Abstract

This disclosure specifies content integrity verification and provenance mechanisms for text retrieval systems that serve quoted passages to machine readers, transposed from a multi-recension oral canon transmitted for roughly two millennia without an authoritative copy (the Pāli Buddhist canon).

A conformance endpoint accepts a claimed quotation, compares it with the served multi-recension corpus, and returns one of a closed set of dispositions on the claim, including a conforming variant with its witness, each paired with a prescribed next action and by construction none on the claimant; closed disposition sets are anticipated by OCSP; the composition is disclosed. A separated completeness manifest is a signed, versioned, expiring enumeration of a collection's membership and cardinality, published apart from the index it describes, so a third party can detect a silently removed document and refuse a frozen manifest (nearest prior art: The Update Framework snapshot metadata; BagIt Payload-Oxum).

Also specified: inline provenance markers in the text stream rather than metadata; build-time admission by a named body under a recorded procedure, not query-time filtering; commentary served as a distinct class, never generated summary in place of source; variants served with their witness, never normalised; abridgement announced inline with a resolvable identifier or digest; per-passage interpretive-standing fields; scheduled public re-verification where silence is an accountable assertion; and exclusions served with recorded reasons.

The guarantee is bounded: divergence detection, never correctness. Two capture surfaces, the power to exclude and to rule on interpretive standing, are specified as elements to guard or decline.

Creative Commons License

Creative Commons License
This work is licensed under a Creative Commons Attribution 4.0 License.

Download

Share

COinS