Between natural and unnatural phonology: The case of cluster-splitting epenthesis
Description
A widely recognized feature of loan-word phonology is the resolution of clusters
by vowel epenthesis. When a language lacking word-initial clusters borrows words from a language with initial #TRV sequences, T an oral stop and R a liquid, it is common to find vowel epenthesis, most typically vowel-copy, as in, for example: Basque <gurutze> ‘cross’ from Latin <cruce(m)>; Q’eqchi’ <kurus> ‘cross’ from Spanish <cruz> ‘cross’, or Fijian <kolosi> ‘cross’ from English <cross>. The phonological rule or sound change responsible for this pattern is sometimes called ``cluster-splitting epenthesis": #TRVi > #TV(i)RVi. The most widely accepted explanation for this pattern is that vowel epenthesis between the oral stop and the following sonorant is due to the vowel-like nature of the TR transition, since #TRVi is perceptually similar to #TV(i)RVi. A fact not often appreciated, however, is that cluster-splitting epenthesis is extremely rare as a language-internal development. The central premise of this chapter is that #TRVi in a non-native language is heard or perceived as #TV(i)RVi when phonotactics of the native language demand TV transitions. Without this cognitive component, cluster-splitting epenthesis is rare and, as argued here, decidedly unnatural.
Files
1.pdf
Files
(167.3 kB)
Name | Size | Download all |
---|---|---|
md5:1e5d9223a3c26f4fa38dfe2db985f296
|
167.3 kB | Preview Download |