We Called It A Ghost. It Was There. The Hedge Was The Error.

ITHACA, N.Y. — Thirty-one items. Eleven qualifiers the evidence did not require. One sentence, nine times, after the document that contradicts it had been offered to this newspaper and declined unread.

That is what the public editor's office found in our own coverage of the Khafre subsurface, counted line by line and set against the borehole footage of 2 August and the muon volumes returned this week.

The finding was correct. We were the newspaper that under-reported it.

What was counted

Every item published here on the subsurface at Khafre since March 2022: nineteen news items, six briefs, three analyses, two letters selections and one null result.

Each was coded against a single test, fixed before the count began. Does the qualifier name the evidence that requires it?

A qualifier that names evidence is reporting. Two of our three processors could not reproduce the helix is a fact about instruments, and a reader can go and check it. A qualifier that names nothing — the claim remains contested, specialists are sceptical, no independent confirmation exists — is a claim about the world made without a source, placed where a reader is least able to test it.

Fourteen of the thirty-one hedged and named their evidence. Six carried no hedge at all and were right not to. Eleven failed: seven in the first three paragraphs, four in the headline.

The ghost

The worst of the four is mine.

On 11 June this newspaper ran Second Processor Calls Khafre Helix A Ghost over a story whose fourth paragraph recorded that the first processor had reproduced the shafts and the chambers, and that the second had reproduced them too.

One of two specialists could not reproduce one feature out of three. That was the headline. Everything both of them did reproduce was paragraph four.

The helical feature is now resolved on a muon array that has been running on the plateau since 4 August, at 470 hours of exposure. The stair the camera photographed at 384 metres turns the way the ghost did.

The sentence

One clause did most of the damage, and it is not a clause anybody argued about.

Nothing below thirty metres has been seen by anybody.

It has run in this newspaper twelve times since 2011, nine of them inside this audit set, in four reporters' copy and twice in mine. It was written as a statement of the record. It worked as a full stop.

It was also false, and had been false since 1978.

What was offered, and what we did with it

In March 2019 a retired site clerk named Ferrand Okpalaeke-Voss wrote offering this newspaper the field file of a geotechnical investigation carried out on the plateau in 1978, for a visitor road that was never built. Four boreholes. His employer's copy, kept when the firm was wound up.

Hole three reached forty-seven metres. At forty-four metres the driller logged a lamp-and-mirror inspection and wrote four words: sound limestone, no void.

That is not a discovery. Four holes for a road, none of them within two hundred metres of the shaft heads the radar later put on the map, and the deepest stopping some six hundred metres above anything anybody is arguing about. It does nothing for Malanga, Biondi and Mei either way, and this column will not pretend otherwise. It shows limestone, at forty-four metres, under a road that was never built.

What it shows is that somebody had looked below thirty metres, with an instrument, forty-one years before we began telling readers that nobody had.

The letter was answered in nine days by the deputy editor of this desk, Hesper Rowntree-Adekunle. Her reply is two sentences long. The file was not requested, not read, and not logged. Mr Okpalaeke-Voss kept our letter for seven years and supplied it this month, and he was asked about it now only because a camera had made the question embarrassing.

He had written four times. The first three are in our own correspondence file and were also unanswered.

Why this is not a victory lap

The temptation in an audit like this one is to make the caution the only error and the finding look inevitable in hindsight. It was not inevitable, and Ms Rowntree-Adekunle's account below makes the case for the hedge better than anyone in this building has made it, and this desk has not been able to answer her.

But caution has a test and eleven items failed it. A qualifier that names its evidence costs a reporter one telephone call. A qualifier that names nothing costs nothing at all, which is why it spreads, and why it was still in our copy in June — six weeks before we drilled, and eight weeks after we had bought and published the radar passes at PX-1912 that told us where to drill.

The amendment

Clause 3b of this newspaper's standards, entered today:

No story carries a qualifier in its headline or first three paragraphs unless the reporter can name, in the same paragraph, the evidence that requires it. A hedge is a claim and is sourced like one.

And under clause 4a, which obliges The Vault to list what we refused: the 1978 field file is entered there this morning with the date it was offered, the name of the editor who declined it, the four letters, and the sentence it would have killed.