of the ring cartilage ; 3,3′ are
the left and right vocal chords respectively. Immediately
above them there is a narrow recess (the ‘sacculus of
Morgagni’) and above it are the so-called ‘false vocal
chords,’ which play no part in the production of speech.
§ 7. The larynx can be closed above by a kind of
lid, the epiglottis (No. 5 on the diagrams, pp. 3 and 5) ;
this is regularly done in eating and drinking in order
to prevent the food from entering the windpipe.
§ 8. In ordinary breathing the epiglottis is raised
and the glottis is open, so as to leave a free passage
for the air. If we require an exceptional amount of
5air, say for blowing out a candle, the glottis is opened
to its fullest extent, as is shown in the following
Laryngoscopic view of the vocal chords opened to their widest extent,
showing the windpipe to its bifurcation.
§ 9. In speaking the larynx plays a far more important
In the case of some sounds it is passive, remaining
in the same position as for ordinary breathing.
Sometimes the glottis is narrow, 1 and the breath
brushes past the edges of the vocal chords, which
gives an [h] sound (see § 24). Or the chords are
pressed closely together, the air is stopped for a
moment and then bursts the closure. Then we have,
according to the amount of energy expended, a
‘glottal stop,’ ‘clearing the throat,’ or ‘coughing’
(see § 32). These sounds do not, however, occur
to any large extent in ordinary speech.
§ 10. Far the most important function of the vocal
6chords is the production of voice, which is an essential
element of the majority of sounds in ordinary speech.
When the vocal chords are brought near to one another
and suitably stretched the current of breath is
able to set them vibrating fast enough to produce a
musical sound, which is here termed ‘voice.’ Those
sounds which are produced by breath which has first
set the vocal chords vibrating, i.e. sounds with
‘voice,’ are called voiced ; and those in the production
of which this vibration plays no part are voiceless.
In whispering the voice glottis (see § 5) is closed,
and the air passes through the whispering glottis
only ; there is hardly any difference of pitch, and the
‘voiced’ sounds of ordinary speech are uttered without
Note 1. — It is well to gain a clear idea of what is meant
by ‘voice’ ; and it is not hard to do so, as the vibration of
the vocal chords may be felt in various ways. It is convenient
to use pairs of sounds like [v, f], [z, s] or [ʒ, ʃ], 1
for we can continue them for some time. If then we utter
the first of each pair, and place a finger on the throat, or
close our ears, or put our hand on the top of our head, we
shall notice quite distinctly that the chords are vibrating.
If we then utter the second sound of each pair we shall
miss the vibrations. The difference becomes very marked
if we alternate between the voiced and voiceless sounds,
saying, e.g. [zszszs], without making a break between
them. When this has been tried with ‘continuants,’
the vibration in the ease of voiced ‘stops,’ e.g. [ɡ, d, b],
will not escape notice ; and the absence of ‘voice’ will also
be recognised as the peculiar feature of whispered speech.
Note 2. — The explanation of the vocal chords given
7above has purposely been made in the simplest terms, and
may suffice to give some idea of their function. It is,
however, not without interest to see the way in which
various muscles act on the adjusting (arytenoid) cartilages,
so as to enable these to draw the vocal chords into
various positions suitable to the exigencies of speech, singing
and so on. The following diagrams are taken from
Professor Alex. Macalister's Textbook of Human Anatomy.
Opening of the glottis, showing the action
of the crico-arytenoideus positions muscle,
which draws the arytenoid cartilages from I,
I to II, II.
Closing of the voice glottis, showing the
action of the thyro-arytenoid muscles, drawing
the vocal chords from II, II to I, I.
Closure of voice glottis and breath glottis
showing the action of the arytenoideus proprius
muscle, drawing the arytenoid cartilages
from the neutral position I, I to the position
Note 3. — The whole range of the human voice is about
four octaves ; but in any individual it is rarely more than
from two to two and a half octaves.
The ‘tone of voice’ as a whole depends on the nature of
the vocal chords : when they are relatively long and thick
(as usually in men), the vibrations are slower and the voice
deeper ; when they are shorter and thinner (as in women
and children), the vibrations are more rapid and the voice
Differences of pitch in the same voice are obtained by
varying degrees of tension in the vocal chords, and of
force in the current of air. A sound may be of uniform
pitch and vary in loudness ; such variation is due to an
increase or decrease in the force of expiration.
§ 11. In discussing the organs of the mouth, it
will not be necessary to give detailed descriptions ;
for the inside of the mouth can be easily examined.
It is surprising how few people have ever looked well
into their mouth. The student of phonetics should always
have a hand-glass by his side, and should use it constantly,
until he is quite familiar with the appearance of the ‘oral’
organs of speech. He will very soon find out the best
angle at which to hold it, so that the mirror may at once
reflect light into the mouth and enable him to watch it.
§ 12. The pharynx is the hollow space above the
larynx, separated from it by the epiglottis
(see § 7). It is at the same time the back
part of the cavern of the mouth, from the rest
of which it is separated (during nasal breathing)
the soft palate or velum, the end of which is
the uvula, which can be clearly seen by means of a
hand-glass. The soft palate passes over into
9the hard palate ; the difference can be easily noticed
by letting the tip of the finger travel along
the roof of the mouth from back to front.
§ 13. In breathing the velum hangs down as a rule,
leaving a free passage for the air on its way through
nose, pharynx, larynx, windpipe, bronchial tubes and
§ 14. In speaking, on the other hand, the velum in
most sounds closes the nasal passage completely by
pressing against the back surface of the pharynx ;
the breath can then pass through the mouth only.
In some sounds, however, the velum is lowered,
and breath passes through the nose as well as the
mouth (“nasal vowels”) ; or the passage through
the mouth is closed, and the breath can pass
through the nose only (“nasals”).
It is easy to determine whether a sound is oral or nasal,
or both. A flat ruler is placed with one edge on the upper
lip, with the other against cold glass. The breath as it
issues will dim the glass ; and therefore if the sound is
nasal, the glass will be dimmed above the ruler. To
ascertain the relative amount of oral and nasal breath in
a vowel we require delicate instruments.
§ 15. When the current of air has caused the vocal
chords to vibrate, and then passes out through the
mouth without encountering any obstruction or check,
the result is a pure or oral vowel ; if the breath has
not produced voice before passing through the mouth,
we obtain the corresponding [h] sound (see §§ 26, 75).10
If in the production of a sound the soft palate has
not closed the passage through the nose, and some
of the breath is allowed to escape that way, the
result is a nasal vowel.
§ 16. The nature of the vowel varies according to
the shape of the oral passage (the passage through
the nose being invariable, except for the space above
the soft palate). The inside of the mouth assumes
different shapes, owing to changes of
the tongue : this is quite the most important instrument
in modifying the current of breath, and
thus causing variety in vowels. It is capable
of assuming many shapes ; some may be seen
in the transverse sections on pp. 28, 29. We
distinguish the point of the tongue, the blade
(above and behind the point when the tongue
lies flat), the front (yet farther behind), and the
back ; also the ridge or dorsum (an imaginary line
drawn along the middle of the top of the tongue
from end to end), and the rims (running down
both sides of the tongue when it lies flat).
The lips also vary in shape, and thus modify the
quality of vowels. As may be seen from the
diagrams opposite p. 27, the corners of the mouth
can be drawn far back, the lips may be left
as in breathing through the mouth, or they
may be rounded till only a very small opening
is left ; there are obviously numberless intermediate
stages between these.
The lower jaw is lowered to various degrees ; usually
this coincides with a change of tongue position.
11The lowering (or the angle of the jaws,
or the distance between the rows of teeth)
is generally greatest when the tongue is flat,
and smallest when some part of the tongue is
raised to its utmost.
In ordinary breathing the rows of teeth and the lips are
closed, and the tongue fills almost the whole of the mouth.
§ 17. The current of air, with or without ‘voice,’
may not have a free passage : it may have to make its
way through a narrowing, or to force its way through
an absolute stoppage. Sounds thus produced are
called (voiced or voiceless) consonants.
There may be merely a narrowing. Then the breath
brushes past ; there is friction ; we may continue
the sound as long as our breath lasts. The result
is a (voiced or voiceless) continuant (or fricative).
The passage may be completely closed. The
breath is stopped for a moment or for a little
time (applosion) ; but then it bursts through the
obstacle with a little explosion. The result is a
(voiced or voiceless) stop (or plosive or explosive).
When the breath issues through the nose, it is a
(voiced or voiceless) nasal.
In a final stop the applosion is the chief sound.
§ 18. This narrowing or closure may take place in
many parts of the mouth. 1 :
between the back of the tongue and the soft palate,
back (velar) continuants, stops, nasals ;
between the front of the tongue and the hard palate,
front (palatal) continuants, stops, nasals ;
between the point (or blade) of the tongue and the
12upper teeth, or the ridge (called ‘alveoles’)
above the upper teeth,
dental (point, lingual) continuants, stops, nasals ;
between the lower lip and upper teeth,
labiodental (lip-teeth) continuants, stops ;
between the lips,
bilabial (lip-lip) continuants, stops, nasals.
§ 19. Little remains to be said about
It has been pointed out that in breathing the velum
is lowered so as to leave a free passage for the breath.
In speaking the nose plays a subordinate part, as it
serves only to give a passage to breath in the nasal
vowels and the nasals.
The resonances of the hollows of the nose are
nearly invariable, as the shape of these hollows is
subject to little change.
In English speech, and generally in singing,
most vowels have some nasal resonance, due to
the passage through the nose not being firmly
closed by the velum. In speech it is not usually
noticed, when it is slight ; when it is marked,
we call it a ‘nasal twang.’ Singers use it
because the additional resonance adds to the
sonority of the voice ; but such voice production
§ 20. A few common misconceptions may here be
No sound is produced ‘in the chest’ or ‘in the
head’ : a man may speak as though his voice came
from his boots, and a ‘ventriloquist’ may pretend to
talk ventrally, but no sound is produced lower than
the ‘Adam's apple.’
We say ‘he is speaking through his nose’ when a
person has a bad cold. The exact opposite is the
case, for the passage through the nose is obstructed
and the breath cannot make its way through it. The
most obvious result is the substitution of ɡ, d, b
for ng, n, m.
It is sometimes maintained that children have an
innate aptitude for learning the language of their
country : but a child will learn the language of the
people who take care of it, even though it be quite
different from that of its parents, and will learn it
just as well as a native child.
The pronunciation of a foreign language cannot be
learnt by means of an ‘imitated pronunciation,’ such
as is supplied in numerous Elementary Grammars,
First Courses, etc. It is impossible to represent the
sounds of one language by means of symbols which
are used in another language to designate sounds only
approximately like them. Thus foh appears in one
of these books for the French faut, and clay for clef ;
14in both cases the reader who has no further help will
pronounce diphthongs in place of the simple long
vowels, and in the second word a cl with c so far
forward as almost to become a t, and certainly with
an English l.15
The Separate Sounds.
Designation of Sounds.
§ 21. To avoid the necessity of constantly repeated
descriptions, we must be able to express each sound
by a clear and unmistakable symbol ; we require a
phonetic alphabet. The current spelling of English,
French and German will not serve our purpose, for
it is inconsistent (e.g. English not and what, French
si and ici, German vor and für), or awkward and
inadequate (e.g. English thin, French mon, German
In this book a phonetic transcript [in square
brackets] is therefore employed, and it is the alphabet
of the International Phonetic Association (the
journal of which is Le Maître Phonétique, edited by
Dr Paul Passy, 20 rue de la Madeleine, Bourg-la-Reine).
The following table contains the symbols
for such sounds as we shall meet with in English,
French and German speech ; the value of those
symbols which are not familiar will become clear as
Clarendon type is employed to represent. letters
and words in the conventional spelling.17
Table of Sounds.
lips | lip and teeth | point of tongue | consonants | with closure | stops | nasals | with narrowing | liquids | lateral | trilled | continuants | vowels | tongue highest | tongue half-way | tongue lowest
̥ (e.g. in m̥, ὴ) means ‘voiceless’ ; French prisme [prism̥].
ː indicates length ; English half [hɑːf], French inné [inːe].
̆ over a symbol means ‘not syllabic’ ; English now [naŭ].
′ (preceding) indicates accent (stress) ; English balloon [bəʹluːwn].18
front of tongue | back of tongue | uvula | glottis
nasal vowels : ã ɛ̃ ɔ̃ œ̃ ; thus French lent [lã], fin [fɛ̃],
long [lɔ̃], un [œ̃].
() indicates lip rounding.
˕ after a vowel means ‘more open.’
˔ after a vowel means ‘more close.’19
Articulation of the Larynx.
1. Full opening of glottis.
§ 22. When the glottis (the interval between the
vocal chords) is fully opened, the breath passes
through without producing any audible sound.
Sound may, however, be subsequently produced
when the breath reaches the articulations of the
mouth, and the nose.
2. Slight opening of glottis and
closure of glottis.
§ 23. Notice the production of ‘voice,’ described
above in § 10 ; it is not a speech-sound of itself, but
an important part of many such sounds, which are
hence called voiced ; the rest are voiceless.
Slight opening of glottis.
§ 24. If the glottis be made sufficiently narrow, the
breath brushing past the edges of the vocal chords
produces a voiceless glottal continuant (or fricative),
which we must not at once identify with the Greek
‘spiritus asper’ or our own h The phonetic symbol
There are many possible varieties of it : the current
of breath may vary in force, the glottis may vary
in narrowness ; or, again, the current of breath may
remain of the same force throughout, or its force
distinctly increase or decrease ; or the glottis may
remain the same throughout, or become narrower
and narrower ; and so on. When the current is very
considerable and the glottis very narrow, we have
what is known as ‘wheezing.’
§ 25. It is possible to produce a voiced [h] sound :
the voice-glottis (see § 5) is allowed to vibrate, and
the cartilaginous glottis is opened as for whispering.
Thus ‘voice’ and the [h] sound are formed at the
same time. Symbol [h̬].
§ 26. Apparently the glottal continuant [h] does
not occur in English, French or German in ordinary
speech : as a rule the various sounds which are
written h start with the glottis open, and it does not
become narrow until the production of the following
vowel. The sound of this h results from the breath
passing through the mouth when its articulations
have already ‘got into position’ for the vowel that
is going to be uttered. Indeed, this h. may be regarded
as unvoiced [ɑ], [e], etc., according as the
following vowel is [ɑ], [e], etc.
§ 27. As, however, h is felt to be a consonant of
uniform value, and as indeed in emphatic diction the
continuant [h] is actually used (in place of, or together
with, the sounds just mentioned), we shall
proceed to discuss [h] here.21
§ 28. [h] varies according to the nature of the
current of breath.
If the current begins with full force for the [h],
and passes undiminished to the vowel, we have the
simple [h] ; this is its common form in German.
If the current distinctly diminishes before the
vowel is sounded, we obtain the usual English form
of [h], more accurately [h>], or better still [<h>].
If on the other hand the current is weak at first,
and does not reach the maximum of force until the
sounding of the vowel, the resulting ‘soft breath’
no longer suggests a distinct [h] ; it may then precede
an initial vowel without being felt as a
separate sound. We may designate it [h<] or
§ 29. English [h] (more accurately [h>] or
[<h>]) : hold [hoŭld] ([h>oŭld] [<h>oŭld]),
occurs only at the beginning of words ; and before
It is lost in enclitics when these are unaccented
(‘weak forms’), except at the beginning of a sentence.
weak | strong | had | has | have | he | her | him | his | [əd hæd] | [əz hæz] | [əv hæv] | [ij hij] | [ə həː] | [im him] | [iz hiz]22
Written h is mute in heir [ɛə], honest [ɔnist], honour
[ɔnə], hour [aŭə], and their derivatives ; also in John
Some say a history, but an historical novel,
a habit, but an habitual action.
A ‘soft breath’ (see § 28) usually precedes the
initial vowel of a word : old [oŭld], more accurately
§ 30. French in regular speech has no [h]. The
written h, whether it be ‘h muette’ or ‘h aspirée,’
is in every case mute ; e.g. héros [ero] and héroïne
§ 31. For German [h] compare § 28. It occurs
only at the beginning of words : herab [hɛʹrɑp], hinein
[hiʹnaĭn] ; elsewhere only before vowels with chief or
secondary accent (‘Hauptton’ or ‘Nebenton’) : Halt
[hɑlt], Anhalt [ʹʔɑnhɑlt] or [″ʔɑnʹhɑlt].
[h] precedes a vowel with secondary accent almost
exclusively in compound words ; particularly
compounds with -heit [hɑĭt], preceded by a consonant,
as Weisheit [ʹvɑĭshɑĭt] or by a vowel, as
Roheit [ʹroːhɑĭt]. Also in Uhu [ʹʔuːhuː], Schuhu
[ʹʃuːhuː] ; and in foreign words, as Alkohol [ʹʔɑlkohɔl],
In all other cases written h is mute, whether it be
etymologically justified, as in sehen [zeːən], sieh [ziː] ;
or a sign of lengthening, as in gehen [geːən], geh [geː].23
Closure of glottis.
