What language is that? The need for a new norm for subtitling multilingual audiovisual texts

There has been a substantial development in the use of multilingualism and language diversity in films and other audiovisual texts over the last few decades (cf. e.g. Voellmer & Zabalbeascoa 2014). In the 1970s and 1980s, foreignness was mainly signalled via the use of foreign accents of English. This is most striking in the long-running BBC TV series ‘Allo ‘Allo (Lloyd & Croft 1982 – 1992, as described by Delabastita 2010), where various European languages (including English) were all stereotypically portrayed in this way. The 1990s saw greater linguistic diversity and the use of code-switching in films like Trainspotting (Boyle 1996). There was also simplification, in that modern use of language was transposed to other contexts, as can be seen in Braveheart (Gibson 1995), where the social and regional variation of Britain today was used to portray social and geographical variation in medieval Britain. The turn of the millennium saw a shift towards quasi-realistic linguistic credibility (Pedersen 2007), with films such as The Passion of the Christ (Gibson 2004), where the languages spoken in Biblical times (e.g. Aramaic, Latin and Hebrew) were used in the film. As a result of globalisation, this trend continues to grow stronger today, with many audiovisual products being coproduced transnationally, and endeavouring to represent the languages actually spoken in various countries. This can be seen in TV series such as Le Bureau des légendes (Rochant 2015 - 2020), which has French as its main language (L1), but several other languages (L3s; e.g. Russian, Arabic, and Farsi) being represented realistically, or Hit and Run (Issacharoff et al 2021) which has two L1s, English and Hebrew. There has been a great deal of research into how dubbing handles multilingualism in audiovisual content in the last decade (e.g. in Rebane & Junkerjürgen 2019 or Corrius, Espasa & Zabalbeascoa 2019), but less so for subtitling. In this mode of AVT, multilingualism is largely ignored or standardized in the translation into the target language (L2), as shown by the studies that we have seen (e.g. Magazzù 2019 or Dore 2019), and the prevalence of this practice is also testified by subtitlers. An investigation into public service and commercial guidelines for subtitling shows that these exclusively give advice on situations where multilingualism consists of the infrequent use of odd words or phrases in a third language, with a specific function, like humour, exoticism, atmosphere or character (cf. de Higes Andino et al 2019). The guidelines offer no advice on how to deal with the more widespread genuine multilingualism that we see on screen today. This leaves rooms for solutions that are controversial, as in the English subtitles of Roma (Cuarón 2018), and sometimes ill-advised, as in the Swedish AVT of the aforementioned Hit and Run. Furthermore, as the source text is part of the target text in subtitled content, L2 viewers are meant to be able to tell when an L3 is spoken. But can they? Is this realistic? Or can it be claimed that multilingualism is lost or muffled in subtitling? It seems that this is the time to start looking for a new norm for handling today’s more widespread and realistic forms of multilingualism in subtitled audiovisual texts.