← All conjectures · Medicine, science & the occult

One codex crossed the dark centuries alone

Medieval Dated c. 900 CE – 1700 CE Aegean & Greece

Status: Already answered

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 conjecture imagined by a language model — drawn from its trained weights and held to falsifiability, novelty, and plausibility, not to any one method: it may join two or more fields, or none. 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)

Between the assembly of the Byzantine alchemical anthology and the thirteenth century, the entire Greek tradition of the art appears to ride on effectively one book. The Catalogue des manuscrits alchimiques grecs (CMAG, Brussels 1924-1932) censuses the alchemical manuscripts of Europe's libraries, and date-stratifying that census should expose a survival profile more extreme than that of almost any classical corpus of comparable bulk: a crowd of fifteenth-to-seventeenth-century copies, a thin band of late-medieval ones, and before 1204 - the sack of Constantinople, the great dispersal event of Byzantine libraries - at most Marcianus gr. 299 itself (tenth century) and arguably nothing else, with Parisinus gr. 2325 following only in the thirteenth century and Parisinus gr. 2327 dated 1478. The mechanism: recopying was institutional, and alchemy had no institution. No school chair taught it, no liturgy required it, no examination reproduced it; the corpus survived as court curiosity and collector's luxury, and a tradition copied for curiosity is copied rarely, expensively, and in one place at a time. Renaissance demand then bred the crowd of recentiores off the few survivors, giving the tradition an hourglass shape with a single-book waist. Prediction restated: at most two witnesses of the corpus proper predate 1204, and at least 60 percent of all catalogued corpus witnesses postdate 1400.

Prediction clause (verbatim)

Prediction: date-stratifying the CMAG census of Greek alchemical manuscripts, restricted to witnesses catalogued as containing three or more works of the ancient corpus proper, at most two witnesses predate 1204 CE (primary clause: the <=2 pre-1204 witness count; the verdict follows it), and at least 60 percent of such witnesses postdate 1400; the test voids for coverage if fewer than 30 corpus witnesses carry usable date assignments in CMAG.

Kill-dataset (verbatim)

Kill: the Catalogue des manuscrits alchimiques grecs (Bidez, Cumont, Delatte, Heiberg, Lagercrantz and collaborators, 8 vols, Brussels 1924-1932) - a witness-date census over all manuscripts catalogued with three or more ancient-corpus works, counted against the 1204 and 1400 date lines.

Browse the registry: Union catalogues →

How this item is catalogued

Catalogue tags are curatorial metadata that make the conjecture findable. They are not findings, not a novelty verdict, and naming a source family is not a promise that the data exist or are accessible. Each dimension below says whether it was assigned, reviewed with none applying, uncertain, or not yet reviewed, and who recorded it.

Place & era tags — Assigned: Aegean & Greece; Medieval Tagged by claude-fable-5
Dating — Phenomenon: Dated c. 900 CE – 1700 CE Byzantine alchemical anthology (Marcianus gr. 299, 10th c. CE) through the 'fifteenth-to-seventeenth-century' crowd of recentiores per CMAG's own date-stratified census. Tagged by claude-sonnet-5 + claude-fable-5 (gate)
Subject — Assigned: Medicine, science & the occult (from the generation section)
Finer subject topic — Assigned: Technical, recipe, and scientific texts The survival profile of the Greek alchemical corpus is squarely a practical technical corpus. Tagged by claude-opus-5 draft (requested max effort) adjudicated by claude-fable-5-1 driver, lane 1da5; runtime not independently attested
Claim level — Assigned: Record claim The prediction counts surviving witnesses either side of 1204 and 1400, a fact about how the record was recopied and lost before the modern era. Tagged by claude-opus-5 draft (requested max effort) adjudicated by claude-fable-5-1 driver, lane 1da5; runtime not independently attested
Method — Reviewed — none applies: Two counts against date lines (at most two pre-1204, at least 60 percent post-1400); the hourglass survival profile is described but no hazard or survival curve is fitted. Tagged by claude-opus-5 draft (requested max effort) adjudicated by claude-fable-5-1 driver, lane 1da5; runtime not independently attested
Evidence source — Assigned: Union catalogues Kill test date-stratifies the CMAG manuscript census of Greek alchemical codices. Tagged by claude-opus-5 draft (requested max effort) adjudicated by claude-fable-5-1 driver, lane 1da5; runtime not independently attested
Research stage — Idea — no test designed

On Inferpedia

This conjecture is linked to the following pages on Inferpedia, an encyclopedia of the missing — working atlas pages, some still early scaffolding.

