Theory and Practice of Specialised Online Dictionaries: Lexicography versus Terminography 9783110349023, 9783110348835

This book is the first comprehensive monograph on the Function Theory of Lexicography, which originated at the Aarhus Sc

204 74 6MB

English Pages 282 Year 2014

Report DMCA / Copyright


Table of contents :
1 Introduction
2 What is Specialised Lexicography?
2.1 Definitions
2.2 Dictionaries versus encyclopaedias
2.3 Typology of specialised online dictionaries
3 Academic Status of Specialised Lexicography
3.1 A discipline with its own subject field
3.2 Relation to linguistics
3.3 Relation to terminology and terminography
3.4 Relation to information science
4 Concept of Lexicographical Theory
4.1 Is lexicography a science?
4.2 A lexicographical theory is possible
4.3 Additional remarks on the concept of lexicographical theory
5 General Theory of Specialised Dictionaries
5.1 Utility tools and user needs
5.2 Concept of user needs
5.3 Data and information
5.4 Lexicographical functions
5.5 General methodology: an example
5.6 Prescription, description or proscription?
5.7 The lexicographical process: the lexicographer’s perspective
5.8 The lexicographical process: the user’s perspective
6 Special Problems Related to Online Dictionaries
6.1 Data filtering
6.2 The type and the individual
6.3 Access to supplementary data
6.4 The use of corpora and the Internet
6.5 Online communication with the user
6.6 Shortening the access route
7 A Critical View of Terminography
7.1 Theories of terminology
7.2 Traditional terminology: The General Theory of Terminology
7.3 Terminology and linguistics
7.3.1 Cabré and the Communicative Theory of Terminology
7.3.2 Temmerman and the Sociocognitive Approach
7.3.3 Martin, Faber and frame-based terminology
7.4 Terminology and knowledge engineering
7.5 Summary
8 An Analysis of Specialised Online Dictionaries
8.1 A functional framework to dictionary criticism: A list of criteria for reviewing specialised online dictionaries
8.2 The Dictionary of Business and Management
8.3 The New Palgrave Dictionary of Economics Online
8.4 The Musikordbogen
8.5 The Glossary of Mortgage and Home Equity Terms
8.6 The Cambridge Business English Dictionary
8.7 TermFinder
8.8 The Business English-Spanish Glossary by A.D. Miles
8.9 The Glossary of FAO Database and Information Systems
8.10 Kicktionary
8.11 IATE, CercaTerm and UNTERM
8.11.1 Inter-Active Terminology for Europe (IATE)
8.11.2 CercaTerm
8.11.3 The United Nations Multilingual Terminology Database (UNTERM)
8.12 The Multilingual Glossary of Technical and Popular Medical Terms in Nine European Languages
8.13 The European Dictionary of Skills and Competences (Disco)
8.14 Genoma and EcoLexicon
8.14.1 Genoma
8.14.2 EcoLexicon
8.15 Summary and The Work Ahead: Working with the Function Theory of Lexicography
9 Designing, Making and Updating Specialised Online Dictionaries
9.1 A Multidisciplinary Activity
9.2 Pre-compilation phase
9.3 Compilation Phase
9.3.1 Lemma Selection
9.3.2 Explanation of meaning
9.3.3 Restricting Meaning
9.3.4 Contextualising Meaning
9.3.5 Offering More Choices
9.4 The Post-Compilation Phase: A Process of Continuous Updating
9.5 Summary and Conclusion
10 Conclusions
Recommend Papers

Theory and Practice of Specialised Online Dictionaries: Lexicography versus Terminography
 9783110349023, 9783110348835

  • 0 0 0
  • Like this paper and download? You can publish your own PDF file online for free in a few minutes! Sign Up
File loading please wait...
Citation preview

Pedro A. Fuertes-Olivera, Sven Tarp Theory and Practice of Specialised Online Dictionaries


Supplementary Volumes to the International Annual for Lexicography Suppléments à la Revue Internationale de Lexicographie Supplementbände zum Internationalen Jahrbuch für Lexikographie Edited by Rufus H. Gouws, Ulrich Heid, Thomas Herbst, Oskar Reichmann, Stefan J. Schierholz, Wolfgang Schweickard and Herbert Ernst Wiegand

Volume 146

Pedro A. Fuertes-Olivera, Sven Tarp

Theory and Practice of Specialised Online Dictionaries Lexicography versus Terminography

ISBN 978-3-11-034883-5 e-ISBN 978-3-11-034902-3 ISSN 0175-9264 Library of Congress Cataloging-in-Publication Data A CIP catalog record for this book has been applied for at the Library of Congress. Bibliographic information published by the Deutsche Nationalbibliothek The Deutsche Nationalbibliothek lists this publication in the Deutsche Nationalbibliografie; detailed bibliographic data are available in the Internet at © 2014 Walter de Gruyter GmbH, Berlin/Boston Druck und Bindung: CPI buch bü GmbH, Birkach ♾ Gedruckt auf säurefreiem Papier Printed in Germany


Thanks are due to the Spanish Ministerio de Economía y Competitividad and to the Junta de Castilla y León for financial support (Grants FF12011-22885 and VA 067A121); also to Aarhus University and the Velux Foundation for financing Dr. FuertesOlivera’s stay at the Centre for Lexicography as Velux Visiting Professor 2011-2012. Our special thanks to Professors Henning Bergenholtz, Theo J. Bothma, Rufus H. Gouws, and Ulrich Heid, to Associate Professors Sandro Nielsen and Patrick Leroyer, and to research student Alejandro García-Aragón for their comments on a draft of this book. Special thanks to Miguel Albillo for his aid for preparing the manuscript. Our special thanks to all the scholars whom we have quoted. All quotations which were not originally written in English have been translated into English by the authors of this book.

Contents 1

Introduction | 1 

2 2.1 2.2 2.3

What is Specialised Lexicography? | 4  Definitions | 4  Dictionaries versus encyclopaedias | 8  Typology of specialised online dictionaries | 12 

3 3.1 3.2 3.3 3.4

Academic Status of Specialised Lexicography | 19  A discipline with its own subject field | 19  Relation to linguistics | 22  Relation to terminology and terminography | 27  Relation to information science | 30 

4 4.1 4.2 4.3

Concept of Lexicographical Theory | 34  Is lexicography a science? | 35  A lexicographical theory is possible | 37  Additional remarks on the concept of lexicographical theory | 40 

5 5.1 5.2 5.3 5.4 5.5 5.6 5.7 5.8

General Theory of Specialised Dictionaries | 44  Utility tools and user needs | 45  Concept of user needs | 48  Data and information | 57  Lexicographical functions | 62  General methodology: an example | 65  Prescription, description or proscription? | 77  The lexicographical process: the lexicographer’s perspective | 84  The lexicographical process: the user’s perspective | 87 

6 6.1 6.2 6.3 6.4 6.5 6.6

Special Problems Related to Online Dictionaries | 91  Data filtering | 92  The type and the individual | 95  Access to supplementary data | 97  The use of corpora and the Internet | 100  Online communication with the user | 101  Shortening the access route | 103 

7 7.1 7.2

A Critical View of Terminography | 104  Theories of terminology | 105  Traditional terminology: The General Theory of Terminology | 106 

VIII | Contents 7.3 7.3.1 7.3.2 7.3.3 7.4 7.5

Terminology and linguistics | 109  Cabré and the Communicative Theory of Terminology | 110  Temmerman and the Sociocognitive Approach | 112  Martin, Faber and frame-based terminology | 116  Terminology and knowledge engineering | 121  Summary | 128 

8 8.1

An Analysis of Specialised Online Dictionaries | 129  A functional framework to dictionary criticism: A list of criteria for reviewing specialised online dictionaries | 130  8.2 The Dictionary of Business and Management | 135  8.3 The New Palgrave Dictionary of Economics Online | 138  8.4 The Musikordbogen | 141  8.5 The Glossary of Mortgage and Home Equity Terms | 143  8.6 The Cambridge Business English Dictionary | 145  8.7 TermFinder | 148  8.8 The Business English-Spanish Glossary by A.D. Miles | 149  8.9 The Glossary of FAO Database and Information Systems | 151  8.10 Kicktionary | 153  8.11 IATE, CercaTerm and UNTERM | 156  8.11.1 Inter-Active Terminology for Europe (IATE) | 157  8.11.2 CercaTerm | 162  8.11.3 The United Nations Multilingual Terminology Database (UNTERM) | 166  8.12 The Multilingual Glossary of Technical and Popular Medical Terms in Nine European Languages | 169  8.13 The European Dictionary of Skills and Competences (Disco) | 172  8.14 Genoma and EcoLexicon | 177  8.14.1 Genoma | 179  8.14.2 EcoLexicon | 185  8.15 Summary and The Work Ahead: Working with the Function Theory of Lexicography | 189  9 9.1 9.2 9.3 9.3.1 9.3.2 9.3.3 9.3.4 9.3.5

