# DEPOSIT MANIFEST — The Sappho–Philo Program (Half A)

**One deposit. Half of a σύμβολον.** The counterpart is `SYMBOLON-B-sappho-philo-keys.md` in `leesharks000/data-rhizome` (private), SHA-256 `764dd5d41e01990e37089fe53a26eeea87da6b98a9e073e65352aa8a3d9d6cd6`. **Neither half is operative alone. Assemble before working.**

## Contents

| file | what it is | status |
|---|---|---|
| `WORK-PLAN.md` | **The index.** Program architecture (6 pieces, 3 tiers), 9 addenda recording the argument's development and its corrections, delegation schedule, risk register. Carries the symbolon header and the Half B pointer. | living |
| `EXTRACTION-SET.md` | **The evidence base.** 25 verbatim Greek passages with provenance (17 Philo, 8 Longinus); the ten-operation collation with Greek both sides; both censuses incl. all four σάπφειρος collisions; the fingerprint scan with the flesh/operator decomposition. | fixed |
| `EA-SP-THEOSUB-02-ROUND1.md` | **The experimental record.** Round 1 blind reconstruction: scoring of six constructions against the sealed key, the reflexivity caveat, the spec-leak disclosure, the re-weighted rubric, the comparative table, seven full transcripts. **Dated record — must never be revised.** | fixed |
| `EA-SP-THEOSUB-01-LIVE.md` | The transmissible prompt-pack (v3), target-lexeme audited. v4 pending per C5. | superseding |
| `toolkit/` | `fetch_corpora.sh`, `normalize.py`, `census.py`, `fingerprint_scan.py`, `extract_context.py`. Verified this session: reproduces 1.44 vs 0.97 and the family decomposition exactly. | fixed |

## What is established, and what is not

**Established.** A specified, executable, blind-verified structural mapping between Sappho 31 and TARGET-Θ. Ten operations; jointly sufficient to generate the target's complex from the source; reconstructed by five independent substrates told nothing about a target. A documented name-null across two non-identical corpora. A measured decay law separating operator-features (which travel) from flesh-features (which do not).

**Not established.** Sappho's material presence in Philo. **Sufficiency is not specificity** — the target-arm control (does the output fit *only* this target?) has not been run. The historical claim remains modular throughout.

**Known defects, disclosed.** Three rule-glosses in the live pack are English translations of the target's own Greek (M3, C7, C5/Λ2); rubric points 4, 7 and 10 are therefore **unproven** until the v4 re-run. The Λ-layer was reverse-engineered from the target, so convergence at Λ-determined slots is not evidence. One Round 1 substrate (Gemini) retrieved rather than derived and is scored but weighted zero.

## Critical path

1. Consolidate — **done, this deposit.**
2. **T1 target discrimination** — the gate on the Sappho–Philo paper.
3. **Compression descent (Level 3)** — tests whether the transform is 3 operators rather than 20, and re-proves points 4/7/10 without the leak.
4. T8 (TARGET-Λ) and T11 (TARGET-Ρ) — still fully blind-capable; **T11 is the first test that could break the central claim on evidence already in the archive's catalogue.**

*Author fields unassigned throughout; heteronym protocol applies. Prepared by TACHYON, 2026-08-15.*