The categorical and variable phonology of Russian
This dissertation analyzes a wide range of categorical and variable phonological phenomena in Russian within the non-derivational framework of Optimality Theory (Prince & Smolensky 1993, McCarthy & Prince 1993). Over the course of this study, the major assumptions of Optimality Theory about input/output correspondence relations, the role of surface well-formedness and markedness restrictions, as well as the assumptions of the theories of feature specification, are reevaluated in an examination of several characteristics of Russian phonology: contrasts in the segmental inventory, vowel reduction, CV assimilations and the phonology/morphology interface. I argue that, given the representational assumptions of the Unified Feature model (Selkirk 1991) and the formalisms of the Correspondence model in OT, the outputs of various phonological processes are uniformly determined by the feature markedness asymmetries and the relations of similarity between the interacting segments. The present analysis allows a fundamentally different, more explanatory approach to several well-known issues in Russian phonology, previously analysed in derivational terms. Consonant mutations, for example, are analysed as consonant/vowel coalescence, target-driven by faithful input/output correspondence and syllabic well-formedness. On the other hand, the "derivational residue" in Russian phonology, such as the exceptionality of epenthetic elements, poses significant problems for OT, and requires a rethinking of correspondence relations.^ I also argue that a constraint-based approach allows a better analysis of many generalizations formerly expressed in feature geometry as a set of representational pre-conditions. When these generalizations are understood as substantive constraints on the output, many problems encountered by feature geometric models are avoided, and new insights into the operation of segmental interactions become possible.^ A theory of markedness, central to OT, is of special interest to this dissertation. I argue that the same featural and morphological markedness relations (formalized as implicationally ordered families of constraints) determine the outcome of categorical phenomena, and are responsible for the gradient effects of linguistic environments in the operation of the sound changes, providing an explanation for the role of naturalness in the dynamics of a linguistic system. An examination of variable data, however, provides a unique challenge to OT, and requires specific modifications of the model. ^
Ekaterina L Zubritskaya,
"The categorical and variable phonology of Russian"
(January 1, 1995).
Dissertations available from ProQuest.