Consonant epenthesis and markedness

In recent years, the role of markedness in shaping phonological patterns has been at the forefront of debate: while in Optimality Theory (OT; Prince and Smolensky 1993 [2004]) and related frameworks such as Harmonic Grammar (Pater 2009), markedness constraints play a major role in circumscribing the space of possible phonologies, a number of recent works in the tradition of Ohala (1981), including several chapters in this volume, argue that markedness is an epiphenomenon of extra-phonological factors. The present work takes the phenomenon of consonant epenthesis as a case study for comparing these two perspectives, showing that the range of consonants which are chosen for insertion cross-linguistically cannot be accounted for in terms of markedness. Specifically, here we focus on the empirical and theory-internal problems which arise for markedness-based accounts of consonant epenthesis. We provide an extensive empirical overview with a focus on English r-insertion and critique several markedness-based approaches to consonant epenthesis, including Lombardi (2002), de Lacy (2006), and Steriade (2009), concluding that a viable theory of synchronic phonology must allow for epenthesis of any segment, no matter how marked. This is consistent with the larger point that “unnatural” phonological patterns are the product of historical processes which may obscure phonetic motivation; specifically in this case, unusual epenthetic consonants may result from hypercorrection and reanalysis of deletion patterns (Blevins 2008).In the OT approaches to consonant epenthesis of which we are aware, epenthetic consonants must be …