SLTES Open common mistakes
SLTES Open · Understand the traps and how to avoid them.
Reading the source text word by word and transferring each lexical item in order, producing target sentences that follow source syntax rather than target grammar.
Candidates who learned translation as vocabulary substitution default to the surface form under time pressure, and the exam's dense passages reward speed over restructuring.
What to do instead
Before writing, identify the main clause and its arguments, then build the target sentence around target-language word order; practise by translating short paragraphs with a deliberate 'read the whole sentence, then look away, then write' cycle.
Failing to notice register and field markers (formal vs. colloquial, technical vs. lay) and rendering an official document in casual phrasing or a chatty text in stiff bureaucratic style.
Candidates focus on propositional meaning and treat style as decoration, often because their practice texts were neutral expository prose.
What to do instead
Annotate each practice passage for register, audience, and text type before translating, and build a personal list of target-language register markers (contractions, honorifics, nominalisation) to check against at the end.
In consecutive interpreting, taking notes as full words and sentences, then reading the notes back and losing the speaker's structure and any content not written down.
Candidates distrust memory and try to transcribe, which splits attention and produces notes too dense to decode quickly.
What to do instead
Train note-taking as a symbol-and-abbreviation system recording ideas, links, and proper nouns only; drill 2-3 minute passages with a strict rule of no more than a few symbols per idea unit, then reproduce from notes alone.
In simultaneous interpreting, continuing to speak after a missed segment and inventing plausible content to fill the gap instead of signalling or omitting.
Fear of silence and the pressure to sound fluent lead candidates to fabricate, which is both an accuracy and an ethics failure.
What to do instead
Practise deliberate omission and compression drills where dropping a redundant modifier is the correct move; rehearse neutral recovery formulas and set a personal rule that unverified content is never voiced.
Accepting a translation or interpreting assignment without confirming the brief: deadline, audience, purpose, deliverable format, and whether revision or review is included.
Candidates treat the request as self-explanatory and assume the client's expectations match their own, especially in short exam-style scenarios.
What to do instead
Memorise a fixed intake checklist (purpose, audience, register, format, deadline, terminology sources, review stage) and apply it to every scenario question, stating the questions you would ask.
Treating confidentiality as satisfied by 'not telling anyone' while discussing the text with a colleague for help, posting a terminology query containing source content, or using a machine engine on client material.
Candidates equate confidentiality only with gossip and overlook disclosure through tools, queries, and informal consultation.
What to do instead
Learn the disclosure pathways explicitly: anonymise queries, strip identifying content before any external tool, and check whether the client's terms permit machine assistance before using it.
Building a glossary as a flat bilingual word list with no definitions, context, part of speech, or source, then applying entries mechanically across different senses.
Candidates equate terminology work with collecting equivalents and skip the conceptual analysis that distinguishes polysemous terms.
What to do instead
Record each term with definition, part of speech, domain, context sentence, source, and status; when a term has multiple senses, create separate entries and decide by context, not by first match.
Using CAT tool leverage and machine output without verifying that placeholders, tags, numbers, and formatting survived, and delivering segments that read fluently but carry wrong figures.
Candidates trust match percentages and pre-translated segments, and QA is often skipped when the deadline is tight.
What to do instead
Run a fixed post-processing QA pass: compare all numbers and units, check tag pairing and placeholders, and read the target alone for internal consistency before delivery.
Rendering idioms, metaphors, and culturally specific references literally, producing target text that is grammatically correct but semantically odd or misleading.
Candidates look up individual words and lack a strategy for non-compositional meaning and culture-bound items.
What to do instead
Classify the item first (idiom, institution, realia, allusion), then choose among equivalent, explanatory, or borrowed rendering based on target audience knowledge; keep a log of culture-bound items with the strategy used and why.
Producing target text that is accurate in content but fails cohesion and punctuation conventions, with run-on sentences, wrong quotation marks, or inconsistent terminology across the document.
Candidates proofread for meaning only and are unaware of target-language typographic and cohesive norms, which the language proficiency domain tests directly.
What to do instead
Do a dedicated monolingual read of the target for punctuation, sentence boundaries, and pronoun reference; maintain a per-language style sheet covering quotes, dashes, numbers, and dates, and apply it consistently.