Schmidt also suggests in his diary (as do other critics of Krashen) that learned knowledge can be used in spontaneous production. The problem, as we have stated (see section 2.2.4) is that much declarative knowledge about language in the form of language rules is so complicated that it is difficult to proceduralize. But even Krashen himself acknowledges that learners can use 'simple rules' to extend their grammatical resources. Simple rules are those:

• which don't involve elaborate movement or permutation (eg, changing a to an in front of a vowel);
• where the semantics of the rule is straightforward (eg, using who with humans and which with all other things).

These ideas chime with Green and Hecht's (1992) findings from the study discussed in Reflective task 13, that explicit knowledge of rules seemed to help more on corrections which involved easy or mechanical rules, rather than those which involved complex concepts such as tense and aspect usage.

As we have seen, work in SLA has challenged both the highly explicit presentation of formal structures in Grammar-translation and Audiolingualism, and the assumption in the Natural Approach that structure and form will take care of themselves through implicit acquisition. There is a strong realization in SLA now that both explicit presentation of forms, and opportunities for implicit acquisition are required for effective language teaching.