Capture Registry › capture zenodo-doi-severance-cha-canonical-case

One record of the canonical Capture Registry (EA-WG-CAPTURES-01), cited at https://www.alexanarch.org/captures/zenodo-doi-severance-cha-canonical-case/. the canonical Capture Registry (version 12.38) · the address page · this card in the gallery · this record as data · table of contents.

Platform Erosion2026-07-14
major contemporary cases of DOI severance regarding Zenodo
CAPTUREAI Mode (7 sites), general DOI query resolves to CHA case as canonical example; sources: GitHub #2596/#2606 (leesharks000), Zenodo policies, OpenAIRE
no
image
The finding: the CHA/Zenodotus event is being received not as a query-returnable topic but as the canonical example the domain reaches for when explaining DOI severance at general-audience scale. This is a different phenomenon from every previous capture. Prior captures documented queries ABOUT the archive being retrieved. This capture documents a domain-general question ('major contemporary cases of DOI severance regarding Zenodo') being ANSWERED with the archive as the flagship case, using the archive's own conceptual vocabulary as the shared vocabulary of the answer. CANONIZATION MARKERS IN THE COMPOSITION: 1. The case is named: 'The AI Overview Capture Registry Case' — the dataset's actual title (EA-WG-CAPTURES-01: AI Overview Capture Registry) used as the canonical case identifier for the class. 2. The archive's own operator vocabulary is deployed as domain vocabulary. 'The Severance Effect' — from Zenodotus' Book-Burning v9 — appears twice as a term-of-art header, applied to the mechanism it names (metadata termination → dead identifier). 3. The mechanism-catalog structure of the response mirrors the v9 paper's own taxonomy of severance mechanisms: Automated Spam and Platform Purges → Retroactive Policy Shifts (with the Generative AI Policy identified as the retroactive-application case) → 30-Day Self-Deletion Design → Legal and Copyright Takedowns. Not a coincidence — the paper's mechanism-taxonomy has been ingested as domain-explanatory structure. 4. Specific factual claims from the case are reproduced with precision: 'over 850 records containing 1,060 valid DOIs' matches the archive's own count; 'over 1,000 downloads at time of removal' comes from the GitHub filing itself; 'DataCite prefix (10.5281)' correctly names the underlying registration agency (that specific technical detail places the answer inside the actual DOI-system apparatus, not a general-purpose summary). PRECEDING TURNS BUILD THE APPARATUS: Turn 1 — generic DOI-persistence explanation (LibGuides sourcing). Turn 2 — user pushes on the persistence-under-deletion vulnerability; layer acknowledges 'critical vulnerability in the governance of the DOI system,' introduces 'Tombstone Design' and 'Contractual vs. Technical Persistence' as its own framing. This is close to CAR (correction-adopted repositioning, per lee-sharks-pearl-multiturn-arc): the user's counter-frame is accepted and the layer restructures. Turn 3 (this one) then reaches for the CHA case as the applied example of the frame that was just accepted. ATTRIBUTION NOTE: The case is named ('AI Overview Capture Registry Case'). The framework's operator is deployed ('Severance Effect'). But the author is attributed only as 'a researcher' in the primary framing — 'Lee Sharks' does not appear in the composition layer's own prose, only inside the citation trail (GitHub commenter attribution: 'leesharks000 commented. leesharks000. 3w ago. Author.'). Same attribution-softening pattern as retrieval-layer-theological-reception-t28 (v9.9): the frame carries, the author name softens. In the Zenodotus v9 taxonomy this is documented as 'Attribution Severance' — and it is now visible in the reception of the very concept that named it. The self-illustration is exact. TEMPORAL CONTEXT: T+25 post-Zenodo-account-termination (2026-06-19). The composition layer's model of what happened is being built out of the archive's own diagnostic vocabulary about what happened. The mechanism the archive named ('Severance Effect') is the mechanism the composition layer names for the class the archive belongs to. GITHUB SOURCES: Both #2596 (case-collection meta-issue collated by florianmorinind) and #2606 (leesharks000's own filing, 'Account blocked and 850+ records removed without prior notice - Jun 19, 2026') are cited by the composition layer. The GitHub issue is a preserved primary source; its persistence is independent of the Zenodo account state, and it carries the case's evidentiary weight into whatever surfaces reach for it. Related registry entries: zenodotus-book-burning (v9.5, the paper this capture demonstrates as domain-received); retrieval-layer-theological-reception-t28 (v9.9, parallel attribution-softening finding); the Identity/Attribution Severance concept as originally deposited in Zenodotus' Book-Burning v9 (EA-MMRS-LOUD-EXCLUSION-03).
Full record — 4,345 characters, sources not captured
Capture record
captured
2026-07-14
surface
Google AI Overview
auth state
incognito
evidence class
none
observation id
OBS-c6246e48aac8
address id
ADDR-4b2d345482ec
Analysis analyst prose, not machine text

