A published page can change.
The harder question is what made it need to change.
Across Clyde Street, corrections arrive from different directions. A knowledgeable reader notices an error. A living witness challenges an attribution. A photograph makes an old reconstruction harder to defend. A newly located source moves an “earliest” date. A second look at the data changes a public forecast before the event it concerns has even happened.
Those are not the same kind of correction.
What connects them is narrower.
Something already public becomes less defensible in its current form. Keith changes the record. And, in the strongest cases, enough of the correction remains visible for a later reader to ask four questions:
What challenged the account?
What changed?
When did it change?
What remains uncertain?
That is more demanding than editability. A page can be technically changeable without ever being corrected. The intellectual issue is whether a changed public claim remains answerable to what made the change necessary.
The boundary matters. Clyde Street also contains public objects that remain open to addition or later development. A missing paper can be added to an already-released archive; a post can acquire new material; a project can continue after publication. Those are forms of revisability. They do not by themselves show that an existing claim has met a challenge and needed correction. Here, the pressure against the existing record is essential.
WHEN SOMEBODY WRITES BACK
In June 2012, Keith returned to an earlier account of Charles Reep and Brentford after contributors to a Brentford supporters’ forum questioned parts of the story.
Their scrutiny exposed problems in the league position and match results Keith had repeated. He went back to the record, corrected the original post and made the prompt for the correction visible.
The repair was local.
Keith did not treat the corrected details as proof that the larger historical question of Reep’s involvement was now settled. The people who challenged the account had made specific claims harder to sustain. That did not give them ownership of the whole history.
The same distinction matters in Keith’s correspondence with Neil Lanham. Neil’s testimony repeatedly changed parts of Keith’s public account, including an earlier attribution around Wimbledon. Neil was a consequential witness. He was not a final court of appeal.
The fuller relationship belongs elsewhere in Clyde Street. Here, the important point is smaller: a public record can become more answerable because the person who knows something about the story can object to how it has been told.
That does not make every objection correct.
It makes the source of the correction part of the record.
WHEN AN OBJECT CHANGES THE ACCOUNT
Sometimes the challenge does not come from a person arguing with the text.
For roughly twenty years, Keith thought Lloyd Messersmith had used a pastry cutter to trace basketball-player movement.
Then Bradley Cook, Curator of Photographs at Indiana University, shared a photograph.
In Cut to the Chase?, Keith looked again. The metal frame around the wheel was thicker than his remembered explanation comfortably allowed. Bradley’s interpretation was that the device was probably a swivel caster from a chair.
Keith changed his reconstruction.
But he did not turn the new interpretation into certainty. Bradley had offered a guess, and Keith retained that probability in the public account.
This is an important kind of correction because it resists the neat story in which an error is simply replaced by truth.
The old explanation had lost warrant.
The new one was better supported.
A question mark still remained.
A correction can therefore improve a public record without closing it.
“EARLIEST” CAN MOVE
In April 2017, a lead from Simon Gleave sent Keith back into the early history of football statistics.
An earlier Clyde Street account had treated June 1937 as the earliest example Keith knew. The new source trail moved that date to January.
That sounds like a simple correction: one date replaced by another.
The source walk made it less tidy.
Keith compared material from a German football-association page, film, a match programme, Wikipedia and Kicker. One Wikipedia link led to a sociolinguist with the same name as a footballer — someone who would have been thirteen at the time of the match. Birth-date information remained incomplete. A postscript added another source.
The priority claim moved, but the source base also became visibly more uneven.
“Earliest” became more provisional, not less.
That is useful because historical correction can easily be mistaken for historical completion. A newly found source can move the boundary of what is currently known without giving the writer permission to announce that the search is over.
CORRECTION HAS A CLOCK
The clearest case is not historical.
On 29 September 2017, Keith published Profiling the 2017 #AFLGF Teams before the Australian Football League Grand Final between Adelaide and Richmond.
He was explicit about the information boundary. He had not watched AFL that season and was using quarter-score medians he had compiled manually.
Then the page changed.
It records that the data were corrected on 30 September. It also records that Keith revised the estimated Adelaide advantage after reviewing the data and before the game.
Richmond later won.
That later result creates a temptation.
Once we know the outcome, it is easy to read every earlier change through hindsight — as though the correction anticipated the result, responded to it, or gained legitimacy from it.
It did none of those things.
The correction and revision belonged to the pre-event state of the page.
The match result belonged to a later state of knowledge.
This is why correction has a clock.
A reader should be able to distinguish what Keith first published, what he corrected before the event, what he revised before the event, and what became knowable only afterwards.
Without that chronology, a public record can remain visibly edited while its intellectual history has quietly disappeared.
WHAT A CORRECTABLE RECORD PRESERVES
Clyde Street contains other forms of public change.
A missing contribution can be added to an already-published archive. A post can remain open to later additions. A word or category can be replaced when it stops fitting. A public project can continue developing after release.
Those neighbouring problems matter, but they are not identical to correction.
The distinctive question here begins when something presses against an existing public claim or state strongly enough that it has to change.
The strongest Clyde Street examples keep four things recoverable.
First, cause: who or what challenged the record?
Second, change: what claim, number, attribution, date or interpretation was actually altered?
Third, clock: when did the change occur relative to what happened later?
Fourth, residual uncertainty: what still cannot responsibly be claimed after the correction?
Keeping those elements visible prevents correction from becoming erasure.
It also prevents a redemption story in which every mistake ends in certainty.
A living witness may correct an attribution and still leave the larger history contested. A photograph may weaken an old explanation without proving the replacement. A new document may move an “earliest known” date while making the archive look more incomplete. A data correction may alter a forecast before an event without being vindicated or invalidated by the result that follows.
A correctable public record is therefore not a mistake-free record.
It is a record whose changes can remain answerable to the evidence and people that made them necessary.
Question to carry
When a public claim changes, can the reader still reconstruct what challenged it, what changed, when it changed — and what remains uncertain?
Sources followed
Keith Lyons, “Charles Reep and Brentford” — 12/06/22
Keith Lyons, “Cut to the Chase?” — 13/06/13
Keith Lyons, “1937 Back Pass” — 17/04/14
Keith Lyons, “Profiling the 2017 #AFLGF Teams” — 17/09/29