Signs of language dysfunction in dementia of the Alzheimer's type (DAT) and in the aphasic syndromes of transcortical sensory aphasia and Wernicke's aphasia are superficially similar. The unresolved question concerning the extent to which the language disturbances of DAT are "aphasic" is linked to a more fundamental question concerning the relation of language to thought, given that aphasia is often defined as language disturbance without disturbance of intellect, and dementia as dissolution of intellectual function, of which language forms an integral part. In this paper we explore the historical roots of today's debate by analyzing the original case studies of Wernicke (1874) and Alzheimer (1907, 1911). Although each of these neurologists described similar patterns of language disturbance, they drew different conclusions. Wernicke argued for a distinction between language and thought and between the language disturbances of aphasia and those of dementia. Alzheimer continued the then dominant paradigm of aphasia in describing the language disturbances of his demented patients as aphasic. Paradoxically his conclusion makes him appear, in contrast to Wernicke, to argue for the identity of the language disturbances of aphasia and dementia. Yet he himself acknowledged that the presence of focal language symptoms arising from diffuse degenerative pathology was indeed problematic. We conclude that today's discussion could profitably be refocused on the question which emerges from the original works of Wernicke and Alzheimer, which Alzheimer himself asked, and which remains unanswered: How can diffuse cerebral pathology give rise to a pattern of language deficit virtually identical to that of a focal lesion?