The morpho-phonological interface in specific language impairment

• Main Author: Marshall, Chloe Ruth
• Published: University College London (University of London) 2005
• Subjects: 401.93
• Online Access: http://ethos.bl.uk/OrderDetails.do?uin=uk.bl.ethos.413671

This thesis investigates the nature of the interface between two components of language - morphology and phonology - in children with Grammatical-Specific Language Impairment (G-SLi), compared to those with typically-developing language. I focus principally on the impact of phonological complexity on past tense inflection, but I also investigate other areas of rnorpnotogy. More specifically, I show that for G-SLI children:- There exists a phonological impairment that is independent of morphology. This impairment is characterised by the simplification of complex syllable structure, and by syllabic and segmental errors when the word starts with an initial unstressed syllable. There exists an impairment in past tense morphology, characterised by suffix omission, that is independent of phonology. Phonological factors affect past tense morphology. Specifically, suffix omission rates are higher when inflection (i) creates clusters at the word-end or (ii) requires the syllabic allomorph lidJ. Phonological factors also affect plural and present progressive formation. Unlike past tense morphology, derivational morphology is not subject to suffix omission. However, non-target derivational forms result when stimuli are morphologically or phonologtcally complex. I argue that grammar has a modular structure, and I propose that deficits in one or more of the following modules - syntax, morphology and phonology - can impact on past tense inflection. This model, termed the 'Computational Grammatical Complexity* (CGC) hypothesis, can account for why tense is an area of exceptional difficulty for children with SLI. This investigation is underpinned by a rigorous theoretical framework. Not only does using a cognitive scientific and linguistic framework further our understanding of the nature of the deficit in SLI, but SLI provides a valuable testing ground for theories of language acquisition and the representation of language in the brain.