What makes a diagnosis honest, rather than another conclusion?
A diagnosis is honest when it can lose. It states one reason the answer leaned the way it did, the evidence for that reason, the evidence against it, and what would count as proof that it was wrong.
Anything that cannot lose — a summary, a hunch, a generated list of suggestions — may still be useful, but it cannot be checked, and this loop only ships what can be checked.
The shape of one
- One reason, in one sentence, tied to a named source.
- What supports it, quoted from what was actually observed.
- What cuts against it, stated with the same care.
- The falsifier: the outcome that would show the reason was wrong.
Why falsifiability is the point
The falsifier is what turns a result into learning. When the change ships and the answer moves, the diagnosis survives; when nothing moves, the diagnosis takes the hit — it is not reworded until it looks right.
Written this way, being wrong is cheap and informative. The steps around it are on the product page.
What this answer does not claim
- A diagnosis is a reading of what sources say. It is never a claim about how any model weighs them.
- Even a diagnosis that survives one pass is supported, not proven. Support and proof are different words on purpose.