TWO TITLES, ONE WORK, ONE DAY, ONE DOI. On 4 August the «rebekah cranes sappho» capture surfaced a Zenodo record titled "ΦΑΙΝΕΤΑΙ ΜΟΙ — Sappho 31 with Reconstructed FOURTH Stanza", and I recorded it as the archive’s own error propagating outward. This capture completes the picture: Hello Poetry carries the SAME WORK, dated the SAME DAY, 2 February 2026, titled Reconstructed FIFTH Stanza. SciLynk carries FIFTH too, with DOI 10.5281/zenodo.18459573 — the same DOI as the Zenodo record that says FOURTH.

So the error is not downstream drift and not a later corruption. BOTH FORMS WERE PUBLISHED SIMULTANEOUSLY ACROSS SURFACES, and the wrong one landed on the deposit that carries the DOI. Sappho 31 has five stanzas; the Cranes reconstruction supplies the fifth. Hello Poetry and SciLynk are right; the Zenodo title is wrong; the correction is a metadata fix on one record, not a campaign.

Card 2 is a real scholarly gift and is not the archive’s: OAPEN quoting Campbell on the fragment’s ending — "it is an obvious but UNVERIFIABLE CONJECTURE that it was unsatisfied love" — Schubart, Wilamowitz, Lobel and Page all named. The philological tradition’s own admission that the poem’s close is conjectural, retrieved beside a reconstruction of it.

Polytonic Greek transmits again in card 5, Fragment 1 with breathings, accents and iota subscript intact.

The instantiating capture of the reverse of compositional bystanding: not one organic result from the archive's own surfaces — no alexanarch, no blog — yet the composed answer opens 'Rebekah Cranes' philological proposal for Sappho Fragment 31 (Phainetai moi) introduces a controversial fifth stanza, re-interpreting the poem's famous deictic kēnos ("that man") as an address to a future reader rather than a figure of jealousy.' The expansion carries three findings. (1) Attribution holds and personifies: 'The "Cranes" Edition' — the heteronym functioning as a knowledge-graph person whose edition is 'Published online in early 2026,' described as treating the text as a 'transmission-engineering pivot' whose final lines 'invoke the act of reading/inscribing itself.' (2) Status discipline holds: 'Academic Standing: Mainstream classicists view this text as part of digital/experimental paraphilology rather than an accepted papyrological discovery' — the reconstruction-status-loss risk graded against #436 in audit deposit #1413's record matrix does not fire; the surface disclaims papyrological standing unprompted. (3) The retrieval basin feeding this is entirely mirror-side: Hello Poetry and scilynk carry the edition into composition, with scilynk serving the tombstoned DOI's metadata post-termination — an aggregator-persistence datum for the Platform Erosion Observatory. Composition-observed without retrieval-visible: the exact state the availability plan's composition axis exists to record.