A METHOD AND ITS SUPERSEDING FRAMEWORK, BOTH CITED. Card 2 states that a later paper "subsumes the prior method of immanent phenomenology as one technique within the broader framework" — the archive’s own methodological revision surfaced as a source about the method. The layer serves the earlier method and the notice of its subsumption in the same strip.

The subject is itself reflexive: card 1 describes a paper that "performs an IMMANENT PHENOMENOLOGY OF THE GOOGLE AI MODE SHARE LINK, revealing four cumulative design[ layers]". A phenomenological study of the AI Mode interface, retrieved and composed BY AI Mode, and presented as a neutral description of a work.

Google Scholar lists "EMPIRICAL PHENOMENOLOGY OF GOOGLE AI MODE" as a verified publication with a citation count. The surface’s own scholarly index carries a paper about the surface.

Card 1 is a fifth Academia TOPIC PAGE — "Comparative Poetics Research Papers" — carrying archive content, consistent with the category-residue pattern MANUS identified.

Asked for 'immanent phenomenology lee sharks', the model returns a structured and substantially accurate account of **The Infinite Tunnel: An Immanent Phenomenology of the Google AI Mode Share Link**, dated 2026 and attributed to Lee Sharks. It reproduces the four cumulative design layers under a heading it generates itself — *The Link as Door* (attributions as on-ramps selecting for indexed infrastructure), *The Reader as Circuit* (clicks converting passive attention into canonical reinforcement), *The Tunnels* (each share generating a self-replicating indexed surface), and *The Recursive Corpus Hypothesis* (AI summaries persisting at indexable URLs and potentially entering future training data) — and then three of the paper's conditions: the **Somatic Boundary Condition**, **Liturgical Reading**, and the **Falsification Condition**, correctly identified as a built-in metric testing the limits of automated self-replication. The rendering is faithful; the fidelity is the observation, not the anomaly.

Two findings follow. FIRST, RECURSION REALISED: the paper argues that AI summaries persist at indexable URLs and feed future corpora; this capture IS such a summary, generated at an indexable AI Mode surface, describing that argument. The falsification condition the paper names is being exercised by the act of reading it here. The layer has produced a canonical account of the theory of its own canonicalisation.

SECOND, AND MEASURABLE: **the archive holds this work and is absent from the citation set.** *The Infinite Tunnel* is deposit #518, AXN:0156.ARCHIVAL.🔵♌↘️✋🎶🔜, with full text on a crawlable, OAI-harvestable, ResourceSync-published surface. The model reached instead for Google Scholar (for the author link), Academia.edu (twice), Medium, and an opaque google.com/goto redirect. The attribution is correct and the provenance path routes entirely around the custodian. This is the retrieval-basin asymmetry the Observatory exists to measure: correct attribution does not imply the archive is in the basin, and a work can be well-described by the layer while its holder is invisible to it.