Designing, Making and Updating Specialised Online Dictionaries | 192  A Multidisciplinary Activity | 192  Pre-compilation phase | 197  Compilation Phase | 198  Lemma Selection | 200  Explanation of meaning | 205  Restricting Meaning | 209  Contextualising Meaning | 217  Offering More Choices | 223 

Contents | IX

9.4 9.5 10

The Post-Compilation Phase: A Process of Continuous Updating | 230  Summary and Conclusion | 239  Conclusions | 241 

References | 245  Index | 269 

1 Introduction Something is rotten in the Kingdom of Lexicography. For more than two hundred years this millenarian cultural practice has suffered from a schism which is becoming still more absurd in the light of our modern information society. In 1704, one of the pioneers of British specialised lexicography, John Harris, wrote a preface to his English Dictionary of Arts and Sciences. Here he explained that he did not aim only at including words but also things in this work; his users could not simply expect to find explications of the “terms of art” but “also those arts themselves”, i.e. the crafts and sciences as they were developed in the early 18th century. Forty-five years later, the British economist Malachy Postlethwayt (1749) published a small booklet in which he told his readers that he planned to “reduce” the knowledge of trade and commerce to “the form of a dictionary, for alphabetical reference”. His motive was that the people in need of such knowledge could not be expected to obtain it elsewhere, as it was scattered throughout many books. This vision materialised two years later in the first of four editions of his Universal Dictionary of Trade and Commerce, a specialised reference work which was rather influential in Britain and beyond in the mid and late-eighteen century. Five years after the publication of Postlethwayt’s booklet, the great French illuminist Jean le Rond d’Alembert (1754) wrote an article titled dictionnaire in the famous Encyclopédie which he edited together with Denis Diderot. Here he initially distinguished between three categories of dictionary: “dictionaries of words, dictionaries of facts, and dictionaries of things”. His concept of “dictionary of things” was more or less the same as the one used by Harris fifty years earlier; it should not be confused with the concept of encyclopaedia which was explained the following year in a separate article written by Diderot (1755). For the great thinkers who preceded our époque there was no doubt that dictionaries were more than just books of words. They were also practical tools that could be used to describe and promote science and progress. However, only one year after d’Alembert’s article on dictionnaire, Samuel Johnson published his Dictionary of the English Language, a milestone in British lexicography. Nobody can ever take away Johnson’s merits. But it is nevertheless surprising to compare his definition of dictionary with the ideas expressed above: “A book containing the words of any language in alphabetical order, with explanations of their meaning; a lexicon; a vocabulary; a word-book.” Here, Samuel Johnson works with a much narrower concept than other authors from the same century. In his view, a dictionary is apparently only about words. The “things” – i.e. the content of arts, sciences and crafts – emphasized by some of his contemporaneous, seem to be completely ignored by the “father” of British lexicography.

2 | Introduction From then on, a vision of lexicography as being a discipline or craft dealing exclusively with words has taken still deeper roots among one group of lexicographers. Today, this group even consider their profession to be a branch of linguistics or something similar. The other group of dictionary makers dealing also with the “lexicography of things” is ignored or looked upon as intruders or strange fellows. This other group, however, has not been impressed by their colleagues’ narrow vision of the joint discipline. They have continued in the footstep of their great precursors and have produced marvellous specialised dictionaries of various sorts during the past centuries – and also a number of not so marvellous ones. In fact, they have been so busy doing so that Leroyer (2011) has calculated that about three quarters of all dictionaries published in 2008-2009 are specialised dictionaries of one type or another. To this, Kilgarriff (2012) has responded that these dictionaries do not have as many readers as general-language dictionaries and, hence, are not so important. The British lexicographer even compares the two groups of dictionaries with airports, the ones in which he himself is engaged being “international airports” and the others being “local airstrips”. Here the old schism is once more reproduced: on the one hand, we have the lexicographical Heathrows and, on the other hand, the provincial stuff. Cecil Rhodes could not have expressed it more clearly. It cannot be denied that specialised dictionaries are most often not honoured with the same number of users as their general-language counterparts, and that general dictionaries of English, in particular, have a much bigger group of potential users. However, the importance of specialised dictionaries in society should not be ignored either. In many aspects, they play a crucial role in economic and social life, in business communication, education, research, and the dissemination of science and knowledge. It is, therefore, not surprising that the 20th century saw a growing body of academic literature related to this branch of lexicography. Especially in the late 1980s and early 1990s, there was a marked interest in the more theoretical aspects of specialised dictionaries. Among the works published in that period was the Manual of Specialised Lexicography (1995). This manual is up to now the only one of its kind and still frequently quoted by many scholars engaged in specialised lexicography. It is, however, a heterogonous work which in various aspects is outdated. It was based upon the lexicographical function theory – originally developed by researchers at the Aarhus-based Centre for Lexicography – but written in a moment when this theory was still in a process of being elaborated. Since then, collaborators from the Centre have participated in more than hundred finished dictionary projects, both general and specialised ones, most of them online. The experience from all these practical projects has been of enormous importance to the continuous development and refinement of the function theory, adapting it to the world of the new technologies. In recent years, the growing collaboration with the Valladolid-based International Centre for Lexicography has fur-

Introduction | 3

ther enriched the practical experience and contributed to the theoretical development. It is, therefore, logical that two researchers from each of these Centres have decided to join together and share their experience, knowledge and ideas with the rest of the community. The book deals with various aspects of specialised lexicography: the discipline itself and its academic status; the function theory and the various arguments against it; competing theories, especially those of terminography; various aspects of methodology; special problems related to online lexicography; and criticism of a number of existing specialised online dictionaries. Finally, the ideas put forward are exemplified by means of a detailed description of a specialised dictionary project from the planning phase to its publication, as well as recommendations for other types of specialised dictionaries. The book should not be considered “the last word” in this respect. Specialised lexicography – both as a theory and a practice – is under constant development. Many aspects have still to be analysed, new problems and challenges are continuously appearing, and solutions have to be found to all of them, the new as well as the old. With the publication of the book, we hope to have taken a relevant step forward, providing ideas and inspiration to other scholars interested in this field of human activity and knowledge – and to people interested in lexicography in general. By saying this, we also hope to have taken steps to end the senseless schism which has reigned in the discipline far too long. Our vision of lexicography is inclusive. Different approaches and confrontation of ideas should be viewed as normal and healthy phenomena within any discipline aspiring to make further progress. But attempts to exclude, ignore or denigrate an important part of the discipline, with whatever excuse, have no place in such a democratic culture. The Kingdom of Lexicography is fascinating. Its practical products represent a mirror of human life for more than four thousand years. The development of crafts, sciences, arts, philosophies, religions and languages are reflected in these works. Specialised dictionaries play a particularly important role in this regard. They should be treated with respect – and this is what we have tried to do with the publication of this book.

2 What is Specialised Lexicography? A frequent and well-known problem within lexicography is the use of various terms to denote one and the same phenomenon, as well as the use of one and the same term to denote various, completely different phenomena. Such a non-systematic terminology may lead to confusion in the discipline and create obstacles to its theoretical and practical development. The problem does not only present itself within general lexicography, but also in relation to its counterpart, specialised lexicography, and it actually starts with the very distinction between these two main branches of the discipline. Consequently, as the topic of this book is specialised lexicography in general and specialised online dictionaries in particular, it is important to begin by defining what is meant by these two terms. In this chapter we will, therefore, first briefly discuss some of the definitions found in the lexicographical literature; then we will discuss some of the terms used to denote lexicographical products; and finally we will make a preliminary typology of specialized online dictionaries with a view to detecting the state-of-the-art of this branch of lexicography.

