Abstract

There is no doubt that we live in an age of very rapid technological advances. The impact of technology is widespread, evident, and far-reaching in translation and interpreting, and in this respect, Translation, Interpreting and Technological Change: Innovations in Research, Practice and Training, edited by Winters, Deane-Cox, and Böser, is a timely volume that probes the evolving intersection between an age-old linguistic activity and modern-day emerging technologies. This edited volume offers a critical, user-oriented, and comprehensive discussion on the human–technology relationship at a time when almost every aspect of human activity is being impacted by technological developments.
Such impact has not occurred without controversy or concern, and this seems especially true in the context of translation and interpreting. Therefore, the need for such a volume is apparent. The relationship between the human and the technological, as well as between two important scholarly areas that intersect in this regard–translation studies and machine translation–has, over the decades, “oscillated between hostility and enthusiasm, indifference and acceptance” (Kenny, 2020, p. 305). Today, the technologies used in translation and interpreting are often praised for their benefits in multiple respects (see, for example, Cronin, 2013), but some of them (e.g., machine translation) have also been described as “implicated in the declining fortunes of many freelance human translators” (Kenny, 2019, p. 428; see also Moorkens, 2017). The human–technology relationship is complex, and this volume reminds us that technology usage is not free from risk and can have negative consequences, which include reduced pay rates, requirements to use specific software, increasing tool complexity, and unsatisfactory product design that neglects translators’ needs.
The editors of this volume are right in stating that the translation and interpreting community is confronted with the challenge of “constructive engagement with the conceptualization and realization of a social practice that may become increasingly technology-mediated or assisted,” which, they stress, “must not be driven by either technology or a primacy of cost-saving concerns” (p. 3). The volume places particular emphasis on the human side of these technologies, adopting a user-driven perspective and aiming at the empowerment of the user rather than merely discussing the technologies themselves. This offers several potential benefits, including “a better understanding of the varied and complex labours of the translator and other stakeholders” (p. 5). Indeed, a user-driven perspective might bring the public perception of technologies closer to reality, “offset any further devaluation of human translation” (p. 5), and offer “a counterpoint to the productivity-oriented agendas” (p. 4).
I have enjoyed reading the chapters of the volume, which are well organised and consistent with the editors’ emphasis on the human side of technology. Each chapter is authored by leading scholars in the field, with both empirical depth and theoretical rigour. The book is divided into three parts to explore technological innovations and their implications, reflecting the multifaceted nature of technological impact on translation and interpreting. The first part focuses on human-technology interaction, addressing the change in cognitive boundaries between human agents and technological tools, as well as the changing scope of technology use in translation and interpreting practice. Part II provides insight into the innovations in methods and models in technology-oriented translation and interpreting studies research. In part III, the focus turns to pedagogy, exploring how translation training is impacted in an era of ubiquitous technological use.
Rogl and Risku’s opening chapter offers a conceptualisation of technological tools as cognitive artefacts and socio-cognitive boundary objects. The chapter draws on a longitudinal study to examine how the tools mediate workflow in translation project management and facilitate collaboration. Their overarching theoretical framework is based on situated approaches to translation and cognition, complemented by the concept of boundary objects. While the longitudinal study is extensive, spanning 12 years from 2002 to 2014, it does not reflect the most recent practice. However, the stated purpose of the chapter is not to identify new technological trends but rather to explore “the social context of technological change in greater depth” (pp. 19–20). In so doing, the authors open new lines of inquiry, helping to emphasise the collaborative role of technology in the processes of translation project management.
