One culture, two garbage cans

Status: Anticipated · untested

Status is derived only from the shepherd-authored triage/prediction data above -- community submissions and claims are a separate overlay and can never change it (see the participation panel below).

This is a proposed connection between two domains, generated by a language model. It is not an article and not evidence: it sits below the evidence/publication boundary. A quantitative prediction and a named kill-dataset are attached (when registered) so the claim stays falsifiable rather than merely evocative.

Claim (verbatim)

Dunhuang Cave 17 was a deliberate deposit inside a shrine; the Turfan finds come largely from ruins, dumps, and graves. Join the two most famous manuscript hoards of Inner Asia into a single instrument: the two sites sampled broadly similar oasis document cultures through opposite filters — curated sacred storage versus accidental refuse — so their genre distributions differ by the deposit filter, not by life, and the ratio between them is a correction factor for what any oasis actually wrote. Monks curate scripture in; rubbish lets contracts and letters in. This is a structural claim about sampling bias, aimed past the caves at the dark interior. If it holds, everyday documentary writing at Dunhuang, and by extension across Inner Asia, was several times more voluminous than the cave suggests, and the correction factor is a number rather than a rhetorical gesture.

Prediction clause (verbatim)

Primary clause (verdict-bearing): the share of documentary genres (contracts, letters, accounts, official papers) among catalogued Turfan items exceeds their share among Cave 17 Dunhuang items by a factor of at least 3. Secondary clause: religious texts show the mirrored excess on the Dunhuang side.

Kill-dataset (verbatim)

Genre and type metadata distributions computed across the Digitales Turfan-Archiv and the IDP Dunhuang Cave 17 holdings.

Nobody has run this test. The kill-data is named above. If you can run it — or you know the paper that already settles it — claim the kill or submit the prior. Kills and priors are credited here, by name, as they come in.

In the atlas

This conjecture is bridged, as an L1 lead, onto these Inferpedia subject pages.

Provenance

Run: Fresh agent generation · model: claude-fable-5

Generated blind in a single Write by a fresh instance with no file reads, web access, or database queries; all context was inline in the launching prompt.

Novelty / leakage triage

anticipated in the literature — this exact test has never been run

The opposed formation processes — Cave 17 as deliberate sealed deposit vs Turfan's ruins, dumps and tomb-recycled paper — are extensively discussed (library-cave nature debate; bundle reconstructions), but nobody has computed the paired genre distributions as a quantitative deposit-filter correction factor.

Predictions

No prediction registered yet.

Weigh in

No community feedback yet.

Add your take

Posted immediately (spam is removed). Community feedback is never an adjudicated verdict and never changes this conjecture's triage label or status above.