Are pivot templates to be trusted? An exploratory multimethod approach to trust in template-centred workflows

Pivot language templates are increasingly used in the global media ecology (Díaz Cintas and Remael 2021, 54). Despite the centrality of this practice in subtitling workflows, research on templates, and more specifically on pivot templates, is recent. So far it has focused on how the use of pivot templates impacts quality (Artegiani and Kapsaskis 2014; Oziemblewska and Szarkowska 2020) or how it constrains translator's work (Georgakopoulou 2019). Other perspectives still await systematic scholarly attention. One recurrent topic among translators’ discussions on social media and other gathering places is trust and, more specifically, how templates should not be trusted. This indicates a need to further understand if and how trust/mistrust influences the decision-making processes of translators working from pivot templates. Adopting an exploratory multimethod approach to the study of trust based on analytical sociology (Barrera, Buskens and Raub 2015), we report on a study which examines trust and mistrust when translating from pivot, templates focusing on the following research questions (RQ): RQ1: Who and what is trusted/ mistrusted in template-centred workflows? RQ2: In what conditions does trust/mistrust occur when translating from pivot templates? RQ3: How technology (e.g., machine translation) is used to build/debilitate trust? The findings are based on the triangulation of: (1) a questionnaire answered by 380 European subtitlers, aimed at eliciting data on subtitlers’ experiences and expectations when translating and creating pivot language templates; (2) a follow-up vignette experiment where six of these subtitlers are asked questions about hypothetical real-life situations related to decision-making processes when translating and creating pivot templates; and (3) a follow-up focus group with the same group of participants to explore beliefs about trustworthiness and mechanisms of trust. Preliminary results suggest that while subtitlers working from pivot templates mainly express mistrust in relation to the work conducted by template-makers, template-makers, when creating templates, implement preemptive mechanisms of trust building (e.g., source- oriented translations, accurate timing and use of annotations). In addition, subtitlers report that their mistrust increases translation time and creates insecurity about where to place their trust (e.g., on automatic speech recognition and machine translation, or on template- makers?). These findings might help provide a fuller picture of the role of trust/mistrust in decision-making processes within template-centred workflows. Our findings are also likely to bring new insights to research on trust within Translation Studies, particularly to research that uses multimethod approaches to trust from the perspective of an analytical sociology.