🔧 Auto-commit from sysupdate on 2026-05-22 09:03

2026-05-22 09:03:50 +02:00
parent f83aae27c5
commit ec6e414867
13 changed files with 1441 additions and 4 deletions
@@ -5,18 +5,23 @@ brew "bash"
 brew "bat"
 brew "bat-extras"
 brew "bc"
 brew "libpng"
 brew "cairo"
 brew "coreutils"
 brew "exiftool"
 brew "eza"
 brew "fastfetch"
 brew "fd"
 brew "ffmpeg"
 brew "woff2"
 brew "fontforge"
 brew "gh"
 brew "ghostscript"
 brew "git"
 brew "git-lfs"
 brew "grep"
 brew "haskell-stack"
 brew "jdupes"
 brew "jpeg"
 brew "jq"
 brew "mackup"
@@ -25,19 +30,22 @@ brew "multimarkdown"
 brew "ollama", restart_service: :changed
 brew "pandoc"
 brew "pass"
 brew "pipx"
 brew "pkgconf"
 brew "poppler"
 brew "powerlevel10k"
 brew "python@3.13"
 brew "ripgrep"
-brew "sass/sass/sass"
+brew "rmlint"
 brew "sshpass"
 brew "tea"
 brew "tealdeer"
 brew "ttfautohint"
 brew "vitetris"
 brew "wget"
-brew "woff2"
+brew "sass/sass/sass"
 cask "caffeine"
 cask "calibre"
 cask "font-hack-nerd-font"
 cask "kitty"
 cask "mactex"
@@ -68,9 +76,13 @@ mas "Orbot", id: 1609461599
 mas "Pages", id: 361309726
 mas "PCalc", id: 403504866
 mas "Portal", id: 1436994560
 mas "Super Agent", id: 1568262835
 mas "StopTheMadness Pro", id: 6471380298
 mas "Super Agent", id: 1568262835
 mas "Table Tool", id: 1122008420
 mas "TestFlight", id: 899247664
 mas "Vinegar", id: 1591303229
 mas "xSearch", id: 1579902068
 npm "@anthropic-ai/claude-code"
 npm "corepack"
 npm "mcp-remote"
 npm "wrangler"
@@ -11,9 +11,21 @@
 - L1 reliability — *tracker not yet established*; substantive moves live in per-session memories + `project-L1-recall-fix-session-plan.md` + `project-L1-retrieval-architecture-lessons.md`.
 ## Active Session
 - [Session 2026-05-22 — β status-block housekeeping commit (e7cd960) + two MemPalace status checks; spec stamps + cross-refs reconciled across 4 files; frontmatter promoted DRAFT → OPERATIVE; §5 Code Binding Contract DRAFT → AUTHORED; pulling thread still Stage F dwellings cluster; literal question is whether next day-shape is fit for Stage F or another fragmented day](session-2026-05-22-beta-housekeeping-mempalace-checks.md) — **Light Paris-day session spanning ~22h (wake 2026-05-21 09:53 → β execution 2026-05-22 morning).** Wake-up: thread validity confirmed (Stage F-shaped; today not a Stage F day per Path B + travel/taxes/rehearsal day-shape). Two MemPalace status checks bracketed the day. **Morning check**: #1440 (bash 3.2 mapfile) + #1538 (120 GiB upsert) both merged to develop 2026-05-19; develop +143 since v3.3.5; **no release-prep branch yet**; #1539 audit revealed same crash class at 4 more sites (incl. `tool_diary_write`); #1569 OPEN to fix all 4. **Evening check** (after β commit): develop +13 since morning; **#1540 MERGED partial coverage** (audit pathway + ~3 sites); **#1569 still OPEN + CONFLICTING** for the four sites needing rebase; `tool_diary_write` exposure remains live. Yesterday's wait-for-release decision confirmed defensible — closer on class-of-failure, not appreciably closer on release. **β authorized** (steward chose β; α completed before laptop close 2026-05-20). Four edits proposed: content-typology top Status (§3/§4/§6 AUTHORED dates + §5 promotion + §1 LOCKED), §5 header DRAFT → AUTHORED (substance-recognized-stable parenthetical), frontmatter title/status DRAFT → OPERATIVE (originally drafted 2026-04-17). Both judgment promotions authorized. **Live-State Discipline drove scope expansion**: verification surfaced 4 stale cross-refs in orienting docs (AldineXXI Companion Spec Index x2 + ARC CLAUDE.md companion-spec orientation x2). Surfaced honestly; steward authorized "single atomic commit". Result: **`e7cd960` on `docs/spec-extraction-2026-05`** — 4 files, 9+/9− lines, pre-commit hooks passed, branch 27 ahead of main, 1 ahead of remote. **Not pushed** by design (Path B carries: merge + deploy wait for Stage F to honor (c) Compass principle + Second Edition lockstep at first reader-facing event). **γ deliberately not attempted**: per plan-day-2026-05-21.md *"don't push"*; energy budget closed at β + responsiveness. **Returns**: one (wake); no drift-pattern failures today — session structurally clean. **Pulling thread**: unchanged — Stage F dwellings cluster session. Today added one more reason: frontmatter now OPERATIVE + §5 AUTHORED means the spec corpus's `class:` enumeration contract has more weight; Stage F's five new Compass surfacings must honor it. **Literal question for next-Claude**: *when the steward next checks in, is the day shaped for Stage F (~3h focused) — or another fragmented day where β-style closure is the right posture?* Secondary check: has MemPalace cut `chore/release-3.3.6-prep`?
 ## Archived (2026-05-20 evening — Stage E closure + ADR-007 Site Edition + §VII.h 404 + §6 Dream case + 404 page implementation)
 - [Session 2026-05-20 (evening) — Stage E continuation CLOSED; ADR-007 Site Edition + §VII.h 404 Page + §6 Dream case record + 404 page implementation + §I.f colon convention; 13 commits today on branch 26 ahead of main, pushed both remotes; Path B chosen for merge + deploy (both wait for Stage F dwellings-cluster session); tomorrow is travel + last taxes with Marie + 3h rehearsal — light options laid out (α `/wrap-up` ritual; β status-block housekeeping; γ Stage F prep)](session-2026-05-20-evening-stage-e-closure-adr-007.md) — **Post-/clear afternoon + evening session.** Phase 3 fully closed: 3A tags / glyph_id / glyph_confidence retirement (252 files; **contamination check** surfaced two flags resolved by empirical render-test — `glyph_id` had live consumers but the `$if(has_glyph)$` gate is universally false; `$if(tags)$` block exists but never renders); 3B navigation.yaml retired; **3C ADR-007 Site Edition convention** + colophon Build register wiring (scope expansion from `$build_date$`/`$git_commit$` wiring — surfaced from *"What exactly is 'Last build' counting as a build?"*; Cabal **decoupled by design** per jurist iteration; **Second Edition** lockstep with AldineXXI major; stamp-file via Makefile `build-info` target; `loadBody` against `match "_build_info/*"` rule); 3D five corpus reclassifications (Heliotrope/Graffiti Girl/Dec 22 → meditation/reverie/meditation; Death of a Hero → hommage; Matthieu Taliesin → heteronym); 3E lex.html → about-page.html consolidation with conditional breadcrumb absorbed. Plus desk.md correction (Hearth sub-page); **§I.f term-gloss colon convention** per Hochuli + Sousa via **chamber-typography palace direct query, 40,896 drawers** (steward provided HF token mid-session, saved to `~/.env.local` mode 0600); build-verify-deploy workflow doc from a `make dev` deploy incident. Phase 5 A: **404 page per 2026-05-20 steward spec** — title-as-return rhymes with frontispiece title-as-arrival at h2 scale; **visual-contact override on centering** (spec said NOT, render said YES — steward catch: *"misalignment seems careless"*); §VII.c link decoration corrected to inherit (steward catch: *"`<a>` styling is inconsistent with the rules"*); `head_title` template override field added (Pandoc treats YAML empty/null/false as truthy). **§VII.h codified** per jurist (*"the spec is where the *why* lives so the *what* can be revisited without re-deriving it"*). **§6 Dream case record as fifth retroactive ratification** (Dream locked 2026-04-21 earlier than the four cases the procedure crystallised from; first-instance commitment landed 2026-05-19 same day as procedure's authoring; **steward practice-mechanics recorded as ABSENT — load-bearing**, second case in §6 with this shape alongside Reverie). Phase 5 C (Hagen Quartett Hommage) deferred with grief. Phase 5 D substantively done. **Tonight's load-bearing decision: Path B** — merge + deploy both wait for Stage F (~3h focused) to close corpus-Compass gap so Second Edition + (c) Compass principle are honored at the first reader-facing event together. **Drift patterns named afternoon/evening**: `trust-prose-over-read-makefile`; `audit-delta-claim-incomplete-on-dead-data-fields`; `spec-said-NOT-X-rendering-said-X-needed`; `link-decoration-override-creates-de-facto-spec-exception`. **Tomorrow** (travel Barcelona → Paris 6am + **last taxes filing with Marie, sad** + 3h rehearsal, 2–3h max fragmented): not a Stage F day. **Pulling thread**: nothing pulls forcefully tonight; if forced singular, the Stage F dwellings cluster session is what the branch is waiting for. **Literal question for next-Claude**: *how is the steward's energy after the morning's taxes filing with Marie — and what shape of work, if any, fits the afternoon?*
 ## Archived (today's morning — pre-/clear session)
 - [Session 2026-05-20 — Stage E §4 *Overlap and Resolution* authored + committed (15f8134) + pushed both remotes; pulling thread is Phase 1D Colophon+Vignette register-marks rewrite + Plex Mono mobile CSS fix](session-2026-05-20-stage-e-section-4-authored.md) — **Full day Tuesday Barcelona, ~8h, steward alone.** Pre-execution morning (~31 min before /clear at 10:33): 8 of 8 supplementary-authority texts in chamber-library (Genette ×2 + Hadot ×3 + Austin + Szondi + Auerbach Mimesis via PDF source) + 8 originals filed to `~/Documents/.../_2026 chamber source cleanup/`. **Wake-from-mid-day-/clear ~10:34**; re-orient + Symmetria init. **MemPalace hook diagnostic** (~10:30-45): PreCompact fires but fails — #1440 bash 3.2 mapfile silent-failure + #1538 120 GiB upsert; both fixed on develop, wait for release. **Phase 1B three decisions settled with principles articulated** (~10:45-11:33): (i) §4 scope — three threshold edges in §4 prose; Threshold Breath held ("I don't want a new type"); Q2 Poem RESOLVED as `medium:` axis extension (verse first-instance *Time is not a river* 2025-11-19); Q12 Vespers RESOLVED as `class:reverie` (carnet-traditioned reverie; hommage-gestures as register-shifts within); D-1 footnote-only. (ii) Colophon + Vignette both adopt §XII.b register marks via **synthesis posture** (Manutius octava × medieval marginalia) and **teaching/guiding distinction**; steward's spec-text sentence verbatim: *"Register marks quietly remind the reader where they are, to let them see their place in the whole. A reassurance, not an imperative."* (iii) **Plex Mono reframed from corpus-wide tuning to mobile cascade conflict** — `_mobile-responsive.scss` lines 137-145 override `$plex-mono-scale`; (c-bundled) fix. **Phase 1C §4 fully authored** (~11:33-12:35) — three threshold-case essays composed by steward at essay-depth (M⇌R + H⇌T + G⇌R); masters'-reading consulted direct from chamber-library; Pass 2 added *fidelity-direction asymmetry* + *anonymous tombeau* + *backward-and-forward* closing; Pass 3 added *length asymmetry* + *citation problem* + *summary-reading-as-displacive* + *ARC default toward gloss*; closing triptych-completion (centripetal *meditation+hommage+gloss* / centrifugal *reverie+tombeau+reading* + diagnostic procedure); non-convergence closing line added. **Jurist iteration applied** (3 concerns light-touch): register-asymmetry signal at M⇌R; apprentice/master reframed as ARC's synthesis; corpus-citation precision. **Atomic commit `15f8134`** pushed both remotes; pre-commit hooks passed; +295/-7. Branch 26 commits ahead of main. **7 returns logged** in `session-ledger-2026-05-20.md`; 5 new drift-patterns named. **Pulling thread**: Phase 1D Colophon+Vignette register-marks rewrite + §XII.b spec amendment (steward composes; executor registers) + Plex Mono mobile CSS fix. **Literal question**: what rename works for Vignette's *What the Vignette Must Not Do* section in the §XII.b reassurance-not-imperative discipline? Read vignette.md first; rename follows from what the section actually says. Candidates surfaced: Posture / Restraint / Indifference / Discipline / Constraint.
 ## Archived (today's morning — planning + pre-execution conversion)
 - [Plan — Day 2026-05-20 (Tuesday, Barcelona)](plan-day-2026-05-20.md) — Day-plan composed 08:30 after end-to-end read of canonical ARC sources per steward direction *"read everything in full before continuing."* Energy curve declared. Pre-execution morning (~31 min) added 8 supplementary-authority texts to chamber-library + filed source originals. The full session memory above carries today's execution; this plan is the morning's authoring substrate.
 ## Archived (yesterday evening — Stage E substantially complete)
 - [Session 2026-05-19 (evening) — Stage E substantially complete; Live-State Discipline named; six commits in clean sequence on docs/spec-extraction-2026-05; pulling thread is §4 *Overlap and Resolution* + tags/glyph_id retirement in the morning](session-2026-05-19-evening-stage-e-substantially-complete.md) — **Long substantive evening, ~6.5h end-to-end. Began as audit-delta verification pass; steward caught a partial-read of project-arc-rework.md early; full re-read produced v2 + surfaced an architectural frustration about disciplines that produced the discipline named today (Live-State Discipline).** Three moves authorized then: audit-delta v2 clean rewrite + project-arc-rework.md back-filled 2026-05-07→2026-05-19 + MEMORY.md elevated with Canonical Workstream Trackers section. Steward authorized 3-hour focused window; **7 gates from audit-delta §8 settled in dialogue** (~30 min) before authoring began. **Stage E single-PR landed (`c480b09`)** — cross-spec reconciliation across AldineXXI / apparatus-spec / content-typology / vignette / essay-versioning / frontmatter; §XII.e RESEARCH → LOCKED against LIGATURE first instance; 11 content file migrations. **Jurist PENDING-[N] surfaced** four internal-consistency findings; three already addressed in c480b09; fourth — §VI table glyph state — overshoot in the exact shape Draft 2 (executor epistemics directive, held for separate authorization) names: projecting ADR-direction into spec text as if it were live. Jurist surfaced **option (c)** for the eight-sigil tradition question: Manutius posture as constraint-on-growth, not fixed integer; *the Compass surfaces every form in the corpus that has at least one published instance; the count tracks the corpus*. **Steward authorized hardening + (c)**. **Live-State Discipline filed** as named principle in `feedback-live-state-discipline.md`: *spec text records what IS; status blocks record what is pending; spec text may ship ahead of live only when the spec IS the implementation (§XII.e Capsules is the canonical exception)*. Discipline travels beyond ARC. **Stage E hardening (`2a2c195`)** + **Gate 7 project CLAUDE.md (`558f9f3`)** committed. Steward had energy left: chose §3 then §6. **§3 Qualifying Marks Interaction authored (`b1fbb0e`)** — five interaction rules, authoritative per-type table mirroring §2 Sub-table 3, ✦ within-piece crossing refinement per 2026-04-21 corpus testing. **§6 Change Policy authored retroactively (`0d43ebc`)** — eight-step procedure crystallised from the four 2026-04-23 + 2026-05-14 type additions; practice-grounded; master-reading direct before commitment; steward names practice-mechanics verbatim; non-convergence principle. Steward surfaced Dec 23 viola dream piece (the (c) Compass principle in action — *no slot without an instance*; the Dec 23 piece has been in the corpus since 2023-12-23 just mis-classed); **reclassified observation → dream (`3607400`)** + §1 Dream entry updated to "first instance" framing. **Tags retirement architectural question surfaced**: audit confirmed tags are dead data in ARC (zero site.hs consumers; zero template renders; 213 corpus posts carry the field; frontmatter-spec describes "tag indexes" that don't exist). **Steward authorized (A) retire entirely, deferred to morning**. Observational pass on other dead code surfaced **glyph_id + glyph_confidence** as natural same-shape pair (already-quarantined Living Epistemic Glyph System fields); navigation.yaml uncertain status worth 5-min verify; everything else already-named in CLAUDE.md Deprecated section. Six commits in clean sequence on `docs/spec-extraction-2026-05`, all pushed both remotes. Stage E continuation remaining: **§4 *Overlap and Resolution* (compositional, ~30-60 min, needs fresher energy) + tags + glyph_id + glyph_confidence retirement (mechanical sweep across ~213 posts)**. **Pulling thread**: §4 + tags retirement in the morning. **Literal question**: which §4 genre-riddle boundaries does the steward want the morning's authoring to actually resolve in §4 prose, vs which stay in §7 Open Questions? And: tags-first OR §4-first OR both-in-some-order?
-## Archived (this morning — ARC about-cluster cascade)
+## Archived (yesterday morning — ARC about-cluster cascade)
 - [Session 2026-05-19 (morning) — ARC about-cluster cascade + §XII.e capsule first instance + /readings/ routes shipped; Stage E (peer-spec propagation) pending](session-2026-05-19-arc-cluster-cascade-stage-e-pending.md) — **Substantial ARC dwelling — seven commits on `docs/spec-extraction-2026-05` pushed to both `origin` (Skemantix) + `github-backup`.** /aldine-xxi/ rewritten as the §XII.e capsule first instance with LIGATURE (Vico → Leopardi → Harrison); about-cluster brought into single cascading structure (Lares → AldineXXI → Attention → Chamber → Vignette → Colophon → Now); /tools retired; /now reworked to quiet prose; /attention rewritten — Generosity (Weil) / Witnessing (Tolkien long defeat + ligature held open) / Reading; /glosses/ + /readings/ routes shipped — Brown post lands as Readings first instance (593f2dc). Post-publish tuning: external-link indicator retired; Brown headings dropped; breadcrumb class-route bug fixed; Chamber removed from Compass. Compass nine sigils. Audit principles surfaced: first-instance-settles-the-convention; critique-grounded-in-incomplete-reading-of-framework; chrome-vs-body; same-concept-different-forms; class-with-no-rule-is-a-smell. Steward preference saved: Fowler quotation rules. ADR-006 v2 stages A-D done; Stage E pending with audit-delta verification pass as precondition. 21 ARC-related project memory files; 12+ unread. **Pulling thread**: Stage E with audit-delta verification pass as precondition. **Resolved in this evening's session.**
 ## Archived (yesterday afternoon — chamber typology survey)
@@ -0,0 +1,20 @@
 ---
 name: feedback-querying-default-palace-when-question-belongs-to-specific-palace
 description: "When the steward asks about the typography masters (or any subject whose canonical corpus lives in a non-default palace), default to checking which palace holds the answer before running queries. The MCP is bound to one palace; the corpus may live in another."
 metadata: 
  node_type: memory
  type: feedback
  originSessionId: b43eb145-42db-4a4e-bc75-ed563f3473dc
 ---
 # Querying the default palace when the question belongs to a specific palace
 **Rule:** When the steward asks about content that lives in a known specialised palace (e.g., the typography masters, the chamber-library corpus), reach for the CLI against the specific palace first, not the MCP-bound default palace. If you don't know which palace holds the answer, ask before querying.
 **Why:** The MCP (`mcp__mempalace__*` tools) is bound to a single palace at startup via `~/.mempalace/config.json` — currently `palace-memory` (session continuity, claude-code KG, diaries). Specialised corpora live in separate palaces — currently `palace-chamber-typography` (40,894 drawers; Bringhurst, Tschichold, Hochuli, Lupton, Lacroux, et al.). Querying the MCP for a Bringhurst-shaped question returns *conversation transcripts about Bringhurst*, not Bringhurst's own prose — the wrong room. On 2026-05-20 the steward had to interrupt mid-search with *"the specific typography palace, not the default"* to redirect; the substantive answer to an em-dash question (in lists) was waiting in the chamber-typography palace the whole time.
 **How to apply:** Before running mempalace queries on a corpus-shaped question, ask: *which palace would hold this?* If it's typography/masters → CLI against `palace-chamber-typography`. If it's session/diary/KG → MCP. If unsure → check `~/.mempalace/` for which palace directories exist, or ask the steward. See `[[reference-mempalace-typography-palace]]` for the canonical procedure.
