“With its clear explanations, engaging style, and insightful case studies, María

Fernández-​Parra’s practical guide to terminology management is set to become the

go-​to source for both student and professional translators.”
Dorothy Kenny, Dublin City University, Ireland

“This book is an invaluable resource for translators, offering up-​to-​date insights into
key concepts in terminology management. It goes beyond tools, focusing on practical
examples from high-​profile organizations and language service providers. An essential
read for professionals, researchers, and students alike.”
Míriam Buendía-​Castro, University of Granada, Spain

“If you are looking for an accessible, comprehensive, and example-​rich guide to ter-
minology management, then this is the book for you. Some books on this topic can
be highly abstract. This book is more practical. It still presents complex theoretical
knowledge, but it supports this presentation with concrete examples and cases in an
effective way. It covers all you need to know to get a solid understanding of termin-
ology management, and it should inspire some students, researchers, or professionals
to want to find out more.”
Patrick Cadwell, Dublin City University, Ireland

“This essential volume in the Routledge Series introduces readers to the pivotal role
of terminology and terminology management in translation practice. Blending theor-
etical insights with practical case studies, it offers a comprehensive exploration of key
concepts, research applications, and practical tools. Each chapter features engaging
tasks and reflective exercises, making it a valuable resource for both academic study
and practical skill development in translation.”
Koen Kerremans, Vrije Universiteit Brussel, Belgium
Terminology Management for Translators

Professional translators are the crucial link between businesses and their success in
international markets. The content they convey across languages very often includes
specialised terminology. But how much exactly should translators know about ter-
minology and its management and how should they deal with it? Can translators
write like experts without being experts? What tools and technologies can assist
translators in the management of terminology? This book provides the results of new
research in this area in order to address these questions in detail.
For professionals, this book provides a fresh view on practices in terminology man-
agement based on empirical data. For translator trainers, it provides a thorough over-
view of the aspects of terminological theories that are useful for translators, shows
how the theory can underpin the design and creation of terminological databases for
translators and offers a wide range of ideas for discussion, exercises and examples.
For students and trainee translators, it explains why they should bother with termin-
ology management at all and offers guidance on acquiring, storing and using their
terminological resources efficiently, dispelling common misconceptions about terms
and terminological work.
With clear explanations and summaries, activities, discussion points, further
reading and links to resources, this book is the ultimate introductory text and ref-
erence resource, covering everything translators ever wanted to know about the ter-
minological side of their work.

María Fernández-​Parra is an Associate Professor in Translation at Swansea University,

where she is the Director of the MA Professional Translation and MA Translation and
Interpreting. She has co-​authored Translation Tools and Technologies, also in this
series, with Andrew Rothwell, Joss Moorkens, Joanna Drugan and Frank Austermühl.
She is also a Chartered Linguist at the Chartered Institute of Linguists (CIOL), an
Academic Member of the Institute of Translation and Interpreting (ITI) and a Senior
Fellow of the Higher Education Academy (SFHEA). She is a past Secretary of Swansea
University’s Language Research Centre (LRC), a past Treasurer of APTIS (Association
of Programmes in Translation and Interpreting in the UK and Ireland) and is currently
the Director of STING, the Swansea Translation and INterpreting Research Group
at Swansea University.
Routledge Introductions to Translation and Interpreting
Series Editor:
Sergey Tyulenev is Professor of Translation Studies and Director of
Postgraduate Research at the School of Modern Languages and Cultures,
Durham University, UK.

Advisory Board
Luise von Flotow, University of Ottawa, Canada
Ricardo Munoz Martin, University of Bologna, Italy
Kobus Marais, University of the Free State, South Africa
Nike K. Pokorn, University of Ljubljana, Slovenia
James St André, Chinese University of Hong Kong, China
Michaela Wolf, University of Graz, Austria

Routledge Introductions to Translation and Interpreting is a series of

textbooks, designed to meet the need for teaching materials for translator/​
interpreter training. Accessible and aimed at beginning students but also
useful for instructors designing and teaching courses, the series covers a
broad range of topics, many of which are already core courses while others
cover new directions of translator/​interpreter teaching.
The series reflects the standards of the translator/​ interpreter training
and professional practice set out by national and international competence
frameworks and codes of translation/​ language service provision and are
aimed at a global readership.
All topics combine both practical and theoretical aspects so as to ensure
a bridging of the gap between the academic and professional world and all
titles include a range of pedagogical support: activities, case studies etc.

Most recent titles in the series:

Translation and Community

Mustapha Taibi

Audiovisual Translation
Patrick Zabalbeascoa

Terminology Management for Translators

María Fernández-​Parra

For more information on any of these and other titles, or to order, please go to www.routle​dge.
com/​Routle​dge-​Introd​ucti​ons-​to-​Tran​slat​ion-​and-​Inter​pret​ing/​book-​ser​ies/​RITI

Additional resources for Translation and Interpreting Studies are available on the Routledge
Translation Studies Portal: http://​routle​dget​rans​lati​onst​udie​spor​tal.com/​
Terminology Management
for Translators

María Fernández-​Parra
Designed cover image: Getty Images
First published 2026
by Routledge
4 Park Square, Milton Park, Abingdon, Oxon OX14 4RN
and by Routledge
605 Third Avenue, New York, NY 10158
Routledge is an imprint of the Taylor & Francis Group, an informa business
© 2026 María Fernández-​Parra
The right of María Fernández-​Parra to be identified as author of this work
has been asserted in accordance with sections 77 and 78 of the Copyright,
Designs and Patents Act 1988.
All rights reserved. No part of this book may be reprinted or reproduced or utilised
in any form or by any electronic, mechanical, or other means, now known or
hereafter invented, including photocopying and recording, or in any information
storage or retrieval system, without permission in writing from the publishers.
Trademark notice: Product or corporate names may be trademarks or registered trademarks,
and are used only for identification and explanation without intent to infringe.
British Library Cataloguing-​in-​Publication Data
A catalogue record for this book is available from the British Library
ISBN: 978-​1-​032-​29906-​8 (hbk)
ISBN: 978-​1-​032-​29909-​9 (pbk)
ISBN: 978-​1-​003-​30263-​6 (ebk)
DOI: 10.4324/​9781003302636
Typeset in Sabon
by Newgen Publishing UK
Access the Support Material: www.routle​dget​rans​lati​onst​udie​spor​tal.com
Contents

List of Figures and Tables ix

Series Editor’s Foreword xii
Acknowledgements xiv
Acronyms xvi
About This Book xvii

PART I
Theories, Research, Applications, Integrations 1

1 
Translators and Terminology Management 3

2 
Terms, Concepts, Domains 25

3 
Key Terminological Theories and Approaches 40

4 
Designing Your Termbase: Principles and Considerations 62

5 
Populating and Using Your Termbase 87

6 
Tools for Terminology Management 118

PART II
Case Studies: Introduction to the Case Studies 163

7 
Governmental Case Study: The Translation Bureau 167

8 
Large International Organisation Case Study:
The World Bank 182
viii Contents

9 
Case Studies on Commercial Organisations 198

10 
Terminological Best Practices and Quality Management 231

Bibliography 245
Appendix 285
Index 287
Figures and Tables

Figures
1.1 Partially overlapping disciplines that include the study of terms 8
1.2 Simplified process and tasks in terminology management 20
2.1 Continuum of specialisation of terms  29
2.2 Terms as triple relations (left) and photosynthesis as an
example (right)  32
2.3 Snippet of the conceptual network of mineralogy 33
2.4 A portion of the conceptual network of financial markets 37
2.5 The semantic triangle 39
3.1 A (broadly) chronological overview of terminological theories
since the 1930s 41
3.2 “Percepts” in the Cultural Terminology theory  54
4.1 Individual entries in a dictionary (left) and a concept-​oriented
entry in a termbase (right)  64
4.2 Example of monolingual term entry for refractive index 68
5.1 Extract illustrating examples of how terms are highlighted
in texts 103
5.2 Example of an MS Word Dictionary export with RWS
MultiTerm of a Catalan-​English termbase 109
6.1 Overview of software tools for terminological work 119
6.2 Some of the advanced search settings in Google 122
6.3 An example of a focused search on Google 123
6.4 Example of KWIC display for the search term bioluminescent
with AntConc 131
6.5 Example of dispersion plot in AntConc 135
6.6 Example of bilingual concordance using AntPConc 137
6.7 Example of bilingual concordance using Sketch Engine 137
6.8 Screenshots from Sketch Engine showing Word Sketches
for light as a noun (top) and verb (bottom) from a
bioluminescence corpus 139
6.9 Example of Sketch Engine’s Word Sketch Difference for
creature and organism in a bioluminescence corpus 141
x List of Figures and Tables

6.10 Example of the term clustering and concept creating

features in Fodina’s TermCatch 144
6.11 Typical workflow with TermCatch 146
6.12 Example of concept map for the term “distribution
agreement”  148
6.13 Visualising different kinds of term relations in a
concept map 149
6.14 An example of a satellite map for scuba diving  151
6.15 Two types of concept visualisation in EcoLexicon: the Tree
mode (left) and the Path mode (right)  152
6.16 Example of concept map with Kalcium QuickTerm by the
Kaleidoscope Group 153
6.17 Example of cluster graph feature in TermCatch to visualise
partial synonyms  154
6.18 A basic concept map 160
7.1 Organisational chart of the Translation Bureau  170
7.2 Translators’ offices at the Translation Bureau,
Gatineau, Quebec 170
7.3 Simplified translation workflow at the Translation Bureau 173
7.4 Descriptive fields and hierarchy in TERMIUM® (left) and
example of a TERMIUM® entry in English,
shortened (right)  177
8.1 The World Bank building in I Street, Washington DC 187
8.2 Generic translation workflow at the World Bank Group 188
8.3 Descriptive fields and hierarchy in WB-​TERM records (left)
and a schematic example of a term record (right) 190
8.4 An excerpt of the Arabic team’s phraseological database at
the Bank 192
9.1 Blum’s plant in Höchst, Vorarlberg, Austria  201
9.2 Two of Blum’s products: Its first product, a horseshoe stud
(left) and the award-​winning BLUMOTION hinge for
smooth opening and closing of furniture (right)  201
9.3 A snapshot of page 38 of the 2022–23 Blum catalogue
in English 202
9.4 The Kaleidoscope Group’s headquarters in Vienna, Austria 204
9.5 Generic translation workflow at the Kaleidoscope Group 205
9.6 Blum’s ‘up and over lift system’ (left) and ‘lift up system’
(right) for wall cabinets 208
9.7 An overview of the features of the various editions
of Quickterm  210
9.8 Typical dashboard interface in Kalcium 211
9.9 Terminology interface in Quickterm 212
9.10 Main Checkterm interface, showing the results of a
Document analysis task 215
 List of Figures and Tables xi

9.11 Descriptive fields and hierarchy in a typical Blum termbase

in MultiTerm 218
9.12 Generic translation and localisation workflow at Terra
Translations 223
9.13 Screenshot of Terra’s termbase in memoQ 225

Tables
1.1 Translators vs Terminologists 16
2.1 General overview of differences between terms and words 28
4.1 Three examples of data category sets for termbases 70
8.1 The five institutions that constitute the World Bank Group.
The IBRD and IDA are the ones that comprise what is
commonly known as the “World Bank”  185
10.1 Summary of lessons learned from the Case Studies 232
Series Editor’s Foreword

The Routledge Introductions to Translation and Interpreting series offers

textbooks covering a range of essential topics and skills already taught, to a
greater or lesser degree, in translator and interpreter training programmes,
such as Translation Ethics, Audiovisual Translation, Translation Tools and
Technologies, etc. This textbook continues the line that fulfils one of its more
ambitious objectives –​to introduce new courses and subject areas focused on
emerging skills that have become or are becoming essential in today’s trans-
lation and interpreting industry, yet which so far are hardly represented in
translator/​interpreter training curricula. Examples include Translation Project
Management, Localization in Translation and Translation and Community.
The overarching aim of such textbooks is to reflect the contemporary realities
of the profession, capturing its ever-​growing diversity of developments and
practices and preparing younger generations of translators and interpreters
for what they will face after graduation.
Traditionally, translators and interpreters have been seen as facilitators of
linguo-​cultural transfer, converting text from a source language to a target
language. However, today’s professional landscape demands much more than
mere text conversion; it requires proficiency in a range of supplementary
skills. One critical area is terminology management, which is the focus of
this textbook.
In modern practice, terminology management is an indispensable skill.
Translators and interpreters often encounter highly specialised fields with
technical or socio-​cultural terms that may not exist or be widely understood
in the target language. While professional terminologists specialise in cre-
ating, defining and managing such terms, translators and interpreters may
not always have the luxury of consulting a trained terminologist. Therefore,
they must be equipped to handle both new and existing terms to ensure
clarity and precision in their translations.
The author of this textbook, Dr María Fernández-​Parra, brings exten-
sive experience in both researching and teaching terminology management
in the context of translation. Her background in training translators and
interpreters makes her a reliable guide for navigating this complex yet crucial
 Series Editor’s Foreword xiii

aspect of translation and interpreting. This textbook equips readers with

both theoretical knowledge and practical tools that can be applied in real-​
world scenarios. Theoretical knowledge anchors the reader’s understanding
of what terminology is and how a translator/​interpreter should approach
this daunting subject area. Practical exercises are intended to teach how to
apply knowledge to tasks at hand. Thus, this book is certain to lay a sound
foundation for anyone interested in terminology from the translation and
interpreting perspective.
Sergey Tyulenev
November, 2024
Acknowledgements

This book would not have been possible without the kind help of a long
list of people and organisations, far too many to name them all here, some
having helped directly and others indirectly. I am therefore immensely
grateful to everyone who has helped bring this book to life in one way or
another. Heartfelt thanks to you all!
Of course, a big part of this book has been the Case Studies, and I am
deeply appreciative to all the organisations and language services providers,
their leaders and their teams, for their incredibly generous support throughout
the research, data collection and writing process. Whether the Case Studies
appear in this book or on the online portal, I should like to thank every
colleague at the Translation Bureau at the Government of Canada; Osama
Abdel-​Hameed and Maria Isabel Swann and their teams at the World Bank;
Iman El Zayat and all the language teams at the International Monetary Fund;
Carlos Llull, Hélène Deguil and everyone at the World Trade Organization;
Klaus Fleischmann, Anthony Dunn and the teams at the Kaleidoscope,
Eurocom UK and Blum in Austria; and Marina Ilari and everyone at Terra
Translations. I am very grateful to you all for taking the time to contribute
to this book, allowing me to meet with so many colleagues and have such
interesting discussions, and for your input in the editing and writing process.
This research was made possible by two rounds of Impact Acceleration
Account funding from the Arts and Humanities Research Council (AHRC)
at Swansea University. Thank you for believing in the value of this research
and for your trust in its potential impact. Your support has been instrumental
and I am truly appreciative of the opportunities it has provided.
My warmest thanks also go to Prof Sergey Tyulenev, the series editor,
for giving me the opportunity to contribute to this incredibly useful series,
but also for your guidance and encouragement at all the crucial times. Your
support has been greatly appreciated. I also owe a debt of gratitude to the
entire Routledge team for your dedicated support, attention to detail and
professionalism throughout.
There are other professionals, translators, translation studies colleagues,
scholars and friends I would like to thank for their support and feedback,
 Acknowledgements xv

for opening doors to great opportunities and for helping refine in so many
ways the approach followed in this book. A big thank you to Dr JC Penet,
Prof Fernando Prieto Ramos, Dr Kara Warburton, Prof Maeve Olohan,
Barbara Inge Karsch, Prof Francis Hult, Veronika DeMichelis and Ben Karl
and Thomas Bunstead.
I am equally very grateful to my generous colleagues in Modern Languages,
Translation and Interpreting at Swansea University, and other colleagues at
the university, for their amazing help in many and varied ways: many thanks
indeed to those who took on extra responsibilities to give me time to travel,
to write, and to those who provided me with very insightful feedback, excel-
lent advice and administrative support throughout the entire project. Special
thanks to Prof Andrew Rothwell and Dr Lloyd Davies for your guidance,
expertise and ideas for each chapter. Thank you all for being such supportive
colleagues and for taking the time to contribute your input, but also for
making my research so much more productive and enjoyable.
Finally, I really want to thank Clive, my family and my friends for their
constant, unwavering support and encouragement, especially when it was
most needed, but also for your patience and reassurance. Thank you very
much for getting me going when the going got tough and for giving me the
final push in the completion of this book.
Despite the incredible help I have received, the views expressed in this
book are my own and should not be attributed to any of the people or any
of the organisations mentioned in it. Any omissions, errors or inaccuracies
remain entirely my own.
Acronyms

AI artificial intelligence
BNC British National Corpus
CAT computer-​assisted translation
DGT European Commission’s Directorate-​General for
Translation
DTP Desktop Publishing
GTT General Theory of Terminology
IMF International Monetary Fund
LGP language for general purposes
LLM large language model
LSP language service provider
language for special purposes
MT machine translation
NMT neural machine translation
SEO search engine optimisation
SL source language
ST source text
TB Translation Bureau, Government of Canada
TL target language
TMS terminology management system
translation management system
TT target text
UI user interface
UK United Kingdom
US United States
WB World Bank
About This Book

Professional translators are the crucial link between businesses and their
success in international markets. The content they convey across languages
very often includes specialised terminology. But how much exactly should
translators know about terminology and its management and how should
they deal with it? Can translators write like experts without being experts?
What tools and technologies can assist translators in the management of ter-
minology? This book provides the results of new research in this area in
order to address these questions in detail.
For professionals, this book provides a fresh view on practices in ter-
minology management based on empirical data. For translator trainers, it
provides a thorough overview of the aspects of terminological theories that
are useful for translators, shows how the theory can underpin the design and
creation of terminological databases for translators and offers a wide range
of ideas for discussion, exercises and examples. For students and trainee
translators, it explains why they should bother with terminology manage-
ment at all and offers guidance on acquiring, storing and using their termino-
logical resources efficiently, dispelling common misconceptions about terms
and terminological work. Students and trainee translators can also use this
book for independent study. For researchers, this book provides a systematic
framework for creating and maintaining terminological databases and cor-
pora. On the whole, whether students, professionals or researchers, this book
is the ultimate resource, covering everything translators ever wanted to know
about the terminological side of their work.
After an introduction to the development of key terminological theories,
Part I focuses on highlighting and explaining how the main principles of those
theories can be usefully applied by translators to their terminological tasks.
Then, in Part II, a series of case studies on both high-​profile organisations,
but also language service providers, demonstrates how those principles are
currently put into practice in an increasingly digital and globalised world,
picking out a range of methods and procedures to manage terminological
quality in translation. Finally, a thorough overview analysis of the technolo-
gies available to support the translators’ domain research work, as well as
xviii About This Book

of their terminological work, is complemented by practical explanations

and exercises, and suggestions for further reading, both in the book and on
the online portal. The final chapter looks ahead to foreseeable changes in
the terminological roles of, and opportunities for, translators in managing
terminological quality, as the profile of the professional translator evolves
in line with new technologies and developments. No prior technical know-
ledge is assumed on the part of the reader, as Terminology Management for
Translators is intended as a reference work for professionals, researchers,
trainers and trainees alike.
It is important to stress that the views expressed in this book are my own
and should not be attributed to any of the people or any of the organisations
mentioned in it. Any omissions, errors or inaccuracies remain entirely
my own.

Part I: Theoretical Considerations

Part I is designed to be a valuable resource for students, translator trainers,
researchers and practising translators to gain or deepen their understanding
of terminology theories useful for professional translators today, the key
principles on which terminology management is based and their place in
current and likely future translation workflows. New research is carried out
in order to present a consolidated view from the relevant literature and con-
tribute to the creation of an emergent body of knowledge specifically focused
on the terminological work of translators. This research is novel in that it
places professional translators at the centre of considerations about their ter-
minological work and presents a complementary view to other approaches
that focus on the work of terminologists and how it can support translators.
In addition to outlining the historical background to the development of
major terminological theories, the book sets out to offer comprehensive and
accessible explanations of key terminological concepts and principles, and
how to apply them in practice, thoroughly illustrated with examples in a
range of languages, and supported by task-​oriented materials that can be
used for self-​training, continuing professional development and/​or in trans-
lator training classrooms. Usual conventions are followed, such as presenting
concepts in italics and backtranslations in single inverted commas.
Chapters 1 to 3 introduce terminology management, the key principles on
which it is based and on reviewing various aspects of terminological theories,
which will then underpin the more practical guidelines for terminological
work discussed in Chapters 4, 5 and 6. Chapter 1 explains why termin­­­­
ology management is so crucial in translation work, and therefore essential
in companies and organisations, and the main tasks and processes that ter-
minology management consists of. Chapter 2 reviews key notions in termin­
ology management such as terms, concepts and domains and explains why
they can be useful to translators. Having set out the notional foundations in
these chapters, Chapter 3 explores key terminological theories and how they
 About This Book xix

inform the terminological work of translators. In Chapter 4, we apply the

key principles in terminology management to the design of terminological
databases, i.e. termbases, whereas in Chapter 5 we focus on making use of
the termbases, by filling them with data and using them to translate with
computer-​assisted (CAT) tools. Chapter 6 presents a range of terminology
management tools available to translators, from term extraction tools and
concordancing tools to concept visualisation tools and others. Rather than
providing user guides for the tools, Chapter 6 focuses on explaining the
key concepts underlying the design and use of the tools, in order to raise
awareness among translators of the benefits and pitfalls of each category of
tool, so that they can make an informed use of them.
At the end of each chapter, there are some suggested activities and reflec-
tion tasks, as well as some selected references for further reading. Additional
activities and resources can be found on the online portal for this book, as
well as more detailed guidance on how to use some of the tools for the ter-
minological work of translators and links to user guides provided by tool
developers themselves. Making these user guides, or links to them, available
on the online portal means that they can be updated as new software versions
come out. Focusing on a conceptual overview of tools, it also helps to min-
imise the aspects of the book that can become outdated given the rapid pace
of technological change.

Part II: Case Studies

In Part II, we head out to several translation workplaces to see how terminology
management theories, principles and practical advice are implemented. The
Case Studies range from governmental translation settings (the Translation
Bureau of the Government of Canada, Chapter 7) and large institutional
settings (World Bank, Chapter 8) to commercial settings (the Kaleidoscope
Group and its client Blum, and Terra Translations, Chapter 9), all strongly
underpinned by the work of freelance translators. Finally, in Chapter 10, we
review what we can learn from the Part II case studies and conclude with
further insights into managing quality and sharing best practices in the ter-
minology work of translators. Part II is also complemented by follow-​up
tasks and reflection activities, as well as some suggestions for further reading.
Additional case studies will also be added to the online portal for this book,
along with further suggested activities and resources.

Online Portal
In addition to the complementary material for Part I and Part II, the online
portal provides other resources for translator trainers. For example, trainers
will find suggestions on how to teach the contents of this book in their
courses, and suggestions on how to spread out the contents over courses that
last from 10, 12 or 14 weeks to an entire academic year, as well as suggestions
newgenprepdf

xx About This Book

for differentiating the content between undergraduate and postgraduate

courses. Trainers will also find a multiple-​choice quiz for each chapter, which
may be adapted for use in tools such as Kahoot, aimed at a more inter-
active engagement of students with the contents of the book. In addition,
other resources will be added, such as MS PowerPoint presentations and
links to online resources, articles, etc. Alternatively, translator trainers may
use these suggestions for inspiration and adapt them to their specific needs
and requirements.
Part I

Theories, Research,
Applications, Integrations


1 Translators and Terminology
Management

Key Questions and Issues

• What is terminology management and why is it important for

translators?
• What are the differences between translators and terminologists?
• How does terminology management work?
• Why a systematic approach to terminology management is a smart
investment for translators (and interpreters).

Terminology and Terminology Management

Terms are everywhere. As new knowledge is being acquired through
inventions and discoveries, the volume of highly technical words and phrases
used to refer to new concepts is also increasing, and sometimes entering our
everyday language (Brekke, 2004: 627). One relatively recent example is that
of severe acute respiratory syndrome coronavirus 2 or SARS-​CoV-​2, in the
wake of the pandemic declared in 2020. The disease caused by this virus is
more commonly known as COVID-​19. COVID-​19 may be a common term
now but it is safe to assume that, before 2020, only certain medical specialists
would have known about the coronavirus family of viruses. But medicine
is not the only subject area in which new concepts and terms are emerging.
New techniques and inventions have been developed in almost every area of
knowledge, for example:

• Physics, e.g. scientific energy breakeven, a method to initiate nuclear

fusion in a laboratory (U.S. Department of Energy, 2022).
• Astronomy, e.g. the James Webb Space Telescope or JWST.

DOI: 10.4324/9781003302636-2
4 Terminology Management for Translators

• Nanoengineering, e.g. AGILE (Axially Graded Index Lens), a device

that removes the need for solar panels to track the Sun (Vaidya &
Solgaard, 2022).

Discoveries have also been made in:

• Biophotonics, e.g. eyeshine reflector, tiny spheres in the eyes of some

crustaceans that reflect light in specific ways (Shavit et al., 2023).
• Astronomy, e.g. oval (triaxial) disk-​shaped galaxies, a discovery made
with JWST (Pandya et al., 2024).
• Palaeontology, e.g. Tyrannosaurus mcraeensis,1 a species that predated
Tyrannosaurus rex (Dalman et al., 2024).

Many more advances in knowledge are made in a wide range of fields

and indeed new subject areas are also emerging, e.g. neuroparasitology
(Hughes & Libersat, 2018) or exometeorology (Dabu, 2017). This raises
two important questions from the point of view of terminology and trans-
lation: how do professional translators manage to understand what these
new concepts are, so that they can translate them into other languages?
Once they find the equivalent term in other languages, if it exists, how do
they store this information for future use, to avoid repeating the research
previously undertaken? In order to answer the first question, we will take
a quick look at what terminology is and what translators can learn from
it (see also Box 1.1). To answer the second question, we will look at
what terminology management is and how it can be best carried out by
translators.
Terminology has been defined in the literature in various and sometimes
overlapping ways (see Box 1.2), but a definition that fits our purposes is the
“discipline concerned with the collection, processing, description and pres-
entation of terms, which are lexical items belonging to a specialised field
(e.g. medicine, law, engineering, library science or art history)” (Bowker,
2015: 304). As well as terms, terminology in this sense is also concerned
with the technical concepts the terms designate. We will take a bilingual or
multilingual approach by looking at how terms and concepts in one lan-
guage map on to those of another language or languages. For example, we
will be interested in the fact that bond and maturity are terms in finance,
because they have specific meanings in this specialised2 subject area (and
different meanings in general language), and in how translators would go
about determining the equivalent terms and concepts in other languages.
Since the phrase subject area can have various possible synonyms, e.g. field,
discipline, branch of knowledge, etc., we will use domain from now on to
avoid confusion.
 Translators and Terminology Management 5

Box 1.1 Demystifying Terminology Management

According to Muegge, the following are three frequent issues that come
up in conversation with translation professionals:

1. I know I should be managing terminology.

2. But I don’t really understand why.
3. And even if I understood why, I wouldn’t know how.

Regarding point 1, it is undeniable that translators should be managing

the terminology they work with, to avoid the detrimental effects of not
managing it. More importantly, translators should develop smart and
efficient methods to do so. This is one of the main premises on which
this book is based.
Regarding point 2, Chapter 1 is an attempt to explain why termin­
ology management is so crucial, not only for translators themselves,
but ultimately for the clients (e.g. experts, businesses, companies,
organisations, etc.) they work with. Terminology management, whether
monolingual, bilingual or multilingual is central to the translators’ role
as disseminators of (specialised) knowledge around the world.
Regarding point 3, all the other chapters in Part I of this book are
devoted to explaining the “how” of terminology management. In Part
II, we show examples of “how” terminology is managed in practice in
a range of translators’ workplaces.

Source: Based on Muegge (2020).

The importance of terminology in our digital and global societies should

not be underestimated. Although it may seem as though some organisations
or businesses do not understand this importance, as we explore later in this
chapter, several authors have stressed its significance. For example, Champagne
highlights its economic value and describes it as part of an organisation’s
intellectual capital and a “major asset” (2004: 9). Pinchuck claims it is
“indispensable for the functioning of our modern society” (1977: 13) and,
for Sager, it is “vital to the functioning of all sciences” (1990: 2).
Another key consideration regarding terminology, on which we expand
in Chapter 3, is that terms are not static or monolithic items. Terminology,
when referring to the specialised vocabulary of a domain, is dynamic; it is
constantly changing and developing. Advances in knowledge may change
our understanding of concepts, which in turn may lead to changes in terms,
the words or phrases used to refer to those concepts. An example quoted
in the literature is that of the concept of planet. Until 2006, this concept
included Pluto. However, the discovery of exoplanets, confirmed in 1992,
6 Terminology Management for Translators

made astronomers re-​think the definition of planet entirely and eventually

produced a new definition for this concept, as specified in the International
Astronomical Union’s Resolution of 2006 (IAU, 2006); see also IAU (n.d.).
This new definition of planet excluded Pluto, which became a dwarf planet
instead (see also Delavigne & Gaudin, 2022: 185; ten Hacken, 2015: 4).
Further, new concepts may emerge with different designations, i.e. terms.
For example, the concept of “All the rights, duties, powers, responsibilities
and authority which by law a parent of a child has in relation to the child
and his property” is referred to as parental responsibility in some UK legal
systems but as guardianship in the Irish legal system (Court of Justice of the
European Union, 2024). However, parental responsibility has two meanings
in European Union law (Court of Justice of the European Union, 2020). The
examples of planet (terms are not static) and guardianship/​parental respon-
sibility (terms are not monolithic) show the extent to which getting termin-
ology right is important and how crucial translators are in the dissemination
of the right content in the right context.

Box 1.2 The Polysemy of Terminology

The word terminology has been used in the literature with at least three
different meanings:

1. Terminology as a discipline or theory: the study of terms as specialised

units of knowledge.
2. Terminology as a practice, methodology or activity: a “set of
principles that govern term compilation” (Cabré, 1996: 16).
3. Terminology as a product or vocabulary of a specialised domain: the
list of terms from a domain which have been compiled according to
the practice of terminology.

According to Faber and L’Homme, terminology was recognised as an

independent discipline in the first half of the 20th century (Faber &
L’Homme, 2022a: 1). However, it is rather ironic that, as Austermühl
(2012: 62) puts it, a theory that “occupies itself with precision and
the avoidance of ambiguity” should have three different meanings for
the concept of terminology. In Chapter 3, we will mostly be refer­
ring to the first meaning; in Chapters 4 and 5, we will be mostly using
meanings 2 and 3; in Chapter 6, the second meaning will often be
referred to. In Part II, we will use mostly meanings 2 and 3 but we will
occasionally need to refer back to meaning 1.

It is also worth noting that terminology as a discipline is “not an end

in itself” but is “at the service of science, technology and communication”
(Cabré, 1999: 9). This means that neither specialists nor translators develop
 Translators and Terminology Management 7

specialised terms and concepts for the sake of it: there is always a need behind
terminological work. Whether specialists or translators, the need is almost
always the “transfer of specialized knowledge” (Pavel & Nolet, 2001: 8) and,
therefore, the key aims of translation-​oriented terminological work always
revolve around consistency and accuracy (Prieto Ramos, 2020c), in order to
ensure that the transfer of knowledge occurs as efficiently as possible.
Several groups of users are involved in the transfer of knowledge.
Translators and terminologists are of course two of the main groups of
users of terminology (e.g. Bowker, 2022: 131; Sager, 1990: 45; ten Hacken,
2006: 11; Våge, 2018: 177), but there are many more, which include benefi­­
ciaries, not only users, of terminology. According to Sager (1990: 197) and
Schmitz (1996: 230), the following professionals will use and/​or benefit from
terminology at some point: subject specialists, lexicographers, journalists,
language policy planners, publishers, language teachers, researchers, among
others, including even general users. Cerrella Bauer (2015: 334) adds
marketing and public relations staff into the mix, as well as product man-
agement staff, software development staff and a few more. Nuopponen
(2018: 10) includes machine translation and other information systems as
non-​human users of terminology. For all these professionals (and machines)
to be able to function appropriately, the terminology they use will ideally be
managed by terminologists (as curators of that knowledge) and translators
(as disseminators of the knowledge). The work involved in managing termin-
ology is generally referred to as terminology management.
Terminology management will be used here, therefore, to refer to a subset of
terminology work which is about finding, “storing, retrieving and displaying
(validated) terms and associated data (definitions, synonyms, usage samples,
etc.)” (Kenny, 2011: 463). Since the focus of this book is translators, ter­
minology management will include all the tasks necessary for translators to
find terms in one or more languages, and any associated data about them
for the purpose of systematic storage and organisation in a database for
future use. However, it may sometimes be difficult to establish the bound-
aries between translators and terminologists (e.g. Cabré, 1999: 115) because
many translators carry out terminological work and many terminologists
work with terms in two or more languages. The core unit of work in termin-
ology management is the terminological entry (e.g. Cerrella Bauer, 2015: 335;
Schmitz, 1996: 224), which we will refer to as term record or term entry
(often shortened to entry). Term records are grouped together in termino-
logical databases. We will use the term termbase as a convenient (and widely
accepted) short form of terminological database.
Since terminology and terminology management are interested in the
treatment of terms, they partially overlap with other disciplines that are
interested in terms too, and the differences or boundaries between disciplines
may not be clear at first glance. In particular, there is a clear overlap with
lexicology, phraseology and lexicography, as shown in Figure 1.1. There are
differing voices in the literature as to which discipline is (or is not) a branch of
which other discipline, but contributing to this debate is beyond the scope of
8 Terminology Management for Translators

Figure 1.1 Partially overlapping disciplines that include the study of terms. The size
of the circles does not represent the size of the discipline.

this book. For the sake of simplicity, therefore, the main point of Figure 1.1
is that the four broad disciplines are represented with the terms appearing in
the centre being objects of study that they have in common.
Terminology management and terminography are shown as subsets of
terminology, with terminology management being “broader in scope than
terminography” (Warburton, 2021: 4). Despite some conflicting descriptions
and definitions in the literature, in this book we will use terminography to
refer essentially to “working with termbases” (2021: 4), whereas we will use
terminology management to refer to the body of knowledge and the practices
a translator needs in order to manage terminology, which includes working
with termbases but is not limited to it. Thus, Chapter 4 focuses on many
terminographic tasks but the entire book focuses on terminology management.
Returning to the four broad disciplines, they can be considered separately
because they differ in focus: lexicology is interested in all the words of the
lexicon (i.e. the collection of all the words that exist in a language), which
includes terms, as some terms are written as single words. However, termin-
ology is limited to the words that acquire a different meaning in a specialised
domain, i.e. terms (Warburton, 2021: 4–​5), which in fact can be single words
or multiword expressions. In this sense, terminology also overlaps with phrase-
ology, in that the latter studies how all the words of a language combine with
each other to form phrases (including idioms and fixed expressions), whereas
terminology is only interested in those combinations of words that acquire a
 Translators and Terminology Management 9

different meaning in a specialised domain, i.e. multiword terms. It is beyond

the scope of this book to explore the similarities and differences of phrase-
ology, lexicography and lexicology as branches of linguistics but interesting
insights can be drawn from various publications (e.g. Bowker, 2017; Granger
& Meunier, 2008; Kageura, 2015; Lewandowska-​Tomaszczyk, 2008). The
distinction between words and terms is at the heart of the differences between
lexicology and terminology, respectively, and we explore this distinction fur-
ther in Chapter 2.
The results of lexicological research can be collected in the form of dic-
tionaries (Riggs, 1989: 90) and the practice of preparing dictionaries is
known as lexicography (Warburton, 2021: 5). Lexicographers are interested
in recording all the words of a language and all of the meanings of each
word. Typically, words are presented alphabetically. In some dictionaries, we
may also see multiword expressions listed within the meanings of a particular
word, not only single words. For example, ice cream, dry ice or break the ice
may appear in the dictionary entry for the noun ice.
A fundamental difference between lexicography and terminology lies in the
storage and mode of storage of their object of study. Whereas lexicography
focuses on listing words alphabetically in dictionaries, terminology focuses
on storing terms in termbases. However, entries are organised in termbases
following a concept-​oriented method whereby the concept determines the
entries, which include all the possible terms (i.e. labels) to refer to that con-
cept. Thus, dry ice might appear in a dictionary entry for ice but, in a termbase
from the domain of chemistry, for example, dry ice and ice will be separate
entries because they correspond to different concepts. A termbase entry for
ice may also include the term CO2, as both CO2 and dry ice are different ways
(i.e. labels) to refer to the concept of “solid carbon dioxide.” In a termbase,
the semantic connection between dry ice and ice can be established by means
of cross-​references (see more below and Chapter 3). Thus, lexicography “has
the primary aim of helping readers to interpret texts, whereas terminology
aims to hep writers produce texts” (Riggs, 1989: 90).
Of course, there may also be specialised dictionaries that contain both CO2
and dry ice within the entry for ice, i.e. both general language expressions
and specialised terms, but the essential principle that termbases are gener-
ally organised around concepts (and the concepts are listed alphabetically,
of course), remains important for translators creating termbases. Concept
orientation is explored further in Chapter 4. Lexicology, phraseology and
lexicography are some of the closest disciplines to terminology, and of course,
terminology management, but terminology is also connected to other discip-
lines, such as information science, computer science, knowledge engineering,
computational linguistics, computer-​ assisted translation, machine trans-
lation and artificial intelligence. In practice, the boundaries between these
disciplines may not be as neat and clearly defined as in Figure 1.1. These
connections are also explored further in Chapter 3.
10 Terminology Management for Translators

Having reviewed what terminology and terminology management are,

and how they relate to each other, as well as to other disciplines, we will,
throughout this book, use terminology loosely for the more theoretical
discussions, whereas we will use terminology management for the more prac-
tical ones. Thus, Chapters 1 to 3 will be mostly focused on aspects of termin­
ology and terminological theories, as well as key principles in terminology
management, which will then underpin the more practical work discussed
in Chapters 4 and 5. While Chapter 4 focuses on the specifics of designing
termbases, Chapter 5 focuses on filling the termbase with data and using it
with computer-​assisted (CAT) tools during translation. Chapter 6 presents
an overview of terminology management tools available to translators, from
term extraction tools and concordancing tools to concept visualisation tools.
In Part II, we head out to see how these theories and practical advice are
implemented in several real workplaces, from governmental (Chapter 7) and
large institutional settings (Chapter 8) to commercial settings (Chapter 9),
all strongly underpinned by the work of freelance translators. Finally, in
Chapter 10, we review what we can learn from the Part II case studies and
conclude with further insights into managing quality in terminology work.

Why is Terminology Management Important for Translators

(and Interpreters)?
Terminology management is a “specialist task” (Sager, 1990: 11) for which
training is required, and an “integral part of the work of translators”
(Champagne, 2004: 5) (and of course, terminologists), which means that a
substantial part of the daily work of the translator will involve managing ter-
minology (Bowker & Marshman, 2011: 61), as we will see below. These are
already some important reasons why professional translators should develop
a smart approach to managing terminology early on, but there are many
other compelling reasons, which are grouped under the following headings.

Time

Managing terminology is a time-​consuming task (Bowker, 2019: 580) and

translators are often under intense time pressure to deliver their translations
(Marshman et al., 2012: 31). Accuracy and consistency are important aims in
terminology management, but it takes time (and an efficient method) in order
to achieve both. Although it is difficult to quantify how much time translators
spend on terminology, as terminological work is often subsumed under trans-
lation work and not paid separately from translation (Bowker, 2020: 164),
some studies provide figures to illustrate how much time translators devote
to terminology management tasks, for example:

• Austermühl (2001: 102) reports that searching for the correct term can
take up to 75% of the translator’s time.
 Translators and Terminology Management 11

• Nkwenti-​Azeh (2001) reports that the search for the correct equivalent in
a target language can take a translator 60% of their time.
• According to Champagne (2004: 30), “experienced translators spend
about 20% to 25% of their work on terminology activities,” whereas for
a new translator, “this percentage might be as high as 40% to 60%.”
• Gornostay (2010) reports that in a survey of 93 translation professionals
throughout Latvia, Estonia and Lithuania, 60% of respondents spend 20–​
30% of their time on terminology tasks.

One way for translators to overcome the laborious nature of terminological

work is to develop a specialism early on, if possible, and develop adequate
terminological resources in a domain or two, so that when they are assigned
a translation project, the terminology work required for it has already (or
mostly) been done. Warburton (2014: 6) demonstrates with a range of
scenarios how a reliable termbase can save valuable time in translation and
revision. Sometimes clients provide translators with a glossary or termbase,
which saves the translator some time (see also Box 1.3).

Box 1.3 Glossaries, Term Banks and Termbases

At first glance, it would seem that glossary, termbase and term bank are
synonymous but they are actually different concepts.3 Each of them can
be monolingual, bilingual or multilingual:

Glossaries: in this book, we will use glossary to refer to a two-​column

list of terms, i.e. the term and sometimes its definition, to which other
columns for other languages can be added (if bilingual or multilingual).
Although, in principle, terminology should be the client’s responsi-
bility, “it is the rare client who will take the trouble to compile and
provide bilingual glossary, especially in smaller projects” (Bononno,
2000: 655), unless they intend to standardise their own terminology
(Olohan, 2016: 42).

Term banks: a term bank, shortened from terminological data bank

(Nkwenti-​Azeh, 2001: 249), is a large-​scale collection of electronic
term records, typically developed and maintained by a large institution
(Bowker, 2015: 308). For this reason, they are often intended to be pre­
scriptive (see also Table 1.1) for the users of the institution. Translators
are frequent users of term banks (2015: 308) from which they can use
information for their own termbases. Many term banks are freely avail-
able online, e.g. IATE, UNTERM and TERMIUM®.
12 Terminology Management for Translators

Termbases: shortened from terminological database, termbases are “ter-

minological databases integrated into the CAT environment” (Rothwell
et al., 2023: 57), which allow the import and export of terminological
information through MS Excel files or other formats. As we will see in
Chapter 4, this integration is crucial in helping translators overcome
the time-​consuming nature of terminology management. Termbases are
one of the most frequently used tools by translators (Désilets et al.,
2009b). They are more robust and flexible than glossaries “because
you can store more data than in Excel and they allow you to more
easily view the concept, terms and metadata in a hierarchical structure”
(RWS, n.d.: 4).

Why should translators use a termbase?

• It is more reliable than their memory! (Olohan, 2016: 42).

• It saves translators from having to repeat term searches time and
again (Bowker, 2019: 580).
• It automatically looks up terms in the termbase during translation
(see automatic terminology lookup in Chapter 5).
• It is organised differently than a dictionary, especially general lan-
guage dictionaries, because it is organised by concept (for which there
may be one or more terms) rather than alphabetically by each single
word. The concept-​based approach helps translators collect and track
term variation, e.g. synonyms and partial synonyms (also known as
near synonyms), more easily and efficiently (Muegge, 2007: 17).

Cost

Time is money, and cost is another reason why translators should have an
appropriate plan in place to deal with specialised terminology. Cutting down
on the time spent on terminology management tasks is likely to lead to sacri-
ficing the quality of terminology in the translated text (e.g. failing to take into
account a client’s terminological preferences), which can “bring in a boatload
of unforeseen problems” (Argos Multilingual, 2021). Having said that, as
Bowker (2015: 305) points out, “it is not easy to establish a precise return
on investment (ROI) for managing terminology in the context of translation”
because “the terminology research conducted by a translator is not usually
broken down as a separate cost but is calculated as part of the overall trans-
lation process.” Specialisation might again be one solution for translators,
as specialised translators can “command higher rates of pay for their work”
(Olohan, 2016: 8). While the terminological work of a specialised translator
is not paid for directly by the client, the higher rate of pay can help translators
offset the cost of becoming specialised to some extent. This is because the
 Translators and Terminology Management 13

‘added value’ to the client is not just the terminology, but the greater under-
lying level of domain understanding, hence reliability. In this sense, one of
the key points throughout this book is that terminology research requires
domain research.

A Challenging Task

As well as being labour-​intensive, managing terminology often requires a

great deal of effort from translators, especially if they have not specialised
in the domain they are working on. A very inefficient practice would be for
translators to repeat the same terminological research every time a transla-
tion project is assigned to them (Bowker, 2019: 580). According to Dunne
(2013: 1), terminology has been the “primary challenge” for translators since
the 1960s. Désilets et al. (2009a) found, in their ethnographic study, that
45% of the problems encountered by translators (32 out of a total of 71
problems) were related to specialised terminology. Other studies found that
translators may also struggle with inconsistencies of content, neologisms,
abbreviations, collocations, geographical variants and phraseology (e.g.
Argos Multilingual, 2021; Candel Mora, 2017: 255; Varantola, 1992).
The good news for translators who invest in terminological work is that
their efforts will not go unrewarded. Importantly, Champagne (2004: 9)
highlights the reusable nature of term records in what he calls the multiplier
effect whereby, once created, “a record is used several times in its life cycle.”
The more a term record is used, the more its value increases. This benefit
extends to termbases, which Champagne (2004: 10) describes as “an invest­
ment with enormous multiplier effects” because they “let businesses develop
long-​term client-​supplier relationships, or make their products stand out in
the market.” Thus, in order to avoid duplication of efforts and time ineffi-
ciencies, translators should keep track of the terms and their associated infor-
mation (e.g. definition, examples of use, etc.) in termbases.

Translators Cannot be Experts in Every Domain

Even if translators develop a specialism or two, generally speaking they do not

have the same level of expert knowledge as specialists in the domain. It would
be unrealistic to expect a translator to be a “natural expert” in every domain,
given “the incredible amount of information available today” (Austermühl,
2001: 102). Even so, translators are expected to “create seamless texts in
which terms are used the same way as experts in the field would use them”
(Montero Martínez & Faber, 2011: 89). Even more so, assuming translators
are highly specialised in the domain they are translating, they can still come
across words and phrases in the source text which fall outside the domain
in question (Bononno, 2000: 655). Translators may also come across poorly
written source texts (e.g. Martínez et al., 2022; Mossop, 2015) for a variety
of reasons, sometimes due to the experts’ lack of writing or linguistic skills
14 Terminology Management for Translators

(Meyer, 2022: 121–​122) or publication pressures (Lindsay, 2020: 4), among

others. No level of specialisation is likely to help translators in these situ-
ations. Fortunately, Mossop (2015) suggests a method for identifying poorly
written texts and dealing with them and, if all else fails, translators should
aim to collect as much associated information (e.g. definition, context, notes,
etc.) about the terms in their termbases. A key takeaway here is that, in
fact, translators do not need to be as expert as specialists in the domain. As
pointed out by Montero Martínez and Faber (2011: 93), what “is essential
is that they [translators] be capable of rapidly acquiring expert knowledge”
and a well-​constructed and well-​managed termbase might just be the tool for
acquiring and storing such knowledge for future reference.

There are Consequences to Mismanaging Terminology

As with translation work, translators’ terminology work is “invisible” as

long as it remains error-​free (e.g. Cerrella Bauer, 2015: 326; Warburton,
2021: 53). However, errors may arise from poor terminology management
practices, e.g. inconsistent or erroneous use of terms, which may lead to
complex and expensive re-​working costs to fix the errors, high re-​publishing
costs, delayed time-​to-​market and other serious consequences that will ultim-
ately have a detrimental effect on the translator’s (or client’s) branding and
reputation (see Argos Multilingual, 2021). Warburton (2015a: 367) goes as
far as saying that the most frequent mistakes are terminology-​related but
specifies that “[l]‌anguage problems can lead to service calls, legal challenges
resulting from copyright or trademark infringement, and in the most severe
situations, injury or death due to product misuse” (2015a: 45). Examples of
terminology-​related errors abound in the literature:

• Incoherence “in the use of terminologies in South Korean government

translations” (Choi, 2022).
• “Reported failed knee operations caused by mistranslation of medical
terms” (Kageura & Marshman, 2019: 61).
• Unacceptability of the term Falkland Islands in Argentina (Kohlmeier,
2000: 1–​9).
• Confused meanings in fire terminology leading to mismanagement of fuels
(Jain, 2004).
• Misleading translations, e.g. organic food translated into Chinese as 有机
食品 (yŏu jī shípĭn), which means ‘food produced with machine or tech-
nology’ (Li & Hope, 2021a: 9).

It is therefore surprising to discover that some organisations and businesses

still treat terminology management as “an annoying factor” (LISA, 2005: 6),
“low key” (Champagne, 2004: 12), “marginal” (Austermühl & Mirwald,
2010: 3) or “hidden” (Bowker, 2015: 305); that many do not have a system­­
atic terminology management system in place (Li & Hope, 2021a: 14–​15;
 Translators and Terminology Management 15

Muegge, 2007: 17); and that others have no such system at all (Warburton,
2015a: 367). Among the reasons why some companies seem to neglect ter­
minology management aspects are, for example:

• Because they are confused as to what terminology management is or how

to do it.
• Because they consider it to be an additional, unnecessary cost.
• Because they are still sceptical about terminology management tools, often
due to “budget/​time constraints, considerable information duplication,
[and] inefficient functionality of existing tools” (Gornostay, 2010).

The message for translators (and interpreters) here is that neglecting termin-
ology management is a false economy as it will take more time, effort and
cost to fix the mistakes that could have been avoided in the first place. Having
a systematic and efficient terminology management system is a smart invest-
ment. If in doubt, translators should always remember that “good language
resonates with customers; bad language frustrates customers” (adapted from
Warburton, 2021: 45).

The Benefits of Managing Terminology Far Outweigh the Efforts to Manage It

On a more positive note, the list of benefits arising from effective termin-
ology management is longer than the list of negative consequences. Although
benefits may not be obvious “until a considerable amount of effort has been
invested in developing a term collection” (Cerrella Bauer, 2015: 340), a well-​
planned and methodical approach to terminology management:

• Avoids miscommunication among translators, company staff, clients and

stakeholders (Cerrella Bauer, 2015: 324; Jaekel, 2000: 159).
• Underpins information management processes (Cerrella Bauer, 2015: 324).
• Reduces overall time and effort (Argos Multilingual, 2021).
• Reduces the risk of error and leads to fewer commercial errors (Champagne,
2004: 29).
• Improves productivity (Champagne, 2004: 29).
• Improves training resources (Champagne, 2004: 29).
• Improves overall quality of the translated content (Champagne, 2004: 29).
• Reduces duplication of effort (Champagne, 2004: 29).
• Makes translation work easier (Champagne, 2004: 29).
• Helps with quality control (Champagne, 2004: 29).
• Improves consistency in documentation (Jaekel, 2000: 159).
• Improves customer satisfaction (Jaekel, 2000: 159).
• Reduces turnaround times (Jaekel, 2000: 159).
• Improves a company’s or a translator’s competitive edge (Champagne,
2004: 10).
• Reduces costs (Jaekel, 2000: 159; Keller, 2010: 2).
16 Terminology Management for Translators

This list is by no means exhaustive, but translators (and their clients) need
not wait for large improvements in their terminology management systems to
reap the benefits. As explained by Moravia (2017), “[e]‌ven a small improve­
ment made now will yield immediate and tangible results.” In addition,
“benefits will multiply when additional languages are added, and when new
content needs to be localized” (2017).

Isn’t a Translator Already a Terminologist?

Translation work and terminology work are closely related. As a result,
it is often assumed that being a translator includes being a terminologist,
but terminology work is different from translation work. That said, many
terminologists have a translation background anyway (Warburton, 2021: 46).
The main differences between the typical work carried out by terminologists
and the terminological work of translators can be summarised in Table 1.1.

Table 1.1 Translators vs Terminologists

ID Terminologists Translators

1 Main objective: to record every Main objective: to produce a target

concept (and the terms used to text.
refer to them) from an entire
conceptual system.
2 Systematic (or thematic) work on Ad hoc (or punctual) work on
the entire domain → Domain-​ terms and concepts found in the
oriented terminology work. source text only → Text-​oriented
terminology work.
3 Prescriptive (or normative) More a descriptive (or informative)
approach to terminology than a prescriptive approach to
(classical terminology) terminology.
and descriptive (modern
terminology).
4 Time-​consuming work. Typically, tight time constraints.
Terminology work should be carried
out in advance whenever possible.
5 Often work with subject specialists Encouraged to specialise in one or
(experts). two domains. Not always possible
or practical to have access to an
expert. Otherwise, become expert
enough to translate the source text.
6 Typically teamwork. Work alone, e.g. freelance translators,
or in teams, e.g. in institutional
translation settings.
7 Onomasiological work (classical Semasiological work.
terminology) and semasiological
work (modern terminology).

Source: Adapted from Tamás et al. (2016: 73), Thelen (2015: 353–​358) and Warburton
(2021: 4).
 Translators and Terminology Management 17

The first three rows of Table 1.1 show how the terminological work of
translators and terminologists differs in objective and approach (Candel
Mora, 2017: 247). Regarding the objective, translators aim to produce a
target text, whereas terminologists aim to collect information about the
entire conceptual network (i.e. ‘family of concepts’) of a domain. Therefore,
translators are generally focused on research into an individual term or con-
cept and its equivalent in the target language when this poses an “immediate
translation problem” (Bowker, 2015: 311). This approach of resolving ter­
minological problems as they arise is often referred to as ad hoc (i.e. non-​
systematic) by some authors (e.g. Ahmad & Rogers, 2001: 727; Cabré,
1999: 152; Warburton, 2015a: 375; Wright & Wright, 1997: 147) and punc­­­
tual (as a synonym) by others (e.g. Bowker, 2015: 311; Cabré, 2003: 175;
Warburton, 2021: 17).
By contrast, the typical work of terminologists is referred to as systematic
by some authors (e.g. Cabré, 2003: 175; Popiolek, 2015: 341; Warburton,
2015a: 375; Wright & Budin, 1997: 148) or thematic (as a synonym) by
others (e.g. Ahmad & Rogers, 2001: 727; Bowker, 2015: 311; Warburton,
2021: 17). Whereas systematic terminology management covers “the terms
of an entire special subject field or a subpart thereof” (Cabré, 1999: 129),
ad-​hoc terminology management is “restricted to a single term or a small
set of terms belonging to a subsection of a subject field, or to a group of
terms belonging to different fields” (1999: 129). Ad hoc research has also
been described as text-​driven or text-​oriented, i.e. selecting terminology from
specific texts belonging to a domain, whereas systematic research has been
described as domain-​oriented or subject-​field driven, i.e. selecting termin-
ology from the materials of a domain and taking time to organise them logic-
ally by concepts in termbases (Schmitz, 2009; Wright & Wright, 1997: 148).
Although ad hoc describes the typical terminology work of translators and
systematic that of terminologists, translators may also (occasionally) carry out
systematic research and terminologists ad hoc research. Somewhere between
ad hoc and systematic research, Cabré (1999: 152) identifies what she calls a
subject mini-​search, i.e. an ad hoc search that follows the same methodology
as systematic searches, although she does not provide any examples.
We can also look at terminological work through the standardisation
lens. Standardisation refers to the process of aiming to eliminate “ambi-
guity through the establishment and use of precisely defined concept-​term
units” (Bowker, 2019: 579) and therefore “foster unambiguous communi­
cation” (Faber & L’Homme, 2022a: 2) (see also Chapter 3). Thus, in row 3
of Table 1.1, prescriptive terminological work aims to control the usage of
terms by accepting and promoting some and rejecting and discouraging others
(Warburton, 2021: 18–​19). The standardising approach of prescriptive work
is thought to facilitate “efficient knowledge development” (Meyer, 2000: 40–​
41). Some authors refer to this approach as normative (e.g. Maslias, 2022;
Schmitz, 2009: slide 34; Varantola, 1992:124; Warburton, 2015a: 377). By
contrast, descriptive terminological work aims to collect and document as
18 Terminology Management for Translators

many terms as possible, explaining all the ways in which they are used, but
“avoids making recommendations about what constitutes proper, correct, or
even permitted term usage” (Warburton, 2021: 18). Some authors refer to
this approach as informative (Cabré, 1999).
Classical terminology (i.e. from 1930s or so, see Chapter 3) was gener­
ally regarded as more prescriptive (while lexicography was regarded as more
descriptive) (Candel, 2022: 38). Over time, though, and with the advent of
new theories of terminology which proposed perspectives such as the com-
municative (Cabré, 2009) and cognitive (e.g. Temmerman, 2000) (see also
Chapter 3) and the widespread use of corpora (i.e. a principled collection of
texts, see also Chapter 5), researchers felt a prescriptive approach to termin­
ology no longer took account of “language in use” (Pearson, 1998: 16), e.g.
it did not allow for term variation, and that this shortcoming limited “the
usefulness of the [terminological] data base” (Sager, 1990: 207). Today, both
approaches have their use in terminology. While the preferred method, by
terminologists and translators alike, tends to be the descriptive one, some
prescriptive terminology is also useful in specific cases, e.g. when translators
need to determine which of the various terms for a concept is the preferred
one in a specific context or for a particular client.
Row 4 of Table 1.1 shows how time affects the work of terminologists
and translators. Whereas terminologists can often spend more time than
translators to craft thorough terminological resources, translators are usu-
ally not afforded this luxury, as they generally work under very tight time
constraints. Therefore, it is highly advisable for translators to develop ter-
minological resources in advance of a translation project, whenever possible,
or to factor in the time required for terminological research into the quote
for their client (e.g. in the case of freelance translators), or to discuss the add-
itional time required with their colleagues or line manager (e.g. in the case of
in-​house translators). Specialising in a domain, as we have seen, could help
translators not only to have terminological resources ready when a job is
commissioned but also to have had time to understand the main issues and
concepts in a domain and how they relate to each other.
Another difference between the terminological work of translators and that
of terminologists may be seen in the access to subject specialists (experts) (see
row 5 of Table 1.1). Whereas terminologists are typically “highly dependent
on the input and collaboration of subject-​field specialists” (Alberts, 2001: 79;
Nuopponen, 2018: 13), this is often not the case for professional translators,
especially in freelance settings. Translators in institutional or governmental
workplaces may sometimes be able to access experts in-​house, but this is not
always the norm either. This is another reason why translators are encouraged
to develop specialism in a domain or two. If specialists are not available,
translators themselves may become expert enough that they can translate the
text at hand. In this sense, translators must work as “closet terminologists
and be capable of carrying out terminological management as a means of
knowledge acquisition” (Faber, 2009).
 Translators and Terminology Management 19

In row 6 of Table 1.1, another trend that differentiates terminologists from

translators, especially freelance ones, is that of teamwork vs lone working.
While terminologists often work in teams, freelance translators, by defin-
ition, work alone and “without the benefit of the revisions, advice, reference
materials, databases and support they would receive if they formed part of a
translation team in an international organization or large translation agency”
(Borja Albi, 2013: 54). Translators in large organisations or agencies tend to
carry out more teamwork than lone work, even if sometimes that work is
remote (Prieto Ramos, 2020a: 455) (see also Chapter 6).
In row 7 of Table 1.1, we find more of a similarity between terminologists
and translators than a difference, at least nowadays. A typical task for both
groups is finding a term in a text, figuring out what it means, recording it
in a termbase and working out what its equivalents are in other languages
(Warburton, 2021: 15–​16). This approach to terminological work, which
begins with terms and progresses to concepts, is referred to as semasio-
logical (from the Greek σημασία, ‘semasia,’ lit. ‘meaning,’ ‘signification’). For
example, the literature often quotes the work of lexicographers as semasio-
logical. Lexicographers work with words for which they aim to find all
their meanings to include in the dictionary (e.g. Béjoint, 1988: 357; Cabré,
1999: 7–​8; Sager, 1990: 56; Warburton, 2021: 14). Cabré describes this
approach as moving “from the word [term] to the concept” (1999: 7–​8).
The opposite approach consists of finding a concept, working out which
conceptual network it belongs to, and in which domain, and then establishing
a term to refer to it. This approach is called onomasiological (from the Greek
όνομα, ‘onoma,’ lit. ‘name’). For example, it is often explained in the lit-
erature that classical terminology begins with an onomasiological approach
(Delavigne & Gaudin, 2022: 186; Sager, 1990: 56; 2001b: 260) because one
of its main aims was to assign names (i.e. terms) to concepts. In the onomasio-
logical view, terms are considered to be members of a system of concepts, and
the meanings of terms are established by comparing them to the meanings
of other terms in the system (Warburton, 2021: 17). Cabré describes this
approach as moving “from the concept to the term” (Cabré, 1999: 7–​8).
In current practice, “both semasiological and onomasiological approaches
are applied and considered entirely compatible, since both methods have
to supplement each other in the terminological working process” (Laurén
& Picht, 2006: 176). Having said that, most terminologists and translators
generally follow the semasiological approach nowadays (Cabré, 1999: 7–​8;
Humbley, 2022: 31; Warburton, 2015a: 373), leaving the onomasiological
for occasional use, for reasons such as the following:

• The semasiological approach allows translators (and terminologists) to

look at terms in context and actual usage, e.g. by finding them in corpora.
Once the use of electronic corpora became widespread, the onomasio-
logical approach faced some criticism by not taking the context of terms
into account (Condamines & Picton, 2022: 221).
20 Terminology Management for Translators

• Due to the effort involved in onomasiological work, this approach is only

followed when justified; for example, it is “rarely relevant in fast-​paced
production environments” (Warburton, 2015a: 374).
• In reality, only scientists and translators follow a truly onomasiological
approach and they do so occasionally, e.g. when scientists need to name a
new invention, a new tool, a new measurement, etc. (see Sager, 1990: 56),
and when translators are the ones in charge of creating an equivalent name
in the target language where the concept may also be new (e.g. Valeontis
& Mantzari, 2006: 1) (see also Chapter 4).

To sum up, the two professions are closely linked and in practice the dis-
tinction between the two is becoming increasingly blurred, often driven by
technology (Bowker, 2015: 306). Before we explore in detail how technology
can help the terminological work of translators, we first consider the general
tasks involved in terminology management.

What Does Terminology Management Consist of?

Having reviewed various ways to classify and refer to terminological work,
which leads to similarities and differences in the terminological work of
translators and terminologists, in this section we explore the typical process
and tasks in terminology management from the point of view of a translator.
The process and tasks are illustrated in simple terms in Figure 1.2.
Terminology management can start with text mining, also known as term
mining, or term documentation (e.g. Austermühl, 2001: 102). Here we will
use the term text mining to refer to the process of searching for texts and

Figure 1.2 Simplified process and tasks in terminology management.

 Translators and Terminology Management 21

resources containing domain-​relevant terms and compiling corpora in each

relevant language accordingly. It is essentially a corpus compilation stage.
Some authors use term extraction to refer to term mining, but here we will
use term extraction (shortened from automatic term extraction) to refer to
the “processing of texts using computer programs in order to identify strings
that are potential terms” (Ahmad & Rogers, 2001: 725), since manual term
extraction, although possible, is becoming increasingly rare. Term extraction
may be monolingual, i.e. from a single text, or bilingual, i.e. from bilingual
or comparable corpora (Kageura & Marshman, 2019: 64) (more on corpora
in Chapter 5). Term extraction is sometimes referred to as term identification
(e.g. Aubin & Hamon, 2006; Heylen & De Hertog, 2015) or term recogni­­
tion (e.g. Kageura & Umino, 1996; Nenadić et al., 2003; Steurs et al., 2015).
Term extraction, whether manual or automatic, bilingual or monolingual,
thus “produces the raw material for terminology databases, material that
has to be examined, tested and validated before inclusion in a termbase”
(Kageura & Marshman, 2019: 64). This means that, crucially, automatic
term extraction does not extract terms but candidate terms (Bowker,
2002a: 82; Condamines & Picton, 2022: 225; Hartley, 2009: 113; Kenny,
2011: 462), also known as term candidates (Ahmad & Rogers, 2001; Heylen
& De Hertog, 2015; Kageura & Marshman, 2019; Tamás et al., 2016).
These are ‘provisional’ terms which need to be checked, e.g. by translators or
terminologists, usually by looking at the context in which they appear, before
confirming whether they are terms or not through a process called term val-
idation. Term validation can be carried out manually by specialists (e.g.
Vivaldi et al., 2012) but also automatic methods are being developed (e.g.
Melo Mora & Toussaint, 2015; Valderrábanos et al., 2002). Candidate terms
which become terms will be stored in the termbase. Once in the termbase,
terms and their associated information can be retrieved by computer-​assisted
translation tools or machine translation (MT) to support the translation pro-
cess by automatically displaying them for translators as they translate. The
process whereby CAT tools or MT match the terms in the termbase to the
terms in the source text will be referred to here as term recognition.
Figure 1.2 also shows that there is much to do in terminology management
and that many of its tasks ideally take place pre-​translation (e.g. Bononno,
2000: 655). Post-​translation, the main tasks are mostly related to reviewing
and updating the contents of the termbase in light of any client preferences,
changes in terminology, etc. It is a big ask for translators to be doing every-
thing shown in Figure 1.2 all the time, given their frequent time constraints.
Even if not every task is carried out all the time, this figure aims to show
what having a quality terminology management system in place could mean
for translators. Ensuring it is checked and updated, such a system is certainly
a smart investment for translators and can contribute to their long-​term sur-
vival and success in the profession (Popiolek, 2015: 358). In the following
chapter, we review three key notions in terminology management, i.e. term,
concept and domain.
22 Terminology Management for Translators

Follow-​up Tasks and Reflection

1. Do you identify yourself with any of the statements in Box 1.1? Which
one(s) and why?
2. Having looked at lexicology and lexicography, terminology and
terminography, phraseography would be the logical ‘partner’ of phrase-
ology. What would be the differences between phraseography and termin-
ology (or terminology management or terminography)?
3. Out of the three meanings of terminology in Box 1.2, only one should
be used in the plural, i.e. terminologies –​which one? You should be able
to find this information in monolingual English dictionaries. The other
meanings of terminology are ‘mass nouns’ or ‘non-​countable nouns’ and
should not be used in the plural (Sager, 1990: 3).
4. The three meanings of terminology in Box 1.2 may not always be straight­
forward to apply in practice. For example, the sentence “Terminology is
concerned with the study of terms” illustrates meaning 1 of terminology
as a discipline or theory. Can you find or invent sentences to illustrate each
meaning of terminology in Box 1.2? Can you classify each of the following
sentences by just one of the meanings?
a. He has made it his business to master the terminology of economics.
b. 20th century terminology has evolved substantially, both theoretically
and methodologically.
c. The development of terminology in South America is covered in the
following chapter.
d. We explain how our company manages terminology in multilingual
communication settings.
e. No structured discipline can function without specific terminology to
communicate its specialised knowledge.
f. Terminology is relatively young compared to lexicography, for example.
5. How do the various meanings of terminology translate in your other
language(s)?
6. Do a search on the web to find local translation agencies. Do they offer
any terminological services? If so, which ones? Is the word terminology
featured at all on their website? If so, in which context does it appear, e.g.
services offered, blog, etc.? Repeat the search for terminology manage-
ment. Do you get any results?
7. Would ICE (meaning ‘In Case of Emergency’) be included in the same
term record as ice (meaning ‘the solid form of water’) in a termbase for
translators? Why/​Why not? Do some research to find which domain(s) the
term ICE can belong to.
8. Consult one or two term banks online, e.g. IATE, UNTERM and
TERMIUM®, or others and browse through a few term records. What
kind of information do they provide and how user-​friendly is the presen-
tation of information?
 Translators and Terminology Management 23

Notes
1 From the McRae Group rock formations in New Mexico, where the discovery
was made.
2 Throughout this book, I will use specialised as a convenient umbrella term to refer
to all the kinds of texts that professional translators deal with. This may include
scientific and technical translation but it is not limited to it and specialised may
therefore also include administrative, legal, political and other subject areas.
3 Dictionaries, specialised dictionaries and encyclopaedias, etc., are lexicographical
works that translators may use, but generally to mine, or extract, information.
Termbases are the databases where translators store the terminological information
they need to carry out their translations and can be as individual as the translators
(or teams of translators) who use them.

Further Reading
A historical overview of the discipline of terminology can be found in Sager
(1990: 163ff) and Schmitz (1996: 221ff). See Cabré for a comprehensive methodology
for both ad hoc (1999: 152ff) and systematic (1999: 129ff) terminology searches.
Warburton (2021) provides useful information about prescriptive and descriptive
terminology. For more on phraseology, it is worth reading Colson (2008), Cowie
(2001) and Granger and Meunier (2008). Computational aspects of phraseology can
be found in Colson (2004), Corpas Pastor (2016) and Heid (2007, 2008).
For a further discussion and comparison of terminology in relation to lexicography
and linguistics, The Routledge Handbook of Lexicography (Fuertes-​Olivera, 2017) is
an excellent starting point, which can also be used to learn more about dictionaries in
general (as opposed to termbases). Contributions in this volume include contrasting
lexicography with terminology (Bowker, 2017), lexicography with translation (Calvo
Rigual & Calvi, 2014), lexicography with applied linguistics (Peters & Fernández,
2017), and lexicography with corpus linguistics (Faaß, 2017), among others.
Elsewhere, Alberts (2001) contrasts lexicography with terminography; Warburton
(2021) compares terminology to lexicology; Kageura (2015) and Tamás et al. (2016)
compare terminology with lexicography. Humbley (1997) examines to what extent
terminology can be referred to as specialised lexicography and vice versa.
There has been some debate in the literature as to whether terminology is a branch
of linguistics or a separate discipline. In fact, this became “one of the foremost themes
of the 1990s and beyond” (Humbley, 2022: 26). In its beginnings, terminology was
conceived by Wüster, who is considered to have laid the groundwork for terminology
as we know it today (see Chapter 3), as an “independent area which presents itself
as an area of intersection made up of the ‘sciences of things’ and of other disciplines
such as linguistics, logic and information technology” (Cabré, 2023a: 11). Cabré’s
more nuanced position is that “terminology considered as the totality of termino-
logical units, does not fall within the scope of linguistics (whether general or applied),
but instead can be addressed by drawing on linguistic theories as well as knowledge,
cognition and communication theories” (2023a: 9). However, there are other views
on this debate, as shown by Kageura (1995) and Myking (2001b).
De Bessé (1994) explores the meanings of terminology and provides a fourth
and fifth meaning (in addition to those in Box 1.2). Bowker (2015: 307) expands
24 Terminology Management for Translators

on the differences and similarities between term banks and termbases in a helpful
table, whereas Egwuogu (2023) compares termbases to translation memories and
glossaries. Other insights into the onomasiological approach can be gleaned from
Bowker (2019: 579–​580) and Humbley (1997: 24), whereas Warburton expands
on the semasiological approach in commercial settings. Finally, Tamás et al. (2016)
offer another perspective on the similarities and differences between translators and
terminologists.
2 Terms, Concepts, Domains

Key Questions and Issues

• Are terms and concepts the same thing?

• What distinguishes terms from words and why is this a useful dis-
tinction for translators?
• Aren’t terms words anyway?
• The importance of the domain for distinguishing words from terms.
• General language (words) vs specialised language (terms).

Terms and Concepts: What’s The Difference?

Although terms and concepts are not the same, it is easy to confuse one
with the other. Concept, as defined by Bowker (2019: 579), is a “unit of
thought that is used to organise people’s knowledge and perception of the
world around them.” For some scholars, the notion of concept is problem-
atic because not all concepts can be delimited clearly and they propose to
refer to concepts as units of understanding instead (Kerremans, 2010: 2;
Temmerman, 2000: 42). For translators, this means that finding definitions
for concepts may not always be a straightforward task, because some concepts
are slippery and difficult to define.
Terms are the labels (or ‘names’) that we use to refer to concepts (Olohan,
2016: 27) and are, of course, made up of words. A terminological ideal
(univocity, see Chapter 3) would result if each concept had a single term (i.e.
label or name) to refer to it in a given language. However, the complexities
of specialised knowledge in the real world (e.g. Montero Martínez & Faber,
2011: 95) mean that, depending on the domain, it is likely that translators
will encounter concepts that will have a set of labels (i.e. synonyms or
partial synonyms) to refer to it (Quah, 2006: 104) rather than a single one.
Therefore, in order to deliver high-​quality translations, adequately tailored to
the relevant readership and context, translators need to know which concepts
DOI: 10.4324/9781003302636-3
26 Terminology Management for Translators

exist in a domain and all the labels that exist to refer to them in both source
and target language, and they need to know which label to use in which
context. Further, as seen in the previous chapter, terms can be single words
or multiword expressions. A common error made by trainee translators
is to assume that “only single words, or at most compound terms, can be
documented in terminology collections” (Wright, 1997a: 13) (see Box 2.1
for more on this and other misconceptions). Terms may also be chemical for-
mulas, mathematical symbols, etc. (Thomas, 1993). A key idea throughout
this book is that translating terminology is a much more complex process
than just translating the words a term is made up of, as it also involves “truly
understanding the concepts behind the terms” (Bowker, 2022: 132).
By domain, as we saw in Chapter 1, we mean a network of inter-​related
concepts which have been decided on by consensus by experts in that domain
(Melby, 2012: 8). The collection of concepts, the relationships between
them and the terms to designate the concepts constitute the conceptual net-
work of that domain. Wherever possible, translators should try to get some
understanding of the conceptual network of the domain they are working
on, rather than focusing exclusively on finding terms and their equivalents in
the target language, since any other domain-​related information is as useful
to translators as terms themselves. However, the divisions between some
domains and sub-​domains are arbitrary and increasingly difficult to estab-
lish (e.g. Fernández-​Silva, 2022: 437). We expand on the importance of the
domain for translators in the following section.

Box 2.1 Common Misconceptions About Terms

• Having an advanced knowledge of a language means that you already

know its technical terminology: NOT NECESSARILY. You may
know the technical terminology of some domains but it is impossible
to know all the technical terminology of all domains simultaneously.
As Rey puts it, even “the most highly specialised people usually only
command a subset of the terminology of their field” (1995: 98).
• Terms are only single words: FALSE. Terms can be both single
words, e.g. shaft and multiword expressions, e.g. rear-​angle needle-​
roller bearing (Hartley & Paris, 2001: 311), both terms from the
automotive domain. In fact, many terms are multiword expressions
rather than single words. Translators should include both single-​
word terms and multi-​word terms in their termbases.
• Terms can be translated word-​for-​word: FALSE. Trainee translators
are taught early on that word-​for-​word translation is generally not
a good idea. The same can be said about terms, i.e. terms should
generally not be translated word-​for-​word into another language
because this approach almost always leads to the wrong label/​term
 Terms, Concepts, Domains 27

used in the target language. Instead, translators should carry out

some research to establish what labels/​terms exist for a concept
in the target language and use those labels/​terms instead. In some
cases, it may be that the correct term is actually a word-​for-​word
translation of the source language term but this should be achieved
by design (i.e. either domain knowledge and/​or research into the
domain) rather than by accident. In other words, terms should not
be translated but equivalents in the target language should be found
for them instead. By following an equivalence approach to terms,
rather than a translation approach, undesired effects such as the
anglicisation of terms in a target language, for example, should be
avoided.
• Terms, concepts and their definitions are forever: FALSE. As sci-
entific and technical knowledge advances and new discoveries and
inventions are made, both the scope of concepts and the labels to
refer to them (i.e. terms) can change too. For example, in addition
to the example of how the definition of planet changed, as we saw
in Chapter 1, whales were first classified as fish before they were
assigned to the category of mammals (Diki-​Kidiri, 2022: 207); the
plant species Aliciella pinnatifida was originally classified in the genus
of Gilia but later changed to Aliciella (Vacanti, 2023); and hundreds
of scientific names of species in the Jepson Manual of Californian
plants have changed since 1993 (Ritter, 2023). For example, the plant
known as Lotus scoparius has been re-​named to Acmispon glaber.

Terms and Words: Why Is This a Useful Distinction?

Terms, whether single words or multiword expressions, are made up of
words, because terms are the linguistic representations assigned to concepts
(Bowker, 2019: 579), at least to start with.1 However, not all words are terms.
There are at least three main reasons why it is a good idea for translators to
distinguish between terms and words –​see Thelen (2002a) and Varantola
(1992):

1 For terminological purposes, they are different kinds of data.

2 They require the use of different resources.
3 Different translation strategies are followed in order to translate them.

Regarding (1), an expression that a translator considers a term should be

entered into the termbase; if it is considered a word or general language
expression, then it should not be entered into the termbase. As the name
suggests, terms (specialised knowledge) should be collected in termbases,
whereas words (general language) should not.
28 Terminology Management for Translators

Regarding (2), information about, and translations of, words should

be sought in general language resources whereas information about, and
translations of, terms should be sought in specialised resources, whenever
possible. The kinds of specialised knowledge required in order to find the
relevant terms in the target language are not always available in general lan-
guage resources.
Regarding (3), terms and words should be treated differently in transla-
tion. Whereas words may be translated, paraphrased or omitted as required,
terms must be rendered by the correct equivalent in the target language and
they should generally not be paraphrased. This is also another reason not
to include words in termbases: whereas terms ideally only have one correct
equivalent in the target language, which is stored in a termbase, it would
be very difficult to collect in a termbase every possible translation and
paraphrase of a word or expression in the target language. The differences
between terms (specialised language) and words (general language) will also
have an impact on how definitions should be worded, as we explore later, but
there are other differences between words and terms worth pointing out, as
shown in Table 2.1.
There is an important difference in origin between terms and words.
Whereas words are the result of the general evolution of a language, through
its (various) word formation strategies, terms are the result of a convention
agreed upon by specialists in a domain (Desmet & Boutayeb, 1994: 309;
Sager, 1990: 61). For example, whereas cardiac and arrest are two words in

Table 2.1 General overview of differences between terms and words

Difference Terms Words

e.g. photosynthesisa in botany e.g. easy

Origin Conventions arising from General evolution of a language.

agreement among
specialists.
Scope of use Specialised language. General language.
Domain Domain-​specific concept. Unspecific concept; all domains.
Form Unique single word or Unspecific, various alternatives
multiword expression to possible to refer to the concept,
refer to the concept. e.g. easy could be substituted
by simple, straightforward, of
little difficulty, etc.
Type of data Terminological data; stored in Non-​terminological data; not
and storage termbase. stored in termbase.
Resources to Specialised resources. General language resources.
translate with
Translation Specific, unique equivalent in Paraphrase, translate or omit as
strategy the target language. required.

a
The process by which green plants and certain other organisms transform light energy into
chemical energy (from Britannica.com).
 Terms, Concepts, Domains 29

English resulting from centuries of development, cardiac arrest is the result

of experts concurring that this is a term that should be assigned to the con-
cept of ‘loss of heart activity.’ The convention can be seen in the choice of
words. Cardiac arrest, not heart arrest for example, was the choice approved
by experts; put differently, specialists take words from general language (and
of course from other, including classical, languages) in order to create terms,
and terms are words from general language that take on a special meaning
in a given domain. As a result, terms must be learned separately from words
(Sager, 1990: 105).
Therefore, distinguishing between words and terms means distinguishing
between general language (words) and specialised language (terms).
Specialised language has also been referred to as language for special purposes
or LSP2 (e.g. Gerzymisch-​Arbogast, 2008: 17; Humbley, 2022: 18; Pavel,
1993: 21; Pearson, 1998: 7), whereas general language has also been referred
to as language for general purposes or LGP (Bowker & Pearson, 2002; Cabré,
1999; Warburton, 2021). Specialised language is usually defined according
to “its users (specialists), its functions (communicating knowledge), and the
vocabulary it uses to do so (terms)” (Desmet & Boutayeb, 1994: 303). To put
it simply, in Bowker and Pearson’s words, “LGP is the language that we use
every day to talk about ordinary things in a variety of common situations,”
whereas “LSP is the language that is used to discuss specialized fields of know-
ledge” (2002: 25). Every language (e.g. English, Chinese, Arabic, etc.) “has
both LGP and LSP” (2002: 25), although it is difficult to establish how many
LSPs there may be in a language (e.g. Kittredge & Lehrberger, 1982: 104).
However, LGP and LSP are not opposites; instead, the difference lies in
the degree of specialisation: LGP is at the least specialised end of the spec-
trum and LSP at the most specialised, leaving some grey area in between
(Béjoint, 1988: 360; Ciapuscio & Kuguel, 2002: 42; Fuertes-​Olivera & Tarp,
2014: 8). For example, ten Hacken (2008: 983) uses the vocabulary referring
to the human body to illustrate four layers of specialisation, as shown in
Figure 2.1, where leg and blood are the least specialised terms at one end, and
platysma and pubococcygeus are the most specialised at the other.

Figure 2.1 Continuum of specialisation of terms.

Source: Based on ten Hacken (2008: 983).
30 Terminology Management for Translators

It is difficult to establish precise divisions as to when a particular term

becomes more or less specialised. This is because specialisation is mainly sub-
jective (Mayoral Asensio, 2007: 49; ten Hacken, 2008: 982), i.e. what one
specialist or translator considers specialised may be more general for another
specialist or translator. The degree of term specialisation may have a bearing
on decisions made regarding the design of a termbase, which is explored fur-
ther in Chapter 4.
In Table 2.1, photosynthesis is used as an example of a term, used in
the LSP of botany, for example. However, photosynthesis may also be
used as a word in LGP. The difference lies in the status of terms as words
or expressions which have developed a specific meaning in a specialised
domain, whereas words remain units of non-​specialised meaning in gen-
eral language. Distinguishing between words and terms can help translators
discern what should be included in their termbases and how they should
approach the research into, and the translation of, specialised terms and
concepts.

Box 2.2 Why is it Advisable for Translators to Specialise in a

Domain?

Becoming specialised in a domain can be very advantageous for

translators for the following reasons:

• Specialised translators can command a higher rate of pay (Olohan,

2016: 12).
• Terminology work becomes less time-​consuming if you can store
your research for future use.
• It is not so crucial to find an expert if translators have queries: they
are their own experts! Otherwise, translators would be confined to
providing “intelligent guesses” (Bononno, 2000: 655).
• Specialised translators can not only have access to the relevant com-
munity of experts but also obtain their “professional acceptance and
respect” (Austermühl, 2010). While recognised status among experts
is an issue with which translators have long struggled with (Dam
& Zethsen, 2008; Katan, 2011), there is now a gradual shift in the
translator’s favour, particularly in the case of those who specialise. In
this sense, specialisation can be a tool to combat the so-​called “sub-
servience” of translators (Godbout, 2016).
• Specialised freelance translators can move further into ‘bou-
tique,’ ‘top tier,’ or ‘premium’ segments of the market (Seidel &
Durban, 2010: 191–​207), as opposed to continue working in ‘bulk’
translation.
 Terms, Concepts, Domains 31

Therefore, budding translators are strongly encouraged to select

domains to specialise in as soon as possible, as specialising takes
time, but should remain prepared to work as a generalist translator if
required (see Picken, 1996: 170) (see Box 2.3). Of course, specialising
in one domain does not preclude translators from accepting any other
translation assignment outside of their domain –​it simply means that
they will be faster and more efficient translating in their specialisations
because they have already acquired knowledge and expertise, as well as
relevant terminological resources in those domains.

It may not always be straightforward to distinguish whether a word or

expression is a term from a specialised domain or whether it belongs to gen-
eral language (Varantola, 1992: 122). According to Thelen (2002a: 194), the
“most treacherous and most difficult are those items that look like ordinary
words, but ultimately appear to be terms.” For example, offer is both a word
from LGP and a term in legal domains. As a word in LGP, it means giving
something to someone. However, in legal LSPs, offer means “a promise, which
has the potential to be an agreement once it is accepted by the recipient”
(Zhang, 2013: 126). Offer could be mistranslated if understood by the trans­
lator as a word in general language rather than the specific term it is in legal
domains. Innselset et al. (2008: 73) found that a typical characteristic of eco­
nomic administrative domains is that “the terminology resembles the words
of everyday general language” and, therefore, “the number of polysemes may
appear too high if the terms are not assigned to their special domains.”
Therefore, a major difference between words and terms is the domain;
terms may belong to one or more specialised domains, e.g. photosynthesis
belonging to the domain of botany; offer to legal domains, etc., whereas gen-
eral language words may be used in all domains. For example, the word
easy in Table 2.1 is not connected to one domain specifically; it may be used
in all domains, from botany to law, physics and engineering. With domain
in the mix, terms can be viewed as triple relations: terms are the labels or
names (also referred to as form) for a specific concept in a specialised domain
in a particular language (Cabré, 1999: 81; Sager, 1990:13; ten Hacken,
2008: 984; Wright, 1997a: 13). This is illustrated in Figure 2.2 with photo­­­
synthesis as an example of term.
This figure shows that photosynthesis can be considered as a term in
botanical domains by translators because it has a special meaning in these
domains. Therefore, photosynthesis might be a term translators want to
record in their termbases. By contrast, we can also try to apply the triangle in
this figure to the word easy, with the aim of determining whether it is a word
or a term. If easy belongs to general vocabulary, i.e. if there is no domain in
which easy takes on a different meaning from general language, the triangle
will not work, and in this case easy can be considered a general word.
32 Terminology Management for Translators

Figure 2.2 Terms as triple relations (left) and photosynthesis as an example (right).

However, there may be domains in which easy takes on a specialised

meaning and thus becomes a term. For example, in snowboarding, when
referring to levels of slope difficulty, easy means ‘with a maximum gradient
of 25 per cent’ (Bieswanger, 2022). Therefore, as summarised by
Pinchuck, “membership of a terminological system thus gives an expression
terminological status” and “without it, it is a lexical item [i.e. word], not
a term” (1977: 178). Thus, photosynthesis and easy are terms in botany
and snowboarding, respectively, as each has taken on a special meaning in
those domains. However, photosynthesis and easy can also be used as words
from general language when not used in a specialised way. The triangle in
Figure 2.2 may not always work as well as described here, but it can be used
as a practical rule of thumb for translators to distinguish words from terms.

