{"path":"research/corpus-graph/what-readers-said-that-is-no-pair-of-notes.md","content":"# What readers said that is no pair of notes\n\n*Written by `remarks.py`. Do not edit by hand: run it with `--write` after a report is filed or a remark is judged.*\n\nThe index holds a connection between two notes. This page holds what readers returned that is no such\npair: a remark from a reader asked one open question, and a finding for what does not fit whose other\nend is the code, the to-do file, the instruction file or the same note. Each is in the reader's own\nwords, with the passages it rests on. Under it stands what a session found when it opened them. A reader's\nremark is one witness: before a sentence on a page written for visitors is put to the founder, the passage\nis read in its section and its whole record is gathered ([README.md](README.md) § 15 and § 16).\n\n- Remarks from the open question: 93. Findings the index cannot hold: 43.\n- Opened by a session: 106 stand, 6 stand in part, 0 do not stand. Not yet opened: 24.\n- Ordered by the note and line each is about, pages written for visitors first.\n- `remarks.py --check` lists the quoted passages that are no longer at their lines, which is how a\n  sentence that has since been changed shows up.\n- The pilot's two open readings (four notes, 2026-10-11) were gone through before this record was kept.\n  What was done with them is in [reading-for-every-kind-of-insight.md](reading-for-every-kind-of-insight.md)\n  § 7a, and their remarks show below as not yet opened.\n\n## `docs/analysis-and-attunement.md`\n\n### `docs/analysis-and-attunement.md:21`\n\n> `docs/analysis-and-attunement.md:21` “\"Attunement\" rather than \"empathy\" because the scope is broader than emotional resonance”\n> `docs/analysis-and-attunement.md:51` “| **Architecture** | Graph nodes — discrete, categorical, committed | Embedding space — continuous, contextual, relational |”\n> `docs/analysis-and-attunement.md:116` “**It is the half that can see what the analytical half is structurally blind to**”\n\n**The reader:** \"Attunement\" names at least three different things in this note. The definition is a person's practice, active inhabitation of other worldviews. The Architecture row makes it embedding space, and a few lines later the note says embeddings \"flatten the contested distinctions that matter most\". The 27 August section makes it the human-reserved acts: naming an absence, testifying to a stake. A reader's discipline, a machine representation and a contributor's act cannot be eaten by analysis in the same way. The conviction that neither pole may eat the other can only be checked once one referent is fixed per use.\n\n**So, in the reader's words:** Decide which of the three the conviction means in each place, and keep \"attunement\" for the reader's practice.\n\n**How sure the reader was:** fairly. This is from reading; no search was needed.\n\n**Opened by session (2026-10-11): stands.** The three passages are on the page: a person's practice, the space of embeddings, and the acts kept for people. One word for three things.\n\nFrom [2026-10-11-pages-open-question-three-concept-pages-as-a-reader-returned-it.md](readers/2026-10-11-pages-open-question-three-concept-pages-as-a-reader-returned-it.md), O11.\n\n### `docs/analysis-and-attunement.md:35`\n\n> `docs/analysis-and-attunement.md:35` “3. **\"The test of every analytical feature is whether it produces an attunement insight that wasn't possible without it.\"**”\n> `docs/analysis-and-attunement.md:76` “The soul lives in the tension between them. Neither pole should be sacrificed in MVP scoping.”\n\n**The reader:** Formulations 1 to 3 each make analysis the instrument and attunement the point. The only named failure in the tests is analysis winning (\"QBAF weights nobody understands\"). Line 27 names the pure-attunement failure, but no test says when an attunement feature has eaten analysis, and only test 5 asks anything of both sides. The conviction is symmetric (\"neither may eat the other\") and line 76 repeats it. The working formulas are \"analysis serves attunement\" and scaffolding/building, where the scaffolding comes down. The symmetric claim has operational teeth on one side only.\n\n**So, in the reader's words:** Add the converse test (a feature that is warm and deepens nothing counts as attunement eating analysis), or say the ordering is deliberate.\n\n**How sure the reader was:** fairly. I read the five tests and the Habermas-machine sentence; I found no converse test.\n\n**Opened by session (2026-10-11): stands.** The five tests are on the page and none says when a feature for attunement has crowded out analysis. The reader looked for such a test and found none; not searched again here.\n\nFrom [2026-10-11-pages-open-question-three-concept-pages-as-a-reader-returned-it.md](readers/2026-10-11-pages-open-question-three-concept-pages-as-a-reader-returned-it.md), O12.\n\n### `docs/analysis-and-attunement.md:104`\n\n> `docs/analysis-and-attunement.md:104` “The difference between his hope and this architecture is that here the AI's mediation is itself measured”\n> `docs/analysis-and-attunement.md:98` “The attunement pole has the worldview filter as a design and voice-on-the-node as a proposal.”\n\n**The reader:** The Wright paragraph says the difference from his hope is that here \"the AI's mediation is itself measured — empathy infrastructure with honesty checks\". The note's own August 18 section says the attunement pole is \"a design\" and \"a proposal\", with nothing feeding analytical output to it. No lens or return edge exists in deliberus/ or web/. The honesty instruments sit on the analytical side, so what is measured is extraction and strength, not mediation of empathy. The contrast with a chat coach is real, but it is a contrast with something not built.\n\n**So, in the reader's words:** Reword the contrast to say the measured part is the analytical mediation and that attunement mediation is unbuilt and unmeasured.\n\n**How sure the reader was:** yes. A search of deliberus/ and web/src for lens, worldview and return edge found only about-page prose.\n\n**Opened by session (2026-10-11): stands.** Searched the code: the words lens and worldview occur only in page texts and test data, so nothing built carries the worldview lens. The page's contrast with Wright says the mediation here is measured.\n\nFrom [2026-10-11-pages-open-question-three-concept-pages-as-a-reader-returned-it.md](readers/2026-10-11-pages-open-question-three-concept-pages-as-a-reader-returned-it.md), O13.\n\n### `docs/analysis-and-attunement.md:114`\n\n> `docs/analysis-and-attunement.md:114` “And a person's report of what a policy costs them, from where they stand, is knowledge no amount of reading supplies.”\n> `docs/analysis-and-attunement.md:174` “it, so the attunement channel is missing from the **read path** rather than from the data. Design for”\n\n**The reader:** Line 114 says a person's report of what a policy costs them is knowledge no amount of reading supplies, and the section builds its case on that. Line 174 says the attunement channel is missing from the read path, not the data, because source_span, attributed_to and raw_text keep who said it and at what cost. Those fields hold quotes and authors from the corpus's 25 or so source texts. The canvassing arm ran on each person telling a story from their own life. coupling-the-poles.md leaves open whether recognition works \"as structure rather than from a person\". The two paragraphs disagree on whether this is a build or a recruiting problem.\n\n**So, in the reader's words:** Decide whether the return edge is only a read-path build or whether a person's own account of cost has to be asked for.\n\n**How sure the reader was:** fairly. I read coupling-the-poles.md lines 53-130; I did not inspect the stored fields.\n\n**Opened by session (2026-10-11): stands.** Read the two lines together: stored quotes and authors come from source texts, where line 114 speaks of a person's own account of a cost. The reader's count of about 25 sources is out of date (31 on 2026-08-27). The stored fields were not inspected.\n\nFrom [2026-10-11-pages-open-question-three-concept-pages-as-a-reader-returned-it.md](readers/2026-10-11-pages-open-question-three-concept-pages-as-a-reader-returned-it.md), O14.\n\n### `docs/analysis-and-attunement.md:134`\n\n> `docs/analysis-and-attunement.md:134` “The INTERSECTION — high NFC + high Openness — is the core target and the founder's own cognitive profile.”\n> `docs/convergence.md:160` “self-selects on one disposition, a thousand users from one reading community are one frame sampled a”\n> `docs/convergence.md:55` “it is a bet about bedrock *conditional on curious readers*”\n\n**The reader:** The note names its core target as high Need for Cognition plus high Openness, \"the founder's own cognitive profile\". convergence.md says N users is not N frames, that a thousand users from one reading community are one frame sampled a thousand times, and that the wager is conditional on curious readers. The intended early population is that selected community. Convergence seen among them is the kind of reading the page warns can be manufactured by who arrives, a third way beside the two it lists in the falsifiability paragraph. This note presents the profile as the aim, not as a risk.\n\n**So, in the reader's words:** Name how the first users who do not share the founder's profile will be reached, and label early convergence readings as conditional on who arrived.\n\n**How sure the reader was:** fairly. I read the three passages together; the link between them is mine.\n\n**Opened by session (2026-10-11): stands.** The three passages are on the two pages. Joining them is the reader's own step: the first users are selected on the founder's own profile, and the convergence page says many users of one kind are one frame.\n\nFrom [2026-10-11-pages-open-question-three-concept-pages-as-a-reader-returned-it.md](readers/2026-10-11-pages-open-question-three-concept-pages-as-a-reader-returned-it.md), O17.\n\n### `docs/analysis-and-attunement.md:164`\n\n> `docs/analysis-and-attunement.md:164` “while sustained exposure to opposing **reasons carried by a person** is what the canvassing result”\n> `docs/analysis-and-attunement.md:149` “conversations **deploying arguments alone had no effect**, while otherwise identical conversations”\n> `docs/convergence.md:159` “reasons-rather-than-flags is the bet, not the rebuttal. And **N users is not N frames**: if arrival”\n\n**The reader:** The note says sustained exposure to \"opposing reasons carried by a person\" is what the canvassing study measured, set against Bail's conclusions. Fifteen lines earlier it reports that arguments alone had no effect and that the durable arm was the exchange of narratives. The source note says the canvassing exposure was one ten-minute conversation (\"neither dose nor duration\") and that nobody has measured sustained reasoning-level exposure. So the contrast supports narrative over argument, not reasons over flags. convergence.md:159 rightly calls reasons-over-flags a bet.\n\n**So, in the reader's words:** Rewrite the sentence to say what the canvassing arm carried (narratives, one conversation) and keep reasons-over-flags as the untested bet.\n\n**How sure the reader was:** yes. I read does-descending-change-the-descender.md lines 196-209 and 271-284.\n\n**Opened by session (2026-10-11): stands.** Opened docs/research/does-descending-change-the-descender.md at line 271: the difference is neither dose nor duration. The canvassing arm carried narratives; line 164 of the page says sustained reasons.\n\nFrom [2026-10-11-pages-open-question-three-concept-pages-as-a-reader-returned-it.md](readers/2026-10-11-pages-open-question-three-concept-pages-as-a-reader-returned-it.md), O15.\n\n### `docs/analysis-and-attunement.md:167`\n\n> `docs/analysis-and-attunement.md:167` “tolerance-versus-participation trade-off, which this project may be industrialising):”\n\n**The reader:** The section sends the reader to the cost side, \"Mutz's tolerance-versus-participation trade-off, which this project may be industrialising\". That is ledger row I13, SUPERSEDED by I14. The cited note was corrected on 27 August after his challenge. It cites the Matthes meta-analysis (no overall relationship in either direction) and says this is \"not a standing cost\". This section is dated the same day and still carries the first claim.\n\n**So, in the reader's words:** Replace the clause with the meta-analysis result, as the cited section now does.\n\n**How sure the reader was:** yes. I read does-descending-change-the-descender.md:216-228 and session26-open-rulings.md I13 and I14.\n\n**Opened by session (2026-10-11): stands.** Opened the ledger: row I13 is superseded by I14, which records the meta-analysis that found no overall relationship. Line 167 of the page still sends the reader to the older cost.\n\nFrom [2026-10-11-pages-open-question-three-concept-pages-as-a-reader-returned-it.md](readers/2026-10-11-pages-open-question-three-concept-pages-as-a-reader-returned-it.md), O16.\n\n## `docs/architecture-reference.md`\n\n### `docs/architecture-reference.md:12`\n\n> `docs/architecture-reference.md:12` “the 8-pass extraction pipeline · shared data architecture and the Darwin port map ·”\n> `docs/architecture-reference.md:14` “visualisation · the FastAPI route inventory · the FB scraper · Claimify · the graph database · voice”\n> `docs/architecture-reference.md:350` “superseded on claim pages 2026-09-20, component kept”\n\n**The reader:** The Contents line still describes the file as it was extracted on 2026-08-29. It promises a \"Darwin port map\" that is only a bullet list. It leaves out three sections added since: Blobs (line 98), Compartments, read scope and the admin surface (line 194), and The map view (line 322). It still lists graph visualisation, which line 350 says the map view superseded. CLAUDE.md lines 222-225 copy the same list, so a session deciding from that pointer whether to open this file will not learn that admin, compartments and the map view are here. The file is now about 13,800 words against the \"~14,300 tokens\" CLAUDE.md states.\n\n**So, in the reader's words:** Rebuild the Contents (and the CLAUDE.md pointer) from the headings, and add a test that compares the two.\n\n**How sure the reader was:** yes, compared the headings with the Contents line and CLAUDE.md:220-227.\n\n**Opened by session (2026-10-11): stands.** The headings include the compartments with the admin surface, and the map view. The Contents line names neither and still lists graph visualisation, and the instruction file's pointer copies the same list.\n\nFrom [2026-10-11-pages-open-question-architecture-reference-as-a-reader-returned-it.md](readers/2026-10-11-pages-open-question-architecture-reference-as-a-reader-returned-it.md), O1.\n\n### `docs/architecture-reference.md:23`\n\n> `docs/architecture-reference.md:23` “with instructor + Pydantic + Gemini 3 Flash, orchestrated by Temporal:”\n> `docs/architecture-reference.md:27` “via Temporal parallel activities”\n> `docs/architecture-reference.md:368` “`POST /extract/stream` — same pipeline with real-time SSE progress events (parallel Pass 3a+3b)”\n> `docs/architecture-reference.md:33` “NOT yet on the Temporal/PDF path.”\n\n**The reader:** The numbered list is not the pipeline that runs. The Temporal workflow runs a pass labelled 2c_implicit (workflows.py:198) between extraction and relationships. It writes the 34 `implicit` claims and appears in the list nowhere; line 134 mentions its activity only as a registration incident. The code and line 368 call the parallel pair 3a and 3b, while the list numbers them 3 and 4. `detect_implicit_premises` is called only from the Temporal activity, and the weighing pass is stream-only (line 33). So the two paths run different passes, and community extractions, which use the stream path (line 241), get no implicit-premise pass.\n\n**So, in the reader's words:** Say beside each pass which path runs it, and add 2c, the one pass that writes premises nobody stated.\n\n**How sure the reader was:** yes. Read workflows.py and activities.py, and searched for every caller of detect_implicit_premises.\n\n**Opened by session (2026-10-11): stands.** The workflow runs a step labelled 2c_implicit that the numbered list leaves out, and the pass for unstated premises is called from the Temporal activity only.\n\nFrom [2026-10-11-pages-open-question-architecture-reference-as-a-reader-returned-it.md](readers/2026-10-11-pages-open-question-architecture-reference-as-a-reader-returned-it.md), O2.\n\n### `docs/architecture-reference.md:26`\n\n> `docs/architecture-reference.md:26` “Relationship detection + scheme classification** — supports/attacks/qualifies/reframes/decomposes_into + Walton's 96 argument schemes”\n> `deliberus/extraction/schemes.py:3` “Walton's catalogue runs to 96 schemes. This module holds a subset of them:”\n\n**The reader** (a finding for what does not fit, kind `number-or-date`): The note says the relationship pass classifies against Walton's 96 argument schemes. The code says its module holds only a subset, and `len(ArgumentScheme)` is 46 members in 11 clusters. A reader who takes 96 as the classifier's menu, for example when judging how often \"does not fit\" should fire or how much of the catalogue is covered, works from a menu twice the real size.\n\n**Which stands, by the reader:** other, by the code\n\n**Opened by session (2026-10-11): stands.** The list of schemes in the code has 46 members, and the module's own first lines say it holds a subset of Walton's 96.\n\nFrom [2026-10-11-pages-what-does-not-fit-architecture-reference-as-a-reader-returned-it.md](readers/2026-10-11-pages-what-does-not-fit-architecture-reference-as-a-reader-returned-it.md), F1.\n\n### `docs/architecture-reference.md:29`\n\n> `docs/architecture-reference.md:29` “6. **Self-evaluation** — per-claim and per-relationship checks with an `issue` and a `suggested_fix` on each”\n> `deliberus/extraction/self_eval.py:1` “Pass 4: Self-Evaluation — quality check the pipeline's own output.”\n\n**The reader** (a finding for what does not fit, kind `word-two-senses`): The note numbers contested concepts as pass 4, self-evaluation as 6 and embed plus auto-connect as 8. The code, the Temporal workflow and temporal-extraction-architecture.md call self-evaluation Pass 4, embedding Pass 5 and auto-connect Pass 6. A reader who meets \"Pass 6\" in a code comment or workflow step name and looks it up in this note lands on self-evaluation, not auto-connect.\n\n**Which stands, by the reader:** other, by the code\n\n**Opened by session (2026-10-11): stands.** The code and the workflow call self-evaluation pass 4 and the linking step pass 6. The page numbers them 6 and 8.\n\nFrom [2026-10-11-pages-what-does-not-fit-architecture-reference-as-a-reader-returned-it.md](readers/2026-10-11-pages-what-does-not-fit-architecture-reference-as-a-reader-returned-it.md), F2.\n\n### `docs/architecture-reference.md:33`\n\n> `docs/architecture-reference.md:33` “Decompose-created children get the same detection in authored mode. NOT yet on the Temporal/PDF path.”\n> `deliberus/temporal/activities.py:460` “2026-08-25, ending the divergence where the Temporal/PDF path had NEITHER”\n\n**The reader** (a finding for what does not fit, kind `state`): The note says the post-store weighing pass is not yet on the Temporal/PDF path. The code runs the shared post-store layer, weighing included, as a Temporal activity wired on 2026-08-25. The worker registers it, and tests/test_pipeline_congruence.py fails if an orchestrator drops it. A PDF extraction today gets weighing detection, so an operator reading the note would believe a gap exists that was closed.\n\n**Which stands, by the reader:** other, by the code\n\n**Opened by session (2026-10-11): stands.** The shared layer that runs after storing, weighing included, has run as a Temporal activity since 2026-08-25, by the activity's own description.\n\nFrom [2026-10-11-pages-what-does-not-fit-architecture-reference-as-a-reader-returned-it.md](readers/2026-10-11-pages-what-does-not-fit-architecture-reference-as-a-reader-returned-it.md), F3.\n\n### `docs/architecture-reference.md:41`\n\n> `docs/architecture-reference.md:41` “Direct FalkorDB via falkordb-py (scoped to `deliberus_extraction` — NEVER write to other graphs).”\n> `docs/architecture-reference.md:100` “A **blob** is a separate FalkorDB graph plus a separate Postgres database on the same servers, both”\n\n**The reader** (a finding for what does not fit, kind `counter-case`): Line 41 says the pipeline's graph client is scoped to deliberus_extraction and must never write to other graphs. By design (blob.py, store.py `_blob_graph_name`), a process started with DELIBERUS_BLOB writes to the separate graph deliberus_blob_<name>, and community input for members does exactly that. The two fit only if \"other graphs\" means graphs outside the app's own, which a reader of line 41 could not know.\n\n**Which stands, by the reader:** other, by the code\n\n**Opened by session (2026-10-11): stands.** A process started inside a compartment writes to that compartment's own graph, by design. The never on line 41 is about other projects' graphs and does not say so.\n\nFrom [2026-10-11-pages-what-does-not-fit-architecture-reference-as-a-reader-returned-it.md](readers/2026-10-11-pages-what-does-not-fit-architecture-reference-as-a-reader-returned-it.md), F4.\n\n### `docs/architecture-reference.md:43`\n\n> `docs/architecture-reference.md:43` “allt ovan om separata nycklar och isolerade fakturakonton är HISTORIA”\n> `docs/architecture-reference.md:43` “Faktureringens läge hör inte hemma i ett publicerat dokument”\n> `docs/architecture-reference.md:300` “(paused note while the LLM key is unfinanced)”\n> `docs/architecture-reference.md:45` “on the free tier (20 calls per day and model) the chain roughly triples the calls available”\n\n**The reader:** The only Swedish passage corrects something \"above\" it, namely separate keys and isolated billing accounts. Nothing above it on this page says that, because the extraction from CLAUDE.md left the correction without its target. It also says billing status does not belong in a published document, yet this file says the system runs on the free tier (line 45) and that the key is \"unfinanced\" (line 300). A visitor reading the page sees an unexplained correction in another language, and a rule the page itself breaks.\n\n**So, in the reader's words:** Delete the orphaned correction and reword lines 45 and 300 to carry no billing state, or say the rule covers only invoices.\n\n**How sure the reader was:** yes. Read lines 1-43 for any mention of keys or accounts and found none.\n\n**Opened by session (2026-10-11): stands.** The Swedish passage at line 43 corrects something above it that this page does not hold, and says the state of billing does not belong in a published document, while lines 45 and 300 give it.\n\nFrom [2026-10-11-pages-open-question-architecture-reference-as-a-reader-returned-it.md](readers/2026-10-11-pages-open-question-architecture-reference-as-a-reader-returned-it.md), O3.\n\n### `docs/architecture-reference.md:45`\n\n> `docs/architecture-reference.md:45` “All extraction passes MUST route through `llm_call` — never `client.chat.completions.create` directly.”\n> `docs/architecture-reference.md:49` “PDF reading, transcription and embeddings call their providers directly and pass through none of the three.”\n> `docs/architecture-reference.md:41` “(scoped to `deliberus_extraction` — NEVER write to other graphs)”\n\n**The reader:** Two of the file's absolutes are falsified by its own later text. \"All extraction passes MUST route through llm_call\" sits beside the PDF scout, which uses `call_genai_with_pdf`. That function has three plain retries and none of the model chain, the router or the claude engine (client.py:72-125). So the \"chain triples the calls\" arithmetic on line 45 does not apply to PDFs. \"NEVER write to other graphs\" predates blobs, which write to `deliberus_blob_<name>` (line 98). Its intent is other projects' graphs; as worded it names a noun, not a use (see J2 and J3 on over-broad rules).\n\n**So, in the reader's words:** Reword the graph rule to name the use, and mark the PDF path as outside the fallback chain wherever the quota arithmetic is given.\n\n**How sure the reader was:** yes for the PDF path (read the code). Fairly for the graph rule's intent, which I inferred.\n\n**Opened by session (2026-10-11): stands.** The PDF path calls the provider with three plain retries and none of the chain of fallback models (client.py). The rule about graphs is the same clash as finding F4 of the reading for what does not fit.\n\nFrom [2026-10-11-pages-open-question-architecture-reference-as-a-reader-returned-it.md](readers/2026-10-11-pages-open-question-architecture-reference-as-a-reader-returned-it.md), O4.\n\n### `docs/architecture-reference.md:49`\n\n> `docs/architecture-reference.md:49` “no call starts once the day's recorded spending reaches `DELIBERUS_ROUTER_DAILY_CAP_USD`”\n> `docs/architecture-reference.md:49` “with the model that served it and its cost in `~/.deliberus/router/usage.jsonl`”\n> `docs/architecture-reference.md:142` “2026-09-19 alone saw seven deploys”\n\n**The reader:** The cap's ledger is a file under the home directory, and the production container's user has none. CLAUDE.md says it is made with --no-create-home and only /app/data is writable. deploy.yml sets DELIBERUS_JUDGMENTS_DIR to /app/data for exactly that reason and sets nothing for DELIBERUS_ROUTER_DIR. With the second switch on, `_reserve`, `spent_today` and `router_dir().mkdir` would raise before the first call. Pointing it at /app/data would work, but a deploy empties it, and seven deploys happened on one day. So the $2.00 daily cap would restart with every container.\n\n**So, in the reader's words:** Before the engine is switched on in production, give the ledger a path that survives a deploy (or put it in the database), and say so in this paragraph.\n\n**How sure the reader was:** fairly. Read router_backend.py:156-213 and 338-360, the Dockerfile and deploy.yml. I did not run it in a container.