|Abstract:|| Sonority-Based Stress
It is well known that the sonority hierarchy plays a major role in
determining the peaks and margins of the syllable. In this paper we
look at several languages in which the relative sonority of syllabic
nuclei determines the optimal stress-bearing units. Our analysis is
couched in the constraints-based Optimality Theory (Prince & Smolensky
1993) as this framework provides a way to formally express the
attraction of stress to the most sonorous vowel in a straightforward
and natural way. In order to extend the OT model to the systems we
consider here, several proposals are made. First, the Peak Prominence
constraint Prince & Smolensky (1993) develop for quantitative
distinctions in Hindi stress is extended to sonority distinctions.
Second, comparable to the Prince & Smolensky (1993) analysis of Berber
syllabification, the Peak Prominence constraint is broken down into a
set of micro constraints for each level of the sonority hierarchy.
It is demonstrated how these constraints can be interleaved with other
constraints that orient prominence with respect to the edges of the
word either directly or indirectly through controlling the size of
the metrical constituent. Finally, in order to express the two
opposing edge orientations in languages such as Mari (Cheremis), it is
suggested that sonority distinctions also optimize the trough
("nonpeak") portions of metrical constituents. This application of
sonority parallels the margin constraints in the Prince & Smolensky
(1993) analysis of Berber syllabification.