Affix copying in Kiranti. An optimality-theoretic approach

In this paper, I argue that affix copying in the verbal paradigms of different Kiranti languages is triggered by defective phonological structure in the representation of morphemes. It is shown that it is impossible to analyse the affix copying as a general repair operation to avoid marked phonotactic structure but it is nevertheless true phonological copying: it avoids elements that are uninterpretable for the phonology. This copying phenomenon is therefore particularly interesting for the discussion of copying and doubling and their respective theoretical accounts in phonology and/or morpholgy (Zuraw 2002; Inkelas 2008; Kirchner 2010).