Rethinking the phonological process of /s/ voicing assimilation in Spanish: An acoustic comparison of three regional varieties

Abstract Spanish is described as having an /s/ regressive voicing assimilation process by which the sibilant is voiced when followed by a voiced consonant. However, experimental studies documenting the nature of the process – including variation in its realization across speech varieties – are limited. The current study presents an acoustic analysis of the phonetic nature of the process, including an analysis of the linguistic and social factors which influence voicing of the /s/. Using an identical controlled phrase elicitation task, rates and location of voicing within the /s/ segment were compared across three varieties of Spanish: Mexican Spanish (Mexico City) and two Peninsular varieties (León and Vitoria). Different voicing rates and different linguistic voicing predictors were found across dialects. The data suggest that /s/ voicing before a voiced consonant is far from a categorical process in Spanish, with variable rates of application between 43% and 63%. We propose that, based on the data, /s/ voicing in Spanish is, in fact, better understood as a progressive voicing process (i. e. continuation of voicing from the previous vowel), and that the linguistic factors that condition voicing can be explained in part through articulatory and aerodynamic mechanisms. We also discuss the phonetics and phonology of the process.