# Claims and their validation

Target length: under 1,300 words.

These claims preserve selected R4 pass 2. The R5 mapping did not change their empirical warrants. Dates are gateway item hours; intervals date messages, not unseen actions or edits. No model claim is validated by the number of coded spans. Return to other cases, held-out validation and the global C5 reader exercise are unavailable in this room route. The accepted R4 critic is a challenge, never a vote or a truth certificate.

| Proposition | Declared kind and primary | Evidence named | Validation of that kind | Strongest rival and decision | Status and scope |
|---|---|---|---|---|---|
| `N01`, loss and revision (`A01`) | words | `c0012` explicitly joins reported loss to a new build; `c0040` and `c0105` vary the defended extent. | Exact cited speaker words, with the still-open-task bound; no frequency claim. | The target was fixed independently before loss. The cited sentence gives loss as its context and proposes a revision, but cannot establish a general rate. | Locally supported; no proper tested-absent comparison. |
| `N02`, located proposal and reported start (`A02`) | both, words primary | `c0012` to `c0013`, `dfp4v0y` to `dfp6ejt`, `62o1bl`, 37.15 minutes. | Matching below-Guude words plus separate item clocks. `E-12` remains unknown rather than a fabricated failure. | The reply names another task or the pair is merely adjacent. Matching location distinguishes this instance; missing responses prevent a sufficiency claim. | Supported local response; completion and general recruitment unestablished. |
| `N03`, accepted correction (`A03`) | both, words primary | `c0099` to `c0100`, `dfr5pxo` to `dfr77ok`, `633yh1`, 29.2166666667 minutes; full thread checks what was being moved. | Exact assent and specific move report with dated order. `F-001` revision 2 preserves the critic's correction. | Ceremonial assent followed by unrelated action. The named task and full context support the local report. `c0088` permission is excluded as a performed move; `E-04` erasure limits control rather than refuting a different consequence. | Locally supported reported correction; final geometry and all-placer compliance unverified. |
| `N04`, invoked obligation (`A04`) | words | `c0091` states reciprocal help and non-expansion; `c0113` explicitly recalls the peace pact in an aid call. | Verbatim stated reason across the dated source items; no same-thread transition asserted. | The promise is unrelated or attributed by the analyst. The later speaker's explicit invocation decides the local reason. `c0129` incomplete rescue cannot test it. | Strong local reason; outcome and wider reach limited. |
| `N05`, closure and accounting (`A05`) | both, words primary | Fixed hour 73; `c0135` frames the FIU offer around the event ending; `c0137` recounts it. Raw `dfs76hx` at 75.48861111111111 qualifies total closure. | Participant framing and order, with `F-002` retaining the uncoded source exception. | Closure caused every recap or ended all organizing. The prospective new-venue call defeats the total-ending claim; no FIU-to-weekly mechanism is inferred. | Supported sequence; broad causal or cessation interpretation rejected. |
| `N06`, consolidation and improved access (`A06`) | words, conjectural | `c0110` through `c0112` establish consolidation only. | No consequence instance or exposure/uptake comparison validates the proposed improvement. | Only the post arrangement changed. The available record cannot decide the conjecture. | Thin and theorised; no drawn arrow and no empirical improvement claim. |

## Supporting scope claims

The final code coverage is 143/143 at receipt 6925, unit codes, population all immutable coded concepts in [0,97], with no endpoint codes. The existing R2 aftermath receipt 1370 is 0/19 distinct coded items strictly after hour 73, with no item exactly at 73. The formal count objects and source qualification are retained in account.md. The latter is not a claim of no source-level coordination; `dfs76hx` is an observed exception outside the closed coded classification. The R6 gateway does not reopen that R2 count operation.

Reported work remained is kept, carried by participant words in `c0120`, `c0121` and `c0133`. It coexists with lost work and incomplete rescue. No direct canvas observation, causal survival effect or comparison to other rooms validates it.

The framework comparison is an interpretive correspondence to the supplied Puranam paper, verified in R5, not a proof of necessity or sufficiency. `T4-001`, `T4-002` and `T4-003` primarily address division, information and local allocation; `T4-007` thinly offers nonmonetary acknowledgment. The unmapped set has functional, relational or temporal limits. Weak absence applies to unknown reward arrangements, private authority and unseen action; none is entered as zero or unsolved.

## What remains unvalidated

The claims do not establish stable membership, actual deployment of scripts, a collective learning process, general effects of praise, permanent fan return, complete feedback or successful migration after closure. The chosen model retains these uncertainties and the seven episode residues. No proper negative comparison establishes failure or tested absence of the narrowly stated mechanisms. Most propositions are claims about bounded instances, not population tendencies. Their external failure observations and boundaries remain explicit in propositions.jsonl.