Extending Finite-State Models of Optimality Theory for Non-Finite State Phenomena

Title: Extending Finite-State Models of Optimality Theory for Non-Finite State Phenomena

Research Question: How can we extend finite-state models of Optimality Theory, like OTP, to handle non-finite state phenomena like reduplication and phrasal stress, while maintaining their computational efficiency?

Methodology: The authors propose a mechanism to extend OTP-like Optimality Theory models to non-finite state phenomena. They use multiple context-free grammars (MCFGs) to model reduplication, following the standard Correspondence Theory account. They also develop an extended version of the Earley chart parsing algorithm to apply constraints to reduplicating candidate sets.

Results: The authors demonstrate how to represent and process higher-level grammar representations (like MCFGs) within the finite-state framework of OTP. They show how this approach can be used to model reduplication and phrasal stress, which are typically non-finite state phenomena.

Implications: This research has several implications. First, it provides a way to extend finite-state models of Optimality Theory to handle non-finite state phenomena, allowing for a more comprehensive model of phonological processes. Second, it offers a method to apply constraints to candidate sets that are generated by higher-level grammar representations, broadening the scope of OTP. Lastly, the extended Earley chart parsing algorithm developed in this research can be applied to other areas of phonology where non-finite state phenomena occur, such as phrasal stress patterns.

Link to Article: https://arxiv.org/abs/0102021v1 Authors: arXiv ID: 0102021v1