{"path":"research/corpus-graph/readers/2026-10-10-wager-incentives-and-vision-pages-as-a-reader-returned-it.md","content":"# What a reader found to bear on eleven notes on the wager, incentives and three public pages (2026-10-10)\n\nA read-only reader (Sonnet), one of those sent through the notes no reader had yet read, after the founder\nasked for the graph's coverage to be widened (README § 12). It was given the brief in\n[../reader_brief.md](../reader_brief.md) and these 11 notes:\n\n- `docs/research/empirical-deliberation-convergence-evidence.md`\n- `docs/research/convergence-wager-typed-residues-synthesis.md`\n- `docs/research/threads-synthesis-and-gaps.md`\n- `docs/research/science-lineage-philosophy.md`\n- `docs/the-missing-layer.md`\n- `docs/research/constellation-candidates.md`\n- `docs/research/lowering-the-cost.md`\n- `docs/research/incentives-analysis.md`\n- `docs/research/societal-integration-outlook.md`\n- `docs/civilizational-vision.md`\n- `docs/analysis-and-attunement.md`\n\nIt made 134 tool calls. Its report is below, unchanged. What was kept from it, after the quoted words were\nchecked by code and both passages of each finding were opened, is in `../found.jsonl` (lines whose `where`\nnames this file).\n\n---\n\n11 of 11 notes read in full.\n\n### F1\ntarget: docs/the-missing-layer.md:29\ntarget_words: \"and every verdict the system issues is itself a challengeable claim in the graph.\"\nolder: docs/research/session26-open-rulings.md:287\nolder_words: \"a verdict is a judge's last word, and under *propose, never assert* plus the Aug 31 no-ratify ruling nothing in the graph is one\"\nkind: later-correction\nwhy: Ledger row O6, ruled 2026-09-16, retired the word verdict: under propose-never-assert nothing in the graph is a judge's last word. What the system issues is a computed reading (a cache); the attackable ground is a separate horizon claim, and the route that minted verdict-claims now answers 410. The July audit also found preservation and completeness readings are API numbers, not claims. The public sentence, and 'challengeable verdicts' two paragraphs later, say more than the record supports and use a retired word.\nstrength: strong\n\n### F2\ntarget: docs/research/constellation-candidates.md:215\ntarget_words: \"the shipped verdict-claim hybrid (`CLASSIFIES_TERMINUS`) as its first instance.\"\nolder: docs/architecture-reference.md:414\nolder_words: \"`POST /claims/{id}/terminus` — retired 2026-09-16, answers 410 pointing at the reading\"\nkind: later-correction\nwhy: Candidate 25 names the shipped verdict-claim hybrid as the first instance of scoring norms standing in the graph as challengeable claims. That route was retired on 2026-09-16: the endpoint that confirmed a kind and minted a CLASSIFIES_TERMINUS claim now answers 410 and points at a computed reading, and only three such claims were ever minted. Anyone wording or ruling on this tile should not cite the hybrid as live; the first instance has to be something else.\nstrength: strong\n\n### F3\ntarget: docs/civilizational-vision.md:88\ntarget_words: \"Most tribal warfare is over semantic confusion and information asymmetry, not genuine irreducible value differences.\"\nolder: docs/research/session26-open-rulings.md:699\nolder_words: \"the word stays on both senses with their names, and *what divides them is who gets the word* is reported as an outcome\"\nkind: narrows\nwhy: Ledger AC3, ruled 2026-10-07 (Blessing all proposals), holds that when two people want one word, showing the senses may not end the dispute: what divides them can be who gets the word, a third outcome beside it was a confusion and they disagree about the world. The sibling note on wanted words (prime-real-estate-words.md) lists freedom, the page's own example, among such words. So mostly semantic confusion, and the worldview filter reveals this, overstate how far clarifying senses dissolves a quarrel.\nstrength: strong\n\n### F4\ntarget: docs/research/constellation-candidates.md:65\ntarget_words: \"is founder-opened deliberation, NOT ratified — noted here, not claimed.\"\nolder: docs/research/parity-and-the-zone-interior.md:49\nolder_words: \"This is the substance the July 10 sacredness-brake re-scope was waiting for\"\nkind: already-answers\nwhy: The candidate's status flag calls the brake re-scope (consent-and-timing, never epistemic exemption) unratified deliberation. This addendum records the founder's sacred-domain statement of 2026-08-25 (no need for stop signs; analysis may go anywhere) as the substance the July 10 re-scope was waiting for, and the weighing design doc's addendum of that day marks it ratified in substance. The flag is stale; the tile and its status note can treat the re-scope as ruled.