§ 32. If the current of breath is stopped by the
glottis being firmly closed, and if it suddenly bursts
through this obstacle, a voiceless stop is produced,
generally called the glottal stop. It is sometimes falsely
called ‘spiritus lenis.’ Phonetic symbol [ʔ].
When the production of [ʔ] is energetic, we have what is
known as ‘clearing the throat’ ; and ‘coughing,’ when it is
spasmodic and involuntary. In these cases, there is usually
§ 33. In English and French the sound hardly
§ 34. In German the glottal stop [ʔ] regularly precedes
the vowel at the beginning of a word, or of the
second part of a compound, when the compound is
still felt as such.
ein [ʔɑĭn], Verein [fɛrʹʔɑĭn] or [fərʹʔɑĭn], but
The presence of this sound is particularly noticeable
There is no symbol for it in the current spelling ;
but as a general rule it occurs before initial vowels.
There are very few exceptions : they comprise words
compounded with particles, as herein [hɛʹrɑĭn], and
foreign compounds, as Adept [ʔɑʹdɛpt].
Note. — It is necessary to insist again and again on
the presence of the glottal stop, in the case of
English pupils learning German. Germans complain
that English people who know German well are often
hard to understand, because they ‘run their words
together,’ i.e. do not make use of the glottal stop.24
[List of Contents]
The diagrams on the following pages illustrate the
section on vowels. That on page 27 represents the ‘vowel
triangle,’ and shows to what extent the lips and the tongue
are used in the production of the principal vowels ; it is a
modification of the diagram in Prof. Viëtor's ‘Kleine
Phonetik.’ The photographs on the opposite page are
new ; they show the position of the lips in uttering normal
[i, e, ɑ, o, u] ; in English the lips are not usually
rounded or retracted quite so much. The diagrams on
pp. 28, 29 are taken from an article by Dr R. J. Lloyd
in ‘Neuere Sprachen’ ; they have been inserted in order
to show the gradual raising of the tongue along the
[a]-[i] and [ɑ]-[u] lines.insert
Vowel Triangle : Lip Positions26
Opening of Lips | smallest | greater | greatest | small | slit | round | hard palate | soft palate | tongue raised | most | least | teeth | front vowels | back vowels | uvula | angle of jaws27
Note. — These diagrams represent
the northern English pronunciation
of Dr Lloyd. In
southern English the vowel of
pat is [æ], i.e. the front of the
tongue is a little higher, and the
tip a little farther from the lower
Articulation of the Mouth.
Free opening of the passages.
Voiced sounds : Vowels.
§ 35. Various tones of ‘voice’ (resulting from varying
vibrations of the vocal chords) are modified by
the articulation of the mouth, which produces certain
resonances within the cavity of the mouth.
The musical pitch of the various resonances can most easily
be observed in whispering, for then there is no ‘voice’ blending
with the resonance of the mouth.
§ 36. The resonance of the mouth varies according to
(a) the position of the tongue, which may be defined
relatively to two places of articulation :
the front or hard palate ; this is the front or
palatal articulation of tongue and palate ;
the back or soft palate ; this is the back or
velar articulation of tongue and palate ;
there is also the intermediate or medio-palatal
articulation, when the middle of the tongue
is slightly raised.
In normal or ‘dorsal’ vowels only the upper surface (dorsum)
of the tongue is raised, and the lower surface remains always
upon the floor of the mouth.
(b) The position of the lips ; these may be :
drawn back at the corners ;
protruded or rounded ;
neutral (as in breathing through the mouth).
(c) The position of the soft palate.
In the case of back vowels there is generally also
lip rounding ; and the higher the tongue is raised
the greater is the lip rounding (i.e. the opening
30becomes very small). If, however, there is no lip
rounding, this may to some extent be compensated
for by the tongue being drawn further back. The
raising of the tongue in the case of front vowels
is similarly accompanied by a drawing back of the
corners of the mouth.
The position of the lower jaw (angle of the jaws,
distance between the rows of teeth) varies, as a rule,
according to the raising of the tongue.
§ 37. We can raise (1) the front or (2) the back of
the tongue, until the interval between it and (1) the
front or (2) the back palate is very small indeed ; so
small that if they were brought any closer, the air
would brush past and produce no longer a vowel
but a continuant consonant. If we just avoid this we
shall obtain (1) the front (or palatal) vowel ‘close’
[i] (diagram 1 on p. 28), and (2) the back (or
guttural) vowel ‘close’ [u] (diagram 8 on p. 29).
At the same time the part of the tongue not in action
will be lowered, i.e. in [i] the back and in [u] the
front ; and the lips will be more or less drawn back
for [i] and rounded for [u]. We have therefore in
[i] a tubular channel with slit opening, and in [u] a
hollow in the front part of the mouth with round
opening ; and so the pitch of [i] is high, and of [u]
§ 38. When the middle of the tongue is slightly
raised we obtain the resonances for [ɑ] sounds. In
the case of ‘pure’ [ɑ] (see diagram 5 on p. 29),
the raising of the tongue is slightly nearer the soft
palate than the hard palate.31
§ 39. Between ‘close’ [i] and [ɑ] there are the
resonances of :
‘open’ i [ɪ] sounds,
‘close’ and ‘middle’ e [e, e] sounds,
‘open’ e [ɛ, æ] and ‘clear’ a [a] sounds.
Between ‘close’ [u] and [ɑ] there are the resonances
‘open’ u [U] sounds,
‘close’ and ‘middle’ o [o, o] sounds,
‘open’ o [ɔ] sounds.
§ 40. Tense (or primary) and lax vowels. In producing
a vowel, we may adjust the articulations so
favourably that the resulting sound is clear and
decided (a ‘primary’ vowel). The vowel will be
less satisfactory, if we adjust the articulations in a
lax way ; in producing the front unrounded vowels
the neglect to arch the tongue towards the point of
greatest constriction, and in producing the back
vowels the neglect to round the inner cavity by
sufficient withdrawal of the tongue-tip, are typical
cases of lax articulation. In English the accented
long vowels are tense, but not as tense as in French,
or even German ; the accented short vowels are lax.
In French, all accented vowels are tense. In
German, the accented long vowels are more tense
than in English, not as tense as in French ; and the
accented short vowels are lax.
For tense and lax the terms ‘narrow’ and ‘wide’ are often
§ 41. If the tongue position of front vowels [i, e] is
combined with the lip rounding of back vowels [u, o],
composite sounds result :
32‘close’ [y], ‘open’ [Y],
‘close’ [ø], ‘open’ [œ].
Similar, but by no means identical, sounds
can be produced by a rounding of the tongue ;
by raising the middle of the tongue rather more
than for [ɑ] ; or by combining front (e.g. [i])
articulation with back (e.g. [u]) articulation, the
middle of the tongue being lowered.
Pure or Oral Vowels
(Without Nasal Resonance).
The u sounds :
close [u], open [U].
§ 42. For [u] the place of articulation is farther
back than in the case of any other vowel. The back
of the tongue is raised quite close to the soft palate
and the front of the tongue is drawn down and back.
In the front of the mouth there is consequently a
fairly large place of resonance, the effect of which is
usually increased by the protrusion of the lips and
the formation of a round opening ; the rows of teeth
being fairly close.
The resonance of [u] has the lowest pitch of all
The symbol [u] is generally used to designate both
the close sound and the open sound (strictly [U]).
§ 43. In English [u] there is usually slight lip
33rounding. There is no ‘close’ [u], generally speaking.
The following are the usual [u] sounds :
The long [u] sound in too (No. 8 on p. 29 and
here) begins with open or middle [u], but towards the
end of the sound the articulation becomes more
tense, and close [u] or consonantal [w] results ; [tUu]
As in the case of other long sounds, the long
[u] is shortened before voiceless consonants,
without, however, losing its diphthongal character :
Before the ‘neutral’ vowel [ə] the diphthong
loses its [w], and the [uː] is somewhat shortened :
cruel [kruːəl] ; the tongue is somewhat lower
before r = [ə] : poor [pùːə], almost [poːə]. (Even
[pɔː] is often heard in London.)
The same sound occurs in the so-called ‘long u,’
preceded by [j], which after voiceless sounds often
passes into the voiceless continuant [ç] or even [ʃ] ;
as in new [njuw], tune [tjuwn] or [tçuwn].
In syllables with slight accent the treatment
of [juw] is like that of [uw] ; in unaccented
syllables we have [jü, jy, jə] : value [vælju] or
[væljü], regular [regjulə], or with [—jü—] or
The short [u] sound in book is open : [buk], more
Before a voiced final consonant the accented
[u], as well as the consonant, is often pronounced
half-long : pull [pul].
Unaccented it becomes [u, ü, ə], or is
dropped : useful [juwsful, —fül, —fəl, —fl 1] ;
should [ʃud, ʃəd, ʃd].
§ 44. In French there is only close [u], and the
lips are much more rounded (the opening is smaller)
than in English or even in German close [u], so that
it is usually accompanied by a brushing of air past
the lips, which sounds somewhat like subdued
whistling. Note the inner rounding of the cavity
by retraction of the tongue-tip.35
Long [u], e.g. in rouge [ruːʒ] ; short [u] in route
At the end of words the u sound, or at least
the end of it, is often uttered without ‘voice,’
but with whisper.
§ 45. In German there are two u sounds which
differ in quality and quantity.
Long [u] in du is close, i.e. the tongue is raised as
far as it can be raised without letting the current of
breath rub past and produce a continuant. The lips
are usually rounded, often very slightly ; du [duː].
In syllables before or after the chief accent
the sound is often shortened : Sudeten [zuʹdeːtən],
Short [u] in und is rather more open, i.e. the raising
of the tongue is not quite so great and not quite so
far back, the articulation is less decided, the lip
rounding weaker ; und [ʔunt], strictly [ʔUnt].
This short [u] is also the first element of the
rare diphthong ui [uĭ] in pfui [pfuĭ].
We may also describe the second element
of the diphthong au [ɑŭ] as a weak open [u],
although an [o] sound is sometimes substituted
The o sounds :
close [o], open [ɔ].
§ 46. These are intermediate between the u sounds
and the a sounds, and there is no clear dividing line
36in either direction. It will suffice to distinguish a
middle o [o], half way between u and a as far as
articulation of tongue and lips is concerned, a
close o [o] — as near again to u as it is to a, —
and an open o [ɔ] — as near again to a as it is
to u. Thus :
close u | open U | close o | middle o | open ɔ | ɑ
There are o sounds in which the raising of the
tongue is more forward or backward than the regular
line joining u and a.
Sounds which closely resemble a very open [ɔ] are
formed by lowering the back of the tongue to the [ɑ]
position, and even lower, and at the same time drawing
it back. It is convenient to treat these sounds
here. They are also designated [ɔ].
§ 47. In English there are three qualities of o in
accented syllables : the two kinds of ‘long o’ in no
and in lord, and the short sound in not.37
The so-called ‘long o’ in no is rather a diphthong,
the first element of which is usually ‘middle’ o [o],
though closer o also occurs. As a rule it is long [oː] ; it
is shortened before voiceless sounds and then becomes
half long. The second element may be regarded as
unstressed [ŭ] ; but the tongue retains its [o] position,
while the lips gradually reach the [u] rounding
Thus : no [no(ː)ŭ], node [no(ː)ŭd], but note [noŭt]
(No. 7 on page 29 and here).
In syllables with weak stress the first part
of the diphthong is a ‘mixed’ sound [ö] 1, and
the second part is usually dropped : fellow
[felöŭ, felö], hotel [höŭʹtɛl, höʹtɛl] (often with
silent h). In affected speech this also occurs in
stressed syllables ; oh [öŭ].38
When the stress is still weaker we have [ö]
becoming [ə] : obey [öʹbeĭ, əʹbeĭ], innocence
The other ‘long o,’ in lord [loːd], is a ‘narrow’
[ɔ]-sound : the back of the tongue articulating nearer
to the velum (see § 12) : the lip rounding is weak. It
is identical with a in all (No. 6 on page 29 and here),
war, or with au, aw in laud, law. This is in standard
speech the only sound of stressed o before r, if it is
spoken as an o sound at all, and if it does not precede
rr as in sorry, or r followed by a vowel as in
authority. Before voiceless consonants the sound is
reduced to half-long, as in short.
Strictly speaking this [ɔː] is not a simple long
vowel, but is followed by a faint [ə]-sound,
which must not be regarded as representing
39the r, as it occurs also where no r follows. —
Only when r is final does it make its presence
felt by strengthening the [ə], which then almost
gives the impression of a full syllable ; compare
law with lore. When a vowel follows, as in
glory, the r is pronounced as at the beginning
When quite unstressed, this sound often
passes into [ɔ], and then also into the same
‘mixed’ sound resembling [ö] as [oŭ] described
above : authority [ɔʹθɔriti, öʹθɔriti 1] ; or stressed
[ɔ:], unstressed [ö, ə].
The usual short o [ɔ] in not [nɔt] is ‘lax,’ as
compared with long [ɔ:] ; the a in what [wɔt, ʍɔt] is
identical with it. Before voiced final consonants the
[ɔ] becomes half-long, as in dog [dɔg]. Before ss [s],
st [st], th [θ], ff [f], long [ɔ:] is frequently substituted
for the short sound, e.g. in cross, lost, cloth, off.
When the stress is weak, this short [ɔ] also
often passes into ‘mixed’ [ö] : October [ɔk-,
ökʹtoŭbə] ; when still weaker, we even have [ə] :
contain [kön-, kənʹteĭn], author [ɔ:θə]. Loss of
vowel, e.g. in lesson [lesṇ]. 2
An open o sound (perhaps = [ɔ:], which is confirmed
by Miss Soames and by Storm, while Sweet
identifies it with [o] of [oŭ]), is also the first part
of the diphthong oi, oy in oil, boy ; the second
40part is [ǐ] (more strictly a sound half-way between
[i] and [e] = [i˕]). Phonetic symbol [ɔĭ] ;
hence [ɔĭl, bɔĭ].
§ 48. In French the o sounds are partly close and
partly open. Close o is long, e.g. in rose [roːz],
short in côté [kote]. Open o is long in or [oːr], short
in robe [rɔb].
The close sound (as we noted in French [u]) is
more forward, and the lip rounding is more energetic,
and the muscles are tenser than in the first part of
English [oŭ] or in German [o]. The open sound is
also tense, being distinguished from the close sound,
not by the muscular action being relaxed, but by the
tongue being lower.
Particularly the short [ɔ] shows a tendency to
pass over into the ‘mixed’ [ö] sound, more
especially before [n] and [m], and in the word
§ 49. In German we have a similar distinction of
quantity and quality, as in the case of the u sounds
(see § 45) ; the long o is close [o], the short o varies
between middle and open [ɔ]. Thus so [zoː], but
Except when it is final, [oː] is shortened to [o]
(not to [ɔ]) if it precedes or follows the chief
accent. Thus somit [zoʹmit], but also [ʔɑlzoː].
Short [ɔ] is also the first part of the diphthong
eu, äu (oi) in the speech of the stage ; the
second part is open [ĭ] or [y̆], often a sound yet
more open. We shall express it by [ɔĭ].41
Intermediate or medio-palatal vowels.
The a sounds :
neutral [ɑ], clear [a].
§ 50. The a sounds occupy a position between the
back and the front vowels.
We regard that a as ‘pure’ or ‘neutral’ in
which the tongue position coincides with the intersection
of the [u]-[ɑ] line and the [i]-[ɑ] line (see
the diagram on p. 27), and the sound of which is
equally untinged by [o] and by [e].
The articulation of [ɑ] consists of a very slight
raising of the middle of the tongue ridge ; as a rule
the front of the mouth is opened farther than in the
case of other vowels.
Owing to this ‘middle’ articulation there is much
scope for variety of a sounds. When the tongue is
raised a little farther back we obtain lower, darker
sounds resembling [ɔ] ; when it is raised a little
farther forward we obtain higher, clearer sounds
resembling [æ] or [ɛ]. ‘Clear’ a will be designated
[a], where the distinction is important.
§ 51. In English standard speech there is the long
[ɑː] in father [fɑːðə] (No. 5 on p. 29 and opposite) ;
and also a short sound, which is very close to neutral
[ɑ], and is here designated [ʌ], as in nut [nʌt], etc.
The long [ɑː] is not always a simple sound ; it
may be a half-long [ɑː] together with a faint ‘mixed’
sound, which is perhaps ‘coronal’ (i.e. containing
42[r], see § 70) and certainly differs very slightly from
[ɑː]. In the case of short [ʌ] the back of the tongue
is raised a little more, and there may be a raising of
the front of the tongue as well, as in mixed vowels
(see § 69) ; thus it comes that — in addition to
the sound like [ɑ] — there are varieties resembling
[ə] or even [æ].
