This paper examines the consequences of the fact that human minds may know more than one language for the poverty-of-the-stimulus argument that speakers know more than they could have learnt. Qualifications to the argument are necessary because not all L2 learners attain the same compe tence as L1 speakers; types of evidence are potentially available that can be ruled out for the Ll. The way in which L2 learning has been concep tualized in terms of access to UG and of a black box metaphor makes the L2 grammar seem separate from the L1 grammar rather than one over all system contained within a single mind. What is needed is the idea of 'multicompetence' - the compound state of a mind with two grammars. Multicompetence is the norm for the human race in that most minds know more than one language. Hence the logical problem of language acquisition is how the mind acquires a grammar with one or more settings for each parameter, rather than the special case of a mind that knows only one language. This has implications for all uses of the poverty-of-the-stimulus argument, not just in L2 learning.
Stored only when both papers are in the corpus — 533 citations exist upstream.