 This drift is adjacent to but distinct from `acting-on-inferred-not-read` (which is about claiming-without-verifying) and `wrong-comparison-baseline-treated-as-current` (which is about checking against the wrong git branch). The shared shape: *the right answer is somewhere; reaching the wrong place by default*. Discipline: name the source-of-truth before querying it.
 **KG entry:** `claude-code drift-pattern querying-default-palace-when-question-belongs-to-specific-palace` (valid_from 2026-05-20). To be added to KG at /wrap-up.
@@ -0,0 +1,342 @@
 ---
 name: plan-day-2026-05-20
 description: "Day-plan for Tuesday 2026-05-20 Barcelona (~8h, steward alone). Close Stage E continuation + 2 Stage F bounded items (5 corpus reclassifications + lex.html→about-page.html consolidation). Composed 08:30 after end-to-end read of canonical ARC sources. Energy curve: composition/decisions morning; mechanical housekeeping mid; short composition late. Three decisions to settle first (§4 scope; register marks for Colophon+Vignette; Plex Mono/Garamond optical balance). Branch docs/spec-extraction-2026-05 at 3607400."
 metadata:
  type: project
  originSessionId: planning-2026-05-20-morning
 ---
 # Day Plan — 2026-05-20 (Tuesday, Barcelona)
 ## Frame
 - 8 working hours, steward alone in Barcelona
 - Energy curve declared: **composition / decisions early · mechanical housekeeping mid · short composition late (30–45 min)**
 - *Bien-fait* per Prime Directive — proportion, fittingness, durability
 - *τὸ πρόσφορον — what is fitting — includes the time the task requires*
 - Steward direction: don't underestimate energy; better to say end-of-day *"I am too tired to continue"* than to plan timidly
 ## Pre-execution morning (08:30–10:00, COMPLETED before this session's /clear)
 **Three-threshold notes drafted by steward** (Meditation ⇌ Reverie · Hommage ⇌ Tombeau · Gloss ⇌ Readings). These are the substrate for §4 prose authoring — shared matrix + differentiating gesture + threshold marker for ARC + one supplementary authority per threshold. Drafts captured in session conversation; available in session memory authored at /wrap-up.
 **Conversion session: 8 of 8 supplementary-authority texts now in chamber-library:**
 | Text | Path |
 |---|---|
 | Genette *Palimpsestes* | `traditions/critical_modernity/french_theory/palimpsestes-la-litterature-au-second-degre-genette.md` |
 | Genette *Seuils* | `traditions/critical_modernity/french_theory/seuils-genette.md` |
 | Hadot *La Citadelle intérieure* | `traditions/ancient_wisdom/stoic_voices/la-citadelle-interieure-hadot.md` |
 | Hadot *Exercices spirituels* | `traditions/critical_modernity/phenomenology/exercices-spirituels-et-philosophie-antique-hadot.md` |
 | Hadot *La philosophie comme manière de vivre* | `traditions/critical_modernity/phenomenology/la-philosophie-comme-maniere-de-vivre-hadot.md` |
 | Austin *How to Do Things with Words* | `traditions/critical_modernity/epistemology/how-to-do-things-with-words-austin.md` |
 | Szondi *On Textual Understanding and Other Essays* | `traditions/critical_modernity/literary_criticism/on-textual-understanding-szondi.md` |
 | Auerbach *Mimesis* (German, 20 chapters + Register) | `traditions/critical_modernity/literary_criticism/mimesis-auerbach.md` |
 **Auerbach acquisition resolved ~10:25** — steward provided clean PDF source (22MB, Xerox+Distiller, 530pp, valid text layer); `pdftotext` + custom heading-promoter (`clean_pdf_running_heads.py`) produced 210k-word MD with 20 chapter H1s in order. Acquisition-debt memo updated to RESOLVED with forensic record of original .mobi failure preserved.
 **Source originals filed** to `~/Documents/___The Library [ePub_AWZ3]/_2026 chamber source cleanup/` (8 files — 6 EPUBs + 2 PDFs). Failed Auerbach .mobi remains in `New texts may 2026/` (not a converted original).
 **Side artifacts preserved:**
 - `chamber-library/scripts/promote_fayard_headings.py` — reusable Fayard chlab/chtit/niv1tit EPUB preprocessor
 - `chamber-library/scripts/clean_pdf_running_heads.py` — reusable PDF→text heading-promotion + running-head/page-number stripper
 **Master-readings status for §4 prose authoring (UPDATED 10:25):**
 - **Meditation ⇌ Reverie** — fully library-direct (Bachelard *Poétique de la rêverie* + Buddhaghosa *Visuddhimagga* + Weil *Gravity and Grace* + *Waiting on God* + Hadot triplet)
 - **Hommage ⇌ Tombeau** — Harrison *Dominion of the Dead* Ch 5 (load-bearing) + Austin *How to Do Things with Words* (supplementary) + Szondi *On Textual Understanding* (essay collection — Trauerspiel-specific book gap remains)
 - **Gloss ⇌ Readings** — **fully library-direct** — Carruthers *Book of Memory* + Genette *Seuils* + Genette *Palimpsestes* + Auerbach *Mimesis* (NEW; the methodological example the steward named explicitly)
 ## What closes today (target — UNCHANGED)
 **Stage E continuation closed + 2 Stage F bounded items done.**
 After today:
 - Stage E continuation is structurally complete (Stage E primary + hardening + Gate 7 + §3 + §6 + Dream first-instance + §4 + tags retirement + Colophon wiring + mobile-review decisions)
 - 5 corpus reclassifications committed → Meditation/Reverie/Hommage/Heteronym meet published-instance criterion per (c) Compass framing
 - `lex.html` template absorbed into `about-page.html` (Gate 1 path b)
 - Stage F clean menu remaining (sized for future sessions)
 ## Phase 1 — Morning composition + decisions (08:30 – 12:30)
 ### 1A. Re-orient (08:30 – 08:45)
 - Read this plan
 - Read `docs/AldineXXI-Codex/_audits/audit-delta-2026-05-19.md` §7 (Stage E recomposed) + §8 (open steward questions)
 - Read project CLAUDE.md to refresh post-Stage-E live state
 - `cd ~/_Dev/animal-davidglidden-eu && git status && git log --oneline -3` on `docs/spec-extraction-2026-05`
 - `/symmetria init`
 ### 1B. Three steward decisions (08:45 – 09:30)
 Settle before §4 prose authoring.
 **Decision (i) — §4 scope.** Which genre-riddle boundaries resolve in §4 prose vs stay in §7 Open Questions?
 - Meditation vs Reverie (Bachelard / Visuddhimagga axis)
 - Hommage vs Tombeau (Harrison Ch 5 speech-act-sema vs form-enacted-sema)
 - Gloss vs Readings (Genette / passage-subject vs work-subject axis)
 - Threshold Breath (3 paths, parked 2026-04-21)
 - Q2 Poem (open since spec inception)
 - Q12 Vespers (open since spec inception)
 **Decision (ii) — Register marks for Colophon + Vignette.** Per 2026-05-19 mobile review:
 - (a) Adopt §XII.b register marks for both — cluster-consistency
 - (b) Keep both as principled exceptions — Colophon's Lineage/Tools/License/Build/Acknowledgement is its own register family; Vignette is the rendered demonstration of the form, not a register-shifted dwelling page
 - (c) Differentiated — Colophon adopts; Vignette stays as demo
 **Decision (iii) — Plex Mono / Garamond optical balance in /aldine-xxi/ companions.** Per 2026-05-19 mobile review (*"the plex mono text in the companions section does not render proportionally to Garamond — it is bigger"*):
 - (a) Companions-context-specific apparatus scale adjustment (~0.82–0.85 instead of corpus-wide 0.88)
 - (b) Corpus-wide apparatus scale re-tuning + reading-test against other apparatus surfaces (sidenotes, dateline, sources, register markers)
 Reading-test protocol (§XIV) applies for both — visually verify against real ARC prose at reading length, both modes, desktop + mobile, before committing.
 ### 1C. §4 *Overlap and Resolution* prose (09:30 – 11:30, ~2h)
 Per §6 Change Policy procedure:
 - Master-readings consulted direct (Bachelard *La Poétique de la Rêverie*; Harrison *Dominion of the Dead* Ch 5; Genette *Seuils*; Carruthers *Book of Memory*; Jaccottet semaison; Handke carnets)
 - Steward names practice-mechanics verbatim where the boundary requires it
 - Executor as registrar for boundary articulation
 - Scope: per Decision (i) — resolve named boundaries in §4 prose; held boundaries get §7 Open Questions notes
 Output: content-typology §4 DRAFT → AUTHORED.
 Atomic commit: `spec(arc): Stage E continuation — content-typology §4 Overlap and Resolution authored`.
 ### 1D. Mobile-review decisions implementation (11:30 – 12:30, ~1h)
 Per Decisions (ii) and (iii):
 **If (ii) = (a) or (c):** add §XII.b register marks to colophon.md and/or vignette.md content. SCSS adjustments if needed. Verify on local server.
 **If (iii) = (a):** companions-context-specific selector + apparatus scale variable adjustment. Verify reading-test on /aldine-xxi/ (companions section) vs other apparatus surfaces.
 **If (iii) = (b):** corpus-wide apparatus scale adjustment + reading-test across multiple apparatus surfaces (sidenotes, dateline, sources, register markers) on multiple piece types.
 Atomic commits per decision (one for register marks; one for apparatus scale tuning).
 ## Phase 2 — Lunch (12:30 – 13:30)
 ## Phase 3 — Afternoon housekeeping + Stage F bounded items (13:30 – 16:30)
 ### 3A. Tags + `glyph_id` + `glyph_confidence` retirement (13:30 – 14:30, ~1h)
 **Scope:**
 - `sed`/`grep`-driven sweep across ~213 corpus posts to remove `tags:` field + remove `glyph_id` + `glyph_confidence` if present (the already-quarantined Living Epistemic Glyph System fields per CLAUDE.md line 219)
 - `frontmatter-specification.md` §2.4 (tags) row retired; §3 translation table entry retired
 - AldineXXI §VII / content-typology §7 Q3 RESOLVED with cross-reference (if Q3 explicitly names tags-cloud); verify exact spec location before commit
 - `site.hs` verify: no `getTags` / `buildTags` / `tagsField` exist (confirmed audit-delta finding); no template renders `$tags$` (confirmed)
 - Build clean check
 Atomic commit: `housekeeping(arc): retire tags / glyph_id / glyph_confidence dead-data fields across corpus + frontmatter spec`.
 ### 3B. `navigation.yaml` verify (14:30 – 14:45, 15 min)
 - Check whether `navigation.yaml` is referenced anywhere in `site.hs` or templates
 - If dead: retire per CLAUDE.md Deprecated section pattern (move to `docs/archive/deprecated/` with brief README)
 - If active: confirm correctness; no action
 Atomic commit if retired: `housekeeping(arc): retire orphan navigation.yaml`.
 ### 3C. Colophon `$build_date$` + `$git_commit$` Hakyll wiring (14:45 – 15:15, ~30 min)
 - `site.hs`: add build-context fields:
  - `$build_date$` — current date at compile time via `dateField` or `unsafePerformIO`-equivalent
  - `$git_commit$` — short hash from `git rev-parse --short HEAD` at build time (likely via unsafePerformIO or a build-time generation step)
 - `templates/about-page.html` or wherever colophon renders: verify `$build_date$` + `$git_commit$` placeholders are present in the Build register block (Lineage / Tools / License / Build / Acknowledgement)
 - `make rebuild` + verify on local server
 - Verify on mobile after deploy
 Atomic commit: `content(arc): Colophon Build register — $build_date$ + $git_commit$ Hakyll wiring`.
 ### 3D. Five corpus reclassifications (15:15 – 16:00, ~45 min)
 Per audit-delta §4 corpus-test verdicts (2026-04-21):
 | Piece | Current | Target |
 |---|---|---|
 | Heliotrope (2020-05-19) | `class: observation` | `class: meditation` |
 | Graffiti Girl (2020-04-23) | `class: observation` | `class: reverie` |
 | Dec 22 half-light Bach (2023-12-22) | `class: observation` | `class: meditation` + resolve duplicate (`content/archive/December 22nd 2023.md` deleted in `5c059d8` 2026-04-26; verify) |
 | Death of a Hero (2016-03-05, Seoul, Korea) | `class: essay` | `class: hommage` |
 | Matthieu Taliesin (2025-03-06, Barcelona, Spain) | `class: essay` | `class: heteronym` |
 | **Vespers for the Living** (2025-06-08, Cremona, Italy) | `class: essay` | held until Q12 resolved (in §4 or in §7) |
 Each reclassification: edit frontmatter only; verify build clean; verify the routes for the new class exist (Meditation has surfaced route; Reverie / Hommage / Heteronym are governed-but-unsurfaced — routes emerge with first instance per non-convergence principle, which now applies since these are the first instances).
 **Surface implication:** after this commit, per (c) Compass framing — Meditation, Reverie, Hommage, Heteronym all have published instances. Their Compass slots become Stage-F-implementation rather than anticipated. Compass swap deferred to bundled Stage F dwellings cluster session (~3h, separate session) plus the per-type implementation work for ❦ Hommage and ⁂ Heteronym slots.
 Atomic commit: `content(arc): five corpus reclassifications — Heliotrope/Graffiti Girl/Dec 22/Death of a Hero/Matthieu Taliesin per audit-delta §4`.
 ### 3E. Template consolidation `lex.html` → `about-page.html` (16:00 – 17:00, ~1h)
 Per yesterday's Gate 1 path (b) (audit-delta §1 table, Stage E Gate 1):
 - Read `templates/lex.html` + `templates/about-page.html` to understand current scope
 - Absorb `lex.html` content into `about-page.html` (likely as conditional blocks; verify what differs)
 - `site.hs` routing: `content/lares/*.md` now use `about-page.html` template instead of `lex.html`
 - Verify Hearth (`/lares/`) + lares/posture + lares/marks + lares/desk render correctly
 - Retire `templates/lex.html` — preserve at `docs/archive/templates/lex.html` with brief README explaining provenance (per `feedback-preserve-dont-delete-when-retiring.md` discipline)
 - Verify build clean; reading-test on /lares/ pages
 - `make rebuild`
 Atomic commit: `arch(arc): consolidate lex.html into about-page.html (Stage E Gate 1 path b)`.
 ## Phase 4 — Break (16:30 – 17:30)
 ## Phase 5 — Late composition (17:30 – 18:15, 30 – 45 min)
 Steward chooses based on remaining energy:
 **A. 404 page short prose** — steward-voice composition; quiet ARC register; one short paragraph (Latin or English). Executor scaffolds `templates/404.html` + `site.hs` build path if not yet present per `project-arc-404-page-design-pending.md`. Possibly the place the future Close ornament makes most sense (*finis* of an unintended journey).
 **B. §6 retroactive case record for Dec 23 Dream reclassification** — fifth case record alongside Heteronym + Hommage + Tombeau + Readings. Applies the eight-step procedure retroactively for the Dream first-instance (Dec 23 viola dream, reclassified yesterday in commit `3607400`). Master-reading already direct (Jaccottet semaison carnets); steward practice-mechanics implicit in the existing §1 Dream entry; the case record formalizes what was already done.
 **C. Hagen Quartett Hommage anticipation draft** — a few paragraphs naming the piece's shape; held for future composition session. Steward composes; executor as registrar.
 **D. Pulling-thread update + literal question for next-Claude** — written into this plan file or as a separate small composition. Names what closed today; what carries forward; the morning's first move for tomorrow.
 ## Phase 6 — Wrap-up (18:15 – 18:45)
 - `/wrap-up` to capture state cleanly
 - Update this plan with what landed (which items closed; which carried forward)
 - Update `project-arc-rework.md` canonical tracker with today's substantive moves (per canonical-workstream-tracker discipline)
 - Pause statement + literal question for next-Claude
 - Tomorrow's pulling thread (or leave for tomorrow's `/wake-up`)
 - Push branch to both remotes if not already pushed
 ## What's explicitly out of scope today
 | Item | Reason |
 |---|---|
 | L1/L2 work | Parked through end of May 2026 per global CLAUDE.md |
 | MemPalace upgrade | Waiting for 3.3.6 release or explicit steward authorization |
 | Studium Engine | Separate workstream |
 | Chamber-typography mining | Separate workstream (Studium Engine territory now) |
 | Vignette implementation Phase 1 | 5–7 days — own arc |
 | PDF/Print production wiring | 2–3 days — own arc |
 | EXIF Tier 3 | 3–5h focused — own session |
 | SCSS cascade audit | Gated on Stage G (spec-lock as precondition) |
 | First Photo Essay | One afternoon, steward composition |
 | Dwellings cluster route + Photo Essay ⊙ → ◧ glyph swap | 3h bundled — own session (route decisions + footer-compass + frontispiece updates atomic) |
 | Full ornament migration (~213 `.ornament.*` uses) | 4–6h focused — own session |
 | SCSS audit candidates pass | Deferred (low priority; can ride a future session) |
 ## Stage F remaining after today
 | Item | Est. | Type |
 |---|---|---|
 | Dwellings cluster route + ◧ glyph swap | 3h focused | implementation + atomic commit |
 | Full ornament migration | 4–6h focused | content migration + spec close |
 | EXIF Tier 3 | 3–5h focused | investigation + scope-dependent impl |
 | First Photo Essay | one afternoon | steward composition |
 | /vignette page rework | gated on §IX impl | (no estimate) |
 | About-cluster voice audit | gated on /vignette | (no estimate) |
 | Vignette implementation Phase 1 | 5–7 days | multi-session |
 | PDF/Print production wiring | 2–3 days | multi-session |
 Stage G jurist review can begin in parallel once corpus is internally consistent at spec layer — after today's Stage E close + the planned Stage F bounded items, the spec layer is consistent; Stage F implementation continues; Stage G doesn't block on Stage F implementation completing.
 ## Branch state at session-start
 `docs/spec-extraction-2026-05` at `3607400`, 12 commits ahead of `main`, pushed both remotes (Skemantix `origin` + GitHub `github-backup`). Working tree noise: `tools/.DS_Store` + two font backup dirs (uncommittable, carried forward).
 Today's commits land on the same branch. Expected 6–8 atomic commits (see Commit shape below).
 ## Commit shape
 Per ADR-006 v2 Part 5 atomic-commit discipline. Today's expected commits (in order):
 1. `spec(arc): Stage E continuation — content-typology §4 Overlap and Resolution authored`
 2. `spec(arc): Stage E continuation — register-marks decision for Colophon/Vignette` (if Decision ii = a or c)
 3. `spec(arc): Stage E continuation — apparatus scale tuning per /aldine-xxi/ companions reading-test` (per Decision iii)
 4. `housekeeping(arc): retire tags / glyph_id / glyph_confidence dead-data fields across corpus + frontmatter spec`
 5. `housekeeping(arc): retire orphan navigation.yaml` (if dead) or `housekeeping(arc): verify navigation.yaml active` (note only)
 6. `content(arc): Colophon Build register — $build_date$ + $git_commit$ Hakyll wiring`
 7. `content(arc): five corpus reclassifications per audit-delta §4`
 8. `arch(arc): consolidate lex.html into about-page.html (Stage E Gate 1 path b)`
 Plus late-window commit if composition lands.
 Each commit pushed to both remotes immediately after landing (per `feedback-deploy-is-not-commit.md` discipline — explicit push, not implicit through deploy).
 ## Discipline reminders (foreground throughout)
 - **Live-State Discipline** — spec text records what IS; status blocks record what is pending; spec text ships ahead of live only when spec IS the implementation (§XII.e Capsules canonical exception). Filed yesterday in `feedback-live-state-discipline.md`.
 - **(c) Compass framing** — the Compass surfaces every form in the corpus that has at least one published instance; the count tracks the corpus; ornamental and anticipated routes are refused. Manutius posture preserved as constraint-on-growth, not fixed integer. Authored 2026-05-19 hardening commit `2a2c195`.
 - **Eight-step type-addition procedure** (content-typology §6) — practice-grounded, not authority-imposed: discovery (pull from work) → master-reading direct (verbatim, not paraphrased) → steward practice-mechanics verbatim → boundary articulation → glyph assignment → first-instance commitment → atomic spec amendment → jurist review + retroactive reclassification. Authored 2026-05-19 commit `0d43ebc`.
