Simultaneous Interpretation: Temporal and Quantitative Data

Abstract

Simultaneous interpretation represents a complex form of activity, about which there is very little in the way of experimental data. In an exploratory study, two professional, two student and two amateur interpreters (one English-dominant and one French-dominant in each case) were required to do the interpretation of several texts from either their weaker into their dominant language or vice-versa. The texts represented four types of material: spontaneous speech, semi-prepared material, prepared " oral " material, and prepared " written " material. Presentation of texts and recording of translations was on tape. The temporal characteristics of the speaker (S) and interpreter or translator (T) delivery patterns and their interrelationship were analysed by computer. Translated speech per se is found to generally present the same patterns and relationships among speech parameters as natural speech, but to be less rhythmical. In omission-free translation T tends to be engaged in speaking for a greater proportion of time than S, largely because of slower articulation. Proportionally more of the T's utterance is delivered while the S is pausing than would be expected were the two delivery patterns independent; thus the T tries to take advantage of the S's pauses and thereby reduce the extent of time during which he must be both speaking and listening at the same time. Characteristically, the T lags behind the S in his delivery by 2 to 3 sec. The temporal data do not in general reveal striking differences in relation to the different categories of T, the various types of material, or the two directions of translation.