ABSTRACT
This chapter explores phonological and phonetic rules and presents a variety of phonological rule ordering hypotheses from the perspective of finite state control. It discusses the major phonological rule types and exemplifies the method of encoding these in a linear fashion. The chapter introduces a class of autosegmental automata that operate directly on autosegmental representations, and exemplifies the method of replacing phonological constraints and rules by such automata. It also discusses generalization of the results to multi-tiered representations. The rules relating underlying representations to surface representations are traditionally divided into three broad classes of morphological, phonological and phonetic rules. Perhaps the simplest possible operations on autosegmental representations are association and delinking, formally defined as the addition of an association line to an association relation. In order to define devices manipulating autosegmental representations, it is instructive to consider devices manipulating simpler data structures such as strings.
