UCLA

This dissertation develops optimality-theoretic analyses of word-level stress assignment in two languages with lexical accent, Cupeño (Takic, Uto-Aztecan) and Hittite (Anatolian, Indo-European); it also assesses the implications of word stress in Hittite and the other Anatolian languages for the reconstruction of stress assignment in Proto-Indo-European.

I argue that stress assignment in Cupeño is governed by the BASIC ACCENTUATION PRINCIPLE (BAP; Kiparsky and Halle 1977): stress is assigned to the leftmost lexically accented morpheme, else to the word's left edge. This analysis is compared to that of Alderete (2001c), who argues that Cupeño shows accentual root faithfulness — i.e., that the accentual properties of roots are privileged over non-root morphemes. I show that the BAP analysis is both simpler and attains greater empirical coverage than the root faithfulness analysis, which fails to account for certain attested stress patterns that are captured under the BAP analysis. Thus reanalyzed, Cupeño has two important typological implications. First, without support from Cupeño, root faithfulness may be unattested as a feature of lexical accent systems. Second, Cupeño provides a clear typological parallel for the ancient IE languages on the basis of which the BAP was posited — in particular, Vedic Sanskrit — as well as for Hittite, where I argue that it is also operative.

The analysis of Hittite stress advanced in this dissertation is the first systematic attempt at a synchronic generative treatment of its word stress patterns. Having established that that stress assignment in Hittite inflection is governed by the BAP, I also adduce evidence for accentual dominance — i.e., morphemes whose accentual specification "overrides" the BAP. I propose that accentual dominance in Hittite is a consequence of morphological headedness: the lexical accent of the word's head morpheme is privileged in Hittite, just as Revithiadou (1999) has argued for other lexical accent systems.

Finally, this dissertation addresses the reconstruction of the Proto-Indo-European word-prosodic system. Hittite and the other Anatolian are not traditionally viewed as important sources for the reconstruction of this system; however, I contend that the BAP is reconstructible for PIE and that — against this traditional view — that this reconstruction depends crucially on the Anatolian evidence, which converges with Vedic Sanskrit in this respect.