This article assesses what standards of safety and certainty of diagnosis need to be met in the determination of brain death. Recent medical, legal, and philosophical developments on brain death are summarized. It is argued that epistemologically adequate standards require the finding of whole-brain death rather than destruction of the cortex. Because of the possibility of positive error in misdiagnosing death, a tutioristic approach of being on the safe side is advocated. Given uncertainties in diagnosis of so-called vegetative states like the apallic syndrome, anything less than whole-brain death, especially given the present state of diagnostic capability, should not qualify as an argument for removing therapy specifically on grounds that the patient is dead.