Nitzke, Jean & Hansen-Schirra, Silvia. A short guide to postediting. Berlin: Language Science Press, 2021, 94 p.

Nitzke, Jean; Hansen-Schirra, Silvia. A short guide to postediting. Berlin: Language Science Press, 2021. 94 p.

Over the past decade, the uptake of machine translation has become increasingly prevalent across the translation profession. Indeed, language technologies continue to advance at a rapid pace, with corresponding implications for the wider industry as a whole. One of the major recent changes has been the rise of post-editing, which in simple terms involves human translators revising a machine-translated text. Though this practice is becoming more and more common, to date there have been comparatively few resources that professional translators and translation students can consult for advice and guidance, especially as machine translation-related technologies become ever more sophisticated. Accordingly, the volume reviewed here is both a highly relevant and much-needed contribution.

A short guide to post-editing is the sixteenth volume in the Language Science Press series ‘Translation and Multilingual Natural Language Processing’, a collection which has included monographs and edited volumes relating to this highly topical area of research. Unlike the others in the series, however, this book is highly practical in its aim and focuses squarely on the needs of professional and putative translators. It is co-authored by two experts in the field: Jean Nitzke, an associate professor of translation with a focus in translation technology at the University of Agder (Norway), and Silvia Hansen-Schirra, a full professor of English linguistics and translation studies at the Germersheim campus of Johannes Gutenberg University in Mainz (Germany). As the authors highlight on the work’s acknowledgements page, the book represents the distillation of relevant research and practice in response to genuine academic and practical demand. Indeed, this clear focus is evident on every page of this highly readable and well-crafted volume, which is divided into ten concise and informative chapters.

The work opens with some brief general remarks regarding artificial intelligence and the advent of machine translation, together with its corresponding effect on the wider translation market. By attracting the reader’s attention through mentioning the commonly-held opinion – surely familiar to every translator and translation student – that the translation profession will surely disappear in the next few years due to technological advances, the authors cite sources both in favour of and against this viewpoint. However, they remain balanced on the issue and instead use it as a point of departure to introduce the rationale underpinning the topic under discussion, noting that machine-translated output is generally not perfect and thus requires post-editing by human translators.

As with subsequent chapters, Chapter 2 opens with a series of learning objectives. It also invites the reader to ask themselves certain questions about post-editing and to bear these in mind as they progress through the book. In the first instance, taking the definition of post-editing as highlighted by O’Brien (2011), this section delineates the structure of the guide as a whole. This includes demystifying relevant machine-translation-related terminology and drawing attention to the differences, which are explained more fully in subsequent chapters, between various aspects of post-editing processes: e.g. between light and full post-editing, as well as between monolingual and bilingual post-editing, and the contrast between post-editing and interactive machine translation. Questions regarding who should perform post-editing, when and where it may be required, as well as the wider role of machine translation within the translation profession are also briefly discussed. Additionally, relevant scholarly research on the issue is highlighted.

The third chapter focuses on the historical development of machine translation, and opens with a chronological table depicting the major events. This includes the first patents for so-called ‘translation machines’ (1933); the seminal Weaver Memorandum of 1949, generally held to be the founding document of machine translation; as well as information about relevant technological developments during the Cold War and the invention of SYSTRAN and its adoption in the relevant directorates of the European Union’s institutions. In the current millennium, milestones featured include the development of Google Translate, as well as the adoption and implementation of neural technologies such as DeepL. Building on the historical overview, the three main architectures (rule-based, statistical, and neural) underpinning machine translation are outlined and defined, with relevant scholarly sources cited for further information.

Chapter 4 tackles the basic but fundamental issue of “how to post-edit.” Accordingly, the guidelines for light and full post-editing from the Translation Automation User Society (TAUS), an independent organisation, are provided, thus offering a useful and practical framework. Subsequently, the main points of ISO 18587, the corresponding international standard for post-editing, are highlighted through a concise and well-thought-out synthesis of the necessary information.

Moving onwards, the fifth chapter focuses on text types and their varying suitability for machine translation. In providing a brief overview, it notes that highly-restricted texts (technical manuals are given as a good example) with simplified language usually obtain the best results, whilst currently acknowledging that machine translation may struggle with literary works such as poetry or texts from domains such as advertising and marketing. In terms of general guidelines (and foreshadowing the content of the subsequent chapter), it notes that, if use of a translation memory would generally not be thought suitable for a given text (for example, in the case of literary translation), then machine translation might not be the best option.

Developing this notion, Chapter 6 examines the interplay between post-editing and various translation software tools. It outlines the strengths and weaknesses of using translation memory systems, before drawing attention to how such systems are integrated with machine translation. Then, definitions of interactive and adaptive machine translation systems are offered; in brief terms, the former attempts to use auto-complete to anticipate the user’s edits; whereas the latter waits for the user’s edits to be made before the system is adapted. As with previous chapters, pointers to academic literature on the topic are again given.