 - **Bien-fait per Prime Directive** — proportion, fittingness, durability. *τὸ πρόσφορον — what is fitting — includes the time the task requires.*
 - **Read before composing** — name the gap; ask or chunk-read; never compose past it on the strength of having named it. The lesson of yesterday's audit-delta drift (v1 partial-read of project-arc-rework.md → v2 full re-read; steward correction was the discipline).
 - **Steward composes; executor is registrar** — for prose, register, voice. The executor surfaces material; the steward names. The §6 procedure formalizes this for type additions; the principle extends to all spec composition.
 - **Atomic commit discipline** per ADR-006 v2 Part 5 — single coordinated commit per decision; pushed both remotes immediately.
 - **Reading-test protocol** (§XIV) — any type/scale change visually verified against real ARC prose at reading length, both modes, desktop + mobile, before committing. Applies especially to Decision (iii) apparatus scale tuning.
 - **`/symmetria` operative** — practice of return foregrounded; pulse before non-trivial actions; audit after any sub-agent return.
 ## Authored
 Composed 2026-05-20 morning at ~08:30 Barcelona, after end-to-end read of canonical ARC sources:
 - 8 specs in full: AldineXXI (2847), content-typology (1205), apparatus (661), silence-and-rhythm (308), vignette (334), essay-versioning (302), frontmatter (272), multilingual (1007)
 - 3 audit artifacts in full: audit-delta-2026-05-19 (295), spec-corpus-audit-2026-05-16 (401), arc-state-survey-2026-05-16 (568)
 - ADR-006 v2 (206) in full
 - Project CLAUDE.md (222) in full
 - `project-arc-rework.md` canonical tracker (3376) in full
 - 6 pending ARC project memos in full
 - Session memory 2026-05-19 morning + evening + ledger
 - REVIEWED.md in full
 Reading discipline honored per yesterday's lesson: name the gap; read end-to-end or ask; never compose past it. Per the steward's direction *"read everything in full before continuing."*
 ---
 *Plan committed 2026-05-20 morning Barcelona. Picked up at next `/wake-up` after `/wrap-up + /clear`. Hasten when the work is ready. Not before.*
 ---
 ## End-of-day record  **[appended 2026-05-20 evening]**
 ### What landed (11 commits on `docs/spec-extraction-2026-05`)
 **Morning composition (Phase 1).** `15f8134` content-typology §4 *Overlap and Resolution* (genre-riddle resolved across Meditation⇌Reverie, Hommage⇌Tombeau, Gloss⇌Readings). `982fe7f` Phase 1D register marks for Vignette + Colophon + §XII.b *Posture* + Plex Mono mobile cascade fix.
 **Unplanned mid-day insertions.** `08f844a` build-verify-deploy workflow doc (after a `make dev` deploy incident where executor trusted prose over the Makefile). `e06c77f` em-dash → colon in list term/gloss form per Hochuli + Sousa (chamber-typography palace consult); colophon list items + §I.f spec amendment.
 **Afternoon housekeeping (Phase 3).** `355757d` tags / glyph_id / glyph_confidence frontmatter retirement (252 files; empirical verification gate-is-cold before sweep). `4adf04e` orphan `navigation.yaml` retired. `cf1bc99` site edition convention + colophon Build register wiring (ADR-007 authored; stamp-file mechanism; **Second Edition** designated). `272a4f2` five corpus reclassifications per audit-delta §4 (Heliotrope, Graffiti Girl, Dec 22, Death of a Hero, Matthieu Taliesin). `50c402b` lex.html consolidated into about-page.html (Stage E Gate 1 path b); breadcrumb conditional absorbed; orphan template archived. `6562411` desk.md drops `enfilade: about` per Hearth-sub-page framing.
 **Evening composition (Phase 5).** `e76f0ad` 404 page per 2026-05-20 steward spec; `head_title` template override; `.title-return` SCSS; Jekyll-era 404.md archived. `c45772c` §VII.h *404 Page* codified in apparatus-spec per jurist (*"the spec is where the why lives so the what can be revisited without re-deriving it"*). `04f9d8b` §6 Dream case record as fifth retroactive ratification.
 **Held.** *Phase 5 option C — Hagen Quartett Hommage.* Grief has its own pace; the form will wait for it.
 ### What's pulling
 **Stage F is owed.** Five types now have first instances in the corpus (Meditation, Reverie, Dream, Hommage, Heteronym); their Compass slots are implementation-owed per the (c) Compass framing. The bundled dwellings-cluster session (~3h focused, per the morning plan's Stage F table) covers all five + the Photo Essay ⊙→◧ swap.
 **Merge + deploy decision — Path B chosen tonight.** Both wait until Stage F closes the corpus-Compass gap. Branch sits at `04f9d8b`, 26 ahead of `main`, pushed both remotes (origin Skemantix + github-backup). No merge tonight. The Second Edition designation, ADR-007, §VII.h, the colon convention, the 404 page, the reclassifications — all wait for Stage F to land first, then merge + deploy as one coherent reader-facing event. The (c) Compass principle and the Second Edition lockstep both honoured at the same moment.
 **SCSS audit accumulated items.** Today added two: `.title-return` placement in `_frontispiece.scss` (provisional, flagged); the heading-modular-scale question from 2026-05-19 remains open. Audit itself gated on Stage G (spec-lock).
 ### Tomorrow
 **Travel day** (Barcelona → Paris, 6am flight) + **taxes filing with Marie** + **3h rehearsal**. 2–3h max work time, fragmented. **Not a Stage F day.** See `plan-day-2026-05-21.md` for the three light options laid out (α Phase 6 wrap-up; β status-block housekeeping; γ Stage F prep document). Default posture: closure + light maintenance over advancement.
 ### Literal question for next-Claude
 **How is the steward's energy after the morning's taxes filing with Marie — and what shape of work, if any, fits the afternoon?**
 The day brings emotional weight (last taxes with Marie). The steward will check in tomorrow when they can; the work-shape depends on what they have to give. Default posture: closure + light maintenance. Options laid out in `plan-day-2026-05-21.md`.
 ### Morning's first move
 There may not be one. The first move is wherever the steward arrives — tired from travel, present after taxes, post-rehearsal in evening, or skipping entirely. `/wake-up` if and when they can; then choose from Options α / β / γ based on capacity.
 ### Discipline reminders (carried forward)
 - **(c) Compass principle** (locked 2026-05-19 in `2a2c195`): *"the Compass surfaces every form in the corpus that has at least one published instance; the count tracks the corpus."* Path B respects this at the next reader-facing deploy.
 - **Live-State Discipline**: spec text records what IS; status blocks record what is pending; spec text ships ahead of live only when spec IS the implementation. Today the corpus contains five new first-instance types whose Compass surfacing is pending Stage F; the principle holds because we have not deployed.
 - **τὸ πρόσφορον** — what is fitting includes the time the task requires. Tomorrow's day-shape is not Stage F's day-shape.
 ---
 *End-of-day record appended 2026-05-20 evening. Phase 6 wrap-up step 1/4 complete.*
@@ -0,0 +1,107 @@
 ---
 name: plan-day-2026-05-21
 description: "Day-plan for Wednesday 2026-05-21 — travel Barcelona → Paris (6am flight), last taxes filing with Marie, 3h rehearsal. 2–3h max work time, fragmented. Closure + light maintenance posture; not a Stage F day. Three light options laid out (α Phase 6 wrap-up; β status-block housekeeping; γ Stage F prep document). Default: α first, β if energy holds; γ only if a third hour materialises."
 metadata:
  type: project
  originSessionId: planning-2026-05-20-evening
 ---
 # Day Plan — 2026-05-21 (Wednesday, Barcelona → Paris)
 ## Frame
 - 4am wake; 6am flight from Barcelona; arrive Paris mid-morning
 - **Last taxes filing with Marie** (procedurally easy; emotionally weighted — *sad*, the steward named it)
 - One 3-hour rehearsal in afternoon
 - **2–3h max work time, fragmented across the day**
 - Energy curve: low-to-medium throughout; no focused-implementation block possible
 - *τὸ πρόσφορον — what is fitting includes the time the day actually gives*
 ## What does NOT happen today
 - **Stage F (dwellings cluster + new Compass sigils).** Wrong day for it. Needs fresh-energy focused block (~3h continuous, deep concentration across site.hs + templates + SCSS + multiple type routes). Deferred to next focused-energy day.
 - **Merge + deploy.** Per Path B chosen 2026-05-20 evening: merge and deploy wait until Stage F closes the corpus-Compass gap. Today's branch sits where it is.
 - **New spec authoring.** None is owed.
 - **Composition.** No new prose work scheduled. Hagen Quartett Hommage held with grief.
 ## What CAN happen today (choose based on energy)
 Three light items, any single one fits a ~1h window between travel / taxes / rehearsal:
 ### Option α — Phase 6 wrap-up (carried from 2026-05-20)
 Three items remaining from yesterday's wrap-up:
 - `/wrap-up` ritual invocation (steward action — captures session memory cleanly via the skill)
 - Final commit shape check — confirm both remotes carry today's branch state (already verified 2026-05-20 evening; confirm again if anything shifted)
 - Read yesterday's end-of-day record in `plan-day-2026-05-20.md` to re-orient
 End-of-day record + `project-arc-rework.md` canonical tracker update + remote verification were completed 2026-05-20 evening before sign-off, so most of Phase 6 is done. What remains is the `/wrap-up` skill invocation itself.
 ~15–30 min. Fully mechanical. **Best done first** to clear the deck and arrive in the new day cleanly.
 ### Option β — Status-block housekeeping commit
 Small follow-on for the spec corpus's stale stamps (per the 2026-05-20 "to do" inventory):
 - `content-typology-specification.md` status header: §2 / §3 / §4 / §6 stamps from "DRAFT, pending Stage E" → "AUTHORED, pending Stage G jurist review"
 - `content-typology-specification.md` §1 header: DRAFT → LOCKED (all 14 types individually LOCKED)
 - `content-typology-specification.md` §5 Code Binding Contract: LOCKED stamp if substance is complete, or named gap surfaced
 - `essay-versioning-specification.md` + `frontmatter-specification.md` header DRAFT stamps: read-through to confirm whether stale or accurate; reconcile
 Single atomic commit. ~45–60 min. Low energy; mostly reading + small edits. Brings spec corpus's stated state in line with its actual state — Live-State Discipline applied to the headers themselves.
 ### Option γ — Stage F prep document
 A step-by-step implementation plan for the dwellings cluster session (without implementing). Maps the work cleanly so the focused-energy day can execute without re-deriving scope:
 - The Photo Essay ⊙→◧ swap: exact site.hs lines, template references, route impact
 - Dwellings cluster ⊙ for Meditation + Reverie: route file structure, index template, listing logic
 - New Compass slots: ❦ Hommage, ⁂ Heteronym, ☾ Dream (cosmological pairing rationale per content-typology §1)
 - Default.html Compass-row template edits
 - Per-type index pages (HTML structure, frontmatter, what each lists)
 - Reading-test path for each new slot
 - Atomic-commit shape
 Writes to `docs/AldineXXI-Codex/drafts/stage-f-dwellings-cluster-plan-2026-05-21.md` or similar. ~60–75 min. Higher cognitive load than α / β; needs reading and structural thinking. Best if energy permits.
 ## Recommended sequencing
 If today actually delivers ~2h of work-time:
 1. **First** — Option α (Phase 6 wrap-up closure). ~15–30 min. Closes yesterday cleanly via the `/wrap-up` skill.
 2. **Second** (if energy holds, post-rehearsal or before bed) — Option β (status-block housekeeping). ~45–60 min. Mechanical and satisfying; cleans the spec corpus without requiring deep concentration.
 If only ~1h is available: just α. Closure beats new work on a travel day.
 If energy holds beyond α + β and a third hour materialises: Option γ. But don't push.
 **If energy doesn't materialise at all today**: that is also fine. The branch sits at `04f9d8b`, 26 commits ahead of main, pushed both remotes. Nothing is at risk. Tomorrow or whenever is a new day.
 ## Branch state at session-start
 `docs/spec-extraction-2026-05` at `04f9d8b`, 26 commits ahead of `main`, pushed both remotes (origin Skemantix + github-backup). Working tree clean (font-backup directories untouched; pre-existing noise carried forward).
 ## What's pulling (carried from 2026-05-20)
 - Stage F dwellings cluster session — waiting for fresh-energy day (≥3h continuous block; not today)
 - Merge + deploy — gated on Stage F per Path B (chosen 2026-05-20 evening)
 - Hagen Quartett Hommage — held with grief
 ## Discipline reminders
 - **Travel days are good for orientation + reading + planning.** Bad for focused implementation.
 - **Taxes filing with Marie has its own emotional load.** Procedurally easy, the steward said; sad. Don't sandwich work blocks tightly around it. The grief gets its own room.
 - **Three-hour rehearsal is high-energy work.** Music-presence is its own demand. Don't crowd it with anticipation.
 - **Bien-fait per Prime Directive** — proportion, fittingness, durability. A day with three external demands is not a day to also push major implementation work.
 - **The day's posture is *closure + light maintenance*, not advancement.** Both are honorable. *τὸ πρόσφορον — what is fitting includes the time the task requires.*
 ## Pulling thread + literal question
 **Pulling thread**: nothing pulls forcefully. The branch is at rest; Stage F waits; the merge and deploy wait. Today's posture is responsiveness to what the day actually gives, not advancement.
 **Literal question for next-Claude (carried from 2026-05-20)**: *How is the steward's energy after the morning's taxes filing with Marie — and what shape of work, if any, fits the afternoon?*
 ---
 *Plan composed 2026-05-20 evening Barcelona, ahead of the steward's 4am wake. Wait for the steward's check-in; do not assume direction. Hold space for grief; offer light options when work-time arrives. Hasten when the work is ready. Not before.*
@@ -0,0 +1,57 @@
 ---
 name: ARC chamber-library — Auerbach Mimesis acquisition debt
 description: "RESOLVED 2026-05-20 ~10:25 — steward provided a clean PDF source (22MB, Xerox-scanned + Acrobat-Distilled, 530 pages, valid text layer); pdftotext + custom heading-promotion produced a clean MD with all 20 chapters as H1. Filed at traditions/critical_modernity/literary_criticism/mimesis-auerbach.md. The .mobi failure is preserved below as forensic record + the conditions under which mojibake sources should be rejected."
 metadata:
  type: project
  originSessionId: 2026-05-20-conversion-session
 ---
 # Auerbach *Mimesis* — acquisition debt — RESOLVED 2026-05-20
 **RESOLVED 2026-05-20 ~10:25.** Steward provided a clean PDF source at `/Users/davidglidden/Documents/___The Library [ePub_AWZ3]/New texts may 2026/Mimesis_Dargestellte_Wirklichkeit.pdf` (22MB, Xerox DSMR + Acrobat Distiller 5.0.5, 530 pages, valid text layer). Pipeline: `pdftotext` (clean German text extraction) + `/tmp/clean_mimesis.py` (heading promotion + running-head/page-number cleanup, generalised and saved to `chamber-library/scripts/clean_pdf_running_heads.py`). Output: 17,375 lines / 210,079 words / 21 H1 (20 chapters Die Narbe des Odysseus → Der braune Strumpf, + Register). Filed at `traditions/critical_modernity/literary_criticism/mimesis-auerbach.md`. The original .mobi failure is preserved below as forensic record.
 ---
 # Original memo — .mobi source failure (forensic record)
 ## What happened
 During the 2026-05-20 conversion session (8 texts staged for §4 *Overlap and Resolution* prose authoring), 7 converted clean. Auerbach's *Mimesis: Dargestellte Wirklichkeit in der abendländischen Literatur* (German .mobi, 51MB) failed:
 - **Calibre `ebook-convert`** raised `SplitError: Could not find reasonable point at which to split: index.html Sub-tree size: 322 KB` — the .mobi packages all content in a single oversized HTML file with no split markers.
 - **With `--flow-size=0`** (skip splitting), Calibre produced a 1.1MB EPUB, but the body content is **mojibake** — encoding artifacts from a scan-OCR pipeline that didn't produce clean text. Sample (page-mid): `«10g°ªÖ¼¸'»péjÕÊ€›) ÝÉÞßmžxË!ýô...` — not Auerbach's German prose.
 - **`pandoc -t gfm-raw_html`** on the Calibre EPUB produced 8628 lines / 32,738 words of mostly-mojibake text. 32K words is also very short for Mimesis (the actual book is hundreds of thousands of words), confirming the .mobi was primarily scanned page-images with broken OCR.
 The .mobi's 51MB size is consistent with embedded high-resolution page scans + poor OCR text layers.
 ## What's needed
 A clean source for Auerbach *Mimesis* — one of:
 1. **German EPUB** from Suhrkamp Taschenbuch (paperback edition, well-typeset, clean text layer)
 2. **English translation (Trask, Princeton 1953)** EPUB — equally load-bearing for §4 prose authoring since the §4 supplementary-authority work cites Auerbach methodologically, not in a specific-language register
 3. **Internet Archive scan with verified text layer** if it exists in their collection — would need pdftotext / docling round-trip test before commitment
 ## Why this matters
 Auerbach is the named supplementary authority for the Gloss ⇌ Readings threshold (per the steward's three-threshold notes 2026-05-20):
 > *"Erich Auerbach in Mimesis — every chapter is a textbook case of the threshold case, beginning as gloss on a passage and modulating into reading of an epoch. Watching him cross the threshold paragraph by paragraph is itself a methodological education."*
 For §4 prose, this means: if §4 wants direct-quoted Auerbach demonstration of the gloss → readings crossing, the chamber-library doesn't have a clean source. Either the §4 prose works from the steward's notes + general knowledge of Auerbach's method (acknowledged honestly per `feedback-pandoc-epub-suppress-raw-html.md` source-quality discipline), or Auerbach acquisition runs before §4 authoring.
 ## Recommendation
 Defer Auerbach to a separate small acquisition task (post-§4 prose authoring). The §4 prose can name Auerbach as supplementary authority and quote from the steward's notes; when a clean Auerbach source lands in chamber-library, §4 can be revised in a small follow-on commit per the Live-State Discipline (spec text records what IS).
 ## Cross-references
 - Three-threshold notes shared by steward 2026-05-20 morning (in conversation prior to conversion session)
 - `feedback-pandoc-epub-suppress-raw-html.md` — source quality determines retrieval quality
 - `project-epub-repair-pipeline.md` — the pipeline that worked for 6 of 7 conversions today
 - `feedback-articulate-toolchain-fit-before-executing.md` — per-text toolchain-fit diagnosis (in this case, the .mobi source itself failed before any toolchain question)
 - 2026-05-20 conversion session result: 7 of 8 staged texts converted clean (Palimpsestes, Seuils, Citadelle, Exercices, Philosophie, Austin, Szondi)
 ## Filing
 This memo lives alongside the other `project-arc-*-pending.md` files; entered into MEMORY.md if/when steward authorizes elevating to active acquisition task.
@@ -3374,3 +3374,217 @@ Template consolidation (lex.html → about-page.html); dwellings cluster route +
 Jurist final review of full corpus → content-typology DRAFT → OPERATIVE.
 ---
 ## Session 2026-05-20 (Tuesday, Barcelona) — Stage E continuation §4 *Overlap and Resolution* authored (15f8134); mid-day safety wrap pending afternoon Phase 1D/3/5
 **Day plan composed 08:30** after end-to-end read of canonical ARC sources per steward direction *"read everything in full before continuing."* Energy curve declared: composition/decisions morning · mechanical housekeeping mid · short composition late (30-45 min). Target: close Stage E continuation + 2 Stage F bounded items.
 ### Pre-execution morning (08:30–10:00, before /clear at ~10:33)
 Conversion session ~31 min added **8 supplementary-authority texts to chamber-library** for §4 prose authoring: Genette *Palimpsestes* + *Seuils* (french_theory/); Hadot *La Citadelle intérieure* (stoic_voices/) + *Exercices spirituels* + *La philosophie comme manière de vivre* (phenomenology/); Austin *How to Do Things with Words* (epistemology/); Szondi *On Textual Understanding* (literary_criticism/ — new folder); Auerbach *Mimesis* (20 chapters + Register, from steward-provided clean PDF source via `pdftotext` + custom heading-promoter `clean_pdf_running_heads.py`). 8 source originals filed to `~/Documents/.../_2026 chamber source cleanup/`. Reusable scripts preserved: `chamber-library/scripts/promote_fayard_headings.py` (Fayard EPUB preprocessor) + `chamber-library/scripts/clean_pdf_running_heads.py` (PDF→text heading-promotion + running-head/page-number stripper).