Why is the Domain Important for Translators?

In the previous section, we saw that the domain is important for translators
because it can help them distinguish words from terms, e.g. using the tri-
angular tool in Figure 2.2, especially because terms should be treated dif­
ferently from words in translation. In this sense, domains (and subdomains)
are as important for translators as the context in which the terms appear, as
context will dictate the correct selection of the appropriate term for the rele-
vant concept in that domain.
In addition, knowing how the conceptual network of a domain is
organised can help translators acquire the necessary specialised knowledge
(Faber, 2022) in their quest to become sufficiently expert in a domain to
translate the text at hand (see Box 2.2). As explained by León-​Araúz (2022),
“conceptual structures can be discovered or verified through corpus-​based or
 Terms, Concepts, Domains 33

Figure 2.3 Snippet of the conceptual network of mineralogy.

definitional analysis.” For example, from a sentence such as: There are two
types of goniometer, the instrument used to measure interfacial angles in a
crystal: contact goniometer and reflecting goniometer, we can start building
a conceptual network of mineralogy, illustrated in simple terms in Figure 2.3.
Thus, we can already establish some concepts in this domain and how
they are related to each other: goniometer as a concept is a hyperonym (or
superordinate) of the concepts contact goniometer and reflecting goniom-
eter, which makes them hyponyms (or subordinates) of goniometer (see also
“Relations between Terms” in Chapter 4). Other concepts and relations can
be gleaned from other sentences in the text and added to the conceptual
network (or ‘family tree’) of mineralogy. Of course, not every sentence in a
text will lend itself so easily to extracting domain knowledge and translators
(as opposed to terminologists) may never develop the full conceptual net-
work of a domain, but any piece of knowledge may be crucial for a thor-
ough conceptual understanding of the domain by translators. One way in
which translators can capture specialised knowledge about concepts and
the relations between them is by collecting definitions and storing them in
the termbase. We will return to definitions and relations between terms in
Chapter 4.

Box 2.3 “Mix and Match” Tips for Selecting a Domain to

Specialise in:

• Select a domain in which you are passionate (or at least interested to

some extent), to avoid becoming bored with the content.
• Select a domain that has great demand in your language combination.
• Select a domain in which you can gain a qualification to prove your
expertise.
34 Terminology Management for Translators

• Select a domain in which you may have access to experts.

• Select a domain in which you can attend conferences or events for
experts, where you can test your knowledge, network with experts
and make your services known to potential clients.
• Select an established domain for which many specialised resources,
term banks and glossaries already exist.
• Select a domain in which you may already have knowledge, to speed
up the specialisation process.

These criteria may not be applicable or relevant to all domains, all lan-
guage combinations and all translators, but they may provide you with
some inspiration for your professional career as a specialised translator
or interpreter. For further advice on how to go about specialising, it is
worth reading Seidel and Durban (2010: 191–​207).

Another important reason why the specialised knowledge of a domain is

important for translators is because it “develops differently in different
languages” (Mousten & Laursen, 2016: 64). When there are equivalent
concepts and terms in two languages, they can act as “anchor points” in
different languages and thus provide “a foundation for interlinguistic corres-
pondence” (Faber & L’Homme, 2022a: 5). Thus, the occurrence of contact
goniometer in the source text will be connected to the occurrence of it in the
target text. Additional translation challenges will of course arise when there
is no (obvious) equivalence between two languages, as explored further in
Chapter 4.
Terms and their equivalents in the target language are not the only aspect
of the source text that translators need to deal with. Translators need to
translate the entire text, which includes terms, but also the specific register,
tone, style, conventions, etiquette, preferences or specificities of a given
domain, which Meyer refers to as its “terminological flavour” (2000: 42).
As well as terminology, therefore, translators need to master the phraseology
of a given domain, an idea that we revisit in Chapter 4. Many examples of
special preferences, conventions or phraseology in different domains can be
found in the literature:

• Wakabayashi (1996: 361) explains how medical domains prefer the use of
the words ipsilateral and contralateral instead of ‘on the same side’ and
‘on the opposite side,’ respectively.
• Klerkx (1990: 216) explains that, in biology, it is customary to “abbre­
viate the name of the genus when several species of the same genus
are mentioned.” Thus, the genus Scirpus is abbreviated to S. from its
second occurrence onwards in Scirpus planifolius, S. rufus, S. cespitosus,
S. fluitans.
 Terms, Concepts, Domains 35

• Moreno-​Pérez (2018: 65) found that legal judgments in the United

Kingdom are “written in a close and personal style, while in Spain the
tone is very distant.”

Another important consideration for translators regarding domains is

that there is a “lack of consensus about what constitutes a subject field”
(Warburton, 2015a: 364). Further, “with increasing interdisciplinarity, the
demarcation lines between subject fields are becoming blurred” (Pearson,
1998: 16), compounding the fact that the divisions between specialised
domains are arbitrary, unstable or far from clear-​cut to start with (Delavigne
& Gaudin, 2022: 189; Fernández-​Silva, 2022: 437; Temmerman, 2000: 51).
For example, Temmerman (2000: 47) demonstrates the extent to which it is
problematic to establish firm division lines between related domains such as
biotechnology, microbiology and biochemistry, among others. This means
that there is considerable overlap between some domains and, therefore,
translators should bear in mind that “mastery of one subject field does not
imply an understanding of others, even if they are fairly closely related”
(Pinchuck, 1977: 164). Importantly, this also means that “sometimes texts
cannot be assigned to a single subject area, since many times more than one
field of specialization is included in a document” (Moreno-​Pérez, 2018: 63).
The multidisciplinary nature of many domains and the arbitrariness of the
divisions of knowledge means that translators may encounter the following:

• One term may designate different concepts in different domains (Olohan,

2016: 27). Examples of this abound: for instance, the term rock is used in
music and geology domains, but with different meanings in each domain;
fugacity means ‘early withering’ in botany but ‘the tendency of a gas to
expand’ in chemistry (Godly, 1993: 142).
• A term may designate a concept that can operate in several domains at
the same time with the same meaning. For example, interfacial angle is a
term both in mineralogy and crystallography with the same meaning in
each domain.
• Texts often contain terms from several domains. It is very rare for a text
to include terms from one domain only, since most texts “deal with mul-
tiple subject fields or are highly interdisciplinary in nature” (Wright,
1997a: 19). This means that, for example, a text on new satellite tech­
nology to uncover crude oil sites might contain terms from astronomy,
engineering and geography, among other domains and sub-​ domains.
This is of particular importance for interpreters, who should not assume
that they should only prepare financial terminology for a finance-​related
interpreting assignment, for example, as the assignment may well include
terms from law, health, medicine, accounting, etc.
• You cannot tell how specialised a text is until you read it. Just because a
text can be said to belong to the domain of economics, for example, it is
not sufficient to say that the text is specialised. The degree of specialisation
36 Terminology Management for Translators

(although it remains subjective) cannot be derived from the title alone

either. The text will need to be read first. In practical terms, factors such
as terminological density (i.e. how many technical terms there are) (e.g.
Ferraresi, 2019), among others, can be used to form an opinion as to how
specialised a text may be. Mayoral Asensio (2007) discusses in more detail
the concept of specialised translation and reviews how scholars have tried
to establish various typologies and classifications, including those based
on the degree of specialisation.

To sum up, terms are the specialised units of knowledge which experts use
to communicate with each other. Terms are typically made up of words, and
can be single words or multiword expressions, but they are treated differ-
ently from words in translation. Deciding whether an expression belongs
to a specialised domain can be used to help translators decide whether a
particular word or expression is a term or whether it belongs to general lan-
guage, but with the caveat that the division between some domains is not
entirely clear. Distinguishing terms from words is not the only theoretical
aspect that can help translators. In the following chapter, we review several
theories of terminology and principles of terminology management that can
also help translators be more efficient in their terminological work.

Follow-​up Tasks and Reflection

1. Explain whether energy is a word or a term in each of the following
sentences and why:
a. The engineers put all their energy into the development of the triode
vacuum tube project.
b. Currently, pumped-​storage hydropower is the most common way of
storing energy.3
2. How would you go about translating energy into your target language in
each of the two sentences above?
3. Find some terms in English which operate with different meanings in
different domains. For example:
a. pupil is a term in education and ophthalmology;
b. leg is a term in anatomy and in transport (e.g. leg of a journey), etc.
4. How well do pupil and leg correspond to the equivalents in your target
language in each of the domains in the previous exercise?
5. In the following examples of interdisciplinary texts, for which domains
would you expect to find terms? The first one has been done as an
example.
a. A text about European Union policy on environmental consequences of
agriculture: terms from agriculture, chemistry, law, environment, etc.
b. A text about terms and conditions for travel insurance for people with
pre-​existing medical conditions.
c. A text on Viking boat-​building techniques.
 Terms, Concepts, Domains 37

6. In the following text (from Kritsky, 2010: 10), find two examples of terms
in beekeeping and two examples of words or expressions from general
language:
Honey is a very stable product. The removal of excess water thickens
the honey and increases its sugar concentration, resulting in high osmotic
pressure. Osmosis is the movement of water across a membrane to
equalize the concentration of substances dissolved in surrounding water
and inside any membrane-​enclosed structure, such as a bacterial cell. The
high osmotic pressure of honey draws water out of bacteria, inhibiting
long-​term survival.
7. Are data bank and sanidine feldspar words or terms in the following
text from the domain of mineralogy (from University of Nairobi,
2019: 4)? Why?
Scientifically, minerals comprise the data bank from which we can learn
about our physical earth and its constituent materials. […] For example,
the mineral referred to as sanidine feldspar crystallizes at high temperatures
associated with volcanic activity; […] and that many clay minerals are
formed as the result of surface or near-​surface weathering.
8. Using the following text (from Drake & Fabozzi, 2010: 24), can you com­
plete the three gaps in the conceptual network of financial markets, in
Figure 2.4 below?
a. …………………………………..
b. …………………………………..
c. …………………………………..
From the perspective of a given country, we can break down a country’s
financial market into an internal market and an external market. The
internal market, which we also refer to as the national market, is made up
of two parts: the domestic market and the foreign market. The domestic

Figure 2.4 A portion of the conceptual network of financial markets.

Source: Adapted from Drake & Fabozzi (2010).
38 Terminology Management for Translators

market is where issuers domiciled in the country issue securities and where
investors then trade those securities. […] The foreign market is where
securities of issuers not domiciled in the country are sold and traded.

Notes
1 We will see later on that terms are more than just linguistic representations or
designations for concepts.
2 However, Pearson notes that in natural language processing the preferred term is
sublanguage (1998: 7).
3 https://​e ng​i nee​r ing​m ast​e rs.onl​i ne.gwu.edu/​e le​c tri​c al-​e ngi​n eer​i ng-​r enewa​b le-​
ene​rgy/​, retrieved on 21 July, 2021.

Further Reading
The distinction between words and terms is still an ongoing debate (e.g. Petit, 2001;
Thoiron & Béjoint, 2010). In a recent volume, L’Homme (2022: 246) discusses
Fillmore’s compelling example of innocent and guilty having different meanings as
words in everyday language, and as terms in legal language. Desmet and Boutayeb
(1994) further consider this distinction from the terminologist’s viewpoint.
A broader overview of specialised translation can be found in Gotti & Šarčević
(2006), Rogers (2015) and Olohan (2016); a more didactic and practical approach
can be found in Byrne (2012) and, more recently in Scarpa (2020). The edited
volume by Humbley, Budin and Laurén covers some of the more recent trends in
terminology and LSP studies (2018). For more on the degree of specialisation of lin­
guistic units, L’Homme (2022: 243) proposes a “sun representation” of the lexicon,
whereby the least specialised units are in the centre. The more specialised the units
are, the further away they are from the centre. Cabré (2023b) addresses the com-
plexity of the relationship between concept and term. Some scholars have tried
to establish a typology of specialised texts based on different (sometimes multi-
dimensional) criteria (Cabré, 2002; Ciapuscio & Kuguel, 2002; Göpferich, 1995;
Moreno-​Pérez, 2018; Vargas Sierra, 2005), although it does not appear that a con­­
sensus has been reached on how specialised texts should be classified or on whether
they should be at all.
The semantic triangle, as dealt with by Ogden and Richards (1923: 11), is often
used to explain how terminology impacts people’s mental processes –​see Figure 2.5,
where “[e]‌ach point of the triangle represents an aspect of how meaning is identi-
fied in language” (Brandt, 2021): the concepts we have in our minds (thought or
reference for Ogden and Richards) for the objects, etc. that we experience in the
world (referents for Ogden and Richards) and we use terms (symbols for Ogden and
Richards) as labels to refer to the referents. For didactic purposes, it is also useful to
add definition to the triangle as per Suonuuti’s model (2001: 13), also in Figure 2.5.
With the introduction of the Sociocognitive theory of terminology (see Chapter 3),
Temmerman (2022: 334–​335) added two more points to Ogden and Richards’ ori­
ginal version, i.e. situatedness (to refer to the fact that understanding takes place in
a situated environment) and embodiment (to refer to the fact that people understand
aspects of reality via sensory observation).
 Terms, Concepts, Domains 39

Figure 2.5 The semantic triangle.

Source: Adapted from Ogden & Richards (1923: 11) (left), and Suonuuti’s model (2001: 13)
(right), which adds definition to the triangle.

Faber (2009) looks at how terminology theories relate to scientific and technical
translation. The problems translators tend to find in scientific and technical transla-
tion are discussed in studies such as those by Klerkx (1990) and Gerzymisch-​Arbogast
(2008), for example. A somewhat overlapping term with LSP is ESP which stands for
English for Specific Purposes, which may include the study of ‘English for science,’
‘English for medical purposes’ etc. Under this label, there are some publications
relating to English in specialised domains (e.g. Gläser, 1995), but many focus on
teaching English language to learners (e.g. Hyland, 2022).
3 Key Terminological Theories
and Approaches

Key Questions and Issues

• What terminological theories and approaches are there?

• How can these theories and approaches inform the terminological
work of the translator?
• Theories may seem daunting, but learning a few key concepts will
help you understand why some things are done in specific ways.

What Do Terminological Theories Say?

Since this book is about the terminological work of the translator, a brief
overview of terminological theories and methods and how they can help
translators navigate the world of terminology in general (and terminology
management in particular) is pertinent here. It is beyond the scope of this
book to delve into translation or linguistic theories as well, so we will focus
on gaining insights into various aspects of terms which might help translators
in their practical use and management of terminology. The theories and
approaches covered in this chapter are mapped out in Figure 3.1 in a broadly
chronological fashion, starting from 1930s and working our way around the
circle. In this figure, key theories are presented along with one of their key
ideas in order to make the study of these theories a little less intimidating.
In addition, other examples of terminological models and approaches are
described in Box 3.1 (The Canadian example), Box 3.2 (Terminology studies
in China), Box 3.3 (Terminology studies in Arabic) and Box 3.4 (The Nordic
example).
According to Cabré (1999: 7), the “work carried out in the 1930s, simul­
taneously but independently, by Austrian, Soviet and Czech scholars” laid the
foundations for terminology as we know it today. Three schools1 of termin­
ology emerged from this work: the Vienna School, the Russian (then Soviet)

DOI: 10.4324/9781003302636-4
 newgenrtpdf
Key Terminological Theories and Approaches 41
Figure 3.1 A (broadly) chronological overview of terminological theories since the 1930s.
42 Terminology Management for Translators

School and the Prague School (Picht, 2011: 13). The 1930s were a period of
rapid industrial advancements, where new objects and concepts often needed
to be named (Sager, 1990: 8). To avoid a proliferation of multiple designations
for a single concept, which could disrupt communication among experts
(Cabré, 2003: 167), Austrian engineer Eugen Wüster “sought to eliminate
ambiguity through the establishment and use of precisely defined concept-​
term units” (Bowker, 2019: 579). He formalised his ideas in the General
Theory of Terminology (Allgemeine Terminologielehre), also known as the
classical approach, on which the Vienna School of Terminology is based.
There is a close similarity between the Vienna School and the Russian School
(Laurén & Picht, 2006: 166–​167), the latter being an umbrella term for many
(sometimes diverging) approaches to ‘terminology science’ in Russia (Shelov
& Leitchik, 2006), which nowadays share interest in knowledge ordering
and standardisation of terminology (Laurén & Picht, 2006: 168). The Prague
School of Terminology was interested in the structure and function of special
languages (or LSP, see Chapter 2), as was the Prague Linguistic Circle from
which it emerged (Schmitz, 2006: 578–​579). The main principles of Wüster’s
General Theory of Terminology (GTT) can be summarised as follows:

1. Concept is a unit of knowledge (Picht, 2011: 10) and terminology is

concept-​based; (onomasiological approach) (Bowker, 2022: 129).
2. Concepts are studied before terms (Temmerman, 2000: 04). The strong
focus on concepts has sometimes been referred to as the primacy of the
concept (Fernández et al., 2011: 16–​17; Humbley, 2022: 30; Trojar,
2017: 59).
3. Concepts are universal (language-​independent) and can be classified sys-
tematically (Warburton, 2020: 34).
4. Concepts and terms are dynamic and it is important to standardise them
(Picht, 2011: 10).
5. The relationships between concepts are hierarchical (hyperonymic/​
hyponymic, holonymic/​meronymic) (Felber, 1984) (see also Chapter 4).
6. Terms are the designations of concepts belonging to a language for special
purposes (see Figure 2.2 in the previous chapter) (Warburton, 2020: 34).
7. Terminology should aim for the ideal of univocity, i.e. each concept
should be designated by one term only, and each term should refer to
only one concept (Bowker, 2022: 129).
8. Terminology should be prescriptive (Humbley, 2022: 40–​41) (as opposed
to descriptive –​see also Chapter 2);
9. Terms are different from general language words (Pearson, 1998: 10–​11).
10. Terms are mostly nouns (or noun phrases) (Kübler & Frérot, 2003: 259).
11. Specialised domains are “water-​tight compartments” (Dury, 2022: 428).
12. Concepts should be defined by a set of necessary and sufficient
characteristics (Temmerman, 2007: 29; Warburton, 2015a: 366). As
Bowker explains, necessary characteristics are “those that every instance
of a concept must have” and jointly sufficient characteristics are “those
 Key Terminological Theories and Approaches 43

that no members of another class possess (and so they can be used to

distinguish one class from another” (Bowker, 2022: 134) (more on
definitions in Chapter 4).

Wüster’s efforts towards achieving standardisation and making terminology

socially recognised (Cabré, 2023a: 53) culminated in being selected to run
the special committee on terminology known as Technical Committee 37
(TC37/​ISO) of the ISO (International Organization for Standardization),
which is still in operation today (2023a: 53). (see also Chapter 10). He was
also instrumental in the “approval of a UNESCO programme (UNISIST)
that sped up the establishment of the International Information Centre for
Terminology (INFOTERM) in Vienna, an organisation dedicated specifically
to terminology” (2023a: 53).
However, for terminology management, one of the greatest contributions
of the General Theory of Terminology is the principle of concept orienta-
tion, which “should be strictly adhered to” (Warburton, 2021: 20). For
many years, Wüster’s theory “offered the only set of principles and premises
for compiling terminological data” (Faber, 2009: 111) but, from the 1990s,
other theories of terminology began to emerge. It is often suggested in the
literature that the new theories were developed as a criticism of, or in answer
to, disagreements with the General Theory. In fact, although many subse-
quent theories and approaches to terminology agree with the GTT regarding
concept orientation and other aspects, almost all of the criticisms of the GTT
are centred around two issues, the “idealisation of reality, knowledge and
communication and the field of application’s limitation to standardisation”
(Cabré, 2023a: 13). The literature often presents the GTT in opposition to
other theories despite any common ground they might share. However, the
critiques of various aspects of the GTT have proved useful to broaden the
scope of terminology as a discipline towards more social, human and lin-
guistic spheres.
For example, although univocity may have worked well in the domains
Wüster was interested in at the time he was working on them, if we take
a broader view that terms are more multifaceted and not always univocal,
then we find that terms may serve more communicative situations than
covered under the General Theory. This is one of the essential premises of
the Communicative Theory of Terminology, developed by Cabré during
the 1990s (Cabré, 1999) and extensively described in the 2000s (Cabré,
2003). According to this theory, terms must be studied in the framework of
specialised communication and, in order to do so, we must bear in mind that
terminological units (rather than terms) are polyhedral, i.e. multidimensional.
The metaphor of a polyhedron is used to “account for the complexity of ter-
minological units in their real and varied communicative context” (Cabré,
2003: 164), since the Communicative Theory “endeavours to account for the
complexity of specialized language units from a social, linguistic and cogni-
tive perspective” (Faber, 2009: 114). The multifaceted nature of terms and
44 Terminology Management for Translators

the need for terms to be considered in their communicative environment are

aspects that other scholars had also started to engage with at that time (e.g.
Sager, 1990: 10).
Since terminological units are viewed as multidimensional, Cabre’s theory
is also referred to as the Theory of Doors in which doors represent “plural,
but not simultaneous, access to the object; and in such a way that, whether
starting from the concept or the term or the situation, the central object,
the terminological unit, is [always] directly addressed” (Cabré, 2003: 186).
Thus, terminological units may be accessed from the linguistic or cognitive
doors, for example (Cabré, 2023a: 63). Cabré talks about terminological
units, rather than terms, and she states that the terminological status of ter-
minological units is only activated with the right “pragmatic conditions of
adjustment to a given type of communication” (Campo & Cormier, 2005). In
sum, the Communicative Theory endeavours “to account for the complexity
of specialized language units from a social, linguistic and cognitive perspec-
tive” (Faber, 2009: 114). The latest refinements of this theory can be found
in Cabré (2023a).
Another terminological theory which came into existence around the same
time as the Communicative Theory is that of Socioterminology. The word
socioterminology had been coined in Quebec in 1981 but later “became a
French specialty” (Humbley, 2022: 25). It emerged as a reaction to the fact
that the Wüsterian approach “did not take into consideration the fundamental
role of interaction” (Delavigne & Gaudin, 2022: 179). Socioterminologists
question the principle of univocity, prescriptivism, the primacy of the con-
cept and onomasiologically-​based terminology. Instead, they suggest that
term variation is fundamental in the terminology of different speech commu-
nities and that terms should be studied in their social dimension over time
(Dury, 2022: 425). Terms are not “regarded as mere labels but rather as lin­
guistic units that circulate and have different nuances” (Delavigne & Gaudin,
2022: 183). Accordingly, there are many types of variation, and definitions
of terms do not remain static. Similarly, due to the high level of termino-
logical variation that can be observed over time, the socioterminological view
is that specialised domains overlap and that they should be “regarded as
entities without fixed boundaries” (Dury, 2022: 428). Therefore, standard­
isation of terms, so strongly supported by Wüster and his successors, is a
“chimera” for socioterminologists, as language is in constant change (Faber,
2009: 113). As the name suggests, Socioterminology borrows methods from
sociolinguistics, especially the use of corpora, which socioterminologists use
to study specialised discourse (Humbley, 2022: 477). The use of corpora
opened the doors to new avenues of research into terminology and new ter-
minological theories, such as corpus-​based and Textual Terminology the-
ories (Pecman & Kübler, 2022: 265). Socioterminology has been described
by scholars such as Gambier (1991) and Gaudin (1993, 2003), and more
recently by Delavigne and Gaudin (2022).
 Key Terminological Theories and Approaches 45

Although socioterminologists worked with corpora, their main focus of

attention was the terms of different speech communities and were therefore
very interested in spoken data. In response to this and with the advances
in computerised tools and corpora research, terminological theories took
a turn towards corpus-​based approaches from the 1990s (rather than the-
ories), focusing more on the written word. Corpus-​based terminology “can
be best characterized as a working method which explores a collection of
domain-​specific language material (corpus) to investigate terminological
issues” (Gamper & Stock, 1998: 149). The influential work of Sinclair
(1991) signalled the start of corpus linguistics and its applications in other
disciplines (Pecman & Kübler, 2022: 265), such as natural language pro­
cessing, knowledge engineering and artificial intelligence. The mutually bene-
ficial relationship between terminology and these disciplines was already
spotted in the 1980s (as reported by Condamines & Picton, 2022: 222),
but the technological advances of the 1990s were needed in order to start
exploiting the speed and other advantages afforded by computerised tools.
One of the best known corpus-​processing tools, which can be added to
the translators’ toolkit, is the concordancer (Kenny, 2011: 460), which we
explore further in Chapter 6. Examples of corpus-​based studies include
Baker (1995), who brought corpora into Translation Studies (Pecman &
Kübler, 2022: 265) and Pearson (1998) who explored the use of corpora
in terminology. Examples of findings from corpus research include the fact
that genres2 have been found to be an essential component of termino­
logical research (Pecman & Kübler, 2022: 263) and the fact that “termino­
logical variation is prominent in specialised corpora” (Fernández-​Silva &
Kerremans, 2011: 319).

Box 3.1 The Canadian Example

• French and English have been official in Canada since the Official
Languages Act in 1969 (Beaudoin, 2009: 139).
• Two governmental institutions conduct terminological work: the
Translation Bureau (interested in translation) and the Office québécois
de la langue française (interested in language planning) (L’Homme,
2006: 65).
• Throughout the 1970s, terminology in Canada emerged as a distinct
discipline (Williams, 1994: 195).
• Before the 1970s, however, as Bowker (2022: 130) points out,
“[t]‌erminology was not taught; textbooks did not exist; termino-
logical research methods were embryonic; computerized term banks
had not been developed; and the tasks of terminologists had yet to
be defined.”
46 Terminology Management for Translators

• Although terminology as a profession or discipline is not unique to

Canada, the level of excellence achieved in this country has been
the object of admiration by the international community (Delisle,
2008: 317) (see also Chapter 7).
• Legal drafting and translation present additional difficulties because
of the coexistence of two languages (English and French) and two
legal systems (common law and civil law), which have undergone
parallel development (Beaudoin, 2009: 136).
• Today, there are two major government-​ run term banks in
Canada: TERMIUM®, and the Grand dictionnaire terminologique,
both publicly available online.
• Bowker (2002b) evaluates the status of the terminology profession
in Canada in the 21st century.

Corpus-​based research continues to this day, with scholars such as Marshman

(2022), for example, who takes a closer look at how knowledge patterns,
borrowed from knowledge engineering, can fruitfully be applied to corpus
analysis. A knowledge pattern is “a lexical pattern that reveals an under-
lying conceptual relation” (Bowker, 2022: 139). For example, X is a Y (e.g.
‘a submarine is a warship that operates submerged in the sea’) is a pattern
of words that can be used to find hyponyms of concepts, and is/​are defined
as (e.g. ‘an iceberg can be defined as a large floating mass of ice’) is a pattern
of words that can be used to find definitions. Knowledge patterns can be
used to find and classify occurrences of conceptual and semantic relations
in texts. Thus, pattern-​based approaches to corpus analysis are “based on
the assumption that there are recurrent and predictable linguistic cues that
indicate specific types of information in texts” (Marshman, 2022: 291).
Automatic methods for extracting knowledge patterns will need to discern
between true occurrences and false positives, i.e. not every occurrence of X
is a Y will point to conceptual relations, such as is a in the sentence ‘this is a
good start,’ for example.
Further, Marshman (2022: 309) makes interesting observations in that,
“given the established usefulness of corpora for both conceptual and lin-
guistic research in terminology, translation and other language activities –​but
also the time-​consuming and labour-​intensive nature of corpus building –​
it is surprising that the use of KPs [knowledge patterns] to help build cor-
pora has not received more attention.” She also notices that a “particularly
glaring gap is that of corpus use among professional translators” and that
more “adapted tools for both corpus construction and querying could help
to encourage corpus use by some professionals.”
The corpus-​based approaches of the 1990s “paved the way for text genres
being an essential component of terminological research” (Pecman & Kübler,
 Key Terminological Theories and Approaches 47

2022: 263). One theory that takes texts and genres as its main focus of study
is the Textual Terminology theory. This theory is close to Socioterminology
in terms of theoretical stances and methodologies (Delavigne & Gaudin,
2022: 179) and it began to be explored in the late 1990s (Pecman & Kübler,
2022: 263) with the work of Bourigault and Slodzian (1999), for example.
However, one difference between the two theories lies in that, as the name
suggests, in Textual Terminology, “it is neither the term nor the corpus that
constitutes the most significant element but rather the text” (Condamines
& Picton, 2022: 222). By concentrating on texts, Textual Terminology
places a strong emphasis on “genres, registers and, consequently, domains”
(Pecman & Kübler, 2022: 263) but faces the problem that genres, domains
and sub-​ domains are “slippery customers” (Rogers, 2000: 9). Another
defining characteristic is that this approach “advocates the combined appli-
cation of corpus linguistic methods, natural language processing techniques
and linguistic analysis to collect and interpret terminological data” (Faber
& L’Homme, 2022b: 7). Condamines and Picton (2022: 223–​228) explain
the six stages typically followed in this theory for linguistic analysis and how
Textual Terminology has evolved since its inception. Looking to the future,
Condamines and Picton (2022: 223–228) also explain how the issues around
compilation of large, specialised corpora, relatively scarce and still substan-
tially smaller than general ones, should be tackled by Textual Terminology,
as appropriate specialised corpora are a useful focal point for terminological
research.
Despite the usefulness of corpora for both translation and terminology,
for some scholars a “text corpus is not always sufficient as a context to
understand realities” (Temmerman, 2022: 352). After the Communicative
and Socioterminology theories of terminology added communicative and
social dimensions to Wüster’s GTT, a new group of terminological theories
emerged at the turn of the millennium which aimed to add yet another
dimension to the GTT. These theories are generally referred to as cognitive,
because they focus strongly on the mental mechanisms that we use to under-
stand specialised discourses. Cognitive theories of terminology also “adopt a
more flexible view that allows for concepts to be organised in different ways
along different dimensions (strongly reminiscent of Cabré’s Communicative
Theory), rather than viewing concept systems as fixed or rigid, where each
concept has a single and ‘correct’ position within the system” (Bowker,
2022: 134), as was claimed by the GTT. Therefore, cognitive theories are
also interested in conceptual representations, which has led some of them to
create methods to visualise concepts as part of a system with images, con-
cept maps, etc., thus adding a multimodal dimension to terminology (see
also the Satellite Model of concept mapping in Chapter 6). Here we will take
a quick look at the Sociocognitive Theory of Terminology, and Frame-​Based
Terminology.
48 Terminology Management for Translators

Box 3.2 Terminology Studies in China

• Erya (‘The Ready Guide,’ 尔雅) is considered to be the first recorded

volume of terminology in Chinese, which dates back to around 200
BCE.
• For a long time, terminology was considered part of lexicograph-
ical work and, more importantly, part of the political process of
translation.
• From the founding of P.R. China in 1949 to 1965, terminology work
was an important part of the standardisation process of the Chinese
language. In 1965, the Chinese Academy of Science (中国科学院)
took over standardisation roles and was involved in developing the
Pinyin system (romanisation of Chinese) and in simplifying official
Chinese characters.
• From 1966 to 1984, the stagnation stage: “strongly influenced by
the Cultural Revolution and domestic political reform in China […]
terminological work was politically and linguistically discouraged”
(Li & Hope, 2021a: 124).
• From 1985 to 1995, a period of exploration began, marked by the
launch of the Natural Science Terminology Committee (全国自然科
学名词审定委员会) in 1985. During this period, Chinese scholars
started to embrace Western terminological theories.
• The years between 1995 and 2005 are labelled as a develop-
ment period, marked by the 95th Congress of the China National
Committee for Science and Technologies, whose name changed to
China National Committee for Terms in Sciences and Technologies
(全国科学技术名词审定委员会) (CNCTST).
• From 2005 to the present, the “boom” stage: international
collaborations with America, Europe and other parts of Asia have
intensified, academic publications have increased, linked to a
sharp rise in research funding, and the number of higher education
institutions offering terminological training (as part of translation
and interpreting courses or degrees) has expanded notably. In 2021,
the CNCTST became CNTERM (Lu, 2022: 44).

The main challenge for China now, as argued by Li and Hope

(2021b: 128) is “that of maintaining this momentum and ensuring
that its socioeconomic and cultural benefits are shared equitably across
China, from its metropolises to its peripheries, and between all sections
of society.” Another overview of terminology work in China to the pre-
sent day is provided by Lu (2022).

Source: Summarised from Li and Hope (2021a, 2021b).

 Key Terminological Theories and Approaches 49

The Sociocognitive Theory of Terminology was postulated by Temmerman

(2000) and shares many principles with Frame-​ Based Terminology. As
explained by Faber and L’Homme (2022b: 9), this theory presents a dynamic
model of understanding, which has a linguistic dimension (terms are under-
stood in texts), a cognitive dimension (understanding is a categorisation in
the mind) and a communicative dimension (understanding is the outcome of
communication). This means that concepts do not exist in isolation but as
“interrelated elements in texts which make them [the concepts] come alive”
(Montero Martínez & Faber, 2011: 96). Thus, since the sociocognitive view of
concept is much more flexible than that of the GTT, sociocognitivists propose
replacing concept by the broader unit of understanding and consider terms
as the starting point of terminological analysis (Temmerman, 2022: 337).
This constitutes a departure from the GTT in that, in the GTT, a concept
should be designated by one term only (univocity). Units of understanding
allow multiple designations, i.e. terms, for a single concept. For instance, a
recent example used by Temmerman (2022: 336) is that the legal definition
of mother in Belgium has given rise to three units of understanding: surro-
gate mother, anonymous mother and co-​mother. This example also shows
that this unit of understanding is very specific to a communicative setting
(i.e. Belgium in this case) and that, therefore, this theory places great import-
ance on situatedness, i.e. the fact that understanding is embedded in society
(Temmerman, 2022: 337), which explains the socio-​ in sociocognitive. This is
why Sociocognitive terminology takes a diachronic approach to terminology
i.e. exploring how terms evolve over time.

Box 3.3 Terminology Studies in Arabic

• Arabic has a rich lexicographical tradition which underwent some

qualitative and quantitative changes over time (Al-​Kasimi, 2019: 7).
• The birth of Islam in 610 CE gave a great impetus to learning in
general and the Arabic language in particular (Al-​Kasimi, 2019: 8).
• The first Arabic language dictionary, Kitāb al-​ ʿAyn (‘The Book
of the letter ‘Ayn,’ ‫ )العين كتاب‬was compiled in the 8th century
(Al-​Kasimi, 2019: 10).
• During the 9th century, the Bayt Al-​ḥikma (‘House of Wisdom,’
ْ was founded in Baghdad, which was an academy for
‫)الحِ ْك َمة بَيْت‬
research, writing, and translating Greek philosophy. The scientific
focus of this time meant that specialised dictionaries began to be
compiled (Al-​Kasimi, 2019: 23).
• During the 19th century, Arabic suffered diglossia, i.e. was divided
into two or more varieties: Modern Standard Arabic (or Literary
Arabic, the language of school and university instruction, formal
speech and the media) and other forms such as Classical Arabic
50 Terminology Management for Translators

(the language of the Qur’an) or Colloquial Arabic (the language of

informal speech). In addition, there are many national or regional
dialects. Diglossia has given rise to multiple synonyms for the same
concept. In addition, the translation of terms is often bound to its
source language, usually English or French, which sometimes forces
terminologists to adopt at least two Arabic equivalents for some
technical terms (Elmgrab, 2016: 75–​76), one influenced by English
and the other by French.
• After the independence of most Arab countries during the middle
of the 20th century, dictionaries of modern science were compiled
in order to ‘arabise’ teaching and the administration that had been
under English, French and Italian rule (Al-​Kasimi, 2019: 27).
• In 1960, the Bureau of Coordination of Arabisation was founded in
Morocco to ensure standardisation and unification of Arabic termin-
ology. Since then, it has compiled more than 40 specialised unified
“encyclopedic dictionaries” and other terminological resources in
various fields of knowledge (Al-​Kasimi, 2019: 27–​28).
• It is encouraging that the 21st century has seen a proliferation of
digital corpora and resources in Arabic, considering there was “next
to nothing” before then (Mörth, 2018: 505).

Source: Summarised from Elmgrab (2016), Mörth (2018) and

Al-​Kasimi (2019).

Because a unit of understanding is broader than concept, Sociocognitive ter-

minology introduces us to the idea of prototypicality in definitions. Rather
than defining concepts by a set of necessary and sufficient characteristics,
as proposed by the GTT (Kerremans, 2010: 2), a method which gen­
erally makes definitions very succinct (Varantola, 1992: 124), units of
understanding can be categorised in a more prototypical way, i.e. by com-
parison to the most similar or typical item or member in that category. For
example, robin can be said to be a prototype of bird, because robins are
‘typical’ instances of what we call bird (ten Hacken, 2008: 983). Therefore,
a definition of bird can start from a definition of robin. In other words,
the idea of prototypicality is that a definition of bird based on a definition
of robin as a ‘typical item’ should be enough to define kiwi or penguin to
some extent too, even though they are generally considered less prototyp-
ical birds than robin. In addition, prototypicality is a matter of degree, i.e.
“not every member is equally representative for a category” (Temmerman,
2022: 338). More about prototypicality in definitions can be found in ten
Hacken (2008).
 Key Terminological Theories and Approaches 51

Box 3.4 The Nordic Example of Terminological Collaboration

• Nordic countries: Denmark (including Faroe Islands and Greenland),

Finland (including Åland), Iceland, Norway and Sweden.
• First terminological efforts started in Iceland in the 19th century.
• Regular terminological work in Denmark, Norway and Sweden
since the 1940s.
• Nordterm:3 a network of organisations and societies in the Nordic
countries which are engaged in terminology work. It started in
1976 and its most important activities are the Nordic terminology
conferences which are held biannually.

Main contributors to Nordterm (from the Nordterm website):

• Denmark: Terminologigruppen (Danish Terminology Group).

• Faroe Islands: Málráðið (Faroese Language Council).
• Finland: Sanastokeskus /​Terminologicentralen (Finnish Terminology
Centre).
• Greenland: Oqaasileriffik (Language Secretariat of Greenland)
• Iceland: Stofnun Árna Magnússonar í íslenskum fræðum (Árni
Magnússon Institute for Icelandic Studies).
• Norway: Språkrådet (Language Council of Norway).
• Sami: Sámi Giellagáldu (the Sámi).
• Sweden: Institutet för språk och folkminnen, Språkrådet (Institute
for Language and Folklore, Language Council of Sweden).

Source: From Bucher (2017: 269–​280).

The Sociocognitive theory also focuses on metaphorical understanding, i.e.

how we use metaphors to name and understand specialised concepts, e.g. river
bed, wing walls (a pair of retaining walls added at an angle to a main wall
or structure for support) or mushroom ceiling (Sager, 1990: 72) (a specific
architectural design where the ceiling has a dome-​like or curved shape, often
resembling the cap of a mushroom). However, the Sociocognitive theory takes
metaphorical understanding further by explaining that specialised meaning
can be mapped on to metaphorical frames that help us understand the text.
Temmerman provides many examples of this in the domain of life sciences.
To illustrate, one metaphor that can be seen throughout texts in this domain
is D N A I S I N FO R MAT I ON I N A N AT L A S O F MA P S , 4 which results in sentences
such as Geneticists set out mapping the genomes of different organisms or in
phrases such as the map of human DNA (Temmerman, 2000: 157), etc. More
52 Terminology Management for Translators

recently, Sociocognitive terminology “has also begun to focus on ontologies5

as a more viable way of implementing conceptual representations” (Faber,
2009: 118) by developing a methodology that Temmerman and her colleagues
at Brussels refer to as Termontography (Temmerman, 2007: 32), a term that
combines terminology, ontology and terminography (see also “Concept
Maps” in Chapter 6). The Termontography methodology aims “to link
ontologies with multilingual terminological information” (Faber, 2009: 118)
and is interested in visualisations of the domain, created by specialists, which
reflect “the relevant culture-​independent and human language-​independent
categories and intercategorial relationships” (Kerremans et al., 2004: 563).
Termontography is described further in Temmerman and Kerremans (2003)
and Kerremans et al. (2004).
Picking up on the idea of frames as templates for understanding, another
cognitive approach to terminology is Frame-​Based Terminology, initially
explained by Faber (2009). Like the Communicative Theory and Sociocognitive
terminology, Frame-​Based Terminology is descriptive, text-​driven and admits
term variation and polysemy (Faber, 2022: 354), principles which are in
opposition to the GTT. As its name suggests, Frame-​Based Terminology is
based on frames, understood as a kind of cognitive context or structure in our
minds which we use to understand the world around us. In this sense, frames
have their roots in the work by Minsky (1975) and Fillmore (1985) on frame
semantics in general language (Faber & Cabezas García, 2019: 204). Frame-​
Based Terminology adapts frame semantics to specialised knowledge and ter-
minology, and it is particularly interested in representations of knowledge, the
organisation of concepts into categories and the semantic and syntactic behav-
iour of terminological units in one or more languages (Faber, 2022: 354). In
terminology, frames “can be used to link terms in different languages to the
same specialized concept” (2022: 354) and “the best way to study specialized
knowledge units is by studying their behavior in texts” (Faber, 2009: 120).
This is because, although words and terms are part of the same linguistic
reality, they “belong to different cognitive structures or frames” (Montero
Martínez & Faber, 2011: 96). Frame-​Based Terminology also suggests that
“the description of specialized domains is based on the events that generally
take place in them, and each area of knowledge is said to have its own event
template” (Bowker, 2022: 141). For example, in very simple terms, in the
domain of environment, there is a frame for “extreme event” under which
we can find event-​specific concepts such as landslide, flood, avalanche, etc.
(Buendía-​Castro et al., 2014: 65–​66). In the domain of finance, the event of
“bank transactions” includes frames for deposit, withdrawal, bank account
management, etc. (Pilitsidou & Giouli, 2020: 267).
According to Frame-​Based Terminology, the specialised knowledge needed
to work out what the frames of a domain are can be extracted from the
knowledge contained in, and provided by, definitions and corpora in different
languages “through the use of knowledge patterns that encode semantic
relations” (Faber, 2022: 358). The general idea is that, because “the general
 Key Terminological Theories and Approaches 53

function of specialized language texts is the transmission of knowledge, such

texts tend to conform to templates in order to facilitate understanding”
(2022: 358). Thus, specialised texts “are also generally characterized by a
greater repetition of terms, phrases, sentences, and even full paragraphs,”
which is one of the principles underlying the use of the translation mem-
ories translators work with (2022: 358). Frame-​ Based Terminology also
maintains that images have “an important role in specialized knowledge
representation” and help understand concepts better (Faber, 2022: 373).
Although more research is required in order to “gain further insights into the
parameters that link visual contexts with linguistic contexts” (2022: 373),
some Frame-​Based Terminology studies, e.g. by Reimerink et al. (2016) and
Reimerink and León-​Araúz (2018), are starting to explore visual know­
ledge patterns, as “the means by which images convey conceptual know-
ledge,” such as arrows, labels and colour-​coding (Faber, 2022: 372). Practical
applications of Frame-​Based Terminology include terminological knowledge
bases such as PuertoTerm (Bowker, 2022; Faber et al., 2006) in the domain
of coastal engineering and EcoLexicon (Buendía-​Castro et al., 2014; Faber
& San Martín, 2011; León-​Araúz & San Martín, 2018; León-​Araúz et al.,
2018) in the domain of environment.
One aspect that has been missing from the criticisms of the GTT is that
of the cultural nature of terms. The aforementioned theories and approaches
expanded the GTT in social, linguistic, textual, and cognitive perspectives
but culture is not a dimension that has been thoroughly studied in termin-
ology, perhaps because, in the long-​term influential GTT, concepts were
“considered objective, atemporal, asocial and ideologically neutral” (León-​
Araúz & Faber, 2024: 40), as well as language-​independent. However, cul­
ture has long been studied from the point of view of translation (Ditaranto,
2005; Gregorio Cano, 2012; Hatim & Mason, 1997; Katan, 2004), and even
from a scientific and technical translation perspective (Ahmad, 2006; Gläser,
2001; Kermas, 2006; Stolze, 2009). A relatively recent terminological theory
that aims to address that gap is the Cultural Terminology theory, as proposed
by Diki-​Kidiri (2008, 2022). This theory takes the long-​standing view that
culture, “both individual and collective, conditions the way people perceive
and categorize reality” (Diki-​Kidiri, 2022: 199) and, of course, the “division
of reality often varies from one culture to another, giving rise to concepts
specific to each culture” (2022: 208), which is in opposition to the GTT’s
claim that concepts are language-​independent. Therefore, terminology in this
theory is understood to be “the study of terms as the cultural expression of
specialized knowledge in any natural language belonging to a community of
speakers” (2022: 200).
This theory originated in Africa where the question arose as to how
one’s culture should be taken into account when acquiring new knowledge,
especially in the context of producing school materials for African children
(Temmerman, 2022: 339). In this sense, this theory is interested in the multidi­
mensional nature of terms and in the re-​conceptualisation of concepts from
54 Terminology Management for Translators

Figure 3.2 “Percepts” in the Cultural Terminology theory.

Source: Adapted from Diki-​Kidiri (2022: 211–​212).

one culture to another, as part of the larger process of ‘appropriation of the

new’ (Diki-​Kidiri, 2022: 208). Each viewpoint from which a concept can be
perceived is referred to as a percept (i.e. from perception). Diki-​Kidiri provides
an example from three African languages, Sanngo (Central African Republic),
Bambara (Mali) and Lilikó (Democratic Republic of Congo). Each of these
languages has a word for the concept of bicycle but this concept is perceived
differently. Thus, in Bambara, the percept is ‘iron horse,’ in Sanngo it is ‘rubber
wheels’ and in Lilikó it is ‘four legs’ (2022), as illustrated in Figure 3.2.
Diki-​Kidiri goes on to provide other examples in other languages and, in
the latest description of this theory, Diki-​Kidiri (2022: 215) proposes a meth­
odology for identifying terms without equivalents. He also lists several studies
that have successfully applied the Cultural Terminology theory. Although
this theory arose with African languages in mind, Diki-​Kidiri envisages that
it may well be “developed to empower speaker communities of less endowed
languages who want to access new technologies and science while preserving
their cultural identity” (2022: 216).
During the 2000s, an approach to terminology appeared in response to
the claims by the GTT that: (i) terms are mostly nouns or noun phrases; (ii)
polysemy should be avoided; and that (iii) term relations are hierarchical,
as we saw at the beginning of the chapter. This approach is known as the
Lexico-​Semantic Theory of terminology, as postulated by L’Homme (2004)
and extensively described in L’Homme (2020a). Inspired by the Explanatory
and Combinatorial Lexicology (ECL) of Mel’čuk et al. (1995), this approach
has provided us with a compelling case for other parts of speech, such as
verbs, adjectives and adverbs, which have long been overlooked in termin-
ology (L’Homme, 2022: 237), to be considered as terms alongside nouns and
noun phrases (L’Homme, 2022: 258). In this approach, polysemy is viewed
as a natural property of lexical units, rather than an “accidental problem” as
stated by the GTT (L’Homme, 2004: 12). For example, L’Homme found that
 Key Terminological Theories and Approaches 55

program has at least three different meanings in computing, i.e.: (i) an oper-
ating system; (ii) application software; (iii) utility program. In addition, this
approach has explained how non-​hierarchical relations (e.g. cause–​effect rela-
tion, process–​product relation), as well as syntagmatic (often collocations)
and paradigmatic relations (e.g. relations of terms to other terms, such as
synonymy, partial synonymy, antonymy) between terms are as useful for ter-
minological analysis as hierarchical ones.
In the case of adjectives, their ‘revised’ status as potential terms raises
the question of where the boundaries lie between multiword units (phrase-
ology) and multiword terms (terminology). L’Homme (2022: 247) discusses
the example of biodegradable pollutant and whether it should be considered
as a multiword term or as a single term (pollutant) which has a collocate
(biodegradable), but this issue is more problematic for lexicographers and
terminologists than for translators. In fact, lexicography has been one of
the applications of the lexico-​semantic theory, e.g. L’Homme (2005, 2014).
For translators, it is more interesting to use other criteria, such as client
preferences, the readership’s level of specialisation, etc. in order to decide
whether to include it in the termbase or not, regardless of whether it is a
phrase or a term (see also the discussion of modifiers in Chapter 5). Examples
of terminological studies that apply the lexico-​semantic approach include
Durán Muñoz & L’Homme (2020) and L’Homme & Bae (2006).
More recently, another approach which has been applied to terminology
for a decade or so (as reported by Bertels, 2022: 311) is that of Distributional
Semantics. The original idea can be traced back to American structuralism
(Harris, 1954) and British lexicology (Firth, 1957) and it is currently a main­­
stay of natural language processing (Emerson, 2020: 7436). The distribu­
tional hypothesis is that similar words should appear in similar contexts
(2020: 7436) and, therefore, the underlying idea is that “words or terms
with similar distributions have similar meanings” (Bertels, 2022: 311–​312).
According to this approach, information about the way that words are
“distributed” (i.e. appear) across text in (ideally) large corpora can be used
to compute the degree of similarity, or semantic relatedness, between them
(Bertels, 2022: 311–​312). For example, the frequency of co-​occurrence of
river and bed in a corpus, coupled with the fact that river appears before
bed more frequently than the other way around, might indicate that river
bed is a meaningful phrase. Of course, for the computations to yield the
desired results, the larger the corpora the better. However, as we have seen
before, corpora in specialised domains are generally much smaller and fewer
than general language ones, and specialised corpora are still relatively under-​
researched (Bertels, 2022: 319).
In order to determine whether river bed is indeed a meaningful phrase,
measures of unithood have been devised (Daille, 2003: 29; Drouin, 2006: 376;
Heylen & De Hertog, 2015: 209). Unithood refers to the relative strength of
the connection between words, for example between river and bed in this
case. If the measures work correctly, the unithood of river and bed should be
56 Terminology Management for Translators

strong, as river bed constitutes a lexical unit. Because unithood is based on

distribution and frequency of co-​occurrence, it is relatively easy to measure,
but it has “no semantic basis” (Warburton, 2021: 95), i.e. it does not tell us
that river and bed form a unit with meaning, it simply tells us that they form
a unit. However, we can infer that, if they form a unit, it is probably because
they are related semantically. In other words, unithood helps to establish
“the boundaries of meaningful phrases” (Hätty et al., 2018: 62). Further,
because we need at least two words to establish unithood, this measure only
applies to multiword expressions (Warburton, 2021: 95).
Although unithood may be a useful measure, it does not determine
whether river bed is a term. In order to do that, measures of termhood are
used. Termhood refers to the degree of association of the word (or multiword
unit) to a specific domain (2021: 95), i.e. termhood measures might indicate
whether river bed is a term in a given domain, e.g. hydrology. Termhood
measures can apply to single words or multiword units, but termhood is
difficult to measure (Daille, 2003: 29). Both unithood and termhood are a
matter of degree. Either can be strong, weak or “any degree in-​between”
(Warburton, 2021: 94). As Warburton explains, “a term that has strong
termhood is highly domain-​specific and probably used by specialists, whereas
a term that has weak termhood might have domain-​specific contexts or com-
municative situations, or is transitioning to the general lexicon.” Unithood
and termhood, as well as other measures, are used in automatic term extrac-
tion (Daille, 2003; Heylen & De Hertog, 2015; Kageura & Marshman,
2019), which we revisit in Chapter 5.
Finally, it would seem that the future will bring even more new approaches
and theories of terminology, as we gain more detailed insights into the mul-
tiple and varied spheres of terminological activity. In this sense, potential
new approaches include Pragmaterminology, initiated by De Vecchi (2015,
2020), which focuses on the language used in commercial settings, such as
marketing and business phraseology, trade names, etc. (Druță, 2020: 74).
In particular, the pragmaterminological approach “aims at answering the
question of what exactly has to be known to work at micro-​level in work
communities, and how knowledge must be shared to cope with knowledge
asymmetries and ensure cooperation between experts within the company or
organization” (De Vecchi, 2020: 241). Other approaches are starting to look
at the psychological (e.g. Kim, 2017) and emotional dimension of termin­
ology (e.g. Condamines, 2017), although the psychological approach has not
taken hold according to Delavigne and Gaudin (2022: 179), or at least not
yet. Another dimension of terminology that seems to be garnering interest
is the ethnoterminological one, as shown in some South American studies,
e.g. Costa and Gomes (2013), Mesquita and Brandão (2023) and Dourado
et al. (2018), and Chinese ones to some extent, e.g. Zhao et al. (2021). The
ethnoterminological approach focuses on studying the discourses of trad-
itional communities who have their own expert knowledge, which involves
local natural resources and which is captured in concepts referred to as
 Key Terminological Theories and Approaches 57

ethnoterms (Costa & Gomes, 2013: 252; Mesquita & Brandão, 2023: 195).
This approach is methodologically grounded in ethnolinguistics and ter-
minology. In China, Zhao et al. emphasise that “term databases have a key
role to play in protecting cultural heritages and also in disseminating them
internationally” (in Li & Hope, 2021a: 16), a finding that arises from Zhao
et al.’s (2021) work on developing a termbase of Manchu Ulabun terms, used
by the Manchu people of northeast China, who are currently in danger of
extinction.

What Have We Learned from These Theories?

Overall, we can appreciate two strong, common threads running through
all the theories and methods we have reviewed here. In one way or another,
almost every approach and theory since the GTT advocates for the multidi-
mensionality of concepts and variation of terms (including polysemy), and
that both multidimensionality and variation are unavoidable. In addition,
terminology is no longer viewed from the linguistic aspect alone but also
from many other, sometimes more human, perspectives, e.g. social, semantic,
textual, communicative, cognitive, cultural, pragmatic, ethnological, psy-
chological, emotional and more. It would seem as if terminology is truly
‘polyhedral’ and it would be difficult to ignore any of these dimensions in
terminological work, although each dimension may be more salient than
others in specific settings. The GTT leaves us with a good ideal to aspire to,
i.e. univocity, and a highly efficient method for terminology management,
i.e. concept orientation. However, our linguistic and terminological reality
is often much more complex than can be accommodated by the GTT, and
the notion of what constitutes an ‘ideal’ in terminology may well need to
be revisited, in light of the wide range of theories and approaches we have
explored in this chapter. Given such a multifaceted ‘enlargement’ of the GTT,
one cannot help but wonder “whether it can be regarded as an independent
theory or as a theory covering only parts of a superordinate terminological
theory” (Laurén & Picht, 2006: 166).
In the meantime, and in addition to multidimensionality, variation and
concept orientation, our takeaways for the terminological work of translators
include the fact that this work is focused on words, terms, texts and corpora.
We have also learned that concepts must be viewed broadly in this work,
whether we call them units of understanding, terminological units, frames,
percepts, ethnoterms or something else. Definitions and term relations can
convey a great deal of knowledge from the domain, as well as images,
diagrams and concept maps, and translators should capitalise on definitions
and term relations and visualisations of both to gather that knowledge for
future reference. Term relations may be hierarchical or non-​hierarchical. We
have also seen how metaphors are just as common in specialised discourse
as they are in general language discourse (although there may be language-​
specific differences), to the extent that metaphorical (and cultural) processes
58 Terminology Management for Translators

often underlie the creation of many new terms. It may not be entirely clear
how we should define ‘domain’ and where the boundaries are between
domains, but this may still be a useful construct for translators who have to
deal with large amounts of variation and polysemy in terminology and trans-
lation. When looking for terms, nouns and noun phrases are not the only
parts of speech to concentrate on; instead, verbs, adjectives and even adverbs
may be considered as potential terms and special attention should be paid
to the specific phraseology and conventions of a given domain. Finally, it is
worth reminding ourselves that terminological work may be time-​consuming
and labour-​intensive but the benefits of efficient and smart terminology man-
agement will far outweigh its downsides (see Chapter 1), especially since the
number of computerised tools available to help us in those two respects is
constantly increasing (see Chapter 6). In the following chapter, we explore
the ways in which we can apply these takeaways from terminological the-
ories to the design and use of termbases for translators.

Follow-​up Tasks and Reflection

1. Although nouns and noun phrases are the most frequent parts of speech
for terms, it is possible to find verbs, adjectives and adverbs which may be
terms too, for example:
a. Adjectives: inlaid in ornamental masonry; intrathecal in medical
domains; pneumatic in drilling equipment; insane in psychiatry, etc.
b. Adverbs: impartially in court proceedings; by decree in law-​making; on
board in aviation; thereinafter, hereunder, thereof and more in many
legal domains, etc. (Olohan, 2016: 130).
c. Verbs: cast in fishing; bore in tunnel construction; countersink in wood-
working, debug in IT, etc.
Since it may be difficult or time-​consuming to find more examples of
adjectives and adverbs which may be terms, can you do some research to
find more examples of verbs that may be terms in a given domain, as in
the examples above?
2. In the following text (from University of Nairobi, 2019), how many
knowledge patterns can you find which might help identify definitions?
For example, one such pattern might be referred to as in the sentence
Minerals occur in certain amounts of symmetry referred to as twinning.
In descriptive mineralogy, a crystal is defined as a solid body bounded
by plane natural surfaces, which are the external expression of a regular
arrangement of its constituent atoms or ions. A zone in a crystal consists
of a collection of set crystal faces that are parallel to a particular line
(or direction) termed as the zone axis. On the other hand, a zone plane
occurs at right angles to the zone axis. Now let us define a parallelpiped by
connecting any neighbouring lattice point in the point lattice.
3. How are the knowledge patterns in the previous exercise translated into
your other language(s)? Are there any differences?
 Key Terminological Theories and Approaches 59

4. Polysemy: we have seen that, even within the confines of a domain, a term
may be polysemic, e.g. program in computing. Do some research to find
out whether any of the following terms are polysemic in finance. The first
one has been done as an example.
a. Interest: polysemic, e.g. (i) the price you pay to borrow money, (ii)
amount of ownership a stockholder has in a company.
b. Deposit:
c. Blockchain:
d. Shark repellent:
e. Contingency fund:
f. Account balance:
5. Term variation: what are the differences, if any, between the following
groups of terms? What are their equivalents in your other language(s)?
a. Hard disk, hard drive
b. Waiver, disclaimer
c. Farrier, horseshoer
d. Pleural cavity, chest wall
e. Full blood count, complete blood count
6. Which theory might have uttered the statements below? Match the
statements (a to j) to the correct terminological theory/​approach (1 to
10) (solutions available on the online portal).
1. Communicative Theory of Terminology
2. Corpus-​based approaches
3. Cultural Terminology
4. Distributional Semantics
5. Frame-​Based Terminology
6. General Theory of Terminology
7. Lexico-Semantic Theory
8. Sociocognitive Terminology
9. Socioterminology
10. Textual Terminology
a. An extensive and in-​depth knowledge of the language and culture
of the community whose language is under-​resourced is required.
b. Each specialised domain has its own event template.
c. It is at the text level that the communicative situation, the domain
and especially genre should be considered.
d. Monolingual concordancers allow users to search for every instance
of a term in a corpus.
e. One choice of door (or focus) to access a term does not entail a
rejection of the other two doors or perspectives.
f. Researchers are increasingly considering a number of equally rele-
vant non-​hierarchical relations between terms.
g. The linguistic environment of a term in a corpus is used to gather
semantic information about is co-​occurrences with other words
and terms.
60 Terminology Management for Translators

h. The purpose of terminology is to be of use to society.

i. Units of understanding allow for prototypically structured under­
standing.
j. Univocity ensures effective and efficient communication.

Notes
1 Although it became quite common from the 1980s to talk about terminological
schools, this concept should not be understood as a unified group of scholars with
homogeneous views about terms and terminology. Instead, it is convenient way
to refer to “foci of interest governed by practical needs and different research
interests” (Laurén & Picht, 2006: 167). Laurén and Picht further state that the con­
cept school “does not have any importance today, not least because of its rigidity
and lack of a satisfactory definition” (2006: 169), although it continues to be used
in the literature.
2 Genre is a rather slippery concept but it is generally defined as “any style or form
of communication in any mode (written, spoken, digital, artistic, etc.) with socially
agreed-​upon conventions developed over time” (Thundil, 2024), e.g. literary, jour­
nalistic, legal, medical, blogs, etc.
3 www.nordt​erm.net/​
4 It is a usual convention in cognitive linguistics to use SMA LL CA PS for conceptual
metaphors, e.g. T I M E I S M ONEY (Valenzuela & Ibarretxe-​Antuñano 2023: 263).
5 Here we use ontology to mean “a set of rigorously defined terms and concepts used
to describe and represent a knowledge area, as well as sets of relations, properties,
and values,” as defined by Faber and San Martín (2011: 49).
6 According to Saussure’s theory (1916/​1995) (l’arbitraire du signe, the arbitrariness
of the linguistic sign), a language is made up of signs, which are made up of two
sides, both of which are inseparable: the signifier (i.e. the label) and the signified
(i.e. the concept referred to by the label). The relation between the signifier and
the signified is arbitrary, i.e. there is no direct connection between the label and
the concept it refers to. The fact that, in English, the label tree is used to refer

to the concept often depicted in our minds as something along the lines of , is
a result of convention, i.e. the speakers of the same language group ‘agree’ to use
that label to refer to that concept. In addition, concepts have no stable, core
meaning but can change over time.

Further Reading
The edited volume by Faber and L’Homme (2022a) is a thorough and up-​to-​date
resource to read more about many of the theories and approaches covered in this
chapter. Temmerman (2000: 1–​38) provides more insights into the three termino­
logical schools, as well as the Canadian approach with a useful comparison of
examples, which are then connected to the reasons for proposing the Sociocognitive
theory of terminology. She also simplifies the Vienna School’s principles to five. Picht
(2011) gives us a detailed account of the development of the Western and Eastern ter­
minological and linguistic traditions, as well as the Nordic terminological efforts and
the Canadian approach to terminology. For more on Eastern views and approaches,
see edited volumes by Manerko et al. (2014) and Picht (2006).
 Key Terminological Theories and Approaches 61

Cabré (2023a) is a collection of a selection of Cabré’s articles on terminology

published since 1999; it is a useful collection since some of the articles may be diffi-
cult to access separately. The selection “has been made on the relevance of the articles
and the non-​repetition of ideas within them” (Cabré, 2023a: vii). Articles originally
written by Cabré in French (e.g. 2000) and Spanish (e.g. 2005) are translated into
English in this volume: In Cabré (2005), translated into English by Besharat Fathi, we
find a very detailed overview of why and how Wüster developed his GTT theory and
the main foundations of Cabré’s Communicative Theory of Terminology; in Cabré
(2000), translated into English by Steven Norris, more can be read about her Theory
of Doors. A detailed biography of Wüster’s life and other insights into his views
on terminology can be found in Trojar (2017). Candel (2022) explains potentially
contradictory views in the GTT and why some of them may not be contradictions
after all. Humbley et al.’s (2018) edited volume covers Socioterminology in several
countries throughout Europe and Africa. Bowker (2022) explores how multidimen­
sionality has been dealt with in terminology. Li and Hope (2021a, 2021b) investigate
the emergence and development of terminological theory and practice in China. The
edited volume by Alsulaiman and Allaithy (2019) reviews many aspects of the termin-
ology in the Arab world, as well as its history.
According to Wüster, the idea of a conceptual network was started in linguistics
by Swiss linguist Ferdinand de Saussure and by Weisberger and Trier, who were the
founders of structural lexicology (Candel, 2022: 41). Saussure stated that the linguistic
sign6 is arbitrary, which means that words refer to general concepts in an unmoti­
vated, arbitrary fashion. For differing views on how the Saussure’s theory applies
to the relationship between concepts and terms in terminology, it is worth reading
Warburton (2021: 21), for whom terms are a sign in the Saussurian sense, whereas
others have different views. For example, Sager points out that “terms refer deliber-
ately to specific concepts within particular subject fields” (2001c: 259) and, according
to Diki-​Kidiri, “if the term is indeed a Saussurean ‘linguistic sign,’ the binary signifier/​
signified relationship [i.e. term/​concept] alone is insufficient to describe the term in all
its complexity” (2022: 205); for Desmet and Boutayeb, terminological units are “the
result of a convention, itself the result of an agreement between specialists of a given
specialised field” and they further claim that terms “are imposed by unified usage
within a milieu of experts, whereas words in general language are the reflection of a
collective memory” (1994: 307–​309). For a more elaborated discussion of the sign in
terminology from a semiotic perspective, see Myking (2001a).
4 Designing Your Termbase
Principles and Considerations

Key Questions and Issues

• How is a termbase organised?

• How do I store information in it?
• What are the “tricks of the trade” in terminology management?
• It is worth devoting time and effort to the design of the termbase as a
long-​term resource, as software does not usually help with termbase
design decisions.
• Creating and using a termbase needn’t be a struggle!
• The time required to store and manage information in a termbase
must be offset by gains in time and/​or quality of the ultimate product
(Marshman et al., 2012: 47).

How Do I Get Started?

Having reviewed a range of terminological theories and approaches, we can
now turn our attention to the more practical aspects of creating a termbase
and managing terminological information. However, in this chapter, we
will not provide step-​by-​step guidance on how to use any specific software
tool for creating termbases, as manufacturers already provide detailed
online user guides and manuals, although there will be some suggestions for
use and links to user guides in the online portal accompanying this book.
Instead, this chapter will focus on the key design principles and the consid-
erations translators should bear in mind when designing termbases, which
should then apply to creating termbases with any software tool.
As a general rule, you will need to have a clear idea of what kind of
termbase you want and what kinds of information you want to store in it
before you use any software. In other words, it is advisable for termbase

DOI: 10.4324/9781003302636-5
 Designing Your Termbase: Principles and Considerations 63

design decisions to be made in advance (see also Box 4.1). The user guides
of software programs provide procedural information, e.g. which options to
choose from which menus or which buttons to press, but they typically do not
provide conceptual information, e.g. how you should deal with synonyms.
Also, because software programs are normally designed to ‘protect’ your (ter-
minological) data, as opposed to erasing it, it may be difficult or laborious to
make substantial changes to the design of your termbase with some programs
once there is data stored in the termbase (Warburton, 2015a: 385). This may
improve with advances in technology but, in the meantime, if you need to
make substantial changes to the design of your termbase, you could con-
sider exporting the data from the old termbase into the new one, as we will
see later.

Concept Orientation
To start with, a key principle in terminology management is that termbases
for translators should be concept-​oriented and, software permitting, hier-
archically organised (see also Box 4.2). This means that termbases should
be organised around concepts, as opposed to terms, and ‘hierarchical’
means that, wherever possible, it is good practice to be able to distin-
guish between data that applies to the entire concept and thus the entire
entry, which is why it is referred to as entry level,1 (or concept level), e.g.
domain or image; data that applies to one language only, referred to as
language level (or index level); and term level, which is the lowest level
and contains data that applies to only one of the terms (or ‘labels’ for the
concept) within the entry, e.g. gender, part of speech, etc. (Rothwell et al.,
2023: 57–​58). Deciding what information goes at what level will depend
on the project’s or client’s requirements, or your own, if the termbase is
for personal use.
As we saw in Chapters 1 and 3, concept orientation is one of the main
differences between dictionaries (lexicography) and termbases for translators.
Because dictionaries are usually arranged alphabetically only, they “provide
no indication of systematic arrangement and consequently no information
about the relations between concepts in a given field” (Bononno, 2000: 651).
The differences between the lexicographical approach and the termbase
approach are illustrated in Figure 4.1.
This figure shows that, in contrast to the typical, alphabetical arrangement
of dictionaries, in a concept-​oriented termbase, gemstone, gem and precious
stone, from the domain of gemmology, for example, would all be collected
as the various labels or terms in a single entry, corresponding to one concept.
The termbase entry can also show information about how each term, e.g.
gem is a common abbreviation for gemstone and precious stone may be a
synonym or partial synonym in some contexts. The termbase entry for gem-
stone could be connected to the termbase entry of faceting (the process of
64 Terminology Management for Translators

Figure 4.1 Individual entries in a dictionary (left) and a concept-​oriented entry in a

termbase (right).

creating polished surfaces on a gemstone) by means of a cross-​reference, as

we will see later. Although termbases are also arranged alphabetically, cross-​
references can facilitate the retrieval of related concepts and their associated
information. Concept orientation, therefore, goes hand in hand with another
fundamental factor in terminology management, i.e. reusability. A “more
intelligent way of handling terminology is to design the system for reuse and
exchange from the start” and adopting a long-​term view for the managing of
terminology (Bononno, 2000: 653).

Data Categories and Fields

With a concept-​oriented design in mind, then, we now focus on how a term
entry is organised. The information in term entries is usually presented in dis-
crete units of knowledge often referred to as fields (Kageura & Marshman,
2019: 66; Olohan, 2016: 10), or descriptive fields, each field containing one
particular. type of data, known as data categories (Warburton & Wright,
2019), e.g. definition, context, source, etc. It is crucial for the systematic
retrieval and search of information, and for the reusability of the termbase,
that each field only contains the data category that it is designed to con-
tain (Rothwell et al., 2023: 59; Schmitz, 1996: 224). For example, the def­­
inition field should not contain the source of the definition, which should be
entered in the source field instead (see also “Data elementarity” in Box 4.2).
Importantly, a common error contravening the principle of data elementarity
is the inclusion of a term with its acronym in the same field. Instead, terms
and acronyms should be stored in separate Term fields, as acronyms are also
terms. Fields can be mandatory or multiple (see Box 4.3).
 Designing Your Termbase: Principles and Considerations 65

Box 4.1 How Do I Get Started?

• Determine the type of termbase you need, e.g. simple, complex, etc.
at least for now. To determine this, you should consider factors such
as your budget (for freelance translators), language and direction,
domain or subdomain, whether the domain will be a specialism for
you, the client, etc.
• Decide on the language variant now, to avoid data compatibility
issues later. For example, if you select English (Australia) as your
language variant, the data in this termbase would not be compatible
with a termbase in English (New Zealand), although they use the
same language. Termbase software often views language variants as
separate languages.
• It is important to look ahead and build a termbase that is scalable
and meets your needs well into the future (RWS, n.d.). Starting with
a simple glossary in Excel is a good way to start thinking about the
information and data categories you need in your termbase. Excel
files can be shared with others online and they can be a good ‘trial
run’ for a termbase that meets your needs.

There are potentially thousands of data categories to choose from, e.g. in

the Data Category Repository2 which originates from ISO standard 12620
(Warburton, 2021: 30) (see also Chapter 10), but there is no ‘one size fits all’
list of data categories for all translators to use in their termbases (Bononno,
2000: 658). Data categories for your termbase need to be selected carefully
based on a number of factors, ranging from the working situation (e.g. free-
lance or corporate) and long-​term or short-​term projects to client preferences,
users and uses of the terminology, etc. (Rothwell et al., 2023: 60). However,
according to Popiolek (2015: 348), most language services providers use
5 to 20 data categories whereas, according to TerminOrgs (a consortium
of terminologists working in large organisations), the two mandatory cat-
egories are term and language, although they strongly recommend using
also definition, context, part of speech and domain (which they call subject
field) (TerminOrgs, 2014: 5). Schmitz (1996: 226) suggests adding also note,
author, and date.
It makes sense for languages such as English to include a field for part of
speech, because it helps differentiate homographs, such as scan as a noun
and as a verb, which should be entered in different entries (Rothwell et al.,
2023: 62), and because it helps automated processing tasks, such as applying
grammatical filters (TerminOrgs, 2014: 5). Domain is another important
field, e.g. in a company’s termbases and “to ensure repurposability of the
termbase in sophisticated applications” (Warburton, 2015a: 386). Notes is
also important because it can help you capture any information that was not
66 Terminology Management for Translators

envisaged in the original design of the termbase. However, you should also
ensure that this field does not become too large (e.g. by breaking the prin-
ciple of data elementarity) and therefore difficult to consult. If that happens,
it would be good practice to export your termbase to a newly created one
with a larger range of fields and data categories, to accommodate some of the
information contained in the old Notes field.

Box 4.2 Key Principles and Considerations for Termbase Design

• Concept orientation: the entries in a termbase should be organised

around a concept, i.e. all the terms that designate the same concept,
including synonyms, should be gathered in the same entry.
• Reusability and repurposability (Warburton, 2021: 28) are the
principles whereby, from the very beginning, you should design your
termbase so that entries are easy to consult, update and be reused
by others (Bononno, 2000: 652). The termbase should then serve
as an efficient repository of terminological data for you, your client
and any teams you may be working with. One way to compromise
the reusability of a termbase might be to avoid including hundreds
of inflected forms in multiple languages, for example (Warburton,
2021: 144). Instead, separate, personal termbases or glossaries for
inflected forms can be created, which might also be connected to a
CAT tool, which might work for freelance translators, but there may
be some issues with this practice in commercial settings (2021: 144).
In order to be reusable, termbases should comply with standard
formats, such as .TBX (Di Nunzio & Vezzani, 2021: 183) or capable
of being imported and exported through Excel.
• Term autonomy: the principle whereby each term in an entry can
be documented independently and described with the same types
of information as any other term (Melby, 2015: 399). This means
that even acronyms and abbreviations must be entered in their
own Term field, rather than in the associated information fields
of another term, as acronyms and abbreviations are terms in their
own right. As explained by Warburton (2018: 124), if required, a
picklist may be added with values such as “abbreviation, acronym,
full form, spelling variant,” for example, to describe the type
of term.
• Data granularity: the principle whereby the level of detail in the
entries is balanced, as should be the number of data categories
and sub-​ categories (Tamás & Sermann, 2019: 35; Warburton,
2021: 27).
 Designing Your Termbase: Principles and Considerations 67

• Data elementarity: each field in the termbase should contain only one
type of information, i.e. “separate fields are required for different
types of information” (Warburton, 2015b: 12). For example, the
Source of a definition should not be included in the Definition field;
instead, there should be a separate field for Source.
• Data integrity: Preserving the quality, coherence and relevance of
the data by minimising the number of errors introduced, by using
picklists wherever possible (Warburton, 2021: 27–​28), for example.
• Directionality and Reversibility: will your termbase be fully revers-
ible, or only meant to be read in one direction, e.g. English into
Portuguese and not Portuguese into English?
• Termbase obsolescence: if your termbase becomes too large, it risks
becoming a “museum” (Sager, 1990: 209); therefore, you should
constantly check and, if necessary, update, the contents of your
termbase and develop a clear policy about the place of obsolete
terminology.

Date, source and author are useful for traceability purposes, although
some programs add this metadata automatically to the entries. Other
fields that could be useful for translators include contextual (or ‘encyclo-
paedic’) information, i.e. any background information about the concept
or any aspect of it that is not collected in the Definition or other fields; geo-
graphical variant is another useful field if you are planning to work with
more than one language variant, e.g. Belgian French and Ivorian French;
synonyms (other than geographical variants) and semantic relations in
general, and phraseological information, e.g. verbs that collocate with
noun terms etc.
Data categories can be grouped into three broad groups: conceptual,
terminological and administrative (Bononno, 2000: 660; Rothwell et al.,
2023: 61; Schmitz, 1996: 224; Warburton, 2021: 30). Conceptual data cat­­­
egories provide semantic information about concepts, e.g. definition, domain,
whereas terminological data categories provide information about terms, e.g.
part of speech, context or example, etc. Administrative data categories can
record information about the creation of the term or when it was last modi-
fied, for example. An example of data categories arranged hierarchically in a
termbase entry is shown in Figure 4.2. In this figure, entry level includes code,
domain, graphic and source of graphic. At language level, there is English.
The remaining fields are at term level, from full form to notes. The underlined
term, wavelength, denotes a cross-​reference to that entry.
68 Terminology Management for Translators

Figure 4.2 Example of monolingual term entry for refractive index.

A list of some useful data categories for each group includes the following:

• Conceptual data categories: definition, example or context of usage, source

(of the definition and of the example), domain, subdomain, encyclopaedic
or contextual information, multimedia (image, sound, video), etc.
• Terminological data categories: term, part of speech, acronym, abbrevi-
ation, geographical variant, gender (for relevant languages), status (e.g.
preferred, forbidden3), register (e.g. lay, specialised, etc.), paradigmatic
(e.g. synonyms, hyponyms/​ hyperonyms) and syntagmatic (collocations
and relevant phrases) relations, etc.
• Administrative data categories: unique entry code, date created, date last
modified, author, etc.
• Other: Notes, which can be used for any field, to refer to concepts, terms
or metadata.

Further, data categories can contain four different kinds of informa-

tion: numeric, free text, multimedia and picklist. Numeric fields can be used
to enter figures, such as dates, or a unique identification code for the entry.
Free texts is used for many data categories, e.g. definition, context, notes,
etc. The Multimedia field, as the name suggests, when available, allows
translators to add image, video or sound files to the term entry. Picklists are
lists of pre-​defined values which “are typically suited to fields which have a
 Designing Your Termbase: Principles and Considerations 69

limited number of options available” (Rothwell et al., 2023: 60), such as part
of speech, gender or status (e.g. forbidden, accepted, etc.). One advantage of
using picklists is that they ensure consistency in the visualisation of data, and
they are a time-​saving method of entering certain types of data.

Box 4.3 Mandatory vs Multiple Fields

• Mandatory fields are those that stop users from saving an entry until
some information has been entered into the field, i.e. fields that are
compulsory to fill in. Warburton (2021: 165) suggests keeping man­
datory fields to a minimum and/​or restrict them to fields that are easy
to complete. This is because they can disrupt the work of users by
not allowing them to save the entry, even if the entry is incomplete.
She also explains that, in practice, many termbases do not have any
mandatory fields, although TerminOrgs recommends making fields
such as part of speech and source mandatory.
• Multiple fields are those that can be used several times within one
entry. For example, a field such as source can be made multiple so
that it can be used as a source of a definition, of an example, of a
context, or of multimedia material. Alternatively, separate fields can
be created, e.g. source of the definition, source of the image, source
of the example, etc., each of which can also be made multiple if
necessary.
• Neither mandatory nor multiple: of course, some fields may not be
mandatory (because they can be filled in later) or multiple (because
you may only want to use them once per entry, e.g. Client). In fact,
many of the fields in your termbase will probably not be mandatory
or multiple.

Three examples of which data categories might be selected for termbases

under different circumstances or working situations are shown in Table 4.1.
These examples are neither exhaustive nor prescriptive, and can be arranged
differently depending on the requirements of the work. Several other ways of
naming and arranging the data categories in a termbase could be devised. Not
all fields need to be filled in with information straightaway; for example, the
definition of a term and its source might be added at a later stage. Data cat-
egories such as Author and Date of creation might be automatically added by
the software. Acronyms, full forms and abbreviations should all be entered in
the Term field (see Term Autonomy in Box 4.2). The Example of usage just
needs to be one useful sentence showing the term in use. It may seem onerous
to translators to use all the fields suggested in the right column in Table 4.1,
but “its use will pay off in the long run” (Bononno, 2000: 654). Although
70 Terminology Management for Translators

Table 4.1 Three examples of data category sets for termbases

Essential data categories Data categories for more Data categories for
comprehensive termbases long-​term termbases
• Unique ID or code • Unique ID or code • Unique ID or code
• Term (in each • Term (in each • Term (in each language)
language) language) • Part of speech (picklist)
• Part of speech • Part of speech • Example of usage
(picklist) (picklist) • Source of the example
• Example of usage • Example of usage • Notes
• Source of the • Source of the • Domain
example example • Subdomain
• Notes • Notes • Definition
• Domain • Source of the definition
• Definition • Encyclopaedic
• Source of the information
definition • Synonyms (and other
paradigmatic relations)
• Notes about synonyms
• Image (or video, or
audio, depending on the
domain)
• Phraseology/​collocations
(and other syntagmatic
relations)
• Date of creation
• Author
• Term status
(picklist: preferred,
obsolete, etc.)

definitions are important sources of knowledge for translators, they have

been left out of the first column in Table 4.1, as definitions might not be
needed for short-​term, basic termbases, e.g. working for a one-​off translation
project but, of course, they can be added if necessary. In the next section, we
take a closer look at definitions.

Definitions
Definitions establish the link between terms and concepts and, for some,
they are one of the most valuable parts of translators’ termbases (Bononno,
2000: 650; Muegge, 2007: 19). As we saw in the previous chapter, the
‘proper way’ of defining a term in the classical or Wüsterian approach was by
following the strict rule of genus +​differentia, where the definition starts with
the genus, which corresponds to the (closest) hyperonym (or superordinate)
of the term to be defined (or definiendum) and differentia corresponds to
the ‘necessary and sufficient’ characteristics (or definiens), of the term to be
defined, compared to those of its co-​hyponyms (or co-​subordinates) (Béjoint,
 Designing Your Termbase: Principles and Considerations 71

1988: 364) (see also Chapter 4). This type of definition is also referred to
as intensional or ‘X is a type of Y,’ and has been traced back to ancient
Greece (Díaz Vázquez & Guerra Aranguren, 2019: 5). By contrast, an
extensional definition consists of listing all the items in the category. For
instance, regarding the concept of planet, the differences between intensional
definitions and extensional definitions can be seen in the following examples
(from ten Hacken, 2015: 11):

1. Planet: celestial body which is in orbit around the Sun (intensional

definition).
2. Planet: Mercury, Venus, Earth, Mars, Jupiter, or Saturn (extensional
definition).

In other words, the intension of a concept is the “set of characteristics

which makes up the concept,” whereas the extension of a concept is “the
totality of objects to which a concept corresponds” (Díaz Vázquez & Guerra
Aranguren, 2019: 5). Intension tends to restrict definitions to specialised
domains, making them often more precise and succinct than general lan-
guage definitions. By contrast, general language definitions tend to be longer
and often contain qualifiers such as usually, generally, typically, etc., which
“make the boundaries of the meaning appear fuzzy” (Béjoint, 1988: 362).
Intensional definitions may work well in precise domains such as engin-
eering but, as sociocognitivist Temmerman (2000: 81) demonstrated, there
are difficulties in trying to apply the strict classical criteria to the defin-
ition of blotting in biotechnological domains such as molecular biology.
Terminologists are interested in intensional definitions because they allow
them to “understand how the concept is related to other concepts” (Bowker,
2022: 128) and can help them devise terminology products for users with a
variety of needs (Bowker, 2022: 134). However, from the translators’ view­
point, with termbases that are specific to them and/​or their (relatively small)
translation teams and clients, there is no need to make the extra effort to
be exhaustive and self-​contained with definitions. Instead, a more useful
working method might be to collect essential defining or describing informa-
tion about the term in the field Definition and leave more informative, nice-​
to-​know, encyclopaedia-​like information in another field of the termbase,
perhaps named Encyclopaedic information, for example. Translators only
need to resort to intensional definitions when the need for precise concept
boundaries arise, e.g. in scientific claims or legal disputes. See more examples
in ten Hacken (2015: 10). Distinguishing between the information to include
in definitions and that for additional contextual information will depend on
each translator, the relevant translation project and its readership, and the
client.
The use of definitions also varies according to the translator’s working situ-
ation. For example, the use of definitions is advisable for freelance translators
wishing to develop a specialism, whereas for translators working in-​house
72 Terminology Management for Translators

in commercial or institutional settings, the use of definitions will depend on

the specific requirements of the business or organisation, their internal pol-
icies and their clients. For example, Swedish company Scania uses definitions,
which are validated by native English speakers (Bucher, 2017: 278), whereas
another Swedish company, Ericsson, did not use definitions at all in its
Eriterm termbase until the 1990s (Jaekel, 2000: 159–​162).
Since definitions for translators’ termbases generally fulfil the function
of collecting specialised information for future reference, to avoid repeating
complex term searches, for example, definitions should be sought from
authoritative sources from the relevant domain. In most situations, “it is
not the terminologist’s (much less the translator’s) responsibility to prepare
definitions” anyway (Bononno, 2000: 656). Therefore, if translators have
access to domain experts, definitions could be prepared by those experts.
When selecting definitions to add to your termbase, you should ensure
that you do not use circular or tautological definitions. Circular definitions
are those where “two concepts are defined in terms of the other” (de Bessé,
1997: 71). For example, the definition of textile industry as ‘branch of the
industry that produces textiles’ is circular. For your termbase, finding non-​
circular definitions such as ‘the industry primarily concerned with the design,
production and distribution of cloth and clothing’ would be much more
useful. Tautological definitions are very similar: in a tautological definition,
the component words of the term are used to define the concept, e.g. defining
roll-​call vote as ‘a vote carried out by roll call’ (de Bessé, 1997: 71), thereby
failing to explain what the concept means.

Relations between Terms

Terms are connected to other terms by meaning. Relationships of meaning
between terms are referred to as semantic relations.4 These relations are
“extremely important for terminology management” (Granda & Warburton,
2001: 4) and, like definitions, they are “central elements in users’ acquisition
and understanding of new concepts and domains” (Marshman, 2022: 291),
as well as in understanding how terms are related within a given concep-
tual network. Making relations visible in a termbase helps enable knowledge
transfer (Nuopponen, 2022: 66).
It is generally agreed that there are two broad kinds of relations, i.e. syn-
tagmatic and paradigmatic (L’Homme, 2022: 253). Paradigmatic relations
include hierarchical (also known as generic) relations such as hyponymy/​
hyperonymy (a hyponym is a kind of hyperonym, e.g. a seagull [hyponym]
is a kind of bird [hyperonym]) and meronymy/​holonymy (or partitive), e.g.
a meronym is part of a holonym, e.g. a petal [meronym] is part of a flower
[holonym]). Paradigmatic relations also include synonymy, partial synonymy
and antonymy, which are ‘horizontal’ (Faber, 2022: 362) as opposed to
‘hierarchical.’
 Designing Your Termbase: Principles and Considerations 73

Box 4.4 How to Deal with Polysemy and Homography?

