New Researchers Forum in Linguistics, University of Manchester and University of Salford, UK
The file is open to annotations via Hypothes.is. Comments are always welcome!
I present an analysis of Welsh ‘svarabhakti’, whereby a copy of a preceding vowel is inserted to break up final clusters of rising sonority (pwdr ‘rotten’ is [pudur] rather than *[pudr]). This pattern is said to coexist with a pattern of deleting consonants in such clusters in polysyllables (perygl ‘danger’ is [perɪɡ] rather than *[perɪɡl]). It has been argued that the two differing outcomes can be reconciled by appeal to a single constraint on the form of feet. I show that this analysis is not tenable for South Welsh, and propose that deletion (as in perygl) is not part of the phonological grammar (at least in South Welsh).