Electrical stimulation of the basal temporal region of the dominant hemisphere before partial temporal lobectomy for epilepsy sometimes produces temporary interruption of language function, but the significance of removal of this area is unknown. We evaluated the correlation between resection of the basal temporal language areas (BTLA) and certain types of postoperative language deficits. In a population of 25 patients, we mapped the inferolateral temporal lobe with cortical electrical stimulation, verifying the positions of the stimulating electrodes with three-dimensional computed tomography (CT). Eighty percent of the patients developed transient language deficits with stimulation, particularly with tests of confrontation naming and comprehension. BTLA was primarily located in the fusiform gyrus, from 1 to 9 cm from the temporal tip. At testing 6-12 months after operation, patients with BTLA resection performed worse (mean 9% decrease) than those with no BTLA resection (mean 4% improvement) on tests of confrontation naming (p = 0.03). Resection size accounted for less of the variance in decrease in confrontation naming than did resection of the BTLA. Tests of performance I.Q. (PIQ), verbal I.Q. (VIQ), or recognition memory for word and verbal learning showed no significant difference between these groups. Most patients do not have language decrease with removal of basal temporal lobe 5-6 cm from the tip, even with removal of BTLA. Some patients, however, have persistent decrease in naming.