Fidelity Assessment in Consecutive Interpretation: An Experiment
In a classroom French-to-English consecutive interpretation experiment, both the speaker and the (student) delegates were found to be unreliable fidelity assessors: they did not detect all interpretation errors on the one hand, and imagined errors that had not been made by the interpreter, on the other. As to their fidelity ratings, they remained surprisingly close to each other in spite of a wide spread in the number of reported errors. The paper also discusses this experiment in the wider context of interpretation research policy.
Interpretation quality can be viewed as a subjectively weighted sum of a number of components: the fidelity of the target-language speech, the quality of the interpreter's linguistic output, the quality of his or her voice, the prosodic characteristics of his or her delivery, the quality of his or her terminological usage, all of them as perceived by the assessor. Indeed, much of the empirical work done by interpretation researchers on quality has been aiming at measuring the weight of each of these components in quality assessment (Bühler 1984 and 1986, Kurz 1989 and 1993, Gile 1990, Kopczynski 1994).