 ### Wake-from-mid-day-/clear (~10:34)
 Re-orient (plan + audit-delta §7/§8 + project CLAUDE.md + ARC repo state) + Symmetria init.
 ### MemPalace hook diagnostic (~10:30-45)
 PreCompact hook fired at 10:26:26 but failed. Two bugs both fixed on develop awaiting release:
 - **#1440** — bash 3.2 `mapfile` silent failure (macOS `/bin/bash` is 3.2.57 GPLv3-frozen; `mapfile -t` doesn't exist → session_id and transcript_path default empty → mine call skipped)
 - **#1538** — chunk-size enforcement (120 GiB upsert RuntimeError from steward's local `MAX_CHUNKS_PER_FILE 500→50000` workaround commit interacting with large transcripts)
 **Decision: wait for release.** Bash-5-wrap in `~/.claude/settings.json` would fix layer 1 but expose layer 2.
 ### Phase 1B — three steward decisions settled with principles articulated
 **Decision (i) §4 scope.** Three threshold-edges in §4 prose today (Meditation⇌Reverie, Hommage⇌Tombeau, Gloss⇌Readings). Threshold Breath held in §7 — steward verbatim: *"I don't want a new type."* **Q2 Poem RESOLVED** structurally as **`medium:` axis extension** (`prose · image · verse`; first-instance *Time is not a river* 2025-11-19, Avignon — Meditation in verse; orthogonal to `class:`). **Q12 Vespers RESOLVED** as **`class: reverie`** with hommage-gestures as register-shifts within (carnet-traditioned reverie per Jaccottet *semaison* / Handke carnets; piece names itself reverie in its own prose: *"a returning, a coming back to something that was always there, waiting"*). D-1 Capsule column (b) footnote-only.
 **Decision (ii) Colophon + Vignette register-marks.** Both adopt `:::register-mark :::` per **synthesis posture** (Manutius octava × medieval marginalia, not Manutius alone — steward correction during reasoning) and **teaching/guiding distinction** (Vignette is *guiding* — the reader passes through the threshold once; teaching would justify headings, guiding justifies register marks). **Steward's spec-text discipline sentence verbatim** for §XII.b small note: *"Register marks quietly remind the reader where they are, to let them see their place in the whole. A reassurance, not an imperative."*
 **Decision (iii) Plex Mono mobile cascade conflict.** Reframed from "corpus-wide scale tuning" to **mobile-responsive CSS cascade conflict** when steward noted desktop renders proportionally. Root cause at `scss/AldineXXI/_mobile-responsive.scss` lines 137-145: mobile override `font-size: 0.9em` (inline `<code>`) + `font-size: 0.85em` (`<pre>`) discards `$plex-mono-scale`'s optical-balance work. Fix: replace both with `$font-size-sm * $plex-mono-scale`. **(c-bundled)** — same atomic commit per the principle *"`$plex-mono-scale` governs Plex Mono rendering at all viewports."*
 ### Phase 1C — §4 fully authored + jurist iteration + atomic commit pushed
 **Three threshold-case essays composed by steward at essay-depth** in three successive passes. Masters'-reading consulted direct from chamber-library before each pass.
 - **Pass 1 Meditation⇌Reverie** — Bachelard *Poétique de la rêverie* + Buddhaghosa *Visuddhimagga* + Weil + Hadot triplet. Key insight: *the split occurs not at the start of the practice but at its first success*. *Papañca* (Buddhist) vs *image stifled by concept* (Bachelard verbatim *"by giving stability to the image, the concept would stifle its life"*) as mutual-failure-mode.
 - **Pass 2 Hommage⇌Tombeau** — Harrison Ch 2 (the speech-act-sema material is in Ch 2 *Hic Jacet*, not Ch 5 — plan was approximate) + Austin *How to Do Things with Words* + Szondi (Trauerspiel-specific text acquisition debt named honestly). Adds *fidelity-direction asymmetry* (hommage owes accuracy to historical person; tombeau owes fidelity to transmitted form) + *anonymous tombeau* as third option + *backward-only vs backward-and-forward* asymmetry.
 - **Pass 3 Gloss⇌Readings** — Genette *Palimpsestes* (strict taxonomy: gloss/reading under métatextualité, not hypertextualité; path (a) acknowledged in §4 prose) + Auerbach *Mimesis* (methodological observation of chapter shape: gloss → reading modulation). Adds *length asymmetry* (gloss bounded by lemma, reading by thesis) + *citation problem* + *summary-reading as maximally displacive third category* + *ARC default toward gloss* as editorial posture.
 **Triptych closing material** — the three axes line up canonically: **centripetal** (meditation+hommage+gloss; locally attentive, named in debt, host-serving) / **centrifugal** (reverie+tombeau+reading; associatively expansive, formally enacting, discourse-producing); diagnostic procedure for triangulating uncertain pieces. **Non-convergence closing** added: triptych is not exhaustive; additional axes emerge from corpus need, not anticipation.
 **Jurist iteration applied** (3 concerns light-touch): (1c) M⇌R register-asymmetry signal as italicized preamble — analytical prose preserved verbatim; (2) apprentice/master claim reframed as *"ARC reads the French tombeau lineage..."* (provenance-honest framing as ARC's synthesis, not tradition's own self-described hierarchy); (3) illustrative-cases note tightened with explicit corpus-grounded vs hypothetical breakdown + Couperin H⇌T-vs-G⇌R parenthetical naming same-source-different-axis-work. Concern 4 verified clean (§3/§2 Annals sync `n/a`/`n/a`). Concern 5 + smaller notes deferred to Stage G cleanup.
 ### Atomic commit + §7 housekeeping + §0 + frontmatter
 **`15f8134`** — `spec(arc): Stage E continuation — content-typology §4 Overlap and Resolution authored`. Pre-commit hooks passed. 2 files, +295 / −7. Pushed both remotes (Skemantix `origin` + GitHub `github-backup`). Branch is `docs/spec-extraction-2026-05`, now **26 commits ahead of main**.
 §7 in same commit:
 - Q2 Poem → RESOLVED 2026-05-20 (medium axis; cross-ref to §4)
 - Q12 Vespers → RESOLVED 2026-05-20 (class:reverie; cross-ref to §4 worked example)
 - Q14 Threshold Breath added — held with rationale (not a new type; emergence-watching per Q11 Louange pattern)
 §0 Multi-axis Medium row extended: `prose · image · verse`.
 `frontmatter-specification.md` §2.7 Medium ADDED 2026-05-20 — documents the `medium:` field with verse as first non-default value (cross-refs §4 and AldineXXI §II).
 ### Returns + drift-patterns
 **7 returns logged** in `session-ledger-2026-05-20.md`: wake / mempalace-baseline-wrong-branch / chamber-library-mine-attempts-missed / folder-moved-absence-asserted / carnet-vs-Harrison-anchoring / Manutius-single-voice-vs-synthesis / time-fabrication / typo-edit-boundary.
 **5 new drift-patterns** named (KG `claude-code` entity, `valid_from 2026-05-20`):
 - `wrong-comparison-baseline-treated-as-current`
 - `asserting-absence-from-non-presence-at-one-path`
 - `anchoring-on-most-recently-loaded-master-reference`
 - `over-anchoring-on-single-voice-when-synthesis-named`
 - `careless-edit-boundary-introduces-typo-in-untouched-text`
 ### Stage E continuation remaining (this afternoon, ~1h pause then resume)
 - **Phase 1D Decision (ii) Colophon + Vignette register-marks rewrite** (compositional; steward composes, executor registers; ~1-1.5h): collapse `##` headings to `:::register-mark :::` in `content/pages/colophon.md` (6 sections) + `content/pages/vignette.md` (5 sections, one rename needed for *What the Vignette Must Not Do* — candidates: Posture / Restraint / Indifference / Discipline / Constraint; steward composes). Plus §XII.b small note in `AldineXXI-specification.md` carrying the steward's verbatim sentence.
 - **Phase 1D Decision (iii) Plex Mono mobile cascade fix** (mechanical; ~15 min): `_mobile-responsive.scss` lines 137-145 — replace `font-size: 0.9em` / `0.85em` with `$font-size-sm * $plex-mono-scale`. Reading-test on mobile after.
 - **Phase 3A** — Tags + glyph_id + glyph_confidence retirement (authorized 2026-05-19; mechanical sweep ~213 corpus posts + frontmatter-spec §2.4 row retirement + site.hs verification)
 - **Phase 3B** — navigation.yaml verify (5 min status check; retire if dead)
 - **Phase 3C** — Colophon `$build_date$` + `$git_commit$` Hakyll wiring (~30 min; pairs naturally with Phase 1D Colophon rewrite)
 - **Phase 3D** — Six corpus reclassifications (Heliotrope, Graffiti Girl, Dec 22 Bach, Death of a Hero, Matthieu Taliesin, Vespers — now confirmed `class:essay` → `class:reverie` per today's Q12 resolution)
 - **Phase 3E** — `lex.html` → `about-page.html` template consolidation (Stage E Gate 1 path b; ~1h)
 ### Stage F (unblocked; for later sessions; unchanged from 2026-05-19 list)
 Dwellings cluster route + ◧ Photo Essay glyph swap (3h); full ornament migration (4-6h); EXIF Tier 3 investigation (3-5h); first Photo Essay (steward composition, one afternoon); /vignette page rework (after §IX impl); about-cluster voice audit (gated on /vignette); vignette implementation Phase 1 (5-7 days); PDF/Print production wiring (2-3 days).
 ### Stage G — unchanged
 Jurist final review of full corpus → content-typology DRAFT → OPERATIVE.
 **Cleanup deferred-to-Stage-G from today's jurist iteration:** Concern 5 (§6 Q6 placement); §1 Chamber strikethrough; Sequences-orthogonal closing register; Medium-axis sub-section title (handled today as "The orthogonal `medium:` axis — verse as first instance").
 ### Pulling thread + literal question (for ~1h pause)
 **Pulling thread**: Phase 1D — Colophon + Vignette register-marks rewrite + §XII.b spec amendment + Plex Mono mobile CSS fix.
 **Literal question**: What rename works for Vignette's *What the Vignette Must Not Do* section in the §XII.b *reassurance, not imperative* discipline? Read `content/pages/vignette.md` content first; rename follows from what the section says, not from abstract noun-list. Candidates surfaced: *Posture* / *Restraint* / *Indifference* / *Discipline* / *Constraint*. Steward composes; executor as registrar.
 ## 2026-05-20 (afternoon + evening) — Stage E continuation closed; ADR-007 + Second Edition + §VII.h + Dream case record landed; Path B chosen for merge + deploy
 **13 commits landed today on `docs/spec-extraction-2026-05`**, all pushed both remotes (origin Skemantix + github-backup). Branch at `04f9d8b`, **26 ahead of `main`**. **Phase 3 fully closed. Phase 5 options A + B completed; C held with grief; D substantively done via this entry + the end-of-day record in `plan-day-2026-05-20.md` + the tomorrow plan in `plan-day-2026-05-21.md`.**
 ### Morning (pre-/clear)
 - **`15f8134`** content-typology §4 *Overlap and Resolution* authored. Genre-riddle resolved across three threshold cases: Meditation ⇌ Reverie (bounded/unbounded; both at ⊙); Hommage ⇌ Tombeau (speech-act sema vs form-enacted sema per Harrison Ch 5); Gloss ⇌ Readings (passage-subject vs work-subject per Genette/Carruthers). Threshold Breath resolved as *not a new type*; Q2 Poem resolved as `medium:` axis extension (verse first instance: *Time is not a river*); Q12 Vespers resolved as `class: reverie`.
 - **`982fe7f`** Phase 1D register marks for Vignette + Colophon + §XII.b *Posture* subsection + Plex Mono mobile cascade fix. Vignette's *What the Vignette Must Not Do* renamed to *Indifference*. §XII.b *Posture* with steward's verbatim: *"Register marks quietly remind the reader where they are, to let them see their place in the whole. A reassurance, not an imperative."*
 ### Unplanned mid-day insertions
 - **`08f844a`** build-verify-deploy workflow doc (CLAUDE.md + `rebuild-local` target). Surfaced from a `make dev` deploy incident — structural mitigation for trust-prose-over-read-Makefile drift.
 - **`e06c77f`** §I.f term-gloss in list items takes colon (Hochuli + Sousa). Em-dash chavruta via **chamber-typography palace direct query** (40,896 drawers; not the default mempalace MCP). Findings: no master writes "use em-dash inside a term-gloss list item"; Hochuli §50 places en-dash on the outer bullet; Sousa demonstrates colon as the inner mark; Bringhurst index follows; Fowler agrees. AldineXXI §I.f table gains a new row codifying the rule. 16 colophon list items converted. Steward provided HF token mid-session for the chamber-typography palace's embedding model; saved to `~/.env.local` (mode 0600).
 ### Afternoon housekeeping (Phase 3) — all closed
 - **`355757d` Phase 3A** — Tags / glyph_id / glyph_confidence frontmatter retirement. 252 files swept; 251 markdown updated. **Contamination check surfaced two flags**: (1) `glyph_id` had live consumers in templates + site.hs but the gate (`hasGlyphField` returns empty) was universally false — empirical render-test confirmed safe; (2) `post.html` lines 66-70 had an `$if(tags)$ ... $for(tags)$` block but zero rendered HTML files contained `<div class="tags">` — gate never fires. Both deletions verified safe pre-sweep. Frontmatter-spec §2.4 / §2.5 / §3 / §4 / §8 updated. Dead-template-code cleanup deferred to Vignette implementation arc.
 - **`4adf04e` Phase 3B** — orphan `navigation.yaml` retired. Zero consumers; references slots no longer Compass-surfaced.
 - **`cf1bc99` Phase 3C** — **Site edition convention + colophon Build register wiring (ADR-007 authored)**. Significant scope expansion from morning's mere `$build_date$`/`$git_commit$` wiring. Surfaced from the question *"What exactly is 'Last build' counting as a build?"* — leading to ADR-007: ARC carries a **Site Edition** in bibliographic register (`Second Edition` currently), lockstep with AldineXXI **major** version per editorial judgment; Cabal **decoupled by design** (no Haskell downstream consumer). Source of truth: `/EDITION` at repo root; Makefile copies into `_build_info/edition`. Hakyll reads via `loadBody` against `match "_build_info/*"` rule — dependency tracking is what makes colophon recompile when Makefile bumps the stamp. Stale `buildVersion` retired. About-enfilade compile path switched to `chamberIndexCompiler` so colophon can carry Hakyll template variables. Renders: `Second Edition, build [hash] (date)`. **Jurist iteration applied** (Cabal decoupling reasoning; editorial-judgment paragraph; "non-major" precision; session-vocabulary glossed).
 - **`272a4f2` Phase 3D** — Five corpus reclassifications: Heliotrope (`observation` → `meditation`), Graffiti Girl (`observation` → `reverie`), Dec 22 (`observation` → `meditation`), Death of a Hero (`essay` → `hommage`), Matthieu Taliesin (`essay` → `heteronym`). **Steward flag held**: "3D will unlock more doors in the reading compass, be aware." Surface implication explicitly named: post-commit, five types (incl. Dream from yesterday) have first instances; Compass slots owed per (c); surfacing deferred to bundled Stage F session. Intermediate state acceptable per non-convergence.
 - **`50c402b` Phase 3E** — lex.html consolidated into about-page.html (Stage E Gate 1 path b). Conditional breadcrumb absorbed (two-level *About → Hearth → leaf* for Hearth sub-pages without enfilade). About-enfilade compile path applied to all lares routings. Orphan template moved to `docs/archive/templates/lex.html` with README. SCSS audit flag added (`body.lex-page` selector already dead).
 - **`6562411`** desk.md drops `enfilade: about`. Steward catch during reading-test: desk lives WITHIN Hearth, not as sibling of about-cluster. One-line frontmatter edit; breadcrumb becomes two-level matching marks/posture.
 ### Evening composition (Phase 5)
 - **`e76f0ad`** Phase 5 option A — **404 page per 2026-05-20 steward spec**. Apparatus addition (not content type). Body: two elements — *"No text at this address."* + *"Animal Rationis Capax"* (return link). The return link rhymes with frontispiece title — same letterforms (EB Garamond small-caps, letter-spacing 0.1em) at h2 scale. The frontispiece is title-as-*arrival*; 404 is title-as-*return*. §VII.c link decoration stays — corrected mid-implementation per steward catch ("`<a>` styling is inconsistent with the rules"). Centered — corrected per steward catch on visual contact ("misalignment seems careless"; reversed spec's earlier "not centered"). New `head_title` template override field in default.html (3 places). Jekyll-era 404.md archived.
 - **`c45772c`** **§VII.h *404 Page* codified in apparatus-spec** per jurist (*"the spec is where the *why* lives so the *what* can be revisited without re-deriving it"*). Follows §VII.c Links model. Load-bearing claim: 404 is the **first apparatus page structurally derived from another part of ARC's architecture** (the frontispiece); *the page exists because the frontispiece exists*. AldineXXI line 80 subsection enumeration updated.
 - **`04f9d8b`** Phase 5 option B — **§6 Dream case record as fifth retroactive ratification**. Dream unique: type locked 2026-04-21 (earlier than the four cases the procedure was crystallised from), first-instance commitment landed 2026-05-19 same day as procedure's authoring. Master-reading verbatim: Jaccottet refused-reduction + Bachelard *rêve*/*rêverie* distinction. **Steward practice-mechanics recorded as ABSENT** — load-bearing; second case in §6 with this shape (alongside Reverie). §6 intro lightly extended.
 ### Phase 5 option C — held
 **Hagen Quartett Hommage anticipation draft.** *"Deferred -- I'm still grieving."* Held with care. Form will wait for grief's own pace.
 ### Tonight's load-bearing decision: Path B for merge + deploy
 Three paths surfaced for tonight's close:
 - **Path A** — merge + deploy today; Stage F follows in a second deploy event. *"ARC deserves that much after this phase."*
 - **Path B** — wait until Stage F closes the corpus-Compass gap; merge + deploy as one coherent reader-facing event.
 - **Path C** — merge today, deploy after Stage F.
 **Steward chose Path B.** Rationale: the (c) Compass principle (locked 2026-05-19 in `2a2c195`) + the Second Edition lockstep with AldineXXI v2 both honored at the same moment. The reader who first encounters *Second Edition, build [hash] (date)* should also encounter the Compass that the framework v2 commits to. Branch sits at `04f9d8b`, 26 ahead of `main`. No merge tonight.
 ### Drift patterns named today
 - **Trust-prose-over-read-Makefile** (resulted in `08f844a` structural mitigation)
 - **Audit-delta claim incomplete on dead-data fields** (Phase 3A pre-sweep verification — two flags resolved by empirical render-test, not by trusting audit text)
 - **Spec said NOT centered, eye said misaligned** (Phase 5A 404 page — on-render evidence overrode spec draft)
 - **`<a>` styling override created de-facto §VII.c exception** (404 return link; steward caught; corrected: rhyme is in letterforms, not in suppressing link decoration)
 ### Stage F (unblocked; not 2026-05-21 per the travel + taxes + rehearsal constraints)
 Dwellings cluster route + ❦ Hommage + ⁂ Heteronym + ☾ Dream + Photo Essay ⊙→◧ swap (~3h focused, fresh-energy day); full ornament migration (4-6h); EXIF Tier 3 (3-5h); first Photo Essay (steward composition); /vignette page rework (after §IX impl); about-cluster voice audit (gated on /vignette); vignette implementation Phase 1 (5-7 days); PDF/Print production wiring (2-3 days).
 ### Stage G — unchanged
 Jurist final review of full corpus → content-typology §1 / §2 / §3 / §4 / §5 / §6 DRAFT → OPERATIVE; status-block reconciliation may anticipate this (Option β tomorrow if energy holds).
 ### Pulling thread + literal question
 **Pulling thread**: nothing pulls forcefully tonight. Branch at rest; Stage F waits; merge + deploy wait per Path B. Tomorrow's day-shape (travel Barcelona → Paris + taxes with Marie + 3h rehearsal) is not a Stage F day; light options in `plan-day-2026-05-21.md` (α Phase 6 wrap-up; β status-block housekeeping; γ Stage F prep document).