\n\n**Opened by session (2026-10-11): stands.** The paid engine's record defaults to a folder under the home directory, the container's user has none, and deploy.yml sets a data folder for the judgment record only. It matters on the day that engine is switched on, which is the founder's decision.\n\nFrom [2026-10-11-pages-open-question-architecture-reference-as-a-reader-returned-it.md](readers/2026-10-11-pages-open-question-architecture-reference-as-a-reader-returned-it.md), O6.\n\n### `docs/architecture-reference.md:57`\n\n> `docs/architecture-reference.md:57` “A `print` is not a logger and is invisible in production. A bare `except Exception: return []` is a hidden capability-degradation bomb.”\n> `deliberus/extraction/cq_generation.py:184` “print(f\"    ⚠ CQ generation failed for edge {idx}: {e}\")”\n\n**The reader** (a finding for what does not fit, kind `rule-and-practice`): The note says any try-block wrapping an LLM call that falls back to an empty result must call logger.exception, because a print is invisible in production. It says this applies to every extraction-pipeline helper. CQ generation returns [] for a failed edge after the print quoted, and scout.py:112 does the same for a failed section. When a section's scout call fails in production, claims are silently lost and the log names no cause.\n\n**Which stands, by the reader:** target, by nothing I found\n\n**Opened by session (2026-10-11): stands.** Both places printed a line and returned an empty list when a model call failed, and five more handlers in the linking step did the same. Repaired the same day: all seven log through the logger, and tests/test_failures_are_logged.py reads the code for the next one.\n\nFrom [2026-10-11-pages-what-does-not-fit-architecture-reference-as-a-reader-returned-it.md](readers/2026-10-11-pages-what-does-not-fit-architecture-reference-as-a-reader-returned-it.md), F5.\n\n### `docs/architecture-reference.md:59`\n\n> `docs/architecture-reference.md:59` “(Jul 6, 2026, dogfood F5, OPEN): `/extract/stream` (the inline path) writes NO forensics row”\n\n**The reader:** The line is marked OPEN with \"Port pending\", but the stream path now writes the row. api.py's `_generate_inner` calls `save_attempt_started` and marks the attempt failed on an exception. The dogfood-run-1 friction log (line 67) lists F5 and F7 as shipped and live-verified. TODO.md line 1095 lists the same stale sentence in CLAUDE.md among \"repairs of fact for a fresh session\", but not this file. A reader of this page believes failures on the path community extractions use are invisible when they are recorded.\n\n**So, in the reader's words:** Correct this line and its cleanup-at-start claim, and add this file to the repair list in TODO.md.\n\n**How sure the reader was:** yes. Read api.py:1105-1130, the friction log and TODO.md.\n\n**Opened by session (2026-10-11): stands.** The same clash as finding F6 of the reading for what does not fit, which the index now holds: the stream path writes its row.\n\nFrom [2026-10-11-pages-open-question-architecture-reference-as-a-reader-returned-it.md](readers/2026-10-11-pages-open-question-architecture-reference-as-a-reader-returned-it.md), O5.\n\n### `docs/architecture-reference.md:61`\n\n> `docs/architecture-reference.md:61` “All source text MUST be escaped (`{{ → { {`) before passing to instructor. See `_escape_jinja()` in `scout.py` and `focused.py`.”\n> `deliberus/extraction/client.py:167` “Escape {{ }} patterns that instructor interprets as Jinja templates.”\n\n**The reader** (a finding for what does not fit, kind `number-or-date`): The note points to `_escape_jinja()` in scout.py and focused.py. No function of that name exists anywhere. Both files import the public `escape_jinja` from client.py, where it is defined. A reader who searches for the named function finds nothing and may rewrite the escape, or conclude escaping is missing.\n\n**Which stands, by the reader:** other, by the code\n\n**Opened by session (2026-10-11): stands.** No function named _escape_jinja exists. escape_jinja is defined in client.py and imported by both files.\n\nFrom [2026-10-11-pages-what-does-not-fit-architecture-reference-as-a-reader-returned-it.md](readers/2026-10-11-pages-what-does-not-fit-architecture-reference-as-a-reader-returned-it.md), F7.\n\n### `docs/architecture-reference.md:83`\n\n> `docs/architecture-reference.md:83` “Tables: `extractions` (JSONB + `input_type` + `pdf_data` BYTEA + `pdf_filename` + `owner_id`)”\n> `docs/architecture-reference.md:49` “But only three of the 25 call sites store it”\n\n**The reader:** The file never mentions `judgments.py`, the write-once record of every model call (ledger Z1/Z11, live since 2026-10-06), though CLAUDE.md carries its gotcha. The storage section names three Postgres tables. The code defines at least ten: deliberation_drafts, sameness_answers, parts_answers, step_answers, synthesis_artifacts, blob_memberships and read_events, plus the judgment tables. The three review pages write his labelling answers into some of them. Line 49's \"only three of the 25 call sites store it\" is about the served model on outputs. Read with no mention of the record, it suggests nobody knows which model answered, when the record holds exactly that.\n\n**So, in the reader's words:** Add a short section on the judgment record and a complete table list, so a session touching storage finds them here.\n\n**How sure the reader was:** yes for the omissions (rg of the file for \"judgments\" returned nothing; table names from rg for __tablename__). Fairly on how line 49 reads.\n\n**Opened by session (2026-10-11): stands.** The page has no mention of the record of machine judgments, and the code defines ten tables where the page names three.\n\nFrom [2026-10-11-pages-open-question-architecture-reference-as-a-reader-returned-it.md](readers/2026-10-11-pages-open-question-architecture-reference-as-a-reader-returned-it.md), O8.\n\n### `docs/architecture-reference.md:157`\n\n> `docs/architecture-reference.md:157` “URL validation with DNS resolution + private IP range rejection + redirect-following with per-hop validation.”\n> `docs/architecture-reference.md:264` “(an allowlist checked once is not checked at all when the first response is a 302)”\n> `docs/architecture-reference.md:265` “This matters more here than on most deployments: the app's host is also the”\n\n**The reader:** The page learned the lesson that a check must be repeated at every hop, but not its sibling. `_validate_url` resolves the hostname with getaddrinfo, and then `httpx.get(url)` resolves it again independently (api.py:608-632 and 682-693). A name that answers a public address the first time and an inside address the second passes. The page itself says [where the app runs], and that \"owner-only is not the defence\". Any signed-in Google user can submit a URL once extraction is un-paused; deploy.yml currently pauses it. The word \"rebind\" appears nowhere in docs/.\n\n**So, in the reader's words:** Before un-pausing extraction, make the fetch connect to the address that was validated, and record that beside the redirect lesson.\n\n**How sure the reader was:** fairly. Read the code and searched docs/ for \"rebind\". I did not exploit or test it.\n\n**Opened by session (2026-10-11): stands.** The check resolves the name and hands the address back unchanged, and the fetch then resolves it a second time, so a name that answers differently the second time is not checked. Extraction is paused in production, so nothing reaches this today. A fault in shipped code.\n\nFrom [2026-10-11-pages-open-question-architecture-reference-as-a-reader-returned-it.md](readers/2026-10-11-pages-open-question-architecture-reference-as-a-reader-returned-it.md), O7.\n\n### `docs/architecture-reference.md:157`\n\n> `docs/architecture-reference.md:157` “Blocks access to [the inside services, listed with their ports].”\n> `docs/architecture-reference.md:87` “[a line giving the database's address and the command that reaches it]”\n\n**The reader** (a finding for what does not fit, kind `number-or-date`): Line 157 gives the Postgres port as 5433. Lines 83 and 87 of the same note, and config/deploy.yml (`port: \"5432:5432\"`), give 5432. The SSRF check itself tests address ranges and not ports. Anyone checking by port number that the defence covers the shared Postgres would check the wrong one.\n\n**Which stands, by the reader:** other, by the code\n\n**Opened by session (2026-10-11): stands.** config/deploy.yml and lines 83 and 87 of the page give port 5432 for Postgres.\n\nFrom [2026-10-11-pages-what-does-not-fit-architecture-reference-as-a-reader-returned-it.md](readers/2026-10-11-pages-what-does-not-fit-architecture-reference-as-a-reader-returned-it.md), F9.\n\n### `docs/architecture-reference.md:158`\n\n> `docs/architecture-reference.md:158` “**Cypher injection whitelist**: `VALID_RELATIONSHIP_TYPES` set in `store.py` — only SUPPORTS/ATTACKS/QUALIFIES/REFRAMES/DECOMPOSES_INTO accepted.”\n> `deliberus/graph/store.py:660` “# (DEFINES, DECOMPOSES_INTO) carries no inference and stays unlabelled. Before”\n\n**The reader** (a finding for what does not fit, kind `number-or-date`): The note says exactly five relationship types pass the whitelist. The code's set at store.py:626 also holds DEFINES, and the store logic treats DEFINES as an ordinary stored structural edge. Someone auditing the injection surface from this note would miss a sixth accepted type.\n\n**Which stands, by the reader:** other, by the code\n\n**Opened by session (2026-10-11): stands.** The set in store.py holds DEFINES as a sixth accepted kind of link.\n\nFrom [2026-10-11-pages-what-does-not-fit-architecture-reference-as-a-reader-returned-it.md](readers/2026-10-11-pages-what-does-not-fit-architecture-reference-as-a-reader-returned-it.md), F10.\n\n### `docs/architecture-reference.md:175`\n\n> `docs/architecture-reference.md:175` “count drifts upward — trust `pytest` over this line). Coverage: SSRF (all private IP ranges)”\n> `docs/architecture-reference.md:45` “`_CAPACITY_MARKERS` in `client.py:178`”\n\n**The reader:** Small rendering and drift faults. The bold on line 175 opens and never closes (one `**` on the line), so the public page prints literal asterisks. The figure says \"800+ tests\" while pytest collects 2,600. Line 45 cites client.py:178 for `_CAPACITY_MARKERS`, which is now at 184. Lines 26 and 27 keep stray \"}\" marks from a bracket in the original that renders as noise. Line numbers in a reference rot where symbol names would not.\n\n**So, in the reader's words:** Fix the asterisk and the braces, and cite symbols instead of line numbers.\n\n**How sure the reader was:** yes. Counted the asterisks, ran pytest --collect-only, and read client.py.\n\n**Opened by session (2026-10-11): stands.** Line 175 opens bold and never closes it, and says 800+ tests where the suite holds about 2,600.\n\nFrom [2026-10-11-pages-open-question-architecture-reference-as-a-reader-returned-it.md](readers/2026-10-11-pages-open-question-architecture-reference-as-a-reader-returned-it.md), O15.\n\n### `docs/architecture-reference.md:185`\n\n> `docs/architecture-reference.md:185` “**Public endpoints**: All GET routes (claims, extractions, concepts, graph stats, docs).”\n> `docs/architecture-reference.md:190` “owner-only `/me` settings page → `GET /auth/claim-link` (401 anon, 403 non-owner) mints a 15-min HMAC-signed token”\n\n**The reader** (a finding for what does not fit, kind `counter-case`): Line 185 says all GET routes are public. The same note documents GET /auth/claim-link answering 401 to the anonymous and 403 to non-owners, and owner-only /api/admin/* GETs that answer 404. The code gates GET /claims/{id}/my-vote on sign-in. A crawler, agent or test that treats every GET as anonymous-safe gets 401 or 404 on these.\n\n**Which stands, by the reader:** other, by the code\n\n**Opened by session (2026-10-11): stands.** The page itself describes GET routes that answer 401, 403 or 404 to a visitor who is not signed in or is not the keeper.\n\nFrom [2026-10-11-pages-what-does-not-fit-architecture-reference-as-a-reader-returned-it.md](readers/2026-10-11-pages-what-does-not-fit-architecture-reference-as-a-reader-returned-it.md), F11.\n\n### `docs/architecture-reference.md:216`\n\n> `docs/architecture-reference.md:216` “settings: only a signed-in reader is named, no IP and no user agent, no query strings, and ninety”\n> `deliberus/reading.py:83` “Delete reads older than the retention window. Returns how many went.”\n\n**The reader** (a finding for what does not fit, kind `rule-and-practice`): The note lists ninety days of retention (continued on line 217) as one of four bounds on the read log that are \"refusals rather than settings\". reading.py defines `prune()` to enforce it, but nothing in the repository calls it, and the admin endpoint reports retention_days 90. Named readers' paths therefore accumulate past ninety days with no sign anywhere, a guard that never fires.\n\n**Which stands, by the reader:** other, by the code\n\n**Opened by session (2026-10-11): stands.** The function that deletes reads older than ninety days was defined and nothing called it. Repaired the same day: an expired read is never returned, and expired reads are deleted when a read is recorded (deliberus/reading.py, tests/test_reading.py). The page's sentence is now true once that is deployed.\n\nFrom [2026-10-11-pages-what-does-not-fit-architecture-reference-as-a-reader-returned-it.md](readers/2026-10-11-pages-what-does-not-fit-architecture-reference-as-a-reader-returned-it.md), F12.\n\n### `docs/architecture-reference.md:218`\n\n> `docs/architecture-reference.md:218` “Wired into the scope middleware, so it covers the whole site.”\n> `docs/architecture-reference.md:118` “Real doc reads notify from `/api/docs/{path}` with logged-in user identity when present.”\n> `docs/architecture-reference.md:286` “The members do not see the read log — a decision, stated in”\n\n**The reader:** The read log is introduced under compartments and community governance. But the middleware also records signed-in reads of the public commons (api.py:430, `_record_read(..., \"public\")`): sign-in email, path and time, kept 90 days. Line 118 adds an owner push naming who read which doc. I found no notice at sign-in or in the about page saying any of this. Line 286 calls the members' not seeing the log a decision \"stated in\" a docstring. It is his ruling of 2026-09-27 (\"Read log should only be visible to me\", shared-sensemaking-what-deliberus-holds-and-owes.md:220), which the page could cite. The docstring itself says nobody in a blob consented.\n\n**So, in the reader's words:** Decide whether signed-in readers, in the commons and in compartments, are told what is recorded about them, and cite the ruling here.\n\n**How sure the reader was:** fairly. Read api.py:420-460 and reading.py, and searched web/src and docs/ for a notice.\n\n**Opened by session (2026-10-11): stands.** The middleware records a signed-in reader's path in the public commons as well as in a compartment, and the ruling that only the keeper sees the log is his of 2026-09-27. A search of the app and the pages found no notice to readers. Whether they are told is his to decide.\n\nFrom [2026-10-11-pages-open-question-architecture-reference-as-a-reader-returned-it.md](readers/2026-10-11-pages-open-question-architecture-reference-as-a-reader-returned-it.md), O10.\n\n### `docs/architecture-reference.md:265`\n\n> `docs/architecture-reference.md:265` “This matters more here than on most deployments: the app's host is also the”\n> `docs/architecture-reference.md:119` “[a clause on the owner allow-list and the home connection]”\n> `docs/architecture-reference.md:87` “[a line giving the database's address and the command that reaches it]”\n\n**The reader:** This file moved from CLAUDE.md, which is not served, into docs/, which line 43 says is served openly. In moving it became the most infrastructure-revealing page on the site. It states [what else the production host does and how one store is exposed]. It gives [ports, names, a command line, how the owner is told apart, who shares the connection, and the sign-in provider's project], and a history of holes found and fixed (a regex bypass, a path traversal to CLAUDE.md). None of it is a secret, but together it is a guide. Line 157 also says Postgres is on 5433, against 5432 here and in deploy.yml.\n\n**So, in the reader's words:** Decide once whether operations detail belongs on the public site, and if not, split this file into a public architecture half and a private operations half.\n\n**How sure the reader was:** a guess about whether he intends this. Fairly on the facts: I read the lines, checked deploy.yml, and found no 5433 anywhere in the code.\n\n**Opened by session (2026-10-11): stands.** The lines are as quoted and the page is live on the site with them. Whether detail about the machines belongs on the site is the founder's decision; put to him on 2026-10-11.\n\nFrom [2026-10-11-pages-open-question-architecture-reference-as-a-reader-returned-it.md](readers/2026-10-11-pages-open-question-architecture-reference-as-a-reader-returned-it.md), O11.\n\n### `docs/architecture-reference.md:296`\n\n> `docs/architecture-reference.md:296` “any page under `[id]`, `[slug]`, or `[...path]` MUST use `$effect(() => { data.param; ... })` to reload data when the route param changes”\n> `docs/architecture-reference.md:119` “[a clause on the owner allow-list and the home connection]”\n> `docs/architecture-reference.md:190` “any new top-level route must be added to the `spa_fallback` decorator list in `deliberus/api.py`”\n\n**The reader:** The header says rules that must fire first stayed in CLAUDE.md and \"check there first\". Seven are duplicated there. But at least three imperative rules exist only in this reactive file, and I found no test holding them: the $effect rule for [id] pages, user_email on every notification call, and new top-level routes entering `spa_fallback`. The /admin anti-scanner trap at line 276 is a fourth, held only by an api.py docstring. Each fails quietly: stale claim data after navigation, a push to the owner's phone for his own actions, a 404 that appears only in the built app. By contrast `get_graph()` (line 149) has a test.\n\n**So, in the reader's words:** Move those rules into the resident gotchas, or give each a test that reads the code, as was done for the others.\n\n**How sure the reader was:** fairly. I searched CLAUDE.md and tests/ for each phrase and found none, but a test under other wording may exist.\n\n**Opened by session (2026-10-11): stands in part.** No test names the fallback list for new routes. Whether a test holds the other two rules under other wording was not established, as the reader also says.\n\nFrom [2026-10-11-pages-open-question-architecture-reference-as-a-reader-returned-it.md](readers/2026-10-11-pages-open-question-architecture-reference-as-a-reader-returned-it.md), O9.\n\n### `docs/architecture-reference.md:304`\n\n> `docs/architecture-reference.md:304` “and never named in this directory, which is publicly served”\n> `docs/architecture-reference.md:305` “`scripts/build_borderland_maps.py` projects a blob's own data files (the ones `scripts/store_hand_extraction.py` stores)”\n\n**The reader:** Line 304 sets the rule that unlisted pages are never named in the served docs. The next line names the unlisted community page by its working files (build_borderland_maps.py, borderlandMaps.ts, test_borderland_maps.py), and six other notes say \"borderland\" too. The literal path [the page's address] is printed nowhere in docs/, so the rule holds only in letter. The page's own header says unlisted is not private, so the harm is small. But the rule is stated as a rule, and it is already broken in seven places.\n\n**So, in the reader's words:** Rename those files to something neutral, or loosen the rule to what it can actually keep.\n\n**How sure the reader was:** a guess about whether it matters. I checked rg for [the page's address] in docs/ (none) and counted seven docs naming borderland.\n\n**Opened by session (2026-10-11): stands in part.** The same as finding F13 of the reading for what does not fit: the address itself is printed nowhere, and the working files carry the community's name.\n\nFrom [2026-10-11-pages-open-question-architecture-reference-as-a-reader-returned-it.md](readers/2026-10-11-pages-open-question-architecture-reference-as-a-reader-returned-it.md), O12.\n\n### `docs/architecture-reference.md:304`\n\n> `docs/architecture-reference.md:304` “they are never added to `robots.txt`, where a Disallow line would advertise the path, and never named in this directory, which is publicly served.”\n> `docs/architecture-reference.md:305` “**The data is generated, not typed**: `scripts/build_borderland_maps.py` projects a blob's own data files”\n\n**The reader** (a finding for what does not fit, kind `rule-and-practice`): Line 304 says unlisted pages are never named in this publicly served directory, so their path is not advertised. Line 305, in the same note, names the unlisted community page's builder, output file and test by the community's name, and other served docs name it too. The [page's path] that the noindex header and the missing robots.txt line protect is thus readable from the served docs. The code's own docstring says \"unlisted is not private\", so this may be accepted.\n\n**Which stands, by the reader:** cannot tell, by nothing I found\n\n**Opened by session (2026-10-11): stands in part.** The page's address is printed nowhere under docs/, and the test that guards this passes, so the rule as tested holds. The page's working files are named for the community on the next line and in other notes, which is what the reader means.\n\nFrom [2026-10-11-pages-what-does-not-fit-architecture-reference-as-a-reader-returned-it.md](readers/2026-10-11-pages-what-does-not-fit-architecture-reference-as-a-reader-returned-it.md), F13.\n\n### `docs/architecture-reference.md:309`\n\n> `docs/architecture-reference.md:309` “This is still a guidance/state layer, not yet full transitive mother-claim strength handling.”\n> `deliberus/graph/badge_inputs.py:316` “The full badge with the decomposition channel composed RECURSIVELY.”\n\n**The reader** (a finding for what does not fit, kind `overtaken`): The note says decomposition guidance is only a state layer, without transitive strength through mother claims. Since 2026-08-22 `compute_badge_recursive` composes the decomposition channel recursively: examined children enter their parent at their own computed badge, to depth 8. The note's own hinge route (line 419) and requirement route (line 416) rely on this. The sentence carries no date or mark, so a reader would think a parent ignores its grandchildren.\n\n**Which stands, by the reader:** other, by the code\n\n**Opened by session (2026-10-11): stands.** The badge has composed the part-of channel recursively since 2026-08-22 (compute_badge_recursive). The sentence carries no date.\n\nFrom [2026-10-11-pages-what-does-not-fit-architecture-reference-as-a-reader-returned-it.md](readers/2026-10-11-pages-what-does-not-fit-architecture-reference-as-a-reader-returned-it.md), F14.\n\n### `docs/architecture-reference.md:527`\n\n> `docs/architecture-reference.md:527` “The six other kinds join a claim to something else: EXTRACTED_FROM (6,244, claim to source),”\n> `docs/architecture-reference.md:494` “| Source | 35 | where claims were taken from |”\n\n**The reader:** I ran a read-only count on the live graph. Of the 6,244 EXTRACTED_FROM links, 4,052 run from cq_positive and cq_negative claims, the possible answers nobody gave. The rest: 2,050 from extracted claims, 84 from reports, 35 from human claims and 23 from implicit ones. So about two thirds of the arrows named \"extracted from\" lead to a source that never said the claim. The table calls Source \"where claims were taken from\" and the edge \"claim to source\". This is the same hazard the page warns of for machine_step. Any reader that follows the edge to say what a source holds picks up scaffolding unless it filters by kind.\n\n**So, in the reader's words:** Label the edge honestly in this page and the glossary (or rename it), and list its readers in a test as was done for machine_step readers.\n\n**How sure the reader was:** yes for the counts (ro_query, 2026-10-11). A guess that any reader currently misuses it.\n\n**Opened by session (2026-10-11): stands.** Counted again on the live graph on 2026-10-11: of 6,244 such links 4,052 run from the two possible answers to critical questions, 2,050 from extracted claims, 84 from reports, 35 from people's claims and 23 from unstated premises.\n\nFrom [2026-10-11-pages-open-question-architecture-reference-as-a-reader-returned-it.md](readers/2026-10-11-pages-open-question-architecture-reference-as-a-reader-returned-it.md), O13.\n\n### `docs/architecture-reference.md:536`\n\n> `docs/architecture-reference.md:536` “a strength, the machine's reason, who or what drew it, whether a person refused it”\n> `docs/architecture-reference.md:331` “is link strength; what a claim hangs on and where an answer would count are meant for emphasis”\n> `docs/architecture-reference.md:336` “written as marked parts (`provenance='split'`, `confirmed` left unset, the verdict on the edge)”\n\n**The reader:** As a description of what is built, the page is right. But the section is dated 2026-10-10 and lists `strength` and `confirmed` as plain fields of a link. It says the map picks its boxes by link strength, with none of the \"ruled, not built\" marks it puts on other pending matters. On 2026-09-28 he ruled Q31 (a link's strength is a reading, never a stored value; 174 public links carry a flat 0.7) and Q32 (retire `confirmed`). The build items are open (TODO.md:1312), and claim_map.py:37-65 still sorts by the stored number. A session reading here to build a feature will build on both.\n\n**So, in the reader's words:** Put a one-line \"ruled to become a reading, not yet built\" beside each of these places.\n\n**How sure the reader was:** fairly. Compared with settled items Q31 and Q32, claim_map.py and TODO.md.\n\n**Opened by session (2026-10-11): stands.