Perception, Production and Markedness in Sound Change: French Velar Palatalization

Abstract

Whereas the palatalization of velar consonants before front high and mid vowels is a common property of the Romance languages (Latin [k]ivitatem ‘city’ resulting in Frenchcite or Italiancitta), French is unique in displaying velar palatalization before the low back vowela (Latin [k]antare‘to sing’ giving Frenchchanter). This change has always remained puzzling, as there is no clear phonetic motivation for palatalization. Whereas perception studies have shown that [ki] is likely to be perceived as [tsi], no perceptual confusion is reported between [ka] and [tsa]. Also, in production, the release of a velar or coronal plosive before a high vowel leads to high turbulency, which is not the case when a plosive is released in a low vowel. In this chapter, we review traditional and more recent accounts of the second French velar palatalization and propose a phonological account framed in a constraint-based OT perspective. Furthermore, we explore to what extent OT with Candidate Chains offers a more restricted way of modeling sound change than classical OT.