How a Good Speaking Coach Responds: 9 Moves From Speech Therapy and Language Research

A good speaking coach responds to what you just said, not to a script. It has nine moves: build, comment, fix, slow down, wait, ask again, echo, hand back and upgrade. Which move fits depends on what just happened: an error, a short sentence, a misunderstanding or a stall.

The moves come from speech-language pathology (SLP) and second language research. They are learning strategies borrowed from speech therapy, not therapy. Learning a language is not a disorder, and neither is an accent.

The 9 moves at a glance

Every example comes from a picture-story retell: The Tortoise and the Hare, The Three Little Pigs or The Lion and the Mouse. The coach's change is in capitals where stress matters.

In Edgewise, the coach is designed around the first eight moves, and Upgrades is planned for advanced learners.

Three jobs: add, fix and make room

The nine moves sort into three jobs. Knowing the job tells you what the coach wants from you next.

Moves that add: Builds, Comments, Upgrades

Builds keeps your words and adds one piece, such as a place, a describing word or a because. In child language intervention, recasts (the family expansions belong to) showed average effects of d = .96 on near measures and .76 on broader ones. The authors call the evidence limited, and study quality varied (Cleave et al., 2015).

Comments adds meaning, not grammar: a reason, a feeling, a guess about what happens next. Among 26 child-care providers, this kind of responsiveness had a strong positive link to preschoolers' language productivity (Girolametto & Weitzman, 2002).

Upgrades swaps a plain word for a Tier 2 word: a high-utility word more typical of written language than of conversation (Beck, McKeown & Kucan, 2013). It suits learners whose sentences are already accurate.

Moves that fix: Fixes, Echoes, Hands back

Fixes is the recast. It's the most common correction in classrooms: 55% of teacher feedback in French immersion lessons (Lyster & Ranta, 1997). It keeps the conversation moving, but it does the repair for you.

Echoes and Hands back are prompts. They point at the problem and leave the repair to you. Across 921 error sequences in immersion classrooms, this "negotiation of form" led to more immediate repair than recasts for word and grammar errors. Pronunciation was the exception: those repairs came mostly from recasts (Lyster, 1998).

Moves that make room: Slows down, Waits, Asks again

Slows down is not dumbing down. In a controlled comparison, comprehension was best when content was repeated and rephrased in interaction. Making the language simpler was not a significant factor (Pica, Young & Doughty, 1987).

Waits gives you time to plan. Teachers typically wait less than one second after a question (Rowe, 1986). When average wait time went past three seconds, both teacher and student talk changed (Tobin, 1987).

Asks again tells you the message didn't land. Clarification requests help both comprehension and production, yet they are rare in classrooms, where teacher and learner roles are unequal (Pica, 1987).

Which move fits? A decision table

The coach reads the last thing you said and picks one move. Here's the logic.

The first two rows rest on a simple idea. A prompt only works if the fix is somewhere in your head. If the form is brand new, hearing it said right is the faster route.

What makes feedback land?

Most coaching fails quietly. The learner never notices the change. Three features make it stick.

One change, stressed, as a statement. In 12 adult ESL classrooms (118 learners), recasts helped at least half the time on later tests. Successful uptake was predicted by stress, declarative intonation and making only one change (Loewen & Philp, 2006). That's why Fixes stresses one word and says it flatly. Echoes uses rising intonation on purpose: it's a question, asking you to repair.

Grammar needs extra help to be noticed. Learners who watched replays of their sessions perceived word, meaning and sound feedback fairly accurately. Grammar was different: "morphosyntactic feedback was generally not perceived as such" (Mackey, Gass & McDonough, 2000). So a fix on an ending or a verb form needs stress, and often a second pass.

Contingency. Good responses are timed to what you just said and build on it. In infants, responsiveness that is prompt and contingent helps map words to things (Tamis-LeMonda, Kuchirko & Song, 2014). A decade of classroom research names contingency, fading and transfer of responsibility as the core of scaffolding (van de Pol, Volman & Beishuizen, 2010). The coach helps, then hands the work back.