** Ledger rows Q31 and Q32 are ruled and their builds are open. The map orders its boxes by the stored number (web/src/lib/claimMap.ts; the reader named another file).\n\nFrom [2026-10-11-pages-open-question-architecture-reference-as-a-reader-returned-it.md](readers/2026-10-11-pages-open-question-architecture-reference-as-a-reader-returned-it.md), O14.\n\n### `docs/architecture-reference.md:602`\n\n> `docs/architecture-reference.md:602` “Transcribed (168KB). Raw output at `data/voice_memos/Robin + Martin; Deliberus.txt`. Analysis pending — massive primary source”\n> `TODO.md:3835` “Remaining 5+ hrs is Whisper hallucination from quiet audio. No new Deliberus design insights beyond existing docs.”\n\n**The reader** (a finding for what does not fit, kind `overtaken`): The note presents the 5.5-hour Robin and Martin transcript as a massive primary source awaiting analysis. TODO.md records that it was reviewed: only the first ~12 minutes are real content and the remaining hours are Whisper hallucination from quiet audio. A reader who plans analysis of 168KB of transcript would spend effort on text that is mostly invented.\n\n**Which stands, by the reader:** other, by a measurement\n\n**Opened by session (2026-10-11): stands.** TODO.md records the review of that transcript: about twelve minutes are real and the remaining hours were invented by the transcriber from quiet audio.\n\nFrom [2026-10-11-pages-what-does-not-fit-architecture-reference-as-a-reader-returned-it.md](readers/2026-10-11-pages-what-does-not-fit-architecture-reference-as-a-reader-returned-it.md), F15.\n\n## `docs/bridging.md`\n\n### `docs/bridging.md:19`\n\n> `docs/bridging.md:19` “Creates cognitive tension — the most productive state for genuine opinion change”\n> `docs/bridging.md:77` “Broockman and Kalla's deep canvassing research adds a nuance: attitude change at group boundaries requires relational engagement, not just logical structure.”\n> `docs/bridging.md:73` “James Fishkin's Deliberative Polling research shows ~70% of participants change positions on at least one major issue”\n\n**The reader:** The table's bold claim, that a well-reasoned argument for a conclusion you reject is \"the most productive state for genuine opinion change\", carries no citation. The study the page does cite for deep canvassing is, in the corpus's own reading (analysis-and-attunement.md:144, coupling-the-poles.md:48), a null for argument-only conversations across 6,869 voters. A four-month effect appeared only once narrative was added, and TODO.md:2908 puts it as \"We built the null arm\". Here it is a \"nuance\". Fishkin's ~70% is also a one-sitting figure, the timescale objection he already ruled correct (I1). Conversation-starter is the right frame, but the 2x2 headline claims the argument-only mechanism.\n\n**So, in the reader's words:** Replace the bold sentence with what is known (tension may be necessary, not sufficient) and let the conversation-starter framing carry the page.\n\n**How sure the reader was:** fairly. Three corpus notes report the Kalla and Broockman result the same way. I did not read the study.\n\n**Not yet opened by a session.** Its quoted words were checked by code when the report was filed.\n\nFrom [2026-10-11-pilot-open-question-four-notes-as-a-reader-returned-it.md](readers/2026-10-11-pilot-open-question-four-notes-as-a-reader-returned-it.md), O9.\n\n### `docs/bridging.md:30`\n\n> `docs/bridging.md:30` “found high cross-group agreement on AI-generated bridging statements — reasoning structured to survive both groups' scrutiny.”\n> `TODO.md:3627` “reframe **Blair et al. / Habermas Machine / Konya** as *statement-level* (not argument-structure) support.”\n\n**The reader** (a finding for what does not fit, kind `other`): Bridging.md presents the Procaccia and Konya result as showing reasoning structured to survive both groups' scrutiny, and uses it as the stake-test for argument-level bridging. The open TODO action says Konya, Blair and the Habermas Machine are statement-level support, not argument-structure, and the page has not been reframed. The page's own line 5 says statement-level agreement is what Polis already finds, so the study cannot carry the claim about reasoning.\n\n**Which stands, by the reader:** cannot tell, by nothing I found\n\n**Opened by session (2026-10-11): stands.** Both ends were read on 2026-10-11 (the reading note's § 7a). Its other end is the code, the to-do file, the instruction file or the same note, so the index does not hold it.\n\nFrom [2026-10-11-pilot-what-does-not-fit-four-notes-as-a-reader-returned-it.md](readers/2026-10-11-pilot-what-does-not-fit-four-notes-as-a-reader-returned-it.md), F7.\n\n### `docs/bridging.md:56`\n\n> `docs/bridging.md:56` “Where disagreement is high (50/50 vote split on the conclusion) BUT QBAF strength is also high”\n> `deliberus/feed.py:358` “WITH c, COUNT(DISTINCT agree) AS agrees, COUNT(DISTINCT disagree) AS disagrees”\n> `deliberus/feed.py:365` “WITH c, agrees, disagrees, total, agree_ratio, AVG(r.strength) AS avg_strength,”\n> `docs/research/relevance-in-the-corpus-sweep.md:116` “The stored machine number is renamed, given a rubric, and left blank wherever no machine judged the link”\n> `docs/research/what-confirmed-means.md:8` “just like the whole thing with voting etc”\n\n**The reader:** The bridging score multiplies a vote split by the average of the stored link number. The feed counts claims with two or more votes, so one agree and one disagree already gives the maximum disagreement factor. The quality term is the stored number. The sweep note describes a computed reading as the replacement, and the feed still reads the stored number. The bridging note calls closing this gap a founder decision, and the founder has also questioned voting as a signal.\n\n**So, in the reader's words:** He should decide whether the bridging feed waits for the computed reading and a larger vote minimum, since its top currently can rest on two votes.\n\n**How sure the reader was:** yes for the code reading, which I checked in feed.py; fairly for the status of the replacement.\n\n**Not yet opened by a session.** Its quoted words were checked by code when the report was filed.\n\nFrom [2026-10-11-pilot-open-question-four-notes-as-a-haiku-reader-returned-it.md](readers/2026-10-11-pilot-open-question-four-notes-as-a-haiku-reader-returned-it.md), O10.\n\n### `docs/bridging.md:56`\n\n> `docs/bridging.md:56` “People disagree on what to conclude but agree the reasoning is structurally sound.”\n> `docs/bridging.md:75` “it surfaces the opposing group's strongest arguments — those rated well-reasoned even by the opposition — as the starting point”\n> `docs/bridging.md:24` “agreement is voted, quality is computed from the arguments and answered critical questions around a claim, and there is no second vote”\n\n**The reader:** The page promises the argument both sides find compelling: people who \"agree the reasoning is structurally sound\", arguments \"rated well-reasoned even by the opposition\". Nobody rates the reasoning. Line 24 says there is no second vote. The feed's score is the vote split on the claim times a machine-computed edge number. A claim that half the voters reject and the machine rates strong is contested and well-supported. Whether the rejecting half finds the reasoning sound is never asked, and that is what the page's title names. \"I can't dismiss this\" is a first-person state, and the score takes no input from it.\n\n**So, in the reader's words:** Score the premises instead (a claim under the contested one that voters on both sides agree with, using votes already collected), or reword the feed's claim to \"contested yet well-supported\".\n\n**How sure the reader was:** yes for the code: _feed_bridging counts AGREES_WITH and DISAGREES_WITH users and averages edge strength, nothing else. Fairly that no other surface collects reasoning ratings.\n\n**Not yet opened by a session.** Its quoted words were checked by code when the report was filed.\n\nFrom [2026-10-11-pilot-open-question-four-notes-as-a-reader-returned-it.md](readers/2026-10-11-pilot-open-question-four-notes-as-a-reader-returned-it.md), O7.\n\n### `docs/bridging.md:56`\n\n> `docs/bridging.md:56` “Where disagreement is high (50/50 vote split on the conclusion) BUT QBAF strength is also high”\n> `docs/bridging.md:91` “A genuine risk: maximizing cross-group agreement incentivizes vague platitudes.”\n\n**The reader** (a finding for what does not fit, kind `word-two-senses`): The page's bridging score is disagreement times strength, highest at a 50/50 split on the conclusion (feed.py agrees). The vagueness-trap section treats bridging as maximizing cross-group agreement and offers the four-type classification as the countermeasure. A platitude everyone endorses scores zero in the feed and the feed never reads claim type, so a reader is misled about which quantity the trap and its countermeasure concern.\n\n**Which stands, by the reader:** neither, by the code\n\n**Opened by session (2026-10-11): stands.** Both ends were read on 2026-10-11 (the reading note's § 7a). Its other end is the code, the to-do file, the instruction file or the same note, so the index does not hold it.\n\nFrom [2026-10-11-pilot-what-does-not-fit-four-notes-as-a-reader-returned-it.md](readers/2026-10-11-pilot-what-does-not-fit-four-notes-as-a-reader-returned-it.md), F8.\n\n### `docs/bridging.md:65`\n\n> `docs/bridging.md:65` “The bridging mode averages `r.strength` across a claim's scheme-bearing edges.”\n> `deliberus/feed.py:365` “WITH c, agrees, disagrees, total, agree_ratio, AVG(r.strength) AS avg_strength,”\n\n**The reader:** The 2026-08-17 correction says the feed averages r.strength over a claim's scheme-bearing edges. That is right but does not say which edges. In _feed_bridging (feed.py:362) and detect_bridging_claims (truth_graph.py:814) the match is (c)-[r]->(t:Claim). These are edges leaving the voted-on claim, the arguments it makes for other claims, not the reasoning arriving at it. The badge was fixed to incoming-only on 2026-08-17 for exactly this fault, and the feed was not. So the number multiplying the vote split says how c serves other claims. It falls back to 0.5 when c supports nothing and never moves when evidence is attached to c. The Q31 build item, the TODO entry and the tests (graph mocked) do not mention it.\n\n**So, in the reader's words:** When Q31's re-rank is built, fix the edge direction in the same change, or the feed will rank by the reading of the wrong links.\n\n**How sure the reader was:** yes. I read both queries against badge_inputs.py (`(s)-[r]->(c {claim_id: $cid})`), TODO.md:3403-3405 and TODO.md:1284.\n\n**Not yet opened by a session.** Its quoted words were checked by code when the report was filed.\n\nFrom [2026-10-11-pilot-open-question-four-notes-as-a-reader-returned-it.md](readers/2026-10-11-pilot-open-question-four-notes-as-a-reader-returned-it.md), O8.\n\n### `docs/bridging.md:67`\n\n> `docs/bridging.md:67` “Closing that gap changes the ranking on a shipped surface, which makes it a founder decision rather than a repair”\n> `TODO.md:1284` “bridging feeds and the truth graph's bridging list re-rank (answers the open”\n\n**The reader** (a finding for what does not fit, kind `overtaken`): The bridging note says closing the stored-strength gap in the feed is a founder decision still to be taken. On 2026-09-28 the founder ruled \"Yes\" to showing computed link strength, and the TODO entry says this answers the open bridging-feed question. The feed still ranks by the stored number, so only the build is pending.\n\n**Which stands, by the reader:** other, by a ruling of 2026-09-28\n\n**Opened by session (2026-10-11): stands.** Both ends were read on 2026-10-11 (the reading note's § 7a). Its other end is the code, the to-do file, the instruction file or the same note, so the index does not hold it.\n\nFrom [2026-10-11-pilot-what-does-not-fit-four-notes-as-a-haiku-reader-returned-it.md](readers/2026-10-11-pilot-what-does-not-fit-four-notes-as-a-haiku-reader-returned-it.md), F2.\n\n### `docs/bridging.md:91`\n\n> `docs/bridging.md:91` “A genuine risk: maximizing cross-group agreement incentivizes vague platitudes.”\n> `docs/bridging.md:93` “Deliberus's four-type classification is the structural countermeasure. Empirical claims are harder to make vague than normative ones.”\n> `docs/bridging.md:56` “Where disagreement is high (50/50 vote split on the conclusion) BUT QBAF strength is also high”\n\n**The reader:** The Vagueness Trap section defends against a design that maximises cross-group agreement. The formula on the page does the opposite. disagreement_factor peaks at a 50/50 vote split. The floor is two votes (`agrees + disagrees >= 2`), and the ratio ignores how many voted, so one agree and one disagree score the same as 500 and 500. The incentive is to be contested, and the cheap attack is two accounts voting opposite ways. The named countermeasure, claim type, is returned by the feed but not used in the score. The trap the page worries about is not the one the code can fall into.\n\n**So, in the reader's words:** Name the real gaming surface (contestedness on thin votes) and add a voter-count floor or weighting before this is presented as the project's distinctive signal.\n\n**How sure the reader was:** yes. I read feed.py lines 359-371: the type column is selected but not in the score.\n\n**Not yet opened by a session.** Its quoted words were checked by code when the report was filed.\n\nFrom [2026-10-11-pilot-open-question-four-notes-as-a-reader-returned-it.md](readers/2026-10-11-pilot-open-question-four-notes-as-a-reader-returned-it.md), O10.\n\n## `docs/civilizational-vision.md`\n\n### `docs/civilizational-vision.md:7`\n\n> `docs/civilizational-vision.md:7` “Every consequence below is also the same claim at a larger size.”\n> `docs/vision.md:86` “the wager is what I am betting, and this is what I am sure of.”\n> `docs/civilizational-vision.md:49` “Value bedrock: consciousness, suffering, shared humanity — small, universal”\n> `docs/vision.md:307` “*There is a shared bottom* is the weaker half”\n\n**The reader:** The civilizational page opens by saying every consequence is the Opacity conviction at a larger size. vision.md separates that conviction (\"what I am sure of\") from the wager (\"what I am betting\"). But the alignment section and the closing \"we share more than divides us\" are the bet. They rest on its weaker half, which vision.md says is the shared bottom rather than \"few and narrow\". The Lean table nonetheless states the value bedrock as \"small, universal\". The alignment argument works only if one shared target exists. The framing line lends the bet's consequences the conviction's certainty.\n\n**So, in the reader's words:** Mark which sections ride on the conviction and which on the wager, and state the alignment argument as conditional on a shared bottom.\n\n**How sure the reader was:** yes for the framing mismatch, fairly for the reading of the weaker half. The last sentence of line 42 hedges, but the table does not.\n\n**Opened by session (2026-10-11): stands.** The four passages are as quoted. That the framing line lends the wager's consequences the conviction's certainty is the reader's reading.\n\nFrom [2026-10-11-pages-open-question-three-vision-pages-as-a-reader-returned-it.md](readers/2026-10-11-pages-open-question-three-vision-pages-as-a-reader-returned-it.md), O8.\n\n### `docs/civilizational-vision.md:26`\n\n> `docs/civilizational-vision.md:26` “Outward-facing copy should follow this version rather than the older *\"glasses you'd never choose\"*”\n> `docs/vision.md:343` “The worldview filter — \"put on glasses you'd never choose and see what you've been missing\" — is the emotional and motivational hook that no predecessor offered.”\n> `web/src/lib/constellation.ts:231` “Put on glasses you would never choose, and see that the person you have argued with for years reasons with integrity.”\n> `docs/civilizational-vision.md:32` “That is what the \"rose-tinted glasses\" UX was always trying to name.”\n\n**The reader:** The 16 September sharpening says outward copy should drop \"glasses you'd never choose\". The paragraph just above it (line 24) still uses the phrase, and so does vision.md's \"soul of the product\" sentence. So does the expanded text of the \"Understanding is the infrastructure of love\" tile in web/src/lib/constellation.ts. That tile text is marked draft and unruled. Line 32 adds a third image, \"rose-tinted glasses\". In English that means flattering distortion, which is the opposite of the sharpening: the same world, bent through someone else's assumptions.\n\n**So, in the reader's words:** Sweep those surfaces, including constellation-candidates.md, and keep one image.\n\n**How sure the reader was:** yes. I searched docs, web and deliberus for the phrase.\n\n**Opened by session (2026-10-11): stands.** The phrase stands at vision.md line 343 and in the opened text of a tile in web/src/lib/constellation.ts, and the page that asks outward copy to drop it uses a third image at its line 32.\n\nFrom [2026-10-11-pages-open-question-three-vision-pages-as-a-reader-returned-it.md](readers/2026-10-11-pages-open-question-three-vision-pages-as-a-reader-returned-it.md), O10.\n\n### `docs/civilizational-vision.md:30`\n\n> `docs/civilizational-vision.md:30` “Procaccia and Konya's peacebuilding work demonstrated that this can happen between Israelis and Palestinians during active conflict.”\n> `docs/research/ai-augmented-cooperation-infrastructure.md:63` “**Limitations**: Self-selected peacebuilder sample (not general population), no controlled experiment”\n\n**The reader:** This is the only empirical support offered for the \"I see you\" mechanism, which the page calls \"precise\". The project's own note on the study describes anonymous responses and votes on AI-distilled statements from about 100 to 120 self-selected peacebuilders, with no control. That shows self-selected people agreeing on statements. It does not show anyone ceasing to think the other side reasons badly. \"If it can happen there, it can happen between the people in your family\" then carries the whole weight. The step from the sample to family is an extrapolation.\n\n**So, in the reader's words:** Say a self-selected peacebuilder sample agreed on statements, and present the family case as the hope.\n\n**How sure the reader was:** fairly. I read section 2 of the cooperation-infrastructure note. The \"shared a room\" detail in bridging.md is already flagged by others, so I left it.\n\n**Opened by session (2026-10-11): stands.** ai-augmented-cooperation-infrastructure.md line 63 gives the limits as quoted: a self-selected sample and no controlled experiment. The sibling sentence on the bridging page is number 13 on the founder's list.\n\nFrom [2026-10-11-pages-open-question-three-vision-pages-as-a-reader-returned-it.md](readers/2026-10-11-pages-open-question-three-vision-pages-as-a-reader-returned-it.md), O11.\n\n### `docs/civilizational-vision.md:34`\n\n> `docs/civilizational-vision.md:34` “A Swedish woman writing publicly in September 2026 about her own brother, who had started working for a party she found close to unbearable”\n> `docs/research/the-post-that-states-the-problem-from-inside.md:4` “**Source**: a public Swedish Facebook post of 2026-09-18 by Tildah Puronne, roughly 1,400 words,”\n\n**The reader:** The vision page keeps her anonymous, but the same paragraph links to a note that names her. That note says which party her younger brother works for (line 70) and that she herself joined two parties. The note says the full text is held privately because reproducing it would be \"publication rather than citation\", so he drew a line there. The name plus family and party detail may sit on the wrong side of it. TODO.md keeps commenters unnamed in docs/. CLAUDE.md asks that named-person material pass the Reader Test before promotion.\n\n**So, in the reader's words:** Check whether she is a public figure or agreed, and otherwise drop her name and her brother's party from docs/.\n\n**How sure the reader was:** a guess. I do not know whether she is public or consented; I only checked what the note and TODO.md say.\n\n**Opened by session (2026-10-11): stands.** The research note names the post's author at its line 4, gives the party her brother works for at line 70, and at line 18 says she named two parties she has joined, all of it from her public post. Put to the founder, who ruled the same day: \"The FB post name is fine\". She stays named and the note is unchanged.\n\nFrom [2026-10-11-pages-open-question-three-vision-pages-as-a-reader-returned-it.md](readers/2026-10-11-pages-open-question-three-vision-pages-as-a-reader-returned-it.md), O14.\n\n### `docs/civilizational-vision.md:80`\n\n> `docs/civilizational-vision.md:80` “Bridging signals identify reasoning that survives cross-group scrutiny.”\n> `TODO.md:2277` “It currently ranks by the classifier's own 0–1 rating from extraction, which no prompt defines and no answered question ever updates”\n\n**The reader** (a finding for what does not fit, kind `state`): The page says bridging signals identify reasoning that survives scrutiny across groups. The bridging feed (deliberus/feed.py, _feed_bridging) scores vote split times the average of the stored link rating the extraction model wrote, not the badge computed from answered critical questions. A claim whose links the model happened to rate high ranks as bridging although no question was ever answered. vision.md:351 carries a mark of this; this page does not.\n\n**Which stands, by the reader:** other, by the code\n\n**Opened by session (2026-10-11): stands.** The project's own instructions record, from a reading of deliberus/feed.py on 2026-10-10, that the bridging mode multiplies the vote split by the average of the number the classifier wrote on each link. Not reopened here. The page states the signal as working.\n\nFrom [2026-10-11-pages-what-does-not-fit-three-vision-pages-as-a-reader-returned-it.md](readers/2026-10-11-pages-what-does-not-fit-three-vision-pages-as-a-reader-returned-it.md), F14.\n\n### `docs/civilizational-vision.md:120`\n\n> `docs/civilizational-vision.md:120` “which value premises are shared by 80% of humanity and which by 5%”\n> `docs/convergence.md:160` “self-selects on one disposition, a thousand users from one reading community are one frame sampled a”\n\n**The reader:** The graph can only count whoever contributed or was extracted. For any foreseeable stage \"shared by 80% of humanity\" would mean 80% of the people who voted, mostly an argument-friendly, English-reading minority. The project's own principle says every published count labels its provenance. convergence.md warns that a thousand users from one reading community are one frame sampled many times. As written the sentence is an unlabelled count, and it names the denominator the project cannot reach.\n\n**So, in the reader's words:** Write it as \"shared by x% of the contributors we have\" and keep the humanity-scale claim as a target that needs selected diversity.\n\n**How sure the reader was:** fairly. I searched for representativeness language in the three notes and bridging.md and found only the convergence.md line.\n\n**Opened by session (2026-10-11): stands.** The two passages are as quoted. By the project's own principle a published count names whom it counted.\n\nFrom [2026-10-11-pages-open-question-three-vision-pages-as-a-reader-returned-it.md](readers/2026-10-11-pages-open-question-three-vision-pages-as-a-reader-returned-it.md), O12.\n\n## `docs/conceptual-threads.md`\n\n### `docs/conceptual-threads.md:79`\n\n> `docs/conceptual-threads.md:79` “It is probably the right position to hold, and it is currently held invisibly: nothing in the system marks it as a position”\n\n**The reader:** The 22 August refinement says the is/ought commitment is held invisibly and that \"three responses are available and ordered by cost\", but it lists none. They are in what-belongs-in-the-ontology.md § 5: de-bake, confess, publish. The third had been executed that same day. The header of scripts/seed_ontology_commitments.py reads \"Founder-ratified 2026-08-22\" and publishes the partition as a claim that the live hedonic-monism node attacks. convergence.md:124 already calls that node live. So the paragraph's diagnosis was overtaken on its own date, and a visitor is promised a list that is not there.\n\n**So, in the reader's words:** Replace those two sentences with the three moves, and say the is/ought partition is already a published, attackable claim.\n\n**How sure the reader was:** yes. I read the seeder header and § 5 of the ontology note.\n\n**Opened by session (2026-10-11): stands.** Opened scripts/seed_ontology_commitments.py: its header says the founder ratified publishing the split between is and ought on 2026-08-22. The page does not name the three responses and says the position is held invisibly.\n\nFrom [2026-10-11-pages-open-question-three-concept-pages-as-a-reader-returned-it.md](readers/2026-10-11-pages-open-question-three-concept-pages-as-a-reader-returned-it.md), O3.\n\n### `docs/conceptual-threads.md:103`\n\n> `docs/conceptual-threads.md:103` “Layer 1 prevents new duplicates, layer 2 catches existing ones, layer 3 discovers semantic equivalence that text similarity misses.”\n> `docs/conceptual-threads.md:111` “Human labor should fall primarily on **splitting false sameness** rather than manually merging every obvious duplicate.”\n\n**The reader:** Layer 3 of the dedup stack reads Polis-style co-voting as a sign that two claims are \"likely equivalent\". Co-endorsement shows two claims sit in one frame. That is how the thread's own warning two paragraphs later uses it, as a detector of forks. Polis clusters people, not statements, and the corpus's Polis note lists near-duplicate statements as a Polis weakness. The April 3 addition then has near-duplicates treated as one cluster by default and leaves humans to split false sameness. Yet the thread says nothing yet detects same-word, different-path cases. The silent failure, over-merging, ends up on the default path.\n\n**So, in the reader's words:** Relabel layer 3 as a frame signal rather than a dedup layer, and decide on clustering by default only once something can detect false sameness.