Abstract
We investigated whether the mean duration of syllables in utterances of different length in Brazilian Portuguese conform to Menzerath–Altmann’s law. Furthermore, by analysing the same spoken material of twenty recorded speakers, we tested for individual differences in the parameters used to formalise the law. The log-transform of the syllable duration was the dependent variable in a linear mixed model with the log-transform of the number of syllables in the utterance as a predictor. Adding a random intercept accounted for the variability across the speakers, but a random slope by speaker did not improve the model. We conclude that Menzerath–Altmann’s law fits our data well, but the decreasing tendency is not different from speaker to speaker.
Notes
1. ‘As an important law of quantity, it may be possible that the speaker can speed up the pace if he is aware that he is to speak a long line of sounds (which is to be spoken preferably in a single train)’.
2. Another piece of work that is relevant in this context is Best (Citation2008). Unfortunately we did not have access to this work.
3. These should be taken here for illustrative purposes only. We analysed the mean syllable duration by the number of syllables per utterance obtained for each single speaker.