Skip to content

Correction and Retraction Notice

Public correction notice — instantiates Transparency for Accountability

Publicly links a corrected or withdrawn record to what was wrong, what changed, and who is affected — without erasing the original from the trail.

A Correction and Retraction Notice is the visible statement that a previously released record was wrong and has been fixed or withdrawn — naming the error, the correction, its scope, and its date, while keeping the original discoverable rather than silently rewritten. Its defining move is closing the loop in the open: an error that was public gets a correction that is equally public and permanently tied to the thing it fixes, so anyone who saw the mistake can find the fix and anyone auditing later can see that a fix occurred. It does not adjudicate blame or impose a consequence; it makes the fact and shape of the correction part of the record.

Example

A newspaper reports that a city official "was charged with embezzlement." Two days later the facts turn out narrower: the official was questioned, not charged. The paper issues a correction notice — a dated, labeled note appended to the original article and listed on a standing corrections page: what the article said, what was wrong, what is now accurate, and that the online version has been updated. Crucially, the original wording is not scrubbed as if it never ran; the correction sits with it, so a reader who shared the false version can trace the fix, and a later reviewer can see both the error and the response. The notice does not punish the reporter or compensate the official — separate matters — but it stops the false claim from circulating as uncorrected fact and puts the paper's willingness to answer for its record on display.

How it works

  • Link the fix to the flaw. The notice points from the corrected item to a plain statement of what was wrong and what changed — not a silent edit.
  • Preserve the original. The superseded version stays discoverable (marked as corrected), so the history is intact and the change is auditable.
  • Scope the error. It says how far the mistake reached — one figure, a whole conclusion, the entire record — so readers can judge what to re-examine.
  • Date and propagate. It is timestamped and pushed to where the error travelled — a feed, a downstream republisher — so the correction chases the mistake rather than waiting to be found.

Tuning parameters

  • Prominence — a buried footnote versus a notice as visible as the original error; low prominence protects reputation but lets the falsehood outrun the fix.
  • Correction vs. retraction threshold — when a flaw is small enough to amend in place versus large enough to withdraw the whole record; drawn too high, it leaves discredited records standing.
  • Original-preservation policy — strike-through and annotate versus replace; preservation supports audit, replacement looks cleaner but hides the trail.
  • Propagation reach — notify only your own page versus every downstream that carried the item; wider reach corrects the record but is costly.
  • Latency — how fast a confirmed error is corrected; every day of delay is more circulation of the wrong version.

When it helps, and when it misleads

Its strength is that it keeps a public record trustworthy over time: errors are inevitable, and a visible, linked correction is what separates an honest record that fixes itself from one that quietly launders its mistakes. It also builds standing — a body that corrects openly is more credible than one that never admits error.

Its failure modes cluster around asymmetry. A correction issued far less prominently than the original error technically closes the loop while ensuring almost no one who saw the mistake sees the fix. Silent editing — changing the record with no notice — is worse: it erases the error and the accountability with it. And a notice can be run backwards, worded to minimize ("clarification") what was in fact a substantive error. The discipline that keeps it honest is the version-of-record norm: the corrected item becomes canonical, but the original remains linked and labelled, so the correction adds to the history instead of overwriting it.[1]

How it implements the components

This mechanism fills the correction-disclosure slice — making a fix visible and traceable, not administering its consequences:

  • remedy_and_correction_link — its core: a visible pointer from the flawed record to the correction, its reason, and its scope.
  • provenance_and_version_link — it supersedes the prior version while keeping the original discoverable, so the record's history stays intact.
  • timeliness_and_update_cadence — it is issued promptly and dated, putting the gap between error and fix on the record too.

It publishes the correction; it does not run the channel through which an error is first challenged (that's Public Comment and Question Channel) nor assemble the evidence that establishes the error (that's Evidence Disclosure Packet); and it stops short of the consequence that answerability may require, which lives beyond disclosure.

Notes

A correction notice is the counterpart to disclosure, not a substitute for a remedy. It makes the record right and the fix visible; whether anyone is answerable for the original error — a consequence, a process change — is a separate step this mechanism only enables. Its own worst failure is the unpublished correction: a fix applied so quietly that the record improves while accountability does not.

References

[1] The version of record is the canonical, citable version of a published work; a correction or retraction updates it while keeping prior versions linked and labelled. Scholarly and journalistic correction norms rest on this: the record is improved openly, not silently replaced, so the trail of what was claimed, and when, remains auditable. ↩