In Chapter 2, the focus turns to the cognitive load of computer-assisted interpreting (CAI). Defrancq, Snoeck, and Fantinuoli present an experimental study where the performance and cognitive load of simultaneous interpreting are examined with respect to the use of automatic speech recognition (ASR) in the interpreting booth. Their aim is twofold: to assess the quality of interpretations, in terms of accuracy and acceptability, that are produced with and without ASR, and to investigate a new way of measuring cognitive load based on the pitch of the interpretation (i.e., ‘fundamental frequency’). The results are not entirely positive, however. The use of ASR did not show much impact on interpreting quality, nor did it affect the fundamental frequency associated with the cognitively challenging task of interpreting numbers. Interestingly, the chapter claims that “on balance,” ASR positively affects performance, and that “the cognitive downsides and upsides of ASR-based support seem to cancel each other out” (p. 54). These claims, of course, would need further evidence from systematic, more tightly controlled experiments. As the authors point out, their study is ‘exploratory’ rather than explanatory. Nevertheless, the chapter provides valuable insights into the methodological approach to quantifying the cognitive load of interpreting while highlighting a crucial, yet often neglected, aspect of translation and interpreting technologies, namely that technological augmentation must be balanced with human cognitive constraints.
Another exciting area in which the boundaries between humans and technology have been shifting quickly is literary translation. This is reflected in Chapter 3 by Kenny and Winters. Once considered far beyond the reach of MT, literary translation is now “seen as a valid object of study in MT circles” (p. 61), and the authors describe literary MT as an emerging interdisciplinary field, embracing a range of phenomena associated with applying MT to the translation of literary texts. In this chapter, various approaches to literary MT are mapped, before the authors discuss the concepts of customisation, personalisation, and style and make a case for adopting a literary studies approach to the issue of style in MT. This literary studies approach clarifies how ‘style’ should be conceptually understood in MT and how it can be empirically operationalised via digital means. Importantly, it resolves the problematic treatment of ‘literary texts’ in MT research as well as the vagueness around the term ‘domain’ in the treatment of ‘style.’ The authors also integrate causality into the discussion, hypothesising that translation, translators, MT, post-editing, and post-editors can all cause stylistic phenomena. On this basis, the post-editing of literary text is characterised as a type of downstream translator-specific personalisation, where, through the editing of the machine-translated text, “a literary translator is shown to assert their own style” (p. 69). The way in which the concept of style (as well as domain) is problematised, critiqued, clarified, and operationalised is comprehensive and succinct, and is a much-needed contribution to MT research. The case of translator-specific personalisation through post-editing is also a complement to the approach to style in literary studies and translation studies, in view of the recent technological influence.
The focus on literary translation continues in Chapter 4 where Hansen discusses the use of corpora as part of computer-assisted literary translation (CALT) and literary machine translation. The study challenges the assumption that literature is entirely incompatible with translation technologies, while seeking ways of empowering literary translators through technology. The author describes the process of constructing a parallel corpus for use in literary translation, analyses users’ opinions on CAT tools in literary translation, and illustrates quality-assessment results from an NMT model which has been fine-tuned with literary texts. The perspective from which the chapter engages with technology is broader than the traditional view, in that the discussion goes well beyond the function of translation memories and encompasses translation technologies as potential tools for text analysis and corpus exploration.
The two chapters in Part II offer a framework for assessing and managing risk in relation to the use of MT (Chapter 5), and a quality assessment of two generations of MT for legal translation (Chapter 6). While the previous chapters describe a somewhat promising future for the technologies used in translation and interpreting, these chapters highlight the importance of examining the consequences of using MT output, the ways in which technologies can be applied more appropriately, and how risks can be better managed. This part, on new models of technology use, presents a refreshing perspective, emphasising that raw MT should not be divorced from a human-centric view.
On this point, Koponen and Nurminen (Chapter 5) investigate various considerations for assessing and managing risks when delivering information to end-users, where the use of MT should be balanced against the potential consequences of not translating the information at all. They argue that applying the risk management framework defined by the International Organisation for Standardization (ISO 31000) would be helpful for the identification, evaluation, and treatment of potential risks in this context, although their approach focuses more on “initiating a conversation than offering final solutions” (p. 132). Indeed, in view of today’s pervasive use of texts produced by large language models (LLMs), such a conversation is not only necessary, but also a significant step towards the practical and ethical use of technologies in translation and interpreting.
