University College London · FACULTY OF LINGUISTICS

CCME0051 Chap.5 Input, Interaction and Corrective Feedback

- one subject, every graph, every model, every mark
7 Chapters4-page Bible
Our own words - no uploaded lecturer files
Updated for this semester
Chapter 5 of 10 · CCME0051

Input, Interaction and Corrective Feedback

Input, Interaction and Corrective Feedback examines how conversation helps a second language develop.

The unit brings together four frameworks: the Input Hypothesis, which stresses comprehensible input; the Interaction Hypothesis, which adds the repair work of negotiating meaning; the Output Hypothesis, which argues that speaking and writing push learners to spot gaps in what they can say; and the Noticing Hypothesis, which holds that forms must be consciously registered before they can be learned.

Corrective feedback is the point where these ideas meet, so much of the empirical research codes how teachers and partners respond to errors, whether through reformulations such as recasts and explicit correction or through prompts that ask the learner to self-repair.

The unit also compares feedback in face-to-face talk with technology-mediated settings such as videoconferencing, text chat and generative AI partners, and asks what the findings imply for classroom practice. For the empirical report, this is one of the most practical areas, because small studies of feedback and uptake can be run with a few pairs of learners and a clear coding scheme.

In this chapter

What this chapter covers

  • 01

    Comprehensible input and the Input Hypothesis

  • 02

    Negotiation of meaning and the Interaction Hypothesis

  • 03

    The Output Hypothesis and the limits of practice

  • 04

    Noticing and the route from input to intake

  • 05

    Recasts, explicit correction and prompts

  • 06

    Coding feedback moves and learner uptake

  • 07

    Feedback in video calls, text chat and chatbot interaction

Worked example · free

Follow corrective feedback through to modified output

Q [5 marks]. During a pair task a learner says: Yesterday we go to the museum. Her partner replies: Go? Yesterday? The learner then says: Yesterday we went to the museum. Identify the feedback move, trace the exchange through the output, feedback, noticing and modified output cycle, and say what a researcher can and cannot conclude from it. Marks here are an illustrative split for practice, not an official UCL allocation.
The loop starts with learner output that contains an error, passes through corrective feedback and noticing the gap, and ends in modified output, which researchThe loop starts with learner output that contains an error, passes through corrective feedback and noticing the gap, and ends in modified output, which researchers record as uptake.Learner outputCorrective feedbackNoticing the gapModified outputerror in messagelearner reformulatesuptake
The loop starts with learner output that contains an error, passes through corrective feedback and noticing the gap, and ends in modified output, which researchers record as uptake.
  • 1The first turn is learner output containing an error: the present form go where the past is required.
  • 1The partner's reply repeats the error and the time word with rising intonation, which is a prompt rather than a recast because no correct form is supplied.
  • 1The prompt highlights the mismatch between go and yesterday, giving the learner a chance to notice the gap.
  • 1The final turn is modified output and is coded as uptake with repair.
  • 1One repaired turn shows the feedback was registered at that moment, but only a delayed test could show that the past form has been learned.
The partner used a prompt, repetition with rising intonation, which led to modified output coded as uptake with repair. The exchange illustrates the full feedback cycle but supports only a claim about immediate repair, not about lasting acquisition.
Sia tip — Code a turn as a recast only when the correct form is actually supplied; a questioning repeat of the error is a prompt.
Glossary

Key terms

Comprehensible input
Language a learner can understand from context and current knowledge, slightly beyond their present level.
Negotiation of meaning
Repair work in conversation, such as clarification requests and confirmation checks, used to resolve a breakdown in understanding.
Uptake
The learner's reply immediately after feedback, coded by whether it repairs the original error.
Clarification request
A move such as asking the speaker to say something again, which signals a problem without supplying the form.
Modified output
A learner's reformulated utterance produced in response to feedback or a breakdown in communication.
FAQ

Input, Interaction and Corrective Feedback FAQ

Is input enough on its own for a second language to develop?

Input is necessary, but the unit's frameworks suggest it is not sufficient. Interaction helps make input understandable and supplies negative evidence, output pushes learners to notice gaps in what they can say, and noticing determines which forms become intake. Most researchers therefore study input together with interaction and feedback.

Why are recasts sometimes ineffective?

A recast supplies the correct form without signalling that an error occurred, so the learner may hear it as simple agreement or a request for confirmation. Whether a recast helps then depends on noticing, the learner's level and the setting, and reviews report that recasts and prompts work differently across classrooms.

What is the difference between a recast and a prompt?

A recast gives the learner the correct version of what they said. A prompt withholds the correct form and pushes the learner to repair, through moves such as clarification requests, repetition of the error, elicitation or a metalinguistic clue about the rule.

Does online interaction work the same way as face-to-face talk?

Not exactly. Video calls can add written or visual cues to a spoken recast, text chat leaves a record learners can scroll back through, and chatbot partners give feedback that depends on how the tool is set up. The unit's readings examine these settings, so a study should treat the medium as a variable.

How do researchers decide whether feedback led to learning?

Uptake in the next turn shows only that feedback was registered. Studies that claim learning usually test the target structure before and after the feedback, often with a delayed posttest, and compare groups that received different feedback types or none at all.

Study strategy

Assessment move

Start by writing each hypothesis as a one-line claim and listing the kind of evidence that would support or challenge it. Then memorise the six feedback moves by what they ask of the learner, and practise coding short exchanges you write yourself, including turns with no feedback and turns with no uptake.

Compare two settings from the unit's readings, such as video calls and text chat, and note what each makes easier to record. If interaction is your report topic, pilot your coding scheme on a five-minute recording, check agreement with a classmate, and decide how you will separate immediate uptake from later learning before collecting your main data, and record that decision in your method section.

Working through Input, Interaction and Corrective Feedback in CCME0051? Sia is AskSia’s AI Linguistics tutor — ask any CCME0051 Input, Interaction and Corrective Feedback question and get a clear, step-by-step explanation grounded in how CCME0051 is taught and assessed. Read this chapter free, then take your hardest questions to Sia.

A+Everything unlocked
Unlocks this Bible + all 32 of your subjects at University College London, plus 1,000+ more Bibles across the AskSia library.
Sia - your CCME0051 tutor, unlimited, worked the way the exam marks it
The full 4-page Bible + practice bank with worked solutions
Chrome extension - sync your LMS so Sia knows your deadlines
Bilingual EN / Chinese on every Bible and every Sia answer
$0.99 Trial
30-day money-back · cancel in one tap · how it works