Unaccented [ʌ] passes into [ə] : but, accented
[bʌt], unaccented [bət].
The first elements of the diphthongs in high
and how may be regarded as a sounds tending
to ‘mixed’ and even palatal [æ] articulation.
The former may be designated [aĭ] (the ĭ lying
between I and e) ; the latter [aŭ] (the ŭ lying
between u and o).43
§ 52. In French there is (1) ‘neutral’ [ɑ], long
in âme [aːm]. short in pas [pɑ] ; (2) palatal or clear
[a], long in rage [raːʒ], short in ma [ma].
But as in the case of other French vowels, the
difference in quantity between the long and the
short sounds is not as marked as in English or
Both a sounds may combine with a preceding
[w] ; we have [wɑː] in croire, [wɑ] in crois,
[waː] in boire, and [wa] in bois (verb). These
diphthongs are usually written oi.
§ 53. In German both long and short ɑ, [ɑː] and
[ɑ], as in Vater [fɑːtər], was [vɑs] may on the whole
be regarded as ‘pure.’ It is true that in the North
particularly the short sound is often rather ‘clear’
(see above, § 50), while in the greater part of
Middle and South Germany both long and short
[ɑ] tend towards [ɔ]. On the stage the tendency
is rather towards [a] than [ɔ]. There is the same
distinction of the long vowel with ‘tense’ formation,
and the short vowel with ‘lax’ formation, as in the
case of the other German vowels.
In unaccented syllables [ɑː] is often shortened
to [ɑ] : Datum [dɑːtum], datieren [dɑʹtiːrən].
The short [ɑ] is also the first part of the diphthongs
ei or ai [ɑĭ] and au [ɑŭ].44
Front vowels, without lip rounding.
The e sounds
(including æ sounds).
Close [e], open [ɛ].
§ 54. The e sounds occupy a similar position
between [ɑ] and [i] as the o sounds between [ɑ]
and [u]. However, the [ɑ]-[i] line representing
the palatal articulations is longer than the [ɑ]-[u]
line representing the guttural articulations, as is
indicated by the diagram on page 27 ; [e] therefore
is farther from [ɑ] than [o] is. 1
This doubtless explains the richer development
of the e sounds in English and German, as
well as the existence of the symbols æ, ä beside
e, as compared with the one symbol o.
The articulating part of the tongue is the front and
middle of the ridge or dorsum. The opening of the
lips and the angle of the jaws are smaller for e
sounds than for [ɑ], and become yet smaller the more
the sound departs from [ɑ] and approaches [i]. It
is not by any means a rule that the corners of the
mouth must be drawn back.
As in the o sounds, we distinguish a close [e] and
an open [ɛ] beside middle [e] ; when the sound is very
open, and indeed closely resembles ‘clear’ [a], it is
The [i]-[ɑ] line will then be as follows :
close i | open ɪ | close e | middle e | open ɛ | more open æ | ‘clear’ a | ɑ
§ 55. In English the e sounds run parallel to the
o sounds, but we note more shades of quality in the
To the very open o sounds correspond :
(i) the long, or commonly half-long, [ɛː] which
lies between [ɑ] and [e] and has ‘tense’ formation;
it is always followed by [ə], which is very
faint before r : there [ðɛːə] (No. 3 on page 28
and opposite), Mary [meː(ə)ri].
There is also a less open pronunciation of the
e sound in these words : [ðɛːə, meː(ə)ri].
(ii) the short [æ], a sound more open than the
one just described, and of ‘lax’ formation :
pat [pæt] (No. 4 on page 28, and opposite, and
see the note on p, 28).46
On the other hand there is no o sound corresponding
(iii) short [e], which (in standard speech) is only
middle [e] or even close : let [let, let].
See note on p. 2847
In unaccented syllables [æ] and [e] pass into [ə] :
strong | weak | that | them | [ðæt ðət] | [ðem ðəm]
Parallel to the diphthongal [oːŭ] in no is [eːĭ] in
pale [peːĭl] (No. 2 on page 28, and here), the second
element of which is half-way between [i] and [e].
Before voiceless sounds the [eː] (like the [0ː] of [oːŭ])
becomes half-long : late [leĭt].
§ 56. In French the open and close e sounds (all
‘tense’) are clearly distinguished. The long and
short open e is usually a full [ɛ(ː)] ; rêve [rɛːv], paix
In unaccented syllables the sound is shortened
(it becomes half-long, indeed almost short)
The close e sound is only found short or half-long :
The combinations of e sounds with a preceding
[j], as in bière [bjɛːr], pied [pje] are instances
of what are sometimes called ‘ascending’
diphthongs (cp. [wɑ, wa] on p. 44) ; see § 160.
§ 57. In German the open sound occurs long [ɛː]
and short (usually only half-open) : Bär [bɛːr], fett
[fɛt]. The long (accented) sound is ‘tense’ ; the
short (accented) sound is ‘lax.’
The close sound is long only : fehlen [feːlən].
It is however more or less shortened when
unaccented, as is [ɛː] in foreign words : Theater
[te(ː)ʹɑːtər], plaidieren [plɛ(ː)ʹdiːrən].49
The i sounds :
close [i], open [ɪ].
§ 58. The i sounds stand in the same relation to
the u sounds as the e sounds do to the o sounds.
In general, the articulation consists in bringing the
middle of the tongue ridge close to the middle of the
hard palate. The drawing back of the corners of the
mouth is not essential, though it may be done in
order to add to the clearness of the sound. If the
tongue is raised so high that the slightest further
raising would produce a continuant (compare close
[u], § 37), we obtain ‘close’ i [i] ; if it be lowered —
along the [i]-[ɑ] line — the sound becomes ‘open’ [ɪ]
and approaches [e].
§ 59. In English there is close parallelism between
u sounds (§ 43) and i sounds.
[uːw] | [iːj] | (before [ə]=r) | [uː] | [iː] | [U] | [ɪ]
The diphthongal [iːj] or [ɪːi], as in me [miːj] (see
50diagram 1 on p. 28 and here), begins with half-open
i ; the tongue is then raised so that the second
element is close [i] or consonantal [j]. The first
element is shortened before voiceless sounds : meet
[mijt]. The long [iː] (strictly half-long) is open ; it
occurs only before [ə] = r: fear [fira].
The short i is also open : fit [fit], strictly [fɪt].
Unaccented short i is still more open ; there are
many shades of it, at the beginning of words and
within them, which cannot be discussed here.
When final, as in very [vere˔, veri], the tongue
is lowered so much that the sound approaches
close or even middle [e]. The symbol for these
unaccented i sounds is [e˔], or roughly [i].51
§ 60. In French the i sounds are close (compare
French u sounds, § 44) and ‘tense.’ The front
of the tongue ridge articulates near the hard palate ;
the corners of the mouth are drawn back. The
sound may be long, as in rive [ri:v], or short, as in
vif [vif], triste [trist]. Sometimes the sound is
‘half-long’ ; but in French the difference between
this and ‘short’ is very slight.
At the end of words the i sound, or at least
the end of it, is often uttered without ‘voice’
but with whisper.
§ 61. In German there is parallelism between
u sounds (§ 45) and i sounds
[uː] | [iː] | [U] | [ɪ]
The long (accented) sounds are ‘tense,’ the
short (accented) sounds are ‘lax.’ (See § 40.)
Examples : mir [miːr], mit [mìt].
In syllables before the chief accent the [iː] is
usually shortened and often becomes [i] : Minute
[minuːtə], or even [mɪnuːtə], which should not
be imitated. The y in words from the Greek is
variously pronounced : as [i] (so regularly in a
few thoroughly acclimatised words, e.g. Gyps
[gips]), or as [y], which is preferred by most
of the educated.52
Front vowels, with lip rounding.
§ 62. These sounds are found in French and German.
According to the usual view they result from
a combination of
the lip rounding of o and u sounds with
the tongue articulation of e and i sounds.
The tongue articulation however appears to be somewhat
modified : the front of the tongue is slightly
lowered (inner rounding) and helps the lip rounding.
In English these sounds do not occur. However,
the ‘mixed’ sounds, frequently used to represent u
in fur or but, and the ‘neutral’ vowel in —er=[ə]
both in articulation and in sound very closely
approach those open [œ] sounds in which the lingual
element prevails over the labial.53
The ö sounds :
close [ø], open [œ].
§ 63. In general, these sounds combine the lip
rounding of o sounds with the tongue position of
e sounds. (The tongue position is modified, compare
§ 64. In French there is close [ø] and open [œ],
and both occur long and short or half-long ; crease
[krøːz], feu [fø] ; peur [pœːr], neuf [nœf].
There is also a ‘middle’ sound between these,
short and unstressed, here designated [ə] : je
§ 65. In German we notice the same distinction
of the close long and open short sounds, as in the o
sounds (§ 49). Long [ø] in schön [ʃøːn] ; short [œ]
in wölben [vœlbən] (this is rather a sound between
close and open, compare short o and e).
The vowels of French words obtain the current
German values in passing into that language
(as in Robe [roːbə], Toilette [toːɑʹlɛtə]) ;
so here also, in the frequent ending —eur,
French [—œːr], the vowel being long becomes
close : Redakteur [redakʹtøːr].54
The ü sounds :
[y] (voiceless [y]).
§ 66. Generally speaking, these sounds combine the
lip rounding of u sounds with the tongue position of
i sounds. (The tongue position is modified, compare
§ 67. In French there is only close ‘tense’ [y] ;
the lip rounding is extreme. Long in ruse [ryːz] ;
half-long or short in durer [dy(ː)re], culte [kylt].
At the end of the word, before a pause, the
sound loses its ‘voice’ after a voiceless consonant;
[sỵ] for [sy] = su. It does not lose it
entirely after a voiced consonant ; [dû] = [dyỵ].
Compare [u] in § 44.
§ 68. In German short ü is rather more open than
long close ü : kühn [kyːn], Sünde [zyndə].
When it has not the chief stress, [yː] is
shortened : amüsieren [ʔamy(ː)ʹziːrən].
The y of words taken from Greek is treated
like German ü : Lyrik [lyːrik], Rhythmus
[rytmus] ; but it is also pronounced as i
(compare § 61).55
Mixed vowels :
§ 69. In the case of these vowels the tongue is
slightly raised in the middle ; the front and back
remain almost idle, though of course they are pulled
up into an inward slant by the rising of the middle.
There may be a slight hollow in the middle of the
tongue, but it is easily abolished without altering the
The articulation of these sounds is not decided,
and the impression on the ear is not clear.
In French [ə] occurs only in weakly stressed syllables (e.g.
in je, me, te) ; it becomes [œ] if emphatic.
§ 70. In English there is such a ‘mixed’ sound,
e.g. in the ur of turn [təːn]. The tongue is practically
in the [ɑ] position. This sound is the current
southern English pronunciation of certain vowels
before a written r, which is pronounced only when
before the vowel of a derivative ending or finally in
‘liaison’ : err [əː], but erring [əːriŋ] ; her [həː], but
her and me [həːr ən miːj].
In northern (and in some parts of southern)
English the combination of vowel and r is pronounced
as a ‘coronal’ əː [əː˕] ; the point of
the tongue is raised almost to the [r]-position,
and the sound closely approaches [r] in value.56
Another ‘mixed’ sound, which differs little from
[əː], is substituted for short (usually back) vowels
when not accented : ivory [aĭvəri], stirrup [stirep].
The sound approaches [əː] yet more closely in those
cases in which the unaccented vowel was (and in
writing still is) followed by final r : better [betə].
Here also there is often transition to the
Another [ə] sound is frequently heard in place of
the southern English [ʌ] sound in but, etc. Further,
there are the first elements of [aĭ] in high and [aŭ]
in now (compare § 51) and the sounds occasionally
substituted for unaccented [oŭ], as in fellow (compare
§ 47), and for [ju] as in value (compare § 43).
§ 71. In German there is only the ‘unaccented
e’ in syllables before or after the stem : Gebot
[gəboːt], Bitte [bitə] ; also in enclitics when unaccented:
The whole tongue seems to be raised a little more
than for [ɑ] ; the raising of front and back and
lowering of middle being very slight. Following
consonants often affect the sound, and it tends to
[e], [ɑ], [o], or [œ]. It may be said that in final e, the
[e] element generally preponderates.57
Vowels with nasal resonance, nasal vowels.
§ 72. In English there are no nasal vowels.
Words borrowed from French, if they really become
part of the language, are generally pronounced in
English fashion, and [ɔn] is used for [ɑ̃] and [ɔ̃], [æn]
for [ɛ̃], [œn] (and perhaps [ʌn]) for [œ̃] ; occasionally
[oːŭn] for [ɔ̃]. Thus envelope [ɔnvəloŭp], rendezvous
§ 73. In French there are four nasal vowels, all
open and with the tongue lowered. The nasal character
of these sounds is so evident because the velum
is lowered considerably.
Nasal [ɔ̃], written on or om, is a little closer than
oral [ɔ] in porte, morte ;
nasal [ɛ̃], written in, ein, or ain, is more open than
oral [ɛ] in perte, mer ;
nasal [œ̃], written un or um, is more open than
oral [œ] in peuple, peur ;
nasal [ɑ̃], written en, em, an, or am, is a little
closer than oral [ɑ] in lâche.
These sounds are short when final (though they
may be long before an emphatic pause) ; before consonants
they are long when they bear the chief
stress, half-long with secondary stress : e.g. short,
[ɔ̃] in rond, long in ronde, half-long in rondeau.58
§ 74. In German nasal vowels occur only in words
recently borrowed from French. They seem, however,
to be always long (for they are in ‘open’
syllables, in which German has no short accented
vowels) : Ballon [bɑʹlɔ̃ː], Nüance [nyʹɑ̃ːsə], Bassin
[baʹsɛ̃ː], Verdun [verʹdœ̃ː]. — In northern, and partly in
middle German, the native sounds [ɔŋ, ɑŋ, ɛŋ, œŋ] are
often substituted : [bɑʹləŋ, nyʹɑŋsə, bɑʹsɛŋ, vɛrʹdœŋ].
Voiceless sounds : h sounds.
§ 75. If the current of breath — without first producing
‘voice’ — passes through the mouth, which
has already prepared itself for the uttering of a
certain vowel, the slight ‘friction’ produced will
appear to the ear as a breath [h]. There are as
many varieties of it as there are of vowels ː for
instance [hɑː] is really voiceless (breathed) [ɑ] followed
by voiced (vocalic) [ɑ], [heː] is voiceless and
voiced [e], and so on. Compare § 26.
representing the articulation
of the tongue against the hard palate.
What is black in the following diagrams indicates
those parts of the hard palate which are touched by
the tongue when the vowels [i, y, e, ø, ɛ] are uttered.
They are the results of actual experiments made
with artificial palates, covered with fine powder, 1 and
59then inserted in the mouth ; the powder being removed
wherever the tongue touches. The results
were obtained by Kingsley for English, by Rousselot
for French, and by Viëtor for German.
English | French | German
French | German
English | French | German
French | German
French | German
The following diagrams show the vowel schemes
of English, French and German. For the [u]-[ɑ]-[i]
lines, see the diagram on p. 27. The asterisk
indicates that the sound occurs both long and short.
The brackets imply lip rounding.
(The antevocalic h sounds of English and German
correspond in articulation with the respective vowels.)61
*i | (ü) | (*u) | *e | (ö) | (*o) | ɔː | *əʌ | *æ | *ə | ɑː
Short vowels pit | pet | pat | ivory | put | but | pot
Long vowels father, fall, firm
slightly diphtongal with [ĭ] bait | with [j] beat |
with [ŭ] boat | with [w] boot |
with [ə] bear, beer, boar, boor
clearly diphthongal with [ĭ] file, foil ; value |
with [ŭ] foul
triphtongal with [ə] fire, employer ; pure
*i (*y) | (*u) | e (*ø) | (*o) | *ɛɛ̃ (*œ*œ̃) | (*ɔ*ɔ̃) | a | *ɑ
Short vowels vif culte | route | blé feu | côté | paix neuf | robe | ma | pas
Long vowels rive ruse | rouge | creuse | rose | rêve peur | or | rage | âme
Nasal vowels banc, bain, bon, un63
iː (yː) | (uː) | ì(ý) | (ù) | eː(øː) | (oː) | *ɛ(œ) | ə | (ɔ) | *ɑ
Short vowels mit Sünde | und | bitte | Sonne | fett wölben | was
Long vowels mir kühn | du | fehlen schön | so | Bär | Vater
Diphthongs drei, Frau, treu64
[List of Contents]
Consonants ; voiced and voiceless (Nasals see § 132).
