The gods do not count

Claim (verbatim)

Angkorian bilingual inscriptions run two languages as two instruments: Sanskrit verse faces the gods — genealogy, praise, merit-theology — while Khmer prose faces the estate: slave rosters, land bounds, rice measures, cattle. If this diglossia is functional rather than ornamental, the arithmetic of the endowment should live almost wholly on the Khmer side, because verse cannot hold audit tables and the audience of enforcement read Khmer; Sanskrit's numerals should be calendrical and doctrinal, not inventorial. The pattern, long described qualitatively since Coedes, has never been cashed as a number. Prediction: across bilingual Cambodian inscriptions of the seventh through thirteenth centuries, at least 90% of inventory numerals — counts of persons, livestock, land and rice — will stand in the Khmer-language portions, while a majority of numerals in the Sanskrit portions will be dates or doctrinal numbers, and the Khmer-side inventory share will hold above 80% in every century with at least five bilingual inscriptions (primary clause: the overall 90% Khmer-side inventory-numeral share; the verdict follows it). Exact computation: tag language sections and numeral tokens by function in the TEI-encoded Khmer epigraphic corpus and control against the printed editions. Kill: Coedes' Inscriptions du Cambodge I-VIII (1937-1966) with the DHARMA project's open TEI editions of Khmer inscriptions (github.com/erc-dharma).

Prediction clause (verbatim)

Prediction: across bilingual Cambodian inscriptions of the seventh through thirteenth centuries, at least 90% of inventory numerals — counts of persons, livestock, land and rice — will stand in the Khmer-language portions, while a majority of numerals in the Sanskrit portions will be dates or doctrinal numbers, and the Khmer-side inventory share will hold above 80% in every century with at least five bilingual inscriptions (primary clause: the overall 90% Khmer-side inventory-numeral share; the verdict follows it). Exact computation: tag language sections and numeral tokens by function in the TEI-encoded Khmer epigraphic corpus and control against the printed editions.

Kill-dataset (verbatim)

Kill: Coedes' Inscriptions du Cambodge I-VIII (1937-1966) with the DHARMA project's open TEI editions of Khmer inscriptions (github.com/erc-dharma).

Novelty / leakage triage

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

The functional Sanskrit/Khmer diglossia — Sanskrit verse for gods, genealogy and eulogy, Khmer prose for inventories of slaves, land and rice — is textbook since Coedes and central to Pollock's account of the Sanskrit cosmopolis; but no numeral-token quantification by language section and function has been published, as the conjecture itself observes.

• G. Coedes, Inscriptions du Cambodge, 8 vols (EFEO, 1937-1966)
• S. Pollock, The Language of the Gods in the World of Men: Sanskrit, Culture, and Power in Premodern India (University of California Press, 2006)

Predictions

No prediction registered yet.