In consecutive interpreting (CI), the interpreter speaks after the source-language speaker has finished speaking. The speech is usually divided into segments, and the interpreter sits or stands beside the source-language speaker, listening and taking notes as the speaker progresses through the message. When the speaker pauses, the interpreter then renders the message in the target language.
Consecutive interpretation can be rendered as "short CI" or "long CI". In short CI, the interpreter relies on memory; each message segment being brief enough to memorize. In long CI, the interpreter takes notes of the message to aid rendering long passages. These informal divisions are established with the client before the interpretation is effected, depending upon the subject, its complexity, and the purpose of the interpretation.
On occasion, document sight translation is required of the interpreter during consecutive interpretation work. Sight translation combines interpretation and translation; the interpreter must read aloud the source-language document to the target-language as if it were written in the target language. Sight translation occurs usually, but not exclusively, in judicial and medical work.
Consecutively-interpreted speeches, or segments of them, tend to be short, generally a few sentences or a couple of minutes. It's good to remember that most audiences don't like to sit through long periods of speech they cannot understand.
Often, if not previously advised, the source-language speaker is unaware that he or she may speak more than a single sentence before the CI interpretation is rendered and might stop after each sentence to await its target-language rendering. Sometimes, however, depending upon the setting or subject matter, and upon the interpreter's capacity to memorize, the interpreter may ask the speaker to pause after each sentence or after each clause. Sentence-by-sentence interpreting requires less memorization and therefore lower likelihood for omissions, yet its disadvantage is in the interpreter's not having heard the entire speech or its gist, and the overall message is sometimes harder to render both because of lack of context and because of interrupted delivery (for example, imagine a joke told in bits and pieces, with breaks for translation in between). This method is often used in rendering speeches, depositions, recorded statements, court witness testimony, and medical and job interviews, but it is usually best to complete a whole idea before it is interpreted. (This disadvantage can be minimized if the interpreter is provided with a copy of the speech in advance.)
Full (i.e., unbroken) consecutive interpreting of whole thoughts allows for the full meaning of the source-language message to be understood before the interpreter renders it in the target language. This affords a truer, more accurate, and more accessible interpretation than does simultaneous interpretation.