A large number of observations since the mid-nineteenth century have shown that damage to the left but not the right hemisphere destroys language function. These observations lead to the formation of the "classic" view that the left hemisphere has language function but not the right hemisphere. It thus came as a considerable surprise in the early 1960's when tests on commissurotomy or split-brain patients suggested the presence of a considerable capacity for reading comprehension in the right hemisphere. It has been suggested that the spared regions in the left hemisphere normally act to prevent the expression of latent language functions within the undamaged right hemisphere. Thus, only after the intact right hemisphere has been released from the disruptive and suppressive influences of the damaged left hemisphere by section of the corpus callosum, can its own residual function become effective. Some investigators have argued, however, that the split-brain evidence is misleading because the language of the split-brain patients is presented in both hemispheres even pre-operatively. Our investigation with functional MRI is in progress to decide which hypothesis is appropriate. In addition to the functional MRI study of split-brain patient, the functional MRI study of aphasia recovery will also provide opportunities to clarify the right hemisphere language function since the right hemisphere is generally intact in the right-handed aphasic.