Topics — What a work is about. Nucleation holds whatever reading the owner files here, whether or not an Anthology of the SOTA topic could also hold it; instruction for ML practice belongs to the anthology (ADR-048). The words are seeded from the clusters the reading-time triage of 2026-09-25 found, and grow by decision.
Translation — translation and its theory — equivalence and its kinds (formal, dynamic, functional), interpretive resemblance, register and quality assessment, translation as skopos-driven action, intersemiotic translation between media, and what a translation preserves or changes (ADR-033).
15 of 191 CASE documents. Back to the full index.
| # | Title | Summary | Status |
|---|---|---|---|
| CASE-038 v2 | TV Tropes: a community-curated catalogue of narrative tropes and their instances, such as the Gilligan Cut | The owner proposed it at U44 as a reference case. A155 read a trope as a relational pattern recognized across changes of characters, medium and wording, and pointed to Gala et al.’s dataset as a large set of instances already annotated. A191 §8 added five manipulations of one trope, the Gilligan Cut. v7 (A178) makes it Study B, and A218 keeps it as Study 5. The first time the thing turns up in the record. The catalogue is documented; the experiments on it are proposals. | Active |
| CASE-055 | The discriminating fidelity study: renderings built to dissociate wording, proposition, static stance and transport structure, judged by named populations in two languages on stated tasks | A proposed study, filed on 2026-10-10 as the record’s repair of CLAIM-233, CLAIM-232, CLAIM-244, CLAIM-226, CLAIM-222 and CLAIM-227. It restores CASE-005’s three-way selection with human receivers, and adds a proposition-plus-force rival, a behavioural task-indexed criterion, predictors rated by other raters, named populations, an interlingual arm and a measurement stage. It has not been run, and its margins are not yet fixed. | Active |
| CASE-056 v2 | J. B. Phillips’s “hearty handshake all round” for Paul’s “holy kiss”, and Nida’s judgement that it keeps the greeting | Phillips renders all five New Testament “holy kiss” directives (Romans 16:16 and four parallels) as handshakes, which changes what the congregation is told to do; Nida (1964, p. 160) holds this up as equivalent effect, and the UBS handbooks call it a brave attempt. The documented case for the necessity half of CLAIM-061, on expert judgement rather than reader evidence. The same handbooks document the sufficiency half: where kissing is amorous, a literal “kiss” changes the act. | Active |
| CASE-059 | Tatiana’s letter in Eugene Onegin III: вы to ты and back, lost in Kline and half kept in Spalding | Tatiana’s letter to Onegin opens with the formal вы, breaks into the intimate ты at “я твоя”, and returns to вы in its last lines; only the pronouns mark the moves. Kline (2009) has “you” throughout, so neither move is in the English; Spalding (1881) switches to “thou” and never returns. A documented candidate for the half of a change of cover in which the source decides a relational question that a rendering makes undecidable. The texts are attested and the loss is visible; native attestation is thin and there is no evidence on English readers. | Active |
| CASE-062 | Simplified texts for restricted and for full readers: a gain and a loss documented on different texts, documented nulls and a reversal, and a design that tests both on the same renderings | A constructed study, built on documented halves. Simplified texts, made from the source alone, raised comprehension for named restricted readers (adults and adolescents with autism; weaker second-graders), and automatic simplifications lowered it for fluent native readers, but in different studies on different texts. Edited subtitles did not help deaf viewers, and text reduction in captions helped the more proficient readers more. The stipulated design puts both populations on the same renderings, with probe items fixed in advance as answerable from both versions or from the original only. | Active |
| CASE-063 | The English localization of Ace Attorney: a Japanese courtroom game moved to an alternate Los Angeles, with “textually accurate” set against “emotionally accurate” by its localizer | Capcom’s Gyakuten Saiban (2001) became Phoenix Wright: Ace Attorney (2005) by moving the courtroom from Japan to Los Angeles. Its second localizer, Janet Hsu, names two fidelity targets in print, “textually accurate” against “emotionally accurate”, and states the correspondence of situations she works to, an alternate Los Angeles where the California Alien Land Law of 1913 “was never passed”. The first, Alexander O. Smith, states the criterion for re-performance. The reception is contested on record. No reader has been asked a report question and a re-performance question about the same lines. | Active |
| CASE-065 | Two satisfaction questions in German and in a back-translated Chinese version: an order effect that reverses across a rendering, with the Clinton–Gore poll as the scale of the 2ε bound | Haberstroh, Oyserman, Schwarz, Kühnen and Ji (2002), Experiment 2, asked German and Chinese students the same two satisfaction questions, in a back-translated Chinese version, in both orders. The part–whole correlation depends on order in the German administration and reverses in the Chinese one, and the authors explain it by a pragmatic mechanism. Only correlations are reported, so the case shows the bound’s contrapositive in kind, not in number. Moore’s Clinton–Gore poll, held through LIT-834, supplies the number: an order effect of total variation about 0.095, against a sampling floor of about 0.03 per order at poll sizes, so 2ε is about as large as the effect it would certify. | Active |
