| Abstract: | This paper takes Idsardi's (1998) discussion of Tiberian Hebrew (TH) spirantization as a point of departure for an investigation of these phenomena within a stratal approach (Kiparsky 2000,2001) to optimality theory (Prince and Smolensky 1993). The paper begins by discussing the larger problem of phonological opacity (Kiparsky 1973) in optimality theory, since the TH facts have played a prominent role in this discussion. The analysis begins with an examination of the phenomena under study (spirantization, epenthesis, laryngeal deletion, vowel deletion and their interaction), followed by a look at the motivation for level-ordering in TH. Unique to the present analysis, in addition to the stratal OT approach in general, is the positing of a `phrasal stratum' of OT phonology in TH, on the basis of Dresher's (1994) observation that several TH phonological processes take the phonological phrase as their domain of application, spirantization among them. The stratal optimality theoretic analysis, aside from accounting for the phenomena under study, raises several interesting theoretical questions for the stratal OT framework that merit further examination. Among these are: the status of what we dub the Constraint Promotion Hypothesis (an idea due to Kiparsky p.c.), whereby promotion of constraints between strata is promotion to undominated status; the status of vacuous Duke-of-York derivations; and the relative ranking of markedness constraints with respect to faithfulness constraints at different strata in the absence of positive evidence. As a preliminary extension of the analysis of the spirantization facts, the paper also includes discussion of epenthesis, umlaut, vowel lowering interaction, which according to Idsardi (2000) are problematic for a level-ordered approach to OT. |
