Run-report — `1882-1886-what-then-must-we-do`

Work-dive on Tolstoy’s What Then Must We Do? (Так что же нам делать?), the fourth Prophet-period panel and the economic/social application of the What I Believe ethic. Run 2026-06-06, in-session unattended (the A-Confession-pilot model), medium visual intensity. Followed the handoff from the What I Believe session.

Scope contract (Phase 0)

• Subject (work dive): What Then Must We Do? (Так что же нам делать? / What to Do? / What Is to Be Done?). Slug 1882-1886-what-then-must-we-do — year-prefixed, slots after 1882-1884-what-i-believe.
• Corpus surface: PSS Tom 25 — work text pp. 182–411 (v25_182_411_Tak_chto_zhe_nam_delat.xml); first printed variant of ch. XVII pp. 503–507; variants pp. 614–652; commentary pp. 740–839. Forerunner «О переписи в Москве» pp. 173–181. Composition-window diaries Tom 49 (1882–87); letters Tom 63 (1880–86) + Tom 64 (1887–89). Composition window 1882–1886 (begun early 1882, written mostly 1884–86).
• Works record: missing — the dive is the source-grounded basis to create it.
• Mode: in-session unattended; no AskUserQuestion gating; judgment calls → needsReview; note draft: true; medium visual intensity.

What was covered

• Deep read of all 40 chapters + the «О переписи в Москве» appeal; keystone passages extracted byte-faithful (the failed-charity narrative & «I sit on a man’s back»; money-as-slavery + the Fiji parable; the science / division-of-labour critique = «захват чужого труда»; the three answers & bread-labour / Bondarev; the contested final chapter on women).
• Genesis triply grounded: the Jan-1882 census + «О переписи»; the 1 Jan 1883 property/violence diary entry and the 1884 estate-reform plan; the Dec-1884 Alekseev bridge and the 1885–86 Bondarev letters. Composition chronology 1882→1886 reconstructed from the apparatus (begun early 1882, not 1884; stop-start; intensive Dec 1884; finishing through 1885–86).
• Publication/suppression (distinct from What I Believe’s clean ban): serialization refused at Русская мысль (Jan 1885); cut fragments in Русское богатство (1885); cut full text in the 5th-ed. vol. 12 (10 Apr 1886) after Pobedonostsev’s Synod refusal + Ivantsov-Platonov’s notes; fuller text abroad (Geneva/Elpidine «Какова моя жизнь?» 1886); complete text 1902 (Chertkov, Christchurch). English: Hapgood 1887 → Wiener 1904 → Maude 1925.
• The family conflict over the book (apparatus): S. A. Tolstaya objected to the named-family episodes, not the argument; Tolstoy conceded the personal forms and held the thought, while she prepared the edition and petitioned the censor — complicating the Sofia-centred frame.
• Scholarship triangulation (7 entries) + the newest reading (Nicolosi 2024, slum-literature).
• Visuals: 8 PD images cached (keystone = Repin’s 1887 Tolstoy Ploughing).

Coverage ledger

Entity work-order (ingestion priority → dependency)

Priority 1 (central; write first): Bread-labour (хлебный труд) [concept]; Money as a form of violence [concept]; 1882 Moscow census [event]; T. M. Bondarev [person]; V. I. Alekseev [person]. Leo Tolstoy, V. G. Chertkov, S. A. Tolstaya all EXIST. Priority 2: Critique of the division of labour [concept]; V. K. Sutaev; Henry George; N. N. Strakhov; K. P. Pobedonostsev; P. I. Biryukov; Lyapinsky house [place]; Rzhanov house [place] (all MISSING). Yasnaya Polyana EXISTS. Priority 3: G. A. Rusanov; K. D. Kavelin; S. A. Yuryev; L. E. Obolensky; A. M. Ivantsov-Platonov; E. M. Feoktistov; M. K. Elpidine; Isabel F. Hapgood (all MISSING).

Visuals work-order

Cached (PD, usable): Repin 1887 Tolstoy Ploughing (keystone); Ge 1884 portrait; LoC ca.1880 photo; Makovsky 1889 Ночлежники (night-shelter); Chertkov 1883; Repin 1887 seated portrait; Chertkov & Tolstoy 1909; Hapgood 1887 What to Do? cover (EN first ed.). To acquire (request from holdings): portraits of V. I. Alekseev (State Tolstoy Museum / Biryukov plates), T. M. Bondarev (Siberian archives; the 1890 English Labour: The Divine Command may carry a frontispiece), V. K. Sutaev (Tretyakov / a Repin study); a Lyapinsky-house photograph (MAMM; Moskovsky Listok illustrated supplements ca. 1882).

