The fear of speaking English is the number 1 block for Brazilians. And it’s not a lack of knowledge — it’s anxiety (what Krashen called the affective filter). Here’s the 5-step method to unlock: accept that you will make mistakes, start speaking alone, move to conversation with AI (zero judgment), then language exchange with humans, and gradually increase complexity. In 4-6 weeks of daily practice, you will feel the difference.
Look, I’m going to say something that I repeat every week to some student: speaking is the fear of 10 out of 10 Brazilians learning English. It gives you a little chill just to think about it, right? You start wondering, "What if I make a mistake?", "What if the person laughs?", "What if I freeze in the middle?" And all this anxiety makes you freeze BEFORE you even start speaking.
The good news is that fear of speaking has a solution. And it doesn’t involve "forcing yourself" or "losing your shame by force." It involves method — a gradual exposure process that reduces fear without drama. In this post, I’ll give you the 5 steps that work, based on what applied linguistics has discovered over the last 40 years and on Lanna (a Brazilian language learning platform with AI), which was built on this method.
Why Fear is So Strong
The fear of speaking English has 3 reinforcing factors:
- Punitive-based education: in school, we learned that making mistakes is shameful. English tests were marked in red, and teachers pointed out mistakes. You internalized that speaking incorrectly is a character flaw. It’s ridiculous but it’s real.
- Lack of active practice: traditional classes are 90% passive — you listen, read, and do exercises. You almost never speak. When it’s time to open your mouth, you have no practical repertoire, just theory.
- Affective filter (Krashen): linguist Stephen Krashen proposed in 1985 that when you’re anxious, the brain literally blocks language acquisition. It’s biology — there’s no point in "trying to force it." You need to lower anxiety first. There’s a complete post about Krashen's theory.
All three reinforce each other: you are afraid → you don’t practice → you don’t evolve → you have more fear. It’s a vicious cycle. To break it, you need to attack the fear directly.
Step Zero: Accept that You Will Make Mistakes
Before the 5 steps, there’s a step zero: accept that making mistakes is part of the process. I always say in videos — the goal of language is to understand and be understood, not to speak perfectly. I’ve lived in São Paulo for 31 years and I make mistakes in Portuguese. If I make mistakes in my own native language, imagine in a second language that I learned as an adult.
When you let go of that pressure of perfection, half of the fear disappears. It’s not a motivational trick — it’s the recognition that the goal you were pursuing (speaking without errors) is impossible even for a native, so pursuing it doesn’t make sense.
The 5 Steps of the Method
Step 1 — Speak Alone, 5 Minutes a Day
It may sound crazy, but it’s the technique most used by polyglots. You speak alone in English about anything — what you ate, what you’re going to do, what happened at work. 5 minutes a day. In the shower, in the car, while cooking, washing dishes. A moment when no one is listening.
Why it works: you train your mouth to produce sounds, your brain to form sentences, your ear to recognize your own voice speaking English. All without social anxiety because there’s no one to judge you. It’s where you move from zero without suffering. I wrote a whole post about speaking alone.
Step 2 — Conversation with AI by Voice, 15 Minutes a Day
After 1 week of speaking alone, move to AI. AI is the perfect conversation partner for those who are afraid: it doesn’t judge, it’s not in a hurry, it doesn’t make a face if you freeze in the middle of a sentence. And it responds instantly, so you train listening too.
Lanna has a real-time voice speaking mode — you speak, the AI understands, responds with natural voice, and the cycle continues. No record button, no waiting 10 seconds, no anxiety. Use it every day for 15 minutes. There’s a complete guide on practicing conversation with AI.
Step 3 — Shadowing with Native Audio
Shadowing is mimicking exactly what the other person is saying. You listen to a phrase in English and try to repeat it exactly — pronunciation, rhythm, intonation. It’s the classic technique of polyglots to gain naturalness in speaking. Do it at a comfortable pace, not at the native speed — then speed up.
Maximum 5 minutes per session. More than that gets tiring and becomes automatic without learning. I explain shadowing in detail in this post.
Step 4 — Language Exchange with Natives Online
After 4-6 weeks of steps 1-3, you’re ready for real conversation with a human. This is where language exchange comes in — apps like italki, Tandem, HelloTalk where you find a native wanting to learn Portuguese and you swap (30 min in English + 30 min in Portuguese).
Your first conversation will be uncomfortable — that’s normal. Start with an easy topic (your routine, your job, the city you live in). Apologize once for freezing, then let it flow. In 3 or 4 conversations, you’ll be speaking well. There’s a complete guide on language exchange with natives.
Step 5 — Gradually Increase Complexity
As you unlock, you will increase the difficulty:
- 15 min conversation on an easy topic → 30 min on a varied topic → 1 hour on anything
- Known topic (your routine) → unknown topic (politics, science, cinema)
- 1 fixed interlocutor → various different interlocutors
- Prepared conversation → spontaneous conversation
In 3 to 6 months of this schedule, you will move from "blocked" to "normal conversation." Not fluent yet — but totally functional for real use.
Mistakes That Keep the Fear
- Wanting to speak perfectly. Half of the fear comes from wanting the impossible. Let go of that pressure.
- Only consuming, never producing. Those who only listen to podcasts never unlock. Active production is mandatory.
- Waiting to be "ready to speak." You will never be ready. Start even if you freeze.
- Comparing with fluent natives. Compare with yourself from a month ago, not with someone who has been speaking for 20 years.
- Skipping the AI step and going straight to human. It works for some, but most get more traumatized and give up. There’s the bridge of AI.
Why This Method Works Scientifically
The 5-step method isn’t my invention — it’s practical application of 3 concepts validated in applied linguistics:
- Affective filter hypothesis (Krashen, 1985): anxiety blocks acquisition. Lowering anxiety = freeing acquisition. Steps 1-3 are exactly that.
- Output hypothesis (Swain, 1985): you need to produce, not just consume. Speaking alone counts as output. Conversing with AI counts as output. Just listening is not enough.
- Noticing hypothesis (Schmidt, 1990): you only learn when you consciously notice a mistake. When you freeze on a sentence and rephrase, you are noticing.
Combined, these three explain why progressive practice in a safe environment is superior to "forced immersion." Going straight to conversation with a native without preparation keeps the affective filter high — you learn little and traumatize more.
Frequently Asked Questions
Why do Brazilians have so much fear?
An education that punishes mistakes + lack of active practice + high affective filter. The three form a vicious cycle.
First step?
Accept that you will make mistakes. The goal is to be understood, not to be perfect.
Does speaking alone help?
Yes, and it’s the best first step. Every polyglot does it. No judgment, pure practice.
How long to unlock?
4 to 6 weeks with daily practice of 15 min. It’s not fluency — it’s the ability to converse without freezing.
Does AI help or is it a crutch?
It helps a lot if you transition to human afterward. Staying only with AI becomes a crutch.
What if I freeze in the middle?
Normal. Breathe, restart more slowly. Freezing is information, not failure.
Let’s Start Today
Today, right now, do step 1: speak in English alone for 5 minutes. In the shower, in the car, while cooking. About anything. With no one to judge. It’s the easiest and most important start — because it breaks the inertia.
Tomorrow, continue. After 1 week, move to AI. The Lanna has a free voice speaking mode in the free plan (3 conversations per month) — you can test the transition without paying anything. In 4-6 weeks, you will look back and not recognize the blocked student you were before.