In common with many other sectors, risk management is of increasing importance across the translation industry. Chapter 7 highlights some potential risks of post-editing, with particular regard to the data security of machine translation-related processes, and also offers guidance in how to approach this sensitive issue. This is done by providing an overview of possible risks, ranging from data breaches to quality issues, whilst underscoring that it is generally the client’s decision whether the benefits of using machine translation – and by extension, post-editing – outweigh any of the latent risks.

The eighth chapter of the book explores the role of post-editing within the wider translation workflow. Aside from the fundamental question as to whether machine translation and post-editing are suitable in a given project, the chapter draws attention to other relevant aspects such as consideration of the text type to be translated, issues of data security, as well as the quality of the machine-translated output provided. The concept of pre-editing is also briefly introduced, which as the name suggests, involves tweaking the source text before the machine translation system is utilised. In addition, issues regarding the potential identification and rectification of errors in the source text are also highlighted. Other relevant factors are also mentioned; for example, the issue of how speedily the text needs to be translated/post-edited, as well as – highly relevant in this digital age – the consideration of how permanent the translated text will be and if it is likely to be superseded at a later date. This information is also helpfully depicted in graphical form, which provides a convenient visual guide to the various stages of the workflow.

Thecprevious chapters in a succinct manner. Interestingly, it also acknowledges information that could have been included in the book but ultimately was not selected. For example, mindful of the multifaceted and often personal nature of the issue, this includes details of pricing for post-editing work, an issue which can sometimes prove to be complex. In addition, noting that it remains an under-researched area, it is also observed that the ethical aspects of machine translation have not been discussed, as well as the use of automatic translators on websites where no post-editing has been done. Finally, the authors also invite the reader to return to the questions that they posed to themselves at the beginning of chapter 2, thus encouraging self-reflection and analysis on the insights gained in the guide.

In concluding this overview of the work, it is clear that Nitzke and Hansen-Schirra’s A short guide to post-editing has numerous strengths. Recognising its primarily practice-based orientation, the volume comes in at just under 100 pages in total, thus making it eminently suitable for busy modern schedules. However, it is important to underline that this brevity does not compromise the high quality of the information contained. An illustration of this can be seen in the work’s reference section, which runs to eight pages and lists several dozen entries, a quantity which testifies to the amount of research and preparation involved. Given that many professional translators and translation students often come from an academic background in languages, humanities, and the social sciences, the book succeeds in providing a clear and concise overview of machine translation for the non-specialist in a highly effective manner.

It is also noteworthy that the work does not shy away from presenting several of the key debates regarding the growing uptake of machine translation in the wider translation industry. Through indicating apposite sources of relevant scholarly information, this allows the reader to search for further insights on these topics, if so wished. In addition, by explaining the relevant technical and practical approaches in straightforward language, complex concepts are able to be digested easily. This is done not only by starting each chapter with a list of learning objectives; each chapter also ends with a crossword puzzle where the solutions are based on the content of the preceding text – in short, a simple but effective method of promoting the learning process and the retention of the necessary information in a fun yet durable way.

Another key attribute of the book is the fact that it is aimed at a broad audience. In maintaining this generalist approach, the work therefore transcends any specific language pair or domain of translatorial activity. This guarantees that the book has a wide appeal; indeed, its clarity, conciseness, and straightforward approach mean that it is highly relevant not only for students of undergraduate and postgraduate translation degrees, but also for professional translators who trained and began their careers before the advent of the post-editing era.

Over the coming years, it is certain that the relevant language technologies will continue to advance and thus prove ever more influential for translators and their professional reality. In short, this pioneering book represents an excellent and meritorious introduction to the practice of post-editing. In addition, by making the electronic version of the work Open Access and thus available to a wider public, the authors and publishers have ensured that A short guide to post-editing will have a significant impact on current and future generations of translation students and professionals across the globe.

Referências

• Nitzke, Jean & Hansen-Schirra, Silvia. A short guide to post-editing Berlin: Language Science Press, 2021. DOI: https://doi.org/10.5281/zenodo.5646896
  » https://doi.org/10.5281/zenodo.5646896
• O’Brien, Sharon. “Towards predicting post-editing productivity”. Machine Translation, 25(3), p. 197–215, 2011. DOI: https://doi.org/10.1007/s10590-011-9096-7
  » https://doi.org/10.1007/s10590-011-9096-7

Publication Dates

• Publication in this collection
  22 Sept 2023
• Date of issue
  2023

History

• Received
  29 Aug 2022
• Accepted
  15 Nov 2022
• Published
  Jan 2023