A note on evidence. Much of this comes from children or classrooms, not adults practicing alone with an app. We use it as a guide to how help should work, not as proof that any app works.

Prompts or recasts? It depends on your level

Researchers have argued about this for decades. The short answer: both help, and your level changes which helps more.

• Prompts often win in classrooms. A meta-analysis of 15 classroom studies (N = 827) found significant, durable effects of oral feedback. Effects were larger for prompts than recasts, and larger for younger learners (Lyster & Saito, 2010).
• Proficiency matters. Among 64 sixth-graders, high-proficiency learners gained equally from both. Low-proficiency learners gained significantly more from prompts (Ammar & Spada, 2006).
• The adult, one-on-one counterpoint. With 42 adult ESL learners in pairs with a teacher, recasts led to more immediate correction than elicitations. More explicit versions of both worked better (Nassaji, 2009).
• Implicit feedback lasts. Across 33 studies, corrective feedback had a medium effect that held over time. Implicit feedback's effects were better maintained than explicit feedback's (Li, 2010).

What to take from it: use prompts on one or two forms you've already studied. Let recasts handle the rest. Whichever move the coach uses, make the change clear enough to hear.

Why comment more and quiz less?

A display question tests what the asker already knows: ¿De qué color es la casa? (What color is the house?) A genuine question asks for something new: ¿Por qué crees que la liebre se durmió? (Why do you think the hare fell asleep?)

The kind of question matters. In a study of 4 teachers and 24 adult ESL learners, teachers were trained to ask more genuine questions. Their classes gave answers that were significantly longer and more syntactically complex, with more connectives (Brock, 1986).

The number of questions matters too. Real conversation isn't an interview. By our own count in the Switchboard corpus of recorded phone calls, only about 1 turn in 6 contains a question. So the Edgewise coach mostly comments, and keeps its questions genuine.

It also never tells the story for you. Classroom research warns teachers against filling in or smoothing over learner turns (Walsh, 2002). The coach can build, fix or wait, but it never supplies the next plot event. You tell the story.

For teachers, tutors and SLPs

Pick one or two target forms for prompts and recast everything else. Make one change per turn and stress it. Wait at least three seconds before rescuing a learner. With adults working one-on-one, don't assume prompts always beat recasts: watch what the learner does with each.

FAQ

What is the best way to respond when a language learner makes a mistake? Change one thing and make it easy to notice. Use a prompt (an echo or a sentence starter) for forms the learner has studied, and a stressed recast for forms that are new.

What's the difference between an expansion and a recast? An expansion keeps the learner's words and adds a missing piece. A recast says the sentence back with an error fixed. Second language research often groups both under "recasts."

Should a tutor correct every mistake? No. One change per turn is easier to notice, and recasts work best when they change one thing. Keep the conversation going and pick one or two targets.

How long should a coach wait before helping? At least three seconds. Teachers typically wait less than one second, and longer wait times change how both teacher and student talk.

Why don't grammar corrections stick? Learners often don't perceive grammar feedback as feedback. A fix needs one clear change, stress on the changed word and often a second chance to use it.

Is this speech therapy? No. These are learning strategies borrowed from speech therapy. Learning a language is not a disorder, and ASHA states that accents are not a communication disorder.

Your next step: Try a story retell in Edgewise. You tell the story out loud, and a voice coach talks with you about it. It builds, comments, fixes, slows down and waits, but you tell the story. Your first retell sets your level. Each round scores story structure and sentence building, then ends with your next piece: one specific thing to add next time.

Join early access

Related: Speech therapy techniques for language learners · What 40 years of language acquisition research says about learning to speak · Do language apps work? · Builds: expansion

