Papers
Topics
Authors
Recent
Search
2000 character limit reached

To MT or not to MT: An eye-tracking study on the reception by Dutch readers of different translation and creativity levels

Published 28 Apr 2025 in cs.CL | (2504.19850v1)

Abstract: This article presents the results of a pilot study involving the reception of a fictional short story translated from English into Dutch under four conditions: machine translation (MT), post-editing (PE), human translation (HT) and original source text (ST). The aim is to understand how creativity and errors in different translation modalities affect readers, specifically regarding cognitive load. Eight participants filled in a questionnaire, read a story using an eye-tracker, and conducted a retrospective think-aloud (RTA) interview. The results show that units of creative potential (UCP) increase cognitive load and that this effect is highest for HT and lowest for MT; no effect of error was observed. Triangulating the data with RTAs leads us to hypothesize that the higher cognitive load in UCPs is linked to increases in reader enjoyment and immersion. The effect of translation creativity on cognitive load in different translation modalities at word-level is novel and opens up new avenues for further research. All the code and data are available at https://github.com/INCREC/Pilot_to_MT_or_not_to_MT

Summary

To MT or not to MT: Evaluating Cognitive Load and Creativity in Translation Modality

The study presented by Gerrits and Guerberof Arenas delves into the reception and cognitive load experienced by Dutch readers as they encounter translations executed through various modalities: machine translation (MT), post-editing (PE), human translation (HT), and the original source text (ST). Through a multi-method approach consisting of eye-tracking, questionnaires, and retrospective think-aloud (RTA) interviews, the study evaluates how translation creativity and errors influence cognitive processing and reader experience.

Methodology and Results

Eight Dutch-speaking participants, who regularly engage with literary texts, were exposed to Kurt Vonnegut's science-fiction short story "2BR02B" under each modality. Human translators professionally curated the HT and PE versions, while a custom NMT-trained engine was responsible for the MT output. The study employed eye-tracking technology to gauge cognitive load and used subsequent interviews to contextualize reading experiences.

Key findings reveal that units of creative potential (UCPs) in both HT and PE resulted in heightened cognitive load relative to MT, suggesting increased cognitive engagement with HT and PE. These effects were similarly observed in readers’ enjoyment and narrative immersion, with HT producing the most engaged reading experience. Surprisingly, error presence demonstrated a minimal effect on cognitive load across modalities, which could be attributed to readers' tendency to skim over large error segments.

The GAM model further elucidated these observations, indicating significant interactions between translation modality and creativity. HT showed the most significant cognitive load related to creative shifts (CSs), contrasting the diminished effect seen with MT. Although MT led to the lowest engagement levels, the comprehension scores ironically registered as the highest, possibly due to the compensatory strategies employed by readers.

Implications for Literary Translation and MT

This study interrogates the rising discourse surrounding the sufficiency of MT for various literary genres. The nuanced findings reveal that while MT may meet basic comprehension criteria, it falls short in delivering the intrinsic literary qualities and reader engagement provided by human-translated texts. The diminished literary style and creative engagement in MT yield pertinent questions about its appropriateness for high-quality literary translation, informing both industry practices and technological development.

Theoretical Development and Future Research

The research proposes a methodological framework for assessing cognitive engagement in translation creativity, advocating for further granularity in examining modality effects at paragraph levels. Such an approach is pivotal to unpacking the complex interactions between translation creativity, reader engagement, and cognitive processing.

Future research could enhance participant diversity and expand genre and language scope, including emerging Large Language Models (LLMs), thus contributing to a comprehensive understanding of MT, PE, and HT in diverse literary contexts. As the industry's reliance on automated translation grows, maintaining a balance between efficiency and creative quality will be vital in shaping the future trajectory of literary translation.

Paper to Video (Beta)

Whiteboard

No one has generated a whiteboard explanation for this paper yet.

Open Problems

We haven't generated a list of open problems mentioned in this paper yet.

Continue Learning

We haven't generated follow-up questions for this paper yet.

Collections

Sign up for free to add this paper to one or more collections.