What does research indicate about pedagogical methods for independent language use through sustained conversational interaction in additional-language acquisition, focusing on interactional feedback, repeated exposure, and explicit instruction?

Multiple reviews and meta-analyses of interaction research conclude that engaging in conversational interaction — including negotiating meaning when communication breaks down — produces measurable gains in second-language ability compared to no interaction, with effects showing up even more strongly on delayed tests given weeks later.⁴

Oral corrective feedback (responses to a learner's speech errors) can meaningfully improve accuracy, but researchers distinguish between feedback that just supplies the correct form (like 'recasts,' which rephrase the learner's error correctly) and feedback that prompts the learner to fix the error themselves (like asking 'Sorry? ' or 'You need the past tense'). Self-correction prompts tend to be viewed as more effective for durable learning and learner motivation, though they only work if the learner already has some underlying knowledge of the correct form.^1,3,5

A widely supported approach is to first try to elicit self-correction from the learner, and only give the correct answer directly if that fails — this balances learner autonomy with the practical need to keep the conversation moving.¹

Feedback can be given in the moment (during a task) or afterward (after the task is finished), and researchers disagree about which is better. Giving corrections during the task allows an immediate side-by-side comparison of wrong and right forms, but giving corrections afterward may prevent learners from becoming so fixated on grammatical form that they neglect the primary goal of communicating meaning.¹

Studies of both immersion classrooms and general L2 classrooms found that explicit attention to grammatical form — noticing rules, discussing language consciously, and getting controlled practice — combined with corrective feedback led to better long-term grammatical development than communicative interaction alone, especially for language features that learners had plateaued on despite years of communicative practice.^6,8,12

When learners perform the same or similar speaking task more than once, their fluency (speed, fewer pauses) reliably improves, but the effects on grammatical accuracy and sentence complexity are mixed or even show trade-offs — improving one area sometimes comes at the expense of another.¹¹

When learners first interact with a teacher who pushes them for clear, complete language, they later perform better — with more output, more information, and wider vocabulary — when repeating a similar task with a peer, compared to learners who hadn't had that teacher interaction first.¹¹

Learners frequently misunderstand what a piece of corrective feedback is actually targeting. In one study, learners correctly identified grammar-focused feedback as being about grammar only 13% of the time, most often mistaking it for a comment about meaning instead.⁷

Corrective feedback tends to help lower- and medium-level learners improve their accuracy more than advanced learners, whose accuracy is often already fairly solid — advanced learners benefit more from work on fluency and complexity instead.³

A qualitative study of Chinese university EFL classrooms found that feedback works best when teachers calibrate its timing and directness to protect students' comfort and social standing (avoiding public embarrassment), not just when it targets the right error — emotional safety and trust shape whether students actually engage with a correction.²

Comparing content-based (subject-matter taught through the L2) and traditional foreign-language classrooms, teachers in more form-focused, grammar-oriented lessons used more explicit correction, while teachers in meaning-focused, content-driven lessons relied more on implicit correction like recasts.⁹

Even in non-classroom digital settings — text chat, voice chat, and online tandem language exchanges with native or near-native speakers — learners engage in negotiation of meaning and receive corrective feedback, though the amount and type varies: text chat produced more explicit correction, and post-conversation feedback (reviewing a chat log afterward) achieved higher correction rates than corrections made in real time.^10,15,16,17

The person a learner talks to changes the interaction: teachers who deliberately push for clearer, more complete speech elicit stronger performance gains than casual peer talk, and adult learners negotiate meaning somewhat differently than child learners, with adults showing more use of confirmation checks.^11,13

How a learner personally views the value of interacting and receiving corrections — their 'interaction mindset' — affects how much they actually engage with corrective feedback and collaborative talk, which in turn affects how much their language ability improves.¹⁴

Are recasts (simply repeating a corrected version back) actually useful, or do learners just ignore them?

Recasts can be beneficial — one classroom study of adult ESL learners found recasts led to successful uptake or improved test scores at least half the time — but their effectiveness depends heavily on how they're delivered. Recasts that used stress, a declarative (statement-like) tone, changed only one thing at a time, and were paired with other feedback moves were more likely to lead to learners actually adopting the correction, whereas vaguer or overly implicit recasts were more likely to be missed entirely, especially since research on learner perceptions shows students often misidentify what a piece of feedback was even correcting.