| CASE-066 | Dolly’s slip into ты in Anna Karenina I.4, and three English translations that carry it three ways | In Anna Karenina I.4 Dolly says one sentence to Stiva with ты amid вы, and the narrator says so. English has no live T/V pronoun. Dole (1886) keeps the Russian word and glosses it; Garnett (1901) adds “Stiva”; the Maudes (1918) add “our children”. The later two add words not in the Russian to carry the relation. Tolstoy himself says вы and its dictionary equivalent vous differ in coldness. A documented instance of the compensation mechanism; the evidence on what readers recover is translators’ and the author’s, not readers’. | Active |
| CASE-067 | The reception of Deborah Smith’s The Vegetarian: the Booker judges’ task, the Korean critics’ task, and one critic’s two verdicts | Deborah Smith’s English version of Han Kang’s 채식주의자 won the 2016 Man Booker International Prize, while Korean critics published lists of errors and charged it with embellishment. Each side’s task is stated in print: the judges evaluate “the finished English-language work on their own terms”, while assessment in Korea “does usually involve comparison”. Charse Yun perceives what both sides perceive and gives two verdicts by changing task. Part of the split is access, not weight: some errors are invisible to readers without Korean. | Active |
| CASE-070 | Voita, Sennrich and Titov’s BLEU-tied translation systems that differ on keeping ты or вы across sentences | Two English-to-Russian subtitle translation systems score the same BLEU (32.40 and 32.38, not significantly different). On a contrastive test built from held-out films, which asks whether a system keeps the T or V form of “you” consistent across sentences, one is at chance (50.0%) and the other is right in 81.6% of cases. A documented instance, on a pragmatic distinction, of two realizations equal under a standard probe family and different under an independently built one. An ML paper that neither record holds. | Active |
| CASE-071 v2 | Kovacs, Deutsch and Freitag’s metric bias in MBR decoding: translations a metric selects, that the metric and its neighbours rate better and human raters do not | Minimum Bayes risk and quality-estimation decoding pick, from 128 sampled translations, the one a neural metric scores best. That metric and other neural metrics rate the picks significantly better than greedy decoding; human MQM raters find them no better, and for four system and language-pair combinations significantly worse. CASE-041’s result in text, by selection rather than training. The distinction lost is accuracy, not a pragmatic one; a stipulated English-to-Russian ты/вы variant states the design that would make it pragmatic. | Active |
| CASE-072 | Hafez’s “Shirazi Turk” in English: one ungendered beloved rendered as a maid, as God and as a boy | The opening of Hafez’s ghazal 3 names a beloved, “that Shirazi Turk”, whose sex and nature the Persian leaves open. English forces a choice at the first possessive. Jones (1771) and Bell (1897) make the beloved a maid, Clarke (1891) makes it God, and Davis (2012) makes it a boy, and Davis says in print that his choice is arbitrary. A documented instance of the half of a change of cover in which the rendering decides what the source left open. The disagreement on the source is among translators and scholars; no reader agreement has been measured. | Active |
| CASE-122 | Warhol’s Orange Prince, licensed to Condé Nast in 2016 from a 1981 Goldsmith photograph: the Supreme Court judged the use by its purpose and the dissent by its meaning, and the same cover came out a substitute on one test and a new work on the other | Andy Warhol Foundation v. Goldsmith (2023), read in NOTE-1019. The experts agreed Warhol’s silkscreen changed the photograph’s meaning, from Prince’s private identity to his public image. The majority held that this did not matter for the use at issue: a magazine cover illustrating a story about Prince serves substantially the photograph’s own purpose, and new meaning counts only as evidence of a different purpose. The dissent measured what the copier added. One rendering, two measures of what it changed, two verdicts. | Active |
| CASE-184 | The central defeat test as stated: human judgements of fidelity or shared kind, predicted by the theory’s judged measures against a text-similarity baseline | The test that the defeat conditions of CLAIM-115, CLAIM-139, CLAIM-005 and CLAIM-078 share, filed as one design so that an argument can start from it. The criterion is human judgements of fidelity or of shared kind. The theory’s predictor is a measure of judged pragmatic observables. The rival is a text-similarity score. It has not been run, and CASE-055 is the design the record filed to repair it. | Active |
| CASE-185 | A proposed study: direct and relay translations of one set of source utterances through three languages, to find whether agreement between the routes, the task-relative loss on each route, and the preservation of response to later framing come apart, and whether correspondences fitted between the communities’ response structures on some items predict receivers’ judgements on held-out items, translators and situations | Put together from A4 §6, A5 §5 and A6’s synthesis of the exchange “Paper assessment” (2026-10-28 to 29), which proposed carrying LIT-1820’s comparison of alternative paths into CASE-055 as independent translation routes. It adds to CASE-055 a relay arm, a path-discrepancy readout set beside the task loss, two negative controls, and an arm that fits correspondences between the communities’ response matrices and scores them on held-out data. Not run; every margin is left to a preregistration. | Active |