A CVCV model of consonant cluster acquisition : evidence from Greek

• Main Author: Sanoudaki, Eirini
• Published: University College London (University of London) 2007
• Subjects: 410
• Online Access: http://ethos.bl.uk/OrderDetails.do?uin=uk.bl.ethos.498349

The aim of this thesis is to develop a model of the acquisition of consonant clusters within the phonological framework of CVCV theory. This is the first attempt to link CVCV to the area of language acquisition. It thus provides a new domain within which CVCV can be evaluated against other phonological theories. The core claim of CVCV is that syllable structure consists solely of onsets and nuclei, without any branching constituents. Consonant clusters are separated by empty nuclei, whose distribution is controlled by binary parameters. The model developed in this thesis is based on the assumption that a central part of the acquisition process is the gradual setting of these parameters to the appropriate value. The model, apart from covering familiar acquisition data, makes a number of predictions about the order of acquisition of consonant clusters. Of particular importance are predictions regarding word initial clusters of non-rising sonority, whose acquisition has attracted little attention. The predictions are tested against experimental data of cluster production by fifty-nine children acquiring Greek as their first language. The experimental results indicate that a CVCV model can account for consonant cluster acquisition. With regard to word initial position, the results support the proposed CVCV analysis by providing evidence for the existence of a word initial Onset-Nucleus unit. Moreover, the notoriously complex issue of s+consonant clusters is examined, and new evidence for the structure and markedness of these clusters is provided. Finally, the results offer a new perspective on a manner dissimilation phenomenon in Greek, whereby clusters of two voiceless fricatives or two voiceless stops turn into a fricative plus stop. A parametric analysis, based on segmental complexity, is proposed, and it is argued that this analysis can explain the acquisition data as well as the historical evolution of Greek clusters.