‘Lost’ in interpreting and ‘found’ in translation: using an intermodal, multidirectional parallel corpus to investigate the rendition of numbers

This paper investigates the rendition of numbers in simultaneously interpreted and translated texts, under the assumption that their processing differs in the two language mediation modes. Exploiting the affordances of EPTIC (European Parliament Translation and Interpreting Corpus), a multialigned, multidirectional, intermodal parallel corpus, we examine accuracy of rendition of numbers in simultaneous interpretations with respect to several source speech parameters: type of number, amount of numerical expressions in the source sentence, delivery rate, language direction, mode of delivery and nativeness of the speaker. Interpreters’ renditions are then compared with those found in translations of the same source texts, to assess similarities and differences in coping with the same source text triggers in the two mediation modes. Conclusions are drawn about the implications of the findings for the education of language professionals.