Work-record work-order (proposed fills for a NEW record)

High confidence: titleRu «Так что же нам делать?»; titleEn “What Then Must We Do?” (alts “What to Do?”, “What Is to Be Done?”); bans[] (3 schema-clean events: imperial-state/serialization-refused, imperial-state/passages-cut, holy-synod/pre-publication-rejected, Pobedonostsev refusal 1885-12-28 NS / 16 Dec OS); censoredVersionExists true; samizdatCirculation true; publishedDuringLifetime / publishedInRussiaDuringLifetime = true; jubileeEdition.volumes 25. Medium/low: dateWritingStarted ~1882 (approximate); dateWritingCompleted ~1886 (approximate); genre essay; dateFirstPublished (which event — see needsReview); dateFirstPublishedInRussia ~1885; relatedWorks [what-i-believe sequel, confession sequel]; excommunicationRelated false (see needsReview). Folder (treatises) → human decision.

needsReview (deferred human judgment)

1. excommunicationRelated value (proposed false — 1901 edict names no works).
2. dateFirstPublished — which event the field carries (refused Russkaya Mysl serialization 1885 / cut fragments Russkoe Bogatstvo 1885 / cut full text 5th-ed. vol. 12 1886 / fuller text Geneva 1886 / complete text Christchurch 1902). No clean “first publication.”
3. New record folder (treatises) + filename.
4. Cross-dive diary ambiguity: the Tom 25 commentary places a «не могу продолжать свою статью. Тонкости, и потому будет ложь» stall-entry at 28 Apr 1884 for THIS book, but the What I Believe dive cited an identically-worded 28 Apr 1883 entry for THAT book; the corpus holds only the 1883 file. Not cited in this dive’s evidence; resolve the year before either dive attributes it.
5. The family-conflict primary quotes (S. A. Tolstaya’s note + Tolstoy’s late-Jan-1885 apology) are cited via the PSS Tom 25 apparatus, not from the primary letter (Tom 83 — not independently extracted).

Verification

• verify_quotes.py → 21/21 verbatim PASS (two superscript-marker fixes: wtmwd-let-03 trimmed to the clean «упомянул…» punchline; wtmwd-let-05 uses an honest […] author-elision around the ⁶+a).
• Phase-5 verifier (opus, fresh context) → PASS-WITH-NITS, 0 must-fix. One voice NIT applied: the evaluative “reactionary” was used in the dive’s own voice in three spine/note spots — softened to factual descriptors, leaving the attributed characterization (Cruise/Lounsbery) in §Scholarly context. NIT 2 (relatedWorks both typed sequel) left as-is (closest controlled value; flagged for the human ingestor). Report: _verifier-report.md.

Evaluation self-assessment (work-dive gate)

• Interlocutor sweep yielded people? Yes — the letter network (Alekseev, Bondarev, Chertkov, Strakhov, Kavelin, Yuryev, Obolensky, Rusanov) + the censorship figures, all carried into entities.
• Russian society/Church reception covered? Partial (honestly graded) — censorship verdicts + letter-network reception grounded; wider press reception and the Synod tie deferred to the queue.
• workRecord fill accurate & provenanced? Yes — every field evidence/commentary-anchored; schema-clean controlled vocab (verifier-confirmed); uncertain fields → needsReview.
• coverage honest? Yes — Redactions & Reception marked partial, later-verdict not-covered; nothing overclaimed (verifier-confirmed).
• --choice=reg extracted cleanly? Yes — pre-reform resolved; diary/letter bodies needed --notes=auto. The diary witness for 1884–86 is genuinely thin (logged as a finding, not a gap).
• Spine stayed bare? Yes after the NIT fix — the contested women chapter and the family conflict presented factually; critical labels attributed, not adopted; neither softened nor whitewashed.

Models & rough cost

Main loop: Opus (in-session). Subagents: 3× Sonnet sweeps (diaries / letters / visuals) + 1× Sonnet scholarship + 1× Opus verifier. Subagent tokens ≈ 436k total (diaries 71k, letters 105k, visuals 58k, scholarship 69k, verifier 134k). Web sweeps: visuals (Commons/IA) + scholarship; no licence-gated downloads to website/src/.

Output paths

• docs/research/1882-1886-what-then-must-we-do/index.md (+ rendered index.html, git-ignored)
• docs/research/1882-1886-what-then-must-we-do/dossier.yaml
• docs/research/1882-1886-what-then-must-we-do/extracts/ (PD primary extracts — work, census appeal, ch.-XVII variant, cited diaries & letters; the swept-but-uncited diary/letter extracts kept as provenance)
• docs/research/1882-1886-what-then-must-we-do/visuals/ (git-ignored cache; 8 PD images)
• website/src/posts/notes/2026-06-06-what-then-must-we-do.md (draft)
• session-log.md, _verifier-report.md, this run-report.md

Wiki ingestion is a separate, human-in-the-loop step — this dossier is the pointer, not the writer.