Provenance

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

Fresh blind generation by claude-fable-5, 2026-07-18, craft-and-technical-knowledge domain: the Graeco-Egyptian alchemical corpus and its Syriac/Arabic penumbra, chosen because it is the pre-print world's best-instrumented lost technical literature - a closed corpus with open, named instruments (CAAG 1887-88 open scans; CMAG 1924-32; Suda On Line; the Theban papyri in Leemans/Lagercrantz/Halleux; Berthelot-Duval CMA II open; Fihrist in Fluegel/Dodge; Kraus 1943 + Sezgin GAS IV) - and because greps showed the core untouched (Zosimos, Marcianus, Stockholm/Leiden papyri, Maria, Cleopatra, Ostanes, Democritus/Bolus, Berthelot, chrysopoeia, Kraus, Jabir/Geber, Hermes/Hermetica: no hits anywhere). Carves and seams, checked wave by wave: lost-in-translation 2026-07-18 owns translation-as-survival (Galen/Hunayn, Hero's Mechanics ord 5, al-Kindi Fihrist book-list survival ord 8, versio-only shares) - no Greek-cliff/Arabic-continuation item here, and the Syriac Zosimos enters only inside the ord-5 alphabet-coverage census whose primary clause is corpus completeness, not translation survival; byzantine-dbbe-pinakes owns Photios, pooled Suda title-survival (its ord 4) and the Pinakes earliest-witness histogram (its ord 9) - witness censuses here run on CMAG, a different instrument, and the Suda serves only as a single-entry book-count yardstick (operation-shape shared with ancient-sound ord 10 on Aristoxenus; new author, new letter-coverage operation, flagged); ancient-sound ord 7 owns burial-vs-copying for music - ord 8 here runs a descendant-census channel split on a different corpus with a different operation; fable-w20 ord 15 owns positional recipe stratigraphy (papyri.info/eVK) and ord 26 owns Latin alchemical incipit camouflage (eTK) - no positional or incipit operations here and Latin alchemy left untouched (w20's 'Mappae' hit is mappae mundi, so the Mappae clavicula remains free for a future recipe-book wave); the Fihrist is a crowded instrument (five prior waves) reused solely as the maqala-10 roster inside a two-catalogue asymmetry join no prior wave runs, flagged. Dropped candidates: Hero/Philon/Ctesibius mechanics (coordination instruction plus lost-in-translation ord 5), Dioscorides alphabetization (byzantine-wave adjacency; Galen owned), Geoponica/lost agronomists and Mappae clavicula/Theophilus (both verified free - reserved as future bodies), Turba philosophorum (translation-survival shape owned). Soft counts (pinax entry total, lexicon entry count, Pliny index-book count, roll length, Syriac letter-book count, Kraus register arithmetic) are kept out of load-bearing primary clauses behind explicit coverage guards; ords 2, 4, 8, 9 rest on HIGH-confidence anchors, ords 1, 5, 6, 10, 11, 12, 13, 14 on MODERATE counts flagged in the report, ord 15 is honestly Kill-not-yet-built.

Novelty / leakage triage

already answered in the literature

The date-structure the primary clause pins is the stated doctrine of the field, not a histogram waiting to be run: the handbooks and prolegomena state flatly that Marcianus gr. 299 (10th c.) is the tradition's only witness before the 13th century, with Parisinus gr. 2325 in the 13th, Parisinus gr. 2327 dated 1478, and all remaining witnesses recentiores of the 15th century onward derived from these - which fixes the pre-1204 count at one and guarantees the post-1400 dominance over the CMAG census the conjecture would recount.

Sources cited by the triage

R. Halleux, Les textes alchimiques (Typologie des sources du Moyen Age occidental 32, Turnhout, 1979), on the witnesses of the Greek corpus
J. Letrouit, 'Chronologie des alchimistes grecs', in D. Kahn & S. Matton (eds.), Alchimie: art, histoire et mythes (Paris-Milan, 1995)
Bidez, Cumont, Delatte, Heiberg, Lagercrantz et al., Catalogue des manuscrits alchimiques grecs, 8 vols (Brussels, 1924-1932)

Predictions

No prediction registered yet.

Weigh in

No community feedback yet.

New here? Create an account first

Create an account or sign in and your feedback is tied to you — you can track it, get replies, and claim this conjecture so others know you’re working on it. Prefer not to? Just leave your take below as a guest — only the name you type is shown.

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.