Quick answer: If you can read English news, write coherent emails, and pass written tests but freeze when you have to speak — you don't have a grammar problem. You have a production problem. The fix isn't another grammar course. It's mouth time: deliberate, sustained speaking practice where your mouth is forced to produce English faster than your brain can audit. SpeakShark was built specifically to deliver mouth time at scale.

I'm a non-native founder who spent two years over-studying grammar before realizing the bottleneck was elsewhere. This post is an argument for tilting your practice the other direction.

The over-grammar trap most learners fall into

The default English learning path looks like this:

Phase 1 (months 0-6): Learn alphabet, basic grammar, common vocabulary
Phase 2 (months 6-18): Drill grammar rules, expand vocabulary, study tenses
Phase 3 (months 18-36): Advanced grammar, idioms, more vocabulary
Phase 4 (somewhere around year 4): "Now you can start speaking"

This path is what most language apps, textbooks, and school curricula reinforce. It feels logical: master the rules, then apply them.

It produces a very specific outcome — learners who score B2 on written tests, understand 95% of English content they read, and cannot order coffee without rehearsing the sentence three times in their head.

I was that learner. After 18 months of Babbel-style "college-level grammar courses" (their description, not mine), I could explain the difference between past simple and present perfect in a written exam. The moment a real conversation forced me to use either, I froze.

The grammar knowledge wasn't the bottleneck. The mouth-to-brain pathway was.

What "mouth time" actually means

Mouth time is the cumulative seconds your mouth spends producing English under cognitive load — meaning you're not reading a script, not repeating after someone, not filling blanks in a textbook. You're constructing sentences in real time, in English, knowing they'll come out imperfect.

The key parameters:

Sustained. A 5-second utterance is not mouth time. Mouth time is measured in minutes-of-continuous-speaking. The first 30 seconds are easy. The struggle starts at minute 2 when your brain runs out of pre-formed phrases and you have to construct new ones in real time.

Pressure. Your brain has to commit to a sentence before fully auditing it. Real conversations don't let you pause for 8 seconds to mentally rehearse. Practice without time pressure doesn't train the same neural pathway.

Variety. Same topic, same prompts, same partner means your mouth shortcuts to the same memorized chunks. Variety forces fresh assembly every time.

Honest feedback at the phoneme level. Not "good job" — actual per-sound correction. Otherwise you reinforce wrong habits while feeling like you're improving.

If your current "speaking practice" doesn't satisfy three of these four parameters, it's not mouth time. It's something else — maybe useful, but not the thing that breaks the plateau.

Why grammar over-study quietly hurts your speaking

This is the counter-intuitive part. Too much grammar study can actively harm your speaking. Three mechanisms.

Mechanism 1: Grammar study cements the auditor

Every grammar drill trains the part of your brain that checks sentences against rules. Useful for writing — you can re-read what you wrote and audit it. Devastating for speaking — the auditor activates mid-sentence, says "wait, should that be past simple or present perfect?", and your mouth pauses.

You've trained the auditor to be vigilant. The auditor doesn't know it's only supposed to be active during writing. So it shows up during speaking too. Result: every sentence triggers an internal grammar-check that interrupts the production flow.

Heavy speakers — native or fluent non-natives — have a much weaker auditor during speech. They make grammar mistakes constantly (listen to any unedited conversation) and they don't notice because they aren't auditing. The fluency comes from low-audit production.

To get there, you need lots of practice producing without the auditor on. Grammar drills do the opposite.

Mechanism 2: Grammar is studied, speech is produced — different skills

Reading a grammar rule and understanding it is receptive. Doing a fill-in-the-blank exercise is cued production. Speaking is uncued production from scratch.

These are three different cognitive operations. Mastering the first two does not automatically transfer to the third. You can be a grammar genius receptively and a beginner productively. They're separate skills.

When learners "study grammar to improve speaking," they're working on the wrong skill. It's like reading books about swimming to improve your butterfly stroke. The information helps marginally; the time spent reading is time not spent in the water.

Mechanism 3: Grammar-heavy apps reinforce drill-as-practice

Most language apps front-load grammar because grammar progresses linearly and gamifies well. The product can show you a path (50 lessons to A2, 80 to B1) and you can complete it. Streaks work. Badges work.

The product can't gamify mouth time the same way. Sustained speaking is messy, variable, hard to score, hard to bound into a "lesson." So apps under-invest in it.

Result: you spend 80% of your "practice" time on grammar drills because that's what's available. After two years, your grammar is excellent and your speaking is still where it started.

The mouth-time alternative — what actually works

Here's the practice shift that breaks the plateau. Five activities ordered by mouth-time density.

Activity 1: Open AI conversation, 20+ minutes sessions

This is the highest-density mouth-time activity available to most learners. SpeakShark Daily Talk mode is designed for exactly this — open conversation with an AI teacher about whatever topic you pick, sustained as long as you want. No mini-exercises. No tap-to-record. Just talking.

In a typical 20-minute SpeakShark session you'll produce roughly 800-1500 words of English speech. Compare to a 20-minute general-app session where you might produce 80-150 words. 5-10x mouth time per minute spent.

The free tier — 3 sessions per day, every day, no card — exists so you can build the habit without paying.