2.1 Definitions The first definition to be discussed here will be the one provided by the Spanish scholar José Martínez de Sousa in his Diccionario de lexicografía práctica, where he initially defines diccionario especializado as a “dictionary which registers the vocabulary of a science, technique or art”, cf. Martínez de Sousa (1995: 144). Short definitions are always problematic because they cannot embrace a complex phenomenon in its totality. Martínez de Sousa, however, does not only have the vocabulary of the various disciplines in mind but also these disciplines themselves, as can be clearly seen in his concluding remarks in the article mentioned: Many sciences and techniques that could be expounded in the form of a treatise or a study are reduced to the form of a dictionary, a formula which is generally well received by the users. (Martínez de Sousa 1995: 144)

As we shall see, this understanding of a specialised dictionary stands relatively alone in the few existing dictionaries of lexicography, although it follows practice in part of the theoretical literature. The Nordisk leksikografisk ordbok (Nordic lexicographical dictionary), for instance, offers two complementary terms – fagordbok and sakordbog – which taken together cover the concept of specialised dictionary as defined by Martínez de Sousa. The first of these terms is defined as a “special dictionary treating the terminology of one or several disciplines”, and the second as a

Definitions | 5

“dictionary primarily meant to give information about the world outside the language” (Bergenholtz et al. 1997: 114, 227). Apart from these definitions, also the proposed English equivalents are interesting. In this connexion, the authors offer the English terms special field dictionary and technical dictionary as translational equivalents of fagordbok, i.e. unlike the terminology employed by Bergenholtz & Tarp (1994a, 1995) who, two years earlier, had translated this last term as specialised dictionary in the very subtitle of the English version of their Manual of Specialised Lexicography. In addition, the authors of the Nordisk leksikografisk ordbok propose dictionary of things and lexicon as equivalents of sakordbok, but at the same time, and quite unexpectedly, they then translate the hypernymous term sakleksikografi as the English specialised lexicography, thus adding to the general confusion. However, if this terminological inconsistency is ignored for a moment, the basic idea, which the Nordisk leksikografisk ordbok seems to transmit, is that a distinction should be made between dictionaries treating the specialised language (fagordbok) and the ones dealing with the speciality or subject field itself (sakordbog). This distinction follows a century-long tradition – most frequently expressed in the opposition linguistic dictionary versus encyclopaedic dictionary – and pervades a large part of the existing literature on lexicography, especially that influenced by linguistic theories, e.g. Kalverkämper (1990a) and Wiegand (1988, 1994). It is even reflected in Bergenholtz and Tarp’s Manual of Specialised Lexicography (although the two editors have subsequently rejected this idea), as well as in Hartmann and James’ Dictionary of Lexicography, the hitherto only English dictionary of this kind and, therefore, rather influential in the English-speaking lexicographical community. In this last publication, specialised dictionary is defined both extensionally and intensionally: The collective term for a range of reference works devoted to a relatively restricted set of phenomena. In contrast to the general dictionary which is aimed at covering the whole vocabulary for the ‘general’ user, specialised (or ‘segmental’) dictionaries concentrate either on more restricted information, such as idioms or names, or on the language of a particular subject field, such as the jargon of the drug scene or the technical terms of mechanical engineering. (Hartmann & James 1998: 129)

As can be seen, specialised dictionary is here, on the one hand, defined in contrast to general dictionary, but with a “line of demarcation” different from the one established by the two former dictionaries of lexicography, as Hartmann and James also include dictionaries of jargon and idioms, among others, in the concept of a specialised dictionary. Although this definition contradicts part of the lexicographical literature, it is nevertheless not surprising because it follows practice in another part of this literature, especially that related to the Anglo-Saxon tradition. The latter is, for instance, also expressed in Cowie (2009), who, in the second volume of his Oxford History of English Lexicography, dedicated to specialised dictionaries, incorpo-

6 | What is Specialised Lexicography? rates various contributions on dictionaries of synonyms, pronunciation, etymology, slang, etc., but only one dealing with scientific and technical dictionaries (Hoare 2009). On the other hand, the above definition suggests that a specialised dictionary is, in the view of Hartmann and James, exclusively devoted to the linguistic or terminological side of the disciplines or areas covered by the respective lexicographical works. In this regard, it is far from clear how the two authors distinguish a specialised dictionary from a LSP dictionary, defined as “a type of reference work intended to describe a variety of a language used by experts in a particular subject field”, a technical dictionary, defined as “a type of reference work devoted to the description of the technical language of a specialised subject field”, and a terminological dictionary, which they define as “a type of reference work which provides information about the language (especially the vocabulary) of a specialist field as defined by its practitioners” (Hartmann & James 1998: 90, 137, 140). Additionally, these authors also provide other lemmata, such as special dictionary and special-purpose dictionary, from which they refer the users to the lemma specialised dictionary without providing any independent definition. Against this background, the German scholar Kurt Opitz hits the nail on the head when, in an article about technical dictionaries in the International Encyclopedia of Lexicography, he points out the ambiguity of the terms applied to denote lexicographical works: Certain types of dictionary, which are conceived to allow for particular situations of application and to alert users to the fact, are variously referred to as technical or specialized. Neither attribute is specific as to content or user. (Opitz 1990: 1505)

Although Opitz himself is interested only in the linguistic aspect, he is fully aware of the fact that the “word dictionary”: can occasionally assume the meaning ‘encyclopedia’ which ordinarily denotes a work providing full factual information on a broad subject, but not necessarily on resulting questions relating to language use. (Opitz 1990: 1505)

By comparison, in Hartmann and James’ Dictionary of Lexicography only one lemma denoting a type of lexicographical work has been found where the definition refers directly to “factual information” about a subject field; this lemma is encyclopedia, defined by the two authors as “a type of reference work which presents factual information in a wide range of subject disciplines” (Hartmann & James 1998: 48). Moreover, encyclopedic dictionary is defined as a “hybrid” or “compromise” type of dictionary with an internal “tension” between the various data categories: A type of reference work which shares features of the general dictionary and the encyclopedia. There is a tension between linguistic information […] and encyclopedic information […], which explains the hybrid or compromise status of this dictionary type. (Hartmann & James 1998: 49)

Definitions | 7

The terminological inconsistency and ambiguity resulting from a comparison of the lexicographical works consulted above could easily be amplified if the contextual definitions scattered throughout the rest of theoretical literature were included in the discussion. The inevitable conclusion is that there is no generally accepted definition of specialised dictionary among lexicographers, although most of them would probably agree with the following brief and almost tautological definition of specialised lexicography: A complex of activities concerned with the design, compilation, use and evaluation of specialised dictionaries. (Hartmann & James 1998: 129)

Hence, the “apple of discord” seems to be how to define specialised dictionary. In this regard, specialised dictionary should not be considered a very specific type of dictionary with rigid and well-established characteristics, but a term denoting a generic concept covering a series of lexicographical works of different characteristics, sizes and names. This also seems to be the opinion of Hartmann & James, who, in the article quoted above, add the following comment to their brief definition of specialised dictionary: There is no uniform framework for this as the nature and scope of such reference works can range widely, from a brief glossary without definitions, through technical dictionaries aimed at lay persons, to large-scale and standardised terminological databases for subject experts and translators. (Hartmann & James 1998: 129)

We shall later argue why terminological database is a highly problematic term when it is used to denote a lexicographical work (cf. section 5.4). At this point, however, it is appropriate to draw attention to two ideas reflected in the hitherto discussion, as both of them may be useful when formulating a definition of specialised lexicography and specialised dictionary for the purpose of this book (and the purpose of lexicography as a scientific discipline). The first of these ideas is the use of the contraposition general versus specialised when defining this branch of lexicography and its practical products; and the second is that these practical products may have a number of different names, sizes, and characteristics despite belonging to the same overall category of lexicographical reference works. With this in mind, specialised lexicography is here defined as the branch of lexicography concerned with the theory and practice of specialised dictionaries, i.e. dictionaries, encyclopaedias, lexica, glossaries, vocabularies, and other information tools covering areas outside general cultural knowledge and the corresponding Language for General Purposes (LGP); it represents mainly, but not exclusively, disciplines related to technology, industry, trade, economic life, law, natural and social sciences, and humanities. Although still frequently referred to as LSP lexicography (see, e.g. Nielsen 2014), specialised lexicography by far transcends a mere description of the various specialised languages and also treats the very substance of

8 | What is Specialised Lexicography? these disciplines themselves in order to provide direct, punctual access to their cognitive achievements. This definition requires a number of comments. First of all, it is relevant to note that the opposition between general cultural and specialised knowledge – and the corresponding opposition between LGP and LSP – should not be regarded as absolute, since the transition from general to specialised knowledge is necessarily a gradual transition with no sharp boundaries, as already argued by Kalverkämper (1989, 1990b). This suggests that a certain grey area may exist between general and specialised lexicography, a phenomenon that does not exclude the existence of these two main branches of lexicography, although it may explain some of the problems concerning their delimitation and definition. The second observation is that specialised language (LSP) is not viewed independently in its opposition to LGP but only as a function of the underlying transition from general cultural to specialised knowledge. In this respect, it is also worth noting that in the above definition no distinction is made between specialised dictionaries dealing with LSP and those concerned with “the world outside the language”, i.e. between linguistic and encyclopaedic dictionaries, as it has been done, among others, by Bergenholtz et al. (1997). As a result, the concept of specialised dictionary includes a wide range of lexicographical works of different names, sizes and characteristics, thereby eradicating the traditional and rigid distinction between dictionaries and encyclopaedias.

