Epistemic class: how we label the sources behind each claim.
Most UAP coverage treats sources as interchangeable. A pilot's first-person account, a peer-reviewed paper, an anonymous leak, and a sworn congressional transcript all get the same visual weight in a typical write-up. They are not the same kind of evidence. This site tags every claim with one or more epistemic-class labels so the reader can see, at a glance, the kind of source behind it.
Note: the class describes the claim's source-and-vetting profile, not whether the underlying factual assertion is true. A sworn claim is sworn testimony; it is not, by virtue of being sworn, also factually verified.
Declassified document
A government document released under FOIA, statutory declassification, or voluntary agency release. Authentic; carries the agency's institutional weight at the time of release.
Institutional statement
Statement issued by an official US or foreign government agency on its own letterhead — press releases, annual reports, official-imagery releases. Carries the agency's institutional position, not just an individual's.
Sworn testimony
Testimony given under oath, typically before a congressional committee or in a deposition. The witness has accepted the legal consequences of false statement; the testimony's factual claims remain to be evaluated on their own evidence.
Named source, on the record
A named individual with verifiable government or contractor service has made the claim publicly and on the record. Excludes anonymous sourcing.
Peer-reviewed publication
Published in a peer-reviewed academic venue. Has passed editorial and reviewer scrutiny for methodological soundness; the underlying claims remain subject to subsequent replication and critique.
Independent analysis
Analysis produced by an independent researcher, journalist, or non-governmental organisation — not commissioned by a government agency. Distinguishes investigative work that meets sourcing standards but is not peer-reviewed.
Contested
There is an active, documented public dispute about the underlying account. Listed because the dispute is part of the record; the underlying factual question remains open.
How a single entry can carry multiple classes
The classes are not mutually exclusive. A House Oversight transcript is both sworn (the witness testified under oath) and institutional (the document is published on the official congressional record). A peer-reviewed paper from an AAWSAP-contracted physicist that discusses recovered material may be peer-reviewed, named-source, and contested simultaneously.
When you see multiple badges on a claim, treat them additively: each adds a separate sourcing dimension. A claim with three orthogonal badges is structurally stronger than one with the same single badge repeated, all else equal.
When a claim has no badge
A small share of entries — typically older incidents whose sourcing predates the modern documentary record — carry no epistemic-class label. Treat the absence as a signal: we have included the event because the historical record references it consistently, but the source-quality dimensions our classes capture are not cleanly applicable. The full body text of those entries explains the documentary situation case by case.
Why we publish the methodology
A site of this kind is only as useful as its editorial honesty. Publishing the class taxonomy lets readers, researchers, and journalists check our work — and lets us be told when a classification is wrong. The corrections inbox is open. Mistaken labels get fixed publicly with a dated note on the affected entry.
Related: About Disclosure Archives · Research library · Briefings.