Borrowed Idea Attribution Scan¶

Provenance audit — instantiates Use-Time Source Attribution Calibration

Sweeps a shared store of notes and ideas for material that arrived from someone else but now feels self-generated, and routes each item back to the source that deserves the credit.

Ideas you read months ago sink into the same notebook as ideas you had yourself, and once the source tag fades they all feel equally yours. Borrowed Idea Attribution Scan works one direction of that confusion: it sweeps a commingled store looking for material that is actually externally sourced but now feels original — the cryptomnesia case — and opens a path to credit it. Its defining move is that it checks origin, not truth: it never asks whether an idea is correct, only whether it came from you, and if not, whose it was. That is what separates it from a fact-check and from its provenance-tracing siblings, which reconstruct a full trail rather than scan a store for un-credited borrowings.

Example¶

A researcher is drafting a review article and reaches for what feels like a crisp original framing — "regulatory capture as an information problem." Before it goes in as their own, the scan runs across their reading log and annotation store and surfaces a near-match: a paper they marked up eight months earlier used that exact framing. The item wasn't invented at the desk; it was absorbed and forgotten. The scan routes it to a credit path — a citation and a line of acknowledgement — so the framing keeps its place in the argument but now carries its origin. A quiet near-miss with unintentional plagiarism becomes a normal citation, caught because the sweep compared feels-like-mine material against a record of what had actually been read.^[1]

How it works¶

The scan is tuned to one asymmetry: it hunts for items whose felt originality outruns their logged provenance. It enumerates the shared store, and for each candidate compares it against an index of external material the author has consumed — reading logs, saved sources, prior correspondence — flagging resemblances that the item's own metadata doesn't account for. A hit is not a verdict of theft; it is a prompt to check memory honestly and, where the borrowing is real and substantive, to open a credit path. Because it matches on ideas and phrasing rather than claims, it catches paraphrase and absorbed structure, not just copied strings.

Tuning parameters¶

Match threshold — how close a resemblance counts as borrowed. Loosen it to catch paraphrase and half-remembered structure; tighten it to avoid flagging common knowledge and genuinely convergent ideas.
Scan scope — the whole store versus only passages headed for publication. Wider scope catches more latent borrowings but spends attention on material that may never ship.
Credit bar — how substantive a borrowing must be before it earns a citation. Set low, everything gets attributed and the credit becomes noise; set high, load-bearing debts slip through uncredited.
Index freshness — how complete and recent the external-source log is. A stale index silently misses everything read since it was last updated.

When it helps, and when it misleads¶

Its strength is catching the borrowing that a truth-check structurally cannot see: a verification pass asks whether a claim is right, never whose it was, so an absorbed idea sails through fluent and un-credited. The scan restores that missing question at the moment the idea is used.

It misleads in two directions. It throws false positives on convergent or common ideas that no one owns, and it is blind to any source that was never logged — an idea from a hallway conversation leaves no trace to match against. Its classic misuse is to run it backwards: a scan performed after the fact to manufacture a tidy paper trail of originality, or to over-cite defensively so that nothing can be called unattributed. The discipline that keeps it honest is to treat a hit as a prompt for an honest memory check rather than a ruling, and to keep the external-source index real and current rather than trusting the sweep to know what it was never shown.

How it implements the components¶

commingled_item_inventory — it enumerates the mixed note-and-idea store and marks which items are candidates for external origin, turning an undifferentiated pile into a scannable inventory.
borrowed_material_credit_path — for each confirmed borrowing it opens the route to proper credit: a citation, an acknowledgement, an attribution line.

It does not reconstruct an item's full handoff trail — that's Chain-of-Custody or Lineage Check — nor score how confident the attribution is (Source Attribution Confidence Rubric), nor keep source and content separated through summarising (Source-Label Preserving Summary Template).

Instantiates: Use-Time Source Attribution Calibration — it supplies the credit-the-borrowed-source half of the appraisal.
Sibling mechanisms: Chain-of-Custody or Lineage Check · Provenance Lookup Before Publication · Generated Content Disclosure Gate · Hallucination Intrusion Triage · Memory Source Probe · Observation Recheck or Replication · Reality Monitoring Checklist · Source Attribution Confidence Rubric · Source Attribution Training Set · Source Confusion Matrix Review · Source-Label Preserving Summary Template

Notes¶

The scan credits origin, and says nothing about whether the borrowed idea is any good — a correctly attributed falsehood is still a falsehood. Keeping attribution separate from verification is deliberate: it lets a team fix a credit gap without re-opening whether the idea was right, and vice versa.

References¶

[1] Cryptomnesia — recalling an idea absorbed from someone else as though it were newly one's own. It is the mechanism behind much unintentional plagiarism, and it is exactly the failure this scan is built to intercept. ↩