Killman’s (Chapter 6) comparative evaluation of statistical and neural MT in legal contexts provides new and perhaps unexpected insight into the performance of the two approaches. This chapter shows that NMT, while impressive in many aspects, may not be as capable as its predecessor when it comes to terminological accuracy—an important concern in legal translation. In addition, NMT seems to struggle more than SMT in handling terminologies that are context sensitive. Unlike Chapter 4, this chapter does not involve such metrics as BLEU, ChrF, or TER but, instead, focuses on a human evaluation of terminological accuracy. The author acknowledges that sentence-level advantage of NMT is beneficial for the translation of legal documents, while highlighting terminological issues as “key areas of challenge facing artificial intelligence approaches to language translation” (p. 150). In this regard, the study could perhaps benefit from incorporating analyses on more recent artificial intelligence approaches, especially LLMs or frameworks such as generative pre-trained transformers (GPTs). Nevertheless, the discussion on terminological accuracy in legal translation is insightful, and such human evaluation would be crucial if new technologies (e.g., AI) are to play an integral role in the workflows of legal translation.
Part III turns to pedagogy, in view of the fact that technologies are beginning to play “an increasingly central, salient and embedded role” (p. 7) in language services and training.
Al Sharou (Chapter 7) pushes the boundaries of traditional pedagogy by urging educators to move from MT use towards MT creation, that is, towards the building of customised MT engines. The chapter advocates a training module based on such open-source software as Moses, run in a Linux environment, where students would be able to acquire more sophisticated computational skills, e.g., using the command line, completing natural language processing tasks, and eventually building their own engines. This is in line with suggestions given by many other scholars that such training should empower translators to avoid passively accepting what technologies offer and to develop capabilities to move beyond the limitations of “after-the-event” roles (e.g., post-editing). In this way, translators’ competence in translation technology would be expanded and their roles would shift from being mere users to creators of MT engines. This is an important notion, especially given the pace at which MT paradigms are changing and the need for translators to be able to transition to new technologies in the future.
In Chapter 8, Bowker demonstrates another pedagogical shift, namely a move towards the active teaching of MT literacy beyond translation and language classes (i.e., for those without a background in translation). The chapter offers a case study in a Canadian university where a course on translation, including machine translation literacy, is provided for non-translation students, enabling them to become critical, informed, and responsible users of MT. The underpinning of MT literacy in higher education beyond translation programmes points to a direction with significant implications for curriculum design, and Bowker’s description of the module with respect to its content, evaluation, and feedback would certainly benefit the community of translation educators worldwide.
Overall, the book is enjoyable, informative, and inspiring. The contributors draw from different areas to paint a multi-dimensional portrait of the technological transformation that is taking place in translation and interpreting. The emphasis on a pro-human and pro-social understanding of the human–technology relationship, and on the notion that technologies will be increasingly salient and embedded, is a central contribution. This is particularly valuable in a field that has long wrestled with changing and sometimes conflicting attitudes towards technology. The book also juxtaposes empirical studies with conceptual frameworks, creating an enriching dialogue between data and theory.
While this edited volume succeeds in its ambition of outlining from different angles the technological change in research, practice, and training of translation and interpreting, a few limitations remain. For instance, most of the chapters focus on MT only, whereas translation and interpreting technologies are not confined to MT at all. The types of technologies discussed in the book could have been more diverse, and although the theme of the book is on both translation and interpreting, there is only one chapter specifically dedicated to interpreting, in which the technology analysed is limited to automatic speech recognition. Readers who are interested in interpreting studies might expect a better balance between translation and interpreting, and perhaps more types of interpreting technologies other than ASR. In addition, issues surrounding bias and ethical concerns in data-driven translation systems would also benefit from greater attention.
Despite these limitations, Translation, Interpreting and Technological Change: Innovations in Research, Practice and Training is a timely and necessary contribution to both scholarly and professional discourses around language technology. Its integrated treatment of theory, practice, and pedagogy in view of the technological turn makes it essential reading for researchers, educators, and practitioners. As the field continues to grapple with AI-driven change, this volume provides a foundational roadmap towards our understanding of the human–technology relationship, and a call for critical, ethically aware engagement.
