Anniversaries and archives corrections: three ways to change an old page

An archive is not a record of what is true. It is a record of what was published, and that difference is its whole value.

If an archive could be silently revised whenever someone learned better, it would stop being evidence of anything. You could no longer show what a paper said in a given week, which is why people consult it.

That principle collides, regularly and painfully, with a family who has found an error in a notice from years ago, or who simply does not want a page about their loss to keep appearing.

Nobody is named in what follows and no death is reported. The situations described are the recurring kinds, not any actual case.

What to take away

Three ways to change an old page, and what each costs

Append. The original text stays; a dated note is added saying what was wrong and what the correct information is. The record of what was published survives, and so does the truth. This is the default, and for most factual errors it is the right answer.

Amend and mark. The text itself is changed and the page carries a visible note recording the change and its date. Appropriate when leaving the error in place would keep actively misleading readers: a wrong name, a wrong survivor, anything a reader would take at face value in the first sentence.

Unpublish. The page is removed or closed to search engines. It is the only remedy that actually stops the page being found, and it is the one that damages the record. Publications reserve it for a narrow set of situations, and even then usually leave a stub saying that something was removed and when.

There is a fourth option that looks like the first three and is not: quietly editing the text and saying nothing. It leaves readers who saw the earlier version with no way to understand what changed, and it makes every other page in the archive slightly less trustworthy.

When a family asks for a notice to be taken down

These requests are not rare, and rarely unreasonable. A search for their own surname returns a notice about their father's death, and they relive it whenever an acquaintance looks them up.

Another finds their address, or their surviving children's names, in a notice they did not place. A notice may name people who never agreed to appear in it.

Publications weigh a few things:

Laws on personal data, publication and removal vary widely by country. A request framed as legal usually goes to a different queue than an editorial one.

Erasure rights regulators describe, like the UK guidance on your right to get your data deleted, cover living individuals. So a relative's own details inside a notice rest on stronger ground than the notice itself.

If that is the ground you are standing on, say so, and expect a slower and more formal reply. Ask the publication for its own policy first. Most that hold archives have one, and it will tell you what is available before you spend energy on a request that was never going to be granted.

A partial remedy is worth asking for explicitly. Removing a photograph, removing an address, or asking that a page be excluded from search indexing while remaining reachable by direct link are all things a publisher may agree to when full removal is refused.

Anniversaries put an old page back in front of people

Anniversary features and "on this day" republication give an archived page a fresh audience. That is a decision to publish, not a decision to display, and it carries the same duties as any other.

Poynter's ethics and trust material collects broader thinking on where harm arises, and most of it comes down to setting the policy in advance rather than case by case.

A named surviving relative may have died, and a named child is now an adult with their own life. Republishing without looking is how a publication apologizes for something it wrote correctly years ago.

The same applies to automated anniversary emails from memorial platforms. Sending someone a reminder of a death on its date is welcome to some people and cruel to others, and the difference cannot be inferred from the record. Where a system does it at all, it should be something a person switched on.

Pages outlive the things that host them

Anyone maintaining a memorial record for the long term runs into a set of problems that have nothing to do with editing.

Hosts close, and the pages go with them. Platforms are bought and migrated, and the migration usually keeps the text and loses the guest book, the photographs, or the dates on the comments.

Addresses change and the links pointing at them rot. A memorial page that everyone linked to ten years ago may now be reachable only through a web archive. Guest books left open for years accumulate spam, and closed ones cannot receive the message someone finally felt able to write.

A publisher alone cannot fix this, but two habits help. First, keep your own copy of anything you care about: the text, the photograph, the messages people left. The platform is not a custodian and has not promised to be one.

Second, where you control the record, keep the addresses stable. Every link someone else made to it is a small act of remembrance you would be breaking.

Where this sits

Correcting a recently published notice is different and quicker. Those steps are in how a correction is handled, while building a career timeline covers assembling a chronology that must hold up over time.

The parent section is anniversaries and archives, and how to read a new notice is in deaths this week.

Common questions

Can I have an old notice taken down?

Sometimes, and it is the hardest thing to ask for. Ask for the publisher's policy first, and ask for a partial remedy in the same message.

What counts as a partial remedy?

Removing a photograph, an address or a child's name; excluding the page from search indexing while leaving it reachable; stopping it being used in anniversary features.

Why will they not just delete it?

Because an archive that can be edited on request stops being evidence of what was published, which is the reason anybody consults it.

Does correcting the page fix the copies?

No. It fixes the version most people will find next, which is usually the one that matters, and it gives the copies something to point at.

Who should send the request?

One person, with one contact address, saying plainly which of the remedies above they are asking for.