§ 76. When the articulation of the mouth is such
that in some part of the passage through the mouth
there is a narrowing or closure, then the current
of exhaled breath produces a certain sound ; the
resonance due to the articulation has some effect
also, but it is of secondary importance.
Certain of these sounds are used for the purposes
of speech ; they are the consonants (which also include
the glottal sounds and the breath sounds, § § 23 foll.).
§ 77. When the sound is due to a rubbing of the air
as it passes (continues) through the narrowing, it is a
continuant 1 or fricative ; if it is due to the checking of
the air by means of a closure or stoppage or to the
explosion caused by breaking through the stoppage,
it is a stop or plosive (sometimes called explosive).
§ 78. When the breath has produced ‘voice’ (by
setting the vocal chords vibrating) before it reaches
the narrowing or the closure, the sound is voiced ;
otherwise it is voiceless.
Generally speaking we have pairs of sounds : a
strongly articulated voiceless sound beside a weakly
articulated voiced sound. They will be treated
together in the following sections.
A sound may be ‘soft,’ i.e. weakly articulated,
without on that account becoming ‘voiced.’ This
is the case in middle and southern Germany,
where the ‘soft’ consonants are actually voiceless.66
The manner of formation may be tense or lax in
the consonants, as well as in the vowels (§ 40). In
English they are lax, in French tense, while in
German the formation is between the two extremes.
Voiced continuants, when adjacent either to a
voiceless sound or to silence, always possess a voiceless
portion, which is characteristically long in
English, and short in French.
Narrowing of the oral passages.
Uvular [ʀ] ;
Back or velar continuants :
voiced [ɡ], voiceless [x] ;
Front or palatal continuants :
voiced [j], voiceless [ç].
§ 79. The narrowing here takes place
between back of tongue ridge and soft palate in
the velars (more strictly : postpalatals) ;
between front of tongue ridge and hard palate in
the palatals (more strictly : antepalatals).
As in the sounds already discussed, where
the passage is open, it is the ridge or dorsum
of the tongue which articulates, the formation
of the sounds is dorsal.
For [ɡ, x] the tongue is raised as for close [u], and
for [j, ç] … close [i] ;
but in each case a little higher, so that the breath
For uvular r, designated [ʀ], the narrowing
is even farther back than for [x].67
Uvular r [ʀ].
§ 80. This is the sound produced when the uvula is
made to vibrate, the back of the tongue ridge being
raised towards the soft palate. If there is no channel
along the tongue to enable the uvula to vibrate freely,
the sound becomes scraping or passes completely into
[ɡ] or [x], as friction of air takes place between the
velum and the tongue ridge. — As a rule the sound is
voiced, but before or after voiceless sounds, and when
final, it may become voiceless.
§ 81. In English it only occurs as a peculiarity in individuals
or in dialect (it is the‘Northumbrian burr’).
Englishmen are warned against trying to acquire the uvular
[ʀ]. “An Englishman easily learns to trill lingual [r] to the
satisfaction of any Frenchman, whether [ʀ] or [r] speaker ; but
I never heard an Englishman produce a decent [ʀ].” — P. Passy.
§ 82. In French it is the regular sound of r. In
Paris and other big towns it is almost exclusively
used, though theoretically the lingual r (see § 102)
is preferred ; in the country as a whole lingual r is
§ 83. In German uvular r (‘Zäpfchen-r’) is much
used, particularly in towns, and seems to be spreading.
In the speech of the stage and in singing only
lingual r is considered correct.
Back or velar continuants.
voiced [ɡ], voiceless [x].
§ 84. By raising the back of the tongue ridge yet
higher than for [u] we obtain a narrowing ; the
breath brushing through it produces velar or back
68continuants. The sound varies slightly according as
the narrowing is further forward or more behind.
§ 85. In English the sound [x] is not a recognised
‘letter,’ but even in southern speech it sometimes
expresses the gh of the interjection faugh (usually
[fɔː]) ; it is becoming more and more common in the
pronunciation of words taken from Scotch (e.g. loch),
Welsh or German. The sound is fairly common in
the popular dialect speech of Scotland and northern
§ 86. In (northern and middle) German the voiced
[ɡ] is used for written ɡ (not final) after back vowels
(ɑ, o, u) : Tage [tɑːgə]. The voiceless [x], the so-called
ach-Laut, represents the ch following ɑ, o, u :
ach [ʔɑx] ; and (in northern and middle German)
also the final g after ɑ, o, u : Tag [tɑːx].
In the speech of the stage and in southern
Germany g after ɑ, o, u is pronounced as a stop
[ɡ], finally [k] : [tɑ:gə, tɑːk]. Foreigners need
therefore not trouble to acquire the other pronunciation ;
see also § 123. The pronunciation
[ɡ, k] seems to be gaining ground.
Front or palatal continuants :
voiced [j], voiceless [ç].
§ 87. [j] and [ç] are in the same relation to [i] as
[ɡ] and [x] to [u].69
§ 88. In English the voiced sound is produced
with so little narrowing that as a rule the brushing
of air is hardly perceptible ; and it is essentially
gliding, not held. It occurs most frequently at
the beginning of words, when it is written y, as
in yes ; and in the combination [juːw], written u, eu,
etc., as in due (compare § 43). After voiceless sounds
it occasionally passes into [ç], or even [ʃ] ; compare
the vulgar don'tcherknow.
French | German
See p. 60.
§ 89. In French we find [j] similarly used : for y
in yeux [jø], for i in vieux [vjø] ; and the voiceless
sound often after voiceless consonants, as in pied
As in these cases vocalic i sounds pass over
into continuants, so French i in general — owing
to its very close articulation — is but little removed
from a continuant ; and indeed for final
[i] we often have [ç], which also takes its place
§ 90. In German [j] is the j at the beginning of
words : ja [jɑː] ; but — unlike English — it is held.
In northern and middle German it is used for g
within words after front vowels or l, r : Siege [ziːjə].
Voiceless [ç], the ‘ich-Laut,’ for ch : ich in [ʔiç|,
and in northern and middle German for final g
after front vowels : Sieg [ziːç].
Here also (see § 86) the speech of the stage
and of southern Germany adopts the pronunciation
of g as a stop : [ziːɡə, ziːk].
Teeth or dental continuants :
voiced [ʒ, z, ð],
voiceless [ʃ, s, θ] ;
liquids [r] [l].
§ 91. The dentals include sounds articulated by
the point of the tongue (including what is
strictly called the ‘blade,’ viz., that part of the
surface which lies behind the point), and
the ridge above the upper teeth (alveolar continuants) ;
the teeth themselves (true dental continuants).
The part of the tongue which helps to form the
the front rim of the tongue apical formation, or
the surface of the tongue behind the front rim (pre)dorsal formation, or
both combined dorso-apical formation.71
As in the case of back and front consonants, the
place of articulation for most dentals is along the
middle line of the mouth (medial formation) ; but
for the l sounds the narrowing is between the side
rim or rims of the tongue and the side teeth (lateral
The ‘hushing’ and ‘hissing’ sounds, or
§ 92. These include the sounds represented in
English by s, z, sh, etc. The characteristic ‘hissing’
seems to be due to a breaking of the current of
breath against the teeth ; the current being diffused
in the sh sounds and concentrated on one point in
the s sounds.
According to Dr Lloyd, the chief force in the production of
English [ʃ] is the resonance of the fore cavity reinforcing that
which comes from the hind cavity.
The hushing sounds :
voiced [ʒ], voiceless [ʃ].
§ 93. In English the simple voiced [ʒ] only occurs,
where in the spelling s is followed by a front vowel
(or by u = original [juː]). Here it has been substituted
for older [zj] : vision [viʒən], pleasure [pleʒə].
The spelling is z in azure [eːǐʒə]. The combination
[ʤ] is much more common. This is the usual
pronunciation of j : joy [ʤɔǐ], June [ʤuːwn], etc.,
and often of g before a written front vowel : age
The voiceless [ʃ] is usually written sh : sheep
[ʃijp], rash [ræʃ], etc. ; also s (after consonants), ss,
c, t, before front vowel (and before u = original
[juː]), when following the chief accent ; as in
tension, passion, vicious, nation. 1 The combination
[tʃ] is very common ; it is usually written
ch : church [tʃəːtʃ] ; also tch : scratch [skrætʃ] ; in
some words [tʃ] has sprung from [tj] = t before front
vowel (or before u = original [juː]), when following
the chief accent, as in righteous [raĭ tʃəs], nature
English | French | German
See p. 60.
[neĭtʃə], etc. Similarly, the combination [kʃ] goes
back to [ksj] = xi in anxious [æŋkʃəs], etc.
The tongue articulation of these sh sounds is
(pre)dorsal 2 and also apical (or dorsal only) and
palatal. There is some friction between the
tongue and the gums ; but that past the front
teeth is more important. The lips are passive.73
§ 94. In French the voiced [ʒ] occurs frequently.
It is written j, and g before e, i (y), ge before a, o, u :
jardin [ʒardɛ̃], loger [lɔʒe], pigeon [piʒɔ̃]. — The
voiceless sound is written ch : chapeau [ʃapo].
The tongue articulation is (pre)dorsal and
(post)alveolar, or apical and (post)dental. The
friction takes place as in English. The lips are
§ 95. In German the voiced [ʒ] occurs only in
borrowed words, mostly from the French, and retains
the foreign spelling j, g : Journal [ʒurʹnɑːl],
Logis [lo'ʒiː]. For the foreign combination [ʤ]
German usually substitutes [ʒ] : Jockey [ʒɔkɑĭ]. —
The voiceless [ʃ] is common ; in native words it is
usually written sch : Schall [ʃɑl], Esche [ʔɛʃə],
Busch [buʃ]. This is also the sound of s in German
words beginning with sp, st (except in a part of
northern Germany) : Spitze [ʃpitsə], sprechen
[ʃprɛçən], Stein [ʃtɑĭn], straucheln [ʃtrɑŭxəln].
The combination [tʃ] occurs for the tsch of German
words, as in Peitsche [pɑĭtʃə] ; also for foreign ch,
though here plain [ʃ] is more frequently spoken :
Check [tʃɛk, ʃɛk].
The tongue articulation is dorsal and alveolar
or postdental. The friction takes place as in
English. The lips are protruded.
§ 95A. The fundamental conditions of the sh sounds are : (1)
post-alveolar friction (for which any available part of the fore
surface of the tongue may be employed) ; and (2) a cavity
behind both rows of teeth.74
The hissing sounds :
voiced [z], voiceless [s].
§ 96. In English both are usually written s. The
voiced sound is also written z, as in zeal [ziːjl].
It is s particularly when it represents an inflection,
and follows a voiced sound : wails [weːĭlz], dances
[dɑːnsiz] ; in some monosyllables it occurs finally :
as, has, is, his, was ; and in certain verbs, whereas
the substantive has voiceless [s], as close [kloːŭz] and
English | French | German
See p. 60.
[kloŭs], use [juːwz] and [juws]. In many positions,
e.g. after dis-, pre-, re- the pronunciation of s is [z] in
one batch of words, [s] in the rest ; as in disease
and disobey, reserve and research. The combination
[ɡz] is written gs, and also x before an accented
vowel : anxiety [æὴɡzaĭəti], etc.
The voiceless [s] is written s initially : set [set] ;
it is an inflection after voiceless sounds :lots [lɔts],
bakes [beĭks] ; it also occurs after voiceless sounds,
as in gipsy, and within the word, in most cases, even
75after voiced consonants, as in intensity ; and also
after vowels (especially in words from the Latin), as
in decisive. It is also written ss : passage, etc. ; and
c or sc before e or i : cite, scene, etc. The letter x
in most cases has the value [ks] : box, etc.
The tongue articulation is (pre)dorso-apical
(cp. § 91, towards the end), and alveolar.
§ 97. In French both sounds occur. The voiced
[z] is written z : zèle [zɛːl], or s (as a rule between
vowels) : maison [meːzɔ], rose [roːz]. Otherwise silent
s, x, z also have the value of [z] in ‘liaison’ ; with
this exception, x rarely = [z]. — In all other cases s is
[s], particularly initially : sel [sɛl], estimer [ɛstime],
Vénus [venys] ; ss always expresses this same sound,
as do also c or sc before e, i (y), ç before other
vowels, and t before i in certain suffixes. — The
combinations [ɡz] and [ks] are also written x, which
usually has the latter value.
The tongue articulation is (pre)dorsal, and
postdental or alveolar.
§ 98. In German s has the value of voiced [z]
initially, and within the word between vowel or
liquid and vowel : so [zoː], reisen [rɑǐzən], Binse
[binzə], etc. — In all other cases s = voiceless [s] (except
in initial sp, st ; cp. § 95) : Skelett [skɛʹlɛt], Erbse
[ʔɛrpsə], List [list], Moos [moːs] ; and ss has the
same value : reissen [rɑǐsən], Fuss [fuːs], Kuss [kus].
The tongue articulation is (dorso)apical or
dorsal, and alveolar.76
§ 98A. The fundamental conditions of the s sounds are the
same as for the sh sounds (§ 95A), but on a smaller scale. The
constriction must be more advanced, (pre)alveolar, and the
intra-dental cavity is therefore much smaller ; but the tongue
may (and does) wander, exactly as in [ʃ, ʒ] though within
The lisping sounds :
voiced [ð], voiceless [θ].
§ 99. If an s sound is articulated, but without a
well-formed channel along the middle of the tongue
and without a well-formed intradental cavity, the
result is a ‘lisping’ sound. These are not found in
French or German, except as a defect of speech.
They are sometimes called interdental sounds ; but
it is by no means an essential of these sounds in
general, or the English sounds in particular, that
they should be produced ‘interdentally,’ i.e. with the
tongue point between the teeth. Indeed as a rule
the English sounds are not so formed, the narrowing
being between the tongue point (with apical articulation)
and the back of the front upper teeth.
§ 100. The voiced sounds occur e.g. in thou [ðaŭ],
this [ðis], brother [brʌðə] ; the voiceless in thin
[θin], breath [breθ]. After a long vowel the final
[θ] of some substantives becomes voiced before the s
of the plural : bath [bɑːθ], baths [bɑːðz].
The liquids :
[r] and [l].
§ 101. These sounds differ in various points from
the other ‘continuants,’ but may be given under this
77heading. According to the English, French and
German system of sounds, we do not regard voiced r
and voiceless r [ṛ], voiced l and voiceless l [ḷ], as independent
sounds, but recognise only one r and
one l, which are both voiced as a rule, but may be
voiceless under certain circumstances.
When there is ‘voice,’ the narrowing may be very
slight, for even then it suffices to give the sound its
characteristic resonance ; the r and l thus formed
have — like the nasals — very much the nature and
effect of vowels. The narrowing only becomes considerable
and the ‘friction’ obvious when the sounds
are produced without ‘voice.’
The r sounds :
voiced [r], voiceless [ṛ].
§ 102. The tongue point [r] (lingual or dental r) is
produced by a narrowing between the raised tongue
point and — as a rule — the ridge of the upper gums
(the alveoles). The French and German r is also
rolled (trilled), i.e. the stream of air causes the tongue
point to vibrate. This is not the case in English r.
(The untrilled r is strictly [ɹ] ; but [r] is usually employed
Meyer has, however, found a single repetition in a purely
southern [r], regularly, when medial.
§ 103. In English alveolar (or : lingual) r, untrilled,
consisting of a single tap of the tongue, or of a
slight narrowing, without strong ‘friction,’ only
occurs before vowels initially and within the word
78(for r, rr, and for rh, rrh in words from the Greek) :
right [raĭt], very [veri], parrot [pærət] ; also after
consonants : grow [groːŭ]. But particularly after
[d] or [t] there is greater narrowing and consequently
‘friction’, and after voiceless sounds there is besides
loss of ‘voice’ : dry [draĭ], try [traĭ] or [tṛaĭ]. Cp.
also e.g. grease [grijs] and increase [inʹkrijs] or
[inʹkṛijs]. These sounds approach the value of
Final r, as well as r within the word before
consonants, have lost their r value in educated
English | French | German
See p. 60.
southern speech, their place being mostly taken
by [ə], compare § 70 ; but for this r, as also for
ur, er, ir the dialects have [r] sounds. In ‘liaison’
before an initial vowel the [r] generally appears :
better [betə], but better and better [betər ən
betə] ; here [hiːə], but here and there [hiːər ən
ðɛːə]. Even here it is often omitted by the
younger generation in southern England.
Before [r] within the word between vowels
and finally in ‘liaison’ there is in most cases
79[ə] ; sometimes not (viz., when the preceding
vowel is short, or after diphthongs, or after
[ɛː] or [ɑː] or, generally, [ɔː]).
