Into B or not into B? The limited impact of interpreting direction on target text fluency and complexity

Understood as a language rarely learned by non-native speakers (Whyatt & Pavlović 2021), Polish might be considered a language of limited diffusion. As a result, interpreters frequently work both into and from Polish. This stands in stark contrast with the approach to interpreting directionality preferred in many international organizations and representatives of the prominent Paris School, according to which retour (A-B) interpreting results in lower quality than B-A interpreting (Seleskovitch & Lederer 1989). As directionality is “one of the oldest issues discussed in the translation and interpreting community” (Gile 2005), it has managed to attract a number of scholars supporting either the ‘into B’ or the ‘not-into B’ approach. This paper aims to contribute to this long debate by focusing on the analysis of fluency and target text complexity as modulated by interpreting directionality. In order to see whether directionality influences interpreting quality, we identified potential differences in interpretations provided by the same interpreters working into A (Polish) or into B (English) and provided into English as either A language (traditional interpreting direction) or B language (retour interpreting). We used three subcorpora from PINC (Polish Interpreting Corpus) (Chmiel et al. 2023) and analysed a set of target text fluency and complexity measures. Our analysis showed mixed results when we compared the same interpreters’ performance in two directions: fluency was higher in interpreting into B based on four measures (shorter silent pauses, fewer and shorter filled pauses, higher speech proportion) but lower based on three measures (more numerous silent pauses, shorter runs, lower articulation rate). When comparing interpreting into English as A or B language, we found lower fluency in interpreting into B based on five measures (longer silent pauses, more filled pauses, shorter runs, lower articulation rate and speech proportion) and no difference in the number of silent pauses and mean length of filled pauses. We also found no difference in target text complexity based on all measures with the exception of dependency distance: retour interpretations were more syntactically complex than interpretations into A language in the case of speeches delivered impromptu. Overall, our study shows that production fluency and complexity in retour interpreting are not consistently lower as compared to interpreting into A language. The findings thus weaken the case for interpreting into one’s native language only. URL: http://www.skase.sk/