Freedom ; anarchy

This paper begins with a review and discussion of Gen and freedom of analysis as presented in the OT foundational works (Prince and Smolensky 1993; McCarthy and Prince 1993). The connection between freedom of analysis and our conception of the richness of the base is also considered. The core of the paper is a discussion of metrical phonology, exploring some specifics of the foot inventory which is available for manipulation by Gen. I begin with a discussion of ternary stress patterns and use this as an opportunity to illustrate the trade-off between restricting the selection of a structure in Gen and restricting it in Con. I’ll argue that (flat) ternary feet will be optimal under certain circumstances, given established constraints in Con, and that their elimination as possible surface structures can therefore be achieved only through limiting the metrical structure alphabet of Gen. My approach stands in contrast to those seen in Elenbaas and Kager (1999) and Green and Kenstowicz (1995). Both of those works suggest restricting ternarity in Eval; Elenbaas and Kager (1999) eliminate all ternary structures with particular constraints and their rankings, while Green and Kenstowicz (1995) are forced to allow mixed binary and ternary feet in the same system. ∗Center for Advanced Study in Theoretical Linguistics (CASTL), University of Tromsø, NO-9037 Tromsø, Norway; [email protected] http://castl.uit.no †This is a handout representing my presentation at the Freedom of Analysis workshop, Tromsø, September 1–2, 2005. Because this is work in progress, the eager critic is kindly requested to refer to the (forthcoming) revised version reflecting, not least of all, the discussion at the workshop; said version will be made available at http://www.hum.uit.no/a/rice .