[Author Login]
Title:Representing OT Grammars
Authors:Alan Prince
Abstract:Analysis proceeds with varying degrees of faithfulness to the theory that it rests on. Within OT, a commonly encountered chokepoint is the representation of grammars, which not infrequently have a form that differs from expectation. A variety of table types and graphical structures are deployed in the literature to handle the eventualities; many are not up to the task and impose their own debilitating artifacts.

This note sets out to determine where the common representational schemes succeed and where they fail in representing the content of OT grammars. Discussion begins by settling on a notion of grammar that accords with the general meaning of the term within Generative Grammar. Definition in hand, we assess the VT (violation tableau), the dashed VT, the Hasse Diagram, the set of Hasse diagrams, the dotted Hassoid, and the CT (comparative tableau), arguing mostly from familiar examples. Concrete conclusions are drawn about best (and non-best) practices.


0 Introductory
1 The VT and its discontents
1.1 Dashed hopes
2 A Gathering of Arcs
3 A Grammar of ERCs
4 Make straight the way
5 Practices make perfect
5.1 The VT
5.2 The CT
5.3 The Hasse Diagram
5.4 Don't Go There
6 Conclusion

Appendix I Grammar and Inversion
Appendix II The legs of (C)V.ins
Appendix III The dotted Hassoid

Type:Paper/tech report
Area/Keywords:formal analysis, analytical technique
Article:Version 4
Version 3
Version 2
Version 1