Research on the role of age in second language (L2) learning, particularly at the level of primary education, is reviewed and discussed. It is concluded that evidence suggests early L2 exposure increases chances of ultimately attaining a high proficiency level in that language, but that in formal educational situations any long-term advantage will be slow to manifest itself and may not do so at all unless articulation between primary and secondary programs is properly managed. Some L2 learners may attain native-like L2 proficiency without an early start. These findings do not resolve the question of whether primary school L2 instruction is good, but do imply some questions for curriculum planners, including: what proficiency level should be required or useful to learners in the long term; what the chances are of ensuring that input at every stage of learning is appropriately focused, abundant, and enhanced; and what degree of coordination is possible between primary-level and secondary-level language programs? Decisions made about primary school language learning must be made with the same planning and foresight and on the basis of broadly the same preoccupations as other aspects of language in the curriculum. Contains 36 references. (MSE)