Methods · Support & cognitive load
A4
Multi-level explanation
Ask for the same idea at three levels of sophistication and use the level where you break as a precise measurement of where your understanding stops.
When to use it
- You are meeting this idea for the first time and need a way in.
- The reading feels dense and you want a first map before the details.
What to ask AI
Type something like: “Explain … three times: to a bright 12-year-old, to a first-year student”.
Ask it to find gaps in what you wrote, without rewriting the whole explanation for you.
Tell it your current level and ask it to stay one step harder than comfortable.
Keep the hard part on your side: you retrieve, attempt, explain, or decide before you ask for help.
How it helps
When something is too hard, the usual advice is to find a simpler explanation.
That's half right and it wastes the interesting information.
Ask for an idea explained at three levels, to a twelve-year-old, to an undergraduate, to a specialist, and read all three.
What matters is not that the simple one is comfortable.
It's where, precisely, you stop following.
That boundary is a measurement, and it's far more useful than the vague feeling of "I don't get this." Because the three versions differ in specific ways, the point of failure names your gap: - Lost at the simple version → you're missing a prerequisite, not this topic.
Fine at simple, lost at intermediate → usually the formalism.
The idea is there; the notation or the mechanism isn't.
How to run it
- Ask for three levels at once, not sequentially. You need them side by side to compare.
- Next: specify the levels concretely: a bright twelve-year-old, a first-year in the subject, someone doing a PhD in it. Vague levels produce three versions of the same paragraph.
- Read all three even if the first is enough. The measurement is the point.
- Name where you stopped. Be exact, which sentence, which term.
- Ask what the simpler version omitted: and whether the omission makes it false or merely incomplete. This distinction matters enormously.
- Work upward from your break point, not from the bottom.
Example exchange
Explain why we use the chain rule at three levels: to a bright twelve-year-old, to a first-year student, and to someone doing a PhD. Make the content actually different, not just the wording.
Here's the twelve-year-old version only: when something depends on another thing that is itself changing, you multiply the rates. Does that feel clear enough to move up a level?
Clear enough. Give me the first-year version.
If y equals f of g of x, then dy/dx equals f prime of g of x, times g prime of x. Which piece of that still feels fuzzy?
Copyable prompt
Explain [CONCEPT] three times: to a bright 12-year-old, to a first-year student
in the subject, and to someone doing a PhD in it. Make them genuinely different
in content, not just in vocabulary.The Tell
Here is how you know this method has flipped: you follow all three and learn nothing. Either the topic is secure, or the levels weren't genuinely different.
In that moment the AI (or the schedule, or the story) did the thinking, and you only recognised a finished product.
Tighten the prompt for multi-level explanation so you attempt, decide, or retrieve before anything is handed to you.
Principle evidence
Strength of the underlying learning idea, not a claim about AI products.
The underlying learning idea is rated moderate. Multiple explanation levels connect to register / audience adaptation and to diagnostic probing of where comprehension breaks, adjacent to formative assessment and scaffolding diagnosis. Less a single named controlled trials tradition than a practical synthesis of explanatory depth matching.
AI delivery evidence
Whether an AI tutor delivers this method well is a separate question.
Treat principle support as moderate; AI multi-level generation quality is speculative.