ABSTRACT
This chapter introduces the role of statistics in phonology. It examines those domains where quantitative effects are modelled in phonological theory and where specific statistical phonological machinery is proposed. The chapter then examines statistical approaches in constraint-based theories of phonology such as Optimality Theory (OT). It also explores various proposals for treating statistical phonotactics and lexical frequency. The chapter focuses on classic variable rule literature in rule-based phonology. Guy combines variable rules with Lexical Phonology to capture the distribution of t/d-deletion in different morphological categories. A domain in which statistical models have made an appearance in phonological theory is the treatment of corpus facts. An extremely influential early paper in this domain is Davis. Coetzee & Kawahara develop a different statistical approach to the role of lexical frequency in variationist data that they test against coronal deletion in English and geminate devoicing in English borrowings in Japanese.
