The loudest silence is Latin

Claim (verbatim)

Rome was loud, and documented its loudness everywhere except in notes. Roman comedy was substantially sung: the transmitted production notices of Terence's six plays name the tibia-player and specify which pipes accompanied each play. Horace's Carmen Saeculare was performed by a trained double choir in 17 BCE, and the commission is carved into the Acta of the Secular Games (CIL VI 32323, which names Horace). The epigraphic habit left on the order of two thousand Latin verse inscriptions in the Carmina Latina Epigraphica, many self-describing as song; emperors performed citharody in packed theatres; instruments survive physically, like the Pompeii tibiae in Naples. Latin literature survives by the shelf-metre. And the count of surviving ancient notated music with a Latin text is zero: every item in Documents of Ancient Greek Music is Greek-texted or textless, and no Latin epitaph among thousands does what the Greek Seikilos stone does at Tralles - carve the tune over the words. The mechanism: Rome imported notation-literate professionals along with the rest of Greek musical culture, and musical literacy stayed inside a Greek-speaking professional guild system, never crossing into Latin schooling, whose curriculum was grammar and rhetoric; Latin song therefore lived exclusively in performance and apprenticeship, and when the guild world dissolved, an empire's entire musical output vanished without leaving even fragments to misread. Medieval neumes added to classical verse a millennium later are reception, not survival, and are excluded by rule.

Prediction clause (verbatim)

Prediction: a language census of every pre-600-CE artefact bearing musical pitch notation - the 61 DAGM items, published supplements, and independent finds - will return exactly zero items carrying a Latin text (primary clause: the 0-of->=61 Latin count; the verdict follows it), while the attestation base on the other side counts at least 1,500 Latin verse inscriptions in the Carmina Latina Epigraphica and tibia specifications in the didascaliae of all six Terence plays; tenth-century-and-later neumations of classical Latin verse are excluded by rule as reception; the test voids only if the notation census cannot reach 60 countable items.

Kill-dataset (verbatim)

Kill: Pöhlmann-West, Documents of Ancient Greek Music (2001) with supplements for the language census, against Bücheler-Lommatzsch's Carmina Latina Epigraphica and the transmitted Terence didascaliae for the attestation base - a zero-numerator language split.

Novelty / leakage triage

already answered in the literature

The Latin zero is stated on both sides of the ledger: the standard treatments of Roman musical life - Wille's Musica Romana and Comotti's survey - register that no notated ancient music with a Latin text survives, and Pohlmann-West's roster, which claims the extant remains entire, is Greek-texted or textless throughout. The attestation base is equally published - the Carmina Latina Epigraphica run to some two thousand verse inscriptions and the transmitted Terence didascaliae specify the tibiae. The primary zero is a stated fact, not an open count.

Predictions

No prediction registered yet.