CoreCHI-P cheat sheet

Performance-Based Interpreting: CoreCHI-P Reference Sheet

• CoreCHI-P assesses interpreting performance in the candidate's language pair; scoring rubrics weight accuracy of meaning transfer above surface polish, so a fluent rendition that drops or adds a proposition scores lower than a plain but complete one.

• Three modes appear in performance tasks: consecutive (speaker pauses, interpreter renders after each segment), sight translation (written source read aloud in target language), and simultaneous (continuous delivery with a lag). Each mode has distinct scoring criteria for delivery, so identify the mode before the task begins.

• Consecutive interpreting scoring typically separates content accuracy, target-language expression, and interpreter conduct (note-taking, turn management, professional demeanor). A single segment error in a critical item (number, negation, medication, legal consequence) can fail the whole segment.

• Note-taking in consecutive mode is a memory aid, not a transcript: symbols, abbreviations, and vertical layout encoding subject-verb-object. Notes are not scored directly, but note-dependent omissions surface as content errors.

• Sight translation requires a fluent read-through before voicing; the interpreter reads a full sentence, then renders it. Literal word-for-word rendering of idioms and fixed phrases is a common accuracy failure; render meaning, not lexicon.

• Numbers, dates, dosages, and proper names must be rendered exactly; when a number is unclear, the interpreter requests clarification rather than guessing. Digits should be noted in Arabic numerals and read back in the target language's conventional number format.

• Register and politeness must be preserved: formal address forms, honorifics, and hierarchical terms in the source have target-language equivalents that must be selected deliberately, not flattened into neutral speech.

• Ethical boundaries are scored as part of performance: the interpreter does not summarize, omit, add advice, or answer questions directed at the client; first-person rendering ('I have pain') is standard for medical and legal dialogue, third-person only when the setting requires it.

• Managing the encounter: pre-session framing (introducing role, confirming the language, stating that everything said will be interpreted), requesting repetition or slowing speech, and correcting one's own error immediately are scored behaviors, not optional courtesies.