Resumen de Sibilant voicing in ecuadorian Spanish
Highland Ecuadorian Spanish has a unique process of /s/ voicing that differs from other dialects of Spanish in that word-final /s/ is realized as [z] intervocalically (Lipski 1989, 1994, Robinson 1979). Ecuadorian /s/ voicing is problematic for serial models of phonology as well as for some output-to-output analyses within Optimality Theory. It is argued that the data can be accounted for by an optimalitytheoretic analysis that incorporates phonetic underspecification as a strategy to satisfy coda-licensing restrictions. Sibilants without a voicing target remain unspecified until the phonetic component, where they adopt the laryngeal configuration of neighboring sounds and can thus be expected to exhibit gradient and variable voicing. Contrary to existing non-derivational proposals (Bradley 2007, Colina 2006), it is shown that it is not necessary to resort to two levels of representation (lexical and postlexical) to explain the behavior of Ecuadorian Spanish. The analysis proposed also accounts for cross-dialectal variation affecting word-final prevocalic /s/.