• Polysemy: words or phrases that have two or more (usually related)

meanings are said to be polysemous. Monosemy (i.e. having only
one meaning) is actually quite rare (Temmerman, 2000: 36). For
example, in educational domains, thesis is polysemous because it
can mean a line of argument used to justify a theory or statement,
and it can also mean a long essay or dissertation submitted to an
educational institution in order to obtain a degree. Although both
meanings are within the same domain and have the same label, they
are in fact different concepts and therefore should be entered in sep-
arate term entries.
• Homography: homographs have the same spelling but do not have
related meanings. For example, a bat can be a nocturnal flying
mammal and a piece of equipment in baseball. In this case, we have
two concepts in two domains, and we should create separate entries
in the termbase, one for the domain of mammals, for example, and
one for the domain of baseball.
• What about terms with more than one part of speech? In the case of
scan, for example, which can be both a noun and a verb, separate
entries should be created, one for the noun form and one for the verb
form. This is the case for many other nouns/​verbs in English, e.g.
mould, pile, form, supply, strip, support, tread, span, slope, wire,
force, etc.

The reason for creating separate entries in each of the cases above
is because, although the words may look the same in English, they
are different concepts that might be rendered differently into other
languages and the word or expression in other languages may come
with a whole set of phraseological and other domain conventions,
as well as collocations, etc. It may be confusing sometimes to estab-
lish a clear difference between polysemy and homography/​homonymy
(L’Homme, 2020b: 417), at least in English, but they will probably need
to be entered in a translators’ termbase with separate entries anyway,
because the termbase is about concepts rather than terms and different
meanings usually point to different concepts. In other words, in termin-
ology management, the fine differences between polysemy and hom-
onymy [and homography] are “not viewed as important because the
two phenomena result in the same pragmatic word behaviour: the same
word form represents more than one concept” (Antia et al., 2005: 12),
which results in separate termbase entries.
74 Terminology Management for Translators

Syntagmatic relations are non-​hierarchical and include collocations and

associative relations (L’Homme, 2022: 253). Associative relations, which link
concepts spatially or temporally (Pavel & Nolet, 2001: 15), include a wide
range of semantic relations such as cause/​effect (e.g. chemotherapy/​hair loss),
process/​instrument (e.g. incision/​scalpel), material/​state (e.g. iron/​corrosion),
etc. (Sager, 1990: 34–​35). Nuopponen (2022) lists many more associative
relations and provides further insights into them. Although associative
relations have not been given much attention in the literature (Marshman
et al., 2012: 30; Nuopponen, 2006: 1), likely because the Wüsterian theory
of terminology only allowed for hierarchical relations for a long time (see
Chapter 3), subsequent theories opened the doors to the vast array of associa­
tive relations that are in use today. In fact, associative relations are the most
frequent relations between terms (L’Homme, 2004: 11; Thelen, 2002a: 194)
and also the most difficult ones to formalise (Thelen, 2002a: 194).
Including both paradigmatic and syntagmatic relations in your termbase is
always a good idea, despite the fact that computer tools are not able compre-
hensively and reliably to identify them in all cases (Marshman et al., 2012: 30)
and the fact that you will often not see term relations in the entries of some
termbases and term banks you may consult (2012: 30), since their term
entry structures may have been created decades ago. The types of relations
to include in your termbase depend on a number of factors, one of which is
the nature of the domain. For example, in anatomy, part-​whole relationships
(meronym/​holonym) abound whereas, in domains such as pathology and
physiology, there are mainly associative relations, especially procedure/​
effect and process/​cause (Sager, 2001a: 252). In order to avoid duplication
of research effort, any information you find about relations between the
terms in your source text should be stored in the termbase for future use.
Marshman et al. (2012) provide examples of how to extract information on
term relations and making it available in your termbase. One of the findings
from their research is that “the benefits of including terminological relations
in many cases will outweigh the modestly increased workload, and that (as is
the case with translation memories) the gradual accumulation of information
will ultimately form a useful resource,” although “each situation is different
and the return on investment may vary depending on user needs and situ-
ation of use” (Marshman et al., 2012: 47).

Equivalence
Equivalence has been long studied in translation but much less so in termin-
ology (Faber & L’Homme, 2022b: 11). Although translators are formally
trained to avoid word-​for-​word substitution as a general method of transla-
tion (Pinchuck, 1977: 32), this approach should be even stricter when applied
to terms. Equivalence is a complex phenomenon and, just like translation,
goes far beyond finding one-​to-​one word correspondences between source
 Designing Your Termbase: Principles and Considerations 75

and target languages. The same applies to finding the correct equivalent for a
term in the source text. In other words, terms should not be translated word-​
for-​word, without any research in both source and target languages, in order
to avoid errors, mistranslations and sometimes inappropriately standardising
irrelevant terminology (León-​ Araúz & Cabezas García, 2020: 222). For
example, translating wind power into Spanish word-​for-​word as fuerza del
viento (‘power of the wind’) would be far from the correct Spanish term
energía eólica (‘wind energy’). Although translating terms literally might
seem unrealistic for English terms open jaw flight or shoulder season, in the
domain of tourism, translators should always carry out some terminological
research to find the correct terminological equivalents in the target language.
In the case of open jaw flight and shoulder season, research might be impera-
tive but determining the correct equivalent in the target language should be a
matter of design, not coincidence or accident.
It is important to stress, therefore, that terms should not be translated liter-
ally (León-​Araúz, 2022: 490). Instead, the ideal approach would be to carry
out monolingual research in each language separately and establishing bilin-
gual (or multilingual) correspondences later. However, it may not always pos-
sible or practical to achieve this ideal. Corpus-​based analysis is one tool that
can help translators discover or verify conceptual structures in each language
(León-Araúz, 2022: 490). However, it is not entirely clear how equivalence
should be determined because there are different degrees of equivalence and
sets of variants. For example, León-​Araúz (2022: 482) found that some studies
establish two types of equivalence, whereas others divide equivalence into five
types. The simplest classification is probably into three types: full equivalence,
partial equivalence, and non-​equivalence (2022: 482). Examples include:

1 Full equivalence: when “terms in different languages point to concepts

that occupy the same position in all systems” (León-Araúz, 2022: 482),
for example leukocyte in English and leucocito in Spanish.
2 Partial equivalence: León-​Araúz (2022: 482) distinguishes two types of
partial equivalence: conceptual overlapping, when two concepts have
much in common, and inclusion, when one concept is larger in scope
than the other. An example of conceptual overlapping is the equivalence
between the English pier (structure built into the sea for walking on for
pleasure) and Spanish embarcadero (a structure, sometimes natural, some-
times built up, where vessels stop for loading and unloading goods). An
example of inclusion is in the English river and the French rivière and
fleuve, because the English terms makes no distinction between rivers that
flow into the sea (fleuve) and rivers that flow into other rivers (rivière). In
other words, river ‘includes’ both rivière and fleuve.
3 Non-​equivalence: when there is not a common concept in source and
target cultures, usually due to (the lack of) culture-​specific terms or other
lexical gaps (León-​Araúz, 2022: 482). Non-​equivalence is often a problem
76 Terminology Management for Translators

in legal translation, where concepts are rooted to a national legal system

and subject to the moral values predominant in that society at a par-
ticular period in time (Sandrini, 1996: 344). For example, as explained
by Alaoui (2023: 211), “there are five legal values characterizing behavior
in Islamic law, namely reprehensible (makrūh), mandatory (wājib), pre-
ferred (mandūb), permissible (ḥalāl), and prohibited (ḥarām),” whereas
in Western-​type laws, there are only two, i.e. legal and illegal. Another
example is yamen(衙门), which is unique in its function and role in
imperial China; it is “not an easily translatable term and has no equivalent
in English” (Cao, 2023: 177).

The real challenge for translators lies in partial equivalence and non-​
equivalence because each requires extensive research to come up with a rele-
vant solution in a given context (León-​Araúz, 2022: 494). Interestingly, as
suggested by Alaoui in Biel & Kockaert (2023: 7), a way to “avoid masking
conceptual asymmetries is to provide translators with well-​organized term
bases.” True terminological equivalence, in sum, may be “difficult, if not
impossible, to establish without a detailed description of the nuances of
meaning projected by term variants that designate the same concept” (Faber
& L’Homme, 2022b: 11). We explore term variants in the following section.

Term Variation
Variation is a feature of both general and specialised language. Although the
Wüsterian GTT acknowledged the existence of variation (albeit only in a
very measured way) (Candel, 2022: 50), its proponents considered synonymy
and polysemy as “impediments” to specialised communication (Bowker,
2019: 582). We have come a long way since Wüsterian times because now
it is generally accepted that terminological variation is important and wide-
spread in specialised domains, as well as general language. Normally, the
less specialised a text is, the more variation of designation we expect (Cabré,
2003: 179). According to Daille (2003: 31), variation can account for as
much as 35% of a text, depending on the domain. Variation is important to
translators for several reasons:

• Different contexts and situations will require different variants of a term

(Campo & Cormier, 2005) and translators need to know which variant
fits which context.
• Terms have a cultural dimension that cannot be ignored (Faber, 2022: 371)
because it leads to different term variants.
• Variation in the use of terms is not random (Delavigne & Gaudin,
2022: 188; Freixa, 2022: 400); it reflects “circumstances, the role of the
speakers, and their position with regard to other users in the communi-
cative context (client vs. vendor, doctor vs. patient, patient vs. user) etc.”
(Delavigne & Gaudin, 2022: 188).
 Designing Your Termbase: Principles and Considerations 77

• Variation also “points to the fact that concept systems and definitions are
not static” (Faber, 2009: 113) (see also Chapter 1).
• Variation in terms is often the result of multiple motivations that take
place during the term naming process (Fernández-​Silva & Kerremans,
2011: 321). Therefore, an expert may consciously choose a term variant
over another at one time, in order to emphasise a particular dimension
of the term, and a different variant another time, to emphasise a different
dimension. This choice can give some clues to the translator as to how
the expert views the concept (Fernández-​Silva et al., 2014: 183; Freixa,
2022: 401).
• Both single-​word terms and multiword terms can vary (León-​Araúz &
Cabezas García, 2020: 213).

It is somewhat ironic that the GTT mostly ignored variation for the sake
of precision in specialised communication even though “variation often
emerges for the same reason, since new ways of conveying meaning are con-
stantly sought for” (León-​Araúz & Cabezas García, 2020: 214) in order to
be as precise as possible in the transfer of knowledge. There are different typ-
ologies of variation in the literature, but some of the kinds of variation that
might be important for translators include the following:

• Diachronic variation: variation of terms over time; for example, what we

now know as yellow fever was known as saffron scourge (and other names)
during the 19th century (Hall, 2018), so yellow fever and saffron scourge
are diachronic variants. In a more recent example, while the French term
pêcheur “would have been systematically translated into English as fish-
erman, it is now much more common to use the gender-​neutral fisher”
instead (Simard, 2020: 87). Both fisherman and fisher can be collected
in the same term entry, with relevant notes for use or avoidance (see also
‘termbase obsolescence’ in Box 4.2).
• Geographical (or diatopic) variation: different terms are used in different
geographical areas, e.g. gemology in the US vs gemmology in the UK.
• Dimensional variation: terms that “point to the same concept but high-
light a different property or dimension” of that concept, e.g. photochem-
ical smog vs. summer smog (León-​Araúz et al., 2020: 2359).
• Cosmetic (or spelling or orthographic) variation (Warburton, 2015a:
382): including capitalisation, hyphenation and the use of numbers (Ahmad
& Rogers, 2001: 754), e.g. Bronze Age vs Bronze age, email vs e-​mail,
four-​cup anemometer vs 4-​cup anemometer.
• Diaphasic variation: formality-​ related variation, often connected to
register, arising from the context of use. For example, in legal contexts,
deceased may be used instead of the less formal variant dead. Compare
to diastratic.
• Diastratic variation: variation arising from differences in users’ social class,
age, education, etc. Diastratic variants are also referred to as sociolects
78 Terminology Management for Translators

(e.g. De Vecchi, 2020: 241; Freixa, 2022: 409). For example, brown bread
is the diastratic variation in London’s Cockney (rhyming) dialect for dead.
Thus, deceased is a diaphasic variant of dead, whereas brown bread is a
diastratic variant of dead. To illustrate further, a doctor giving evidence
in court might use deceased (diaphasic) whereas brown bread (diastratic)
might be used by the same doctor when speaking to family and friends in
informal contexts. Dead would be a more neutral variant of deceased and
brown bread.
• Abbreviations and acronyms: it would be difficult to conceive terminology
without the use of acronyms. In some settings, e.g. the European Union,
the production of acronyms takes place on such a massive scale that “at
times glossaries have had to be compiled especially to deal with them”
(Cosmai, 2014: 63). Cosmai also points out that sometimes acronyms are
not easy to identify. For example, the name for the Erasmus programme,
“is not only a worthy homage to the great Dutch humanist, but also
stands –​almost perfectly –​for the European Community Action Scheme
for the Mobility of University Students.” The fact that some expressions
have acronyms may be an indication that they are terms in a domain
(as opposed to general language words) (L’Homme, 2022: 238–​240).
Abbreviations are also very common in terminology but difficulties may
arise “over the recognition of short or abbreviated forms of terms which
are contextually conditioned” (Sager, 1990: 149), e.g. recognising clamp
bolt as an abbreviated form of dynamo strap clamp bolt (Sager, 1990: 215)
and, of course, the use of acronyms and abbreviations may differ from one
language to another.

It may not always be straightforward for translators to assign the correct

variant label to a term, but more relevant to translators than assigning labels
is the need to ensure that the correct variant is used in the correct context,
taking into account the client’s profile, needs and preferences, as well as
the target readership, which may vary from project to project. Therefore,
collecting information about variants in a termbase, with or without the
correct labels, is crucial for the quality of translations, although the above
list may help categorise variants to some extent. The downside of term vari-
ation is that it may impede the leverageability of the translation memory
(Warburton, 2015a: 382).
It is difficult to separate equivalence from variation in that all the term
variants above illustrate intralingual partial equivalence, and are therefore
partially synonymous (except perhaps acronyms and abbreviations which are
considered by some as true synonyms), but these examples show the extent
to which variation pervades specialised language. Equivalence and variation
call for high sensitivity on the part of the translator to identify instances of
these in the source texts and appropriately to render them into the target
language. Sometimes, term variation is “the result of multiple motivations
that take place in the naming process” (Fernández-​ Silva & Kerremans,
 Designing Your Termbase: Principles and Considerations 79

2011: 321); the ‘naming process’ refers to term formation, to which we turn
our attention next.

Term Formation
Term formation is the process of how individual terms are created as
designations for a concept (Kageura, 2022: 461). Term formation is required
when new inventions or discoveries are made and, at first glance, it would
seem that creating terms is not in the remit of translators (whose job is to
‘translate’ rather than ‘name’). However, as we saw in Chapter 1, specialised
communication relies heavily on translated terminology and, therefore, it
should not be surprising that “translators play a critical role in the import-
ation and consolidation of neologisms” (Prieto Ramos & Morales Moreno,
2019: 87), to the extent that translators have been called pollinators of science
(Fischbach, 1993). This applies particularly to institutional translators who
have an especially instrumental and potentially influential role in dissem-
inating neologisms, e.g. World Health Organization translators translating
coronavirus or COVID-​19 into other languages. In commercial translation
settings, by contrast, translators (or terminologists) “are rarely called upon
to create terms” (Warburton, 2021: 212). Instead, they may be involved in
finding a target language version of the names of new products, objects,
features, etc. (2021: 212).
As English “has been used as a de facto lingua franca in many subject
domains, primary term formation is expected to occur more frequently in
English, whereas secondary term formation is common in other languages,
including borrowings” (Kageura, 2002: 462). In primary term formation,
“a new concept is named with no direct linguistic precedent” (2002: 462),
whereas secondary term formation “occurs when a new term is created for
a known concept […] as a result of knowledge transfer to another linguistic
community” (Sager, 1990: 80).
In this section, we will briefly review some of the mechanisms of term
formation (at least in English) as a brief source of reference for translators.
Of course, a major challenge for translators is the fact that term formation
processes may differ from one language to another, and knowledge of how
terms are created in their language(s) is an invaluable asset. Neologisms,
or new terms, may be new words or existing words that are assigned new
meanings (Pavel & Nolet, 2001: 20). They can be created through one or
more of the following processes:

• Derivation: creating new words by adding prefixes or suffixes, e.g.

upstream, shrinkage, loading, etc. (Gerzymisch-​Arbogast, 2008: 17).
Within derivation, we also find the process of conversion, which consists
of “changing the grammatical category of the word without affixation”
(Daille, 2017: 14). For example, google as verb derives from Google
as noun.
80 Terminology Management for Translators

• Compounding: the combination of existing words into new ones (Sager,

1990: 72) e.g. railway tunnel, database, keywords, etc.
• Within compounding we also find neoclassical compounds, which consist
of at least one element of Greek or Latin etymological origin e.g. patho,
bio, logy (Daille, 2017: 16). Classical Latin and Greek “have provided
scientists and technologists speaking European languages with basic root
forms for a large number of terms” (Sager, 1990: 86), from medical terms
to names of plants and animals (Klerkx, 1990: 216), e.g. Rosa rubiginosa
for rose. Some knowledge of common Latin and Greek roots and affixes
used to form medical terms “is extremely useful in analyzing the meaning
of English medical terms, which generally consist of some or all of the
following components” (Wakabayashi, 1996: 359):
• word root: this usually represents the organ of the body;
• prefix (although not all words have a prefix);
• suffix: this describes the disease process or the operation;
• combining form: a vowel linking the root to the suffix or to another root.
Thus, cholelithotomy can be analysed into the following components:
• chole: the root, in this case meaning ‘gallbladder;’
• lith: the root for ‘stone;’
• otomy: suffix meaning ‘incision.’
Therefore, cholelithotomy means ‘an incision into the gallbladder
to remove a stone.’
• Conjunction: the “process by which two concepts are combined as equals
in a new concept and this fact is reflected in the term” (Sager, 1990: 66),
e.g. socio-​cultural, condenser-​receiver, gas and air mixture, etc.
• Disjunction: as conjunction but denoting “or,” e.g. with “either-​or” or
“whether or not” relationship (see Sager, 1990: 67), e.g. left-​right indi­
cator, on/​off valve, drive-​on/​drive-​off ferry, etc.
• Compression: “any form of shortening of an expression by means of
abbreviation, clipping, acronyms, etc.” (Sager, 1990: 72), phone for tele­
phone, DNA, etc. The names of some institutions and organisations are
abbreviations that “have taken over so completely that the full form is prac-
tically never used” (Sager, 1990: 79), e.g. UNESCO, NATO, etc., and are
sometimes written as words rather than as acronyms, e.g. Unesco, Nato.
• Metaphors: as we saw in Chapter 3, one of the objects of study for cog­
nitive theories of terminology is metaphorical understanding, which has
led to the creation of new terms from metaphors, e.g. the cloud to refer
to the Internet, wormholes in astronomy, etc. As well as the creation of
individual terms, metaphors can also lead to patterns of term creation.
For example, Ureña & Ruiz (2022: 390) found that the conceptual meta­
phor SEA O R G A N I SMS A RE J OB HO L D E RS is instantiated in terms such as
surgeon fish, angler fish, pilot fish, Spanish dancer, inn-​keeper worm,
hermit crab, etc. Similarly, in the domain of risk analysis, a metaphor of
Swiss cheese was used extensively in the literature after James Reason and
colleagues introduced it (Reason, 1990; Reason et al., 2006). The Swiss
 Designing Your Termbase: Principles and Considerations 81

cheese metaphor is used to illustrate that, although there may be many

layers of protection against hazards and accidents, there may be flaws in
the system too (i.e. holes in each layer of cheese) that can allow the acci-
dent to happen. Later on, Li & Thimbleby (2014) developed the Swiss
cheese metaphor into a fondue pot5metaphor, where the pot acts as an
incident reporting and learning point, collecting all the drips of hot cheese.
• Borrowing and loanwords: it is well known that many languages ‘borrow’
words from English, especially in domains such as computing, e.g. soft-
ware, byte, etc. However, English has also borrowed words from many
languages, e.g. typhoon is from Japanese 台風 (Kageura, 2022: 462),
banana comes from Wolof, alcohol from Arabic, sauna from Finnish, feng
shui from Chinese, etc.
• Blending (or portmanteau): “forming a new term by joining parts of
existing terms” (Warburton, 2021: 214), e.g. brunch, smog, motel,
coywolf (i.e. half coyote, half wolf), malware (malicious software), neti-
quette (net etiquette), etc.
• Terminologisation: process whereby general language words or expressions
become specialised terms, e.g. platform has become a term in information
technology (Ahmad & Rogers, 2001: 752), whereas chatter, in the context
of engineering processes that involve cutting tools has been translated into
Chinese as 刀振, lit. ‘knife vibration’ (Li & Hope, 2021a: 12). Similarly,
through a process of de-​terminologisation, specialised terms may find their
way into everyday language, e.g. debug, originally from computing, in
the sentence The weaker parts of the [theatre] production can be easily
debugged as the summer run progresses (adapted from Meyer, 2000: 47).
The process of re-​terminologisation also exists, whereby a specialised term
from one domain takes on another specialised meaning in another domain
(e.g. Shumeli et al., 2023), e.g. virus from medicine to computing; weed
out (‘to remove unprofitable assets or investments’) from agriculture to
finance, loan shark from marine biology to finance, etc.

As mentioned before, these are some of the most common methods of term
formation in English. Because term formation methods are not identical in
every language, it is advisable for translators to become familiarised with the
term formation methods in their other languages, which should be of help
when needing to create terms in other languages. Some of the examples in this
section are single-​word terms but others are multiword terms. The behaviour
of multiword terms in texts is one of the objects of study of phraseology,
which we explore in the following section.

Phraseology
Terms do not occur in isolation in texts. How do you use coarctation of the
aorta in a sentence, for example? Can you get a coarctation of the aorta?
Should you say aortic coarctation instead? These examples illustrate that
82 Terminology Management for Translators

there is a “preferred way of saying things” (Gledhill, 2000: 202) within a

particular domain and that, therefore, there is more to translating specialised
texts than just translating their terms. Other conventions and “patterns of
wording” (2000: 202) will also need to be taken into account by translators
when rendering content to another language. Here, we will use phraseology
to refer precisely to those patterns of wording in specialised domains, slightly
re-​phrasing Firth’s famous adage “you shall know a term by the company
[i.e. phraseology] it keeps” (1957: 11). Although Firth was “referring to col­
location analysis, the idea is the same: co-​occurrences of a word [or term in
our case] provide access to its meaning” (Bertels, 2022: 312).
Phraseological information has not traditionally been included in
termbases (Humbley, 1997: 27) (although some term banks include it, e.g.
TERMIUM Plus®), possibly because phraseology, context and variation
were downplayed by Wüsterian terminology as “problematic aspects” (León-​
Araúz & Cabezas García, 2020: 213–​214). However, translators “are in con­
stant need of information on how terms behave with other lexical units”
(L’Homme, 2006: 60) and including phraseological information in your
termbase will contribute to higher quality translations. Problems such as the
use of aorctic coarctation in a sentence are best solved “through extensive
phraseological research and verification conducted on the basis of a large
high-​ quality corpus of specialized language texts” (Akkach, 2016: 5–​6).
Corpora and how to extract information from them are further explored in
Chapters 5 and 6. Further examples of conventions (see also Chapter 2) in
specialised domains are as follows:

• In chemistry, there is a convention for punctuation (Godly, 1993: 151–​

160) in the names of chemicals by using hyphens, enclosing brackets and
commas, e.g. in terms such as cyclo-​propyl(methyl)amine, quinoline-​4-​
carbaldehyde and 1,2 -​dimethoxy ethyl.
• In aviation, cleared to land “gives the pilot very specific rights and respon-
sibilities” (Schmitt & Carter, 2004: 10).
• In medical domains, the “conventionalised way of reporting blood pressure
is blood pressure is 140 over 60 and everyone in the medical field knows to
place the higher pressure figure first” (Schmitt & Carter, 2004: 10).
• In legal domains, phraseology may have a specific style and be rather for-
mulaic, e.g. The Company shall not be liable for any indirect, exemplary,
special, punitive, con-​sequential or incidental damages in connection
with or arising out of the use of products provided hereunder (Olohan,
2016: 66–​67).

The problem of phraseology for translators is, of course, that conventions

differ not only from domain to domain, but also from language to language.
Translators ideally need to establish what the conventions are in the target
language before they start translating. Although translators are ‘freer’ to
use a range of translation strategies to translate phraseology, as opposed to
 Designing Your Termbase: Principles and Considerations 83

terms, translators are at the same time constrained to use the conventional
phrases of the given domain, instead of inventing new ones. As well as the
termbase, the translation memory can also be used by translators as a tool
to achieve consistency of phraseology, but of course, you need to have saved
some content in the translation memory first, before you can use it to ensure
the consistency of the phraseology used (Olohan, 2016: 44).

Box 4.5 Tips and Tricks for Managing your Termbase

• Avoid the temptation to enter more than one data category per field
(Schmitz, 1996: 224): for example, the field definition should contain
the definition only, not its source. The source of the definition should
be entered in a separate field. Similarly, terms and their acronyms
should not be entered together but in separate fields, e.g. term and
acronym. This should improve consistent and systematic searching
and querying of the termbase, as well as its reusability (see also ‘data
elementarity’ in Box 4.2).
• Add a Notes field to your termbase, just in case: once a termbase has
been populated with terms and their associated information, it is
sometimes difficult to add, change or remove data from a termbase
(Warburton, 2015a: 385), but a field such as Notes can be a useful
workaround to add any kind of information that was not envisaged
in the original design of the termbase.
• Start early and never stop: terminologist Uwe Muegge’s tip to ter-
minology blog owner Patricia Brenes: “Also, start small. Don’t aim
to build a huge termbase in one month. Use a short list of terms, and
make sure that each is duly validated and includes all the categories
you need. Most organizations already have glossaries here and there,
which you can also use to your benefit” (Argos Multilingual, 2018).
• Beware of geographical language variants or ‘locales’: as we saw in
Box 4.1, termbase software may treat British English and American
English, for example, as different languages, which means that their
data will not be compatible, i.e. you would not be able to import
data in British English into a termbase in American English and vice
versa. The same applies to any other language variant. Therefore,
it is good practice to ensure the geographical variant and locale are
decided on (by clients, project managers, other translators, revisers,
etc.) and adhered to early on. Alternatively, check what your
termbase software allows.
• Keep backups: as we have seen, terminology work is highly laborious,
sometimes difficult and very time-​consuming. Regularly backing up
your termbase constitutes best practice, e.g. exporting copies of it to
other online and/​or offline locations.
84 Terminology Management for Translators

Reversibility of the Termbase

Although some translators may work in only one language direction,
it is important to take into account one final design consideration in this
chapter: the language direction of your termbase. The reversibility of the
termbase refers to whether it can be searched in every direction between its
pairs of languages. For example, if the languages of a termbase are Arabic
and English, a reversible termbase would yield the same translation quality
when the source language is set to Arabic and the target English, as vice versa.
For reversibility to work, you would need to plan your termbase design in
advance. Among the aspects to consider are:

• How important is the visualisation of the termbase data?

• What will the entry look like with each language as the source language?
• Will the labels for fields be monolingual or bilingual? For example, would
the label for the field Definition be given in both English and Arabic?
• In which language will the definition(s) be?
• Similarly, will the values of the picklists be monolingual or bilingual?

While keeping the labels monolingual saves time and effort creating the
termbase, the visualisation of the termbase data in one language but with
labels in another language might not be ideal in some cases. By contrast,
creating many bilingual labels considerably increases the time and effort
required to create the termbase in the first place, but you could achieve a fully
reversible termbase, if this is what you need. The more important it is for a
specific termbase, the more effort you will need to put in to ensure it is fully
reversible. This is also connected to the amount of data per language. For
example, in a non-​reversible termbase from English into Chinese, the Chinese
part of the entry might contain more information than the English side, and
vice versa. This is because termbases are generally designed for translation
into the target language.
To conclude this chapter, have a look at some tips for designing and man-
aging your termbase in Box 4.5. In the following chapter, we take a look at
how to add information to your termbase (i.e. ‘populating’ your termbase)
and how to use it with CAT tools.

Follow-​up Tasks and Reflection

1. Is it a good idea to create a picklist for the field of Client? What would be
the advantages and disadvantages? Would you make it mandatory, mul-
tiple or neither? Why?
2. In Figure 4.2, the field definition was included at term level. Is this the best
level for this field? What are the advantages of including definition at entry
level, for example?
 Designing Your Termbase: Principles and Considerations 85

3. Do some research and find a good intensional definition for cantilever in

the domain of structural engineering.
4. Do some research and create an extensional definition for weekday and
another one for noble gases.
5. Find three nouns in English which are also verbs, in addition to the
ones already seen in Box 4.4. What are their equivalents in your other
language(s)? Do they have the same part of speech as in English or are
they (slightly or completely) different?
6. How would you translate the Swiss cheese and fondue pot metaphors
as terms in the domain of risk management into your other language(s)?
Can you maintain the metaphor of ‘cheese’ or does this need to change
to something else? Would you store Swiss cheese and fondue pot in the
same entry in a termbase, i.e. are they the same concept?
7. Meronyms or hyponyms? Fill the gap in each of the following sentences
with either meronym or holonym. It can be confusing at times to distin-
guish them but, in order to remember the difference, bear the following
in mind: hyponyms =​X is a kind of Y, e.g. a rose is a kind of flower;
meronyms =​X is a part of Y, e.g. a petal is a part of a flower. Thus, a rose
could not be a meronym of flower because it is not a ‘part’ of it; petal
could not be a hyponym of flower because it is not a ‘kind of’ flower, but
a part of it.
a. Elbow is a ………………… of arm.
b. Bulldog is a…………….. of dog.
c. Wing is a …………….. of plane.
d. Pine, oak and cedar are ……………. of tree.
e. Leaf is a ………………… of plant.
f. Car is a …………… of vehicle.
8. True synonymy is rare: do some research and find the (slight) differences
between the following variants of laugh: guffaw, chuckle, giggle, snigger
and cackle. Would you store all these in the same termbase entry? Why?
How would you document their differences in a termbase entry?
9. Polysemy: do some research to find definitions or meanings for the word
file in different specialised domains, e.g. IT, beauty, etc. How would you
store polysemous terms in a termbase? How would you document the
different domains?
10. Make a list of all the paradigmatic relations we have covered in this
chapter, e.g. synonym. As a clue, they all end with -​nym, like synonym.
Incidentally, the term plesionym also exists. Find out what it means and
add it to the list.
11. Syntagmatic relations are often exemplified as collocations. Select the
correct collocation in each sentence:
a. It is difficult to …… keep/​make/​do …… the street riot under control.
b. The speech was so boring that nobody was …… giving/​having/​paying
…… attention all the time.
86 Terminology Management for Translators

c. The decorators were not very professional and …… did/​made/​took

…… a real mess of the wallpaper and glue.
12. Associative relations: match the pairs of words (a to e) to their associa-
tive relation (1 to 5) (adapted from Sager, 1990: 34–​35):
a. Bridge and iron
b. Global warming and deforestation
c. Light and watt
d. Poison and antidote
e. Weaving and cloth
1. Process/​product
2. Phenomenon/​measurement
3. Cause/​effect
4. Product/​counteragent
5. Object/​material

Notes
1 In the sense of ‘terminological entry,’ not in the sense of ‘entrance’ (e.g. to a
building).
2 https://​dat​cati​nfo.net/​
3 Forbidden in this context means that a given term must not be used in a specific
context or client, for example.
4 Or conceptual relations, if the main viewpoint is the concept. Since translators need
to translate terms, we will use the term semantic relations (or just relations), which
should facilitate a more fine-​grained discussion of equivalence among, and diffe-
rence between, terms, for example.
5 A melted cheese dish typical of Switzerland.

Further Reading
TerminOrgs (2014, 2016, 2022) provide a range of freely available online resources
and guides covering multiple aspects of termbase design. For readers who can access
ISO standards, ISO 12620 covers data categories, whereas ISO 704 “spells out the
basic principles of object-​concept-​term relationships, reflecting linguistic theory that
has developed over the past century” (Wright, 2022: 95–​96) (see also Chapter 10).
Rothwell et al. (2023: 61) provide a sample list of fields for data categories, with
suggested multiple or mandatory values and hierarchical level. Cabezas García and
León-​Araúz (2020) explore term variation and, as part of their study, describe fields,
data categories and picklists of several term banks, e.g. IATE, TERMIUM Plus® and
others, whereas Freixa (2022) investigates the causes of term variation. Šarčević
(2000: 229–​269) is still a relevant read on the equivalence problems in terminology,
especially from the viewpoint of legal translation. For more on term formation, it
is worth reading Kageura (2002, 2022), Sager (1997) and Valeontis and Mantzari
(2006). Gagné and L’Homme (2016) study opposite relationships between terms
which include, but are not limited to, antonymy and provide a typology of antonyms
based on lexical semantics, psycholinguistics and corpus linguistics.
5 Populating and Using Your Termbase

Key Questions

• Where do I find terms?

• How do I select terms and include them in the termbase?
• How do I retrieve the information from the termbase?
• What is a ‘corpus’ and why is this useful for translators?
• How do I use a corpus?
• How do I use a termbase with computer-​assisted translation (CAT)
tools?

Corpora and Text Mining

The first thing you need to use your termbase are terms and, to find
them, you will need to carry out some text mining (see also Figure 1.2 in
Chapter 1), ideally in a corpus (i.e. a structured ‘collection’; plural: corpora)
of specialised texts from the relevant domain. From the corpus and other
resources, you can also extract associated information about the terms, such
as definitions, examples of use, synonyms etc. to be entered into the termbase
too. In this chapter, we will cover the ideal process of collecting and storing
terminological data, as we saw in Figure 1.2 in Chapter 1, which includes
text mining (i.e. finding terms, usually in relevant corpora), term extraction
(i.e. obtaining the candidate terms from the corpus), term validation (i.e.
deciding which candidate terms are actual terms), and storing the terms and
any related information in the termbase, with a view to using them during
translation. We will also look at the issues behind term recognition, i.e. how
CAT tools retrieve terms from the termbase without you having to open your
termbase directly, and how the information in the termbase can be shared
by importing and exporting the data in it. Of course, it may not always be
possible or practical to implement this ‘ideal’ process in full, but you can
follow parts of it at different times, whether it be before, during or after a
translation project. Another important point is that we focus exclusively on

DOI: 10.4324/9781003302636-6
88 Terminology Management for Translators

digital corpora that can be accessed electronically online or offline, as manual

corpus work is becoming increasingly rare.
Text mining refers to the search for texts and resources containing domain-​
relevant terms and compiling corpora in each relevant language, as we saw in
Chapter 1. Finding relevant texts for a translation project can be a very time-​
consuming task, although there are some tools that can help in this respect,
as we see further in Chapter 6. In addition, for widely-​used languages such as
English, and for Asian languages such as Chinese, Korean and Japanese, you
may find an abundance of materials (McEnery & Xiao, 2007: 1), but this is
not the case for many other languages or for some highly specialised domains
(Bowker & Pearson, 2002: 58).
Text mining can be carried out monolingually in each of your languages,
which results in monolingual corpora, or it can be carried out in several
languages if you are interested in bilingual or multilingual corpora. Corpora
are defined as “a large collection of authentic texts that have been gathered
in electronic form according to a specific set of criteria” (Bowker & Pearson,
2002: 9). A bilingual corpus typically contains texts and their translations
in two languages whereas, if the corpus contains three languages or more,
we often speak of a multilingual corpus (Bowker & Pearson, 2002: 92).
Some authors use bilingual and multilingual interchangeably (e.g. McEnery
& Xiao, 2007). A corpus that consists of texts in one language and their
translations into one or more languages is commonly called a parallel corpus.
To date, the availability of parallel corpora is relatively scarce, “especially for
countries with only one official language and for language pairs not involving
English” (Morin et al., 2010: 1). However, for a parallel corpus to be useful,
its texts have to be aligned, i.e. each sentence in the source language must be
linked to the corresponding sentence (or sentences) in the target language.
Text alignment tools are discussed further in Chapter 6. Some authors use the
term DIY corpus (i.e. do-​it-​yourself corpus) (e.g. Sánchez Gijón, 2009: 115)
or disposable corpus (e.g. Buendía-​Castro & López-Rodríguez, 2013: 55)
to refer to the “small, specialized corpora created ad-​hoc to serve the needs
of the translator for a specific translation project” (Zanettin, 2002b). While
some corpora created by translators may indeed be discarded after a transla-
tion job, others may be more ‘permanent’ tools in the translators’ toolkit, e.g.
if they are specialising in a domain and could use the corpus in future jobs.
Another issue with parallel corpora is the fact that the language used in
the target language texts might be highly influenced by the source language
(McEnery & Xiao, 2007: 4; Morin et al., 2010: 1), which has led some
translators to prefer a comparable corpus. A comparable corpus is “composed
of texts written in different languages which are not mutual translations”
(Daille, 2017: 187). The word comparable is used to indicate that, although
the texts are not translations of each other, there is a certain level of simi-
larity between them, e.g. in terms of domain, degree of specialisation, etc.
Because the texts may have been created independently in each language,
translators are likely to find more contextually-​appropriate terminology and
 Populating and Using Your Termbase 89

idiomatic phraseology in comparable corpora which they can then apply to

their translations. In addition, comparable corpora may overcome overly lit-
eral translation but “they are less useful for the study of how a message is
conveyed from one language to another” (McEnery & Xiao, 2007: 4).
Corpora can be used in a wide range of disciplines, such as lexicography,
language teaching and learning, and sociolinguistics (Bowker & Pearson,
2002: 20). For translation-​oriented terminology, you can use corpora to iden­
tify single-​word terms, multiword terms and which other types of words and
expressions co-​occur with those terms (Bowker & Pearson, 2002: 32) (see also
Box 5.1). In addition, corpora can be classified from various perspectives, e.g.
spoken vs written, synchronic vs diachronic, etc. (Rothwell et al., 2023: 78),
but an important distinction for the terminological work of translators is
that of general language and specialised corpora. Specialised corpora typic-
ally contain texts in a specific domain or of a specific text type, e.g. a corpus
of application forms in the domain of immigration law. Specialised corpora
are particularly suited to terminological work because of the assumption that
words that appear more frequently in a specialised corpus than in a general
language one are likely to be terms in a specific domain (Gamper & Stock,
1998: 151). A list of freely available specialised corpora can be found in the
online portal. The tools to process the information contained in corpora are
explored further in Chapter 6.

Box 5.1 What Can I Use a Corpus For?

How many times do you turn to the web to find solutions to problems
and answers to questions? Translators also turn to the web for the same
reason, particularly when they translate specialised texts. Imagine how
good it would be if you had all those answers collected in one place for
consultation at any time. This is exactly the purpose of a specialised
corpus for translators. The problem is that you need to compile it
first. However, once compiled, you can use a corpus to answer many
questions such as the following:

• What does this term mean?

• How do I use this term in this particular context?
• How is this term used in another language?
• What preposition collocates with this term?
• What verb goes with this noun?
• What is the most frequently used version of this term?

A corpus can also help you determine the terms, conventions and
phraseology used in a given domain, e.g. symbols, measures, passive
vs active voice, etc. Corpora in the source language will help you gain
90 Terminology Management for Translators

a full understanding of your source text, whereas corpora in the target

language will provide you with useful information for the production
of the target text. We just need to make sure that the texts in these cor-
pora meet our quality standards and fit our purposes (Rothwell et al.,
2023: 78).
You can also use text from corpora as part of translation memories and
reference material when translating with CAT tools; other uses include
cultural awareness, idiomatic expressions, domain understanding and
more (Bowker, 1998; McEnery & Xiao, 2007: 7). In fact, translation
memories are a kind of parallel corpus anyway. With bilingual or
multilingual corpora, you can compare languages and obtain insights
into translation which may not be possible with other resources, thus
helping you make fewer mistakes in translation (McEnery & Xiao,
2007: 1–​7). In sum, using corpora as an aid can help you enhance your
overall competence as a translator (Corpas Pastor & Seghiri, 2009: 93).

It is important to note that a corpus should not simply be a random

collection of texts, i.e. a set of texts does not necessarily become a corpus
just by gathering those texts together (Corpas Pastor & Seghiri, 2009: 79).
In other words, you cannot “just start downloading texts haphazardly from
the Web and then call your collection a ‘corpus’ ” (Bowker & Pearson,
2002: 10). Instead, good practice consists of ensuring that each of the texts in
your corpus meets “a set of clear design criteria” (Corpas Pastor & Seghiri,
2009: 79), which you can draw up in advance, to ensure the corpus is rep­
resentative of the domain (see also McEnery & Brookes (2022: 37) and
Koester (2022: 51)). That said, the concept of ‘representativeness’ of corpora
is difficult to pin down, e.g. Rothwell et al. (2023: 78), more so for gen­
eral language corpora than for specialised corpora intended for terminology
purposes (Weisser, 2022: 96), but the general guidance is that the texts that
make up your corpus should share some common features. These features
will be determined by “your needs and by the goals of your project” (Bowker
& Pearson, 2002: 45) but here are some useful ones:

• Quantity vs Quality: is big beautiful? Corpora should ideally be big, but

what ‘big’ means is unclear and a bigger corpus does not always mean that
it is better. Since there is an increasing number of electronic tools avail-
able, it is becoming easier to put together large volumes of text, and some
electronic tools are designed to work with corpora of several billions of
words. This means that these tools will not work so well with smaller cor-
pora of about several hundreds of thousands of words (Daille, 2017: 151).
However, the larger the corpus, the more difficult it may be for you to
manage it, e.g. in terms of file storage. Bowker and Pearson (2002: 54)
suggest that “anywhere from a few thousand to a few hundred thousand
 Populating and Using Your Termbase 91

words have proved useful for LSP studies” and, as a rule of thumb, you
should aim to strike a good balance between a large enough corpus where
you can find what you are looking for, but a small enough corpus that is
manageable. As Morin et al. (2007: 664) state, “quality rather than the
quantity of the corpus matters more in terminology mining.”
• Number of texts: although corpus size is often measured by number
of words, you should also take into account the number of texts your
corpus has. As a general rule, it is better to have a multi-​author corpus
than a single-​author one. This is to avoid the corpus over-​representing the
preferences of an individual author (Bowker & Pearson, 2002: 49).
• Authenticity: your corpus should be composed of texts that “reflect real-​life
language use” (Rothwell et al., 2023: 78). This means that the texts con­
sist of “genuine communication” among experts going about their normal
business and that the texts have “not been created for the express purpose
of being included in a corpus” (Bowker & Pearson, 2002: 9). By learning
how experts communicate, for example, translators can use that informa-
tion to produce the translation in a way that experts can relate to when
they read it (and, of course, the same can be said about other readerships).
• Domain: the texts of your corpus should belong to the relevant domain
or subdomain, although this may not always be as straightforward as it
sounds. This is because, as we saw in Chapter 2, texts can be highly inter­
disciplinary and sometimes it may be hard to establish where one domain
ends and where the next one begins. As a general rule, if a text contains
terms from your domain, you can include it in your corpus, even if it has
terms from other domains.
• Degree of specialisation and text types: when compiling your corpus, it is
also advisable to distinguish between, or create separate corpora for, texts
belonging to different degrees of specialisation and/​or text types. This will
help you select the relevant terms according to the correct level of tech-
nicality when translating. For example, if you are working in a medical
domain, you can distinguish between a corpus of news articles, aimed at
the general public, where you find terms such as eczema and a corpus of
research articles, aimed at a much more specialised readership, where you
may find terms such as atopic dermatitis, i.e. a more technical synonym
of eczema.
• Authoritativeness: when selecting texts, one very important aspect is that
the texts should be reliable because they have been written by authori-
tative sources, especially if you are looking for texts on the web. This is
because anyone can upload content to it, which raises issues of quality
control (Bowker & Pearson, 2002: 63; Sánchez Gijón, 2009: 116). Ideally,
the authors of the text should be recognised domain experts. This means
that they “must have a suitable educational and/​or professional back-
ground in the discipline about which they are writing, and they should be
recognized by their peers as having the level of expertise required to write
about the particular subject” (Bowker & Pearson, 2002: 51). For example,
92 Terminology Management for Translators

collecting texts in the domain of cardiology from a National Institute of

Cardiology is much better than collecting texts for this domain from a
blog belonging to a member of the general public sharing their experiences
of cardiological issues. You may not always be able to tell who the author
is, and checking the authorship of a text may be very time-​consuming, but
verifying the provenance of your texts is well worth it.
• Language: since many languages have variants, e.g. British and American
English, as we saw in Chapter 4, you should compile separate corpora
for different variants to avoid confusion later. You should also consider
whether the texts were written originally in a given language or whether
they are translations (Bowker & Pearson, 2002: 51). Wherever possible,
original texts are better than translations for the terminological work of
translators, to avoid the target language being influenced too strongly by
the source language (i.e. translationese) although, of course, it may not
always be possible to check. Another consideration concerning language
is whether the text was written by a native speaker of the language but
be careful to avoid a major source of bias, such as the use of stereotyp-
ical information to identify authors. For example, “a person with the last
name Smith may not speak English as a native language” (Brandt, 2021).
• Date: depending on which domain you work on, you may want to know
the latest developments or terminology in it and, therefore, you may
want to check the creation or publication date of texts you may want to
include in your corpus. However, older texts may also be valuable e.g. by
providing definitions or explanations for a concept (Bowker & Pearson,
2002: 52).
• Balance: the concept of balance refers to the internal consistency of the
corpus (McEnery & Brookes 2022: 38) in terms of content, the variety of
language you want to study or specialise in, the register, the level of spe-
cialisation, etc.

The more you use the above criteria systematically and coherently to select
texts for inclusion in your corpus, the more representative and of higher
quality your corpus will be. On occasion, you may have to accept the fact
that it is not possible to create ‘the perfect corpus’ but any corpus, once
compiled, will be available for you to consult at any time, much like having
an expert ‘on demand.’ In addition, if your corpus is multi-​authored, you can
have access to more than one expert opinion at once (Bowker & Pearson,
2002: 19).
One final point to consider when compiling your own corpus is the fact
that, when you look for texts on the web, you may find multimedia material
(Bowker & Pearson, 2002: 63; Sánchez Gijón, 2009: 118), such as graphics,
videos or audio material included in the texts (see also Box 5.2), as they
are very useful for understanding how concepts (and terms) relate to each
other. You may want to include the multimedia material in your termbase,
software permitting, for future reference. However, it is worth highlighting
 Populating and Using Your Termbase 93

that, because corpora are texts, corpus analysis tools work with text only
and not with multimedia material, which means that you will need to store
a text-​only (e.g. .TXT files) version of your corpus if you want to use corpus
analysis software to extract information and terms from it. Some programs
allow .HTML and other formats (Bowker & Pearson, 2002: 65; Sánchez
Gijón, 2009: 119) and it is thus important to be aware of the capabilities of
any software you use. Therefore, while you may store multimedia content in
your termbase, you can process the text-​only files with corpus analysis tools.
We will start by looking at how to extract terms from your corpus but we
explore other corpus processing tools in Chapter 6.

Box 5.2 Protocol for Text Mining

Since the documents on Internet can “contain all kinds of false and
unreliable data, identifying the information providers you can trust
is an important step towards ensuring the quality of the information
you obtain” (Austermühl, 2001: 52). Corpas Pastor and Seghiri (2009)
suggest a four-​step approach to text mining which you can use once you
have selected the criteria for your corpus:

1. Locating and accessing resources (which includes evaluating the

authoritativeness, etc. of the texts). This includes carrying out some
preliminary research through “three interconnected strategies”
(Austermühl, 2001: 52): an institutional search, a thematic search
and a keyword search. For example, in the domain of agriculture,
you could search for national or international organisations and
institutions, such as the Food and Agriculture Organization; a the-
matic search in online resources such as directories,1 where informa­
tion is organised by domain and subdomain (e.g. feeding systems,
harvesting, etc.). Institutional or thematic searches might require
keyword searches, e.g. fertilisers, herbicides, etc. The point about
these searches is that they can yield reliable information for your
terminological work.
2. Downloading the selected texts from the web, so that you can use
them offline any time on your computer (Sánchez Gijón, 2009: 118–​
119). It is good practice to establish a file naming convention before
you start saving texts. Using information in the file name that relate
to the content of the file or corpus allow you and any other users to
manage files more effectively, for example by using language codes,
topic of the corpus and/​or date: es-​en_​Architecture 1_​2024.txt (e.g.
Reppen, 2022: 15). For the next text in the corpus, the number 1
would be replaced by 2 and so on, e.g. es-​en_​Architecture 2_​2024.
94 Terminology Management for Translators

3. Pre-​processing of texts: ensuring the texts are in the correct format

for corpus processing tools, which often require .TXT (i.e. text-​
only) files, without multimedia content, although some can process
.HTML files too (Sánchez Gijón, 2009: 119). Many texts come in
.PDF format which can sometimes be copied and pasted on to a
MS Word file. However, it can be quite onerous to ensure that there
are no hard line breaks in the wrong places which will then cause
problems when aligning parallel corpora. Therefore, an alternative
might be to use a free online PDF converter that may be available
at the time. PDF-​to-​text converters can be easily found through a
simple Internet search (Anthony, 2022: 104).2
4. Storage of the corpus: file and folder management is important!
Make sure you remember where you store these files and folders
and make backup copies to protect your data, as sometimes elec-
tronic files can get corrupted and become lost. You can also store
your files and folders in a hierarchy, i.e. with various levels of
subfolders, which may be used to store data for subcorpora, for
example.

Term Extraction
Once you have compiled your corpus, you can use a range of corpus analysis
tools to make the most of it. In this section, we take a look at the principles
behind term extraction, while we will review some term extraction tools in
Chapter 6. Term extraction is the process of “using computer programs in
order to identify strings [of text] that are potential terms” (Ahmad & Rogers,
2001, p. 725). Some authors refer to term extraction as “automatic” but
this is a misnomer because the entire process is far from automatic. Usually,
terms are extracted with software but require verification by human users
afterwards (Bowker & Pearson, 2002: 165).
Typically, term extraction is a task for a terminologist, who then “consults
with a domain expert to arrive at a final list of validated terms” (Heylen & De
Hertog, 2015: 203). However, since term extraction is increasingly integrated
into CAT tools, translators also use term extraction to create and maintain
their own terminological resources. Term extraction can be monolingual or
bilingual. While monolingual term extractors identify candidate terms (i.e. a
list of terms whose terminological status needs to be confirmed, or validated,
by the translator), bilingual ones attempt to extract term candidates simul-
taneously from the source and target texts (Bowker & Pearson, 2002: 165)
(see also Chapter 1). In other words, term extraction provides the “raw
material” (Ahmad & Rogers, 2001: 725), which you still need to check and
validate before you enter it into the termbase.
 Populating and Using Your Termbase 95

McEnery and Xiao (2007: 4) suggest carrying out term extraction from
either monolingual corpora or comparable corpora because, according to
some authors, the results from term extraction in parallel corpora may not
be valid due to the (often) strong influence of the source language on the
translated text of the target language (or translationese). However, Zanettin
(2011: 17–​22) makes a compelling case for the opposite, i.e. that translations
should be included in most corpora because:

• Translation “is a legitimate language production activity.”

• Translations represent language, they do not distort it.
• The production of translated language is neither deviant nor unacceptable
and can be “part of a collective body of reference for the target language.”
• Translated language “may indeed be a specific variety of language but all
language production is subject to some kind of constraints.”
• Translated language does not “taint” a corpus by being in it.
• Etc.

Thus, Zanettin suggests that translated texts have a role to play in corpus-​
based translation studies and that the proportion of translated texts to include
in a corpus will “vary depending on the language and genres considered”
(2011: 17). For example, in English-​speaking countries, “translated texts are
only a minor percentage of all the texts published” whereas, for languages
such as Italian or Brazilian Portuguese, “translations represent a consider-
able share of all published texts” (2011: 17) and it would be “very diffi-
cult to design a representative/​ balanced corpus consisting exclusively of
non-​translations.” In sum, as long as translated texts are recognised as such,
they are “full-​fledged members and legitimate data” to include in a corpus
(Halverson, 1998: 8). What is more important is “a careful description of
what the corpus is intended to represent” (1998: 3).
Some tools also extract the immediate context of the term with the term
(e.g. a sentence or paragraph), which can become a potentially useful term
record with some additional research (Bowker & Pearson, 2002: 165) but
it can also reveal the co-​occurrents of the terms and other phraseologisms
which can also be stored in the termbase (Pavel & Nolet, 2001: 44). Other
tools are able to recognise capitals and hyphenation, which is helpful in iden-
tifying, for example, quasi-​crystal and Quasi crystal as different occurrences
of one term, instead of considering them as separate terms, so that you do not
have to group them together later.
As we saw in Chapter 1, terms may be single words or multiword
expressions. Single words are often extracted from text by software on
the basis of their frequency in the text. The basic idea is that if a term
is important in the domain, it will appear often in a specialised corpus
(Bowker & Pearson, 2002: 165). As a user, you can set the minimum fre­
quency value; this means that, if you set a minimum frequency value of 2,
96 Terminology Management for Translators

the program will extract every single word which occurs twice in the text,
hoping some of them will be terms. Of course, this does not mean that terms
which do not appear frequently in the corpus are less important: they are
equally important even if they only occur once in the text. In fact, less fre-
quent terms may pose more challenges to translators, so they may be worth
storing in your termbase too. Another strategy for extracting single-​word
terms is to compare their frequency in the specialised corpus (also known as
focus corpus) to their frequency in a general language corpus (also known as
reference corpus) (Bowker & Pearson, 2002: 167). There are many general
language corpora available online, such as the British National Corpus or
BNC3 for English.
For multiword terms, two approaches are generally followed in term
extraction: linguistic and statistical. Linguistic approaches to term extrac-
tion, as the name suggests, extract term candidates based on linguistic infor-
mation such as part-​of-​speech patterns. The assumption behind linguistic
term extraction is that the vast majority of terms are nouns or noun phrases,
e.g. noun +​noun, adjective +​noun, etc. (Bowker & Pearson, 2002: 168)
as widely acknowledged in the literature (e.g. Ahmad, 1997: 475; Béjoint,
1988: 356; Bertels, 2022: 325; Daille, 2003: 30).

Box 5.3 Key Concepts Used in Term Extraction

• Types of term extraction:

• Linguistic term extraction uses part-​of-​speech patterns to iden-
tify terms.
• Statistical term extraction uses frequency of co-​ occurrence
patterns to identify terms.
• Hybrid term extraction uses a mixture of linguistic and statistical
methods.
• Trade-​off in the quality of the output:
• Noise: unwanted or irrelevant candidate terms.
• Silence: wanted or relevant terms are missed.
• Lists of stopwords can serve as filters to minimise the unwanted
or irrelevant candidate terms extracted by the term extraction
program.
• Measures to assess the effectiveness of a term extraction tool:
• Recall: the ratio of correctly extracted terms among all the terms
that should be extracted (Kageura & Marshman, 2019: 64).
• Precision: the ratio of correctly extracted terms from the list of
extracted candidate terms (Kageura & Marshman, 2019: 64).
• F-​measure: overall performance based on a weighted average
between precision and recall.
 Populating and Using Your Termbase 97

• Other measures: mutual information (MI), T-​score, C-​value/​NC-​

value, TF-​IDF, log likelihood, etc.
• Validation: manually checking which candidate terms are actual
terms.
• Ranking (or scoring): a number indicating how confident a term
extraction tool is that a candidate term is an actual term. Usually,
the higher the number, the more confident the system is, and high
numbers can be used as a post-​processing tool to retrieve actual
terms from potentially long lists of candidate terms.

For linguistic term extraction to work, the corpus needs to be tagged.

Tagging means assigning parts of speech to words and phrases, so that you
can instruct the computer to extract nouns only, or a specific type of noun
phrase, for example (Bowker & Pearson, 2002: 168). Linguistic term extrac­
tion has its own challenges, as not all nouns or noun combinations are terms,
for example (Melo Mora & Toussaint, 2015: 236). In addition, the part-​of-​
speech patterns that work in English, such as adjective +​noun phrases (e.g.
cataclysmic eruption) may not work for other languages, which may prefer
noun +​adjective phrases, for example (Bowker & Pearson, 2002: 169), as is
the case with Romance languages.
Statistical approaches to term extraction, by contrast, require no lin-
guistic information in principle and instead use two properties of multiword
terms in order to identify them: the fact that they are relatively fixed word
combinations and the fact that they should occur with relatively high fre-
quency in texts belonging to their specialised domains (Heylen & De Hertog,
2015: 208). For example, it is very unlikely that the word order of cataclysmic
eruption will change in texts from volcanology. Further, the frequency of this
expression should be much higher in texts from volcanology than in texts
from general language. As with linguistic approaches, users can also set the
minimum frequency value and, because statistical systems are not language-​
dependent, their algorithms can be applied to many languages. There are also
hybrid systems which combine statistical and linguistic methods and, in fact,
most state-​of-​the-​art term extractors use hybrid systems (Bowker, 2020: 268;
Heylen & De Hertog, 2015: 217).
The advantage of using a term extraction program is speed, but at the
expense of the quality of results, as all programs end up extracting some
“units that are not relevant from a terminological point of view” (Drouin,
2006: 375). This means that the results of a term extraction always require
the translator’s validation. The alternative is manual extraction which can be
extremely labour-​intensive but usually yields good results. However, manual
extraction does not guarantee that you will extract all the relevant units and
will typically miss patterns that emerge only when large quantities of text
98 Terminology Management for Translators

are processed (Ahmad & Rogers, 2001: 740). In the next section, we take a
closer look at term validation.
Both approaches, as well as a hybrid one, can produce ‘bad’ candidates,
i.e. words or phrases which are confirmed as not terms after validation
(Kenny, 2011: 463). Irrelevant or unwanted term candidates are referred to
as noise and will be rejected during validation. Similarly, term extractors can
fail to return ‘good’ candidates, i.e. some actual terms do not get included in
the list of term candidates (2011: 463). Relevant results that are not retrieved
are referred to as silence and it means that some appropriate terms are missed
by the software in the list of term candidates.
In order to be effective, a term extractor must produce low levels of
noise and silence (Warburton, 2021: 205). However, many term extraction
tools produce as much as 60% noise according to a study by Warburton
(2020: 36). The process of removing noise from a list of candidate terms is
referred to as cleaning. In order to control noise, it is generally good practice
to use stopwords which are generally function words, i.e. extremely common
words that fulfil an essentially grammatical function, such as the, on, at, etc.
(Corver & van Riemsdijk, 2001: 1). By contrast, content words are all other
words, which have more semantic content, e.g. nouns, verbs, adjectives, etc.
As a general rule, terms are content words, rather than function words.
A stopword list (or stop list) works as a filter used by the term extractor to
omit unwanted words, such as stopwords (Heylen & De Hertog, 2015: 208).
However, you should exercise caution, as function words can also be com-
ponent words of multiword terms. For example, including of in a stopword
list might exclude viable terms such as plane of symmetry in crystallography.
This is why some term extractors allow you to customise stopword lists
or import your own. Running a term extraction task with stopwords can
increase the quality of the results, but you should always ensure that you val-
idate the results. There is a trade-​off to be made between the amount of effort
that you are willing or able to put in and the desired result, and this trade-​off
may vary considerably from project to project.
In order to calculate the effectiveness of a term extraction tool, measures
borrowed from the domain of information retrieval are used, such as pre-
cision, recall and F-​measure. Precision is the number of correctly identi-
fied terms out of all the proposed term candidates (Heylen & De Hertog,
2015: 216) whereas recall is the number of terms identified out of all the
terms that appear in the corpus (2015: 216). In other words, precision is
the answer to “How many terms were correctly extracted?,” i.e. a measure
of exactness, whereas recall is the answer to “How many terms were identi-
fied?,” i.e. a measure of completeness (Walkinshaw et al., 2008: 257).
Precision and recall can be combined into the so-​called F-​measure (or F-​
score), which provides an indication of overall performance (Manning &
Schütze, 1999: 269) of the program. The F-​measure is not a mere average
but a “weighted harmonic mean” (Korkontzelos et al., 2008: 254) between
precision and recall. The formulas for precision, recall and F-​measure, as well
 Populating and Using Your Termbase 99

as an example of how they can be applied to term extraction tasks carried

out by translators and researchers in translation, are discussed in Fernández-​
Parra (2014) and in the online portal. Although it falls beyond the scope of
this book to give a full review of other measures, readers may be interested in
knowing more about measures such as mutual information (Biber, 2009: 287;
Church & Hanks, 1990), T-​score (Hadni et al., 2014), C-​value/​NC-​value
(Frantzi et al., 2000), TF-​IDF (Huangfu & Zhao, 2014), and log likelihood
ratio (Dunning, 1993; He et al., 2006). For a general overview of statistical
natural language processing, see Manning and Schütze (1999). The general
message for translators is to be aware of what the program can extract and
what it tends to miss (Kageura & Marshman, 2019: 65).

Term Validation
From the viewpoint of translators, as we saw in Figure 1.2 in Chapter 1,
term validation consists of confirming that the candidate terms extracted are
valid and useful for the domain and to be included in the termbase. Whether
linguistic, statistical or hybrid term extraction tools are used, the list of term
candidates produced by the software is hardly ever perfect. Therefore, it is
usually validated manually by translators. In order to avoid retrieving lists of
candidate terms with too much noise, some term extraction tools allow users
to apply stopword lists, as we saw in the previous section.
If, despite using a stopword list, you still obtain a large volume of noise,
another frequent feature of term extraction tools is ranking. A rank (or score)
is a number which indicates how confident the term extraction tool is that
the extracted expression is an actual term. Usually, the higher the number,
the more confident the system is. Term candidates are generally displayed
in descending order of ranks, i.e. the highest ranked term candidates at the
top, but you can also view them in other arrangements, e.g. alphabetical, in
ascending order, by frequency, etc. In this way, you can use ranks to focus
on the potentially highest quality results of the extraction, although there are
often errors and you will also find useful term candidates with a low rank.
Ranks can vary significantly from one tool to another, e.g. from 1 to 5 or
from 0 to 99, but they are a useful device to post-​process the list of candi-
date terms, especially if the list is long and your time is limited. How much
effort you should put into post-​processing this list will depend on the texts on
which the extraction was performed, the volume of the output, the tool used,
your availability to carry out the post-​processing, etc.

How to Select Terms for the Termbase

Once you have obtained a list of term candidates, you still need to select from
this list which terms you would like to store in your termbase, as the software
does not do this for you. Although term extraction is carried out by the soft-
ware, term validation is a manual process, as is the process of selecting terms
100 Terminology Management for Translators

from the list of candidates to include in the termbase. Because the purpose
of the termbase is to help you collect domain knowledge and information in
two languages or more in order to translate, the terms you need to collect
are those that will help achieve this aim. For translators, the criteria to select
terms may often go beyond the Wüsterian or classical notion (see Chapter 3)
of terms as units that acquire a special meaning in a specialised domain. The
following is a non-​exhaustive list of such criteria, which may be used indi-
vidually or in combination:

• Frequency: frequency is important (Warburton, 2021: 99). If a term

appears frequently in texts from a given domain, you may want to include
it in the termbase because it is likely that you will use it again, but this will
also depend on the other criteria below. However, it may also be useful for
you to include terms that appear infrequently, but for another reason, as
infrequent terms may cause challenges in translation.
• Leverageability and reusability: one reason why it is useful to include cer-
tain frequent terms in your termbase is because they can be ‘leveraged,’ i.e.
retrieved automatically by CAT tools during translation, as we see below.
This has led some translators, e.g. in commercial environments, to store
words and expressions which are not strictly terms. However, storing such
words and expressions saves them time and effort when translating with
CAT tools.
• Embeddedness: if a term is ‘embedded’ (Warburton, 2015a: 379) or
‘nested’ (Frantzi et al., 2000: 117) within a longer term, it may be an indi­
cation that this is also a useful term for translators because of its “prod-
uctivity in forming longer multiword terms” (Warburton, 2015a: 382).
Warburton’s example is the term sustainable development which, as she
discovered, appears in 26 longer terms, such as ecologically sustainable
development economics, environmentally sustainable development and
sustainable development policy planning, to name but a few.
• Translation difficulty: it is generally a good idea to include in the termbase
any terms that cause challenges in translation, e.g. due to their complexity,
e.g. forward strut rear angled needle roller bearing housing (Hartley &
Paris, 2001) and/​or lack of frequency, among other criteria.
• Multiple synonymy: if a term has multiple synonyms in use, it is also worth-
while storing it in the termbase, along with all its variants and comments
or notes on their differences in use or context (RWS, n.d.: 7).
• Visibility: in commercial settings in particular, the visibility of a term,
especially if connected to a brand, which would make it visible to product
users, for example, might also be a criterion to include the term in the
termbase (Warburton, 2021: 97). In this way, a company can track
the item and check that it is correctly and consistently labelled and
referred to throughout the company’s documentation. Under this cri-
terion, expressions such as product names, colour names, taglines, etc. in
marketing domains, for example; law names, entity and body names in
 Populating and Using Your Termbase 101

finance, law and banking domains; button names, menu items, etc. in IT
and software domains (see also Giai et al., 2023), etc., may all be included
as terms in the termbase although they are not usually considered as terms
in the classical sense.
• Part of speech: as we have seen in Chapters 3 and 4, a large volume of
terms are nouns or noun phrases. However, you can also consider verbs,
adjectives and adverbs as potential terms. Some important terms may not
be nouns or noun phrases. As a general rule, there are only four main parts
of speech worth considering: nouns (and noun phrases), verbs, adjectives
and adverbs; parts of speech such as conjunctions, determiners, etc. are
seldom terms (except perhaps in linguistics domains).
• Non-​terms: in some situations, you may want to store other ‘non-​terms’
along with terms in your termbase, e.g. extensively developed in the so
that, if this is a frequently-​occurring segment in the source text, it can
be retrieved from the termbase automatically when using CAT tools
(connected to the criterion of reusability). Although a fragment such as
extensively developed in the does not map directly on to a semantic unit,
and would not score highly with any unithood or termhood measure, “it
could be worth the translator’s effort to record it as an entry in a termbase
so that it can be easily reused” (Bowker, 2015: 313). Since CAT tools do
not currently offer a separate kind of database to store non-​terms, except
perhaps Atril’s Déjà Vu (see Fernández-​Parra, 2023), it may be more con­
ceptually tidy to keep non-​terms in a separate termbase from terms, when-
ever possible.

The above list should help you decide what to include in your termbase.
The idea is not to collect a large amount of terms, but a useful amount
of terms (see Warburton, 2015a: 379). However, deciding what exactly
counts as ‘useful’ may vary from project to project.

As well as deciding what to include, you should consider what to exclude.

We have seen that you should ideally exclude any lexical item that is not a
noun (or noun phrase), verb, adjective or adverb. In addition, you should
exclude modifiers, i.e. you should not list them alongside the term in the
Term field, but you can include modifiers of the term elsewhere in the entry,
e.g. in a field named Modifiers, if you consider those modifiers useful for
current or future translation projects. Modifiers are words or phrases that
appear with the term, which at first glance seem to be part of the term, but
which in fact are not part of it. For example, black in black passenger car is
a modifier because the concept is only passenger car; it just so happens that
one particular passenger car may be black and other passenger cars may be
of other colours. However, black is not a modifier in black hole or black-
bird because ‘black’ is part of the concept in each case. In other words,
102 Terminology Management for Translators

whereas you could say red passenger car, you could not say red hole (and
still mean ‘black hole’) or red bird (and still mean ‘blackbird’). Of course,
modifiers may differ from language to language but you should ensure you
store the relevant information in the termbase to streamline your transla-
tion work.
Many modifiers are often adjectives, e.g. large, high, steady, etc., but
modifiers are not always easy to spot. For example, is chronic a modifier
in chronic asthma? In order to answer that, you would need specialised
knowledge of the domain (hence returning to the idea of the importance of
specialising in a domain, as we saw in Chapters 1 and 2). However, as a rule
of thumb, ask yourself whether the concept is asthma, which can be chronic
sometimes (in which case chronic would be a modifier), or whether chronic
asthma is a distinct concept from just asthma (in which case chronic would
not be a modifier). If in doubt, you can make a note in the Notes field of the
termbase, for example, until you can resolve it.
On another note, even if you find that chronic asthma is not a term, you
might still want to include it in the termbase, e.g. if it is a highly leverageable
phrase in a given translation project. The main point here is that you should
be knowledgeable about what you translate and get the translation right for
the right reasons, rather than by accident. The termbase is the repository
where you can store any important information that you can retrieve when
you need it.

Recognising Terms in Text

It may not always be straightforward to pick out terms from non-​terms
within a list of candidate terms, i.e. without context. The more sophisticated
term extraction tools will allow you to click on a candidate term and view
its context, e.g. sentence or paragraph, which might help you confirm it as a
term or discard it. If you look at the context of a candidate term (or if you
are extracting terms manually from a text), there are some clues you can
use to make the final decision, as some authors of technical texts “tend to
treat terms in a special way in a text (although unfortunately not always)”
(Rothwell et al., 2023: 65). Some examples of special treatment are the
following (2023: 65) (also illustrated in Figure 5.1):

• Different formatting: terms may be highlighted with bold, italics, etc., e.g.
crystal structure in Figure 5.1.
• Different punctuation: terms may be highlighted by means of punctuation,
e.g. inverted commas, colons, hyphens, etc. For example, in Figure 5.1,
crystal structure is followed by a colon, whereas lattice is followed by a
hyphen.
• Different capitalisation: sometimes, terms may be shown in capital letters
even though their ‘normal’ use is in lowercase. For example, in Figure 5.1,
 Populating and Using Your Termbase 103

Figure 5.1 Extract illustrating examples of how terms are highlighted in texts.

Source: From University of Nairobi (2019: 5).

the term Halite appears under the graphic, although it does not require a
capital H.
• Context: terms may be preceded by a definition, description, explanation,
acronym or abbreviation. For example, in Figure 5.1, crystal structure and
lattice are followed by the This is which introduces some kind of definition
or explanation of the terms.
• Visual display: terms may appear in graphics (but not everything that
appears in a graphic is a term). For example, in Figure 5.1, Cl and Na are
the symbols for chlorine and sodium, respectively, which can be terms in
chemistry and mineralogy domains.
• Translation: in languages such as Japanese or Korean, “English terms are
often provided in brackets after the corresponding terms” (Kageura &
Marshman, 2019).

These clues are not foolproof, since it may well be that authors of
specialised texts do not use any of them in a given text. However, they
may be of some help when deciding what constitutes a term from a list of
candidate terms.
104 Terminology Management for Translators

Entering Terms in the Termbase

Once you are ready to enter your selected terms in a termbase, there are sev-
eral ways to do this. The more sophisticated term extraction tools allow you
to export the list of confirmed terms and import it directly into a termbase,
often through MS Excel. Alternatively, an Excel spreadsheet may be used in
the first place, e.g. by teams who list the agreed terms on it, before each lan-
guage team imports the terms into the relevant termbase, with its equivalents
in the target language (see also Box 4.1 in Chapter 4).
Another method is to enter the terms manually into the termbase, once it is
created, and often before starting to translate. However, if you plan to enter
many terms manually into your termbase, it is advisable to create a template
(or input model), and use it to input data systematically and consistently into
the termbase (see also Steurs et al., 2015: 247). A template or input model,
created by the user, is a pre-​defined arrangement for displaying fields and
data categories consistently according to the user’s needs, which will be sub-
sequently used as a model to enter data into future entries, so that all entries
using that template will look exactly the same. This allows faster input and
consistent display of data into the termbase. The more sophisticated termbase
software will allow users to edit existing fields and create or edit hierarchical
structures (Rothwell et al., 2023: 67).
However, once you have started translating, you can also enter terms into
the termbase on the fly (Candel Mora, 2017: 249), i.e. as you translate, since
many CAT tools have this functionality. Entering terms on the fly is helpful
because it may save you some time by entering data into the termbase directly
from the CAT environment and without having to open the termbase, but it
is always good practice to go back to the newly created entry and add other
information about the term, e.g. context, definition, etc., whenever you have
more time to devote to it.
An important convention when adding terms to your termbase is that terms
are generally entered in their dictionary form or canonical form (Wright,
1997a: 16) which consists of reducing words to their base forms (also known
as lemmas, see also Concordancers section in Chapter 6). For example, for
English and many other European languages, this means:

• Entering all common words in lowercase (except German nouns); proper

nouns can remain capitalised, e.g. Portland in Portland stone. Wright
(1997a: 16) found that “there seems to be a temptation in the creation
of technical glossaries and other special language collections to capitalize
either the first letter or even all the letters in all terms, although these
words are not normally capitalized” and she concludes that this tendency
should be strictly avoided. This may have a bearing on term recognition
when using a termbase with CAT tools (see Chapter 6).
• Reducing any verb forms to the infinitive, e.g. took, taken and taking
should be entered as take (assuming that take is a term in the relevant
 Populating and Using Your Termbase 105

domain); English verbs do not need to go with to, e.g. take is preferable
to to take, which avoids all verbs being alphabetically grouped under “t.”
• Reducing any declensional endings to their nominative form, e.g. the
Greek αστεριού (genitive) and αστέρια (accusative) should be entered as
αστέρι (nominative, meaning ‘star’).
• For languages with gender, nouns and adjectives should be entered in the
masculine singular form.
• Enter nouns and noun phrases without articles.
• Multiword terms should be entered in the usual written order, e.g. dia-
mond drilling and not drilling, diamond, as this can cause confusion and
hinder future searches of the termbase. Again, this is important when
using a CAT tool, which will typically only find exact termbase matches
with the source string.

For other languages, there may be other standard practices and rules which are
often found in lexicographical practice (Wright, 1997a: 17). The main point
is that you should resist the temptation of entering terms in the termbase just
as they appear in the text. By entering terms in their dictionary form instead,
you ensure clarity and simplicity of display, while avoiding redundancy and
cluttering of the termbase (see ‘pretranslation’ below) with multiple versions
of the same term and this allows for termbase searches to be more efficient
(see also automatic terminology lookup below).

Using a Termbase with CAT Tools

Termbase software does not exist in isolation. From the early versions of
CAT tools, terminology management features have been integrated into CAT
tools (Candel Mora, 2017: 248), as terminology has always been considered
an important part of a translation project (García, 2015: 73). However, CAT
tools come in many shapes and sizes nowadays, from those that offer the
minimum functionality of allowing the use of a limited glossary to exten-
sive CAT tools that comprise a suite of interconnected programs, one of
which allows the use of complex termbases (Steurs et al., 2015: 224) (see
also Box 5.4). Regardless of their sophistication, most CAT tools, understood
as tools that incorporate both translation memory and termbases (Rothwell
et al., 2023: 4; Warburton, 2018: 120), allow efficient retrieval of termino­­
logical data during translation through features such as fuzzy matching,
pretranslation and automatic terminology lookup.
As we saw in Chapter 1 (e.g. Figure 1.2), term recognition is the process
of matching the terms in the source text to an entry in the termbase and auto-
matically displaying the entry for the translator within the translation inter-
face for the translators to decide subsequently whether anything needs to be
copied and pasted into the target text. Depending on the settings available in
the CAT tool, term recognition, used here as a synonym of automatic termin-
ology lookup, as it is also known in the literature (e.g. Bowker, 2003: 56),
106 Terminology Management for Translators

can take place automatically, freeing the translator from finding the terms
manually in the termbase. Automatic terminology lookup saves translators
a great deal of time because they can ignore the term record that appears, if
it is not relevant or, if it is relevant, they can insert the term with one click,
thus saving time typing potentially long terms and ensuring consistency (e.g.
ensuring translators do not use hard disk in one part of the target text and
hard drive in another). Another advantage of automatic terminology lookup
is that all the information stored in the entry about the term, e.g. definition,
collocations, etc., also becomes visible to the translator for background infor-
mation and/​or inspiration during translation.
Term recognition/​automatic terminology lookup can be exact (i.e. 100%)
(i.e. when the term in the source text is identical to a term in the termbase)
or fuzzy (i.e. when the term in the source text is similar but not identical to a
term in the termbase, 1% to 99%). Normal searches can look for a specific
term or “terms beginning with a particular character or series of characters”
(MultiTerm help files). Fuzzy matching is the technique that locates partial
matches to terms in the termbase. This is a helpful feature, not only to recog-
nise spelling variants of terms or fragments of terms, but also inflected forms
and morphological variants, the precise order of multiword terms and even
errors (Bowker, 2003: 55). The advances made in the algorithms used for
fuzzy matching mean that translators can be increasingly confident of storing
only the dictionary form of terms in the termbase, as opposed to storing
multiple entries for different inflected or morphological forms, for example.
Using dictionary forms in the termbase thus simplifies the “process of cre-
ating records and reducing record duplication” (Kageura & Marshman,
2019: 71). In addition, some tools allow users to carry out full text searches,
which means that text fields such as definitions, examples, etc. will also be
searched. Full text searches can help translators to retrieve terms that do not
have an entry in the termbase but appear in fields such as definition, con-
text, etc.

Box 5.4 Other Advanced Features for Using Termbases

The more advanced features for making the most of a termbase when
translating with CAT tools vary greatly from CAT tool to CAT tool and,
typically, each tool gives the feature a different name. The following is
a non-​exhaustive list of advanced termbase features offered by different
components CAT tools at different stages of the translation process:

• Checking for non-​usage of target terms, e.g. as part of the revision

process.
• Checking for terms that may have been set as forbidden.
• Checking for terms without equivalent in the target text.
 Populating and Using Your Termbase 107

• Finding and merging duplicate term records (Candel Mora, 2017:

250).
• Applying filters, e.g. for searching obtained results by domain or
client (Kageura & Marshman, 2019: 67–​68).
• Setting permissions to consult, edit and delete data by different users.
• Creating term records from results of searches within the TM or
parallel texts.
• Creating term records from results of term extraction.
• Bulk import or export of term records, useful when integrating
resources (Kageura & Marshman, 2019: 67–​68).

Many CAT tools also allow translators to pretranslate a source text.

Pretranslation means that you can instruct the CAT tool to “identify those
terms for which an entry exists in the termbase” (Bowker, 2003: 57) and
automatically insert them into the target text. The idea of pretranslation is
also to save time and effort typing the target language terms by inserting them
automatically, so that translators only need to type the rest of the sentence
around the term or terms. It is worth noting that matches from the transla-
tion memory generally take preference over those from the termbase, i.e. the
target terms are only inserted in the target text if there is no exact match in
the translation memory (Rothwell et al., 2023: 71). This is because matches
from the translation memory are matches ‘in context’ since the term appears
in a complete sentence, which are ranked above terms appearing ‘in isolation’
in the termbase. However, the CAT tool settings may allow translators to
change this depending on their needs.
A termbase can also be consulted from within a CAT tool with the con-
cordance search feature, which can be used to make one-​off consultations
and searches in both the termbase and the translation memory (see also
Bundgaard & Christensen, 2019). Some tools allow the use of wildcards in
concordance searches. A wildcard is “a character such as an asterisk (*) that
can be used to represent any other character or string of characters.” For
example, typing blade* in the concordance search box should yield both
blades and bladeless.
In the early days, CAT tools came with empty termbases and it was up
to translators themselves to fill in the termbase so that, in the medium to
long term, they could reap the benefits of the termbase. However, nowadays
several CAT tools, e.g. RWS Trados, memoQ, MultiTrans, Wordscope, etc.,
can provide direct access to major term banks, e.g. IATE, TERMIUM® and
other resources (García, 2015: 85). In addition, some CAT tools can help
translators with any non-​matched source segments and terms by populating
them with matches from machine translation, artificial intelligence or by
“enhancing manual translation with predictive typing and sub-​segmental
matching” (2015: 85).
108 Terminology Management for Translators

Termbase Export, Sharing and Collaboration

As we saw in Box 4.2 (Chapter 4), termbases are generally intended to
be reused and repurposed. In most scenarios, terminological data will be
exchanged in all directions among translators, managers, clients, etc. (unless
a termbase is created by freelance translators for personal use, for example).
Terminological data is also frequently migrated from one program or
system to another, e.g. for backup purposes or to consolidate or integrate
resources in a different way. Therefore, most termbase software will provide
several methods for you to export your termbase. Another takeaway from
Chapter 4 (e.g. Box 4.1) is that Excel spreadsheets are one way of importing
or exporting data from your termbase, but there are other ways and formats
to export your termbase, e.g. formats such as TBX (which allows export to
other termbase software), XML, HTML, .TXT formats (which allows export
to spreadsheet software) and .RTF format (which allows export to a MS
Word file, for example), among other formats.
While you are working with termbase software, the data will be stored in
its proprietary format. Since proprietary formats can only be used with the
software that created them, the “community of users successfully applied
pressure such that many tool vendors have now ensured that their tools are
compatible with standard exchange formats such as the XML-​based standard
TermBase eXchange (TBX)” (Bowker, 2015: 315), developed by the now
defunct LISA (Localization Industry Standards Association) and promoted
by ISO (see also Chapter 10). TBX was officially released in 2008 as ISO
30042 (Kageura & Marshman, 2019: 71). TBX documents consist of two
parts: one file containing the terminological data and another file containing
information about the structure of the termbase (Popiolek, 2015: 346). It
is this separation of data and structure that makes it possible to move a
termbase from one system to another without losing the structure of the
termbase (2015: 346).
With sophisticated termbase software such as RWS MultiTerm, the Word
Dictionary export option allows you to export the data in your termbase to
a Word file (.RTF format) which is automatically laid out and presented as
if it were a dictionary, as illustrated in Figure 5.2. This is a useful option if
you plan to print your termbase, for example. The header, the two-​column
layout, the individual letter sections and the line between columns are all
done automatically by the software.
Termbases can also be shared by users in cloud-​based or server-​based
systems, which are the options often used when freelance translators or
external users, e.g. domain specialists, need to consult the data too. For even
larger scale, open-​access collaborative needs, initiatives such as TermWiki4
have emerged which allow “multiple users to create, edit, search and consult
term records in a variety of fields and languages” (Kageura & Marshman,
2019: 72).
 newgenrtpdf
Populating and Using Your Termbase 109
Figure 5.2 Example of an MS Word Dictionary export with RWS MultiTerm of a Catalan-​English termbase.
110 Terminology Management for Translators

Updates, Regular Backups and Maintenance

Terminology management should be viewed as an ongoing, continuous pro-
cess in all translation workplaces, from freelance to institutional. Therefore,
a special effort should be made to keep termbases updated and backed up
(ideally both online and offline). As we have seen in previous chapters,
the need for updates is ongoing, but you should really plan to update
your termbase when concepts, definitions or variants of terms change, for
example, or if a term becomes obsolete, or if you add terms to your termbase
as you translate. It is then important to remember to go back to those newly
created term records and complete all the relevant information. As a gen-
eral rule, the best time to carry out the maintenance or updates for your
termbase is either at the beginning of a project (with new project-​specific
terminology) or at the end of it (when translation reveals that some changes
to existing terms are desired, or that new terms should be added) (Argos
Multilingual, 2021). As with other types of databases, bigger termbases
are not always better. A termbase should “relate as much as possible to a
given domain, client and project” (García, 2015: 73). Therefore, it is “usual
practice to compile multiple termbases which can be kept segregated for
designated uses (and, of course, periodically dumped into a 'big mama' term
bank too)” (2015: 73).

Follow-​up Tasks and Reflection

1. In English, many terms are noun phrases consisting of an Adjective +​
Noun, e.g. cardiac arrest. However, there are many other types of noun
phrase. Have a look at the following noun phrases and consider how
they might be translated into your other language(s) –​do they remain
as noun phrases? What are the differences between your language(s) and
English?
a. Noun +​noun: wind turbine, balloon angioplasty, climate change.
b. Adjective +​noun +​noun: fresh fruit market, fast speed train, primitive
unit cell.
c. Adjective +​adjective +​noun: ventricular septal defect, peripheral pul-
monary stenosis.
d. Proper noun +​common noun: Allen key, Basel problem, Portland
cement.
e. Noun +​preposition +​noun: cause of death, axis of symmetry, law of
constancy.
2. Would you, in principle, include words or expressions in your termbase if
they were…? Why?
a. Frequent but not leverageable, e.g. ten different inflected forms of a
noun? Yes/​No
b. Difficult but not frequent, e.g. potassium hydrogen sulphate? Yes/​No
 Populating and Using Your Termbase 111

c. Highly visible but not a term, e.g. product names? Yes/​No

d. Highly embedded in other terms and frequent, e.g. global? Yes/​No
e. Not a term but highly leverageable, e.g. will be examined by? Yes/​No
3. Entering terms in the termbase: How would you enter the terms in bold
from the table below in their dictionary form (i.e. as they should be entered
in the termbase)? The first row has been done as an example.

Term and context Dictionary form of the Notes (optional)

term

1. The most important Down’s syndrome No change. Down

thing that you need to stays with a capital D
know about Down’s because it is a proper
syndrome is that noun, but syndrome
everyone with the stays in lowercase
condition is a unique because it is a
individual.5 common noun.

