The trough the empire fell into

Claim (verbatim)

Plot dated cuneiform by period and a canyon opens after the Old Babylonian collapse. Tablets thin dramatically across the First Sealand Dynasty and the early Kassite 'dark age' (roughly 1730-1450 BCE) before the Middle Babylonian recovery. This discontinuity can be production - a real political-economic collapse suppressing the making of documents - or survival - the relevant southern and Sealand sites are drowned, unexcavated, or looted. The histogram cannot by itself separate these, but its sharpness constrains the choice, and its sharpness is measurable in the catalogue right now. The mechanism at issue is precisely the ambiguity the coverage guard is built to adjudicate: a genuine ancient absence versus an accident of where the spade has fallen.

Prediction clause (verbatim)

Prediction (in-house CDLI copy, 126,000 rows): the dated-tablet count falls from the Old Babylonian bin (in-house: 14,975, plus 1,573 Early Old Babylonian) to the Middle Babylonian bin (in-house: 1,790) by a factor >= 5 (in-house actual ratio about 8; primary clause: Old-Babylonian:Middle-Babylonian dated-tablet ratio >= 5); and the trough is a genre-general collapse rather than one archive vanishing - the ratio remains >= 3 within the administrative genre alone. Disambiguation: 'Old Babylonian' and 'Middle Babylonian' are CDLI's own period labels; count distinct catalogued tablets, not fragments. Coverage guard: report alongside the number of distinct proveniences per period - if the Middle Babylonian site count is itself >= 5 times smaller, the drop is excavation rather than history, and the item resolves toward survival-accident rather than ancient absence.

Kill-dataset (verbatim)

Kill: the CDLI period field, dated-tablet histogram with per-period provenience counts. Runs THIS WEEK in-house, confirmable on the full CDLI catalogue.

Novelty / leakage triage

already answered in the literature

Self-measurement by construction: the primary clause's ratio was run in-house during design - the item and the register both state the in-house OB:MB dated ratio of about 8 against a threshold pinned at >= 5. The field has also published the shape of the trough: Streck's quantification of the cuneiform corpus by period and Brinkman's catalogue of dated Kassite material document the post-Old-Babylonian documentary collapse. Only the within-administrative ratio and the per-period provenience guard remain un-run.

• In-house CDLI catalogue copy (126,000 rows, about 30 percent of CDLI's 421,501 artifacts, harvested 2026-07-04), the wave's own registered aggregates: Old Babylonian 14,975 (+1,573 early OB) vs Middle Babylonian 1,790 dated rows
• M.P. Streck, 'Grosses Fach Altorientalistik: Der Umfang des keilschriftlichen Textkorpus', MDOG 142 (2010)
• J.A. Brinkman, Materials and Studies for Kassite History, vol. 1 (Chicago, 1976)

Predictions

No prediction registered yet.