G8 Guru / Judging Someone You Cannot Check

Test them on the ground you already stand on

Whenever a claim touches something you know, treat it as a free calibration sample.

The hardest case is the ordinary one: someone makes a claim in an area where you have no ability to evaluate the substance. Direct verification is off the table. What remains is indirect evidence, and the most useful piece is their performance on the small patch of ground where you can check. Everybody has one. It may be your own trade, a place you have lived, a process you have been through, a topic you happen to have read properly. When a confident general speaker touches that patch, pay very close attention.

What you are looking for is not whether they get every detail right. Nobody does, and a minor error is not damning. You are looking for whether they know the difference between what they know and what they are improvising. Somebody who ventures onto your ground and gets it roughly right, with appropriate caution and a visible sense of where their footing ends, is showing you how they behave everywhere. Somebody who is fluently, confidently wrong about the thing you happen to know is also showing you how they behave everywhere, and the fluency was no protection.

This is one of the few genuinely reliable tests available to a non-specialist, and it is nearly free. It works because the trait you are measuring, epistemic care, is a stable habit rather than a subject-specific fact. Collect these samples deliberately. Notice when someone passes and when they fail. Over time you will build a small private record that is worth more than any credential list, because you produced it yourself from evidence you could actually assess.