Committee: Adamantios Gafos (thesis advisor), Arto Anttila, John Singler, Ray Dougherty, Greg Guy, Elan Dresher
This dissertation presents the application of a constraint-based approach to phonology to historical sound change in English. I employ Optimality Theory (Prince and Smolensky 1993) (henceforth 'OT') in the analysis of diphthong-related sound changes that affected the vowel system in Old and Middle English. I pay special attention to Old English Breaking, a variety of Early Middle English developments and Middle English Breaking.
Past studies do not answer why these diphthong-related sound changes tend to recur before /x, r, l/ throughout the history of English. In addition, such studies analyze each change individually but do not provide a unified account that captures structural coherence among these changes. A general claim of this dissertation is that the combination of, on the one hand, an enhanced theory of representations reflecting temporal relations of the segments involved and, on the other hand, a constraint-based theory of grammar (OT), makes possible a unified and natural account of these sound changes.
With an enhanced theory of representations, I argue that the ambiguity of phonetic substance in the sequence 'V + /x, r, l/' is responsible for the oscillation between diphthongization and monophthongization throughout the history of English. Assuming that OE Breaking and related processes represent different strategies to avoid the Obligatory Contour Principle violations between two vowel-place nodes in 'V + /x, r, l/' sequence, I present a unified account that characterizes each process via re-ranking of the one set of OT constraints.
Through detailed OT analysis of OE Breaking, I show next that the emerging partial grammar not only characterizes the breaking process but derives the asymmetric patterns among breaking contexts for OE Breaking in a quantitative fashion as well. I then construct a partial grammar of EME developments, with invariant rankings that derive dialectal variation as well as circumscribe the space of possible changes in a specific historical period of English. Through an analysis of ME Breaking, specifically as it compares with OE Breaking, I show that the difference between these processes comes down to a simple permutation of two faithfulness constraints.