Memory for generalities: Access to higher-level categorical relationships in amnesia

This paper presents results from four studies in which the distinction between exemplar and higher-level categorical knowledge was investigated in amnesic patients and non-memory-impaired controls. After studying a series of face-occupation associations it was found that patients could only discriminate faces at a level of knowledge that was higher in a specified occupational hierarchy than the one studied. This effect was illustrated most clearly in Experiment 1, where patients discriminated the faces of educators from those of tradespeople accurately, but were unable to discriminate between the types of educator and types of tradespeople that they had originally studied. This differentiation in performance between exemplar and more general level knowledge in amnesia is referred to as preserved memory for generality. A similar tendency to discriminate faces at the higher occupational level was observed in Experiment 2, where the study context was changed to include subtypes of teacher and lecturer. For one patient there was some evidence of improvement in ability to discriminate ata level of categorisation that was not accessible in the first experiment (i.e. at the level of teacher/lecturer). In Experiment 3 we investigated the contribution of direct and indirect testing to results of the previous studies. It was found that indirect testing facilitated access to higher-level knowledge. Results of a final study indicated that non-memory-impaired controls also produced superior memory for generalities under suboptimal learning conditions. However, unlike patients, controls performed equally well on direct and indirect tests of higher-level knowledge. These results are discussed with reference to theories of amnesia and to explanations of the phenomenon in dementia.