Stratal OT: A synopsis and FAQs

By modeling phonology as a system of ranked violable constraints, Optimality Theory (OT) succeeded in bringing substantive universals and typological generalizations to bear on the analysis of individual phonological systems, and uncovered important generalizations that escaped classical generative phonology, such as top-down effects and the emergence of the unmarked, to name just two (Prince and Smolensky 1993, 2004). Another fundamental principle of classic OT, that all constraints are evaluated in parallel on output representations, initially contributed much to the theory’s conceptual appeal, but it soon became clear that the price for maintaining it is prohibitive. In order to account for phonology/phonology and phonology/morphology interactions under parallelism, numerous new computationally and learning-theoretically intractable constraint types had to be devised — Output-Output constraints, Paradigm Uniformity constraints, Base-Reduplication constraints, Sympathy constraints, Precedence constraints, among others. They vastly expanded the factorial typology and, separately or in any combination, failed to do the empirical job they were intended for.