Repository path: workshop/translations/miserere/rulings-R07.md · rendered 2026-09-09

R07 interpretive rulings for the «El Miserere» rendering

R07-fluency.md §"Binding condition, added 2026-07-31 (S073)" requires that any work translating under R07 settle IR1, IR2 and the missing rules M1–M5 in its own text before it opens the source, and say which reading it took. Committed before the source was opened for translation. The design that uses the rendering cites this page rather than restating it.

IR1 — the scope of F10 ("Nothing that calls attention to the language")

Reading taken: the narrow one — F10 governs the translator's own latitude, not the reproduction of figures the source contains. A simile, a repetition or an apostrophe that is in Bécquer is rendered; what F10 forbids is a word choice whose noticeability is the translator's addition.

Stated against the evidence, not around it. R07's own §"Measured defects" 2 records that three non-lead readers, given F10 verbatim, returned 8 of 12 cells saying F10 requires the plain rendering — i.e. the literal reading, which this ruling rejects. The narrow reading is taken anyway, for one reason: on this span the literal reading would delete the source's entire figural apparatus (P2's phosphorescent-skeleton simile, P8's four-term music simile), and a rendering that deletes the content is not a register pole, it is a different text. Consequence to be borne: F10 does very little work in this log, and where a decision is coded on F10 alone it is flagged IR1.

IR2 — the scope of F6 ("Idiomatic syntax before close syntax")

Reading taken: F6's text, literally — it licenses departure from the source's order and forbids nothing. F6 may therefore never be cited to exclude a paraphrase, and no decision in this log is coded D on F6. The three published runs read F6 as forbidding paraphrase; R07 § "Measured defects" 6 records that the text does not.

Missing rules — declared before they fire

• M1 typography/layout — the Latin verse [P7] is set as its own paragraph in the source. F1–F10 are silent on whether it is italicised, indented, or run in. Held as the copy-text has it, and the decision is coded S.
• M2 sequence of tenses — Spanish preterite/imperfect against English past. Silent in F1–F10; every such site is S unless F7 bears on the clause independently.
• M3 F4's heading is wider than F4's clauses — the span carries one word of a third language (Latin, [P7]). F4's heading ("No foreign matter") reaches it; F4's clauses, which name source-language words, do not. Ruling: the Latin stands untranslated in both renderings, because it is a liturgical citation whose identity as a citation is the sentence's subject, and the site is coded S, not D.
• M4 steadiness of a lexical choice across a text — silent. ara occurs twice in [P3] and versículo(s) twice across [P6]/[P8]; each is held to one English word, coded S.
• M5 second-person address systems — does not arise; the span has no second-person address.
• M6 proper-noun knowledge (added S077) — Rey Salmista in [P8] presumes the reader knows David wrote Psalm 50. F4 reaches proper-noun conventions, not proper-noun knowledge. Ruling: no gloss, no footnote, in either rendering — a footnote would be a difference between the arms that has nothing to do with register. Coded S.

Coverage statistic

Only the strict §5 test is computed — a rule names the feature and exactly one live option survives it. The exclusion-test rate is not computed, per the binding condition's last clause. The mean number of live options per logged site is reported beside the strict rate.

Dialect

F5: American English, declared here and held. The same declaration governs the R08 rendering, so dialect is not a difference between the arms.