Anticipation in the Sequence:
Vowel and Consonant-Group *
Having shown, in the course of my investigation of vowel duration in French that the
length of the vowel [ɛ] before consonants decreases as the force of articulation of the
following consonants increases, I published an article 1 in which I tried to make a
pedagogical application of this principle. I advised French teachers not to use any
length marks to differentiate the vowels of words like serre and sec, the durations of
which I considered the result of equal intentions differing however because of the
shortening or lengthening influence of the subsequent consonant, a mechanical and
unconscious influence. Professor Daniel Jones objected to my theory. Moreover, his
remarks, which followed my articles, actually denied the influence of the nature of a
consonant on the duration of the preceding vowel. Having in mind the duration of
[ɛ] before [r] in words like serre, Daniel Jones said: “The fact that the lengthening is
not a consequence of the nature of [r] itself is shown by the existence of words such
as orge, herbe, corse, in which the vowel is not lengthened.”
This being entirely in contradiction to the principle mentioned before of the inverse
variation of vowel length and of the articulatory force of a following consonant, I had
to answer Daniel Jones and try to demonstrate that the lengthening or shortening of
a vowel can definitely be the consequence of the nature of the following consonant.
Professor Jones finally seemed to accept my views and printed my answer to his
The influence of a consonant on the preceding vowel can best be explained by the
action of anticipation.
The importance of anticipation in historical phonetics is well known. One perfect
example of it is metathesis, which is generally the anticipation of a combined speech
sound the articulation of which presents some difficulty in its original place. Epenthesis
also is often caused by anticipation. The same can be said of vowel nasalization
before nasal consonants; of palatalization of velars before front vowels; of umlaut;
and in general of numerous cases of assimilation.
In the same way, in modem French, when a consonant following a vowel requires
a certain effort of articulation, it is the anticipation of that effort while the vowel is
122being pronounced which brings about too soon the articulation of the consonant,
not allowing the vowel its average duration. But when the consonant is weak, the
anticipated effort is smaller, the consonant shows no hurry to be heard, and the vowel
can prolong itself to a greater extent. For instance, if the [ɛ] of sec is much shorter
than the [ɛ] of serre, it is because the anticipated effort of articulation of a [k], a velar
stop, is much greater than that of an [r]. It appears to be a compensatory process:
a short vowel precedes a strong consonant, and a long vowel precedes a weak consonant.
Its cause is the course of action of anticipation.
In an effort to make this quite evident, I propose to study here some aspects of
anticipation by comparing duration of vowels before groups of consonants with
duration of vowels before the single consonants that enter into the groups. The
following tables present a series of such comparisons. Each one of the figures gives
in hundredths of a second the approximate average duration of five kymograph
recordings made under equal conditions by the same person. The [ɛ] sounds are all
in stressed position at the end of a sentence or sense-group of 5 to 7 syllables.
Groups: r + Consonant
r | rʒ | rv | ʒ | v | rd | rb | rg | d | b | g | rɲ | rn | rm | ɲ | n | m | rf | rs | rʃ | f | s | ʃ | rk | rt | rp | k | t | p | rl | l
In the middle column, this table shows the durations of [ɛ]'s before groups comprising
an [r] and another consonant. On the left are the durations of [ɛ]'s before [r]; and
on the right the durations of [ɛ]'s before the second consonants of the groups.
Let us make comparisons.123
All [ɛ]'s have less duration before the groups than before [r] alone. Therefore,
the anticipation of the group does not correspond to that of the groups' first
In comparing from top to bottom the durations before the groups with the durations
before the second consonants of the groups, there appears between these two
columns a striking correspondence: the shortening effect of the groups increases
parallelly with that of the single consonants. Therefore, it is certain that anticipation
reaches the second consonant of the groups.
The duration of all [ɛ]'s is smaller before the groups than before the second consonant
of the groups. Therefore, the anticipation of the group does not correspond
to the anticipation of the second consonant of the group.
Since both consonants are involved in the process of anticipation, and it corresponds
to neither one of them, we must understand that there is anticipation, not of one or
the other of the two consonants, but of their combination.
We have just seen that the shortening effect of groups beginning with a liquid is
stronger than that of the second consonant of these groups. According to the theory
of vocalic duration mentioned before, it means that one anticipates a greater effort
in the pronunciation of serve, certes, than in that of sève, sept, and that the pronunciation
of [rv], [rt], in the same syllable, is comparatively difficult. It is interesting to
notice a complete agreement between this fact and what happens in phonetic syllabication:
in words like cerveau, certain, where the two consonants do not have to
belong to the same syllable, they choose to be divided, observing the principle of
economy of effort, the [r] belonging to one syllable and the [v] or [t] to the following
one. (Without entering details of the problems of syllabication, let us recall that the
main reason for the separation of these two consonants is to be found in their difference
of aperture, the first having the greater.)
