To the question of whether Chomsky's hypothesized Language Acquisition Device (LAD) in young children is an adequate and feasible model of language acquisition, this paper answers that LAD should be reformulated so as to include semantics; that "informant presentation" rather than "text presentation" is responsible for language acquisition; and that nonlinguistic knowledge may be crucial in syntax acquisition. The first argument is supported by experiments which show that learning strategies differ according to whether or not the learner has semantic information. The second argument is supported by experiments which show that parents present semantic-specific and phonology-specific feedback to the learner, and, indirectly, syntactic feedback. Finally, situations are discussed in which nonlinguistic knowledge may be the basis for linguistic hypotheses. The paper's conclusion is that this approach may lead to insight into the link between cognitive development and language acquisition. (AM)