Abstract
Thirty university students learned a miniature artificial language by watching linguistic inputs consisting of forty example sentences, presented one by one, in a subject-paced procedure where the time each subject spent for each exemplar was measured and recorded. The twenty subjects exposed to the linguistic inputs which contained 20% of mistaken examples acquired the linguistic rules as much as the other ten who were exposed to the correct exemplars only. Twelve of the 20 subjects noticed the presence of mistaken examples in their linguistic inputs. They tended to get higher scores on acquisition tests and spent a longer time watching the mistaken exemplars than their counterparts.

This publication has 1 reference indexed in Scilit: