Self-Proving Documents: A Format-Provenance Fusion Architecture for Tamper-Evident Knowledge Artifacts

Every existing system for recording the provenance of creative or decision-intensive work maintains an external ledger — a data structure stored and maintained separately from the artifact whose provenance it records. Version control repositories, blockchain timestamps, C2PA manifests, and laboratory notebook databases all share this fundamental design: the proof lives somewhere other than the work. This creates three structural vulnerabilities: distribution strips provenance, ledger-artifact desynchronization, and ledger dependency. We present the Self-Proving Document — a structured text format in which the provenance chain is embedded as human-readable content within the document itself. On each save event, a save hash cryptographically binds a content hash, a decision context record, and the predecessor save hash. The resulting chain entry is appended as readable text, making the document simultaneously a container for creative work and a tamper-evident ledger of the decision process that produced it. A four-tier verification system operates on the embedded chain. We address five adversarial conditions: chain breaks, concurrent saves, unbounded chain growth, provenance laundering, and retroactive fabrication. Cognitive Rhythm Analysis computes a quantitative provenance score distinguishing genuine human-directed decision processes from fabricated ones. The architecture is model-agnostic, application-agnostic, format-agnostic, and currency-agnostic. To our knowledge, no prior system provides a document format in which the provenance ledger IS the document. Patent Support: Patent 3 — Self-Proving Document Format (34 claims, 7 independent). USPTO App# 64/022,423, filed March 30, 2026.