|Title:||Compensatory lengthening in phonological representations: nature, constraints and typology|
|Comment:||PhD dissertation at UniversitÃ© Paris 3 (Sorbonne-Nouvelle) [in French]|
|Abstract:||This PhD dissertation deals with Compensatory Lengthening (henceforth CL) and its formal expression in phonological representations.
Compensatory lengthening is a fairly widespread process across the world's languages in which a segment lengthens to compensate for the deletion or migration of an adjacent segment.
The most represented case, that is the 'classical' CL case, is triggered by the loss of a consonant in the syllable coda position and followed by the subsequent lengthening of the preceding vowel (cf. chapter 1).
While classical CL is relatively frequent, CL triggered by the loss of a consonant in the onset position is claimed to be inexistent (cf. Hayes 1989 among other). Nevertheless, chapter 2 provides some examples of various languages undergoing CL after an onset consonant deletion, what I call 'exotic' CL cases.
Chapter 3 states that moraic phonology (McCarthy & Prince 1986, Hayes 1989) fails to account for this kind of process while, obviously, the segmental theories of the prosodic tier (CV theory, X-slot) make correct predictions. In this chapter, the CVCV theory (Lowenstamm 1996, Scheer 2004), a lateral approach of syllabic constituents that comes from Government Phonology, is also discussed.
Chapter 4 is intended to give a short introduction/basis to Optimality Theory (Prince & Smolensky 1993, McCarthy & Prince 1995).
The chapter 5 examines several representations of compensatory lengthening in Optimality Theory framework. There, I show that an implementation of the moraic theory in a constraint-based framework such as OT faces several problems (opacity is one of them).
Finally, chapter 6 tries to solve these problems and introduces three Optimality Theory based accounts of compensatory lengthening. Among them, one deserves particular attention since it assumes input moraicity revival (cf. Hyman 1985).