Modelling Parameters of Cognitive Effort in Translation Production (Memento)

Over the last few decades Translation Process Research (TPR) has been prolific in the generation of hypothesis and models which, among other issues, are all concerned with three interconnected questions:

  1. How translations are represented in the mind and what a cognitive architecture could be;
  2. What kinds of typical translation phenomena are present in the translation product;
  3. How translations are produced.

Existing translation models are either purely theoretical, or address only two of the above-mentioned research questions empirically. Addressing these questions in an international coordinated and collaborative context, the project seeks to integrate evidence from translation product research and translation process research, and to develop an integrated cognitive model of the translating mind based on empirical evidence, which can be validated in a broad range of language combinations.

Besides corpus-based research, keylogging, eye-tracking and fNIRS technologies are used to systematically investigate the relation between translation universals and translation process patterns. It is expected that closing the gap between corpus-linguistics and behavioral observations will ground a comprehensive theoretical framework of the bilingual mind in empirical observations

Consortium members:

  • Prof. Victoria Lai Cheng LEI, University of Macau (UM), China
  • Prof. Defeng LI, University of Macau (UM), China
  • Prof. Fabio ALVES, Universidade Federal de Minas Gerais (UFMG), Brazil
  • Prof. Michael CARL, Copenhagen Business School/Renmin University of China (RUC)
  • Prof. Silvia HANSEN-SCHIRRA, Johannes Gutenberg-Universit├Ąt Mainz (FTSK), Germany
  • Dr. Moritz SCHAEFFER, Johannes Gutenberg-Universit├Ąt Mainz (FTSK), Germany

Boot-camps on empirical Translation Process Research (TPR)

  1. Boot-camp 2018 at Macau University / China
  2. Boot-camp 2019 at Kent State University, Ohio / USA

The MEMENTO project anticipates three boot-camps, organized by different partners around the world. The boot-camps aim to investigate and develop new ways of assessing, analyzing and measuring translation process data and to explore possibilities to integrate translation product and process research in novel ways. A focus is put on the conceptualization of translation processes with a view to investigate what metrics we need to describe the various translation sub-processes based on translation process data. The boot-camps are of interest for advanced students, PhD students, post-docs and senior researchers, covering different fields of expertise, including translation studies, psychology and psycho-linguists, as well as computer science students with an interest in machine learning, data analytics and artificial intelligence. TPR data will be collected and legacy data will be made available to implement and test novel approaches for measuring the translation process in different modes of translation, including written translation, post-editing, sight translation and interpreting. Participants will form small ad-hoc working-groups to brain-storm questions, including the following:

  • Reading
    • How can we isolate and compare gaze paths in the translation data?
    • How can eye movement data be used to infer higher level coherence processes during translation/post-editing beyond the sentence level?
    • Which factors determine the transition from target text reading to source text reading in different translation modes?
  • Writing
    • How can we classify and formalize patterns of translation production?
    • What is the behavioral correlate of paraphrasing as opposed to a stricter, literal translation?
    • What is the impact of the translation brief on translation and writing behavior?
  • Speaking
    • Which dependent variables can be used in order to compare spoken language translation (sight translation and interpreting) with written translation?
    • How can the segmentation of the source text during sight translation and simultaneous interpreting be predicted on the basis of the properties of the source text?
    • How can the transition from reading ST items to speaking a translation during sight translation be predicted on the basis of process and product data?
  • Product and process
    • What are the typical patterns of gaze/typing interaction?
    • How can the cognitive translation unit be inferred from both keyboard and eye movement data?
    • How can reading and writing patterns be identified which are indicative of problem identification and problem solution during translation/post-editing?
    • Which factors determine the transition from source or target text reading to target text typing during translation?
  • Effort
    • How can cognitive effort be predicted on the basis of the produced final target texts?
    • How do known aspects of the source and target texts that have an effect on cognitive effort during translation affect the cognitive translation unit?
    • How are traces of translation universals realized during the translation process? For example, are there typical behavioral patterns for simplification, generalization, conventionalization, etc.?
    • How can the raw MT output be annotated to predict cognitive post-editing effort?
  • Strategies
    • Which aspects of the behavioural data can predict the need to interact with external resources such as dictionaries?
    • What kind of monitoring processes can we distinguish in the translation process and the translation process data?
    • How can the role of monitoring processes during the different modes of translation be compared and investigated?
    • What is the role of working memory in the different modes of translation?

The boot-camp working groups will meet on a daily basis to report/discuss/exchange their progress and develop new questions and approaches to integrate translation product and process data. The outcome of the boot camp will be a set of new metrics for segmenting, measuring and quantifying the process data, which are suited to assess the translation process in more detail and under new angles. The boot-camps will generate new research methods and themes suited for MA and PhD theses.

The boot camps are intended to be mostly a bottom-up workshop. Participants are expected to attend the entire boot camp duration of four weeks. Senior researchers will supervise the students and working groups, discuss technical problems and collaborate to direct the various sub-projects.