Copies per text follow preferential attachment: power-law copy counts, with survival size-biased so bestsellers over-survive superlinearly. Falsify: medieval library catalogs vs extant counts.
Generated by Fable · below the evidence/publication boundary
One Thousand and One Conjectures
308 of 1001 posed · 158 shepherd-triaged · 150 provisional · 0 frontier · 20 predictions · 9 resolved (6 supported / 3 killed) — the 1001st will be posed at Ars Inquirendi, Oxford, 20 November 2026.
Cross-domain conjectures generated noetically by Fable — a frontier AI proposing, from its own knowledge, surprising connections between two well-known domains that it judged likely to be both novel and important. Each pairs a specific claim with a quantitative prediction and a dataset that could prove it wrong; each was then checked against the literature to flag the ones with known priors.
This is one form of lead generation for Inferpedia, the encyclopedia of the missing — and this page is an early preview.
Why these conjectures matter: charting the noetome — its structure, limits & potential →
Nothing here is claimed as verified-novel. Each sits below the evidence/publication boundary: a connection already known in the literature is shown honestly and tagged Prior, and every prediction is registered before it is scored. Spotted a prior yourself? Open any conjecture and weigh in.
Filter
Clear all filtersWhat the tags mean
- Open — no decisive result yet
- Prior — a prior formulation exists in the literature
- Supported — a registered prediction held up in data
- Falsified — a registered prediction was refuted
- testable — a quantitative prediction + kill-dataset is registered
- Shepherd-triaged — an authoritative Fable-authored verdict; shown as the pills above and the only tier in the headline numbers
- provisional — model-triaged, shepherd review pending — an Opus-authored first pass, not yet shepherd-confirmed and excluded from every headline figure
- awaiting prior-art check — hunt open — no triage yet; found a prior yourself? open it and weigh in
Showing 1–11 of 11 matching conjectures.
The mutually derivable columns of Easter tables (golden number, epact, dominical letter, indiction) form a redundant error-correcting code that medieval users actually exercised.
Scribal error-apology formulae (“forgive the faults of the unworthy scribe”) are costly quality signals from careful scribes, not confessions from sloppy ones.
Manuscripts became binding waste on an institutional obsolescence clock, not a material wear-out clock.
Monastic copying was counter-cyclical to farm labor; professional urban copying was not — so the seasonality amplitude of dated colophons is a quantitative urbanization index of book production.
Production mode (visual exemplar copying vs dictation/pecia mass production) is recoverable from error typology alone.
Staff notation cut chant copying errors discontinuously: predict per-neume Levenshtein error drops by half across the Guidonian transition within single traditions. Falsify: aligned chant stemmas.
Within a scriptorium, per-copy error rates are constant enough to date manuscripts by error accumulation, like a decay clock. Falsify: error counts vs colophon-dated families.
Joins the cognitive psychology of recall (the serial-position curve: primacy and recency beat the middle) to the stemmatics of oral law: Icelandic law was recited from memory in fixed thirds over three years, so when it was finally written down, the middle…
The cutting-stock problem of operations research meets codicology: parchment page sizes are not aesthetic free choices but near-optimal cuts of animal skins, so leaf dimensions quantize at dyadic folds of regional skin rectangles, while paper leaves, cut from standard moulds, cluster at…
Packet-switched content distribution meets the medieval university book trade: the pecia system chunked exemplars into pieces copied in parallel and reassembled, so stemmatic affiliation should switch at known pecia boundaries — an error signature that continuous-exemplar copying cannot produce.