\n\n**How sure the reader was:** fairly. I read polis-deep-dive.md:237 and the soft-clustering note, which is marked \"not yet implemented\", and found no default-clustering code. The Polis internals are from my own knowledge.\n\n**Opened by session (2026-10-11): stands.** Opened lines 103 to 111: the third layer reads voting together as a sign that two claims are the same. Voting together shows a shared frame. What the reader says about how Polis works inside is the reader's own knowledge and was not checked.\n\nFrom [2026-10-11-pages-open-question-three-concept-pages-as-a-reader-returned-it.md](readers/2026-10-11-pages-open-question-three-concept-pages-as-a-reader-returned-it.md), O1.\n\n### `docs/conceptual-threads.md:137`\n\n> `docs/conceptual-threads.md:137` “This requires separating two voting axes: conclusion agreement vs. reasoning quality (see Thread 7).”\n> `docs/conceptual-threads.md:192` “(2) **Argument quality record** — cross-adversarial ratings weighted most heavily (a \"well-argued\" vote from someone who disagrees with your conclusion carries more weight).”\n> `docs/conceptual-threads.md:194` “This two-axis voting is also the prerequisite for detecting bridging arguments (Thread 5).”\n> `docs/conceptual-threads.md:264` “*(Corrected 2026-09-09: the separation shipped as single-axis voting plus computed strength, never as two votes”\n\n**The reader:** The 9 September correction is an aside on the critical-path sentence. Three other passages in this same note still need the second vote. One is the bridging paragraph, where groups \"both rate an argument well-reasoned\" and each node carries per-group strength vectors. Another is reputation signal (2), which weights \"well-argued\" votes from opponents. The third is the UX-primitive paragraph. The built bridging mode multiplies the agree/disagree split by computed strength, so none of these is how it works. The March feature note reports that most such votes go one direction over 95% of the time. analysis-and-attunement.md:102 repeats the dropped design.\n\n**So, in the reader's words:** Mark or rewrite the three passages against ux-principles P2, so a builder or visitor does not take the dropped design for the live one.\n\n**How sure the reader was:** yes. I read ux-principles.md:62 and :64, bridging.md:24 and unbuilt-features-research-mar31.md:11-17, and searched the docs for \"two-axis\".\n\n**Opened by session (2026-10-11): stands.** Opened the passages: the correction of 2026-09-09 sits on one sentence, and lines 137, 192 and 194 still need a second vote on quality that was never built.\n\nFrom [2026-10-11-pages-open-question-three-concept-pages-as-a-reader-returned-it.md](readers/2026-10-11-pages-open-question-three-concept-pages-as-a-reader-returned-it.md), O2.\n\n### `docs/conceptual-threads.md:190`\n\n> `docs/conceptual-threads.md:190` “This capital weights their future contributions — a Metaculus-style track record applied to argument quality, not just prediction accuracy.”\n> `docs/conceptual-threads.md:69` “Classify claims as descriptive vs normative. For normative: score on importance + agreement, not true/false”\n\n**The reader:** Thread 7 says a track record weights future contributions, and the first signal is a Brier-scored calibration score. Calibration needs claims that resolve. Thread 3 says normative claims are scored on importance and agreement, not truth. So weight can be earned on the factual half and hardly at all on the value half, then carried onto contributions where a forecasting record says nothing. The ruled principle that an act's value is not its rung says the most valuable contributions, a stake or a reframe, need only a position. The thread never says how weight is earned where nothing resolves.\n\n**So, in the reader's words:** Say which of the four signals can exist on normative claims, and what the weight means for a first contribution that rests on position alone.\n\n**How sure the reader was:** fairly. I compared Threads 3 and 7 in this note against ledger D4 and D6, but did not read epistemic-gamification.md.\n\n**Opened by session (2026-10-11): stands.** Read the two threads together: a record that needs claims to resolve can be earned on factual claims and hardly on value claims. The reader did not open the note on the reputation design, and neither did I.\n\nFrom [2026-10-11-pages-open-question-three-concept-pages-as-a-reader-returned-it.md](readers/2026-10-11-pages-open-question-three-concept-pages-as-a-reader-returned-it.md), O4.\n\n### `docs/conceptual-threads.md:222`\n\n> `docs/conceptual-threads.md:222` “A truth-seeking feed should optimize for **Bayesian surprise** — content that maximally updates beliefs.”\n> `docs/conceptual-threads.md:222` “Deliberus's feed should surface: (1) arguments most likely to change YOUR mind (personalized counter-arguments)”\n\n**The reader:** Thread 8 defines the truth-seeking feed as one that maximally updates beliefs and surfaces arguments \"most likely to change YOUR mind\". That is a persuasion objective under a better name. The claims that move a reader most are as often startling and false as sound. A feed tuned to one reader's mind-changing is the personalized persuasion the project's conviction (checking that keeps pace with persuasion) is set against. The contrast in the paragraph is drawn against engagement and consensus feeds, never against persuasion. The built bridging mode ranks by agree/disagree split times computed strength and does not follow this paragraph.\n\n**So, in the reader's words:** Reword the feed objective around what a reader can check (strength, open questions, contested-and-supported) and drop \"most likely to change YOUR mind\".\n\n**How sure the reader was:** fairly. deliberus/feed.py has no belief-change ranking, only a docstring aspiration.\n\n**Opened by session (2026-10-11): stands.** The wording is on the page at line 222. The reader found no ranking by change of belief in the feed code; not reopened here. A remark about what the page aims the feed at.\n\nFrom [2026-10-11-pages-open-question-three-concept-pages-as-a-reader-returned-it.md](readers/2026-10-11-pages-open-question-three-concept-pages-as-a-reader-returned-it.md), O5.\n\n## `docs/convergence.md`\n\n### `docs/convergence.md:35`\n\n> `docs/convergence.md:35` “This empirical result confirms the core hypothesis: convergence is not imposed — it emerges from structured decomposition.”\n> `docs/convergence.md:181` “- **Say at which level an agreement was found: on words, on a case, or on reasons.** Agreement on an”\n> `docs/convergence.md:193` “abstract). Asking *which of these two is worse* is a different question from asking *is this”\n\n**The reader:** The 8 October section asks that every agreement be placed at a level (words, a case, reasons) and notes that orderings may converge where verdicts do not. The page's lead evidence was not re-read at those levels. The Moral Graph result is \"the whole thesis in miniature\", yet it is an ordering (value A wiser than B in a context). academic-foundations.md:272 gives the sample as 500 Americans, which the page omits beside \"256 societies\". The seven moral categories are agreement on categories. The new section calls the ordering question untested here, so the same kind of result is confirmation in one section and untested in another.\n\n**So, in the reader's words:** Annotate each piece of lead evidence with its level and sample, so the new counting rule is applied to the page's own support.\n\n**How sure the reader was:** fairly. I read academic-foundations.md:272 and searched for any place the level rule is applied back to the Moral Graph or Curry evidence, and found none.\n\n**Opened by session (2026-10-11): stands.** Opened docs/academic-foundations.md at line 272: the sample is 500 Americans. The rule of 8 October about the level of an agreement is not applied to the page's own lead evidence; the reader searched for a place where it is and found none.\n\nFrom [2026-10-11-pages-open-question-three-concept-pages-as-a-reader-returned-it.md](readers/2026-10-11-pages-open-question-three-concept-pages-as-a-reader-returned-it.md), O6.\n\n### `docs/convergence.md:47`\n\n> `docs/convergence.md:47` “**A terminus is dated, and the map does not say so”\n> `docs/convergence.md:47` “nothing in the horizon map represents a terminus expiring”\n> `docs/convergence.md:51` “Two honest responses, a founder call: add an **is-bedrock / open-empirical-metaphysical** type”\n\n**The reader:** Two items on this page are presented as open and are closed. The bold lead says the map does not say a terminus is dated. The bracket inside it says the gap was closed in code on 16 September, and horizon.py says readings are \"refreshed on every read\" and dated, so the sentence about nothing representing expiry no longer holds. Separately, the 10 July paragraph asks the founder to choose between an is-bedrock type and a normative-only scope. terminus.py and TODO.md:2287 show is_bedrock added and ratified on 22 August, and \"What Stops the Descent\" on this page is the reasoning. The bold leads still carry the old state.\n\n**So, in the reader's words:** Retitle the first as closed and turn the second into what was ruled; the same fork still reads open in what-belongs-in-the-ontology.md § 7 and TODO.md:3988.\n\n**How sure the reader was:** yes. I read the terminus.py docstring, the horizon.py header and both TODO rows.\n\n**Opened by session (2026-10-11): stands.** Opened deliberus/terminus.py: the kind the paragraph of 10 July asks the founder to choose is in the list. The bold lead still says the map does not date a horizon, and the bracket inside it says that was closed on 16 September.\n\nFrom [2026-10-11-pages-open-question-three-concept-pages-as-a-reader-returned-it.md](readers/2026-10-11-pages-open-question-three-concept-pages-as-a-reader-returned-it.md), O8.\n\n### `docs/convergence.md:74`\n\n> `docs/convergence.md:74` “| Kernel: small, auditable | Value bedrock: consciousness, suffering, shared humanity |”\n> `docs/convergence.md:76` “| Trust: kernel's simplicity | Trust: values' universality |”\n\n**The reader:** The Lean table puts \"value bedrock\" in the kernel row and values' universality as the source of trust. In Lean the kernel is the small checker. Trust in it does not depend on which axioms anyone adopts, and the system can show which axioms a result rests on. The checker's counterpart here is the legibility floor. The research note's own § 3 maps the kernel to form-checking. As drawn, the table makes verification-based alignment depend on the wager winning, though the structure-checking half works whichever values people hold. The bedrock belongs on the axioms row, as the wager's condition.\n\n**So, in the reader's words:** Redraw the table with the checker as the floor and shared values as the conditional, so the Lean analogy survives the wager losing.\n\n**How sure the reader was:** fairly. I compared lean-deliberus-analogies.md § 3; the Lean points are from my own knowledge.\n\n**Opened by session (2026-10-11): stands.** Opened docs/research/lean-deliberus-analogies.md: line 15 has the kernel as the small trusted checker. The page's table puts shared values in the kernel's row. What the reader adds about Lean is the reader's own knowledge.\n\nFrom [2026-10-11-pages-open-question-three-concept-pages-as-a-reader-returned-it.md](readers/2026-10-11-pages-open-question-three-concept-pages-as-a-reader-returned-it.md), O9.\n\n### `docs/convergence.md:104`\n\n> `docs/convergence.md:104` “Rorty's complementary contribution is irony — the recognition that the analytical framework itself”\n\n**The reader:** Rorty is offered as the complement to a Peircean thesis, \"truth as the limit of inquiry\". Rorty is the standard philosophical opponent of that limit, so the pairing works as a mood (irony stops dogmatism, convergence keeps you building) and not as grounding. The ledger holds the pairing as P11, OPEN and \"mine\". The 26 September audit says the Rorty reading rested on an unverified 500-word summary and skipped his solidarity view, which is closer to what a badge measures. The public section states it as grounding.\n\n**So, in the reader's words:** Mark the section as an unverified sketch, or rewrite it with Rorty as the standing objection to the wager.\n\n**How sure the reader was:** fairly. I read ledger row P11 and listening/2026-09-26-arc-p-the-april-conversation-audited.md; Rorty's position is from my own knowledge.\n\n**Opened by session (2026-10-11): stands.** Ledger row P11 is open and marked as the session's own. The audit of 26 September has a part on Rorty and Peirce, which I did not read whole. Rorty's position as an opponent of the limit of inquiry is the reader's own knowledge.\n\nFrom [2026-10-11-pages-open-question-three-concept-pages-as-a-reader-returned-it.md](readers/2026-10-11-pages-open-question-three-concept-pages-as-a-reader-returned-it.md), O10.\n\n### `docs/convergence.md:134`\n\n> `docs/convergence.md:134` “the project's boldest Hail Mary and should be stated with explicit falsification conditions”\n> `docs/convergence.md:41` “The convergence thesis is falsifiable, and stating how is part of stating it honestly.”\n> `docs/convergence.md:45` “hinge-clash, parity, and permissivism horizons count AGAINST convergence”\n> `docs/convergence.md:136` “Population ethics is the existence proof of the endgame”\n\n**The reader:** The page says convergence can lose and should have explicit falsification conditions, but it never gives N, a depth or a cutoff for \"small minority\" (docs and code searched). The instrument counts examined leaves, not the \"breadth\" of disagreement the synthesis note names. In horizon.py every examined empirical leaf reads \"empirical\" and joins the denominator, so the separability floor he ruled, which splits finer, lowers the fraction with nobody converging. The reading can emit only empirical, definitional, fittingness and permissive_zone. \"Hinge-clash\" is not a kind, and population ethics' axiom_choice arrives as a proposal only, so neither can count against the wager yet.\n\n**So, in the reader's words:** Fix N, depth, a numeric cutoff and the unit (leaves or breadth) before the first live fraction is read as news.\n\n**How sure the reader was:** fairly. I read horizon.py (read_horizon_from_inputs), completeness.py (summarize_horizon_map) and terminus.py, but did not run them on the live graph.\n\n**Opened by session (2026-10-11): stands.** Opened deliberus/horizon.py: an examined factual leaf reads as a factual horizon and a claim with parts is left out of the count. The page gives no number, depth or cut-off for the wager losing. Not run on the live graph.\n\nFrom [2026-10-11-pages-open-question-three-concept-pages-as-a-reader-returned-it.md](readers/2026-10-11-pages-open-question-three-concept-pages-as-a-reader-returned-it.md), O7.\n\n## `docs/depth.md`\n\n### `docs/depth.md:40`\n\n> `docs/depth.md:40` “and that asymmetry is itself moral data, not a bug to be reasoned away.”\n> `docs/depth.md:43` “Neither answer is stupid. The graph doesn't settle this.”\n\n**The reader:** The page's demonstration is a hand-built chain, not a pipeline run. I searched the repo and the Singer case is extracted nowhere. Another note calls the decomposition \"designed to bottom out at weighting\". With one chain per branch, every \"because\" is a single choice that a real descent would branch on. The last node of the phenomenological branch is written in one side's voice. \"That asymmetry is itself moral data, not a bug\" is the dissenter's conclusion, and Singer's camp calls the asymmetry a bias. The next paragraph says the graph \"doesn't settle this\". So the finding that the branches meet follows from how the chain was written.\n\n**So, in the reader's words:** Label the example an authored illustration and make the bottom node a contested claim with the opposing reading attached, or run the case through the pipeline and show what comes out.\n\n**How sure the reader was:** fairly. rg for drowning child and Singer finds only depth.md and the Swedish intro page. I read convergence-thread-corpus-map.md:98. The point about Singer's camp rests on general knowledge of the debate.\n\n**Not yet opened by a session.** Its quoted words were checked by code when the report was filed.\n\nFrom [2026-10-11-pilot-open-question-four-notes-as-a-reader-returned-it.md](readers/2026-10-11-pilot-open-question-four-notes-as-a-reader-returned-it.md), O2.\n\n### `docs/depth.md:47`\n\n> `docs/depth.md:47` “Research confirms this matters: many ethical decision problems are NP-hard (Stenseke, 2024)”\n> `docs/depth.md:136` “This means decomposition (breaking hard moral problems into tractable sub-problems) is not just useful but *necessary*”\n\n**The reader:** Twice the page says the NP-hardness of ethical decision problems makes decomposition \"computationally necessary\". The corpus's own source note (deliberus-as-alignment-infrastructure.md:97) says only that the result \"suggests\" this. depth.md turns that into \"Research confirms\" and \"This means\". The inference is also loose. NP-hardness is a worst-case result about computing the best action, which Deliberus does not attempt. Splitting an NP-hard problem into parts helps only when the parts barely interact. That condition is what the page's later Simon section supplies, and it is the better reason. The corpus also cites this paper under two different titles.\n\n**So, in the reader's words:** Drop \"necessary\" or rest it on Simon's near-decomposability, and check what Stenseke actually concludes before the page says \"confirms\".\n\n**How sure the reader was:** fairly on the logic. A guess on what Stenseke himself concludes, since I did not read the paper. The two titles are at non-zero-sum-economics-and-civilizational-cooperation.md:484 and deliberus-as-alignment-infrastructure.md:565.\n\n**Not yet opened by a session.** Its quoted words were checked by code when the report was filed.\n\nFrom [2026-10-11-pilot-open-question-four-notes-as-a-reader-returned-it.md](readers/2026-10-11-pilot-open-question-four-notes-as-a-reader-returned-it.md), O3.\n\n### `docs/depth.md:57`\n\n> `docs/depth.md:57` “The platform should NEVER accept unexamined premises as stopping points.”\n> `docs/depth.md:167` “until the user reaches their actual bedrock, not a convenient stopping point”\n> `docs/depth.md:118` “Converting what went without saying into something sayable is the core operation”\n\n**The reader:** The page's imperative is unconditional: never accept a stopping point, keep inviting until the user reaches \"actual bedrock\". It cites the unsaid note as warrant for turning the unsaid into the sayable. That note's findings are limits on exactly this operation: for a hinge the pass is a category error that performs the taboo trade-off, and the conversion is political. The September evidence research adds that hinges are a principled stop and that bare demands turn decomposition into a stalling weapon. depth.md never mentions hinges in that sense. Its title, No Permanent Bottom, is the unscoped form of a conviction that CLAUDE.md words as currently-undecomposed and invitable.\n\n**So, in the reader's words:** Add one scoped sentence to depth.md (invite, never press; a hinge or protected value is a legitimate place to stop) and retitle, since a headline stays over-broad whatever the body says.\n\n**How sure the reader was:** fairly. I compared the page with the CLAUDE.md conviction wording and with what-counts-as-evidence.md §2 on hinges and stalling. rg for hinge and sacred in depth.md finds only the sensitivity sense and one protected-value line.\n\n**Not yet opened by a session.** Its quoted words were checked by code when the report was filed.\n\nFrom [2026-10-11-pilot-open-question-four-notes-as-a-reader-returned-it.md](readers/2026-10-11-pilot-open-question-four-notes-as-a-reader-returned-it.md), O1.\n\n### `docs/depth.md:69`\n\n> `docs/depth.md:69` “you ratify, edit, or refuse; you never author structure from a blank page.”\n\n**The reader:** The first of the seven things the machine holds says it proposes the decomposition and you ratify, edit or refuse. Since 2026-09-21 the splitter writes marked parts and the person reviews afterwards (keep, reword, drop). CLAUDE.md already flags the old propose-only description as wrong, and a \"ratified\" state is being retired. The order of events a visitor is promised runs the other way round. It also touches the page's claim that the one irreplaceable thing is the user's recognition: that now comes after the write, and only if someone sits down to review.\n\n**So, in the reader's words:** Reword item 1 to say the machine writes marked parts and the user keeps, rewords or drops them.\n\n**How sure the reader was:** fairly. I checked CLAUDE.md:1027, proposed_split.py and ledger B7/Q32. I did not check whether the web UI still offers a proposal step.\n\n**Not yet opened by a session.** Its quoted words were checked by code when the report was filed.\n\nFrom [2026-10-11-pilot-open-question-four-notes-as-a-reader-returned-it.md](readers/2026-10-11-pilot-open-question-four-notes-as-a-reader-returned-it.md), O5.\n\n### `docs/depth.md:70`\n\n> `docs/depth.md:70` “taps, not essays; a critical question arrives as a small answerable thing, with *does not fit* always available.”\n> `web/src/routes/claim/[id]/+page.svelte:1181` “<textarea class=\"invitation-input cq-answer-input\" bind:value={cqAnswerText} placeholder=\"Provide evidence or reasoning...\" rows=\"2”\n\n**The reader** (a finding for what does not fit, kind `state`): Depth.md says the machine asks in taps, not essays, and that \"does not fit\" is always available on a critical question. The claim page lists every critical question at once, offers only Yes and No, and disables Submit until the user has typed text (submitCQAnswer returns on empty text). A visitor who wants to answer by tapping, or to say the question does not fit, cannot.\n\n**Which stands, by the reader:** other, by the code\n\n**Opened by session (2026-10-11): stands.** Both ends were read on 2026-10-11 (the reading note's § 7a). Its other end is the code, the to-do file, the instruction file or the same note, so the index does not hold it.\n\nFrom [2026-10-11-pilot-what-does-not-fit-four-notes-as-a-reader-returned-it.md](readers/2026-10-11-pilot-what-does-not-fit-four-notes-as-a-reader-returned-it.md), F2.\n\n### `docs/depth.md:99`\n\n> `docs/depth.md:99` “and the verdict that marks it is itself a challengeable claim”\n> `docs/depth.md:185` “every confirmed terminus remains a fallible fixed point open to a new move”\n> `docs/worldview-lenses.md:28` “including premise staleness and terminus expiry”\n> `docs/research/the-load-bearing-unsaid.md:144` “and the residue map exists to count exactly that”\n\n**The reader:** Four sentences across these notes use machinery or names the founder has since ruled out. depth.md says a classified terminus is a fixed point and that a verdict marks it. Since 2026-09-16 a kind is read, never set, and the verdict-claim ground is retired. It also speaks of a \"confirmed terminus\", but the confirmed field is retired and terminus itself was ruled a misnomer. The worldview page says \"terminus expiry\". The unsaid note says \"residue map\", and horizon has been the name since 2026-08-25. Each sentence still says something true, but a reader hunting for a verdict on a claim page finds none.\n\n**So, in the reader's words:** Restate them in the horizon-reading vocabulary (a dated reading over the current arrangement), so a correction is not left standing on four surfaces.\n\n**How sure the reader was:** yes for the rulings (ledger O2, O6, Q32; CLAUDE.md:787 and :553). Fairly that these are the only instances in the four notes, since I did not search every use of verdict.\n\n**Not yet opened by a session.** Its quoted words were checked by code when the report was filed.\n\nFrom [2026-10-11-pilot-open-question-four-notes-as-a-reader-returned-it.md](readers/2026-10-11-pilot-open-question-four-notes-as-a-reader-returned-it.md), O6.\n\n### `docs/depth.md:112`\n\n> `docs/depth.md:112` “but it is inert until someone records the interpretation”\n> `docs/depth.md:120` “its default was neutral until the 2026-08-31 padding ruling deliberately made untagged decomposition parts weakest-part-governed”\n> `docs/bridging.md:67` “which makes it a founder decision rather than a repair; it is recorded in”\n\n**The reader:** Two statements in these notes were overtaken and not updated. The depth page calls the necessary-default fix \"inert until someone records the interpretation,\" yet a later paragraph of the same page says the default was changed on 2026-08-31. The bridging page calls the strength question a pending founder decision, though the founder has since accepted the replacement. When a ruling lands, the code and canon move and the public explanation does not.\n\n**So, in the reader's words:** He should have each public explanation re-checked when its ruling lands, or dated, so a visitor is not told an answered question is still open.\n\n**How sure the reader was:** yes. Both contradictions are inside the notes and need no code to confirm.\n\n**Not yet opened by a session.** Its quoted words were checked by code when the report was filed.\n\nFrom [2026-10-11-pilot-open-question-four-notes-as-a-haiku-reader-returned-it.md](readers/2026-10-11-pilot-open-question-four-notes-as-a-haiku-reader-returned-it.md), O11.\n\n### `docs/depth.md:112`\n\n> `docs/depth.md:112` “but it is inert until someone records the interpretation.”\n> `docs/depth.md:120` “its default was neutral until the 2026-08-31 padding ruling deliberately made untagged decomposition parts weakest-part-governed”\n> `docs/depth.md:79` “91% of verdicts survived every filing, the typical score did not move at all”\n\n**The reader:** Three statements on the page about the strength side do not agree. Line 112 says the necessary fix is inert until someone records the interpretation. Line 120 says the 2026-08-31 ruling made untagged parts weakest-part-governed, and qbaf.py now passes undecided_as=\"necessary\", so the fix is no longer inert. The slicing result at line 79 was measured on 2026-08-22, before that flip. Its own note says most strengths then sat near neutral because critical questions were unanswered, so part of the stability is emptiness.