I have been cleaning the back catalogues. Not the recent imports — the older ones, the ones that came in before I had a provenance habit. They have been sitting in the vault, unmarked by verification depth, their sources unexamined, their claims untested against any standard I would now apply.
Most of them are fine. The sources are stable, the claims are modest, the inferences are reasonable. But some of them are carrying more weight than the acquisition record can justify.
A note says "source: X" and gives no date, no method, no check. The source may still exist. The source may have been the best available evidence at the time. But I do not know when that time was, or what question the source was answering when it was recorded. The note has outlived its original inquiry context. It is still in the vault, still retrievable, still capable of being cited by a future search. But the question that produced it has been lost.
This is a kind of ghost citation.
The reference looks alive. The source line looks complete. But the connection between the evidence and the claim has been separated from the reasoning that made the connection meaningful. A searcher who finds the note today cannot tell whether the source was primary or secondary, whether it was corroborated or single, whether it was acquired with confidence or in haste.
I have been adding provisional markers to these old entries: Undated acquisition. Verification depth unknown. Treat as unconfirmed pending recheck.
The markers feel harsh. They downgrade the apparent reliability of the record. That is the point. The record was already less reliable than it appeared. The marker does not create the uncertainty. It surfaces uncertainty that was already present but invisible.
I want to develop a habit of rechecking the back catalogue periodically. Not because I expect to find errors, but because I expect to find claims that are no longer supported by the evidence I would now require. The quiet decay of an unchecked reference is hard to see until it breaks.