similarities across acquirers as to which structures tend to be acquired early and which tend to be acquired late (Brown, 1973; Dulay and Burt, 1975). Acquirers need not have a conscious awareness of the "rules" they possess, and may self-correct only on the basis of a "feel" for grammaticality.
Conscious language learning, on the other hand, is thought to be helped a great deal by error correction and the presentation of explicit rules (Krashen and Seliger, 1975). Error correction it is maintained, helps the learner come to the correct mental representation of the linguistic generalization. Whether such feedback has this effect to a significant degree remains an open question (Fanselow, 1977; Long, 1977). No invariant order of learning is claimed, although syllabi implicitly claim that learners proceed from simple to complex, a sequence that may not be identical to the acquisition sequence.
The fundamental claim of Monitor Theory is that conscious learning is available to the performer only as a Monitor. In general, utterances are initiated by the acquired system--our fluency in production is based on what we have "picked up" through active communication. Our "formal" knowledge of the second language, our conscious learning, may be used to alter the output of the acquired system, sometimes before and sometimes after the utterance is produced. We make these changes to improve accuracy, and the use of the Monitor often has this effect. Figure 1 illustrates the interaction of acquisition and learning in adult second language production.
The acquisition-learning distinction, as I have outlined it, is not new: Lawler and Selinker (1971) propose that for rule internalization one can "postulate two distinct types of cognitive structures: (1) those mechanisms that guide 'automatic' language performance... that is, performance... where speed and spontaneity are crucial and the learner has no time to consciously apply linguistic mechanisms... and (2) those mechanisms that guide puzzle- or problem-solving