\nstrength: strong\n\n### F5\ntarget: docs/research/threads-synthesis-and-gaps.md:105\ntarget_words: \"The available resolution is real but written nowhere: the brake gates *who may decompose and when*, not *whether the node is decomposable in principle*,\"\nolder: docs/research/parity-and-the-zone-interior.md:49\nolder_words: \"This is the substance the July 10 sacredness-brake re-scope was waiting for\"\nkind: already-answers\nwhy: The synthesis says the way out of the depth.md contradiction is written nowhere, and wants NEVER amended with a consent carve-out. The founder's statement of 2026-08-25 (no stop signs, analysis may go anywhere, the brake keeps only consent and timing) is that resolution, ratified in substance that day. Measured too: the brake's shipped lexicon matched 0 of 1,678 claims, so the named collision was latent, and the confirmed live oracle case had another cause (dogfood run 4).\nstrength: strong\n\n### F6\ntarget: docs/research/constellation-candidates.md:47\ntarget_words: \"The founding image (pre-2011): put on glasses you'd never choose and see that the person you've argued with for years reasons with integrity.\"\nolder: docs/research/prime-real-estate-words.md:319\nolder_words: \"such words entirely: *Understanding is the infrastructure of love* and *Worldviews divide us. Our\"\nkind: narrows\nwhy: Ruling AC4 (2026-10-07, Blessing all proposals) adds a rider to the stranger test: ask whether a key word is one readers already bring a meaning of their own to, since a wanted word passes the test falsely. This note's first pass over the wall names two drafts that rest entirely on such words, this tile on love and the heart-level one (candidate 4). Nothing was changed pending a stranger's reading, so the tile cannot be treated as settled wording.\nstrength: strong\n\n### F7\ntarget: docs/research/science-lineage-philosophy.md:270\ntarget_words: \"Deliberus extends science only if it replaces the Iron Rule with an equally strict rule for public reasoning.\"\nolder: docs/research/legibility-under-power-red-team.md:33\nolder_words: \"The Strevens comparison the project leans on makes the problem sharper, not softer.\"\nkind: narrows\nwhy: The red team says the Iron Rule worked because nature adjudicates and can refuse to cooperate, while the Legibility Rule has no external oracle, so controlling the form controls the outcome. It adds that a graph-legible official record makes oral, narrative and embodied reasoning second-class by definition (the Delgamuukw case). The lineage note offers the rule as the equally strict replacement without this objection; anyone ruling on the rule needs both.\nstrength: strong\n\n### F8\ntarget: docs/civilizational-vision.md:42\ntarget_words: \"leads to a thesis about human nature: **when you decompose any value system far enough, worldviews converge.**\"\nolder: docs/research/convergence-wager-typed-residues-synthesis.md:26\nolder_words: \"decomposition converts vague weighting-disagreements into *typed, narrow, maximally legible residues* — and the wager is that\"\nkind: later-correction\nwhy: The July 5 synthesis explicitly replaced \"values converge when decomposed far enough\" with a weaker, falsifiable claim: decomposition turns weighting disagreements into typed, narrow residues (fittingness, partiality, axiom choice, risk posture) that may remain. The page still states the old thesis flat and describes a small universal bedrock of consciousness, suffering and shared humanity. It should carry the transformed claim, as the public open-problems page already does.\nstrength: strong\n\n### F9\ntarget: docs/civilizational-vision.md:92\ntarget_words: \"And research suggests most disagreement is the former, not the latter.\"\nolder: docs/research/empirical-deliberation-convergence-evidence.md:81\nolder_words: \"**A single universal number is not supported by the literature, and any doc claiming one is overreaching.**\"\nkind: contradicts\nwhy: The empirical record says a single universal figure for how much disagreement dissolves is unsupported. It finds large movement in misperception and affect, some in feasibility, very little in worldview priors, with the residue disproportionately in the last. \"Research suggests most disagreement is the former\" states the wager as a finding; the page does not cite this note. Safer wording is \"much of it\", with the residue mapped rather than assumed.