A notation for the structure of cognitive processes, with application to patterns of verb-processing impairment

Abstract

A method is proposed for processing datasets that show whether or not each of several patients was impaired on each of several tests, and expressing conclusions about them. Its advantages are that (1) results from the patterns of impairment alone are shown, uninfluenced by theories, previous empirical work, knowledge of lesions, or ideas about what the tests are really testing, and (2) the presentation to the reader is briefer and clearer than a narrative account. The method is applied to data on the patterns of impairment of 30 subjects on six tests, each requiring the processing of verbs (Kemmerer et al., Language and Cognitive Processes , 2001). A new interpretation of this dataset is given, in terms of a set of three systems, all of which are essential. These systems resemble three of those that Kemmerer et al. hypothesised that the six tests might utilise.