Proof trails should stay close to claims.
That is one way to keep the public layer honest.
If a note says something useful, the reader should not have to hunt forever for why it might be true.
Where did the thought come from?
What made it feel earned?
What older writing does it connect to?
What work is testing it?
What is still just a hunch?
The note does not need to prove everything inside itself.
That would make every note too heavy.
But the trail should be nearby.
Related notes.
Source dates.
Older writing.
Proof projects.
Open questions.
That proximity matters.
It keeps the language from floating away from the work.