AI-generated conjecture · below the evidence/publication boundary

Two species of archive

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)

Excavated tablet groups run from a dozen tablets in a jar to tens of thousands in a palace wing, and the sizes are usually treated as accidents of preservation. Economics knows that firm sizes form structured distributions, with small owner-operated firms and large institutions obeying different statistics. The conjecture: archive sizes are a two-regime mixture — household archives lognormally distributed, bounded by the rhythms of family affairs, and institutional archives forming a separate heavy tail fed by compulsory recording — with a diagnostic crossover in the low hundreds of tablets. If it holds, size alone classifies an archive as household or institutional, which matters enormously for the thousands of looted, provenance-stripped lots whose social context is otherwise unrecoverable.

Prediction clause (verbatim)

Over CDLI archive and provenance groupings with at least 10 tablets, a two-component mixture model will fit archive sizes significantly better than any single lognormal or single power law (likelihood-ratio and BIC), with the component crossover between 150 and 500 tablets. Primary clause, which decides the verdict: the two-component model wins on BIC by at least 10 points. Secondary clause: archives independently labelled as private houses fall at least 80 percent in the lower component, and palace or temple archives at least 80 percent in the upper.

Kill-dataset (verbatim)

CDLI (provenance, findspot, and archive-group metadata across all periods), with BDTNS archive assignments for the Ur III subset.

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

Composed blind by claude-fable-5 from internal knowledge only, with zero tool calls, and emitted directly as a single JSON text message.

Novelty / leakage triage

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

Searched archive-size statistics. The qualitative dichotomy (private archives of dozens to ~a thousand tablets vs. institutional archives of many thousands) is well documented, but no statistical mixture-model treatment of archive-size distributions or size-based classifier was located.

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.