Spoken word production and its breakdown in aphasia

ABSTRACT This thesis examines the processes involved in spoken word production. The disorders of naming observed in fifteen aphasic subjects are analysed in detail and the results related to current models of spoken word production. The data is used to support and refute predictions made by these models and the model is presented which best fits the patterns observed. Each of the fifteen aphasic subjects was tested using a range of tasks for assessment of both comprehension and production deficits. Their performance on a picture naming task was examined in detail. Statistical analyses were performed to study the effects of a number of variables on naming performance for the group as a whole and for individual subjects. These variables included written word frequency, familiarity, imageability, concreteness, word length (in terms of number of syllables and number of phonemes) and articulatory/ phonological complexity. Although there was no significant relationship between overall naming performance and comprehension, the production of semantic errors was significantly correlated with semantic errors in the comprehension of high imageability items. There was also found to be no relationship between the production and detection of phonological errors and input processing deficits. In contrast to many previous studies, there was not found to be a significant effect of written word frequency on picture naming for this group of aphasic subjects. However, a significant effect of familiarity was found and it is argued that this serves as a more accurate measure of spoken word frequency than the measures of written word frequency commonly used. Similarly, imageability was found to be a better measure of the influence of a semantic deficit on naming than concreteness. The majority of patients showed a significant effect of word length on the production of phonological errors and this length effect was predominantly found to be due to the number of phonemes, rather than the number of syllables, present in the word. The theoretical implications of these results are discussed.