Definition:
Conversational repair refers to the practices and sequences by which interlocutors manage and resolve communication problems in ongoing talk — including mishearings, misunderstandings, formulation difficulties, speaking errors, and reference failures. The term comes from Conversation Analysis (CA), where repair is analyzed as a pervasive, systematic, and cooperatively managed feature of all interaction. Repair is not primarily about grammar correction — it is about maintaining mutual understanding and intersubjectivity in conversation. In SLA research, repair — especially interactional feedback through repair sequences — has been studied extensively as a mechanism through which learners receive input about errors and refine their interlanguage.
Organization of Repair
Schegloff, Jefferson and Sacks (1977) established the fundamental organization of repair along two dimensions:
Dimension 1: Who initiates repair?
- Self-initiated: The speaker who produced the trouble source initiates repair
> “I went to the — um — the market yesterday.” (self-initiated mid-utterance)
- Other-initiated: Another participant signals a problem and prompts repair
> A: “She was at the, uh…” B: “At the what?” (other-initiation)
Dimension 2: Who carries out the repair?
- Self-repair: The speaker who produced the trouble source fixes it
- Other-repair: Another participant completes or corrects
This produces four main repair types:
| Type | Example |
|---|---|
| Self-initiated self-repair | “I’ll see you — let me rephrase — I’ll meet you at five.” |
| Self-initiated other-repair | A: “His name is…?” (pause, looks to B) B: “Takahashi.” |
| Other-initiated self-repair | A: “She went with Kim.” B: “With who?” A: “With Kim, my coworker.” |
| Other-initiated other-repair | A: “She went to Paris last year.” B: “London, wasn’t it?” |
Preference for Self-Repair
A key CA finding: self-repair is preferred over other-repair in natural conversation. Speakers are given maximal opportunity to initiate their own repairs:
- Other-initiation gives the speaker a chance to self-repair before the listener corrects
- Direct other-correction is a dispreferred sequence that threatens face and intersubjectivity
- Simple repetition-with-rising-intonation (“She went when?”) initiates repair while allowing self-correction
Repair in SLA Contexts
In SLA, repair sequences are studied as a primary mechanism of interactional feedback:
Recast (implicit corrective feedback):
The teacher/NS reformulates the learner’s erroneous utterance correctly without explicitly flagging it as a correction.
> Learner: “He go to school yesterday.”
> Teacher: “He went to school? What did he do there?”
The teacher recasts go → went while maintaining conversational flow.
Negotiation of meaning:
When communication breaks down due to comprehension failure, the interlocutors negotiate:
> Learner: “I went to hanabi yesterday.”
> NS: “Sorry, hanabi?”
> Learner: “Yes, fireworks — it’s a Japanese festival.”
This negotiation pushes the learner to produce modified output (Long’s Interaction Hypothesis).
Explicit correction:
> Learner: “She is married since five years.”
> Teacher: “Actually, we say: she has been married for five years.”
Repair Sequences and L2 Interaction
CA-informed SLA research (Schegloff; Markee; Hall) shows that:
- Repair in NS–NNS interaction is often organized differently than in NS–NS interaction
- Learners often interpret recasts as mere repetition rather than corrective reformulation
- The explicitness of repair initiation affects whether learners notice and uptake the targeted form
Conversational Repair in Japanese
Japanese repair practices include:
- Particle repairs: Correcting a particle choice mid-utterance
> “駅に — じゃなくて — 駅から来た。” (“Coming from, not to, the station”)
- Register repair: Switching from casual to polite form mid-sentence when context shifts formality
- Repetition-based other-initiation: Simple repetition with rising intonation (“〜に？”) prompting clarification
- あの、えっと、なんか as filled pause repair initiators — equivalent to English “um, uh”
History
The systematic study of conversational repair originated in conversation analysis (CA) with Schegloff, Jefferson, and Sacks’s (1977) foundational paper on self-correction and repair sequences in American English conversation. Their analysis documented the sequential organization of repair: the trouble source (what prompts repair), the repair initiation (where repair begins), and the repair itself. This work established that repair is highly organized and socially meaningful — the timing and manner of repair initiation reveals whether a speaker is engaged, understanding, or experiencing difficulty. Cross-linguistic CA research subsequently documented repair patterns in numerous languages, revealing both universal tendencies and language-specific organization. In SLA, repair was studied as a mechanism through which learners signal comprehension problems and obtain comprehensible input (Varonis and Gass, 1985).
Common Misconceptions
“Conversational repair is just correcting mistakes.” Repair in the conversation analytic sense is much broader than correcting grammatical errors. It encompasses any process a speaker or hearer uses to address trouble in speaking, hearing, or understanding within ongoing conversation — including mishearings, topic divergences, unclear reference, and word searches, not just linguistic errors.
“Only less proficient speakers do a lot of repair.” Repair frequency in L2 communication reflects the mismatch between current L2 proficiency and communicative demand, but native speakers also perform repair constantly — to reformulate ideas more precisely, to hedge claims, or to initiate other-repair when they don’t understand an interlocutor. Repair is an observable marker of communicative competence at all levels.
Criticisms
Conversation analytic research on repair has been criticized for its limited generalizability — detailed analysis of small conversational samples in specific contexts may not represent broader patterns across genres, modalities, and language communities. The application of CA repair findings to L2 learning research has been questioned by researchers who note that the conditions of natural conversation differ substantially from classroom interaction, so repair patterns in classroom discourse require independent analysis rather than direct extension from natural conversation research.
Social Media Sentiment
Conversational repair is discussed in language learning communities primarily as a practical skill — how to ask “wait, what did you say?” or “could you repeat that?” and how to signal non-understanding without stopping the conversation entirely. Japanese learners discuss the use of あの (ano), えーと (eeto), and other repair initiators as markers of genuine conversational competence. Strategies for initiating repair in real conversations rather than textbook scripts are popular content in intermediate-advanced learner communities who are beginning conversations with native speakers.
Last updated: 2026-04
Practical Application
Learning L2 repair strategies is essential for conversational survival. Beginners should master basic self-repair tools (backtracking, rephrasing in simpler terms) and other-repair initiators (clarification requests, confirmation checks) in their target language. Authentic conversation exposure — language exchange, immersion, conversation courses — allows learners to encounter and practice repair sequences in real communicative contexts.
Related Terms
- Adjacency Pair
- Turn-Taking
- Negotiation of Meaning
- Corrective Feedback
- Interaction Hypothesis
- Discourse Analysis
See Also
Research
Schegloff, E. A., Jefferson, G., & Sacks, H. (1977). The preference for self-correction in the organization of repair in conversation. Language, 53(2), 361-382.
The foundational conversation analytic paper documenting the systematic organization of conversational repair, establishing the structure of repair sequences and the social preference for self-initiation — the basis for all subsequent CA repair research across languages.
Varonis, E. M., & Gass, S. (1985). Non-native/non-native conversations: A model for negotiation of meaning. Applied Linguistics, 6(1), 71-90.
Applies conversation analytic insights to L2 interaction, documenting the repair and negotiation sequences that arise when non-native speakers encounter comprehension difficulties — foundational for the interaction hypothesis and research on how interactional trouble resolution supports vocabulary and grammatical acquisition.
Nakayama, T. (1997). Conversational repair strategies among Japanese learners of English. JALT Journal, 19(2), 189-206.
Documents how Japanese learners of English use and initiate conversational repair in cross-linguistic interaction, comparing learner repair patterns with native speaker norms and examining what repair strategies are most effective for maintaining conversation flow.