Abstract
This paper discusses corpus-based analyses of English-Japanese simultaneous interpreters’ speech rates. We used the Simultaneous Interpretation Database of Nagoya University in order to conduct the quantitative analyses that require a large scale corpus. We analyzed a total of 88 sets of simultaneous interpretations by 17 interpreters covering 22 different lectures given in English. As a result, we confirmed that the lecturers' speech rates had little relevance to the interpreters' speech rates. Moreover, it became clear that the interpreters' average speech rate was faster after each source utterance was finished than before it was finished. We also confirmed that the interpreters' average speech rate was faster while the lecturers were not speaking than while the lecturers were speaking.
