Narrating Creativity: Human Translation vs. Machine Translation Post-Editing in Subtitling Training

This study investigates the effects of machine translation post-editing (MTPE) training on the creativity of fourth-year Translation and Interpreting students in subtitling narrative, in comparison with conventional human translation (HT) practice. Employing a pre-test/post-test experimental design where creativity was measured in terms of accuracy and novelty, the study analysed performance changes before and after the instructional intervention in order to assess the pedagogical effects of MTPE training. Statistical analyses revealed no significant differences between the HT and MTPE groups in overall Creativity Scores. However, the HT group produced more Creative Shifts, while the MTPE group tended to produce fewer errors. ANCOVA analyses confirmed that these between-group differences were attributable to pre-existing baseline differences rather than training mode. Within-group analyses using proportional scores further revealed that both groups significantly improved in acceptability after training with large effect sizes, and both showed a non-significant trend toward increased Creative Shifts. The HT group additionally showed a statistically significant improvement in overall Creativity Scores. The MTPE group also showed greater variability in performance, suggesting the possible influence of individual factors. These findings indicate that HT and MTPE support different dimensions of creativity rather than one having a greater impact. The study contributes to the limited research on creativity in Turkish subtitling and suggests that an integrated training model combining both approaches may be pedagogically beneficial.