SOURCE_MANIFEST — Goethe KB v1

Proof work: Maxims and Reflections (Maximen und Reflexionen — the prose
maxims, published partly in Wilhelm Meisters Wanderjahre 1829 and in the
posthumous Nachgelassene Werke 1833; Saunders' selection of 590 maxims).

Layer Edition Where fetched PD status
EN (primary evidence pool) The Maxims and Reflections of Goethe, tr. T. Bailey Saunders, Macmillan 1906 Project Gutenberg #33670, fetched 2026-07-07 Goethe d. 1832 → PD worldwide; Saunders d. 1928 → EU-PD since 1999; text pre-1930 → US-PD
DE (term arbiter only) Maximen und Reflexionen (original) consulted for term sense, not ingested Goethe d. 1832 → PD worldwide
RU (reading layer) перевод наш — our translation of each cited passage, EN Saunders base with the German term sense as arbiter produced in this build our text; released with the KB

Files

  • sources/maxims_en_units.json — canonical English evidence pool, 619
    units: mx.1–mx.590 (590 numbered maxims) + nat.1–nat.29 (the closing
    "NATURE: APHORISMS" fragment, disputed authorship — see below). Every
    verbatim EN: line in this vault must be a literal substring of this file.
  • sources/maxims_sections.json — unit → section map (LC.I–LC.VII = Life
    and Character subsections I–VII, LA = Literature and Art 403–514, SCI =
    Science 515–590, NAT = Nature fragment).
  • sources/build_corpus.py — deterministic parser from the Gutenberg .txt to
    the units JSON.
  • sources/check_quotes.py — the verbatim gate.
  • sources/gb_33670.txt — raw Gutenberg source.

Unit keys

  • mx.<n> — maxim n in Saunders' continuous numbering (1–590).
  • nat.<n> — paragraph n of the "Nature: Aphorisms" fragment.

Normalization applied (documented, deterministic)

  • Gutenberg header/footer (*** START/END ***) stripped; translator's preface
    and index not ingested.
  • Paragraphs joined to single lines; runs of whitespace collapsed.
  • Source misprint fixed: the Gutenberg text numbers maxim 265 as a second
    "263" (right after 264). The parser deterministically renumbers a duplicate
    non-sequential number to the expected sequential one, restoring mx.265
    («politeness of the heart»). No wording changed.
  • No other changes; the English is otherwise byte-verbatim to Gutenberg.
  • Gate normalization additionally folds --/dashes to spaces, strips quote
    characters and _-italics markers before substring-matching.

The nat.* units — disputed authorship

The "Nature: Aphorisms" fragment (Die Natur, Tiefurter Journal 1782/83) is
of disputed authorship: modern scholarship attributes the wording to
Georg Christoph Tobler, writing down Goethe's conversation. Goethe himself
(1828, letter to Chancellor von Müller) did not remember composing it but
acknowledged the ideas as his own at that stage. Saunders prints it as an
appendix; we ingest it as nat.* so quotes from it can be identified, but
concept/principle pages do not use nat.* as primary evidence for Goethe's
voice — any use must be marked disputed_authorship.

The RU translation policy

No public-domain Russian translation aligned to Saunders' selection exists.
The Russian in evidence blocks is our translation of the specific cited
passage, made from the Saunders English with the German original consulted as
term arbiter (Tätigkeit, Entsagung, Bildung, die Forderung des Tages, das
offenbare Geheimnis …), and labelled «перевод наш» on the pages. It is a
reading aid; the English Saunders line is the verbatim evidence and the
only thing the gate checks. Where a term is contested, TERMINOLOGY.md
records the German arbiter form.

Not fetched in v1 (next per recipe)

Faust (Anster/Taylor EN are PD; RU Kholodkovsky PD), Conversations with
Eckermann (Oxenford EN PD), Dichtung und Wahrheit, Wilhelm Meister
(Carlyle EN PD), West-östlicher Divan. Add as separate unit sets; do not mix
into the Maxims evidence pool. Only these would let the outside-corpus
concepts (demonicheskoe, prafenomen) carry their own primary quotes
instead of not_in_ingested_corpus.