Uptake, in Lyster and Ranta's (1997) influential framework, is the learner's immediate response following the teacher's corrective feedback. It is a discourse-level construct, an observable move in classroom interaction, not a direct measure of acquisition, though it may facilitate it by pushing learners to reprocess the problematic form.
Lyster and Ranta coded uptake into two broad categories:
| Category | Description | Subtypes |
|---|---|---|
| Repair | Learner successfully corrects the error | Repetition, incorporation, self-repair, peer-repair |
| Needs-repair | Learner responds but does not correct | Acknowledgment, same error, different error, off-topic, hesitation, partial repair |
A third possibility, no uptake, occurs when the teacher's feedback is followed by topic continuation with no learner response to the correction.
Lyster and Ranta's (1997) study of four French immersion classrooms found that recasts, despite being the most frequent CF type (55% of all feedback), generated the lowest rates of learner repair. Feedback types that withheld the correct form, namely elicitation, metalinguistic clues, clarification requests, and repetition (later grouped as prompts), were far more effective at generating student-produced repair.
This finding sparked a productive debate: does uptake matter for acquisition? Lyster (2004) argued that uptake involving modified output (self-repair) creates deeper processing and stronger memory traces. Long (2007) countered that recasts work through different mechanisms, providing positive evidence and implicit negative evidence, and that low uptake rates do not mean low acquisitional value.
Uptake connects Corrective Feedback to the Output Hypothesis: when learners repair their own errors, they are engaged in pushed output, testing hypotheses about the L2 and noticing the gap between their production and the target form.