Zymet (2019) – Malagasy OCP targets a single affix: implications for morphosyntactic generalization in learning
Malagasy OCP targets a single affix: implications for morphosyntactic generalization in learning
direct link: http://ling.auf.net/lingbuzz/004540
Despite claims of the existence of phonological processes that apply strictly in limited morphosyntactic domains, recent corpus and experimental findings suggest that such processes are unproductive and difficult to learn. Martin (2011) and Chong (2017) consequently posit learning models having a generality bias, such that they cannot acquire morphosyntactically specific restrictions without supporting statistical tendencies from other domains. This article presents a corpus study of Malagasy backness dissimilation, showing that it applies consistently and exclusively to the passive imperative suffix, the only suffix eligible to undergo it. Dissimilation was extended to loanwords having the suffix and has persisted for multiple generations, but is entirely unsupported by the lexicon, which instead displays a phonotactic harmony tendency. These findings suggest that no degree of generality is necessary for learning. Rather, if the bias is active, then it must be overridable.
|Format:||[ pdf ]|
(please use that when you cite this article)
|Published in:||To appear in Linguistic Inquiry|
|keywords:||generalization, bias, dissimilation, harmony, malagasy, morphology, phonology|