Summary¶
inferred The pipeline below applies this bundle's claim-status policy as an editorial gate rather than only a research convention. Its purpose is that every published claim can be defended on the record, which is the one differentiator the competition cannot copy cheaply.
The gates¶
| Gate | Question | Fail action |
|---|---|---|
| 1. Tip intake | Is there a document, a named person, or an observable event? | Hold; do not assign |
| 2. Sourcing | Can the claim be attributed to a named source or a retrievable record? | Downgrade to inferred and label it in the copy |
| 3. Status labelling | Is this documented, inferred, or contested? |
State it in the draft, not just in the notes |
| 4. Interest check | Does the source have an interest in this being true? | Name the interest in the copy |
| 5. Right of reply | Has the subject had a fair chance to respond, with time? | Do not publish |
| 6. Corrections | Is there a published, findable corrections policy? | Publish the policy before the story |
| 7. Post-publication | Are follow-ups and errors logged in the open? | Log them; the log is the asset |
The specific hazard¶
documented Kingston documents organised online harassment of journalists and academics in Japan, which means a published story can be met with a coordinated response aimed at the reporter rather than the reporting.2 inferred A pipeline that pre-commits to publishing its own errors is harder to attack on accuracy grounds than one that does not — the correction log becomes a defensive asset rather than a liability.