One afternoon at Cakranagara

Claim (verbatim)

In November 1894 the Dutch army stormed Cakranagara on Lombok, and the philologist J.L.A. Brandes, attached to the expedition, carried some four hundred lontar out of the burning palace of its Balinese dynasty. That single afternoon's salvage — the Lombok collection, taken to Batavia and thence largely to Leiden — turned out to hold the sole old witness of the Desawarnana, the poem that anchors the whole modern picture of Majapahit, returned to Indonesia only in 1973. The conjecture generalizes the anecdote into a mechanism: in a recopying culture, texts of the dead political center survive at the periphery that still venerates them, concentrated in the few court libraries rich enough to keep copying the old repertoire; destroy or disperse those libraries and whole shelves of a literature ride on single seizures. Cakranagara was precisely such a peripheral hoard — a Balinese royal house on Lombok curating Majapahit-era court literature four centuries after Majapahit fell — so its catch should not be one lucky unicum but a systematic stratum: the collection should dominate the oldest-or-only witness slots for Old and Middle Javanese court texts out of all proportion to its share of the world's lontar. Prediction: working through Pigeaud's census of Javanese manuscripts, at least ten Old or Middle Javanese works — kakawin and kidung of the Majapahit court sphere beyond the Desawarnana itself — will have their oldest known witness, or their only known witness, in a codex of Brandes' 1894 Lombok haul, even though the Lombok collection is under five percent of catalogued Javanese-Balinese lontar worldwide (primary clause: the ten-works floor for oldest-or-only Lombok witnesses; the verdict follows it). Kill: Pigeaud, Literature of Java I-IV (1967-1980), whose codex descriptions record Lombok-collection provenance, read with Brandes' own expedition-era reports on the Cakranagara library.

Prediction clause (verbatim)

Prediction: working through Pigeaud's census of Javanese manuscripts, at least ten Old or Middle Javanese works — kakawin and kidung of the Majapahit court sphere beyond the Desawarnana itself — will have their oldest known witness, or their only known witness, in a codex of Brandes' 1894 Lombok haul, even though the Lombok collection is under five percent of catalogued Javanese-Balinese lontar worldwide (primary clause: the ten-works floor for oldest-or-only Lombok witnesses; the verdict follows it).

Kill-dataset (verbatim)

Kill: Pigeaud, Literature of Java I-IV (1967-1980), whose codex descriptions record Lombok-collection provenance, read with Brandes' own expedition-era reports on the Cakranagara library.

Novelty / leakage triage

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

Pigeaud's Literature of Java records Lombok-collection provenance codex by codex, and the Desawarnana's dependence on the 1894 Cakranagara haul is famous, so the materials for the test are wholly in print. But the primary clause is a tally - at least ten Old or Middle Javanese works whose oldest or only witness sits in Brandes' Lombok haul, against the collection's under-five-percent share of catalogued lontar - and no one has worked through Pigeaud's census to produce that oldest-or-only count. The disproportion is plausible and gestured at, but the ten-works floor is un-run. Adjacent.

Predictions

No prediction registered yet.