Creativity and narrative engagement of literary texts translated by translators and neural machine translation

Projektinformationen

CREAMT

ID Finanzhilfevereinbarung: 890697

DOI

10.3030/890697

Projekt abgeschlossen

EK-Unterschriftsdatum 23 März 2020

Startdatum 1 September 2020

Enddatum 31 August 2022

Finanziert unter

• EXCELLENT SCIENCE - Marie Skłodowska-Curie Actions

Gesamtkosten

€ 175 572,48

EU-Beitrag

€ 175 572,48

Koordiniert durch

RIJKSUNIVERSITEIT GRONINGEN
Netherlands

Periodic Reporting for period 1 - CREAMT (Creativity and narrative engagement of literary texts translated by translators and neural machine translation)

Berichtszeitraum: 2020-09-01 bis 2022-08-31

Zusammenfassung vom Kontext und den Gesamtzielen des Projekts

The main objective of CREAMT is to explore what is meant by creativity in translation in literary translation, how to define it and quantify it, and how technology, especifically machine translation (MT) impacts creativity.

The focus of the project is on the textual elements that determine creativity in translated literary texts as well as to assess the experience of the reader, to assess how effective MT is in literary translation.

With this aim in mind, the following questions are asked:

1. Can we quantify creativity in texts translated by humans as opposed to those produced with the aid of machines?
2. And, since the main aim of the translation of a literary text is to preserve the reading experience of the original, what is the reader’s experience when faced with machine-translated texts? Do users exposed to different translation modalities have different reading experiences?

This information is a key consideration for training future translators that will have to compete with machine-generated output, but also to train MT systems that aim to emulate the “creativeness” of human translations. Ultimately, further knowledge on readers' narrative engagement and translators' creativity in relation to MT will help us understand how technology can be of benefit and where it needs improvement.

To pursue the project goals, I use the concepts of creative shifts in texts translated using MT and I triangulate this data with the readers’ narrative engagement and translation reception to understand better how creativity and technology in translation might influence readers' engagement and their enjoyment.

The following project objectives were defined accordingly:

• O1) To quantify creative shifts in a fictional text translated in different modalities (MT, MT post-editing and human translation)
• O2) To further develop a narrative engagement scale and a reception questionnaire for translated texts
• O3) To obtain quantitative data on the readers’ narrative engagement and reception of translated literary texts, and
• O4) To explore if there is a relationship between creative shifts and reader’s narrative engagement scores

Arbeit, die ab Beginn des Projekts bis zum Ende des durch den Bericht erfassten Berichtszeitraums geleistet wurde, und die wichtigsten bis dahin erzielten Ergebnisse

CREAMT is articulated in two main axes. The first axis proposes to identify creative shifts (O1).
The second axis seeks to identify reader’s narrative engagement and gather data on translation reception (O2 and O3). The two axes come together in a final analysis phase that compares their results (O4).

The project is divided in six work packages (WP) that include Mangagement (1), Acceptability and creative shifts (2),Questionnaires (3), Experiments (4), Dissemination and Engagement (5) and Training and knowledge transfer (6).

In addition to meeting the general and specific objectives of the research project, the implementation of the research project resulted in the achievement of the following milestones:

Milestone 1.1 was achieved with approval of the research by University of Groningen Ethics Committee

Milestone 2.1 was achieved with the completion of creative shifts index which meant delivering the translations and reviews into Catalan and Dutch, creative shift scores, acceptability report, and inter-annotator agreement between reviewers

Milestone 3.1 was achieved with the completion of Narrative and reception questionnaire in Catalan and Dutch from English

Milestone 4.1 was achieved with the participants’ results and the analysis of these results which meant uploading the questionnaires in Qualtrics, and creating a report describing the experiments.

Milestone 5.1 was achieved with an extensive activity to disseminate results on CREAMT and on collaborations with other researchers that exceeded the number initially plan except in Public engagement activities due to the lockdown imposed due to the COVID-19 pandemic. The following activities were completed or are in progress of completion: six talks at Scientific meetings, four public engagement activities, two non-academic articles submission, eight submissions and presentations to international conferences, there have been two articles and one submission pending approval directly related to CREAMT and one article and two chapters resulting from collaborations during this time. Also there are five chapters approved for submission.

Milestone 6.1. the completion of training activities, as well as supervision of students and teaching activities, and learning Dutch. The researcher supervised 1 PhD and 1 Master student in Translation Studies at University of Surrey (UK), 1 PhD at Universitat Autònoma de Barcelona and 2 Master students in Digital Humanities at University of Groningen from September 2020 to August 2022. As regards to training activities, I have attended and completed various training activities throughout the fellowship to improve my statistical skills, programming skills, as well as the Dutch language (I have reached level B2 level in the two year period). For my career development plan I have attended courses on grant proposal and interviewing. Since I submitted an ERC Consolidator Grant proposal in March 2022, I have passed Step 1 of the proposal, and I am doing an interview on November 28th.

Fortschritte, die über den aktuellen Stand der Technik hinausgehen und voraussichtliche potenzielle Auswirkungen (einschließlich der bis dato erzielten sozioökonomischen Auswirkungen und weiter gefassten gesellschaftlichen Auswirkungen des Projekts)

The CREAMT project has opened a new line of research that has never been explored before which is the intersection between creativity, translation and technology. The project has shed light on how MT impacts translators' creativity and the users reception of translated content, but also has given the scientific community instruments to measure creativity and users' reception by means of taxonomies and surveys, and the combination of these instruments.

On the one hand the project has measured creativity in translation and interview translators and found that translators' creativity is constraint when using MT as a default (that is that MT is presented as a proposal to the translators to post-edit) which points as the need to create alternative methods.

And on the other hand, the project has found that readers are less engaged, enjoyed less and think the translation is worse when reading MT, and higher when translators intervened (human translation and post-editing). However, the differences in the readers' reception between Catalan and Dutch readers suggest that when looking at creativity and translated texts there are other factors such as reading patterns, reading language and language status in a given country, and language, that impact also the readers' engagement with that text.

Further, the project has established that the element of engagement that readers find more difficult in MT is Narrative Understanding, i.e. to understand the thread of a story, but to a lesser extent Attentional Focus, Narrative presence, Emotional Engagement and Visual imaginary. These findings are novel in Translation Studies but also in Machine Translation studies where creativity has not really been dealt with and reception studies are truly scarce.

CREAMT presentation at Trinity College Dublin

Seminar at Leiden University

Article in Slator Magazine

Presentation at NeTTTT Conference