Five books for a city, thousands for its echo

Claim (verbatim)

Rome sang its own chant, the dialect scholars call Old Roman, and it survives in about five notated manuscripts: the graduals Vatican lat. 5319, Rome San Pietro F 22, and Bodmer 74 (the Santa Cecilia gradual of 1071), and the antiphoners Rome San Pietro B 79 and London BL Add. 29988. The Gregorian dialect that overwrote it survives in thousands. The paradox is that the metropolitan repertory of the papal city - the presumed fountainhead - is among the least written chant in Latin Europe, while the provincial, Frankish-transmitted Gregorian was copied into every choirbook from Iberia to Poland. The mechanism is that Old Roman was an oral urban practice sustained by living institutions - the basilican and monastic clergy of the city singing from memory - and it was committed to parchment late and defensively, in the eleventh century, precisely as reformed and Gregorianizing pressure threatened to erase it. Writing was a symptom of endangerment, not vitality; a repertory that everyone could hear every day in Rome had no reason to be inscribed until it was dying. So the survival record inverts the prestige: the chant of the center is a near-unicum, and the chant of the periphery is ubiquitous. Prediction restated: in the aggregating chant instruments, Old Roman is carried by fewer than ten indexed sources and by fewer than one source for every hundred that carry the cognate Gregorian item, a source-count geometry no other major Latin repertory approaches.

Prediction clause (verbatim)

Prediction: in the Cantus Index network, the number of indexed sources transmitting the Old Roman melodic tradition is under ten, and for chant texts shared between the two Roman dialects the Old-Roman-to-Gregorian source-count ratio is below 1:100 (primary clause: the sub-1:100 source-count ratio for shared items; the verdict follows it). A source counts as Old Roman only if it carries the Old Roman melodic dialect, not merely a Roman-use Gregorian book; coverage guard: because the instruments index the survivors, the ratio measures differential inscription and survival and the true oral-era disparity was larger still, so the test voids only if fewer than three Old Roman sources are indexed at all.

Kill-dataset (verbatim)

Kill: the Cantus Index cross-source concordance counts, filtering the Old Roman witnesses (Vat. lat. 5319, San Pietro F 22, Bodmer 74, San Pietro B 79, BL Add. 29988) against the Gregorian concordance breadth per shared chant ID; the operation is a source-count ratio, dialect against dialect, on the same texts.

Novelty / leakage triage

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

The corpus-size half is fully published - the survey literature lists exactly these five notated Old Roman manuscripts, so the under-ten-indexed-sources clause is guaranteed, and the witnesses have print editions and inventories (Stablein-Melnicki for Vat. lat. 5319, Lutolf for Bodmer 74, Baroffio-Kim for San Pietro B 79). But the designated primary clause is a concordance-count ratio inside the Cantus Index network, and no published work computes per-chant Old-Roman-to-Gregorian source-count ratios for shared chant identities; whether shared items clear 1:100 depends on how much of the Gregorian mass and office corpus the network indexes, and on whether three of the five Old Roman witnesses are network-indexed at all - the packet's own void condition, unresolved in the literature.

• D. Hiley, Western Plainchant: A Handbook (Oxford, 1993), Old Roman section listing the five notated sources
• B. Stablein & M. Landwehr-Melnicki (eds.), Die Gesange des altromischen Graduale Vat. lat. 5319 (Monumenta Monodica Medii Aevi 2; Kassel, 1970)
• M. Lutolf (ed.), Das Graduale von Santa Cecilia in Trastevere (Cod. Bodmer 74) (Cologny-Geneve, 1987)
• G. Baroffio & S.J. Kim (eds.), Biblioteca Apostolica Vaticana, Archivio S. Pietro B 79: Antifonario della Basilica di S. Pietro (Rome, 1995)

Predictions

No prediction registered yet.