The group [rl], having a liquid for second consonant as well as for first, must be
Vowel duration before the group is here greater than before the second consonant
of the group. The main reason lies in the importance taken here by the factor of
“combination of the two consonants” which is generally secondary, although never
absent. The two liquids blend so easily that the anticipated effort is smaller for their
combined pronunciation than for the pronunciation of the second one alone.
We shall examine next the groups ending with an [r]. It is now the second consonant
of the group which has the greater aperture and the smaller force of articulation.
The figures again show that there is anticipation neither of the first consonant alone
nor of the second alone but of the combination of the two consonants. This fact is
much more marked here than in groups beginning with a liquid, for it is the first
articulation which dominates. The two consonants appear almost as a single articulation.
We notice that when the first consonant has a definitely stronger articulation, it takes124
Groups: Consonant + r
b | d | g | br | dr | gr | r | p | t | k | pr | tr | kr | v | vr
a dominant place in the anticipation of the group. A comparison of the three columns
indicates that there is a close relationship between duration before groups and duration
before single consonants corresponding to the first one of the groups, except
when that first consonant is a fricative.
We also notice that when the first consonant of a group has a definitely stronger
articulation, the anticipated effort is smaller for the group than for its first consonant
alone. In more concrete terms we can say that the groups stop + r are easier to
articulate than the stop alone.
Groups Without Liquid
s | st | sk | t | k | p | ps | ks | g | gm | kt | m | pt
Finally, let us examine the groups without liquids.
In these groups, the difference of aperture is small (in first four) or negligible (in
Excepting the case of [pt], which will be taken up later, the anticipated effort of
articulation is greater for the groups than for their first consonant alone, or at least
125equal to it. In other words, a group of consonants without a liquid is generally more
difficult to pronounce than the first consonant of the group alone.
In the first two groups, [st], [sk], the dominant articulation is in second place.
However, anticipation does not operate in this case as for the groups r + consonant;
in fact it is the opposite: the anticipated effort is smaller for the group than for the
second consonant alone (the dominant one). This is because the difference of aperture,
and consequently the tendency for the two consonants to separate, is less marked:
it is not impossible to say [st], [sk], in the same syllable (pesté, presqu'île) but it is
almost impossible to do it for [rt], [rk], (parti, marqué).
The two next groups, [ps], [ks], have their dominant articulation in first place. It
is rather remarkable that the anticipated effort of the group remains equal to that of
the first consonant (the dominant one). One should not have expected it to be smaller,
as for groups ending with a liquid, for the second consonant, [s], has not a sufficient
aperture and is not weak enough to allow the first consonant to dominate very
The two next groups, [gm], [kt], have no dominant articulation. It is normal,
therefore, to find that the anticipated effort is greater for the groups than for one or
the other of the consonants that compose it.
The particular case of the last group, [pt], in which the anticipated effort is smaller
for the group than for either one of its consonants, can be explained, as I did for
[rl], by the importance of the factor “combination”. The facility with which these
two consonants, [p] and [t], combine is exceptional because their places of articulation
are close, yet not too much, and their articulations are such that they do not
interfere with each other when they follow the order [p-t]. (It would not be the same
in the order [t-p].)
During the articulation of a vowel, anticipation of the following group of consonants
reaches the second consonant of the group as well as the first.
The consonant with the stronger articulation takes a dominant place in the anticipation
of the group articulation only if it is very definitely dominant and at the
beginning of the group.
Anticipation of a group of two consonants concerns neither the first nor the
second separately but the combination of the two.
Groups for which the anticipated effort of articulation is greater than that of the
consonant of the group having the smaller aperture are those beginning with a liquid
(except [rl]), those composed of two stops (except [pt]), and the groups fricative +
liquid. For the others (stop + liquid, fricative + stop, stop + fricative, [rl], [pt]) it
is the opposite: the anticipated effort of articulation of the group is smaller than, or
at the most equal to, that of the consonant having the smaller aperture.126
In a more general way, the anticipated effort of articulation of a group always
depends on the facility for combination shown by the consonants: groups beginning
with a liquid show a minimum of facility; groups ending with a liquid a maximum.
To the latter groups we must add the groups [rl] of two liquids and [pt] of two stops,
the articulations of which fit each other exceptionally well when uttered in this
* Originally published in The French Review, XIII, 4 (February, 1940).
1 Le Maître Phonétique, 3ème série, 64 (Dec, 1938), 66-7.
2 Ibid., 3ème série, 67 (Sept., 1939), 41-4.