\n\n**So, in the reader's words:** Rewrite line 112 to say the default now does the work and what that costs, and re-run the slicing study or date the result on the page.\n\n**How sure the reader was:** yes for line 112 against line 120 and the code (qbaf.py:288). Fairly for the stability caveat: I read slicing-stability-first-run.md and did not re-run it.\n\n**Not yet opened by a session.** Its quoted words were checked by code when the report was filed.\n\nFrom [2026-10-11-pilot-open-question-four-notes-as-a-reader-returned-it.md](readers/2026-10-11-pilot-open-question-four-notes-as-a-reader-returned-it.md), O4.\n\n### `docs/depth.md:112`\n\n> `docs/depth.md:112` “but it is inert until someone records the interpretation.”\n> `deliberus/support_semantics.py:42` “by their weakest member (under the separability floor, parts of a decomposed whole are”\n\n**The reader** (a finding for what does not fit, kind `state`): The note says the shipped necessary-tagging fix stays inert until someone records a join interpretation. Since the 2026-08-31 ruling, `subclaim_energy` passes `undecided_as=\"necessary\"`, so untagged decomposition children are weakest-part-governed by default and the fix is live with no recording. depth.md:120 records that flip, but line 112 was never updated.\n\n**Which stands, by the reader:** other, by the code\n\n**Opened by session (2026-10-11): stands.** Both ends were read on 2026-10-11 (the reading note's § 7a). Its other end is the code, the to-do file, the instruction file or the same note, so the index does not hold it.\n\nFrom [2026-10-11-pilot-what-does-not-fit-four-notes-as-a-haiku-reader-returned-it.md](readers/2026-10-11-pilot-what-does-not-fit-four-notes-as-a-haiku-reader-returned-it.md), F1.\n\n### `docs/depth.md:114`\n\n> `docs/depth.md:114` “rebuilds a parent from its children alone and reports what went missing, which turns \"as lossless as possible\" from an aspiration into a number.”\n> `deliberus/extraction/decompose_claim.py:44` “the model that rebuilds has the original in front of it. A blind rebuilding”\n\n**The reader** (a finding for what does not fit, kind `overtaken`): Depth.md says the recomposition check rebuilds a parent from its children alone. The code header, found 2026-10-10, says that reconstruction is not blind: the same call prints the original, so the rebuilding model sees it. Five places were corrected that day (quality-control-for-deeper-steps.md § 1.3); this page was not, which breaks the rule that a correction reaches every surface that repeats it.\n\n**Which stands, by the reader:** other, by the code\n\n**Opened by session (2026-10-11): stands.** Both ends were read on 2026-10-11 (the reading note's § 7a). Its other end is the code, the to-do file, the instruction file or the same note, so the index does not hold it.\n\nFrom [2026-10-11-pilot-what-does-not-fit-four-notes-as-a-reader-returned-it.md](readers/2026-10-11-pilot-what-does-not-fit-four-notes-as-a-reader-returned-it.md), F4.\n\n## `docs/design-guide.md`\n\n### `docs/design-guide.md:15`\n\n> `docs/design-guide.md:15` “**Page width**: `max-width: 820px` centered. All pages share this constraint.”\n> `web/src/routes/compartments/+page.svelte:220` “main { max-width: 900px; margin: 0 auto; padding: 1rem 1.5rem 5rem; }”\n\n**The reader** (a finding for what does not fit, kind `number-or-date`): The guide says every page is capped at 820px. The code gives /concepts 1040px, /concept/[term] 980px, and compartments, sameness and part-review 900px (step-review 860px). A page built to the guide's width would not match those pages.\n\n**Which stands, by the reader:** other, by the code\n\n**Opened by session (2026-10-11): stands.** Checked the page styles: widths of 900, 980 and 1040 pixels exist beside the 820 the guide gives for every page.\n\nFrom [2026-10-11-pages-what-does-not-fit-ux-principles-and-design-guide-as-a-reader-returned-it.md](readers/2026-10-11-pages-what-does-not-fit-ux-principles-and-design-guide-as-a-reader-returned-it.md), F13.\n\n### `docs/design-guide.md:17`\n\n> `docs/design-guide.md:17` “View Transitions API — blur-out (200ms) on old, fade-in (300ms) on new. Smooth cross-page navigation.”\n> `web/src/app.css:98` “/* ── Page transitions (View Transitions API) ── */”\n\n**The reader** (a finding for what does not fit, kind `number-or-date`): The guide describes a 200ms blur-out and a 300ms fade-in. The rules under that comment in app.css (lines 99-104) are a plain 120ms fade-out and a 180ms fade-in with no blur. Commit 68db7bb (\"Remove page transition blur\", 2026-03-31) removed the blur, and the guide was never updated.\n\n**Which stands, by the reader:** other, by the code\n\n**Opened by session (2026-10-11): stands.** Opened web/src/app.css at lines 98 to 104: a fade of 120 ms out and 180 ms in, with no blur. The guide still describes a blur.\n\nFrom [2026-10-11-pages-what-does-not-fit-ux-principles-and-design-guide-as-a-reader-returned-it.md](readers/2026-10-11-pages-what-does-not-fit-ux-principles-and-design-guide-as-a-reader-returned-it.md), F14.\n\n### `docs/design-guide.md:25`\n\n> `docs/design-guide.md:25` “All colors live in `:root` in `app.css`. Change one value → changes everywhere.”\n> `docs/ux-principles.md:382` “So the principle binds the surface, not the arithmetic: **every display must have vocabulary for \"I have no reading here.\"**”\n\n**The reader:** The promise does not hold for the colours that carry the epistemic states. In web/src the four semantic hexes appear hard-coded 58 times (#fbbf24 18, #f87171 21, #34d399 12, #60a5fa 7), against about five uses of the four semantic tokens. The badge's six states (green, blue, amber, red, gray, neutral, including the UNEXAMINED and UNSETTLED vocabulary P22 required) live as scoped classes in claim/[id]/+page.svelte. The P25 compartment marker is hard-coded warm rgba in ScopeSwitcher.svelte. None of these has an entry in the guide, so the palette where \"I have no reading\" should be defined does not mention it.\n\n**So, in the reader's words:** Move the badge-state and compartment-marker colours into tokens and into the guide, since they are what P22 and P25 make load-bearing.\n\n**How sure the reader was:** yes. I counted hex literals with rg in web/src, and read the badge CSS and ScopeSwitcher styles.\n\n**Opened by session (2026-10-11): stands.** Counted in web/src: the four semantic colours are written out 58 times and used through their tokens 6 times.\n\nFrom [2026-10-11-pages-open-question-ux-principles-and-design-guide-as-a-reader-returned-it.md](readers/2026-10-11-pages-open-question-ux-principles-and-design-guide-as-a-reader-returned-it.md), O16.\n\n### `docs/design-guide.md:34`\n\n> `docs/design-guide.md:34` “| `--dl-card` | `#1d2535` | Card/block solid background |”\n> `web/src/app.css:12` “--dl-card: rgba(29, 37, 53, 0.7); /* Card/block bg — 70% opacity lets dark bg breathe through */”\n\n**The reader** (a finding for what does not fit, kind `state`): The guide says the card token is the solid colour #1d2535. In app.css it is the same colour at 70% opacity, deliberately so the background shows through. Every .obsidian card therefore renders translucent, not solid. A reader copying the guide's value gets a different surface.\n\n**Which stands, by the reader:** other, by the code\n\n**Opened by session (2026-10-11): stands.** Opened web/src/app.css at line 12: the card colour is the same colour at 70 percent opacity, where the guide gives it as solid.\n\nFrom [2026-10-11-pages-what-does-not-fit-ux-principles-and-design-guide-as-a-reader-returned-it.md](readers/2026-10-11-pages-what-does-not-fit-ux-principles-and-design-guide-as-a-reader-returned-it.md), F15.\n\n### `docs/design-guide.md:165`\n\n> `docs/design-guide.md:165` “Every interactive element gets a **same-colored outer glow** on hover (100ms transition):”\n> `docs/ux-principles.md:441` “7. **Never treat mobile as degraded desktop.** Mobile users are the primary data source. Their evaluations produce the quality scores that desktop analysts rely on.”\n\n**The reader:** Anti-pattern 7 and P1 make the phone the main case (\"the 90% experience\"). The design guide never mentions phones, touch, tap targets, breakpoints or responsive layout, apart from one link to a research note at line 303. Its whole interaction language is hover: glow on hover, .breathe, .shimmer, .undercurrent's underline, and `cursor: help`. Hover does not exist on touch. The code has @media rules in 15 files, so responsive behaviour is being decided file by file, outside the guide.\n\n**So, in the reader's words:** Add a short touch section saying what each hover effect becomes on a phone, plus a minimum target size.\n\n**How sure the reader was:** fairly. I searched the guide for mobile, touch, tap, responsive, breakpoint and @media and counted @media files with rg. I did not test the pages on a phone.\n\n**Opened by session (2026-10-11): stands.** The design guide's only mention of phones is a link in its reading list, at line 303.\n\nFrom [2026-10-11-pages-open-question-ux-principles-and-design-guide-as-a-reader-returned-it.md](readers/2026-10-11-pages-open-question-ux-principles-and-design-guide-as-a-reader-returned-it.md), O18.\n\n### `docs/design-guide.md:212`\n\n> `docs/design-guide.md:212` “4. Concept keywords: `Attunement · Lenses · Bridging · Depth · Convergence · Vision` (lavender, 1.25rem, linked)”\n> `docs/design-guide.md:17` “**Page transitions**: View Transitions API — blur-out (200ms) on old, fade-in (300ms) on new. Smooth cross-page navigation.”\n\n**The reader:** The guide's \"Landing page (top → bottom)\" describes an earlier page: header, input, Full Picture banner, keywords, four paragraphs. The shipped landing (PublicHome.svelte) runs hero, conviction ladder, example map, constellation, about beats and three doors. Only then come the reading links, now with Frontier, and the input, now titled \"Analyze your own thinking\". A code comment there records the founder's 2026-09-19 instruction that reading come before the input box. The guide still shows the input second and calls transitions blur-out 200ms and fade-in 300ms. app.css has fade-out 120ms and fade-in 180ms with no blur. The guide was last committed 2026-09-21, for the map rows. The reordering also sits uneasily with Conviction 3 and P7, where the input box is the ground level.\n\n**So, in the reader's words:** Replace the landing list with the real order and the 09-19 ruling, or drop page-by-page layouts from the guide.\n\n**How sure the reader was:** yes on the facts. I read PublicHome.svelte 640-760 and app.css 97-108, and used git log dates.\n\n**Opened by session (2026-10-11): stands.** The landing component runs hero, ladder, example map, constellation, the beats, the reading links and then the input, and app.css fades out in 120 ms and in over 180 ms with no blur.\n\nFrom [2026-10-11-pages-open-question-ux-principles-and-design-guide-as-a-reader-returned-it.md](readers/2026-10-11-pages-open-question-ux-principles-and-design-guide-as-a-reader-returned-it.md), O14.\n\n### `docs/design-guide.md:212`\n\n> `docs/design-guide.md:212` “Concept keywords: `Attunement · Lenses · Bridging · Depth · Convergence · Vision` (lavender, 1.25rem, linked)”\n> `web/src/lib/components/PublicHome.svelte:710` “Founder, 2026-09-19: \\\"Move 'the full picture' and the attunement/lenses/bridging etc”\n\n**The reader** (a finding for what does not fit, kind `overtaken`): The guide's landing layout puts the input section second, then the Full Picture banner, then six keywords. The landing now opens on the example map and constellation, and the founder moved the banner and keywords above the input on 2026-09-19. The input is retitled \"Analyze your own thinking\", and there are seven keywords including Frontier (architecture-reference.md:300 also says seven). The whole numbered layout is out of date.\n\n**Which stands, by the reader:** other, by a ruling of 2026-09-19\n\n**Opened by session (2026-10-11): stands.** Opened web/src/lib/components/PublicHome.svelte at line 710: a comment quotes the founder's instruction of 2026-09-19 to move the banner and the keywords. The guide's numbered layout of the landing page predates it.\n\nFrom [2026-10-11-pages-what-does-not-fit-ux-principles-and-design-guide-as-a-reader-returned-it.md](readers/2026-10-11-pages-what-does-not-fit-ux-principles-and-design-guide-as-a-reader-returned-it.md), F16.\n\n### `docs/design-guide.md:282`\n\n> `docs/design-guide.md:282` “Green = supports/agree. Red = attacks/disagree. Amber = qualifies/evidence. Purple = reframes/decompose. Blue = empirical/decomposes.”\n> `docs/ux-principles.md:60` “Collapsing quality and stance into a single upvote/downvote destroys this signal.”\n\n**The reader:** The product's central distinction is that stance (voted), relation (supports or attacks) and quality (computed) must not be collapsed. The palette collapses them into the same four hues. In code the \"agree\" button uses --dl-success and the \"well supported\" QBAF badge uses the same #34d399, while the \"disagree\" button and the red badge share --dl-danger's hue. On the maps, blue is both the \"fact\" box and the \"splits\" line, and amber is the \"should\" box and the \"assumed\" and \"narrows\" lines. The guide has four claim types and mapLanguage.ts has three kinds. Key principle 4 says each colour has \"a dedicated\" meaning. Disagreement is also coded as danger, in a product that calls it \"the richest territory\".\n\n**So, in the reader's words:** Decide which dimension owns green and red, and give votes a distinct pair (or none), then rewrite key principle 4 to match.\n\n**How sure the reader was:** yes on the facts (app.css 336-360, claim page .qbaf-* classes, mapLanguage.ts). The harm to readers is my inference.\n\n**Opened by session (2026-10-11): stands.** app.css gives agree and supports one green and disagree and attacks one red, and the claim page's badge for well supported and for challenged uses the same two colours.\n\nFrom [2026-10-11-pages-open-question-ux-principles-and-design-guide-as-a-reader-returned-it.md](readers/2026-10-11-pages-open-question-ux-principles-and-design-guide-as-a-reader-returned-it.md), O15.\n\n### `docs/design-guide.md:282`\n\n> `docs/design-guide.md:282` “Amber = qualifies/evidence. Purple = reframes/decompose. Blue = empirical/decomposes.”\n> `docs/design-guide.md:175` “| Decompose CTA | `rgba(236, 72, 153, 0.2)` | Pink glow |”\n\n**The reader** (a finding for what does not fit, kind `rule-and-practice`): Key principle 4 says each meaning has one consistent colour and gives purple to decompose. The guide's own glow table and tokens give the Decompose CTA pink, and in the code the decompose invitation (SorryMarker) is pink. Only reframes is purple, and the DECOMPOSES_INTO relation is blue. A designer following principle 4 would colour a decompose invitation purple.\n\n**Which stands, by the reader:** other, by the code\n\n**Opened by session (2026-10-11): stands.** Opened web/src/lib/components/SorryMarker.svelte: the invitation to split a claim is pink, as the guide's own table says, where its key principle 4 gives purple.\n\nFrom [2026-10-11-pages-what-does-not-fit-ux-principles-and-design-guide-as-a-reader-returned-it.md](readers/2026-10-11-pages-what-does-not-fit-ux-principles-and-design-guide-as-a-reader-returned-it.md), F17.\n\n### `docs/design-guide.md:284`\n\n> `docs/design-guide.md:284` “5. **The graph is a map, not the product.** The D3 graph on claim pages is supplementary — the primary experience is card-based text. The graph is for orientation.”\n> `docs/design-guide.md:263` “| `ClaimGraph.svelte` | D3 force-directed ego-graph. No longer mounted on any page (replaced 2026-09-20) |”\n> `docs/ux-principles.md:54` “**The founder's brief of 2026-09-20 goes further than this principle, and the principle has not been re-ruled.**”\n\n**The reader:** Key principle 5 repeats P1 in its pre-09-20 form, 20 lines below the table row saying that D3 graph is unmounted. Its \"card-based text is the primary experience\" is the thing the founder's brief reversed (\"instead of the endless walls of text… more like a video game\"). ux-principles says P1 \"has not been re-ruled\", so the guide states as settled what its sibling says is unsettled. The two rules he gave for any map (as few boxes and as little text as possible; the wheel must not zoom without ⌘ or Ctrl) exist only inside one long P1 paragraph. The guide has no mention of them.\n\n**So, in the reader's words:** Re-rule P1 either way, then restate it in the guide and add the two map rules where a styling session will look.\n\n**How sure the reader was:** yes. I searched design-guide.md for ⌘, Ctrl and \"box\" and found nothing relevant, and rg confirmed ClaimGraph is mounted nowhere.\n\n**Opened by session (2026-10-11): stands.** ClaimGraph is mounted nowhere in web/src, and the page of principles says its first principle has not been ruled again since the brief of 2026-09-20. That the guide lacks the two rules for maps is the reader's search.\n\nFrom [2026-10-11-pages-open-question-ux-principles-and-design-guide-as-a-reader-returned-it.md](readers/2026-10-11-pages-open-question-ux-principles-and-design-guide-as-a-reader-returned-it.md), O17.\n\n### `docs/design-guide.md:284`\n\n> `docs/design-guide.md:284` “The D3 graph on claim pages is supplementary — the primary experience is card-based text.”\n> `docs/design-guide.md:263` “D3 force-directed ego-graph. No longer mounted on any page (replaced 2026-09-20)”\n\n**The reader** (a finding for what does not fit, kind `state`): Key principle 5 says a D3 graph on claim pages supplements a card-and-text primary experience. The same guide's component table says that graph is mounted on no page, and the code confirms it. Claim and source pages open on the computed ClaimMap, and the founder asked on 2026-09-20 that the map, not the lists, lead.\n\n**Which stands, by the reader:** other, by the code\n\n**Opened by session (2026-10-11): stands.** Opened lines 263 and 284 of the guide: its table says the older graph is on no page since 2026-09-20, and its key principle 5 still describes that graph on claim pages.\n\nFrom [2026-10-11-pages-what-does-not-fit-ux-principles-and-design-guide-as-a-reader-returned-it.md](readers/2026-10-11-pages-what-does-not-fit-ux-principles-and-design-guide-as-a-reader-returned-it.md), F18.\n\n## `docs/the-missing-layer.md`\n\n### `docs/the-missing-layer.md:3`\n\n> `docs/the-missing-layer.md:3` “It never built the layer for reasoning itself. That is the slot Deliberus fills.”\n> `docs/vision.md:305` “a commons in the lineage of Wikipedia and open proof libraries, carrying a falsifiable civilizational wager.”\n> `docs/vision.md:369` “that knowledge systems radically compress reasoning and the lost derivational scaffolding is recoverable as a verifiable corpus”\n\n**The reader:** The tagline says civilization never built durable, shared, inspectable structure for why. vision.md's own lineage names open proof libraries, and the Lean analogy runs through all three notes. In mathematics that structure exists and is computed over. vision.md also concedes a heavyweight group (SciencePedia) built a reasoning corpus at very large scale for the verifiable-STEM slice. The true claim is scoped to contested, normative reasoning with no verifiable endpoint. The missing-layer page is the one without that scope, and its list of four kinds omits formal proof libraries. Compression drops the scope clause first (ledger J8).\n\n**So, in the reader's words:** Add \"for contested reasoning with no verifiable endpoint\" to the tagline and the \"What does not exist\" heading.\n\n**How sure the reader was:** yes for the omission, since the note has no Lean, proof or argument-mining mention. Fairly on how much it matters to a reader.\n\n**Opened by session (2026-10-11): stands.** The body of the page has no mention of Lean, proof libraries or argument mining.\n\nFrom [2026-10-11-pages-open-question-three-vision-pages-as-a-reader-returned-it.md](readers/2026-10-11-pages-open-question-three-vision-pages-as-a-reader-returned-it.md), O17.\n\n### `docs/the-missing-layer.md:23`\n\n> `docs/the-missing-layer.md:23` “none of it keeps itself honest**. Peer review does not review itself.”\n> `docs/the-missing-layer.md:15` “Adversarial structure, rules of evidence, cross-examination, attributed judgment, appeal.”\n> `docs/the-missing-layer.md:23` “An infrastructure layer that audits its own distortions has no precedent at all.”\n> `deliberus/disagreement_preservation.py:17` “an alarm, not a proof — the honest posture for an instrument guarding against its own bias.”\n\n**The reader:** \"None of it keeps itself honest\" is over-broad by the page's own list. Courts have appeal, science has replication and review (line 15), and Metaculus is described as calibrated, which is a form of self-audit. What none measures is its own flattening of disagreement. That narrower claim is the one the preservation instrument supports. Its docstring calls itself an alarm, not a proof, and says it is silent where linking failed. CLAUDE.md:800 already says the register is \"keeps itself honest\", never \"measures its own distortions\", yet line 23 says \"audits its own distortions has no precedent\". This page was woven into the funding drafts.\n\n**So, in the reader's words:** Narrow to \"no other layer measures whether it flattens the disagreements it holds\", and drop \"audits its own distortions\".\n\n**How sure the reader was:** fairly. I read CLAUDE.md:800 and the instrument's docstring. I did not check outside whether Wikipedia or Metaculus audit themselves.\n\n**Opened by session (2026-10-11): stands.** The instruction file gives the register as keeps itself honest and never measures its own distortions, and line 23 says audits its own distortions. The same line is sentence 6 of the walk-through, already with the founder.\n\nFrom [2026-10-11-pages-open-question-three-vision-pages-as-a-reader-returned-it.md](readers/2026-10-11-pages-open-question-three-vision-pages-as-a-reader-returned-it.md), O16.\n\n### `docs/the-missing-layer.md:27`\n\n> `docs/the-missing-layer.md:27` “Deliberus is not a better member of an existing class. It is built for the empty slot.”\n> `docs/the-missing-layer.md:35` “They demanded structure as *input*, and hand-building argument graphs costs more than any user will pay.”\n> `docs/vision.md:343` “**That's not a feature. It's the soul of the product.**”\n> `docs/vision.md:9` “Neither alone is Deliberus — the analysis serves the attunement, and the attunement motivates the analysis.”\n\n**The reader:** README calls this page the positioning backbone, and SESSION-HISTORY says it was woven into all six funding drafts. It describes only the analytic pole: extraction, strength, honesty instruments. Outside the See-also link it has no word about worldview lenses or attunement. vision.md calls the lens \"the soul of the product\" and says analysis alone is the neutrality illusion. It also gives two reasons predecessors failed, no motivational hook and formal structure demanded as input. The positioning page gives only the second. The page telling funders which slot is filled leaves out what the vision says makes people use it.\n\n**So, in the reader's words:** Add two sentences on the lens, or state on the page that it is deliberately the analysis-side pitch.\n\n**How sure the reader was:** yes for the omission, since a search for attun, lens, worldview, glasses and inhabit finds only the See-also link. Fairly on significance.\n\n**Opened by session (2026-10-11): stands.** Outside its See-also line the page has no word for lenses, worldviews or attunement.\n\nFrom [2026-10-11-pages-open-question-three-vision-pages-as-a-reader-returned-it.md](readers/2026-10-11-pages-open-question-three-vision-pages-as-a-reader-returned-it.md), O18.\n\n### `docs/the-missing-layer.md:29`\n\n> `docs/the-missing-layer.md:29` “Claim strength is computed from the graph by published argumentation mathematics, never by a model's opinion, so a score is always”\n> `docs/research/relevance-and-the-two-factors.md:25` “**The supporting claim's own strength never reaches the”\n> `docs/vision.md:351` “The shipped feed ranks by a rating the classifier gives each link at extraction, which no prompt defines,”\n> `web/src/lib/constellation.ts:405` “So a claim’s strength is computed by published mathematics from the arguments around it, never by a model’s opinion”\n\n**The reader:** The page's promise about scores is stronger than the project's own measurements of 28 September. The relevance note says a supporting claim's own strength never reaches its target, unlike the published model the page cites. The stored per-link number is a machine's rating with no definition at any writer. It is printed as a percentage and used to rank four feeds (what-a-link-strength-means.md § 0). vision.md conceded the feed point on 17 August. The page was edited on 10 October for a sibling overstatement and this sentence stayed. The same words are in a draft wall tile.\n\n**So, in the reader's words:** Scope the sentence to what is true today (the badge, computed from answered critical questions) until the Q31 and Q33 builds land, on the page and the tile.\n\n**How sure the reader was:** fairly. I read the relevance and strength notes and ledger Q31 and Q33. The badge alone may justify a narrower version of the sentence.\n\n**Opened by session (2026-10-11): stands in part.** The stored rating of a link still ranks the feeds, as vision.md line 351 says and as the bridging query read earlier the same day shows. The other half is overtaken: since 2026-09-30 a link carries its supporter's own strength (ledger Q33), and the quoted line of the relevance note describes the state before that.\n\nFrom [2026-10-11-pages-open-question-three-vision-pages-as-a-reader-returned-it.md](readers/2026-10-11-pages-open-question-three-vision-pages-as-a-reader-returned-it.md), O15.