Fluency parameters in the Polish Interpreting Corpus (PINC)

The following chapter introduces PINC — the Polish Interpreting Corpus, a Polish-English and English-Polish corpus of short European Parliament speeches and their interpretations. The uniqueness of PINC, apart from its language combination, consists in careful balancing of mode of delivery, in rich metadata, interpreter identification and availability of a strictly controlled subcorpus of retour interpretations. The chapter also briefly presents custom-built tools used in the making of the corpus, especially for transcription, text-audio alignment at word level and interpreter identification. To showcase PINC's potential for analysing various aspects of simultaneous interpreting, we examined fluency parameters, such as speaking rate and pauses, in the Polish-English subcorpus. We found that interpreting speed was modulated by the source text speaking and articulation rate and the target text compression rate. Target texts had fewer but longer silent pauses and more numerous and longer filled pauses. Together with shorter runs, understood as utterances uninterrupted by pauses, this suggests more fragmented delivery of interpretations. We also found interesting individual differences in compression rate with the majority of interpreters producing interpretations longer than the source texts.