Phonological Opacity and the General Schema FAITH-to-INPUT *

Abstract

A species of phonological opacity (Kiparsky 1971, 1973a) is taken care of by the level condition INPUT that destines markedness constraints as a whole to be satisfied at input level(Lee 1996, In prep.). Besides, there are cases in which at first glance it would seem to be plausible to rely on this level condition to tackle the problem of phonological opacity. Naturally, they are considered to give rise to complex and intricate problems not susceptible to a satisfactory solution in Optimality Theory (OT; Prince & Smolensky 1993). In this paper, I will investigate these cases cited from various languages within the framework of OT, specifically, the Correspondence Theory (McCarthy & Prince 1995). Eventually, this study will prove that the instantiations of the following schema play a crucial role in accounting for the complex and intricate data, which constitute another species of phonological opacity: (1) General Schema FAITH-to-INPUT FAITH-to-INPUT (C₁) Element E in the satisfied domain of constraint C₁ has a correspondent K in input.