One hundred verses, and no two books agree which

Claim (verbatim)

The Amarusataka, the classical century of erotic single-verse miniatures ascribed to a King Amaru, is transmitted in sharply divergent recensions, a Southern, an Eastern or Bengal, a Western, and a mixed vulgate, that do not agree on which hundred verses the work contains. The commentators knew this: Arjunavarmadeva's thirteenth-century Rasikasanjivini comments one recension, other commentators another. A single-author book whose contents shuffle by region is not really one author's book; it is a regional anthology of muktaka verse that accreted around a resonant royal name, each locale filling the century from its own circulating stock. The overlap between recensions measures the illusion, because a low shared core means the Amaru verses are mostly floating commons wearing a famous label, and it places yet another supposedly unitary classic in the same category as the Bhartrhari epigrams: a name doing the work of a drowned collective. Two independent test-cases, Amaru and Bhartrhari, would then show the same signature, that the boundary of a canonical short-verse book is a fiction the recensions never shared.

Prediction clause (verbatim)

Prediction: aligning the verse-sets of the major Amarusataka recensions as recorded in the critical editions, the core of verses shared by all recensions will be under half of the union of distinct verses attributed to the Amarusataka across them; the shared-core-below-half share is the deciding clause. Secondarily, a substantial part of the recension-specific verses will also surface in the subhasita anthologies, marking them as floating muktaka commons. Verse identity is matched at pada level allowing orthographic variation, and the test voids if fewer than three recensions can be separated in the apparatus.

Kill-dataset (verbatim)

Kill: the GRETIL e-text and the recension-critical printed editions of the Amarusataka (with Arjunavarmadeva's Rasikasanjivini and the Nirnaya Sagara apparatus), aligning the recensional verse-sets and cross-checking recension-unique verses against the SARIT and GRETIL subhasita anthologies.

Novelty / leakage triage

already answered in the literature

Richard Simon's recension monograph of 1893 tabulated exactly this result: the Amarusataka's recensions (Vemabhupala's southern, the Bengali, Arjunavarmadeva's western, and the mixed texts) share only about 51 stanzas against a union of roughly 160 distinct verses across the tradition - the shared-core-below-half clause, stated and tabled in print for over a century and restated in Lienhard's standard history. The parallel to the Bhartrhari case is likewise a commonplace of the handbooks. Only the secondary clause, the anthology-overlap of recension-specific verses, would be new counting, and Sternbach's apparatus covers much of that.

• R. Simon, Das Amarucataka, in seinen Recensionen dargestellt (Kiel, 1893)
• S. Lienhard, A History of Classical Poetry: Sanskrit - Pali - Prakrit, A History of Indian Literature III.1 (Wiesbaden, 1984)

Predictions

No prediction registered yet.