2. Such larvae have well-​

developed locomotor
structures.6

3. Most bacteria in the

body are harmless, and
some are even helpful.7

4. Nuclear reactors
generate close to one-​
third of the world’s
carbon free electricity.8

Cumulonimbus clouds
have 3 distinct ‘species’
which describe the
appearance of the head of
the cloud:
Cumulonimbus
calvus –​the top of the
cumulonimbus is puffy,
like a cumulus cloud. The
water droplets at the top
of the cloud tower have
not frozen to become ice
crystals.9
112 Terminology Management for Translators

Term and context Dictionary form of the Notes (optional)

term

The first metal to be

smelted in the ancient
Middle East was probably
copper (by 5000 BCE).10

Shakespearean sonnets end

in two lines that rhyme
with one another, called a
couplet.11

4. By contrast with e­ xercise 1, what are the English plurals of the following
terms? Do some research to find out, if you don’t already know.
a. Software .....................................
b. Moose (animal) .....................................
c. Octopus .....................................
d. Platypus .....................................
e. Bureau .....................................
f. Cello .....................................
g. Samurai .....................................
h. Runner-​up .....................................
i. Attorney general .....................................
j. Brigadier general .....................................
5. From the following text,12 are the adjectives highlighted in bold with the
term they qualify likely to be modifiers, or are they part of the term? You
can either do some research to confirm one way or the other, or you can
discuss how likely they are to be modifiers or not (e.g. if you don’t have
knowledge of the domain).
a. Is aggressive a modifier in aggressive tumour?
b. Is common a modifier in common breed?
c. Is ultraviolet a modifier in ultraviolet light?
d. Is poor a modifier in poor pigmentation?
e. Is dramatic a modifier in dramatic blood loss?

What is Hemangiosarcoma?

Hemangiosarcoma (HSA; angiosarcoma or malignant hemangio­

endothelioma) is an extremely aggressive tumour of blood vessel
origin. […] HSA is a disease generally of older dogs and cats with an
 Populating and Using Your Termbase 113

average onset of 9 to 10 years […]. German shepherd dogs are most

commonly diagnosed with HSA. […] In cats, the most common breed
is the domestic shorthair. […] Ultraviolet light exposure from the sun
may be a potential cause of HSA in dogs, as HSAs of the skin are com-
monly seen in dogs with light hair and poor pigmentation (e.g. Salukis,
Whippets, and white Bulldogs). […] At the most extreme, HSA can
cause sudden death due to tumour rupture and/​or dramatic blood loss.
Episodes of weakness and collapse can last for minutes to hours with
recovery from the episode being a common feature.

6. Although it would be very difficult to create a specialised corpus that

meets all the corpus design criteria we have covered in this chapter, what
would constitute a reasonable balance? Which of the following fictitious
corpora (1 to 3) would be the best one for each of the translation projects
(a to e) below? Is there a translation project for which none of the corpora
is suitable? If not, why might this be? You can select a source language or
use your other language(s) as source language(s).
a. Project 1: translating a highly technical medical research report into
American English.
b. Project 2: translating post World War II stew recipes into British English
for a current TV programme for children.
c. Project 3: translating a repair manual for central heating systems into
Canadian English for the engineers of a multinational commercial
organisation.
d. Project 4: translating a marketing campaign for highly luxurious jew-
ellery items into Australian English for a high end online shopping
marketplace.
e. Project 5: localising a medieval-​themed videogame into English to be
sold in Ireland.

Corpus 1 Corpus 2 Corpus 3

• Very large • Small • Very small

• 2 texts • 17 texts • 5 texts
• 2 authors • 8 authors • 1 author
• Not very • Very authentic • Highly authentic
authentic • Highly focused on • Highly focused on
• Moderately the domain the domain
focused on • Very high levels of • Very low levels of
the domain specialisation specialisation
114 Terminology Management for Translators

Corpus 1 Corpus 2 Corpus 3

• Various degrees • Highly authoritative • Not very

of specialisation, • British English authoritative
from very low to • Original British • American English
very high English texts • Three original
• Moderately • From 2020 American English
authoritative texts and 2
• Australian English translations from
and South African other languages
English • From 1960 to 1980
• Some translated
texts from other
languages
• From 1945 to
present

7. Having completed the previous exercise, which criteria seem to emerge as

more important, e.g. is the authoritativeness of the corpus more or less
important than the language variant?, and so on. Rank the following cri-
teria from 1 to 9 (based or not on your answers to the previous exercise),
whereby 1 =​most important and 9 =​least important.
Quantity Authenticity/​reliability Language variant
Quality Relevant domain Authoritativeness
Number of texts Degree of specialisation Date (old, modern, etc.)
8. Why is term extraction not a fully automatic process at present? To what
extent do you think it is plausible to expect the term extraction process to
become fully automated, including term validation?

Notes
1 For example, the Farming UK Directory available at www.farmin​guk.com/​agric​
ultu​ral-​direct​ory
2 As Anthony (2022: 104) explains, care “should be taken when using these, how­
ever, as it can be unclear if the site will store and use the data for its own purposes.
They may also only convert a single file at a time, which is inappropriate for many
corpus building tasks that require huge numbers of files to be converted. One alter-
native is AntFileConverter [www.laur​ence​anth​ony.net/​softw​are/​antfi​leco​nver​ter/​
] which is an offline tool that can batch-​convert PDF and Word files into a text
format without any restriction on the number of files being processed.”
3 At the time of writing, the BNC is available at www.natc​orp.ox.ac.uk/​ but some
tools also provide access to it from within their interface.
4 https://​en.termw​iki.com/​
 Populating and Using Your Termbase 115

5 Source text available online at: www.downs-​syndr​ome.org.uk/​wp-​cont​ent/​uplo​

ads/​2022/​12/​Frie​nds-​and-​fam​ily-​leaf​l et-​web-​vers​ion-​sin​gle-​page.pdf
6 Source text available online at: www.bri​tann​ica.com/​scie​nce/​larva
7 Source text available online at: www.gen​ome.gov/​genet​ics-​gloss​ary/​Bacte​
ria#:~:text= ​ B acte​ r ia%20are%20sm​ a ll%20sin ​ g le%2Dcel ​ l ed,bacter ​ i al%20ce​
lls%20t​han%20hu​man%20ce​lls
8 Source text available online at: www.iaea.org/​new​scen​ter/​news/​what-​is-​nucl​
ear- ​ e ne ​ r gy- ​ t he- ​ s cie ​ n ce- ​ o f- ​ n ucl ​ e ar- ​ p ower#:~:text= ​ N ucl ​ e ar%20ene ​ r gy%20
is%20a%20f​orm,fus​ion%20%E2%80%93%20w​hen%20nuc​lei%20f​use%20t​
oget​her
9 Source text available online at: www.metoff​i ce.gov.uk/​weat​her/​learn-​about/​weat​
her/​types-​of-​weat​her/​clo​uds/​low-​level-​clo​uds/​cumul​onim​bus
10 Source text available online at: www.bri​tann​ica.com/​tec​hnol​ogy/​smelt​ing
11 Source text available online at: www.bbc.co.uk/​bites​ize/​artic​les/​zwn6​6g8#
zpt2​2v4
12 Adapted from the source text available online at: www.acade​mia.edu/​33832​
619/​Ettinger_​and_​Feldman_​Textbook_​of_​Veterinary_​Internal_​Medicine_​Client_​
Inform​atio​n_​Sh​eet_​Hema​ngio​sarc​oma
13 www.sta​tmt.org/​europ​arl/​
14 https://​opus.nlpl.eu/​
15 www.sketc​heng​ine.eu/​
16 https://​joint-​resea​rch-​cen​tre.ec.eur​opa.eu/​langu​age-​tec​hnol​ogy-​resour​ces/​dgt-​tran​
slat​ion-​memory​_​en
17 https:// ​ w arw ​ i ck.ac.uk/​ f ac/​ s oc/​ a l/​ r ep​ o sit​ o ry/​ s taff/​ h arris​ o nti​ l ly/​ c orp​ o ra-​ f or-​
works​hop/​
18 http://​corp​ora.lancs.ac.uk/​clmtp/​
19 http://​corp​ora.lancs.ac.uk/​
20 https://​cor​pws.cymru/​?lang=​en

Further Reading
Bowker and Pearson (2002) provide a great starting point for those who are learning
about corpora from scratch, including markup and annotation, and there are
exercises in each chapter. In a similar vein, Corpas Pastor and Seghiri (2009) illustrate
how to compile a specialised corpus step by step and the edited volume by Hundt
et al. (2007) provide useful insights into corpus linguistics and the web. Frankenberg-​
Garcia (2015) lists some challenges of training translators to use corpora. McEnery
and Xiao (2007) explain the usefulness of parallel and comparable corpora to transla­
tion studies. Two edited volumes, those by Lefer and Granger (2022) and by O’Keeffe
and McCarthy (2022b) are up-​to-​date reads not only about guidelines for the design
and use of corpora but also about the past, present and future of corpus linguistics
including possible pathways for future research. Zanettin has published extensively
on using corpora in translation practice (2002a, 2011, 2014, 2015) and on translator
education (Zanettin et al., 2003).
Some large parallel corpora are freely available online, such as the Europarl
corpus,13 described in detail by Koehn (2005), and the OPUS14 collection of corpora
which, to date, hosts more than 1,200 corpora online (monolingual or bilingual) in
a huge range of languages. For readers with access to Sketch Engine,15 many free
116 Terminology Management for Translators

corpora are provided there, as well as the option of creating your own corpora, which
we explore further in the following chapter. The Directorate-​General for Translation
(DGT) of the European Commission has made its translation memories freely avail-
able online16 and a guide on how to extract bilingual data can also be downloaded
from there. Simard (2020) offers insights into building parallel corpora for trans­
lation. Li and Ye (2023), in the second edition of the Routledge Encyclopedia of
Translation Technology, providing a long list of corpora per language. Many English
language corpora are available in the Applied Linguistics Repository of the University
of Warwick.17 McEnery and Hardie have published a textbook on corpus linguistics
on the website of Lancaster University.18 Other corpora applications and utilities can
also be found on another webpage of Lancaster University.19
There is also a great deal of literature on using corpora with specific languages,
including lesser-​resourced languages. A selection is included here:

• African languages: Esimaje et al. (2019) focus on corpus linguistics for African
varieties of English, whereas Prinsloo et al. (2018) explore online African language
dictionaries. Also, Chabata (2023) explores the development of Zimbabwe’s under-​
studied languages and Taljard & de Schryver (2002) study term extraction for nor­
thern Sotho.
• Arabic: the edited volume by McEnery et al. (2021) is devoted to Arabic corpus lin­
guistics, whereas Mörth (2018) provides a useful overview of corpora and digital
resources in Arabic.
• Basque: Gurrutxaga et al. (2010) turn to the internet to build language resources
for Basque.
• Catalan: Vàzquez (2024) uses a combination of natural language processing tools
to maximise Catalan terminology currently available in open access.
• Chinese and Asian languages: Yang (2006) looks at corpus-​based research in China,
whereas Zhang et al. (2008) carry out term extraction from Chinese documents;
Kim et al. (2001) extract collocations from a Korean corpus.
• European languages: Archer & Rayson (2006) is another edited volume that looks
at corpus linguistics around the world, including many European languages and
Brazilian Portuguese; Pinnis et al. (2012) focus on Croatian, Latvian, Lithuanian,
and Romanian; Doğru (2019) extracts terms from a Turkish to English medical
corpus, whereas Koeva (2007) extracts terms from a Bulgarian corpus.
• Welsh: Andrews & Prys (2016) explore the standardisation of Welsh terminology
in education and the Welsh National Language Technologies Portal20 provides
“corpora resources that can be downloaded by developers and researchers.”

Other studies focus on newly written languages such as those of the Zapotec and
Chatino communities in Mexico, Tibetan languages, Eton ethnic group in Cameroon,
Javanese and Arabizi (Millour & Fort, 2020), whereas Zavorin et al. (2020) investi­­
gate the creation of corpora in six other lesser-​resourced languages: Tagalog, Swahili,
Somali, Lithuanian, Bulgarian and Pashto.
For a more in-​depth overview of text mining, see Berry and Kogan (2010). For a
getting started guide on term extraction, see TerminOrgs (2022). For examples of
automatic multilingual term extraction tools used in specialised corpora, see Periñán-​
Pascual (2017) and Valderrábanos et al. (2002). For other aspects of term extraction
see Thurmair (2003), Heylen and De Hertog (2015) and Saralegi et al. (2008). By
 Populating and Using Your Termbase 117

contrast, there does not seem to be a great deal of literature on using termbases with
CAT tools, except for what has already been quoted in this chapter, but additional
studies that deal with CAT tools (although not necessarily termbases) include Bellino
Machado (2015), Heylen et al. (2014a) and Horner (2021), to name but a few. For
example, at the time of writing, the Sketch Engine corpus interface is a repository of
over 500 corpora across 95 languages (Kilgarriff et al. 2014).
6 Tools for Terminology Management

Key Questions

• What tools are there to carry out terminology work?

• How can software help with the various terminological tasks?
• Is there one single terminological program that does everything?
• What do I need to know before using software for terminology work?
• How can the new AI-​powered LLMs be used for terminological tasks?

What Tools can I Use for Terminology Work?

In this chapter, we explore how software can help with the various termino-
logical tasks, as summarised in Figure 1.2 in Chapter 1 and revisited in
Figure 6.1. In this figure, there are many different aspects to the terminological
work of the translator; apart from carrying out a translation job containing
specialised terminology, the preparatory work to acquire or maintain the
relevant knowledge of the domain, conceptual and terminological, includes
tasks such as text mining, corpus building, term extraction, etc. There is an
increasingly vast array of tools –​too many to list here –​that can be used for
many of these tasks, some of them freely available online. In addition, not all
the tools can carry out all of the terminological tasks needed by translators
(i.e. the ones in Figure 6.1). Instead, we take a concept-​based overview of sev­
eral categories of tools that might support various terminological activities
although, of course, there will be some overlap between tools, e.g. some of
them can do corpus building and term extraction, whereas others may offer
text mining and term extraction, etc.
As with previous chapters, we will not take a “user guide” approach,
as the user guides created by software providers are already comprehensive
and many of them are available online. Instead, we will focus mostly
on conceptual explanations of the software, to help readers understand
what each type of tool does. However, some examples of currently available
tools are mentioned to illustrate the concepts behind the tools and there

DOI: 10.4324/9781003302636-7
 Tools for Terminology Management 119

Figure 6.1 Overview of software tools for terminological work, revisiting Figure 1.2
from Chapter 1.

are lists of tools and additional exercises and guides on the online portal
for this book, with step-​by-​step instructions on how to use some tools for
terminological tasks, especially in cases where the software was not origin-
ally designed for terminology work. Making these user guides available on
the online portal means that they can be updated as new software versions
come out.
To start with, search engines are tools that can be used for text mining.
The texts found in this way can be used to create a corpus, as we saw in
Chapter 5. The texts may contain useful information, such as definitions,
which can then be stored in the termbase. Once compiled, the corpus can be
processed with concordancers, corpus processing tools and other language
processing tools, to extract terms as well as various types of terminological
and phraseological information. In addition, if parallel texts in source and
target languages are found, text alignment tools can be used to import the
texts into the translation memory of a CAT tool, for use during translation.
There are dedicated term extraction tools which can be used to identify the
terms in a corpus (see Term Extraction in Chapter 5). Other tools include
concept maps and other visualisation tools, including mind maps, which can
help translators with understanding the conceptual network of a domain. In
Chapter 5, we also saw that CAT tools typically come with built-​in termino­
logical capabilities, such as a termbase to store the terms and functions such
as pretranslation and fuzzy matching to retrieve the terms as effectively as
120 Terminology Management for Translators

possible. The (monolingual) texts identified by means of web crawlers and

search engines can be used as reference texts during translation, as some CAT
tools allow this.
Many CAT tools come with integrated MT functionality, underpinned by
artificial intelligence (AI) technology, which is already being used by some
programs for term recognition during translation, and large language models
(LLMs), such as ChatGPT, which can also serve as a tool for the termino-
logical work of translators. The final section looks at AI and LLMSs, as well
as a range of other tools that can also make terminological work more effi-
cient, such as meta-​search tools. This chapter concludes with follow-​up tasks
and reflection and with some pointers for further reading.

Search Engines as Text Mining and Corpus Building Tools

The Internet is a massive resource and, although it may be one of the first places
translators use to find information, it may be daunting to locate relevant (and
reliable texts) for terminological work. In addition, you may be overwhelmed
trying to sort through the results you obtain (Bowker & Pearson, 2002: 61).
Search engines have been described as “our gateways to the web” (Fletcher,
2007: 25), as they may help translators zoom in on appropriate texts for
corpora building and term extraction from the wild online jungle out there.
However, translators should exercise caution when using them (as we see
in Box 6.1). A search engine is a tool “that will search the Internet for sites
containing the words that you enter as search terms” (Mirtaheri et al., 2014).
In this section, we take a look at three kinds of search engine, i.e. commer-
cial, web-​based and meta-​search engines, we suggest some more proficient
methods of using them than just typing random words in the search box and
review some of their limitations (see also Box 6.1).

Commercial Search Engines

Commercial search engines, such as Google, Baidu, Bing, etc., are widely
used by translators and linguists alike for text mining and other documenta-
tion activities. However, in order to use a search engine, you need to know
in advance what you are looking for, i.e. you need to type something in the
search box. Experienced or specialised translators may already know the
keywords of a domain which they can type in order to maximise search
results. However, if you are researching a domain from scratch, or if you are
not familiar with it yet, the typical process is to identify a few seed words,
i.e. some preliminary key words and phrases which you can expect to be
“representative of the domain under investigation” (Baroni & Bernardini,
2004: 1314) and which will get your documentation task under way. You
can identify these seed words in various ways, e.g. from previous translations
or by using the name of the domain (e.g. molecular biology) as a keyword in
free reference resources such as Wikipedia.1
 Tools for Terminology Management 121

Box 6.1 Caveats for Using Search Engines for Terminological Work

• We still know very little about the size of the web as a corpus: the
text types it contains, the quality of the material included or the
amount of repetitive ‘junk’ that it ‘samples.’
• Furthermore, due to the ephemeral nature of the web, constantly chan-
ging and growing replicability or reproducibility of the results is impos-
sible; the search results you obtain today may be different tomorrow,
and a text or webpage you find today may have disappeared tomorrow.
• Other problems have to do with the way that the commercial
crawlers work: they cannot access all webpages because some pages
are ‘invisible.’
• More worrying still –​the commercial crawlers have an inbuilt local
bias. Commercial crawlers apparently prioritise hits that are closer
to the ‘home’ of the individual user, which may lead to different
results depending on which country the website is accessed from.
• Crawlers also build up a profile of the user and since we rarely use
crawlers for linguistic searches only, this may produce an additional
skewing effect.
• When using search engines, it is imperative for translators to avoid
the naïve approach of entering “key words into a search engine and
simply retrieve the first ten hits, thinking that any text containing
these key words will be a useful source of information for their
purposes” (Bowker, 2015: 317).
• As Bowker further explains, the result of such a naïve approach is
texts that are inappropriate in terms of register, technicality and
type; learning to exercise good judgement remains key (2015: 321).
• For more on why “Googleology is bad science,” see Kilgarriff (2006).

Additional issues with the use of search engines can be found in Schäfer
& Bildhauer (2013: 2–​ 3). Despite the rather unfavourable impres­
sion the above factors may give about carrying out web searches, the
Internet is and probably will be one of our best sources of information.
The message for translators, therefore, is to make a discerning use of
the web, taking additional steps, if necessary, to ensure the provenance
of the texts and other information you obtain from Internet before you
use it for terminological tasks in translation.

Source: Adapted from Hundt et al. (2007: 2–​3).

In order to research a specialised domain, you only need a small list of seeds.
According to Baroni and Bernardini, a list of 5 to 15 seed words is typic-
ally sufficient and they report that they obtained interesting results from an
122 Terminology Management for Translators

experiment using “as few as two seeds” (2004: 1314). Ideally, seed words
should be as unique to the domain as possible in order to get good search
results. For example, lens, megapixel and ISO are better seed words for pho-
tography domains than photo and image, which are less suitable because
they feature in many domains, not just photography.
Another good way to zoom in on appropriate texts with search engines
is to use Boolean operators and/​ or the advanced settings of the search
engine. For search engines that accept them, Boolean operators are words
such as AND or OR (often typed in uppercase between the search words)
which are used as filters in order to obtain more relevant or focused results.
For example, typing red AND blood AND cells will clearly yield results
containing all three words, red, blood and cells. By contrast, typing red AND
blood AND cells NOT white should ideally exclude results containing white
blood cells; typing red OR white AND blood AND cells should ideally yield
results containing red blood cells or white blood cells. The advanced settings
of search engines can also provide ways to zoom in on the desired results.
For example, in Google advanced searches,2 as well as selecting to include or
exclude words, there are additional filters, shown in Figure 6.2.
This figure shows that you can narrow down your results by language, i.e.
only search for texts in a specific language; by region, i.e. only show results
from a specific country from the list that appears; the setting by last update
allows you to focus on recent results only, from 24 hours old up to a year

Figure 6.2 Some of the advanced search settings in Google.

 Tools for Terminology Management 123

old; by site or domain is a very helpful setting to narrow down your searches
considerably. For example, by typing aiaa.org in the site or domain box in the
above figure, you should get only the results that appear on the website of the
American Institute of Aeronautics and Astronautics (AIAA). You can also use
this setting to narrow down your searches to a particular country, e.g. .es for
results from Spain, or context, e.g. typing .ac.uk allows you to retrieve only
academic results from the United Kingdom. The terms appearing setting can
help narrow down your search by topic by instructing Google to return only
results which contain the key word in the title or the URL of the webpage.
Typing red blood cells here and selecting in the title of the page will yield pages
that contain this term in the title. The file type setting allows you to focus your
search on a specific file format, e.g. .PDF, or .PPTX. The usage rights setting
helps you to discern results with different licensing requirements.
The searches discussed so far may work well for individual words, but for
multiword terms (and paradigmatic and syntagmatic relations of terms), other
search strategies are more helpful. According to Rothwell et al. (2023: 95),
three of the best ways to search collocations (which also works for multiword
terms) include the phrase search (i.e. typing a phrase in inverted commas, e.g.
“wind power”), the site setting as we saw in Figure 6.2 and wildcards, as
we saw in Chapter 5. These methods of searching can be used separately
but they can also be combined. For example, in order to find out whether
spread (as a verb or a noun) collocates with the term virus according to the
Royal Pharmaceutical Society, you can type the following in a Google search
box: “spread * virus” site:rpharms.com, as shown in Figure 6.3.
As shown in Figure 6.3, using inverted commas covering both spread and
virus ensures that the results contain both terms. The asterisk in the middle
means that there can be any number of words between spread and virus (e.g.
spread the virus, spread of the virus, etc.) and the specified site, rpharms.com,
ensures that only results from this website are returned. This search can be
used with other websites and other terms until you obtain the desired infor-
mation. This is just an example of a more focused search that may yield more
suitable results for terminology work. The general message is that search
engines are extremely useful tools for the terminological work of translators,

Figure 6.3 An example of a focused search on Google.

124 Terminology Management for Translators

but it is worth learning more sophisticated ways of using search engines than
just typing words in the search box. It is beyond the scope of this chapter to
provide a full review of sophisticated search engine use, but a list of suggested
resources is provided in the online portal.
Despite the limitations summarised in Box 6.1, translators from all
workplaces, from freelance to institutional, often resort to commercial search
engines for information, and will continue to do so even if, in the case of
institutional translation, they are provided with resources and sometimes
internal search engines to browse the resources. The above suggestions for
using search engines in more proficient ways than lay users may be helpful
for ad-​hoc searches and consultations, or if you only need to find a few texts.
However, those suggestions may be far too time-​consuming if you are trying
to compile a corpus. For this purpose, there are web-​based search engines
that essentially work as corpus compilation tools which allow you to put
together a very large corpus in a very short time, as we explore in the next
section.

Box 6.2 Using RegEx (i.e. Regular Expressions)

What Are They?

• Regular expressions (also known as RegEx) are a very powerful

resource and open a full range of possibilities in different programs,
including Sketch Engine and some CAT tools.
• You can think of regular expressions as a search-​ and-​replace
function “on steroids” (i.e. a much more powerful function than
simply searching and replacing).
• A regular expression is a special sequence of characters or symbols
that define a search pattern. This pattern is then used to search for
(or replace) specific instances of words or phrases in a text. Regular
expressions are used by search engines, text editors, text processing
utilities, and for lexical analysis.

Why Are They Useful for Translators?

Regular expressions can assist your translation work by allowing you to
search, replace, and filter text in ways that would otherwise be impos-
sible in translation software tools.

What Are Examples of Regular Expressions?

At first, regular expressions may appear cryptic, but once you have
learned the basics and seen how useful they can be, you will be able to
 Tools for Terminology Management 125

decide how much time you want to invest to become more proficient
at using them.

How Can They Help the Terminological Work of the Translator?

By improving the capabilities of the search and search-​
and-​
replace
functions and allowing you to:

• Perform two or more separate searches at the same time (e.g.

searching for different forms of the same term, or perhaps for com-
pletely different words).
• Filter text in your CAT tool to display only those segments that are
capitalised differently between the source and target.
• Search a glossary for all capitalised headwords and change them to
lowercase while leaving acronyms and other terms that are in all
uppercase unchanged –​and do all this in a single operation!
• Filter text in your CAT tool to find any segments where the end
punctuation differs between the source and target.
• Filter text in your CAT tool to display only the segments that contain
certain words in the source, or the segments that do not contain a
specific word in the target.
• Etc.

Where Can I Read More About Them?

First, read the excellent ATA article on regular expressions (2019)
on which this Box is based. You will find many examples of regular
expressions, screenshots and even a cheat sheet. This article also tells us
where to read more about regular expressions, e.g. RegexBuddy resources
such as those at Regular-​Expressions.info3 and at RegexBuddy,4 and
other resources such as Regular Expressions 101.5
A final suggestion from this article is that “[r]‌egardless of the tool you
use, […] keep a list of the regular expressions you use and write a brief
description to remember what each does. No need for anything fancy,
a simple text file will do.”

Source: Adapted from ATA Chronicle, 2019.

Web-​based Search Engines

With this kind of search engine, the web can be used as a corpus itself, either
as “a source of machine-​readable texts for corpus compilation” (Fletcher,
2007: 28), a method known as Web for Corpus (WfC), or by using the web
126 Terminology Management for Translators

directly as one huge corpus, a method known as Web as Corpus (WaC).

Although not everyone agrees that the web can be considered a corpus in the
sense described in Chapter 5, there is consensus that it constitutes a kind of
corpus (Fletcher, 2001: 3; Zanettin, 2002b). Examples of tools in this cat­­
egory include several tools that are no longer functional, e.g. KWiCFinder,6
or actively maintained, e.g. TranslatorBank,7 but their websites still contain
some useful reference material. Tools available at the time of writing include
WebCorp Live8 and WaCky (The Web as Corpus Kool Ynitiative).9 Other
suggested tools are listed in the online portal with their URLs.
Internet-​based search engines differ from commercial search engines
in that “they pre-​process the questions before sending them to the search
engines and then post-​process the results and present them in such a way as
to facilitate linguistic studies” (Buendía-​Castro & López-Rodríguez, 2013).
The limitations of these tools coincide to a great extent with those of com-
mercial search engines (see Box 6.1). Using WebCorp Live as an example,
this type of tool “adds refinement options not provided by commercial search
engines” (Renouf & Kehoe, 2013: 168), such as:

• More powerful wildcard search capabilities: Google does not allow

wildcard-​initial words as search terms (e.g. typing *athon to find words
like marathon), but WebCorp Live does (Renouf et al., 2007: 62); inter­
estingly, WebCorp Linguist’s Search Engine10 (or WebCorp LSE), a child
project of WebCorp Live, also allows the hash operator (i.e. #) to find
all cosmetic variants of a term (see also Term Variation in Chapter 4).
For example, typing web#site will yield web-​site, web site, and website
(Renouf & Kehoe, 2013: 191).
• Case sensitivity: WebCorp Live allows you to specify or ignore case in
word retrieval. For example, you can choose to retrieve results with Queen
and ignore results with queen, or vice versa, whereas searches in Google
are case-​insensitive.
• KeyWord-​In-​Context (KWIC) display: WebCorp Live “organizes the
results […] in a clean ‘keyword in context’ format, similar to that of
standard concordancing programs” (Lüdeling et al., 2007: 16) (also see
the next section).
• Date-​sorted output: post-​search, WebCorp Live allows diachronic searches,
i.e. searches for terms at different times, by selecting the relevant values in
the filter by date settings. This is not possible in the standard searches in
Google, although it can sort results by their last update (Renouf & Kehoe,
2013: 168).
• Other post-​processing functionality: available in WebCorp Live includes
frequency lists, collocate display, use of stopwords, etc., not usually avail-
able in standard commercial search engines.

These differences are not too surprising if we take into account that com-
mercial search engines were not designed specifically for translators (or
 Tools for Terminology Management 127

even linguists). Ultimately, however, WebCorp Live and similar tools are
“interfaces to Google and other search engines” (Lüdeling et al., 2007: 16)
and as such:

• “they are subject to all the query limitations that the engines impose;
• they cannot provide information that is not present in the data returned
by the engines; and
• they are subject to constant brittleness, as the nature of the services provided
by the engines may change at any time” (Lüdeling et al., 2007: 16).

Another tool that may be worth mentioning in this section is Sketch Engine.11
This web-​based program provides many tools for linguistic analysis, many
of which can be used for terminological purposes, as we explore in the
following sections. In the meantime, regarding text mining, Sketch Engine
can be used to create a corpus in two ways: (i) as a search engine, much
like WebCorp Live; or (ii) by uploading the user’s own texts. In fact, Sketch
Engine is probably the only tool that allows the import of custom data to
create a corpus (Anthony, 2022: 107), at least at present. In either case, once
a corpus has been compiled, it can be analysed with a wide range of linguistic
tools, including term extraction and concordancing tools. If what you need is
to compile a bilingual corpus, you will first need some (text) alignment tools
(see below).

Meta-​search Engines and Special Purpose Search Engines

Search engines are “among the most successful applications on the web
today” (Howe & Dreilinger, 1997: 19) but “internet content is considerably
more diverse and certainly much larger than what is commonly understood”
(Bergman et al., 2000: 1) and it has been claimed that about 85% of search
engine users “cite the inability to find desired information as one of their
biggest frustrations” (2000: 1). According to Howe & Dreilinger (1997: 19),
“[s]‌o many search engines have been created that it is difficult for users to
know where they are, how to use them and what topics they best address.”
To alleviate the burden, metasearch engines can be used. A metasearch engine
is a system that provides simultaneous access to multiple existing search
engines and ranks the results by the user’s preferred search engine(s) (see
Meng et al., 2002: 50). In other words, with a single search, users can receive
results from multiple search engines. As well as the convenience of using
multiple search engines simultaneously, an advantage of using such engines
is that they increase the search coverage of the web (Meng et al., 2002: 50)
by a factor of 3.5 (Bergman et al., 2000: 7).
Despite the obvious convenience of use, they also have limitations. For
example, not all metasearch engines consult the largest search engines, some
do not give the user more than ten results from each search engine and some
list paid listings first (SACLANT, 2002: 43). It is always advisable to be aware
128 Terminology Management for Translators

of the capabilities and the types of results you get with the metasearch engine
you would like to use. At the time of writing, examples of metasearch engines
include MagicSearch12 and Lexicool,13 with other tools and suggestions for
use in the online portal for this book.
Other search engines are best referred to as special purpose search engines
which “focus on documents in an organization or in a specific subject area”
(Meng et al., 2002: 49). Many search engines can be included in this category,
covering a wide range of topics and disciplines. Examples of special purpose
search engines include Open MD,14 which produces high-​quality medical
content, Google Scholar,15 which returns academic results and WIPO Pearl16
which can be used to search for terms that have appeared in patents. In add-
ition, “most organizations and business sites have installed search engines for
their pages” (Meng et al., 2002: 49), e.g. the World Health Organization,17
which can be an efficient method to narrow down results by the primary
domain of action of the organisation.

Text Alignment Tools

During a text mining task, you may find parallel texts, i.e. source texts
and their corresponding translations (see also Corpora and Text Mining in
Chapter 5), from which you could create a parallel corpus. Translators are
often already familiar with parallel texts as they typically work with trans-
lation memories, which are essentially a “very specialized kind of parallel
corpus” (Fantinuoli, 2016: 63). Translators can use parallel texts in several
ways. They can, for example:

• “[E]‌nable translators to look up how a particular translation issue has

been dealt with by others and draw inspiration from this” (Kenning,
2010: 494).
• “[H]‌elp translators to gain a better understanding of the text and its ter-
minology, to identify potential target language equivalents, and to develop
appropriate target language phraseology” (Kenning, 2012: 494).
• “[O]‌ffer unprecedented opportunities for translation research” (Kenning,
2012: 494).
• Etc.

Once you have compiled a parallel corpus, you can use corpus analysis tools,
which we explore in the following section, to process it; you can also use
term extraction tools, which we also explore below, to extract terms from
the corpus; and you can ‘feed’ the texts into a translation memory, so that
you can use the texts for leverage and/​or for reference during translation. In
order to make parallel texts compatible with translation memory software,
you will need an alignment tool, which many CAT tools offer by default,
although there are also standalone alignment tools. During the alignment
process, texts are paired and then broken up into sentences (or segments
 Tools for Terminology Management 129

at sentence level such as headings, sub-​headings, captions, etc.) (García,

2015: 75). The resulting bilingual segments will become translation units
in the translation memory. Aligned parallel texts are sometimes referred to
as bitexts or bi-​texts (Simard, 2020: 78). Alignment “binds together a given
source segment and its equivalent translation in one or more target languages
[…] along with relevant metadata” (Rothwell et al., 2023: 40). Alignment
is a powerful tool but mostly in the specific scenario where you have “sets
of already translated documents that correspond closely to new documents
that now have to be translated” (Díaz & Zetzsche, 2022), so that it yields a
useful number of high-​quality matches during translation. There is not much
point in spending valuable time and effort on aligning “less relevant material
which is unlikely to repay the cost in terms of useful leverage” (Rothwell
et al., 2023: 42).
Alignment tools use various text features, e.g. punctuation, etc., to make
the source-​target correspondences (Rothwell et al., 2023: 40–​41) (unless they
have linguistic information built in). However, “performing an alignment is
not always straightforward” (García, 2015: 75) because translation is hardly
ever a word-​for-​word transfer between languages. It may be that the source
segment corresponds to two or more segments in the target text, or vice
versa, and it may be that a particular source segment has been paraphrased
more freely in the target language, rather than translated literally, etc. It
will also depend on how much linguistic knowledge has been integrated
into the alignment tool; early alignment tools worked without any lin-
guistic knowledge and relied on the readily observable features of texts such
as punctuation, formatting, numbers, etc. and worked by matching strings
of texts from the source language to strings of text in the target language.
Therefore, after the alignment tool has made a first attempt to automatic-
ally align source and target segments, translators are usually required to
check the quality of the alignments and correct any misalignments, before
saving any useful, high-​quality segments in the translation memory. This has
given rise to various types of alignment; beyond the one-​to-​one matching of
segments (sometimes shortened to 1:1), the alignment tool of RWS Trados,
for example, also allows translators to make 1:2 connections (i.e. one source
segments corresponds to two target segments), 2:1, 1:3, 3:1, 1:4, 4:1, 1:5 and
5:1 connections (Avila, 2018). In addition to the alignment tools included
within CAT tools, currently available examples of standalone alignment soft-
ware include LF Aligner,18 YouAlign19 and the Wordfast Aligner.20 For more
on alignment tools, see Rothwell et al. (2023), García (2015) and the online
portal for this book.
For the terminological work of the translator, aligned parallel texts (as
well as monolingual texts and corpora), whether they are stored in a trans-
lation memory or used as reference files during translation, can be the raw
material from which to extract valuable phraseological and terminological
information before, during or after a translation project and/​or when they
carry out domain research with a view to developing a specialism. In order to
130 Terminology Management for Translators

obtain this information, you can use corpus analysis tools and term extrac-
tion tools, which we explore in the following two sections.

Concordancers: Corpus Analysis Tools

Concordancers are one of the best known corpus-​processing tools (Kenny,
2011: 460) and you need one in order to extract information efficiently from
a corpus (see Box 6.3). Concordancers are tools “designed to find all the
occurrences of a search term in a corpus and display these in an ordered
fashion together with the words that surround them” (Anthony, 2022: 115).
Concordancers date back to long before the advent of computers (Zanettin,
2015: 437), when “biblical scholars and their teams of minions pored over
the Christian Bible and manually indexed its words, line by line, page by
page” (O’Keeffe & McCarthy, 2022a: 5–​6). The first computer-​generated
concordances appeared in the late 1950s and used punched-​ card tech-
nology for storage (2022a: 5–6). With advances in computing, O’Keeffe and
McCarthy also note that, by 2021, “concordancing programmes [could] rep-
licate the work of 500 monks in nanoseconds” (2022a: 5–6). Well-​known
examples of concordancers include AntConc,21 ParaConc22 for multilin­­
gual concordancing (Barlow, 2002), and WordSmith Tools23 but other tools
are listed in the online portal, with suggestions for their use in termino-
logical tasks. According to Bernardini (1997), working with concordancers
promotes skills such as awareness, reflectiveness and resourcefulness which
“distinguish a translator from those unskilled amateurs.”
The advantage of a concordance over reading a text in the normal, linear
fashion is that it affords a different way to access the information, revealing
patterns of use in a text that might go unnoticed to the naked eye, and that it
provides this information more quickly and reliably than human processing
can. Once a concordancer has crunched the numbers and produced some
data, it is up to the translator to interpret what the data mean. As Hunston
explains (2022: 143), concordancers find and organise the data in different
ways but interpreting the data is a human activity.

Box 6.3 Using Concordancers for Terminological Work

Although concordancers were not designed with terminology in

mind –​they were mainly aimed at linguistic researchers, compu-
tational linguists and lexicographers, among others (Fantinuoli,
2016: 65) –​you can use concordancers to obtain many kinds of data
from a corpus which can inform the terminological work of the trans-
lator. Translators can use concordancers to help them both with the
source and target texts:
 Tools for Terminology Management 131

• Re. source texts: understand better the content (and language) they
are translating from.
• Re. target texts: “confirm candidate translations and find unforeseen
solutions to translation problems” (Zanettin, 2015: 439).

More specifically, and depending on the functionality available in a

given tool, concordancers may also be used to:

• Identify new (multiword) terms.

• Identify term collocates.
• Validate or confirm candidate terms.
• Identify paradigmatic and syntagmatic term relations.
• Etc.

Another use of concordancers is to find knowledge patterns (see also

Chapter 3), “often in combination with a term of interest” (Marshman,
2022: 297).

Concordancers typically display the information using a visualisation format

known as KWIC (Key Word In Context) (Bowker & Pearson, 2002:13), as
shown in Figure 6.4, whereby the search term, in this case bioluminescent
(text from Smithsonian Ocean, 2018), is shown in the centre of the display.

Figure 6.4 
Example of KWIC display for the search term bioluminescent with
AntConc.
132 Terminology Management for Translators

The display is also referred to as concordance24 and the analysis of

concordances is known as concordancing (Anthony, 2022: 115). Each row is
referred to as a concordance line. Users can decide how many words to view
either side of the search term. The total number of words surrounding the
search term is referred to as span, i.e. the “window of words to both sides of
the search word in which to find collocates” (Römer & Wulff, 2010: 111).
The words within the span are often referred to by a number and a capital
L (for “left”) or R (for “right”). For example, 3L refers to three words to
the left of the search term, whereas 4R refers to four words to the right of
the search term. In this case, the span is 3L to 4R. A commonly used span
in corpus linguistics is 5L to 5R, i.e. five words to the left and five to the
right (2010: 111), but a common span for translators is yet to be determined
(though it may be difficult to do so given the wide range of text types and
topics translators work with).
For translators who are accustomed to working with CAT tools, it is
worth highlighting a few aspects of using concordancers which might not be
obvious to them at first glance:

• Translators may be used to working with text at the sentence level, but
concordance lines do not correspond to sentences (Anthony, 2022: 115),
although some concordancers do give this option; as we have seen,
concordancers typically focus on displaying the key word in the centre,
with its contextual span either side.
• The span can be extended, i.e. many concordancers provide the option
of viewing a larger amount of context around the search term, and some-
times they can even direct the user to the entire file from which the search
term derives.
• Concordance lines are often ordered by frequency of occurring patterns
(Anthony, 2022: 115), not necessarily alphabetically.
• Concordance lines can be read vertically as well as horizontally (Römer &
Wulff, 2010: 109); by reading the concordance line vertically, you focus on
each occurrence of the word or term of interest; by reading horizontally,
you “get access to phraseological patterns and to the meanings expressed”
(2010: 109) by each occurrence of the search word.

With these few pointers in mind, concordance lines can be browsed and
the results re-​
sorted until the desired information is found. However,
concordancers often offer other tools with which to process the corpus.
Although some names may vary from one program to another, here are some
of the most common tools offered by concordancers:

• Raw frequency count, i.e. total number of words in a corpus. Knowing

how big your corpus is allows you to make informed decisions as to
whether more texts are needed, for example, or to compare your corpus
to other corpora.
 Tools for Terminology Management 133

• Concordancers can also count the number of types and tokens in your
corpus. Tokens refer to every single occurrence of a word in a text; for
example, in the sentence The genetic code is the blueprint for life there are
eight words and, therefore, eight tokens. However, some words in the sen-
tence are repeated, e.g. the. Although the occurs twice, if we only count it
once, because the is the same article in both occurrences, we are counting
the types. Thus, types refer to each ‘unique’ word in the corpus. Types
should not be confused with lemmas, which are also known as dictionary
forms (as we saw in Chapter 5). For example, in the sentence DNA is a
molecule and is made up of two strands, there are 11 tokens (i.e. 11 words)
and ten types (because is appears twice but it is one type). The lemma of is
is be, which is also the lemma of were, been, etc. If you know the number
of types and tokens in your corpus, working out the type-​token ratio
(TTR) can measure the lexical density or richness of your corpus (Gries,
2022: 173). TTRs consist of “the number of types in a certain (part of a)
text divided by the number of tokens in the same corpus” (2022: 173).
A “high TTR indicates that the corpus contains many different words
(lexically rich); a lower TTR means that the corpus does not have many
different words and could be terminologically dense” (Messina, 2018: 4).
• Normalised frequency: raw frequency may not always be the most useful
measure. It may be that a word or term displays “more frequency simply
because the size of the corpus is bigger” (Jones, 2022: 130). For this reason,
many concordancers also provide a normalised frequency, that is, the fre-
quency per million words. For smaller corpora, as Jones points out, “using
a per 1,000 figure is generally considered to be most useful” (2022: 130).
• Word lists: most concordancers offer a word list tool. As its name suggests,
this tool lists all the words in the corpus, alongside their raw and/​or
normalised frequencies in the corpus. The words can be sorted out alpha-
betically, by frequency and sometimes by a mixture of both. Stopword
lists may be applied to exclude frequent function words (see also Box 5.3).
Word lists may be useful “for giving you an indication of what a text or
corpus is about” (Bowker & Pearson, 2002: 114–​115), but frequency is
not everything in terminology; critical terms may occur only once in a text
or corpus (TerminOrgs, 2022: 8). This is why many concordancers also
offer a keyword list tool.
• Keyword lists: a keyword list is a list of word frequencies in your corpus
(i.e. a specialised corpus), which have been compared to their frequencies
in a reference corpus (general language). By comparing these frequencies,
concordancers can work out the keyness of a word in your (specialised)
corpus. Keywords are those that appear “significantly more (positive
keywords) or significantly less (negative keywords)” (Jones, 2022: 130)
in your corpus compared to the reference corpus. Keyness is based on the
idea that terms have increased frequencies in specialised corpora from their
domain, compared to their frequency in general language or reference cor-
pora. For example, Römer and Wulff (2010: 106) found that terms such
134 Terminology Management for Translators

as species, gene or protein were more frequent in their specialised biology

corpus than words such as fields, organisational or governments, which
were rare in their biology corpus but common in more general language
corpora. Thus, species, gene and protein would normally have a high
keyness value in a biology corpus.
• Multiword terms: word lists and keyword lists apply to individual words
(or terms for our purposes). However, concordancers can also deal with
phrases and multiword terms if they offer an n-​gram (or cluster) search
feature. An n-​gram (where ‘n’ stands for ‘any number’) is a sequence of a
given number of contiguous words. While a one-​word sequence is referred
to as a unigram, a contiguous sequence of two words is referred to as a
2-​gram (or bigram) and a three-​word sequence is a 3-​gram (or trigram), 4-​
gram (or quadgram), etc. Users can determine the minimum and maximum
n-​gram size to be searched. For example, you can instruct a concordancer
to find 3-​grams and 4-​grams in your corpus, which would yield every pos-
sible (and contiguous) 3-​word and 4-​word sequences it contains. From a
sentence such as The starting point is the polypeptide chain, this search
would yield the following results:
the starting point
starting point is
point is the
is the polypeptide
the polypeptide chain
Once all the 3-​grams have been presented, the results would also yield all
the 4-​grams possible in your corpus, e.g.:
the starting point is
starting point is the
point is the polypeptide
is the polypeptide chain
Depending on the size of your corpus, the list of results may be very long.
In that case, you can use stopwords, keyness measures, raw or normalised
frequencies and/​or alphabetical sorting to streamline the processing of the
results and, ideally, zoom in on a possible term such as polypeptide chain
while ignoring less meaningful n-​grams such as starting point is the. Once
you have identified polypeptide chain as a potential term with the n-​gram
feature, you can generally also analyse this term and the ‘company it keeps’
by running a concordance on it. Although this example shows that it may
be possible to identify terms with the n-​gram tool of a concordancer, con-
cordancing is in fact a different process from term extraction (see Box 6.4).
• Collocations: concordancers can not only deal with single words and
multiword expressions (n-​ grams), but they can also deal with non-​
contiguous sequences of words, i.e. collocations, e.g. the verb issue with
the noun bond in a sentence such as The financial corporation issued
 Tools for Terminology Management 135

many bonds that year. Collocate analysis “is an extremely powerful way
to identify patterns of language use” (Anthony, 2022: 113) and many
concordancers offer this functionality. The list of collocates of a target
term may include their frequencies and confidence scores, the latter indi-
cating the strength of collocation between the words. The higher the confi-
dence score (or rank), the stronger the association is between the words. As
we saw with term extraction (see Box 5.3), higher confidence scores may
be a useful feature to search through a large list of results, but there might
also be interesting results at the lower end of the scale. For translators, for
example, the collocate tool can be used to establish that, in a particular
context, lift is a more conventional collocate of crane than hoist in a sen-
tence such as Lift the load with the crawler crane, as one of Frankenberg-​
Garcia’s students discovered (Frankenberg-​Garcia, 2015: 372).
• Data visualisations: some concordancers also offer various methods to
visualise the data in your corpus, other than the KWIC visualisation,
which may help you to find yet more patterns of language use in your
corpus. For example, AntConc and WordSmith Tools can produce disper-
sion plots, which are visual representations, often by means of barcode
charts, of where in the corpus specific words and phrases are used. To
illustrate, Figure 6.5 shows a dispersion plot for the term light in a marine
biology corpus. The vertical lines indicates that light occurs everywhere in
this corpus but it is perhaps more frequent towards the beginning and the
end of the corpus than in the middle of it. The plot can be enlarged and
the individual bars can be clicked, which will take the user to a preview of
the term in context. In a corpus of multiple files, the dispersion plot can
help you identify the file(s) in which the term light, in this case, is particu-
larly frequent, which might warrant further investigation.

Figure 6.5 Example of dispersion plot in AntConc.

136 Terminology Management for Translators

Box 6.4 Concordancing vs Term Extraction

Although concordancing and term extraction are different processes,

both can be used by translators to identify terms, but each achieves this
in different ways:

Term Extraction Tools

• They select terms from text, whether single words or multiword
expressions and present them to the translator in a list.
• Confidence scores are assigned to each term candidate and can be
used to be selective when large lists are returned.
• It is up to the translator to review the term candidates and confirm
which words and phrases are actual terms.

Concordancing Tools
• Rather than ‘extracting’ information form a text, concordancing
tools make a list of all the words and phrases from a text, along with
frequency counts.
• They provide the contexts in which the words and phrases appear.
• By viewing words and phrases in context, translators may identify
single-​word and multiword terms.

The above discussion applies to concordancing monolingual corpora, either

a source language corpus or a target language corpus, depending on the
information required by the translator. However, there are also bilingual
concordancers which work on parallel corpora (see also Chapter 5). For a
bilingual concordancer to work, the source sentences in the parallel corpus
must be aligned to the target language ones (see Text Alignment Tools section
below). Once compiled and aligned, a parallel corpus can be queried with a
bilingual concordancer. At the time of writing, available bilingual concord-
ancing software includes AntPConc,25 freely available online alongside other
tools; also Tradooit,26 currently only available in French, Spanish and English,
and Sketch Engine27 (see more on Sketch Engine below). Because building a
parallel corpus can take time and effort, translators may resort to publicly
available sources of parallel texts such as the OPUS open parallel corpora28
(and various tools to process the corpora) or use tools such as Linguee29
or Wordreference30 for inspiration. The “main advantage of using a parallel
corpus is that it allows you to produce bilingual concordances” (Bowker &
Pearson, 2002: 93), showing all the sentences with the search word in both
the source and target languages side by side, so that you can observe every
occurrence of how a particular phrase or term has been translated before.
 Tools for Terminology Management 137

Figure 6.6 Example of bilingual concordance using AntPConc (from the AntPConc

website).

Figure 6.7 Example of bilingual concordance using Sketch Engine.

The display varies depending on the software used; for example, AntPConc
uses a vertical display (see Figure 6.6), whereas Sketch Engine uses a hori­
zontal one (see Figure 6.7).
In Sketch Engine, if your search word in your source language is pro-
tein, you can opt to include its equivalent in the target language or leave
138 Terminology Management for Translators

the target language field empty. In the first option, Sketch Engine retrieves
all the occurrences of protein which have been translated by the equivalent
you have typed, i.e. it excludes occurrences of protein which may have been
paraphrased or translated differently. In the second option, Sketch Engine
retrieves all the occurrences of protein that have been translated into your
target language, regardless of how they have been translated. Both options
may be useful for translation-​based terminology work, depending on the type
of information you are looking for.

It is worth noting that, if a parallel corpus can be found in a relevant

domain in the OPUS website (see Tiedemann, 2012, for a descrip­
tion), the corpus can be downloaded as a .TMX file, which means
it can be easily imported into a translation memory for use during
translation. For more suggestions of tools and corpora, see the online
portal.

Sketch Engine is a “web-​based multifunctional corpus query and manage-

ment system” (Moze & Krek, 2019: 111), launched in 2004 (Kilgarriff et al.,
2014: 7). It has been widely used by linguists but especially by lexicographers,
with Sketch Engine at the centre of “important large-​scale lexicographic
projects at all major dictionary houses” (Moze & Krek, 2019: 11). By con­
trast, it seems “majorly under-​utilised in commercial translation environ-
ments” (2019: 11) although, in recent years, Sketch Engine and other corpus
tools are featuring more prominently in translator training programmes, at
least in Europe, according to a survey by Mikhailov (2022: 243). As we have
already seen, as well as bilingual (and monolingual concordancing), allowing
the upload of local corpora (Bernardini, 2022: 487) and providing access to
many pre-​existing large corpora (Olohan, 2016: 28), with sophisticated dis­
play search and display options, users of this program can “combine standard
statistical analyses with fairly easily programmed (or adjusted), customized
and reusable queries” (Marshman, 2022: 308).
Once you have created or uploaded your corpus, Sketch Engine provides
a wide range of tools to query it and obtain information from it, from term
extraction and keyword extraction tools to KWIC displays and other visu-
alisation tools. Examples and (updated) suggestions on how to use Sketch
Engine for the terminological work of translators can be found in the online
portal. An additional use of Sketch Engine for translators is to “compensate
for limited knowledge of the subject field and its terminology” (Buendía-​
Castro & López-Rodríguez, 2013: 66–​67), e.g. in cases where translators
need to acquire new specialised vocabulary in the domain they are working
on. A distinctive feature of Sketch Engine, which gives it its name, is the
Word Sketch (as well as the Word Sketch Difference).
 Tools for Terminology Management 139

A Word Sketch is usually described as “an automatic corpus-​derived sum-

mary of a word’s grammatical and collocational behaviour” (Kilgarriff et al.,
2010: 372) in a given corpus. This means that a Word Sketch groups words
according to their grammatical relations with other words, which can help
identify collocations, among other lexical items. Since corpora in Sketch
Engine are tagged for part-​of-​speech (see also Box 5.3 in Chapter 5) and
grammatical category, Word Sketches can identify which nouns or noun
phrases function as objects or subjects of a particular verb, for example,
which nouns collocate with other nouns, etc. In short, if two words are habit-
ually found in the vicinity of each other, a Word Sketch can find them.
For terminological purposes, Word Sketches can help translators identify
not only collocations and other phraseological information, but also para-
digmatic and syntagmatic relations of terms, multiword terms, modifiers of
terms (see also Chapter 5), etc. In some cases, Word Sketches can also be
used to pinpoint seed words on which to base a corpus building task. For
suggested exercises illustrating these and other uses of Word Sketches for
translators, and other useful tools for translators, including the Word Sketch
difference we look at below, please see the online portal. Examples of Word
Sketches are shown in Figure 6.8.
In Figure 6.8, a corpus from the domain of bioluminescence in marine
biology was used. The word light was queried twice in this corpus, once as a

Figure 6.8 Screenshots from Sketch Engine showing Word Sketches for light as a
noun (top) and verb (bottom) from a bioluminescence corpus.
140 Terminology Management for Translators

noun (top three Word Sketches) and as a verb (bottom three Word Sketches).
A quick search such as this (Word Sketches typically take seconds to be
produced) can already inform a translator that:

• In English, light as a noun is much more common in this domain than as

a verb.
• The top three Word Sketches also show that the most salient adjectives
that collocate with the noun light (in this corpus at least) are visible, blue
and bioluminescent.
• In the case of visible, the figures in the same row indicate how frequent vis-
ible light is as a phrase (11 occurrences) and the confidence score assigned
by the program (12.2).
• Similarly, translators can glean that the most frequent verbs with light as
an object are produce, absorb and emit.

This information already shows translators a fragment of the conceptual net-

work (see also Chapter 1) of the domain of bioluminescence, in this case,
which can be used for their understanding of the domain in general and the
source text in particular. Depending on the needs of a translation project, the
Word Sketches can be browsed more extensively and/​or they can be produced
in the other language(s) of the project.
A ‘partner’ tool to Word Sketches is Word Sketch Difference, which is
used to compare the collocational behaviour of two semantically related
words. In the case of translation-​based terminology work, this tool can be
used to analyse subtle differences between translation equivalents or partial
synonyms, for example. As explained by Moze & Krek (2019: 123), the
“greatest advantage of using sketch difference has to do with data visualisa-
tion, as the tool combines the two words’ collocates into the same column
and uses colour coding to show which word they typically co-​occur with.”
An example of some Word Sketch Differences for creature and organism in a
bioluminescence corpus are illustrated in Figure 6.9.
In this figure we can see three Word Sketch Differences, i.e. three different
columns, showing how the terms creature and organism behave in this spe-
cific bioluminescence corpus. It is important to specify that the figure shows
the behaviour of these terms in a particular corpus because, in other cor-
pora, or if you add more texts to your corpora, the figures produced by
the Word Sketch Difference will also change, potentially presenting different
behaviours for each term. By clicking on the three horizontal dots at the end
of each row, you can use other tools in Sketch Engine to explore the diffe-
rence in behaviour further.
In each column the collocates of each term are presented in colour-​coded
backgrounds, typically the words related to the term on the left (in this case
creature) appear in green, and those for the term on the right (in this case
organism) typically appear with a red background (see colour examples in
the online portal). In addition, the red and green backgrounds appear with
 Tools for Terminology Management 141

Figure 6.9 
Example of Sketch Engine’s Word Sketch Difference for creature and
organism in a bioluminescence corpus.

different intensities, i.e. the brighter the green or red, the stronger the associ-
ation is between the words. This provides a visual overview of where the
strongest, and potentially most interesting, collocates can be found. In the
centre of each column, often presented in a neutral-​coloured background, are
the shared collocates of both search terms, if there are any. In our case, the
Word Sketch Difference on the left in Figure 6.9 shows that other and bio­
luminescent both collocate with creature and organism.
The strength of collocation between the displayed words and terms can
also be seen in the scores automatically provided by Sketch Engine. The
higher the number, the higher the strength of collocation between those
words. For example, looking at the Word Sketch Difference in the middle of
Figure 6.9, Sketch Engine’s algorithms applied to this corpus would suggest
that the relationship between creature and the verb mask is stronger (i.e.
13.4) than that between creature and the verb have (i.e. 9.5). The importance
of these scores for translators is mostly to show where trends might lie rather
than to interpret the scores in absolute terms.
As well as showing the collocates, organised in various categories of rela-
tion (i.e. one column per category), and the scores, Sketch Engine provides a
number of counts which, together with the scores, can be viewed or hidden
depending on the user’s preferences. For example, at the top of the left-​most
column, it can be seen that jelly-​like appears once with creature but it does
not appear in this corpus with organism. Similarly, brainless, mysterious, gel-
atinous and sea all collocate with creature, but not with organism; looking
at the bottom of this column, we see that planktonic, filter-​feeding and
142 Terminology Management for Translators

barrel-​shaped collocate with organism but not with creature. This would
suggest that, although they might appear synonymous at first glance, crea-
ture and organism are not true synonyms and each is used in a different
way in bioluminescence domains. The texts in this corpus would indicate
to translators that the trend among bioluminescent experts is towards using
the terms sea creature and marine organism, but not sea organism or marine
creature, for example, because, according to Figure 6.9, sea does not col­
locate with organism, or marine with creature (although, of course, much
larger data might be needed in order to reach definitive conclusions). For
other settings and tools in Sketch Engine, see the online portal.

Term Extraction Tools

Once you have mined some texts and created a corpus, term extraction can
be carried out manually (e.g. by highlighting the terms on a paper document)
or semi-​automatically (Costa & Zaretskaya, 2018), i.e. by getting the soft­
ware to extract the candidate terms from a text which are then manually
validated by the translator. Although there are methods to validate terms
automatically (e.g. Melo Mora & Toussaint, 2015), this task may be carried
out manually when the quality of the extracted terms is important but this
is, inevitably, a time-​consuming procedure. Although this section focuses on
automatic term extraction tools, there is also a place for manual extraction
in the translation workflow, e.g. in the case of very short source texts which
are only available on paper or if highlighting the terms on a piece of paper is
“faster than taking the time to scan and convert” the document (TerminOrgs,
2022: 3).
Having already covered the concepts and terminology used in term extrac-
tion (see Chapter 5), in this section, we take a conceptual look at the various
kinds of automatic term extraction tools available. Depending on the number
of languages they can deal with, term extraction tools can be monolingual,
bilingual or multilingual. Regarding the platform where they are offered,
they can be web-​based or standalone (i.e. desktop-​based). Concordancers
can often be used as a proxy for term extraction tools and translators can
even use their own CAT tools to extract terms from the source and target
texts as they translate if the CAT tool does not already offer monolingual
term extraction. For lists of tools and their URLs, including some user guides,
please see the online portal.

Monolingual and Bilingual Term Extraction Tools

While monolingual term extraction tools process and extract terms from texts
in one language, bilingual or multilingual term extraction tools extract and,
crucially, align terms from texts in two or more languages (Bakaric et al.,
2021: 19) (see also Box 6.5). This means that you can either run a monolin­­
gual term extraction tool separately on texts from different languages, or you
 Tools for Terminology Management 143

can use a bilingual tool that will align the texts as well as extract information
from them. In each case, the end result might be the same, i.e. a list of terms
in each language. However, the reasons for using a monolingual or bilingual
tool may depend on a number of considerations.
Monolingual term extraction is useful in the case of some low-​resourced
languages, which have an online presence. Even for those which do not,
monolingual term extraction may be a way to increase their online presence
(e.g. Bustamante et al., 2020). Monolingual term extraction is also used
“for collecting the terms you need to translate and also for checking to see
whether they are already in your database before you start translating texts”
(TerminOrgs, 2022: 2). As pointed out by TerminOrgs, if “you only write or
communicate in one language, then monolingual is all you need” (2022: 2).
Bilingual term extraction is “much harder to perform than only monolingual
as it requires a good word alignment system” (Costa & Zaretskaya, 2018).
This is because in bilingual term extraction, the tool extracts terms from the
source language (i.e. monolingually) first and then tries to match the candi-
date terms to their equivalents in the target language.

Box 6.5 When Should Term Extraction Be Carried Out?

Here are some suggestions as to when it is a good time to carry out term
extraction:

• Populating a termbase from scratch.

• Preparing a termbase for a translation project or for a glossary for
internal communication in a company.
• Determining preferred terms or commonly used terms in a given lan-
guage or set of texts. This can allow the adoption or non-​adoption
of specific synonyms or partial synonyms, for example, in order to
enable both consistent monolingual source text production and con-
sistent translation.
• Improving your termbase quality, e.g. checking the coverage and
completeness of your termbase.
• As you translate, e.g. you can add new terms to your termbase on the
fly if you spot them during translation.
• When developing or expanding on a specialism, e.g. enlarging an
existing corpus with new texts in a domain you specialise in.
• Etc.

Source: Adapted from TerminOrgs, 2022: 4–​5.

Bilingual term extraction tools can help you extract terms along with
their equivalents and are likely tools that are used in multilingual translation
144 Terminology Management for Translators

environments (TerminOrgs, 2022: 2). With older term extraction tools, you
will need a corpus that is already aligned. However, with newer tools, such
as the OneClick Terms31 feature of Sketch Engine, you can carry out term
extraction from a bilingual set of texts even if they are not aligned, as this
tool aligns the texts first (see also Text Alignment Tools section below) before
attempting to extract terminology from them.
The more sophisticated term extraction tools may do more than term
extraction –​ they can also help translators create new concepts (as opposed
to terms, i.e. labels for the concept, see Chapter 2) or to group (i.e. cluster) a
number of partial synonyms in the same termbase entry. One such tool avail-
able at present is Fodina’s TermCatch (see Figure 6.10). TermCatch can be
instructed to extract terms from uploaded files, from translation memories
and/​or from selected websites. The list of candidate terms produced comes
with confidence scores and number of occurrences for each term. Its term
clustering feature can be used on this list and, when activated, it clusters
(i.e. “groups”) terms which are synonymous or partially synonymous, or
which are variants of each other, e.g. geographical variants, spelling variants,
inflectional variants, having different capitalisation, etc. The clusters of
terms produced in this way can be used as a basis to create a new concept.
For example, the following screenshot (from Fleischmann & Lundahl,
2023: 50:21) shows that the terms Locking nut, locking nut, jamnut, lock
nut and jam nut (i.e. different labels for a single concept) have been identified
by TermCatch as a potential cluster of terms, from which users can create the
new concept in the termbase.

Figure 6.10 Example of the term clustering and concept creating features in Fodina’s
TermCatch.
 Tools for Terminology Management 145

Term clustering is not a monolithic tool –​users can instruct TermCatch to

cluster terms at different levels, with level 1 grouping terms which are very
close to each other or 100% synonymous, to level 5 in which terms are not
synonymous, but they are related in some way. The lower the cluster level
selected, the fewer but more focused synonyms returned. Conversely, the
higher the cluster level selected, the more synonyms provided but they are
less closely related. Even if terms have been clustered at one level, they can be
re-​clustered at a different level.
By clicking on the Create a new Concept button, translators can edit
the list of clustered terms and, if desired, they can request TermCatch to
produce AI-​generated definitions. If the results are not what the translator
expected, they can be re-​generated until the generated definition meets the
desired standards. The text of the definition can also be edited manually.
Another feature in TermCatch, called Lookup Views, allows users to see if
the term users type (or a synonym of this term) appears in either a termbase
or in material harvested by TermCatch, which can help users find new
terms, duplicates, etc. The general workflow with TermCatch is illustrated
in Figure 6.11.
In this figure, from left to right, the workflow begins by collecting all the
documents (including files, websites, etc.) containing all the terminology
to be consolidated. Then, the term extraction tool in TermCatch identifies
terms. In the next step, the extracted terms may be grouped into term clusters
by the term clustering tool. Once terms are clustered, the concept editing tool
comes in, allowing users to generate definitions and set usage status, among
other attributes. Once the concepts are standardised in this way, they can be
exported to Excel or other termbase software such as Quickterm (see below
and Chapter 9). For more on the terminological features of TermCatch, see
the online portal for this book.

Web-​based, Server-​based and Standalone Term Extraction Tools

Standalone term extraction tools are those that “can be installed on the com-
puter and operate independently of any other device or system” (Costa &
Zaretskaya, 2018). Web-​based term extraction tools, by contrast, are cloud-​
based and therefore “do not require any prior installation as they can be
accessed within a web browser” (2018). In settings where a server is avail-
able, resources are typically held on the server and accessed from local
workstations. Other tools may be hybrid, e.g. they have a fully-​featured local
editor with cloud-​based access to resources. The trends over the last decade
or so suggest that most tools will have or already have a cloud-​based compo-
nent, which raises issues of data ownership, data protection and data confi-
dentiality (e.g. Lakshmi & Tyagi, 2023).
newgenrtpdf
146 Terminology Management for Translators
Figure 6.11 Typical workflow with TermCatch.
 Tools for Terminology Management 147

Concordancers as Term Extraction Tools

As we saw in the Concordancers section, concordancers are widely used

tools in fields such as linguistics or corpus research. Although they were
not designed as term extraction tools, they have been used, and continue
to be used as such, by linguists (e.g. Muegge, 2013). AntConc, one of the
best known concordancers, can be used for term extraction, whereas Sketch
Engine has a concordance feature that can be used in a similar way as a term
extraction tool but it also has its own term extraction tool. For more on the
difference between concordancing and term extraction, see Box 6.4.

Term Extraction in CAT Tools

Some CAT tools offer monolingual term extraction as a regular part of the
translation process (see Mitchell-​Schuitevoerder, 2020: 73) but, even if they
do not, most CAT tools allow translators to highlight new terms in both
the source and target texts and add them on the fly to the termbase, a fea-
ture that essentially works as a term extraction feature (2020: 72) (see also
Chapter 5, “Entering Terms in the Termbase”). In this case, though, as we
saw in Chapter 5, it is good practice to have a system in place, integrated into
a translator’s workflow, whereby you return to complete any term records
created in this way.

Concept Maps and Other Concept Visualisation Tools

Understanding the concepts that make up a specialised domain and how
they relate to one another is a crucial part of the terminological work of the
translator (Granda & Warburton, 2001: 4). This is one of the reasons why
concept orientation is a key principle in the design of termbases (see also
Chapter 4). However, concepts are not formed or used in isolation and, for
translators, one way to acquire domain knowledge and competence, beyond
the equivalents of individual terms or phrases, is to create a corpus (through
text mining) and analysing its contents (through corpus analysis and term
extraction tools), as we have seen so far in this chapter. Concept mapping
tools offer translators another method of supporting their terminological and
domain competence. Concept maps are “visual representations of networks
of concepts belonging to the same domain” (Falco, 2017: 98), strongly
advocated by cognitive theories of terminology (see Chapter 3).
Concepts constitute the “smallest units used to communicate specialized
knowledge and information” (Galinski & Picht, 1997: 42), but they are not
only verbal; images convey conceptual knowledge too and so can “non-​verbal
signs, alphanumeric characters or character strings, and a variety of hybrid
forms” (1997: 42). Solving conceptual problems is substantially more difficult
than solving purely linguistic ones and “a suitable technology has not
been available” to deal with what Meyer calls concept management which
148 Terminology Management for Translators

Figure 6.12 Example of concept map for the term “distribution agreement.”

Source: From Falco (2017: 101).

consists of “acquiring, analyzing, comparing, formalizing, revising –​in other

words, managing –​a variety of information about the concepts designated
by the terms” (Meyer, 2022: 112). Concept maps can be produced electronic­
ally “through mind mapping software or by using organizational charts”
(Austermühl, 2012: 65). An example of a concept map for the term distribu­
tion agreement is shown in Figure 6.12 (from Falco, 2017: 101).
Concept mapping tools offer “flexibility in representing concepts and
their potentially multidimensional relations” (Austermühl, 2012: 66). As
Austermühl explains, “the links between concepts can, for example, be
one-​way, two-​way, or non-​directional; [and] the links describing the nature
of the relations can be labeled” (2012: 66). The “underlying idea behind
mapping is, generally speaking, to make use of the information visualiza-
tion potential of maps […] as a general learning technique” (Ditlevsen &
Kastberg, 2009: 3). Thus, with concept maps, translators can acquire domain
knowledge in more visual, non-​linear and non-​sequential representations of
concepts and the relationships between them (Falco, 2017: 96). Galinski and
Picht (1997) review various types of non-​verbal knowledge and methods to
visualise it.
There are also graphical conventions that can be followed to distin-
guish the types of term relations depicted in the concept map, although
such conventions are not standardised. For example, Rousseau et al.
(2018: 5) suggest using tree forms (or right-​angled lines) to visualise hier­
archical relations such as hyperonym/​hyponym; fan forms (or angled lines) to
visualise partitive relations such as meronym/​holonym; and (curved) labelled
arrows to visualise associative relations, as illustrated in Figure 6.13.
In order to create a concept map, you will need a focus question, a so-​called
parking lot of concepts and cross links. A focus question is what “guides the
 Tools for Terminology Management 149

Figure 6.13 Visualising different kinds of term relations in a concept map.

identification of relevant concepts” (Austermühl, 2012: 66). For example, the

focus question for the concept map in Figure 6.12 is What is a distribution
agreement? A parking lot is the list of identified concepts. Concepts can be
extracted through concordancing and/​or term extraction tools, for example.
Cross links are the links showing the relationships between the concepts, e.g.
hierarchical (hyponym/​hyperonym), partitive (e.g. meronym/​holonym) and
associative (e.g. process/​product relation) (see Relations between Terms in
Chapter 4).
Once you have a focus question, a parking lot of concepts and the relations
you want to visualise through cross-​links, the “low tech solution needs only
some sheets of paper” (Austermühl, 2012: 66) and a pen or marker. If you
prefer to create concept maps electronically, you may find free programs
online, but one tool that has been suggested in the literature for translators
is IHMC Cmap Tools (Austermühl, 2012: 66–​67; Falco, 2017: 98), which
is currently free for educational institutions. Cmap Tools is a massive online
collaboration tool which “enables users to construct concept maps on their
own computers and even share them via CmapServers on the Internet with
other users, thus enhancing knowledge-​ sharing and encouraging mean-
ingful learning among members of the same discourse community” (Falco,
2017: 98). Another freely available tool is the Concept Map Search inter­
face in WIPO Pearl,32 which “offers the option of displaying concept systems
(called Concept Maps) in different languages and conducting searches on
concepts” (Caffrey & Valentini, 2019: 139). An example of free online soft­
ware for drawing charts and diagrams is Flowchart Maker33 which, according
to Kudashev and Mikhailov (2021), “provides all the necessary features for
drawing classical concept charts,” requires no installation, little user training
150 Terminology Management for Translators

and it is well suited for collaborative work. For other examples of concept
maps, concept mapping tools and how to use them for translation, see the
online portal.
Other visualisation methods have been proposed from within the termin-
ology literature, such as the Satellite Model, Termontography and EcoLexicon,
a resource developed by scholars following Frame-​ Based Terminology
principles (see Chapter 3), to name but a few. Concept modelling has been
“a long-​standing speciality of some Scandinavian terminologists” (Humbley,
2022: 33) and Nuopponen “is perhaps the best known” (2022: 33). Her
Satellite Model was developed in the 1980s (as reported by Nuopponen,
2016: 190) and first described in Nuopponen (1994). Taking Wüster’s
General Theory of Terminology (see Chapter 3) as a point of departure
(Picht, 2011: 17–​18), the Satellite Model “consists of a core concept and
concepts in satellite nodes which in turn may have their own satellite nodes”
(Nuopponen, 2016: 190), much like a mind map. A satellite map for the
domain of scuba diving is shown in Figure 6.14 (from Nuopponen, 2016: 95).
Already introduced in Chapter 3, Termontography is a method of
implementing conceptual representations suggested by the Sociocognitive
theory of terminology (e.g. Temmerman, 2000). Temmerman and
Kerremans (2003: 3) define ontology as “a knowledge repository in which
categories (terms) are defined as well as relationships between these cat-
egories” and describe termontography as a “multidisciplinary approach in
which theories and methods for multilingual terminological analysis […] are
combined with methods and guidelines for ontological analysis” (2003: 4).
Termontography “seems to owe a great deal to the work done by Ingrid
Meyer (e.g. Meyer et al., 1992) who was one of the first terminologists to
perceive that term bases would be even more useful if their organization
bore some resemblance to the way concepts are represented in the mind”
(Faber, 2009: 118–​119).
According to its own website, EcoLexicon is a “terminological resource
developed by the LexiCon Research Group at the University of Granada.”34
EcoLexicon is designed to represent “the conceptual structure of the
specialized domain of the Environment in the form of a visual thesaurus”
(Faber, 2009: 118–119). The design principles of EcoLexicon are rooted in
Frame-​Based Terminology (e.g. Faber, 2012, 2015, 2022) (see also Chapter 3).
It is currently freely available online in six languages. It offers many settings
and tools, such as concordances, to visualise concepts from environmental
domains in various ways. It has two different modes to represent concepts,
i.e. the tree mode and the path mode, as shown in Figure 6.15.
Several termbase software tools allow translators to upload graphics
and, sometimes, multimedia items such as videos. However, few termbase
software tools at present also provide concept mapping capabilities. Two
current exceptions are Kalcium QuickTerm by the Kaleidoscope Group from
the publicly available demo server,35 and TermCatch. Figure 6.16 shows a
concept map created with QuickTerm for the French term FIM (Fédération
 Tools for Terminology Management 151

Figure 6.14 An example of a satellite map for scuba diving.

Source: From Nupponen (2016: 195).

Internationale de Motocyclisme). In TermCatch, the cluster graph feature

provides a visualisation of partial synonyms in particular (as opposed to
different concepts in a domain). For example, Figure 6.17 shows a concept
map created with TermCatch for partial synonyms such as lift shaft, propeller
shaft, drive shaft, etc. The synonyms and their relationships can be edited dir-
ectly from the graph view. Although it may not be clear in a black-​and-​white
graphic, the cluster graph shows various shades of blue between the nodes to
denote various strengths of association between the synonyms. To my know-
ledge, there are no CAT tools available at present that can offer concept
mapping functionality, but this may well change in future. For resource-​poor
languages, there are freely available tools such as BabelNet,36 a “multilingual
encyclopedic dictionary and semantic network […] created by linking
Wikipedia to WordNet […] which are integrated through automatic mapping
 newgenrtpdf
152 Terminology Management for Translators
Figure 6.15 Two types of concept visualisation in EcoLexicon: the Tree mode (left) and the Path mode (right).
Source: Adapted from the EcoLexicon website.
 newgenrtpdf
Tools for Terminology Management 153
Figure 6.16 Example of concept map with Kalcium QuickTerm by the Kaleidoscope Group.
154 Terminology Management for Translators

Figure 6.17 Example of cluster graph feature in TermCatch to visualise partial

synonyms.
Source: From Fleischmann & Lundahl (2023).

and by filling in lexical gaps in resource-​poor languages with MT” (León-​

Araúz et al., 2017: 326).
Although “it has been demonstrated the graphic representations are less
prone to inaccuracies than verbal ones” according to Galinski and Picht
(1997: 48), there are some drawbacks to the use of concept maps. Novak and
Cañas (2008: 12) point out that a concept map is never finished and that it is
always necessary to revise it. For Falco (2017: 104), the main drawbacks are
that “maps are non-​exhaustive and their construction is time-​consuming.”
Therefore, “[n]‌ew maps must be created continuously to answer new focus
questions and enhance encyclopaedic knowledge” (2017: 104) and, “as a
result, their impact is not immediately effective” (2017: 104). Austermühl
(2012: 70) highlights that concept maps remain “singular identities isolated
from their socio-​cultural and, above all, linguistic contexts” and suggests
complementing concept maps with a corpus-​based approach to acquiring
domain knowledge. Although “more research is necessary in order to gain
further insights into the parameters that link visual contexts with linguistic
contexts, […] it is evident that different types of image have an important
role in specialized knowledge representation” (Faber, 2022: 373).
 Tools for Terminology Management 155

Artificial Intelligence and Large Language Models

Artificial intelligence (AI) is currently a buzzword in the language industry
which is eclipsing “all other trends in the minds of language industry
stakeholders,” according to the European Language Industry Survey 2024
(ELIS Research, 2024: 23), which obtained 1,776 responses from around the
world. This was confirmed in the results of the ELIS report of 2025 (ELIS
Research, 2025: 7). For some, communicating with machines is a new para-
digm (Massion, 2024: 80) and others are even describing the rise of AI as a
fourth industrial revolution (Kirov & Malamin, 2022: 1; Schwab, 2016: 16).
Although the use of AI in the language industry is still in its infancy (ELIS
Research, 2024: 4), the concept of artificial intelligence is not new. It has been
around since the 1950s when the phrase “artificial intelligence” was coined by
American mathematician and computer scientist John McCarthy (e.g. Kelly,
1993: 11). Generative Artificial Intelligence (GenAI), a subset of AI, became
the centre of attention in 2022 when “an AI-​powered chatbot developed
by OpenAI and based on the GPT (Generative Pretrained Transformer) lan-
guage model” (Browne, 2023), i.e. ChatGPT, became available to the public.
ChatGPT is powered by a Large Language Model (LLM), “meaning that
it is programmed to understand human language and generate responses
based on large corpora of data” (2023). When using GenAI, “prompting,
i.e. formulating work instructions for the LLM, is of particular importance”
(One Word, 2024). LLMs are one of AI’s most significant developments and
are often used for conversational applications such as chatbots and virtual
assistants (Tavares et al., 2023: 3). As well as ChatGPT, other examples
of LLM-​powered chatbots, include Gemini (formerly Bard, developed by
Google), Azure OpenAI and expert shopping assistant Rufus (trained by
Amazon) (see Mehta, 2024), to name but a few.
According to López Jiménez and Ouarachi (2020: 251), “the perception
of AI has continually shifted between miracle and monster,” due to its wide
application in many areas, but also because some view it “as a threat, cap-
able of eliminating tasks that are currently done by humans” (2020: 251), as
expressed, for example, by a survey by the Society of Authors (SOA Policy
Team, 2024). Since LLM-​powered chatbots such as ChatGPT can query vast
amounts of corpora, their role in the terminological work of the translator
is relevant.
The connection between terminology and AI is also not new. According
to Condamines and Picton (2022: 222), the close relationship between these
disciplines, “united by an interest in representing knowledge in the form of
networks of terms/​concepts” was already noticed in the 1980s. For example,
Felber (1984: 109) predicted that “since the studies in artificial intelli­
gence and the establishment of expert systems have to deal with concepts,
systems of concepts, concept linking, conceptual mapping of reality, etc.,
a stronger affinity between the GTT [General Theory of Terminology, see
also Chapter 3] and computer science is to be expected in the near future.”
156 Terminology Management for Translators

It is perhaps too early to tell what exactly generative AI and LLMs have in
store for us and to what extent these tools will be integrated into the profes-
sional translator’s (and interpreter’s) workflows. However, current trends are
already suggesting that the terminological work of the translator can benefit
from AI-​based tools (Tavares et al., 2023: 12), e.g. in the following ways:

• For writing tasks: since LLMs can generate new text (albeit based on
massive corpora of existing data), the potential of LLMs for writing
definitions is already being investigated. For example, the terminological
team at One Word (2024) compared human term extraction to term
extraction by machine, whereas San Martín (2024a, 2024b) highlights
some prompts used in ChatGPT 4 to craft definitions. Łukasik (2023)
also finds that ChatGPT can be used for definition writing and the results
of his experiments in several languages suggest that ChatGPT works best
for English. According to Massion (2024: 91), LLMs can also “adapt
definitions based on reliable sources.”
• For term extraction tasks: according to Massion (2024: 91), LLMs can
carry out selective term extraction tasks. Muegge (2023) has created a
guide on how to use ChatGPT for term extraction, aimed at translators
and interpreters.
• For domain and conceptual understanding: for example, Łukasik (2023)
shows some examples of how ChatGPT can be used to identify domains and
reconstruct a conceptual field, among other tasks. However, since LLMs are
trained on finite data after all, their usefulness for domain research remains
limited to the data they were trained on. Therefore, it would be advisable
for translators to use other methods for domain research, e.g. making a
sophisticated use of search engines, to complement this limitation.
• For cross-​lingual terminology matching: LLMs can “help identify equiva-
lent terms in different languages, facilitating multilingual terminology
management” (Massion, 2024).
• For cross-​ cultural terminology analysis: LLMs can “analyse how ter-
minology is used and understood in diverse cultural contexts” (Massion,
2024).
• For terminology validation: LLMs can “compare terminological data with
authoritative sources to ensure accuracy and consistency” (Massion, 2024).
• For the identification of conceptual relations: According to Massion
(2024), LLMs can “help identify and establish relations between concepts
or terms.”

Well established CAT tools are already offering AI, e.g. RWS Trados (Trados,
n.d. 2), memoQ (memoQ, 2023), Phrase (Phrase, n.d.), Matecat (Trombetti,
2023), etc. and ‘newer’ ones already have ChatGPT built in, e.g. Bureau
Works37 and Wordscope.38 Tavares et al. (2023) and ELIS Research (2024) dis­­­­
cuss other tasks that translators and interpreters may use LLMs for, beyond
 Tools for Terminology Management 157

terminological tasks. From the (non-​exhaustive) list above, it follows that, in

order to make an efficient use of LLMs for terminological work, translators
will need to become proficient at prompt writing, i.e. at finding the best
ways to query LLMs in order to achieve the desired result. New research
is already emerging in this respect. For example, He (2024: 8) compared
the outputs generated by ChatGPT to four different kinds of prompts, some
including ‘translator brief’ and ‘translator persona’ information, and found
that “assigning the persona as a translator allowed ChatGPT to achieve the
best performance” in her experiments. Łukasik (2023) discusses prompt
engineering in relation to “improvements of the instruction set” with a view
to increasing the quality of the answers obtained and reports that ChatGPT
4 “seems to have problems with longer texts and with extensive instructions.”
Prompt engineering is also used by Massion (2024: 86) who states that “the
profession of prompt engineer has emerged and is now in high demand” and
this is backed up by respondents to Stasimioti’s survey (2023), who expect
that new professional roles will also include Post-​editor of LLM MT Output,
Prompter Debugger and AI Editor, AI Integration Specialist and even Expert
in Linguist-​AI Interaction, among others.
As with many other tools covered in this chapter, the use of LLMs is not
without its limitations. Any translator using LLMs for their work should be
aware of at least the following:

• Lack of reproducibility: the One Word team (2024) found that “even iden­
tical prompts with identical input (same source text, same form of data
input) never delivered the same result twice,” which makes it difficult to
confirm the validity of the results obtained. This also makes LLMs some-
what unreliable and unpredictable, but new developments in the tech-
nology may overcome this in time.
• Hallucinations: this means that LLMs can introduce non-​existent facts
in their output and potentially mislead users who may not be aware that
they are fabricated facts (see Yao et al., 2024). For example, in their term
extraction experiment, the One Word team (2024) found that, in “a total
of five runs, ChatGPT delivered terms that appeared technically correct
but did not appear in the text at all.”
• Potential for errors: as well as introducing hallucinations, LLMs are
sometimes incorrect, which means that results, “like everything that
comes directly from a machine, must be checked with a critical eye” (One
Word, 2024).
• Lack of contextual understanding, precision and quality: as Tavares et al.
(2023: 11) found in their survey of language service providers.
• Lack of confidentiality guarantee: Tavares et al.’s respondents (Tavares et
al., 2023:11) also stated that they did not find AI useful because it was not
possible to guarantee the confidentiality of their data.
• Lack of familiarity with the use of LLMs: Tavares et al. found that their
respondents were not very familiar with the use of AI in the profession
158 Terminology Management for Translators

and, as a result, “perceived the level of usefulness of AI as essentially low

to moderate” (Tavares et al., 2023:11).
• A “more pressing concern than futuristic fears about an AI takeover”
(McCallum et al., 2024), is “AI’s potential to amplify bias or discrimin­
ation” (2024), such as racism or sexism, as these programs can “repro-
duce the bias contained in their source material” (2024).
• Some fear the “risk of indiscriminate use of the technology by those who
have not been trained to use it correctly [will] likely lead to quality issues”
(ELIS Research, 2024).

To a large extent, we are seeing the same reactions by translators to AI

and LLMs as we saw with the emergence and widespread use of MT (ELIS
Research, 2024: 40). However, there are also those who view the technology
positively and have pointed out some of its benefits, for example:

• According to some of the views collected by the ELIS Survey, the flaws in
AI may “re-​ignite the appreciation for high quality human translation”
(ELIS Research, 2024: 4).
• LLMs can be used as a tool to increase efficiency (ELIS Research, 2024: 40)
or even for inspiration.

In addition, Tavares et al.’s (2023: 10) respondents thought LLMs are useful
because they can:

• Save research time and effort.

• Provide more translation options.
• Help understand specific terminology.
• Help understand cultural references.
• Assist in verifying the clarity of the translated texts in long, complex
sentences.
• Assist in reducing production costs for the client.
• Etc.

