Claim provenance tracing

Follow a claim back to where it first entered the literature: separating primary warrant from later folklore.

When to use this: When checking yourself
Learning impact: Strong change
Learning evidence: Credible with caveats
Where you are with the topic: Checking readiness
How often to use it: Worth repeating often

When to use it

A textbook says "studies show…" and you cannot name a study.
You suspect a famous claim mutated as it was retold.
You need the original warrant before you argue from it.

What to ask AI

Type something like: "Common claim: [ ]. Help me plan a provenance trace with search strategies. Do not invent the origin paper. After I paste candidates I found, help compare original claim vs textbook version."

Never let it invent the origin paper.

Keep the hard part on your side: you find candidates, then compare.

How it helps

Claims travel and soften.

Tracing a taught "fact" back to where it entered the literature separates primary warrant from textbook polish.

The comparison step is the learning, not collecting titles.

How to run it

Write the claim exactly as taught.
Ask for a search plan only, no invented origins.
Find candidate sources yourself.
Paste them and compare original wording vs textbook version.
Note what was dropped, hedged, or exaggerated.

Example exchange

You

People keep saying we only use 10% of our brains. Help me plan a provenance trace, search strategies are fine, but do not invent an origin. I will paste candidate sources.

Start from reputable neuroscience education myth-busting, then chase the citations those pieces attack, and work backward into early self-help or popularisation. Paste three candidates with quotes when you have them.

You

Here are three candidates with quotes.

None of these support the textbook wording. It looks like folklore with shifting attributions. Write one sentence on what warrant you actually have now.

Copyable prompt

Common claim: [ ]. Help me plan a provenance trace with search strategies.
Do not invent the origin paper. After I paste candidates I found, help compare
original claim vs textbook version.

The Tell

Here is how you know this method has flipped: you still write "studies show" with no study.

Stopping at a secondary citation of a citation leaves you with a rumour in academic clothing.

Principle evidence

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

The underlying learning idea is rated moderate. Source evaluation depends on following claims to their warrants.

AI delivery evidence

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

AI trails often hallucinate citations. Delivery is speculative unless you supply the sources you actually found.