For more than half a century now the idea has been that children are better language learners than adults and that therefore second language learning in school should start early. Ironically, until fairly recently, the SLA literature on age effects, which was used to advocate foreign language in the elementary school, comprised hardly a single classroom study. The more recent literature does include a number of interesting classroom studies that show that children are not necessarily better language learners in the classroom, and actually worse learners for the most part; this literature has now been used to argue against the concept of the critical period. In this article I argue that these two bodies of literature are not contradictory and that age effects cannot be fully understood without reconciling them, showing that children are better at one particular kind of learning. This has profound implications for theories of cognitive development, theories of second language learning, and second language curriculum design.