Neural subsystems for object knowledge.

Critical issues in the cognitive neuroscience of language are whether there are multiple systems for the representation of meaning, perhaps organized by processing system (such as vision or language), and whether further subsystems are distinguishable within these larger ones. We describe here a patient (K.R.) with cerebral damage whose pattern of acquired deficits offers direct evidence for a major division between visually based and language-based higher-level representations, and for processing subsystems within language. K.R. could not name animals regardless of the type of presentation (auditory or visual), but had no difficulty naming other living things and objects. When asked to describe verbally the physical attributes of animals (for example, 'what colour is an elephant?'), she was strikingly impaired. Nevertheless, she could distinguish the correct physical attributes of animals when they were presented visually (she could distinguish animals that were correctly coloured from those that were not). Her knowledge of input stimulus. To explain this selective deficit, these data mandate the existence of two distinct representations of such properties in normal individuals, one visually based and one language-based. Furthermore, these data establish that knowledge of physical attributes is strictly segregated from knowledge of other properties in the language system.