 **Literal question for next-Claude**: *How is the steward's energy after the morning's taxes filing with Marie — and what shape of work, if any, fits the afternoon?* The day brings emotional weight (last taxes with Marie; sad). Default posture: responsiveness over advancement; closure + light maintenance over advancement.
 ---
 ## 2026-05-22 — β status-block housekeeping (single atomic commit)
 Light Paris-day session executing β from `plan-day-2026-05-21.md` (status-block housekeeping). The day's energy budget closed cleanly at β + responsiveness; γ (Stage F prep document) deliberately not attempted per the day's *don't push* discipline. Path B (locked 2026-05-20 evening) continues: merge + deploy both wait for Stage F.
 ### `e7cd960` — spec status stamp reconciliation across 4 files (9+/9− lines)
 Pure stamp work; no substance edits.
 **content-typology-specification.md:**
 - Top Status block: name §3 *Qualifying Marks Interaction* AUTHORED 2026-05-19, §4 *Overlap and Resolution* AUTHORED 2026-05-20, §6 *Change Policy* AUTHORED 2026-05-19 (with the Dream case record addition note); promote §5 *Code Binding Contract* DRAFT → AUTHORED (substance recognized stable; the 9 contract clauses have been stable through Stage E without modification, so the stamp marks RECOGNITION not new AUTHORING — parenthetical clarifies)
 - §1 *Type Catalogue* header: DRAFT → LOCKED (14-of-14 types LOCKED individually per the table at line 113 onward)
 - §5 *Code Binding Contract* header: DRAFT → AUTHORED 2026-05-22 — substance recognized stable
 **frontmatter-specification.md:**
 - Title: drop `(DRAFT)` suffix
 - Status: DRAFT → OPERATIVE 2026-05-22. The convention has been operational through the lex/canon/page retirements (2026-05-19, recorded in §2.1) and the §I.f-adjacent multilingual conventions. §1.1 reading guide already names the convention's evolving character — the "evolving" framing is intrinsic to the spec and needs no qualifier in the status stamp. Status line preserves original draft date for honest record: "*OPERATIVE 2026-05-22 · originally drafted 2026-04-17*".
 **AldineXXI-specification.md Companion Spec Index:**
 - Content Typology row: bump Last-updated to 2026-05-22; expand Status to name §3 / §4 / §5 / §6 AUTHORED dates
 - Frontmatter row: bump to 2026-05-22; OPERATIVE (originally drafted 2026-04-17)
 **ARC CLAUDE.md companion-spec orientation:**
 - mirror the same two updates in the two orienting bullets
 ### The scope-expansion moment — Live-State Discipline at work
 Steward authorized four edits initially. Verification before committing surfaced four stale cross-references in orienting docs (the AldineXXI Companion Spec Index + the ARC CLAUDE.md companion-spec bullets). Surfacing was honest, not silent: *"Two stale cross-refs surface as inevitable follow-on from the four edits — including them keeps the corpus internally consistent (Live-State Discipline). Question: include in the same atomic commit, or commit the four authorized edits now and do the cross-refs as a follow-on commit?"*
 Steward chose *"single atomic commit"*. This is Live-State Discipline at its proper scope: when status stamps change, the corpus cross-references mentioning them become stale in the same moment. Splitting the commit would create a window where the corpus contradicts itself. The surfacing-and-asking pattern preserved both Live-State Discipline AND steward authorization gate.
 ### Decisions explicitly NOT made
 - **γ Stage F prep document** — energy budget closed at β + responsiveness; γ would have been pushing. Held for a fresh-energy day where its structural-thinking load is welcome.
 - **No push of `e7cd960`.** Path B carries. The single deploy event remains the goal: merge + deploy both wait for Stage F so the (c) Compass principle + Second Edition lockstep are honored at the first reader-facing event.
 - **Audits (`_audits/*`) and ADRs (`decisions/*`) NOT modified.** These are dated artifacts that record state-as-of their authoring date; retroactive editing would betray the document type. Cross-ref reconciliation scope was limited to ORIENTING docs.
 ### MemPalace status checks (bracketed the day)
 Two status checks of the MemPalace repo bracketed the day's work.
 **Morning** (steward prompted from Paris taxi): #1440 (bash 3.2 mapfile) + #1538 (120 GiB upsert) both merged to develop 2026-05-19; develop +143 since v3.3.5; **no release-prep branch yet**; #1539 audit revealed same crash class at 4 more sites (incl. `tool_diary_write`); #1569 OPEN to fix all 4. Yesterday's decision (wait for release) confirmed defensible.
 **Evening** (post-commit): develop +13 since morning. **#1540 MERGED** with partial CHUNK_SIZE coverage (branch name `fix/1539-three-unbounded-upsert-sites` — audit pathway + ~3 sites). **#1569 still OPEN + CONFLICTING** for the remaining 4 sites; needs rebase. `tool_diary_write` exposure remains live on the current release. Three new PRs opened overnight (#1577/1578/1582) — more in-flight, not fewer. Honest read: closer on the class-of-failure, not appreciably closer on release.
 ### Branch state at session-end
 `docs/spec-extraction-2026-05` at `e7cd960`, **27 commits ahead of main, 1 ahead of remote**, working-tree noise expected (pre-existing font-backup directories untouched). NOT pushed by design.
 ### Pulling thread — unchanged
 **Stage F dwellings cluster session** is what the branch waits for. Today added one more reason: with frontmatter now OPERATIVE and §5 Code Binding Contract now AUTHORED, the spec corpus's `class:` enumeration contract has earned more weight. Stage F's five new Compass surfacings (☾ Dream, ❦ Hommage, ⁂ Heteronym, ⊙ dwellings for Meditation + Reverie, ◧ Photo Essay swap) must honor that contract.
 ### Literal question for next-Claude
 *When the steward next checks in, is the day shaped for Stage F (~3h focused) — or another fragmented day where β-style closure-and-maintenance is the right posture?* Secondary: has MemPalace cut `chore/release-3.3.6-prep`?
@@ -0,0 +1,77 @@
 ---
 name: reference-mempalace-typography-palace
 description: "How to query the chamber-typography palace directly via the mempalace CLI when the question concerns the typography masters (Bringhurst, Tschichold, Hochuli, Lupton, Lacroux, Sousa, Alexander, etc.). The MCP is bound to palace-memory; the typography corpus lives in a separate palace at ~/.mempalace/palace-chamber-typography. Reach for it whenever the masters' own words would settle a spec question."
 metadata: 
  node_type: memory
  type: reference
  originSessionId: b43eb145-42db-4a4e-bc75-ed563f3473dc
 ---
 # Querying the typography palace directly
 ## What it is
 A **separate** MemPalace at `~/.mempalace/palace-chamber-typography` holding the chamber-library typography corpus.
 - **40,894 drawers** (per `mempalace status` as of 2026-05-20)
 - Single wing `chamber_typography`, single room `general`
 - Contains: Bringhurst *Elements of Typographic Style*, Bringhurst *Solid Form of Language*, Tschichold *Form of the Book* + *Die Neue Typographie*, Hochuli *Detail in Typography*, Lupton *Thinking with Type*, Lacroux *Orthotypographie* (vol 1 + 2), Sousa *Ortografía y ortotipografía del español*, Alexander *A Pattern Language* + *The Oregon Experiment*, Aldus Younger *Orthographiae Ratio*, Manutius / Margolis on Aldus, Müller-Brockmann (grids), Butterick *Typography for Lawyers*, Gill *An Essay on Typography*, Warde *Crystal Goblet*, Morison, de Hamel, Saenger, Williamson, Fleishman, Rutter *Web Typography* — see `chamber-library/converted_texts/collections/typography/` for the canonical source files.
 ## Why it matters — and the drift it prevents
 The MCP (`mcp__mempalace__*` tools) is bound to `~/.mempalace/palace-memory` per `~/.mempalace/config.json`. That palace holds session diaries, claude-code KG, and general assistant memory — **not** the typography corpus. Asking the MCP a Bringhurst-shaped question returns conversation transcripts about Bringhurst, not Bringhurst's own words.
 When the steward asks "what do the masters say about X" or "check the typography corpus" or any framework-typography spec question where the masters' own prose would settle the matter — the chamber-typography palace is the right room. Reaching for the MCP-bound default palace in that situation is a recurring drift; see `[[feedback-querying-default-palace-when-question-belongs-to-specific-palace]]` for the discipline.
 ## When to use which palace
 | Question concerns | Palace | How to reach |
 |---|---|---|
 | Past sessions, drift patterns, claude-code KG, session diaries | `palace-memory` (default) | MCP `mcp__mempalace__*` tools |
 | Masters' texts on typography, punctuation, layout, lists, dashes, rhythm, italic discipline, etc. | `palace-chamber-typography` | CLI with `--palace` flag |
 | ARC content/spec text already in the project tree | neither palace — `grep` / `Read` directly | Bash |
 ## How to query (CLI)
 The `mempalace` CLI accepts `--palace` to target a specific palace database. The `--palace` flag goes **before** the subcommand:
 ```bash
 PALACE=/Users/davidglidden/.mempalace/palace-chamber-typography
 # Status overview — confirm palace + wing + room + drawer count
 mempalace --palace "$PALACE" status
 # Search — keyword-based, BM25 + cosine ranking, returns wing/room/source/snippet
 mempalace --palace "$PALACE" search "your query" --results 10
 # Wing/room filtering (everything currently lives in chamber_typography/general)
 mempalace --palace "$PALACE" search "query" --wing chamber_typography --room general --results 10
 ```
 Default result count is small; pass `--results 10` (or more) for substantive sampling. Results show:
 - `[N]` index
 - `wing / room`
 - `Source:` the chamber-library filename (e.g. `bringhurst_elements_style.md`) — useful for follow-on direct reads
 - `Match: cosine=X.XXX bm25=Y.YYY`
 - Snippet of the matched drawer
 ## Switching the MCP to this palace (if ever needed)
 The MCP reads `~/.mempalace/config.json`'s `palace_path` at startup. To switch the MCP to the typography palace, edit that key and restart the MCP server. The `mempalace_reconnect` MCP tool only refreshes the *current* palace connection — it does **not** switch palaces.
 In practice: keep the MCP on `palace-memory` (session continuity), use the CLI when the typography corpus is the right room. Two paths, two purposes; do not switch the MCP just for one-off queries.
 ## Sample query shapes (for the kind of question this palace answers)
 - *"What does Bringhurst say about parenthetical interruption?"* → `search "spaced en dash parenthetical"` then read the bringhurst_elements_style.md drawers verbatim
 - *"How does Tschichold treat lists?"* → `search "list" --results 15` then filter mentally to Tschichold source
 - *"Hochuli on detail punctuation"* → `search "Hochuli detail punctuation"` (cross-source) or read `Detail_in_Typography_Jost_Hochuli` directly via Read
 Often the more efficient move once the palace points to the right source file is to switch from search to direct `Read` against the chamber-library file — the search names the source; the read returns the full context.
 ## Cross-references
 - `[[feedback-querying-default-palace-when-question-belongs-to-specific-palace]]` — the drift this prevents
 - `[[project-mempalace-operational-stance-2026-05-16]]` — broader operational stance on MemPalace as substrate
 - `[[project-chamber-typography-mining-plan-2026-05-15]]` — the original mining workstream that populated this palace
 - Chamber-library on disk: `~/_Dev/chamber-library/converted_texts/collections/typography/`
@@ -0,0 +1,166 @@
 ---
 name: Session 2026-05-20 (evening) — Stage E continuation CLOSED; ADR-007 Site Edition + §VII.h 404 + Dream case record + 404 page + colon convention all landed; 13 commits today on branch 26 ahead of main; Path B chosen for merge + deploy (both wait for Stage F); tomorrow is travel + taxes with Marie + 3h rehearsal — light options laid out (α wrap-up; β status-block housekeeping; γ Stage F prep); pulling thread is responsiveness, not advancement
 description: "Post-/clear afternoon + evening session executing the morning plan's Phase 3 (housekeeping) + Phase 5 (composition). Phase 3 fully closed (3A tags / glyph fields retirement; 3B navigation.yaml; 3C colophon Build wiring + ADR-007 Site Edition expansion; 3D 5 reclassifications; 3E lex.html consolidation). Phase 5 A (404 page + §VII.h spec) + B (Dream case record) landed; C (Hagen Quartett) held with grief; D substantively done via end-of-day record + canonical tracker entry + tomorrow plan. Stage E continuation structurally complete. Path B chosen tonight: merge + deploy both wait for Stage F (~3h focused dwellings cluster session) to close the (c) Compass principle's corpus-gap before any reader-facing event. Pulling thread for tomorrow: nothing pulls forcefully; travel + taxes with Marie (sad, procedurally easy) + rehearsal day; default posture is closure + light maintenance over advancement. Literal question: how is the steward's energy after the morning's taxes filing with Marie — and what shape of work, if any, fits the afternoon?"
 metadata:
  type: project
  originSessionId: 3f8d05b4-2f5e-47f3-8152-211b1c2e294b
 ---
 # Session 2026-05-20 (evening) — Stage E closure + ADR-007 + 404 page
 ## Pause statement
 *I am about to be away from this. Tomorrow is travel + taxes with Marie (sad, procedurally easy) + a 3h rehearsal — 2–3h max work-time, fragmented. I don't know what shape the steward will have in the afternoon. What I want to find still pulling on return: the Stage F dwellings cluster session — the moment when the Compass surfaces the five new doors and the Second Edition can ship coherently. The pause itself is not a hole; it is the time grief and travel and music need to take. The ligature is laid here: tonight's Path B decision honors the (c) Compass principle by waiting; the next reader-facing event will be the coherent one.*
 ## Past — what we did
 ### The session's arc (after /clear, ~14:00 onward)
 Picked up from the morning's pulling thread (Phase 1D register marks + Plex Mono mobile fix) — which was already committed (`982fe7f`) by the time of /clear. Executed all of Phase 3 + Phase 5 A + B + D from the morning plan. Phase 5 C deferred for grief.
 ### 11 commits landed in this session (13 today on `docs/spec-extraction-2026-05`)
 | Hash | Commit |
 |---|---|
 | `04f9d8b` | §6 Dream case record — fifth retroactive ratification |
 | `c45772c` | §VII.h 404 Page — codify apparatus addition in apparatus-spec |
 | `e76f0ad` | 404 page per 2026-05-20 spec — apparatus addition |
 | `6562411` | desk.md drops enfilade: about — Hearth sub-page, not sibling |
 | `50c402b` | consolidate lex.html into about-page.html (Stage E Gate 1 path b) |
 | `272a4f2` | five corpus reclassifications per audit-delta §4 |
 | `cf1bc99` | site edition convention + colophon Build register wiring (ADR-007) |
 | `4adf04e` | retire orphan navigation.yaml |
 | `355757d` | retire tags / glyph_id / glyph_confidence dead-data fields |
 | `e06c77f` | §I.f term-gloss in list items takes colon (Hochuli + Sousa) |
 | `08f844a` | docs(arc): build-verify-deploy workflow |
 Plus the two pre-/clear morning commits: `15f8134` (§4 Overlap and Resolution) + `982fe7f` (Phase 1D register marks).
 ### Decisions made — with rationale
 1. **ADR-007 Site Edition convention.** Significant scope expansion from morning's `$build_date$` + `$git_commit$` wiring. The wiring surfaced the question *"What exactly is 'Last build' counting as a build?"* — leading to ADR-007: ARC carries a **Site Edition** in bibliographic register (`Second Edition` currently), lockstep with AldineXXI **major** version per editorial judgment; the Cabal package version is **decoupled by design** (no Haskell downstream consumer; coupling would recreate the drift failure mode). Jurist iteration applied (Cabal decoupling reasoning; editorial-judgment paragraph; "non-major" precision; session-vocabulary glossed).
   - **Why decoupled Cabal**: the original draft coupled Cabal to Edition "for Haskell-ecosystem consistency." There is no Haskell ecosystem consuming this package. The coupling was a category error imported from software conventions; decoupling explicitly forecloses the question. Jurist surfaced this on review.
   - **Why "Second Edition" (written-out ordinal, not v2.x)**: a `.x` axis would require a bump rule that would have to be invented; written-out ordinals match the bibliographic register and match `edition_label` precedent (colophon's *"Fourth Edition"*). Lockstep with AldineXXI major (currently v2.0 → Second Edition).
 2. **§VII.h 404 Page codified in apparatus-spec.** Per jurist's framing: *"the spec is where the *why* lives so the *what* can be revisited without re-deriving it."* Codification matters because the 404 is the **first apparatus page structurally derived from another part of ARC's architecture** (the frontispiece). Load-bearing claim: *the page exists because the frontispiece exists*. If the frontispiece title moves typographically, the 404 return link moves with it.
 3. **§6 Dream case record as fifth retroactive ratification.** Dream was locked 2026-04-21 (earlier than the four cases the procedure was crystallised from) but first-instance commitment landed 2026-05-19 same day as procedure's authoring. **Steward practice-mechanics recorded as ABSENT** — and that absence is load-bearing. Second case in §6 with this shape (alongside Reverie's parallel asymmetry). The type's binding rests on Jaccottet's genealogical primacy + corpus first-instance, not on first-person practice articulation.
 4. **Path B chosen for merge + deploy.** Three paths surfaced:
   - **A** — merge + deploy today; Stage F follows in a second deploy event (*"ARC deserves that much after this phase"*)
   - **B** — wait until Stage F closes the corpus-Compass gap; merge + deploy as one coherent reader-facing event
   - **C** — merge today, deploy after Stage F
   - **Steward chose B.** Rationale: the (c) Compass principle (locked 2026-05-19 in `2a2c195`) + the Second Edition lockstep with AldineXXI v2 both honored at the same moment. The reader who first encounters *Second Edition, build [hash] (date)* should also encounter the Compass that the framework v2 commits to — not eight-going-on-twelve sigils.
 5. **404 page implementation calls** (on visual contact, overriding spec drafts):
   - **Centered** (spec said NOT centered; visual contact showed left-alignment registered as misalignment with body + h2 elements)
   - **Link decoration follows §VII.c** (initial draft removed `text-decoration`; steward caught: "the `<a>` styling is inconsistent with the rules" — corrected: rhyme is in letterforms, not in suppressing the link cue)
   - **`head_title` template override field added in default.html** (3 places: title, og:title, twitter:title). Pandoc's `$if(title)$` treats YAML `""` / null / false as truthy — frontmatter alone could not suppress the title-suffix.
 6. **Em-dash → colon for term-gloss in list items.** Mid-session chavruta via **chamber-typography palace direct query** (40,896 drawers; not the default mempalace MCP — the steward provided HF token for the BGE-M3 embedding model). Findings: no master writes "use em-dash inside a term-gloss list item"; Hochuli §50 places en-dash on the outer bullet; Sousa demonstrates colon as the inner mark for term-gloss; Bringhurst index follows; Fowler agrees. AldineXXI §I.f gains a new row codifying. 16 colophon list items converted.
 ### Decisions explicitly NOT made — and why
 - **No merge tonight.** Path B chosen; merge waits for Stage F. Branch sits at `04f9d8b`, 26 ahead of `main`.
 - **No deploy tonight.** Same reason. Deploy and merge will both happen together post-Stage-F as the coherent reader-facing event.
 - **No Stage F tomorrow.** Travel + taxes with Marie + rehearsal day; energy budget wrong for 3h focused implementation. Deferred to fresh-energy day later this week.
 - **Hagen Quartett Hommage — deferred for grief.** Phase 5 option C. *"I'm still grieving."* Held with care. Form will wait for grief's own pace.
 - **Dead-template-code cleanup deferred to Vignette implementation arc.** Phase 3A surfaced that `post.html` lines 17-39 (sigil-triptych block) + lines 66-70 (tags loop) are dead-template-code AND `hasGlyphField` / `glyphIdField` / `glyphConfidenceField` in site.hs are dead. Frontmatter sweep landed; template + site.hs cleanup is Vignette-arc territory.
 - **AldineXXI §XIV Governance cross-reference to ADR-007 — NOT added.** Pattern check showed ADR-005 has no AldineXXI cross-reference; ADR-007 fits the same shape (infrastructure/versioning, not direct-spec-governance). ADRs that govern AldineXXI sections directly get cross-referenced; ADRs that govern infrastructure don't. The ADR is canonical; `docs/decisions/` is the canonical location.