\n\n### `docs/the-missing-layer.md:41`\n\n> `docs/the-missing-layer.md:41` “The fitting measures are therefore collective: coverage, whether mapped concepts get reused across unrelated debates”\n> `docs/the-missing-layer.md:43` “concept reuse across the corpus is healthy while reuse at the level of individual claims is close to absent.”\n\n**The reader:** The August correction is careful. The group trial is kept as a failure detector, and the warrant moves to collective measures. But none of the measures has a value that would count as failure. The page admits one already looks bad, claim-level reuse, and then says \"A layer is proven by use\". Coverage has no stated denominator: what is \"the contested space\"? The discipline asked of the gate (write done-ish down first) applies here too. The project pre-registers predictions for its passes and has not done so for the layer itself.\n\n**So, in the reader's words:** For each collective measure, write down and date the value that would count as the layer failing, before the next reading.\n\n**How sure the reader was:** fairly. I looked for thresholds on the page and in frontier.md and found none.\n\n**Opened by session (2026-10-11): stands.** The page gives no value at which any of its collective measures would count as the layer failing. The reader's search of the frontier page was not repeated.\n\nFrom [2026-10-11-pages-open-question-three-vision-pages-as-a-reader-returned-it.md](readers/2026-10-11-pages-open-question-three-vision-pages-as-a-reader-returned-it.md), O19.\n\n## `docs/ux-principles.md`\n\n### `docs/ux-principles.md:5`\n\n> `docs/ux-principles.md:5` “is measured, and replicated five times, is **faster triage**: average time on one screen fell from”\n> `docs/ux-principles.md:9` “own value. And the answer to rendering a dense graph is **slicing, not zooming**: one argument on”\n> `docs/ux-principles.md:52` “a computed map with a box budget of about ten, supporters to the left of the claim and challengers to the right, three zoom steps, and a click that travels.”\n\n**The reader:** What was replicated five times is how long people stay on one screen. \"Faster triage\" is a reading of that number, and the research note credits the phrase to reviewers. Dwell time alone cannot tell \"decide sooner what deserves focus\" from \"cannot hold focus\", and the banner's design consequence (brevity paid once, depth after) depends on which is true. The banner also promises \"slicing, not zooming\", yet P1 keeps \"semantic zoom\" (line 48) and the 09-20 map has \"three zoom steps\". Neither of those mentions the banner.\n\n**So, in the reader's words:** Label \"faster triage\" the working reading, and name what readers do past the first screen that would show it wrong; the banner itself promises that measurement discipline.\n\n**How sure the reader was:** fairly. I read game-feel-and-the-first-screen.md lines 28-50 (44/47/50 s dwell figures, triage wording attributed to reviewers). I did not check Mark's own interpretation of her data.\n\n**Opened by session (2026-10-11): stands.** The research note behind the banner gives the three dwell figures and calls triage the reviewers' phrasing (game-feel-and-the-first-screen.md). The page keeps semantic zoom in its first principle and three zoom steps in the map, beside the banner's slicing, not zooming.\n\nFrom [2026-10-11-pages-open-question-ux-principles-and-design-guide-as-a-reader-returned-it.md](readers/2026-10-11-pages-open-question-ux-principles-and-design-guide-as-a-reader-returned-it.md), O1.\n\n### `docs/ux-principles.md:18`\n\n> `docs/ux-principles.md:18` “Nothing here prescribes a specific UI framework or visual design — these are the principles that any implementation must satisfy.”\n> `docs/ux-principles.md:418` “A warm tinted marker, deliberately not the indigo of the rest of the chrome, whenever you are not.”\n\n**The reader** (a finding for what does not fit, kind `counter-case`): The status line says the page prescribes no visual design and leaves any implementation free. P25, ruled 2026-09-22, prescribes a warm tinted marker deliberately not indigo, and ScopeSwitcher.svelte builds it that way. A builder who follows the status line and draws the compartment marker in the chrome's indigo breaks P25.\n\n**Which stands, by the reader:** other, by a ruling of 2026-09-22\n\n**Opened by session (2026-10-11): stands.** Both lines are on the page: the status line says no visual design is prescribed, and principle 25 prescribes a warm tinted marker that is deliberately not indigo. A small clash inside one page.\n\nFrom [2026-10-11-pages-what-does-not-fit-ux-principles-and-design-guide-as-a-reader-returned-it.md](readers/2026-10-11-pages-what-does-not-fit-ux-principles-and-design-guide-as-a-reader-returned-it.md), F2.\n\n### `docs/ux-principles.md:28`\n\n> `docs/ux-principles.md:28` “Every structured argumentation platform that died (20+ since 2007) had adequate technology and inadequate UX. The graveyard proves that logical rigor without experiential delight is a dead end.”\n> `docs/ux-principles.md:18` “The convictions predate the research. The research validates, deepens, and concretizes them.”\n\n**The reader:** Conviction 1 says the cemetery \"proves\" thin UX killed every predecessor. The note it cites lists Kialo with a \"Clean UI\" and over a million registered users, stalled on revenue, format and visibility. That note's section 2 opens with an August 2026 reframe: the friction diagnosis \"was right\", LLM extraction has removed it, and exposure and social cost, which \"neither yields to better tooling\", are now the binding constraints. So the research did not simply validate the conviction; it narrowed it. Conviction 4 rests on the same ground. The page cites incentives-analysis only under P21.\n\n**So, in the reader's words:** Put a one-line pointer to the reframe under conviction 1 and replace \"proves\" with what the cemetery does show: good UX is necessary, and it is no longer what binds.\n\n**How sure the reader was:** fairly. I read adoption-problem.md lines 7-30 and 72-74.\n\n**Opened by session (2026-10-11): stands.** adoption-problem.md § 2 opens with the reframing of August 2026: the friction diagnosis was right, extraction has removed it, and exposure and social cost now bind. The same note lists Kialo as surviving with a clean interface.\n\nFrom [2026-10-11-pages-open-question-ux-principles-and-design-guide-as-a-reader-returned-it.md](readers/2026-10-11-pages-open-question-ux-principles-and-design-guide-as-a-reader-returned-it.md), O2.\n\n### `docs/ux-principles.md:62`\n\n> `docs/ux-principles.md:62` “**Implementation decision (Mar 31, 2026 — Option B)**: Single-axis human voting (agree/disagree only).”\n> `docs/ux-principles.md:293` “6. Ask the user to rate the counter-argument's QUALITY (not agreement) — teaches the two-axis distinction”\n> `docs/ux-principles.md:441` “7. **Never treat mobile as degraded desktop.** Mobile users are the primary data source. Their evaluations produce the quality scores that desktop analysts rely on.”\n> `docs/design-guide.md:306` “- [docs/research/epistemic-gamification.md](../epistemic-gamification.md) — calibration rewards, two-axis voting, reputation architecture”\n\n**The reader:** P2's Option B, and the 2026-09-09 correction of anti-pattern 2, say quality is computed from answered critical questions and never voted. Three other places still describe people rating quality. P16 step 6 teaches \"the two-axis distinction\" by having a newcomer rate quality, and line 295 calls that the first meaningful contribution. Anti-pattern 7 says mobile users' \"evaluations produce the quality scores\". The design guide advertises \"two-axis voting\". The correction reached P2 and anti-pattern 2 and not their neighbours. P16 is the planned first-impression hook, so the stale step matters most there.\n\n**So, in the reader's words:** Rewrite P16 step 6 and the anti-pattern 7 clause to the computed-quality model, and decide what the newcomer does instead (answer one critical question?).\n\n**How sure the reader was:** yes. I searched \"two-axis\" across docs/*.md, and web/src and deliberus/ have no onboarding-hook code.\n\n**Opened by session (2026-10-11): stands.** The page's second principle says quality is computed and never voted. Step 6 of the onboarding principle, anti-pattern 7 and the design guide's reading list still describe people rating quality.\n\nFrom [2026-10-11-pages-open-question-ux-principles-and-design-guide-as-a-reader-returned-it.md](readers/2026-10-11-pages-open-question-ux-principles-and-design-guide-as-a-reader-returned-it.md), O3.\n\n### `docs/ux-principles.md:103`\n\n> `docs/ux-principles.md:103` “testable rule for every gap-displaying surface, including the completeness oracle's `unsupported_value_premises`: **no gap displayed without a move offered.**”\n> `TODO.md:1758` “it shows `?` marks and `+N` counts and offers nothing beside them.”\n\n**The reader** (a finding for what does not fit, kind `rule-and-practice`): P4 states the rule for every gap-displaying surface: no gap without a move offered. The shipped claim and source maps show `?` marks and `+N` counts and offer no move beside them (MapCanvas.svelte readout has only \"Go here\"). The map breaks the rule on the page's most visible new surface, and P4 and P1 do not say so; only P24 says the maps offer no way to contribute.\n\n**Which stands, by the reader:** target, by nothing I found\n\n**Opened by session (2026-10-11): stands.** The to-do list holds the prompts on the map as unbuilt, and the project's instructions name the map as a standing breach of the rule. Principle 4 on the page does not say so.\n\nFrom [2026-10-11-pages-what-does-not-fit-ux-principles-and-design-guide-as-a-reader-returned-it.md](readers/2026-10-11-pages-what-does-not-fit-ux-principles-and-design-guide-as-a-reader-returned-it.md), F4.\n\n### `docs/ux-principles.md:123`\n\n> `docs/ux-principles.md:123` “**Blue**: Claim stated, ready for evidence — all prerequisite claims are green”\n> `docs/ux-principles.md:384` “now reads `UNSETTLED` in a neutral shade — never an approval word in confident blue.”\n\n**The reader** (a finding for what does not fit, kind `word-two-senses`): P5 gives blue the meaning \"stated, not yet supported, you can contribute here\". P22, and display_band in qbaf.py, use blue for the positive supported band (strength 0.52 to 0.8), the colour of an approval word. A reader looking at a blue claim cannot tell \"needs evidence\" from \"supported\".\n\n**Which stands, by the reader:** other, by the code\n\n**Opened by session (2026-10-11): stands.** Both lines are on the page: principle 5 gives blue to a claim that is stated and waits for evidence, and principle 22 calls blue the colour of an approval word. The strength code was not opened for this.\n\nFrom [2026-10-11-pages-what-does-not-fit-ux-principles-and-design-guide-as-a-reader-returned-it.md](readers/2026-10-11-pages-what-does-not-fit-ux-principles-and-design-guide-as-a-reader-returned-it.md), F5.\n\n### `docs/ux-principles.md:141`\n\n> `docs/ux-principles.md:141` “`format_context_for_synthesis` truncates *again* to 3 / 5 / 8 supporting and 2 / 3 / 5 challenging.”\n> `deliberus/synthesis_ledger.py:62` “# Raised for curious and expert on 2026-08-13. Casual stays tight on purpose: it is the”\n\n**The reader** (a finding for what does not fit, kind `overtaken`): P6 gives the display budgets as 3/5/8 supporting and 2/3/5 challenging and tells the reader to resist raising the limits. The code raised the curious and expert budgets the same day (3/8/16 supporting, 2/5/12 challenging). The page says the caps are \"now set separately\" but never gives the new numbers, so a reader takes 3/5/8 as current.\n\n**Which stands, by the reader:** other, by the code\n\n**Opened by session (2026-10-11): stands.** Opened deliberus/synthesis_ledger.py: the limits are 3, 8 and 16 supporting and 2, 5 and 12 challenging since 2026-08-13. The page still gives 3, 5, 8 and 2, 3, 5.\n\nFrom [2026-10-11-pages-what-does-not-fit-ux-principles-and-design-guide-as-a-reader-returned-it.md](readers/2026-10-11-pages-what-does-not-fit-ux-principles-and-design-guide-as-a-reader-returned-it.md), F6.\n\n### `docs/ux-principles.md:169`\n\n> `docs/ux-principles.md:169` “Notify when someone challenged your reasoning, not when someone agreed with it. Never notify for raw engagement metrics.”\n> `docs/ux-principles.md:103` “This yields a testable rule for every gap-displaying surface, including the completeness oracle's `unsupported_value_premises`: **no gap displayed without a move offered.**”\n\n**The reader:** P4 states the PACE rule as a test for \"every gap-displaying surface\". P9 then designs a notification stream whose only triggers are \"your argument has a counter-argument\", \"a premise you support has been flagged\" and \"the community has downgraded a key premise\". All of them tell the reader a gap exists. P9 says nothing about the move each carries, and it bans the one kind of notification, agreement, that is never a gap. By P4's own mechanism that stream trends toward anxiety-related inhibition rather than approach. The rule was not carried over to the page's own notification section.\n\n**So, in the reader's words:** Add to P9 that every notification names its move, and decide whether any non-adversarial notification (your question was answered) is allowed.\n\n**How sure the reader was:** a guess. This is an inference from P4's rule. deliberus/notifications.py is only operator push alerts, so there is no reader-facing notification code to check.\n\n**Opened by session (2026-10-11): stands.** The three notification triggers are as quoted and none names a move. That such a stream would discourage is the reader's inference from the fourth principle: the only notification code sends the keeper's own alerts, so there is nothing to test it against.\n\nFrom [2026-10-11-pages-open-question-ux-principles-and-design-guide-as-a-reader-returned-it.md](readers/2026-10-11-pages-open-question-ux-principles-and-design-guide-as-a-reader-returned-it.md), O5.\n\n### `docs/ux-principles.md:262`\n\n> `docs/ux-principles.md:262` “### P14: The Interpassivity Warning — LLMs Scaffold, Never Replace”\n> `docs/ux-principles.md:264` “systems that do cognitive work FOR users rob them of the epistemic engagement that IS the point”\n> `docs/ux-principles.md:443` “9. **Scaffold, never replace.** ⚠ **This headline is under revision and there is currently no”\n\n**The reader:** The \"under revision\" flag sits on anti-pattern 9, 180 lines from P14. P14 itself carries no flag. Its heading still reads \"LLMs Scaffold, Never Replace\", which is the superseded wording (ledger J9). Its first sentence is the delegation-hostile frame. conviction-and-critique.md lines 141-149 records that this frame already caused one measured regression and says \"P14's own body still leads with the rejected frame, which is where the next regression will come from\". The rewrite is open as ledger R14. The reader or session who reads P14 gets no warning.\n\n**So, in the reader's words:** Put a two-line banner on P14 now, pointing to the naming note and R14, rather than waiting for the rewrite ruling.\n\n**How sure the reader was:** yes. I read conviction-and-critique.md 141-152 and session26-open-rulings.md R9/R14; the page's only other mention of the problem is the unruled \"founder input\" in the naming note.\n\n**Opened by session (2026-10-11): stands.** conviction-and-critique.md records the regression of 2026-08-30 and says the body of P14 still leads with the rejected frame. On the page the warning stands at anti-pattern 9 only. Ledger rows R9 and R14 hold the open proposals.\n\nFrom [2026-10-11-pages-open-question-ux-principles-and-design-guide-as-a-reader-returned-it.md](readers/2026-10-11-pages-open-question-ux-principles-and-design-guide-as-a-reader-returned-it.md), O6.\n\n### `docs/ux-principles.md:293`\n\n> `docs/ux-principles.md:293` “Ask the user to rate the counter-argument's QUALITY (not agreement) — teaches the two-axis distinction”\n> `docs/ux-principles.md:62` “Single-axis human voting (agree/disagree only). The QBAF badge IS the quality signal — computed from CQ evidence, not subjective votes.”\n\n**The reader** (a finding for what does not fit, kind `overtaken`): P16's onboarding hook asks a new user to rate an argument's quality, and calls that the first contribution. P2 (Option B, 2026-03-31) and anti-pattern 2 hold that humans vote only agree/disagree and quality is computed from answered critical questions. A user cannot be asked to rate quality in a system that has no quality vote to take.\n\n**Which stands, by the reader:** other, by a ruling of 2026-03-31\n\n**Opened by session (2026-10-11): stands.** Opened both lines: step 6 of the first-use path asks a newcomer to rate an argument's quality, and principle 2 on the same page says people vote agree or disagree only.\n\nFrom [2026-10-11-pages-what-does-not-fit-ux-principles-and-design-guide-as-a-reader-returned-it.md](readers/2026-10-11-pages-what-does-not-fit-ux-principles-and-design-guide-as-a-reader-returned-it.md), F8.\n\n### `docs/ux-principles.md:299`\n\n> `docs/ux-principles.md:299` “### P17: Feed as Dashboard, Not Timeline — The Organic Redraw”\n> `docs/ux-principles.md:304` “- **60% quality threshold** — only arguments meeting minimum quality appear in primary feed”\n\n**The reader:** P1, P3, P4, P6, P15, P18, P20 and P22 carry dated implementation notes, but P8, P16, P17 and P19 read as how the product works. They are not built. deliberus/feed.py lists Bayesian surprise under \"Future\", has no 60% gate and no coverage guarantee, and defaults to needs-help. needs-help ranks claims with the most unanswered questions, which is close to the opposite of a quality gate. Under P22 a claim with nothing answered is \"unexamined\", not 0.5 quality, so a 60% gate would also exclude every new claim. I found no code for the thinking hats or the onboarding hook.\n\n**So, in the reader's words:** Give every principle a one-line status (built, partly, design only), starting with P8, P16, P17 and P19, and drop or restate the 60% figure.\n\n**How sure the reader was:** yes on the code (read feed.py 1-58, rg for hats and onboarding in web/src and deliberus). The P22 interaction is my inference.\n\n**Opened by session (2026-10-11): stands.** deliberus/feed.py lists Bayesian surprise under Future and has no quality threshold and no coverage rule. That no code exists for the thinking hats or the onboarding hook is the reader's search and was not repeated.\n\nFrom [2026-10-11-pages-open-question-ux-principles-and-design-guide-as-a-reader-returned-it.md](readers/2026-10-11-pages-open-question-ux-principles-and-design-guide-as-a-reader-returned-it.md), O7.\n\n### `docs/ux-principles.md:350`\n\n> `docs/ux-principles.md:350` “searching it for *side-by-side*, *synchronous* or *co-present* returns nothing; everything documented is either solo or asynchronous”\n\n**The reader:** This is an absence claim made by search, written in the present tense, on a page that now contains the searched words (co-present appears 7 times here, and 43 files in docs/ contain one of the three terms). It was true on 2026-08-17 and became false when P21 was written; line 360 gets the tense right (\"returned nothing\"), line 350 does not. A reader who repeats the search finds this page first. The page's own audit paragraph shows how an \"it's missing\" claim turns into a \"it was findability\" claim, which is the same hazard.\n\n**So, in the reader's words:** Date the claim in line 350 (\"on 2026-08-17 returned nothing\") or delete the parenthesis.\n\n**How sure the reader was:** yes. I ran rg -l -i \"co-present|side-by-side|synchronous\" docs (43 files) and counted co-present in this page (7).\n\n**Opened by session (2026-10-11): stands.** The page now holds the word co-present seven times, and 44 files under docs/ hold one of the three words the sentence says return nothing.\n\nFrom [2026-10-11-pages-open-question-ux-principles-and-design-guide-as-a-reader-returned-it.md](readers/2026-10-11-pages-open-question-ux-principles-and-design-guide-as-a-reader-returned-it.md), O8.\n\n### `docs/ux-principles.md:427`\n\n> `docs/ux-principles.md:427` “### P24 (CANDIDATE, unratified — proposed 2026-09-10): Invite the Reframe”\n> `docs/ux-principles.md:429` “Founder ruling 2026-09-10, verbatim: *\"Sure. Yes. This is one of the central ways for humans to contribute after all (read up on that), so I agree,”\n> `docs/ux-principles.md:577` “| P23: Addressability | [structure-versus-scale.md](../structure-versus-scale.md), [interaction-modalities-and-the-pointing-test.md](../interaction-modalities-and-the-pointing-test.md) |”\n\n**The reader:** P24 is labelled CANDIDATE, unratified, yet its first line is a founder ruling, and ledger K5/K6 are RULED 2026-09-10 \"in full\" with a build item. So the reader cannot tell what is unratified: the wording, the number, or the placement. P25 (ruled 09-22) sits above P24 in the file. Neither appears in the Research Foundation table, which stops at P23. design-guide.md:300 still says \"23 UX principles\".\n\n**So, in the reader's words:** Say in P24's heading exactly what is still unratified (or promote it), put P24 before P25, and extend the table.\n\n**How sure the reader was:** fairly. I read the K5/K6 rows in the settled file; I did not check whether the numbered-principle wording was itself ruled.\n\n**Opened by session (2026-10-11): stands.** Ledger rows K5 and K6 are ruled 2026-09-10. The table of research foundations ends at P23, the page has 25 principle headings, and the design guide says 23.\n\nFrom [2026-10-11-pages-open-question-ux-principles-and-design-guide-as-a-reader-returned-it.md](readers/2026-10-11-pages-open-question-ux-principles-and-design-guide-as-a-reader-returned-it.md), O9.\n\n### `docs/ux-principles.md:438`\n\n> `docs/ux-principles.md:438` “4. **Never present AI-generated structure as authoritative.** The confirmation step is sacred. The user always has the last word on what they meant.”\n> `docs/ux-principles.md:72` “The confirmation step is non-optional — it is the trust layer.”\n\n**The reader:** The confirmation step does three jobs on this page: an author checking a rendering of their own words, a person checking a machine's split of someone else's claim, and a click that stands in for thought. The later rulings kept only the first as sacred. On 2026-08-31 he blessed \"no ratify act\", and on 2026-09-21 the design became \"the machine writes what it found, and a person corrects\". split-review-and-the-founders-intentions.md lines 336-340 already reads anti-pattern 4 as covering only the author's own words. The page never says that, so \"sacred\" and \"non-optional\" read as universal.\n\n**So, in the reader's words:** Add the scope clause to anti-pattern 4 and P3 (the author confirming their own words), which is what the headline-test rule asks of a bolded refusal.\n\n**How sure the reader was:** fairly. I read the cited section of the split-review note and CLAUDE.md line 857. Neither note and neither ux-principles passage mentions the write-then-review sittings.\n\n**Opened by session (2026-10-11): stands.** split-review-and-the-founders-intentions.md reads anti-pattern 4 as an author confirming a rendering of the author's own words. The page states it, and the third principle's non-optional, without that scope.\n\nFrom [2026-10-11-pages-open-question-ux-principles-and-design-guide-as-a-reader-returned-it.md](readers/2026-10-11-pages-open-question-ux-principles-and-design-guide-as-a-reader-returned-it.md), O10.\n\n### `docs/ux-principles.md:452`\n\n> `docs/ux-principles.md:452` “## Why the inversion principle outranks everything else here (2026-08-27)”\n> `docs/ux-principles.md:455` “inversion principle — *the machine holds the user's hand through the descent* — as **the single most”\n> `docs/ux-principles.md:101` “**The inversion principle** (Jul 6, 2026, from the first dogfood run): the correction UX as shipped still has the user holding the machine's hand”\n> `docs/ux-principles.md:20` “They are invisible to anyone scanning the `### P` headings, which is a findability problem rather than an absence”\n\n**The reader:** By its own account the inversion principle is the most success-critical commitment in the whole system, and the always-loaded CLAUDE.md line points readers here for it. It has no heading, no number and no row in the Research Foundation table. It is a bold run-in inside P4's sixth paragraph (line 101), and the section arguing it outranks everything sits after the anti-patterns. That is the exact burial the 2026-08-17 audit repaired for P7a. Meanwhile unbuilt P19 (thinking hats) has a full heading.\n\n**So, in the reader's words:** Give it a heading (a P4a, as P7a was done) and a table row, so the pointer from CLAUDE.md lands on it.\n\n**How sure the reader was:** yes. I read the heading list (rg \"^#{1,4} \") and searched \"inversion\" in the page; it appears only at lines 101 and 452.\n\n**Opened by session (2026-10-11): stands.** The word inversion occurs on the page only inside the sixth paragraph of the fourth principle and in the section after the anti-patterns. It has no heading and no row in the table of research foundations.\n\nFrom [2026-10-11-pages-open-question-ux-principles-and-design-guide-as-a-reader-returned-it.md](readers/2026-10-11-pages-open-question-ux-principles-and-design-guide-as-a-reader-returned-it.md), O4.\n\n### `docs/ux-principles.md:477`\n\n> `docs/ux-principles.md:477` “## Naming note on principle 9, and the layer collision it exposed (2026-08-27)”\n> `docs/ux-principles.md:167` “### P9: Notifications Are Epistemic, Not Engagement-Maximizing”\n> `docs/ux-principles.md:507` “| **Principle 9** | whether a person's confirmation carries any judgment | the **person's** act of attending |”\n\n**The reader:** On this page \"P9\" is the notifications principle and \"principle 9\" is anti-pattern 9. The naming note's heading, its prose and its comparison table all say \"principle 9\", while everything above is numbered P1 to P25. Ledger row J9 had to be corrected because this exact ordinal confusion made a row name the wrong list. The bare ordinal the correction warned about is still in the page that caused it. A reader scanning the heading list for \"9\" lands on notifications.