Public statement corrections rumors: four positions, four responses

A request arrives asking you to correct something. You check, and you never published it. Or you published only that a claim was circulating, which is a different sentence from the claim itself, and the request treats them as the same thing.

This is a common situation and it has a bad default: say nothing, because there is nothing to correct. That answer is usually wrong, and the better answers all involve deciding what you are correcting and where. For a relationship timeline there are four positions your page can be in, each with its own response, and two well-known celebrity cases show how they play out.

What to take away

What you actually published

Establish this before anything else, because the request will not have made the distinction.

The third row generates most of these requests and most of the bad outcomes. A page saying a claim is circulating has made a true statement about circulation, but the requester experiences it as the site carrying the claim.

Both readings are reasonable, a sign the sentence probably should not have run. The reasoning on when circulation is worth reporting is in notes on repeating a claim.

Denial amplifies, and the effect is measurable in your own logs

A correction is a page. It has a heading, it names a subject, and it is indexed. If the original claim was thinly circulated, your denial may be the most visible thing about it.

That is no reason never to deny. It is a reason to choose where the denial lives and how it is framed.

A note on the page that carried the material reaches the people who saw it. A standalone article reaches everyone, including readers meeting the claim for the first time, in your framing.

The longest-running example in celebrity coverage shows the cost. In April 2005, one month after Jennifer Aniston filed for divorce from Brad Pitt, paparazzi photographs of Pitt and Angelina Jolie emerged, and the press read them as evidence of a relationship. According to Angelina Jolie's published biography, she was accused of causing the divorce of Pitt and Aniston, who had been living apart since January 2005. Jolie and Pitt later stated that they fell in love on the set and that there was no infidelity. Every repetition of that denial since has carried the accusation with it, which is the trade-off the rules below manage.

Two rules follow, and they are close to mechanical.

Where a correction has to be made at all, its anatomy and its placement are set out in the guide to corrections.

The request that is really a scope complaint

Many of these requests are not about accuracy. They are about whether the subject should be written about at all, and they arrive in the language of correction because that is the language that gets a response.

Answer the accuracy question first, honestly, separately.

Then answer the scope question on its own terms: whether the entry meets your published criteria, whether the public role test is met for each person named, whether it would be published today.

Those are legitimate questions with real answers, and treating them as a correction request produces a reply that satisfies nobody.

Whether a person belongs in an entry rests on what they did to enter public life, not on who they are connected to. This follows the terms set out in the treatment of a public figure.

Correcting the record about a claim with no origin

Sometimes the claim being complained about has no author. It grew out of an inference somebody made in public and hardened through repetition, and there is nothing at the bottom of it.

That is worth establishing before responding, because it changes what a truthful answer looks like. You cannot say the claim was retracted, since nobody made it, but you can say it circulates, no source could be found, and your page does not carry it.

The mechanism that produces sourceless claims and makes them look corroborated is described under circular reporting.

Jason Momoa and Lisa Bonet's wedding date is the cleanest celebrity example. It was previously believed that they had married on November 15, 2007; they did not marry until October 2017. A date with no traceable origin had been repeated for a decade. A timeline that carried the 2007 date in its own voice is in the first row of the table above and owes a dated correction on that page, keeping the earlier wording visible. A timeline that said only that a 2007 date was widely reported is in the third row: accurate as written, but it should now carry the October 2017 date beside it. A timeline that never carried a wedding date owes nothing public at all.

The internal record matters more here than the public one. Log what was checked, when, and what was found, so the same claim recirculating in a year is answered from the file rather than from scratch. What that log should hold, and what it should never publish, is bounded by the notes on what a record should hold.

Replying without publishing

A direct reply to a requester is not a publication and it is frequently the whole of the correct response.

Say what your page carries and what it does not. Say what you checked. Say what you will change, if anything, and when. Keep the exchange in your log with dates. None of that circulates the claim, and it answers the person who actually contacted you, which a public note does not necessarily do.

One exception: your page carried something wrong. Then a private reply is not enough, because the error is public and readers will not read your email.

What a public note must accomplish then, and why the earlier state is preserved rather than overwritten, is in the notes on how entries age.

Common questions

Should a correction name the claim being corrected?

Only as narrowly as is needed for a reader of the original page to understand what changed. A correction that restates a claim in full has republished it to a new audience.

Someone demands a public retraction of something we never published. What do we say?

That the page does not carry it, in a direct reply, with the date. A public retraction of something you never said would be your first publication of it.

Does a denial belong in the entry itself?

If the entry exists and the denial is relevant to it, yes, attributed and dated as a statement. What it must not do is turn the entry into a running exchange about a private matter.

Should a request ever be published?

No. The request is a dated fact for your log. Publishing who complained is a new disclosure about a person who was trying to reduce one.