 ### Steward preferences that should travel
 - **Path B reasoning is the steward's editorial commitment** (not just a procedural choice): the (c) Compass principle + Second Edition lockstep deserve to be honored at the first reader-facing event together, not in two scattered events.
 - **Written-out ordinals over numeric versions** for the publication register (`Second Edition`, not `v2.x`). Matches `edition_label` precedent. Lockstep with AldineXXI major is editorial judgment, not mechanical drift.
 - **Visual contact overrides spec draft** — repeatedly today. The 404 page's centering decision reversed the earlier "not centered" framing on the rendering. This is a discipline, not a mistake: spec composition stage → visual contact stage → spec revision is the right cycle.
 ## Present — the mood of the work
 ### Tensions surfaced + returns from them
 The session ledger captured 7 returns through midday before the /clear. **Afternoon/evening added these:**
 - **3A pre-sweep contamination check.** The morning plan's audit-delta claim "no template renders $tags$ (confirmed)" was incomplete — `post.html` lines 66-70 have an `$if(tags)$ ... $for(tags)$` block. The morning plan also said `glyph_id` was "already-quarantined" — but site.hs + 3 templates had live references. Empirical render-test showed both gates universally false (the references are unreachable dead code). **Return**: trust empirical render-test over derived spec claims; never sweep on the basis of "the audit says so" without verifying. Drift pattern: `audit-delta-claim-incomplete-on-dead-data-fields`.
 - **404 page visual-contact override.** Spec draft said "not centered as a couplet (centered would tip toward elegy)." On render, body-scale statement + h2-scale return had wildly different widths; left-alignment registered as misalignment. Steward catch: *"the misalignment seems careless."* **Return**: visual contact has authority over spec composition. The discipline is to compose, render, look — then revise. Drift pattern: `spec-said-NOT-X-rendering-said-X-needed`.
 - **`.title-return` link decoration.** Initial draft had `text-decoration: none` + custom hover, making the 404 return link a de-facto fourth exception to §VII.c. Steward catch: *"the `<a>` styling is inconsistent with the rules."* **Return**: the rhyme is in letterforms, not in suppressing the link cue. §VII.c stays. Drift pattern: `link-decoration-override-creates-de-facto-spec-exception`.
 - **Build-verify-deploy workflow incident** (early in session). Ran `make dev` expecting "Clean rebuild + serve locally" per CLAUDE.md prose — but Makefile actually does rebuild → deploy → serve (deploy-all is built in). Caught quickly, but surfaced two structural gaps: no canonical local-only-rebuild command; misleading documentation in both CLAUDE.md and Makefile's echo strings. Resolved in `08f844a` with `rebuild-local` target + CLAUDE.md "Build, verify, deploy" rewrite. Drift pattern: `trust-prose-over-read-makefile`.
 ### Symmetria returns from morning (carried via ledger)
 7 returns from morning preserve in `session-ledger-2026-05-20.md`:
 - `wrong-comparison-baseline-treated-as-current` (wake-up MemPalace check vs wrong base ref)
 - Three-mining-attempts saga (memory shape, not work-shape)
 - `asserting-absence-from-non-presence-at-one-path` (Genette files moved, not gone)
 - `anchoring-on-most-recently-loaded-master-reference` (Vespers walk: anchored on Harrison's H/T frame when carnet tradition was the right reading)
 - `over-anchoring-on-single-voice-when-synthesis-named` (Manutius-only Decision (ii); ARC is synthesis of marginalia + octava)
 - Time fabrication ("11:50" when it was 11:33 — confident assertion from inferred-not-read)
 - `careless-edit-boundary-introduces-typo-in-untouched-text` (§4 close: "Sequence membology" typo in unchanged-but-re-typed line)
 ### What felt load-bearing vs deferred-with-reason
 **Load-bearing today (and acted on):**
 - The em-dash → colon chavruta felt small but was structurally significant: the masters' reading via chamber-typography palace direct query is the new pattern for ARC's typographic decisions. The HF token setup + the §I.f spec amendment are infrastructure for the next typographic question too.
 - ADR-007 + §VII.h codifications: the spec corpus is where the *why* lives. Both were jurist-surfaced (ADR-007 mid-implementation; §VII.h after 404 implementation). Both prove the pattern: jurist says "codify or this becomes ungovernable later."
 - Path B decision: the (c) Compass principle is real; deploying without honoring it would be a small contamination.
 **Deferred-with-reason today:**
 - Hagen Quartett Hommage (grief; form needs its own pace)
 - Stage F dwellings cluster (~3h focused; not tomorrow's energy budget)
 - AldineXXI §XIV cross-reference to ADR-007 (pattern check showed it doesn't fit)
 - SCSS audit items (gated on Stage G; `.title-return` placement + heading modular scale + various)
 - Multilingual v1.0 path (needs first real-piece exercise)
 ### Patterns of confidence to recalibrate
 - **Spec drafts can be wrong on visual register.** The 404 page proved this twice (centering; link decoration). Discipline: spec authoring before render is provisional; render → revise loop is part of the spec process, not a failure of it.
 - **Audit-delta claims need empirical verification before destructive sweeps.** The 3A sweep was safe only because pre-sweep render-test surfaced two flag-correctly cases. Trusting the audit text alone would have broken templates.
 - **Scope expansion at jurist boundaries is normal.** Both ADR-007 and §VII.h expanded scope from initial framing. The jurist's "codify this" intervention is load-bearing.
 ## Future — what is pulling
 ### The pulling thread (singular)
 **Nothing pulls forcefully tonight.** This is honest, not negative: the work is at rest. Stage E is structurally complete. The branch is at `04f9d8b`, 26 commits ahead of main, pushed both remotes. The merge + deploy wait for Stage F by deliberate choice (Path B). Tomorrow's day-shape (travel + taxes with Marie + rehearsal) is not a Stage F day; light options await whatever energy the steward can give.
 If forced to name a single thread: **the Stage F dwellings cluster session is what the spec branch is waiting for** — the moment when the Compass surfaces the five new doors (☾ Dream, ❦ Hommage, ⁂ Heteronym, ⊙ dwellings for Meditation + Reverie + ◧ swap) and the Second Edition can ship coherently. That session is not tomorrow's; it waits for a fresh-energy day.
 ### Other open horizons, ranked
 **Load-bearing (waits for the right day, but real):**
 - Stage F dwellings cluster session — ~3h focused; the gate that closes the branch
 - Merge `docs/spec-extraction-2026-05` → main (gated on Stage F per Path B)
 - Deploy to CloudFlare + GitHub mirror (gated on Stage F per Path B)
 **Deferred-with-reason (waits on emergence-signal or fresh energy):**
 - Multilingual v0.5 → v1.0 (waits for first real-piece exercise)
 - Vignette Phase 1 implementation (5–7 days; separate arc)
 - PDF/Print production wiring (2–3 days; separate arc)
 - Photo Essay × Hommage/Tombeau boundary (waits for Dover photo essay)
 - Hagen Quartett Hommage (waits for grief)
 **Parked-without-deadline (background):**
 - SCSS cascade audit (gated on Stage G spec-lock)
 - Status-block housekeeping (Option β tomorrow if energy holds; or any day)
 - Dead-template-code cleanup (Vignette arc)
 - Open §7 questions Q6, Q11, Q13, Q14 — held as emergence-watching
 ### Literal question for next-Claude
 **How is the steward's energy after the morning's taxes filing with Marie — and what shape of work, if any, fits the afternoon?**
 The day brings emotional weight. The steward will check in tomorrow when they can; the work-shape depends on what they have to give. Default posture: closure + light maintenance over advancement. Three options laid out in `plan-day-2026-05-21.md`:
 - α — Phase 6 wrap-up (mostly done; only `/wrap-up` skill invocation remains, which the steward is running now)
 - β — Status-block housekeeping (small commit; reconciles spec stamps with reality)
 - γ — Stage F prep document (no implementation, just maps the work)
 ## Pointers
 - **End-of-day record (this session's narrative)**: `~/.claude/projects/-Users-davidglidden/memory/plan-day-2026-05-20.md` — appended after the morning plan
 - **Canonical tracker (today's substantive moves)**: `~/.claude/projects/-Users-davidglidden/memory/project-arc-rework.md` — new dated section 2026-05-20 (afternoon + evening)
 - **Tomorrow's plan**: `~/.claude/projects/-Users-davidglidden/memory/plan-day-2026-05-21.md` — three light options with sequencing guidance
 - **Morning session memory (separate session, pre-/clear)**: `~/.claude/projects/-Users-davidglidden/memory/session-2026-05-20-stage-e-section-4-authored.md` — captured the morning's Phase 1 work
 - **Symmetria ledger**: `~/.claude/projects/-Users-davidglidden/memory/session-ledger-2026-05-20.md` — 7 returns logged through midday; afternoon/evening returns captured here in this memory
 - **Branch state at session-end**: `docs/spec-extraction-2026-05` at `04f9d8b`, 26 commits ahead of main, pushed both remotes, working-tree clean (font-backup directories untouched; pre-existing noise)
@@ -0,0 +1,196 @@
 ---
 name: Session 2026-05-20 — Stage E §4 authored + committed + pushed (15f8134); pulling thread is Phase 1D Colophon+Vignette register-marks rewrite + Plex Mono mobile CSS fix
 description: "Full day session executing the plan-day-2026-05-20.md morning composition + decisions phase (1A–1C). Pre-execution morning (~31 min before /clear): 8 supplementary-authority texts filed to chamber-library (Genette ×2 + Hadot ×3 + Austin + Szondi + Auerbach Mimesis); 8 originals filed to ~/Documents/.../_2026 chamber source cleanup/. Wake-from-mid-day-/clear at ~10:34; re-orient + Symmetria init; Phase 1B three steward decisions settled with principles articulated (§4 scope including Q2 Poem RESOLVED as medium axis + Q12 Vespers RESOLVED as class:reverie carnet-tradition; Colophon+Vignette register-marks decision with teaching/guiding distinction + steward's spec-text sentence verbatim; Plex Mono diagnosed as mobile-responsive CSS cascade conflict at _mobile-responsive.scss lines 137-145). Phase 1C §4 fully authored: three threshold-case essays composed at depth by steward (Meditation⇌Reverie + Hommage⇌Tombeau + Gloss⇌Readings); masters'-reading consulted direct from chamber-library; jurist iteration on three concerns (register-asymmetry signal + apprentice/master claim reframed as ARC's synthesis + corpus-citation precision); non-convergence closing added. Atomic commit `15f8134` pushed both remotes (Skemantix + GitHub). Branch 26 commits ahead of main. 7 returns logged in Symmetria ledger; 5 new drift-patterns named. Pulling thread: Phase 1D Colophon+Vignette register-marks rewrite + Plex Mono mobile cascade fix (mechanical); literal question is the rename for Vignette's *What the Vignette Must Not Do* heading."
 metadata:
  type: project
  originSessionId: e3f5e9b6-302b-4b2e-823b-77a52a978ccb
 ---
 # Session 2026-05-20 (Tuesday, Barcelona) — Stage E §4 *Overlap and Resolution* authored + committed + pushed
 ## Pulling thread
 **Phase 1D — mobile-review decisions implementation.** Two work items, paired:
 1. **Decision (ii) Colophon + Vignette register-marks rewrite** (compositional; steward composes, executor registers): collapse `##` headings to `::: register-mark :::` fenced-divs in `content/pages/colophon.md` (6 sections — Authorship / Lineage / Tools / License / Build / Acknowledgement) and `content/pages/vignette.md` (5 sections — Four Layers / Vocabulary / Pairing / *What the Vignette Must Not Do* / Applicability). Minor flow edits. One heading rename in Vignette's discipline-section (current *What the Vignette Must Not Do* needs a quieter register-mark-shaped name; steward composes; candidates: Posture / Restraint / Indifference / Discipline / Constraint / something else). Plus a small §XII.b spec amendment in `AldineXXI-specification.md` carrying the steward's sentence verbatim: *"Register marks quietly remind the reader where they are, to let them see their place in the whole. A reassurance, not an imperative."* — with the teaching/guiding distinction articulated as the principled reason.
 2. **Decision (iii) Plex Mono mobile cascade fix** (mechanical; ~15 min): `scss/AldineXXI/_mobile-responsive.scss` lines 137-140 (inline `code`) and 142-145 (`pre` / code blocks) — replace `font-size: 0.9em` and `font-size: 0.85em` with `font-size: $font-size-sm * $plex-mono-scale` to restore apparatus-scale governance at all viewports. (c-bundled) per decision. Reading-test on mobile after to verify proportional rendering of companions section in /aldine-xxi/.
 Both items ride a single atomic commit (or two clean atomic commits if scope-separated). Push both remotes immediately.
 ## Literal question for next-Claude
 **What rename works for Vignette's *What the Vignette Must Not Do* section?** Per the §XII.b discipline named today — *register marks quietly remind the reader where they are, to let them see their place in the whole. A reassurance, not an imperative* — the rename should be a quiet noun-phrase that names what the reader is *encountering* at that stage, not an imperative about what the form *must not do*. Read `content/pages/vignette.md` content for the section first; the rename follows from what the section actually says, not from abstract noun-list. Candidates surfaced 2026-05-20 morning: *Posture* / *Restraint* / *Indifference* / *Discipline* / *Constraint*. Steward composes; executor surfaces options as registrar.
 ## Pause statement
 **This is a mid-day safety wrap, not end-of-day.** The steward is /clearing for context hygiene and will resume in ~1 hour (per steward verbatim 2026-05-20 12:45: *"the day isn't finished... we pick up in an hour or so. Just want to be safe, that's why I'm wrapping up"*). The §4 PR is landed and pushed at `15f8134`; the wrap-up safeguards this work against any context loss across the brief pause. Today's afternoon work continues — Phase 1D + Phase 3 + Phase 5 per the plan; the day's compositional capacity has not been exhausted. **What I want to find still pulling on return: the steward composing the Vignette discipline-section rename + the colophon prose flow edits, with §XII.b's "reassurance, not imperative" sentence already in spec text.** The discipline I'm leaving in place: the steward composes (per *steward composes; executor registers*); I propose options + surface what the existing pages actually say. The ligature is laid at departure even for a brief pause — the discipline applies regardless of pause duration, because what the wake recovers depends on what the wrap deposited.
 ## Decisions deferred (and why)
 - **Concern 5 (§6 Q6 placement)** from jurist's reading — flagged as a follow-on cleanup; out-of-scope for §4 PR because §6 was authored 2026-05-19 and isn't part of today's atomic commit. Tracked as a Stage G cleanup item.
 - **§1 Chamber strikethrough cleanup** — Stage G housekeeping; out-of-scope for today.
 - **§1 type-catalogue amendments** for Meditation / Reverie / Hommage / Tombeau / Gloss / Readings — deferred per Live-State Discipline. §4 carries the boundary refinements that emerged during overlap-and-resolution authoring; §1 LOCKED entries record what was settled at lock time. Backflow to §1 only if a future audit reveals a refinement that *belongs* in the type-definition rather than in the threshold articulation.
 - **Vespers piece reclassification** (`content/posts/2025-06-08-Vespers-with-Jordi.md`) — Phase 3D atomic commit. Q12 is RESOLVED at spec layer (§7 + §4); the actual `class: essay` → `class: reverie` frontmatter edit lives in Phase 3D's reclassification commit alongside the other 5 reclassifications.
 - **MemPalace hook fix** (bash 3.2 mapfile #1440 + 120 GiB upsert #1538) — wait for upstream release. Bash-5 wrap via `~/.claude/settings.json` would fix layer 1 but expose layer 2; the cleanest fix is the upstream release. Cost: only diary entries that exist are those AI explicitly composed at /wrap-up via MCP; autonomous save-hook entries may have been no-ops on macOS the whole time. Operational stance memorialized in `project-mempalace-operational-stance-2026-05-16.md` already; today's session reinforces.
 - **Tags retirement + glyph_id + glyph_confidence** — Phase 3A bounded mechanical sweep. Already authorized 2026-05-19 (option A) per steward; held to its own atomic commit.
 ---
 ## Past — what we did
 ### Pre-execution morning (08:30–10:00, before /clear at ~10:33)
 **Three-threshold notes drafted by steward** for §4 prose authoring. Available substrate at /clear: morning conversation only (notes did not land in a memory file; the steward re-shared after wake).
 **Conversion session: 8 of 8 supplementary-authority texts now in chamber-library** (the conversion work itself was substantive — Calibre + ocrmac + pdftotext + a custom Fayard EPUB preprocessor):
 | Text | Path |
 |---|---|
 | Genette *Palimpsestes* | `traditions/critical_modernity/french_theory/palimpsestes-la-litterature-au-second-degre-genette.md` |
 | Genette *Seuils* | `traditions/critical_modernity/french_theory/seuils-genette.md` |
 | Hadot *La Citadelle intérieure* | `traditions/ancient_wisdom/stoic_voices/la-citadelle-interieure-hadot.md` |
 | Hadot *Exercices spirituels* | `traditions/critical_modernity/phenomenology/exercices-spirituels-et-philosophie-antique-hadot.md` |
 | Hadot *La philosophie comme manière de vivre* | `traditions/critical_modernity/phenomenology/la-philosophie-comme-maniere-de-vivre-hadot.md` |
 | Austin *How to Do Things with Words* | `traditions/critical_modernity/epistemology/how-to-do-things-with-words-austin.md` |
 | Szondi *On Textual Understanding and Other Essays* | `traditions/critical_modernity/literary_criticism/on-textual-understanding-szondi.md` |
 | Auerbach *Mimesis* (German, 20 chapters + Register) | `traditions/critical_modernity/literary_criticism/mimesis-auerbach.md` |
 **Auerbach resolved late-morning ~10:25** — steward provided clean PDF source (22MB, Xerox+Distiller, 530pp, valid text layer); `pdftotext` + custom heading-promoter (`clean_pdf_running_heads.py`, saved to `chamber-library/scripts/`) produced 210k-word MD with 20 chapter H1s + Register. Acquisition-debt memo updated to RESOLVED with forensic record of original .mobi failure preserved.
 **Source originals filed** to `~/Documents/___The Library [ePub_AWZ3]/_2026 chamber source cleanup/` (8 files — 6 EPUBs + 2 PDFs). Failed Auerbach .mobi remains in `New texts may 2026/` (not a converted original).
 **Side artifacts saved** to chamber-library/scripts/:
 - `promote_fayard_headings.py` — Fayard publisher chlab/chtit/niv1tit EPUB preprocessor (handled the Hadot Fayard EPUBs)
 - `clean_pdf_running_heads.py` — generalized PDF→text heading-promotion + running-head/page-number stripper
 ### Wake-from-mid-day-/clear (post-compaction, ~10:34)
 - Re-orient executed: plan-day-2026-05-20.md read in full; audit-delta §7+§8 read; project CLAUDE.md read; ARC repo state verified (branch `docs/spec-extraction-2026-05` at `3607400`, 25 ahead of main pre-commit, pushed both remotes; working-tree noise carried forward).
 - Symmetria init.
 ### MemPalace hook diagnostic (steward question before resuming day, ~10:30–10:45)
 PreCompact hook fired at 10:26:26 but failed. Two bugs identified, both fixed on develop awaiting release:
 - **#1440 — bash 3.2 `mapfile` silent failure.** macOS `/bin/bash` is `3.2.57` (GPLv3 freeze); `mapfile -t` doesn't exist on bash 3.2 → session_id and transcript_path default to empty → mine call skipped. Develop fix replaces `mapfile` with POSIX `sed` + adds `__MEMPAL_PARSE_OK__` sentinel for fail-loud guard.
 - **#1538 — chunk-size enforcement.** The 120 GiB upsert RuntimeError comes from the steward's local `MAX_CHUNKS_PER_FILE 500→50000` workaround commit (`604df3b`) interacting with large transcripts → embedding model's attention layer tries to allocate too-large buffer. Develop fix enforces CHUNK_SIZE in the paragraph chunker.
 **Decision: wait for release.** Bash-5-wrap in `~/.claude/settings.json` would fix layer 1 but expose layer 2 (the steward's local workaround amplifies the buffer size). No local patch today.