2.2 Dictionaries versus encyclopaedias Most people have an intuitive idea of what is meant by dictionaries and encyclopaedias. Even so, the history of lexicography has been witness to a protracted discussion on how these two kinds of work should be defined. Some scholars, like Henriksen (1992), consider “encyclopaedology” to be a discipline different from lexicography but subordinated to a joint reference science (“referensology”), whereas others have a more balanced perception of their internal relation and possible differences. The discussion goes back at least to the reflections by d’Alembert (1754) and Diderot (1755) in their respective articles on the subject in their Encyclopédie. In this regard, it is worth bearing in mind that this monumental work also included the word Dictionnaire in its title. The same did a number of other lexicographical works like Chamber’s Cyclopædia (1728) and the first edition of the Encyclopædia Britannica (1768-1781), both of which were subtitled dictionary of arts and sciences. Whitelaw (1846) even used the three terms dictionary, lexicon and encyclopaedia in the very impressive title of his work The popular encyclopedia: or, ‘Conversations lexicon’: being a general dictionary of arts, sciences, literature, biography, history, ethics, and political economy.

Dictionaries versus encyclopaedias | 9

Many similar examples can be found in several European languages. Daniele Besomi, a Swiss expert in the history of economics, for instance, has studied more than 650 encyclopaedias and dictionaries of economics from the past three centuries and uses words like motley and fuzzy to describe the situation within this branch of lexicography: Lexicographical works concerning economics (exclusively or not) are a motley set of objects. As we shall see below…, they include a wide range of sizes, purposes, languages, editorial histories and intended audiences. And they have different names: we have dictionaries, encyclopaedias, encyclopaedic dictionaries, lexica, vocabularies and glossaries, which correspond to different kinds of reference works. The distinction between these kinds of works in practice is rather fuzzy. Several of them – both of general scope and specifically addressed to economics or the social sciences – actually carry more than one of these denominations in their title. (Besomi 2011: 4)

In lexicographical literature there is a deep-rooted tradition according to which dictionaries describe or define the words of a language, whereas encyclopaedias (or encyclopaedic dictionaries) provide factual information on various or all branches of knowledge. Some scholars defend this distinction claiming that “dictionaries are about words, encyclopedias are about things” (Landau 2001: 6), whereas others oppose the idea with the argument that “dictionaries are encyclopedias” (Haiman 1980: 331). However, even scholars accepting the distinction recognise that there is an unavoidable grey area between the two extremes. Landau himself “admits of many exceptions if we try to apply it to every item in each type of work”, as does Ladislav Zgusta in his well-known Manual of Lexicography: It must also be remembered that the division of dictionaries into encyclopedic and linguistic ones is not necessarily an either-or-matter. We shall have occasion to show that there are elements of encyclopedic character in almost all dictionaries. Some of these encyclopedic elements are unavoidable, some are introduced because the compiler of the dictionary wishes to give his work a certain character. (Zgusta 1971: 199)

Most authors, like Zgusta and Landau, who have tried to distinguish between the two dictionary types, base their ideas more on intuition than on convincing scientific arguments. The only known attempt to provide a more precise and scientifically acceptable definition of the concepts of linguistic and encyclopaedic dictionaries has been that of the German scholar H.E. Wiegand, who strongly criticises the methods used by some of his colleagues: It is not necessary from the theoretical standpoint of lexicography, does not correspond to pretheoretical experience in the handling of dictionaries, and is thus subject to criticism of method if the differentiation of language lexicography and encyclopaedic lexicography and thus a clear delineation of language and encyclopaedic dictionaries is made exclusively dependent on some one of the linguistic, philosophical or psychological concepts or even eclectic, composite concepts from these research areas regarding the relationship between so-called linguistic and encyclopaedic knowledge. Neither language dictionaries nor encyclopaedic dictionaries are

10 | What is Specialised Lexicography? made to implement any of the various concepts regarding the relationship of the two types of knowledge, which in the meantime are rather numerous and quite different in detail. This means that an empirically adequate, metalexicographical delineation of language lexicography and encyclepaedic lexicography can be made without recourse to such concepts. (Wiegand 1988: 739)

Wiegand’s criticism is convincing but the method which he himself proposes is rather disappointing. His basic idea is that dictionaries are utility tools produced with the genuine purpose of meeting human needs. However, instead of taking these needs as the point of departure, he starts with a phenomenological analysis of the objects of the dictionaries assigning these objects to two main classes, C1 and C2, where “C1 is the class of linguistic objects and C2 the class of non-linguistic objects” (Wiegand 1988: 745). Using these two classes as a basis, he then divides the genuine purpose of dictionaries into a language-lexicographical purpose and an encyclopaedic-lexicographical purpose: From the preceding argument regarding C1 and C2, it is apparent that one can also assign to two classes the characteristics of a reference work to have at least one genuine purpose, these being C1, the class of true language-lexicographical purposes, and C2, the class of true encyclopaedic-lexicographical purposes. There is a true language-lexicographical purpose if the intention of the lexicographer was that a potential user should derive information from the lexicographical text data regarding a linguistic object from C1 and if this is actually possible. Correspondingly, there is a true encyclopaedic-lexicographical purpose if the intention of the lexicographer was that a potential user should derive information from the lexicographical text data regarding a subject from C2 and if this is actually possible. (Wiegand 1988: 745f)

Accordingly, Wiegand formulates his basic dictionary typology, comprising three types of lexicographical reference work: 1) language dictionaries (with genuine linguistic-lexicographical purpose), 2) encyclopedic dictionaries (with genuine encyclopaedic-lexicographical purpose), and 3) the so-called all-round dictionaries, which mix elements of the other two types – a typology which in a specialised lexicographical perspective translates into specialised language dictionaries (fachliche Sprachwörterbücher), specialised encyclopaedic dictionaries (fachliche Sachwörterbücher), and specialised all-round dictionaries (fachliche Alwörterbücher), respectively, cf. Wiegand (1994). Wiegand’s attempt to classify dictionaries in general, and specialised dictionaries in particular, according to a distinction between linguistic and encyclopaedic elements has been criticised by various scholars including Rossenbeck (1994) and Bergenholtz (1998). The main criticism is that the distinction is problematic or even impossible (the point also made by Haiman 1980). The discussion seems endless, mainly because the point of departure is basically irrelevant. Although Wiegand rejects “the linguistic, philosophical or psychological concepts” as a basis for a lexicographical typology and introduces the intention of the respective lexicographers, i.e. the genuine purpose of the lexicographical work, he himself turns to lin-

Dictionaries versus encyclopaedias | 11

guistic-philosophical criteria in his attempt to distinguish between linguistic and encyclopaedic knowledge, and the reason for this is that his real starting point is a phenomenological analysis of the content of existing dictionaries instead of a much more relevant determination of the users’ specific needs. (For a detailed discussion of Wiegand’s method, see Bergenholtz & Tarp 2004.) Users of lexicographical works consult them with certain needs which arise in specific social situations and contexts, and their main interest is, therefore, that these needs are satisfied by means of the information they can retrieve from the pertinent publications. From this point of view, it is completely irrelevant whether the information is retrieved from data classified as either linguistic or encyclopaedic according to some linguistic or philosophical criteria; the most important thing is that the necessary data are available (as to the use of the terms data and information, see section 5.3). A much more relevant classification is, therefore, the one based upon the functions of the works in terms of the user needs to be covered, i.e. needs determined both by the type of social situation or context where they occur and by the relevant characteristics of the type of user in question. This is the fundamental approach defended in this book, and it seems to be more or less similar to the one which Diderot used in his article on Encyclopédie, where he initially defined this kind of work based upon its purpose: In truth, the aim of an encyclopedia is to collect all the knowledge scattered over the face of the earth, to present its general outlines and structure to the men with whom we live, and to transmit this to those who will come after us, so that the work of past centuries may be useful to the following centuries, that our children, by becoming more educated, may at the same time become more virtuous and happier, and that we may not die without having deserved well of the human race. (Diderot 1755: 635)

