Translation Study Guide

📖 Core Concepts Translation – communication of a source‑text’s meaning by an equivalent target‑text. Interpreting – oral or signed rendering of speech; deals with spoken language, not written text. Fidelity – how accurately the translation reflects the source meaning (faithfulness). Transparency (idiomaticity) – how naturally the translation reads in the target language. Equivalence Spectrum – formal equivalence (metaphrase, word‑for‑word) ↔ dynamic equivalence (paraphrase, meaning‑for‑meaning). Back‑translation – re‑translating the target text back into the source language to check accuracy. CAT tools – dictionaries, translation memories, terminology‑management systems that aid human translators. Machine Translation (MT) – automatic generation of a target text; always needs human pre‑/post‑editing. BLEU & NIST – automatic metrics that compare MT output to reference translations (n‑gram precision). --- 📌 Must Remember Translation ≠ Interpreting – written vs. oral. Fidelity vs. Transparency trade‑off – higher fidelity can reduce transparency, and vice‑versa. False friends – look‑alike words that differ in meaning; a common pitfall. Borrowing – used when the target language lacks a term; enriches the lexicon. Back‑translation purpose – quality check, especially in clinical research and surveys. TRAPD model – Translation → Review → Adjudication → Pretest → Documentation for survey translation. BLEU formula (simplified): $$\text{BLEU}=BP \times \exp\!\Big(\sum{n=1}^{N} wn \log pn\Big)$$ where \(pn\) = n‑gram precision, \(BP\) = brevity penalty. Schleiermacher’s two methods – foreignization (preserve source culture) vs. domestication (adapt to target culture). --- 🔄 Key Processes Full Source‑Text Analysis – read, note semantics, style, cultural references. Draft Translation – aim for literal equivalents where possible, paraphrase where needed. Terminology Check – use glossaries, borrow if no target term exists. Human Post‑Editing (MT) – resolve ambiguity, ensure idiomatic flow. Back‑Translation (if required) – translate target back, compare to original, revise. Survey Translation (TRAPD) Translation: initial draft. Review: peer/sme review. Adjudication: resolve conflicts. Pretest: pilot with target respondents. Documentation: record decisions. --- 🔍 Key Comparisons Metaphrase vs. Paraphrase – word‑for‑word (formal) ↔ re‑expressed meaning (dynamic). Foreignization vs. Domestication – keep source culture visible ↔ make text feel native. Human Translation vs. Machine Translation – nuanced, cultural accuracy ↔ speed, consistency but limited on ambiguity. Sworn (Certified) Translation vs. Regular Translation – legal equivalence, notarized ↔ general purpose. --- ⚠️ Common Misunderstandings “Translation is just word substitution.” – ignores cultural context, connotation, and ambiguity. “Machine translation can replace humans.” – MT cannot reliably handle pronoun disambiguation, idioms, or cultural nuance. “Back‑translation proves a translation is perfect.” – it is an approximate check; differences may stem from legitimate alternatives. “Formal equivalence = better fidelity.” – excessive literalism can sacrifice readability and transparency. --- 🧠 Mental Models / Intuition “Bridge Model” – think of translation as building a bridge: the pillars are fidelity (source side) and transparency (target side); the deck must be stable enough for readers from both banks. “Signal vs. Noise” – the translator’s job is to preserve the signal (meaning) while filtering out noise (source‑language syntax that doesn’t belong in the target). --- 🚩 Exceptions & Edge Cases Free‑word‑order languages (e.g., Russian) may allow more syntactic flexibility than fixed‑order languages (English). Poetry – meter, rhyme, and imagery often force creative deviation; “hypertranslation” (chōyaku) may intentionally improve the original. Legal/Medical documents – require literal precision; borrowing is limited, and certified translation may be mandatory. Religious texts – often treated as “interpretations” (e.g., Quranic translation) because of doctrinal nuance. --- 📍 When to Use Which Choose Formal Equivalence when the field demands exact wording (legal contracts, technical manuals). Choose Dynamic Equivalence for literary, marketing, or audience‑engaging texts where feel matters more than exact phrasing. Use Back‑Translation for consent forms, survey items, or any material where cross‑cultural validity is critical. Apply CAT tools for repetitive terminology‑heavy projects (software localization, technical docs). Rely on Human Post‑Editing for any MT output destined for publication, especially in literary, legal, or medical domains. --- 👀 Patterns to Recognize False‑friend pattern – similar‑looking words across languages that have different meanings (e.g., English actual vs. Spanish actual “current”). Pronoun ambiguity – sentences with “he/she/it” often need contextual clues; MT will usually fail here. Cultural‑specific concepts → gloss or footnote rather than forced literal translation. Technical term repetition – indicates a term that should be stored in a translation memory for consistency. --- 🗂️ Exam Traps “Translation is only about words.” – exam may present a definition lacking cultural/contextual aspects; the correct answer must include connotation and cultural references. Choosing BLEU as the sole quality metric. – BLEU measures surface similarity, not meaning; high BLEU does not guarantee a good translation. Confusing “foreignization” with “literal translation”. – foreignization preserves cultural distance, not necessarily word‑for‑word fidelity. Assuming back‑translation equals perfect accuracy. – differences can be legitimate; the trap is to pick “back‑translation proves correctness” as a statement. Mixing up sworn vs. certified translation requirements. – some jurisdictions require notarization; others accept a simple declaration. ---