Proto-Romance Stress Shift Revisited

This paper argues that the notoriously problematic Late Latin stress shift in words with a light penultimate syllable before consonant-liquid clusters cannot be adequately explained by double prosody (an additional mora, projected to the syllable, but not realized by the vowel) which allows a light syllable to count as heavy before consonant-liquid clusters (Bullock 2001). Double prosody leads to questioning the very nature of syncope, is not obvious in the cases of desyllabification where it leads to a far more elaborate account, is not helpful in dealing with the stress shift from prefix to stem, and, does not prevent syncope from rendering the stress system opaque. Sharing Bullocks (2001) intuition of a relation between syncope and stress shift, an OT account is proposed which expresses that relation in a more direct and less abstract way and obviates the need for double prosody in accounting for the stress shifts.