We have already seen how various terms were frequently used in the title of one and the same lexicographical work during the past centuries. Today, in fact, dictionary, encyclopaedia, encyclopaedic dictionary, lexicon and others terms have lost much of their original meaning and are, above all, names given by publishers in order to sell their products, cf. Tarp (2012d). Although there is a clear tendency for the largest lexicographical works, especially the ones designed to transmit factual knowledge, to be called encyclopaedias, and the smallest works, especially the ones conceived to solve communication problems, to be referred to as dictionaries, there is also quite a few other publications in which the opposite picture is visible. In fact, this is also the conclusion Besomi (2011) reaches when he speaks about a continuum with easily discernible extremes, after studying more than 650 specialised dictionaries and encyclopaedias: In practice, the spectrum of the reference works concerning economics and related disciplines […] forms a continuum [...]. Yet its extremes are easily discernible. On one side, we have the writings sharing the features of specialized encyclopaedias, carrying more or less detailed entries not only defining the concept under discussion but also reporting the economic under-

12 | What is Specialised Lexicography? standing concerning it, supplying essential (occasionally rich) bibliographies and offering a systematization of the subject within the corpus of economic knowledge. On the other side, there are the writings limiting themselves to a definition of the word under examination, or very little more. The latter naturally offer short entries for each heading, so that the former kind of works result in much larger numbers of pages; the length of these volumes […] offers a rough proxy for their position in the spectrum. (Besomi 2011: 5)

Specialised lexicographical works may be published under a whole range of different names such as dictionary, encyclopaedia, encyclopaedic dictionary, lexicon, vocabulary, glossary, terminological database, knowledge bank, resource, tool, etc. However, at the end of the day it is not the names but the functions of the respective products that determine whether or not they can be considered specialised lexicographical works. In this book, all these products will be viewed under the umbrella term specialised lexicographical works or simply specialised dictionaries, including traditional encyclopaedias or encyclopaedia-like works such as the International encyclopaedia of social sciences (2008) and the online New Palgrave Dictionary of Economics (2012). Of course, this inclusion of a wide range of lexicographical products under the same umbrella does not mean that the differences between them should be ignored. It is evident that the production of dictionaries conceived to solve communication problems, dictionaries designed to provide quick information about a specific subject, and dictionaries (or encyclopaedias) permitting a profound study of one or more subject fields, present different challenges and require different practical solutions. Dictionaries or encyclopaedias with particularly long articles – sometimes dozens or even hundreds of printed pages – require, for instance, special lexicographical techniques in order to provide easy access to specific data incorporated in and dispersed throughout these articles. However, the different publications mentioned are nonetheless united by the essential fact that they are all highly specialised reference works designed for consultation purposes. From the point of view of specialised lexicography, and as Besomi (2011) rightly observed, they form a continuum, and it is, therefore, counterproductive to establish an artificial and non-existent Chinese wall between them. In this book, we will argue that the differences in terms of size and other characteristics are eventually determined by the different needs which they intent to meet and that these needs, in turn, are determined both by the characteristics of their potential users and the social situations with which they are intimately related.

2.3 Typology of specialised online dictionaries Specialised electronic lexicography, or e-lexicography, is the branch of lexicography concerned with the theory and practice of specialised e-dictionaries in general. Nowadays the Internet seems to have consolidated itself as the dominant electronic

Typology of specialised online dictionaries | 13

platform for dictionaries and other lexicographical tools, and (although it is difficult to make predictions in the present scientific and technological rush) there are no visible signs that this situation will change in the near future. Therefore, despite being fully aware of the fact that a number of specialised dictionaries are still being published on other electronic platforms such as CD-ROMs and various handheld devices, in this book we will focus exclusively on online dictionaries. When a deep-rooted millenarian cultural practice like lexicography passes from one medium to another, one would expect such a gigantic step to be more than a mere change of platform, and that it would also involve improvements in terms of quality. In order to get an idea of the extent to which this expectation has become a reality, in this section we will present a preliminary typology of online dictionaries in accordance with the technology employed. For this purpose we will make use of a historical analogy. About a hundred years ago, a similar great technological transformation took place in relation to the medium for the transportation and carriage of passengers and goods. According to a famous anecdote, the inventor of the Model T Ford, Henry Ford, was asked, when introducing the new technological wonder, if he had consulted the people before creating this model. His laconic answer was: If we had asked people what they wanted, they would have said faster horses.

Henry Ford, apart from being a tough capitalist, was also an audacious inventor due to his ability to go beyond the usual boundaries and satisfy people’s needs in a completely new way. If we maintain his terminology and way of thinking and apply it to lexicography, this will allow us to classify existing – and even future – online dictionaries into five main categories according to how far they incorporate the available – and as yet unavailable – technologies and techniques. These five categories are: 1. Copycats, 2. Faster Horses, 3. Stray Bullets 4. Model T Fords, 5. Rolls Royces This terminology permits, as we shall see, a useful distinction between genuine edictionaries and electronically-adapted paper dictionaries with a view to detecting problems and promoting new and more advanced lexicographical solutions. Copycats are photographed or scanned copies of already existing printed dictionaries which have been placed on the Internet, most frequently as PDF files. There are two subtypes of copycat: old dictionaries which are now merely used for research purposes and no longer as consultation tools, and modern dictionaries which are supposed to be used as reference works.

14 | What is Specialised Lexicography? As to the first type, i.e. old specialised dictionaries, today a growing number of such dictionaries are made freely available on,, and other Internet sites. Among these hundreds or thousands of historical works are John Harris’ Lexicon Technicum (1704), Savary des Bruslons’ Dictionnaire universel de commerce (1723), Postlethwayt’s Universal Dictionary of Trade and Commerce (1774), Ganilh’s Dictionnaire analytique d’économie politique (1826), Johnson and Emerson’s The American farmer’s encyclopedia and dictionary of rural affairs (1844), and Boccardo’s Dizionario della economia politica e del commercio (1857). This means of making old, historical dictionaries available constitutes a gigantic step forward, especially for researchers who only a few years ago had to arm themselves with patience, travel far and cope with all sorts of difficulties in order to get access to the required research material. In this regard, the electronic copying of already printed dictionaries and their publication on the Internet is fully legitimate. The same can hardly be said about the electronic copycats of modern dictionaries. It may not be surprising that this sort of copying was frequent in the infancy of Web-based lexicography, i.e. back in the twenties century, where even large general dictionaries such as the Diccionario de la Lengua Española (Dictionary of the Spanish Language), edited by the Royal Spanish Academy, were placed on the Internet as PDF files. Although today it might be expected that this way of providing online access to modern dictionaries would be gradually disappearing, the fact is that there is a continuous flow of new publications of this kind, especially dictionaries placed on the Internet by individuals with few financial and technological resources. Similarly, a number of public and international institutions, including ones in a strong financial situation, also place primitive online PDF versions of their lexicographical works on the Internet; such examples are the Illustrated glossary on stone deterioration patterns, published by the International Council of Monuments and Sites (ICOMOS), and the Illustrated Glossary for Transport Statistics, published in collaboration between the Eurostat, the International Transport Forum (ITF) and the United Nations Economic Commission for Europa (UNECE). The only real advantage of this class of specialised online dictionaries is that they become available – and most often freely – to all users with Internet access; however, in terms of data accessibility and user-friendliness in general, there are essentially no advantages at all, because they are basically still printed dictionaries and as a rule even more difficult to access. The second type of these specialised online dictionaries is represented by Faster Horses, which may be either electronic versions of existing paper dictionaries or completely “new” ones. Their fundamental characteristic is that the articles visualized on the screen are static and modelled on the corresponding articles in printed dictionaries; the only thing which is really different is that some of the new technologies and techniques available have been used, albeit in a very restricted way, in order to provide quicker data access by means of more or less advanced search and link techniques. In this respect, they are basically Faster Horses, i.e. conventional

