Results of recent studies, some of which appear in this volume, suggest that the incorporation of some degree of focus on form (FonF) into meaning‐centered instruction can lead to improved performance in processing input and increased accuracy in production. Long (1983) originally defined FonF as a brief turning of attention to some formal feature while the overriding focus of the interaction remains on meaning. Ideally, this focus should arise incidentally, in response to some problem in comprehension or production. The incidental nature of the interaction would be more likely to result in focus on a form to which the learner is attending and is ready to notice.
However, to ascertain the effect of such a focus in an experimental setting, it becomes necessary to manipulate the situation somewhat, in particular, to predetermine the formal focus and to ensure that it will occur. As a result, in most studies the focus is provided by the researcher and delivered intentionally, and usually extensively, through the teacher or special materials. Thus, the focus no longer arises incidentally, potentially vitiating the effectiveness of the intervention.
This study differs from previous studies in two ways. First, it is set in intact classrooms, so all episodes involving focus on form arise incidentally. Second, it examines only such episodes as are initiated by learners. The justification for this is an extension of the argument for incidental FonF. If the effectiveness of FonF is ultimately determined by learner need, then it is essential to examine the episodes in which the learners themselves choose to focus on formal aspects of language.
The study addresses the following research questions: Do learners initiate attention to form, as requests for assistance, feedback on error, modeling, or fine tuning, that is, repetitions, recasts, and requests for clarification? If so, on what kinds of forms do they choose to focus? The results suggest that learners can and do attend to form, though relatively infrequently. The most frequent way that they do this is t o request assistance from their teachers. Learner‐generated attention to form increases considerably with rising proficiency and during specific activities. In general, the likelihood of learner‐generated attention to form seems to be linked to learners' perception of the goals of the activity. In addition, as learners' proficiency increases, they are less likely to rely on their teachers for help. Finally, episodes that involve learners‐generated attention to form revolve largely around lexis. More than anything else, learners want to know about words.