§ 104. In French the lingual [r], if used at all, is
trilled. Like the uvular [ʀ], which generally takes
its place, it is as a rule voiced : rire [riːr], porter
[pɔrte] ; final re after consonants is often voiceless
(and dropped altogether in colloquial speech) : offre
[ɔfṛ], sucre [sykṛ]. The written rr (rh, rrh) has the
same value as r ; but in a number of words rr = [rr]
or [rː], the rolling decreasing and then increasing
again ; as in mourrai.
§ 105. In German the lingual [r] is as a rule clearly
trilled, except when it is final. It is generally voiced
and produced without any distinct ‘friction.’ Before
or after voiceless sounds, it often loses its ‘voice,’
partly or altogether. Finally the rolling is sometimes
reduced to a single tap of the tongue (as in
English initial [r]) ; or [ə], [ɑ], and [æ] sounds are
substituted (but this should not be imitated). The
written r and (after short vowels) rr, as also rh and
rrh, have the value of [r] in all cases.
The l sounds :
voiced [l], voiceless [ḷ].
§ 106. Here we have a narrowing between the side
rims (or one only) of the tongue and the molars and
side gums (lateral articulation) ; the tongue point
80touching the top of the mouth somewhere along the
middle line ; in English, French and German it is the
middle of the ridge over the upper teeth.
§ 107. In English there is a concave lowering of
the front of the tongue, and in connection with it a
raising of the back of the tongue. This gives the [l]
a low sound, particularly noticeable when it is final.
Before and after voiceless sounds it may lose its
‘voice,’ as in play, felt, etc., and — the friction becomes
English | French | German
See p. 60.
§ 108. In French, [l] (like French [r]) is generally
voiced, but voiceless when final after a consonant :
lire [liːr], aller [ale] ; table [tabl], etc. — In the combinations
ll, ill, ille, il, ille after a vowel the [j]
sound (‘l mouillée’) is often used 1 ; rarely for l alone.
§ 109. In German [l] is usually voiced and has no
obvious ‘friction.’ Usually the back of the tongue
Lip or labial continuants
lip-teeth or labiodental :
voiced [v], voiceless [f] ;
lip-lip or bilabial :
voiced [w], [v], [ɥ], voiceless [ʍ], [F], [ɥ̊].
§ 110. The labiodental sounds are produced by the
breath passing between the lower lip pressed against
the upper teeth, the bilabial sounds by its passing
between the lips.
The labiodentals and the bilabials have both a
tongue articulation which is always very open, and
resembles that of the following vowel : the [w] sounds
have a tongue articulation like [u] or somewhat more
compressed ; the [ɥ] sounds have a tongue articulation
like [y] or somewhat more compressed. Still,
purely consonantal bilabial [v] and [f] are not rare.
§ 111. In English the voiced labiodental [v] is
usually written v : very [veri], liver [livə], serve
[səːv]. The voiceless labiodental [f] is usually written
f or ff, but also ph : find [faĭnd], differ [difə], philosophy
There is also a voiced bilabial [w], with the tongue
in the [u] position, but with rapidly vanishing and
therefore indistinct friction, usually written w : wet
[wet]. The corresponding voiceless sound [ʍ] is
written wh, as in which ; in natural southern English,
however, the voiced sound is used here also, though
82attempts are often made to reintroduce the voiceless
sound, which is generally used in northern
English and in Scotch. — After voiceless sounds [ʍ]
sometimes takes the place of [w] : twenty [twenti]
Both these sounds [w, ʍ] are essentially gliding,
§ 112. In French there are also labiodental voiced
[v] and voiceless [f] : vin [vɛ̃], vive [viːv] ; fin [fɛ̃],
French has further two kinds of bilabials : a [w]
with lips well rounded and the tongue raised to the
[u] position, as in roi [rwɑ] ; and a [ɥ], also with lips
well rounded and the tongue raised to the [y] position,
as in ruine [rɥin]. After voiceless sounds both tend
to become voiceless [ʍ] and [ɥ̊], as in point [pwɛ̃] or
[pʍɛ̃], puis [pɥi] or [pɥ̊i].
§ 113. In German the voiced labiodental [v] is represented
by w in writing, also by v in words borrowed
from Latin or the Romance languages ; like all
other voiced consonants in German it does not occur
finally ; examples : was [vɑs], Klavier [klɑʹviːr]. The
friction is much less distinct than in English and
In middle and southern Germany w is generally
pronounced as a bilabial [v], produced
without lip rounding and without much friction,
also without raising of the tongue ; it is
not voiced except sometimes when it passes over
83into the next sound, so that it is perhaps better
designated as a weak [F].
In these parts of Germany foreign v is pronounced
[f], not only when it is final, as in
kursiv [kurziːf], but also initially, as in Vikar
[fikɑːr], and medially, as in Klavier [klɑfiːr].
For w in the combinations schw, zw, (sw, tw), and
for u in qu, bilabial [v] is often spoken instead of
labiodental [v], and in these cases both often lose
some or all ‘voice’ : Qual [kvɑːl], Schwester [ʃvɛstər]
almost [kfɑːl, ʃfɛstər], or [kvɑːl, ʃvɛstər] almost
The voiceless labiodental [f] is written f or ff, sometimes
v, and ph (almost exclusively in words from the
Greek) : fiel and viel [fiːl], laufen [lɑŭfən], Philosoph
Closure of the oral passages :
Stops without nasal resonance.
§ 114. As in the case of continuants, the stops (or
plosives) generally go in pairs, one sound being
voiced and weakly articulated, the other voiceless
and strongly articulated (compare § 78). In grammars
they are sometimes called ‘media’ and ‘tenuis’ ;
and both kinds of stops are known as ‘mutes.’
§ 115. In the articulation of a simple stop we must
distinguish (1) the making of a stoppage, and (2) the
bursting through the stoppage by the breath ; we may
leave out of account the momentary pause between
(1) and (2). Often at the beginning of words only (2)
is important, and finally only (1) ; for the other half
of the process is slow and imperceptible. Within
the word both parts are distinct.
If the pause between (1) and (2) is noticeable, the
stop is ‘long’ or ‘double.’
§ 116. In English and German voiced stops occurring
initially ‘voice’ comes a moment after mouth
articulation (i.e. the sound begins voiceless and
then becomes voiced) ; in English voiced stops
occuring finally ‘voice’ ceases a moment before
mouth articulation (i.e. the sound begins voiced
and then becomes voiceless) ; in French the voiced
85stops remain unchanged wherever they occur, except
that finally the off-glide is perhaps oftener
whispered than voiced.
§ 117. In middle and southern German we notice
a treatment of stops similar to that of continuants
(compare § 78). In place of voiced sounds we here
find weakly articulated voiceless sounds ; and these
are often substituted also for p, t, k.
§ 118. Aspirates, 1 i.e.. aspirated voiceless stops,
(strictly [pʰ, tʰ, kʰ]) occur regularly in English and
German before accented vowels, and often elsewhere
in accented position, especially when final ;
in this position aspiration is not unknown in French
§ 119. When a stop is followed by the corresponding
nasal, as in [bm, pm, dn, tn], etc., the stoppage for the
first sound often remains for the nasal. Instead of
the bursting through the oral stoppage we then have
the lowering of the velum as the breath passes
through the nose ; and this is heard. This kind of
stop is a ‘velar’ or ‘faucal’ stop.86
Back or velar stops
Front or palatal stops :
voiced [ɡ], voiceless [k].
§ 120. The stoppage is effected between the ridge
of the tongue and
the back or soft palate, giving back or velar
stops ; or
the front or hard palate, giving front or palatal
As a general rule, however, the distinction is unimportant;
for the difference in articulation is much
less than in the case of the continuants, and the
acoustic effect is practically identical. Both back and
front stops may be designated [ɡ, k] for our present
§ 121. In English voiced [ɡ] is usually written g,
and also gg within the word and finally : gift [ɡift], beg
[beɡ], beggar [beɡə], egg [eɡ] ; sometimes gu, usually
before front vowels : guest [ɡest], rarely gh : ghost
[ɡoŭst]. The voiceless [k] is variously written,
usually c or k : keen [kiːjn], sick [sik], can [kæn],
accuse [əʹkjuːwz], chord [kɔːd], queen [kwiːjn], tax
[tæks] ; also gh in a few words.
The place of articulation varies according to the
neighbouring sounds. The [k] is usually aspirated
[kʰ] before accented vowels, more rarely when final.87
§ 122. In French the voiced sound is written g,
and gu before front vowels (e, i, y) : gant [ɡɑ̃ː],
guerre [ɡɛːr]. The voiceless [k] is never aspirated
when initial or within the word ; it is usually written
c (cc), qu (cqu) : camp [kɑ̃], qui [ki]. — When there
is ‘liaison’ of (otherwise silent) g, it is pronounced
[k] : long usage [lɔ̃k yzaːʒ], rang élevé [rɑ̃k elve].
The place of articulation varies according to
neighbouring sounds ; sometimes it is very far forward
§ 123. In German the voiced sound is represented
by g 1 (except when final) and also by gg (within the
word) : Tage [tɑːɡə], Siege [ziːɡə], gut [ɡuːt], Gift
[ɡift], Egge [ʔɛɡə]. The voiceless [k] (aspirated —
much more than in English — [kʰ] before accented
vowel, or after it when final) is written k, after a
short vowel ck, and finally also g 1 : Kohle [koːlə],
Birke [birkə], keck [kɛk], Tag [tɑːk], Sieg [ziːkj.
In foreign words there are various other spellings :
Clique [klikə], Accord [ʔɑʹkɔrt], Chor [koːr]. Notice
also in German words qu = [kv] (compare § 113) and
chs = [ks], for which x also occurs : Quelle [kvɛlə],
Hexe [hɛksə], Fuchs [fuks].
The place of articulation is usually the middle of
the soft palate ; but before or after ‘front’ sounds it
is usually more forward.88
Teeth or dental stops :
voiced [d], voiceless [t].
§ 124. The place of articulation of the dental stops
practically coincides with that of the various dental
continuants (see § 91).
§ 125. In English the voiced [d] is written d (also
dd within the word) : do [duːw], adder [ædə], bed
[bed]. The voiceless [t] — aspirated [tʰ] initially, and
elsewhere, before accented vowel, and often finally
— is written t (tt) : tell [tel], matter [mætə] ;
English | French | German
See p. 60.
also d in the -ed of verbs after voiceless sounds :
tapped [tæpt] ; rarely th.
The articulation is usually apical 1 and alveolar
(‘supradental’). English [t d] is a stoppage of [ɹ̣, r].
§ 126. In French the voiced sound is always
written d : dire [diːr], bande [bɑ̃ːd]. The voiceless
[t] is not aspirated initially or within the word, and
is written t (tt) : tirer [tire], pɛrte [pert], attendre
[atɑ̃ːdṛ] ; in borrowed words also th. — When there is
89‘liaison’ of (otherwise silent) d, it is pronounced [t] :
grand homme [grɑ̃t ɔm], répond-il [repɔ̃t i(l)].
The articulation appears to be usually dorsal and
alveolar. Fr. [t d] is the stoppage of a somewhat
retracted (but still dental) [θ ð].
§ 127. In German the voiced [d] is written d (dd) :
da [daː], leiden [lɑĭdən], Kladde [klɑdə]. The
voiceless [t] — aspirated [tʰ] initially, and elsewhere,
before accented vowels, often also when final — is
usually written t (tt), th, sometimes dt, and finally
also d : treu [trɔĭ], retten [rɛtən], Stadt [ʃtɑt], und
[ʔunt], Thron [troːn]. The combination [ts] is often
written z (tz), in foreign words also c : zu [tsu],
Satz [zɑts], Cäsar [tsɛːzɑr].
The articulation is either apical and pre-alveolar,
or dorsal and alveolar, i.e. intermediate between the
Engl. and Fr. In the North of Germany it tends to
be apical, elsewhere dorsal.
Lip or labial stops :
voiced [b], voiceless [p].
§ 128. In general the lip stops are bilabial, i.e.
formed by a closure of both lips.
§ 129. In English the voiced sound is written
b (bb) : but [bʌt], babe [beːĭb], ebb [eb]. The voiceless
sound — aspirated [pʰ] initially, and elsewhere,
before accented vowels, and often finally — is written
p (pp) - put [put], proper [prɔpə], happy [hæpi] ;
very rarely ph.
§ 130. In French the voiced [b] is written b (bb) :
bon [bɔ̃], robe [rɔb], abbaye [abei]. The voiceless
90[p] is written p (pp) : point [pwɛ̃], troupe [trup],
§ 131. In German [b] is written b (bb), and does
not occur finally : bei [bɑĭ], aber [ʔɑːbər], Ebbe [ʔɛbə].
The voiceless [p] — aspirated [pʰ] initially, and elsewhere,
before accented vowels, and often when final
— is written p (pp), finally also b : Paar [pɑːr],
Rappe [rɑpə], ob [ʔɔp].
In the combination [pf], as in Pferd, Kopf, the
[p] is often formed labiodentally, i.e. by pressing
the lower lip against the upper teeth.
Stops with nasal resonance :
nasal stops or nasals.
§ 132. The stoppage is practically identical with
that of the ‘oral’ or ‘pure’ stops [ɡ, k], [d, t],
While the latter are momentary sounds, in the
case of the nasals the breath can pass freely through
the nose, and they can consequently be continued as
long as continuants or vowels.
As a rule they have ‘voice.’
Back or velar nasal :
and front or palatal nasal :
§ 133. In English the place of articulation of [ŋ] is
the same as for [ɡ, k] — viz., the middle of the soft
palate, except before front vowels, when it is farther
91forward, roughly speaking, at the line between soft
and hard palate. It is written ng, and n before g [ɡ]
and k [k] : long [lɔŋ], strictly [lɔŋː], singer [siŋə],
finger [fiŋɡə], think [θiŋk]. Within the word ng is
either [ŋ] (singer) or [ŋɡ] (finger) (or [nʒ] (strange)).
Before voiceless sounds [ŋ] often loses its ‘voice,’ at
any rate towards the end, as in length [leŋθ].
§ 134. In French [ŋ] is not recognised as an independent
sound. But there is a palatal nasal [ɲ],
never formed farther back than the line between
hard and soft palate ; it is often followed by [j],
becoming [ɲj]. It is written gn (when final, it
is always followed by e) : régner [reɲe], Cologne
[kɔlɔɲ]. It sometimes becomes partly voiceless
before voiceless sounds, as in enseignes-tu [ɑ̃sɛɲʹty].
By assimilation we obtain the velar nasal sometimes,
as in une longue main [yn lɔ̃ŋ mɛ̃].
§ 135. In German [ŋ] is written ng, and also n
before k and — in foreign words — g : lange [lɑŋə], lang
[lɑŋ], denken [dɛŋkən]. In foreign words g before n
is sometimes pronounced [ŋ] : Agnes [ʔɑŋnɛs], Signal
[ziŋʹnɑːl] ; better [ʔɑgnɛs, ziʹgnɑːl].
In northern German the French nasal vowels
in borrowed words are often pronounced thus :
[ɑŋ] for an and en, [ɛŋ] for in, [ɔŋ] for on, [œŋ]
for un. In southern and middle German and
on the stage the French vowel is usually retained.
See above, § 74.92
Teeth or dental nasal :
§ 136. In English [n] the stoppage is the same as
in English [d, t]. Before and after voiceless sounds,
as in hint, sneer, [n] often loses its ‘voice’ in part.
It is written n (nn) : nine [naĭn], manner [mænə].
§ 137. In French [n] the stoppage is the same as
in French [d, t]. It is written n (nn) : neuf [nœf],
tourner [turne], bonne [bɔn]. Long, or double,
[n] is pronounced in ‘learned’ words, inné [inːe,
inne] ; and in proper names.
When final and before a consonant (except n) the
n, together with a preceding vowel, indicates a nasal
vowel ; but in ‘liaison’ the [n] appears, in which case
the vowel is less strongly nasalised : un [œ̃], but un
ami [œ̃n ami], almost [œn ami]. Often the preceding
vowel is then not regarded as nasal at all, and [yn
ami] is pronounced.
In foreign words final n often has the value of [n],
as in amen [amɛn], hymen [imɛn]. Rarely there is
a nasal vowel within the word before a vowel or
before a second n, as in enivrer [ɑ̃nivre] (also
[anivre]), ennui [ɑ̃nɥi]. Sometimes [n] loses its
‘voice’ before or after voiceless sounds : (je) ne sais
pas [ṇ se pɑ], hanneton [anṇtɔ̃], des tenailles [de tṇɑːj].
§ 138. In German [n] the stoppage is the same as
in German [d, t]. After voiceless sounds it loses its
‘voice’ wholly or in part. It is written n (nn) : nein
[naĭn], nennen [nɛnən].93
Lip or labial nasal :
§ 139. In English [m] is bilabial, like [b, p]. It
is written m (mm) : my [maĭ], summer [sʌmə], name
[neːĭm]. It loses its ‘voice’ towards the end — compare
[ŋ] and [n] — before voiceless sounds, as in lamp
§ 140. In French [m] is bilabial. It is written m
(mm) : mon [mɔ̃], blâmer [blɑːme], ferme [fɛrm]. In
some ‘learned’ words and proper names mm is pronounced
as a ‘double,’ i.e. long, sound : immortel
[imːɔrtɛl]. It serves to indicate a nasal vowel, like n.