Typology of specialised online dictionaries | 15

dictionaries allowing faster consultation. Examples of this category of dictionaries include the current online versions of A Glossary of Archival and Records Terminology, Kemisk Ordbog and Bilteknisk Ordbog, all originally printed, later digitalized and then made available on the Internet, the two latter for payment on, Denmark’s biggest provider of Web-based dictionaries. This way of “producing” online dictionaries represents a relatively easy and cheap solution that may to a certain extent be justified in the case of already printed dictionaries, especially those with a relatively small group of users. However, an increasing number of the dictionaries belonging to the category of Faster Horses are not electronic versions of former printed ones, but have been made from scratch according to traditional models and concepts adopted uncritically from the era of printed lexicography. One emblematic example of this kind is the Inter-Active Terminology for Europe, probably the world’s most expensive dictionary project ever (cf. section 8.11.1). Other much smaller and less ambitious examples are Wells Fargo’s Glossary of Mortgages and Home Equity Terms and Kodak’s Glossary of Photographic Terms. Most of the specialised online dictionaries discussed by Caruso (2011) also belong to this category. The result of this somewhat unimaginative and restricted use of the new technological possibilities is lexicographical products where the data included are still structured in static articles which over and over again, consultation after consultation, are displayed in the same way without being adapted to the specific needs of the user in each consultation. The large majority of specialised online dictionaries are either lexicographical Copycats or Faster Horses, an indication of the fact that specialised lexicography has still a long way to go until it takes full advantage of the new technologies and techniques available. However, there are a small – but growing – number of online projects which have started to make use of such technologies and techniques. Roughly speaking, these projects are moving in two very different directions: the right way and the wrong way. The third type of specialised online dictionaries, the Stray Bullets, is among the latter. These dictionaries are characterised by new visions and the incorporation of new techniques, but “everything is going in the wrong direction”, as Mick Jagger and Keith Richard wrote in their song Connection. There are basically two kinds of Stray Bullets. The first one has made use of new techniques to a limited degree, but only with a view to offering users the option between more or less data, thus taking a purely quantitative step “forward”. These dictionaries are relatively “harmless drudges”, to use the famous expression by Samuel Johnson (1755), but in no way do they clear the way for the great prospects regarding online lexicography which lie ahead. In contrast, the second kind of Stray Bullets has made much more extensive use of the new techniques, but in this case technology has taken over and relegated lexicography and its fundamental objective of guaranteeing maximum user need satisfaction to a second level. A significant example of this sort of online dictionary

16 | What is Specialised Lexicography? is the Spanish-English-Greek EcoLexicon, where lexicography seems to have been replaced by lexicotainment (cf. section 8.14.2). It is important to bear in mind that the new technologies should not be used just to be used at the risk of creating fancy lexicographical products that may entertain their users without providing better solutions to their real problems and needs. The application of the new technologies should above all contribute to an improved quality, which in the case of specialised lexicography – apart from coverage of the relevant areas and disciplines – can be translated into quicker, more accurate and personalised satisfaction of the corresponding user needs. In this respect, it is necessary – as happened in the case of Henry Ford – to leave old habits behind and make full use of the available technology in order to invent new advanced solutions to old problems, in other words, to satisfy people’s lexicographical needs in a completely new way. These new solutions are represented by the lexicographical Model T Fords and Rolls Royces, respectively. Both of them have used the available technology with a view not only to providing quicker access, but also to creating dynamic articles with dynamic data, that is, articles that vary from consultation to consultation in terms of their lexicographical content. In the case of the Model T Fords, the articles and the visualised lexicographical data are adapted to the various functions displayed by the dictionary, frequently assisted by different types of interactive options where the users may define themselves and the activity for which they need information. These lexicographical tools exploit, at least to a certain extent, available information science techniques like user profiling, filtering, and adaptive hypermedia (cf. chapter 6), whilst they also provide a frequent link to the Internet, where already existing data is reused in order to satisfy the users’ specific needs. Today, only very few specialised online dictionaries merit the title of Model T Ford, among them the Diccionarios de Contabilidad, produced in cooperation between researchers from the International Centre for Lexicography at the University of Valladolid, Spain, and the Centre for Lexicography at Aarhus University, Denmark. Going a step further, the Rolls Royces are individualized lexicographical tools which permit the satisfaction of the individual information needs which a concrete and individual person may have in a concrete and individual situation. In fact, the term Rolls Royce may not sufficiently reflect the really new and revolutionary characteristics of these highly sophisticated tools, which perhaps should be compared with driverless vehicles or any other means of transport that, equipped with all the comfort adapted to the demands of each individual, could transport the individual to any place in the world without them having to do anything other than inform the system of the final destination. These advanced lexicographical tools may also combine access to selected data in a prepared database with search on the Internet (and in existing corpora) in order to take in relevant data that can be recreated and rerepresented in dynamic solutions different from the ones provided by the lexico-

Typology of specialised online dictionaries | 17

graphical Model T Fords, which link to specific and selected Web pages in order to reuse their data. Although some interesting experiments going beyond the principles of the Model T Ford may already take place, no specialised dictionary has yet been found that qualifies to the category of lexicographical Rolls Royce; this, therefore, continues to be an empty, but nevertheless highly relevant category, pointing as it does to the future of lexicography and to the new horizons gradually opening up thanks to the new computer and information technologies and techniques. In this regard, it is worth noting that while the available information science technologies already seem to offer users individualized lexicographical solutions such as the one described above, it is, however, still necessary to develop certain new technologies before the recreation and re-presentation of the data automatically acquired from the Internet can become a lexicographical reality. To sum up, present-day specialised dictionaries may be assigned to three main groups: 1) printed paper dictionaries, 2) paper dictionaries or paper-like dictionaries placed on the Internet or another electronic platform, and 3) genuine e-dictionaries designed from scratch to work on one of these platforms. The fact that the lexicographical Copycats and Faster Horses are just conventional paper dictionaries or paper-like dictionaries located on the Internet leads inevitably to the conclusion that the only authentic specialised online dictionaries are the Model T Fords and the Rolls Royces, whereas the Stray Bullets represent an inevitable and, hopefully, temporary aberration on the long road towards the future. Both the Model T Fords and the Rolls Royces have left behind the static articles inherited from printed lexicography and provide dynamic solutions designed either to meet the types of need which a specific type of user may have in a specific type of situation or activity (Model T Fords) or the individual needs of an individual user in each consultation (Rolls Royces). The challenge facing visionary specialised lexicography today is to take the step from the lexicographical Model T Ford to the Rolls Royce, i.e. the step towards individualisation of needs satisfaction. However, in this book – apart from studying the experience of existing printed and electronic dictionaries – we will discuss not only the problems relating to the conception and production of Rolls Royces but also those regarding Model T Fords, as both classes of dictionaries mark a big step forward in comparison to the present state-of-the-art of specialised online lexicography. The design and production of these two classes of dictionaries clearly require an advanced theory capable of guiding the lexicographers and other specialists, e.g. subject-field experts, programmers, Web designers, etc., who, in one way or another, intervene in their creation. At the end of the day, it is not merely a question of giving users the option to choose between longer or shorter articles, that it, more or less data displayed on the screen. The problem is not only quantitative but mainly qualitative, and it is, therefore, necessary to determine not only the amount (a quan-

18 | What is Specialised Lexicography? titative criterion) but also the types (a qualitative criterion) of data required. This cannot be done without a theory providing detailed guidelines that permit a meticulous study and classification of users’ needs and the corresponding amount and types of lexicographical data needed to satisfy them in each and every situation where they may occur. Such a theory, the function theory, will be sketched out and discussed in the following chapters.

3 Academic Status of Specialised Lexicography In this book, lexicography in general and specialised lexicography in particular is understood as both a millenarian cultural practice and an independent academic discipline with its own system of scientific theories. This point of view is polemic. On the one hand, the independent nature of lexicography is questioned by many linguists engaged in lexicographical practice, and, on the other, the very existence of specialised lexicography is rejected by a number of terminographers involved in the production of terminological reference works. In this chapter we will discuss some of the arguments put forward to defend the various positions and, upon this basis, argue in favour of the discipline’s independent status and the real relations which exist – or ought to exist – between general lexicography and linguistics, between specialised lexicography and terminology, as well as between lexicography in general and any other discipline, especially information science.

3.1 A discipline with its own subject field One of the most beautiful and intriguing characteristics of lexicographical works is that they, during the millennia, have covered almost all spheres of human activity and knowledge. If we concentrate on the European experience, during the past few centuries lexicographical works have constituted themselves into a mirror of the rapid development of science, law, economy, industry, culture, and other social phenomena. Mikkelsen (1994), for instance, have traced the development of Danish trade, navigation, science and society in general since the sixteenth century by means of the various kinds of specialised subject-field dictionary published in the period. In this sense, lexicography is essentially characterised by its big interdisciplinary vocation. There is hardly any academic discipline or branch of human knowledge that has not left its traces in lexicographical works. It is, therefore, logical that dictionaries have played an important role in the diffusion and popularisation of scientific knowledge: Specialized and general dictionaries played an important role in the dissemination of scientific knowledge at a time when the opportunities for acquiring a knowledge of science through a formal education were limited. Whilst the extension of the popular science movement to the artisan and working classes had scarcely begun in the eighteenth century, the existence of scientific dictionaries helped to make possible the self-education of many scientists of humble origins, and directed the attention of men to a field of activity which was often neglected in an age when formal education was still predominantly classical. (Layton 1965: 222)

During certain periods, lexicographical works were even the preferred means of communication for vanguard researchers of various disciplines. This is, for instance, the case within economics where “authorship of articles is normally expert

