Rhythm Types and the Speech of Working-Class Youth in a Banlieue of Paris 111V (10%), CCV (8%), and CCVC (4%). This suggests that speakers in the two
groups did not differ with respect to the type and complexity of the syllables
that they used. If vowel elision, insertion, or reduction occurred in AF speak-
ers’ speech, it did not lead to substantial inter-group differences. Numerical
differences between the two groups were, in fact, so small that inferential
statistics were not computed.
And yet, near-dichotomous differences exist between the two groups at a
lower level of phonetic contrast. This level of allophonic differences usually
remains unexplored in studies of rhythm type distinctions, because contrast-
ing patterns are often too scarce to be modeled statistically. But, as we shall
see, even a few percent of variability at perceptually salient points in the
acoustic signal can carry important information.
Figure 4.5 shows that four out of seven acoustic variables, although
numerically not substantial to account for statistically signi¿ cant effects,
Figure 4.4 Types of syllables in the readings of French speakers of North African
(AF) (N=713) and European (EF) (N=742) descent.