Corps de l’article

Computers and Translation is a collection of 17 articles edited by Harold Somers, a renowned scholar in the areas of machine translation and computational linguistics. Somers is Professor of Language Engineering in the School of Informatics at the University of Manchester. The articles are written by experienced professionals, including academics, employees of software companies, and translators, who all have a keen interest in the field of computers and translation. In his opening introduction, Somers delimits the scope of the book and clearly describes his target audience. He writes:

This book is, broadly speaking, and as the title suggests, about computers and translators. It is not, however, a Computer Science book, nor does it have much to say about Translation Theory. Rather it is a book for translators and other professional linguists (technical writers, bilingual secretaries, language teachers even), which aims at clarifying, explaining and exemplifying the impact that computers have had and are having on their profession. It is about Machine Translation (MT), but it is also about Computer-Aided (or -Assisted) Translation (CAT), computer-based resources for translators, the past, present and future of translation and the computer.

The articles can be grouped as having two main areas of focus. The first section, which includes the first seven chapters, looks at the various tasks a computer might perform, with a strong emphasis on CAT tools. Chapters eight to eighteen form the second section and focus on various aspects of MT. All the articles contain a relevant introduction and many include a historical overview to provide the reader with useful background information. Tables, charts, illustrations and screen captures of computer programs are often included to illustrate the examples provided. The articles are well cross-referenced, when applicable, and often contain information for further research. Reference lists at the end of each article are also very useful. A thorough index at the end of the book helps readers easily locate information within the articles.

The first section contains the book’s introduction as well as six logically organized articles on CAT. The second chapter, The translator’s workbench by Harold Somers, is particularly useful in giving an overview of technology in translation and providing an excellent lead into the topics subsequent chapters examine in greater detail. Somers provides a historical sketch and then discusses the current state, limitations, and future outlook of the main areas of computers in translation. He touches upon all major translation tools including word processing, dictation software, desktop publishing, localisation aids, online and electronic lexical resources, term banks, machine translation, and corpus-based resources such as translation memories and aligned parallel texts. Chapter four, titled Terminology tools for translators by Lynne Bowker, is an excellent resource for anyone interested in terminology. Bowker begins by giving a brief history of the use of computer tools in terminology, from the early 1960’s to current day applications. She defines a term record and termbase in the context of human information needs and contrasts these needs with the requirements of a computer that can not understand definitions, contextual examples, or learned usage. Terminology databases must therefore be prepared differently when used by a computer in coordination with other CAT tools. Bowker provides a very useful overview of what she calls the “new generation of terminology-management tools”. Pulling examples from CAT vendors such as Trados and Multicorpora, she outlines the features that one should consider when deciding on a specific program. She concludes the article with an introduction to term extraction tools, a concept that has been gaining popularity in recent years. Much like chapter four, the other articles in this section provide a detailed analysis of a specific CAT tool that was introduced by Somers earlier in the book. In addition to the chapters already mentioned, the remaining articles include information on translation memory systems, localization and corpora. The final chapter of the first section looks back at the tools discussed in the previous six chapters but within the context of “minority” languages. It quickly becomes evident that there is a significant gap in resources available to these languages when compared to languages that hold a more dominant position on the socio-economic scale.

The articles that constitute the second section are also thematically arranged starting with system-based descriptions of how MT systems work. The first two articles examine why translation is difficult for computers and the relevance of linguistics to machine translation. Understanding why machines have difficulty translating is essential when deciding whether and how to use MT in a given environment. John Hutchins, who has successfully collaborated with Harold Somers on several occasions, is the author of chapter ten titled Commercial Systems: The state of the art. This article is particularly useful in helping the reader obtain an overview of the MT systems that exist to meet three types of translation demand: dissemination, assimilation, and interchange. Hutchins explains that these demands serve different purposes and can therefore accept varying degrees of translation quality. For example, texts for dissemination require the highest level of quality and must often be revised before and after computer processing. Translations for assimilation can accept a lower level of quality because they are meant for people wanting quick access to basic information rather than having nothing at all. This area has seen tremendous growth in recent years. Translation for interchange meets the needs of people requiring instant translation even in rough form especially for the purpose of electronic communication such as email, chats and other Internet-based communication tools. The differences between translation for dissemination, assimilation and interchange have been outlined before in other sources and are again in later chapters of the book, but they are essential to understanding the information needs people have and how MT can contribute. MT should therefore not been seen in simple terms of “successful” and “unsuccessful” based on a few translated sentences, but rather in context of the end user’s information needs. Hutchins continues by explaining the types of systems that exist for various users from large corporations to independent professional translators to non-professional home users. He cross-references his research with previous articles in the book by explaining how MT has been integrated with other tools such as memories and localisation aids. In addition to articles that focus on MT systems, several chapters look at how texts can be altered so they work better with automated systems. Controlled language for authoring and translation, for example, is the title of chapter 14 which examines how restrictions on natural language, specifically constraints on lexicon, grammar and style, can reduce or eliminate ambiguities difficult for a computer to process. The authors, Eric Nyberg, Teruko Mitamura and Willem-Olaf Huijsen, pull specific examples from MT at Caterpillar Inc. to explain how controlled languages can bring about more favourable translation results. Chapter 15 titled Sublanguage by Harold Somers has a similar focus. Somers explains that sublanguage, a natural form of controlled language, appears in specific domains such as recipes, medicine and stock-market reports. To illustrate his argument he refers to Météo, the well-known Canadian MT system that has been successfully translating weather bulletins and reports for decades. The final chapter, titled Machine translation in the classroom again by Harold Somers, is the only article in the book devoted entirely to computers and translator training.

Although the title of the book is quite vague, Computers and Translation does an excellent job at exploring the most important aspects of translation technology without giving too much detail about any one system. It serves as an excellent resource to academics as well as translation and language professionals at large.