Methods · Language & communication · Short guide
G1
Comprehensible input generation
Get reading/listening material pitched just beyond your current level, so input stays understandable enough to learn from, hard enough to grow.
When to use it
- You need reading or listening practice in a language you are learning, but textbooks feel too hard or too boring.
- You keep opening “authentic” articles and quitting after two paragraphs of noise.
- You finished a graded chapter and want one more text on a topic you actually care about, pitched just past comfort.
What to ask AI
Type something like: “Generate a ~200-word text on weekend markets at roughly A2. Prefer high-frequency vocabulary; allow about six stretch items. Do not include a gloss yet. After I retell, ask comprehension questions, then gloss only what blocked meaning.”
Name your level, topic, and length up front. Ask for no parallel translation until after you retell.
Keep the hard part on your side: you read, retell, and answer before you ask for vocabulary help.
How it helps
Language grows on input you can mostly understand, with a little stretch left in.
Too easy and nothing new sticks. Too hard and the page becomes noise you “cover” without learning.
Researchers call this pitching just beyond your current level. You do not need the theory name to use the idea.
AI can write short texts on topics you chose, at a level you named.
That only helps if you check comprehension afterwards, a retell or questions, instead of feeling exposed to language and calling it practice.
How to run it
- Declare your level, topic, and length in one sentence.
- Get a text or audio script with no side-by-side translation yet.
- Read or listen once. Cap how many unknown words you look up (five to eight is enough).
- Retell the content in your own words, or answer comprehension questions, before you binge the dictionary.
- Ask the AI to adjust the next text up or down based on how clean that retell was.
- Tomorrow, reuse the same content as a quick recall check before you ask for a new text.
Example exchange
Can you write me a short Spanish story about a rainy morning commute? I'm around A2, everyday words mostly, just a few stretch ones. Please don't give me a gloss yet.
Here's about 150 words. Read it once without translating. When you're ready, retell what happened. English or Spanish is fine.
Someone leaves late, misses the bus, waits under a shop awning, shares an umbrella with a stranger, and gets to work damp but on time.
Good skeleton. Who offered the umbrella first?
Copyable prompt
Generate a ~[N]-word text on [TOPIC] at roughly [LEVEL]. Prefer high-frequency
vocabulary; allow ~5-8 stretch items. Do not include a gloss yet. After I
retell, ask comprehension questions, then gloss only what blocked meaning.The Tell
Here is how you know this method has flipped: you have collected lots of “input” and never retold a single text.
Exposure without a comprehension check feels productive and leaves almost nothing you can use tomorrow.
Tighten the session so you attempt a retell before any gloss or translation arrives.
Principle evidence
Strength of the underlying learning idea, not a claim about AI products.
The underlying learning idea is rated moderate. Comprehensible input, material pitched just beyond what you can already handle, is influential in second-language teaching. It is contested as a complete theory of how languages are acquired, but there is broader agreement that slightly-above-level input supports growth better than material you cannot parse at all.
AI delivery evidence
Whether an AI tutor delivers this method well is a separate question.
There is little direct research on AI writing level-pitched texts for you. Whether the model actually hits your level, and whether the language sounds natural, is still speculative. The benefit rests on the input idea and on whether you still force a retell.