Modality-specific operations in semantic dementia

A patient suffering from semantic dementia is described who consistently demonstrated the preserved ability to support specific types of semantic judgements from visual, but not from verbal, input. In addition the representations accessed from visual input were found to trigger complex behavioural schemata, while with verbal materials the patient performed almost invariably at chance level. A preliminary description is given of the nature of visual semantic representations, and the privileged relationship between this modality of input and some aspects of semantic knowledge is also explored. The richness of the semantic representations accessed from visual input can be accommodated within the "Multimodal Semantics" framework; alternative views, derived from the Identification Semantics and the Organized Unitary Content Hypothesis, are also discussed.