This experimental study addresses one of the least explored audiovisual modes: reverse subtitling (L1 audio, L2 subtitles). Specifically, it investigates the effects of different translational choices on learners’ noticing and memorisation of lexical items and grammatical structures. The participants were English (L1) native speakers learning Italian (L2) at an upper-intermediate level (CEFR B2). Formal similarity (literal transfer) and discrepancy (non-literal transfer) between L1 and L2 were compared to establish if and how they affect the learners during subtitle processing and recall. Does one of the two translation conditions yield a better recall rate in a verbatim memory post-test? This constitutes the main research question addressed in this study. The main hypothesis was that there would be a difference in recall by translation condition, with formal equivalence having a facilitative effect on memory and literal subtitles therefore being more accurately recalled by learners. To determine how the different subtitle translations were processed, attention allocation and noticing were investigated through triangulation of eye-tracking, the recall post-test and an open questionnaire, which allowed the subjects’ thought processes to also be recorded. Subtitle-specific variables such as corpus frequency and linguistic category (lexicon vs. syntax) were also analysed. While participants watched the reversely subtitled clip, their eye behaviour was recorded using a Tobii X120 eye-tracker. After watching, participants answered the recall post-test followed by the open questionnaire and took part in a working memory control test.
Translation condition was found to influence recall, with literal translations yielding superior recall performance than non-literal ones. The data also showed that participants did notice a translation discrepancy. Eye-tracking findings reveal a complex relationship between language elaboration and memory, whereby comparable amounts of visual attention to two subtitle versions can result in significantly different recall. Moreover, considerable insights were drawn from the open questionnaire, indicating that qualitative data can provide a richer picture of processing and memory attainment and should more regularly support experimental studies. The results demonstrate that subtitle-specific factors like translation can indeed influence the viewer and should therefore be taken into consideration in the design of future subtitle reception studies. The mnemonic potential of reverse subtitles for foreign language learning is also confirmed, strongly suggesting that this subtitling mode should be reconsidered as a valuable tool in language learning and deserves a place in the foreign language classroom.
Future research could build on this study by using a larger sample size and more advanced statistical techniques, such as multilevel modelling. The results obtained highlight the complexity of the language faculty and call for additional reception studies where more fine-grained analyses further assess the effects of translation during the consumption of subtitled material. For instance, using a higher-frequency eye-tracker and considering more eye movement measures in the future will provide more precise insights in the reading process, enabling deeper understanding of information processing and memory retention, both crucial aspects in the development of foreign language skills.
