Aspects of Tone and Voice in Phuthi

1157 p.Thesis (Ph.D.)--University of Illinois at Urbana-Champaign, 2007.This dissertation seeks to explore the architecture of tonal structures in a formal phonological analysis of the tone patterns found in Phuthi, a Bantu language. The focus of the work is Phuthi high tone (H), and its interaction with tonal 'depression domains', often triggered by breathy voiced segments, which condition low tone within the lexical phonology. Both theoretical and empirical claims are made. Within a general Optimality Theory framework, the Phuthi data requires a distinction between the planning and execution of a H tone with a tonal domain, modeled here with distinct PARSE and EXPRESS constraints in the tone grammar, as there are frequently very significant mismatches between H tone domains---established from lexical tone placement, chiefly in the highly productive verb paradigms (though nouns are explored too)---and domains of tonal depression (consonant-triggered, or morphologically imposed). A strong case is made for the presence of tonal low (L) domains, which can be multiply nested even within a single H domain. In such instances, a single H tone does not fission into two separate tone domains, nor does its interruption constitute a significant violation of locality. Phuthi instantiates a wide range of L and H domain edge conflicts. When these conflicts coincide with a tone domain-head (a notion developed in the work), the language attempts to force tone shift, a phenomenon widely commented on for Nguni languages. Phuthi displays a number of unprecedented variations on tone shift and tone block, arising from a dine of tone-depression interactions not observed in other languages. Phuthi is shown to be a language that prohibits or severely restricts the coincidence of H and L tone domains over the same interval. And yet the grammar conspires with the lexicon to assign both H and L features to syllabic nuclei in head positions of phonological domains. The simultaneous assignment of H and L tones introduces a feature conflict which is resolved in a striking variety of ways across verbal and nominal paradigms. Optimal Domains Theory (ODT) is argued to be a theoretical framework capable of sufficiently expressing this range of voice and tone data, in contrast to any type of optimal constraint model that lacks the augmented domains architecture.U of I OnlyRestricted to the U of I community idenfinitely during batch ingest of legacy ETD