Infixes really are (underlyingly) prefixes/suffixes: Evidence from allomorphy on the fine timing of infixation

While infixation and allomorphy have long been investigated as independent phenomena (see, e.g., Ultan 1975, Moravcsik 1977, Yu 2007 on infixation, and Carstairs 1987, Paster 2006, Veselinova 2006 on allomorphy), relatively little is known about what happens when infixation and allomorphy coincide. This paper presents the results of the first cross-linguistic study of allomorphy involving infixation, considering 49 case studies from 40 languages (13 language families). Allomorphy and infixation interact in consistent, systematic ways, suggestive of a universal architecture of the morphosyntax-phonology interface. More specifically, the findings support the type of serial architecture proposed by Distributed Morphology and related approaches (Halle and Marantz 1993, 1994, Embick 2010, Bye and Svenonius 2012), and run counter to fully parallel models (e.g., McCarthy and Prince 1993a,b, Prince and Smolensky 1993) and those that take infixation to be “direct” (e.g., Inkelas 1990, Yu 2007, Wolf 2008).