ABSTRACT
Investigators who study language acquisition from widely different theoretical perspectives rarely talk to one another. They have also been divided by their attitudes toward child language data. The problem of how children avoid an overly general grammar was first raised in 1971 by Martin Braine, who used it to argue against the nativist position espoused by Chomsky and in favor of the idea that language is learned largely from scratch. It was later revived by Baker who, in an interesting turn-around, made it the cornerstone of the argument that children must be guided by innate constraints in their acquisition of language. According to Chomsky, children construct an internalized grammar by using incoming language data, together with innate linguistic knowledge, to formulate hypotheses about possible grammatical rules. Braine pointed out that in order for a hypothesis-testing procedure to work, learners must get feedback about the correctness of their predictions.
