Letters Concerning Poetical Translations, and Virgil's and Milton's Arts of Verse

What it is about

Four letters, dated from 1736, on translating poets and on verse. The first is the one about translation. A translator must first find the author's particular style, or the work hardly deserves the name. Benson defines two styles by their syntax. A style is rapid when several relative clauses hang on one antecedent, or several verbs on one subject, to the close of the period: Homer's anger of Achilles "which brought … and sent … and gave". It is majestic when the matter is delayed by inversion and parenthesis: Virgil's Arms and the man I sing, with the fugitive of heaven set in. He shows a paraphrase of the Iliad's opening that keeps the sense and loses Homer's speed by adding and expanding (Pluto's gloomy empire, devouring dogs). He also sets side by side two renderings of the Aeneid's opening by the same translator, Pitt: one in Virgil's majestic manner, one in Homer's rapid one. The later letters compare Virgil's and Milton's versification. They argue that Milton's auxiliary verbs, as in did + verb, add strength, and they treat pauses, monosyllables and the absence of rhyme as tools of style.

Ideas

• I1. Style as counts over the UD tree, as Benson defines it. Rapidity is the number of verbs that share one subject in a coordination chain (conj under the same nsubj), and the number of relative clauses on one antecedent. Majesty is the number of fronted dependents before the verb (obj, obl, advmod) and of parenthetical appos and parataxis. Count both per author on the Bibles and the school registers. These are style features that prag and en/src/lin.rs can compute without new annotation.
• I2. Hebrew narrative is the rapid style by Benson's definition: wayyiqtol chains of verbs on one subject. Measure, per English Bible, how often a Hebrew chain of k verbs stays one English chain with and rather than being split into sentences or subordinate clauses. This is a construction-level feature of each delta that the token-level residual hides.
• I3. One translator, two styles: Pitt's two versions are a natural control for the style deltas. Encode the same meta-world with a rapid and a majestic setting (school-talk against school-exam, or WEB against KJV) and check that the I1 counts move in the expected direction. A delta that changes words but not these counts is a lexical delta, not a style delta.
• I4. Periphrastic do as an author feature. KJV and Milton use did eat, did go in affirmative clauses, while modern English does not. Count affirmative do-support per author and check that the KJV delta writes it where KJV does. It is a function word that carries style and no sense (open problem 4).

For Mova

Benson gives an early operational definition of style in English, and it is syntactic. That makes it measurable from the trees tools/translate already parses. I1–I3 give the style deltas (open problem 8) features beyond word choice. I4 is a small, checkable override class for profile.rs.

Topics: Machine translation

Cite

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

@book{benson1739letters,
  title = {{Letters Concerning Poetical Translations, and Virgil's and Milton's Arts of Verse}},
  author = {Benson, William},
  year = {1739},
  url = {https://www.gutenberg.org/ebooks/17548}
}

This page in Markdown → https://md.mova.elamur.ai/free/en/docs/papers/benson-1739-letters-concerning-poetical-translations/