Prosodic Prominence in Karuk
- Author(s): Sandy, Clare Scoville
- Advisor(s): Garrett, Andrew
- Inkelas, Sharon
- et al.
This study focuses on word-level prosodic prominence in Karuk (kyh), a Hokan isolate of Northern California. Prosodic prominence in Karuk is made up of sparse tone and stress, and there are two main influences on its placement: the alignment of high tone and certain syllable structures, and the use of prosodic prominence to mark stem edges. These influences are at times in conflict, with the resolution depending on criteria specific to particular sets of morphology. The study is based on analysis of a corpus combining recent fieldwork and historical data. Specific findings include: 1) the placement of prominence in a Karuk word is largely dependent on CV-skeleton syllable structure and far more predictable than previ- ously thought; 2) while one tone-syllable alignment is the unmarked output of constraints, a different tone-syllable alignment on the input blocks its surfacing; 3) various sets of morphol- ogy interfere with the basic placement of prominence by triggering stem-final prominence; and 4) the predictable placement of basic prominence only applies within the prosodic stem, from which certain morphemes are excluded.
The prosodic system of Karuk bears some striking resemblances to those of other lan- guages known as accent languages, but has important differences and added complexities. An unmarked high-tone-before-long-vowel alignment in Karuk is typologically unusual, as is a dispreferred high-tone-on-CVC alignment. Although tone and stress generally coincide, reference to both tone and metrical structure is required in defining preferred and avoided structures.
The complexity of Karuk word prosody poses a challenge for approaches which seek to limit the role of morphology to either the phonology or the syntax, as well as for those which seek to implement a wholly parallel or a wholly derivational model of the phonology- morphology interface. The Karuk case shows that phonology and morphology must be interleaved, and I find that a cophonology approach best accounts for the complex patterns seen. Other challenges for an Optimality Theoretic approach raised by the Karuk data are opacity in a non-surface-apparent constraint driving placement of prosody, and the need for special faithfulness to account for the protected status of a particular tone-syllable alignment.