Abstract:
Explores the relationships between comprehensibility (well-formedness) and magistrates' credibility judgements, the ethnicity of witnesses and their court roles. The first 36 cases that included undefended accused who had pleaded not guilty and who had led evidence were chosen. An equal number of testimonies judged by magistrates to be credible and not credible, and an equal number presented by 'coloured' and 'white' witnesses and by complainants and accused were included. Three common users assessed the credibility of 4 cases in random order. They defined 9 types of incomprehensibilities: irrelevant and delayed information, direct and empirical contradictions, expectation conflicts, ambiguous wording, missing and inadequate justifications, and other problems. A qualitative analysis showed that encoding problems were caused by 'defects' in the semantic structure of testimonies. It was noted that each type of encoding problem was expected to affect credibility differently. The common users concluded that they had used incomprehensibilities as indicators of deception. Three different common users noted incomprehensible propositions in the remaining 32 cases in random order. These measures were operationalized as percentages of the total number of propositions. It was found that testimonies judged by magistrates not to be credible and those presented by coloured witnesses, were perceived by common users to be less comprehensible than those that magistrates had judged to be credible and presented by whites. The relative proportions of each type of encoding problem did not differ according to whether the testimonies were credible or not. nor according to whether they were coloured or white, accused or complainants. It is inferred that magistrates' credibility judgements and verdicts are influenced by the well-formedness of testimonies, and that coloured witnesses may be disadvantaged in interracial cases.