Semantic feature representations for normal and aphasic language

Aphasic and non-neurological patients grouped nouns on the basis of similarity of meaning. These word groupings served as input matrices for hierarchical clustering and multidimensional scaling analyses. The emergent structures suggest that, while the normal adult has a number of levels upon which to organize his lexicon, the adult aphasic's lexicon can be characterized as a set of partial entries that are tied to affective and situational data. The results also suggest that semantic feature representations derived from similarity-of-meaning judgments are of relevance in the study of factors which influence actual language performance.