Before a second m there is sometimes a nasal vowel,
as in emmener [ɑ̃mne] ; and final m has the value
[m] in foreign words and proper names : Jerusalem
[ʒeryzɑlɛm], album [albɔm]. Voiceless [ṃ] is the
rule in certain combinations, such as sme : prisme
[prisṃ] ; it also occurs occasionally elsewhere : monsieur
[ṃsjø], hameçon [aṃsɔ̃], (perhaps it is here
§ 141. In German [m] is bilabial. It is written m
(mm) : mein [maĭn], kommen [kɔmən], ihm [ʔiːm].
It may become partly voiceless after initial sch [ʃ], as
in Schmerz [ʃmɛrts], or in foreign words after s, as in
Smaragd [smɑʹrɑkt], or t, as in Tmesis [tmeːzis].94
lips | lip-teeth | teeth | front—of tongue—back | uvula | glottis | English | continuants | liquids | stops | nasals | French | German95
[List of Contents]
Sounds in Connected Speech.
§ 142. So far we have considered each sound
separately. If, however, we consider all the sounds
of a language, comparing them with all the sounds of
any other language, we find that there is something
characteristic in the mode of articulating them.
We might say, that they all presuppose the same
basis of articulation ; and this we shall have to
determine for each language. Again the sounds of
a language do not occur separately, but as members
of little groups, which again form the units of larger
groups ; and we must examine the ‘timbre,’ length,
force and pitch of each sound, as compared with other
sounds ; as well as the relative loudness (audibility),
which plays an important part in the formation of
syllables. Finally we shall consider the changes of a
sound which result from its position, according as it
is influenced by a preceding or following sound, or
The Basis of Articulation.
§ 143. We can only suggest the various bases of
articulation by showing in what respects those of
other languages differ from our own.98
Strictly speaking there is no uniform basis of
articulation for all English speaking people, nor
for Frenchmen, nor for Germans, just as the
pronunciation is not uniform ; we shall therefore
single out the speech of educated natives of
London, Paris and Berlin, which will enable us
to notice some characteristic differences.
§ 144. The French mode of articulation is more
definite, more tense 1 than ours : the tongue is in
general much farther forward in the mouth. The
lips are very active : they are strongly rounded or
protruded, or the corners of the mouth are well
drawn back ; and the mouth is smartly opened.
The timbre of the voice is bright and clear ; and
there is enough modulation to make us easily distinguish
the musical intervals. The exhalation of
breath is more uniform than in English (or German),
and indeed tends to increase in force as it goes on.
Hence there are no lax vowels, in particular
no lax [U], [I] and [Y] sounds (for
the half-wide weak [ɑ], [o], [e] and [ø]
sounds, compare § 166) ; also no lax consonants.
On the other hand the [d, t]
sounds and [z, s] sounds are farther forward
than in English ; the [l] is also farther
forward and has a clearer sound than ours ;
and front [ɲ] takes the place of back [ŋ]. —
A further peculiarity is the nasalising of
certain vowels, the passage to the nose not
99being closed by the velum ; and the presence of
§ 145. In German the tongue is on the whole not
quite so forward in the mouth as in French. The
mode of articulation is between that and the English :
it is more definite and precise than our own, which
appears indolent to French and German ears. The
lips are also distinctly more active than in English ;
the angle of the jaws is greater. The expiration is
fairly uniform ; and there is more modulation of the
voice than in English.
Hence there are more rounded vowels, and
there is less variety of indefinite (neutral) vowels.
The [d, t] sounds and [z, s] sounds are farther
forward ; the [l] is also farther forward, and has a
clearer sound than ours ; the [ʃ] is strengthened
by protrusion of lips. A further peculiarity is (at
any rate in middle and northern Germany) the
frequency of front and back continuants ; and
the presence of uvular [ʀ].100
The Sounds in their Relation to
§ 146. We distinguish three fundamental properties
of the sounds of speech : length (duration, quantity),
force (stress, emphasis, ‘dynamic’ accent), and pitch
(intonation, ‘musical’ accent).
When the formation of a sound is perfect, the
force of the expiration is identical with the force
of the mouth articulation ; there is an ‘equilibrium
of the active and reactive forces of articulation’
(Techmer). In order to maintain this equilibrium,
long sounds are as a rule more tensely
(firmly) articulated than short sounds, i.e. the
long are tense, the short are lax. Again,
when the expiration is strong, the pitch is
higher ; when the expiration is weak, the pitch
is lower. Still, as deviations from the normal
inter-relation of these fundamental properties
are not uncommon, length, force and pitch are
best considered separately.
§ 147. If we regard the sounds in the speech of an
individual as a whole, it will depend on their length
whether he speaks fast or slowly (‘drawls’) ; it will
101depend on their force, whether he speaks in a loud
or low (‘inaudible’) voice ; it will depend on their
pitch, whether he speaks in a high or in a deep
voice. A clear and distinct speech depends on all
three, but principally on the force with which the
sounds are uttered. — The peculiar character of each
voice (which enables us to distinguish individuals by
their speech) is largely due to the quality of the
larynx, and partly due to the basis of articulation
(see § 142).
§ 148. Connected words are uttered in groups,
according to the length and force of each expiration.
When the groups depend on the length of the
exhalation, they are also called breath groups (French
‘groupes de souffle,’ German ‘Atemtakte’ or ‘Atemgruppen’) ;
the line of separation between two consecutive
groups lying between the end of one
expiration and the beginning of the next.
When the groups depend on the force of the
exhalation, they are called stress groups (French
‘groupes de force,’ German ‘Nachdruckstakte or
-gruppen’) ; the line of separation between two consecutive
groups lying at a point where the stress,
after falling to a minimum, begins to rise again.
In practice the breath groups and stress groups
cannot always be considered separately ; they
may either coincide or not ; and further,
there is variety in the usage of individual
speakers. The division of connected speech
into groups is therefore to a large extent
subjective, depending on each individual ; except
102in so far as these groups coincide with
sentences or clauses (or even single words) which
are separated by pauses (of breath, and at the
same time logical ; usually indicated in writing
by means of punctuation).
§ 149. The length of sounds is here indicated by the
addition of [ː] or of [ˑ] to indicate half-length ; the force
of sounds is generally only indicated where a syllable
has the chief stress, here by [ʹ] preceding the first
sound of that syllable.
The pitch of sounds is sometimes indicated by [∕]
for rising and [∖] for falling tone, [∧] and [∨] representing
combinations of these.
The general tone of voice is sometimes indicated
by [⌈] for a high and [⌊] for a low voice. The rate
of speech and the degree of loudness is not usually
The ‘groups’ are in some transcriptions separated
by a vertical line [|] ; in others the words are either
not separated or separated by hyphens [-], and blank
spaces are left between the groups.103
§ 150. For general purposes it will suffice to use the
old, purely relative, distinction of short sounds, e.g.
[u], and long sounds, e.g. [uː] ; now and then we
shall also require the term half-long, indicating length
intermediate between ‘short’ and ‘long.’
Although the term ‘length’ or ‘quantity’ is generally
applied to vowels only, we find differences in
length in the ease of consonants also, including even
the stops (sometimes called ‘momentary’ sounds).
§ 151. The length of a stop is the time that elapses
between the formation of the closure and its opening.
Only when the stop is at the beginning of a word
can we really say that the sound is ‘momentary’ ;
for in this case the formation of the closure is not
audible, and the sound as such does not come into
existence until the closure is opened. On the other
hand, we may (it is not essential) produce a stop at
the end of a word without any audible opening of
the closure ; in that case the sound as such results
from the formation of the closure, which is audible
when it follows another sound.
Long stops easily give the impression that they are
double sounds, especially owing to the fact just
mentioned, that under certain circumstances the
104formation of the closure by itself and the opening
of it by itself may have the value of a stop, and also
to the fact that in long stops the closure is usually
opened with increased force.
Long continuants and vowels may also be regarded
as double when there is an increase of force in the
course of their production ; see the examples in
§ 152. In English the rules for the quantity of
vowels are, according to Sweet :
Long vowels only occur in syllables with strong
stress (rarely where stress is not quite strong) ;
finally, e.g. [əː] in err, fur, and before voiced consonants,
e.g. [ɑː] in hard, [ɔː] in all. Except these
three [ɑː ɔː əː] there are no simple long vowels ; they
are followed by [ə], or by [w] (in the case of ‘long
u’), by [j] (in ‘long e,’ as in seen), by [ŭ] (in ‘long
o’), by [ĭ] (in ‘long a,’ as in pale). The second
element of these diphthongs shares the length of the
first, so that in the [uːw] of who, the [iːj] of he, the
[oːŭ] of rode, and the [eːĭ] of pale, each element is
Before voiceless sounds the long vowels [ɑː ɔː əː],
as well as the diphthongs [uːw iːj oːŭ eːĭ] are reduced
to half-length ; and in syllables which have not the
105chief accent they are also half-long, or short.
Examples : [ɑː] in hart, [oːŭ] in wrote, [ɔː] in
When a ‘short’ vowel precedes a final voiced
consonant, it often becomes half-long ; e.g. bad, his.
In slow and emphatic speech the final [e˔] in pity,
steady, etc., and sometimes the [ə] in better, etc.,
are half or quite long.
Accented short vowels occur in final syllables (or
in monosyllables) before voiceless consonants, as in
bat, hiss ; often also in other than final syllables, as
in better, natural, etc. Unaccented short vowels are
The relative value of ‘long’ and ‘short’ vowels
in English has not yet been satisfactorily determined.
§ 153. In French the following rules have been
given by P. Passy :
Vowels are long in syllables with chief accent
(before a pause) before the voiced continuants
[r z ʒ v j]; e.g. [ɛː] in faire, [yː] in ruse, [aː] in
nage, [iː] in rive, [œː] in feuille ; similarly where
-je is added to a verb with final vowel, as in dis-je
[diːʒ]. The sounds [ɑ o ø], and the nasal vowels are
long before other consonants also. But before the
consonant sound due to ‘liaison’ the vowel is not
lengthened : pas ici = [pɑz isi] with short [ɑ] ; in
the same way, méchante is [meʃɑ̃ːt], but méchant et
vilain is [meʃɑ̃t e vilɛ̃] with short [ɑ̃]. — Before consonants
that are not voiced continuants the other
106vowels (except [ɛ]) are rarely long, even when a
pause follows. All final vowels are also short as a
§ 154. In German the vowels of accented syllables
are long or short.
The long vowels (except [ɑː] and [ɛː] they are all
close) occur finally, with or without accent, e.g. [ɑː]
in ja, da, Emma. The only final vowels that are
short are the unaccented [ə], and [ɑ] in interjectional
na, da, ja. The vowels are long also before a single
consonant or before a combination of consonants
which could begin a word, e.g. [iː] in ihr, Mitra
(the syllable is open when it ends in a vowel,
as in ihrig = ih-rig, Mitra = Mi-tra). It is less
common for a vowel to be long when a combination
of consonants follows which could not begin a
word, e.g. [oː] in Mond (then the syllable is closed and
remains so even when a vowel follows the consonants :
Mon-de ; in many cases, however, such syllables were
open in an earlier stage of the language).
In syllables with secondary accent, virtually long
vowels become half-long or even short, e.g. [e] in
Sekretär, [i] in Militär.
Accented short vowels, which were short also in the
earlier stages of the language, as [ɛ, ɔ], are almost
confined to ‘closed’ syllables, and are ‘open’
In the ordinary spelling the consonant is doubled
after a short vowel within the word, and often finally,
as in füllen, satt ; and the length of a vowel is often
indicated by adding h, by doubling the vowel, or —
107sometimes in the case of i — by adding e : Mehl,
That, Saat, dieser, etc.
The relative duration of German ‘long’ and ‘short’
rowels has been variously estimated ; it naturally
differs according to dialects and even individuals.
Brücke obtained the proportion 5:3, Kräuter 3:2,
Ph. Wagner (Reutlingen dialect) 3:2.
Viëtor's measurements of his own speech (Nassau)
gave the following results :
Relation of long to short 2 (average 0.3 second) :
1 (average 0.15 second) ; and he arrived at the following
results, taking this scale : 0.1 second, over-short ;
0.15 second, short ; 0.2 second, half-short ;
0.25 second, half-long ; 0.3 second, long ; 0.35 second,
The ‘long’ vowels are — I. ‘over-long’ in final
syllables with the chief stress before a pause (1) when
not followed by a consonant, (2) when followed by a
single consonant (or by a liquid and a consonant) ;
II. ‘long’ (a) in syllable with chief stress (1) when
the syllable is final and the vowel is followed by a
combination of consonants, (2) in the last syllable
but one, or in a syllable that was formerly ‘open,’
and (b) in the syllable following a syllable with chief
stress and ‘over-long’ vowel ; III. ‘half-long’ (a) in
syllable with chief stress, preceding the last syllable
and ending in a combination of consonants, (b) in
the syllable following a syllable with chief stress
and ‘long’ vowel ; IV. ‘half-short’ if in a syllable
preceding the chief stress.
The ‘short’ vowels are — I. half-short’ in syllables
with chief stress (1) in last syllable before a single
108stop, (2) in last but one before a double (i.e. long)
stop ; II. ‘short’ (a) in syllables with chief stress,
(1) in last syllable when no consonant follows, (2) in
last but one, or in a formerly ‘closed’ syllable, (b) unaccented,
final e [ə] ; III. ‘over-short’ : accented
in last syllable, before liquid or liquid and consonant.
The diphthongs are on the whole treated like the
§ 155. In English there is a marked difference
between long and short consonants. According to
Meyer, final stops after long vowels are decidedly
short ; the shortest final is [d] after a long vowel : it
may be called over-short. The length progressively
increases, not only when the preceding vowel becomes
short, but still more when the consonant ceases
to be explosive, ceases to be voiced, and ceases to be
dental (apical). But [l, m, n, ŋ], though voiced, are
always long, and over-long after a short vowel. The
prolongation of these consonants before voiced final
consonants (build, pens) is still greater ; but before
voiceless final (built, pence) they cease to be long,
without becoming short. When by the addition of
an unaccented vowel, the word becomes dissyllabic,
and the consonant medial, all these differences shrink
considerably. Initial consonants, if single, are nearly
always of middle length.
Long consonants also occur where words are closely connected
or compounded, as in wild duck, ripe pear : here the
formation of the closure is the ‘first d (p)’ and the opening of
it is the ‘second d (p)’ ; or in slim man, full list, with them :
here the sound gains a fresh value through an increase of force,
and the consonant becomes ‘double,’ or the [m, l, ð] are simply
lengthened. In somewhat careless speech the simple consonants
§ 156. In French (according to Ph. Wagner) the
consonants are long or over-long before a pause,
the quantity of the preceding vowel being immaterial.
In cases like mont(ent) tous, la nett(e)té, la d(e)dans,
the [tː, dː] are over-long, and long when they are
some distance from the final syllable of the group,
as in le temps d(e) transmettre (d(e)t = [tː]). Long
also is the initial [sː] in sais pas (for je ne sais pas),
and [tː] in temps en temps (for de temps en temps).
In cases like illégal, illusion, etc., and irrité, irrégulier,
etc., there appears to be no lengthening
(Ph. Wagner). Rousselot gives the following rules
as the result of his own experiments :
Stops are a little shorter than continuants. Voiced
sounds are often shorter than voiceless sounds. The
length of the consonants is in inverse proportion to
the length of the word. There is an ‘accent of time
or duration’ which lengthens certain consonants. —
The ‘doubled sounds’ are single consonants with
greater force and double length.
§ 157. In German the length of consonants (especially
continuants, including nasals) is to some extent
affected by their position ; e.g. [m] is short initially,
but long when it is final and follows an accented
Long consonants also occur (as in English, compare
§ 155) when words are closely connected or
compounded ; e.g. mitteilen, Packkorb and Tauffeier,
Baummeise, Stillleben. Here, too, the simple
consonants are substituted in careless speech.110
§ 158. Tie sounds and small or large groups of
sounds in ordinary speech are not uttered with, uniform
force ; there is either a decrease [>] (decrescendo)
or an increase [<] (crescendo) in the force with
which the breath is exhaled ; and the force with
which the consonants are articulated (i.e. the force
employed for forming a ‘narrowing’ or ‘closure’)
usually decreases or increases in the same proportion
(see § 146). We can therefore distinguish various
degrees of force if we compare the parts of a sound
or of a group of sounds. This is more easily done if
we whisper the sound or sounds, as our ear is then
not confused by the addition of ‘voice.’ Accurate
results can obviously be obtained only by mechanical
§ 159. Differences in the force of exhalation are
most important in the case of vowels. A relatively
greater or greatest degree of force is called stress ; the
vowels concerned are stressed.