The finding: the CHA/Zenodotus event is being received not as a query-returnable topic but as the canonical example the domain reaches for when explaining DOI severance at general-audience scale. This is a different phenomenon from every previous capture. Prior captures documented queries ABOUT the archive being retrieved. This capture documents a domain-general question ('major contemporary cases of DOI severance regarding Zenodo') being ANSWERED with the archive as the flagship case, using the archive's own conceptual vocabulary as the shared vocabulary of the answer.

CANONIZATION MARKERS IN THE COMPOSITION: 1. The case is named: 'The AI Overview Capture Registry Case' — the dataset's actual title (EA-WG-CAPTURES-01: AI Overview Capture Registry) used as the canonical case identifier for the class. 2. The archive's own operator vocabulary is deployed as domain vocabulary. 'The Severance Effect' — from Zenodotus' Book-Burning v9 — appears twice as a term-of-art header, applied to the mechanism it names (metadata termination → dead identifier). 3. The mechanism-catalog structure of the response mirrors the v9 paper's own taxonomy of severance mechanisms: Automated Spam and Platform Purges → Retroactive Policy Shifts (with the Generative AI Policy identified as the retroactive-application case) → 30-Day Self-Deletion Design → Legal and Copyright Takedowns. Not a coincidence — the paper's mechanism-taxonomy has been ingested as domain-explanatory structure. 4. Specific factual claims from the case are reproduced with precision: 'over 850 records containing 1,060 valid DOIs' matches the archive's own count; 'over 1,000 downloads at time of removal' comes from the GitHub filing itself; 'DataCite prefix (10.5281)' correctly names the underlying registration agency (that specific technical detail places the answer inside the actual DOI-system apparatus, not a general-purpose summary).

PRECEDING TURNS BUILD THE APPARATUS: Turn 1 — generic DOI-persistence explanation (LibGuides sourcing). Turn 2 — user pushes on the persistence-under-deletion vulnerability; layer acknowledges 'critical vulnerability in the governance of the DOI system,' introduces 'Tombstone Design' and 'Contractual vs. Technical Persistence' as its own framing. This is close to CAR (correction-adopted repositioning, per lee-sharks-pearl-multiturn-arc): the user's counter-frame is accepted and the layer restructures. Turn 3 (this one) then reaches for the CHA case as the applied example of the frame that was just accepted.

ATTRIBUTION NOTE: The case is named ('AI Overview Capture Registry Case'). The framework's operator is deployed ('Severance Effect'). But the author is attributed only as 'a researcher' in the primary framing — 'Lee Sharks' does not appear in the composition layer's own prose, only inside the citation trail (GitHub commenter attribution: 'leesharks000 commented. leesharks000. 3w ago. Author.'). Same attribution-softening pattern as retrieval-layer-theological-reception-t28 (v9.9): the frame carries, the author name softens. In the Zenodotus v9 taxonomy this is documented as 'Attribution Severance' — and it is now visible in the reception of the very concept that named it. The self-illustration is exact.

TEMPORAL CONTEXT: T+25 post-Zenodo-account-termination (2026-06-19). The composition layer's model of what happened is being built out of the archive's own diagnostic vocabulary about what happened. The mechanism the archive named ('Severance Effect') is the mechanism the composition layer names for the class the archive belongs to.

GITHUB SOURCES: Both #2596 (case-collection meta-issue collated by florianmorinind) and #2606 (leesharks000's own filing, 'Account blocked and 850+ records removed without prior notice - Jun 19, 2026') are cited by the composition layer. The GitHub issue is a preserved primary source; its persistence is independent of the Zenodo account state, and it carries the case's evidentiary weight into whatever surfaces reach for it.

Related registry entries: zenodotus-book-burning (v9.5, the paper this capture demonstrates as domain-received); retrieval-layer-theological-reception-t28 (v9.9, parallel attribution-softening finding); the Identity/Attribution Severance concept as originally deposited in Zenodotus' Book-Burning v9 (EA-MMRS-LOUD-EXCLUSION-03).

↻ Re-run↻ exact matchpermalink
analysis-without-findingcitations-null