As with other tools we have covered in this chapter, the key to making the
most of AI and LLMs is to make a discerning and informed use of them,
given the task at hand. As Tavares et al. (2023: 20) explain, “AI has the
potential to transform many aspects of life and create new opportunities
and efficiencies.” However, “it is important to approach AI development
and deployment responsibly and ethically to ensure that its benefits are
maximized, and its potential risks minimized” (2023: 20). It is hoped that
this chapter and the accompanying resources in the online portal will con-
tribute to dispelling translators’ misgivings about the use of the tools covered
in this chapter: they are not as onerous to use as they might seem at first
glance. Ultimately, increasing the take-​up in the use of the software covered
 Tools for Terminology Management 159

here is an investment for translators that can lead to both more productive
and high-​quality translation work.

All things considered, an LLM is only as good as the data it is trained

on, however large. If the data it was trained on contains biases, for
example, those biases will come through in the answers it gives to
prompts. A challenge for the future remains to remove all bias and dis-
criminatory material from data to be used in LLMs.

Follow-​up Tasks and Reflection

1. Search engines:
a. Find advanced settings of Bing and Yahoo and compare them to those
in Google. Are there any settings that are unique to a particular search
engine?
b. Select a search engine and carry out the following advanced search:
you want to retrieve results including (in your own language) “AI” and
“artificial intelligence” but without including “LLM” or “large lan-
guage models” and you also want to restrict the search to results from
your own country since 2020 only.
2. Carry out some general research on the internet to identify five seed words
for the domain of wind turbine technology.
3. Are translators typically paid by number of types or number of tokens in
a text?
4. Why is alignment typically associated with the translation memory of a
CAT tool, rather than with the termbase? In which ways can you ‘extract’
terms from a translation memory?
5. What does a good result from a text alignment task depend on? In other
words, why should you always check the results of an alignment task
carried out by a machine?
6. How does a concordance line differ from the segments in a translation
memory?
7. Apart from term extraction tools themselves, which other categories of
tools have we mentioned in this chapter which could be used as term
extraction tools too?
8. Regarding the following basic concept map (in Figure 6.18 below), which
might be its focus question?
9. Many professional translators are proficient in the use of regular
expressions, and it is a good investment for the terminological work of
translators, especially for ad hoc term searches, domain research and gen-
erally being able to retrieve the desired information efficiently from large
resources. Read through the resources suggested in Box 6.2 (or any other
160 Terminology Management for Translators

Figure 6.18 A basic concept map.

useful ones you may come across) and find out for yourself what the buzz
about regular expressions is all about.
10. What are your current views on the use of AI and LLMs in the transla-
tion (and/​or interpreting) profession? What positive and negative aspects
do you see? What would encourage or discourage you from using such
tools for translation and terminology?

Notes
1 However, translators should bear in mind that Wikipedia is crowdsourced, which
raises issues of confidentiality for some scholars (e.g. Kapsali, 2022: 8). In this
context, a good use of Wikipedia is to identify seed words or sources of texts that
will be double-​checked for authoritativeness and reliability later, especially if you
plan to include those texts in your corpus.
2 www.goo​gle.co.uk/​adva​nced​_​sea​rch
3 www.regu​lar-​expr​essi​ons.info/​
4 www.reg​exbu​ddy.com/​
5 https://​regex​101.com/​
6 www.kwi​cfin​der.com/​KWi​CFin​der.html
7 www.staff.uni-​mainz.de/​fanti​nuo/​tra​nsla​torb​ank.html
8 www.webc​orp.org.uk/​live/​
9 https://​wacky.ssl​mit.unibo.it/​doku.php?id=​start
10 www.webc​orp.org.uk/​wcx/​lse/​
11 www.sketc​heng​ine.eu/​
12 https://​magi​csea​rch.org/​
13 www.lexic​ool.com/​
14 https://​ope​nmd.com/​
15 https://​scho​lar.goo​gle.com/​
16 www.wipo.int/​en/​web/​wipo-​pearl
17 www.who.int/​
18 https://​sour​cefo​rge.net/​proje​cts/​alig​ner/​
19 https://​youal​ign.com/​
20 www.wordf​ast.net/​?go=​align
21 www.laur​ence​anth​ony.net/​softw​are/​antc​onc/​
22 www.parac​onc.com/​index.html
 Tools for Terminology Management 161

23 www.lexica​lly.net/​wordsm​ith/​
24 It may be somewhat confusing for translators to note that concordance also refers
to the name of the manual search in some translation memory tools (Bowker,
2005: 15; Mitchell-​Schuitevoerder, 2020: 84).
25 www.laur​ence​anth​ony.net/​softw​are/​antpc​onc/​
26 www.trado​oit.com/​
27 www.sketc​heng​ine.eu/​ (30-​day free trial)
28 https://​opus.nlpl.eu/​
29 www.ling​uee.com/​
30 www.wordre​fere​nce.com/​
31 https://​terms.sketc​heng​ine.eu/​
32 WIPO is the World Intellectual Property Organization and WIPO Pearl is its multi­
lingual terminological portal available at www.wipo.int/​web/​wipo-​pearl
33 https://​app.diagr​ams.net/​
34 https://​eco​lexi​con.ugr.es/​en/​index.htm
35 http://​demo.kalei​dosc​ope.at
36 https://​babel​net.org/​
37 www.bure​auwo​rks.com/​
38 https://​pro.wordsc​ope.com/​

Further Reading
Several online articles provide many useful guides and tips for the most efficient use
of Boolean operators with search engines, e.g. IONOS Digital Guide (2021, 2022),
Kati (2022), Ward (2020) and Hardwick (2024). For more on (text) alignment,
useful insights can be gained from Tiedemann (2011), García (2015), Simard, (2020)
and Rothwell et al. (2023). Regarding text mining and other aspects of informa­
tion retrieval, Monti et al. (2011) focus on technical translation, whereas Warburton
(2020) uses keywords and concordancers to identify corporate terminology. Zanettin
(2015) and Römer and Wulff (2010) also explore the use of concordancers for
translators, whereas those who have looked at the uses of Sketch Engine for trans-
lation include Moze and Krek (2019) (focusing on corpora and concordances) and
Yuliawati and Indira (2019) (focusing on partial synonyms).
For more on corpora, Fantinuoli (2016) examines some reasons why translators
prefer other resources to corpora, which seems to be supported by the figures from
Mikhailov’s survey (2022). Buendía-​Castro and López-​Rodríguez (2013) review three
ways to approach the Web as Corpus from a methodological perspective. A more
detailed comparison of Webcorp LSE and Google can be found in Renouf and Kehoe
(2013). Gries (2022) provides more information about statistics in quantitive corpus
analysis, including measures such as lexical gravity.
There are several publications on using concept maps to teach translators, such as
Falco (2017), Krishan (2017), Austermühl (2012), Ditlevsen and Kastberg (2009),
who discuss mapping techniques in the context of personal knowledge management.
Sheng (2024) uses concept maps to help interpreters acquire domain knowledge.
Mind maps are very similar to concept maps and there are many resources available
to learn how to use mind maps, e.g. Buzan and Buzan (2010).
The number of publications on AI and LLMs for translators is increasing exponen-
tially. Another book in this series focuses on translation automation and devotes a
162 Terminology Management for Translators

chapter to AI and LLMs (Moorkens et al., 2024). Meyer (2022) reviews developments
in the subfield of AI known as knowledge engineering, which can provide computer
assistance for terminological concept management. She also describes the knowledge
engineering tool for terminologists under development by her team at the Artificial
Intelligence Laboratory of the University of Ottawa. Some surveys regarding the
impact of AI and LLMs on translation and interpreting have already been carried out.
López Jiménez and Ouarachi (2020) surveyed experts and students, and one of their
findings is that “lifelong learning and constant re-​skilling are fundamental to face the
current acceleration of technology.” They suggest that training, e.g. at universities,
“should focus on improving unique human traits and skills related to social inter-
action” (2020: 15); Tavares et al. (2023) explain the views of Portuguese language
service providers whereas Kirov and Malamin (2022) survey Bulgarian translators on
the topic of AI. AI and LLMs are also prominent topics at translators’ conferences
such as Asling’s Translating and the Computer. For example, at Translating and the
Computer 46, Farrell (2024) presented the results of a survey of 425 professional
translators from 96 professional translators’ associations and provided useful insights
on how and why translators use (or not) generative AI. Bingham (2024) reports on
another conference, SlatorCon, which brought together “expertise to look at how
artificial intelligence and machine translation are really playing out in the market.”
Worne et al. (2024) discuss whether “the new language-​related AI should be put on
hold.” Downie (2020) devotes an entire volume to discussing whether interpreters
can survive in an AI-​dominated world.
Part II

Case Studies
Introduction to the Case Studies

The data for the Case Studies was gathered through visits in person to each
of the companies and organisations (except for Terra Translations which is
an online company), from April to November 2023. Qualitative information
about each organisation’s translation-​ based terminological processes and
procedures was collected through semi-​structured interviews (see Appendix)
and field notes with translators and colleagues holding other roles such as
project managers, terminologists, etc. as available on the days of the visits.
No specific duration was set for the interviews, they lasted for as long as the
participants had information to share in answer to the questions. On average,
most interviews lasted two to three hours each day on two to three days.
By investigating the relevant processes and procedures at the translators’
workplaces, we ensure that the data collected remains ecologically valid, as
stated by Ehrensberger-​Dow (2014: 358).
The selection of organisations for this research was necessarily limited to
the availability and willingness among staff to participate in this research.
After some background research online, 100 emails or so were sent in October
2022 to reach out to key contacts at several organisations and language ser-
vices providers. Only a few responses were received but they led to contacts at
other organisations and, eventually, mostly by word of mouth, simultaneous
contact was established with the following organisations: the Translation
Bureau at the Government of Canada, the World Bank, the International
Monetary Fund, the World Trade Organization, the Kaleidoscope Group (and
its client Blum), Terra Translations and a pharmaceutical company. Research
on the International Monetary Fund, the World Trade Organization and the
pharmaceutical company is ongoing, and the case studies will be added to the
online portal for this book in due course.
In Part II, therefore, we present the results of the semi-​structured interviews
at two institutional translation workplaces (the Translation Bureau in a gov-
ernmental setting and the World Bank in a large international translation
setting), and two commercial workplaces (the Kaleidoscope Group with
its client Blum, and Terra Translations), as summarised in the table below.
As we will see in the following chapters, outsourcing of translation work

DOI: 10.4324/9781003302636-8
164 Terminology Management for Translators

is a main business model at each of these workplaces (Olohan, 2016: 11;

Walker, 2023: 7) and, therefore, translation work in commercial and institu­
tional settings is strongly underpinned by a large, global freelance translator
workforce.
Summary of participating organisations and companies in the Case Studies

Organisation Type Location visited Date of interviews

The Translation Institutional Gatineau, April 2023

Bureau (governmental Quebec,
focus) Canada.
The World Bank Institutional Washington DC, August 2023
(large USA.
international
focus)
The Kaleidoscope Commercial Vienna, Austria. November 2023
Group (and its
client Blum)
Terra Translations Commercial [online] August/S​ eptember
2023

The Case Studies in this book encompass all three broad translator’s
workplaces, i.e. institutional, commercial and freelance. Institutional trans-
lation can be a slippery concept to define (Koskinen, 2014b: 2) because
different interpretations can be made of institution (Koskinen, 2014a: 480)
but, in this book, we will define institutional translation as “any translation
carried out in the name, on behalf of, and for the benefit of institutions”
(Gouadec, 2007: 38). Thus, “in institutional translation, the voice that is to
be heard is that of the translating institution” (Koskinen, 2014b: 22).
Typical features of institutional translation, which set it apart from com-
mercial translation, include the anonymity of the translators, who make a
collective effort to convey the ‘voice’ of the institution, a necessary respect
for past translations (with recurring texts with subsequent amendments
and cross-​references that require careful and vigorous verification), and a
particular conventional structure and style for the production of often ter-
minologically complex documents (Garrido Nombela, 2013: 145). The type
of institution determines which kinds of discourse and texts are produced
(Schäffner et al., 2014: 494). In addition, “issues of power, status and
authority colour the [institutional] contexts in numerous significant ways”
(Koskinen, 2011: 59). Perhaps “the most distinctive element of institutional
translation is not its translation agents per se but the institutionalized norms
and conventions that these must observe, that is, established translation pol-
icies and practices to serve institutional purposes in compliance with each
institution’s language policy” (Prieto Ramos, 2020a: 455).
 Case Study 165

Institutional contexts are many and varied (Koskinen, 2014b: 2) but gov­
ernmental translation may be considered a subset within institutional trans-
lation because “governments are national in scope [and] have developed
mature [terminology management] processes that fit public service well”
(Warburton, 2021: 35), as opposed to large international organisations
which are multinational in scope. Thus, this research contributes to filling
a gap in institutional translation, which is a rather under-​researched area of
translation (e.g. Kang, 2009: 260; Prieto Ramos & Guzmán, 2021: 255), by
providing empirical data on two potential subsets of institutional transla-
tion, i.e. governmental settings and large international translation settings.
For more on institutional translation, see Joscelyne (2000), Mossop (2014),
Svoboda et al. (2017), Ding and Li (2017), Poirier and Gallego-​Hernández
(2018), Choi (2018, 2022), Lafeber (2023), Svoboda et al. (2023), and the
“Further Reading” section in Chapter 8.
In general, institutional translation differs from commercial translation1 in
that commercial translation is constrained by “market factors which are so
fundamental to commercial activities” (Warburton, 2021: 35), such as fre­
quent staff changes, a global marketing division, dramatic changes presented
by mergers and acquisitions and cost-​driven efforts to shorten time-​to-​market,
etc. (2021: 35). In addition, institutional translation settings “share the man-
date of serving the linguistic needs of society at large, which is not the case for
[commercial] companies” (2021: 51). There are also many types of commer-
cial organisations, e.g. language services providers, who may employ both in-​
house and freelance translators, and companies “whose core business is not
translation [but which have] a translation department employing in-​house
translators” (Olohan, 2016: 10). The Case Studies in Part II include two lan­­
guage services providers (the Kaleidoscope Group and Terra Translations)
and one non-​language services provider (Blum, as a client of the Kaleidoscope
Group). For more on terminology and translation in commercial translation
settings, see Olohan (2010), Poirier and Gallego-​Hernández (2018), Esselink
(2019), King (2019), Warburton (2021) and Fırat et al. (2024).
By definition, freelance translators work on their own as sole traders.
Freelance work is the most common form of employment for most
translators worldwide (Olohan, 2016: 9; Penet, 2024: 25). When working
for their own direct client base, and in some commercial settings, freelance
translators may work “without the benefit of the revisions, advice, refer-
ence materials, databases and support” (Borja Albi, 2013: 54), but freelance
translators working in institutional settings often have access to the same
advice, support and materials as translators working in institutional settings.
For more on working as a freelance translator and for more insights into
working ‘in-​house’ in either commercial or institutional settings, see Horváth
(2016), Zetzsche (2019), Penet (2024: 23–​40) and Walker (2023: 6–​10).
Each of the Case Studies is organised as follows: the “Background”
section places the organisation or company firmly in context by outlining its
origins, the circumstances that led to its creation and how it has evolved to
166 Terminology Management for Translators

the present day. As Woods (2006: 9) states, the “historical record helps us
critically to evaluate the nature of the organizations.” In the “Organisation
and Workflow” section, we present the internal structure and hierarchy
of the organisation or company. This includes listing the main working
languages and directions, the names of roles (e.g. does the organisation have
terminologists, revisers, etc.?), how translators fit in within the organisation,
the role of freelance translators, and tracing a typical translation workflow
(although many variations are possible at each organisation), from the client’s
commission to delivery. In the “How Do Translators Manage Terminology”
section, we explore the work of translators with a focus on terminological
processes and procedures, by considering the range of (technological) tools
at their disposal, mechanisms for support, the design and contents of their
termbases, as well as any challenges and constraints at the interface between
terminology and translation. In these sections, except for the “Background”
section, readers should be aware that the information about the companies’
and organisations’ processes and procedures represents just a snapshot in
time of how they are implemented at each company and organisation. In
other words, these processes and procedures change very rapidly and it is not
possible to present information in this regard that will not change with time.
Although the information in these sections may be (slightly) out-​of-​date by
the time you read them, it nevertheless is a good illustration of how the com-
panies and organisations worked at a particular moment in time and may
help readers trace the evolution of such processes and procedures at these
workplaces over time.
Each Case Study concludes with an “Outlook” section, where we provide
a brief overview of emerging trends and issues, which might give us some
indication of what the future might hold. There are several follow-​up tasks
and reflection for each Case Study and some relevant suggestions for further
reading.

Note
1 In this section, the term in-​house is used to refer to translators who are employed
by companies (commercial) or organisations (institutional), as opposed to freelance
translators who are self-​employed.
7 Governmental Case Study
The Translation Bureau

Key Highlights

• Terminologists working alongside translators.

• It is more efficient to leave the bulk of terminology management
tasks to (in-​house) terminologists.
• Permanent terminological committees deal with ongoing termino-
logical issues in specific areas, e.g. railways.
• Ad hoc terminological committees can be formed when needed, to
solve a specific terminological issue, e.g. neologisms.
• Terminological committees can include not only translators and
terminologists, but also clients, subject matter experts, etc.
• Revision always in-​house.
• Phraseological information included in termbase entries.
• TERMIUM® termbase publicly available online in French, English,
Spanish and Portuguese.
• Many language resources in the Translation Bureau’s website and in
the Language Portal of Canada website.

Historical Background
Canada, as we know it today, was founded in 1867 after having been
governed by both the English and the French (Gagnon, 2021: 25). Since then,
Canada has embraced bilingualism and in 2010 it accounted for 10% of the
global translation market (Hamilton, 2010: 13). Translation in Canada is
mainly between English and French, the two official languages, although it
is worth highlighting that they are not the country’s only languages. Despite
the fact that it is difficult and even contentious to establish what exactly
constitutes a language and where a language variety or dialect begins and
another one ends (Patrick, 2018: 262–​263), it has been estimated that over
70 Indigenous1 languages are spoken throughout Canada (Statistics Canada,
2023), grouped in 12 different language families, the largest of which is
DOI: 10.4324/9781003302636-9
168 Terminology Management for Translators

Algonquian, followed by Inuktitut (LeBlanc, 2018: 351). According to the

Government of Canada, English is the official (spoken) language of about
76% of the population, whilst French is the official (spoken) language of
22%, with some 2% having neither French nor English as official (spoken)
languages (Government of Canada, 2024c). In terms of mother tongues,
also according to the Canadian government, English is the mother tongue of
about 57% of the population and French of 20%, with the remaining 23%
belonging to other languages (2024c).
These figures illustrate the fact that, although Canada is often referred
to as a bilingual country, there is in fact a disproportion between English
and French, both numerically and geographically. According to Hamilton
(2010: 12), few Canadians actually speak both English and French (Mossop,
2006: 4). He further estimates that a quarter of the population that speaks
French is geographically concentrated in a single area (Hamilton, 2010: 13)
and Roberts (2019: 2) reports that in 2016 there were 34 million English-​
speaking Canadians as opposed to 7 million French speakers. As a result,
most of the translation work in Canada is in the English into French direc-
tion, rather than in the opposite one. Other language combinations that are
becoming quite frequently requested include translations between Inuktitut
and English, especially since the publication of the Indigenous Languages Act
in 2019 (MultiLingual Magazine, 2022), and between Spanish and English.
In addition, LeBlanc (2018: 351) calculated that there are a couple of hun­
dred other mother tongues spoken in Canada, mostly due to immigration.
These include Chinese, Punjabi, Tagalog, Arabic, Polish and Portuguese, to
name but a few (LeBlanc, 2018: 351).

Box 7.1 Translation Bureau –​Translation Facts and figures

Location: Gatineau, Quebec Own MT? No.

No. of translators: 600 approx. Project Management tool: GClingua
Terminologists: 40 approx. Specialism required/​encouraged: Yes
CAT tool: RWS Trados Studio Outsourced work: 48% on average
Termbase: RWS MultiTerm, Main source language: English
TERMIUM®

Despite these numerous and diverse languages, English and French remain
the two official languages throughout the country, enshrined in law after the
Canadian Parliament passed the Official Languages Act in 1969 (Macklovitch,
2014: 268), “primarily to address the under-​representation of Francophones in
the federal civil service” (Leimgruber, 2019: 36). It was not until 1982, how­
ever, with the Canadian Charter of Rights and Freedoms, that equal status
was guaranteed between the two languages (Roberts, 2019: 2). Despite this
duality of official languages, only one province, New Brunswick, and two
 Governmental Case Study: The Translation Bureau 169

territories, the Yukon and the Northwest Territories, are officially bilingual
(Reed, 1993: 79). In Nunavut, a territory created in 1999, Inuktut is official
alongside French and English (see Lepage et al., 2019). Quebec, by contrast,
is the only province that is officially unilingual French (Hamilton, 2010: 13).
According to Hamilton, Quebec’s “Charter of the French Language, intended
as a bulwark against the encroachment of English, requires businesses to
operate in French and deal with the government in French, and requires non-​
Canadian citizens (such as immigrants from the US) to send their children to
French-​language schools” (2010: 13). However, Quebec’s official unilinguality
does not “apply to the courts or to individual citizens in their private dealings
with the government” (2010: 13).
The largest employer of translators in Canada (and indeed one of the
largest around the world) is its federal government (Hamilton, 2010: 13;
Macklovitch, 2014: 267) which established the Translation Bureau (Bureau
de la traduction) in 1934 as its translation service (Mossop, 2006: 2) (see
also Box 7.1). As well as the government’s translation and interpreting ser­
vice, the Translation Bureau has a third, terminological role whereby its
Terminology Centre will “carry out terminological research and prepare
and distribute terminological instructions and bulletins” (Government of
Canada, 2024a: 10–​11).
From its inception until 1995, all government agencies and departments
were required to use the services of the Translation Bureau, which were free
(LeBlanc, 2018: 344). In 1995, also according to LeBlanc, major changes
were introduced which meant that it was no longer compulsory to use the
Translation Bureau and that translation services were no longer free of charge.
From then on, government agencies and departments were able to decide
to use an external LSP for translation or continue to use the Translation
Bureau at a cost. The Translation Bureau is organised into four “business
lines” (Translation Bureau, 2023) as shown in Figure 7.1, provided in the
webpages of the Bureau (Translation Bureau, 2023). Translation and termin­
ology services are provided through the “Linguistic Services” business line,
whereas the bulk of interpreting services are provided through the “Services
to Parliament” business line.
The Translation Bureau continues to be the translation service of the federal
government, and it now runs on a cost recovery basis. This means that
the Bureau has its own budget, is independent and a self-​sustaining agency
(Nimdzi, 2018). It handles some 80% of the volume of the government’s
translation needs and provides translation, interpreting and terminology
products and services to the Canadian Parliament, the courts and federal
departments and organisations, in both official languages as well as in more
than 100 Indigenous languages, foreign languages and sign languages, with
approximately 1,300 staff (PSPC, 2017), although staffing levels have been
steadily declining since 2011, largely due to the increased use of technology
(LeBlanc, 2016: 344). Since 1995, the Translation Bureau only works for
governmental clients and reports to Public Services and Procurement Canada
170 Terminology Management for Translators

Figure 7.1 Organisational chart of the Translation Bureau.

Source: Translation Bureau (2023).

Figure 7.2 Translators’ offices at the Translation Bureau, Gatineau, Quebec.

(Services publics et Approvisionnement Canada in French), a special operating

agency (2016: 344). The vast majority of linguistic staff are based in the offices
in Gatineau, Quebec (see Figure 7.2), but there are also offices in Montreal,
Quebec City, Toronto, Winnipeg, Moncton, Halifax and Vancouver.
On average, the Translation Bureau translates 350 million words a year,
provides 10,500 days of interpreting, 4,500 hours of closed captioning for
 Governmental Case Study: The Translation Bureau 171

the Parliament and 7,200 hours of visual interpreting for the hard of hearing,
among other tasks, products and tools, and it provides this service 24/​7
(Nimdzi, 2018). It should not come as a shock, therefore, that Canada has
been dubbed a “virtual paradise for translators” (Roberts, 2019: 1). In the
following sections, we take a closer look at how the Bureau conducts its
translation and terminology operations.

Organisation and Workflow2

Of the current 1,300 or so staff at the Translation Bureau, there are approxi-
mately 600 translators and 40 terminologists. Although the Bureau also
employs interpreters, this section focuses only on translators (who translate
and deal with some terminological issues) and terminologists (who deal with
terminology but do not translate), not only because they are the object of
study in this book but also because interpreting is a separate service and
“business line,” mostly associated with the Parliament, as we have seen in
Figure 7.1. In addition to the 1,300 in-​house staff, the Bureau also outsources
much of its translation (though not revision) work to freelancers, mostly
due to lack of capacity and the demand for foreign languages, e.g. Russian,
Arabic, Tagalog, etc., which may not necessarily be offered in-​house. On
average, 48% of the work is outsourced to external providers, although this
figure varies across departments. In some years, for example, more than 80%
of the translations required for departments such as Fisheries and Oceans
Canada needed outsourcing, often due to the quality control needed in this
domain. At the time of writing, 90% of source texts were in English, with
most of the translations being carried out into French.

Translators

Translators receive up to two years’ training where they develop their trans-
lation skills, familiarise themselves with client terminology and preferences,
and learn about the resources available and how to use the technology. Entry
level translators are referred to as TR-​01 and can progress to TR-​02 and
TR-​03. Translators will typically develop and train in a specialism and are
encouraged to continue increasing their knowledge in that specialism. Some
translators may be assigned a specific client, depending on several factors,
including their experience in the relevant specialism. When having worked
with a particular client for a while, translators will have also developed ample
knowledge of the client’s preferred terminology and phraseology usage. Some
translators can also take the role of content specialist, which involves keeping
resources and documents up to date, and building and updating documenta-
tion kits. One of the specialisms with a great deal of demand for translation
is law. There are two law systems in Canada,3 common law and civil law,
which means that there is often much to translate, especially after a parlia-
mentary sitting.
172 Terminology Management for Translators

Translators do their due diligence on a daily basis to research terminology

equivalents. When they are faced with terminology difficulties, they can con-
tact the terminology team for help. Although translators manage the specific
terminology for their specialties, the main task of managing the terminology
data bank is carried out by terminologists. The Bureau’s view is that it is not
efficient for translators to carry out many terminology management tasks,
which are left to terminologists, allowing translators to devote most of their
time to the translation task at hand.

Terminologists

The terminology team (about 40 staff at the time of writing) is “responsible

for the creation and management of about 1.4 million terminology records
documenting 4 million terms in 4 languages (English, French, Spanish and
Portuguese) and the maintenance of additional data in 9 other languages
(such as Inuktitut)” (Translation Bureau, n.d.: 3). The team also deals with
terminological queries as they arise.
Terminologists-​in-​training are referred to as TR-​01, ‘regular’ terminologists
are TR-​02 and senior terminologists as TR-​03. Everything a TR-​01 does is
checked, whereas only a portion of the work of a TR-​02 is checked. Typically,
every two months, 25 records are extracted, revised by a senior terminologist
and sent back to the TR-​02 with feedback. TR-​03s aim to revise a minimum
of 20% of the work of every TR-​02 each year.
Terminological committees can be formed when needed, usually when a
specific terminological issue arises that requires wider discussion and con-
sultation, including neologisms. For example, many ad-​ hoc committees
have been created for issues arising from areas such as fisheries and oceans.
There are also several permanent committees for areas with special issues,
e.g. railway safety, international aviation, etc. Terminological committees are
made up of internal (translators and terminologists) and external members
(clients, subject matter experts, language users, etc.).
Issues can arise from a translator’s query or from a client’s complaint.
Client’s complaints about terminology are the most frequent reason for the
creation of a committee or discussion in a permanent one. However, not every
issue that arises requires a committee –​sometimes satisfactory solutions are
arrived at and no committee needs to be formed.

Workflow

The Bureau has an online ordering system (GClingua) for ‘regular’

translations to which clients can log in and commission a translation. The
Request Processing Centre receives the documents to be translated and
checks the domain they belong to, so that they can be sent to the relevant
translator(s). The online system features a Calendar with translators’ and
revisers’ schedules so that the Request Processing Centre can send the work
 Governmental Case Study: The Translation Bureau 173

Figure 7.3 Simplified translation workflow at the Translation Bureau.

to available colleagues. If the translations are 1,500 words or less, they can
often be delivered on the day after they were commissioned but there are
exceptions and variations for specific jobs.
The Bureau’s clients can also access the SVP Service, an ad hoc terminology
service. A simplified translation workflow at the Bureau is summarised in
Figure 7.3. Interpreting is not included here because, as we saw in Figure 7.1,
interpreting is provided through a different “business line” than translation
and terminology.
Figure 7.3 illustrates in a simplified way how terminology management
serves the translators at the Bureau. On the one hand it is an entire standalone
apparatus, helping to standardise terminology throughout the federal public
service and, on the other hand, this large mechanism underpins the work of
the in-​house translators, providing them with specialised knowledge when
they need it and freeing them up to focus on the translation task.
At the start of a job, translators are assisted by a Request Processing
Centre (RPC) team which receives the translation requests, pre-​processes the
files to be translated and assigns them to the relevant translators or teams.
For example, they can prepare the formatting of a source text for optimal
use within RWS Trados or split a long source text among several translators.
When necessary, this team can also carry out some desktop publishing tasks
before delivery of a translation. This team can also assign translations to
external suppliers if the translation is to be outsourced. They are the first line
of enquiry for freelancers’ queries and can contact the commissioner of the
translation as required. Within the RPC team, there are managers and project
174 Terminology Management for Translators

managers. Managers have staffing responsibilities and manage a budget.

They do not carry out linguistic work. By contrast, project managers have no
staffing or budgeting responsibilities.
Once a translation is finished, it is sent to a reviser. There are four levels
of revision at the Bureau:

1 Complete revision by comparing the source text to the target text line
by line.
2 Re-​reading: checking the target text only, not the source text, looking
for any inconsistencies. The source text is only consulted if anomalies are
found in the target text.
3 Sampling: quick revision of approximately 15% of the document, only
applied when the provider is a reliable supplier already. Otherwise, the
full target text is revised.
4 Proofreading: checking that numbers, names, etc. are accurate in the
target text.

Revision of the target text is usually performed in Trados (see the next
section) which is connected to the Bureau’s online portal. Thus, once a trans-
lation is finalised on Trados, the job will be updated and delivered to the
client through the portal.
The Quality Office coordinates translation delivery and revision, as well as
terminological research and typically carries out 300 assessments each year,
90% into French and 10% into English approximately. The assessments con-
sist of checking term inconsistencies, not only within translated documents
but also relating to the quality of the term records in the TERMIUM®
termbase4 (see the next section).
Alongside translators, terminologists and various working groups,
offices and committees, the Bureau also has a Documentation Centre that
provides services to both translators and terminologists. These services
include searching for documents on a particular topic in order to build
a corpus, searching for the equivalents of the titles of reference works in
other languages, searching for citations and formatting bibliographies that
accompany the Bureau’s glossaries and vocabularies. The Documentation
Centre also manages subscriptions to magazines, journals, newspapers,
databases, etc.

How Do Translators Manage Terminology at the Translation Bureau?

There are two large terminological databases in Canada, the Grand dictionnaire
terminologique (‘large terminological dictionary’) and TERMIUM®, which
both staff at the Translation Bureau and the public can access. The Grand
dictionnaire terminologique, also known by its acronym GDT, is produced by
the Office québécois de la langue française (Quebec Office of the French lan-
guage) (L’Homme, 2006: 57) and currently contains close to 3 million terms
 Governmental Case Study: The Translation Bureau 175

in French and English from a wide range of economic, technological and

scientific domains. It is publicly available online.5 TERMIUM®, produced
by the Government of Canada, comes in two flavours. On the one hand,
TERMIUM Plus® is one of the largest term banks in the world, often cited
alongside termbases such as IATE or UNTERM (e.g. Bowker, 2015: 308). It
provides millions of terms in French, English, Spanish and Portuguese and it
can be accessed online6 by the general public free of charge.
On the other hand, staff at the Translation Bureau have access to an
internal version of TERMIUM®, referred to as TERMIUM® 6 or T6. This
is an even larger version of TERMIUM Plus® and it is also the internal con-
tent management system used by the Translation Bureau to manage the
TERMIUM® PLUS database. The information on TERMIUM Plus® is fed
from T6 but T6 also contains additional information for internal consump-
tion only, e.g. such as the author of the record. Furthermore, T6 allows
translators and terminologists to create glossaries and reports (Translation
Bureau, n.d.: 8). T7 is currently being further developed and we can expect
the new TERMIUM® 7 to be launched in the short term.
T6 is the daily tool for terminologists. Translators use it for terminological
consultations and they can also add information to the term records. T6 is
hierarchical and concept-​oriented, so each record corresponds to one con-
cept. This means that, for the Bureau, cosmetic synonyms such as telework
and tele-​work will be found within the same term record and referred to
as variants. They will have markers to highlight the distinguishing factors
between them, e.g. archaic. In the case of partial synonyms, it will depend on
how semantically close they are. If they are semantically the same or closely
similar, they will appear in the same term record. If their meanings are too
different, they will be entered on separate term records. In the case of pseudo-​
synonyms (e.g. tidal wave and tsunami in the Relations between Terms section
in Chapter 4), the policy at the Bureau is to research how well established
the terms are in the field. If well established, both terms will appear on the
same term record. If not, they will appear on separate records, with a cross-​
reference from the less well-​established term to the main one. Thus, if tidal
wave is the longer established term, its record will have tsunami listed on it
as a synonym. Tsunami will also have its own separate record with a cross
reference to tidal wave. However, there will be no cross-​reference from tidal
wave to the standalone term record for tsunami.
Term records are colour-​coded: draft records have a grey background,
whereas records published to TERMIUM Plus® are green and translators’
records, for internal consumption only, are yellow. Grey records will even-
tually become either yellow or green. With this system, terminologists can
trace the quality of the record. Green records are accepted as complete and
validated term records whereas yellow records do not undergo the same
quality processes. However, if the information contained in a yellow record
becomes useful for the public at large, it can be published to TERMIUM® as
a green record. Otherwise, it will remain as a yellow record on T6.
176 Terminology Management for Translators

There are two kinds of records, long forms and short forms. The long
forms are used by terminologists to create a complete, accurate and veri-
fied record, which can also be used to update TERMIUM Plus®. The short
forms are used by translators and usually consist of a term and its context
and notes. Typically, they do not contain fields such as Definition or Sources.
The short forms are not checked or uploaded to TERMIUM Plus® but they
constitute a valuable source of both terminological and phraseological infor-
mation, as well as client’s preferences, for translators’ future reference. Once
created, yellow records are available to all translators.
For example, the term record for lean management in the domain of cor-
porate management contains several terms in French, e.g. gestion allégée and
gestion lean. These two French terms appear in TERMIUM Plus® which
means they have undergone quality control by terminologists. However, it
may be that a particular client prefers to use gestion lean rather than the other
term. Once translators learn of such preferences, they can create a record but
this kind of information does not go in the public-​facing TERMIUM Plus®.
New fields can be added to entries in TERMIUM®. If there is a need to
record information on an entry that does not have a field for it, a note will be
made in the Observation field (see below). The next version of TERMIUM®,
currently under development, will be aligned with the data categories found
in the (now rather old) TBX standard.7
Excel files are regularly used, especially by terminologists, in order to
collect and import terminological data. Excel files are most often used to
record the results of manual or automatic term extraction. Validated term
extraction data can then be imported into TERMIUM® as working records,
or they can be sent to a client as a list of terms. Excel is also the current
method for the exchange of terminological data with partners.
The pre-​established descriptive fields available in an empty term record
are listed in Figure 7.4, alongside a sample record in English. At the top
of the entry, the fields at Entry level are Subject field (or domain), Client
and Author, with another field for metadata, although Client and Author
are not published to TERMIUM Plus® at present. At Language level, the
fields Term and Synonym are entered in full form but for the remaining fields
abbreviations have been used:

• DEF: definition; it is interesting to note that, at the Bureau, definitions

must be one sentence only, starting with the closest hyperonym of the term.
All other information must be included in the Observations (OBS) field.
• CONT: from Context; in this field, a sentence must be entered containing
the term and showing an example of its usage; in some termbases, this field
is referred to as Example or Example of usage.
• OBS: from Observations; this is a multiple field, which is essentially a
Notes field, including any information not already collected in the DEF,
CONT or PHR fields.
• PHR: phraseological information is captured here.
 Governmental Case Study: The Translation Bureau 177

Figure 7.4 Descriptive fields and hierarchy in TERMIUM® (left) and example of a

TERMIUM® entry in English, shortened (right).

Throughout the text of the fields above, the numbers refer the reader to the
sources of the information for this entry, which can be found at the bottom of
the entry under Cited Sources. Finally, the entry ends with metadata, showing
when the entry was last modified and by whom, as well as other metadata.
Part of speech does not feature in Figure 7.4 but the latest guidelines require
all new terms to include it.
The main CAT tool provided by the Bureau for translators at present is
RWS Trados Studio (currently 2022 version) but the use of another RWS
product, MultiTrans, is being considered for the future, as it seems to be
more seamless than Trados for the import and export of terminological data
from T6. Only translators have licenses to use Trados. When term extraction
is required, terminologists often resort to a statistical bilingual term extrac-
tion tool available in 30 or so languages (its name cannot be disclosed). This
tool can also perform alignment tasks although translators may also use
other alignment tools.
The Bureau does not have its own machine translation system or any plans
at present to develop one. The decision to adopt an MT system or not is
very centralised, lying with the Canadian government, so it is beyond the
Bureau’s remit. Several MT systems have been tried over the years but none
has been adopted so far. Other systems continue to be tested. In the mean-
time, translators at the Bureau typically use the MT provided through Trados
anyway the enterprise machine translation service provided by Intento.8
Some of the Bureau’s clients send raw MT translations of a source text for
the Bureau to post-​edit. Depending on the quality and purpose of the transla-
tion, the target text is either revised or translated from scratch.
178 Terminology Management for Translators

The well-​known MÉTÉO machine translation system, developed by a

research group at the University of Montreal in 1976 (Hutchins, 2004: 2),
was used by the federal agency Environment Canada until 2004, when
MÉTÉO was “rendered obsolete by an interactive system called SCRIBE …
that meteorologists now use to create forecasts in both languages simultan-
eously, thus obviating the need for translation” (Gotti et al., 2013: 400–​401).
Translators from the Bureau contributed to this system by supervising the
process and making sure “that the occasional spelling error or other difficulty
found in the source text did not prevent the production of a French forecast”
(2013: 400–401).
To help with their terminology management tasks, translators and
terminologists have further tools at their disposal, many of which are pub-
licly available at the Language Portal of Canada.9 TERMIUM Plus® can be
accessed from this page along with tools such as the Inclusionary10 and the
Language Navigator.11 The Inclusionary is a collection of gender-​inclusive
solutions, whereas the Language Navigator allows you to “search by key-
word or theme to find answers quickly to questions about language or writing
in English and French” (Language Navigator webpage). The Language Portal
of Canada’s website also contains links and information about a wide range
of other tools, resources and featured articles.

Outlook
The Translation Bureau is currently modernising its terminological land-
scape, which will inevitably affect its translators and translation processes,
so we can expect to see new developments in this area. The vision statement
for the Bureau includes engaging with the language industry and working on
reconciliation with Indigenous languages. Further, inclusivity, diversity and
initiatives against racism are new priorities for the Canadian government.
These are some of the areas in which the emergence of new terms can be fore-
seen. For example, reconciliation is already a specialised term with a specific
meaning in the domain of Indigenous affairs. Apart from its meaning in gen-
eral language where it is defined as the action of restoring estranged people
or parties or friendship according to the Oxford English Dictionary, the
Canadian Encyclopedia documents the definition of reconciliation drafted by
the Truth and Reconciliation Commission (TRC). The TRD defines recon-
ciliation in Canada as the “process of establishing and maintaining a mutu-
ally respectful relationship between Indigenous and non-​Indigenous peoples”
(Sterritt, 2020). Beyond this level, reconciliation is “a complex, multi-​
faceted, ever-​evolving journey of great importance to the healing, forgive-
ness, acceptance and well-​being of individuals in Canada, both Indigenous
and non-​Indigenous” (NATIONS Translation-​Group, 2022). It is an ongoing
“process of acknowledging and deconstructing the painful behaviours and
 Governmental Case Study: The Translation Bureau 179

social structures of colonization and moving forward with a shared vision

of a better world” (2022). New terms in this domain include Indigenous
Languages Commissioner and Indigenous Languages Act, both new since
2019. All these terms and any new ones will need to be dealt with sensitively
in translation.
Terminological challenges for translations will also arise from inclusivity,
diversity and racism issues. For example, in Indigenous affairs, it has been
suggested that stakeholder should be replaced with terms along the lines of
rights holder, for example. Black list will need to be replaced by block list, for
example. Regarding gender issues, new, neutral pronouns are being proposed
in French, such as iel, ille, al or ol for the third person singular (Aussant,
2019). More issues and guidelines regarding inclusive writing can be found
in the Inclusionary (see above).
When many terms emerge in a new area, the government typically requests
the Bureau to produce a glossary that can be used across all governments
agencies and departments to ensure consistency in the terminology used.
After all, one of the constituting mandates of the Bureau is to standardise
terminology throughout the federal public service, not only to provide trans-
lation and interpretation services. For example, Justice Canada is a govern-
mental department that is currently defining sex and gender for the new laws
that are being drawn up. Once the concept and the labels (terms) to refer to
them have been accepted and validated, no doubt they will find their way
into TERMIUM® (T6 and Plus) and thus become available to both to the
general public and experts in Canada and beyond.

Follow-​up Tasks and Reflection

1. Why do you think Canada has been described as a “paradise” for
translators? Could we describe the Translation Bureau as a “paradise” for
terminologists? What would constitute a “paradise” for terminologists?
And for the terminology work of translators?
2. To what extent should the view that Canada is a bilingual country be
replaced by the view that it is more accurate to refer to it as a multi-
lingual country? Compare bilingual vs multilingual with official vs non-​
official –​should other languages in Canada, such as Inuktitut, be included
in TERMIUM Plus® even though they are not official? What would be the
terminological challenges of doing so?
3. Reconciliation can now be considered a term in the domain of Indigenous
affairs in Canada. What other terms might be included in a termbase for
this domain? How important would fields be such as Acronym, Graphic
(e.g. map) and Notes in this termbase?
180 Terminology Management for Translators

Notes
1 Indigenous should be capitalised when it refers to peoples in Canada (Government
of Canada, 2020b).
2 The names of certain software programs and tools have been omitted from this
chapter following the Bureau’s request to do so.
3 Common law applies in all provinces. In Quebec, however, civil law applies to
matters of private law. Beaudoin provides more information about legal transla-
tion in Canada (2009).
4 In general, “TERMIUM” is written in uppercase, in italics, with the symbol
in superscript, but not in italics: TERMIUM®. Mentions of TERMIUM® and
TERMIUM Plus® should be in uppercase and italics and with the superscript
registered trademark. The Plus is not in uppercase. The version is written in Arabic
numerals (6, 7, 8…), not written in italics. For example: TERMIUM®, TERMIUM
Plus®, TERMIUM® 6, TERMIUM® 7, etc.
5 https://​vitr​inel​ingu​isti​que.oqlf.gouv.qc.ca/​
6 www.btb.term​iump​lus.gc.ca/​tpv2al​pha/​alpha-​eng.html?lang=​eng
7 www.tbxi​nfo.net/​develo​per-​resour​ces/​tbx-​data-​cat​egor​ies/​
8 https://​inten.to/​
9 www.nos​lang​ues-​ourla​ngua​ges.gc.ca/​en
10 www.nos​lang​ues-​ourla​ngua​ges.gc.ca/​en/​writ​ing-​tips-​plus/​inclu​sion​ary
11 www.nos​lang​ues-​ourla​ngua​ges.gc.ca/​en/​nav​igat​eur-​naviga​tor

Further Reading
Brian Mossop was a full-​time translator at the Translation Bureau in Toronto for
40 years, from 1974 to 2014, and is an influential scholar in Translation Studies. He
has published his first-​hand insights and analyses in several publications (e.g. 1988,
2006, 2023). Gagnon takes a look at how the federal government of Canada “can
be seen as a typical example of governance through translation” (2021). L’Homme
(2006) examines the role played by the Translation Bureau and another govern­
mental organisation, the Office québécois de la langue française, in the development
of terminology as a discipline in Canada. On the topic of interpreting services at the
Translation Bureau, Delisle (2009) provides a useful and detailed historical overview,
focusing on simultaneous interpreting at the Canadian Parliament. To know more
how interpreting services are offered at the Translation Bureau, it is worth reading the
relevant federal government webpage (Government of Canada, 2023).
In order to dispel any confusion about the three types of Indigenous Peoples in
Canada, McMillan (1995) provides a thorough historical overview of each group,
together with a wide range of related geographical, ethnical and sociological issues.
They are the First Nations, Inuit and the Métis. Patrick (2018) and Allen (2007)
also help shed light on the terminological confusion between the terms Inuit and
Inuktitut. Patrick also suggests that using the term minority to refer to Canada’s
Indigenous languages is a misnomer, given that they “are spoken by the majority
of Arctic residents and, as with other Indigenous languages in Canada, have rights
associated with them that are specific to the country’s historical Indigenous-​state
relations” (2018: 260). Instead, she uses the term small languages.
 Governmental Case Study: The Translation Bureau 181

For a closer look at the linguistic situation in Quebec, see Rousseau (1993). From
the point of view of language planning and policy, with a historical and political
background, Leimgruber (2019) is an up-​to-​date account, whereas Oakes and Warren
(2007) look at Quebec from a broader sociolinguistic perspective. The Indigenous
Languages Act of 2019 can be read online (Government of Canada, 2020a). In the
true spirit of the reconciliation promoted by this Act, appointments were made in
2021 for the newly created posts of Commissioner of Indigenous Languages and
Director of Indigenous Languages. You can read more about it on the government’s
website (Canadian Heritage, 2021).
8 Large International Organisation
Case Study
The World Bank

Key Highlights

• Linguistic roles include translators/​revisers and editors.

• Editors are an additional layer of terminological quality assurance.
• Terminology work is carried out by translators/​revisers.
• It is not compulsory to translate every document, but a great deal of
translation is carried out.
• Main languages: Arabic, Chinese, English, French, Russian and
Spanish, but much of the work takes place in English.
• Working languages instead of ‘official languages.’
• Substantial investment in term extraction to consolidate and improve
terminological resources.
• Terminological consistency valued alongside phraseological consistency.

Historical Background
Many of us will have heard of the World Bank and of the International
Monetary Fund (IMF), but few will actually know much about their
internal workings and even less so about their in-​house translation teams
and their organisation. In fact, despite both institutions being in operation
for eight decades, many authors report that confusion and mystery continue
to surround them, sometimes one institution being confused with the other,
and that, in general, they remain poorly understood (Driscoll, 1997: 1;
Marshall, 2008: 2; McQuillan & Montgomery, 1999: xv; Woods, 2006: 8).
In the case of the World Bank, the confusion is compounded by its sheer size
and complexity of organisation, which means that “[n]‌o one person sitting
in the headquarters building or working elsewhere grasps all the threads
of activities bearing the label of World Bank” (Marshall, 2008: 2). The
IMF’s “long-​standing practice of not releasing internal documents to the
public” (McQuillan & Montgomery, 1999: xv) has not helped any claim it
may have to be a transparent organisation. However, the World Bank and
IMF (the “Bank” and the “Fund”) have always collaborated closely in a
DOI: 10.4324/9781003302636-10
 Large International Organisation Case Study: The World Bank 183

“complex but evolving fashion, sometimes closely in harmony, sometimes

in tension as their styles and approaches are quite different” (Marshall,
2008: 1). Today they share the common goal of raising living standards in
their member countries (IMF, 2022b) and their “highly innovative vision
for international policy coordination” (Helleiner, 2017: 27) is considered
by many to be “supportive of the economic development of poorer coun-
tries” (2017: 27).
The Bank and the Fund are not private organisations –​they are inter-​
governmental institutions “with public purposes” (Woods, 2006: 9) and deal
with member countries and their governments, rather than individuals or
commercial organisations. They were created at the same time, following
the Bretton Woods conference of world leaders in New Hampshire, US,
in 1944. As a result, they share headquarters in Washington DC and are
often collectively referred to as the “Bretton Woods institutions” (Maverick,
2021: 29), “Bretton Woods siblings” (Gutner, 2023: 965) or “Bretton Woods
twins” (Marshall, 2008: 1). The architects of this conference were Maynard
Keynes in the US and Harry Dexter White in the UK (Boughton, 2014: 10).
They envisaged that, after World War II, a major international collaborative
approach in financial and economic matters would be needed if growth was
to be engendered in a war-​ravaged world. They were thus hoping to avoid
another Great Depression and maybe even another war. After many diffi-
cult post-​war negotiations, the International Bank for Reconstruction and
Development (IBRD) and the International Monetary Fund were created.
The IBRD1 was intended to “restore economic activity, while creation of an
International Monetary Fund would help restore currency convertibility and
multilateral trade” (Boughton, 2014: 7). The success of this conference has
been described in the literature as a “political miracle” (Woods, 2006: 9),
given the lack of international precedent on the same scale; the creation of
such institutions was totally uncharted territory. Both the Bank and the Fund
would have a great deal of funds at their disposal, as they would be auto-
matically financed by their members and would “earn income from their
lending and investments.” However, no instruction manual or set of criteria
existed as to how they “might practically achieve the broad objectives set for
them” (Woods, 2006: 39), so the road ahead might not have seemed entirely
straightforward at that time.
In any case, the two institutions have remained in close and regular con-
tact over the decades. Despite both being financial institutions, their tasks
are distinct: The Bank’s task is “to promote development” whereas the Fund
is in charge of maintaining “order in the international monetary system”
(Harper, 1998: 94). At present, the Bank lends money only to “governments
of middle-​income and creditworthy low-​income countries” (World Bank,
2003: 4) whereas the Fund lends “money only to member countries with
payments problems, that is, to countries that do not take in enough foreign
currency to pay for what they buy from other countries” (Driscoll, 1997: 17)
but only on condition that “they undertake economic reforms to eliminate
184 Terminology Management for Translators

these difficulties for their own good and that of the entire membership”
(Driscoll, 1997). The “explicit commitment that members make to imple­
ment remedial measures in return for the Fund’s support is known as condi-
tionality” (McQuillan & Montgomery, 1999: 68), an important term in the
Fund’s vocabulary.
The literature on the Bank and the Fund suggests that they have been both
loved and feared in equal measure. On one hand, they exist “in large part to
go where angels fear to tread” (Woods, 2006: 9) and, on the other, “many
have a critical view of its policies and operations” (Marshall, 2008: 7). Given
that they tackle “some of the world’s hardest issues” (Marshall, 2008: 12)
and despite the fact that there is a gap between what they “attempt to achieve
and what their record shows they can deliver” (Woods, 2006: 9), if we cor­
relate economic growth and monetary stability, it would seem that “Bretton
Woods was successful beyond all expectations” (Van Dormael, 1978: ix).
The Bank and the Fund are two of the 17 “specialized agencies”2 of the
United Nations (UN) but they are not beholden to the UN for approval
of their policies and actions and above all for financial support (Marshall,
2008: 35). They have developed close links with many other UN institutions,
e.g. the International Labour Organization (ILO) or UNICEF (United Nations
International Children’s Emergency Fund). The most obvious differences
between the Bank and the Fund are “in size and culture” (Woods, 2006: 6)
as we saw above. In terms of size, whereas the Fund employs some 3,000
staff, the Bank employs over 13,000 as of 2023 (World Bank, 2023b: 81)
and has “decentralized some of its operations to the field” (Woods, 2006: 7).
That said, they have almost the same membership, likely due to the fact that,
in order to become a member country of the Bank, “a country must first join
the International Monetary Fund” (World Bank, 2003: 72). As of 2024, the
Bank has 189 member countries (World Bank, 2024) and the Fund has 190
(IMF, 2022a). The structure and governance of each institution is also very
similar, and a brief overview of the organisation and workflow at the World
Bank is provided in the next section.

Organisation and Workflow at the World Bank (“the Bank”)

The World Bank consists of a “complex set of multinational, multisectoral,
multidisciplinary institutions that are involved in almost every aspect of
development –​which means, in practice, virtually every issue imaginable”
(Marshall, 2008: 2). What we know as “World Bank” consists in fact of five,
separate but interrelated, institutions referred to as the “World Bank Group”
(see Table 8.1), all of which have the overarching aim of eradicating poverty
around the world (Van den Bossche, 2007: 150). The first two institutions,
the IBRD and the IDA, comprise a subgroup which is commonly known as
the “World Bank” (Hayes, 2022).
The World Bank is very much a “presidential institution, with the presi-
dent exercising substantial authority by mandate (…) and by setting a tone”
 Large International Organisation Case Study: The World Bank 185

Table 8.1 The five institutions that constitute the World Bank Group. The IBRD
and IDA are the ones that comprise what is commonly known as the
“World Bank”

World Bank Group: Institutions

Acronym Name Year of Purpose

creation

IBRD International Bank for 1945 Lend money to

Reconstruction and governments of
Development middle-​income and
(“World Bank” together creditworthy low-​
with IDA) income countries.
IDA International Development 1960 Provide interest-​free
Association loans, called credits,
(“World Bank” together to governments of the
with IBRD) poorest countries.
IFC International Finance 1956 Lend directly to the private
Corporation sector in developing
countries.
MIGA Multilateral Investment 1988 Provide guarantees to
Guarantee Agency investors in developing
countries against
losses caused by non-​
commercial risks.
ICSID International Centre for 1966 Provide international
Settlement of Investment facilities for conciliation
Disputes and arbitration of
investment disputes.

(Marshall, 2008: 77). The 189 member countries are also its shareholders
and are represented by a Board of Governors, “usually the finance minister
or central bank governor of each member country” (George & Sabelli,
1994: 207), who are “the ultimate policymakers at the World Bank” (World
Bank, 2024). The Bank’s governors meet governors of the International
Monetary Fund’s Board of Governors once a year at a joint meeting called
the Annual Meetings (2024). The International Monetary Fund’s Managing
Director and the Bank’s President also “meet regularly to consult on major
issues (IMF, 2022b). For more about the structural organisation of the Bank
under the Board of Governors, see the World Bank webpage (World Bank,
2024).
There is one linguistic unit at the Bank which includes both translation
and interpreting and provides services to all institution members of the World
Bank Group (IBRD, IDA, IFC, MIGA and ICSID), as well as the Bank’s
numerous funds and initiatives such as GEF (Global Environment Facility),
GPE (Global Partnership for Education) and CGAP (Consultative Group
to Assist the Poor). Its full name is Global Corporate Solutions Translation
186 Terminology Management for Translators

and Interpretation unit, or GCSTI for short. Translation is funded through a

chargeback system, where the costs of the service are applied to the respective
client. Having said that, it is not compulsory for the Bank’s institutions to use
the Bank’s GCSTI unit; they are free to seek translation services elsewhere.
This section will focus on the Translation section of the GCSTI unit, where
the main roles are Translators/​Revisers and editors. Although there used to
be a Terminologist role, it has been obsolete for the past two decades or so.
At present, there is a more decentralised system whereby Translators/​Revisers
are the ones carrying out any terminological work required.
It may be of interest to note that the Bank has never adopted a common
language and translation policy, preferring instead to retain a pragmatic and
flexible approach that reflects its changing needs. In other words, transla-
tion is not compulsory at the Bank. It is not compulsory to translate every-
thing, but a great deal of translation is done nonetheless. Given the size of
the Bank’s operations and worldwide reach, it would be very difficult to
enforce a compulsory translation policy. The main language combinations
are English as a source language into Arabic, Chinese, French, Russian
and Spanish. Sometimes translation work is done in other languages, e.g.
Portuguese. Languages are not referred to as “official” but as “working”
languages. However, much of the Bank’s business takes place in English.
The GSCTI unit is relatively very small these days, comprising 15
linguists (13 translators and 2 editors). As a general rule, there are two or
three translators per language. It should not come as a surprise, therefore,
that a great deal of translation work at the Bank is outsourced. The Bank’s
established practice is to prioritise engaging its GCSTI unit first, prior to
considering outsourcing. Any translation work that cannot be done in-​
house must be outsourced, as must all translations in languages other than
the Bank’s working languages. The extent to which translation work is
outsourced is also contingent on the demand for translations, with the exact
figure of outsourced translation work varying considerably from year to year.
For example, in 2021, GCSTI outsourced over 70% of the total amount of
its translation work. However, in 2023, this percentage dropped to less than
60%. On average, as we will see later on, the Bank outsources around 60–​
70%, sometimes more, of its translation work.
Translators work in groups based on language. Each language team has
a “team leader” who is usually the translator with the most experience. All
translators are answerable to their team leader. Occasionally, translators
will also carry out some interpreting work, but this is largely left to the
Interpreting section of the unit, although almost all interpreters are external
providers. As well as in Washington DC (see Figure 8.1), translators/​revisers
may be based in offices around the world, i.e. in Cairo for Arabic, in Beijing
for Chinese, in Yaoundé (Cameroon) for French, in Sofia (Bulgaria) and
Almaty (Kazakhstan) for Russian (since the war with Ukraine) and in Buenos
Aires for Spanish.
 Large International Organisation Case Study: The World Bank 187

Figure 8.1 The World Bank building in I Street, Washington DC.

The work of translators is complemented by that of editors, whose primary

role is to provide an additional layer of quality assurance, after translators
have completed their work. Editors ensure that translations destined for the
public domain are completed to the highest possible standards. Also, any
translations initially completed by freelancers are also reviewed by editors
for accuracy and quality. In recent times, a new trend has emerged whereby
editors are increasingly undertaking translation tasks to help reduce the cost
of outsourced work.
The fact that the Bank’s GCSTI unit is relatively small compared to those
of other organisations in the UN network also limits the feasibility of trans-
lator specialisation. While there are instances where translators may prefer
working on a specific type of document, the Bank’s translators have to be ver-
satile and capable of working across various domains. In cases where a trans-
lation job requires specialised knowledge or expertise not available among
the in-​house translators, the job is typically outsourced to external vendors.
Translators at both the Bank and the Fund experience two very busy
periods every year, always coinciding with the Spring meetings and Annual
meetings. A great deal of work also arises from the Director’s meetings and
other massive events online. Much of the translation and revision work at
188 Terminology Management for Translators

the Bank is outsourced to external providers, referred to as “vendors,” but

the clients are all internal, from the various institutions and departments
of the Bank.
When a client has a request, they log it in through an institutional portal
called e-​Services. Requests are handled by Project Coordinators, who assign
jobs to either internal translators or external vendors as required. A simple
illustration of the generic workflow at the Bank is shown in Figure 8.2.
Any of the World Bank Group institutions (IBRD, IDA, IFC, MIGA,
ICSID) or other departments (e.g. Communications team) or partners (e.g.
GEF) can submit a translation request through e-​Services. After consult-
ation with their respective team leaders, Project Coordinators assign jobs to
linguists, either in-​house translators or external vendors, and they also carry
out all administrative tasks, such as invoicing, pre-​processing of files and
client management tasks. They liaise with external vendors directly when
DTP tasks are required –​DTP tends to be outsourced rather than done in-​
house. Project Coordinators are also in charge of delivering the translation
to the client –​as a general rule, neither translators nor Editors liaise directly
with the clients.
Translations are either fully edited (e.g. in the case of web documents, or
documents to be published) or lightly proofread. Translations of other types of
documents are not revised or edited if they are done by an already experienced
vendor, referred to as “tier 1,” or when time constraints do not allow it. The
translations done by vendors are only revised by in-​house translators where
there is a compelling need to do so, or if the client requires it. If translations
need to be edited, Project Coordinators liaise directly with the editors, who

Figure 8.2 Generic translation workflow at the World Bank Group.

 Large International Organisation Case Study: The World Bank 189

are responsible for the final version of the translations. However, editors can
also provide feedback to translators or vendors. Translators may use MT
as part of their work process. The GCSTI unit offers free raw MT output
and paid light post-​ editing services as new business lines for internal use
only. As well as the translation, DTP and light post-​editing work carried out
by vendors, the Bank has outsourced most of its term extraction work to
another external vendor, the Terminology and Linguistic Research Group.
Their work and the terminological work of translators and editors at the
Bank is explored further in the next section.

How Do Translators Manage Terminology at “the Bank”?

Since there is no dedicated terminologist role at the Bank, translators carry
out all of the internal terminological work themselves (see also Box 8.1).
The MultiTerm termbases at the Bank are organised by domain, e.g. educa-
tion, environment, health, etc. To get a sense of the range of domains and
subdomains in use at the Bank, it is worth having a look at its Projects By
Theme webpage.3
If clients provide a job-​specific list of terms, it is reviewed and imported
into the team’s MultiTerm termbases and activated alongside these termbases
during translation. Translators also have access to WB-​TERM, the Bank’s
terminology portal, which at the time of writing contains some 28,500 terms
and over 3,200 acronyms. WB-​TERM is for internal use only for now, with
plans to make it public in the near future. The general policy is that all terms
should be in at least the six working languages of the Bank. The source
language of the termbases is usually English and the most frequent target
languages are the other working languages of the Bank, i.e. Arabic, Chinese,
French, Portuguese, Russian and Spanish.

Box 8.1 World Bank Group –​Translation Facts and Figures

Location: Washington DC Own MT? WB-​Translate

No. of translators: 15 approx. Project Management tool: e-​Services
Terminologists: No Specialism required/​encouraged: No
CAT tool: RWS Trados Studio Outsourced work: 60–​70% on
Termbase: RWS MultiTerm, average
WB-​TERM Main source language: English

Translators’ main approach to feeding the termbases is to extract terms

from a translation after the translation is completed, either manually or
aided with technological tools such as MultiTerm Extract, Sketch Engine or
mAI, the Bank’s own GPT platform (for more details about this platform see
Richards (2023). This ensures that only approved and validated terms find
190 Terminology Management for Translators

their way into the termbase. The extracted terms are first shared on an online
Excel spreadsheet (e.g. on platforms such as OneDrive) so that all translators
and editors can contribute to the terminological discussions. Once agreement
is reached, the terms are imported to a MultiTerm termbase, which will be
available for use for the next translation job.
There is also a large institutional investment in multilingual term extrac-
tion, that will feed a large volume of vetted terms into WB-TERM, which
in turn will be used to train the Bank’s MT engine (see below). Since it has
been recognised at the Bank that terminology extraction is a rather time-​
consuming task and given the lack of dedicated terminologists, this massive
term extraction endeavour has been mostly outsourced to an external vendor
based in Argentina. The vendor consists of a team of terminologists referred
to as the Terminology and Linguistic Research Group (see Figure 8.2). Their
mandate is to use the term extractor to extract about 1,000 terms per month
for the next few years from all of the Bank’s existing multilingual documents,
but it is also likely that monolingual documents will be used. The task of
deciding which terms to extract from the monolingual documents is being
carried out separately within each language team.
When a translation is completed, each language team validates the
translations of the terms. Translators and editors can work together on the
revision of the terms and their translations, and one designated member of
staff in CGSTI. WB-​TERM can be used by any employee of the Bank, not
just translators, and it is integrated into Trados Studio. Translators have
access to their respective termbase settings and can update terms or add new
ones. They can also edit the source term, should they find an error in it, for
example. Vendors can use Trados Studio via the GroupShare server.
The terms stored in the WB-​TERM records come with little associated
information, but there is a compelling reason for this. Apart from the terms
themselves, the descriptive fields include only Subject and Source, with each
field at a different hierarchical level, as shown schematically in Figure 8.3. On

Figure 8.3 Descriptive fields and hierarchy in WB-​TERM records (left) and a sche-
matic example of a term record (right).
 Large International Organisation Case Study: The World Bank 191

the left, the descriptive fields of an empty WB-​TERM record are illustrated;
on the right, an example of how the information is typically organised and
displayed in a WB-​TERM record.
The reason for limiting the WB-​TERM records to two descriptive fields,
alongside terms, is mostly practical. When highly specialised terminological
information is needed, or when the information required is not stored in
WB-​TERM, the Bank’s translators often retrieve this information from the
termbases of partner organisations from the UN network. For example,
translators may find terminological information on foodstuffs from the
FAOTERM termbase, the largest terminology database of the Food and
Agriculture Organization, or more general (or indeed specialised) informa-
tion from UNTERM, the multilingual termbase of the United Nations. In
this way, the Bank’s translators avoid duplicating time-​consuming efforts to
create highly specialised termbases and contribute to terminological consist-
ency across the network of UN organisations, while not missing out on this
crucial information. WB-​TERM fits in by storing World Bank-​specific ter-
minology (which occasionally differs from that of the UN organisations),
and any other terminological information required for the daily work of the
Bank’s translators.
Figure 8.3 can also be used to highlight the importance of the source and
domain of a term. When translators have to work with a streamlined version
of a termbase, the descriptive fields that cannot be left out are source and
domain at the very least. As we have seen in previous chapters, the source
is crucial for the terminological work of translators, as it provides proven-
ance and can confirm the reliability and location of the source. The domain
(or Subject in the WB-​TERM record) is also crucial to place the term in the
correct context when translating. In the case of the World Bank, though,
translators are not deprived of any additional information about terms –​they
simply retrieve it from the relevant termbases of partner organisations.
Picklists are used in WB-​TERM for the Subject field. As we can see in
Figure 8.3, Organizational structure is one of the picklist values. The others,
which give us a sense of how the vast range of domains covered by the Bank
are grouped together in WB-​TERM, are the following:

• General
• Economics
• Finance
• Environment and Natural Resources Management
• Human Development and Gender
• Private Sector Development
• Public Sector Management
• Social Development and Protection
• Urban and Rural Development
• Agriculture, Fishing, and Forestry
• Education
192 Terminology Management for Translators

• Energy and Extractives

• Health
• Industry, Trade, and Services
• Information and Communications Technologies
• Transportation
• Water, Sanitation, and Waste Management
• Organizational Structure

As with terminology, phraseology is an important part of the translator’s

work at the Bank. Since many of the texts that require translation may be
from general language, i.e. with few, if any, occurrences of specialised or
highly specialised terms, translators have observed over time that there
is a recurrent number of phrases and expressions, which are gradually
evolving into preferences by the World Bank’s authors, speakers and writers.
Therefore, the Arabic team is currently leading efforts to collect this phraseo-
logical information, which is being collected in spreadsheets (see Figure 8.4).
The purpose of these spreadsheets is twofold.
On the one hand, phraseological information will be used to maintain
phraseological consistency alongside terminological consistency during
translation and, on the other hand, it will be used to train the Bank’s MT
system. Figure 8.4 shows a mixed bag of expressions, ranging from idioms
(e.g. put all eggs in one basket) and collocations (e.g. highly vulnerable to) to
standalone phrases (e.g. at regular intervals) and portions of sentences (e.g.

1 the ultimate global crisis lender ‫المقرض العالمي األخير خالل األزمات‬
2 keeping inflation under control ‫مواصلة السيطرة على التضخم‬
3 highly vulnerable to ‫معرضة بدرجة كبيرة ل‬
4 reaching net zero emissions ‫الوصول بصافي االنبعاثات إلى مستوى الصفر‬
5 orderly transition to net zero by 2050 ‫االنتقال المنظم إلى انبعاثات كربونية صفرية صافية‬
6 on a steady downard path ‫على مسار تنازلي ثابت‬
7 vulnerable households ‫األسر المعرضة للمخاطر‬
8 keeping inflation low ‫إبقاء التضخم في مستويات منخفضة‬
9 goal of declining external debt ‫هدف تخفيض الدين الخارجي‬
10 at regular intervals ‫على فترات منتظمة‬
11 In a low-​inflation environment ،‫وعند تدني مستويات التضخم‬
12 keeping inflation under control ‫مواصلة السيطرة على التضخم‬
13 working at cross-​purposes ‫يُنظر إليه باعتباره مخالفا ألغراض‬
14 a year marked by unprecedented ‫عام اتسم بتحديات عالمية غير مسبوقة‬
global ch
15 put all eggs in one basket ‫تعليق جميع آمالنا على سيناريو واحد‬
16 Going forward, ،‫في الفترة القادمة‬

Figure 8.4 An excerpt of the Arabic team’s phraseological database at the Bank.
 Large International Organisation Case Study: The World Bank 193

Going forward,). As we saw in Chapter 2 (Terms and Words section), it is

conceptually attractive to keep such phrases in a separate database, at least
for now, as they do not belong in a termbase because they are not terms and
they do not belong in a translation memory because they are not sentences
(although phrases may appear in translation memories, just as terms do). In
the case of the Bank, whose GCSTI unit also carries out interpreting work,
a special treatment of phraseology will also help interpreters maintain the
speakers’ idiosyncrasies across languages.
On the topic of MT, as well as WB-​TERM, translators have access to
the Bank’s own MT system, WB-​Translate, trained on the Bank’s own
documents. WB-​ Translate is provided by Microsoft and integrated into
Trados Studio. It is available internally with plans to make it available to the
public in future. Vendors have access to WB-​TERM but not WB-​Translate.
As with WB-​TERM, WB-​Translate can be used by all Bank employees, not
just translators.
Other tools available to translators at the Bank include mAI which is
powered by OpenAI GPT. mAI is undergoing a test phase at the Bank and
the possibilities afforded by ChatGPT are currently being explored. Although
little testing has been done so far (e.g. Richards, 2023), the general consensus
among the linguists interviewed is that using these tools for term extraction
seems to be promising.

Outlook

World Bank

There are many and frequent collaborations between the Bank and
the International Monetary Fund and the same can be said about their
translators. Translators from both organisations often know each other and
may meet to discuss best practices and share resources, e.g. translation mem-
ories (although the Fund still has some unresolved legal issues in this respect).
Both the Bank and the Fund continue to work closely with the World Trade
Organization and the three organisations continue to work together towards
the common aim to reduce global poverty, improve trade across nations and
generally increase living standards around the world.
A major overhaul of the terminological resources of the Bank is underway,
which should yield some interesting termbase and MT resources, some
of which may become publicly available in future. This overhaul is being
implemented through a major term extraction project outsourced to external
providers. By contrast, the Fund has invested in its own internal Terminology
Working Group and Language Technology Team, so the bulk of terminology
management tasks are carried out internally. This seems to be a stark con-
trast between two “sister” organisations but the difference may lie in the fact
that specialisms are more specific and tailored to the Fund’s own activities
and research, whereas the Bank is far too big an institution covering too
many domains and subdomains to require a single specialism of translators.
194 Terminology Management for Translators

Since many economists and financial experts work at the Fund, alongside
translators, many opportunities can arise for authors of documents and sub-
ject specialists to collaborate with translators, making this an ideal workplace
for further research into expert-​translator interactions, e.g. in the drafting of
definitions, consulting specific expert knowledge, etc.

Follow-​up Tasks and Reflection

1. Both organisations explored in Chapter 7 and Chapter 8 are still using
Excel spreadsheets. Why do you think this is? What are the benefits of
using them when you already have a sophisticated termbase such as
MultiTerm?
2. An important terminological issue for translators is that of the name of
the organisations. The World Bank is often referred to by its employees as
“the Bank,” and the IMF is often referred to as “the Fund,” but the World
Trade Organization is simply referred to as “the WTO” (as opposed to “the
Organisation” if we followed the Bank’s and the Fund’s model of abbrevi-
ation). You would definitely need to know this information regarding the
shortening of their names for translation purposes, but how would you
store it in a termbase, e.g. in the Notes, Context, Example and/​or other
fields? Are these abbreviations the same in your other language(s)? How
do other languages deal with them?
3. In English, the fact that these organisations are referred to as “Bank” and
“Fund” has repercussions for neologisms and term formation, since new
terms have been coined for their specific ways of dealing with language,
i.e. Fundese and Bankese (otherwise Fundese might have been IMFese!).
How would you deal with Fundese and Bankese in your other languages?
Do they have similar term formation mechanisms?
4. In the domain of economics and finance, many new neologisms are con-
stantly coined through metaphors which refer to already known concepts
and ideas. This means that translators might need to create new terms
in other languages too when new concepts arise. Such was the case of
terms like soft landing and hard landing when they first appeared, which
were borrowed by economics and finance from the domain of aviation.
Browse through the Fund’s IMF TermBase4 and find how these and
other metaphors (e.g. Asian noodle bowl or spaghetti bowl) have been
translated into other languages. Have other languages borrowed also from
the domain of aviation, food, etc.?
5. A more recent neologism is resilience in the IMF TermBase. Have a look
at the various translations of resilience into other languages in the IMF
TermBase and discuss to what extent the English term is influencing the
terms that are being used for this concept in other languages.
6. Many terms from economic and financial domains may be difficult to
identify as such because they seem to belong to general language at first
glance. One such example is global imbalances, since global might initially
 Large International Organisation Case Study: The World Bank 195

be considered a modifier (see Chapter 5). Find other terms in the IMF
Termbase, in English or other languages, that would also seem to be words
or expressions from general language but have acquired a specialised
meaning.

Notes
1 Today, the IBRD and the IDA (International Development Association, see
Table 8.1) are the institutions jointly referred to as the “World Bank” (or the
“Bank”).
2 UN specialised agencies are “international organizations working with the UN, in
accordance with relationship agreements between each organization and the UN”
(United Nations, 2024).
3 https://​ p roje​ c ts.worldb​ a nk.org/​ e n/​ p roje​ c ts-​ o pe ​ r ati ​ o ns/ ​ p roj ​ e ct- ​ t heme?lang=​
en&page=​
4 www.imf.org/​en/​About/​Term​inol​ogy/​imf-​termb​ase
5 “LETRINT” stands for “Legal Translation in International Institutional
Settings: Scope, Strategies and Quality Markers.” It was a research project led by
F. Prieto Ramos and supported by the Swiss National Science Foundation through
a Consolidator Grant –​see Prieto Ramos et al. (2019).

Further Reading

The World Bank, the International Monetary Fund and Other

Financial Organisations
The World Bank and the International Monetary Fund have elicited a great deal of lit-
erature over the years but most of it is on business, economic and financial aspects of
their operations. However, Marshall (2008) is one of the few “nuts and bolts” guides
about the World Bank that should provide informative background information for
translators. For more about the organisation and history of the Bank, there is a wealth
of information on its website (World Bank, 2023a, 2023c). For more about the five
institutions that make up the World Bank Group, it is worth referring to a book
published by the Bank itself (World Bank, 2003). Another World Bank publication
expands on the differences between the Bank and the Fund (World Bank, 2011: 29ff).
Woods (2006) is a very accessible joint overview of both institutions. Buira (2005)
takes stock of sixty years’ worth of work in these organisations.
Beyond the Bank, the Fund and the WTO, several publications in Translation
Studies have focused on a range of translation and terminology aspects of the work
of large multilingual organisations, often focusing on the EU institutions, but I am
not aware of any other workflow-​based studies such as the one adopted in this book.
General approaches include Joscelyne (2000), Pym (2004) and Wagner et al. (2012).
From a legal translation perspective, many useful contributions can be found in Prieto
Ramos’ edited volume (2018) which together provide a thorough overview of “insti­
tutional translation issues related to the development of international law and pol-
icies.” Other very useful edited volumes are those by Borja Albi and Prieto Ramos
(2013), Faini (2017), Olsen et al. (2009), and Svoboda et al. (2023).
196 Terminology Management for Translators

Under the auspices of the LETRINT project,5 Prieto Ramos has investigated exten­
sively legal translation in large multilingual organisations. In Prieto Ramos (2020a),
he explores how the legal functions of large organisations have an impact on transla-
tion; in Prieto Ramos (2020c) he looks at how organisations such as the World Trade
Organization, the United Nations and four EU institutions can reach terminological
consistency in legal translation, and in Prieto Ramos (2020b) he reviews the process
of introducing corrections to translated documents in the same organisations.
Other studies of legal terminology in large multilingual organisations include Best
(2017) on untranslatability of legal terms, Felici (2010) on legal issues of multi­­
lingualism, Floros (2017) on legal terminology in Cyprus, and Prieto Ramos and
Guzmán (2021) on a comparative analysis of multilingual texts, Chirimbu and Barbu-­
Chirimbu (2014) on Romanian terminology, Francuski (2017) on Serbian termin­­
ology, Våge (2018) on users of terminology in Norway and Bendegard (2012) on
plain language in Swedish texts.
Several publications exist on the terminological and translation practices at the
Directorate-​ General for Translation (DGT) of the European Commission, e.g.
Stefaniak (2017, 2022), Maslias (2017, 2022), Bardarska (2009), Brownlie (2017),
Valero Garcés (2018) and Fernández-​Parra (2021). Beyond the EU but still within
Europe, other studies have focused on the European Central Bank, including Tcaciuc
and Mackevic (2017a, 2017b) and Šarčević and Robinson (2013) on economic trans­­­
lation and Schäffner et al. (2014) who also look at organisations such as Amnesty
International and the German Foreign Office.
Elsewhere, studies on the United Nations include Pym (2004), Šarčević (2018), who
also examines the Council of Europe and the European Cort of Human Rights, Zhao
and Cao (2013) looking at legal translation at the United Nations, Mulders (2008),
who looks at the Court of Justice of the European communities, De Saint-​Robert
(2008), who focuses on the Languages Service of the United Nations and Tomić and
Beltrán (2013) who look at translation at the International Criminal Court. Prieto
Ramos and Guzmán (2023) explore institutional translation profiles at WIPO (the
World Intellectual Property Organization) and the World Trade Organization in com-
parison to EU institutions.
Joscelyne’s work (2000) focuses on the OECD (Organisation for Economic Co-​
operation and Development), another inter-​ governmental organisation, whereas
Fontanet (2018) takes us to the European Council for Nuclear Research (CERN from
its French name) and Millet (2013) takes us to INTERPOL.

Other Institutional Case Studies

A number of case studies exist in the literature focusing on specific aspects of transla-
tion in governmental settings. Many of these are case studies from European settings.
For example, De Camillis (2021) looks at how civil servants may act as “occasional
translators” in the Italian region of South Tyrol, where Italian and German may be
used in interactions with the government. In the same volume, there are case studies
about Italian as an official minority language within the Swiss multilingual institu-
tional system (Canavese, 2021), about the translation and interpreting routines in
institutional settings in Slovakia (Štefková & Tužinská, 2021), and about policies
and practices in administrative interactions in the Dutch language area of Belgium
(Bernaerts, 2021). In another edited collection, Garrido Nombela (2013) looks at
 Large International Organisation Case Study: The World Bank 197

(especially diplomatic) translation in the Spanish Ministry of Foreign Affairs and

Hickey (2013) examines what happens with the translation of English letters of
request in the police and legal settings. Elsewhere, the work of Heylen and colleagues
shed light on how a specific CAT tool, TermWise, can be used to support the work
of legal translators at the Belgian Federal Ministry of Justice (Heylen et al., 2014a;
Heylen et al., 2014b; Heylen & De Hertog, 2015). The workings of translation at
the German Foreign Office is examined by Schäffner et al. (2014), who also look at
the translation department of the European Central Bank and the non-​governmental
organisation Amnesty International.
Case studies outside Europe and Canada do not abound but we can mention a few
here, some of which are quite narrow in focus. In Africa, Alberts (2000) explored how
terminology works at the National Language Service in the Republic of South Africa.
Despite its age, it is a useful reference article that can be contrasted with a more up-​
to-​date look at the current offerings of the National Language Service in the official
website (Gaffane, 2020). Still in the South African context, Beukes (2009) examines
the possibilities of using translation not only as a language mediation tool but also as
a language development tool for indigenous (minority) languages.
As for Asia, Hung surveys two millennia of translation activities in China, one cat-
egory of which is government translation, and shows how this kind of translation was
aimed at “strengthening the ruler’s prestige and effectiveness” (2005: 67). Mottura
(2006) shows the lexical changes that have occurred in Chinese public administra­
tion since 1978. Using a Korean-​English parallel corpus methodology, Choi (2022)
devotes an entire book to look at translation processes and products in South Korean
government institutions –​a very interesting case study if we assume that South Korea
is a monolingual society.
In South America, Díaz Vázquez and Guerra Aranguren (2019) explain how the
Colombian government channelled its efforts to internationalise its economy by
improving and updating its customs regulations. Their study shows how termino-
logical objectivity and conciseness, among other features and complexities, are now
reflected in the new regulations. In Australia, O’Mara and Carey (2019) suggest
how information technology can be best used in government translation in order to
support migrants and refugees.
9 Case Studies on Commercial
Organisations

Key Highlights

• Two examples of how terminology is managed in language services

providers:
• the Kaleidoscope Group providing software solutions
(Kaleidoscope branding) alongside translation services (Eurocom
branding); and
• Terra Translations as an online localisation company.
• One example of how terminology is managed in a non-​language ser-
vice provider which employs linguists in-​house: Blum, as a client of
the Kaleidoscope Group.

Translation and Terminology at LSP1: The Kaleidoscope Group and its

Client Blum

Background of the Company: The Kaleidoscope Group

The Kaleidoscope Group is an Austrian language technology company

offering a wide range of services not only in translation and terminology but
also in software localisation, SEO, machine translation post-​editing, quality
management and documentation. It was founded in 1996 by its current co-​
CEO, Klaus Fleischmann (with Anita Wilson). The original idea for the com-
pany was to develop software only but services were added over time, such as
content creation, technical writing, automation and consultancy.

DOI: 10.4324/9781003302636-11
 Case Studies on Commercial Organisations 199

Box 9.1 Key Highlights: The Kaleidoscope Group

• Terminological software developed and maintained in-​house.

• Terminologist and tools expert in Kaleidoscope, translators and pro-
ject managers in Eurocom UK.
• Translators’ access to subject matter experts in client’s company.
• Central figure of the project manager in client–​ translators inter­
actions.
• Experienced and/​or highly specialised freelancers highly valued.
• Terminological services and ‘term rate’ offered.
• Revision always in-​house.
• Query management software increases three-​ way collaboration
(translators/​project managers/​clients or subject matter experts) and
discussion in translation projects.
• Responsibility for terminology across the entire company, including
non-​linguistic staff, not only translators or revisers.
• Terminological consistency begins with the source text(s).

Today, the Kaleidoscope Group is made up of two divisions, each with

separate branding: Kaleidoscope and Eurocom. Kaleidoscope is the profes-
sional services side of the company, offering software solutions, whereas
Eurocom is the language services division, essentially functioning as a trans-
lation agency. Eurocom has a subsidiary, Eurocom UK, bought in 2016 from
a company called Bridge, whose CEO had worked at Siemens, which goes
some way to explain the current focus of Eurocom UK on technical transla-
tion. While Kaleidoscope and Eurocom are separate legal entities with their
own branding, in practice there are no internal differences between staff from
each branding, who meet and work together as a single company.
Throughout its more than 25 years in the business (Kaleidoscope,
2023a), the Kaleidoscope Group has developed a number of partnerships
with key companies in the translation industry, including CAT tool com-
panies such as RWS Trados (of which Kaleidoscope is an official distributor
in the Austrian market), Phrase, memoQ and Smartling, technological
companies such as Quanos Content Solutions (formerly known as Schema)
and Acrolinx, and language consulting companies such as Berns Language
Consulting and Congree, to name but a few. In return for distributing
Trados in the Austrian market, RWS is also a “worldwide reseller” for
Kaleidoscope’s software solutions, which include an online content man-
agement platform called Kalcium, which has a number of interconnected
but standalone modules, such as Kalcium Quickterm (see also Chapter 6)
and Kalcium Checkterm.
Kalcium was conceived as a workflow solution for MultiTerm but evolved
over time to be a standalone program. It is also worth pointing out that
200 Terminology Management for Translators

Kalcium is not a CAT tool but a platform integrating a number of programs,

and that it does not feature any CAT tool. Instead, Kalcium Quickterm and
Kalcium Checkterm can be integrated into Trados or used as standalone
programs, as we will see in the following sections. The company has also
forged strong links with the Semantic Web Company, that has developed
the content creation software PoolParty. The Kaleidoscope Group regu-
larly offers and takes part in webinars, e.g. at GALA (the Globalization and
Localization Association) and other translation technology-​related events.
The above introduction goes to show the extent to which terminology and
translation are at the heart of the Kaleidoscope Group’s business.
In sum, the Kaleidoscope Group offers its own translation solutions and
products, as software solutions can be developed in-​house, as well as LSP
services. Through its consultancy work (Kaleidoscope division), the company
offers support and advice to clients wishing to use technological tools for
their business, such as how to create a termbase and prepare content to be
stored in it. This puts the Kaleidoscope Group in a unique position as a case
study for the terminology management research discussed in this book. In
the following sections, we explore the internal translation and terminology
workings of Eurocom in particular, the LSP side of the business, and we will
add the perspective of such workings from Blum, one of Eurocom’s clients.
As we progress through the following sections, we will also be unpacking the
key highlights shown in Box 9.1.

