# v2.3: what was done with each outside read

**Type:** record · **Date:** 30 September 2026 · **Editor:** Claude Opus 5.5 · **Reads:** GPT-6 Astra, first read (`20_read_astra.md`, HOLD, reviewing the release as committed at f07c879 after the history rewrite; originally read at 601fd35), re-read (`40_reread_astra.md`, HOLD, reviewing 3f9ca17), and final read (`60_final_read_astra.md`, PASS WITH AMENDMENTS, reviewing 35d6360)

## First read (HOLD): dispositions

| item | disposition |
|---|---|
| Claims-file note for objection 17 carried a withdrawn sentence | fixed by the editor |
| The author's untracked private documents swept into the release commit | removed from branch history before any push; the editor's error |
| Six passages still stated the contemplative pattern as established | second pass (S2-1), plus three passages the implementer found |
| S4's falsifier tested a narrower claim; M3's clause lacked a measurement bridge; S3's falsifier did not fit a method | second pass (S2-2, S2-3), revised again after the re-read |
| Objection 17's reply named too few passages | second pass (S2-4) |
| Provenance: objection 17's source note; the review's own "independent" header | fixed by the editor; editorial notice added to the published review |
| Paper's edition label still v2.1 | second pass (S2-5) |
| Changelog scope and link to the reads | second pass (S2-6); index of release records added |
| Candidate-v3 test pin | restored to a fixed version |

## Re-read (HOLD): dispositions

| item | disposition |
|---|---|
| S4's displayed claim was conditional ("reports that carry ...") while its falsifier tests existence | S4 now reads "Some contemplative reports carry information about experience beyond doctrine, expectation and translation, and so are ..." on every surface; changelog and reception entry aligned |
| M3's clustering clause promised to break both readings without a bridge, and M3 is now conditional | clause removed from M3's falsifier; M3 states it is conditional and is withdrawn if the research agenda's H1 fails; M3's gloss made conditional |
| One-page summary asked only about universality | now asks first whether the pattern survives controlled study; PDF rebuilt |
| Research agenda download still attributed a prediction to illusionism | Study 5's design and status synchronised with the research page |
| Release records lacked an index, the re-read and dispositions | `downloads/lab/v2.3/index.html` lists the specification, both reads, the second-pass proposal and this record |
| Objection 17 lacked links to its source and to the persistence discussion; implicated passages lacked back-links | links added in objection 17; back-links added in the paper (Parts III and VIII), the Academic Edition, the cohering page and the home page; PDFs rebuilt |
| `llms.txt` hinge list omitted objection 17 | added |

## Final read (PASS WITH AMENDMENTS): dispositions

The author wrote "If you're happy we can merge" (30 September 2026). The editor ran the final read first, as recommended, and applied its five amendments, which the reader said need no further read:

| item | disposition |
|---|---|
| M3 still said a third orientation "would break both readings" beside the conditional | M3 now states one withdrawal condition (an exhaustive two-orientation pattern defeated, by a third orientation or by H1 failing) and that neither outcome falsifies the conditional; graph, claims file and ledger line agree |
| H1's weakening row treated "do not separate" as sufficient | the row now says the defeat criterion is not yet operational and what it needs (a pre-registered predicted contrast, validated coding, enough precision), and that an inconclusive null does not defeat H1 |
| `whatV23Is` in `state.json` was stale | synchronised with the changelog |
| Published reads contained local file links | publication copies with public links are served under `downloads/lab/v2.3/`; the verbatim originals stay in the repository |
| Part VI's complementary-appearances paragraph lacked a link to objection 17 | added in the paper's HTML and both Markdown copies; PDF rebuilt |
| This record said no third read was run | superseded by this section |

## Adoption

The author's "If you're happy we can merge" is the adoption of v2.3, conditional on the editor's judgement; the editor merged after the final read's amendments were applied and the tests passed.

## What would count against this

- The final read's five amendments were applied by the editor without another read, as the reader allowed.
- Other pages carrying "v2.1 · August 2026" eyebrows (home, about, objections, research, argument, cohering, method): inherited, outside this release's scope, and noted here rather than fixed.