Etiology should be fully addressed by regulatory agencies, researchers, and the pharmaceutical industry as a variable in the design of antiepileptic drug trials. A greater emphasis on etiology, and greater efforts to determine etiology, should in the future improve clinical practice. In epilepsy, the most important advances have been made in relation to the symptomatic epilepsies of metabolic or congenital origin. The clinical data have been provided largely by MRI, and the rate of incrementation of useful new information in this area is now relatively slight. The environmental causation of epilepsy remains largely terra incognita, and yet is most intriguing not least because of the tantalizing possibility of fresh approaches to therapy. A classification scheme based on etiology would be a very different animal from the current schemes based on semiology or electroencephalography. There are significant drawbacks to current schemes for clinical work and for research and conceptualization in epilepsy.