Sources

1. American Speech-Language-Hearing Association. (n.d.). Accent modification [Practice Portal]. https://www.asha.org/practice-portal/professional-issues/accent-modification/
2. Ammar, A., & Spada, N. (2006). One size fits all? Recasts, prompts, and L2 learning. Studies in Second Language Acquisition, 28(4), 543–574. https://doi.org/10.1017/S0272263106060268
3. Beck, I. L., McKeown, M. G., & Kucan, L. (2013). Bringing words to life: Robust vocabulary instruction (2nd ed.). Guilford. https://www.guilford.com/excerpts/beck8_ch2.pdf
4. Brock, C. A. (1986). The effects of referential questions on ESL classroom discourse. TESOL Quarterly, 20(1), 47–59. https://doi.org/10.2307/3586388
5. Cleave, P. L., Becker, S. D., Curran, M. K., Owen Van Horne, A. J., & Fey, M. E. (2015). The efficacy of recasts in language intervention: A systematic review and meta-analysis. American Journal of Speech-Language Pathology, 24(2), 237–255. https://doi.org/10.1044/2015_AJSLP-14-0105
6. Girolametto, L., & Weitzman, E. (2002). Responsiveness of child care providers in interactions with toddlers and preschoolers. Language, Speech, and Hearing Services in Schools, 33(4), 268–281. https://doi.org/10.1044/0161-1461(2002/022)
7. Li, S. (2010). The effectiveness of corrective feedback in SLA: A meta-analysis. Language Learning, 60(2), 309–365. https://doi.org/10.1111/j.1467-9922.2010.00561.x
8. Loewen, S., & Philp, J. (2006). Recasts in the adult English L2 classroom: Characteristics, explicitness, and effectiveness. The Modern Language Journal, 90(4), 536–556. https://doi.org/10.1111/j.1540-4781.2006.00465.x
9. Lyster, R. (1998). Negotiation of form, recasts, and explicit correction in relation to error types and learner repair in immersion classrooms. Language Learning, 48(2), 183–218. https://doi.org/10.1111/1467-9922.00039
10. Lyster, R., & Ranta, L. (1997). Corrective feedback and learner uptake: Negotiation of form in communicative classrooms. Studies in Second Language Acquisition, 19(1), 37–66. https://doi.org/10.1017/S0272263197001034
11. Lyster, R., & Saito, K. (2010). Oral feedback in classroom SLA: A meta-analysis. Studies in Second Language Acquisition, 32(2), 265–302. https://doi.org/10.1017/S0272263109990520
12. Mackey, A., Gass, S., & McDonough, K. (2000). How do learners perceive interactional feedback? Studies in Second Language Acquisition, 22(4), 471–497. https://doi.org/10.1017/S0272263100004010
13. Nassaji, H. (2009). Effects of recasts and elicitations in dyadic interaction and the role of feedback explicitness. Language Learning, 59(2), 411–452. https://doi.org/10.1111/j.1467-9922.2009.00511.x
14. Pica, T. (1987). Second-language acquisition, social interaction, and the classroom. Applied Linguistics, 8(1), 3–21. https://doi.org/10.1093/applin/8.1.3
15. Pica, T., Young, R., & Doughty, C. (1987). The impact of interaction on comprehension. TESOL Quarterly, 21(4), 737–758. https://doi.org/10.2307/3586992
16. Rowe, M. B. (1986). Wait time: Slowing down may be a way of speeding up! Journal of Teacher Education, 37(1), 43–50. https://doi.org/10.1177/002248718603700110
17. Tamis-LeMonda, C. S., Kuchirko, Y., & Song, L. (2014). Why is infant language learning facilitated by parental responsiveness? Current Directions in Psychological Science, 23(2), 121–126. https://doi.org/10.1177/0963721414522813
18. Tobin, K. (1987). The role of wait time in higher cognitive level learning. Review of Educational Research, 57(1), 69–95. https://doi.org/10.3102/00346543057001069
19. van de Pol, J., Volman, M., & Beishuizen, J. (2010). Scaffolding in teacher–student interaction: A decade of research. Educational Psychology Review, 22(3), 271–296. https://doi.org/10.1007/s10648-010-9127-6
20. Walsh, S. (2002). Construction or obstruction: Teacher talk and learner involvement in the EFL classroom. Language Teaching Research, 6(1), 3–23. https://doi.org/10.1191/1362168802lr095oa

Switchboard note: the "about 1 turn in 6 contains a question" figure is Edgewise's own count from the Switchboard corpus of telephone conversations, not a published finding.