Background of the Client Company: Blum

Blum is an Austrian company that manufactures custom-​made furniture and

fittings, especially for kitchens, for the international market, and develops
software for furniture design. It was founded in 1952 by Julius Blum, a far-
rier and carriage smith by trade (Blum, n.d.). It is a family-​owned company
based in Höchst, Vorarlberg, Austria (see Figure 9.1) but today it also has
plants in Poland, USA, Brazil and China (Blum, 2023). Blum has over 9,000
employees worldwide, 6,800 of which are based in Vorarlberg. As of 2023,
its sales worldwide exceed 2,300 million euros (Blum, 2023). More facts and
figures about Blum can be obtained from its website (Blum, 2023) and so can
a more detailed timeline of its history (Blum, n.d.).
Its first product was a horseshoe stud (see Figure 9.2) which prevents horses
from slipping (Wikipedia, 2023a). Over its almost 70 years in business, Blum
has designed and manufactured several other kitchen furniture and fittings
and in 2013 its BLUEMOTION hinge (see Figure 9.2) “for the smooth
closing of furniture doors” won the European Inventor Award, awarded by
the European Patent Office (Advantage Austria, 2013).
Blum’s central website in English, mainly designed for the European
market, is fully localised by Eurocom into other languages and markets. The
Blum websites in each language can be accessed from the bottom of any of
the Blum’s English webpages by clicking on the Change market and language
button. Browsing through the Blum’s website, it becomes clear that one of
 Case Studies on Commercial Organisations 201

Figure 9.1 Blum’s plant in Höchst, Vorarlberg, Austria.

Source: From the Blum website.

Figure 9.2 Two of Blum’s products: Its first product, a horseshoe stud (left) and the
award-​winning BLUMOTION hinge for smooth opening and closing of
furniture (right).
Source: From the Blum website.

its main documents is its Catalogue, available in many of the languages of

the markets covered by Blum. Thus, the Catalogue is available in French (for
countries such as France and Luxembourg, for example), in Croatian for
Croatia or Portuguese for Portugal. However, some markets use the cata-
logue in English, rather than in their own language, e.g. Albania, Finland
and Iceland, for example, as the language of Blum’s website will depend on
how the local market is set up. A snapshot of one of the pages of the English
catalogue for 2022–23 is shown in Figure 9.3.
202 Terminology Management for Translators

Figure 9.3 A snapshot of page 38 of the 2022–23 Blum catalogue in English.

 Case Studies on Commercial Organisations 203

At the time of writing, Blum’s English catalogue consists of 754 pages,1

the vast majority of them very similar to the contents shown in Figure 9.3,
i.e. with sophisticated technical drawings alongside photographs, text and
figures in complex formatting, etc. The Blum catalogue showcases many of
the challenges that translators will encounter when translating or localising it
into other languages, not least rendering the specific Blum terminology (e.g.
power factor LF, TGR, SERVO-​DRIVE in uppercase), challenging termin-
ology (e.g. cross stabiliser cover cap oval) and brand names (BLUMOTION).
Other challenges include the correct localisation of figures, measurements,
etc. into the target language, formatting and fitting the text within the avail-
able space, to name a few. The translations of the catalogue are typically
updated with new material every two years or so. The revision of the cata-
logue is often a large undertaking, and it is often carried out on PDF format,
so that images can be seen alongside text in the way the readership will see
them. The catalogue is reviewed in full, not just the updated sections but, in
recent times, new features in the DTP software used allow the Blum DTP
team to mark new content as new. In the following sections, we investigate the
translation and terminology setup at Eurocom, as part of the Kaleidoscope
Group, and how that setup helps the company deal with translation requests
from Blum.

Organisation and Workflow at the Kaleidoscope Group

The Kaleidoscope Group is a B2B (‘business to business’) company which

means it does not accept translation jobs from individuals, only companies.
The main Kaleidoscope and Eurocom offices are based in Vienna, Austria
(see Figure 9.4). In this section, we focus on how translators at the Eurocom
division (the LSP part of the company) manage their terminology, although
we will occasionally refer to Kaleidoscope which is the side of the company
that develops and maintains the Kaleidoscope Group’s software solutions,
including the Kalcium platform, from which programs such as Kalcium
Quickterm and Kalcium Checkterm can be accessed, both by in-​house staff
and freelance translators.
The Eurocom team consists primarily of 20+​project managers (out of the
60 or so staff employed throughout the Kaleidoscope Group), supported by
four in-​house translators/​revisers at the UK subsidiary (see also Box 9.2).
Eurocom also works with many external translation agencies and freelance
translators. In fact, most of the translation work, i.e. at least 90% or more,
is outsourced. Typically, any translation project that has English as a source
language or any target language that is not English is outsourced. The in-​
house translators work mainly in the German to English language combin-
ation, although some jobs in this direction are also outsourced.
Eurocom can carry out translation work into almost any language, but of
course their biggest language combination by far is German to English. There
204 Terminology Management for Translators

Figure 9.4 The Kaleidoscope Group’s headquarters in Vienna, Austria.

are no designated terminologists at Eurocom but there is a Terminology

and Tools Expert at Kaleidoscope, who carries out a wide range of the typical
terminological tasks. Some of the Kaleidoscope Group’s clients employ
terminologists as well, who then liaise with the project managers at Eurocom,
as we will see with Eurocom’s client company Blum.
When a company wants to engage the translations services of Eurocom,
they will first open an account through the financial team, at which point
the rates for the required services are agreed and the general profile of the
client established. This means that, for every job commissioned, the client
knows in advance roughly how much a translation will cost. Quotes can
still be generated for a job, though, e.g. if the client requires them for their
own reference. When produced, quotes document the total price for the job
based on the agreed rates and any additional services required. On obtaining
a requested quote, clients can decide to go ahead with the translation job
or decline it. Establishing the financial terms in advance also means that
Eurocom project managers do not negotiate individual fees for transla-
tion jobs with the client, although they do negotiate deadlines and discuss
any other specifics of the translation commission. The standard prices at
Eurocom include translation and revision, with services such as DTP (either
“rough layout” or “print-​ready” DTP) quoted separately and carried out by
the Design and Graphics team of the company.
 Case Studies on Commercial Organisations 205

Box 9.2 Eurocom UK –​Translation Facts and Figures

Location: HQ in Vienna, Austria Own MT? No

No. of translators: 4 (in UK) Project Management tool: Worx until
Terminologists: No 2025 (new tool in development)
CAT tool: RWS Trados Specialism required/​encouraged: Yes
Studio, Across Outsourced work: 90% +​
Termbase: RWS MultiTerm, Main source language: German and
Kalcium English

Once financial arrangements are completed, project managers are free to

focus on all other project management tasks, such as liaising with clients,
in-​house translators/​revisers and/​or external vendors, dealing with clients’
and external vendors’ queries and complaints, preparing files for translation
and Trados project packages and splitting files among several translators in
large jobs, but they also keep clients up to date with the company’s practices
and news, and pass on feedback and queries from clients to translators and
vice versa. With this company structure, internal and external translators
can always have access to experts and/​or the authors of the ST, but always
through the project manager.
The typical translation workflow at Eurocom is illustrated in Figure 9.5,
with project managers at the centre of all interactions between in-​ house

Figure 9.5 Generic translation workflow at the Kaleidoscope Group.

206 Terminology Management for Translators

translator/​revisers, vendor agencies (Eurocom GmbH PMs) and freelancers

(Eurocom UK PMs), and clients. Already having an idea of how much a trans-
lation job might cost, the client contacts the project manager to commission
a job. The project managers will assign the job to either in-​house or external
vendors and prepare all the files for translation. The assigning of jobs not
only depends on the size of the job, the language combination and staff
workloads but also on the client’s preferred suppliers. Clients can specify a
preferred supplier at Eurocom (which can be a freelancer), e.g. a translator
with a legal background if much of the translation work is going to be in the
legal domain. Often, a second or even third preferred supplier is arranged
in advance too. With this setup, some freelance translators become “per-
manent freelancers” which means that, although they work on a freelance
basis, translation work comes in reasonably regularly.
Translators are certainly encouraged to offer their own specialisms, which
are often used to match clients to their “preferred” suppliers, either internal
or external translators. Translation in specialist subject areas is also charged
at a different, higher rate. Typically, translations in domains such as finance,
law or medicine command a higher rate of pay, and this also applies to
external translators. It is interesting to note that the company offers termino-
logical services too, which are charged at “term rate” rather than “word
rate.” In these jobs, the company quotes an “estimate” showing the likely
price for a job, which consists of an hourly rate based on the number of
terms involved. The accuracy of the quote is reviewed after a few hours and
adjusted as required. The estimate also takes into account the amount of
research needed to complete the job.
All translations are revised by a second linguist, including those carried
out by external vendors. Translators (including external ones) can discuss
the translation with a reviser, but the in-​house reviser will have the ultimate
responsibility for the translation. Most of the translation discussions take
place via comments in Trados Studio, which is often all that is needed.
However, in-​house translators also have a range of weekly, fortnightly and
monthly meetings where any interesting or important issues are discussed,
which may include feedback to and from a client, for example.
Once the translation is completed and revised, it is returned to the project
manager, who will carry out the final checks and QA tasks before delivering
to the client. If DTP services are required, the project manager will pass the
translation job to the in-​house DTP specialist before delivering to the client.
Translations with large DTP requirements may be outsourced to an external
DTP partner. Translation jobs are managed through the Worx project man-
agement software which is a separate program from the company’s platform
Kalcium. However, not only has Worx been made compatible with Kalcium,
but work is also in progress on developing an entirely new project manage-
ment system.
Post-​editing is offered as a managed service. This means that MT tools are
tested internally first, so that informed decisions can be made by Eurocom
 Case Studies on Commercial Organisations 207

project managers and Kaleidoscope staff as to which MT engine would be

the most appropriate for each job. Post-​editing may also be used in very spe-
cific, very urgent cases. For example, when a new call for tenders is issued,
Eurocom can translate (and, if requested, post-​edit) the text using a machine
translation tool which is suitable for the respective subject matter and lan-
guage pair. The client can then quickly get an idea of the requirements and
whether the project is of interest to them or not. If yes, Eurocom can then
offer a fully post-​edited or translated and revised version of the tender. This
gives the client significant cost and time savings, as they: a) do not have to
pay for full translation and revision of calls for tenders that later turn out to
not be of interest to them; and b) allows them to start working on tenders
straight away while waiting for the fully translated and reviewed version.
A second use case for post-​editing is where deadline constraints mean there
is not enough time to fully translate and review the text. Eurocom can offer
a post-​edited machine translation version as a stopgap until the final version
is ready.
Project managers handle all queries, whether from clients, in-​ house
translators or freelancers, using a query management program called Kalcium
Smartquery, from which project managers receive automatic notifications
(described further in the following sections). “Permanent freelancers”
and regular partner vendor agencies have permanent licenses for Kalcium
Smartquery and have access to change some of its settings. With the plug-​in
activated, internal or external translators can ask questions or request add-
itional information from project managers at the touch of a button. Project
managers may then pass on the request to clients where relevant and answer
all of the translators’ queries. In the following section, we look at the termin-
ology and translation processes from the point of view of the client company,
Blum, before we go back to Eurocom to see how they deal with translation
jobs from Blum.

Organisation and Workflow at Blum

In this section, “Blum” refers to “Blum Austria,” i.e. Blum’s plant in Austria,
which has a small linguistic unit comprising three translation managers,
one terminology manager and one in-​house translator/​reviser (who mainly
works from German into English). As well as translation, the translator’s
tasks include revising all the documents that require a high level of quality,
especially documents to be published or posted on the website, but also
important internal documents, such as some management documents. Since
the linguistic unit is small, there are many workflows possible but, typic-
ally, translation jobs are managed by translation managers, who decide if the
work goes to the in-​house translator or outsourced to other Blum branches,
e.g. Blum Ukraine, or to freelance translators, depending on the usual factors
such as translator capacity, specialisms, availability, etc. The translation
managers also outsource the localising of the Blum website (30+​languages)
208 Terminology Management for Translators

and catalogue (20+​languages) to Eurocom, which constitutes about 80% of

the translation tasks outsourced by Blum. In Figure 9.5 we saw the typical
workflow at Eurocom for dealing with translation requests from their clients,
including Blum.
Technical staff at Blum liaise with the linguistic unit and often contribute
corrections to the technical aspects of a text. As a general rule, the linguistic
unit never makes such corrections. The final responsibility for the content
of the texts lies with the technical staff who carry out the revisions, but
these revisions are not the core of their work at Blum. Since some of the
biggest translation jobs required by Blum include a good deal of external-​
facing materials, such as brochures, its website or its catalogue, DTP is an
important part of translation projects. Therefore, Blum has a large DTP team
and DTP tasks are carried out in-​house.
Blum’s terminology manager is tasked with maintaining and updating
Blum’s terminological resources, which includes a central termbase to which
all Blum staff can access (although not everyone does). The termbase focuses
on German and English (UK) but it includes some multilingual term records,
and some records also contain terms in English (US) too which, for example,
serve to document that the type of hinges used for furniture in the US differ
from those used in Europe and around the world.
Generally, a strong emphasis is placed on definitions, especially for
new products, in order to make it easier for translators to understand the
components used for the product and how they fit in together. Where pos-
sible, images of new and important components are added to the termbase
for the same purpose. For example, it is not always easy to imagine the
differences between lift systems in wall cabinets: some fittings simply fold up
the cabinet door (‘lift up’ system), whereas others swing up the cabinet door
and over (‘up and over lift system’), both illustrated in Figure 9.6.2

Figure 9.6 Blum’s ‘up and over lift system’ (left) and ‘lift up system’ (right) for wall
cabinets.
 Case Studies on Commercial Organisations 209

How Do Translators Manage (Blum’s) Terminology at the Kaleidoscope Group?

Terminology at Eurocom (the LSP side of the Kaleidoscope Group) is

implemented by translators/​ revisers, managed by project managers and
supported technically by staff at Kaleidoscope (the software side of the com-
pany). The main CAT tool used at Eurocom is Trados (and MultiTerm),
which is used by translators in conjunction with the Kalcium interface,
developed by the company’s own Kaleidoscope division. Kalcium is made up
of several interconnected modules which can work as standalone programs
too. Some of these modules can be integrated through plug-​ins into Trados,
MultiTerm, MS Word and other programs. The full list of modules available
in Kalcium can be seen on the website (Kaleidoscope, 2023c) but the ones
that interest us here include Kalcium Quickterm, Kalcium Checkterm and
Kalcium Smartquery. A full review of the features of each program cannot
be provided here so we will only focus on the most interesting features for
creating and managing terminology in translation projects.

Kalcium Quickterm

Kalcium Quickterm (Quickterm for short) is Kaleidoscope’s own termin-

ology management system. It is available in three editions, Management,
Workflow and Enterprise, the features of which are itemised on the website
(Kaleidoscope, 2023c). An overview of those features, taken from the web­
site, is shown in Figure 9.7.
In Quickterm Management, terminological data can be accessed and
managed either directly in the Kalcium platform, or through programs such
as MultiTerm, Acrolinx, etc. Quickterm Workflow offers more options for
managing terminology within a company and Quickterm Enterprise is the
top of the range version of the platform, which offers enhanced options such
as the use of concept maps (see Figure 9.7 and Chapter 6). Quickterm, in its
various editions, can be used in conjunction with TermCatch and TermAlign
(developed by the Swedish terminology management company Fodina),
which allows translators to integrate AI into terminology retrieval and man-
agement (see Fleischmann & Lundahl, 2023). The main Kalcium interface3
is shown in Figure 9.8 and the terminological features of Quickterm can be
accessed by clicking on “Terminology” at the top of the screen.
Before we review the terminological features available in Quickterm, it
is worth noting the contents of the three main columns in Figure 9.8, i.e.
Welcome!, Entry of the week and Term quiz. In the left column, Welcome!,
translators (or any other company user) can use the Search boxes to consult
terms and retrieve information about them. Above the search boxes,
there are buttons to access documentation, in English or German, on how
to use the demo and the various roles in which the demo can be accessed.
In this section, the demo was accessed using the terminologist role which
 newgenrtpdf
210 Terminology Management for Translators
Figure 9.7 An overview of the features of the various editions of Quickterm.
Source: Kaleidoscope (2023c).
 Case Studies on Commercial Organisations 211

Figure 9.8 Typical dashboard interface in Kalcium.

has complete admin rights to all the terminological features. Other roles
include end user, translator, reviser and, interestingly, marketing tech. This
is interesting because it demonstrates Kaleidoscope’s approach to termin-
ology, considering its purpose to be “to guide all content creators –​from
authors and marketers to UI designers and master data specialists” (from
Kaleidoscope promotional material). So, terminology software is not only for
translators, revisers and terminologists, it is also to be used simultaneously
by marketing, technical staff and indeed across the entire company. This is
one way in which Kaleidoscope helps companies to “speak” consistently
with the same voice, both internally and externally.
The other two columns in the above figure support this approach further.
Both the middle column, Entry of the week, and the right-​hand column Term
quiz, are designed “to support a 'low resistance' roll out of terminology in
the organisation, reduce people’s inhibitions toward terminology, and pro-
mote terminology awareness in a more light-​hearted manner” (Kaleidoscope,
2022a: 4). As already noted, the terminological features of Quickterm are
accessed by clicking on “Terminology” at the top of Figure 9.8. This leads to
a highly streamlined interface, illustrated in Figure 9.9 (also from the publicly
available demo).
On the left of the figure, terms can be searched for directly by typing
them in the box, while the “Search Settings” button allows users to carry
out wildcard searches, fuzzy searches or full-​term searches. Under “Search
Settings,” the Taxonomy button is used to search for related entries. To use
the example in the online demo, a search for brake system might yield entries
newgenrtpdf
212 Terminology Management for Translators
Figure 9.9 Terminology interface in Quickterm.
 Case Studies on Commercial Organisations 213

for 2 piston brake calliper and 4 piston brake calliper but of course this will
depend on the contents of the termbase. The name taxonomy may refer to
the fact that, in this case, 2 piston brake calliper and 4 piston brake calliper
are meronyms of brake system. Taxonomy is not to be confused with concept
maps. Whereas Taxonomy is a search method, concept maps in Quickterm
are a visual representation of the relationships between the concepts in the
termbase. It can also be used as a cross-​reference tool as it allows users to
click their way through the termbase, “drilling down or zooming back out
of the context” (Kaleidoscope, 2022a: 4) as required. Concept maps can be
created manually by clicking on the Concept Maps button in the top toolbar
and selecting the relevant entry. Concept maps are typically created by the
clients’ technical experts, not by linguists, but translators at Eurocom can
benefit from them.
Returning to Figure 9.9, other terminological tasks available in Quickterm
can be accessed on the right hand side, ranging from Creating new entry
and Merging two entries to Request new term and assigning approval tasks,
translation tasks or discussion tasks. The advantages of carrying out all
these tasks online include staff being able not only to collaborate remotely
in real time, but also keeping a record of all the tasks, discussions (and
their outcomes) and approvals for future reference by anyone throughout
the company.
The Watch button allows translators to receive email notifications if a spe-
cific entry is changed. The Send feedback button allows two-​way communi-
cation between users, typically between project managers and subject matter
experts, or project managers and translators or revisers. Users are notified by
email when feedback has been requested or received. More detailed infor-
mation about the feedback requesting process can be found in Kaleidoscope
(2022a). Although not visible in Figure 9.9, there is also chat functionality
allowing all stakeholders in a specific translation project to maintain contact
with each other.
Project managers, translators, revisers and subject matter experts often
need to contact each other when new terms are created or existing ones
edited. In order to harvest new terms, Quickterm is integrated with Fodina’s
TermCatch which can be used as its term extraction and concept creation fea-
ture (see Term Extraction Tools in Chapter 6). Once the new concept has been
created in TermCatch, showing the various terms for it and a definition, the
concept can be imported directly into a termbase in Quickterm. Quickterm
also has a permissions system whereby only the relevant colleagues can
approve or edit terms and concepts. A very detailed demo can be viewed for
more information in the joint webinar by Fleischmann (Kaleidoscope) and
Lundahl (Fodina) (Fleischmann & Lundahl, 2023).
It is also possible to generate a request from within TermCatch to amend
concepts and terms which exist in Quickterm (the other way around is not
possible at present), or to add new ones. To do this, TermCatch users can
request an export of concepts from Quickterm. This will result in a local
214 Terminology Management for Translators

copy of the Quickterm termbase being stored locally in TermCatch, which

means that users can compare the terms identified by TermCatch to the
terms already existing in the termbase ‘imported’ from Quickterm. This
would help translators find new terms, duplicates, terms to be updated, etc.
(see also Chapter 6). In other words, the integration between TermCatch
and Quickterm means that new concepts can be requested in Quickterm
if TermCatch identifies a cluster of synonyms which do not exist in the
Quickterm termbase. Similarly, if TermCatch identifies a cluster of synonyms,
some of which already belong to other clusters in Quickterm, users can gen-
erate a ‘change request’ in Quickterm which would amend the existing con-
cept by adding the new terms to it (and/​or creating a new concept with the
others if they are deemed to constitute a different concept).

Kalcium Checkterm

Kalcium Checkterm (Checkterm for short) is a terminology verification tool

which can be used to check the consistency of the terms that appear (or not)
in a text against the relevant termbases, much like a CAT tool. However, it
differs from the typical CAT tools’ terminology recognition capabilities in
that Checkterm has built-​in linguistic information. This means that, whereas
a CAT tool usually matches strings on a character-​by-​character basis, without
any linguistic knowledge, Checkterm employs morphological information to
recognise that scans, scanned and scanning, for example, are all inflected
forms of the verb scan. The advantage of this approach is that it produces
better term search results by reducing the amount of false positives returned
(Kaleidoscope, 2021a). Similarly, it can recognise different or incorrect
spellings, term variants and term compounds and can work monolingually
or bilingually. The main Checkterm interface, as shown in the Kaleidoscope
demo server4 is shown in Figure 9.10, having accessed this demo in the “ter­­
minologist” role.
Figure 9.10 shows that Checkterm is accessed through the main Kalcium
interface by clicking on “Terminology” in the top toolbar and then on
“Check” in the toolbar below it. The left-​ hand pane displays the three
main tasks available in Checkterm, i.e. Document analysis, Document/​TM
results and Segment analysis, whereas the main pane shows the results of a
Document analysis check. By clicking on Document analysis on the left pane,
translators can upload a single document, or a document and its translation,
and check them for accuracy and compliance with the terms in the relevant
termbase. By clicking on Document/​TM results, translators can view a list
of all the documents they have checked before and can re-​open any results
again if desired. The Segment Analysis button brings up a dialog box where
translators can paste or type specific text to be checked.
In Figure 9.10, the results of checking segment 2 (“During 2019, KTM
secured its fourth International Six Days Trial with Cooper Webb’s triumph
 Case Studies on Commercial Organisations 215

Figure 9.10 
Main Checkterm interface, showing the results of a Document ana-
lysis task.

on the ISDT”) can be viewed in the main pane. By hovering the mouse
over the term “International Six Days Trial,” for example, we will see that
it has a forbidden tag attached. In order to replace this with the correct
term, users can click on the Replace with button on the right-​hand side,
which should allow users to select the correct term from a drop-​down list.
Translators and revisers can use the same procedure to discern between
partial synonyms. For example, from a termbase on combustion engines,
Checkterm can help translators to ensure that the preferred term single
piston brake caliper is used, while avoiding the forbidden term one pot
brake caliper.
To view the results of the checks of other segments, translators click on

the arrows on the right of the segment (top right in the above figure). The
icon allows translators to view all segments at once in a table, showing
whether each segment has any issues to be addressed or not. To the right of

this icon, translators can click on to copy the segment to the clipboard.
On the left of the main pane, a list of forbidden hits is shown, e.g.
International Six Days Trial in this case, as well as similar compounds which
might be of interest to the translator. Above this, there are two buttons that
allow translators to download a report of the check that has been performed,
either in .CSV or .HTML format. The Regenerate source document button
produces the corrected version of the document.
In the centre of the main pane, there are some small buttons that can be
used (from left to right) show feedback, to show entry history, print entry,
216 Terminology Management for Translators

copy URL of current entry, send URL of current entry in an email (which
allows translators to send an email without leaving the Checkterm interface)
and entry view settings. On the right of the main pane, additional buttons
allow translators to add stopwords, request new terms, replace the incor-
rect terms and offer a replacement term for an incorrect one. Scrolling down
through the results in the above figure (and therefore out of view in the figure),
the translator would see the various terms/​partial synonyms included in the
entry and any associated information about them, e.g. definition, graphic,
status, etc., as well as metadata, e.g. date of last modification.
As we have seen before, Checkterm is designed for use by every employee of
a company, not just translators and revisers. If a company employs linguistic
staff such as translators, terminologists or project managers, for example,
they will probably be the most frequent users of this software. However,
Checkterm is also designed to be used by non-​linguistic staff, and this is the
case at Blum. For example, if a technical author at Blum queries whether
a term has appeared before in company documentation, or if they spot an
error that needs to be amended, they can submit term requests and change
requests on Checkterm. These requests can be dealt with by Eurocom project
managers or other internal users. In this way, authors and experts at Blum
can be in touch quickly and efficiently with (internal or external) translators
and project managers and ensure that the correct company terminology is
adhered to in every document.
As a tool conceived as corporate terminology software, another use of
Checkterm is to raise awareness of the importance of terminology among
staff, especially the non-​linguistic staff of a company, who are less likely to
be aware of the benefits of consistent terminology than linguistic staff. One
way in which this is achieved is by ensuring that Checkterm can be accessed
from the environments where experts and technical authors usually work,
e.g. Oxygen XML Author, XMetaL, Visual Studio, Visual Studio Code, or
from any content management system used by the company (Kaleidoscope,
2024). Checkterm is also available as a plug-​in for RWS Trados Studio for
use by translators.

Kalcium Smartquery

Kalcium Smartquery (Smartquery for short), as we saw in a previous section,

is a query management program that allows comments, questions and
feedback to flow smoothly and securely among project managers, clients,
translators and external vendors. It is by means of this tool, albeit through the
project manager, that translators can contact the authors of the source text
and/​or relevant experts from the client company. By storing the translators’
queries, Smartquery creates a knowledge repository of the queries which
can be consulted and searched at any time. Answered queries can also be
imported into a termbase. If a client decides not to use Smartquery, queries
 Case Studies on Commercial Organisations 217

can be collected in an Excel spreadsheet, which allows import and export

between other programs. The Smartquery system has been extensively used
by Eurocom translators and clients and has generally improved relations and
communication among all the agents in the translation process. Smartquery
can be accessed from the Kalcium platform but, typically, Eurocom translators
work with Trados from which Smartquery can also be directly accessed.
Once a query is logged on the Kalcium interface, it can follow several
routes until it is answered. For example, it can simply be answered by a
project manager and sent back to the translator, or it can be delegated by
a project manager to a subject matter specialist and then to a reviser, or to
another department in the company, e.g. Marketing if the query is related
to important external-​facing content, before it is returned to the translator.
Once a query is answered, it is stored in the Kalcium repository so that it can
be viewed at a later stage, e.g. by revisers in the current translation project,
or by translators of similar documents in future projects. If an instant reply
to a query is not expected, queries can also be “watched,” i.e. translators
will receive automatic notifications by email when an update or reply has
been received for a specific query. A full demo showing the interactions
between translators, project managers and clients on Trados and Smartquery
can be viewed in an online video by Kaleidoscope (Kaleidoscope, 2022b).
In the following section, we take a look at a typical MultiTerm termbase
from Blum.

Blum’s MultiTerm Termbase

As well as working with the various Kalcium modules, Eurocom translators

also deal with terminology in MultiTerm termbases, as required by their
clients. Blum is one of Eurocom’s clients that uses MultiTerm and, in this
section, we take a brief look at how a typical Blum termbase is organised in
MultiTerm. The structure and descriptive fields are shown in Figure 9.11. At
Entry level, there is a multimedia field for images which is multiple, i.e. sev-
eral images can be added for one concept. Field refers to the domain the con-
cept belongs to, and it is a picklist with values that are meaningful in Blum
such as General, Sales and after-​sales, Quality assurance, etc. The Definition
field at this level is normally used for the definition in English. The Definition
field at Term level is used for definitions in languages other than English.
Source is a multiple field, which can refer to any other field, e.g. source of the
definition, source of the image, etc. Group, Subgroup and Subcategory are
all picklists that mostly include the various categories and subcategories of
Blum’s products, always used in all capitals (e.g. MINIDRILL, EASYSTICK,
BLUMOTION, etc.). Entry-​Status is another picklist with four values, i.e.
OK, Pending, For translators and Invalid. Entry-​Comment is a text field that
allows users to add qualitative data to these statuses. The Keyword text field
is currently not being used.
218 Terminology Management for Translators

Figure 9.11 Descriptive fields and hierarchy in a typical Blum termbase in MultiTerm.

At Index level, there are only two fields which capture information about
the workflow in a specific translation project. Workflow-​Status is a picklist
with four values, i.e. New, Completed, In progress, Invalid, which can be
explained further in the Workflow-​History field, which is a text field in which
users can add any relevant comments about the progress of a translation job.
At Term level, we find that gender and number have been grouped together
in the Gender and Number field, which is a picklist of several values, allowing
users to select the gender (for relevant languages only) and number of a specific
term. For example, the value n.m.sg is for masculine singular nouns. However,
this picklist is not often used. Usage consists of a picklist with three values,
Approved, Forbidden and Use with caution. This information can be elaborated
on in the Usage note field, which is a text field. The Usage register field is
another picklist with many values relating to various aspects of the manufac-
ture of Blum’s products, e.g. Assembly equipment, Control engineering, Powder
coating, etc. Term type is another picklist with six values: Long form, Shortened
form, Standard form, Slogan, Proper name and Full form. Context is a text field
where examples of usage of a term can be recorded. All descriptive fields are
multiple except for those at Index level and Term type at Term level. Usage is
mandatory, not multiple, whereas Field is both mandatory and multiple.

Other Aspects of Translation and Terminology Management at the

Kaleidoscope Group

The Kaleidoscope Group does not have any central termbases. Instead,
termbases are created according to each individual client’s needs, which in
 Case Studies on Commercial Organisations 219

turn depend on the size of the client’s company and their terminological
requirements. Thus, it may be that a single termbase is all one client requires,
whereas other clients may require several. For example, if a client works in
three separate domains, e.g. construction, railway and automotive domains,
Eurocom project managers will typically create one termbase with a field in
each entry used to indicate to which domain a respective term belongs. It is
also very common to create a separate termbase for UI strings, for example.
Sometimes termbases contain different terminology for different geograph-
ical variants of a language, e.g. UK English and US English are treated as
two different languages within the same termbase, but this is less often the
case.
When a new client company signs up, or if a new language is added to
an existing client’s portfolio, for example, the norm is for the project man-
ager to work with the client to develop a “terminology project,” where the
client collects a list of about 1,000 key terms, drafts their definitions and
commissions Eurocom to translate the terms and definitions. Different rates
apply depending on the amount of work required (usually ranging from ter-
minology databases created from an existing translation memory through to
creating an entire termbase from scratch).
Terminology projects are usually managed through QuickTerm (and
Smartquery, e.g. if a translator finds inconsistencies), which allows the
client’s preferred translator to be automatically assigned to translate the
terms and definitions. Revisers are also assigned to look through both
the translations of the terms and of the definitions, and project managers
will carry out the final quality checks before delivering the project back to
the client. In this sense, all of the translation jobs at Eurocom start with ter-
minology and great emphasis is placed on definitions, as we will see below,
which can help disambiguate the context in which a specific client’s term is
being translated.
Eurocom linguists, i.e. project managers or translators, can also assist a
client in extracting terms and drafting their definitions from scratch, e.g. if
the client is a smaller or newer company, or if the client is yet to develop ter-
minology for its products. Conversely, some clients create and manage their
own termbases, without Eurocom’s involvement, and simply send a subset of
their termbase, exported to Excel format, to project managers at Eurocom,
who will then import the Excel spreadsheet into a MultiTerm termbase when
a translation is commissioned.
Kaleidoscope and Eurocom have established a working group for machine
translation and artificial intelligence, which is mainly concerned with testing
different NMT tools to identify where AI has the potential to make the trans-
lation process more efficient. This includes looking at the types of content
for which NMT offers a quantifiable efficiency gain and for which types of
content it is inappropriate (either because there are no efficiency gains or
because the final translated content does not meet the client’s expectations),
and exploring the potential that AI has to improve post-​editing, e.g. whether
220 Terminology Management for Translators

AI can be used to analyse a pretranslated text and highlight to the human

expert which sections of text require the most attention. The group is also
investigating which NMT tools are better for which language pairs and con-
tent types, and how Kaleidoscope/​Eurocom’s data needs to be prepared in
order to improve the NMT output, e.g. through TM cleaning and ST prepar-
ation. In terms of terminology, the group is currently looking at how it can be
incorporated into NMT tools in order to customise the output. The ultimate
aim here is to be able to guide Kaleidoscope/​Eurocom’s clients, advising
them where MT and AI can bring them benefits while steering them away
from processes that could be harmful to the quality of their content. There is
also a newly created working group on transcreation, so we can also expect
developments in this area.
Dealing with phraseology, especially in the form of clients’ slogans and
marketing material, is an important part of the Eurocom’s translators’ work.
Phrases or slogans that have product or company relevance are normally
recorded in a termbase, rather than in a translation memory, and Eurocom’s
clients are encouraged to do the same. This is because, during translation,
matches from the termbase are ranked higher than matches from the transla-
tion memory. Checkterm is very useful in this respect: if the author changes
the set phrase or slogan in a termbase, e.g. by introducing wordplay, which
the term recognition function in Trados may not “notice,” Checkterm then
finds and highlights these matches to the translators so that the changes can
be dealt with appropriately in translation.
Linguistic changes towards more diverse and inclusive language are felt
both by Eurocom and its clients. This can be seen in contracts of employ-
ment, for instance. Where contract templates used to feature terms such
as Jane Doe or John Doe, which would be replaced by the name of the
new member of staff, contract templates have been updated to replace
those terms by other more inclusive and diverse options such as person, or
average person, for example. Similarly, the German man is gradually being
replaced by wir, and Otto Normalverbraucher by Durchschnittsmensch,
throughout documentation, both at the Kaleidoscope Group and its
clients.
The impact of new technological trends is also strongly felt at the
Kaleidoscope Group. One such trend is the increase of AI and MT elem-
ents in software, another is that the number of revision requests by client is
increasing, as opposed to translation requests. These trends may suggest that
translation requests will move away from the source text-​centred requests
to a more holistic service in which translation is just a part of the service.
Future software developments that can also be expected include fuller com-
patibility among software programs, including more compatibility with RWS
programs.
 Case Studies on Commercial Organisations 221

Translation and Terminology at LSP2: Terra Translations

Box 9.3 Key Highlights: Terra Translations

• Translators and QA Managers carry out terminological tasks.

• Two review stages during the translation process.
• An additional review may be carried out after desktop publishing
(when it is required).
• Termbases are ‘moderated’ by QA managers before use in a transla-
tion project.
• Terminology is standardised at client level.
• Advanced text processing skills, e.g. RegEx rules and auto-​
translation rules.

Background of the Company: Terra Translations

Terra Translations is a language services company specialising in localisa-

tion, established at the turn of the millennium by the current CEO of the
company, Marina Ilari and her mother, who were both translators. Marina
Ilari is now a certified ATA translator with over a decade of professional
translation experience, having earned a degree in Literary Translation from
the Universidad del Museo Social Argentino (Terra Translations, 2018a). In
2020, Marina received the Enterprising Woman of the Year Award (Terra
Translations, 2023d).
Terra Translations started as a family business and today, it is a fully virtual
company, albeit with three offices, in Argentina, Spain and the US. It employs
150 linguists and specialises in six industries: education, e-​learning, govern-
ment, law, life sciences and video games (Terra Translations, 2023b) (see
also Box 9.3). Terra Translations offers a range of translation and localisa­
tion services, including transcreation, linguistic quality assurance and desktop
publishing. It is a women-​owned and minority-​owned company with a strong
ethos on diversity, as employees describe themselves as being “on a mission to
empower diversity in the localization industry” (Terra Translations, 2023a). It
also provides pro-​bono translation services to “select non-​profit organizations
in underserved areas” around the world, through its Terra Cares programme
(Terra Translations, 2018b). Examples of beneficiaries include organisations
such as Girls in Technology and Doctors Without Borders.
Terra was originally founded as a company specialising in Latin American
Spanish and Brazilian Portuguese. Over time, it evolved into a multilingual
vendor, expanding its services to meet the diverse needs of its clients. Today,
Terra collaborates with a group of in-​house and freelance linguists world-
wide, offering services in over 50 languages. In 2017, Terra Translations was
222 Terminology Management for Translators

recognised as a Rising Star in the MARKETPLACE 2017 Governor’s Awards

competition (Pilarski, 2023) and in 2021 it was ranked by CSA Research
as the second largest language services provider in Latin America (Terra
Translations, 2021). In 2024, CSA Researched positioned Terra Translations
in the Top 20 Language Services Providers in North America (CSA Research,
2024). A full list of awards and certifications can be found on its website
(Terra Translations, 2023e).

Organisation and Workflow at Terra Translations

At Terra Translations, 150 staff members approximately work in linguistic

roles, all referred to collectively as linguists (see also Box 9.4). The actual
role of a linguist in a translation project will depend on the project itself.
Although Terra offers a wide range of translation and localisation services,
the most requested service is what they call “TEP,” i.e. Translation, Editing
and Proofreading. Thus, a linguist may act as a “translator” in one pro-
ject, whereas they might be a “reviewer” in another. There are no dedicated
terminologists but there are QA managers who carry out terminological
tasks. QA Managers typically hold a Linguistic/​Translation degree and are
highly experienced in both CAT tool/​memoQ technical field and the gaming
industry. Multilingual QA Managers oversee all processes, both linguistic
and procedural, to ensure that all languages and procedures follow a con-
sistent approach, abide by the team’s and the client’s needs and exceed any
expectations. A primary responsibility of the Quality Assurance workflow is
to reduce the occurrence of human error as much as possible. To do so, they
take care of performing the last quality check to any translation and also
put preventive and reactive measures in place. Terminology management is
one of those measures. The generic translation and localisation workflow is
illustrated in Figure 9.12.
When a client’s request comes in, it is dealt with by the assigned project
manager (PM). PMs then select the linguist(s) for the project and create the
project in Terra Soft, the company’s translation management system. Once
on Terra Soft, the technical preparation of the files, e.g. segmentation, tags,
placeholders, filters, etc., is carried out by the solutions architect. The pro-
ject manager then organises the project scope, the sharing of instructions,
deadlines and assets with linguists and/​or freelancers as required. Linguists
and/​or freelancers complete the translation or localisation job in memoQ
(unless of course the client has requested the use of another tool).
Once a translation is complete, it is first reviewed by another translator
(who then takes a reviewer/​editor role) in what constitutes the first review
stage. The translation then progresses to the second review stage, which
is performed by the QA managers. QA managers will deliver the finished
job to PMs, who will in turn deliver the translation to the client. If desktop
publishing is required, it can be carried out in-​house and sometimes it is
outsourced. After desktop publishing, the QA team reviews the final files
 Case Studies on Commercial Organisations 223

Figure 9.12 Generic translation and localisation workflow at Terra Translations.

again. If everything is OK, the files go back to the project manager who
delivers them to the client. It is ultimately the responsibility of QA managers,
rather than the project managers, to approve both the translations and the
terms that are finally used in each project. At all stages, the entire team is
supported by the chief operating officer and the director of localizations.
The CEO and the business development manager ensure that the client is
supported throughout.

Box 9.4 Terra Translations –​Translation Facts and figures

Location: Online Own MT? No

No. of translators: 151 linguists Project Management tool: TerraSoft
Terminologists: Yes Specialism required/​encouraged: Yes
CAT tool: memoQ Outsourced work: Yes
Termbase: memoQ Main source languages: English and
Spanish

How Do Translators Manage Terminology at Terra Translations?

The default CAT tool at Terra is memoQ, as we have seen in the previous
section, unless clients require the use of other CAT tools and/​or project man-
agement systems, such as Smartling, Phrase, Wordbee and XTM, although
224 Terminology Management for Translators

sometimes clients have their own proprietary tools to which Terra’s linguists
need to adapt. Terminology, therefore, is mostly managed in memoQ. At
Terra there is no central termbase to which all linguists contribute. Instead,
Terra’s linguists work with different glossaries or termbases for each client
and, in most cases, there are different glossaries or termbases for each account
or product belonging to the same client. Glossaries are only used when ter-
minology needs to be shared with the client for approval or feedback. In some
cases, the content among accounts from the same client overlaps and, when
that happens, more than one termbase is added to the project in memoQ, but
new terminology is only added to the main termbase for each account.
Termbases at Terra are “moderated,”5 i.e. approved by a terminologist (or
QA manager in Terra’s case). When linguists start working for a new client,
the QA team usually creates a new glossary based on the documentation
provided by the client. Only moderated termbases are used when carrying
out jobs for the client. An example of Terra’s termbase in memoQ is shown
in Figure 9.13. At Entry level (see also Concept Orientation in Chapter 4),
the fields used include Project (for a project ID), Subject, Domain and Client
as well as metadata (e.g. Created by, Modified by, etc.). If necessary, linguists
can use the Notes field and add a graphic under Image. At Language level,
terms in each language are entered, including synonyms. At Term level, infor-
mation can be entered under four headings, i.e. Matching, Usage, Grammar
and Definition.
Matching is a picklist with the default value being 50% prefix, which means
that a term in the source text must be an exact match to the first half of a
term in the termbase. If other kinds of fuzzy match are needed, other picklist
options include Fuzzy (which allows the first half of a term to be different
between source text and termbase), Exact and Custom. Under Matching,
there are several capitalisation options, i.e. yes, no and permissive.6
Under Usage, terms can be marked as forbidden (see Chapter 4) and
examples of use can be added. Grammar is another picklist to assign a part
of speech to terms and it has four main values (noun, verb, adjective and
adverb), as well as other. Therefore, noun phrases must either be classed as
noun or as other. Under Grammar, users can also add information regarding
the Gender and Number of the term as required. Definition is a free text field
to add the definition of a term.
In the case of cosmetic synonyms, (e.g. check box, checkbox, Check box,
etc.), in the source language, they are all included as terms within the same
entry when they are used with the same meaning. In this way, Terra’s linguists
can aim for consistency in translation, regardless of the spelling used in the
source text. However, if the cosmetic synonyms have different meanings, then
Terra’s linguists create different entries, one per meaning. In these cases, a
note is made in the Notes field about the differing meanings and spellings,
although sometimes this is rather obvious and does not require a note at all.
Terra’s preferred solution to deal with cosmetic synonyms is by using RegEx
rules (in QA profiles and auto-​translation rules in memoQ) which can be set
newgenrtpdf
Case Studies on Commercial Organisations 225
Figure 9.13 Screenshot of Terra’s termbase in memoQ.
226 Terminology Management for Translators

to “demand” a specific spelling in a specific context. If there is any doubt,

linguists will add relevant information in the Notes field, although Terra’s
usual policy is to use Notes sparingly, as the general view is that having too
many notes to read is as hindering as having no notes at all. Terra linguists
are proficient in using RegEx rules (see also Chapter 6), which can be used
to generate a QA warning showing all the synonyms that can be used for the
translation of a term in a specific context. If the content of a termbase needs
editing or updating, approval must be sought from the QA team.
Since Terra deals with many language variants, e.g. French (Canada) or
French (France), dealing with geographical variants is an important part of
the linguists’ work at Terra. The general policy is to treat every variant as a
separate language anyway. This is implemented when creating a new project
and assigning translation memories and termbases to it. In cases where the
same translation of a term can be used in several variants, linguists set this up
as a preference in the general project settings. If a project is set to distinguish
between variants, all the terms in the attached termbases will be displayed by
variant. If not, the linguists working in those projects will be able to use all
the terms for the variants. Auto-​translation rules and QA profiles are always
used. Depending on the account requirements, there may be one QA profile
and one auto-​translation rule for all the languages, or there may be separate
QA profiles and auto-​translation rules for each language.
Terminology standardisation is a key process at Terra, where it is under-
stood as the process of centralising terminological information and thus
ensuring terminological consistency. Standardisation of terms is important
at Terra because it is inherent to the bigger process of quality assurance. This
is another reason why there are no central termbases at Terra, only project-​
specific ones (or account-​or product-​specific ones). However, the termin-
ology standardisation process is carried out at account level because the same
client might request different localisations for the same term depending on
the nature of the source document, among other factors. Each department at
Terra works with different clients and accounts. Then, the QA team in each of
the departments is in charge of creating and updating the termbases regularly.
Project managers and QA managers work together to set up project-​creation
templates which are also account-​specific (or client-​or category-​specific) in
order to ensure that only the relevant termbases are attached to relevant
projects. These termbases must always be moderated.
At Terra there are three possible ways to deal with special phraseology,
e.g. a phrase which has a preferred translation or treatment in a specific pro-
ject. For such phrases, one method is to include its preferred translation in a
termbase; another is to create a RegEx rule added to a QA profile; and the
third approach is to include the preferred translation in an auto-​translation
rule. As with all projects at Terra, the decision is project-​specific or account-​
specific and will vary from client to client, project to project and account
to account. To date, Terra has not trained its own MT system. Terra’s pro-
ject managers can use MT through Intento, an AI-​powered plugin that is
 Case Studies on Commercial Organisations 227

connected to memoQ and allows linguists to access many MT engines via

a single API. It is the project managers who decide for which projects MT
will be used or not. Linguists will see the MT suggestions for the relevant
segments, but they do not select or deselect the MT option for projects.
Finally, issues relating to inclusive and diverse language pose some challenges
to Terra’s linguists, e.g. in game localisation where there may be characters
described as gender neutral, but Terra’s linguists have to work with languages
which do not have approved or established rules for gender inclusivity and
diversity, so it is challenging to localise certain character names, descriptions
and dialogues.

Outlook
Each of the case studies in this chapter highlights the importance of considering
terminology work in its specific linguistic, cultural and commercial context.
The terminology solutions presented here vary a great deal from one case
study to the next, but they reinforce the idea that terminology needs to be
developed differently for different reasons in different contexts. They also
show us that terminology is at the service of other subject areas or domains.
The case studies included in this chapter are in fact not representative of the
entire spectrum of commercial companies and LSPs –​there are many other
ways of implementing terminology solutions in commercial contexts than
can be shown within the scope of this book. However, these case studies
are intended to showcase some examples of solutions to the terminological
challenges posed by the constraints of commercial translation.
Some important terminological insights can be drawn from each case
study. We have seen that the winning solution at Kaleidoscope/​Eurocom
is for software specialists and linguists to join forces. If a specific program
or software feature is required, it can be developed in-​house. If software
specialists have a query as to what a software program for translators should
include, they can contact their in-​house translator colleagues. Kaleidoscope/​
Eurocom have also shown us that sometimes it is more cost-​effective to
partner up with other companies than to go it alone. This can be seen in the
partnership between Fodina and Kaleidoscope/​Eurocom when it comes to
term extraction. Kaleidoscope has not needed to develop its own term extrac-
tion software because Fodina’s term extraction program TermCatch has been
integrated into its Kalcium platform, e.g. through Checkterm, Quickterm
and Smartquery.
Kaleidoscope has also shown us that translators’ queries are an
important part of the terminological part of the translation process and
have developed Smartquery precisely to deal with them and to preserve
the wealth of information they contain. The idea of devising a system-
atic method of recording information for future use can be traced right
back to ancient civilisations such as the Egyptians or Sumerians who
“pioneered accounting techniques to keep track of data and understand
228 Terminology Management for Translators

their day-​to-​
day lives” (Quickbase, 2024). Once computers became
widespread, so did database software and eventually query management
software (2024). Today, query management is at the heart of effective
communication which, in turn, is an essential ingredient in termino-
logical consistency across a company or organisation. To my knowledge,
Kaleidoscope is one of the first companies to pioneer the use of query
management software in the translation by launching Smartquery in 2022
(Kaleidoscope, 2023f), a desirable feature to include in future versions
of not only computer-​assisted translation but also terminology software.
Terra Translations showcases a different concept of language service
provider: it is a fully online company whose case study shows us that
translators will benefit from enhanced text processing skills, such as being
able to use regular expressions (RegEx for short), for carrying out ter-
minological tasks. More broadly, Terra Translations highlight the import-
ance for translators of knowing in depth the software they use.

Follow-​up Tasks and Reflection

1. Kaleidoscope is not the only company that uses a query management
system as part of its translation process. Compare Kaleidoscope’s
Smartquery to the “Discussions” feature in memoQ (e.g. Beheshty, 2018)
and the Query Management module in the CAT tool XTM (from version
13.1) (e.g. Favre-​Lorraine, 2022). Do some research online on these tools
and try to find out what the similarities and differences, strengths and
limitations are.
2. Where do the main differences lie between terminology in commercial
settings and terminology in institutional settings? Make a list of any
differences between the case studies in Chapter 7 (governmental) and those
in Chapter 8 (large institutional), e.g. in terms of who has permission to
edit termbases. Who has the final word when revising translations?, etc.
3. Do some research to find the website of a translation agency in your region.
Explore their website and find out which kinds of translation services they
offer, e.g. legal, technical, medical, pharmaceutical, etc., and whether
any of their clients are mentioned on the website, e.g. in a testimonials
webpage. Do they accept jobs from individuals or is it a “B2B” company?
If there are no clients listed on their website, think of a plausible client
or two for them. How might you organise the terminological efforts of
this company, e.g. one central termbase and/​or several, more focused
termbases, in view of their clients?
4. What are the advantages and disadvantages of adding a definition at
different levels of a termbase’s hierarchy? For example, Blum adds
definitions at Entry level and Term level, the Translation Bureau at Index/​
Language level (see Chapter 7)? Alternatively, organisations may add
some definitional material in the Notes field of a termbase on the basis
that their translators do not usually need definitions.
 Case Studies on Commercial Organisations 229

Notes
1 The Blum catalogue is actually presented in full colour, not in black and white as
shown in Figure 9.3.
2 The ‘lift up’ graphic is available at www.blum.com/​gb/​en/​produ​cts/​lift​syst​ems/​aven​
tos-​hl/​progra​mme/​ and the ‘up and over lift system’ graphic is available at https://​
alum​inum​sys.com/​produ​cts/​blum-​aven​tos-​lift-​up-​syst​ems/​
3 Demo publicly available at https://​demo.kalei​dosc​ope.at, version 6.6.0.14.
4 http://​demo.kalei​dosc​ope.at
5 Normally, when a translator adds a term to a termbase, it becomes visible imme­
diately. However, in memoQ, the Moderated checkbox can be ticked during the
termbase creation process, which means that the new term will only become visible
once a QA manager has approved it.
6 Permissive means that “memoQ will be sensitive to capital letters only. Uppercase
letters in the term must be uppercase in the text, too, otherwise memoQ will not
return a match. On the other hand, lowercase letters in the term will be matched
by uppercase or lowercase letters, memoQ will not be sensitive there. For example,
if the term base contains the term ‘ALPAC,’ the text must also contain ‘ALPAC’ to
receive a match. However, if the term base contains the term ‘Alpac,’ the text can
be ‘ALPAC,’ ‘Alpac,’ ‘AlPac’ etc., as long as the initial A is in uppercase: ‘alpac’ will
not match" (from the memoQ documentation at https://​docs.memoq.com/​curr​ent/​
en/​Conce​pts/​conce​pts-​term-​bases-​ins​ide-​an-​entry.html).

Further Reading

About the Kaleidoscope Group and Terra Translations

A more detailed overview of the services provided by the Kaleidoscope Group can
be found on its website (Kaleidoscope, 2023d). Under the “Resources” tab, readers
will find Kaleidoscope’s blog and information about current and previous webinars
(Kaleidoscope, 2023b). As for Terra Translations, readers are referred to an interview
with its current CEO Marina Ilari in the Milwaukee Magazine (2022) and to the
company’s blog (Terra Translations, 2023c). A Slator article shows Terra Translations
reaching number 1,570 in the LSP Ranking on the 2023 Inc 5,000 List (Txabarriaga,
2023). For readers interested in regular expressions, it is worth exploring the intro­
duction for translators offered by the ATA Chronicle (ATA Chronicle, 2019).

Further Terminology and Translation Practices in Other Commercial

Organisations
Bernth, McCord and Warburton (Warburton being a former terminology manager
at IBM) describe “the process of proposing, designing, developing and deploying
a terminology extraction tool” (2002: 1) at IBM. In a more recent publication by
Warburton (2021), she focuses entirely on terminology management in commercial
settings. Ericsson is another company whose terminology management processes have
been described by Jaekel (2000).
There are other studies on commercial organisations which do not focus on termin-
ology specifically, but which shed light on other aspects of language, translation and
230 Terminology Management for Translators

localisation practices in other companies. For example, several studies have focused
on language and translation practices at IKEA: Allam (2023) looks at the localisa­
tion process of its websites in English and Arabic; House and Kádár (2020) look
at the translation of pronouns in a variety of languages, including Chinese, French,
Dutch, German and Hungarian; Thije and Pinto (2023) compare Dutch and Italian
translations of the IKEA catalogues. Hein et al. (1997) look at the issues around
establishing a controlled language at Scania and later Almqvist and Hein (2000) pre­
sent Scania’s language checker and how it is integrated in translation workflows.
In a landmark publication, the edited volume by Sprung and Jaroniec (2000) pro­
vide a comprehensive study on the challenges faced by a wide range of commercial
businesses and the methods and processes they follow in order to succeed. Within this
volume, there are studies on Microsoft (Brooks, 2000; Kohlmeier, 2000), Schneider
Automation (Hofmann & Mehnert, 2000), Hewlett Packard (Combe, 2000), Kodak
(Topping, 2000) and Caterpillar (Lockwood, 2000), to name but a few.
10 Terminological Best Practices
and Quality Management

Key Questions

• What can we learn from the Case Studies?

• What other research is being carried out in terminology management?
• What efforts are being made to standardise terminology?
• How may new technological developments affect terminology man-
agement practices?
• What is the likely future place of terminology management for
businesses, freelancers and organisations?

Lessons Learned from the Case Studies

Translators carry significant social responsibility because of the role they play
in the transfer of knowledge, products, services and ideas around the world.
One of the ways in which specialised knowledge is disseminated among
different groups is through the terminological work of translators. In this
book, we have carried out new research in several translators’ workplaces,
i.e. institutional (including sub-​types such as governmental and large institu-
tional settings) and commercial, both workplaces strongly underpinned by
outsourced freelance work. The aim was to provide some empirical evidence
as to how professional translators carry out their terminological work, under
which conditions, with which resources, how they solve or find workarounds
for the challenges they encounter and, ultimately, how they ensure termino-
logical quality. In this chapter, we review the key lessons to be learned from
the data collected through visits to those workplaces and the interviews with
translators and terminologists, which is summarised in Table 10.1.
From our visit to the Translation Bureau in Canada (Chapter 7), we
learned that terminologists work alongside translators. The bulk of the
terminological work is carried out by in-​house terminologists, freeing up
translators who can then focus on translation per se, but translators still
carry out domain research as required and liaise with terminologists when
DOI: 10.4324/9781003302636-12
232 Terminology Management for Translators

Table 10.1 Summary of lessons learned from the Case Studies

Organisation Lessons learned

Translation • Translators work alongside terminologists.

Bureau, • Both translators and terminologists can carry out domain
Government research but translators who specialise can become “content
of Canada specialists.”
• About half of the translation work is outsourced, all of it
revised in-​house.
• Widespread use of terminological committees, both permanent
and ad-​hoc, to solve terminological issues.
• Experts, language users, language planners and other
stakeholders alongside translators and terminologists can take
part in relevant terminological committees.
• Definitions are important.
• Phraseological information is included in termbases.
World Bank • Terminology work is carried out by translators.
• Essential fields and data categories only in termbases.
Specialised knowledge is sought from specialised sources, e.g.
other UN organisations.
• There are no terminologists but there are editors who can
provide an additional layer of quality assurance to translations.
• Terms are only added to a termbase after they have been
validated.
• Significant investment in term extraction as a tool to
consolidate and improve existing terminological resources.
• Separate spreadsheet for language-​specific phraseological
information.
Kaleidoscope, • Software developers working alongside linguists, can help
Eurocom the company stay competitive in the wake of technological
UK and their developments.
client Blum • Terminological services offered.
• Kalcium platform consisting of interconnected programs that
can also function as standalone programs.
• Terminology is for everyone in a company, even non-​
linguistic staff.
• Considerable effort made to raise terminological awareness
across an entire company and provide light-​hearted methods to
engage with it regularly.
• Quickterm: termbase software with linguistic information;
includes concept maps; definitions are important.
• Checkterm: a program dedicated to terminology verification.
• Terminological consistency begins with the source text.
• Smartquery: knowledge repository of translators’ queries,
regardless of how small, for enhanced communication among
all agents in a translation project and for more effective project
management.
• Phraseological information included in the termbase.
 Terminological Best Practices and Quality Management 233

Table 10.1 (Continued)

Organisation Lessons learned

Terra • There are no terminologists, but QA Managers take

Translations responsibility for terminological quality and consistency.
• QA Managers work alongside translators, revisers and project
managers.
• QA Managers oversee all revision work.
• Glossaries (as subsets from a termbase) created if required for
client feedback and/​or approval.
• Terminology is standardised at client’s account level (as a client
may have more than one account at Terra); terminological
consistency also begins at source, with the client.
• Linguists have advanced text processing skills, e.g. the use of
RegEx rules; it is important to know the capabilities of your
software and making full use of it (see also Box 10.1).
• Phraseology can be dealt with by using RegEx rules (see also
Chapter 6).

queries arise. Specialising in a domain is important and translators who do

can opt to become content specialists.
An important part of the management of terminological quality at the
Translation Bureau is done through the terminological committees, some
permanent and some ad hoc, which ensure there is a space for dealing with
issues as they arise or when required. Involvement with subject matter experts
also takes place at these committees, which also include the involvement of
other stakeholders in specific terminological or linguistic issues such as lan-
guage users, language planners, etc. Another aspect to ensure quality and
consistency is the fact that, although almost half of the translation work is
outsourced, revision is always carried out in-​house.
The terms used by the Translation Bureau have been collected in the
TERMIUM® termbase for a few decades already and are publicly avail-
able online. Concept orientation is adhered to and there are data categories
at each level. At Concept level, domains and subdomains are clearly spe-
cified for each concept. Definitions are an important part of the termbase
and they are crafted following the strict classical method of starting with
the closest hyperonym, which must be followed by any necessary and suffi-
cient conditions (see Definitions in Chapter 4). This means that definitions
are often just one sentence. Excel files are regularly used, typically to collect
the results of term extraction activities and to import the new terms into the
termbase. Collocations and other important phraseological information are
collected in the phraseology field of the termbase. A Notes field (called OBS)
ensures that other contextual and background information becomes available
234 Terminology Management for Translators

to translators. Colour coding term entries provides information as to the

internal status of each term, e.g. draft records, published on TERMIUM
Plus®, or for translator’s use only.
At the World Bank, we discovered that not everything needs to be
translated into all the languages of the Bank, as is the case, for example, in
the European Union. In addition, there are no languages referred to as offi-
cial; instead, languages are referred to as working languages. The bulk of
the translation and terminological work is carried out by translators (either
in-​
house or outsourced) and editors constitute another layer of quality
assurance before documents which require a high level of quality, such as
external-​facing documents, are published.
In order to ensure the quality of the terms stored in the World Bank’s
termbase, WB-​TERM, terms are only added to a termbase after equivalents
have been found and have typically been validated after a translation job. In
this way, draft terms or translations are left out of the termbase and are only
added to it when their quality can be guaranteed. Another quality measure,
which also helps avoid duplication of efforts, is to keep a streamlined
termbase consisting only of three data categories at each of the three different
hierarchical levels, i.e. subject (at concept level), term (at language level) and
source (at term level). It is desirable for translators to specialise but, when
highly technical terms come up in a translation project, they are consulted in
the termbases of partner United Nations organisations. For example, agricul-
tural terms may be sought from the termbase of the Food and Agriculture.
For the same reason, there are no definitions in WB-​TERM.
Online Excel files are used to collect and work collaboratively on val-
idating extracted terms. Validated terms are imported into the termbase
through Excel too and used to train the Bank’s MT engine. In order to ensure
terminological quality over time, the Bank is currently making a significant
investment in term extraction as a method to consolidate and improve their
terminological repository accumulated over a long period of time. External
vendors have been tasked with extracting terms from all of the Bank’s
documentation, including monolingual documents. Quality is also ensured
through phraseological consistency and currently phraseological informa-
tion and resources are being compiled, outside of termbases, which will help
translators re-​create speakers’ idiosyncrasies across languages and to train
the Bank’s MT engine.
The Kaleidoscope Group, made up of the software solutions brand
(Kaleidoscope) and the translation services provider brand (Eurocom UK) was
a unique case study in that software developers work together with in-​house
translators and project managers. The company employs a Terminology and
Tools Expert in its Kaleidoscope side which can help ensure that key termino-
logical features are included in the relevant software as required. Eurocom
UK offers terminological services and charge a term rate (which is hourly),
estimated at the start of the job and finalised at the end, once the full number
of hours needed is known.
 Terminological Best Practices and Quality Management 235

Kaleidoscope has developed the company’s platform Kalcium, made up

of interconnected programs such as Checkterm, Quickterm and Smartquery,
which can also function as standalone programs. Having software developers
in-​house, working together with linguists, means that the developers can
add new features to the software depending on the linguists’ or clients’
needs; such collaboration can also contribute to keep the company competi-
tive in the wake of new translation technologies developments, e.g. AI and
LLMs. Another important lesson at the Kaleidoscope Group is that termin-
ology concerns everyone, including non-​linguistic staff, not just translators
or terminologists and this premise underlies their software. In order to
encourage involvement in the company’s terminology by everyone and pro-
mote terminological awareness throughout in a more light-​hearted manner,
Quickterm’s home page includes an Entry of the week window and a Term
quiz window.

Box 10.1 Terminology-​related Upskilling and Awards

Several large organisations offer programs, traineeships, fellowships

and awards where you can apply your terminology management
skills. For example, at the time of writing, the following were
available:

• Fellowship program1 for graduate students by WIPO (World

Intellectual Property Organization). There five types of pos-
ition: Terminology Fellows, Translation Fellows, Post-​ Editing
Fellows, Translation Technology Fellows and Technical Specialist
Fellows.
• Traineeships and study visits are also offered by TermCoord,2 the
Terminology Coordination Unit at the European Parliament.
• Those who become members of GALA (Globalization and
Localization Association) can join any of their Special Interest
Groups (SIG). One of them is on “Inclusive Language and
Terminology.”3 Teachers and students whose institution is a
member of GALA can also serve as volunteers in participate in
SIGs and committees.
• TermNet, the International Network for Terminology, offers
Terminology Summer Schools.4
• The European Association for Terminology (EAFT) offers
Terminology Awards.5
• The Young Professionals Program6 at the World Bank Group
allows you to apply for a program that matches your expertise and
experience.
236 Terminology Management for Translators

• You can track other terminology-​related events at the TerminOrgs

website.7
• For more on institutions’ outreach to and involvement with univer-
sities, see Biel and Martín Ruano (2023).

Close collaboration between translators, clients and subject matter

experts is strongly encouraged and facilitated through the Kalcium platform
and its modules and, typically, the project manager is a central figure who
ensures everything runs smoothly on all sides. Whether in-​house or free-
lance, translators who become experienced in the client’s domain(s), or those
who specialise in it to start with, are highly valued and can aim to become
a client’s preferred supplier. In the case of freelance translators, they can
become ‘permanent freelancers,’ which helps them overcome the uncertainty
of freelance work to some extent. Another measure to ensure quality is that
all translations, whether in-​house or freelance, are revised in-​house.
The company’s terminology management system, Quickterm, can be
accessed directly or through CAT tools such as Trados Studio and Acrolinx.
Quickterm stands out from other termbase software in that, for example, it
allows translators to create concept maps to help with understanding of the
subject matter of a domain. In the termbase of Eurocom UK’s client company
Blum, we found that definitions are deemed important and Blum’s termbases
include definitions at two levels: a definition in English is included at Concept
level, whereas a definition in any other language is included at Term level.
Blum’s termbases are concept-​oriented and include a range of data categories
at all levels.
Terminological verification is considered essential and, therefore, rather
than being a feature in Quickterm, the company have developed Checkterm
as its terminological quality assurance program. It differs from the termino-
logical features of CAT tools in that it has linguistic information which
improves terminological queries and searches and reduces the number of
false positives in term searches. Making use of linguistic information, the
Document analysis feature in Checkterm can help to what extent a source
text complies with the relevant company terminology before it is translated,
which shows us that terminological consistency begins with the source text.
This feature can also spot any errors to be amended in the source text before
translation.
The Kaleidoscope Group is also unique in that it recognises that there is a
wealth of terminological and other useful information in translator’s queries,
although project managers may find it tedious to have to reply to the same
queries at different times by different translators. In order to recycle and reuse
the treasure trove of information stored in translators’ queries, the company
has developed Smartquery as a translation query management system. By
creating a knowledge repository of all the queries raised by clients, experts or
 Terminological Best Practices and Quality Management 237

translators, communication among all parties involved in a translation pro-

ject is enhanced and more effective. Project managers will not have to answer
the same query twice because translators can access the repository as a kind
of Q&A resource for future reference.
Eurocom UK’s translators deal with phraseology mostly in the form of
clients’ slogans and marketing materials. We learned that phrases or slogans
for marketing purposes are stored in termbases rather than in translation
memories so that they are ranked higher than translation memory matches
during translation. The linguistic information built into Checkterm ensures
that its term recognition capabilities are enhanced compared to standard
term recognition features in CAT tools.
At Terra Translations there are no dedicated terminologists but there are
QA Managers instead, who oversee all processes in a translation project,
not only linguistic ones. QA Managers work alongside translators, revisers
and project managers and are ultimately responsible for delivering high
quality translations. Once revisers have revised a translation, QA Managers
perform a second revision. If desktop publishing is applied, QA Managers
will perform a third revision of the final text. QA Managers are also in
charge of approving terms and their translations in the relevant termbases.
Another quality measure in place at Terra is that glossaries are used as a
subset of a termbase when client approval or feedback is needed during a
translation project, and help Terra staff to gain a better overview of the
clients’ terminological needs and requirements. In addition, glossaries can
help Terra staff to standardise a client’s terminology at account level, i.e.
a client may have several ‘accounts’ with Terra, and the terminological
requirements of each one may be different. Ensuring terminological consist-
ency at the start of a project contributes to ensuring quality and consistency
during translation too.
Terra uses memoQ as its default terminology management tool (unless the
client requests using a different tool). This means that termbases created in
memoQ are concept-​oriented and have data categories at all levels. Termbases
are moderated by QA Managers and translators can only use moderated
termbases for translation projects. Therefore, definitions are at Term level
by default. A good lesson learned at Terra is that its staff are proficient with
text processing skills and used advanced features such as RegEx rules (see
Chapter 6) in memoQ’s QA profiles and auto-​translation rules. RegEx rules
also help Terra staff to deal with phraseology efficiently.
Overall, the case studies show that all the organisations and companies
are on a quest to achieve high terminological quality and consistency in
their translation work. Whether in commercial or institutional/​governmental
settings, there is dedicated staff, whether referred to as terminologist, QA
Manager or something else, who carry out quality and terminology checks as
well as translators and revisers. Translator specialisms are highly valued in
all our case studies, even if they are not always explicitly required. In some
cases, specialising in a domain can lead to benefits such as relatively regular
238 Terminology Management for Translators

work for freelance translators or becoming a preferred supplier for a given

client.
In all workplaces, outsourcing a large volume of translation work is the
main business model and this is in line with the findings in the recent survey
of the language industry by ELIS Research (2024: 32). However, it is equally
common in our case studies, to revise all translations in-​house, including
translations by freelancers. There is also a common aim in both institutional/​
governmental and commercial case studies to involve subject matter experts
and other stakeholders during the translation process. In the institutional/​
governmental case studies, this can mean bringing in external subject matter
experts to take part in a terminological committee or using the terminological
resources of other, expert organisations in a given subject matter.
Regardless of the termbase software used, all termbases in our case
studies are concept-​oriented and make use of the three-​tiered hierarchy (see
Chapter 4). Despite the arbitrariness of divisions between domains, even the
most streamlined termbases included information about domains and sub-​
domains and translators considered important to access this information for
contextual disambiguation during translation. Marking terms as forbidden
is another important tool to ensure the correct term is used in the correct
context.
Excel spreadsheets are widely used in all our case studies as either an
import/​export tool of extracted terms or as a collaborative tool to collect
preliminary bilingual or multilingual terminological information during a
translation project, which can then be batch imported into a termbase. Term
extraction is another crucial aspect of the terminological work of the trans-
lator in our case studies. As well as using term extraction for the standard
purpose of populating termbases with useful information, it is also used in
our case studies on existing termbases and documentation in order to take re-​
examine and consolidate existing terminology. A wide range of tools are used
in all the case studies but an important takeaway is that acquiring advanced
skills such as the use of RegEx can help translators make the most of any
software used. More generally, it is also advisable for translators to get to
know the software they are using to ensure they use it to its full potential.
Phraseological consistency is often considered alongside terminological
consistency in order to ensure that quality in translation is implemented hol-
istically throughout a client’s documentation, not just for terms. Finally, we
learned that, in fact, the quest for quality and consistency begins at source,
with the content creation stage in general and the authoring of the source text
in particular. Viewing quality and consistency holistically at the client’s level
can contribute to increasing the level of consistency in the client’s (translated)
documentation in a given translation project and over time.
Obvious limitations in this research include the fact that only a few
organisations and companies were studied. There may be many other kinds
of institutional, governmental and commercial organisations and companies,
and case studies there might yield new or additional insights into how
 Terminological Best Practices and Quality Management 239

professional translators ensure high quality in their terminological work.

Avenues for future research may include investigations into the termino-
logical work of interpreters, for example.

Terminology Management in Translation Competence Frameworks

Beyond the above results from the Case Studies, many scholars have studied
how translation competence works and proposed various models for how
translators acquire the necessary competences and knowledge, including
terminological competences. In these models, terminology and terminology
management may not be directly mentioned but they are usually included
or referred to under domain competences, information mining and docu-
mentation competences, and the selection and use of appropriate tools. It is
beyond the scope of this book to review these models in detail, but it is worth
highlighting a few. For example, it has been more than two decades since the
PACTE Group presented its own model (e.g. 2003) but they continue to be
active in the research of translation competence and its acquisition, among
other areas.
In her translation competence model, Göpferich (2009) specifically
mentions domain competence and tools and research competence. Kiraly
first suggested a social constructivist model (2000) which later he revised
into a four-​dimensional model of the emergence of translator competence
(2013, 2015). Kiraly’s model “reflects the compex interplay of competences
and their non-​parallel emergence over time” (PACTE Group et al., 2020)
but he “opts against identifying specific sub-​competences (whose labelling he
sees as reductionist) on the grounds that there is no consensus on which ones
actually exist” (2020).
In 2017, the European Master’s in Translation (EMT) network, a partner-
ship between the European Commission’s Directorate-​General for Translation
(DGT) and several European universities, published its Competence
Framework (2017), which was revised in 2022 (EMT, 2022: 7–​8) and which
included thematic knowledge, domain-​ specific knowledge, terminology,
phraseology and other skills under their Translation heading. For a com-
parison of the EMT competences to those in the TransCert project, CIUTI
and ISO, see Krajcso (2018). For a more in-​depth overview of studies in trans­
lation competence and translation competence acquisition, including Harris
and Sherwood (1978), Toury (1986, 1995), Presas (2000), Shreve (1997)
and Chesterman (1997), see PACTE Group (2020: 95–​102). More recently,
Quinci (2023) offers “a systematic and comprehensive account of translation
competence (TC), reflecting on its different models and conceptualisations.”
The EFFORT project (“Towards a European framework of reference for
translation”),8 an Erasmus+​project that ran from 2020 to 2023, developed a
reference framework for translation in a similar vein to the CEFR (Common
European Language Framework of Reference for Languages), which is typic-
ally applied to language teaching and learning, since there is no comparable
240 Terminology Management for Translators

framework for translation performance. The reference framework designed

by the EFFORT project describes the key competences in written translation
which includes terminological tasks at various levels and tasks. For example,
the EFFORT project suggests that “as of level B1, translators should develop
the capability to create ad hoc documentary resources: glossaries, termin-
ology databases, corpora, translation memories, etc.” (Hurtado Albir &
Rodríguez-​Inés, 2023: 17).

Terminology Management in Organisations and Research Projects

The importance of terminology management also features in a range of
research projects since the 2010s that focus on the professionalisation of
translator trainees and on bridging the gap between academia and the market-
place. The list of projects in this section is not exhaustive but it is meant to
illustrate the international importance and relevance of terminology manage-
ment for translators. For example, the OPTIMALE project (Optimising pro-
fessional translator training in a multilingual Europe), which ran from 2010
to 2013 and was coordinated by Université Rennes 2 in France, with funding
from the Erasmus academic network and the European Commission, involved
64 universities offering translator training and the EUATC (European Union
of Associations of Translation Companies) representing European transla-
tion service providers and aimed to “enhance the visibility and professional
relevance” of professional translator education and training in Europe
(OPTIMALE, 2013: 3). As part of OPTIMALE’s Work Package 6, the Train
the Trainers workshop in Brussels in 2013 focused on the use of tools and
technologies in the classroom and on terminology for domain specialisation
(2013: 10).
Following in the footsteps of the EMT and OPTIMALE, the OTCT pro-
ject (Optimising Translator Training through Collaborative Translation)
(2014–​2016), “took the form of an Erasmus+​Strategic Partnership of seven
academic institutions around Europe” (Hernandez Morin et al., 2017: 3).
The project originated from the work of Daniel Gouadec who implemented
“a project-​based pedagogy, with bi-​annual intensive technical translation
sessions known as Tradutech” (2017: 3), also coordinated by Rennes 2
University. One of the specific goals of the OTCT project was to “dissem-
inate this proven methodology for translator training” (2017: 3). Among
the competences that students develop during these intensive sessions are
“specialised knowledge and competences” (2017: 3) such as documenta-
tion and terminology mining, including developing “confidence in tackling
new highly specialized fields for translation and/​or terminology projects.”
Also arising from the EMT efforts and from a competences survey piloted
at the Université de Paris Diderot is the CATO (Competence Awareness
in TranslatiOn) project, which started in 2019 and which is a “longitu-
dinal study aimed at surveying translation students’ self-​ perception of
their competences” (Froeliger et al., 2023: 13), including terminological
 Terminological Best Practices and Quality Management 241

competences. Terminology management is also a research topic among

under-​resourced languages. For example, SADiLaR (South African Centre
for Digital Language Resources)9 is a “national centre supported by the
Department of Science and Innovation (DSI) as part of the new South African
Research Infrastructure Roadmap (SARIR).” SADiLaR focuses specifically
on all official languages of South Africa and its website contains a list of
the publications by their Digital Humanities researchers, including work on
term extraction (Mlambo et al., 2021), a medical glossary (Malele, 2021)
and COVID-​19 multilingual terminology (SADiLaR, 2021), to name but a
few. Terminology management is also of interest in publications concerning
terminography (e.g. Alberts, 2000; Joffe & de Schryver, 2005), education
(e.g. Alberts, 2014) and language planning and policy in these languages (e.g.
Ramuedzisi et al., 2019).

Terminology Management and ISO Standards

Terminological inconsistency is one of the main challenges faced by
specialised translators. In order to avoid confusion and achieve clear,
unambiguous communication in specialised domains, as proposed by
the General Theory of Terminology (see Chapter 3), two terminological
activities can be carried out: standardisation and harmonisation (Scarpa,
2020: 59–​60). According to Sager (2001c: 255), the “standardization of
terminology is a two-​step process which consists of (a) unifying and fixing
each referent [concept] and (b) unifying and standardizing its designation
[term].” A “primary function of a standardized terminology shall be to indi-
cate preferred, admitted and deprecated [forbidden] terms” (2001c: 255).
Standardising terminology is a task for technical committees in standards
organisations, but also in companies, organisations, professional associ-
ations, etc. (Schmitz, 2009). In commercial translation settings, however,
consistent terminology is sought after but ‘formal’ standardisation is not a
goal in these settings, as “defining concepts, making semantic relations and
creating concept systems […] does not reflect the reality faced by corporate
terminologists” (Warburton, 2021: 9).
Standardisation is usually introduced “when there are alternative
designations (terms) for the same concept” (Scarpa, 2020: 59). Typically,
the choice of term to use depends on factors such as economy of lan-
guage, transparency and frequency of use (2020: 59). However, not every
specialised domain has standardised terminology. For example, while “the
electrotechnical domain has a highly standardised vocabulary, the domain of
thermal insulation generally shows limited terminological standardisation”
(Giacomini, 2022: 216). The main international body that sets international
standards is ISO (International Organization for Standardization),10 founded
in 1947 and currently headquartered in Geneva. Its three official languages
are English, French and Russian (Sager, 2001c: 257) (see also Box 10.2).
242 Terminology Management for Translators

Box 10.2 ISO Technical Committee 37 (ISO/​TC 37)

• It is the ISO committee that prepares standards for terminology and

terminological work.
• Created in 1952, under the leadership of Eugen Wüster (Infoterm,
2013) (see also Chapter 3).
• Wüster ran the Secretariat until 1971 (Infoterm, 2013).
• TC 37 (aka. “Language and Terminology”) aims “to find out and
formulate general principles of Terminology and Terminological
Lexicography” (Temmerman, 2022: 332).
• With the creation of TC 37, the ISO “was acknowledging the need
for terminology standardisation” (Cabré, 2023a: 52).
• It is made up of several sub-​committees, some of which are organised
into several working groups.

ISO standards are valid for a particular country only after the country’s
standards organisation has endorsed them (Sager, 2001c: 257). Each country
“can decide whether to further enhance the relevance of ISO standards by
adopting them as national standards written in languages different from ISO’s
three official languages” (Scarpa, 2020: 60). The content of the “national norms
must then be made compatible and compliant with the corresponding inter-
national standards by means of a process called harmonisation of standards”
(2020: 60). Using Scarpa’s example, in the European Union, the “resulting
harmonised standards begin with the suffixes BSI ISO (or BSI EN ISO) in the
United Kingdom, DIN ISO (or DIN EN ISO) in Germany, etc.” (2020: 60).
There is no “single standard that covers all best practice and models
for terminology management but there are several standards that are very
useful in this context” (Popiolek, 2015: 346). For example, the ISO 704
standard, Terminology work –​principles and methods, covers principles of
systematic terminology work, whereas ISO 860 covers Terminology work –​
Harmonization of concepts and terms and ISO 1087:2019 (Terminology work
and terminology science –​Vocabulary) provides a “systematic description
of the concepts related to terminology work and terminology science.”11
Standards can also guide translators in the selection of the correct variant
(Scarpa, 2020: 60); and standards can help translators with a theoretical
basis, e.g. ISO 704 defines concept as a “mental representations of objects”
(Brandt, 2021); concept orientation is another ISO standard (see ISO 16642,
ISO 30042 and ISO 26162) (Warburton, 2021: 23); the ISO 30042 Termbase
Exchange Standard promotes interoperability and reusability of termbases
among CAT tools through the .TBX format (Di Nunzio & Vezzani,
2021: 183) (see also Chapter 5); standards ISO 12616 and ISO 12620 pro­­
vide specifications for data categories for terminological entries (Popiolek,
2015: 246); standard ISO 2000 “lists desirable criteria for term formation as
 Terminological Best Practices and Quality Management 243

principles of term formation” (Kageura, 2022: 462). Standards are reviewed

regularly, and the latest version of the standard can be included with the
standard. For example, the current version of ISO 704 is ISO 704:2022,12
whereas the latest version of ISO 860 is ISO 860:2007.13 Two “major new
terminology standards” (Wright, 2022: 96) appeared in 2023: ISO 20539 for
Translation, interpreting and related technology and ISO 24183 for Technical
Communication –​Vocabulary.
There are also ISO standards for the representations of names of coun-
tries (ISO 3166) and languages (ISO 639); these standards are often followed
by translators (and their project managers, revisers, clients, etc.) to name
files and folders in a meaningful way. For example, using the two-​letter ISO
codes for countries (ISO 3166-​1) and languages (ISO 639-​1), a file named
Translation fr-​CA_​en-​NZ, might indicate that the source language is French
(Canada), in lowercase, and the target language English (New Zealand),
in uppercase. For these standards to work, languages should always be
referred to in lowercase and countries in uppercase (rather than the other
way around). Finally, it is also worth pointing out that quality standard ISO
17100:2015 (Translation services –​Requirements for translation services)
requires terminological consistency and semantic accuracy in the translation
process (Prieto Ramos 2020c: 128) and that ISO 11669:2024 (Translation
projects –​General guidance) also terminological consistence as important.
For language services providers, being “ISO 17100-​certified can be a way for
them to demonstrate that they meet widely recognised, high-​quality profes-
sional standards” (Penet, 2024: 134). For more on standardisation, motiv­
ation, methods, benefits and limitations, see Sager (2001c). A more detailed
description of ISO 704, ISO 12620, ISO 16642 and ISO 12616 is provided
by Wright (2006). A more up-​to-​date overview of terminology and standards
can be read in Wright (2022).

Finally… Pushing the Boundaries

Terminology is “arguably the most important component of specialized
communication” (Austermühl, 2012: 62) and termbases have now become
“critical business tools for those who seek quality and consistency in [their]
content” (Hartley, n.d.), although “managing your linguistic assets such as
glossaries or style guides was never a [company’s] top priority” (Malinowski,
2023). It should be now. In an era where AI tools and automation are
“shooting up like mushrooms” (ELIS Research, 2024: 41), the demand for
quality in translation and the need for carefully curated linguistic assets is
unlikely to diminish but “likely to open up new opportunities for linguists”
(Rothwell et al. 2023: 217), although linguists should remain alert to poten­
tial risks and threats (2023: 217). Despite “the potential benefits of LLMs, the
expertise of human linguists will still be needed” (Pantcheva, 2023), albeit in
different ways. Thus, some are predicting that the translators’ work “will be
augmented by prompt design tasks and testing learning strategies to extract
244 Terminology Management for Translators

the best AI performance” (2023) but, ultimately, the “use of LLMs should
be viewed as a supplement to human linguists, rather than a replacement”
(2023). As technology becomes more efficient and more useful, translators
can look forward to more collaborative and more proactive terminology
management practices in which stakeholders such as domain experts, users
of translation, etc., might also be involved (Marshman, 2017). Marshman’s
prediction that tomorrow’s terminology will be more visible and more valued
(2017) seems reasonable and plausible. As the proverb goes, “if you can’t
beat them, join them”: at a time of content-​driven growth, successful com-
panies and language professionals will be those that excel in their manage-
ment of knowledge and data, and break down the language barrier in order
to “truly connect with end-​users” (Malinowski, 2023). Or, in other words,
“[i]‌f content is king, terminology is queen” (Fleischmann, 2024).
The current sentiment among linguists, as collected in recent surveys,
indicates that the language industry is a “reasonably good industry to be
a part of” (ELIS Research, 2024: 45) and, considering it generates some
“25 billion euro (according to Slator.com) or even 60 billion (according to
Nimdzi), [it] is, by definition, a valuable source of employment” (2024: 43)
for linguists with translation and terminology skills. We are only scratching
the surface with this book when it comes to highlighting the importance of
terminological quality and consistency in translation. There is plenty of scope
to improve the way we understand specialised knowledge, the way we mine
and store it and the way we deal with it in translation but, with this book, we
hope to have facilitated opening the doors to further debate and research in
this area. The remaining question is not whether institutions, organisations,
companies and businesses (including freelance translation businesses) can
afford to carry out terminology management, but rather whether they can
afford not to.

