This paper reviews some of the predictions and assumptions made by different versions of the Critical Period Hypothesis which assumes that language acquisition will in principle be more successful in younger subjects. Findings from investigations of early vs. delayed acquisition of both a first and a second language which support this claim are presented. The question is then addressed whether the overall difference in ultimate attainment is a consequence of a loss of neural plasticity, or of the stabilizing of neural connections through the language learning process itself. The findings from a study of the L1 attrition of Korean adoptees in France are presented in support of the latter explanation.