translation.
13 of 1538 NOTE documents. Back to the full index.
| # | Title | Summary | Status |
|---|---|---|---|
| NOTE-614 | On Linguistic Aspects of Translation | Names the three kinds of translation (intralingual, interlingual, intersemiotic) and argues that interlingual translation renders whole messages, since code-units rarely have full equivalents; that every cognitive content can be conveyed in any language; and that languages differ in what their grammar obliges a speaker to say, so translation adds and drops information and a chain of translations can erode a message. Poetry, where grammar and sound carry meaning, allows only creative transposition. | Read |
| NOTE-750 | Abdulhai et al. — How LLMs Distort Our Written Language | Three studies of LLM-assisted writing. Heavy LLM use in an essay task made judge-labelled neutral stances 68.9% more common. LLM revisions of 2021 student essays moved them further, and all in one direction, in sentence-embedding space than the students’ own revisions did. LLM-written ICLR 2026 reviews named clarity and relevance less often. “Meaning” is measured by embedding position and by an LLM judge, never by the writers. | Read |
| NOTE-804 | Goldwasser et al. — A Theory of Unsupervised Translation Motivated by Animal Communication | With a plausibility prior over the target language and no parallel data, maximum likelihood learns a translator whose error is bounded like a supervised Occam bound divided by the common ground. In the common-nonsense model err = O(log|Θ| / (α·min(m,|T|))), tight up to a constant; in the knowledge-graph model more knowledge per node needs less agreement. Information-theoretic only, lossless, uniform languages. | Read |
| NOTE-1016 | Tushnet — Worth a Thousand Words: The Images of Copyright | Copyright law takes the word as its model and treats images as either transparent or opaque, with incoherent results in copyrightability, infringement, authorship and fair use. Realism is a style, not a fact. Describing an image changes how it is seen and remembered, so a test that asks factfinders to dissect works in words and then judge their gestalt cannot work. The proposed cure is a reproduction right against near-exact copies plus a narrowed derivative-works right. | Read |
| NOTE-1019 | Sotomayor — Andy Warhol Foundation for the Visual Arts, Inc. v. Goldsmith | Under the first fair-use factor, transformation is a difference in the purpose of a particular use, judged by degree and weighed against commerciality. New meaning or message counts only as evidence of a different purpose, and copying is justified most clearly when the source is the target. AWF’s licence of Orange Prince to illustrate a Prince magazine shared the photograph’s purpose and was commercial, so the factor favours Goldsmith. Kagan’s dissent would measure the factor by added expression and meaning. | Read |
| NOTE-1336 | Alvarez-Melis & Jaakkola — Gromov-Wasserstein word alignment | Unsupervised word translation is cast as a Gromov–Wasserstein problem between the two languages’ intra-language cosine-similarity matrices, solved by entropic projected gradient (each step a Sinkhorn problem), with no anchors. It matches adversarial-plus-refinement methods on related languages trained on comparable corpora at a fraction of the cost, and does markedly worse on English–Russian and English–Finnish. The objective sees only within-space similarities, so it cannot distinguish a correspondence from one composed with a symmetry of either space; the paper does not say so. | Read |
| NOTE-1337 | Søgaard et al. — Limits of unsupervised dictionary induction | The adversarial unsupervised method of Conneau et al. (2018) induces bilingual dictionaries only when the two embedding spaces are alike: it fails for English paired with Estonian or Finnish, for corpora from different domains even within English–Spanish, and for spaces trained by different algorithms. A seed of identically spelled words does as well or better. A spectral dissimilarity of small nearest-neighbour graphs tracks success across eight language pairs (|ρ| = 0.89). The failures are partly optimiser instability (some seeds succeed), and the method tested is adversarial, not GW. | Read |
| NOTE-1341 v2 | Guerreiro et al. — xCOMET | xCOMET is a learned metric for machine translation, a 3.5B or 10.7B XLM-R encoder with a sentence-score head and a token-severity head, trained on DA scores and then on MQM error spans with synthetic hallucinations. On WMT22 expert MQM it reaches segment-level Kendall τ of 0.39–0.47 and span F1 of 0.24–0.32, at or above the open metrics and GPT-4 prompting of its time, and it would have won WMT23’s metrics task. Its two heads agree with each other far better than either agrees with human annotators, partly by construction. It flags most injected critical errors and ranks fully detached hallucinations low. | Read |
| NOTE-1344 | Osmelak et al. — PETra | PETra is a corpus of about 3,000 English-source sentence pairs from TED-Multi and Europarl, in twelve target languages, in which a translation adds or replaces material to make culture-bound knowledge explicit: unit conversions, entity descriptions and specifications, translator glosses. Candidates come from null alignments filtered to pairs with a named entity, and a BERT classifier trained by active learning on German and Spanish finds the rest. Measurement conversion is the largest category. No inter-annotator agreement is reported, the text points to a single annotator, and three quarters of the pairs were found by the classifier, so the labels are not a validated measure. | Read |
| NOTE-1345 | Singh et al. — VakyArth | VakyArth is 578 items written by four native speakers, one per language, for Hindi, Punjabi, Tamil and Malayalam, over five phenomena and three formats. Five LLMs do better on MCQ than NLI in every model–language pair, and when wrong they lean to the literal reading. COMET tracks one native rater’s pragmatic-adequacy scores at r = 0.81. The translation items are not the MCQ and NLI items: in the release, 1 of 204 translation sources overlaps an inference item. So the finding that translation “does not track” pragmatic understanding compares different items. No agreement between people is reported anywhere. | Read |
| NOTE-1435 | Gutt — Translation and Relevance (thesis, 1989) | Gutt’s thesis puts translation under Sperber and Wilson’s interpretive use. Utterances resemble each other to the degree that the assumptions each communicates in its own cognitive environment overlap, so identical propositional form does not guarantee identical interpretation. A translation is faithful when it resembles the original closely enough in the respects that make it relevant to its audience, which is a context-dependent standard. Direct translation is complete interpretive resemblance in the context the original envisaged; its communicative clues, from word order to connectives and greetings, are fixed by their effect on interpretation, not by their form. | Skimmed |
| NOTE-1448 | Nord — Translating as a purposeful activity: a prospective approach | Nord states functionalist translation in brief. A text is an offer of information that different receivers take differently, so translation aims at the purpose the target text is to serve for a particular audience under the brief. Phatic, referential, expressive and appellative functions each need conditions in the receiving culture. Where those differ, the translator either adapts the function to work (instrumental) or explains how it worked (documentary). Neither is better in general, and equivalence is often neither required nor desired. | Read |
| NOTE-1451 | Nida — Principles of Correspondence | Nida sets out formal and dynamic equivalence. Formal equivalence matches the message’s form and content and suits gloss translations with notes. Dynamic equivalence follows the principle of equivalent effect: the receptor’s relation to the message should be substantially what the original receptors’ was, as in Phillips’s handshake for the holy kiss. It is “the closest natural equivalent” to the source message, natural to the receptor language, the message’s context and the audience. It must still reflect the source’s meaning and intent, with meaning before style. There is no single right answer. | Read |