We studied the spoken language of 49 children who had undergone hemispherectomy as part of the UCLA Pediatric Epilepsy Surgery Research Program and analyzed, among a number of clinical factors, the relation between acquired vs developmental pathology and spoken language outcomes. In this paper we will briefly review the results of our study and attempt to explain (1) why "the early" is not always better, (2) why so many right hemispherectomies fail to develop language, and (3) why some left hemispherectomized children develop remarkably good language despite removal of the "language" hemisphere. This account will rest on the proposed model of brain maturation and progressive lateralization.