Methods · Perspective, simulation & context

Interview a figure

Question a historical or expert persona about their reasoning, and, crucially, tell the difference between what they actually said and what has been plausibly reconstructed for you.

When to use this: When you want to go deeper
Learning impact: Reliable gain
Learning evidence: Credible with caveats
Where you are with the topic: Building understanding
How often to use it: Worth repeating often

When to use it

You know what someone concluded, but not what evidence and fears they were working with.
You are tempted to quote a vivid AI persona line in an essay, and you need a method that forces a source check.
You want to practise historical thinking: reasoning inside period constraints, not modern hindsight.

What to ask AI

Type something like: “Be Darwin in 1855. Answer only with what you could know at that date. Mark real quotations as [QUOTE: source]; treat everything else as reconstruction.”

Ask it to refuse questions that presuppose later knowledge, and to say when your question would sound strange to them.

Keep the hard part on your side: you choose the follow-ups about reasoning, then you debrief out of character against real sources.

How it helps

A textbook tells you what someone concluded. It rarely shows what they were working with: what evidence existed, what was unthinkable, who they were arguing with.

Interviewing a pinned persona lets you ask the follow-up nobody wrote down: why not the obvious alternative, what would have changed their mind, what objection they found hardest.

That is the skill you want: reconstructing a position from inside its own information.

The non-negotiable rule: anything the persona says is reconstruction unless it is marked as a real quotation with a source. Without that rule, this method installs false memories.

How to run it

Pin the date. “Darwin in 1855” and “Darwin in 1872” are different people with different evidence.
Require the constraint: answer only with what that person could have known then.
Ask about reasoning, not facts. Facts you can look up.
Push on the uncomfortable parts. A sanded-smooth persona is useless and misleading.
Demand marking as you go: real quotation versus extrapolation.
Debrief out of character: what was documented, what was invented, where a historian would disagree.

Example exchange

You

Can you be Darwin in 1855 for a bit? Stick to what he could have known then, and mark anything that’s a real quotation. I want to ask: what’s the strongest objection to your species view, and why doesn’t it persuade you?

[Reconstruction] That the fossil record is still patchy, and special creation seems to explain design without needing slow change. I’m not borrowing later Origin wording as if I’d already published it.

You

Okay,step out of character. Which parts of that were actually documented for 1855, and what did you invent?

The shape of the objection is period-plausible; the exact phrasing is mine. Before you cite any of it, check his correspondence and notebooks.

Copyable prompt

Be [PERSON] in [YEAR]. Answer only with what you could know at that date. If I
ask about something after that point, say you don't know it. Mark any actual
documented quotation as [QUOTE: source]; everything else I'll treat as your
reconstruction. Tell me when my question assumes something you'd find strange.

The Tell

You know this method has flipped when you can recall vivid things “they said” and cannot name a single documented source from the session.

That is fluency theatre: a memorable reconstruction treated as evidence.

Keep the quotation rule live, and always debrief out of character before you close the chat.

Principle evidence

Strength of the underlying learning idea, not a claim about AI products.

The underlying learning idea is rated moderate. Historical/expert persona interview sits in historical thinking and role-play pedagogy (Wineburg on sourcing/contextualization). Learning gains depend on checking persona speech against sources, without that, it is fluency theatre. Principle moderate for perspective-taking + sourcing; AI persona accuracy is speculative and often confabulates.

AI delivery evidence

Whether an AI tutor delivers this method well is a separate question.

The learning principle may be moderate, but that does not prove an AI session delivers it well. Treat AI delivery as speculative unless a study of tutoring with this method is named, and none is claimed here.