An Optimality-Theoretic Split-Margin Analysis of the Evolution of Consonant Clusters in Historical French Phonology

Description

Despite the highly marked consonant clusters attested in present-day French (Dell 1995), French's early history exhibited an evolutionary trend towards syllable-geometric simplification as regards the sonority contour of licensed syllable- and word-internal clusters. In this diachronic study, I analyze the evolution of the Latin phonological system, exhibiting relatively marked clusters, into the more prohibitive phonology of later Old French, which essentially banned all codas (Gess 1999) and allowed only the least marked onset clusters. This diachronic progression towards unmarked syllable structure retracts during Middle French, when previously banned clusters begin to resurface. In this dissertation, I present a unified optimality-theoretic analysis of a diverse range of changes within onset and syllable-contact clusters from early French diachrony, drawing on implicational structural relationships between margin segments in the syllable and phonological word as formalized in the Split Margin Approach to the Syllable (Baertsch 2002; Baertsch & Davis 2003 and later work). This dissertation formalizes systemic connections among numerous diachronic developments affecting consonant clusters, traditionally viewed as independent in French diachrony. In Late Latin and Gallo-Romance (1st-10th century), I examine /s/-obstruent prosthesis, coda obstruent deletion in syllable-contact clusters, and repairs on clusters resulting from Late Latin syncope (deletion, rhotacization, epenthesis). In Old French (11th-13th centuries), I extend my analyses to the incremental sonority-based loss of codas (Gess 1999): /s/ deletion, vowel nasalization and /N/ deletion, and /l/ vocalization. The result is a comprehensive representation of several stages of French with respect to cluster licensing, providing strong evidence for viewing these developments as interrelated, interdependent manifestations of a broader diachronic trend. Finally, I present preliminary analyses for the resurfacing of previously banned clusters during Middle/Renaissance French (14th-16th centuries), demonstrating how loanword integration likely triggered the reverse trend towards syllable-geometric complexification, setting the stage for the highly marked clusters of contemporary French.