The Strict Locality of Phonological Processes

This paper addresses the question of ‘what is a possible phonological process’ from a computational perspective. Many previous studies have offered explanations for why certain processes are attested and/or common while others are unattested or rare  (see Hume & Johnson 2001, Hayes et al. 2004, Blevins 2004, among others). Following work on phonotactics by Heinz (2007, 2009, 2010), the goal of the present study is to demonstrate the extent to which computational properties can distinguish the subset of what is phonologically possible from the larger set of logically possible processes.  Specifically, I identify a strong computational property of the mapping from underlying representation (UR) to surface representation (SR) in local phonological processes. This property is called Input Strict Locality (ISL) after the well-studied Strictly Local formal languages (McNaughton & Papert 1971, Rogers & Pullum 2011, Rogers et al. 2013). I demonstrate  that this property has broad empirical coverage and describe its  utility in phonological learning.