Institut numerique

1.2.3. French vs. English: syllable- and stress-timing theory

The terms “syllable-timing” and “stress-timing” are used in the theory of isochrony,
according to which some languages have isochronous syllable-durations, and others have
isochronous inter-stress intervals (Pike, 1945; Abercrombie, 1967, among others). The aim
here is not to debate over the existence or not of isochrony, but rather to enhance the
difference between the prosodies of French and English, which are often regarded as the
prototypical examples of, respectively, syllable-timed languages and stress-timed languages
(Bertrán, 1999)(7). An account of this theory will help us understand why French EFL learners
have difficulty in acquiring the English rhythm.

The notion of isochrony has often been used as an attempt to characterize and classify
languages according to how their rhythm comes into being, i.e. whether rhythmic stresses
occur at relatively regular intervals (stress-timing), or stressed and unstressed syllables are
treated similarly (syllable-timing). However, no experiment has actually proved the existence
of strict isochrony in any language so far. Roach (1982) conducted an experiment with six
speakers of six different languages, among which three were supposed to be syllable-timed
like French, and three stress-timed like English. The first claim of the author was that syllable
length is more variable in stress-timed languages. The second claim concerned the presence
of regular stress beats in stress-timed languages, and their absence in syllable-timed
languages. Roach measured the duration of tone groups without preheads and tails; the
duration was divided by the number of feet in order to reach the ideal isochronous interval,
and it was then compared with the actual durations of feet. The results of the experiment
gave no support to the classification of these languages as stress-timed or syllable-timed. On
the basis of the measurement of time intervals in speech, the main conclusion was that the
distinction between stress-timing and syllable-timing is auditory and subjective – a language
is classified as syllable-timed if it sounds syllable-timed. Isochrony tends to be apparent
rather than real.

Another interesting experiment has been carried out by Bertrán (1999) on speakers of
seven languages, including French and English. The stimuli used were utterances with the
same kinds of stressed vowels and consonants, but the distance between the stressed
syllables varied. The author then measured the absolute duration of the feet, concluding:

Languages considered stress-timed, and others considered syllable-timed give a
rather similar response to the tests, with results that openly contradict the
typological models they are supposed to represent. There is no compensation at
all to balance the duration of the units composed of different number of
elements. On the contrary, the rhythmic units not only demonstrate a strong
temporal inequality, but even certain parallels with their morphological
inequality, a phenomenon which is the antithesis of both rhythmic types. (125)

In the same way as Roach (1982), a lack of accentual or syllabic isochrony was detected in all
seven languages. There seems to be no phenomena of compensation or compression in the
feet, nor in the syllable. More particularly, the measurements of French revealed that syllabic
duration was not uniform, while English did not fit either of the rhythmic schemas.

Since no clear evidence of strict, “strong” isochrony has been provided, the term “weak
isochrony” has emerged to refer to the relative, seemingly equal amount of time between
stresses or syllable durations. This term enables one to get an insight into the rhythmic
structures of languages, without going as far as advocate a perfect equality between inter-
stress intervals or syllable lengths. The analysis given by Dauer (1983) is now widely
accepted (Nava & Zubizarreta, 2009), because it offers a solution to the contradiction between
the perceived isochrony and the measured lack of isochrony. She established a timing
continuum thanks to comparisons of data from continuous texts in English, Thai, Spanish,
Italian, and Greek. These showed that inter-stress intervals in English are no more
isochronous than inter-stress intervals in, e.g., Italian, which is supposed to be syllabletimed.
Instead, Dauer claims that the tendency for stresses to occur regularly is more a
language-universal property, no matter if the language is traditionally considered as stresstimed
or syllable-timed. The difference between stress-timed and syllable-timed languages
has to do with such things as differences in syllable structure complexity, vowel reduction,
and the phonetic realization of stress and its influence on the linguistic system. That is why
languages should be treated as more or less stress-based, depending on their characteristics.
Therefore, the timing continuum is a scale that goes from maximally stress-timed to
maximally syllable-timed, and each language has its place on it.

As Ploquin (2009) rightly points out:

After all, we don’t expect to find categories of languages according to their
segmental inventories. Trying to find rhythmic categories might [sic.] the same as
calling a language ‘nasal’ because it includes nasal vowels or ‘fricative’ because it
makes use of more fricatives than any other type of consonants. (49)

Even if there is no such thing as strong isochrony, the theory gives an interesting insight into
how English rhythm is structured, and it highlights the difference with French and the
difficulties that French EFL learners have to face. In fact, the stress-timing/syllable-timing
theory seems to account for many suprasegmental difficulties for L2 speakers. Auer (1993)
explains that one of the main consequences of stress-timing is vocalic reduction, while in
syllable-timed languages, there is no phonemic reduction and very little phonetic reduction.
That is why in a syllable-timed language like French, assimilation – of place of articulation,
especially – is rarer than in such a stress-timed language as English. According to this
reasoning, some segmental problems are due to the fact that French learners do not realize
the English stress-timed rhythm properly. As for the problem of nucleus misplacement, Nava
and Zubizarreta (2009: 175) specify that “in order to acquire the Germanic NS [nuclear stress]
algorithm, the L2 learner must have moved from a syllable-timed to a stress-timed rhythm”.
Once again, one of the major difficulties for French learners, i.e. English tonicity, may find its
origin in the rhythmic difference between the two languages, and it may be good that
English teachers introduce this notion of weak isochrony and rhythmic typology to learners.

7 As the objective here is to compare French and English and highlight their specific differences, the
third type of language, i.e. mora-timed languages, is not discussed.

Page suivante : 1.3. Conclusion

Retour au menu : EXPERIMENTAL RESEARCH INTO THE ACQUISITION OF ENGLISH RHYTHM AND PROSODY BY FRENCH LEARNERS