§ 160. If we take the stress of a syllable, we can
investigate the nature of this stress ; and, if the
syllable contains more than one vowel, we must
determine which has the stronger stress. If the
111first of two vowels has the greater sonority (loudness),
the diphthong is ‘descending’ ; if the second, the diphthong
is ‘ascending’ (see § 56) ; and ‘level,’ if they have
the same stress. The symbol for level stress is [=].
§ 161. We determine the stress of a word by finding
out the vowel in it which has the chief (principal,
strongest) stress ; the other vowels either have
secondary stress, or they are unstressed. In accordance
with this, we distinguish ‘strong,’ ‘half-strong’
and ‘weak’ syllables. When several syllables have
the same stress there is ‘level stress.’
When a language shows a wide range between
the extremes of ‘strong’ and ‘weak’ expiration,
the unaccented vowels tend to change into the
indefinite vowel sounds [ə], etc. On the other
hand, there is considerable variety in the force
of the secondary stresses ; and the formation of
compound words is easier.
§ 162. There is further the stress of the sentence, obtained
by ascertaining the most strongly pronounced
vowel of the sentence ; and there are also secondary
stresses of varying force.
§ 163. This stress of words and sentences serves
to emphasise certain ideas ; it is ‘emphatic’ stress.
This explains why the stress of the word must yield
to the stress of the sentence. Some words (enclitics
and proclitics) lose their stress under certain circumstances.
§ 164. The force of expiration is a matter of some
importance in the case of consonants. It generally
112depends on the degree of force of the expiration
possessed by the preceding or following vowel at the
moment when the two sounds meet.
Among the consequences of strong expiration
are : the change or ‘shifting’ of voiceless stops to
aspirates and affricates, e.g. [p] to [pʰ] and [pf] ;
the shifting of voiced sounds to voiceless sounds,
e.g. [b] to [p] ; the development and retention of
distinct [h] sounds ; the development and retention
of the glottal stop (compare § 32) ; partly
also the retention of final ‘unaccented e.’
§ 165. English. Taking first the stress of the
syllable, we find that generally the diphthongs are
descending’ or decrescendo diphthongs : [aĭ] = i,
[aŭ] = ou, [ɔĭ] = oi, [oːŭ] = o, [eːĭ] = a, etc.
In the word also the accentuation is usually descending ;
the chief stress falls upon the syllable
which is logically the most important, i.e. the stem
syllable, which usually comes first in the word. This
principle has been adopted even in the case of foreign
words ; compare the native holiness [ʹhoːŭlinis]
and the foreign memory [ʹmeməri]. In unstressed
syllables an original a, o, u, and also er, etc., regularly
becomes the ‘neutral’ vowel [ə] : Arab [ærəb],
There are many compound wards ; frequently they
are separated in writing. The rule used to be that
the chief stress was on the first part [>] ; according
to Sweet the following now holds good, if we consider
the accentuation of compounds with regard to
their meaning : level stress [=] contrasts, and uneven
113stress ([>], less frequently [<]) unites the ideas expressed
by the compound words.
Hence (according to Sweet) there is level stress
(a) in the combination of substantive and substantive,
the first having attributive force : steel pen, garden
wall (but > in the case of natural objects when the
two ideas have become blended into one : butterfly,
blackbird (but black bird with =) ; similarly compounds
with street have > accentuation, but those
with road, square, etc., =) ; (b) adjective (or adverb)
and adjective : good-looking, twenty-five ; (c) inseparable
compounds in which the second part has
a clearly marked meaning : undo, thirteen ; (d) interjections :
hallo, bravo ; also in the foreign amen ;
(e) foreign names : Hongkong, Chinese.
The > accentuation is found : (a) when the compound
denotes a casual relation, an action or a
natural phenomenon : rainbow, walking excursion,
earthquake ; (b) in an otherwise = compound, if it
is employed as an attribute before a substantive :
The < accentuation is found (a) in ‘constant’
combinations of substantives with of : bill of fare ;
(b) in title and name : Mr Smith ; (c) in forms of
address : good morning !
There is often aspiration of initial and final voiceless
stops before and after the vowel which bears the
chief accent ; thus two [tʰuːw], not [notʰ].
Initial [h] is weak, and often in danger of being
lost ; it regularly disappears, even in educated speech,
in the ‘weak’ forms of certain monosyllables (see
The final ‘unstressed e’ has long been lost ; and
the final [ə], even where it is not = er, or, etc., in
uneducated speech often takes a [r] sound : idea of
[aĭdiːər əv]. This is now sometimes heard even in
Enclitics are common ; here a, o, u (and sometimes
other vowels) are reduced to [ə]. This leads to
numerous pairs of (‘strong’ and ‘weak’) forms :
and [ænd] and [ənd] or [ən], from [frɔm] and [frəm],
us [ʌs] and [əs], etc.
The accentuation of the sentence has also a logical
basis. But — particularly in educated speech — there
is not a very great range between the extremes
(chief stress and no stress), so that sometimes
subject and predicate (or even subject, predicate
and object) appear to have the same accent, and the
sentence appears to have = rather than < accentuation
(which is clearly the case in attributive adjective
and substantive). The syllables with secondary
stress very often become quite unstressed.
Where there is a contrast, a strong stress is regularly
employed to mark the contrasted words.
There is, besides, a distinct tendency to rhythmic
accentuation in polysyllables, in compounds, and in
§ 166. In French the stress is far less strongly
marked than in English ; and the tendency as regards
its movements is directly opposed to the Teutonic.
In the syllable there are the ‘ascending’ diphthongs
[wɑ, wa] = oi, [je] = ie, [wɛ̃] = oin, [jɛ̃] = ien,
etc., with < accentuation.115
In the word there is the same ascending accentuation ;
the last sonorous syllable bears the stress ;
jamais [ʒaʹmɛ], pardonner [pardɔʹne], considération
[kɔ̃siderɑʹsjɔ̃]. This is the rule in theory ; but in
practice other influences frequently cross it. The
first, or some other preceding syllable usually has a
secondary stress, which becomes particularly obvious
to the ear when it is combined with a higher pitch of
the voice. It is not uncommon for this stress to
have greater force even than the normal stress of
the word ; this is especially the case when the last
syllable but one is long : thus baron is [ʹbɑːrɔ̃] rather
than [bɑːʹrɔ̃], and beaucoup is [ʹboːku] rather than
We find very much the same influences at work in
the sentence. The rhetorical element is far more
important than in English ; and the rhythmic tendency
is very marked. For instance, in le roi Jean
the stress is on Jean, but in le roi Théodoros the
syllables roi and -ros are stressed.
Where there is a contrast the stress shifts to the
syllables which mark it : se soumettre ou se
démettre, with strong stresses on sou- and dé-.
The absence of stress causes long vowels to
be reduced to half length (e.g. rase and raser),
and changes the short [ɔ, a, e, ɛ] to sounds
approaching [ə], as in the vowels preceding the
accent in comment, mardi, méchant ; when the
syllable is weaker still these vowels (as also [ø]
and [œ], and occasionally others) actually become
[ə], as in peut-être, déjeuner. This [ə], unless its
position preserves it, may drop out altogether (as
116[ə] of le in je le dis) ; on the other hand when it is
accented it may become a full [œ] (as e in dis-le).
— Consonants also frequently disappear in weak
syllables ; thus in ordinary conversation the 1 of
il is dropped before a consonant. Instances of
‘strong’ and ‘weak’ forms are, e.g. le [lə] and
la [la] before consonants, and l' [l] before vowels.
§ 167. In German there is mainly > stress in the
syllable as well as in the word.
The stress of a long vowel in an ‘open’ syllable
assertion | question | anger | warning.
(e.g. du) has been ascertained by means of the
cymagraph. The results are shown in the accompanying
diagram, which gives the characteristic
curves for the word du according as it is pronounced
in a tone of (1) assertion, (2) question, (3) anger, (4)
The diphthongs [aĭ] = ai, ei, [aŭ] = au, [oĭ] = eu, äu
are stressed on the first element.
In the word (in almost all cases where it is really
German) the stress is on the stem syllable, which is
117logically the most important, and — as in English —
this is the first syllable in simple (uncompounded)
words : Héiligung, árbeitete ; exceptions : Forélle,
Holúnder, Hornísse (but also Hórnisse), lebéndig,
Wachólder ; also luthérisch. There is a secondary
stress on the second syllable of Élènd, élènd ; both
[ʔeːlɛnt]. Foreign words usually retain their original
accent : Dóktor, pl. Doktóren, Hotél ; old loan words,
or foreign words that are much used, receive German
accentuation : Fénster, Dútzend, Kógnak. Foreign
suffixes, e.g. -ei, -ieren, -ur are stressed, even when
they have been added to German stems : Betteléi,
hausíeren, Glasúr. The foreign accentuation is also
found in the originally Slavonic names in -in of some
north German towns : Berlín.
The stress varies in some words which are frequently
contrasted with certain other words that have
the same or similar terminations : Síngular-Plúral,
Ínfanterie-Kávallerie, instead of Singulár, Infanteríe,
etc. Similarly, an otherwise quite unstressed
syllable may receive the chief stress, if it is to be
particularly emphasised, especially for purposes of
contrast : gégàngen [ʹgeːgɑŋən], nicht vérgàngen.
There is a greater range between the extremes
(chief stress and no stress) than in English. The
syllables with ‘chief stress’ are very strongly marked.
On the other hand the unstressed vowels have in
many cases passed into ‘unstressed e,’ i.e. [ə]. The
great range of accentuation makes it possible to
have much variety of force in the secondary stresses ;
hence German lends itself particularly well to the
formation of long compounds.118
The first element of compound words usually has
the chief stress, as being the determining word, as in
Sónntagshèiligung ; the stress in Jàhrhúndert follows
the same logical principle. Level stress is much less
common : stéinréich ; it usually indicates that the
two parts of the compound are logically equal.
Ascending stress is found in the following cases :
In substantives : (a) in words like Hòhepríester
(for hohe Priester) ; (b) in geographical names and in
names of festivals, etc., the first part of which is an
adjective or a genitive : Nèuwíed, Èberswálde (but
Kárlsbàd and some others) ; also Nòrdóst, etc. ; and
Grũndónnerstag, Frònléichnam (Fron originally ‘of
the Lord’) ; (c) in many compounds, the second part
of which is itself a compound and ‘heavier’ than the
first part : Òberfórstmeister (but Óberfö̀rster) ; (d)
in compounds which have the chief accent far from
the end : (sometimes) Làndgeríchtsdirèktor, as though
it were Land + Gerichtsdirektor (compare (c)), and
not Landgerichts + Direktor.
In adjectives : (a) in words like bàldmöglich (for
so bald wie möglich), and in derivatives like hòhepríesterlich ;
(b) when the first part is all- (giving
emphasis) : àllgeméin ; (c) when the first part is alt-,
and the second refers to a nation or country ;
àltíndisch. (but usually áltdèutsch, and áltfrä̀nkisch
in the sense of áltmòdisch) ; (d) in several adjectives
in -lich and -ig (in most cases the middle syllable is
‘heavy’) : àbschéulich, bàrmhérzig (in some the
accentuation varies, e.g. absichtlich, armselig) ; (e)
un- + verbal adjective : ùnhö́rbar ; (f) in lèibéigen,
vòllkómmen, wìllkómmen ; also in àusgezéichnet.119
In verbs : (a) with miss- : mìssbráuchen (míssdeùten
is rare) ; (b) with inseparable voll- : vòllénden ;
(c) with inseparable durch-, hinter-, über-, um-, unter-,
wider- : dùrchdríngen, hìnterbríngen, ǜbersétzen,
etc. (in the verbs with separable and accented durch,
etc., the prefix is still distinctly adverbial ; thus
dúrchdrìngen = durch dringen, hindurch dringen).
In particles (instead of level stress) : bèrgáuf, vòrhér.
Level stress is found in the following cases :
In substantives : (a) in names like Héssen-Nássau
(for Hessen und Nassau), but here frequently < ;
(b) in compounds with Erz-, where this prefix merely
serves to give emphasis and has lost its original force
(= Greek archi-) : Érzschélm (but Érzbìschòf) ; (c) in
long compounds, the two parts of which are themselves
compounds : Rèálschùlóberlèhrer.
In adjectives : (a) when the first part merely
serves to emphasise the second : stéinréich (see above) ;
(b) when the word is so long that one accent seems
insufficient : únwìederbrínglich (but also ùnwìederbrínglich).
The accentuation of particles is usually =, but
sometimes becomes < or >, owing to the rhythm of
neighbouring words : bérgáuf and bèrgáuf ; wéitáus,
wèitáus, and wéitàus.
Such variation occurs in adjectives also : er ist
stéinréich, but often ein stéinrèicher Mann. On the
other hand words naturally accented > or < may
receive level stress if they are strongly emphasised ;
eine fúrchtbáre Kā́lte (Kā́lte) ; in such cases > may
also become < ; in the last example we might also
These are examples of a tendency to rhythm ; as
other instances we may take words like Landgerichtsdirektor
and Realschuloberlehrer mentioned above.
Owing to the force of the word stress, the voiceless
stops initially before, and finally after, a stressed
vowel pass over into aspirates : Kind [kʰintʰ], tot
[tʰoːtʰ], Pack [pʰɑkʰ]. The force of the stress also
explains the energetic initial [h], as in Halt [hɑlt],
hier [hiːr], and the retention of the glottal stop :
echt [ʔɛçt], alle [ʔɑlə]. This applies to the secondary
no less than to the chief stress : Schósskìnd with
[kʰ], Ánhàlt with distinct [h], únècht with [ʔɛ].
Words often lose their stress in the sentence ; and
some (enclitics and proclitics) have ‘strong’ and
‘weak’ forms according to the force of the stress :
er = [ʔeːr] or [ʔɛr] or [ʔər] or even [ər] ; der = [deːr]
or [dɛr] or [dər], etc. An extreme case of this is the
complete loss of words, e.g. Mórgen ! for guten
In the sentence we observe an accentuation which
obeys very much the same principles as that of the
word. Often the groups of a sentence (see § 148)
coincide with the words, and then have > accentuation.
But taking the sentence as a whole the <
(crescendo) is the rule ; for the subject usually goes
before the predicate, and it is the predicate which, as
a rule, supplies the ‘determining’ element of the
sentence and therefore has the chief stress, while the
subject usually has a secondary stress : “ich schréibe,”
“der Hùnd béllt,” “er ist kránk.” If the predicate
has an object, then the object is the ‘determining’
121word, and therefore has the chief stress of the
sentence : “ich schrèibe einen Bríef.” In compound
tenses, in inverted clauses, and in secondary clauses,
the accentuation changes also, and < becomes > or
< >. The attributive adjective has a rather weaker
stress than the word to which it belongs : “die
kìndliche Líebe” (but “die Lìebe der Kínder,” “zu
den Kíndern”). Where there is a contrast, the
emphasis on the contrasted words counteracts the
usual tendencies of sentence accentuation : “ist ér
krànk (oder síe) ?”122
§ 168. All voiced sounds necessarily have a certain
musical pitch (or intonation) ; and, though this is
most important in the case of vowels, the term
‘voiced sounds’ here, as usually, includes voiced
In ordinary speech the voice does not remain for
any appreciable time on one note, but is continually
gliding to a higher or lower note. If we dwell on
each separate sound, giving it a uniform musical
value, speech becomes a recitative or a song.
§ 169. Uniform tone does indeed occur in speech,
especially in the short vowels ; we here have to
distinguish at least three varieties, a high tone [—], a
middle tone [—] (or no symbol), and a low tone [—].
But far more frequent — not only in the sentence
or in polysyllables, but also in the syllable — is rising
[∕] or falling [∖] tone, and of compound tones the
commonest are the rising-falling [∧] and the falling-rising
These terms correspond very roughly to the
actual facts. The musical interval between the
highest and lowest note in a case of a rising or
falling tone may be of very varying length ; and
it may not be a simple rising or falling, but a
series of rises and falls, which can only be
123accurately determined by mechanical means, the
human ear being unable to distinguish more
than the general result of these infinite variations
Apart from the intonation of the sentence we must
take into account the tone of the voice as a whole,
which may be high [⌈], middle, or low [⌊] ; in the
case of women and children the tone of voice is
higher than in men (see § 10, note 3).
§ 170. In English the simple rising or falling tone
has the same value as punctuation in written speech.
