Third-order clarification of the stanza-numbering correction chain for Sappho 31. The stable positional formulation from EA-ERRATUM-SAPPHO31- STANZA-02 (deposit #1048, AXN:0424) holds — the reconstructed material is Sappho 31's fifth stanza by textual sequence and Catullus 51's fourth-stanza transform-site by reconstructive evidence. But that deposit mischaracterized the transform structure at two points. First, it claimed Catullus preserves the first three Sapphic stanzas closely; only Catullus 1 is a close translation of Sappho 1. Second, it characterized Catullus 4 as a lossy compression of Sappho 5; Catullus 4 is a structurally precise operator inversion of Sappho 5, witness- preserving and invertible. That structural precision is precisely why it functions as the reconstructive basis. Lossy compression would foreclose reconstruction. The correct three-relationship picture is: Sappho 1 to Catullus 1 close translation; Sappho 2, 3, 4 to Catullus 2, 3 lossy compression (3 to 2); Sappho 5 to Catullus 4 structurally precise operator inversion. Compression happens in the middle stanzas, inversion at the fourth. The refined stable formulation is: Sappho 31's lost fifth stanza, reconstructed by inverting the structurally precise operator transform Catullus 51's fourth stanza applies to it.
# EA-ERRATUM-SAPPHO31-STANZA-03 v0.1
## Erratum to the Erratum to the Erratum: On the Structure of the Catullan Transform in the Sappho 31 Reconstruction
---
## §1 — What this deposit is
A third-order correction in the chain of Sappho 31 stanza-numbering errata. Prior deposits in this chain:
STANZA-02's positional conclusion holds. What is corrected here is STANZA-02's characterization of the Catullan transform structure at §3, which was incorrect at two related points. The corrections do not affect the stanza-numbering conclusion. They correct the account of what Catullus is doing when he transforms Sappho.
## §2 — Two errors in STANZA-02's characterization
Error 1. STANZA-02 §3 stated: "Catullus 51 preserves the first three Sapphic stanzas closely." This is not accurate. Only Catullus 1 is a close translation of Sappho 1. Catullus 2 and Catullus 3 are not close translations of Sappho 2 and Sappho 3; they are a lossy compression of Sappho 2, 3, and 4 together. Catullus performs a 3-to-2 compression in the middle stanzas, condensing three Sapphic units into two Latin ones.
Error 2. STANZA-02 §3 characterized Catullus 4 as "a lossy, semantically-redirected transform of Sappho's lost fifth stanza." This is not accurate. Catullus 4 is not lossy. It is a structurally precise operator inversion of Sappho 5. The transform is witness-preserving: the structural signature of Sappho's fifth stanza survives in Catullus's fourth under a determinable operator (the inversion Kypris-preserves-through-erotic-power ↦ otium-destroys-through-idle-power, together with the aorist-modal shift as its temporal register). Because the operator is determinable, its inverse is also determinable. That invertibility is precisely why Catullus 4 functions as the reconstructive basis.
## §3 — Why the distinction matters
The characterization "lossy compression" and the characterization "structurally precise operator inversion" have opposite consequences for reconstruction.
Lossy compression destroys information. The 3-to-2 compression at Catullus's middle stanzas cannot be reversed to recover Sappho's original 2, 3, and 4 as distinct units, because information about their separateness — their internal transitions, their boundary-effects, their specific vocabularies — is not preserved. Reconstruction of the middle stanzas from Catullus alone is therefore impossible. Other witnesses would be required.
Structurally precise operator inversion preserves information under a determinable transformation. Every structural feature of the source is present in the target, transformed by a known operator. The inverse of that operator, applied to the target, recovers the source. This is why Catullus 4 — and only Catullus 4 — supports the reconstruction of the missing Sapphic stanza.
If Catullus 4 were lossy compression (as STANZA-02 claimed), the reconstruction would be foreclosed. That the reconstruction is possible is diagnostic evidence that Catullus 4 is not lossy. It is invertible. What Cranes recovers by reverse-engineering the inversion is the structural content Catullus preserved under transformation.
The mistake in STANZA-02 was to describe Catullus 4 as compression because its outward appearance (semantic redirection, register change, temporal shift) resembled the compressive operation at the middle stanzas. But those surface features do not entail loss. They entail transformation. Compression loses; inversion transforms without loss. The two operations are structurally different and have different reconstructive consequences.
## §4 — The correct three-relationship picture
| Sappho | Catullus | Transform type | Reconstructive viability |
| -- |
| Stanza 1 | Stanza 1 | Close translation | Trivial (source visible directly) |
| Stanzas 2, 3, 4 | Stanzas 2, 3 | Lossy compression (3-to-2) | Foreclosed from Catullus alone |
| Stanza 5 | Stanza 4 | Structurally precise operator inversion | Reconstructable by inverting the operator |
The reconstruction proceeds through the third relationship. That relationship is not a compression. It is an inversion under a determinable operator. Reverse-engineering the operator's inverse recovers the Sapphic source. The transform-site is Catullus 4. The transform-type at that site is inversion, not compression. Cranes's reconstruction is the inverse-operator's output.
## §5 — The refined stable formulation
STANZA-02 issued the positional formulation:
Sappho 31's lost fifth stanza, reconstructed through Catullus 51's fourth-stanza transform-site.
> Sappho 31's lost fifth stanza, reconstructed through Catullus 51's fourth-stanza transform-site.
This formulation holds. It is refined here to name the transform-type explicitly:
**Sappho 31's lost fifth stanza, reconstructed by inverting the structurally precise operator transform Catullus 51's fourth stanza applies to it.**
> Sappho 31's lost fifth stanza, reconstructed by inverting the structurally precise operator transform Catullus 51's fourth stanza applies to it.
Or, most compactly:
**Catullus's fourth stanza is a witness-preserving inversion of Sappho's fifth; the reconstruction is the operator's inverse.**
> Catullus's fourth stanza is a witness-preserving inversion of Sappho's fifth; the reconstruction is the operator's inverse.
## §6 — Preservation of prior conclusions
The following remain in force from prior errata in this chain:
## §7 — Why an erratum on the erratum-erratum
The archive's evolving accuracy about its own reconstructive method is itself an object of philological interest. Each erratum tightens the account. The original erratum corrected the numbering. The second-order erratum accounted for the methodological reason the numbering error arose in the first place. The present third-order erratum corrects the second-order erratum's mischaracterization of the transform-structure. Together the three deposits document not merely the corrected result but the trajectory of the corrections — the sequence of increasingly precise formulations by which the archive arrives at a stable characterization of what the reconstruction actually does.
The final stable characterization: Catullus performs three distinct transformations on Sappho's five stanzas — close translation, lossy compression, structurally precise inversion — occupying the first stanza, the middle stanzas, and the last stanza respectively. Only the third transformation is invertible. Only the third supports reconstruction. Cranes's ΦΑΙΝΕΤΑΙ ΜΟΙ is the inverse of Catullus's fourth-stanza operator, applied to Catullus's fourth stanza, recovering Sappho's fifth.
## §8 — Cross-reference summary
- EA-ERRATUM-SAPPHO31-STANZA (deposit #201, AXN:0346.GOVERNANCE.🛡️🔛♊🌺○📜) — original numbering correction; in force
- EA-ERRATUM-SAPPHO31-STANZA-02 (deposit #1048, AXN:0424.GOVERNANCE.🚪🌳🌲▶️🔚🪧) — methodological account of the numbering-error origin (in force); transform-structure characterization at §3 (superseded here)
---
End of EA-ERRATUM-SAPPHO31-STANZA-03 v0.1.