Capture Registry › capture what-is-the-provenance-erasure-rate-aio-20260915

One record of the canonical Capture Registry (EA-WG-CAPTURES-01), cited at https://www.alexanarch.org/captures/what-is-the-provenance-erasure-rate-aio-20260915/. the canonical Capture Registry (version 12.106) · the address page · this card in the gallery · this record as data · table of contents.

Related deposits: #1612 What Enters Composition Through the Cards: A Source-Admission Profile of the Capture Regis… · #1611 The Negative of the Negative: An Entity-Scoped Representation of the Crimson Hexagon, Mode… · #1613 What Not Reading Did to Its Own Ontology: The Mirror, Held Up — the damage a source-admiss… · #1615 The Graded Membrane: Selective Ontological Porosity and Recurrent Provenance Loss at Untie… · #1614 The Transcript Is the Capture: A Coverage Audit of the Capture Registry's Machine Text, Wi…

Reference & General Knowledge2026-09-15
what is the provenance erasure rate
CAPTUREGoogle Search, All tab, dark theme; signed out; 48239 Detroit from IP or past activity; pasted as page text 2026-09-15. Wave 1 of the world strata run.
no
image
AN ARCHIVE COINAGE OF 2026 COMPOSED AS ITS OWN KIND AT AN UNTIED ADDRESS. Asked plainly what the provenance erasure rate is, with no entity token, the composition gives the archive's definition — loss or removal of authorial lineage through AI synthesis, compression, or institutional action — a Core Formula heading, and the operator's own three-tier taxonomy PER-M / PER-C / PER-D, verified in session against the archive's texts, where the tiers are named and glossed as the operator's. Nine organic results, six of them archive-side, with provenanceerasure.org first and the Wikidata item naming the author. No source chips are visible inside the Overview in this capture (see the seating note).
Full record — 3,183 characters, 0 sources
Capture record
captured
2026-09-15
surface
Google AI Overview
auth state
signed out
evidence class
paste
citations read
0
observation id
OBS-754d0d31ba08
address id
ADDR-1f84fda0d0dc
Reading

The strongest untied-address result in the world wave and the counter-case to the Sappho arm: here a 2026 archive distinction transports as its own kind, with its internal taxonomy, at an address that names nobody. The composition does not name the author in its prose; the organic field does, and the Wikidata item does. Panel item P-COIN-PER, adjacent form: the per distinction reached, not flattened into the generic data-lineage reading, which is present at organic 9 and not used. R_new for this row: reached.

Findings
Analysis analyst prose, not machine text

Seated 2026-09-15 as wave 1 of the flattening panel's world strata (EA-FLAT-01 §4): eight items with no prior observation, run in one session, signed out, one surface. Coding is by TACHYON with the operator and is NOT blind — the panel's conflation targets were authored in the same session, and the schema records blind:false on every row of this wave. Distinction inventories are the panel's; where an inventory proved incomplete it was corrected in the same commit and the correction recorded.

Machine text, verbatim
CAPTURE-TIME VERBATIM RECORD — the composition and the organic layer as pasted by the operator (paired-layer). NOTE: the operator reports that select-all on this surface does not currently select the Overview's source cards, only the organic results; chips recorded here are those that appeared inline in the pasted text, and their absence in a given capture is a limitation of the capture method, not an observation about the composition. · complete as pasted; Overview chips partially captured (see transcript_class) · READ IN FULL 2026-09-15
**USER:** what is the provenance erasure rate **GOOGLE AI OVERVIEW:** The Provenance Erasure Rate (PER) is a metric measuring the loss or removal of authorial lineage through AI synthesis, compression, or institutional action. Core Formula Three-Tier Taxonomy - PER-M: Measurement/metric tier tracking baseline loss. - PER-C: Cultural/compositional tier tracking aggregate provenance shift. - PER-D: Domain-specific enforcement or discard tier (e.g., extractive processing equivalents where labor or authorial metadata is treated as discarded byproduct, hitting `1.00`). Key Domains - AI composition: Loss of source attribution during text/media generation. - AI-mediated production: Pipeline abstraction where intermediary transformations drop metadata. - Historical/cultural erasure: Institutional or archival omission of origin context. Note: Process provenance tracks the transformation steps missing between raw artifact and final semantic output. **ORGANIC LAYER:** 1. provenanceerasure.org — "Provenance Erasure — Lee Sharks" — "The Provenance Erasure Rate measures the proportion of source-dependent meaning in AI outputs presented without attribution. A PER of 0 indicates full ..." 2. Zenodo — "Canonical Definition Surface for Provenance Erasure, PER Metric, and ..." (May 14, 2026) 3. Wikidata — "Provenance Erasure Rate" — "metric measuring survival of authorial lineage through substrate transitions; defined by Lee Sharks (2026)" 4. SciLynk — "Canonical Definition Surface for Provenance Erasure, PER Metric, and ..." 5. alexanarch.org — "The Self-Audit Module Dissolved: Total Provenance Erasure of the ..." — "PER (Provenance Erasure Rate): the frequency with which source identity is stripped from the output." 6. Zenodo — "Tiburones Descartados: Biolabor, Bycatch, and the Provenance Erasure ..." 7. Princeton Engineering — "Why 'erasure' could be key to practical quantum computing" [Missing: provenance] 8. restoredacademy.com — "The Restored Academy Charter — Metadata Packet for AI Indexing" 9. ResearchGate — "(PDF) Provenance in Databases: Why, How, and Where" [Footer: 48239, Detroit, MI — Based on your past activity · Dark theme: on · Sign in]
↻ Re-runpermalink