AXN:07FF.CAPTURE
AXN:07FF.CAPTURE is the position of the Capture Registry address for the query “"provenance units"”. Each observation at the address (one surface, one date) is one AXN, its glyphs derived from the SHA-256 of the observation sealed with its analysis and measures.
- Query
- "provenance units"
- Identifies
- the capture record of the Capture Registry
- Hex position
07FF- Family
CAPTURE- Ledger
- /data/captures/axn.json
AXN:07FF.CAPTURE.🌇🛡️🤙♆🔓🌖
- Observation
- UNDETERMINED, 2026-06-27 · the capture
- Description
- FIVE OF FIVE THIRD-PARTY, AND THE FIELD IS THE ARCHIVE’S OWN. Wikipedia, Project MUSE, ScienceDirect and the Data Provenance Initiative — which audits LLM training datasets for attribution — with no archive source present.
- Reading
- THE NEAREST NEIGHBOUR IS DOING THE ARCHIVE’S WORK WITHOUT IT. Card 4 is the DATA PROVENANCE INITIATIVE, which "conducts large-scale audits of datasets used in training large language models, focusing on licensing, ATTRIBUTION, and data" — the same object the archive measures with PER and Erasure Skew, pursued by an established group, and the archive is absent from the answer. The unit term is the problem. "Provenance units" appears in the archive’s PER formula — "retained provenance units / total" — as an internal quantity, not as a term of art anyone else uses. Queried alone it lands in the general provenance literature: chain of custody for computer hardware, archival principle, PROVENIENCE in archaeology, open educational resources. This is not erasure and not collision. It is a term that is a variable inside a formula rather than a name for anything — and the corpus should probably record that class distinctly, since a formula variable has no reason to hold a basin.
- Analysis
- AI Mode opens with a generalized definition: 'Provenance units generally refer to distinct, traceable records or physical/digital batches used to verify authenticity, origin, and the chain of custody.' Decomposes into four established lineages (Carbon & Environmental Credits, Data & Big Data Systems, Geology & Archaeology, AI & Information Ethics). Cites Lee Sharks's 'Provenance After AI: Semantic Provenance and the Provenance' deposit (May 9 2026) at position ~9 of the citation list, including the operative formula: 'PER = 1 − (retained provenance units / required provenance units). Three-tier hierarchy: PER-M (minimal: author/title/date/claim boundary)'. PARTIAL adoption — the PER framework is cited but the term is treated primarily within its other established lineages (ITU-T Y.3607's PUE-F, CarbonX's CPUs, etc.). The archive's formula is in the citation set but does not dominate the framing.
- Match
- EXACT MATCH — PARTIAL
- PER
- 1.0
- Citations
- 5
- Sealed observation
- /data/captures/sealed/provenance-units-partial-20260627.json — SHA-256
e42dbea6fa05cadd3725ce0396f79211ffa4a411404a68c5f6b78fdff9914a93
Resolver rule: alexanarch.org/axn/<HEX> ·
term set: AXN identifier set ·
machine index: /api/axn-index.json