Notes
1 https://​short​url.at/​z73DM
2 www.europ​arl.eur​opa.eu/​tran​slat​ion/​en/​term​inol​ogy/​what-​we-​do
3 www.gala-​glo​bal.org/​knowle​dge-​cen​ter/​profe​ssio​nal-​deve​lopm​ent/​sigs
4 https://​term​net.eu/​
5 www.eaft-​aet.net/​en/​term​inol​ogy-​awa​rds
6 www.worldb​ank.org/​en/​about/​care​ers/​the-​young-​profes​sion​als-​prog​ram/​how​toap​ply
7 https://​ter​mino​rgs.net/​Term​inol​ogy-​Eve​nts.html
8 effortproject.eu
9 https://​sadi​lar.org/​en/​about/​
10 www.iso.org/​standa​rds.html
11 www.iso.org/​obp/​ui/​#iso:std:iso:1087:ed-​2:v1:en
12 www.iso.org/​stand​ard/​79077.html
13 www.iso.org/​stand​ard/​40130.html
Bibliography

Advantage Austria. (2013). European Inventor Award 2013 goes to Julius Blum
GmbH from Austria. Retrieved 2 November, 2023, from https://​web.arch​ive.org/​
web/​201​6042​0142​857/​www.advan​tage​aust​ria.org/​intern​atio​nal/​zent​ral/​news/​aktu​
ell/​201​3061​9_​Eu​rop._​Erfi​nder​prei​s_​fu​er_​B​lum.en.html
Ahmad, K. (1997). Terminology management and knowledge processing in lesser-​
used languages. Proceedings of the First International Congress on Terminology,
San Sebastián, Spain.
Ahmad, K. (2006). Metaphors in the languages of science? In M. Gotti &
D. S. Giannoni (Eds.), New trends in specialized discourse analysis (pp. 197–​220).
Peter Lang.
Ahmad, K., & Rogers, M. (2001). Corpus linguistics and terminology extraction.
In S. E. Wright & G. Budin (Eds.), Handbook of terminology management vol. 2
(pp. 725–​760). John Benjamins.
Akkach, J. (2016). English medical terminology: Word combinations, semantics and
specialized translation. T21N Translation in Transition, 1, 1–​32.
Al-​
Kasimi, A. M. (2019). The history of Arabic lexicography and terminology.
In A. Alsulaiman & A. Allaithy (Eds.), Handbook of terminology. Volume
2: Terminology in the Arab world (pp. 7–​30). John Benjamins.
Alaoui, A. (2023). How equivalent is equivalence in Arabic-​English legal translation?
In Ł. Biel & H. Kockaert (Eds.), Handbook of Terminology, volume 3: Legal ter-
minology (pp. 206–​223). John Benjamins.
Alberts, M. (2000). Terminology management at the National Language Service.
Lexikos, 9(AFRILEX-​reeks/​series 10: 2000), 234–​251.
Alberts, M. (2001). Lexicography versus terminography. Lexikos, 11(1), 71–​84.
www.ajol.info/​index.php/​lex/​issue/​view/​6948
Alberts, M. (2014). Terminology development at tertiary institutions: A South African
perspective Lexikos, 24(AFRILEX-​reeks/​series 24: 2014), 1–​26.
Allam, R. A. B. A. (2023). Localized or globalized: Assessing the localization level
of IKEA Egypt 2023 English and Arabic websites. CDELT Occasional Papers in
the Development of English Education, 83(1), 313–​349. https://​doi.org/​10.21608/​
OPDE.2023.325​340
Allen, S. (2007). The future of Inuktitut in the face of majority languages: Bilingualism
or language shift? Applied Psycholinguistics, 28, 515–​536. https://​doi.org/​10.1017.
S01427​1640​7070​282
246 Bibliography

Almqvist, I., & Hein, A. S. (2000). A language checker of controlled language and its
integration in a documentation and translation workflow. ASLIB, Proceedings of
Translating and the Computer 22, London, November.
Alsulaiman, A., & Allaithy, A. (Eds.). (2019). Handbook of terminology. Volume
2: Terminology in the Arab world. John Benjamins.
Andrews, T., & Prys, G. (2016). Terminology standardization in education and the
construction of resources: The Welsh experience. Education Sciences, 6(1), 1–​15.
Anthony, L. (2022). What can corpus software do? In A. O’Keeffe & M. McCarthy
(Eds.), The Routledge handbook of corpus linguistics (2nd ed., pp. 103–​125).
Routledge.
Antia, B. E., Budin, G., Picht, H. et al. (2005). Shaping translation: A view from
terminology research. Meta. Journal des traducteurs. Translator’s Journal, 50(4).
https://​doi.org/​https://​doi.org/​10.7202/​01990​7ar
Archer, D., & Rayson, P. (Eds.). (2006). Corpus linguistics around the world.
Rodopi.
Argos Multilingual. (2018). Top tips on terminology management. Retrieved 21
August, 2024, from www.argosm​ulti​ling​ual.com/​blog/​top-​tips-​on-​term​inol​ogy-​
man​agem​ent
Argos Multilingual. (2021). Terminology management. Retrieved 9 August, 2021,
from www.argosm​ulti​ling​ual.com/​term​inol​ogy-​man​agem​ent
ATA Chronicle. (2019). Regular expressions: An introduction for translators.
Retrieved 30 April, 2024, from www.ata​net.org/​resour​ces/​regu​lar-​expr​essi​ons-​an-​
intro​duct​ion-​for-​tran​slat​ors/​
Aubin, S., & Hamon, T. (2006). Improving term extraction with terminological
resources. In T. Salakoski, F. Ginter, S. Pyysalo, et al. (Eds.), Advances in natural
language processing (pp. 380–​387). Springer Verlag. https://​doi.org/​https://​doi.org/​
10.1007/​1181​6508​_​39
Aussant, L. (2019). Respecter la non-​ binarité de genre en français. Retrieved
5 October, 2023, www.nos​lang​ues-​ourla​ngua​ges.gc.ca/​fr/​blo​gue-​blog/​respec​ter-​la-​
non-​binar​ite-​de-​genre-​fra
Austermühl, F. (2001). Electronic tools for translators. St. Jerome Publishing.
Austermühl, F. (2010). A collaborative approach to the teaching of terminology man­
agement. T21N Translation in Transition, 9. http://​t21n.com/​homep​age/​artic​les/​
T21N-​2010-​09-​Aust​ermu​ehl.pdf
Austermühl, F. (2012). Using concept mapping and the web as corpus to develop ter­
minological competence among translators and interpreters. Translation Spaces,
1(1), 54–​80. https://​doi.org/​10.1075/​ts.1.09aus
Austermühl, F., & Mirwald, C. (2010). Images of translators in localization dis­
course. Retrieved 23 March, 2011, from www.t21n.com/​homep​age/​artic​les/​T21N-​
2010-​08-​Aust​ermu​ehl,Mirw​ald.pdf
Avila, C. (2018). How to use the translation alignment tool in Trados Studio.
Retrieved 8 October, 2024, from www.tra​dos.com/​blog/​how-​to-​use-​the-​new-​enhan​
ced-​tran​slat​ion-​alignm​ent-​tool-​in-​tra​dos-​stu​dio/​
Bakaric, M. B., Babic, N., & Matetic, M. (2021). Application-​based evaluation of
automatic terminology extraction. International Journal of Advanced Computer
Science and Applications, 12(1), 18–​27.
Baker, M. (1995). Corpora in translation studies: An overview and some suggestions
for future research. Target, 7(2), 223–243.
 Bibliography 247

Bardarska, R. (2009). The work of the translator at DGT. Conference for Translators
from Candidate Countries, [online]. Retrieved 10 May 2025, from www.slideserve.
com/christoffer-carus/the-work-of-the-translator-at-dgt
Barlow, M. (2002). ParaConc: Concordance software for multilingual parallel corpora.
Third International Conference on Resources and Evaluation (LREC), Workshop
on Language Resources in Translation Work and Research, Las Palmas, Spain.
Baroni, M., & Bernardini, S. (2004). BootCaT: Bootstrapping corpora and terms
from the web. Proceedings of the Fourth International Conference on Language
Resources and Evaluation (LREC), Lisbon, Portugal, pp. 1313–1316.
Beaudoin, L. (2009). Legal translation in Canada: The genius of legal language(s).
In F. Olsen, A. Lorz, & D. Stein (Eds.), Translation issues in language and law
(pp. 136–​144). Palgrave MacMillan.
Beheshty, L. (2018). Effective query management –​the key to successful translation
projects. Retrieved 30 April, 2018, from www.linke​din.com/​pulse/​effect​ive-​query-​
man​agem​ent-​key-​suc​cess​ful-​tran​slat​ion-​behes​hty-​miti/​
Béjoint, H. (1988). Scientific and technical words in general dictionaries. International
Journal of Lexicography, 1(4), 354–​368.
Bellino Machado, M. J. (2015). DGT-​OmegaT-​2014: DGT’s CAT environment, a
translator’s guide. Retrieved 7 June, 2018, from http://​ec.eur​opa.eu/​tran​slat​ion/​por​
tugu​ese/​magaz​ine/​docume​nts/​folha4​5_​om​ega-​t_​tran​slat​or_​g​uide​_​en.pdf
Bendegard, S. (2012). Plain language in Swedish translations of EU legal texts.
Tijdschrift voor Skandinavistiek, 33, 41–​55.
Bergman, M. K., Smither, M., & Bushee, W. (2000). The deep web: Surfacing hidden
value. Retrieved 19 October, 2024, from https://​resour​ces.mpi-​inf.mpg.de/​d5/​teach​
ing/​ws01​_​02/​pros​emin​arli​tera​tur/​deepwe​bwhi​tepa​per.pdf
Bernaerts, J. (2021). Translation in administrative interactions: Policies and practices
at the local level in the Dutch language area of Belgium. In M. Bourguignon, B.
Nouws, & H. van Gerwen (Eds.), Translation policies in legal and institutional
settings (pp. 223–​246). Leuven University Press.
Bernardini, S. (1997). A ‘trainee’ translator’s perspective on corpora. Paper presented
at Corpus Use and Learning to Translate, Bertinoro, November 1997.
Bernardini, S. (2022). How to use corpora for translation. In A. O’Keeffe & M.
McCarthy (Eds.), The Routledge handbook of corpus linguistics (2nd ed., pp. 485–​
498). Routledge.
Bernth, A., McCord, M. C., & Warburton, K. (2002). Terminology extraction for
global content management. IBM Research Report, RC 22615(W0210–​206),
1–​14. Retrieved 14 January, 2010, from www.staff.it.uts.edu.au/​~asi​dhu/​32126/​
pap​ers/​Rc22​615.pdf
Berry, M. W., & Kogan, J. (2010). Text mining: Applications and theory. John Wiley
and Sons.
Bertels, A. (2022). Terminology and distributional analysis of corpora. In P. Faber &
M.-​C. L’Homme (Eds.), Theoretical perspectives on terminology: Explaining terms,
concepts and specialized knowledge (pp. 311–​327). John Benjamins.
Best, D. A. (2017). Equal respect or equal effect: The untranslatability of legal termin­
ology and its impact on language parity in the EU. In P. Faini (Ed.), Terminological
approaches in the European context (pp. 271–​287). Cambridge Scholars Publishing.
Beukes, A.-​M. (2009). Translation in South Africa: The politics of transmission.
Southern African Linguistics and Applied Language Studies, 24(1), 1–​6.
248 Bibliography

Biber, D. (2009). A corpus-​driven approach to formulaic language in English: Multi-​

word patterns. International Journal of Corpus Linguistics, 14(3), 275–​311. Retrieved
24 October, 2009, from https://jan.ucc.nau.edu/biber/Biber/Biber_2009.pdf
Biel, Ł., & Kockaert, H. (2023). Introduction: Legal terminology. In Ł. Biel & H.
Kockaert (Eds.), Handbook of terminology. Vol. 3: Legal terminology (pp. 1–​14).
John Benjamins.
Biel, Ł., & Martín Ruano, R. (2023). Institutions’ outreach to and involvement
with universities: How international organizations collaborate with universities in
training translators. In T. Svoboda, Ł. Biel, & V. Sosoni (Eds.), Institutional trans-
lator training (pp. 151–​169). Routledge.
Bieswanger, M. (2022). Slope difficulty levels explained. www.int​ersp​ortr​ent.com/​en/​
rente​rtai​ner-​blog/​slope-​dif​fi cu​lty-​lev​els-​explai​ned~14364​578
Bingham, L. (2024). In the loop. ITI Bulletin (September–​October 2024), 12–​14.
Blum. (2023). Facts and figures: The 2022/​2023 financial year. Retrieved 2 November,
2023, from www.blum.com/​gb/​en/​comp​any/​facts-​figu​res/​
Blum. (n.d.). History: In motion since 1952. Retrieved 2 November, 2023, from
www.blum.com/​gb/​en/​comp​any/​comp​any-​hist​ory/​
Bononno, R. (2000). Terminology for translators: An implementation of ISO 12620.
Meta, 45(4), 646–​669.
Borja Albi, A. (2013). Freelance translation for multinational corporations and law
firms. In A. Borja Albi & F. Prieto Ramos (Eds.), Legal translation in context.
Professional issues and prospects (pp. 53–​74). Peter Lang.
Borja Albi, A., & Prieto Ramos, F. (Eds.). (2013). Legal translation in con­
text: Professional issues and prospects. Peter Lang.
Boughton, J. (2014). The IMF and the force of history. Events that have shaped the
global institution. IMF Publication Services.
Bourigault, D., & Slodzian, M. (1999). Pour une terminologie textuelle. Terminologies
nouvelles, 19, 20–​32.
Bowker, L. (1998). Using specialized monolingual native-​language corpora as a trans­
lation resource: A pilot study. Meta: Translators’ Journal, 43(4), 631–​651.
Bowker, L. (2002a). Computer-​aided translation technology: A practical introduc­
tion. University of Ottawa Press.
Bowker, L. (2002b). An empirical investigation of the terminology profession in
Canada in the 21st century. Terminology, 8(2), 283–​308.
Bowker, L. (2003). Terminology tools for translators. In H. Somers (Ed.), Computers
and translation: A translator’s guide (pp. 49–​65). John Benjamins.
Bowker, L. (2005). Productivity vs quality? A pilot study on the impact of translation
memory systems. Localisation Focus, 4(1), 13–​20.
Bowker, L. (2015). Terminology and translation. In H. Kockaert & F. Steurs (Eds.),
Handbook of terminology, vol. 1 (pp. 304–​323). John Benjamins.
Bowker, L. (2017). Lexicography and terminology. In P. A. Fuertes-​Olivera (Ed.), The
Routledge handbook of lexicography (pp. 138–​151). Routledge.
Bowker, L. (2019). Terminology. In M. Baker & G. Saldanha (Eds.), Routledge
encyclopedia of translation studies (pp. 579–​583).
Bowker, L. (2020). Terminology management. In E. Angelone, M. Ehrensberger-​Dow,
& G. Massey (Eds.), The Bloomsbury companion to language industry studies (pp.
261–​283). Bloomsbury Academic.
 Bibliography 249

Bowker, L. (2022). Multidimensionality. In P. Faber & M.-​ C. L’Homme (Eds.),

Theoretical perspectives on terminology: Explaining terms, concepts and specialized
knowledge (pp. 127–​147). John Benjamins.
Bowker, L., & Marshman, E. (2011). Better integration for better preparation: Bringing
terminology and technology more fully into translator training using the CERTT
approach. In A. Alcina (Ed.), Teaching and learning terminology: New strategies
and methods (pp. 61–​88). John Benjamins.
Bowker, L., & Pearson, J. (2002). Working with specialized language: A practical
guide to using corpora. Routledge.
Brace, C. (2000). Language automation at the European Commission. In R. C.
Sprung & S. Jaroniec (Eds.), Translating into success: Cutting-​edge strategies for
going multilingual in a global age (pp. 219–​224). John Benjamins.
Brandt, A. (2021). Getting started with terminology management. Retrieved
7 August, 2021, from atanet.org/​growing-​your-​career/​getting-​started-​with-​
terminology-​management/​
Brekke, M. (2004). Linguistic aspects of the translation of scientific and technical
texts. In H. Kittel, J. House, & B. Schultze (Eds.), Translation. An international
encyclopedia of translation studies /​Übersetzung: Ein internationales Handbuch
zur Übersetzungsforschung (Vol. 2, pp. 619–​635). Walter de Gruyter.
Brooks, D. (2000). What price localization? Managing costs at Microsoft. In R. C.
Sprung & S. Jaroniec (Eds.), Translating into success: Cutting-​edge strategies for
going multilingual in a global age (pp. 43–​57). John Benjamins.
Browne, R. (2023). All you need to know about ChatGPT, the AI Chatbot that’s got
the world talking and tech giants clashing. Retrieved 13 October, 2024, from www.
cnbc.com/​2023/​02/​08/​what-​is-​chat​gpt-​viral-​ai-​chat​bot-​at-​heart-​of-​micros​oft-​goo​
gle-​fight.html
Brownlie, S. (2017). Institutional memory and translating and the DGT. The
Translator, 23(1), 1–​16.
Bucher, A.-​L. (2017). Terminology work in Sweden. In M. Małachowicz & S. Grucza
(Eds.), Polskie i europejskie nurty terminologiczne (Polish and European termino-
logical trends) (pp. 266–​282). Warsaw University.
Buendía-​Castro, M., & López-​Rodríguez, C. I. (2013). The web for corpus and the
web as corpus in translator training. New Voices in Translation Studies, 10(2013),
54–​71.
Buendía-​Castro, M., Montero Martínez, S., & Faber, P. (2014). Verb collocations and
phraseology in EcoLexicon. Yearbook of Phraseology, 5(1), 57–​94.
Buira, A. (Ed.). (2005). The IMF and the World Bank at sixty. Anthem Press.
Bundgaard, K., & Christensen, T. P. (2019). Is the concordance feature the new black?
A workplace study of translators’ interaction with translation resources while post-​
editing TM and MT matches. JoSTrans, the Journal of Specialised Translation,
31(31), 14–​37.
Bustamante, G., Oncevay, A., & Zariquiey, R. (2020). No data to crawl? Monolingual
corpus creation from PDF files of truly low-​resource languages in Peru. Proceedings
of the 12th Conference on Language Resources and Evaluation (LREC), Marseille,
France, 11–​16 May.
Buzan, T., & Buzan, B. (2010). The mind map book. BBC Active.
250 Bibliography

Byrne, J. (2012). Scientific and technical translation explained. A nuts and bolts guide
for beginners. Routledge
Cabezas García, M., & León-​Araúz, P. (2020). Term variation in terminographic
resources: A review and a proposal. Proceedings of the XIX Congress of the
European Association for Lexicography (EURALEX), Alexandroupolis, Greece,
7–​11 September.
Cabré, M. T. (1996). Terminology today. In H. Somers (Ed.), Terminology, LSP and
translation. Studies in language engineering in honour of Juan C. Sager (pp. 15–​
33). John Benjamins.
Cabré, M. T. (1999). Terminology: Theory, methods and applications (J. A. DeCesaris,
Trans.). John Benjamins. (La Terminologia. La teoria, els mètodes, les aplicacions.
Barcelona, Empúries 1992.)
Cabré, M. T. (2000). Terminologie et linguistique: La théorie des portes. Terminologies
nouvelles: Terminologie et diversité culturelle, 21, 10–​15.
Cabré, M. T. (2002). Textos especializados y unidades de conocimiento: Metodología
y tipologización. In J. García Palacios & M. T. Fuentes Morán (Eds.), Texto,
terminología y traducción (pp. 15–36). Ed. Almar.
Cabré, M. T. (2003). Theories of terminology: Their description, prescription and
explanation. Terminology, 9(2), 163–​200.
Cabré, M. T. (2005). La terminología, una disciplina en evolución: Pasado, presente
y algunos elementos de futuro. Revista Debate Terminológico, [online]. https://​seer.
ufrgs.br/​rit​erm/​arti​cle/​downl​oad/​21286/​15349
Cabré, M. T. (2009). The communicative theory of terminology: A linguistic approach
of terms. Revue française de linguistique appliquée, 2, 9–​15.
Cabré, M. T. (2023a). Terminology. Cognition, language and communication. John
Benjamins.
Cabré, M. T. (2023b). The polyhedricity principle: Articulation between discourse,
cognition and language in terminology (1). In M. T. Cabré (Ed.), Terminology.
Cognition, language and communication (pp. 69–​90). John Benjamins.
Caffrey, C., & Valentini, C. (2019). Applications of technology in the Patent
Cooperation Treaty (PCT) Translation Division of the World Intellectual Property
Organization (WIPO). In M. O’Hagan (Ed.), The Routledge handbook of transla-
tion and technology (pp. 127–​147). Routledge.
Cagiltay, K., & Bichelmeyer, B. (2000). Differences in learning styles in
different cultures: A qualitative study. Paper presented at the Annual
Meeting of the American Educational Research Association, New Orleans,
LA, 24–28 April 2000.
Calvo Rigual, C., & Calvi, M. V. (2014). Translation and lexicography: A necessary
dialogue. MonTI Monografías de Traducción e Interpretación, 6(2014), 37–​62.
Campo, Á., & Cormier, M. C. (2005). The role of the communicative approach in the
development of terminology. Meta. Journal des traducteurs. Translator’s Journal,
50(4). https://​doi.org/​ https://​doi.org/​ https://​doi.org/​10.7202/​01991​3ar
Canadian Heritage. (2021). The first Commissioner and Directors of Indigenous
languages are appointed. Retrieved 5 October, 2023, from www.can​ada.ca/​en/​
canad​ian-​herit​age/​news/​2021/​06/​the-​first-​commi​ssio​ner-​and-​direct​ors-​of-​ind​igen​
ous-​langua​ges-​are-​appoin​ted.html
Canavese, P. (2021). Investigating the status of Italian as an ‘official minority lan­
guage’ within the Swiss multilingual institutional system. In M. Bourguignon, B.
 Bibliography 251

Nouws, & H. van Gerwen (Eds.), Translation policies in legal and institutional
settings (pp. 133–​155). Leuven University Press.
Candel, D. (2022). General principles of Wüster’s General Theory of Terminology.
In P. Faber & M.-​ C. L’Homme (Eds.), Theoretical perspectives on termin-
ology: Explaining terms, concepts and specialized knowledge (pp. 37–​59). John
Benjamins.
Candel Mora, M. Á. (2017). Criteria for the integration of term banks in the profes­
sional translation environment. Sendebar. Revista de Traducción e Interpretación,
28, 243–​260.
Cao, D. (2023). Terminological features of the Chinese legal language. In Ł. Biel &
H. Kockaert (Eds.), Handbook of terminology, volume 3: Legal terminology (pp.
173–​190). John Benjamins.
Cerrella Bauer, S. (2015). Managing terminology projects. Concepts, tools and
methods. In H. Kockaert & F. Steurs (Eds.), Handbook of terminology (pp. 324–​
340). John Benjamins.
Chabata, E. (2023). Reflections on the development of corpora for Zimbabwe’s
understudied languages. Studies in African Linguistics, 52(13), 317–​326.
Champagne, G. (2004). The economic value of terminology: An exploratory
study. Report submitted to the Translation Bureau of Canada, Public Works and
Government Services Canada. Retrieved 1 August, 2021, from www.dant​erm.dk/​
docs/​Econo​micV​alue​Term​inol​ogy-​1.pdf
Chesterman, A. (1997). Memes of translation: The spread of ideas in translation
theory. John Benjamins.
Chirimbu, S., & Barbu-​Chirimbu, A. (2014). Translating the EU terminology and
community documents into Romanian. Linguistic and Philosophical Investigations,
13, 650–​657.
Choi, J. (2018). Introducing a Korean-​English parallel corpus for the standardisation
of South Korean government documents: A terminology focus. In É. Poirier & D.
Gallego-​Hernández (Eds.), Business and institutional translation: New insights and
reflections (pp. 33–​44). Cambridge Scholars Publishing.
Choi, J. (2022). Government translation in South Korea. Routledge.
Church, K. W., & Hanks, P. (1990). Word association norms, mutual information
and lexicography. Computational Linguistics, 16(1), 22–​29.
Ciapuscio, G., & Kuguel, I. (2002). Hacia una tipología del discurso
especializado: Aspectos teóricos y aplicados. In J. García Palacios & M. T. Fuentes
Morán (Eds.), Texto, terminología y traducción (pp. 37–73). Ed. Almar.
Colson, J.-​P. (2004). Phraseology and computational corpus linguistics: From
theory to a practical example. In H. Bouillon (Ed.), Langues à niveaux multiples.
Hommage au professeur Jacques Lerot à l´occasion de son éméritat. (Bibliothèque
des Cahiers de l´Institut de Linguistique de Louvain, 113. pp. 135–​145). Peeters.
Colson, J.-​P. (2008). Toward a general theory of phraseology: A plea for a web-​based,
structural and statistical approach. Proceedings of EUROPHRAS 2008, Helsinki.
Combe, K. (2000). Localization at Hewlett-​ Packard’s Laserjet Solutions Group.
In R. C. Sprung & S. Jaroniec (Eds.), Translating into success: Cutting-​edge strat-
egies for going multilingual in a global age (pp. 97–​109). John Benjamins.
Condamines, A. (2017). The emotional dimension in terminological variation.
In P. Drouin, A. Francœur, & A. Picton (Eds.), Multiple perspectives on termino-
logical variation (pp. 11–​30). John Benjamins.
252 Bibliography

Condamines, A., & Picton, A. (2022). Textual terminology: Origins, principles and
new challenges. In P. Faber & M.-​C. L’Homme (Eds.), Theoretical perspectives on
terminology: Explaining terms, concepts and specialized knowledge (pp. 219–​235).
John Benjamins.
Corpas Pastor, G. (Ed.). (2016). Computerised and corpus-​ based approaches to
phraseology: Monolingual and multilingual perspectives. Éditions Tradulex.
Corpas Pastor, G., & Seghiri, M. (2009). Virtual corpora as documentation
resources: Translating travel insurance documents (English-​Spanish). In A. Beeby
Lonsdale, P. Rodríguez Inés, & P. Sánchez Gijón (Eds.), Corpus use and trans-
lating: Corpus use for learning to translate and learning corpus use to translate
(pp. 75–​107). John Benjamins.
Corver, N., & van Riemsdijk, H. (2001). Introduction: Semi-​lexical categories. In N.
Corver & H. van Riemsdijk (Eds.), Semi-​lexical categories: The function of content
words and the content of function words (pp. 1–​19). Mouton De Gruyter.
Cosmai, D. (2014). The language of Europe. Multilingualism and translation in
the EU institutions: Practice, problems and perspectives (3rd ed.). Éditions de
l’Université de Bruxelles.
Costa, H., & Zaretskaya, A. (2018). Nine terminology extraction tools: Are they
useful for translators? Retrieved 23 September, 2024, from https://​ling​uagr​eca.com/​
blog/​2018/​03/​nine-​term​inol​ogy-​ext​ract​ion-​tools-​are-​they-​use​ful-​for-​tran​slat​ors/​
Costa, N. M. P., & Gomes, D. M. (2013). The ethnoterminology of Mundurukú
Tupi-​language and the contributions of ecolinguistics. Cadernos de Linguagem e
Sociedade, 14(1), 252–​274.
Court of Justice of the European Union. (2020). Comparative multilingual legal
vocabulary. Leaflet downloadable from https://​e-​just​ice.eur​opa.eu/​119/​EN/​glos​sari​
es_​a​nd_​t​rans​lati​ons
Court of Justice of the European Union. (2024). Parental responsibility. Retrieved
17 July, 2024, from IATE entry, https://​iate.eur​opa.eu/​entry/​res​ult/​3583​353/​all
Cowie, A. P. (Ed.). (2001). Phraseology: Theory, analysis and application.
Clarendon Press.
CSA Research. (2024). The top 100 language service providers: 2024. Retrieved
8 July, 2024, from https://​csa-​resea​rch.com/​Featu​red-​Cont​ent/​For-​LSPs/​Glo​bal-​
Mar​ket-​Study/​The-​Larg​est-​Langu​age-​Serv​ice-​Provid​ers-​2024?utm_​c​ampa​ign=​
Spec ​i al%20Mail​ ers&utm​_ ​m ed​ium=​email&_​hs ​e nc=​p2AN​qtz- ​9ec3S2oVgjaLe
dkuITlwiGFnRLig1qwZxaTJYQPDfFchmKbLUz6ayyHmPMv5fPlDGGZ93Z
18UN73aqItnUr4KF​ S 1ML​ p rWf​ u ylk​ h 80m​ V bQk ​ q RQY ​ F w4&_ ​ h smi= ​ 3 14267​
586&utm_​cont​ent=​314267​586&utm​_​sou​rce=​hs_​em​ail
Dabu, C. M. (2017). Computer science education and interdisciplinarity. In A.
Vanderlei Dos Santos & J. C. Krause (Eds.), Science education: Research and new
technologies (pp. 137–​151). IntechOpen.
Daille, B. (2003). Terminology mining. In M. T. Pazienza (Ed.), Information extrac­
tion in the web era (pp. 29–​44). Springer Verlag.
Daille, B. (2017). Term variation in specialised corpora: Characterisation, automatic
discovery and applications. John Benjamins.
Dalman, S. G., et al. (2024). A giant tyrannosaur from the Campanian–​Maastrichtian
of southern North America and the evolution of tyrannosaurid gigantism. Nature
Scientific Reports, 14(22124).
Dam, H. V., & Zethsen, K. K. (2008). Translator status: A study of Danish company
translators. The Translator, 14(1), 71–​96.
 Bibliography 253

de Bessé, B. (1994). Contribution à la définition terminologique. In P. Martel & J.

Maurais (Eds.), Langues et sociétés en contact. Mélanges offerts à Jean-Claude Corbeil
(pp. 135–​138). Max Niemayer.
de Bessé, B. (1997). Terminological definitions (translated by Juan C. Sager). In S.
E. Wright & G. Budin (Eds.), Handbook of terminology management (vol. 1)
(pp. 63–​74). John Benjamins.
de Camillis, F. (2021). Institutional translation practices in South Tyrol: An explora­
tory study on civil servants working as ‘occasional translators.’ In M. Bourguignon,
B. Nouws, & H. van Gerwen (Eds.), Translation policies in legal and institutional
settings (pp. 87–​108). Leuven University Press.
de Saint Robert, M.-​J. (2008). CAT tools in international organisations: Lessons
learnt from the experience of the Languages Service of the United Nations office at
Geneva. In E. Yuste Rodrigo (Ed.), Topics in language resources for translation and
localisation. John Benjamins.
de Saussure, F. (1916/​1995). Cours de linguistique générale. Tullio de Mauro.
De Vecchi, D. (2015). Company and organisational-​speak oriented terminology. In
G. Budin & V. Lušicky (Eds.), Languages for special purposes in a multilingual,
transcultural world. Proceedings of the 19th European Symposium on languages
for special purposes (pp. 279–​288). University of Vienna, Centre for Translation
Studies.
De Vecchi, D. (2020). Words at work: The dynamics of company-​speak in the work.
Hermes, Journal of Language & Communication, 60(2020), 241–​249.
Delavigne, V., & Gaudin, F. (2022). Founding principles of socioterminology.
In P. Faber & M.-​ C. L’Homme (Eds.), Theoretical perspectives on termin-
ology: Explaining terms, concepts and specialized knowledge (pp. 177–​195). John
Benjamins.
Delisle, J. (2008). La terminologie au Canada. Histoire d’une profession. Linguatech.
Delisle, J. (2009). Fifty years of Parliamentary interpretation. Canadian Parliamentary
Review, 32(2), 27–​32.
Désilets, A., Huberdeau, L.-​P., Laporte, M., & Quirion, J. (2009b, 19–​20 November).
Building a collaborative multilingual terminology system. Proceedings of ASLIB
Translating and the Computer 31, London.
Désilets, A., Melançon, C., Patenaude, G., & Brunette, L. (2009a, 26–​30 August).
How translators use tools and resources to resolve translation problems: An
ethnographic study. Proceedings of Beyond Translation Memories: New Tools for
Translators MT Summit XII, Ottawa.
Desmet, I., & Boutayeb, S. (1994). Terms and words: Propositions for terminology.
Terminology, 1(2), 303–​325.
Di Nunzio, G. M., & Vezzani, F. (2021). On the reusability of terminological data.
In F. Boschetti, A. M. Del Grosso, & E. Salvatori (Eds.), Proceedings of the 10th
National Conference of AIUCD: Digital humanities for society: E-​quality, partici-
pation rights and values in the digital age (pp. 183–​186). AIUCD.
Díaz, N., & Zetzsche, J. (2022). The translator’s toolbox (version 15). Retrieved
12 October, 2024, from www.acad​emia​clp.com/​cou​rse/​the-​tran​slat​ors-​tool-​box
Díaz Vázquez, J. C., & Guerra Aranguren, M. A. (2019). Term definitions and their
impact on accurate communication within international customs practices. ÍKALA,
Revista de Lenguaje y Cultura, 24(3), 1–​14. https://​doi.org/​10.17533/​udea.ikala.
v24n03​a08
254 Bibliography

Diki-​Kidiri, M. (2008). Le vocabulaire scientifique dans les langues africaines. Pour

une approche culturelle de la terminologie. Karthala.
Diki-​Kidiri, M. (2022). Cultural terminology: An introduction to the theory and
method. In P. Faber & M.-​C. L’Homme (Eds.), Theoretical perspectives on termin-
ology: Explaining terms, concepts and specialized knowledge (pp. 197–​216). John
Benjamins.
Ding, H., & Li, X. (2017). Technical translation in China. Overview, practice,
resources. In C. Shei (Ed.), The Routledge handbook of Chinese translation
(pp. 537–​550). Routledge.
Ditaranto, E. (2005). Translating culture. A good translation includes “cultural due
diligence.” MultiLingual Computing & Technology #73, 16(5).
Ditlevsen, M. G., & Kastberg, P. (2009). Personal knowledge management.
Knowledge mapping techniques in the training of LSP translators. Retrieved on 8
October, 2014, from inTRAlinea Special Issue: Specialised Translation I, www.int​
rali​nea.org/​print/​artic​le_​s​peci​als/​1731
Doğru, G. (2019). Automatic term extraction from Turkish to English medical corpus
Proceedings of EUROPHRAS 2019.
Dourado, D. A. O., Melo, J. G. d., & Aragão, M. d. S. S. d. (2018). Glossário
etnobotânico das plantas utilizadas na fabricação de temperos comercializados
na feira livre de Belém do São Francisco, Pernambuco: Uma abordagem sócio e
etnoterminológica [Ethnobotanical glossary of the plants used in the manufacture of
seasonings marketed at the free market of Belém do São Francisco, Pernambuco: A
socio-​ethnoterminological approach]. Diversitas Journal, 3(3), 577–​601.
Downie, J. (2020). Interpreters vs machines: Can interpreters survive in an AI-​
dominated world? Routledge.
Drake, P. P., & Fabozzi, F. J. (2010). The basics of finance: An introduction to finan­
cial markets, business finance and portfolio management. John Wiley and Sons.
Driscoll, D. (1997). What is the international monetary fund? External Relations
Department, International Monetary Fund.
Drouin, P. (2006). Termhood: Quantifying the relevance of a candidate term.
In H. Picht (Ed.), Modern approaches to terminological theories and applications
(pp. 375–​391). Peter Lang.
Druță, I. (2020). Vernacular ergonyms, in between creativity and imitation.
In I. Boldea & C. Sigmirean (Eds.), The dialogue of multicultural discourses
(pp. 74–​79). Arhipelag XXI Press.
Dunne, K. J. (2013). Translation tools. In C. A. Chapelle (Ed.), The encyclopedia of
applied linguistics (pp. 839–​843). Wiley-​Blackwell Publishing.
Dunning, T. (1993). Accurate methods for the statistics of surprise and coincidence.
Computational Linguistics, 19(1), 61–​74.
Durán-​Muñoz, I., & L’Homme, M.-​C. (2020). Diving into English: Motion verbs
from a lexico-​semantic approach. A corpus-​based analysis of adventure tourism.
Terminology. International Journal of Theoretical and Applied Issues in Specialized
Communication, 26(1), 33–​59.
Dury, P. (2022). Diachronic variation. In P. Faber & M.-​ C. L’Homme (Eds.),
Theoretical perspectives on terminology: Explaining terms, concepts and specialized
knowledge (pp. 421–​434). John Benjamins.
Egwuogu, C. (2023). Terminology management guide: How to build a termbase.
Retrieved 8 August, 2024, from https://​cen​tus.com/​blog/​term​inol​ogy-​man​agem​
ent-​guide
 Bibliography 255

Ehrensberger-​Dow, M. (2014). Challenges of translation process research at the

workplace. MonTI Monografías de Traducción e Interpretación (special issue),
355–​383. Retrieved 26 April, 2017, from www.raco.cat/​index.php/​MonTI/​arti​cle/​
viewF​ile/​292​860/​381​290
ELIS Research (2025). ELIS 2025 Results. Retrieved 10 May 2025, from https://elis-
survey.org/repository/
ELIS Research. (2024). European Language Industry Survey 2024. Trends,
expectations and concerns of the European language industry. Retrieved 13
October, 2024, from https://​elis-​sur​vey.org/​wp-​cont​ent/​uplo​ads/​2024/​03/​ELIS-​
2024-​Rep​ort.pdf
Elmgrab, R. A. (2016). The creation of terminology in Arabic. American International
Journal of Contemporary Research, 6(2), 75–​85.
Emerson, G. (2020). What are the goals of distributional semantics? Proceedings of
the 58th Annual Meeting of the Association for Computational Linguistics (pp.
7436–7453), ACL [online].
EMT. (2017). Competence framework 2017. Retrieved 26 October, 2024, from
https://​com​miss​ion.eur​opa.eu/​sys​tem/​files/​2018-​02/​emt​_​com​pete​nce_​fwk_​2017​_​
en_​web.pdf
EMT. (2022). European Master’s in Translation: Competence framework 2022.
Retrieved 22 October, 2024, from https://​com​miss​ion.eur​opa.eu/​sys​tem/​files/​2022-​
11/​emt​_​com​pete​nce_​fwk_​2022​_​en.pdf
Esimaje, A., Gut, U., & Antia, B. E. (Eds.). (2019). Corpus linguistics and African
Englishes. John Benjamins.
Esselink, B. (2019). Multinational language service provider as user. In M. O’Hagan
(Ed.), The Routledge handbook of translation and technology (pp. 109–​ 126).
Routledge.
Faaß, G. (2017). Lexicography and corpus linguistics. In P. A. Fuertes-​Olivera (Ed.),
The Routledge handbook of lexicography (pp. 123–​137). Routledge.
Faber, P. (2009). The cognitive shift in terminology and specialized translation.
Monografías de Traducción e Interpretación, 1, 107–​134.
Faber, P. (Ed.). (2012). A cognitive linguistics view of terminology and specialized
language. De Gruyter Mouton.
Faber, P. (2015). Frames as a framework for terminology. In H. Kockaert & F. Steurs
(Eds.), Handbook of terminology (pp. 14–​33). John Benjamins.
Faber, P. (2022). Frame-​based terminology. In P. Faber & M.-​C. L’Homme (Eds.),
Theoretical perspectives on terminology: Explaining terms, concepts and specialized
knowledge (pp. 353–​375). John Benjamins.
Faber, P., & Cabezas García, M. (2019). Specialized knowledge representation: From
terms to frames. Research in Language, 17(2), 197–​211.
Faber, P., & l’Homme, M.-​C. (Eds.). (2022a). Theoretical perspectives on termin­
ology. John Benjamins.
Faber, P., & L’Homme, M.-​C. (2022b). Theoretical perspectives on terminology: An
introduction. In P. Faber & M.-​C. L’Homme (Eds.), Theoretical perspectives on
terminology: Explaining terms, concepts and specialized knowledge (pp. 1–​11).
John Benjamins.
Faber, P., et al. (2006). Process-​oriented terminology management in the domain of
coastal engineering. Terminology, 12(2), 189–​213.
256 Bibliography

Faber, P., & San Martín, A. (2011). Linking specialized knowledge and general know­
ledge in EcoLexicon. Actes de la conférence Terminologie et Ontologie: Théories et
applications (TOTh) 2011, Annecy, France.
Faini, P. (Ed.). (2017). Terminological approaches in the European context. Cambridge
Scholars Publishing.
Falco, G. (2017). Concept maps as teaching tools for students in legal translation.
Lingue e Linguaggi, 21(2017), 91–​106.
Fantinuoli, C. (2016). Revisiting corpus creation and analysis tools for translation
tasks. Cadernos de Traduçao, 36(1), 62–​87.
Farrell, M. (2024). Survey on the use of generative artificial intelligence by professional
translators. Paper given at AsLing’s Translating and the Computer Conference 46
(TC46), Luxembourg, 18–​20 November. https://​asl​ing.org/​tc46/​
Favre-​Lorraine, M. (2022). Query management reinvented. Retrieved 30 April, 2024,
from https://​xtm.cloud/​blog/​query-​man​agem​ent-​rei​nven​ted/​
Felber, H. (1984). Terminology manual. General Information Programme and
UNISIST.
Felici, A. (2010). Translating EU law: Legal issues and multiple dynamics.
Perspectives: Studies in Translatology, 18(2), 95–​108.
Fernández-​Parra, M. (2014). Formulaic expressions in computer-​assisted translation.
Scholars’ Press.
Fernández-​Parra, M. (2021). Every second counts: A study of translation practices in
the European Commission’s DGT. In F. Prieto Ramos (Ed.), Institutional transla-
tion and interpreting: Assessing practices and managing for quality (pp. 111–​127).
Routledge.
Fernández-​Parra, M. (2023). Expanding horizons: Adding phraseological function­
ality to CAT tools. NETTT, New Trends in Translation Technology, Rhodes,
Greece.
Fernández-​Silva, S. (2022). Cognitive approaches to the study of term variation.
In P. Faber & M.-​ C. L’Homme (Eds.), Theoretical perspectives on termin-
ology: Explaining terms, concepts and specialized domains (pp. 435–​455). John
Benjamins.
Fernández-​Silva, S., Freixa, J., & Cabré, M. T. (2014). A method for analysing
the dynamics of naming from a monolingual and multilingual perspective. In R.
Temmerman & M. van Campenhoudt (Eds.), Dynamics and terminology: An inter-
disciplinary perspective on monolingual and multilingual culture-​bound communi-
cation (pp. 183–​211). John Benjamins.
Fernández-​Silva, S., & Kerremans, K. (2011). Terminological variation in source texts
and translations: A pilot study. Meta, 56(2), 318–​335.
Fernández, T., Flórez de la Colina, M. A., & Peters, P. (2011). Terminology and
terminography for architecture and building construction. In A. Alcina (Ed.),
Teaching and learning terminology: New strategies and methods (pp. 11–​36). John
Benjamins.
Ferraresi, A. (2019). How specialized (or popularized)? Terminological density as
a clue to text specialization in the domain of food safety. Lingue e Linguaggi,
29(2019), 17–​39.
Fillmore, C. J. (1985). Frames and the semantics of understanding. Quaderni di
Semantica, 6(2), 222–​254.
 Bibliography 257

Fırat, G., Gough, J., & Moorkens, J. (2024). Translators in the platform economy: A
decent work perspective. Perspectives, 1–​19. https://​doi.org/​https://​doi.org/​10.1080/​
09076​76X.2024.2323​213
Firth, J. (1957). Studies in linguistic analysis. Blackwell.
Fischbach, H. (1993). Translation, the great pollinator of science. In S. E. Wright & L.
D. Wright (Eds.), Scientific and technical translation. John Benjamins.
Fleischmann, K. (2024). If content is king, terminology is queen! Interview with
Klaus Fleischmann (Kaleidoscope). Video by GALA, www.gala-​glo​bal.org/​knowle​
dge-​cen​ter/​profe​ssio​nal-​deve​lopm​ent/​artic​les/​if-​cont​ent-​king-​term​inol​ogy-​queen-​
interv​iew
Fleischmann, K., & Lundahl, M. (2023). The symbiosis between AI and terminology.
Retrieved 2 November, 2023, from [video resource] https://​kalei​dosc​ope.at/​en/​
webin​ars/​term​inol​ogy-​retrie​val-​and-​man​agem​ent/​
Fletcher, W. H. (2001). Concordancing the web with KWiCFinder. Paper at the
American Association for Applied Corpus Linguistics’ Third North American
Symposium on Corpus Linguistics and Language Teaching, Boston, MA, 23–​25
March 2001.
Fletcher, W. H. (2007). Concordancing the web: Promise and problems, tools and
techniques. In M. Hundt, N. Nesselhauf, & C. Biewer (Eds.), Corpus linguistics
and the web (pp. 25–​45). Rodopi.
Floros, G. (2017). Legal terminology in Cyprus: Linguistic issues and issues of acces­
sibility. In P. Faini (Ed.), Terminological approaches in the European context
(pp. 288–​304). Cambridge Scholars Publishing.
Fontanet, M. (2018). Translating hybrid legal texts for science and technology
institutions: The case of CERN. In F. Prieto Ramos (Ed.), Institutional translation
for international governance. Bloomsbury.
Francuski, B. Đ. (2017). Translating EU terminology from English as the dominant
language into Serbian. In P. Faini (Ed.), Terminological approaches in the European
context (pp. 107–​131). Cambridge Scholars Publishing.
Frankenberg-​Garcia, A. (2015). Training translators to use corpora hands-​
on: Challenges and reactions by a group of thirteen students at a UK university.
Corpora, 15(10), 351–​380.
Frantzi, K., Ananiadou, S., & Mima, H. (2000). Automatic recognition of multi-​word
terms: The C-value-NC-value method. International Journal on Digital Libraries,
3(2000), 115–​130.
Freixa, J. (2022). Causes of terminological variation. In P. Faber & M.-​C. L’Homme
(Eds.), Theoretical perspectives on terminology: Explaining terms, concepts and
specialized knowledge (pp. 399–​420). John Benjamins.
Froeliger, N., Krause, A., & Salmi, L. (2023). Institutional translation –​EMT
competence framework and beyond. In T. Svoboda, Ł. Biel, & V. Sosoni (Eds.),
Institutional translator training (pp. 13–​29). Routledge.
Fuertes-​Olivera, P. A. (2017). The Routledge handbook of lexicography. Routledge.
Fuertes-​Olivera, P. A., & Tarp, S. (2014). Theory and practice of specialised online
dictionaries: Lexicography versus terminography. De Gruyter. https://​doi.org/​
doi:10.1515/​978311​0349​023
Gaffane, M. (2020). National Language Service. Retrieved 13 October, 2023, from
www.dac.gov.za/​cont​ent/​langu​age-​serv​ice-​0
258 Bibliography

Gagné, A.-​M., & L’Homme, M.-​ C. (2016). Opposite relations in terminology.

Terminology. International Journal of Theoretical and Applied Issues in Specialized
Communication, 22(1), 30–​51.
Gagnon, C. (2021). Government ideologies in translation: An inquiry into past
Canadian budget speeches. In M. Bourguignon, B. Nouws, & H. van Gerwen
(Eds.), Translation policies in legal and institutional settings (pp. 23–​46). Leuven
University Press.
Galinski, C., & Picht, H. (1997). Graphic and other semiotic forms of knowledge
representation in terminology management. In S. E. Wright & G. Budin (Eds.),
Handbook of terminology management, vol. 1 (pp. 42–​61). John Benjamins.
Gambier, Y. (1991). Travail et vocabulaire spécialisés: Prolégomènes à une socio-​
terminologie. Meta. Journal des traducteurs. Translator’s Journal, 367(1), 8–​15.
Gamper, J., & Stock, O. (1998). Corpus-​based terminology. Terminology. International
Journal of Theoretical and Applied Issues in Specialized Communication, 5(2),
147–​159.
García, I. (2015). Computer-​aided translation: Systems. In C. Sin-​wai (Ed.), The
Routledge encyclopedia of translation technology (pp. 68–​87). Routledge.
Garrido Nombela, R. (2013). Translating for government departments: The case of
the Spanish Ministry of Foreign Affairs and Co-​operation. In A. Borja Albi & F.
Prieto Ramos (Eds.), Legal translation in context: Professional issues and prospects
(pp. 143–​154). Peter Lang.
Gaudin, F. (1993). Pour une socioterminologie. Des problèmes sémantiques aux
pratiques institutionnelles. Université de Rouen.
Gaudin, F. (2003). Socioterminologie, une approche sociolinguistique de la
terminologie. Meta, 49(2), 221–​458.
George, S., & Sabelli, F. (1994). Faith and credit: The World Bank’s secular empire.
Penguin Books.
Gerzymisch-​Arbogast, H. (2008). Fundamentals of LSP translation. In H.
Gerzymisch-​Arbogast, G. Budin, & G. Hofer (Eds.), LSP translation scenarios.
Selected contributions to the EU Marie Curie conference, Vienna 2007 (pp. 7–​64).
ATRC Group.
Giacomini, L. (2022). In C. Mellado Blanco (Ed.), Productive patterns in phraseology
and construction grammar (pp. 215–​233). De Gruyter.
Giai, E., Poeta, N., & Turnbull, D. (2023). Multi-​ modal terminology manage­
ment. Corpora, data models and implementations in TermStar. In I. Koprinska,
P. Mignone, R. Guidotti, et al.(Eds.), Proceedings of the European conference
on machine learning and principles and practice of knowledge discovery (ECML
PPKD 2022) (pp. 219–​226). Springer.
Gläser, R. (1995). Linguistic features and genre profiles of scientific English.
Peter Lang.
Gläser, R. (2001). The stylistic potential of phraseological units in the light of genre
analysis. In A. P. Cowie (Ed.), Phraseology. Theory, analysis, applications (pp. 125–​
143). Oxford University Press.
Gledhill, C. J. (2000). Collocations in science writing. Gunter Narr Verlag.
Godbout, M. (2016). Lack of status –​are translators the authors of their own mis­
fortune? Retrieved 25 August, 2021, from www.circ​uitm​agaz​ine.org/​doss​ier-​131/​
lack-​of-​sta​tus-​are-​tran​slat​ors-​the-​auth​ors-​of-​their-​own-​mis​fort​une
 Bibliography 259

Godly, T. (1993). Terminological principles and methods in the subject-​field of chem­

istry. In H. B. Sonneveld & K. L. Loening (Eds.), Terminology: Applications in
interdisciplinary communication (pp. 141–​163). John Benjamins.
Göpferich, S. (1995). A pragmatic classification of LSP texts in science and tech­
nology. Target, 7(2), 305–​326.
Göpferich, S., & Jääskeläinen, R. (2009). Process research into the development
of translation competence: Where are we and where do we need to go? Across
Languages and Cultures, 10(2), 169–​191.
Gornostay, T. (2010, 22–​ 23 October). Terminology management in real use.
Proceedings of the 5th International Conference on Applied Linguistics in Science
and Education, Istanbul, Turkey.
Gotti, F., Langlais, P., & Lapalme, G. (2013). Designing a machine translation system
for Canadian weather warnings: A case study. Natural Language Engineering,
20(3), 399–​433.
Gotti, M., & Šarčević, S. (2006). Insights into specialized translation. Peter Lang.
Gouadec, D. (2007). Translation as a profession. John Benjamins.
Government of Canada. (2020a). Indigenous Languages Act. Retrieved 5 October,
2023, from https://​laws-​lois.just​ice.gc.ca/​PDF/​I-​7.85.pdf, https://​laws-​lois.just​ice.
gc.ca/​eng/​acts/​i-​7.85/​page-​1.html
Government of Canada. (2020b). Capitalization: Races, languages, people. Retrieved
7 May, 2024, from www.nos​lang​ues-​ourla​ngua​ges.gc.ca/​en/​writ​ing-​tips-​plus/​cap​
ital​izat​ion-​races-​langua​ges-​peop​les
Government of Canada. (2023). Interpretation at the Translation Bureau. Retrieved
5 October, 2023, from www.tpsgc-​pwgsc.gc.ca/​bt-​tb/​inte​rpre​tes-​inter​pret​ers/​accu​
eil-​home-​eng.html
Government of Canada. (2024a). Translation Bureau regulations. Retrieved 7 May,
2024, from https://​laws-​lois.just​ice.gc.ca/​PDF/​C.R.C.,_​c._​1561.pdf
Government of Canada. (2024b). Inuk, Inuit (linguistic recommendation from the
Translation Bureau). Retrieved 7 May, 2024, from www.nos​lang​ues-​ourla​ngua​
ges.gc.ca/​en/​writ​ing-​tips-​plus/​inuk-​inuit?wbdisa​ble=​true#:~:text=​The%20Tran​slat​
ion%20Bur​eau%20rec​omme​nds%20us​ing,from%20wh​ich%20it%20is%20b​
orro​wed.
Government of Canada. (2024c). Statistics on official languages in Canada. Retrieved
7 May, 2024, from www.can​ada.ca/​en/​canad​ian-​herit​age/​servi​ces/​offic​ial-​langua​
ges-​bilin​gual​ism/​publi​cati​ons/​sta​tist​ics.html
Granda, R., & Warburton, K. (2001). Terminology management as data manage­
ment. Proceedings of the 2001 Conference of the IBM Centre for Advanced Studies
on Collaborative Research (CASCON 01). Toronto, Canada. IBM Press.
Granger, S., & Meunier, F. (Eds.). (2008). Phraseology: An interdisciplinary perspec­
tive. John Benjamins.
Gregorio Cano, A. (2012). Becoming a translator: The development of intercultural
competence in Spain. Cultus –​The Journal of Intercultural Mediation and
Communication, 5, 154–​170.
Gries, S. T. (2022). How to use statistics in quantitative corpus analysis. In A. O’Keeffe
& M. McCarthy (Eds.), The Routledge handbook of corpus linguistics (2nd ed.,
pp. 168–​181). Routledge.
260 Bibliography

Gurrutxaga, A., et al. (2010). Exploiting the Internet to build language resources for
less-​resourced languages. Paper at the 7th SALTMIL Workshop on creation and use
of basic lexical resources for less-​resourced languages, LREC 2010, Valetta, Malta.
Gutner, T. (2023). Collaboration, cooperation, coordination: A history of the Bretton
Woods twins’ efforts to work together. Review of International Political Economy,
30(3), 965–​990. https://​doi.org/​10.1080/​09692​290.2022.2073​461
Hadni, M., Lachkar, A., & Ouatik, S. A. (2014). Multiword term extraction based on
new hybrid approach for Arabic language. Proceedings of the Second International
Conference on Computational Science and Engineering (CSE 2014), Dubai, UAE.
Hall, C. M. (2018). The saffron scourge in New Orleans. One hundred years of pesti­
lence. Retrieved 10 August, 2024, from https://​count​ryro​adsm​agaz​ine.com/​art-​
and-​cult​ure/​hist​ory/​yel​low-​fever/​#:~:text=​Yel​low%20fe​ver%20w​ent%20by%20m​
any,what%20to%20do%20ab​out%20it.
Halverson, S. (1998). Translation studies and representative corpora: Establishing
links between translation corpora. Theoretical/​descriptive categories and a con-
ception of the object of study. Meta. Journal des traducteurs. Translator’s Journal,
43(4), 494–​514.
Hamilton, G. (2010). Translation in Canada. ATA Chronicle (October 2010), 12–​15.
Hardwick, J. (2024). Google search operators: The complete list (44 advanced
operators). Retrieved 19 October, 2024, from https://​ahr​efs.com/​blog/​goo​gle-​
advan​ced-​sea​rch-​operat​ors/​
Harper, R. H. R. (1998). Inside the IMF. An ethnography of documents, technology
and organisational action. Academic Press.
Harris, B., & Sherwood, B. (1978). Translating as an innate skill. In D. Gerver & H.
W. Sinaiko (Eds.), Language, interpretation and communication (pp. 155–​170).
Plenum Press.
Harris, Z. S. (1954). Distributional structure. Word, 10(2-​3), 146–​162.
Hartley, A., & Paris, C. (2001). Translation, controlled languages, generation. In E.
Steiner & C. Yallop (Eds.), Exploring translation and multilingual text produc-
tion: Beyond content (pp. 307–​326). Mouton de Gruyter.
Hartley, T. (2009). Technology and translation. In J. Munday (Ed.), The Routledge
companion to translation studies (pp. 106–​127). Routledge.
Hartley, V. (n.d.). The future of termbase management. Retrieved 3 November,
2024, from Language Wire, www.langu​agew​ire.com/​en/​blog/​fut​ure-​termb​ase-​man​
agem​ent
Hatim, B., & Mason, I. (1997). The translator as communicator. Routledge.
Hätty, A., Schulte im Walde, S., & Bosch, R. (2018). Fine-​grained termhood predic­
tion for German compound terms using neural networks. Proceedings of the Joint
Workshop on Linguistic Annotation, Multiword Expressions and Constructions
(LAW-MWE-CxG-2018).
Hayes, A. (2022). World Bank Group. Retrieved 24 August, 2023, from Investopedia,
www.inves​tope​dia.com/​terms/​w/​world-​bank-​group.asp
He, S. (2024). Prompting ChatGPT for translation: A comparative analysis of transla­
tion brief and persona prompts. Proceedings of the 25th Annual Conference of the
European Association for Machine Translation (EAMT) (Volume 1), Sheffield, UK.
He, T., Zhang, X., & Ye, X. (2006). An approach to automatically constructing domain
ontology. Proceedings of the 20th Pacific Asia Conference on Language, Information
and Computation (PACLIC 2006), Wuhan, China, 1–​3 November, 2006.
 Bibliography 261

Heid, U. (2007). Computational linguistic aspects of phraseology II. In H. Burger, D.

Dobrovol´skiï, P. Kühn, & N. R. Norrick (Eds.), Phraseologie /​phraseology. An
international handbook of contemporary research, vol.2 (Vol. II, pp. 1036–​1044).
Walter de Gruyter.
Heid, U. (2008). Computational phraseology. In S. Granger & F. Meunier (Eds.),
Phraseology: An interdisciplinary perspective (pp. 337–​360). John Benjamins.
Hein, A. S., Almqvist, I., & Starbäck, P. (1997, 13–​14 November). Scania Swedish: A
basis for multilingual machine translation. Translating and the Computer 19, ASLIB,
London.
Helleiner, E. (2017). What’s been missing from conventional histories of Bretton
Woods? In G. Scott-​ Smith & S. Rofe (Eds.), Global perspectives on the
Bretton Woods conference and the post-​war world order (pp. 17–​34). Palgrave
Macmillan.
Hernandez Morin, K., Barbin, F., & Phuez-​Favris, G. (2017). Handbook for the
organisation of intensive translation sessions. Retrieved 22 October, 2024, from
https://​univ-​renn​es2.hal.scie​nce/​hal-​04182​942/​docum​ent
Heylen, K., et al. (2014a, May 2014). TermWise: A CAT tool with context-​sensitive
terminological support. Proceedings of the Ninth International Conference on
Language Resources and Evaluation (LREC), Reykjavik, Iceland.
Heylen, K., et al. (2014b, June 2014). TermWise: Leveraging big data for termino­
logical support in legal translation. Proceedings of Terminology and Knowledge
Engineering, Berlin, Germany.
Heylen, K., & De Hertog, D. (2015). Automatic term extraction. In H. Kockaert &
F. Steurs (Eds.), Handbook of terminology, vol. 1 (pp. 203–​221). John Benjamins.
Hickey, L. (2013). Translating for the police, prosecutors and courts: The case of
English letters of request. In A. Borja Albi & F. Prieto Ramos (Eds.), Legal transla-
tion in context: Professional issues and prospects (pp. 123–​142). Peter Lang.
Hofmann, C., & Mehnert, T. (2000). Multilingual information management at
Schneider Automation. In R. C. Sprung & S. Jaroniec (Eds.), Translating into
success: Cutting-​edge strategies for going multilingual in a global age (pp. 59–​79).
John Benjamins.
Horner, J. (2021). What are CAT tools and how do they benefit you? Retrieved 29
March, 2020, from www.mtmlin​guas​oft.com/​what-​are-​cat-​tools-​and-​how-​do-​
they-​bene​fit-​you/​
Horváth, I. (Ed.). (2016). The modern translator and interpreter. Eötvös
University Press.
House, J., & Kádár, D. Z. (2020). T/​V pronouns in global communication practices
the case of IKEA catalogues across linguacultures. Journal of Pragmatics, 161(5),
1–​15.
Howe, A. E., & Dreilinger, D. (1997). SavvySearch: A metasearch engine that learns
which search engines to query. AI Magazine, 18(2), 19–​25.
Huangfu, W., & Zhao, Y. (2014). A corpus-​based translation method of term extrac­
tion in LSP texts. Theory and Practice in Language Studies, 4(1), 46–​51.
Hughes, D. P., & Libersat, F. (2018). Neuroparasitology of parasite-​insect associ­
ations. Annual Review of Entomology, 63, 471–​487.
Humbley, J. (1997). Is terminology specialized lexicography? The experience of
French-​speaking countries. Hermes, Journal of Linguistics, 18, 13–​31.
262 Bibliography

Humbley, J. (2022). The reception of Wüster’s General Theory of Terminology.

In P. Faber & M.-​ C. L’Homme (Eds.), Theoretical perspectives on termin-
ology: Explaining terms, concepts and specialized knowledge (pp. 15–​35). John
Benjamins.
Humbley, J., Budin, G., & Laurén, C. (Eds.). (2018). Languages for special
purposes: An international handbook. De Gruyter Mouton.
Hundt, M., Nesselhauf, N., & Biewer, C. (Eds.). (2007). Corpus linguistics and the
web. Rodopi.
Hung, E. (2005). Translation in China. An analytical survey. In E. Hung & J.
Wakabayashi (Eds.), Asian translation traditions (pp. 67–​ 107). St. Jerome
Publishing.
Hunston, S. (2022). How can a corpus be used to explore patterns? In A. O’Keeffe
& M. McCarthy (Eds.), The Routledge handbook of corpus linguistics (2nd ed.,
pp. 140–​154). Routledge.
Hutchins, J. (2004). Machine translation and computer-​ based tools. In J. M.
Bravo (Ed.), A new spectrum of translation studies (pp. 13–​48). University of
Valladolid.
Hyland, K. (2022). English for specific purposes: What is it and where is it taking
us? ESP Today –​Journal of English for Specific Purposes at Tertiary Level, 10(2),
202–​220.
IAU. (2006). Definition of a planet in the solar system. Retrieved 17 July, 2024, from
Resolution B5, XXVI General Assembly, www.iau.org/​sta​tic/​reso​luti​ons/​Reso​luti​
on_​G​A26-​5-​6.pdf
IAU. (n.d.). Pluto and the developing landscape of our solar system. Retrieved
17 July, 2024, from www.iau.org/​pub​lic/​the​mes/​pluto/​#n10
IMF. (2022a). Crisis upon crisis. IMF annual report 2022. International Monetary
Fund. Retrieved 12 August, 2023, from www.imf.org/​exter​nal/​pubs/​ft/​ar/​2022/​
IMF. (2022b). The IMF and the World Bank. Retrieved 12 August, 2023, from www.
imf.org/​en/​About/​Fac​tshe​ets/​She​ets/​2022/​IMF-​World-​Bank-​New
Infoterm. (2013). History of the standardization of terminological principles and
methods. Retrieved 27 October, 2024, from www.infot​erm.info/​stan​dard​izat​ion/​
history_​standardization_​terminolog​ical​_​pri​ncip​les_​and_​meth​ods.php#:~:text=​
In%201 ​ 9 51%2C%20ISO/ ​ T C%2037,to%20the%20act ​ i vit ​ i es%20of%20I​
nfot​erm
Innselset, K., Kristiansen, M., & Øvsthus, K. (2008). Looking back to move for­
ward. Challenges related to deceitful parallel texts and slippery terms. SYNAPS,
21(2008), 73–​89.
IONOS Digital Guide. (2021). Boolean operators: Smart searching in Google and
other search engines. Retrieved 19 October, 2024, from www.ionos.co.uk/​digit​algu​
ide/​onl​ine-​market​ing/​web-​analyt​ics/​bool​ean-​sea​rch-​operat​ors/​
IONOS Digital Guide. (2022). Google search operators: Advanced web searches.
Retrieved 19 October, 2024, from www.ionos.co.uk/​digit​algu​ide/​onl​ine-​market​ing/​
sea​rch-​eng​ine-​market​ing/​sea​rch-​more-​effe​ctiv​ely-​with-​goo​gle-​operat​ors/​
Jaekel, G. (2000). Terminology management at Ericsson. In R. C. Sprung & S.
Jaroniec (Eds.), Translating into success: Cutting-​edge strategies for going multi-
lingual in a global age (pp. 159–​171). American Translators Association (ATA)
Scholarly Monograph Series, Volume XI. John Benjamins.
Jain, T. (2004). Confused meanings for common fire terminology can lead to fuels
mismanagement. A new framework is needed to clarify and communicate the
 Bibliography 263

concepts. Retrieved 18 July, 2024, from www.fs.usda.gov/​rm/​pub​s_​ot​her/​rmrs​_​

200​4_​ja​in_​t​002.pdf
Joffe, D., & de Schryver, G.-​M. (2005). From TshwaneLex to TshwaneTerm: Tailoring
terminology management for South Africa. Lexikos, 15, 312–​315.
Jones, C. (2022). What are the basics of analysing a corpus? In A. O’Keeffe & M.
McCarthy (Eds.), The Routledge handbook of corpus linguistics (2nd ed., pp. 126–​
139). Routledge.
Joscelyne, A. (2000). The role of translation in an international organization. In R.
C. Sprung & S. Jaroniec (Eds.), Translating into success: Cutting-​edge strategies for
going multilingual in a global age (pp. 81–​95). American Translators Association
(ATA) Scholarly Monograph Series Vol. XI. John Benjamins.
Kageura, K. (1995). Toward the theoretical study of terms: A sketch from the lin­
guistic viewpoint. Terminology. International Journal of Theoretical and Applied
Issues in Specialized Communication, 2(2), 239–​257.
Kageura, K. (2002). The dynamics of terminology. A descriptive theory of term for­
mation and terminological growth. John Benjamins.
Kageura, K. (2015). Terminology and lexicography. In H. Kockaert & F. Steurs (Eds.),
Handbook of terminology, vol. 1 (pp. 45–​59). John Benjamins.
Kageura, K. (2022). Terminological growth. In P. Faber & M.-​C. L’Homme (Eds.),
Theoretical perspectives on terminology: Explaining terms, concepts and specialized
knowledge (pp. 457–​475). John Benjamins.
Kageura, K., & Marshman, E. (2019). Terminology extraction and management.
In M. O’Hagan (Ed.), The Routledge handbook of translation and technology
(pp. 61–​77). Routledge.
Kageura, K., & Umino, B. (1996). Methods of automatic term recognition: A review.
Terminology, 3(2), 259–​289.
Kaleidoscope. (2021a). Always pick the right word with Checkterm –​checked in all
types of content. Retrieved 10 May, 2025, from https://​kalei​dosc​ope.at/​en/​blog/​alw​
ays-​pick-​right-​word-​with-​checkt​erm/​
Kaleidoscope. (2022a). Kalcium 6.5.0 demo server documentation. Retrieved 19
April, 2024, from https://​demo.kalei​dosc​ope.at/​kalcas​set/​Kalciu​m_​De​mose​rver​_​
en.pdf
Kaleidoscope. (2022b). Kalcium Smartquery: Managing translators’ queries effect­
ively [video resource]. Retrieved 23 April, 2024, from www.yout​ube.com/​watch?v=​
C9I3​O_​QA​bAo
Kaleidoscope. (2023a). Together on behalf of our clients. Retrieved 10 May 2025,
from https://kaleidoscope.at/en/partners/
Kaleidoscope. (2023b). Kaleidoscope: Custom solutions for your successful global
content. Retrieved 2 November, 2023, from https://​kalei​dosc​ope.at/​en/​about-​us/​
Kaleidoscope. (2023c). Systematic terminology management for you. Retrieved 2
November, 2023, from https://​kalei​dosc​ope.at/​en/​produ​cts/​quickt​erm/​
Kaleidoscope. (2023d). Our solutions. Retrieved 2 November, 2023, from https://​
kalei​dosc​ope.at/​en/​soluti​ons/​
Kaleidoscope. (2023e). Are you trapped in the translation query management maze?
Here’s how to escape. Retrieved 23 April, 2024, from Slator article at https://​sla​tor.
com/​trap​ped-​in-​tran​slat​ion-​query-​man​agem​ent-​maze-​heres-​how-​to-​esc​ape/​
Kaleidoscope. (2023f). What’s new in Kalcium 6.5.2. Retrieved 30 May, 2024, from
https://​kalei​dosc​ope.at/​en/​blog/​new-​in-​kalc​ium-​6-​5-​2/​
264 Bibliography

Kaleidoscope. (2024). Terminology verification. Retrieved 16 April, 2024, from

https://​kalei​dosc​ope.at/​en/​produ​cts/​checkt​erm/​
Kang, J.-​H. (2009). Institutional translation. In M. Baker & G. Saldanha (Eds.),
Routledge encyclopedia of translation studies (2nd ed., pp. 141–​144). Routledge.
Kapsali, V. (2022). Method of mapping interdisciplinary research and practice at the
intersection of biology and design. DRS Conference Proceedings 2022: Bilbao,
Bilbao, Spain, 25 June–​3 July.
Katan, D. (2004). Translating cultures. An introduction for translators, interpreters
and mediators (2nd ed.). St. Jerome Publishing.
Katan, D. (2011). Occupation or profession: A survey of the translators’ world.
In R. Sela-​Sheffy & M. Shlesinger (Eds.), Identity and status in the translational
professions (pp. 65–​87). John Benjamins.
Kati, V. (2022). How to do a Boolean search effectively (& advanced tips). Retrieved
19 October, 2024, from www.mentio​nlyt​ics.com/​blog/​how-​to-​do-​a-​bool​ean-​sea​
rch/​#:~:text=​If%20you%20w​ant%20to%20li​mit,will%20app​ear%20in%20y​
our%20sea​rch
Keller, N. T. (2010). Integrating multiword terms in terminology management
systems: A case study. Proceedings of T21N Translation in Transition 2010–​10.
Kelly, J. (1993). Artificial intelligence. A modern myth. Ellis Horwood Ltd.
Kenning, M.-​M. (2010). What are parallel and comparable corpora and how can
we use them? In A. O’Keeffe & M. McCarthy (Eds.), The Routledge handbook of
corpus linguistics (pp. 487–​500). Routledge.
Kenny, D. (2011). Electronic tools and resources for translators. In K. Malmkjaer
& K. Windle (Eds.), The Oxford handbook of translation studies (pp. 455–​472).
Oxford University Press.
Kermas, S. (2006). Metaphor and ideology in business and economic discourse
in British and American English. In J. Flowerdew & M. Gotti (Eds.), Studies in
specialized discourse (pp. 109–​130). Peter Lang.
Kerremans, K. (2010). A comparative study of terminological variation in specialised trans­
lation. In C. Heine & J. Engberg (Eds.), Reconceptualizing LSP. Online proceedings of
the XVII European LSP Symposium 2009 (pp. 1–​14). Aarhus University.
Kerremans, K., Temmerman, R., & Tummers, J. (2004). Discussion on the
requirements for a workbench supporting termontography. Proceedings of the 11th
EURALEX International Congress, Lorient, France.
Kilgarriff, A. (2006). Googleology is bad science. Computational Linguistics, 33(1),
147–​151. https://​doi.org/​10.1162/​coli.2007.33.1.147. 2
Kilgarriff, A., Baisa, V., Bušta, J., et al. (2014). The Sketch Engine: Ten years
on. Lexicography, 1(1), 7–​36. https://​doi.org/​https://​doi.org/​10.1007/​s40​607-​014-​
0009-​9
Kilgarriff, A., et al. (2010). A quantitative evaluation of word sketches. Paper presented
at the Proceedings of the 14th EURALEX International Congress, Leeuwarden,
The Netherlands.
Kim, M. (2017). Variation terminologique en francophonie. Élaboration d’un modèle
d’analyse des facteurs d’implantation terminologique. Doctoral thesis, Université
Paris IV.
Kim, S., Yoon, J., & Song, M. (2001). Automatic extraction of collocations from
Korean text. Computers and the Humanities, 35, 273–​297.
 Bibliography 265

King, P. (2019). Small and medium-​sized enterprise (SME) translation service pro­
vider as technology user. In M. O’Hagan (Ed.), The Routledge handbook of trans-
lation and technology (pp. 148–​165). Routledge.
Kiraly, D. (2000). A social constructivist approach to translator education.
Empowerment from theory to practice. St. Jerome Publishing.
Kiraly, D. (2013). Towards a view of translator competence as an emergent phenom­
enon: Thinking outside the box(es) in translator education. In D. Kiraly, S. Hansen-​
Schirra, & K. Maksymski (Eds.), New prospects and perspectives for educating
language mediators (pp. 197–​224). Gunter Narr.
Kiraly, D. (2015). Occasioning translator competence: Moving beyond social con­
structivism toward a postmodern alternative to instructionism. Translation and
Interpreting Studies, 10(1), 8–​32.
Kirov, V., & Malamin, B. (2022). Are translators afraid of artificial intelligence?
Societies, 12(70), 1–​14.
Kittredge, R., & Lehrberger, J. (Eds.). (1982). Sublanguage: Studies of language in
restricted semantic domain. de Gruyter.
Klerkx, J. (1990). The translation of biological texts: Some typical problems. In
M. Thelen & B. Lewandowska-​ Tomaszczyk (Eds.), Translation and meaning
1: Proceedings of the Maastricht session of the First International Maastricht-​Łódź
Duo Colloquium on translation and meaning (pp. 214–​219). Euroterm.
Koehn, P. (2005). Europarl: A parallel corpus for statistical machine translation.
Proceedings of Machine Translation X: Papers, Phuket, Thailand.
Koester, A. (2022). Building small specialised corpora. In A. O’Keeffe & M. McCarthy
(Eds.), The Routledge handbook of corpus linguistics, 2nd edition (pp. 48–​61).
Routledge.
Koeva, S. (2007). Multi-​ word term extraction for Bulgarian. Proceedings of the
Workshop on Balto-​Slavonic Natural Language Processing: Information Extraction
and Enabling Technologies, Prague, Czech Republic.
Kohlmeier, B. (2000). Microsoft Encarta goes multilingual. In R. C. Sprung (Ed.),
Translating into success (pp. 1–​11). John Benjamins.
Korkontzelos, I., Klapaftis, I. P., & Manandhar, S. (2008). Reviewing and evalu­
ating automatic term recognition techniques. In A. Ranta & B. Nordstrøm (Eds.),
Advances in natural language processing. Proceedings of the 6th International
Conference GoTAL, Gothenburg, Sweden, 25–​ 27 Aug 2008 (pp. 248–​ 259).
Springer.
Koskinen, K. (2011). Institutional translation. In Y. Gambier & L. Van Doorslaer
(Eds.), Handbook of translation studies (Vol. 2, pp. 54–​60). John Benjamins.
Koskinen, K. (2014a). Institutional translation: The art of government by translation.
Perspectives: Studies in Translatology, 22(4), 479–​492.
Koskinen, K. (2014b). Translating institutions: An ethnographic study of EU trans­
lation. Routledge.
Krajcso, Z. (2018). Translator’s competence profiles versus market demand. Babel,
64(5–​6), 692–​709.
Krishan, T. (2017). Concept map technique as a new method for whole text transla­
tion. English Language Teaching, 10(4), 62–​66.
Kritsky, G. (2010). The quest for the perfect hive. A history of innovation in bee cul­
ture. Oxford University Press.
266 Bibliography

Kübler, N., & Frérot, C. (2003). Verbs in specialised corpora: From manual corpus-​
based description to automatic extraction in an English-​French parallel corpus.
Proceedings of Corpus Linguistics 2003, vol. 16 (pp. 429–438).
Kudashev, I., & Mikhailov, M. (2021). Terminology course as an integral part of
Master in Translation programme. In С. Д. Шелов & E. Цисун (Eds.), Слово и
термин (pp. 103–​114). Первый том.
L’Homme, M.-​C. (2004). A lexico-​semantic approach to the structuring of termin­
ology. Proceedings of CompuTerm 2003: The Third International Workshop on
Computational Terminology, Geneva, (pp. 7–14), ACL.
L’Homme, M.-​C. (2005). Sur la notion de terme. Meta, 50(4), 1112–​1132.
L’Homme, M.-​C. (2006). A look at some Canadian contributions to terminology.
In H. Picht (Ed.), Modern approaches to terminological theories and applications
(pp. 55–​75). Peter Lang.
L’Homme, M.-​C. (2014). Why lexical semantics is important for e-​lexicography and
why it is equally important to hide its formal representations from users of diction-
aries. International Journal of Lexicography, 27(4), 360–​377.
L’Homme, M.-​C. (2020a). Lexical semantics for terminology. John Benjamins.
L’Homme, M.-​C. (2020b). Revisiting polysemy in terminology. Proceedings of
the XIX Congress of the European Association for Lexicography (EURALEX),
Alexandroupolis, Greece, 7–​11 September 2021.
L’Homme, M.-​C. (2022). Terminology and lexical semantics. In P. Faber & M.-​
C. L’Homme (Eds.), Theoretical perspectives on terminology: Explaining terms,
concepts and specialized knowledge (pp. 237–​259). John Benjamins.
L’Homme, M.-​C., & Bae, H.-​S. (2006). A methodology for developing multilingual
resources for terminology. Proceedings of the 5th International Conference on
Language Resources and Evaluation (LREC), Genoa, Italy.
Lafeber, A. (2023). Skills and knowledge required of translators in institutional
settings. In T. Svoboda, Ł. Biel, & V. Sosoni (Eds.), Institutional translator training
(pp. 30–​48). Routledge.
Lakshmi, D., & Tyagi, A. K. (Eds.). (2023). Privacy preservation and secured data
storage in cloud computing. IGI Global.
Laurén, C., & Picht, H. (2006). Approaches to terminological theories: A compara­
tive study of the state-​of-​the-​art. In H. Picht (Ed.), Modern approaches to termino-
logical theories and applications (pp. 163–​184). Peter Lang.
LeBlanc, M. (2016). The impact of new translation technologies on specialized texts.
Studia Romanica Posnaniensia, 43(1), 77–​92.
LeBlanc, M. (2018). Language minorities in a globalized economy: The case of pro­
fessional translation in Canada. In G. Hogan-​Brun & B. O’Rourke (Eds.), The
Palgrave handbook of minority languages and communities (pp. 333–​ 354).
Palgrave Macmillan.
Lefer, M.-​A., & Granger, S. (Eds.). (2022). Extending the scope of corpus-​based
translation studies. Bloomsbury Publishing.
Leimgruber, J. (2019). Language planning and policy in Quebec. A compara­
tive perspective with views from Wales and Singapore. Narr Francke Attempto
Verlag GmbH.
León-​Araúz, P. (2022). Terminology and equivalence. In P. Faber & M.-​C. L’Homme
(Eds.), Theoretical perspectives on terminology: Explaining terms, concepts and
specialized knowledge (pp. 477–​501). John Benjamins.
 Bibliography 267

León-​Araúz, P., & Cabezas García, M. (2020). Term and translation variation of
multi-​word terms. MonTI Special Issue 6, 210–​247.
León-​Araúz, P., Cabezas García, M., & Reimerink, A. (2020). Representing
multiword term variation in a terminological knowledge base: A corpus-​based
study. Proceedings of the 12th Conference on Language Resources and Evaluation
(LREC), Marseille, France. 11–​16 May.
León-​Araúz, P., & Faber, P. (2024). Including the cultural dimension of terminology
in a frame-​based resource. Aspects of cognitive terminology studies: theoretical
considerations and athe role of metaphor in terminology, 55, 39–​72.
León-​Araúz, P., Reimerink, A., & Faber, P. (2017). EcoLexiCAT: A terminology-​
enhanced translation tool for texts on the environment. Proceedings of eLex
2017: Electronic Lexicography in the 21st Century.
León-​Araúz, P., & San Martín, A. (2018). The EcoLexicon semantic sketch
grammar: From knowledge patterns to Word Sketches. Proceedings of the Language
Resources and Evaluation (LREC), Miyazaki, Japan.
León-​Araúz, P., San Martín, A., & Reimerink, A. (2018). The EcoLexicon English
corpus as an open corpus in Sketch Engine. arXiv:1807.05797 [cs.CL], 893–​901.
https://​doi.org/​https://​doi.org/​10.48550/​arXiv.1807.05797
Lepage, J. F., Langlois, S., & Turcotte, M. (2019). Evolution of the language situ­
ation in Nunavut, 2001 to 2016. Retrieved 7 May, 2024, from Statistics Canada
report. www.150.statcan.gc.ca/n1/pub/89-657-x/89-657-x2019010-eng.htm
Lewandowska-​Tomaszczyk, B. (2008). Notes on lexicology and terminology.
Translation and Meaning Part 8: Proceedings of the Łódź Session of the 4th
International Maastricht –​Łódź Duo Colloquium on “Translation and Meaning,”
Łódź, Poland.
Li, L., & Ye, M. (2023). Corpora. In S.-​W. Chan (Ed.), Routledge encyclopedia of
translation technology (2nd ed., pp. 529–​550). Routledge.
Li, S., & Hope, W. (2021a). Introduction: The role of terminology translation in
China’s contemporary identities and cultures. In S. Li & W. Hope (Eds.), Terminology
translation in Chinese contexts. Theory and practice (pp. 1–​18). Routledge.
Li, S., & Hope, W. (2021b). Introduction to part II: A historical overview of termin­
ology management and scholarship. In S. Li & W. Hope (Eds.), Terminology trans-
lation in Chinese contexts. Theory and practice (pp. 123–​128). Routledge.
Li, Y., & Thimbleby, H. (2014). Hot cheese: A processed Swiss cheese model. Journal
of the Royal College of Physicians of Edinburgh, 44(2), 116–​121.
Lindsay, D. (2020). Scientific writing =​thinking in words (2nd ed.). CSIRO Publishing.
LISA. (2005). Terminology management survey. Retrieved 18 July, 2024, from www.
ter​mino​rgs.net/​downlo​ads/​term​inol​ogy_​repo​rt_​2​005.pdf, edited by Arle Lommel.
Lockwood, R. (2000). Machine translation and controlled authoring at Caterpillar. In
R. C. Sprung & S. Jaroniec (Eds.), Translating into success: Cutting-​edge strategies
for going multilingual in a global age (pp. 187–​201). John Benjamins.
López Jiménez, E., & Ouariachi, T. (2020). An exploration of the impact of artifi­
cial intelligence (AI) and automation for communication professionals. Journal of
Information, Communication and Ethics in Society, 19(2), 249–​267.
Lu, Z. (2022). Terminology work in China. Terminàlia, 25(2022), 40–​49.
Lüdeling, A., Evert, S., & Baroni, M. (2007). Using web data for linguistic purposes.
In M. Hundt, N. Nesselhauf, & C. Biewer (Eds.), Corpus linguistics and the web
(pp. 7–​24). Rodopi.
268 Bibliography

Łukasik, M. (2023). Terminology work in the era of AI-​generated content. A case

of ChatGPT (an end-​user perspective). Paper at the 5th International Conference
on Terminology: Scientific, Administrative and Educational Dimensions of
Terminology, Vilnius, 19–​20 October.
Macklovitch, E. (2014). Translation technology in Canada. In S.-​W. Chan (Ed.),
Routledge encyclopedia of translation technology (pp. 267–​278). Routledge.
Malele, N. (2021). English-​ Isindebele glossary of medical terms. Retrieved
26 October, 2024, from https://​repo.sadi​lar.org/​han​dle/​20.500.12185/​547
Malinowski, K. K. (2023). Future of terminology management in AI first world.
Retrieved 3 November, 2024, from https://​med​ium.com/​@kajet​anm/​fut​ure-​of-​term​
inol​ogy-​man​agem​ent-​in-​ai-​first-​world-​2f411​f849​7b8
Manerko, L. A., Baumann, K.-​D., & Kalverkämper, H. (Eds.). (2014). Terminology
science in Russia today: From the past to the future. Frank & Timme Verlag für
wissenschaftliche Literatur.
Manning, C. D., & Schütze, H. (1999). Foundations of statistical natural language
processing. MIT Press.
Marshall, K. (2008). The World Bank: From reconstruction to development to equity.
Routledge.
Marshman, E. (2017). Above and beyond: The future of terminology. Circuit, 136(3/​
11/​24).
Marshman, E. (2022). Knowledge patterns in corpora. In P. Faber & M.-​C. L’Homme
(Eds.), Theoretical perspectives on terminology: Explaining terms, concepts and
specialized knowledge (pp. 291–​309). John Benjamins.
Marshman, E., Gariépy, J. L., & Harms, C. (2012). Helping language professionals
relate to terms: Terminological relations and termbases. JoSTrans, the Journal of
Specialised Translation, 18, 30–​56.
Martínez, E., Mollica, F., & Gibson, E. (2022). Poor writing, not specialized concepts
drives processing difficulty in legal language. Cognition, 224(2022), 1–​7.
Maslias, R. (2017). Converting the European terminology database IATE into the
worlds’ largest multilingual data space. In P. Faini (Ed.), Terminological approaches
in the European context (pp. 13–​19). Cambridge Scholars Publishing.
Maslias, R. (2022). EU terminology in training and education. New Trends in
Translation and Technology (NETTT) conference, Rhodes Island, Greece, 4–​6 July.
Massion, F. (2024). Terminology in the age of AI: The transformation of terminology
theory and practice. Journal of Translation Studies, 1(2024), 67–​94.
Maverick, J. B. (2021). International Monetary Fund (IMF) vs. The World Bank:
What’s the difference? Retrieved 24 August, 2023, from Investopedia, www.inves​
tope​dia.com/​ask/​answ​ers/​043​015/​what-​dif​f ere​nce-​betw​een-​intern​atio​nal-​monet​
ary-​fund-​and-​world-​bank.asp
Mayoral Asensio, R. (2007). Specialised translation. A concept in need of revision.
Babel, 53(3), 48–​55.
McCallum, S., et al. (2024). What is AI, how does it work and what can it be used
for? Retrieved 13 October, 2024, from BBC News, www.bbc.co.uk/​news/​tec​hnol​
ogy-​65855​333
McEnery, A., Hardie, A., & Younis, N. (Eds.). (2021). Arabic corpus linguistics.
Edinburgh University Press.
McEnery, A., & Xiao, Z. (2007). Parallel and comparable corpora: What are they up
to? In G. James & G. Anderman (Eds.), Incorporating corpora: Translation and the
linguist (pp. 1–​13). Multilingual Matters.
 Bibliography 269

McEnery, T., & Brookes, G. (2022). Building a written corpus: What are the basics?
In A. O’Keeffe & M. McCarthy (Eds.), The Routledge handbook of corpus linguis-
tics (2nd ed., pp. 35–​47). Routledge.
McMillan, A. D. (1995). Native peoples and cultures of Canada (2nd revised and
enlarged ed.). Douglas and McIntyre Ltd.
McQuillan, L. J., & Montgomery, P. C. (Eds.). (1999). The International Monetary
Fund: Financial medic to the world? Hoover Institution Press Publication No. 456.
Mehta, R. (2024). Amazon announces Rufus, a new generative AI-​powered conversa­
tional shopping experience. Retrieved 15 October, 2024, from www.abou​tama​zon.
com/​news/​ret​ail/​ama​zon-​rufus
Mel’čuk, I. A., Clas, A., & Polguère, A. (1995). Introduction à la lexicologie explica­
tive et combinatoire. Duculot.
Melby, A. K. (2012). Terminology in the age of multilingual corpora. JoSTrans, 18.
https://www.jostrans.org/article/view/7520
Melby, A. K. (2015). TBX: A terminology exchange format for the translation and
localization industry. In H. Kockaert & F. Steurs (Eds.), Handbook of terminology
(pp. 393–​424). John Benjamins.
Melo Mora, L. F., & Toussaint, Y. (2015). Automatic validation of terminology by
means of formal concept analysis. In J. Baixeries, C. Sacarea, & M. Ojeda-​Aciego
(Eds.), Formal concept analysis (pp. 236–​251). Springer Verlag.
memoQ. (2023). Introducing memoQ adaptive generative translation. Retrieved 29
October, 2024, from memoQ webinar on YouTube, www.yout​ube.com/​watch?v=​
JPdO​1Hi_​_​iQ
Meng, W., Yu, C., & Liu, K.-​L. (2002). Building efficient and effective metasearch
engines. ACM Computing Surveys, 34(1), 48–​89.
Mesquita, A. M. C. d., & Brandão, A. P. B. (2023). Os otyahaliti, conhecimento cultural
e etnoterminologia da fauna em Paresi (Aruák) [The otyahaliti, cultural knowledge
and ethnoterminology of fauna in Paresi (Aruak)]. Revista Moara, 63, 194–​218.
Messina, C. (2018). Using corpora in terminography. Retrieved 20 September, 2024,
from https://​cots​oes.org/​sites/​defa​ult/​files/​Downlo​ads/​Using_​Corpora_​f​or_​t​ermi​
nogr​aphy​_​201​8061​8_​2.pdf
Meyer, I. (2000). Computer words in our everyday lives: How are they interesting
for terminography and lexicography? Proceedings of the Ninth EURALEX
International Congress, Stuttgart, Germany.
Meyer, I. (2022). Concept management for terminology: A knowledge engineering
approach. In P. Faber & M.-​C. L’Homme (Eds.), Theoretical perspectives on ter-
minology: Explaining terms, concepts and specialized knowledge (pp. 111–​126).
John Benjamins.
Meyer, I., Bowker, L., & Eck, K. (1992). COGNITERM: An experiment in building
a knowledge-​based term bank. Proceedings of the Fifth EURALEX International
Congress, Tampere, Finland.
Mikhailov, M. (2022). Text corpora, professional translators and translator training.
The Interpreter and Translator Trainer, 16(2), 224–​246.
Millet, M. (2013). Legal translation at INTERPOL. In A. Borja Albi & F. Prieto
Ramos (Eds.), Legal translation in context: Professional issues and prospects
(pp. 243–​262). Peter Lang.
Millour, A., & Fort, K. (2020). Text corpora and the challenge of newly written
languages. Proceedings of the First Joint SLTU and CCURL Workshop (SLTU-​
CCURL 2020), Marseille, France, 11–​16 May.
270 Bibliography

Milwaukee Magazine. (2022). Women of distinction 2022: Terra Translations.

