The lost poet's name drifts

Claim (verbatim)

The subhasita corpus is notorious for cross-attribution: the same verse turns up in different anthologies under different poets' names, a phenomenon Ludwik Sternbach mapped exhaustively in his concordances and in his studies of the subhasita-samgrahas as a historical source. The instability is not uniform, and its structure is diagnostic. When a verse belongs to a surviving work, a stanza of Kalidasa's Meghaduta say, the anthologists can be checked against the text and the attribution holds firm. When the source-work is gone there is nothing to check against, so the name floats: later anthologists guess, harmonize, or default to a famous author, and the verse migrates from poet to poet across the collections. Attribution conflict should therefore concentrate on exactly the verses whose works are lost, which makes the disagreement rate a proxy for drowning. This turns a textual nuisance into an instrument, because the more a verse's attribution wobbles across independent anthologies the more likely its home text is gone. The wobble itself, measured against traceability, becomes a way of counting the invisible.

Prediction clause (verbatim)

Prediction: among verses that recur in at least two major anthologies (matched via Sternbach's Maha-subhasita-samgraha concordance), the attribution-conflict rate, meaning distinct poet-names assigned to the same verse, will be at least twice as high for verses untraceable to any extant work as for verses traceable to a surviving work; the twofold conflict ratio is the deciding clause. A verse counts as traceable only on a located occurrence in a GRETIL or SARIT full text and not on a mere titular echo, anonymous attributions are set aside rather than counted as a rival name, and the test voids for coverage if fewer than 300 recurring verses resolve.

Kill-dataset (verbatim)

Kill (not yet built): Sternbach's Maha-subhasita-samgraha concordance and Descriptive Catalogue of Poets Quoted in Sanskrit Anthologies and Inscriptions for the recurring verses and their competing attributions, with traceability to an extant work determined against the GRETIL and SARIT full-text corpora.

Novelty / leakage triage

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

Sternbach's Maha-subhasita-samgraha apparatus tabulates, verse by verse, both the competing attributions and the occurrences in extant works, so the raw conflict-and-traceability data is published, and Sternbach stated repeatedly that anthology ascriptions are unreliable. What has never been run is the stratification of the conflict rate by traceability, and its direction is genuinely uncertain: the unreliability was discovered in the first place by checking anthology ascriptions of traceable verses against the extant classics and finding them wrong wholesale, so conflict may not concentrate on the untraceable at all. The twofold conditional ratio over GRETIL/SARIT traceability is therefore a live test, not a restatement.

• L. Sternbach, Maha-subhasita-samgraha (Hoshiarpur, 1974- ), apparatus of attributions and source-occurrences
• L. Sternbach, A Descriptive Catalogue of Poets Quoted in Sanskrit Anthologies and Inscriptions, 2 vols. (Wiesbaden, 1978-1980)
• L. Sternbach, Subhasita, Gnomic and Didactic Literature, A History of Indian Literature IV.1 (Wiesbaden, 1974)

Predictions

No prediction registered yet.