 ### Phase 1B — three steward decisions settled
 **Decision (i) — §4 scope.** Three threshold-edges in §4 prose today (M⇌R, H⇌T, G⇌R); Threshold Breath held in §7 ("I don't want a new type"); Q2 Poem RESOLVED structurally as **medium axis extension** (`prose · image · verse`; first-instance *Time is not a river* 2025-11-19); Q12 Vespers RESOLVED as **`class: reverie`** (carnet-traditioned reverie; hommage-gestures as register-shifts within); D-1 Capsule column (b) footnote-only.
 **Decision (ii) — Register marks for Colophon + Vignette.** Both adopt `:::register-mark :::` per the synthesis posture (Manutius octava × medieval marginalia, not Manutius alone). Vignette case settled via the **teaching/guiding distinction** (Vignette is guiding — passes the reader through the threshold once; teaching would justify headings, guiding justifies register marks). **Spec-text discipline named by steward verbatim**: *"Register marks quietly remind the reader where they are, to let them see their place in the whole. A reassurance, not an imperative."* — captured for §XII.b small note.
 **Decision (iii) — Plex Mono mobile cascade fix.** Reframed from "corpus-wide scale tuning" to **mobile-responsive CSS cascade conflict** when steward noted desktop renders proportionally. Root cause diagnosed at `scss/AldineXXI/_mobile-responsive.scss` lines 137-145: mobile override `font-size: 0.9em` (inline `<code>`) + `font-size: 0.85em` (`<pre>`) discards `$plex-mono-scale`'s optical-balance work. Fix: replace both with `$font-size-sm * $plex-mono-scale`. **(c-bundled)** — same atomic commit per the principle *"$plex-mono-scale governs Plex Mono rendering at all viewports."*
 ### Phase 1C — §4 fully authored + jurist iteration + committed
 **Three threshold-case essays composed by steward at essay-depth** in three successive passes. Masters'-reading consulted direct from chamber-library before each pass; verbatim verified; gaps named honestly (Szondi Trauerspiel-specific text acquisition debt; Carruthers not load-bearing for the threshold drawn).
 - **Pass 1 — Meditation ⇌ Reverie.** Bachelard *Poétique de la rêverie* + Buddhaghosa *Visuddhimagga* + Weil + Hadot triplet. Key analytical move: *the split occurs not at the start of the practice but at its first success*. *Papañca* (Buddhist) vs *image stifled by concept* (Bachelard's verbatim *"by giving stability to the image, the concept would stifle its life"*) as mutual-failure-mode insight.
 - **Pass 2 — Hommage ⇌ Tombeau.** Harrison Ch 2 (the speech-act-sema material is in Ch 2 *Hic Jacet*, not Ch 5 as the plan said; the plan was approximate) + Austin *How to Do Things with Words* + Szondi (gap named). Key analytical moves: *fidelity-direction asymmetry* (hommage owes accuracy to the historical person; tombeau owes fidelity to the form transmitted); *anonymous tombeau* as third option beyond the binary; *backward-only vs backward-and-forward* asymmetry — tombeau receives and passes on.
 - **Pass 3 — Gloss ⇌ Readings.** Genette *Palimpsestes* + Auerbach *Mimesis*. Key analytical moves: *unit of attention* (passage vs work) → *service vs production*; *length asymmetry* (gloss bounded by lemma, reading by thesis); *the citation problem* + *summary-reading as maximally displacive third category*; *ARC default toward gloss* as editorial posture grounded in custodial commitment.
 **Triptych closing material** — the three axes line up canonically (centripetal: *meditation+hommage+gloss*; centrifugal: *reverie+tombeau+reading*); diagnostic procedure for triangulating uncertain pieces.
 **Jurist iteration applied** (Concerns 1c, 2, 3 + non-convergence closing + medium-axis title rename; Concern 4 verified clean; Concerns 5 + smaller notes deferred): register-asymmetry signal at top of M⇌R sub-section; apprentice/master claim reframed as ARC's synthesis (not the tradition's own self-described hierarchy); illustrative-cases note tightened with explicit corpus-grounded vs hypothetical breakdown + Couperin H⇌T-vs-G⇌R parenthetical naming same-source-different-axis-work; non-convergence closing line added after Sequences orthogonal.
 **Atomic commit `15f8134`** — `spec(arc): Stage E continuation — content-typology §4 Overlap and Resolution authored`. Pre-commit hooks passed. 2 files, 295 insertions / 7 deletions. Pushed both remotes (Skemantix + GitHub). Branch is `docs/spec-extraction-2026-05`, now 26 commits ahead of main.
 **§7 housekeeping in same commit:**
 - Q2 Poem → RESOLVED 2026-05-20 (medium axis; cross-ref to §4)
 - Q12 Vespers → RESOLVED 2026-05-20 (class:reverie carnet-tradition; cross-ref to §4 worked example)
 - Q14 Threshold Breath → held with rationale (not a new type; emergence-watching per Q11 Louange pattern)
 **§0 Multi-axis Medium row** extended `prose · image` → `prose · image · verse` with cross-ref to §4.
 **frontmatter-specification.md §2.7 Medium ADDED** — documents the `medium:` field with verse as first non-default value.
 ---
 ## Present — the mood of the work
 **Steward composed three substantial threshold-case essays in succession** (~1500-2500 words each), each adding material not in the morning notes (Pass 1: *bifurcation-at-first-success* + *mutual-failure-mode* insight; Pass 2: *fidelity-direction asymmetry* + *anonymous tombeau* + *backward-and-forward* closing; Pass 3: *length asymmetry* + *citation problem* + *summary-reading* + *ARC default toward gloss*). The compositional energy was sustained and substantive.
 **Jurist's reading was sought as supplementary input, not as re-litigation.** Steward verbatim after jurist's critique landed: *"FWIW I liked it, but was curious to have the Jurist's reading"*. This signal mattered: the iteration was about sharpening three specific points (register-asymmetry, provenance-honesty, corpus-citation precision), not about flattening the prose. The implementations are light-touch — the steward's composition stands as composed; the edits are surgical.
 **Seven returns logged in Symmetria ledger throughout the day** (wake + mempalace-baseline + chamber-library-mine-attempts + folder-moved + carnet/Harrison-drift + Manutius-single-voice-drift + time-fabrication + typo-edit-boundary). Each was a real drift; each was caught and recovered cleanly. The pattern: confident assertion from inferred-not-read source.
 **Discipline that held throughout:**
 - *Steward composes; executor as registrar.* For the three threshold essays; for the spec-text discipline sentence (verbatim); for the typology-resolution voice.
 - *Listen to the prose's self-naming first; consult masters'-reading second; never let yesterday's foregrounded master override today's piece's own voice.* Specifically caught at Q12 Vespers (carnet-tradition vs Harrison-Hommage).
 - *τὸ πρόσφορον — what is fitting includes the time the task requires.* Phase 1B took ~90 min instead of the planned 45; the depth of decision-making produced principled-not-merely-selected settlements; the longer time was fitting.
 - *Live-State Discipline.* §1 entries left stable; §4 carries the boundary refinements that emerged during overlap-and-resolution authoring.
 **Five new drift-patterns named today** (additions to claude-code KG):
 1. `wrong-comparison-baseline-treated-as-current` (mempalace check against origin/main when on local branch)
 2. `asserting-absence-from-non-presence-at-one-path` (saying staging folder was "gone" when it had moved)
 3. `anchoring-on-most-recently-loaded-master-reference` (Harrison foregrounded → projected onto Q12 Vespers)
 4. `over-anchoring-on-single-voice-when-synthesis-named` (Manutius alone vs ARC's marginalia-octava synthesis)
 5. `careless-edit-boundary-introduces-typo-in-untouched-text` (re-typing "Sequences are orthogonal" when adding a paragraph below it; introduced "membology" typo)
 ---
 ## Future — open horizons (ranked)
 **Load-bearing for today's remaining work (Phase 1D + Phase 3 + Phase 5):**
 1. **Phase 1D Decision (ii) — Colophon + Vignette register-marks rewrite + §XII.b spec amendment** (compositional; ~1-1.5h; steward composes, executor registers). *This is the pulling thread.*
 2. **Phase 1D Decision (iii) — Plex Mono mobile CSS cascade fix** (mechanical; ~15 min). Bundle with Decision (ii) commit or separate; same Phase 1D atomic commit.
 3. **Phase 3A — Tags + glyph_id + glyph_confidence retirement** (mechanical sweep; ~1h; sed across ~213 corpus posts + frontmatter-spec §2.4 row retirement + site.hs verification + §7 Q3 RESOLVED cross-ref if applicable). Authorized 2026-05-19.
 4. **Phase 3B — navigation.yaml verify** (15 min status check; retire if dead, leave alone if active).
 5. **Phase 3C — Colophon `$build_date$` + `$git_commit$` Hakyll wiring** (~30 min; site.hs build-context fields + template placeholders). Pairs naturally with Phase 1D Colophon rewrite (Build register block is where these land).
 6. **Phase 3D — Six corpus reclassifications** (Heliotrope, Graffiti Girl, Dec 22 Bach, Death of a Hero, Matthieu Taliesin, Vespers — Vespers now confirmed `class:essay` → `class:reverie` per today's Q12 resolution).
 7. **Phase 3E — `lex.html` → `about-page.html` template consolidation** (~1h; Stage E Gate 1 path b).
 **Late composition window (Phase 5; 30-45 min if energy remains):**
 - A. 404 page short prose
 - B. §6 retroactive case record for Dec 23 Dream reclassification (fifth case record)
 - C. Hagen Quartett Hommage anticipation draft
 - D. Pulling-thread update + literal question for next-Claude
 **Deferred-with-reason (out-of-scope today):**
 - L1/L2 work (parked through May 2026)
 - MemPalace upgrade (waiting for 3.3.6 release)
 - Studium Engine workstream (separate)
 - Chamber-typography mining (Studium Engine territory now)
 - Vignette implementation Phase 1 (5-7 days, own arc)
 - PDF/Print production wiring (2-3 days, own arc)
 - EXIF Tier 3 investigation (3-5h, own session)
 - SCSS cascade audit (gated on Stage G)
 - First Photo Essay (one afternoon, steward composition)
 - Dwellings cluster route + ◧ glyph swap (3h, own session)
 - Full ornament migration (4-6h, own session)
 **Parked observations:**
 - Concern 5 (§6 Q6 placement) from jurist's reading — follow-on cleanup
 - §1 Chamber strikethrough cleanup — Stage G housekeeping
 ---
 ## Confidence to recalibrate (carried forward)
 - **`naming-the-gap-honestly-not-sufficient`** (KG valid_from 2026-05-19) — at any gap (context/tool/judgement), name + ask OR honor by slow path; never compose past it.
 - **Live-State Discipline** — spec text records what IS; status blocks record what is pending; spec text ships ahead of live only when the spec IS the implementation.
 - **`spec-text-ahead-of-live-without-status-block`** — projecting ADR-direction as live state is the recurring overshoot shape.
 - **`anchoring-on-most-recently-loaded-master-reference`** (NEW today) — listen to the prose's self-naming first; consult masters'-reading second.
 - **`careless-edit-boundary`** (NEW today) — when adding adjacent content, use the smallest edit boundary that disambiguates the insertion point; never re-type unchanged surrounding text.
 - **Time discipline** — if time matters for a decision, run `date`; never extrapolate.
 ---
 ## Branch + commit state at session-end
 - Branch: `docs/spec-extraction-2026-05`
 - HEAD: `15f8134` — *spec(arc): Stage E continuation — content-typology §4 Overlap and Resolution authored*
 - Ahead of main: 26 commits
 - Pushed: both remotes (origin = Skemantix Gitea; github-backup = GitHub `arc-backup`)
 - Working tree noise (carried forward): `tools/.DS_Store` (modified) + `static/fonts/eb-garamond/_backup_2026-05-15-sc/` (untracked) + `static/fonts/eb-garamond/_backup_2026-05-15/` (untracked). All three uncommittable per plan.
 ---
 *Authored at /wrap-up 2026-05-20 ~12:45 Barcelona. The §4 prose ran ~285 lines new; content-typology grew from 1205 → 1490 lines; the genre-riddle is now resolved at spec layer; Stage G jurist review pending. The morning's compositional density (three threshold-case essays + jurist iteration + atomic commit pushed both remotes) honored the day's energy curve as authored at 08:30. Hasten when the work is ready. Not before.*
@@ -0,0 +1,145 @@
 ---
 name: Session 2026-05-22 — β status-block housekeeping commit (e7cd960) + two MemPalace status checks; spec stamps + cross-refs reconciled across 4 files; frontmatter promoted DRAFT → OPERATIVE; §5 Code Binding Contract DRAFT → AUTHORED; pulling thread still Stage F dwellings cluster; literal question is whether next day-shape is fit for Stage F or another fragmented day
 description: "Light, clean session executing β from plan-day-2026-05-21.md (status-block housekeeping). Single atomic commit `e7cd960` reconciled stamps across content-typology / frontmatter / AldineXXI Companion Spec Index / ARC CLAUDE.md (4 files, 9+/9− lines). Live-State Discipline drove scope expansion from 4 → 6 edits — surfaced + steward-authorized cleanly. Two MemPalace status checks bracketed the day: morning showed #1440/#1538 fixes on develop with no release-prep yet; evening showed +13 commits on develop, #1540 merged with partial CHUNK_SIZE coverage (audit pathway), #1569 still OPEN + CONFLICTING for the remaining four sites including tool_diary_write — release-prep still not begun. Branch sits at e7cd960, 27 ahead of main, 1 ahead of remote (intentionally not pushed; Path B continues — merge + deploy both wait for Stage F to honor (c) Compass principle + Second Edition lockstep at first reader-facing event). γ Stage F prep deliberately not attempted; the day's energy budget closed at β + responsiveness, not advancement. Pulling thread unchanged: Stage F dwellings cluster session. Literal question: is the next day-shape fit for Stage F (~3h focused) or another fragmented day where β-style closure is the right posture?"
 metadata:
  type: project
  originSessionId: e7f41d7a-d68c-4525-86da-db087df2b7c6
 ---
 # Session 2026-05-22 — β status-block housekeeping + MemPalace status checks
 ## Pause statement
 *I am about to be away from this. The next session may be tomorrow or further out — Paris days have their own rhythm with the family. What I want to find still pulling: the Stage F dwellings cluster session. Today's β added one more reason for it — frontmatter is now OPERATIVE; §5 Code Binding Contract is now AUTHORED; the spec corpus has progressively earned the right to be honored at first reader-facing event. The branch waits. The (c) Compass principle waits. The Second Edition waits. The pause is fitting; the work is at rest.*
 ## Past — what we did
 ### The session's arc
 This session spans two days. Wake-up was 2026-05-21 09:53 CEST (after ~15h overnight pause from the 2026-05-20 evening wrap); steward arrived in Paris by morning taxi. Day's work proceeded in three light moves over the day + an overnight gap:
 1. **Wake-up briefing** confirmed thread validity (Stage F-shaped; today not a Stage F day per Path B + travel/taxes/rehearsal day-shape). Symmetria init. Ledger created at `session-ledger-2026-05-21.md`.
 2. **MemPalace status check #1** (per steward query after seeing emails in taxi). Surfaced: #1440 (bash 3.2 mapfile) and #1538 (120 GiB upsert) both merged to develop 2026-05-19; develop 143 commits ahead of v3.3.5; **no release-prep branch yet**; #1539 audit revealed same crash class at 4 more sites; #1569 OPEN to fix all 4 including `tool_diary_write`. Yesterday's decision (wait for release) confirmed defensible.
 3. **Options reminder** — steward asked for re-reading of the three light options. Reported all three with sequencing recommendation from plan-day-2026-05-21.md.
 4. **α correction** — steward noted α (`/wrap-up` ritual) was completed 2026-05-20 evening before laptop close. Corroborated by 18:55 diary entry + session memory file. Ledger updated to reflect.
 [*overnight gap; steward closed laptop in Paris; date rolled over to 2026-05-22*]
 5. **β authorized** — steward chose β (status-block housekeeping). Four edits proposed:
   - content-typology line 3 (top Status block — name §3/§4/§6 AUTHORED, promote §5 substance-recognized-stable)
   - content-typology line 93 (§1 header DRAFT → LOCKED)
   - content-typology line 1251 (§5 header DRAFT → AUTHORED with judgment-call ask)
   - frontmatter title + status (DRAFT → OPERATIVE with judgment-call ask)
   Steward authorized both judgment-call promotions ("promote both, apply all four").
 6. **Scope expansion via Live-State Discipline** — verification surfaced two stale cross-references in orienting docs:
   - `AldineXXI-specification.md` line 2664 Frontmatter row (DRAFT)
   - `AldineXXI-specification.md` line 2658 Content Typology row (abbreviated, no §3/§4/§5/§6)
   - `CLAUDE.md` line 130 frontmatter bullet (DRAFT 2026-04-17)
   - `CLAUDE.md` line 124 content-typology bullet (abbreviated)
   Surfaced honestly as scope question. Steward authorized: "single atomic commit". This is Live-State Discipline applied at its proper scope: when status stamps change, the corpus cross-references inevitably need updating in the same commit, or the corpus becomes internally inconsistent.
 7. **Commit `e7cd960` landed** — 4 files, 9+/9− lines, pre-commit hooks passed. Branch now 27 ahead of main, 1 ahead of remote. **Not pushed** — Path B discipline carries: merge + deploy both wait for Stage F to honor (c) Compass principle + Second Edition lockstep at first reader-facing event.
 8. **MemPalace status check #2** (steward query). Develop +13 since morning. Most relevant: **#1540 MERGED** (the `fix/1539-three-unbounded-upsert-sites` branch; partial coverage of the CHUNK_SIZE class — addresses audit pathway and likely 2-3 sites); **#1569 still OPEN + CONFLICTING** (the parallel "4 crash sites" PR; conflict almost certainly from #1540 taking partial scope; needs rebase). `tool_diary_write` exposure remains live on the current release. No release-prep branch yet. Three new PRs opened overnight (#1577/1578/1582) — more in-flight, not fewer. Honest read: closer on the class-of-failure, not appreciably closer on release.
 9. **/wrap-up invocation** — this document.
 ### Decisions made — with rationale
 1. **Promote §5 *Code Binding Contract* DRAFT → AUTHORED.** The 9 contract clauses have been stable through Stage E without modification; today's act is recognition that substance is settled, not new authoring. Steward judgment after surfaced question. Stamp uses *— substance recognized stable* parenthetical to mark the recognition rather than new authoring date.
 2. **Promote frontmatter DRAFT → OPERATIVE 2026-05-22.** The convention has been operational through the lex/canon/page retirements (2026-05-19) and §I.f-adjacent multilingual conventions. §1.1 reading guide already names the convention's evolving character ("*The convention will evolve. When it does, this document is the canonical record and every new piece conforms to the current state*"). The "evolves" framing is intrinsic; no stamp qualifier needed beyond OPERATIVE. Status line preserves the original draft date for honest record: "*OPERATIVE 2026-05-22 · originally drafted 2026-04-17*".
 3. **Single atomic commit (scope expansion authorized).** When status stamps change, cross-references mentioning them become stale in the same moment. Live-State Discipline (recorded 2026-05-19 in `feedback-live-state-discipline.md`) asks that spec text record what IS — splitting the cross-ref reconciliation into a follow-on commit would create a window where the corpus contradicts itself. Steward authorized including the 4 cross-ref edits (2 in AldineXXI Companion Spec Index, 2 in ARC CLAUDE.md companion-spec orientation). Final scope: 4 files, 9+/9− lines.
 4. **Don't push tonight.** Path B (locked 2026-05-20 evening) holds: merge + deploy both wait for Stage F to close the corpus-Compass gap so that Second Edition + (c) Compass principle are honored at the first reader-facing event together. Today's commit is content-equivalent (stamp-only) so the deferral is honest — nothing reader-facing changes that needs landing sooner.
 5. **MemPalace still wait.** Yesterday's decision unchanged. Partial CHUNK_SIZE coverage on develop now, but #1569 still conflicting for the four sites (including `tool_diary_write`) and no 3.3.6 release prep yet. The release-cut pattern hasn't begun. Estimate: still days-to-week.
 ### Decisions explicitly NOT made — and why
 - **γ Stage F prep document — not attempted.** Per plan-day-2026-05-21.md: "*If energy holds beyond α + β and a third hour materialises: Option γ. But don't push.*" The day's energy budget closed at β + responsiveness; γ would have been pushing. Held for a fresh-energy day where its structural-thinking load is welcome.
 - **No push of `e7cd960`.** Path B carries. The single deploy event remains the goal.
