| Summary: | The present thesis concerns the acquired reading disorder of deep dyslexia. Semantic errors (semantic paralexias) in reading aloud (e.g. reading 'ring' as 'wedding' or 'ruler' as 'rubber') constitute the cardinal symptom of deep dyslexia. Semantic errors of oral reading by aphasic patients have been said to be comparatively rare in languages with a shallow (transparent) orthography (e.g. Spanish and Italian). Miceli et al. (1994) argued in relation to reading aloud and writing that 'transparent orthographies are relatively protectedfrom the production o f semantic paralexias and paragraphias' (p.331). Thus the first part of the thesis reports a series of investigations of this claim in three bilingual readers of two orthographies, one deep, one shallow, namely English and Welsh. On a picture naming task, each of the three brain damaged patients made a similar proportion of semantic errors in the two languages as expected. However, contrary to the predictions of Miceli et al. (1994), in oral reading of the corresponding words no patient produced proportionally more semantic errors in English than in Welsh. Indeed, two of the patients made proportionally more semantic errors in Welsh. Therefore the findings of this thesis do not support the view that semantic errors are rare in a shallow orthography. It was concluded from the data that the patients could be considered deep dyslexic in both Welsh and English. Regression analyses revealed that age of acquisition influenced the production of semantic errors in Welsh reading for all three bilingual deep dyslexic patients and in English reading for two of the patients. This supports others findings (e.g. Gerhand & Barry 2000) that age of acquisition is the most salient factor that predicts a participant's response. The data were also in agreement with the viewpoint expressed by Morrison and Ellis (1995), among others, that a major component of what has been reported in the literature as frequency effects in lexical processing is in fact due to a confound with age of acquisition, as frequency was not found to exert an independent effect on the patients' responses. The semantic errors generated by the patients were earlier acquired, more frequent and were shorter in length than the target words to which the errors were made, supporting Gerhand and Barry's (2000) finding. Studies of bilingual aphasia considering the cognate status of words are extremely rare. It was examined whether cognate status influenced the accuracy of the patients' naming and reading responses. However, when the cognate items were removed from the analysis, it had little effect on the findings from the multinomial regression. No cognate facilitation effect was found in either language. The majority of theories of deep dyslexia attribute the occurrence of semantic errors to a lack of sub-lexical phonological ability. However, Katz and Lanzoni (1992) and Buchanan et al. (1994) claim that at least some deep dyslexics patients are sensitive to implicit phonological information. The second part of this thesis examined phonological decoding ability in deep dyslexia using pseudohomophones as stimuli. Implicit phonological ability was found in terms of the Stroop effect (increased reaction times to incongruent stimuli compared to congruent stimuli) using pseudohomophones but no effect was found with orthographically controlled nonwords. Patients were also significantly better at reading pseudohomophones than orthographic controls and showed the standard 'pseudohomophone effect' (extended reaction times) in lexical decision. However, no evidence o f semantic priming using pseudohomophones was found in the three deep dyslexics, even though the control participants did show an effect with the same priming stimuli as was used with the patients.
|
|---|