Activity 2: Self-narration in English

Talk to yourself, out loud, in English, about your day, your work, what you ate, what you're thinking. Sounds silly. Works enormously. Twenty minutes of self-narration walking to lunch generates more mouth time than three days of typical app practice.

The downside: no feedback. You won't catch your own pronunciation errors. Use this as volume-builder, not as the only practice.

Activity 3: Voice memo journaling

Record yourself summarizing the day in English, 5 minutes daily. Listen back. You'll be horrified the first week. By week three you'll hear your own filler patterns, dropped consonants, weird intonation. Self-awareness alone fixes 30% of issues.

Bonus: you build a portable archive of your speaking that you can compare across months. Massively motivating when you hear month-one-you and month-three-you back to back.

Activity 4: Shadowing native podcasts/videos

Pick a podcast you like. Play 30 seconds. Pause. Repeat the 30 seconds aloud, matching pace and intonation as closely as possible. Continue.

This is the highest-leverage pronunciation activity in the toolkit. Not pure mouth time — there's a script — but you're forced to produce at native pace, which trains the mouth muscles. Combine with Activity 1 for full coverage.

Activity 5: Free conversation with humans

If accessible: language exchange partners, Cambly tutors, EngVarta — anywhere you talk to a human in English. Higher mouth-time density than apps but expensive at scale ($50-200/month for Cambly, $1-2/session for EngVarta).

Recommended weekly mix: Activity 1 daily (20+ min on SpeakShark), Activity 2 ad hoc, Activity 3 every 2-3 days, Activity 4 weekly. Activity 5 if budget allows for 1-2 sessions per week.

"But I still need grammar for writing/tests"

Yes. Two clarifications.

One: I'm not saying never study grammar. I'm saying don't expect grammar study to fix your speaking. They're different skills. Keep your grammar practice for writing and exam prep. Add separate, dedicated mouth time for speaking.

Two: Grammar matters most at A1-A2 levels and stops mattering much past B1. If you're already B2+ on grammar and your speaking is A2, more grammar will deliver diminishing returns. Switch the time allocation.

A reasonable mix for an intermediate learner: 80% mouth time, 20% targeted grammar/vocab for gaps you specifically notice during speech. Inverse of what most apps push.

What changed when I made the switch

Personal data point. In 2024 I spent 9 months on grammar-heavy apps (~30 min/day) and felt zero improvement in spoken fluency. In early 2025 I switched to a roughly inverse mix — 80% mouth time, 20% grammar — and within 8 weeks:

I could start a sentence without rehearsing it mentally
I stopped freezing in the first 30 seconds of meetings
My filler word density dropped from constant "um, uh, like" to occasional
Native speakers stopped asking me to repeat myself

The mouth-time tools I used: SpeakShark Daily Talk for daily sustained practice, voice memo journaling 3x/week, and shadowing the All-In podcast (don't judge, the cadence is gold).

Nothing in that practice involved studying a single grammar rule. The grammar I already knew receptively just needed more reps producing it before it could flow at speaking speed.

Diagnostic: how to know if this is you

If 3+ of these apply, you're in the over-grammar / under-mouth-time trap:

You can read English news without a dictionary
You can write a coherent paragraph in English in under 10 minutes
You score B1 or above on written placement tests
You freeze inside the first 30 seconds of real conversations
Your "speaking practice" consists mostly of app exercises under 1 minute each
You've been "studying English" for over a year with grammar-focused tools
You catch yourself mentally rehearsing sentences before saying them
You feel like your speaking has plateaued for 3+ months

The pattern is consistent. Receptive skills are good. Production is stuck. The fix isn't more receptive practice (grammar, reading, listening). It's more production practice (mouth time).

What this looks like for the next 30 days

Days 1-7. Stop adding new grammar. Whatever you currently know is enough to start. Add 15 minutes of daily mouth time via SpeakShark free tier (3 sessions a day, no card). Track minutes-of-speaking-output, not "lessons completed."

Days 8-14. Add voice memo journaling 3 days this week — 5 minutes recording yourself summarizing the day in English. Listen back the next morning. You'll cringe. It works.

Days 15-21. Add one shadowing session — pick a podcast you actually like (mine: All-In, Lex Fridman for engineering, Acquired for slower business pace). Shadow 10 minutes per session, 3 times this week.

Days 22-30. Hold the routine. By now you should notice: starting sentences faster, fewer freezes, less mental rehearsal time. If you're not noticing changes, increase mouth time. Anything below 60 minutes/week of sustained speaking output is too little to move the needle.

Bottom line

Grammar knowledge is not your bottleneck. Producing English at speaking speed is. The training that fixes the bottleneck is mouth time — sustained, varied, with honest feedback — not more rules.

The English learners I've watched break through plateaus all share one transition: they switched their daily metric from "lessons completed" to "minutes spent producing speech." The grammar that used to feel inaccessible during conversation suddenly came out, because the bottleneck wasn't knowledge — it was production reps.

The SpeakShark free tier exists for exactly this practice. 3 full conversational sessions a day, every day, no card required. You can test the mouth-time hypothesis on your own speaking inside one week. If it doesn't change anything, you've lost nothing. If it does, you've found what you've been missing.

I'm biased, but the science is clear: speaking improves through speaking, not through studying. Add mouth time. Hold the grammar steady. Watch the wall move.