The Case for More Speaking Practice (and Less Grammar Instruction) in ESL Classes

The challenge most ESL teachers face is a syllabuse dominated by grammar instruction at the expense of speaking practice. Grammar instruction dominates most ESL syllabuses. A typical general English coursebook devotes more time to teaching and practising grammar than to any other skill. Teachers who feel uncertain about a lesson default to grammar. Parents and institutional managers often feel that grammar is what "real" language learning looks like.

The research does not support this emphasis. Not because grammar doesn't matter - it does - but because the time allocation is badly calibrated, and the methods used to develop grammatical competence are often less effective than the time devoted to them suggests.

YapYapGo is a classroom speaking practice tool for ESL and EFL teachers built on the insight that communicative speaking practice is the highest-leverage use of classroom time for most language goals. Here's the research case.

What the research actually says about grammar instruction

1. Explicit grammar knowledge doesn't transfer directly to speaking

The interface debate in applied linguistics has gone on for decades, but the current consensus is clear: knowledge of grammar rules does not automatically produce correct grammatical use in speaking. A student who can state the rule for present perfect formation and complete a gap-fill exercise correctly will still make present perfect errors in free conversation.

This is because speaking requires implicit, automatic grammatical knowledge - the ability to select and produce correct forms in real time without conscious reflection. Explicit, declarative rule knowledge is useful for revision and correction but doesn't directly build the implicit knowledge that fluent speaking requires.

Research by DeKeyser (1997) and others suggests that implicit grammatical knowledge develops through meaningful, communicative use rather than through explicit instruction. Grammar instruction can help - particularly for adult learners - but its role is to raise awareness and provide a framework that communicative practice then automatises.

2. The time allocation is wrong

Consider a typical 45-minute lesson with the following time allocation:

• Grammar presentation: 15 minutes


• Controlled grammar practice (gap-fill, transformation): 10 minutes


• Speaking activity based on grammar: 10 minutes


• Admin and transitions: 10 minutes

Students speak for approximately 10 minutes. Of that, perhaps 6-7 minutes in pair work. Individual speaking time: 3-4 minutes.

A lesson with the same goals but restructured:

• Grammar focus (brief explanation + examples): 5 minutes


• Speaking activity using target grammar: 25 minutes


• Reactive grammar feedback on what emerged: 5 minutes


• Admin: 10 minutes

Students speak for 25 minutes. Individual speaking time: 12+ minutes. The grammar is practised in context rather than in isolation. The feedback is on real production rather than exercises.

The second approach produces more grammatical development as well as more fluency development, because the grammar is encountered in meaningful use.

3. Implicit learning through speaking is underused

Nation and colleagues have documented the conditions under which incidental grammatical acquisition occurs: through meaningful input, through meaningful output, through noticing forms in context. All of these conditions are better met by communicative speaking activities than by explicit grammar instruction.

When students are engaged in genuine discussion - forming arguments, responding to partners, expressing ideas they actually have - they encounter grammatical challenges in context. A student who needs to express a counterfactual to make their argument work is more receptive to instruction about conditionals than a student who is completing a gap-fill exercise.

4. The research on speaking-as-acquisition

Swain's Output Hypothesis (1985, 1993) argued that output - speaking and writing - drives acquisition in ways that input alone does not. When students produce language, they encounter gaps in their ability and develop strategies to fill them. This "pushed output" is a direct driver of grammatical development.

This means well-designed speaking activities develop grammar, not just speaking. Debate tasks that require hypothetical language develop conditional structures. Role plays that require professional register develop modal verbs and hedging language. Opinion tasks that require qualification develop concession structures.

What this means practically

Teach grammar, but spend less time on it

Grammar instruction has a role - particularly for adult learners with explicit learning styles, for raising awareness of forms students are ready to acquire, and for correcting fossilised errors. But 30% of lesson time on grammar, 10% on speaking is badly inverted. 10-15% on grammar focus, 60%+ on speaking is closer to what the research supports.

Use delayed, targeted grammar correction

Correct grammar in the debrief after speaking activities, not during them. Note patterns during observation, address them after the activity, and target instruction to what actually came up in real production. Reactive grammar teaching is more efficient than predictive grammar teaching.

Design speaking activities that push grammar

The best speaking activities are not grammatically neutral. A debate task that requires hypothetical reasoning pushes conditional forms. A narrative task requires past tense. A planning task requires modal verbs. Design speaking activities to contextually push the grammar you want students to develop.

Tool tip: YapYapGo provides structured speaking activities across six modes - all of which generate natural grammar use in context. The Debate mode specifically pushes students to use complex grammatical structures (conditionals, concession language, hedging) that explicit instruction rarely produces in spontaneous speech. A conversation topic generator provides quick discussion topics for teachers who want to supplement with their own content.

Be honest about what grammar tests measure

Grammar tests measure grammar test performance. They correlate somewhat with general language ability, but they don't measure what actually matters in communication: the ability to use grammatical knowledge fluently, in real time, for real communicative purposes. A student who scores 95% on a grammar test and speaks haltingly in English has not achieved the goal of language teaching.

The counter-argument (and why it's partly right)

The case for explicit grammar instruction is not baseless. Research by Spada and Tomita (2010) found that explicit grammar instruction, when combined with communicative practice, produces better outcomes than communicative practice alone. The issue is not that grammar instruction is worthless - it's that it's overweighted relative to what the research supports.

The optimal balance for most adult ESL learners is probably something like: 15-20% explicit language focus (including grammar, vocabulary, and discourse), 70-75% communicative production practice, 10% feedback and reflection. Not 50% grammar instruction and 20% speaking, which is what many coursebook-driven syllabuses produce.

For the research on communicative language teaching more broadly, see our post on communicative language teaching in 2026. For specific speaking activities that push grammatical development, see task-based language teaching for speaking. A classroom countdown timer and random student picker make the rebalancing practical - visible time in speaking blocks signals to students (and institutional observers) that speaking is the main work of the lesson.

---

Sources:

• Swain, M. (1985). Communicative Competence. Input in Second Language Acquisition. - The Output Hypothesis: speaking as a driver of grammatical development.
• DeKeyser, R. (1997). Beyond Explicit Rule Learning. Studies in Second Language Acquisition. - The interface between explicit and implicit grammatical knowledge.
• Spada, N. & Tomita, Y. (2010). Interactions Between Type of Instruction and Type of Language Feature. Language Learning. - Meta-analysis: explicit + communicative beats either alone.
• Krashen, S. (1982). Principles and Practice in Second Language Acquisition. - The comprehensible input hypothesis and its implications for grammar instruction.
• Nation, I.S.P. (2001). Learning Vocabulary in Another Language. Cambridge University Press. - Incidental learning through meaningful use.