Literary Machine Translation as a Human-Machine Dialectic

 

6 October 2022

 

The popularization and advances of machine translation (MT) in the recent years have made this technology an integral part of translation. Helped by the proliferation and trading of data, by the shift towards fragmented and perishable content, by the massive investments from public bodies and private businesses into translation technologies, and by the continued growth of the language industry, machine translation tools have consequently given additional momentum to some of the transformations that were already underway since the advent of the first computer tools, e.g. translation memory systems, and that have progressively changed the working processes, training, role, status, and power dynamics of translators (O’Brien 2012; Kenny 2017; Desjardins et al. 2020; O’Hagan 2020). The biggest of these changes, however, is perhaps the wider access to translation beyond the professional sphere and the increased presence of MT in public debate since most people now have access to online systems and translations that are, if not perfect, at least much more comprehensible and acceptable than they used to be (Way 2013). As a result, some people have been tempted to break the traditional dichotomy between literature and computers, in an attempt to assess the usefulness of MT for creative texts.

Although there has been a long-standing interest in the convergence of machines and literary arts, as manifested in the computerization of the Oulipian Raymond Queneau’s Cent mille milliards de poèmes, in the subsequent experiments of the ALAMO (Atelier de Littérature Assistée par la Mathématique et les Ordinateurs), or more recently in the organization of the CLfL (Workshops on Computational Linguistics for Literature), the last years have seen a sharp rise in the amount of research conducted on literary machine translation. With the arrival of neural machine translation (Bahdanau et al. 2014), and the Transformer architecture in particular (Vaswani et al. 2017), various researchers have thus started to investigate the possibility of having literature translated by or with the help of computers, developing this area of research in a number of scientific gatherings and publications that keeps growing today, starting with:

Some of these initiatives echo the wishes expressed by professionals, who are either curious or keen to discover new tools that would be better suited to the literary domain, whereas others simply reflect the desire to explore a field of research that was not often discussed before then. The tools themselves can range from applications that aim to facilitate the work of translators, such as corpus exploration environments (Loock 2016), computer-assisted translation (Rothwell 2018) or collaborative platforms (Goncharova and Lacour 2011), to programmes that focus on a specific translation task (Miller 2019) or methods that can help translators and scholars to reveal new understandings of a creative text (Youdale 2020; Bories et al. 2022). Among these, MT now seems to have carved out a place for itself in the larger reflexion on computer-assisted literary translation (CALT) and established itself as a potential new tool.

If the scientific relevance immediately comes to mind, the progress made by MT since the turn of the millennium might also be of use to literary translators. Even though MT systems are still evidently incapable of “understanding” the texts themselves or the translation choices that they are asked to reproduce ‒ proof that human intervention will remain an essential part of translation for years to come ‒ this does not necessarily mean that they should be entirely useless nonetheless. It might help, for instance, by offering suggestions, speeding up the translation of certain fragments and freeing up some time for more demanding and creative segments, or by stimulating creative thinking and encouraging different reinterpretations of the same text, especially if MT intervenes as an interactive suggestion tool during or after the initial translation, rather than a pre-processing step that churns through the text and spoon-feeds a raw draught in its entirety.

Halfway between a productivity tool and a computer-assisted literary creation tool, machine translation systems, contrary to traditional CAT tools, have the advantage of offering a fresh look, or a second reading of either the source or the target text. Similarly:

The merit of NLP research is that it inevitably makes us aware of the complexity and subtlety of natural language. There lies the enthralling and thereby increasingly fascinating aspect of this line of research: to bring out the subtleties of natural language, even those that we did not expect. (Chaty 1998)

As pointed out by the same author:

It is always a matter of communication! Between humans, between humans and machine, between machines. (Ibid.)

Still, the question of literary machine translation (LMT) remains very much unresolved. While studies have shown that it is indeed possible to adapt MT systems to the literary domain (Toral & Way 2018; Kuzman et al. 2019; Matusov 2019), this realization raises many more issues and interrogations:

  • What does this mean for creativity and reader experience (Guerberof and Toral 2020)?
  • What impact would it have on property rights and work conditions (Taivalkoski-Shilov 2019)?
  • What are translators’ thoughts on technology (Daems 2021)?
  • How can we ensure that the machine does not erase translators’ voice (Kenny and Winters 2020)?
  • How does a human translation differ from post-editing and revision of a literary text (Macken et al. 2022)?
  • Would it be possible to tailor systems not just to literary data, but to each translator as well (Hansen et al. 2022)?
  • Could proper post-editing experience and tailored environments help mitigate the priming and constraining effects of MT (Hansen 2021)?
  • Could we ever hope to see MT systems capable of maintaining discursive continuity and contextual information (Poncharal 2021)?
  • …

To address some of these questions, learn more about the various aspects of this topic, open up new research avenues and discuss the possible future directions for LMT, we invite you to join a leading panel of experts for an entire day of presentations and round tables centred specifically around the subject of LMT.