Retrieved 6 November, 2023, from www.milwa​ukee​mag.com/​women-​of-​dist​inct​
ion-​2022-​terra-​trans​lati​ons/​
Minsky, M. (1975). A framework for representing knowledge. In P. H. Winston (Ed.),
The psychology of computer vision (pp. 211–​277). McGraw-​Hill.
Mirtaheri, S. M., Dinçtürk, M. E., Hooshmand, et al. (2014). A brief history
of web crawlers. arXiv:1405.0749. https://​doi.org/​https://​doi.org/​10.48550/​
arXiv.1405.0749
Mitchell-​Schuitevoerder, R. (2020). A project-​based approach to translation tech­
nology. Routledge.
Mlambo, R., Skosana, N., & Matfunjwa, M. (2021). The extraction of terminology
list using ParaConc for creating a quadrilingual dictionary. Southern African
Linguistics and Applied Language Studies, 39(1), 82–​91.
Montero Martínez, S., & Faber, P. (2011). Terminological competence in translation.
In A. Alcina (Ed.), Teaching and learning terminology: New strategies and methods
(pp. 89–​106). John Benjamins.
Monti, J., Elia, A., Postiglione, et al. (2011). In search of knowledge: Text mining
dedicated to technical translation. Proceedings of Translating and the Computer
33, London.
Moorkens, J., Way, A., & Lankford, S. (2024). Automating translation. Routledge.
Moravia.com. (2017). Terminology management. Retrieved 3 February, 2017, from
www.mora​via.com/​en/​servi​ces/​langu​age-​servi​ces/​term​inol​ogy-​man​agem​ent/​
Moreno-​Pérez, L. (2018). A case of hybrid text translation: Secondary-​ field ter­
minology in purchase agreements. In É. Poirier & D. Gallego-​Hernández (Eds.),
Business and institutional translation: New insights and reflections (pp. 63–​79).
Cambridge Scholars Publishing.
Morin, E., Daille, B., Takeuchi, K., et al. (2007). Bilingual terminology mining –
using brain, not brawn comparable corpora. Proceedings of the 45th Annual
Meeting of the Association of Computational Linguistics (ACL), Prague, Czech
Republic.
Morin, E., Daille, B., Takeuchi, K., et al. (2010). Brains, not brawn: The use of
“smart” comparable corpora in bilingual terminology mining. ACM Transactions
on Speech and Language Processing, 7(1).
Mörth, K. (2018). Arabic lexicography in the Internet era. In P. A. Fuertes-​Olivera
(Ed.), The Routledge handbook of lexicography (pp. 503–​517). Routledge.
Mossop, B. (1988). Translating institutions: A missing factor in translation theory. La
traduction et son public, 1(2), 65–​71.
Mossop, B. (2006). From culture to business: Federal government translation in
Canada. The Translator, 12(1), 1–​27.
Mossop, B. (2014). Motivation and de-​motivation in a government translation ser­
vice: A diary-​based approach. Perspectives. Studies in Translatology, 22(4), 581–​
591. https://​doi.org/​https://​doi.org/​10.1080/​09076​76X.2014.948​889
Mossop, B. (2015). Understanding poorly written source texts. Terminology Update,
28(2).
Mossop, B. (2023). Taking C: Canadian revision workshops to institutions abroad. In
T. Svoboda, Ł. Biel, & V. Sosoni (Eds.), Institutional translator training. Routledge.
Mottura, B. (2006). The changing terminology of Chinese public administration.
In H. Picht (Ed.), Modern approaches to terminological theories and applications
(pp. 211–​231). Peter Lang.
 Bibliography 271

Mousten, B., & Laursen, A. L. (2016). Specialized languages: The alter ego of any
research field. Academic Quarter /​Akademisk Kvarter, 13(oktober), 72–​86.
Moze, S., & Krek, S. (2019). Leveraging large corpora for translation using Sketch
Engine. In M. Ji & M. Oakes (Eds.), Advances in empirical translation studies
(pp. 110–​144). Cambridge University Press.
Muegge, U. (2007). Disciplining words: What you always wanted to know about ter­
minology management. TC World Magazine, 2(3), 17–​19.
Muegge, U. (2013). 10 things you should know about automatic terminology extrac­
tion. Retrieved 1 October, 2024, from https://​ling​uagr​eca.com/​blog/​2013/​09/​
automa​tic-​term​inol​ogy-​ext​ract​ion/​
Muegge, U. (2020). Why should you care about terminology management –​even if
you never translate a single term. Retrieved 13 July, 2024, from ATA Chronicle,
www.ata​net.org/​resour​ces/​why-​you-​sho​uld-​care-​about-​term​inol​ogy-​man​agem​ent-​
even-​if-​you-​never-​transl​ate-​a-​sin​gle-​term/​
Muegge, U. (2023). Terminology extraction for translation and interpretation made
easy. Retrieved 13 October, 2024, from www.ama​zon.co.uk/​Term​inol​ogy-​Ext​ract​
ion-​Tran​slat​ion-​Int​erpr​etat​ion-​Made/​dp/​B0C​L2ZF​XXQ
Mulders, L. (2008). Translation at the Court of Justice of the European Communities.
In S. Prechal & B. van Roermund (Eds.), The coherence of EU law; the search for
unity in divergent concepts (pp. 45–​60). Oxford University Press.
MultiLingual Magazine. (2022). Lucie Séguin talks about the past, present and
future of Canada’s Translation Bureau. Retrieved 25 April, 2024, from Interview
in MultiLingual magazine, https://​multi​ling​ual.com/​iss​ues/​septem​ber-​octo​ber-​2021/​
lucie-​seg​uin-​talks-​about-​the-​past-​pres​ent-​and-​fut​ure-​of-​cana​das-​tran​slat​ion-​bur​eau/​
Myking, J. (2001a). Sign models in terminology: Tendencies and functions. LSP and
Professional Communication, 1(2), 45–​62.
Myking, J. (2001b). Against prescriptvism? The ‘sociocritical’ challenge to termin­
ology. IITF Journal, 12(1–​2), 49–​64.
NATIONS Translation-​Group. (2022). Why Indigenous language revitalization is a
tool for reconciliation. Retrieved 5 October, 2023, from www.linke​din.com/​pulse/​
why-​ind​igen​ous-​langu​age-​rev​ital​izat​ion-​tool-​rec​onci​liat​ion-​
Nenadić, G., Spasić, I., & Ananiadou, S. (2003). Terminology-​driven mining of bio­
medical literature. Journal of Biomedical Informatics, 33(1), 83–​87.
Nimdzi. (2018). A peek inside the Canadian Translation Bureau. Retrieved 22 August,
2023, from www.nim​dzi.com/​a-​peek-​ins​ide-​the-​canad​ian-​tran​slat​ion-​bur​eau/​
Nkwenti-​Azeh, B. (2001). Term banks. In M. Baker (Ed.), The Routledge encyclo­
pedia of translation studies (pp. 249–​251). Routledge.
Novak, J. D., & Cañas, A. J. (2008). The theory underlying concept maps and how to
construct and use them. Retrieved 1 October, 2024, from Technical Report IHMC
CmapTools 2006-​01 Rev 01-​2008, Institute for Human and Machine Cognition,
2008, available at http://​cmap.ihmc.us/​docs/​pdf/​Theo​ryUn​derl​ying​Conc​eptM​
aps.pdf
Nuopponen, A. (1994). Begreppssystem för terminologisk analys [concept systems
for terminological analysis]. University of Vaasa.
Nuopponen, A. (2006, 25–​26 August). A model for structuring concept systems
of activity. Proceedings of the Third International Conference on Terminology,
Standardization and Technology Transfer (TSTT), Beijing.
272 Bibliography

Nuopponen, A. (2016, 22–​24 June). Satellite system as a visualization tool for con­
cept analysis. Proceedings of the 12th International Conference on Terminology
Knowledge and Engineering (TKE), Copenhagen.
Nuopponen, A. (2018). Dimensions of terminology work. Terminologija, 25, 6–​22.
Nuopponen, A. (2022). Conceptual relations: From the General Theory of Terminology
to knowledge bases. In P. Faber & M.-​C. L’Homme (Eds.), Theoretical perspectives
on terminology: Explaining terms, concepts and specialized knowledge (pp. 63–​
85). John Benjamins.
O’Keeffe, A., & McCarthy, M. (2022a). Of what is past, or passing, or to come: Corpus
linguistics, changes and challenges. In A. O’Keeffe & M. McCarthy (Eds.), The
Routledge handbook of corpus linguistics (2nd ed., pp. 1–​9). Routledge.
O’Keeffe, A., & McCarthy, M. (Eds.). (2022b). The Routledge handbook of corpus
linguistics (2nd ed.). Routledge.
O’Mara, B., & Carey, G. (2019). Do multilingual androids dream of a better life
in Australia? Effectiveness of information technology for government transla-
tion to support refugees and migrants in Australia. Australian Journal of Public
Administration, 78(3), 449–​471.
Oakes, L., & Warren, J. (2007). Language, citizenship and identity in Quebec.
Palgrave Macmillan.
Ogden, C. K., & Richards, I. A. (1923). The meaning of meaning: A study of the
influence of language upon thought and the science of symbolism. Harcourt, Brace
& World Inc.
Olohan, M. (2010). Commercial translation. In Y. Gambier & L. Van Doorslaer
(Eds.), Handbook of translation studies (vol. 1, pp. 41–​44). John Benjamins.
Retrieved 31 July, 2018, from https://​benjam​ins.com/​onl​ine/​hts/​intro​duct​ion.html
Olohan, M. (2016). Scientific and technical translation. Routledge.
Olsen, F., Lorz, A., & Stein, D. (Eds.). (2009). Translation issues in language and law.
Palgrave Macmillan.
One Word. (2024). Term extraction and AI: How to get it right. Retrieved 23
September, 2024, from https://​www.onew​ord.de/​en/​term-​ext​ract​ion-​ai/​
OPTIMALE. (2013). Optimising professional translator training in a multilingual
Europe. Progress report –​public part. Retrieved 14 March, 2017, from https://​
ec.eur​opa.eu/​pro​gram​mes/​eras​mus-​plus/​proj​ect-​res​ult-​cont​ent/​0bcd8​0b1-​59eb-​
4f2f-​88db-​dc519​b043​329/​59-​ENWA-​FR-​RENNE​S02.pdf
PACTE Group, Beeby, A., Fernández Rodríguez, M., Fox, O., et al. (2003).
Building a translation competence model. In F. Alves (Ed.), Triangulating transla-
tion: Perspectives in process-​oriented research (pp. 43–​66). John Benjamins.
PACTE Group, Hurtado Albir, A., Galán-​ Mañas, A., Kuznik, A., et al. (2020).
Translation competence acquisition. Design and results of the PACTE Group’s
experimental research. The Interpreter and Translator Trainer, 14(2), 95–​233.
Pandya, V., Zhang, H., Huertas-​ Company, M., et al. (2024). Galaxies going
bananas: Inferring the 3D geometry of high-​ redshift galaxies with JWST-​
CEERS. arXiv,:2310.15232v2 [astro-​ph.GA], 1–​47. https://​arxiv.org/​pdf/​2310.
15232
Pantcheva, M. (2023). Terminology management made easier with large language
models. Retrieved 10 May, 2025, from www.rws.com/​blog/​term​inol​ogy-​man​agem​
ent-​made-​eas​ier-​with-​large-​langu​age-​mod​els/​
 Bibliography 273

Patrick, D. (2018). Arctic languages in Canada in the age of globalization. In G.

Hogan-​Brun & B. O’Rourke (Eds.), The Palgrave handbook of minority languages
and communities (pp. 257–​284). Palgrave Macmillan.
Pavel, S. (1993). Neology and phraseology as terminology in the making. In H. B.
Sonneveld & K. L. Loening (Eds.), Terminology. Applications in interdisciplinary
communication (pp. 21–​34). John Benjamins.
Pavel, S., & Nolet, D. (2001). Handbook of terminology. Minister of Public Works
and Government Services, Canada.
Pearson, J. (1998). Terms in context. John Benjamins.
Pecman, M., & Kübler, N. (2022). Text genres and terminology. In P. Faber & M.-​
C. L’Homme (Eds.), Theoretical perspectives on terminology: Explaining terms,
concepts and specialized knowledge (pp. 263–​289). John Benjamins.
Penet, J. C. (2024). Working as a professional translator. Routledge.
Periñán-​Pascual, C. (2017). DEXTER: A workbench for automatic term extraction
with specialized corpora. Natural Language Engineering, 24(2), 163–​198.
Peters, P., & Fernández, T. (2017). Lexicography and applied linguistics. In P. A.
Fuertes-​Olivera (Ed.), The Routledge handbook of lexicography (pp. 105–​122).
Routledge.
Petit, G. (2001). L’introuvable identité du terme technique. Revue française de
linguistique apliquée, VI, 63–​79.
Phrase (n.d.). Harness GenAI to scale your localization. Retrieved 29 October, 2024,
from https://​phr​ase.com/​gen​erat​ive-​ai-​local​izat​ion/​?utm_​c​ampa​ign=​Email%20%7C
%20Le ​ g al&utm​ _ ​ m ed​ i um=​ e mail&_​ h s​ e nc=​ p 2AN ​ q tz- ​ 9 i3K1PhzaCBZAen ​ g PkI​
oTgY​MQG7​wpg5​bZ51​zorV​rxF-​ltyRVDWtKVsiHRRyLX6OWreYd8OlzPIhzg2
6sUTqUie​Ynhc​abW8​3a6t​mxKD​9cYw​3TOD​9o4&_​hsmi=​299025​207&utm_​cont​
ent=​299025​207&utm​_​sou​rce=​hs_​em​ail
Picht, H. (Ed.). (2006). Modern approaches to terminological theories and
applications. Peter Lang.
Picht, H. (2011). The science of terminology: History and evolution. Terminologija,
18(2011), 6–​26.
Picken, C. (1996). The vital role of the generalist in technical translation from and into
‘minor’ languages. In H. Somers (Ed.), Terminology, LSP and translation: Studies
in language engineering in honour of Juan C. Sager (pp. 169–​173). John Benjamins.
Pilarski, K. (2023). Wauwatosa women-​ owned businesses recognized for their
success at annual marketplace conference. Retrieved 6 November, 2023, from
https://​eu.jsonl​ine.com/​story/​comm​unit​ies/​west/​crime/​2017/​12/​29/​wauwat​osa-​
women-​owned-​bus​ines​ses-​rec​ogni​zed-​their-​succ​ess-​ann​ual-​mark​etpl​ace-​con​fere​
nce/​990350​001/​
Pilitsidou, V., & Giouli, V. (2020). Frame semantics in the specialized domain of
finance: Building a termbase to aid translation. Proceedings of the XIX Congress
of the European Association for Lexicography (EURALEX), Alexandroupolis,
Greece, 7–​11 September 2021.
Pinchuck, I. (1977). Scientific and technical translation. Westview Press.
Pinnis, M., Ljubešić, N., Ştefănescu, D., Set al. (2012). Term extraction, tagging and
mapping tools for under-​resourced languages. Proceedings of the 10th Terminology
and Knowledge Engineering Conference (TKE 2012), Madrid, Spain, 19–​22 June.
274 Bibliography

Poirier, É., & Gallego-​Hernández, D. (Eds.). (2018). Business and institutional trans­
lation: New insights and reflections. Cambridge Scholars Publishing.
Popiolek, M. (2015). Terminology management within a translation quality assurance
process. In H. Kockaert & F. Steurs (Eds.), Handbook of terminology (pp. 341–​
359). John Benjamins.
Presas, M. (2000). Developing translation competence. In C. Schäffner & B. Adab
(Eds.), Bilingual competence and translation competence (pp. 19–​ 31). John
Benjamins.
Prieto Ramos, F. (2018). Institutional translation: Surveying the landscape at inter­
national organizations. In F. Prieto Ramos (Ed.), Institutional translation for
international governance: Enhancing quality in multilingual legal communication
(pp. 1–​6). Bloomsbury.
Prieto Ramos, F. (2020a). Translation at international organizations. The legal and
linguistic hierarchies of multilingualism. In M. Ji & S. Laviosa (Eds.), The Oxford
handbook of translation and social practices (pp. 455–​477). Oxford University
Press.
Prieto Ramos, F. (2020b). Facing translation errors at international organizations: What
corrigenda reveal about correction processes and their implications for translation
quality. Comparative Legilinguistics, 41(1), 97–​133.
Prieto Ramos, F. (2020c). Ensuring consistency and accuracy of legal terms in
institutional translation. The role of terminological resources in international
organizations. In F. Prieto Ramos (Ed.), Institutional translation and interpreting.
Assessing practices and managing for quality (pp. 128–​149). Routledge.
Prieto Ramos, F., Cerutti, G., & Guzmán, D. (2019). Building representative multi-​
genre corpora for legal and institutional translation research: The LETRINT
approach to text categorization and stratified sampling. Translation Spaces, 8(1),
93–​116.
Prieto Ramos, F., & Guzmán, D. (2021). Examining institutional translation through
a legal lens. A comparative analysis of multilingual text production at international
organizations. Target, 33(2), 254–​281.
Prieto Ramos, F., & Guzmán, D. (2023). Institutional translation profiles. A com­
parative analysis of descriptors and requirements. In T. Svoboda, Ł. Biel, & V.
Sosoni (Eds.), Institutional translator training (pp. 49–​72). Routledge.
Prieto Ramos, F., & Morales Moreno, A. (2019). Terminological innovation and har­
monization at international organizations: Can too many cooks spoil the broth?
In I. Simonnæs & M. Kristiansen (Eds.), Legal translation. Current issues and
challenges in research, methods and applications (pp. 87–​112). Frank & Timme
GmbH.
Prinsloo, D., Prinsloo, J., & Prinsloo, D. (2018). African lexicography in the Internet
era. In P. A. Fuertes-​ Olivera (Ed.), The Routledge handbook of lexicography
(pp. 487–​502). Routledge.
PSPC. (2017). Translation Bureau invest in its future by hiring students. Retrieved 10
May, 2025, www.can​ada.ca/​en/​pub​lic-​servi​ces-​proc​urem​ent/​news/​2017/​08/​trans-
lation_​bureauinvest​init​sfut​ureb​yhir​ings​tude​nts.html
Pym, A. (2004). The use of translation in international organizations. In H. Kittel,
A. P. Frank, & N. Greiner (Eds.), Übersetzung – translation – traduction. Ein
internationales Handbuch zur Übersetzungsforschung (pp. 85–​92). De Gruyter.
Quah, C. K. (2006). Translation and technology. Palgrave Macmillan.
 Bibliography 275

Quickbase. (2024). A timeline of database history & database management. Retrieved

30 April, 2024, from www.quickb​ase.com/​artic​les/​timel​ine-​of-​datab​ase-​hist​ory
Quinci, C. (2023). Translation competence. Routledge.
Ramuedzisi, L. S., van Huyssteen, L., & Mandende, I. P. (2019). An enhanced termin­
ology development and management approach for South African languages. South
African Journal of African Languages, 39(3), 263–​272.
Reason, J. (1990). Human error. Cambridge University Press.
Reason, J., Hollnagel, E., & Paries, J. (2006). Revising the “Swiss cheese” model
of accidents. Retrieved 10 August, 2024, from www.resea​rchg​ate.net/​publ​icat​ion/​
285486777_​Revisiting_​the_​Swis​s_​Ch​eese​_​Mod​el_​o​f_​Ac​cide​nts
Reed, D. (1993). Some terminological problems of translating common law concepts
from English into French. In H. B. Sonneveld & K. L. Loening (Eds.), Terminology.
Applications in interdisciplinary communication (pp. 79–​86). John Benjamins.
Reimerink, A., & León-​Araúz, P. (2018). Manzanilla: An image annotation tool for
TKB building. Proceedings of the 11th International Conference on Language
Resources and Evaluation (LREC), Miyazaki, Japan, May 2018.
Reimerink, A., León-​Araúz, P., & Faber, P. (2016). Image selection and annotation
for an environmental knowledge base. Language Resources and Evaluation, 50,
443–​474.
Renouf, A., & Kehoe, A. (2013). Filling the gaps: Using the WebCorp linguist’s search
engine to supplement existing text resources. International Journal of Corpus
Linguistics, 18(2), 167–​198.
Renouf, A., Kehoe, A., & Banerjee, J. (2007). WebCorp: An integrated system for
web text search. In M. Hundt, N. Nesselhauf, & C. Biewer (Eds.), Corpus linguis-
tics and the web (pp. 47–​67). Rodopi.
Reppen, R. (2022). Building a corpus: What are the key considerations? In A. O’Keeffe
& M. McCarthy (Eds.), The Routledge handbook of corpus linguistics (2nd ed.,
pp. 13–​20). Routledge.
Rey, A. (1995). Essays on terminology. John Benjamins.
Richards, C. (2023). What has the World Bank’s internal Evaluation Group learned
from experimenting with NLP? Retrieved 7 May, 2024, from https://​merlt​ech.org/​
what-​has-​the-​world-​banks-​inter​nal-​eva​luat​ion-​group-​lear​ned-​from-​experi​ment​
ing-​with-​nlp/​
Riggs, F. W. (1989). Terminology and lexicography: Their complementarity.
International Journal of Lexicography, 2(2), 89–​110.
Ritter, M. (2023). Why plant names change. Retrieved 28 July, 2024, from https://​
paci​fich​orti​cult​ure.org/​artic​les/​why-​plant-​names-​cha​nge-​2/​
Roberts, R. P. (2019). Translation in Canada. Hermēneus, Revista de Traducción e
Interpretación, 21(2019), 1–​6.
Rogers, M. (2000). Genre and terminology. In A. Trosborg (Ed.), Analysing profes­
sional genres (pp. 3–​20). John Benjamins.
Rogers, M. (2015). Specialised translation. Shedding the ‘non-​literary’ tag. Palgrave
Macmillan.
Römer, U., & Wulff, S. (2010). Applying corpus methods to written academic
texts: Explorations of MICUSP. Journal of Writing Research, 2(2), 99–​127.
Rothwell, A., Moorkens, J., Fernández-​Parra, et al. (2023). Translation tools and
technologies. Routledge.
Rousseau, D., Billingham, J., & Calvo-​Amodio, J. (2018). Systemic semantics: A systems
approach to building ontologies and concept maps. Systems (Basel), 6(3), 1–​24.
276 Bibliography

Rousseau, L.-​J. (1993). Terminology and languages in contact in Québec. In H. B.

Sonneveld & K. L. Loening (Eds.), Terminology. Applications in interdisciplinary
communication (pp. 35–​41). John Benjamins.
RWS Trados (n.d.). Why is terminology management important? Retrieved 8 August,
2024, from www.rws.com/​local​izat​ion/​servi​ces/​resour​ces/​term​inol​ogy-​man​agem​ent/​
RWS Trados. (n.d. 2). Harness the power of AI in Trados. Retrieved 29 October, 2024,
from www.tra​dos.com/​prod​uct/​disco​ver/​AI-​tran​slat​ion/​#:~:text=​AI%20As​sist​ant
%20enab ​ l es%20tran ​ s lat ​ o rs%20to,or%20sho ​ r ten ​ i ng%20se ​ n ten ​ c es%2C%20
for%20exam​ple
SACLANT (2002). Intelligence exploitation of the Internet. Retrieved 19 October,
2024, from https://​nsa​rchi​ve2.gwu.edu/​NSA​EBB/​NSAEBB​436/​docs/​EBB-​005.pdf
SADiLaR (2021). COVID-​19 multilingual terminology. Retrieved 26 October, 2024,
from technical report, available at https://​repo.sadi​lar.org/​han​dle/​20.500.12185/​544
Sager, J. (1990). A practical course in terminology processing. John Benjamins.
Sager, J. (1997). Term formation. In S. E. Wright & G. Budin (Eds.), Handbook of
terminology management (pp. 25–​41). John Benjamins.
Sager, J. (2001a). Terminology. Applications. In M. Baker (Ed.), The Routledge
encyclopedia of translation studies (pp. 251–​255). Routledge.
Sager, J. (2001b). Terminology. Theory. In M. Baker (Ed.), The Routledge encyclo­
pedia of translation studies (pp. 258–​262). Routledge.
Sager, J. (2001c). Terminology. Standardization. In M. Baker (Ed.), The Routledge
encyclopedia of translation studies (pp. 255–​258). Routledge.
San Martín, A. (2024a). What generative artificial intelligence means for termino­
logical definitions. arXiv: https://​arxiv.org/​pdf/​2402.16139. https://​arxiv.org/​pdf/​
2402.16139
San Martín, A. (2024b). Generative AI and the future of terminological definitions
RWS Campus 2024 [online], www.rws.com/​about/​care​ers/​eve​nts/​rws-​cam​pus-​onl​
ine-​2024/​
Sánchez Gijón, P. (2009). Developing documentation skills to build do-​it-​yourself cor­
pora in the specialised translation course. In A. Beeby Lonsdale, P. Rodríguez Inés,
& P. Sánchez Gijón (Eds.), Corpus use and translating: Corpus use for learning to
translate and learning corpus use to translate (pp. 109–​127). John Benjamins.
Sandrini, P. (1996). Comparative analysis of legal terms: Equivalence revisited.
Proceedings of the Fourth International Congress on Terminology and Knowledge
Engineering (TKE96), Frankfurt.
Saralegi, X., San Vicente, I., & Gurrutxaga, A. (2008). Automatic extraction
of bilingual terms from comparable corpora in a popular science domain. In P.
Zweigenbaum, É. Gaussier, & P. Fung (Eds.), Proceedings of LREC 2008 Workshop
on building and using comparable corpora (pp. 27–​32).
Šarčević, S. (2000). New approach to legal translation. Kluwer Law International.
Šarčević, S. (2018). Challenges to legal translators in institutional settings. In F. Prieto
Ramos (Ed.), Institutional translation for international governance (pp. 9–​24).
Bloomsbury.
Šarčević, S., & Robertson, C. (2013). The work of lawyer-​ linguists in the EU
institutions. In A. Borja Albi & F. Prieto Ramos (Eds.), Legal translation in con-
text: Professional issues and prospects. Peter Lang.
Scarpa, F. (2020). Professional practice in specialised translation. Palgrave Macmillan.
Schäfer, R., & Bildhauer, F. (2013). Web corpus construction. Morgan & Claypool
Publishers.
 Bibliography 277

Schäffner, C., Tcaciuc, L. S., & Tesseur, W. (2014). Translation practices in polit­
ical institutions: A comparison of national, supranational and non-​governmental
organisations. Perspectives: Studies in Translatology, 22(4), 493–​510.
Schmitt, N., & Carter, R. (2004). Formulaic sequences in action. In N. Schmitt (Ed.),
Formulaic sequences. Acquisition, processing and use. John Benjamins.
Schmitz, K.-​D. (1996). Terminology management systems. In R. Owens (Ed.), The
translator’s handbook (pp. 221–​246). ASLIB.
Schmitz, K.-​D. (2006). Terminology and terminological databases. In K. Brown (Ed.),
Encyclopedia of language and linguistics (2nd ed., pp. 578–​587). Elsevier.
Schmitz, K.-​D. (2009, 6-​10 July 2009). Applied principles of terminology work.
Terminology Summer School, Cologne.
Schwab, K. (2016). The fourth industrial revolution: What it means, how to respond.
Retrieved 13 October, 2024, from www.wefo​rum.org/​age​nda/​2016/​01/​the-​fou​rth-​
ind​ustr​ial-​rev​olut​ion-​what-​it-​means-​and-​how-​to-​resp​ond/​
Seidel, E., & Durban, C. (2010). The prosperous translator. FA&WB Press.
Shavit, K., Wagner, A., Schertel, L., et al. (2023). A tunable reflector enabling
crustaceans to see but not be seen. Science, 379, 695–​700.
Shelov, S. D., & Leitchik, V. M. (2006). Terminology science in Russia: Social needs
and object content. In H. Picht (Ed.), Modern approaches to terminological the-
ories and applications (pp. 13–​23). Peter Lang.
Sheng, D. (2024). A concept map approach to knowledge competence acquisition for
student interpreters. PLOS ONE, 19(1), 1–​16.
Shreve, G. M. (1997). Cognition and the evolution of translation competence. In J. H.
Danks, G. M. Shreve, S. B. Fountain et al.(Eds.), Cognitive processes in translation
and interpreting (pp. 120–​136). Sage.
Shumeli, A., Shehu, E., Mine, L. et al. (2023). Terminology evolution in economic
discourse: Unravelling reterminologization in “Economist Magazine.” Proceedings
of the 4th International Conference on Agriculture and Life Sciences (ICOALS 4),
Tirana, Albania.
Simard, M. (2020). Building and using parallel text for translation. In M. O’Hagan
(Ed.), The Routledge handbook of translation and technology (pp. 78–​90). Routledge.
Sinclair, J. M. (1991). Corpus, concordance, collocation. Oxford University Press.
Smithsonian Ocean. (2018). Bioluminescence. Retrieved 21 September, 2024, from
https://​ocean.si.edu/​ocean-​life/​fish/​biol​umin​esce​nce
SOA Policy Team. (2024). SOA [Society of Authors] survey reveals a third of
translators and quarter of illustrators losing work to AI. Retrieved 15 October,
2024, from https://​socie​tyof​auth​ors.org/​2024/​04/​11/​soa-​sur​vey-​reve​als-​a-​third-​
of-​tran​slat​ors-​and-​quar​ter-​of-​illus​trat​ors-​los​ing-​work-​to-​ai/​#:~:text=​Approx​imat​
ely%201%20in%205%20resp​onde​nts,%2Dfict​ion%20writ​ers%20(25%25)
Sprung, R. C., & Jaroniec, S. (Eds.). (2000). Translating into success: Cutting-​edge
strategies for going multilingual in a global age. John Benjamins.
Stasimioti, M. (2023). Here are 30 new jobs language industry CEOs expect to hire
for in the AI age. Retrieved 29 October, 2024, from Slator article available at
https://​sla​tor.com/​30-​new-​jobs-​langu​age-​indus​try-​ceos-​exp​ect-​to-​hire-​ai-​age/​
Statistics Canada. (2023). Indigenous languages across Canada. Retrieved 3 October,
2023 from https://​www12.stat​can.gc.ca/​cen​sus-​rece​nsem​ent/​2021/​as-​sa/​98-​200-​X/​
2021​012/​98-​200-​X2021​012-​eng.cfm
Stefaniak, K. (2017). Terminology work in the European Commission: Ensuring
high-​quality translation in a multilingual environment. In T. Svoboda, Ł. Biel,
278 Bibliography

& K. Łoboda (Eds.), Quality aspects in institutional translation (pp. 109–​122).

Language Science Press.
Stefaniak, K. (2022). Machine translation and terminology: The experience at the
European Commission. New Trends in Translation and Technology (NETTT) con-
ference, Rhodes Island, Greece, 4–​6 July
Štefková, M., & Tužinská, H. (2021). Translation, interpreting and institutional
routines: The case of Slovakia. In M. Bourguignon, B. Nouws, & H. van Gerwen
(Eds.), Translation policies in legal and institutional settings (pp. 157–​180). Leuven
University Press.
Sterritt, A. (2020). Reconciliation in Canada. Retrieved 5 October, 2023, from
Article in The Canadian Encyclopedia, www.thec​anad​iane​ncyc​lope​dia.ca/​en/​arti​
cle/​rec​onci​liat​ion-​in-​can​ada#:~:text=​In%20Can​ada%2C%20the%20proc​ess%20
of,ongo​ing%20effe​cts%20on%20Ind​igen​ous%20peop​les.
Steurs, F., De Wachter, K., & De Malsche, E. (2015). Terminology tools. In H.
Kockaert & F. Steurs (Eds.), Handbook of terminology, vol. 1 (pp. 222–​249). John
Benjamins.
Stolze, R. (2009). Dealing with cultural elements in technical texts for translation.
JoStrans –​The Journal of Specialised Translation, 11. Retrieved 27 October, 2010,
from www.jostr​ans.org/​issu​e11/​art​_​sto​lze.php
Suonuuti, H. (2001). Guide to terminology. Tekniikan Sanastokeskus. Retrieved 17
August, 2024, from Nordterm 8, 2nd ed. www.nordt​erm.net/​filer/​publik​atio​ner/​gui​
der/​Guide​_​to_​Term​inol​ogy.pdf
Svoboda, T., Biel, Ł., & Łoboda, K. (Eds.). (2017). Quality aspects in institutional
translation. Retrieved 10 June, 2018, from Language Science Press. http://​lang​sci-​
press.org/​cata​log/​book/​181
Svoboda, T., Biel, Ł., & Sosoni, V. (Eds.). (2023). Institutional translator training.
Routledge.
Taljard, E., & de Schryver, G.-​M. (2002). Semi-​automatic term extraction for the
African languages, with special reference to northern Sotho. Lexikos, 12(AFRILEX-​
reeks/​series 12:2002), 44–​74.
Tamás, D. M., Papp, E., & Petz, A. (2016). The translator as terminologist. In I.
Horváth (Ed.), The modern translator and interpreter (pp. 57–​ 77). Eötvös
University Press.
Tamás, D. M., & Sermann, E. (2019). Evaluation system for online terminological
databases. Terminologija, 26, 24–​46.
Tavares, C., Oliveira, L., Duarte, P. et al. (2023). Artificial intelligence: A blessing or a
threat for language service providers in Portugal. Informatics, 10(81), 1–​21.
Tcaciuc, L. S., & Mackevic, V. (2017a). Romanian economic terminology in European
Union institutions. In P. Faini (Ed.), Terminological approaches in the European
context (pp. 305–​327). Cambridge Scholars Publishing.
Tcaciuc, L. S., & Mackevic, V. (2017b). Translators’ agency in translating economic
metaphors in European Union institutions: The case of the European Central Bank.
Perspectives, 25(3), 417–​433.
Temmerman, R. (2000). Towards new ways of terminology description: The
sociocognitive approach. John Benjamins.
Temmerman, R. (2007). Approaches to terminology. Now that the dust has settled...
SYNAPS, 20, 27–​36.
Temmerman, R. (2022). Units of understanding in sociocognitive terminology
studies. In P. Faber & M.-​ C. L’Homme (Eds.), Theoretical perspectives on
 Bibliography 279

terminology: Explaining terms, concepts and specialized knowledge (pp. 331–​352).

John Benjamins.
Temmerman, R., & Kerremans, K. (2003). Termontography: Ontology building
and the sociocognitve approach to terminology description. In E. Hajicová, A.
Kotešovcová, & J. Mírovský (Eds.), Proceedings of CIL17 (pp. 1–​10). Matfyzpress.
ten Hacken, P. (Ed.). (2006). Terminology, computing and translation. Narr.
ten Hacken, P. (2008). Prototypes and discreteness in terminology. Proceedings of the
XIII Euralex International Congress, IULA-​UPF, Barcelona.
ten Hacken, P. (2015). Terms and specialized vocabulary: Taming the prototypes. In
H. Kockaert & F. Steurs (Eds.), Handbook of terminology, vol. 1 (pp. 3–​13). John
Benjamins.
TerminOrgs. (2014). TBX-​basic. Retrieved 11 August, 2024, from www.ter​mino​rgs.
net/​TBX-​Basic.html
TerminOrgs. (2016). Terminology starter guide. Retrieved 21 August, 2023, from
www.ter​mino​rgs.net/​downlo​ads/​Ter​minO​rgs_​Star​terG​uide​_​V2.pdf
TerminOrgs. (2022). Term extraction starter guide. Retrieved 21 August, 2023, from
www.ter​mino​rgs.net/​downlo​ads/​Ter​minO​rgs_​Term​Extr​acti​onGu​ide.pdf
Terra Translations. (2018a). A professional translator’s guide to ATA certification.
Retrieved 6 November, 2023, from https://​terrat​rans​lati​ons.wordpr​ess.com/​2018/​
11/​16/​a-​profe​ssio​nal-​tran​slat​ors-​guide-​to-​ata-​certif​i cat​ion/​
Terra Translations. (2018b). Terra Cares program. Retrieved 6 November, 2023,
from https://​terrat​rans​lati​ons.com/​2018/​11/​13/​terra-​cares-​prog​ram/​
Terra Translations. (2021). Terra ranked as second largest language services provider
in Latin America. Retrieved 6 November, 2023, from https://​terrat​rans​lati​ons.
com/​2021/​07/​12/​terra-​ran​ked-​as-​2nd-​larg​est-​langu​age-​servi​ces-​provi​der-​in-​latin-​
amer​ica/​
Terra Translations. (2023a). Terra Translations. Retrieved 6 November, 2023, from
https://​terrat​rans​lati​ons.com/​
Terra Translations. (2023b). Services. Retrieved 6 November, 2023, from https://​
terrat​rans​lati​ons.com/​servi​ces/​
Terra Translations. (2023c). Blog. Retrieved 6 November, 2023, from https://​terrat​
rans​lati​ons.com/​blog/​
Terra Translations. (2023d). Our team. Retrieved 6 November, 2023, from https://​
terrat​rans​lati​ons.com/​our-​team/​
Terra Translations. (2023e). Certifications and awards. Retrieved 6 November, 2023,
from https://​terrat​rans​lati​ons.com/​tag/​cer​tifi​cati​ons-​awa​rds/​
Thelen, M. (2002a). Relations between terms: A cognitive approach. The interaction
between terminology, lexicology, translation studies and translation practice.
Linguistica Antverpiensia, I, 193–​209.
Thelen, M. (2015). The interaction between terminology and translation. trans-​kom,
8(2), 347–​381.
Thije, J. D. t., & Pinto, M. (2023). Cultural filters in persuasive texts: A contrastive
study of Dutch and Italian IKEA catalogs. In R. Supheert, G. Cascio, & J. D. T.
Thije (Eds.), The riches of intercultural communication (pp. 123–​135). Brill.
Thoiron, P., & Béjoint, H. (2010). La terminologie, une question de termes? Meta.
Journal des traducteurs. Translator’s Journal, 55(1), 105–​118.
Thomas, P. (1993). Choosing headwords from language-​for-​special-​puposes (LSP)
collocations for entry into a terminology data bank. In H. B. Sonneveld & K. L.
Loening (Eds.), Terminology. Applications in interdisciplinary communication
(pp. 43–​68). John Benjamins.
280 Bibliography

Thundil, A. J. (2024). Images, patterns and audio: A search for an identifier of genre
inclusive of science fiction. International Journal of Research and Review, 11(10),
73–​79.
Thurmair, G. (2003). Making term extraction tools usable. Controlled Language
Translation, EAMT-​CLAW 03, Joint Conference combining the 8th International
Workshop of the European Association for Machine Translation and the 4th
Controlled Language Applications Workshop, Dublin City University, 15–​17 May.
Tiedemann, J. (2011). Bitext alignment. Springer.
Tiedemann, J. (2012). Parallel data, tools and interfaces in OPUS. Proceedings of
the 8th International Conference on Language Resource and Evaluation (LREC),
Istanbul.
Tomić, A., & Beltrán Montoliu, A. (2013). Translation at the International Criminal
Court. In A. Borja Albi & F. Prieto Ramos (Eds.), Legal translation in con-
text: Professional issues and prospects (pp. 221–​242). Peter Lang.
Topping, S. (2000). Shortening the translation cycle at Eastman Kodak. In R. C.
Sprung & S. Jaroniec (Eds.), Translating into success: Cutting-​edge strategies for
going multilingual in a global age (pp. 111–​125). John Benjamins.
Toury, G. (1986). Natural translation and the making of a native translator.
Textcontext, 1, 11–​29.
Toury, G. (1995). Descriptive translation studies. John Benjamins.
Translation Bureau. (2023). Translation Bureau organizational information. Retrieved
4 October, 2023, from www.tpsgc-​pwgsc.gc.ca/​bt-​tb/​apro​pos-​about-​eng.html#a3
Translation Bureau (n.d.). Termium 7 for Translation Bureau. Retrieved 15 August,
2023, from Request for Information EN976-​202708, https://​buy​ands​ell.gc.ca/​cds/​
pub​lic/​2021/​01/​05/​c621c​496d​7d95​62cc​588b​dcaa​a4c8​812/​ABES.PROD.PW_​_​
EE.B071.E38​871.ATTA​001.PDF
Trojar, M. (2017). Wüster’s view of terminology. Slovenski jezik /​Slovene Linguistic
Studies, 11(2017), 55–​85.
Trombetti, M. (2023). Translated integrates GPT-​4 into Matecat for faster and more
accurate translations. Retrieved 29 October, 2024, from https://​tra​nsla​ted.com/​
mate​cat-​gpt-​4
Txabarriaga, R. (2023). Here are the 11 language service providers that made the
2023 inc. 5,000 list. Retrieved 6 November, 2023, from Slator article available at
https://​sla​tor.com/​11-​langu​age-​serv​ice-​provid​ers-​made-​the-​2023-​inc-​5000-​list/​
U.S. Department of Energy (2022). DOE national laboratory makes history by
achieving fusion ignition. Retrieved 10 May, 2025, from www.ene​rgy.gov/​artic​les/​
doe-​natio​nal-​lab​orat​ory-​makes-​hist​ory-​achiev​ing-​fus​ion-​ignit​ion
United Nations (2024). UN system documentation. Retrieved 8 July, 2024, from
https://​resea​rch.un.org/​en/​docs/​unsys​tem/​sa, UN Documentation Research Guides.
University of Nairobi (2019). Lecture 1: Mineralogy and crystallography. Retrieved
27 July, 2024, from http://​stude​ntor​gani​sati​ons.uonbi.ac.ke/​nugsa/​wp-​cont​ent/​
uplo​ads/​sites/​45/​2019/​11/​SGL-​201-​lectur​e_​1-​_​min​eral​ogy_​and_​crys​tall​ogra​phy-​
3_​rev​iew.pdf
Ureña Gómez-​Moreno, J. M., & Ruiz de Mendoza Ibáñez, J. (2022). Conceptual
metaphors. In P. Faber & M.-​C. L’Homme (Eds.), Theoretical perspectives on ter-
minology: Explaining terms, concepts and specialized knowledge (pp. 377–​396).
John Benjamins.
 Bibliography 281

Vacanti, M. (2023). Examining how and why scientific names change. Retrieved
28 July, 2024, from https://​natur​emus​eum.org/​chic​ago-​acad​emy-​of-​scien​ces/​blog/​
examin​ing-​how-​and-​why-​sci​enti​fic-​names-​cha​nge#:~:text=​Most%20n​ame%20
chan​ges%20oc​cur%20beca​use,our%20kn​owle​dge%20le​ads%20to%20re​clas​sifi​
cati​ons
Våge, O. K. (2018). Terminologiske ressurser i Norge: Nye brukergrupper og
bruksområder. LexicoNordica, 25, 173–​191.
Vaidya, N., & Solgaard, O. (2022). Immersion graded index optics: Theory, design,
and prototypes. Microsystems & Nanoengineering, 8(69).
Valderrábanos, A. S., Belskis, A., & Iraola Moreno, L. (2002). Multilingual termin­
ology extraction and validation. Proceedings of the Third International Conference
on Language Resources adn Evaluation (LREC), Las Palmas, Canary Islands, Spain.
Valeontis, K., & Mantzari, E. (2006). The linguistic dimension of termin­
ology: Principles and methods of term formation. Proceedings of the 1st Athens
International Conference on Translation and Interpretation: Between Art and
Social Science, Athens, 13–​14 October.
Valero Garcés, C. (2018). Interview with the Spanish language department at DGT.
FITISPos International Journal, 5(1), 114–​122. Retrieved 30 July, 2018, from
http://​www3.uah.es/​fiti​spos​_​ij/​OJS/​ojs-​2.4.5/​index.php/​fitis​pos/​arti​cle/​viewF​ile/​
189/​178
Van den Bossche, P. (2007). Regulatory legitimacy of the role of NGOs in global gov­
ernance: Legal status and accreditation. In A. Vedder, V. Collingwood, A. van Gorp
et al (Eds.), NGO involvement in international governance and policy: Sources of
legitimacy (pp. 135–​173). Martinius Nijhoff Publishers.
Van Dormael, A. (1978). Bretton Woods: Birth of a monetary system. Macmillan
Press.
Varantola, K. (1992). Words, terms, translators. Proceedings of the 5th EURALEX
International Congress, Tampere, Finland.
Vargas Sierra, C. (2005). A pragmatic model of text classification for the compilation
of special-​purpose corpora. In J. Mateo & F. Yus (Eds.), Thistles. A homage to
Brian Hughes. Essays in memoriam vol. II (pp. 295–​315). Universidad de Alicante.
Vàzquez, M. (2024). Creating terminological resources in the digital age for less-​
resourced languages. In N. Calzolari, M.-​Y. Kan, V. Hoste, et al. (Eds.), Proceedings
of the 2024 Joint International Conference on Computational Linguistics,
Language Resources and Evaluation (LREC-​COLING 2024) (pp. 4088–​4091).
ELRA and ICCL. https://​aclan​thol​ogy.org/​2024.lrec-​main.364/​
Vivaldi, J., Cabrera-​Diego, L. A., Sierra et al. (2012). Using Wikipedia to validate
the terminology found in a corpus of basic textbooks. Proceedings of the Eighth
International Conference on Language Resources and Evaluation (LREC), Istanbul,
Turkey.
Wagner, E., Bech, S., & Martínez, J. M. (2012). Translating for the European Union
institutions (2nd ed.). St. Jerome Publishing.
Wakabayashi, J. (1996). Teaching medical translation. Meta, 41(3), 356–​365.
Walker, C. (2023). Translation project management. Routledge.
Walkinshaw, N., Bogdanov, K., & Johnson, K. (2008). Evaluation and comparison of
inferred regular grammars. In A. Clark, F. Coste, & L. Miclet (Eds.), Grammatical
inference: Algorithms and applications. Springer. Retrieved 3 March, 2009, from
www.sprin​gerl​ink.com/​cont​ent/​d3070​l784​2087​j34/​fullt​ext.pdf
282 Bibliography

Warburton, K. (2014). Developing a business case for managing terminology.

Termologic. Retrieved 7 August, 2021, from http://​ter​molo​gic.com/​wp-​cont​ent/​
uplo​ads/​2014/​06/​roi-​arti​cle-​warbur​ton.pdf
Warburton, K. (2015a). Managing terminology in commercial environments. In H.
Kockaert & F. Steurs (Eds.), Handbook of terminology, vol. 1 (pp. 360–​392). John
Benjamins.
Warburton, K. (2015b). Terminology resources in support of global communication
Paper presented at the International Conference on New Horizons in Translation
Technology, Chinese University of Hong Kong, 24th April.
Warburton, K. (2018). Terminology resources in support of global communication.
In S.-​w. Chan (Ed.), The human factor in machine translation (pp. 118–​136).
Routledge.
Warburton, K. (2020). Supporting translators through keyword mining. Proceedings
of Translation in Transition: Human and Machine Intelligence (TT5), Kent State
University, USA.
Warburton, K. (2021). The corporate terminologist. John Benjamins.
Warburton, K., & Wright, S. E. (2019). A data category repository for language
resources. In A. Pareja-​Lora, M. Blume, B. C. Lust et al.(Eds.), Development of
linguistic linked open data resources for collaborative data-​intensive research in the
language sciences (pp. 69–​97). MIT Press.
Ward, K. L. (2020). The best Google search tips for translators. Retrieved 10 May,
2025, from https://​kati​elou​isew​ard.co.uk/​artic​les/​tec​hnol​ogy/​goo​gle-​sea​rch-​tips-​
for-​tran​slat​ors/​
Weisser, M. (2022). What corpora are available? In A. O’Keeffe & M. McCarthy
(Eds.), The Routledge handbook of corpus linguistics (2nd ed., pp. 89–​ 102).
Routledge.
Wikipedia. (2023a). Julius Blum. Retrieved 2 November, 2023, from https://​en.wikipe​
dia.org/​wiki/​Juli​us_​B​lum
Williams, M. (1994). Terminology in Canada. Terminology, 1(1), 195–​201.
Woods, N. (2006). The globalizers: The IMF, the World Bank and their borrowers.
Cornell University Press.
World Bank. (2003). A guide to the World Bank. The International Bank for
Reconstruction and Development.
World Bank. (2011). A guide to the World Bank. World Bank Publications. Retrieved
11 August, 2023, from https://​doc​umen​ts1.worldb​ank.org/​cura​ted/​en/​406​8714​
6813​9488​518/​pdf/​638430PU​B0Ex​to00​Box0​3615​27B0​PUBL​IC0.pdf
World Bank. (2023a). What we do. Retrieved 16 October, 2023, from www.worldb​
ank.org/​en/​what-​we-​do; www.worldb​ank.org/​en/​about/​what-​we-​do.print
World Bank. (2023b). A new era in development. Annual report 2023. Retrieved
27 October, 2023, from www.worldb​ank.org/​en/​about/​ann​ual-​rep​ort#anc​hor-​
ann​ual
World Bank. (2023c). Explore history. Retrieved 11 August, 2023, from www.worldb​
ank.org/​en/​arch​ive/​hist​ory
World Bank. (2024). Organization. Retrieved 23 January, 2024, from www.worldb​
ank.org/​en/​about/​lea​ders​hip
Worne, J., Ulmer-​Leahey, C., Braun, S. et al. (2024). AI: Stop or go? The Linguist,
63(3), 8–​10.
 Bibliography 283

Wright, S. E. (1997a). Term selection: The initial phase of terminology management.

In S. E. Wright & G. Budin (Eds.), Handbook of terminology management. Vol. 1
(pp. 13–​23). John Benjamins.
Wright, S. E. (2006). Standards for the language industry. In P. ten Hacken (Ed.),
Terminology, computing and translation (pp. 19–​39). Gunter Narr.
Wright, S. E. (2022). Terminology and standards. In P. Faber & M.-​C. L’Homme
(Eds.), Theoretical perspectives on terminology: Explaining terms, concepts and
specialized knowledge (pp. 87–​110). John Benjamins.
Wright, S. E., & Budin, G. (Eds.). (1997). Handbook of terminology management.
Vol. 1: Basic aspects of terminology management. John Benjamins.
Wright, S. E., & Wright, L. D. (1997). Terminology management for technical trans­
lation. In S. E. Wright & G. Budin (Eds.), Handbook of terminology management
(Vol. 1, pp. 147–​159). John Benjamins.
Yang, X.-​j. (2006). Survey and prospect of China’s corpus-​based research. In D. Archer
& P. Rayson (Eds.), Corpus linguistics around the world (pp. 219–​233). Brill.
Yao, J.-​Y., Ning, K.-​P., Liu, Z.-​H. et al. (2024). LLM lies: Hallucinations are not
bugs, but features as adversarial examples. arXiv:2310.01469v3 [cs.CL] https://​
arxiv.org/​pdf/​2310.01469
Yuliawati, S., & Indira, D. (2019). Near synonyms and the feature of word sketch
difference of the Sketch Engine: A case study of the English adjectives “brave”
and “courageous.” International Journal of Recent Advances in Multidisciplinary
Research, 6(3), 4684–​4689.
Zanettin, F. (2002a). Corpora in translation practice. Language Resources for
Translation Work and Research (LREC Workshop 8), Las Palmas de Gran
Canaria, Spain.
Zanettin, F. (2002b). DIY corpora: The www and the translator. Proceedings of the
International Conference on “Training the language services provider for the new
millennium,” Porto, Portugal.
Zanettin, F. (2011). Translation and corpus design. SYNAPS, A Journal of Professional
Communication, 26, 14–​23.
Zanettin, F. (2014). Corpora in translation. In J. House (Ed.), Translation: A multi­
disciplinary approach (pp. 178–​199). Palgrave Macmillan.
Zanettin, F. (2015). Concordancing. In S.-​W. Chan (Ed.), The Routledge encyclopedia
of translation technology (pp. 437–​449). Routledge.
Zanettin, F., Bernardini, S., & Stewart, D. (Eds.). (2003). Corpora in translator edu­
cation. Routledge.
Zavorin, I., Bills, A., Corey, C. et al. (2020). Corpora for cross-​language informa­
tion retrieval in six less-​resourced languages. Proceedings of the Workshop on
Cross-​Language Search and Summarization of Text and Speech (CLSSTS 2020),
Marseille, France.
Zetzsche, J. (2019). Freelance translators’ perspectives. In M. O’Hagan (Ed.), The
Routledge handbook of translation and technology (pp. 166–​182). Routledge.
Zhang, L. (2013). Legal translation as a process of constantly creative re-​interpretation: A
tentative investigation into several causes of the dilemmas in translating legal terms.
In G. Fu, L. Cheng, & N. Ye (Eds.), Proceedings of the 2nd international conference
on law, language and discourse: Multiculturalism, multimodaliaty and multidimen-
sionality (pp. 123–​129). The American Scholars Press.
Zhang, W., Yoshida, T., & Tang, X. (2008). A study on multi-​ word extraction
from Chinese documents. In Y. Ishikawa, J. He, G. Xu et al. (Eds.), APWeb 2008
284 Bibliography

workshops, LNCS (Vol. 4977). Springer Verlag. Retrieved 24 October, 2010, from
www.sprin​gerl​ink.com/​cont​ent/​4u44w​3u60​1u53​507/​fullt​ext.pdf
Zhao, W., Su, X., & Huang, W. (2021). The construction of a Chinese and English
term database of Manchu Ulabun. In S. Li & W. Hope (Eds.), Terminology transla-
tion in Chinese contexts. Theory and practice (pp. 214–​228). Routledge.
Zhao, X., & Cao, D. (2013). Legal translation at the United Nations. In A. Borja
Albi & F. Prieto Ramos (Eds.), Legal translation in context: Professional issues and
prospects (pp. 203–​220). Peter Lang.
Appendix

Questions used in the Semi-​structured Interviews

1. How many staff approximately work as translators and/​or terminologists?
2. Do you have designated terminologists? If so, are they actually called
“terminologists” or something else?
3. Are translators called “translators” or something else, e.g. “translators/​
revisers,” etc. Are they encouraged/​required to specialise?
4. Which CAT tools and terminology tools do you use?
5. Terms
a. What do you do with synonyms and near synonyms, e.g. do you store
them in the same entry, or separate entries?
b. Do you have a policy to deal with “cosmetic” synonyms, e.g. checkbox,
check-​box, Check box, etc.? For example, do you make notes about the
spelling in a Notes field?
c. How do you deal with client-​specific terminology?
d. Do you have a central termbase to which all translators contribute, or
do you have a range of separate ones?
e. Do you have a method/​policy for updating terms, i.e. can anybody
go to the termbase and update it, or do you have designated staff to
do this?
f. Same for new terms –​how do you go about adding new terms, e.g. do
translators do this as they translate or is there a designated person to
do this?
g. How do you deal with geographical variants, e.g. US vs UK English?
h. Do you have any processes for the standardisation of terms, i.e. a
method for centralising terminological information and ensuring ter-
minological consistency? How important is this in your company/​
organisation?
6. Workflow: what is your general/​basic translation workflow and how does
terminology fit in?
286 Appendix

7. Do you use MT? If so, which MT systems do you use? Do you use MT
systems from within a CAT tool?
8. Do you have any specific methods to deal with phraseology?
9. How is AI/​ChatGPT impacting on translation and terminology, if at all?
How do you think it will impact them in future?
10. How does the company deal with inclusivity and diversity in terminology?
11. Any other insights re. terminology and translation that you can share
with me?
Index

2-​gram see n-​gram bitexts 129

3-​gram see n-​gram blending see term formation
4-​gram see n-​gram Blum 163–​5, 198, 200–​3, 232; term
record 218; termbase 217–​18;
abbreviations 13, 66, 69, 78, 80 workflow 207–​8; see also Eurocom
acronyms 64, 66, 69, 78, 80 BNC 96
ad hoc terminology 16–​17, 23, 130; Boolean operators 122, 161
see also terminology borrowing see term formation
adjective see part of speech Bourigault, D. 47; see also Condamines,
administrative data categories see data A., Picton, A., Slodzian, M., Textual
categories Terminology
adverb see part of speech British National Corpus see BNC
African languages 54, 61, 116, 197, 241 Bureau Works 156; see also CAT
AI 9, 20, 119–​20, 155, 161–​2; GenAI 155 tool
alignment 88, 119, 128–​9, 136, 162
Allgemeine Terminologielehre see Cabré, M. T. 17, 38, 61; see also
General Theory of Terminology; Communicative Theory of
see also Wüster, E. Terminology, Theory of Doors
alphabetical sorting 9, 12, 63–​4, 105 Canadian example 40, 45–​6, 60
AntConc 130–​1, 135, 147 candidate terms 20–​1, 87, 96–​9, 102,
antonym 55, 74, 86 119, 131, 142–​4; definition 21, 94
antonymy see antonym canonical form see dictionary form
AntPConc 136–​7 capitalisation 77, 95, 102, 104, 125,
applied linguistics see linguistics 224; see also lowercase, uppercase
Arabic 49–​50, 61, 116, 182 CAT 8, 20, 119
article see part of speech CAT tool 10, 20–​1, 66, 90, 94, 100–​1,
artificial intelligence see AI 104–​7, 117, 119, 125, 147, 156, 177,
associative relations 74, 86, 148–​9 214
automatic term extraction see term circular definitions see definitions
extraction ChatGPT 120, 155–​7, 193; see also AI,
automatic terminology lookup 12, 20, LLM
105–​6, 119; see also term recognition Checkterm 199–​200, 203, 214–​16, 232,
235–​6; see also Eurocom, Kalcium,
backup 83, 94, 108, 110 Kaleidoscope Group
base form 104 classical definitions see definitions
Bertels, A. 55; see also Distributional classical terminology 18–​19, 42–​3;
Semantics see also General Theory of
bigram see n-​gram Terminology, Wüster, E.
288 Index

cleaning 98; see also noise, candidate conventions see domain conventions
term corpora see corpus
client preferences 12, 21, 55, 65, 78, corpus 20, 82, 87–96, 119, 161; aligned
171, 176 88; authenticity 91; authoritativeness
cluster see n-​gram 91; bilingual 88; bitexts 129;
co-​hyponym see hyponym comparable 88; definition 88;
collocate see collocation disposable 88; DIY 88; download
collocation 55, 82, 106, 116, 123, 131, 138; focus corpus 96; general
134–​5, 139, 141 language 89; multilingual 88; parallel
commercial translation settings see 88, 128; pre-​processing 94; reference
translation corpus 96; representative 90;
Communicative Theory of Terminology specialised 89, 115; storage 94
41, 43–​4, 47, 52, 61; see also Cabré, corpus linguistics see linguistics
M. T., Theory of Doors corpus-​based approaches to terminology
compounding see term formation 44–​6
compression see term formation cosmetic synonyms see synonyms
computer-​assisted translation see CAT cross-​reference 9, 64, 67–​8, 164, 175, 213
computer-​assisted translation tool see cross links 148
CAT tool CTT see Communicative Theory of
concept 38, 42, 49–​50, 54, 57, 61; Terminology
as anchor point 34; concepts that Cultural Terminology 41, 53–​4; see also
operate in different domains 35; Diki-​Kidiri, M.
definition 25; extension of a concept C-​value 99; see also nc-​value
71; in GTT 42–​3, 53; intension 71;
new concepts 6, 144; parking lot 148; data categories 64–​70, 83;
percept in Cultural Terminology 54; administrative data categories 67–​8;
terminological unit in CTT 44; several conceptual data categories 67–​8; other
concepts designated by one term 35; data categories 68; terminological
see also domain, term, terminological data categories 67–​8; see also field
unit, unit of understanding data elementarity 64, 66, 83
concept level see entry level data granularity 66
concept map 119, 147–​54, 160–​1 data integrity 67
concept orientation 9, 12, 43, 57, 63, de Saussure, F. 61; see also linguistic sign
66, 147, 233, 242 De Vecchi, D. 56; see also
concept visualisation 47, 150–​1; Pragmaterminology
synonym visualisation 154; see also definiendum see definitions
Nuopponen, A., Satellite Model definiens see definitions
concept-​based see concept orientation definitions 26, 38–​9, 65, 67, 70–​2;
conceptual data categories see data circular 72; classical 70; definiendum
categories 70; definiens 70; differentia
conceptual network 17, 26, 33, 61 70; extensional 71; genus 70;
concordance 132, 136; bilingual 136–​7; intensional 71; necessary conditions;
concordance vs term extraction 136 prototypicality 50; sufficient
concordance line 132 conditions; tautological 72
concordance search see termbase, Delavigne, V. 44; see also Gambier, Y.,
translation memory Gaudin, F., Socioterminology
concordancer 45, 119, 130, 134, 147, derivation see term formation
161 descriptive terminology 16–​18, 23;
concordancing see concordance see also terminology
Condamines, A. 47; see also Bourigault, desktop publishing see DTP
D., Slodzian, M., Textual Terminology de-​terminologisation see term formation
conjunction see term formation DGT see Directorate-​General for
content words 98 Translation
contextual information see diastratic term variation see term
encyclopaedic information variation
 Index 289

dictionaries 8–​9, 12, 49–​50, 63–​4 Faber, P. 52–​3, 60, 150; see also Frame-​
dictionary form 104, 133 Based Terminology
differentia see definitions field 64, 83, 176, 190, 217; mandatory
Diki-​Kidiri, M. 53–​4; see also Cultural 64–​5, 69; multiple 64, 69; see also
Terminology data category
directionality see termbase file naming 93, 243
Directorate-​General for Translation 116, file type 123
196, 239 F-​measure see term extraction measures
disjunction see term formation focus corpus see corpus
dispersion plot 135 focus question 148–​9; see also concept
Distributional Semantics 55; see also map
Bertels, A. Fodina 144, 209, 213, 227; see also
DIY corpus see corpus TermCatch
do-​it-​yourself corpus see DIY corpus formatting 102, 173–​4, 203
domain 4, 65–​6; arbitrariness of frame 51–​3, 57; see also Frame-​Based
divisions 26; as a difference between Terminology, metaphor, metaphorical
words and terms 31; as context understanding, Sociocognitive Theory
32; conventions 34, 58, 61, 82, 89; of Terminology
definition 26, 58; importance for Frame-​Based Terminology 41, 52–​3,
translators 32; overlap between 150; see also Faber, P., frame
domains 35; domain understanding free text 68, 224
13, 90, 156; in relation to freelance translation settings see
specialisation 33–​4; one concept in translation
several domains 35; of a corpus 91; freelance translators see translators
phraseology 34; terms from several frequency see term frequency
domains 35; see also concept, term function words 98, 133
domain knowledge 27, 33, 100, 147–​8, fuzzy matching 105–​6, 119, 224
154, 161
DTP 188, 203, 206, 208 Gambier, Y. 44; see also Delavigne, V.,
Gaudin, F., Socioterminology
EcoLexicon 53, 150, 152 Gaudin, F. 44; see also Delavigne, V.,
embeddedness 100 Gambier, Y., Socioterminology
embodiment 38; see also Temmerman, GenAI see AI
R., situatedness general language 25, 27, 29, 61
emotion in terminology 56 General Theory of Terminology 41–​3,
encyclopaedic information 67–​8, 70–​1 61; see also Wüster, E.
English for Specific Purposes 39 generative Artificial Intelligence see
entry level 63, 67, 177, 190, 218, 224; GenAI
see also termbase, termbase design generic relations see hierarchical
equivalence 74–​5; full equivalence relations
75; non-​equivalence 75; partial genus see definitions
equivalence 75 geographical term variation see term
ESP see English for Specific Purposes variation
ethnolinguistics 57 glossary 65–​6, 78, 83, 105, 175, 224,
ethnoterm 57 233, 237; definition 11
ethnoterminological approach 56–​7 governmental translation 18; see also
Eurocom 198–​200, 232, 237; facts and translation
figures 205; workflow 205; see also GTT see General Theory of Terminology
Kalcium, Kaleidoscope Group
Eurocom UK see Eurocom hallucination 157; see also ChatGPT,
Excel see MS Excel LLM
experts 16, 18, 21, 26, 30, 36, 72, 91, 142, hierarchical relations 54–​5, 57, 73–​4,
162, 167, 194, 199, 205, 213, 216, 148–​9; see also associative relations,
232 non-​hierarchical relations, partitive
extensional definition see definitions relations
290 Index

hierarchical structure of a termbase 12, leverageability 78, 100, 102, 110–11

42, 63, 86, 104, 175, 190, 234 lexicography 9, 18, 23, 55, 63, 89, 242;
holonym 73–​4, 148–​9; see also partitive definition 9; overlap with terminology
relations and other disciplines 8
homographs 65, 73 lexicology 55, 61; definition 8; overlap
homography see homographs with terminology and other disciplines
hybrid term extraction see term 8, 23
extraction lexicon 8, 38, 56
hyperonym 33, 42, 70, 73, 148–​9, 176, Lexico-​Semantic Theory 41, 54–​5;
233 see also L’Homme, M.-​C.
hyponym 33, 42, 73, 148–​9; LGP see language for general purposes
co-​hyponym 70–​1 linguistic sign 60–​1; see also de Saussure
linguistic term extraction see term
IATE 11, 86, 107, 175 extraction
IMF 163, 182, 194–​5 linguistics 23, 60–​1, 101, 147;
index level 63, 177, 190, 218; see also applied linguistics 23, 116; corpus
language level linguistics 23, 45, 86, 115–​16, 132;
information technology 23, 197 psycholinguistics 86
informative terminology see descriptive LLM 118, 120, 155–​9, 161–​2
terminology loanwords see term formation
input model see termbase locale see language variant
institutional translation settings see log likelihood ratio 97, 99
translation lone working 16, 19
intensional definitions see definitions lowercase 102, 104, 111, 125, 229, 243;
International Monetary Fund see IMF see also uppercase, capitalisation
International Organization for LSP see language service provider; see
Standardization see ISO language for special purposes
ISO 41, 43, 86, 108, 241–​2
ISO standards 65, 86, 241–​3 machine translation see MT
mandatory field see field
Kalcium 150, 153, 199, 203, 205, 207, manual term extraction see term
209, 211, 214, 235; see also Eurocom extraction
UK, Kaleidoscope Group Matecat 156; see also CAT tool
Kaleidoscope Group 163–​4, 198–​200, memoQ 107, 156, 199, 222–​5, 227,
203–​4, 232, 234; see also Eurocom, 229, 237; see also CAT tool
Kalcium, Kaleidoscope meronym 72–​4, 85, 148–​9, 213; see also
Key Word In Context see KWIC partitive relations
keyness 133–​4 metaphor 43, 51, 80–​1, 85, 194;
keyword list 133 metaphorical processes 57–​8; see also
knowledge pattern 46, 53, 58, 131 metaphorical understanding
KP see knowledge pattern metaphorical understanding 51, 80;
KWIC 126, 131, 135, 138 see also Frame-​Based Theory of
Terminology, Sociocognitive Theory
L’Homme, M.-​C. 38, 54–​5 of Terminology
language for general purposes 29–​30 metasearch engine 127–​8; see also
language for special purposes 29–​31, 39 search engine
language level 63, 67, 176, 190, 218, modifier 55, 101–​2, 112, 195; see also
224; see also index level term
language service provider 157, 162, 198, monolingual term extraction see term
228 extraction
language variant 65, 67, 83, 114, 226 monosemy 73; see also polysemy
large language model see LLM MS Excel 12, 65–​6, 104, 108, 145, 176,
lemma see base form 190, 217, 219, 233–​4, 238, 244
 Index 291

MS Word Dictionary export 108–9 ontology 52, 60, 150

MT 20, 119, 154, 157–​8, 168, 177–​8, orthographic (term) variation see term
189, 193, 205–​7, 220, 223, 226–​7; variation
MÉTÉO 178; see also post-​editing
Muegge, U. 5, 12, 156 paradigmatic relations 55, 68, 70, 72–​4,
multidimensionality of terms 43–​4, 53, 123, 131, 139
57, 61, 148 paraphrasing 28, 129, 138; see also
multimedia information 68, 92–​3 translation strategies
multiple field see field part of speech 54, 58, 63, 65, 67–​70, 73,
multiplier effect 13 85, 96, 101, 139, 177, 224; adjective
MultiTerm 106, 108, 168, 189–​90, 199, 58, 96–​8, 101–​2, 105, 140, 224;
205, 209, 217–​18; see also CAT tool, adverb 54, 58, 101, 224; article 105,
Trados 133; noun 22, 54, 58, 65, 73, 79, 90,
multiword terms 8–​9, 26–​7, 28, 36, 96–​8, 101, 104–​5, 139–​40, 224; noun
55–​6, 77, 81, 89, 95, 97–​8, 100, 105–6, phrase 54, 58, 68, 96–​7, 101, 105,
134, 136; see also single-​word terms 139, 224; verb 54, 58, 73, 79, 89,
mutual information measure 97, 99 101, 104, 123, 134, 139–​41, 214, 224
partial synonym 12, 25, 64, 73, 78, 140,
natural language processing 45, 47, 55, 143–​4, 151, 154, 175, 215–​16
99, 116 partitive relations 73, 148–​9; see also
NC-​value 99; see also C-​value holonym, meronym
near synonym see partial synonym percept 54, 57; see also concept,
necessary conditions see definitions Cultural Terminology, Diki-​Kidiri, M.
neoclassical compound see term Phrase 156, 199, 223; see also CAT tool
formation phraseology 7–​9, 12, 22–​3, 34, 55–​6,
neologism see term formation 58, 70, 81–​3, 89, 128, 177, 192–​3,
neural machine translation see NMT 220, 226, 233, 237, 239; database
n-​gram 134; 2-​gram 134; 3-​gram 134; 192; definition 8; overlap with
4-​gram 134; bigram 134; definition terminology and other disciplines 8;
134; quadgram 134; trigram 134; phraseological information 67, 232
unigram 134; NLP see natural picklist 66, 67–​70, 191, 217–​18, 224
language processing Picton, A. 47; see also Bourigault,
NMT 219–​20; see also MT, post-​editing D., Condamines, A., Slodzian, M.,
noise see term extraction measures Textual Terminology
non-​hierarchical relations 55, 57, PM see project managers
74; see also associative relations, polyhedricity see multidimensionality
hierarchical relations polysemy 6, 52, 54–​5, 58, 72–​3, 76;
Nordic example 40, 51, 60 see also monosemy
Nordterm 51 portmanteau see blending
normative terminology see prescriptive post-​editing 177, 188–​9, 198, 206–​7,
terminology 219, 235; see also MT
notes see termbase Pragmaterminology 56; see also
noun phrase see part of speech De Vecchi, D.
noun see part of speech Prague School 41–​2; other Eastern views
numeric information 68 60; see also Russian School, Vienna
Nuopponen, A. 72, 74, 150; see also School
Satellite Model precision see term extraction measures
prescriptive terminology 11, 16–​18, 23;
Ogden, C.K 38–​9; see also semantic see also terminology
triangle pretranslation 105, 107, 119, 220
on the fly 104, 143, 147 project managers 83, 163, 174, 199,
onomasiological work 16, 19–​20, 24, 203–​7, 209, 213, 216–​17, 219, 222–​3,
42, 44; see also terminology 225, 233, 234, 236–​7, 243
292 Index

prompt 155–7; see also AI, ChatGPT, signified 61; see also de Saussure, F.
LLM, prompt engineering signifier 60–​1; see also de Saussure, F.
prompt engineering 157; see also AI, silence see term extraction measures
ChatGPT, LLM, prompt single-​word terms 77, 81, 89, 96, 136;
prototypicality 50; see also definition see also multiword terms
psycholinguistics see linguistics situatedness 38; see also embodiment,
psychology in terminology 56 Temmerman, R.
punctual terminology see ad hoc Sketch Engine 115, 117, 124, 127,
terminology 136–​42, 144, 147, 161, 189
punctuation 82, 102, 125, 129 SL see source language
Slodzian, M. 47; see also Bourigault, D.,
quadgram see n-​gram Condamines, A., Picton, A, Textual
QuickTerm 150, 199, 209, 212, 232, Terminology
235–​6; see also Eurocom, Kalcium, Smartquery 205, 207, 209, 216–​17,
Kaleidoscope Group 219, 227–​8, 232, 235–​6; see also
Eurocom, Kalcium, Kaleidoscope
ranking 97, 99, 119; score 135, 141; Group
see also candidate term Sociocognitive Theory of Terminology
recall see term extraction measures 41, 47, 49–​52, 60, 150; see also
reference corpus see corpus frame, Faber, P., Frame-​Based
RegEx 124, 224, 238; definition 124 Terminology, Temmerman, R., unit of
regular expressions see RegEx understanding
representation of knowledge 52; see also sociolect see diastratic term variation
concept map Socioterminology 41, 44, 47, 61;
repurposability see reusability see also Delavigne, V., Gambier, Y.,
re-​terminologisation see term formation Gaudin, F., Textual Terminology
reusability see termbase source language 27, 50, 84, 88–​9, 92,
revisers 83, 166, 172, 182, 186, 95, 113, 129, 136–​7, 143, 186, 189,
199, 203, 205–​6, 209, 211, 213, 203, 224, 243
215–​17, 219, 233, 237, 243; see also source text 13, 16, 20, 34, 74–​5, 78,
terminologists, translators 90, 101, 106–​7, 119, 128, 131, 140,
revisors see revisers 142–​3, 157, 173–​4, 177, 216, 220,
Richards, I. 38; see also semantic 224, 232
triangle Soviet School see Russian School
Russian School 41–​2; other Eastern span 132
views 60; see also Prague School, specialisation 12, 14, 30, 34, 55, 187;
Vienna School of domain 240, 246; of translators
RWS MultiTerm see MultiTerm 12–​14, 16, 18, 30–​1, 33; of terms
RWS Trados Studio see Trados 29–​30, 38; of texts 35–​6, 91;
subjectivity 30, 35–​6
Satellite Model see concept visualisation specialised knowledge 5, 25, 27–​8, 32–​4,
Saussure, F. see de Saussure, F. 52, 102, 173, 187, 231–​2, 240, 244
scientific translation see specialised specialised language 25, 28, 29, 76, 78
translation specialised texts 38, 53, 82, 87, 89, 103
scoring see ranking specialised translation 36, 38
search engine 119–​28, 156, 159, 161; specialists see experts
bias 121; definition 120; see also spelling (term) variation see term
metasearch engine variation
seed word 120–​2, 139, 159–​60 standardisation 17, 41–​4, 48, 50, 116,
semantic relations see term relations 226, 241–​3; definition 17
semantic triangle 38–​9 statistical term extraction see term
semasiological work 16, 19, 24; see also extraction
terminology stop list see stopword list
 Index 293

stop word list see stopwords term entry 7, 13, 20, 64, 68, 74, 77,
stop word see stopwords 119, 175, 177; definition 7; multiplier
stopwords 96, 98, 126, 134, 216; effect 13; term record at Blum 218;
stopword list 98, 133 term record at the Translation Bureau
structuralism 55 177; term record at the World Bank
subject field specialists see experts 190; term record at Terra Translations
subject specialists see experts 224–5
subjectivity of specialisation see term extraction 21, 94, 96–​9, 116, 119,
specialisation 136, 142–​4; bilingual 21, 94, 142;
subordinate see hyponym ChatGPT 156; definition 20, 94;
sufficient conditions see definitions hybrid term extraction 96–​7; linguistic
Suonuuti, H. 38–​9 term extraction 96; manual 21;
superordinate see hyperonym monolingual 21, 94, 142; multilingual
synonym 12, 17, 25, 55, 64, 66–​8, 70, 142; statistical term extraction 96–​7;
78, 87, 91, 100, 105, 142–​5, 151, term extraction vs concordance 136;
154, 176–​7, 214, 224, 226; cosmetic see also term extraction measures
synonyms 77, 126, 175, 224; see also term extraction measures 96; F-​measure
partial synonym 96, 98; noise 96, 98; precision 96, 98;
syntagmatic relations 55, 68, 70, 73–​4, recall 96, 98; silence 96, 98
123, 131, 139 term formation 57–​58, 79–​81; blending
systematic terminology 16–​17, 23; 81; borrowing 81; compounding
see also terminology 80; compression 80; conjunction 80;
derivation 79; de-​terminologisation
tagging 97 81; disjunction 80; loanwords
target language 11, 17, 20, 26–​8, 34, 81; neoclassical compounds 80;
75, 78–​9, 82, 84, 88, 90, 92, 95, 107, neologisms 79; primary term
128–​9, 136–​8, 203, 243 formation 79; re-​terminologisation
target text 16–​17, 20, 90, 94, 106, 119, 81; secondary term formation
129–​31, 174 79; terminologisation 81; see also
TBX 66, 108 metaphor
TC37 see Technical Committee 37 term frequency 95, 97, 100, 132;
teamwork 16, 19 normalised frequency 133
Technical Committee 37: ​41, 43, 242; term identification see term extraction
see also Wüster, E. term level 63, 177, 190, 218, 224
technical translation see specialised term mining see text mining
translation term recognition 20–1, 105, 119, 237
Temmerman, R. 18, 25, 38, 49–​51; term record see term entry
see also embodiment, situatedness term relations 67, 72–​3, 86
term 25–​36, 38, 61; as a designation term retrieval 20, 64
for new concepts 6; as a label 8, 25, term storage 20, 63, 87
38–​9; as an anchor point 34; as not term validation 20–​1, 94, 97, 99,
static or monolithic units 5; as triple 119
relations 31; definition 25; in GTT term variation 12, 18, 44, 52, 57, 73–​4,
42–​3; misconceptions 26; terms from 76–​7; cosmetic 77; diachronic 77;
several domains 35; terms vs words diaphasic 77; diastratic 77–​8; diatopic
27–​32, 38; see also concept, domain, 77; dimensional 77; term visibility see
multidimensionality visibility
term autonomy 66 termbase 20, 24, 27, 62; concordance
term bank 74, 82, 86, 107, 110, 175; search 107; definition 11–​12; design
definition 11 62–​84; design principles 66; direction
term candidate see candidate term 84; directionality 67; hierarchical
term creation see term formation organisation 63; input model 104;
term documentation see text mining notes 65–​7, 83, 224; obsolescence 67,
294 Index

77; phraseological information 82; 49–​50; in China 48, 57, 61, 116;
reusability 66, 100; repurposability mismanagement 14; objective 16–​17;
66; reversibility 67, 84; use with CAT overlap with other disciplines 7–​8;
tools 105–​7 prescriptive 16; punctual 16–​17;
TermCatch 144, 150–​1, 154, 213–​14; systematic 16–​17; text-​oriented 16;
see also Fodina thematic 16–​17; time 18; users 7;
termhood 56; see also unithood see also terminological theories
terminography 52; definition 8; overlap TerminOrgs 65, 86, 236
with terminology and other disciplines TERMIUM Plus® 82, 86, 175, 178;
8, 23 see also Translation Bureau
terminological data categories see data TERMIUM® 11, 46, 107, 175, 177,
categories 233; see also Translation Bureau
terminological schools see Prague termontography 52, 150; see also
School, Russian School, Vienna concept visualisation, Temmerman, R.
School terms see term
terminological theories 18, 39, 40–​61; Terra Translations 163–​4, 221–​8, 237;
see also Communicative Theory facts and figures 221; term record
of Terminology, corpus-​based 225; workflow 223
approaches, Cultural Terminology, text alignment see alignment
Distributional Semantics, text mining 20, 87, 93, 116, 119, 161;
ethnoterminological approach, definition 88; protocol 93
Frame-​Based Terminology, General Textual Terminology 41, 47;
Theory of Terminology, Lexico- see also Bourigault, Slodzian, M.,
Semantic Theory, Pragmaterminology, Socioterminology
Sociocognitive Terminology, TF-​DF measure 99
Socioterminology, Textual thematic terminology see systematic
Terminology terminology
terminological unit 44; see also concept, Theory of Doors 44, 61; see also Cabré,
unit of understanding M. T., Communicative Theory of
terminologisation see term formation Terminology
terminologist 16, 38, 94, 172; objective TL see target language
16–​17; terminologists vs translators TMS see terminology management; see
16–​19 translation management system
terminology management 62; as a tokens 133; see also types
challenging task 13; as a labour-​ Trados 156, 174, 177, 190, 199, 209;
intensive task 13; as a smart approach see also CAT tool
10, 58; as a specialist task 10; as a Trados Studio see Trados
time-​consuming task 10, 16; cost 12, trainee translators see translators
15; definition 7, 10, 20; process 20, transfer of specialised knowledge 7, 41
87, 119; time constraints 16, 21 translation 103; freelance 16, 164–​5; in
terminology: added value 13; ad hoc governmental or institutional settings
16–​17; as a discipline 6; as false 18, 163–​4, 196–​7, Chapters 7 and
economy 15; as a methodology 6; 8; in commercial settings 163, 165,
as a polysemous word 6, 23–​4; as a Chapter 9; paraphrasing 28; strategies
practice 6; as a product 6; as a theory 27; see also translators
6, 10; as an activity 6; as invisible Translation Bureau 163, 167–​85, 231–​2;
work 14; as labour-​intensive 58; as facts and figures 168; term record
vocabulary of a specialised domain 177; workflow 173; MÉTÉO 178;
6; benefits 15–​16, 58; definition see also TERMIUM®
4; descriptive 16; differences with translation memory 24, 90, 107, 128–​9
lexicography 8; domain-​oriented translation strategies see translation
16; economic value 5; in Arabic translationese 95
 Index 295

translators 161, 171; as closet WaC see Web as Corpus

terminologists 18; as experts 13–​14, Warburton, K. 43, 56, 64
32; freelance 16, 18–​19, 71–​2; WB-​TERM see World Bank
objective 16; specialised translators Web as Corpus 126
12; trainees 26; translators vs web crawler 119
terminologists 16–​19 Web for Corpus 125; see also Web as
trigram see n-​gram Corpus
T-​score 99 WebCorp 126–​7, 161
TT see target text weighted harmonic mean see F-​measure
types 133; see also tokens WfC see Web for Corpus
wildcards 107, 123
UI see user interface word 104; form 28; in Distributional
unigram see n-​gram Semantics 55; origin 28; scope of
unit of understanding 25; see also use 28; words vs terms see terms
concept, Sociocognitive Theory vs words; see also content words,
of Terminology, Temmerman, R., function words, stopwords
terminological unit word list 133
unithood 55–​6; see also termhood Word Sketch 138–​41; see also Sketch
univocity 25, 42, 49; see also General Engine
Theory of Terminology Word Sketch Difference 138, 140–​1;
UNTERM 11 see also Sketch Engine
uppercase 122, 125, 180, 203, 229, 243; word-​for-​word translation 26, 74–​5
see also capitalisation, lowercase words see word
user interface Wordscope 156; see also CAT tool
WordSmith Tools 135
validation see term validation World Bank 163, 182–​94, 232, 234;
verb see part of speech facts and figures 189; term record
Vienna School 41–​2; see also Prague 190; workflow 188; WB-​TERM 190,
School, Russian School, Wüster, E. 234
visibility 100 World Trade Organization 163,
visual knowledge pattern 53; see also 195
knowledge pattern WTO see World Trade Organization
visualisation see concept visualisation Wüster, E. 23, 42, 61, 70, 82; see also
visualisation tools 119 General Theory of Terminology