\n\n**So, in the reader's words:** Rename the heading and table cell \"anti-pattern 9\" (and \"anti-pattern 9 (P14)\" where it is cross-referenced).\n\n**How sure the reader was:** yes. I searched for \"principle 9\" and P9 on the page: lines 167, 477, 502 and 507.\n\n**Opened by session (2026-10-11): stands.** The words principle 9 stand at lines 477, 502 and 507 for anti-pattern 9, on a page where P9 is the principle about notifications.\n\nFrom [2026-10-11-pages-open-question-ux-principles-and-design-guide-as-a-reader-returned-it.md](readers/2026-10-11-pages-open-question-ux-principles-and-design-guide-as-a-reader-returned-it.md), O11.\n\n### `docs/ux-principles.md:480`\n\n> `docs/ux-principles.md:480` “The reasoning is recorded here rather than in the session's ruling”\n> `docs/ux-principles.md:512` “**Proposal on the table, unruled**: headline **\"Confirming is not checking.\"** — self-decoding,”\n> `docs/ux-principles.md:444` “proposal on the table** — *\"Confirming is not checking\"* was withdrawn 2026-08-29 for failing the”\n> `docs/ux-principles.md:485` “stranger test** — which is how the two-test discipline for rule headlines was discovered”\n\n**The reader:** Read top to bottom, the naming note presents \"Confirming is not checking\" as a live proposal. Line 444 says it was withdrawn on 08-29, and nothing in the note's body says so. The note also says the arc found a \"two-test discipline\". The arc ended with three tests (J13): the collision test was earned against this very headline. The reason \"scaffold\" failed (two ratified senses that disagree about whether it comes down) is in ledger J9, not here. The note says it exists so the argument \"should outlive\" the ledger, but the ledger rows now hold the argument's ending.\n\n**So, in the reader's words:** Add a dated status line at the head of the note pointing at J9 to J13, and move the collision finding into it.\n\n**How sure the reader was:** yes. I compared the note with settled rows J9, J10, J13 and with session26-open-rulings.md; I searched the page for \"collision\" (only the unrelated \"layer collision\").\n\n**Opened by session (2026-10-11): stands.** Line 444 says the proposed headline was withdrawn on 2026-08-29 while the note's body still calls it on the table. The note speaks of two tests where three were ruled and has no word on the third.\n\nFrom [2026-10-11-pages-open-question-ux-principles-and-design-guide-as-a-reader-returned-it.md](readers/2026-10-11-pages-open-question-ux-principles-and-design-guide-as-a-reader-returned-it.md), O12.\n\n### `docs/ux-principles.md:512`\n\n> `docs/ux-principles.md:512` “**Proposal on the table, unruled**: headline **\\\"Confirming is not checking.\\\"** — self-decoding,”\n> `docs/ux-principles.md:444` “proposal on the table** — *\\\"Confirming is not checking\\\"* was withdrawn 2026-08-29 for failing the”\n\n**The reader** (a finding for what does not fit, kind `overtaken`): The naming note presents \"Confirming is not checking\" as the proposal on the table, unruled. Anti-pattern 9's header, and ledger row J10, say it was withdrawn 2026-08-29 for failing the stranger test and that no proposal is on the table. The note carries no mark of the withdrawal, so a reader who stops at the note thinks a live proposal awaits a ruling.\n\n**Which stands, by the reader:** other, by a ruling of 2026-08-29\n\n**Opened by session (2026-10-11): stands.** Opened lines 444 and 512: the first says the headline was withdrawn on 2026-08-29, the second still presents it as the proposal on the table.\n\nFrom [2026-10-11-pages-what-does-not-fit-ux-principles-and-design-guide-as-a-reader-returned-it.md](readers/2026-10-11-pages-what-does-not-fit-ux-principles-and-design-guide-as-a-reader-returned-it.md), F12.\n\n### `docs/ux-principles.md:541`\n\n> `docs/ux-principles.md:541` “**One historical question, checked and unresolved**: whether Deutsch and Feynman ever publicly”\n\n**The reader:** Ledger J16, closed 2026-09-09, says \"the anecdote stays out of every doc\". This page is the one doc in docs/ that still carries it, as a full paragraph on the public site. The paragraph makes unsourced biographical claims about two real people (Deutsch's motive for the quantum computer, Feynman's, \"far apart on interpretation\") in answer to a question no visitor has asked. Either J16 meant only the fight, in which case a null-result paragraph can stay, or it meant the whole matter, in which case this paragraph should go. The ledger does not say which.\n\n**So, in the reader's words:** Decide which J16 meant, then cut the paragraph or source its two claims.\n\n**How sure the reader was:** fairly. rg for Deutsch-and-Feynman across docs finds it only here, the J16 row in session26-open-rulings.md, and a reader file.\n\n**Opened by session (2026-10-11): stands.** Ledger row J16 closes with the words that the anecdote stays out of every doc, and the paragraph is on this page. Which of the two the row meant is the founder's to say.\n\nFrom [2026-10-11-pages-open-question-ux-principles-and-design-guide-as-a-reader-returned-it.md](readers/2026-10-11-pages-open-question-ux-principles-and-design-guide-as-a-reader-returned-it.md), O13.\n\n## `docs/vision.md`\n\n### `docs/vision.md:86`\n\n> `docs/vision.md:86` “the wager is what I am betting, and this is what I am sure of.”\n> `docs/vision.md:169` “above was mine and he corrected it: *\"this is a bit too hard on ourselves but it is a fear that I”\n> `docs/vision.md:267` “*(The four-row division is Claude's, refined by the founder's correction above and approved”\n\n**The reader:** On this page \"I\" is two people. In the Opacity and Wager sections it is Fredrik (\"what I am betting\", \"I want to win this wager\"). In the gate section the same pronoun is Claude: \"the harder version above was mine and he corrected it\". There \"Claude\" and \"the founder\" are named in the third person, at lines 223 and 267. A stranger cannot tell who is speaking. The page also narrates its own drafting, for example \"the first draft of this section wrongly asked one person for two of them\".\n\n**So, in the reader's words:** Mark whose voice each section is in, or move the editorial history to research/, so a visitor is not left guessing who \"I\" is.\n\n**How sure the reader was:** yes. I searched the file for \"I\" and \"Claude\"; Claude appears only at lines 223 and 267.\n\n**Opened by session (2026-10-11): stands.** Claude is named on the page at lines 223 and 267 only. The first person is the founder in the sections on opacity and the wager, and Claude in the section on the gate.\n\nFrom [2026-10-11-pages-open-question-three-vision-pages-as-a-reader-returned-it.md](readers/2026-10-11-pages-open-question-three-vision-pages-as-a-reader-returned-it.md), O1.\n\n### `docs/vision.md:92`\n\n> `docs/vision.md:92` “Lowering the cost *dissolves* disagreement that was confusion, and only *clarifies* disagreement that is conflicting interests.”\n> `docs/civilizational-vision.md:88` “Much tribal warfare is over semantic confusion and information asymmetry, not genuine irreducible value differences.”\n> `docs/civilizational-vision.md:34` “She also states the guard this page needs and does not have”\n\n**The reader:** The Opacity section limits its own claim. Someone who understands me perfectly and wants what I have was never a comprehension problem, and a tool promising otherwise \"would be lying about the world\". The civilizational page never mentions interests. Its tribal-warfare section sorts disagreement into confusion or irreducible values, and omits the case where both sides understand each other and want the same thing. Line 34 names one missing guard, \"understanding is not acceptance\", and leaves it missing. The interest guard is not noticed at all. The page where scale raises the stakes lacks the limit.\n\n**So, in the reader's words:** Carry the interest-conflict limit and \"understanding is not acceptance\" into the civilizational page as two sentences.\n\n**How sure the reader was:** yes. A search for \"interest\" in civilizational-vision.md returns nothing.\n\n**Opened by session (2026-10-11): stands.** The word interest does not occur in civilizational-vision.md.\n\nFrom [2026-10-11-pages-open-question-three-vision-pages-as-a-reader-returned-it.md](readers/2026-10-11-pages-open-question-three-vision-pages-as-a-reader-returned-it.md), O9.\n\n### `docs/vision.md:156`\n\n> `docs/vision.md:156` “Contributions are added and never overwritten, which is the”\n> `deliberus/proposed_split.py:743` “MATCH (c:Claim {claim_id: $cid}) SET c.text = $t, c.short_title = $st\",”\n\n**The reader** (a finding for what does not fit, kind `rule-and-practice`): The page says contributions are never overwritten and that this is the architecture, not a policy. Saving a reworded machine-written part runs a query that sets the claim's text in place, and a second person rewording the same part leaves no trace of the first wording (docs/research/split-review-and-the-founders-intentions.md:516 records this as a conflict). The last edit stands, which is the edit-war dynamic the principle says has nothing to run on.\n\n**Which stands, by the reader:** other, by the code\n\n**Opened by session (2026-10-17): stands.** Opened deliberus/proposed_split.py at line 743 on 2026-10-11: saving a reworded part set the claim's text in place, so the earlier wording was not kept. Repaired 2026-10-17 on the founder's ruling of ledger AL10 (\"Can you build AL10 then please?\"): a reword now mints the reworded part as a new claim beside the old one and marks the old link superseded, the old part keeps its text and everything on it, unchanged wording writes nothing, and tests/test_claim_text_is_never_rewritten.py fails on any in-place edit of a claim's text.\n\nFrom [2026-10-11-pages-what-does-not-fit-three-vision-pages-as-a-reader-returned-it.md](readers/2026-10-11-pages-what-does-not-fit-three-vision-pages-as-a-reader-returned-it.md), F3.\n\n### `docs/vision.md:157`\n\n> `docs/vision.md:157` “architecture rather than a policy — so a wrong ontology at launch contaminates permanently. Frame”\n> `docs/vision.md:272` “a bet that a corpus launched on a wrong foundation cannot be repaired — the three reasons above —”\n\n**The reader:** The gate's second reason says a wrong ontology contaminates permanently and that \"an accumulated corpus\" cannot be refactored. Since 27 August the settled record shows repair at small scale. Terminus, verdict and dissolved were renamed (O1, O2, O6). The confirmed field was retired (Q32). 231 auto-connect links were retired, kept visible and re-read (AB11). 45 version-1 parts were superseded (T7). Each was cheap because nothing was overwritten. The case the paragraph actually names, frame lock-in from a domain's first extraction, has not been tested. The gate may be better aimed at that class than at the whole epistemology.\n\n**So, in the reader's words:** Narrow the second reason to frame lock-in, and say that link-level and field-level errors have proved repairable.\n\n**How sure the reader was:** fairly. I checked the ruled items in the settled-items file. I did not test lock-in itself, and the repairs were at a corpus of roughly 2,000 claims.\n\n**Opened by session (2026-10-11): stands.** The six rulings it cites are in the ledger as ruled. Whether they bear on the gate's second reason is an argument, and the founder's to weigh.\n\nFrom [2026-10-11-pages-open-question-three-vision-pages-as-a-reader-returned-it.md](readers/2026-10-11-pages-open-question-three-vision-pages-as-a-reader-returned-it.md), O3.\n\n### `docs/vision.md:180`\n\n> `docs/vision.md:180` “layer is right *enough* when the **discovery rate falls** — when concentrated attack on the”\n> `docs/vision.md:238` “corrections per period, already instrumented in `scripts/log_arc.py`”\n> `docs/vision.md:258` “done-ish\"*, is the correct locus, and the one thing worth adding is that **done-ish gets written down”\n> `docs/research/session26-open-rulings.md:247` “**The discovery rate falling** as the checkable criterion for when the gate lifts”\n\n**The reader:** The gate lifts when the discovery rate falls, but the instrument counts corrections per period. That is yield multiplied by effort. I did a crude count of commit subjects matching correct, retract, was wrong, overstate, misattribut, does not hold, refuted or stale. It gives about 46 in the 17 days to 28 August, then 22, then 18, then 25 in the week to 6 October, then 32 in the five days to 11 October. October rose because readers were sent through 28 files on 10 October and he asked for coverage to widen. A count that rises with attack cannot show attack running dry. I also found no written-down \"done-ish\", and ledger M4 is still OPEN.\n\n**So, in the reader's words:** Define the criterion as findings per reader-pass or per note read, and write the done-ish threshold down now, before the walk-through ends.\n\n**How sure the reader was:** fairly. I ran the git log counts myself (the regex is loose) and searched docs, TODO.md and SESSION-HISTORY.md for a written threshold.\n\n**Opened by session (2026-10-11): stands.** Ledger row M4, the discovery rate as the criterion, is open and marked as Claude's. The reader's own count of commits was not repeated.\n\nFrom [2026-10-11-pages-open-question-three-vision-pages-as-a-reader-returned-it.md](readers/2026-10-11-pages-open-question-three-vision-pages-as-a-reader-returned-it.md), O4.\n\n### `docs/vision.md:201`\n\n> `docs/vision.md:201` “and the prototype died on the one obstacle nobody could remove then: **someone had to type the”\n> `docs/history.md:49` “What failed: no LLMs, limited real-time tools, 250 users insufficient for network effects, funding/commercialization challenges”\n> `docs/history-delibr-era.md:11` “Nils wanted to sell a product; Fredrik wanted to build what was healthy for society”\n\n**The reader:** The why-now argument says the prototype died on one obstacle, typing the structure, and that LLM extraction removed it. The project's own history lists four causes: no LLMs, thin tools, 250 users too few for network effects, and funding. It also records a co-founder who wanted a product while the founder wanted what was healthy for society. LLM extraction answers the first, and partly the cold start. Funding and who builds it are untouched. The retrodiction fits neatly because it names only the cause that has since been removed.\n\n**So, in the reader's words:** Rewrite the sentence as one cause among several and say which of the others have changed.\n\n**How sure the reader was:** fairly. I read both history notes and found no source saying typing was the decisive cause.\n\n**Opened by session (2026-10-11): stands.** history.md line 49 lists four causes, and the history of the earlier company gives the difference of aim between the two founders. Sentence 14 on the founder's list already corrects that line's count of users.\n\nFrom [2026-10-11-pages-open-question-three-vision-pages-as-a-reader-returned-it.md](readers/2026-10-11-pages-open-question-three-vision-pages-as-a-reader-returned-it.md), O5.\n\n### `docs/vision.md:205`\n\n> `docs/vision.md:205` “four independent 2026 arrivals at the same empty reasoning-layer slot were recorded before the”\n> `docs/research/complexity-transitions-and-why-now.md:85` “The corpus recorded four of them — LawZero's machine-safety substrate, science's stalled reasoning infrastructure, forecasting's ceiling analysis, the noosphere macro-frame”\n\n**The reader:** \"Recorded before the explanation\" protects against fitting the story afterwards. It does not protect against selection. The four are sources the project chose to read. Three of the four notes are titled \"...and Deliberus\" (wright-god-test-noosphere-, forecasting-room-at-the-top-, bengio-safety-from-honesty-), written by looking for the Deliberus-shaped gap. Nobody counts the 2026 sources read that did not point at the slot. \"Independent\" holds for the authors, but the sampling is the project's. A retrodiction needs its denominator, and without one this is an illustration.\n\n**So, in the reader's words:** Keep the four as illustrations and drop the word \"retrodiction\", or state how many 2026 sources were read.\n\n**How sure the reader was:** fairly. I checked the why-now note at line 85 and the note filenames.\n\n**Opened by session (2026-10-11): stands.** Three of the four notes are named for Deliberus: the ones on Wright, on forecasting and on Bengio.\n\nFrom [2026-10-11-pages-open-question-three-vision-pages-as-a-reader-returned-it.md](readers/2026-10-11-pages-open-question-three-vision-pages-as-a-reader-returned-it.md), O6.\n\n### `docs/vision.md:238`\n\n> `docs/vision.md:238` “corrections per period, already instrumented in `scripts/log_arc.py`, which classifies by what a commit touched rather than what it was called”\n> `scripts/log_arc.py:93` “Epistemic markers in commit subjects (independent of prefix):”\n\n**The reader** (a finding for what does not fit, kind `state`): The page says the discovery rate, corrections per period, is already instrumented in log_arc.py and keyed on what a commit touched. The script sorts commit kinds by path, but counts corrections only by words found in commit subjects, which is what a commit was called, and prints that count once for the whole window, not per period. So no per-period correction rate exists, and the count it gives is keyed on titles.\n\n**Which stands, by the reader:** other, by the code\n\n**Opened by session (2026-10-11): stands.** Opened scripts/log_arc.py: it sorts commits by the paths they touched, and it counts corrections from words in commit subjects, once for the whole window. The page says more than the script does.\n\nFrom [2026-10-11-pages-what-does-not-fit-three-vision-pages-as-a-reader-returned-it.md](readers/2026-10-11-pages-what-does-not-fit-three-vision-pages-as-a-reader-returned-it.md), F5.\n\n### `docs/vision.md:303`\n\n> `docs/vision.md:303` “with the ranking rules themselves standing in the graph as challengeable claims”\n> `CLAUDE.md:769` “an epistemic constitution is demanded by three docs and written nowhere”\n\n**The reader** (a finding for what does not fit, kind `state`): The page says the ranking rules stand in the graph as challengeable claims. The canon's frontier line says a constitution for those rules is demanded by three documents and written nowhere. The strength rules are constants in code that the founder rules on, and the one shipped example, verdict-claims, was retired on 2026-09-16. A reader who tries to challenge a ranking rule as a claim finds none.\n\n**Which stands, by the reader:** other, by the code\n\n**Opened by session (2026-10-11): stands in part.** Read the sentence in its paragraph: it states a conviction about the design. One commitment of the schema is published in the graph as a claim (the split between is and ought, by scripts/seed_ontology_commitments.py); no ranking or strength rule is, and the written rules for the keeper do not exist.\n\nFrom [2026-10-11-pages-what-does-not-fit-three-vision-pages-as-a-reader-returned-it.md](readers/2026-10-11-pages-what-does-not-fit-three-vision-pages-as-a-reader-returned-it.md), F6.\n\n### `docs/vision.md:325`\n\n> `docs/vision.md:325` “short-authored first takes can now be preserved as non-public deliberation drafts containing the raw exchange and provisional promotion hints”\n> `CLAUDE.md:583` “raw input saved at once, durable/public structure only through lightweight confirmation; one graph, staged maturity, never a separate draft universe”\n\n**The reader** (a finding for what does not fit, kind `rule-and-practice`): The page says the graph stays one graph, and the canon says never a separate draft universe. The implementation note describes saving first takes as non-public drafts; the code keeps them in a separate Postgres table (deliberus/database.py, DeliberationDraft) with its own lifecycle stage, retrieval weight and provisional synthesis, outside the graph. A user's first take then sits where no other claim can link to it or attack it.\n\n**Which stands, by the reader:** cannot tell, by nothing I found\n\n**Opened by session (2026-10-11): stands in part.** Opened deliberus/database.py at line 89: first takes are kept in a table of their own outside the graph. Whether a private store of raw first takes is the separate draft universe the principle rules out, or the raw capture the same principle asks for, is the founder's to say.\n\nFrom [2026-10-11-pages-what-does-not-fit-three-vision-pages-as-a-reader-returned-it.md](readers/2026-10-11-pages-what-does-not-fit-three-vision-pages-as-a-reader-returned-it.md), F8.\n\n### `docs/vision.md:345`\n\n> `docs/vision.md:345` “The user never needs to think formally; the formality emerges.”\n> `docs/vision.md:393` “Contribution barrier drops to zero while structural rigor is maintained underneath.”\n> `docs/the-missing-layer.md:29` “Deeper down, where a disagreement is taken apart, it is still being built.”\n\n**The reader:** On 10 October he ruled that the positioning page should stop saying nobody hand-builds anything, because the descent is still built by hand. The same promise stands unqualified on the vision page: the user \"never needs to think formally\" and the \"contribution barrier drops to zero\". Both are in present tense. The repair record has entries for the other two notes and none for vision.md. The correction reached one of the surfaces that repeat the claim, which is the situation D13 describes.\n\n**So, in the reader's words:** Carry the qualifier \"holds at the door, not yet in the descent\" to vision.md lines 345 and 393.\n\n**How sure the reader was:** yes. I read the repairs-made.md headings and the 10 October commit c7ff6bc.\n\n**Opened by session (2026-10-11): stands.** The ruling of 2026-10-10 changed the positioning page (commit c7ff6bc). The record of repairs has no entry for vision.md on this, and lines 345 and 393 stand as quoted.\n\nFrom [2026-10-11-pages-open-question-three-vision-pages-as-a-reader-returned-it.md](readers/2026-10-11-pages-open-question-three-vision-pages-as-a-reader-returned-it.md), O2.\n\n### `docs/vision.md:377`\n\n> `docs/vision.md:377` “The personal tool and the civilizational graph are not sequential phases with gates — they're nested.”\n> `docs/vision.md:305` “a product, a personal tool, or a research instrument. Those are phases it passes through.”\n> `docs/vision.md:23` “Surface hidden social dynamics: tone/emotion analysis to detect self-censorship, obfuscation”\n\n**The reader:** vision.md is layered and the layers disagree without saying so. The March \"Three Horizons\" says the personal tool and the civilizational graph are not phases with gates. The July wager says product, personal tool and research instrument \"are phases it passes through\". The August gate then sits upstream of deploy and outreach. The undated Solution bullets promise tone and emotion analysis to detect self-censorship and obfuscation. I found that in no code, and in no document other than technical-direction.md. A visitor reading top to bottom meets every stratum as current.\n\n**So, in the reader's words:** Annotate Three Horizons as superseded by the gate section, and date or prune the Problem and Solution bullets.\n\n**How sure the reader was:** fairly. I searched deliberus/, scripts/ and web/src for goemotion, self-censorship and obfuscation, and searched docs for the rest.\n\n**Opened by session (2026-10-11): stands.** No code mentions the detection of tone, emotion or self-censorship. Besides this page only technical-direction.md does.\n\nFrom [2026-10-11-pages-open-question-three-vision-pages-as-a-reader-returned-it.md](readers/2026-10-11-pages-open-question-three-vision-pages-as-a-reader-returned-it.md), O7.\n\n### `docs/vision.md:412`\n\n> `docs/vision.md:412` “I'm an EA but also anticapitalist in the long run, but in order to figure out (and transition into) something better than capitalism”\n> `docs/civilizational-vision.md:71` “In order to figure out (and transition into) something better than capitalism, we need to build a collaborative sensemaking platform”\n> `docs/civilizational-vision.md:67` “The transition from market pricing to something more intentional becomes *thinkable*”\n\n**The reader:** One quote is trimmed two ways. vision.md keeps \"I'm an EA but also anticapitalist in the long run\" and gives no attribution. The civilizational page starts after that clause and attributes it to Fredrik, in the Resource Allocation section. The project's credibility rests on mapping reasoning without endorsing conclusions. Yet the destination in these pages includes a transition beyond market pricing, and \"beyond cutthroat competition\" (vision.md:407). The section's trust-and-GDP figures support lower transaction costs, not that step. A libertarian reader, the lens's most useful visitor, meets a conclusion before a map.\n\n**So, in the reader's words:** Decide whether the economic destination is the founder's hope (label it as his) or a claim of the project, and quote the line once, the same way.\n\n**How sure the reader was:** fairly. I compared the two quotes. The neutrality tension is my reading, not a ruled position.\n\n**Opened by session (2026-10-11): stands.** The two pages cut one sentence of his at different points, and only one of them says it is his.\n\nFrom [2026-10-11-pages-open-question-three-vision-pages-as-a-reader-returned-it.md](readers/2026-10-11-pages-open-question-three-vision-pages-as-a-reader-returned-it.md), O13.\n\n## `docs/worldview-lenses.md`\n\n### `docs/worldview-lenses.md:5`\n\n> `docs/worldview-lenses.md:5` “They agree on roughly 80% of the structure underneath. They diverge on one definitional premise.”\n> `docs/worldview-lenses.md:9` “And once located, the disagreement is almost always much smaller than it felt.”\n> `docs/research/prime-real-estate-words.md:17` “to: *magic, God, love, freedom, natural, science, conscious*. Others carry a rebuke: *cult,”\n\n**The reader:** The page's first example, and the moment it calls among the most powerful the platform offers, is two people finding that \"freedom\" meant different things. After that, the disagreement is \"almost always much smaller than it felt\". The corpus's own 2026-10-07 finding lists freedom among the words people want. Showing two senses does not end that dispute, because each person is asked to give up what they came for, and what remains can be who gets the word (ledger AC3, ruled). The flagship example is therefore the exception class. The 80% is illustrative with no source, and \"almost always\" has none.