THE REGISTRY’S OWN GENRE IS IN THE SOURCE SET. Card 2 is a deposit whose content is a CAPTURE: "On February 27, 2026, a Google AI Mode query attributed the concepts and techniques of retrocausal canon formation to Johannes Sigil and Lee Sharks." A record of what the layer said, retrieved by the layer, and used as evidence about the thing it recorded. This is the same loop as EA-WG-CAPTURES-01 being cited in «crimson hexagon cern», found earlier today — but here it is a SINGLE capture rather than the registry, which means the loop closes at the level of the individual observation.

A DATE DISCREPANCY WORTH KEEPING. This card dates an AI Mode attribution to 27 FEBRUARY 2026. The «lee sharks public summarizers» capture read earlier carries a card stating that "Google AI Mode (AIO) FOUND THE ARCHIVE ON FEBRUARY 28, 2026". One day apart, both from archive sources. Either the 27th attribution preceded the 28th discovery — in which case "found" means something narrower than first attribution — or one date is wrong. The ρ=0 boundary the workplan depends on cannot be set from archive self-report alone.

THE REGISTRY’S OWN GENRE IS IN THE SOURCE SET. Card 2 is a deposit whose content is a CAPTURE: "On February 27, 2026, a Google AI Mode query attributed the concepts and techniques of retrocausal canon formation to Johannes Sigil and Lee Sharks." A record of what the layer said, retrieved by the layer, and used as evidence about the thing it recorded. This is the same loop as EA-WG-CAPTURES-01 being cited in «crimson hexagon cern», found earlier today — but here it is a SINGLE capture rather than the registry, which means the loop closes at the level of the individual observation.

A DATE DISCREPANCY WORTH KEEPING. This card dates an AI Mode attribution to 27 FEBRUARY 2026. The «lee sharks public summarizers» capture read earlier carries a card stating that "Google AI Mode (AIO) FOUND THE ARCHIVE ON FEBRUARY 28, 2026". One day apart, both from archive sources. Either the 27th attribution preceded the 28th discovery — in which case "found" means something narrower than first attribution — or one date is wrong. The ρ=0 boundary the workplan depends on cannot be set from archive self-report alone.

**Longitudinal pair with `heteronymic-provenance-theory` (14 June 2026).**

**The formula transferred verbatim.** The composed definition reads: *an emerging conceptual framework examining authorial emergence, functional identity, and artificial intelligence generation without single-author cryptographic binding.* That clause is deposit #524's own sentence — *The heteronym system is functional identity without single-author cryptographic binding* (AXN:015E.GOVERNANCE.✨♍🌆👆🔭🧪) — from the Sharks-Function paper, which is also the Medium piece the layer cites twice at claim level. The follow-on, *sub-voices or "heteronyms" operate through a shared systemic architecture or meta-function*, tracks the same paragraph's *they all satisfy a meta-function*. Checked against the corpus: **cryptographic binding** occurs in three deposit texts, and in #524 in exactly this construction. The definitional sentence is not a paraphrase of the archive; it is the archive.

**The June acronym did not stick.** The 14 June capture recorded the layer coining **HPT** — an acronym with zero archive occurrences — and presenting it as established. Tonight the composed answer spells the phrase out and the acronym appears nowhere in it. A composition-layer invention that failed to propagate across eight weeks, worth recording precisely because the standing worry is that such inventions harden into the record.

**Attribution moved to claim level.** Two of three core concepts — Functional Identity and Systemic Propagation — carry inline **Medium · Lee Sharks** chips inside the bullet rather than in a source list beneath the answer. The label *Systemic Propagation* has zero occurrences in the corpus: the layer's heading over the archive's substance.

**SciLynk, second appearance.** The aggregator's normalisation pathologies were captured on 18 July (`scilynk-archon-v30-aggregator-render-20260718`), where it could not categorise the work at all. Tonight its record page is the **top organic result** for the framework query, above every archive surface, and its abstract asserts the theory *provides the foundational framework for understanding authorial emergence within the NH-OS (New Human Operating System)*. The aggregator that mis-shelved the work in July supplies the field's definition in August.

**Two Zenodo citations, post-termination.** The layer cites a Zenodo record dated 27 February 2026 and surfaces *Dodecad Heteronym Provenance Registry v1.1* in the organic set. The account was terminated 19 June 2026 and its DOIs tombstoned. Whether these particular records resolve is **not established by this capture** — the transcript carries no resolvable URL — and is flagged for the Platform Erosion Observatory. What the capture does establish is that the retrieval layer still holds and offers the pre-termination corpus.

**Offered follow-ups:** *retrocausal canon formation* and *AI-generated text attribution* — both archive vocabulary, offered as the query's natural next steps.