References

Bahdanau, Dzmitry, et al. “Neural Machine Translation by Jointly Learning to Align and Translate”. 3rd International Conference on Learning Representations: Conference Track Proceedings, edited by Yoshua Bengio and Yann LeCun, 2015, pp. 1–15.

Bories, Anne-Sophie, et al., editors. Plotting Poetry: On Mechanically-Enhanced Reading. Presses Universitaires de Liège, 2022.

Chaty, Guy. “Dialogue entre sciences et literature : perspectives et limites à travers notamment l’expérience d’ALAMO”. ALAMO, 1998.

Daems, Joke. “Wat denken literaire vertalers echt over technologie?”. WEBFILTER, 2021.

Desjardins, Renée, et al., editors. When Translation Goes Digital: Case Studies and Critical Reflections. Palgrave Macmillan, 2020.

Goncharova, Yuliya, and Philippe Lacour. “TraduXio : nouvelle expérience en traduction littéraire”. Traduire, no. 225, 2011, pp. 86–100, doi: 10.4000/traduire.94.

Guerberof-Arenas, Ana, and Antonio Toral. “The Impact of Post-Editing and Machine Translation on Creativity and Reading Experience”. Translation Spaces, vol. 9, no. 2, 2020, pp. 255–282, doi: 10.1075/ts.20035.gue.

Hansen, Damien. “Les lettres et la machine : un état de l’art en traduction littéraire automatique”. Actes de la 28e Conférence sur le Traitement Automatique des Langues Naturelles, edited by Pascal Denis et al., vol. 2, ATALA, 2021, pp. 28–45.

Hansen, Damien, et al. “La traduction littéraire automatique : Adapter la machine à la traduction humaine individualisée ?”. Journal of Data Mining and Digital Humanities, forthcoming.

Kenny, Dorothy, editor. Human Issues in Translation Technology. Routledge, 2017.

Kenny, Dorothy, and Marion Winters. “Machine translation, ethics and the literary translator’s voice”. Translation Spaces, vol. 9, no. 1, 2020, pp. 123–149, doi: 10.1075/ts.00024.ken.

Kuzman, Taja, et al. “Neural Machine Translation of Literary Texts from English to Slovene”. Proceedings of the Qualities of Literary Machine Translation, edited by James Hadley et al., EAMT, 2019, pp. 1–9.

Loock, Rudy. La traductologie de corpus. Presses Universitaires du Septentrion, 2016.

Macken, Lieve, et al. “Literary Translation as a Three-Stage Process: Machine Translation, Post-Editing and Revision”. Proceedings of the 23rd Annual Conference of the European Association for Machine Translation, EAMT, 2022, pp. 101–110.

Matusov, Evgeny. “The Challenges of Using Neural Machine Translation for Literature”. Proceedings of the Qualities of Literary Machine Translation, edited by James Hadley et al., EAMT, 2019, pp. 10–19.

Miller, Tristan. “The Punster’s Amanuensis: The Proper Place of Humans and Machines in the Translation of Wordplay”. Proceedings of the Second Workshop on Human-Informed Translation and Interpreting Тechnology (HiT-IT 2019), Incoma Ltd., 2019, pp. 57–64, doi: 10.26615/issn.2683-0078.2019_007.

O’Brien, Sharon. “Translation as Human–Computer Interaction”. Translation Spaces, vol. 1, no. 1, 2012, pp. 101–122, doi: 10.1075/ts.1.05obr.

O’Hagan, Minako, editor. The Routledge Handbook of Translation and Technology. Routledge, 2020.

Poncharal, Bruno. “La TA à l’épreuve du texte littéraire : d’une (im)possible restitution de l’expérience de lecture ?”. La main de Thôt, no. 9, 2021.

Rothwell, Andrew. “CAT Tools and Creativity: Retranslating Zola in Stereo”. Translation and Creativity: Readers, Writers, Processes, edited by Mariagrazia De Meo and Emilia Di Martino, Aracne editrice, 2018, pp. 169–181.

Taivalkoski-Shilov, Kristiina. “Ethical Issues Regarding Machine(-Assisted) Translation of Literary Texts”. Perspectives: Studies in Translation Theory and Practice, vol. 27, no. 5, 2019, pp. 689–703, doi: 10.1080/0907676X.2018.1520907.

Toral, Antonio, and Andy Way. “What Level of Quality can Neural Machine Translation Attain on Literary Text?” Translation Quality Assessment: From Principles to Practice, edited by Joss Moorkens et al., Springer, 2018, pp. 263–287.

Vaswani, Ashish, et al. “Attention is All you Need”. NIPS’17: Proceedings of the 31st International Conference on Neural Information Processing Systems, edited by Ulrike von Luxburg et al., Curran Associates Inc., 2017, pp. 6000–6010.

Way, Andy. “Traditional and Emerging Use-Cases for Machine Translation”. Proceedings of Translating and the Computer 35, Aslib, 2013, pp. 1–12.

Youdale, Roy. Using Computers in the Translation of Literary Style: Challenges and Opportunities. Routledge, 2020.

Partagez cette page