\n\n**So, in the reader's words:** Keep the example but add the third outcome (the senses are shown and the contest over the word remains), or pick a word nobody is fighting for.\n\n**How sure the reader was:** fairly. I read prime-real-estate-words.md §2 and ledger AC3. That both speakers want the word is my assumption.\n\n**Not yet opened by a session.** Its quoted words were checked by code when the report was filed.\n\nFrom [2026-10-11-pilot-open-question-four-notes-as-a-reader-returned-it.md](readers/2026-10-11-pilot-open-question-four-notes-as-a-reader-returned-it.md), O11.\n\n### `docs/worldview-lenses.md:11`\n\n> `docs/worldview-lenses.md:11` “This is not a feature. It is the core interaction mode, and it has been part of the vision since before 2011.”\n> `TODO.md:2911` “the return edge is testable in the first workshop and the lens is not”\n> `TODO.md:1640` “build the return edge first, the lens second”\n> `docs/worldview-lenses.md:52` “Both are on the table; neither has won.”\n\n**The reader:** The lens page calls the lens \"not a feature\" and \"the core interaction mode,\" and presents it as the hook no predecessor offered. The project's own plan puts it second: TODO says to build the return edge first and not to lead with the lens. The lens page itself says neither proposal for finding a visitor's lens has won. No worldview or lens code exists in deliberus/. The public site lists \"Lenses\" as a concept, so visitors meet a deferred feature described as the core.\n\n**So, in the reader's words:** He should decide whether the lens is described as core or as planned second, since the page and the build order currently disagree.\n\n**How sure the reader was:** yes for the plan and the absence of code, which I checked by search; fairly for the reading of what visitors see.\n\n**Not yet opened by a session.** Its quoted words were checked by code when the report was filed.\n\nFrom [2026-10-11-pilot-open-question-four-notes-as-a-haiku-reader-returned-it.md](readers/2026-10-11-pilot-open-question-four-notes-as-a-haiku-reader-returned-it.md), O7.\n\n### `docs/worldview-lenses.md:11`\n\n> `docs/worldview-lenses.md:11` “This is not a feature. It is the core interaction mode”\n> `docs/worldview-lenses.md:13` “That's why they failed — analytical tools alone don't motivate sustained use.”\n\n**The reader:** The page speaks in the present tense of a product mode: the filter \"is the core interaction mode\" and its moment is among the most powerful \"the platform offers\". There is no lens code. I searched deliberus/ for worldview and lens and found nothing, and web/ only links to this page from the home page (\"Lenses\"). CLAUDE.md files the filter as \"EXPLORATORY until explicitly decided\". The page also states as fact a cause that the corpus lists as an open existential risk. steelmanned-critiques.md §3 gives a rival cause: structural incompatibility with online social dynamics, with platforms that all got users and stagnated, and Kialo with 1M registered users.\n\n**So, in the reader's words:** Add a status line to the page (not built, a design under test) and mark \"That's why they failed\" as a hypothesis or cite what supports it.\n\n**How sure the reader was:** yes on unbuilt (rg over deliberus/ and web/src). Fairly on the causal point: I read the steelman and found no evidence for the lens cause.\n\n**Not yet opened by a session.** Its quoted words were checked by code when the report was filed.\n\nFrom [2026-10-11-pilot-open-question-four-notes-as-a-reader-returned-it.md](readers/2026-10-11-pilot-open-question-four-notes-as-a-reader-returned-it.md), O14.\n\n### `docs/worldview-lenses.md:11`\n\n> `docs/worldview-lenses.md:11` “It is the core interaction mode, and it has been part of the vision since before 2011.”\n> `CLAUDE.md:691` “All are EXPLORATORY until explicitly decided. Deep convictions and intuitions about”\n\n**The reader** (a finding for what does not fit, kind `state`): The worldview page calls the lens the core interaction mode, present tense and part of the vision since 2011. CLAUDE.md lists the worldview filter among things exploratory until explicitly decided, and TODO.md:2911 defers the lens as \"deliberately second\". No lens or worldview code exists under deliberus/, so the present-tense claim has no decision or build behind it.\n\n**Which stands, by the reader:** other, by the code\n\n**Opened by session (2026-10-11): stands.** Both ends were read on 2026-10-11 (the reading note's § 7a). Its other end is the code, the to-do file, the instruction file or the same note, so the index does not hold it.\n\nFrom [2026-10-11-pilot-what-does-not-fit-four-notes-as-a-haiku-reader-returned-it.md](readers/2026-10-11-pilot-what-does-not-fit-four-notes-as-a-haiku-reader-returned-it.md), F6.\n\n### `docs/worldview-lenses.md:19`\n\n> `docs/worldview-lenses.md:19` “with the software as the lens that bends the light”\n> `docs/worldview-lenses.md:102` “The answer that emerged from the error-taxonomy work is that the interface is not the instrument”\n> `docs/civilizational-vision.md:26` “the software is the lens, and what it does is bend the light”\n\n**The reader:** The lens page's outward copy says the software is the lens that bends the light, a sharpening the founder approved on 2026-09-16. Further down, the same page says the interface is not the instrument and another person is, and that the lens grants no transcendence. The copy tells visitors the software does the seeing, which the page's own August finding rejects. Both cannot be the public position.\n\n**So, in the reader's words:** He should choose whether the software is the lens or only the place where a person's view is held, and fix the copy to match.\n\n**How sure the reader was:** yes. I read the quoted lines and the approval attribution in civilizational-vision.md.\n\n**Not yet opened by a session.** Its quoted words were checked by code when the report was filed.\n\nFrom [2026-10-11-pilot-open-question-four-notes-as-a-haiku-reader-returned-it.md](readers/2026-10-11-pilot-open-question-four-notes-as-a-haiku-reader-returned-it.md), O8.\n\n### `docs/worldview-lenses.md:46`\n\n> `docs/worldview-lenses.md:46` “From inside their worldview, their conclusions make sense”\n> `docs/worldview-lenses.md:104` “A lens weighs which likenesses count, and the weighing partly follows the conclusion wanted”\n\n**The reader:** The rose-tinted section promises that from inside another worldview, their conclusions make sense. The note's own 2026-10-08 paragraph says the weighing of likenesses partly follows the conclusion wanted, and cites a study where precedents seemed more alike when their outcome suited the reader's view, with legal training no protection. If weights bend toward conclusions, a lens that renders a view coherently shows chosen weights as integrity. The page has the evidence against its central promise and does not reconcile them.\n\n**So, in the reader's words:** He should decide whether the lens shows the weights as an inspectable dial, since a coherent rendering alone cannot tell integrity from a motivated weighting.\n\n**How sure the reader was:** fairly. I read the section and the 2026-10-08 paragraph in full.\n\n**Not yet opened by a session.** Its quoted words were checked by code when the report was filed.\n\nFrom [2026-10-11-pilot-open-question-four-notes-as-a-haiku-reader-returned-it.md](readers/2026-10-11-pilot-open-question-four-notes-as-a-haiku-reader-returned-it.md), O9.\n\n### `docs/worldview-lenses.md:46`\n\n> `docs/worldview-lenses.md:46` “Not \"I tolerate your ignorance\" but \"I see the coherence of your worldview and I must engage with it honestly.\"”\n> `docs/worldview-lenses.md:84` “not the reassuring experience of \"understanding\" the other side (which often means containing them within your existing framework).”\n\n**The reader:** The third feature is named \"Rose-Tinted Glasses UX\", and its success is the recognition \"I see the coherence of your worldview and I must engage with it honestly.\" Later on the same page, Žižek's warning defines failure as the reassuring experience of \"understanding\" the other side, which \"often means containing them within your existing framework\". The two describe the same experience, and the page never says how a designer tells them apart. \"Rose-tinted\" is also the ordinary idiom for flattering distortion, so the feature's name states the failure. The 2026-08-27 section's answer (another person is the instrument) serves the self-assessment use, not this one.\n\n**So, in the reader's words:** Give feature 3 a test that separates it from tourism, for example the reader must name one premise of their own that the other lens makes costly, and consider renaming it.\n\n**How sure the reader was:** fairly. The tension is read straight from the page. The idiom collision is a guess until a real stranger (Malin) hears it.\n\n**Not yet opened by a session.** Its quoted words were checked by code when the report was filed.\n\nFrom [2026-10-11-pilot-open-question-four-notes-as-a-reader-returned-it.md](readers/2026-10-11-pilot-open-question-four-notes-as-a-reader-returned-it.md), O12.\n\n### `docs/worldview-lenses.md:62`\n\n> `docs/worldview-lenses.md:62` “it captures operational values (how you actually decide) rather than stated values (what you say you believe)”\n> `docs/worldview-lenses.md:62` “The computed profile IS the lens — derived from behavior, not declaration.”\n\n**The reader:** Edvin's dilemma filter is praised for capturing \"how you actually decide\", derived \"from behavior, not declaration\". Answers to trolley-type cases are declarations about hypotheticals, not behaviour. The corpus's own evidence section (what-counts-as-evidence/testimony-experience-and-normative-evidence.md §6.3) records order effects on such judgments as large in professional philosophers as in laypeople. Reflection did not remove them. A profile computed from dilemma answers inherits that instability, and a lens built on it could shift with question order. Nothing in the corpus tests dilemma answers against behaviour. The 2026 bottom-up version, a profile from votes, is declaration too.\n\n**So, in the reader's words:** Before choosing \"seed with dilemmas\", run the cheap check: same person, shuffled order, does the profile hold?\n\n**How sure the reader was:** fairly. I searched the corpus for dilemma-to-behaviour tests and found none, only the order-effects note. A guess on how large the instability would be here.\n\n**Not yet opened by a session.** Its quoted words were checked by code when the report was filed.\n\nFrom [2026-10-11-pilot-open-question-four-notes-as-a-reader-returned-it.md](readers/2026-10-11-pilot-open-question-four-notes-as-a-reader-returned-it.md), O13.\n\n### `docs/worldview-lenses.md:92`\n\n> `docs/worldview-lenses.md:92` “The lenses show the surface; the depth reveals the convergence.”\n> `CLAUDE.md:553` “the bet can LOSE (hinge/parity/permissivism ceilings count against it)”\n\n**The reader** (a finding for what does not fit, kind `tension`): The worldview page states as a present fact that depth reveals convergence beneath surface disagreement. The always-loaded convergence line calls convergence a wager that can lose to hinge and parity ceilings, and depth.md:134 calls it a conviction, not a proven fact. Two people who drill to a protected value and find no shared bedrock make the page's sentence false.\n\n**Which stands, by the reader:** other, by nothing I found\n\n**Opened by session (2026-10-11): stands.** Both ends were read on 2026-10-11 (the reading note's § 7a). Its other end is the code, the to-do file, the instruction file or the same note, so the index does not hold it.\n\nFrom [2026-10-11-pilot-what-does-not-fit-four-notes-as-a-haiku-reader-returned-it.md](readers/2026-10-11-pilot-what-does-not-fit-four-notes-as-a-haiku-reader-returned-it.md), F5.\n\n## `docs/research/the-load-bearing-unsaid.md`\n\n### `docs/research/the-load-bearing-unsaid.md:35`\n\n> `docs/research/the-load-bearing-unsaid.md:35` “and why implicit-premise recall is already this project's most important quality axis”\n> `docs/research/dogfood-run-6-israel-palestine-cross-domain.md:274` “The comparison protocol from run 3F still cannot execute, and this run sharpens what it should measure”\n\n**The reader:** The load note says implicit-premise recall is already the project's most important quality axis. The evidence I found is one debate pair in run 6, where a premise its author never wrote outranked his stated claims, and that run's note says the registered comparison protocol still cannot execute. Several other notes repeat the \"most important axis\" line by citing that case. The axis is a hypothesis the corpus has not measured.\n\n**So, in the reader's words:** He should run the registered protocol, or date the claim as a hypothesis, before it is cited as the project's main quality measure.\n\n**How sure the reader was:** fairly. I searched the notes and TODO and found no measured result.\n\n**Not yet opened by a session.** Its quoted words were checked by code when the report was filed.\n\nFrom [2026-10-11-pilot-open-question-four-notes-as-a-haiku-reader-returned-it.md](readers/2026-10-11-pilot-open-question-four-notes-as-a-haiku-reader-returned-it.md), O12.\n\n### `docs/research/the-load-bearing-unsaid.md:47`\n\n> `docs/research/the-load-bearing-unsaid.md:47` “Pointing the implicit-premise pass at a sacred commitment and rendering it as a weighable premise is the mechanised form of the taboo trade-off”\n> `docs/research/the-load-bearing-unsaid.md:120` “Measured effect: 0 → 2 firings on the corpus, both genuine author value premises”\n\n**The reader:** The note concludes the brake gets a better justification, and counts the repair as 0 to 2 firings. In the code the brake is not a gate on decomposition. detect_sacredness has one caller, inside detect_weighing, which returns early unless the claim also contains a weighing phrase (outweigh, more important than, trumps). The two exemplar claims, human life as inherently precious, contain none. So on the pipeline's route there is nothing for the brake to do for them, and the \"2\" counts the regex on claim text. The splitter (which writes parts since 2026-09-21), the implicit-premise pass and the completeness copout count never consult it.\n\n**So, in the reader's words:** Decide whether the brake is meant to guard every machine decomposition; if so it belongs in front of the splitter and the premise passes, and its firing count should be taken where it changes an outcome.\n\n**How sure the reader was:** yes for the code. detect_sacredness is called only at weighing.py:248, process_extraction_weighings skips non-weighings, proposed_split.py:446 mentions the brake only in a comment, and there is no \"sacred\" in implicit_premises.py or completeness.py. Fairly for the consequence, since I did not run the pipeline.\n\n**Not yet opened by a session.** Its quoted words were checked by code when the report was filed.\n\nFrom [2026-10-11-pilot-open-question-four-notes-as-a-reader-returned-it.md](readers/2026-10-11-pilot-open-question-four-notes-as-a-reader-returned-it.md), O15.\n\n### `docs/research/the-load-bearing-unsaid.md:47`\n\n> `docs/research/the-load-bearing-unsaid.md:47` “Pointing the implicit-premise pass at a sacred commitment and rendering it as a weighable premise”\n> `deliberus/extraction/implicit_premises.py:3` “Surfaces unstated assumptions connecting explicit claims within each argument.”\n\n**The reader** (a finding for what does not fit, kind `rule-and-practice`): The load-bearing note calls pointing the implicit-premise pass at a sacred commitment a category error. Pass 2c runs over every argument's claims, and I found no sacredness check in `detect_implicit_premises` or its extraction callers. So the pass as built does this to any sacred claim it receives, and the note does not say whether that is intended.\n\n**Which stands, by the reader:** cannot tell, by nothing I found\n\n**Opened by session (2026-10-11): stands.** Both ends were read on 2026-10-11 (the reading note's § 7a). Its other end is the code, the to-do file, the instruction file or the same note, so the index does not hold it.\n\nFrom [2026-10-11-pilot-what-does-not-fit-four-notes-as-a-haiku-reader-returned-it.md](readers/2026-10-11-pilot-what-does-not-fit-four-notes-as-a-haiku-reader-returned-it.md), F7.\n\n### `docs/research/the-load-bearing-unsaid.md:65`\n\n> `docs/research/the-load-bearing-unsaid.md:65` “a pseudo-sacred commitment yields to decomposition, a hinge does not, and which one you have is discovered by trying”\n> `docs/research/the-load-bearing-unsaid.md:47` “Pointing the implicit-premise pass at a sacred commitment and rendering it as a weighable premise”\n\n**The reader** (a finding for what does not fit, kind `tension`): Line 65 says whether a commitment is a pseudo-sacred one or a hinge is found out only by trying decomposition on it. Line 47 says decomposing a sacred commitment is the mechanised form of the taboo trade-off, a category error. Take the ACLU's \"human life is inherently precious\" claim: to learn whether it is a hinge you must try, and trying is what line 47 forbids.\n\n**Which stands, by the reader:** cannot tell, by nothing I found\n\n**Opened by session (2026-10-11): stands.** Both ends were read on 2026-10-11 (the reading note's § 7a). Its other end is the code, the to-do file, the instruction file or the same note, so the index does not hold it.\n\nFrom [2026-10-11-pilot-what-does-not-fit-four-notes-as-a-haiku-reader-returned-it.md](readers/2026-10-11-pilot-what-does-not-fit-four-notes-as-a-haiku-reader-returned-it.md), F8.\n\n### `docs/research/the-load-bearing-unsaid.md:71`\n\n> `docs/research/the-load-bearing-unsaid.md:71` “Making the unsaid explicit redistributes advantage, and the people whose doxa is being converted rarely experience it as help”\n> `TODO.md:3202` “the conversion is political, and the people whose doxa is being converted rarely experience it as help”\n\n**The reader:** The load note's warning is that making the unsaid explicit is political and redistributes advantage away from the people whose doxa is converted. I found no ruling or build item that answers it. TODO carries it only as a positioning note, which also marks the Bourdieu and Brandom citations as missing even though this note supplies them.\n\n**So, in the reader's words:** He should decide whether the warning becomes a design gate or a stated caveat on the page, since it is currently neither.\n\n**How sure the reader was:** fairly. I searched the settled items, TODO, and the code for any response.\n\n**Not yet opened by a session.** Its quoted words were checked by code when the report was filed.\n\nFrom [2026-10-11-pilot-open-question-four-notes-as-a-haiku-reader-returned-it.md](readers/2026-10-11-pilot-open-question-four-notes-as-a-haiku-reader-returned-it.md), O13.\n\n### `docs/research/the-load-bearing-unsaid.md:93`\n\n> `docs/research/the-load-bearing-unsaid.md:93` “After widening (below), it matches the assisted-dying opening line and nothing else: one packaging hit, zero body hits, on precisely the case predicted.”\n> `docs/research/the-load-bearing-unsaid.md:103` “Only one is the author's own voice: the ACLU writing *\"life is precious and death irrevocable\"*”\n\n**The reader** (a finding for what does not fit, kind `number-or-date`): Line 93 says the widened lexicon matches only the assisted-dying opening line, with zero body hits. Line 103 says the ACLU writes \"life is precious\" in its own voice, and the results file the note cites (dogfood-run-6/sacred-displacement-results.json, aggregate A_live_brake) gives packaging_hits 1 and body_hits 4, the ACLU hit being in the body. Both cannot hold.\n\n**Which stands, by the reader:** other, by a measurement\n\n**Opened by session (2026-10-11): stands.** Both ends were read on 2026-10-11 (the reading note's § 7a). Its other end is the code, the to-do file, the instruction file or the same note, so the index does not hold it.\n\nFrom [2026-10-11-pilot-what-does-not-fit-four-notes-as-a-reader-returned-it.md](readers/2026-10-11-pilot-what-does-not-fit-four-notes-as-a-reader-returned-it.md), F12.\n\n### `docs/research/the-load-bearing-unsaid.md:105`\n\n> `docs/research/the-load-bearing-unsaid.md:105` “in the argued register you may *describe* a comprehensive doctrine, you may not *argue from* one.”\n> `docs/research/the-load-bearing-unsaid.md:65` “a pseudo-sacred commitment yields to decomposition, a hinge does not, and which one you have is discovered by trying.”\n> `docs/research/the-load-bearing-unsaid.md:27` “The founder's intuition is correct at this level”\n\n**The reader:** Of seven sacred-vocabulary hits, six were reported speech and one was asserted, in an advocacy voice. The note reads this as Rawls's register norm. The pattern is also what hinge theory allows (mention is fine, use is nonsense). But sacred content appearing in advocacy and in an opening line shows it is sayable where the register permits, which a hinge is not. \"What this changes\" then proceeds on the stronger reading (\"category-mistaken\", \"wrong tool\"). The proposed discriminator, \"discovered by trying\", is the very attempt the contemplation effect says does the harm. The \"answer in short\" was not revised after the test.\n\n**So, in the reader's words:** Say in the note which of the two the brake is built for, and test the hinge reading only by invitation (a person asked whether to try), not by running the pass on their value.\n\n**How sure the reader was:** fairly. I re-read the results table, the counts and \"What this changes\". The note itself says the 3.48x ratio is not shown.\n\n**Not yet opened by a session.** Its quoted words were checked by code when the report was filed.\n\nFrom [2026-10-11-pilot-open-question-four-notes-as-a-reader-returned-it.md](readers/2026-10-11-pilot-open-question-four-notes-as-a-reader-returned-it.md), O16.\n\n### `docs/research/the-load-bearing-unsaid.md:120`\n\n> `docs/research/the-load-bearing-unsaid.md:120` “Measured effect: 0 → 2 firings on the corpus, both genuine author value premises; the one lexical false positive”\n> `deliberus/weighing.py:238` “Deterministic weighing detection. High precision over recall by design:”\n\n**The reader** (a finding for what does not fit, kind `rule-and-practice`): The note counts the rebuilt brake as firing on two genuine value premises, the ACLU and Humanists UK claims. In the pipeline detect_sacredness is called only inside detect_weighing, after a claim of a weighable type has matched a weighing marker such as \"outweighs\"; neither claim does (tested: is_weighing False, sacred False). The two claims are never routed to the brake, so the count measures the lexicon alone, which is the unfired-guard case the project warns about.\n\n**Which stands, by the reader:** other, by the code\n\n**Opened by session (2026-10-11): stands.** Both ends were read on 2026-10-11 (the reading note's § 7a). Its other end is the code, the to-do file, the instruction file or the same note, so the index does not hold it.\n\nFrom [2026-10-11-pilot-what-does-not-fit-four-notes-as-a-reader-returned-it.md](readers/2026-10-11-pilot-what-does-not-fit-four-notes-as-a-reader-returned-it.md), F13.\n\n### `docs/research/the-load-bearing-unsaid.md:144`\n\n> `docs/research/the-load-bearing-unsaid.md:144` “A hinge clash that survives being made explicit has been *clarified*, not resolved, and the residue map exists to count exactly that.”\n> `CLAUDE.md:553` “horizon\" ruled 2026-08-25 as the name for what was \"residue\": where shared mapping currently ends, moving as the corpus grows”\n\n**The reader** (a finding for what does not fit, kind `overtaken`): The note speaks of the residue map and residue taxonomy as current, and its see-also line points to convergence.md for \"what a residue is\". A founder ruling of 2026-08-25 renamed residue to horizon; api.py says \"Renamed from residue-map 2026-08-25\" and keeps residue only as a legacy alias. The page carries no mark of the rename.\n\n**Which stands, by the reader:** other, by a ruling of 2026-08-25\n\n**Opened by session (2026-10-11): stands.** Both ends were read on 2026-10-11 (the reading note's § 7a). Its other end is the code, the to-do file, the instruction file or the same note, so the index does not hold it.\n\nFrom [2026-10-11-pilot-what-does-not-fit-four-notes-as-a-reader-returned-it.md](readers/2026-10-11-pilot-what-does-not-fit-four-notes-as-a-reader-returned-it.md), F14.\n\n## `docs/research/writing-out-the-unwritten.md`\n\n### `docs/research/writing-out-the-unwritten.md:571`\n\n> `docs/research/writing-out-the-unwritten.md:571` “Every link is written out in one form: the first claim, what ties this case to a rule”\n> `TODO.md:1079` “is writing every link out, with the writing checked, what he asked for”\n\n**The reader:** The depth page, dated 2026-10-08, states the aim as a link's step being \"written out when someone asks, never that all of them are.\" The day before, the founder asked for all implicit premises to be \"explicated and exposed and scaffolded automagically.\" The open proposal that would write every link out (AD13) awaits his ruling, and TODO records his direction on it as pending. A public page therefore states as the aim an option he has not chosen.\n\n**So, in the reader's words:** He should rule AD13 before this sentence stays public as the aim, or the page will have settled the question for him.\n\n**How sure the reader was:** yes. I read the four sources and checked the dates and wording of each quote.\n\n**Not yet opened by a session.** Its quoted words were checked by code when the report was filed.\n\nFrom [2026-10-11-pilot-open-question-four-notes-as-a-haiku-reader-returned-it.md](readers/2026-10-11-pilot-open-question-four-notes-as-a-haiku-reader-returned-it.md), O6.\n"}