20 | Academic Status of Specialised Lexicography and often authoritative”, and where some works even include “Nobel laureates among their contributors” (Besomi 2011: 16). And the Swiss researcher adds: These reference works played an important role in the popularization but also in the systematization of knowledge during the nineteenth century and the early twentieth century, and – judging from the continuing exponential increase in their publication – are still widely used in the support of teaching and, to a lesser extent, research. What is recorded in dictionaries is therefore rather influential, in particular for those works recognized by contemporaries to be authoritative. (Besomi 2011: 4)

In the same vein, and after analysing two well-known dictionaries from different centuries – namely, the Lexicon Technicum (1704) and the McGraw-Hill Dictionary of Scientific and Technical Terms (1974) – the American researcher Olga Menagarishvili concludes that such specialised dictionaries “participate in the scientific knowledge economy”, cf. Menagarishvili (2012: iii). Almost two hundred years ago, the French economist Charles Ganilh expressed a similar idea in the Preface to his Dictionnaire analytique d’économie politique: Dictionaries are the best means to disseminate sciences, speed up their progress, and make them quickly move to the highest stage they can reach. The greatest perfection of human thought is in its proliferation. (Ganilh 1826: xxvii)

The American farmer’s encyclopedia and dictionary of rural affairs (1844) is another example of a lexicographical work that was meant to propagate the newest results of science, in this case in order to educate farmers with a view to improving the techniques used in American agriculture and, thereby, raising its productivity. According to its subtitle, this dictionary was specifically “adapted to the comprehension of unscientific readers”; in their Introduction, the American Editors even described the foreseen user type as a “book-farmer”, i.e. a farmer “searching into books containing modern information”: The success with which science has developed the agencies concerned in the various stages and processes of vegetation, and the certainty with which deficiencies of soil can now be detected and remedied, have suddenly elevated agriculture from the condition of an art under the guidance of common observation and empirical experiment, to a science regulated by recognised principles of induction. We are indeed much mistaken if the day has not arrived when the successes of the book-farmer shall cause his incredulous brother farmer of the old routine system, to cease his taunts and spend some of his leisure hours in searching into books containing modern information in regard to matters of husbandry. (Johnson & Emerson 1844: 5)

Similar examples can be found in thousands of dictionaries and other lexicographical works. Two very different and opposed conclusions could be drawn from this fact. The first is that lexicography should simply be considered a means of propagating the ideas and practical results of various disciplines and sciences, and that the corresponding works are but sub-products of these disciplines and sciences. Such a

A discipline with its own subject field | 21

view is basically schizophrenic because it implies that lexicography is, simultaneously, a sub-discipline of hundreds or even thousands of different disciplines and sciences. This conclusion contradicts the very fact that we can speak of dictionaries, encyclopaedias, and lexicographical works in general. If all these works with their very different content can be grouped under the same umbrella concept of lexicographical products, then there must be something that unites them and justifies their classification as such. In fact, this “something – i.e. the aspects and elements common to all these multifaceted works – may be considered the very essence or core of lexicography. It is the “something” that distinguishes it and makes it different from every other discipline; the “something” that is left when an abstraction is made from all the specific content incorporated from a huge number of different disciplines according to the topic of the respective works. In chapter 5, we will discuss in detail what should be considered the core of lexicography. At this point, the above reflections lead immediately to a second conclusion which is defended here, namely, that lexicography should be viewed as an independent discipline with a great interdisciplinary vocation, i.e. independent but not isolated from other disciplines. The interdisciplinary vocation has already been illustrated above, whereas the independent status is mainly, but not exclusively, based on the fact that lexicography as a discipline has its own subject field which is distinct from the subject fields of all other disciplines. This subject field comprises dictionaries and other lexicographical reference works whatever they are called and covers all aspects related to the production, design, usage, criticism, and history of these works. This understanding of lexicography is disputed, especially in the Anglo-Saxon lexicographical community. Nevertheless, it is supported by a considerable number of scholars with their roots in traditions hailing from outside this community, among whom are to be found Wiegand (1989, 2013), Svensén (2009), and the supporters of the function theory, cf. Tarp (2008, 2010). According to the Russian scholar Fedor Petrovich Sorokoletov, this also seems to be the general view in his country, at least in the recent past: In the Soviet period lexicography developed into an independent discipline with its own theory, own tasks and own methods for their solution. (Sorokoletov 1982: 79)

Sorokoletov’s understanding of lexicography may differ considerably from the one presented in this book, as he considers this independent discipline to be a “special branch of linguistic science”. However, his reference to lexicography’s “own theory, own tasks and own methods” is nevertheless highly relevant because it points to the special and complex relation that exists between lexicography and other disciplines with which it interacts. Although it is frequently ignored, this necessary interaction does not imply that lexicography can automatically and uncritically take over the

22 | Academic Status of Specialised Lexicography arsenal of concepts, theories, and methods used by these other disciplines. The fact that the subject field is delimited, and the fact that lexicography has its own independent core distinguishing it from other disciplines, mean that all these concepts, theories and methods must be subjected to critical analysis with a view to determining what should be rejected, what can be used as it is, and what can be used only after being adjusted and adapted to suit the particular nature of lexicography. This is a fundamental principle of methodology valid for both general and specialised lexicography.

3.2 Relation to linguistics The ideas expressed in the previous section are, as already mentioned, polemic; as such, they are opposed with a wide range of different arguments, especially those of a number of linguists engaged in the production of various types of general dictionary. It is by no means strange that the lexicographical function theory originally was developed at the Aarhus-based Centre for Lexicography, which, in conformity with its statutes, from the very outset was almost exclusively devoted to the study and production of specialised dictionaries. Through their practical and theoretical work within this branch of lexicography, the researchers at the Centre soon became aware that linguistically inspired reflections on lexicography largely contradicted the practice observed, and were of little or no assistance when solutions had to be found to the complex problems related to the production of a new generation of high-quality specialised dictionaries. Although the cognition was gradual and took several years to be completed (cf. Tarp 2008: 33f), a fundamental idea which crystallised little by little was that lexicography in general – and not only specialised lexicography – should be redefined as an independent discipline with its own subject field, theories, methods, and tasks in society – a discipline which in order to be successful and play its social role had to cultivate a fruitful relation to other relevant disciplines. With this vision the supporters of the function theory started to work in two directions: on the one hand they engaged in a number of dictionary projects in order to apply and test their ideas in practice, and on the other they became very active within the lexicographical community, producing a large number of publications and papers for different conferences where they argued in favour of their theoretical and practical discoveries. Very soon they experienced an unexpected and contradictory reaction. When they started to cooperate with experts from various other disciplines in the conception and production of a number of specialised dictionaries, they immediately observed that these experts, who were very conscious of their own subject-field knowledge, were at the same time very well aware that – despite possessing some lexicographical experience – they were not experts in lexicography; they therefore readily agreed to engage in creative interdisciplinary cooperation with people who knew relatively little about their respective disciplines but a lot

Relation to linguistics | 23

about lexicography. In this way, it was easy to apply and test the function theory with a view to improving it and taking new steps towards creating high-quality lexicographical products. However, the reaction was quite different when the ideas developed by the supporters of the function theory were presented to the established lexicographical community, which, even today, is mainly composed by people with a background in linguistics. The majority simply preferred to do business as usual, ignoring or rejecting the new ideas presented by the “intruders”. Some even felt offended and interpreted these ideas as an attack on linguistics, despite the fact that the main purpose of the outstretched hand was to establish new and even more fruitful relations between the two disciplines. However, following a deep-rooted tradition within general lexicography, where linguistics obviously has contributed to the compilation of quite a number of fine dictionaries, almost all these scholars considered – and many still consider – lexicography to be a sort of sub-discipline of their own scientific mother discipline. This idea is very aptly expressed by Hans H. Meier, in an article titled Lexicography as applied linguistics: Dictionary making, though often regarded as ‘a special technique rather than a branch of linguistics’, may be considered an instance of applied linguistics. (Meier 2003: 307)

It is obvious that the discussion on the status of lexicography is not a purely academic matter; rather, it has serious theoretical and practical implications. The rejection of the independent character of lexicography and its classification as a linguistic sub-discipline immediately brings to the forefront the question of theory: – Is there a theory of lexicography? (or can it be developed?) – Should the available linguistic theories instead be applied to lexicography? – Should the existence and relevance of theory be completely rejected within this branch of human activity, which some scholars – like Landau 2001 – regard simply as “art and craft”? In the lexicographical community there are supporters of all three positions, although there are no well-defined borders between them. We have already seen how scholars such as Sorokoletov consider lexicography to be an independent discipline within linguistics and, thus, advocate the need to formulate independent theories. Similar, though not identical, views are expressed by authors who, like Reinhard Hartmann (2009), refer to a so-called meta-lexicography, admitting the existence and need for lexicographical theories. A similar position is held by the Polish researcher Tadeusz Piotrowski, who does not reject a “new theory on lexicography”, but only on the condition that it is strongly embedded in modern linguistics: A new theory of lexicography would be one that would take seriously what we know about pragmatics and discourse, text structure, and would account for the contribution of particular

