Incomplete Resyllabification and Ambisyllabic Gestural Coupling in Spanish

In the generative literature, the pattern of coronal fricative lenition found in the traditional Chinato Spanish dialect is commonly cited as a phonological argument that the resyllabification of word-final prevocalic consonants is complete, in the sense that onsets derived by resyllabification are structurally identical to canonical (word-level) onsets. However, recent acoustic studies of Northern-Central Peninsular Spanish have problematized the completeness of resyllabification with experimental evidence that /s̺/ is shorter and more voiced as a derived onset than as a canonical onset. Using a split-gesture, competitive, coupled oscillator model of the syllable in Articulatory Phonology, which divides consonants into a separate constriction and release gesture, we propose a novel representation of ambisyllabicity that predicts the phonetic behavior of derived onset /s̺/ in Northern-Central Peninsular Spanish. We then show that ambisyllabic coupling permits a simpler phonological analysis of coronal fricative lenition in Chinato Spanish as compared to alternative accounts. Our analysis makes typological predictions that are confirmed by patterns from other contemporary Spanish varieties. Lastly, we examine the consequences of ambisyllabicity for the analysis of Spanish rhotic consonants, which have also been argued to support complete resyllabification. We offer an analysis of rhotics that is entirely compatible with an ambisyllabic representation of incomplete resyllabification.