The kings wrote only in stone

Claim (verbatim)

Before the Timbuktu chronicles of the seventeenth century, the earliest securely dated written record of the Middle Niger and its Saharan approaches is not on paper but on stone: the royal funerary stelae of Gao-Saney and the grave and rock inscriptions of Essuk-Tadmakka and Bentyia/Kukiya, the corpus assembled by Paulo de Moraes Farias (Arabic Medieval Inscriptions from the Republic of Mali, 2003) after Sauvaget's study of the Gao royal epitaphs. Some Gao stelae are imported marble, cut and inscribed in the Iberian workshops of Almería and carried across the desert — an object whose very existence proves a literate court commissioning texts abroad in the twelfth century. The mechanism is substrate triage under a hostile climate: Sahelian paper is mortal (termite, Harmattan, damp) and must be recopied or is lost, but cut stone persists, so for the earliest centuries the epigraphic record is not a supplement to the manuscript record — it is very nearly the whole surviving written record, and it drastically undercounts production, because only the monumental and lapidary genres had a survival substrate. Prediction: among all written items from the Middle Niger and its Saharan edge that are securely datable before 1450 by an internal date or archaeological context, at least nine in ten will be inscriptions on stone rather than manuscripts on paper or parchment, and no in-situ Malian paper manuscript will be securely datable before circa 1400 (primary clause: the nine-in-ten epigraphic share of pre-1450 datable writing; the verdict follows it). Kill: P. F. de Moraes Farias, Arabic Medieval Inscriptions from the Republic of Mali (British Academy/Oxford, 2003), with Jean Sauvaget, 'Les épitaphes royales de Gao' (Bulletin de l'IFAN, 1950), counted against the earliest securely dated paper manuscripts recorded in the AMMS and SAVAMA-DCI Timbuktu inventories.

Prediction clause (verbatim)

Prediction: among all written items from the Middle Niger and its Saharan edge that are securely datable before 1450 by an internal date or archaeological context, at least nine in ten will be inscriptions on stone rather than manuscripts on paper or parchment, and no in-situ Malian paper manuscript will be securely datable before circa 1400 (primary clause: the nine-in-ten epigraphic share of pre-1450 datable writing; the verdict follows it).

Kill-dataset (verbatim)

Kill: P. F. de Moraes Farias, Arabic Medieval Inscriptions from the Republic of Mali (British Academy/Oxford, 2003), with Jean Sauvaget, 'Les épitaphes royales de Gao' (Bulletin de l'IFAN, 1950), counted against the earliest securely dated paper manuscripts recorded in the AMMS and SAVAMA-DCI Timbuktu inventories.

Novelty / leakage triage

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

Moraes Farias' 2003 corpus is the standard edition of the Gao-Saney, Essuk-Tadmakka and Bentyia inscriptions, and the observation that the earliest securely dated writing of the Middle Niger is lapidary rather than on paper — including the imported Almería-marble stelae — is his and Sauvaget's published ground. But the primary clause is a quantified census: that at least nine in ten of all pre-1450 securely datable written items are inscriptions on stone. That epigraphic-share proportion has never been computed against an enumerated early paper corpus, and the secure dating of the earliest in-situ Malian paper manuscripts is itself unsettled rather than fixed at c.1400. The materials and the qualitative claim are in print; the 90% count is not.

Predictions

No prediction registered yet.