Borrowed-Credential Sourcing — a real archive identifier attached to the wrong claim
mechanism
Attach a real archive document number to a false claim, and anyone who checks gets a yes and stops reading.
Who they are
A named failure mode, with two worked examples found in one research report this week[1][2].
What they do
In the engine this is how a claim survives being checked without ever being true.
How it works
The report cited Nuremberg evidence code NO-446 about a Nazi doctor. Harvard's Nuremberg archive does hold NO-446. It is the appeals judgment in the case of a man named Karl Kieling, dated 10 October 1946, filed under the euthanasia programme, and it has nothing to do with the doctor. The same report cited the 1950 Der Spiegel article as linking three men. The article is real. Two of the three names do not appear in it anywhere[1][2].
Why it matters
It matters because the citation is not invented, which is what makes it hard to catch. The container is genuine and the contents are not what the citation says. A web address and a date pass inspection. A quote pulled from the document itself does not — and this engine's own record on that same doctor failed the same test, which is written into the correction.
The engine's record — word for word
[Report #195] **In plain terms:** a way of making a false claim survive a spot-check. You attach a real document number from a real archive to a claim that document does not support. Anyone who checks whether the document exists gets a yes and stops there. Checking the document's CONTENTS is a different and much rarer act.
THE WORKED EXAMPLE, established this pass. A research artifact cited Nuremberg evidence code NO-446 in support of a claim about Hubertus Strughold. The Harvard Law School Library Nuremberg Trials Project holds NO-446. It is, verbatim from the catalogue, 'Im Namen des Volkes [reversal of a sentence against Karl Kieling]', also catalogued as 'Judgment of Circuit Court of Appeals in the case of Karl Kieling', dated 10 October 1946, author 'Circuit Court of Appeals', held at HLSL item numbers 715501, 715511, 1758 and 1759, filed under the trial issue 'Euthanasia program (Action 14 f 13)'. It is a real Nuremberg document. It has nothing to do with Strughold. The citation is not fabricated; it is misattached, which is harder to catch.
THE SECOND INSTANCE IN THE SAME ARTIFACT: a citation to the March 1950 Der Spiegel article as linking Schriever to Habermohl and Miethe. The article is real and was retrieved. Neither name occurs in it. Again the container is genuine and the contents are not what the citation says.
WHY THE ENGINE HOLDS THIS. The engine's own standing practice is to carry a URL, a date and a quote per source. A URL and a date alone would have passed both of these. The quote field is what catches them, and only when the quote is taken from the document rather than from the citing work. This record exists so the failure mode is named and checkable, and because this engine's own Strughold record failed a version of the same test (see the correction at hubertus_strughold in this report).
SCOPE. This is a named failure mode, not an allegation of intent about any author. Nothing here establishes whether any given misattachment was careless or deliberate, and none is asserted.
Follow the trail
Walk this on the live map →