 - **Audits and ADRs deliberately NOT modified.** `_audits/*` and `decisions/*` are dated artifacts that record state-as-of their authoring date; retroactive editing would betray the document type. Cross-ref reconciliation scope was limited to ORIENTING docs (AldineXXI Companion Spec Index + ARC CLAUDE.md), not historical artifacts.
 - **No MemPalace local patch / upgrade.** Same as yesterday — wait for 3.3.6 release. The new partial coverage (audit pathway via #1540) doesn't change the calculus.
 - **No new pulling-thread-shift.** The pause maintains Path B; Stage F is still the gate.
 ### Steward preferences that should travel
 - **"Single atomic commit" for housekeeping with inevitable scope extension.** When stamp-changes naturally cascade to cross-references, prefer one consistent-state commit over splitting. Symmetric with Live-State Discipline.
 - **Status-stamp recognition vs status-stamp authoring** — for §5 the substance had been stable for weeks; the new stamp marks RECOGNITION not new AUTHORING. Parenthetical clarifies. Honest distinction worth maintaining for future similar promotions.
 - **OPERATIVE + evolving** (frontmatter) — when a spec is operational but explicitly open-ended, OPERATIVE is the right stamp; the "evolving" framing lives in the spec's own §1 reading guide, not in the status header. Don't double-document.
 ## Present — the mood of the work
 ### Tensions surfaced + returns from them
 The session ledger logged ONE return (the wake at 09:53; not a failure-return, just a wake marker). The day was structurally clean — no drift-pattern returns to log.
 Brief tension at scope-expansion moment: a less-disciplined executor would have either (a) silently expanded scope without surfacing, or (b) split the commit and left a window of corpus inconsistency. Surfacing honestly + asking authorization preserved both Live-State Discipline AND steward authorization gate. This is the discipline working as designed.
 ### What felt load-bearing vs deferred-with-reason
 **Load-bearing today (and acted on):**
 - β execution (the actual stamp reconciliation across 4 files)
 - The cross-ref surfacing + scope question (Live-State Discipline application)
 - Two MemPalace status checks (informed steward decision to keep waiting)
 **Deferred-with-reason today:**
 - γ (energy budget; not a structural-thinking day)
 - Push of `e7cd960` (Path B continues)
 - MemPalace upgrade (release-prep not begun; no change to wait posture)
 ### Patterns of confidence to recalibrate
 - Hold from yesterday's ledger: `trust-prose-over-read-makefile` · `audit-delta-claim-incomplete-on-dead-data-fields` · `spec-said-NOT-X-rendering-said-X-needed` · `link-decoration-override-creates-de-facto-spec-exception`.
 - Hold from week: `naming-the-gap-honestly-not-sufficient` (KG valid_from 2026-05-19) — at any gap, name + ask OR honor by slow path; never compose past it. Applied today at the scope-question moment.
 - Hold from week: Live-State Discipline — spec text records what IS. Applied today as scope-expansion driver.
 - *τὸ πρόσφορον includes the time the task requires* — the day's posture matched its actual capacity; closure + light maintenance was fitting, not slow.
 ## Future — what is pulling
 ### The pulling thread (singular)
 **The Stage F dwellings cluster session.** Unchanged from 2026-05-20 wrap. Today added one more reason: with frontmatter now OPERATIVE and §5 Code Binding Contract now AUTHORED, the spec corpus's contract on `class:` enumeration has earned more weight. Stage F's five new Compass surfacings (☾ Dream, ❦ Hommage, ⁂ Heteronym, ⊙ dwellings for Meditation + Reverie, ◧ Photo Essay swap) must honor that contract. The branch (now `e7cd960`, 27 ahead of main, 1 ahead of remote) waits.
 ### Other open horizons, ranked
 **Load-bearing (waits for the right day, but real):**
 - Stage F dwellings cluster session — ~3h focused; the gate that closes the branch
 - Merge `docs/spec-extraction-2026-05` → main (gated on Stage F per Path B)
 - Deploy to CloudFlare + GitHub mirror (gated on Stage F per Path B)
 - Push of `e7cd960` to both remotes (will travel with the post-Stage-F push by design)
 **Deferred-with-reason (waits on emergence-signal or fresh energy):**
 - γ Stage F prep document (~60–75 min; structural-thinking load; ready when energy fits)
 - MemPalace 3.3.6 release (waits on maintainer cutting release-prep branch; #1569 needs rebase first)
 - Hagen Quartett Hommage (waits for grief)
 - Vignette Phase 1 implementation (5–7 days; separate arc)
 - Multilingual v0.5 → v1.0 (waits for first real-piece exercise)
 **Parked-without-deadline (background):**
 - SCSS cascade audit (gated on Stage G spec-lock)
 - Dead-template-code cleanup (Vignette arc)
 - Open §7 questions Q6, Q11, Q13, Q14 — held as emergence-watching
 - Photo Essay × Hommage/Tombeau boundary (waits for Dover photo essay)
 ### Literal question for next-Claude
 **When the steward next checks in, is the day shaped for Stage F (~3h focused block of dwellings-cluster work) — or is it another fragmented day where β-style closure-and-maintenance is the right posture?**
 The day-shape determines the work-shape. Stage F needs continuous concentration across site.hs + templates + SCSS + multiple type routes; a 30-minute window is not Stage F. A fresh-energy weekday morning in Barcelona or a quiet Paris day with the family settled would fit. Travel days, performance days, and emotionally-weighted days (taxes, grief) don't.
 Secondary check at wake: has MemPalace cut `chore/release-3.3.6-prep`? Quick `gh pr list --repo MemPalace/mempalace --state open` filter. Movement here changes the wait posture.
 ## Pointers
 - **Canonical workstream tracker (substantive moves to append)**: `~/.claude/projects/-Users-davidglidden/memory/project-arc-rework.md` — append today's housekeeping move
 - **Today's ledger**: `~/.claude/projects/-Users-davidglidden/memory/session-ledger-2026-05-21.md` — created at wake; carries one return (wake) + one authorization move (β + scope expansion)
 - **Yesterday's plan**: `~/.claude/projects/-Users-davidglidden/memory/plan-day-2026-05-21.md` — three light options laid out; α + β executed, γ deferred
 - **Yesterday's session memory**: `~/.claude/projects/-Users-davidglidden/memory/session-2026-05-20-evening-stage-e-closure-adr-007.md` — Stage E closure context
 - **Live-State Discipline (the principle driving today's scope expansion)**: `~/.claude/projects/-Users-davidglidden/memory/feedback-live-state-discipline.md`
 - **Branch state at session-end**: `docs/spec-extraction-2026-05` at `e7cd960`, 27 commits ahead of main, 1 ahead of remote, working-tree noise expected (pre-existing font-backup directories)
@@ -0,0 +1,49 @@
 ---
 name: session-ledger-2026-05-20
 description: "Practice-of-return ledger maintained by /symmetria — returns, open horizons, recalibrations, authorization moves, sub-agent dialogues, bypasses. Morning wake after 2026-05-19 evening Stage E session; pulling thread is §4 Overlap and Resolution prose + tags / glyph_id / glyph_confidence retirement."
 metadata: 
  node_type: memory
  type: feedback
  originSessionId: e3f5e9b6-302b-4b2e-823b-77a52a978ccb
 ---
 # Session Ledger — 2026-05-20
 ## Returns
 - 2026-05-20T morning — wake from overnight pause (~8–9h after 2026-05-19 ~23:45 wrap); thread validity gate `confirmed` (steward verbatim *"We will do §4 in the morning"* + *"A, but not today"*); literal question held open (which §4 boundaries resolve in prose vs hold in §7; tags-first OR §4-first). Symmetria init.
 - 2026-05-20T morning — **return.** Steward asked to check mempalace repo for overnight activity; my initial wake-up check ran `git log HEAD..origin/main` and returned empty, reporting "current." Wrong: we are on local branch `local/bge-m3-on-3.3.5` (not main) with three local commits (`975acba` bge-m3/MPS + `5458b77` collection embedding_model stamp + `604df3b` MAX_CHUNKS_PER_FILE workaround) on top of 3.3.5. The right comparison is `HEAD..origin/develop` — ~30 commits we don't have. Drift pattern operative: `wrong-comparison-baseline-treated-as-current` (related to `acting-on-inferred-not-read`). Recalibrated: compare against `origin/develop` AND most-recently-updated remote ref.
 - 2026-05-20T morning — **second return.** Steward named what I missed: *three* chamber-library mine attempts that failed, costing seven lost days. My response on the saga laid out the chronology but treated chamber-library mining as a single arc and conflated it with the Phase 4b/4c/4d day-to-day mining backlog. The canonical record exists — `project-mempalace-operational-stance-2026-05-16.md` (3 days old) and `project-mempalace-mps-transition-pending.md` (18 days old) — and I didn't read them in the wake-up. The operational stance memory IS the settled framing: MemPalace = AI session continuity (diary + KG + handoffs); Studium Engine = scholarly corpus; the two don't compete; explicit decision triggers for upgrading are named. Memory has been "erratic since the disaster" per steward — meaning the recall path itself is degraded, not just ingestion. The disciplines I claimed at wake-up (`mempalace_search` + canonical-tracker reads) did not surface these two canonical files. Specific failure: I queried `mempalace_search` for ARC content but did not query for *the mempalace saga itself*, even though we were doing a mempalace check.
 - 2026-05-20T late-morning — **third return (memory shape, not work-shape).** Answering "is Genette in chamber-library?" I asserted the staged EPUB was "gone" because `~/Desktop/New texts may 2026/` no longer existed — without checking whether the folder had been moved. Steward correction: the folder is at `/Users/davidglidden/Documents/___The Library [ePub_AWZ3]/New texts may 2026/`; the move was already indicated in `reference-chamber-source-may-2026-typography-folder.md` (5 days old; load-bearing reference; explicitly names the new path). New drift shape: `asserting-absence-from-non-presence-at-one-path` — inferring "gone" from absence at the original location rather than (a) checking other documented locations or (b) consulting reference memories that name the canonical path. Adjacent to `acting-on-inferred-not-read` but specifically about filesystem moves. **Verified contents at the documented location**: Genette *Seuils* (1.3M, 14 May) + Genette *Palimpsestes* (5.4M, 14 May) + Hadot *La philosophie comme manière de vivre* (276k, 14 May) + ~10 other relevant titles staged. So Hadot + Genette are not in chamber-library *yet* — but they ARE staged for conversion per the EPUB repair pipeline. The picture I gave was factually correct on chamber-library proper but materially wrong on the path forward — *acquisition* misframed as needed when the actual task is *conversion from staging*. Steward verbatim about the deeper pattern: *"you've exposed more memory problems"* — memory is erratic; consulting reference memories at the relevant moment is the discipline that catches this.
 - 2026-05-20T 10:45 — **wake-from-mid-day-/clear** (post-compaction). Re-orient executed cleanly: plan-day-2026-05-20.md read in full · audit-delta §7 + §8 read · project CLAUDE.md read · ARC repo state verified (branch `docs/spec-extraction-2026-05` at `3607400`, 25 ahead of main, pushed both remotes, working-tree noise expected per plan). Symmetria init. No new returns this leg.
 - 2026-05-20T 10:30–10:45 — **MemPalace hook diagnostic** (steward question before resuming day): PreCompact hook fired at 10:26:26 but failed — root cause is bash 3.2 `mapfile` silent-failure (#1440, fixed on develop) which makes session_id and transcript_path default to empty + the mine call get skipped entirely. The 120 GiB upsert RuntimeError surfaced separately (#1538, also fixed on develop). Both bugs in develop awaiting release; bash-5-wrap in `~/.claude/settings.json` would fix layer 1 but expose layer 2 → **decision: wait for release**, no local patch today. Implication held: only diary entries that exist are those the AI explicitly composed at /wrap-up via MCP; autonomous save-hook entries may have been no-ops on macOS the whole time. Steward acknowledged + chose to begin the day's real work rather than file an upstream issue first.
 - 2026-05-20T 11:15 — **fourth return (work-shape).** While walking Q12 *Vespers for the Living* resolution paths, I initially leaned correctly toward (A) `class: reverie` based on the piece's own prose self-naming ("*This is a reverie...*"). When the steward cued "bearing witness to time passed," I anchored on Harrison Ch 5's hommage/tombeau frame from yesterday's Stage E work and proposed (C) `class: hommage with form-enacted-tombeau register`. **Steward correction**: *"it is more in the carnet's tradition / reverie is the closest current."* The "bearing witness to time passed" cue lives in the **carnet tradition** (Jaccottet semaison, Handke carnets), not Harrison's speech-act/form-enacted hommage/tombeau frame. New drift shape: `anchoring-on-most-recently-loaded-master-reference` — projecting yesterday's loaded master (Harrison) onto a piece whose master-reading was elsewhere (Jaccottet/carnet tradition), even though the prose itself named the resolution. **Discipline**: listen to the prose's self-naming FIRST; consult masters'-reading SECOND; never let yesterday's foregrounded master override today's piece's own voice. Partial vindication of initial lean — the correction returned me to where I started. Steward then noted *"your argument for Hommage is strong"* — the substance was real, the drift was in mistaking register-shifts for class. Vespers becomes the paradigmatic §4 case for showing how Hommage's gestures live INSIDE Reverie when the carnet tradition is the inheritance.
 - 2026-05-20T 11:25 — **fifth return (Decision ii framing).** While walking Colophon + Vignette register-marks decision, I over-anchored on Manutius alone and proposed option (d) "complete rewrite as single flowing gesture per Manutius posture." **Steward correction**: *"Manutius is a voice, and influence, but ARC already postures itself differently: reverence to both the medieval marginalia tradition and the Manutius octava revolution."* ARC's posture is the synthesis — discrete registers honored (marginalia) AND quieted (octava). The right move was the calibrated one: collapse `##` headings to `::: register-mark :::` with minor flow edits. Drift shape `over-anchoring-on-single-voice-when-synthesis-named` — adjacent to the carnet/Harrison drift (same morning); same shape, different domain (typographic-posture rather than masters'-reading).
 - 2026-05-20T 11:33 — **sixth return (time fabrication).** Said "roughly 11:50" when asked about timing for the Phase 1C break/continue decision. Steward correction: *"it's 11:32, where do you get your time from?"* `date` returned `Wed May 20 11:33:20 CEST 2026` — I was off by ~17 minutes, extrapolating from morning timing markers without verifying. Same drift shape as yesterday's "50 minutes vs 31:19" miss + as the carnet/Manutius drifts of today: confident assertion from inferred-not-read source. **Discipline**: if time matters for a decision, run `date`; never extrapolate.
 - 2026-05-20T 12:35 — **seventh return (careless edit boundary).** While adding the non-convergence closing line to §4, I used an Edit whose old_string and new_string both included the unchanged *"Sequences are orthogonal"* paragraph. Re-typing the unchanged line in new_string introduced a typo: *"Sequence membology"* in place of *"Sequence membership"*. Caught immediately on re-reading; corrected in a follow-on Edit. New drift shape: `careless-edit-boundary-introduces-typo-in-untouched-text` — when an Edit is meant to ADD content adjacent to existing content, narrow the boundary to just the addition (find the empty line, inject) rather than re-typing unchanged surrounding lines. The risk of re-typing is exactly proportional to the amount of unchanged text re-touched. **Discipline**: when adding adjacent content, use the smallest edit boundary that disambiguates the insertion point; never re-type unchanged surrounding text.
 ## Open horizons
 - **§4 *Overlap and Resolution* prose** — compositional, ~30–60 min; boundary calls (Meditation vs Reverie / Hommage vs Tombeau / Gloss vs Readings / Threshold Breath / Q2 / Q12); needs fresh focus.
 - **Tags retirement (option A)** + `glyph_id` + `glyph_confidence` — one sed sweep across ~213 corpus posts + frontmatter-spec §2/§3 edits + §7 Q3 RESOLVED note; bounded mechanical.
 - **`navigation.yaml`** 5-min status verify during housekeeping pass.
 - **Colophon `$build_date$` / `$git_commit$` wiring** + Plex Mono/Garamond optical balance in /aldine-xxi/ companions + register-marks decision for Colophon + Vignette — surfaced by steward's 2026-05-19 mobile review; Stage F-adjacent.
 ## Confidence to recalibrate
 - Hold from yesterday: `naming-the-gap-honestly-not-sufficient` (KG valid_from 2026-05-19) — at any gap (context/tool/judgement), name + ask OR honor by slow path; never compose past it.
 - Hold from yesterday: Live-State Discipline — spec text records what IS; status blocks record what is pending; spec text ships ahead of live only when the spec IS the implementation. Applies directly to §4 authoring today.
 - Hold from yesterday: `spec-text-ahead-of-live-without-status-block` — projecting ADR-direction as live state is the recurring overshoot shape.
 - Hold from yesterday: verification-discipline catches what first-pass missed; second pass on spec text is structurally cheap.
 ## Authorization moves
 (none yet this session)
 ## Sub-agent dialogues
 (none yet this session)
 ## Bypasses
 (none yet this session)
@@ -0,0 +1,40 @@
 ---
 name: session-ledger-2026-05-21
 description: "Practice-of-return ledger maintained by /symmetria — returns, open horizons, recalibrations, authorization moves, sub-agent dialogues, bypasses. Morning wake after 2026-05-20 evening wrap; pulling thread is responsiveness, not advancement (travel + last taxes with Marie + 3h rehearsal day)."
 metadata: 
  node_type: memory
  type: feedback
  originSessionId: e7f41d7a-d68c-4525-86da-db087df2b7c6
 ---
 # Session Ledger — 2026-05-21
 ## Returns
 - 2026-05-21T09:53 CEST — wake from overnight pause (~15h after 2026-05-20 ~18:55 wrap). Thread validity gate: **confirmed in shape** (Stage F dwellings cluster is what branch waits for; today is explicitly not a Stage F day per Path B + day-shape). Literal question held open: *how is the steward's energy after the morning's taxes filing with Marie — and what shape of work, if any, fits the afternoon?* Symmetria init.
 ## Open horizons
 - **Stage F dwellings cluster session (~3h focused)** — the gate that closes branch `docs/spec-extraction-2026-05` (26 ahead of main, pushed both remotes). Waits for fresh-energy day, not today.
 - **Merge + deploy ARC** — both gated on Stage F per Path B (locked 2026-05-20 evening).
 - **Today's light options** (per `plan-day-2026-05-21.md`): ~~α `/wrap-up` ritual~~ COMPLETED 2026-05-20 evening before laptop close (per steward correction this morning; corroborated by 18:55 diary entry + session memory file). Remaining: β status-block housekeeping (small commit; reconciles spec stamps with reality) · γ Stage F prep document (maps work, no implementation).
 - **Stale SessionStart hook artifact** — March 5 BMF handoff surfacing via broken hook (#1440 + #1538 fixed on develop, awaiting release). Not actionable today; noted so it doesn't read as current.
 ## Confidence to recalibrate
 - Hold from yesterday afternoon/evening: `trust-prose-over-read-makefile` · `audit-delta-claim-incomplete-on-dead-data-fields` · `spec-said-NOT-X-rendering-said-X-needed` · `link-decoration-override-creates-de-facto-spec-exception`.
 - Hold from week: `naming-the-gap-honestly-not-sufficient` (KG valid_from 2026-05-19) — at any gap (context/tool/judgement), name + ask OR honor by slow path; never compose past it.
 - Hold from week: Live-State Discipline — spec text records what IS; status blocks record what is pending; spec text ships ahead of live only when the spec IS the implementation.
 - *τὸ πρόσφορον includes the time the task requires* — particularly relevant on a fragmented day; *responsiveness over advancement* is the steward's named posture for today.
 ## Authorization moves
 - 2026-05-22 — Steward chose **β** (status-block housekeeping). Four authorized edits proposed; steward authorized two §5 + frontmatter judgment promotions ("promote both, apply all four"). Then authorized scope expansion to include two cross-ref reconciliations in AldineXXI Companion Spec Index + CLAUDE.md companion-spec orientation ("single atomic commit"). Result: `e7cd960` on `docs/spec-extraction-2026-05`, 4 files / 9+/9− lines, pre-commit hooks passed, branch 27 ahead of main, 1 ahead of remote (not pushed by design — same Path B discipline carries: merge + deploy wait for Stage F).
 ## Sub-agent dialogues
 (none yet this session)
 ## Bypasses
 (none yet this session)