\nstrength: strong\n\n### F10\ntarget: docs/research/constellation-candidates.md:96\ntarget_words: \"therefore the adversarial *structure* — diverse contributors attacking each other's premises under shared rules — must carry the load that individual open-mindedness cannot.\"\nolder: docs/research/the-group-premise-evidence.md:52\nolder_words: \"a graph is a group only when frames engage in it, and the one measurement we have shows they need not\"\nkind: already-answers\nwhy: The group-premise note finds Mercier's premise holds only under five conditions and the graph supplies three; frames that differ must be selected for, and reasons exchanged inside one frame push a group further out. The founder ruled on 2026-09-09 that the tile read \"Minds that disagree beat one mind, when every reason is on the table.\" This candidate's \"the graph is the group\" is the over-broad form that ruling replaced (a 2026-10-10 line caveats the one measurement).\nstrength: strong\n\n### F11\ntarget: docs/analysis-and-attunement.md:102\ntarget_words: \"and the two-axis voting (agree vs well-argued) is the strongest defense against cynical distance.\"\nolder: docs/research/cognitive-bias-codex-and-human-contribution.md:257\nolder_words: \"**Belief bias is the named mechanism by which that assumption would fail**\"\nkind: narrows\nwhy: The page calls two-axis voting the strongest defense against cynical distance. The bias note ranks two-axis separability the corpus's most novel and least tested assumption, with zero data, and names belief bias as the mechanism by which it would fail. The shipped design has no rigor vote at all (UX principles record Option B, agree/disagree only). The sentence names a defense that is untested and, as a vote, not built.\nstrength: medium\n\n### F12\ntarget: docs/research/convergence-wager-typed-residues-synthesis.md:99\ntarget_words: \"or decomposition routinely *widens* disagreement (the motivated-reasoning backfire regime).\"\nolder: docs/research/does-descending-change-the-descender.md:50\nolder_words: \"**reverses** — positional extremity was *higher* when citizens articulated their understanding\"\nkind: supplies-a-missing-piece\nwhy: The synthesis lists decomposition widening disagreement as the unfavorable outcome without saying how likely it is. The replication history is worse than mixed: making people explain replicated the collapse of felt understanding but not moderation; protected values shield intuitions; and in a post-conflict sample of 5,139, articulating understanding raised extremity. The backfire regime is measured, not hypothetical, and the residue typing should expect it on protected values and identity-laden cases.\nstrength: strong\n\n### F13\ntarget: docs/research/empirical-deliberation-convergence-evidence.md:99\ntarget_words: \"Deliberus is building the right instrument for the largest and most tractable slice of the problem.\"\nolder: docs/research/does-descending-change-the-descender.md:206\nolder_words: \"conversations **deploying arguments alone had no effects**, while otherwise identical conversations\"\nkind: narrows\nwhy: The empirical record concludes Deliberus is building the right instrument, resting on contact, depolarization and perception-gap results. Kalla and Broockman's three preregistered field experiments (6,869 voters) found conversations deploying arguments alone had no effect, while the same conversations with a non-judgmental exchange of narratives lasted at least four months. The evidence backs contact and attunement formats; an argument-structure instrument alone resembles the null arm unless it carries the second half.\nstrength: medium\n\n### F14\ntarget: docs/the-missing-layer.md:29\ntarget_words: \"structure is output, never input, so nobody hand-builds anything.\"\nolder: docs/research/threads-synthesis-and-gaps.md:129\nolder_words: \"It is false today at the descent layer, which is the layer where the wager is tested.\"\nkind: narrows\nwhy: The synthesis checks the slogan against the dogfood record: it holds at entry (text in, claims out, zero user structuring) but is false today at the descent layer, where the typed residue was reached by hand and each answer generated six to thirteen new items. It advises scoping the public sentence: structure is output at entry, and the descent is where the promise is still being built. The page states it unconditionally.\nstrength: medium\n\n### F15\ntarget: docs/research/lowering-the-cost.md:59\ntarget_words: \"the fan-out hinge is what identifies which divergences are load-bearing rather than incidental.\"\nolder: docs/research/dogfood-run-2-orthogonal-experiments.md:23\nolder_words: \"with QEM's additive subclaim energy over trees this size, hinge differentiation requires heavy engagement.