Specificity in the development of language structure

The nature of children's productive structures can be investigated via two types of structural error which are manifest in their surface data: the overextension of a very general adult rule to a context from which it is excluded, and the introduction of a novel rule in a specific context. Examples of these two types of errors are shown, on the one hand, to undermine theories which treat language development largely as a process of projecting and testing innately specified hypotheses; on the other hand, they provide support for more recent theories which conceive of abstract linguistic structure as the outcome of a process relating specifically acquired items and combinations of items. The implications of this conception for the goals and tasks of developmental psycholinguistics are explored.This paper is based on research undertaken for the Ph.D degree under the supervision of Dr. N. V. Smith, University College, London, and with the financial support of the Social Science Research Council of Great Britain.     

Phoneme frequency effects in jargon aphasia: a phonological investigation of nonword errors

This study investigates the nonwords produced by a jargon speaker, LT. Despite presenting with severe neologistic jargon, LT can produce discrete responses in picture naming tasks thus allowing the properties of his jargon to be investigated. This ability was exploited in two naming tasks. The first showed that LT's nonword errors are related to their targets despite being generally unrecognizable. This relatedness appears to be a general property of his errors suggesting that they are produced by lexical rather than nonlexical means. The second naming task used a set of stimuli controlled for their phonemic content. This allowed an investigation of target phonology at the level of individual phonemes. Nonword responses maintained the English distribution of consonants and showed a significant relationship to the target phonologies. A strong influence of phoneme frequency was identified. High frequency consonants showed a pattern of frequent but indiscriminate use. Low frequency consonants were realised less often but were largely restricted to target related contexts rarely appearing as error phonology. The findings are explained within a lexical activation network with the proposal that the resting levels of phoneme nodes are frequency sensitive. Predictions for the recovery of jargon aphasia and suggestions for future investigations are made.     

Why Mikey's right and my key's wrong: The significance of stress and word boundaries in a child's output system

The systematic errors children make in the course of phonological development, like adult production errors and adult phonological processes, can provide evidence of language production mechanisms. A detailed investigation of the environments in which velar stops are fronted by a phonologically delayed child reveals that fronting is dependent on both word stress and word boundaries; that it shows lexical exceptions; and that it occurs in output only. This distribution suggests that the child has output lexical representations which are independent of input lexical representations, and that the fronting error occurs in these output representations. It also suggests that prosodic features are crucial to the identification of articulatory features within these representations. Such an analysis has implications for theories of lexical access, and for the development of lexical access in children.