Voicing and Devoicing Assimilation of French /s/ and /z/

The present acoustic-phonetic study explores whether voicing and devoicing assimilations of French fricatives are equivalent in magnitude and whether they operate similarly (i.e., complete vs. gradient, obligatory vs. optional, regressive vs. progressive). It concurrently assesses the contribution of speakers’ articulation rate to the proportion of voicing (i.e., voicing ratios) in /s/ and /z/ embedded in fricative#stop sequences. Data analyses show that voicing and devoicing assimilation are similar in many regards: the absolute amounts of voicing change are equivalent in magnitude (0.77, 0.78) for the two processes: changes in voicing ratios are accompanied by changes in fricative and preceding vowel durations. These concomitant alterations result in the increased acoustic-phonetic similarity between the assimilated and the non-assimilated forms, suggesting that the two processes might be complete. In addition, the two processes operate regressively and across word-boundary. However, data show that the voicing assimilation of /s/ is not rate dependent, which suggest that it might be obligatory, while the devoicing assimilation of /z/ is rate dependent, which suggest that it might be optional.