What are language activities in Natural Decoding?
In the first three steps of the Natural Decoding Method, you absorbed language: through decoding, active listening, and passive listening. Step four reverses the direction — now you produce language yourself, in real or simulated contexts.
Natural Decoding always emphasized: Speaking and writing should follow from comprehension, not the other way around. No child is forced to speak before they understand. But once they have had enough input, they begin to communicate naturally — and that process is fluid, not forced.
Krashen's i+1 Principle
Stephen Krashen, one of the most influential language acquisition researchers, formulated the "Input+1" hypothesis: optimal language learning happens when the input is about 70–80% familiar and 20–30% new. DopaSpeak's activities are built exactly on this principle.
What activities does DopaSpeak generate?
DopaSpeak uses the local Gemma AI model to create personalized activities. These are always based on material you have already decoded and listened to — no unfamiliar vocabulary, no surprise tests.
Everyday dialogues and scenarios
The AI generates short role-plays and real-life scenarios:
- Conversation in a restaurant — placing an order
- Giving directions in your target language
- Shopping at the market
- Small talk about weather, hobbies, travel
- Making or cancelling a phone appointment
Content-based activities
Based on decoded songs and texts:
- Fill-in-the-blank: Insert missing words from songs
- Comprehension questions: Answer questions about decoded texts
- Retelling: Summarize a dialogue in your own words
- Word search: Find all occurrences of a specific word in your material
Grammar discovery (without a grammar book)
The AI highlights grammatical patterns you have already encountered in your material and asks: "What rule do you notice here?" You discover grammar inductively — from examples to rules, not the other way around.
AI and Personalization: A Learning Path Just for You
The greatest advantage of DopaSpeak over static language courses is dynamic personalization. The on-device AI knows:
- Which words you have seen and how often
- Which grammatical structures you are still uncertain about
- Which scenarios you haven't practiced yet
- Which content you have listened to the most
Based on this data — which stays exclusively on your device — the AI generates the optimal next activity. No teacher, no fixed curriculum — just your personal AI language companion.
Output vs. Input: When do I start speaking?
A common question: "When do I actually start speaking?" The Natural Decoding Method's answer:
When you want to, not when you're supposed to. Step four is not an obligation — it's an invitation. If you're ready to speak, speak. If you're not ready yet, keep decoding and listening. Both are valid.
In DopaSpeak you can complete all activities in writing instead of speaking — type your answers, and the AI provides feedback. No voice recording required.
How to get started with activities
-
Prepare your material. Decode at least 3–5 songs or texts completely and listen to them both actively and passively.
-
Open the Activities tab. DopaSpeak automatically shows you activities based on your material.
-
Choose a scenario. Pick an everyday scenario or a content-based activity — whatever you feel like.
-
Receive feedback. The AI gently corrects mistakes, explains errors, and suggests alternative phrasings.