Non-transparent Constraint Effects in Gere: From Cycles to Derivations

Abstract

Abstract This chapter addresses the case of a constraint in Gere which holds across the board—that is, an undominated (unviolated) constraint in Optimality Theory‘s (OT) terms—but whose effects are none the less non-transparent on the surface in some cases, although easily recoverable in a derivational framework. I will show that the often non-transparent effects of this constraint evidence the need for a cyclic application of constraints, and support more generally a derivational approach as advocated in Paradis (1995, 1996). More precisely, I aim at showing that phonological rules / constraints may refer to phonological properties that never come to the surface, which is in essence one of the claims of the cyclicity hypothesis in Lexical Phonology (see e.g. Booij, Chapter 8 above). We will see that evidence in favor of such a hypothesis questions the adequacy of non-derivational constraint-based theories. The argumentation will be partially based on a comparison between the Theory of Constraints and Repair Strategies (TCRS; see e.g. Paradis, 1988a; 1988b; 1993), a derivational framework, and OT (Prince and Smolensky, 1993; McCarthy and Prince, 1993), a non-derivational one.