Abstract
The present work developed and explored the use of a multidimensional scheme of phonetic transcription and analysis that included conventional featural and segmental as well as syllabic and suprasegmental levels of description. In addition, the scheme incorporated the underlying infrastructure, or infraphonological level of the phonological system. The newly developed system sought to provide a basis for context-sensitive rule-based interpretations of errors in child productions that considers phonological universals and developmental processes through weighted measures of the adequacy of children's speech. The multidimensional analysis method was empirically assessed utilizing speech samples from children with severe-to-profound hearing-impairments to determine its reliability and to examine the relationship of the constituent components of the multidimensional phonological system to speech intelligibility. Results indicated that a global analysis measure that reflected the structural and segmental components of syllables was the best predictor of intelligibility.