# Claims against their evidence

Target length: under 1,400 words.

## Proposition audit

All propositions are word-carried, element-level claims with exact bottom-tier instances. Their dated instances establish local relations; none claims how often the mechanism runs. No dimension-level proposition imports a mechanism from its lower arrows. None rests wholly on theorised arrows or a from-literature mechanism. The prior's distinction between words and frequencies remains intact. Future local/global return verdicts are unavailable in this standalone run, not passing validations.

| Proposition and kind | Mechanism and named evidence | Strongest rival and what decides | Validation of the declared kind | Limit and falsifying observation |
|---|---|---|---|---|
| `N01`, words, rests on `A01`, held `F-001` | Correcting admitted misrecognition removes the respondent's stated reason to resist. `c0019`, `c0160` say old reading and new intention in the respondent's turn; full quotes and metadata in account.md. | Polite apology may leave pixels unchanged. The words settle only the stated intention; the rival remains for behaviour. | Cite resolution and full source contexts in `E-02`, `E-18`; `E-09` bounds general cooperation, `E-17` lacks the warning author's adjustment. No claim of an independent outcome test. | Acknowledgment followed by explicit same-reason opposition would challenge the mechanism. Deliberate opposed goals fall outside it. No prevalence denominator is owed for this local claim. |
| `N02`, words, rests on `A02`, held `F-002` | Expressed organizational/overview difficulty becomes a reason for a venue response. `c0258` gives the because relation; `c0095`, `c0096`, `c0270`, `c0271` distinguish practical requests, creation and offer. | Venue creation may have independent reasons; an offer may never become access. The explicit replacement reason supports an instance, not general sufficiency. | Resolved quotes and full `E-04`, `E-05`, `E-20` contexts. `E-13` and `E-15` comparisons are not testable negative instances. Negative check remains not-addressed; this is fair, not strong. | Evidence that the cited response was independent of the problem, or no proposal/creation occurred, would defeat that instance. Implementation and upkeep remain unknown for the last offer. |
| `N03`, words, rests on `A03`, held `F-003` | A collection destination supplies somewhere to send posts under `K04` and `R01`. `c0136` announces it, `c0082` reports removals, `c0199`, `c0292` are actual directions. | Repeated announcements could exaggerate enforcement. Reported removal plus concrete directions supports bounded post governance, while the Oslo mismatch limits targeting. | Resolved exemplars and dated order, not a count of obeying placers. `E-05` defeats automatic enforcement after any venue creation; `E-13` and `E-20` show routing uses. | Earlier enforcement of an unrelated destination would contradict the sequence. Voluntary venues without post authority fall outside it. No inferred pixel command. |

## Other load-bearing claims

`F-004` and `C01` are bounded attribution claims: `c0145` supplies practical planning, while `c0229` denies whole-room support and `c0197` denies official status. Their coexistence does not require or prove absence of leaders. The strongest contrary material is the timed planning and fuller raw planning claims retained in R4; command could work in particular campaigns. A supported decision/compliance chain for the same full population would change the limit. Dissent alone does not falsify Puranam's system-level goal.

`A04`, `A05`, `A06` remain excluded conjectures with no proposition. Missing whole-room acceptance, criticism-to-decision and recognition-to-continuation links are weak absences in retained public text. No absence is quantified as prevalence. `E-09` supplies a positive loss report, not proof that criticism never matters. The later rebuilding report for Slovakia remains a participant's retrospective report without a linked restraint decision. `E-04`, `E-09`, `E-10`, `E-16` retain automation, loss, consent and decision residues. Naming them does not solve them.

`F-005` is an inferred, literature-prompted comparison, not an empirically observed fifth mechanism. mapping-C3.jsonl and the verified Puranam pages distinguish division/information facets from spatial claims, standing and retrospective accounting. Allocation and reward remain thinly visible as complete arrangements. The two omitted shelf legs are disclosed in literature-C3.md and account.md. No new theory name or explanatory mechanism entered after R5. There is no words-to-pattern switch to audit, and no recalled borrowed declaration was silently verified.

## Measurements and scope

The aftermath claim uses instrument receipt 1767, preserved in organizing_end.json: 17 coordinating distinct coded items over 64 distinct coded items strictly after hour 73 in the window ending before 97, share 0.265625. Boundary item sets are empty at 73 and 97. The prior gate admits a prospective practical move in event work, discussion governance or an explicitly connected successor occasion; mixed items count once. It excludes retrospective praise/explanation alone. This validates that descriptive selection/count, not an effect of praise or a prediction of survival. The last practical item `c0271` at 93.86194444444445 is a last observation. The bracket is the prior's 0/73/97; edited OP creation times do not date later edits.

The figure's eleven supporting codes over 302 live coded concepts, [0,97), is coverage receipt 4671, with empty boundary ID sets. It measures the included-arrow register only. The zero cross-dimension support register (receipt 4674, same population and boundaries) is no estimate of absent interaction. T4 membership footprints in literature-C3.md are likewise not quantities of solved problems. No claim about most participants, typical episodes or outcome rates is made.

## Figure, table and record audit

The native image inspection confirms every drawn arrow has its proposition tag, the closure box states hour 73, and excluded arrows are not drawn. Disconnected boxes disclose incomplete explanation. The table source was read through the gateway: six columns, verbatim material, original labels, separate optional names, range including both parts of `t07`, contextual `t06` and `t09`, undimensioned `T4-005`. Figure-C5.md states the instances and omissions.

No coding file was changed. Findings and stamped histories remain append-only. R6 changes presentation and adds a final account, with no tier recut, element change or new mechanism; no dependent scientific pass or fresh critic is owed. The accepted critic's alternative remains a defensible competing partition, with the lead's disagreement recorded, not a rejected verdict. The room route has no C5 independent readers or later return in this turn. The full close checker tests the supplied local schemas and provenance; its passing result will not establish interpretive truth.

This is a partial model with descriptive residue, not a novel concept or typology. The model does not explain why one drawing survived another; it preserves task-specific reports without attributing their cause to a form of organization. No empirical canvas, atlas or unserved literature was consulted.