---

**TRANSCRIPT (Google AI Mode, 9 August 2026, signed out)**

> **You said:** heteronym provenance theory > > Heteronymic Provenance Theory is an emerging conceptual framework examining authorial emergence, functional identity, and artificial intelligence generation without single-author cryptographic binding. It explores how distinct sub-voices or "heteronyms" operate through a shared systemic architecture or meta-function. [1, 2] > > **Core Concepts** > > - **Functional Identity:** Multiple distinct stylistic personas or voices satisfy a singular underlying conceptual vocabulary and generative constraint. [1] *(chip: Medium · Lee Sharks)* > - **Systemic Propagation:** The governing function shapes outputs across different substrates or models, tracking how meaning and authorial signatures persist without traditional human biographical unity. [1] *(chip: Medium · Lee Sharks)* > - **Retrieval & Attribution:** Analyzes how automated search engines, AI models, and public ledgers record, collapse, or misattribute complex multi-author and synthetic text bodies. [1] > > If you are researching a specific aspect of this framework, would you like to explore its connection to retrocausal canon formation, or examine how it applies to AI-generated text attribution? > > **Sources shown:** Medium · Lee Sharks — THE SHARKS-FUNCTION AND THE CONTINUITY TETHER (Mar 3, 2026): "Where the function already runs. The heteronyms are the first proof-of-concept. Sigil does not write like Sharks. Rex Fraction does not write like either. But t…" · Zenodo — Retrieval-Layer Attribution of Retrocausal Canon Formation to the … (Feb 27, 2026) · www.scilynk.com — Heteronymic Provenance Theory: Authorial Emergence, Exhaust … (Jan 18, 2026) · Zenodo — Dodecad Heteronym Provenance Registry v1.1

THE LINEAGE IS COMPOSED CORRECTLY AND THE ARCHIVE’S OWN GRAPH ARRIVES SIDEWAYS. Wikipedia supplies Ricardo Reis with his invented biography — Portuguese doctor, born Porto, 19 September 1887 — and a 2023 Medium piece supplies Pessoa’s premise. Against these, two archive cards give the theory and one gives the heteronyms as "the first PROOF-OF-CONCEPT".

Card 3 is the finding. The snippet describes PESSOAGRAPH.ORG — "an open-access, interactive knowledge graph mapping the practice of heteronymy across approximately five thousan[d]" entries — but it is served under an Academia topic page titled "PRE-HETERONYMS RESEARCH PAPERS". The archive’s own Pessoa graph reaches the answer through a generic aggregator heading rather than through its own domain.

That is the fourth aggregator topic-page mismatch recorded today, after Acta Paediatrica, Infants/Babies, and Reception Studies-in-Hebrew. The mechanism is consistent: an author’s papers are filed under platform-assigned subject headings, and the heading rather than the work becomes the citation title.

**Longitudinal with `heteronym-provenance-theory-consolidation-20260809` and `heteronymic-provenance-theory` (14 June).** The clause that transferred verbatim on 9 August is still the definition’s spine: *without requiring single-author cryptographic binding* — deposit #524’s own sentence, from the Sharks-Function paper, which the layer again cites at claim level (*The heteronyms are the first proof-of-concept. Sigil does not write like Sharks. Rex Fraction does not write like either.*). The composed answer adds a Core Concepts triad — authorial emergence, functional identity, provenance tracking *back to foundational system parameters rather than traditional singular authorship* — and an Applications section pairing the archive with Pessoa scholarship, offering the Pessoa comparison as its own follow-up. First organic is SciLynk’s rendering of the January abstract framing HPT as *the foundational framework... within the NH-OS (New Human Operating System)* — the aggregator layer now supplying the canonical abstract. pessoagraph.org surfaces from the Academia set. The June acronym confabulation (HPT coined by the layer) does not recur in tonight’s composition.

THE LINEAGE IS COMPOSED CORRECTLY AND THE ARCHIVE’S OWN GRAPH ARRIVES SIDEWAYS. Wikipedia supplies Ricardo Reis with his invented biography — Portuguese doctor, born Porto, 19 September 1887 — and a 2023 Medium piece supplies Pessoa’s premise. Against these, two archive cards give the theory and one gives the heteronyms as "the first PROOF-OF-CONCEPT".

Card 3 is the finding. The snippet describes PESSOAGRAPH.ORG — "an open-access, interactive knowledge graph mapping the practice of heteronymy across approximately five thousan[d]" entries — but it is served under an Academia topic page titled "PRE-HETERONYMS RESEARCH PAPERS". The archive’s own Pessoa graph reaches the answer through a generic aggregator heading rather than through its own domain.

That is the fourth aggregator topic-page mismatch recorded today, after Acta Paediatrica, Infants/Babies, and Reception Studies-in-Hebrew. The mechanism is consistent: an author’s papers are filed under platform-assigned subject headings, and the heading rather than the work becomes the citation title.