On Translating Homer; with F. W. Newman's "Homeric Translation in Theory and Practice" and Arnold's "Last Words"

What it is about

Three Oxford lectures by Arnold, Newman's reply, and Arnold's last lecture, debating how an ancient text should sound in English. Arnold says a translator must feel four qualities of Homer. Homer is rapid in movement, plain and direct in words and syntax, plain and direct in ideas, and noble. He judges the existing versions by them: Cowper is slow and elaborate, Pope artificial, Chapman full of conceits, Newman odd and ignoble. Arnold's judges are scholars who can compare with the Greek. He warns against building a special vocabulary for translation and against anything "odd". He keeps the familiar Castor rather than Kastor, and repeats fires where Homer repeats πυρά, giving up verbal elegance to keep the plainness. Newman had chosen a ballad metre and a deliberately antiquated diction, "a plausible aspect of moderate antiquity, while remaining easily intelligible". He answers that the judge of taste is the educated but unlearned public, and that a critic should set a prose version of the Greek beside each rendering. Arnold grants that Newman's syntax is Homeric but faults his vocabulary.

Ideas

• I1. An author's style as a few measurable qualities, compared with the source's. For each author, compute rapidity (mean clause length in tokens, tokens per event of prag frames) and plainness (dependency depth, share of rare lemmas, function-word density). Compute archaism too (share of forms marked by en/src/emode.rs). Compute the same for the Hebrew source via MACULA trees. An author delta can then be described, and checked, as a move in this small space.
• I2. Two tribunals, two scores, never merged. The scholar's tribunal is agreement with the source: reader and world level in measure.rs. The public's tribunal is how native the output reads: the stack residual and the parser's confidence on the output, against native text of the same register. Report both for every system.
• I3. Newman's "moderate antiquity" as a strength setting of a delta. Rank an archaic author's overrides by count and apply only the top k. A small k gives thou and hath; a large k gives the rare forms Arnold calls odd. Sweep k on the school registers and measure archaism (I1) against the public score (I2). This gives a dial between WEB and KJV for the style deltas (open problem 8).
• I4. Repetition keeping as an author feature. When a concept occurs twice in a source verse, does the author write the same word twice or vary it? Count it per author from the meta-world. Arnold repeats where Homer repeats; elegant translators vary. Hebrew narrative repeats heavily, so this splits KJV-like from paraphrasing authors.
• I5. Newman's three-way comparison (source, prose literal version, verse rendering) is the translator's baseline table: MACULA glosses word by word, a literal Bible and the system. Keep all three in every Hebrew report, as tools/translate already does for the glosses.

For Mova

The debate is the archaic-versus-modern question inside the translator: KJV and JPS against WEB, and the standard that leans to the Bibles (open problem 6). I1 and I3 give a way to describe and tune a style delta; I2 is a reporting rule for measure.rs.

Topics: Machine translation

Cite

The paper itself, not this card. BibTeX from the library's record of the work.

@book{arnold1861translating,
  title = {{On Translating Homer; with F. W. Newman's "Homeric Translation in Theory and Practice" and Arnold's "Last Words"}},
  author = {Arnold, Matthew and Newman, Francis William},
  year = {1861},
  url = {https://www.gutenberg.org/ebooks/65381}
}

This page in Markdown → https://md.mova.elamur.ai/free/en/docs/papers/arnold-1861-translating-homer-f-w/