SOURCE_MANIFEST — Napoleon Hill

Hill died 1970. Under life+70 (EU / most rule-of-life countries) his work is
NOT public domain until 2041. So the choice of work is a copyright decision,
not a convenience — it is fixed here before any extraction.

Work used: The Law of Success in Sixteen Lessons (1928)

work file edition source confidence
The Law of Success in Sixteen Lessons law_of_success_1928.txt The Ralston University Press, Meriden, Conn., 1928 (© 1928 Napoleon Hill) Internet Archive item 1928HillLawOfSuccess, OCR full text high (OCR; normalized)

law_of_success_1928.txt is the normalized corpus (see "OCR normalization"
below) and is the single evidence pool. los_extracted.txt is the raw OCR page
text kept for provenance; do not quote from it directly.

PD basis — why this work and not "Think and Grow Rich"

We use The Law of Success (1928), NOT Think and Grow Rich (1937).

  • The Law of Success was published in 1928. In the United States a 1928 work
    lost copyright by term expiration on 1 January 2024 (95-year term; US
    Copyright Office Circular 15A — a work published in year N enters the public
    domain on 1 Jan of year N+96, so 1928 → 2024). This is the strongest possible
    PD basis: nothing hinges on a renewal record, a registration, or a contested
    fact. Clean US public domain.
  • Think and Grow Rich (1937) is widely treated as US public domain, but its
    status rests entirely on the claim that the 1909-Act 28-year copyright was
    not renewed in its 28th year (c. 1965) and so lapsed into PD in 1966. That
    non-renewal is now well documented (a Thomson CompuMark search is commonly
    cited), but it is a non-renewal argument, not term expiration, and it is still
    described as contested by some. The Napoleon Hill Foundation also holds a
    trademark on the title. Under life+70 abroad it is protected until 2041
    regardless.
  • Per the build brief: prefer the safe path if in doubt. Term-expiration PD on
    a pre-1931 work is unambiguous; a non-renewal argument on a 1937 work is not.
    Therefore the KB is built from The Law of Success (1928) only. Think and Grow
    Rich
    is treated as out of corpus (see CORPUS_BOUNDARY.md), and its
    signature terms are handled in QUOTE_AUTHENTICITY.md.

Sources for the PD reasoning:

  • Public Domain Day 2024 (Duke Center for the Study of the Public Domain) — US works
    published in 1928 entered PD on 1 Jan 2024 (95-year term). US Copyright Office
    Circular 15A confirms the 95-year term for pre-1978 published works.
  • Internet Archive 1928HillLawOfSuccess — the 1928 first edition full text.
  • The "not renewed → PD 1966" claim for Think and Grow Rich (zerolimits.org
    summary of the Thomson CompuMark non-renewal report) — noted as the weaker,
    non-term basis we deliberately do not rely on.

OCR normalization (reproducible)

The 1928 scan is OCR, not a clean Gutenberg etext. Two deterministic transforms
produce law_of_success_1928.txt from the raw page text:

  1. Join soft-hyphen line breaks: remove ¬ and any following whitespace
    (Philos¬ ophy → Philosophy).
  2. Collapse every whitespace run to a single space.

The self-check (scripts/verify_quotes.py) applies the same whitespace
normalization to every evidence blockquote before substring-matching, so quotes
are matched against the normalized corpus exactly. OCR error note: page-number
tokens like -33- and residual scanner slips (e.g. bum/burn) survive in the
corpus; evidence quotes are chosen to avoid those spans.

RU layer

No clean PD Russian translation of The Law of Success is used. Every evidence
blockquote is verbatim English from law_of_success_1928.txt; every Russian
rendering in the notes is our own translation, marked «перевод наш». English
is the primary evidence pool.