Half a cosmos, whole only in Arabic

Claim (verbatim)

Ptolemy's Planetary Hypotheses is the book where mathematical astronomy becomes a physical cosmos - and Greek preserves only the opening stretch of Book I (Heiberg's text in the Opera astronomica minora, 1907). The rest of Book I, with the nesting-spheres computation that converts the Almagest's ratios into an absolute universe - each planet's shell beginning where the previous ends, from the Moon's least distance out to the sphere of the fixed stars at roughly twenty thousand Earth radii - and the whole of Book II survive in Arabic (the London manuscript published in facsimile and analysis by Bernard Goldstein, TAPS 57.4, 1967; the Arabic Book I edited by Régis Morelon) and in the Hebrew made from the Arabic. That lost-in-Greek remainder is the part the Middle Ages actually lived in: the distance-and-size ladder that al-Farghani's Elements broadcast westward, the cosmos of scholastic textbooks and of Dante. Until Goldstein printed the Arabic, moderns knew Ptolemy's own version of his cosmology mostly through its medieval echoes. The mechanism is curricular again: Byzantium recopied the Almagest, the classroom text with its teaching apparatus, and let the off-syllabus physical appendix dwindle to its opening; Baghdad's astrologers and cosmographers wanted precisely the world-picture, and kept all of it. The two streams each preserved the half their institutions had a use for - and only the Arabic half contains the universe.

Prediction clause (verbatim)

Prediction: measured against the Arabic version as published in Goldstein's facsimile, the portion of the Planetary Hypotheses with any surviving continuous Greek parallel will amount to less than 40% of the whole work, and the planetary distance-and-size section will fall entirely within the Greek-less portion (primary clause: the <0.40 Greek-coverage fraction plus the location of the distance section; the verdict follows them); secondarily, the Earth-radii distance ladder transmitted by al-Farghani's Elements will match the scheme of the Arabic-only section and correspond to nothing in the extant Greek stretch; Greek parallel means Heiberg's continuous text, quotations in later Greek authors do not extend it, and the test voids if the Arabic witness proves incomplete for Book II.

Kill-dataset (verbatim)

Kill: Goldstein, The Arabic Version of Ptolemy's Planetary Hypotheses (TAPS 57.4, 1967) with Heiberg's Greek (Opera astronomica minora) and Morelon's Arabic Book I - a coverage-fraction measurement and section-location check, with al-Farghani's Elements as the downstream echo test.

Novelty / leakage triage

already answered in the literature

This is Goldstein's published discovery stated as such: his 1967 study says explicitly that the Greek (Heiberg's text in the Opera astronomica minora) covers only the opening portion of Book I, that the nesting-spheres distance-and-size computation falls in the Arabic-only remainder, and that Book II survives only in Arabic and the Hebrew made from it; Morelon's edition fixes the Arabic Book I's extent. The <0.40 coverage fraction is page-arithmetic on those printed extents with the threshold far from the boundary, and the al-Farghani distance-ladder echo is the standard account of the section's medieval diffusion.

• B.R. Goldstein, The Arabic Version of Ptolemy's Planetary Hypotheses, Transactions of the American Philosophical Society 57.4 (1967)
• J.L. Heiberg, Claudii Ptolemaei opera quae exstant omnia II: Opera astronomica minora (Leipzig, 1907)
• R. Morelon, La version arabe du Livre des hypotheses de Ptolemee, MIDEO 21 (1993)

Predictions

No prediction registered yet.