Music survived by burial, not by copying

Claim (verbatim)

Greek literature and Greek music came down by different roads, and the difference is countable inside one edition. Greek texts survive overwhelmingly because medieval scriptoria kept recopying them; papyrus finds are the supplement. Greek music inverts this. Of the sixty-one items in Documents of Ancient Greek Music, the medieval manuscript channel transmits only a small theory-adjacent parcel - the notated Mesomedes hymns and the exercise scraps embedded in the Anonymus Bellermanni treatise compilation - while everything else was dug up, unrolled, or read off stone: the Euripides papyri, the Berlin and Michigan and Oxyrhynchus fragments, the two great paeans carved at Delphi for the Pythaïdes of the late second century BCE, the complete little song on the Seikilos stele from Tralles. The mechanism: scriptoria copied what schools taught and readers read; notation had no reading public - by late antiquity almost nobody could sing from the page - so scores fell out of the copying stream within a generation unless embedded inside a theory treatise, the one genre whose readers needed specimens. Music reached the Middle Ages only as fossil inclusions in technical prose; the rest of the corpus exists because Egypt is dry and stone is heavy. A corollary is also countable: completeness tracks the channel - the accidental channel (papyrus) yields scraps, while the deliberate channels (stone monuments, treatise inclusions) supply the corpus's only complete pieces.

Prediction clause (verbatim)

Prediction: classifying the 61 DAGM items by carrier from the edition's source descriptions, at least 75% will be archaeologically recovered objects (papyri, ostraca, inscriptions) and at most 20% transmitted by continuous medieval copying (primary clause: the >=75-percent-archaeological / <=20-percent-manuscript channel split; the verdict follows it), with every manuscript-transmitted item traveling inside a music-theory compilation rather than a literary anthology, and at most one complete piece in the entire papyrus subcorpus while stone and manuscript together supply at least four complete pieces; the test voids if the edition describes carriers for fewer than 55 items.

Kill-dataset (verbatim)

Kill: Pöhlmann-West, Documents of Ancient Greek Music (2001) - a carrier-and-completeness tabulation over all 61 items from the edition's own source descriptions.

Novelty / leakage triage

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

The channel split is stated qualitatively in the edition - its source descriptions present the corpus as papyri, ostraca, and stone against a small manuscript parcel (the Mesomedes hymns and the Anonymus Bellermanni exercises) - and West's handbook draws the same burial-versus-copying contrast. But the pinned percentage bars and the completeness cross-tabulation are un-run arithmetic and not trivially guaranteed: how the Bellermanni exercise scraps and the other late manuscript pieces are itemized decides whether the manuscript share lands under the 20 percent line, and no published tabulation was located that counts carriers and complete pieces over all 61 items.

Predictions

No prediction registered yet.