Whamit!

The Weekly Newsletter of MIT Linguistics

Phonology Circle 12/15 - Sam Zukoff

Speaker: Sam Zukoff
Title: Repetition Avoidance Effects in Indo-European Reduplication
Date: Dec. 15 (M)
Time: 5 – 6:30
Location: 32D-461

Fleischhacker 2005 develops a theory of cluster-reduction under partial reduplication based on principles of perceptual similarity. The Indo-European languages Ancient Greek, Gothic, and Sanskrit, each of which have a default CV- prefixal reduplication pattern, play significant roles in demonstrating the typology predicted by her theory. In each of these languages, there are differences in copying patterns in reduplicative categories dependent on the sonority profile of initial clusters, generally with stop + sonorant clusters patterning with single-consonant-initial roots to the exclusion of obstruent + obstruent roots, which undergo some special treatment.

In this paper, I propose that the primary data from these languages admits also of an account based on repetition avoidance in poorly-cued contexts. The proposal hinges on the idea that local repetition of consonants is perceptually dispreferred (Walter 2007), and this dispreference is exacerbated when the second consonant lacks significant phonetic cues. Stop + sonorant sequences pattern with consonant + vowel sequences because both contexts permit significant phonetic cues to the root-initial consonant to surface, whereas fewer cues are available to the root-initial consonant in other environments.

This account yields equivalently satisfactory explanation of the basic Ancient Greek and Gothic facts, but allows for more complete coverage of the Sanskrit facts. There are two relevant patterns that do not follow directly from a similarity-based approach: (i) root-initial s-stop clusters copy the stop, contrary to normal leftmost copying, and (ii) certain CVC roots in categories where the root vowel is deleted show a vowel change rather than reduplication. The first type can be accommodated with Fleischhacker’s theory, but admittedly does not follow from principles of similarity. The latter type is not discussed by Fleischhacker, and does not obviously follow from her account. Both of these patterns can be analyzed in the repetition avoidance framework as avoidance strategies for what would be poorly-cued environments if reduplicated normally. A pattern almost exactly equivalent to the CVC pattern in Sanskrit can be reconstructed for an earlier stage of Gothic, providing an explanation for the Germanic “Class V” preterite plurals (Sandell & Zukoff 2014).