\"\nkind: narrows\nwhy: Dogfood run 2 measured the hinge on a 17-node tree: perfectly flat when unengaged, and after uneven engagement the ranking moved the right way but one evidence item shifted it by about 0.001. Differentiation needs heavy engagement; the instrument never lied, it is just quiet. The position fingerprint ranks the divergences that matter with the hinge, from a reader's own votes, which are light engagement. On a young graph that ranking would not be reliable.\nstrength: medium\n\n### F16\ntarget: docs/research/incentives-analysis.md:116\ntarget_words: \"A rational actor is legible exactly where legibility helps and vague everywhere else.\"\nolder: docs/research/science-reasoning-infrastructure.md:115\nolder_words: \"found 96 percent positive results in standard reports but 44 percent positive results in Registered Reports\"\nkind: same-mechanism-elsewhere\nwhy: Selective legibility is the reporting problem science has measured. Standard psychology papers show 96 percent positive results against 44 percent in Registered Reports, where scope is committed before outcomes are known and publication does not depend on them. That gives a magnitude for an actor who shows only what helps, and a precedent for the defense: remove the actor's control over what appears afterward, as \"let opponents propose your premises\" does. The incentive note names no precedent.\nstrength: medium\n\n### F17\ntarget: docs/research/societal-integration-outlook.md:30\ntarget_words: \"a premise cannot go stale and a terminus cannot expire.\"\nolder: docs/research/session33-the-horizon-vocabulary.md:61\nolder_words: \"never set, compute the reading from the arrangement, refresh on change, classifier proposes structure only.\"\nkind: narrows\nwhy: The outlook says a terminus cannot expire. On 2026-09-15 the founder ruled \"The second\" on horizon kinds: never set, computed from the arrangement of claims, refreshed when it changes, classifier proposing structure only. Hand-set undated termini are gone and every reading is dated, so staleness from rearranged claims is covered. His caveat stands (a re-read can look final when it is not), and ageing of the evidence base is still unrepresented.\nstrength: medium\n\n### F18\ntarget: docs/the-missing-layer.md:23\ntarget_words: \"An infrastructure layer that audits its own distortions has no precedent at all.\"\nolder: docs/research/threads-synthesis-and-gaps.md:182\nolder_words: \"The defensible claim is narrower and stronger: no infrastructure audits its own *mediation* of the reasoning it carries\"\nkind: contradicts\nwhy: The synthesis tests this sentence against the sibling study of science's infrastructure, which calls Registered Reports science retrofitting a confession channel for method and incentives. It offers a defensible replacement: no infrastructure audits its own mediation of the reasoning it carries, since Registered Reports audit researcher incentives, not the medium's distortion. \"No precedent at all\" is stronger than the record supports and invites an easy counterexample.\nstrength: medium\n\nNoticed: Not reported for lack of slots, all read at both ends: typed-residues §5.2 \"directly counters the most underappreciated hole\" vs dogfood-run-6-israel-palestine-cross-domain.md:154 (\"an instrument built for one flattening mechanism measures nothing about the others\"); lowering-the-cost §6.1 (claim reuse \"cannot be settled\" from outside) vs auto-connect-upgrade.md:224 (253 of 435 candidates never surfaced by cosine); incentives §4.5 \"Maintenance has none\" vs ledger P13 (the 2026-09-13 ruling that confident housekeeping is the machine's job, session26-open-rulings.md:321); threads-synthesis A4 vs dogfood-run-7 §5.5 (0 of 9 Swedish sentences fired any instrument, the brake included); threads-synthesis C7's \"absent entirely\" for pragma-dialectics vs what-a-link-rests-on/outside-reading-argumentation-theory.md:72 (van Eemeren and Garssen 2026, full text read); analysis-and-attunement.md:9 (\"something was missing, the other pole\") vs conviction-and-critique.md:27 (the twenty platforms failed on input friction) and incentives §1; science-lineage \"Deliberus already embodies this through Walton schemes\" vs ledger AG1 (1 of 1,569 live links has anything under it). The founder's tile rulings that overtake constellation candidates 1, 8, 9, 14 and 26 (Show the work, Being vague is safe, Supporting is easy, You never owe us a contribution, Only you can say whether a retelling of your view is fair) are recorded only in specs/landing-redesign/constellation-map.md, outside docs/.\n"}