Einleitung zu Aeschylos Agamemnon metrisch übersetzt (Introduction to the verse translation of Aeschylus' Agamemnon)
What it is about
In the preface to his German Agamemnon, Humboldt argues that such a poem cannot, strictly, be translated. Apart from words for bodily objects, no word of one language is fully equal to a word of another. Languages are like sets of synonyms: each expresses a concept with its own side meaning, a step higher or lower on the scale of feeling. He proposes a comparative synonymy of Greek, Latin and German. He then notes a paradox: the harder a translation strives for fidelity, the more it diverges, because it imitates fine particulars and can only set a different particular against each one. Still, translation is needed, both to bring foreign forms to a nation and to widen the target language itself, as Voss's hexameters widened German. Fidelity should aim at the true character of the original, not its accidents. The reader should feel "the foreign" (das Fremde), never "foreignness" (die Fremdheit). The latter shows a translator not equal to the original. A translation is not a commentary: where the original only hints, skips linking ideas or uses hard metaphors, the translator must not add a clarity of his own.
Ideas
- I1. Concreteness as a merge-risk feature. Humboldt predicts that words for physical objects match across languages and abstract words do not. Split the lexicon's merges in
meta.rsand the shuffled-lexicon control by concreteness: concrete nouns versus abstract nouns, verbs of feeling and judgement. The prediction is that merges of concrete concepts lower the bits and merges of abstract ones raise them, which would focus open problem 2. - I2. The lexicon as a "synonymy of languages". A concept should keep its members with their side meanings, not discard them. Each member of a concept carries features (era, register, number and person, connotation class) that the encoder can choose by. This extends open problem 5 from FEATS to register and era.
- I3. Test the fidelity paradox. Compare the reader residual by class for a literal and a free English Bible against the same Hebrew. Today wlc→KJV and wlc→WEB are within 10 bits per verse. The paradox predicts the literal one is lower on order and function classes and not lower on lexical ones.
- I4. Fremde versus Fremdheit as two measures. "The foreign" is source content kept: concepts and constructions with no native counterpart. "Foreignness" is target text the target grammar rejects. Measure the second as the stack residual and parse failures on the encoder's output, against the same measure on native target text. A good translation keeps the first and drives the second to the native level.
- I5. No commentary. Count the encoder's inserted connectives and linking clauses (because, so that, then) that have no meta token behind them, per author. For a literal author, each is a violation. For a sense author, they are part of the delta (compare Jerome's §7 in
jerome-1893-letter-57-pammachius-best).
For Mova
Humboldt gives the reason the meta-world cannot be a plain one-to-one lexicon, and the rule for when to keep the foreign. I1 and I2 bear on meta.rs (open problems 2 and 5). I4 gives measure.rs a split between the content kept and the strangeness added. I5 is a per-author count from what the encoder already logs.
Cite
@book{humboldt1816einleitung,
title = {{Einleitung zu Aeschylos Agamemnon metrisch übersetzt (Introduction to the verse translation of Aeschylus' Agamemnon)}},
author = {Humboldt, Wilhelm von and Leitzmann, Albert},
year = {1816},
url = {https://archive.org/details/gesammelteschrif08humbuoft}
}This page in Markdown → https://md.mova.elamur.ai/free/en/docs/papers/humboldt-1816-einleitung-zu-aeschylos-agamemnon/