The falling tone is used for making or confirming a
statement ; the rising tone asks a question or prepares
for the making or confirming of a statement ;
in the second part of an alternative question there
is falling tone. The compound tones on the other
hand usually serve to express various states of emotion ;
the rising-falling tone suggests stubbornness
or surliness, and we use the falling-rising tone in order
to give a warning. Certain monosyllables, such as
yes, no, hm, oh, etc., may represent a whole sentence,
and the intonation of that one sentence will be
crowded into the tone in which the monosyllable is
The range between the highest and lowest note
varies according to the dialect of the speaker, his
temperament, and the particular emotion to which
he is giving expression. It may be as much as two
octaves in the same syllable ; sometimes the most
important members of a sentence do not differ by
124more than half a tone, or are even quite of the same
The following diagram is a facsimile of the cymagraphic
representation of du, showing the waves of sound.
§ 171. In French the intonation is more musical ;
the notes are clearer and the intervals more distinct.
In the use of falling and rising tone, etc., French
agrees on the whole with English ; the compound
tones are particularly common. The average pitch
of voice is higher than in English.
§ 172. In German the intonation is very similar to
our own. There is rather more variety of musical
tone ; and the average pitch of the voice is higher,
though not as high as in French.125
Sonority ; Formation of Syllables.
§ 173. The degree of sonority (loudness) of a sound,
or in other words the distance to which it will carry,
depends on various factors : whether there is ‘voice’
or not ; whether the ‘voice’ is the more or the less
important part of the sound — which again depends on
the accompanying resonance or noise (friction, etc.) ;
and if there is no ‘voice,’ whether the noise is more
or less distinct — which depends on the articulation
§ 174. The vowels possess sonority in the highest
degree ; for they are voiced, have resonance, and the
passage through the mouth is free. As the passage
is widest for [ɑ], this sound has greater sonority than
any other. Then come the voiced liquids and nasals ;
they are voiced sounds with resonance and with no
great narrowing of the mouth passage, or with free
passage through the nose, the mouth being closed.
After these are the voiced, and then voiceless continuants
and stops ; the voiceless [s, ʃ] being particularly
marked owing to their sharp sound.
§ 175. In sounds otherwise identical there may be
varying degrees of sonority according to the force of
expiration, the pitch, the length (especially in the
case of vowels), and the energy of articulation (in the
case of consonants).126
§ 176. According to its sonority a sound is more or
less distinctly audible. The least sonorous, viz.,
[k, t, p], could not be heard even at a short distance ;
[ɡ, d, b] carry a little further, and so on. Sounds
that are weak in this respect may become more
distinct if associated with more sonorous sounds ;
and thus the vowels, being most sonorous of all, are
not only themselves audible at a very great distance,
but can also make a number of preceding or following
sounds carry far.
§ 177. Apart from their own sonority, they therefore
form a kind of support for weaker sounds : they
may be said to carry the syllable. The term ‘syllable’
is also applied to an isolated vowel, or to a vowel
standing beside other vowels equally sonorous, sometimes
also to other sonorous sounds. If a syllablecarrying
vowel has had as much as it can bear joined
on to it, the rest is dropped, unless it can obtain the
help of some auxiliary vowel, or unless it possesses
sufficient sonority to become a syllable itself (compare
As a rule, it is vowels that ‘carry’ the syllables ;
but this can also be done by liquids and nasals (not,
however, in syllables with chief stress in English,
French and German, except perhaps in interjections
like the German brr [br]), and by the voiceless [s, ʃ],
also in interjections : [pst], [hʃ], etc.
§ 178. A consonant between two vowels (in English
and German) usually belongs to both syllables. It
is impossible, therefore, always to divide olf the
syllables according to their pronunciation ; and
127where the exigencies of writing or printing demand
it, we often find inconsistencies. Thus we
divide lead-ing, but in German it is lei-dend ; the
pronunciation of the d being the same in each
case, as it belongs to both syllables, the closure
to the first, the opening to the second.
If, however, there is a change of force, i.e. a
weakening and then again a strengthening of pressure
in the current of breath, we can be in no doubt
as to the division between the syllables ; for here it
coincides with the change of force. Thus the d in
English ado or German Geduld belongs to the second
Influence of other Sounds.
Assimilation and Dissimilation.
§ 179. Initial sounds are often influenced by the
sounds which follow, final sounds by those that precede,
and sounds occurring within the word by those
on both sides. In dealing with the resulting changes,
we must take into account that a sound may begin
or end a written word, and yet not be treated as an
initial or final sound, because it is within a breath
group. In I saw Tom and Jack d is the final of
and if we consider the word by itself ; but if we take
it as part of the breath group, we can no longer
regard it as a final.
§ 180. The general rule is, that when two sounds
come together, those movements of articulation which
are common to both are executed once only, as far as
that is feasible (Winteler).
§ 181. Sometimes certain movements of articulation
are made a little sooner or kept up a little longer
than the sound really requires, and this somewhat
modifies a preceding or following sound (or even
Thus a sound may become voiced (or voiceless)
through the influence of neighbouring voiced (or
voiceless) sounds ; a sound may be ‘mouille’ (palatalised)
129by the front of the tongue being raised
before the [i] sound is actually pronounced ; or
rounded (labialised) by the lips being rounded before
the [u] sound is actually produced.
A large number of the changes recorded in
Historical Grammars are due to these tendencies ;
they are to some extent reflected in
the conventional spelling, but become clear
only when the nature of the spoken sounds
§ 182. In English the most striking changes have
been caused by a following [r], now [ə(r)] ; compare
far, fare with fat, fate. The same tendency is still at
work, e.g. in poor [puːə, pɔːə]. Here there has been
change in the place and manner of articulation. Its duration
has also been changed in some cases ; compare,
for instance, the frequent lengthening of [ɔ] before
[f, ft, s] : off, oft, loss. Again short vowels before
final voiced consonants become half-long or long :
bad, dig ; and long vowels (particularly diphthongs)
become half long before voiceless consonants : fate,
poke. Place, manner and duration of articulation
are altered in the change from [æ] to [ɑː] before a
dental, as in past, dance, which is not yet universal.
The double forms [ði, ðə] = the, and [tu, tə]= to,
vary according to the following sounds ; the first of
each pair being used before vowels, the second before
As an example of dissimilation we may take the
130diphthongising of original long vowels which is still
in progress ([eːĭ] in day is going on to [ɛːĭ, æːĭ, ɑːĭ]).
Final stressed vowels are long or diphthongal :
ah, day ; unstressed [i] or [e'] and [ə] may become
half-long : steady ! never !
§ 183. In French the place wad manner of articulation
is also influenced by a following [r] ; and the
duration too, as in fer, mort, heure, except in -er with
liaison, as in premier âge. The sound [v] has a similar
effect to [r] : fleuve ; [z] and [ʒ] lengthen vowels :
ruse, neige ; so do [r] and [v] in certain cases, without
producing any other change : lire, prouve. Sometimes
a vowel is changed owing to the anticipation of
a following front vowel, the front of the tongue being
lowered too soon, as in [ɛtɛ], a common pronunciation
for [etɛ] = était, or being advanced too soon, as in
[ʒœli] for [ʒɔli] = joli.
When [u], [i] or [y] ends a word within a breath
group, and before a vowel, it usually changes to [w],
[j] or [ɥ], e.g. [i] in qui est là ? An [ə] in the same
position is dropped, as also the [a] of la ; in writing
this is, as a rule, only indicated in the case of la,
and in the monosyllables me, le, etc.
Other unaccented vowels also disappear no w and then.
The [ə] may be inserted in order to avoid the
clashing of three consonants, e.g. in porte bien [pɔrtə
bjɛ̃] ; arc boutant [arkə butɑ̃], l'arc de Triomphe
[larkə d triɔ̃ːf].
Final vowels are short (amie in pronunciation =
ami), though a little longer than short vowels preceding
§ 184. In German most long vowels undergo a
change in the place of articulation when followed
by r = [r (ər, ər, ə)]. In conversational German the
close [e] and [o] sounds become more open, as in her,
vor ; and the long [i, y, ø] become somewhat deeper
and duller, as in mir, für, Stör.
The short vowels are little affected by [r]. On the
other hand their quality depends very much on the
following consonants, which bring the vowel to a
sudden close. This is particularly obvious before
a final consonant, as in ab, where there is often no
distinct consonantal sound at all, but only the onglide.
There are no stressed vowels at the beginning of
German words ; the glottal stop always precedes.
With the exception of [ɑ] in the interjections da,
ja, na and of ‘unaccented e’ = [ə], all final vowels
are long : ja [jɑː], Anna [ʔɑnɑː], Juni [juːniː].
§ 185. In English there is often assimilation of
the place of articulation. Thus in [k, ɡ] the closure
is effected at various points of the tongue ridge and
palate, as in lack, lick, or log, leg. Frequently [s] or
[z] becomes [ʃ] before [ʃ], as in this sheep ; there is
partial assimilation in the change of [s] to [ʃ] before
[j], as in this year. Sometimes a nasal is partially
assimilated to a following sound ; thus [n] may become
[ŋ] in can go. — Most of these changes are confined
to rapid colloquial speech.132
The manner of articulation is variously affected by
neighbouring sounds. When two identical stops
meet, as in coat-tail, only one is spoken, and it is
long only in emphatically distinct speech. Such
combinations as [t] + [d], as in sit down, are often
reduced to a simple sound also, the second sound
being preferred as a rule. When two different stops
come together, e.g. [k] + [t], as in acting, the second
closure is usually made before the first is opened ;
this opening being then unaccompanied by an explosion.
In cases like open, the [n] closure is made
before the explosion of the [p]. When the stops [t]
and [k] are followed by [l], as in atlas, clean, they
often explode laterally, i.e. the opening is at the sides
only. When three consonants come together, the
middle sound often disappears, as in jumped, acts
(in careless speech) ; particularly when a related
nasal precedes and follows a stop, as in don't know.
Voiceless stops are aspirated before stressed
vowels ; but the aspiration is weak when another
sound precedes the stop, which is closely connected
with it ; compare take and stake, peak and speak.
Assimilation of voiceless to voiced, and voiced to
voiceless, sounds is a marked feature in the inflection
of substantives and verbs ; note the [d, t] in lagged,
lacked, and the [z, s] in bids, bits. The plural
forms in [vz], e.g. loaves, and in [ðz], e.g. baths, are
Before and after voiceless sounds, the liquids often
tend to lose their ‘voice’ partially, as in plot, pelt.
When voiced consonants occur initially they are
usually voiceless at the very beginning ; and when
133they occur finally they become voiceless at the very
end ; zeal almost = [sziːjl], is almost = [izs].
Final r is, as a rule, preserved by ‘liaison’ ; compare
there and there is, and see § 103.
§ 186. In French the difference in the place of
articulation of [k, ɡ] is more marked than in English;
compare cas with ‘guttural’ [k] and qui with
In the manner of articulation French shows features
which differ from our own. When two identical
stops, e.g. [tt], come together they are pronounced
long. When different stops meet, e.g. [kt], each has
its own explosion, as in acte.
When there is voiced + voiceless consonant or
voiceless + voiced consonant, we usually find progressive
assimilation, i.e. in the first case both become
voiceless, e.g. [bs] becomes [ps] in absolu, and in the
second case both become voiced, e.g. [sg] becomes
[zg] in second. — Liquids and semi-vowels tend to
adapt themselves to a preceding voiceless sound, e.g.
[pi, pj] becomes [pç] in pied.
Before a nasal consonant [b, d, ɡ] are often nasalised,
especially if a nasal vowel precedes ; e.g. [d]
becomes [n] in point de mire.
Initially we find long consonants as the result
of contraction ; thus [d + t] becomes [t + t] and then
[tː] in de temps en temps.
Finally the liquids [l] and [r], following a consonant,
lose their ‘voice’ and are often dropped, as
in peuple, chambre. Emphasis may lengthen a final
consonant, e.g. [lː] in seul !134
Many otherwise silent consonants are preserved by
‘liaison’ ; e.g. [t] in petit a petit, attend-il. The [l]
of il is — in ordinary conversation — only retained in
‘liaison’ ; compare il a and il va. Inconvenient
combinations of consonants at the end of words are
sometimes made easier by the addition of [ə], as in
§ 187. In German the effects of assimilation are
very similar to those in English.
Owing to a difference in the place of articulation
we distinguish ch [ç], as in ich, and ch [x], as in ach.
(Compare also the treatment of ɡ in lagen, liegen,
when it is pronounced as a continuant [lɑːgən, liːjən].)
The stops [ɡ, k] also vary according to the next vowel
sound, as in kund, Kind.
A [p] before [f], as in Pferd, is occasionally labiodental;
and [s] is sometimes assimilated to [ʃ] as in
Hausschlüssel. These changes are, however, confined
to rapid colloquial speech ; as also the assimilation of
nasals in cases like Anbau, glauben, which become
[ʔambɑŭ], [glɑŭbm], or Unkunde, merken, which
become [ʔuŋkundə], [mɛrkŋ].
As far as the manner of articulation is concerned,
German agrees with English in the following cases :
[t] + [t], as in not tun ; [tj + fd], 1 as in hast du ;
[k] + [t], as in Akte ; [t] + [l] and [k] + [l], as in
When there is a stop + related nasal, the explosion
of the stop in the oral passages is usually replaced
135by the opening of the velum, i.e. by a nasal (or velar,
faucal) explosion (compare § 119).
When there is nasal + related stop and consonant,
the middle sound is often lost, as in pumpte, especially
if the third sound is a related continuant, as
in Lanze [lɑn(t)sə].
German agrees with English in the aspiration of
stops before accented vowels ; compare [tʰ] in Tier
and [t] in Stier.
In a great part of Germany g is a voiced continuant
between vowels, as in lagen, liegen ; in some
parts b also, as in leben.
The assimilation of voiceless sounds to voiced
neighbouring sounds is regarded as incorrect, but
is often found, e.g. ss in Wasser, Strasse is pronounced
[z] ; the opposite is more frequent, e.g. [ps]
for [pz] in Absicht. Liquids before or after voiceless
sounds, as in kalt, klein, lose their ‘voice’ partially
A nasalising — due to a premature lowering of the
velum — is often found in words like Signal, when
gn = [ŋn] instead of [ɡn].
When voiced consonants occur initially, they are
usually voiceless at the very beginning ; thus so
almost = [szoː].
The only voiced consonants which occur finally are
[l, m, n, ŋ, r] ; hence d in Land is [t], though in
Lande it is [d]. The same applies to the consonant
ending a syllable, before a suffix beginning with a
consonant ; therefore b in liebte (and in lieblich also,
though there is some variation here) = [p], while in
136lieble it is [b] because l here belongs to the stem of
the word. In lieb' ich we have [b], for it is not
really final, inasmuch as it is within the breath
group ; but here we may also have [p], if the two
words are spoken slowly and distinctly.137
1 As a matter of fact they have roughly the shape of a three-sided
pyramid, though this is not obvious from the diagrams.
1 The way in which the adjusting cartilages can change
the size of the glottis is shown in the note on p. 8.
1 [ʒ] is the sound written s in leisure ; [ʃ] is the sound
usually written sh.
1 For examples of these sounds see the table on pp. 18, 19.
1 [ü] = [u] advanced towards the middle position.
1 Vocalic or syllabic [l].
1 [ö] = [o] advanced towards the middle position.
1 The final [i] is open ; it may also be written [e˔].
2 [ṇ] indicates vocalic or syllabic [n].
1 “I think the greater variety of sounds is due to the far
greater mobility, adaptability and muscular activity of the front
of the tongue, and also to the fact that in these articulations
a small variety of position produces a more noticeable change of
sound than in the case of the back vowels.” — H. W. Atkinson..
1 So Viëtor ; others covered the artificial palate with paint.
The artificial palate is now made of black vulcanite, and kaolin
1 This term is not strictly correct in the case of all sounds
to which it is applied. English [ɹ], [w] and [j] gradually open
or shut (they are ‘gliding,’ not ‘held’), and are not continuants ;
but they are fricatives.
1 Here [ʃ] arose from [aj].
2 (Pre)dorsal = dorsal or predorsaL
1 In provincial French [lj] often represents the ‘l
1 Continuants are sometimes called ‘spirants,' and this term
has been confused with ‘aspirates’ ; perhaps mainly because
in Greek φ, θ, χwere at first aspirates, and then passed into
continuants. The Latin transliteration ph, th, ch suggests
the earlier stage ; the English pronunciation of φ and θ as [f]
and [ð, θ] represents the later.
1 See however § 86 for the pronunciation of g as a continuant.
1 Dorsal, according to Sweet ; in which case we have a
stoppage of a point-blade [s].
1 For tense and lax as applied to vowels, see § 40 ; as applied
to consonants, see § 78.
1 The mark over [ĭ] and [ŭ] must not be taken to indicate
that the vowel is short ; it implies that the vowel is unable to
bear a syllable, that it if a voyelle consonnante, not a voyelle
1 ‘Long’ in German means a little more than ‘long' in English,
1 Here [t(ː)] and [dː] are both frequently heard.