24 | Academic Status of Specialised Lexicography textual elements to the dynamic meaning of a text, in short, the dynamics of meaning construction both in the receptive and the productive mode. (Piotrowski 2009: 485)

Other scholars, however, advocate the opinion that the only theories which should be used in lexicography are those developed within the framework of linguistics. An almost emblematic example of this position is the classical publication of Haensch et al. (1982), which, translated into English, is simple entitled Lexicography: From Theoretical Linguistics to Practical Lexicography. A similar position is taken by Sue Atkins and Michael Rundell, who “do not believe that such a thing exists”, i.e. a theory of lexicography, although they do not deny that “there is an enormous body of linguistic theory which has the potential to help lexicographers to do their jobs more effectively and with greater confidence” (Atkins & Rundell 2008: 4). This position has been reaffirmed by Rundell (2012). In the same vein, Ten Hacken (2009: 399) argues that linguistics, and especially Chomskyan linguistics, should be “taken as a background” in order to consider lexicography “a scientific activity”. Finally, there is a relatively small group of authors who do not believe in any theoretical guidance to practical dictionary-making claiming that “lexicography has no theoretical foundation, and even the best lexicographers, when pressed, can never explain what they are doing, or why” (Wierzbiecka 1985: 5). In a certain sense, Henri Bejoint (2010) sums up the position of those scholars who do not even believe that a theory of lexicography can exist: I simply do not believe that there exists a theory of lexicography, and I very much doubt that there can be one. Those who have proposed a general theory have not been found convincing by the community, and for good reasons. A theory is a system of ideas put forward to explain phenomena that are not otherwise explainable. A science has a theory, a craft does not. All natural phenomena need a theory, but how can there be a theory of the production of artefacts? There are theories of language, there may be theories of lexicology, but there is no theory of lexicography. Lexicography is about all a craft, the craft of preparing dictionaries, as well as an art, as Landau (2001) says. It may be becoming more scientific, but it has not become a science. (Bejoint 2010: 381)

We will return to this statement in chapter 4 and define the lexicographical concept of theory as it is used in the function theory. It is important that everybody has the possibility to present and defend their point of view; this is part of the democratic culture within academic disciplines, but the consequences of the various positions should not be silenced either. It is, therefore, appropriate to stress that the classification of lexicography, in one form or another, as a sub-discipline of linguistics constitutes a flagrant contradiction of lexicographical practice, should specialised dictionaries and encyclopaedias like the ones mentioned in the previous sections be also taken into account. An early bird among these specialised dictionaries is Malachy Postlethwayt’s Universal Dictionary of Trade and Commerce, published in various editions in the mid-eighteenth century. The author was a British economist and publicist of certain

Relation to linguistics | 25

renown during this époque; the dictionary was originally a translation and adaptation of a similar French dictionary, Jacques Savary des Bruslons’ Dictionnaire universel de commerce, published first in 1723 and subsequently in various editions and translations. The Universal Dictionary of Trade and Commerce appeared in instalments in 1751-1754 and later in three other editions (1757, 1766, and 1774) with the incorporation of a considerable amount of new material. As such, it was rather influential in Britain and beyond in the mid and late eighteenth century, although it, quite surprisingly, is not even mentioned in Cowie’s Oxford History of English Lexicography from 2009. Postlethwayt’s dictionary is interesting for various reasons, one of them being that the author two years before its first appearance published a 52-pages essay where he explained his motives for this excursion into the world of lexicography (Postlethwayt 1749). Here the future “lexicographer” writes that the objective of the planned dictionary is to “more particularly accommodate the Fame to the Trade and Navigation of the British Empire”; this is done in the light of a serious problem he has observed, namely, that the relevant people frequently do not have a “satisfactory knowledge of Facts in complicated matters of a commercial nature”, and that, in addition, these people have neither time nor the possibility to obtain this knowledge because it is scattered throughout an infinity of volumes (Postlethwayt 1749: 2). He then writes: Foreign and domestic trade admitting of so infinite variety of matter, and the knowledge communicated to the world, by those skilled and experienced therein, being scattered in an infinity of volumes, it is no easy matter to have immediate recourse to what may be occasionally requisite… A subject of this extensive nature therefore being reduced to the form of a Dictionary, for alphabetical reference, seems the most naturally adapted to answer these desirable purposes, and especially so, as the compilers can have no motive to deceive. (Postlethwayt 1749: 2)

As an example of the practicability of this idea, he refers explicitly to Savary des Bruslons’ Dictionnaire universel de commerce, a “celebrated work” which has proven “how far an universal knowledge of commerce is capable of being reduced into the like form” (Postlethwayt 1749: 3). It was these ideas that he himself applied shortly afterwards in his own dictionary. If we bear all this in mind, it seems nothing short of impossible to meaningfully characterise Postlethwayt’s Universal Dictionary of Trade and Commerce and Savary des Bruslons’ Dictionnaire universel de commerce as practical results of applied linguistics. Both authors were subject-field experts within trade, commerce and economics, and no specialised knowledge of linguistic theory was required to produce these famous dictionaries, although both authors most certainly dominated the general as well as specialised economic language in their respective mother tongues. Hence, when it is frequently claimed that “dictionary making […] may be considered an instance of applied linguistics” (Meier 2003: 307), then a whole range of specialised dictionaries are in fact disregarded as such and, de facto, excluded

26 | Academic Status of Specialised Lexicography from the world of lexicography with all the negative consequences that this may have. As a result it means that future authors of specialised dictionaries have nowhere to go when they look for guidance to their work, or at least that they cannot with any benefit seek assistance in the established lexicographical community. A variant of this is the view of the American linguist William Frawley, who, well aware that many specialised dictionaries have been authored by experts from their respective subject fields, goes so far as to advocate that these experts should desist from dictionary making and content themselves with being mere “informants”; specialised lexicography, according to him, should “remain the province of lexicographers and linguists since they are the ones who, in the long run, know how lexicons operate and how to represent meaning” (Frawley 1988: 196). It is interesting to see how an almost identical criticism is made by a terminologist who is against dictionaries “compiled by specialists who know their own subject field well, but […] are neither lexicographers nor Terminologists” and who try to follow “the lexicographical format […] without knowing how to do it properly” (Riggs 1989: 90). Linguistics is until now the only external discipline that has tried to claim property of lexicography. This classical example of not being able to see the wood for trees is probably due to the fact that linguists have participated in a large number of dictionary projects of various types. Some of these projects have resulted in specialised dictionaries where linguistics has been the subject treated; for instance Theodor H. Lewandowski ‘s Linguistisches Wörterbuch, David Crystal’s Dictionary of Linguistics and Phonetics, and Peter H. Matthews’ Concise Oxford Dictionary of Linguistics. Most of these projects have, however, given rise to general dictionaries conceived with a view to documenting a language (e.g. the Oxford English Dictionary), tracing the origins of the words (etymological dictionaries), assisting text reception, production and translation (communicative dictionaries), or supporting the acquisition of a first or second language (school and learner’s dictionaries), of which the two latter classes constitute the big majority. Some of these dictionaries are excellent lexicographical works while others are more problematic, yet all of them can undoubtedly be improved. However, the rejection of lexicography as an independent entity, and the resulting rejection of the need to develop independent lexicographical theories and methods distinct from those used within linguistics, may seriously hamper the conception of even better general dictionaries; this is probably one of the main obstacles to a smoother transition from lexicographical Faster Horses to Model T Fords in the online environment. To sum up, lexicography and linguistics have traditionally interacted in the compilation of a large number of dictionaries. There are, nevertheless, a considerable number of other dictionaries in which knowledge of linguistic theory has been neither used nor required. This fact alone leaves no room for doubt that lexicography cannot reasonably be considered an instance of applied linguistics, and that theories and methods developed within linguistics cannot automatically be used to guide lexicography, whether general or specialised. The inevitable conclusion is

Relation to terminology and terminography | 27

that lexicography must develop its own system of theories and methods while it maintains, at the same time, its traditionally fruitful cooperation with linguistics or any other relevant discipline.

3.3 Relation to terminology and terminography Terminology can be traced back at least to Whewell (1847), although it only developed into a scientific discipline as a result of the work initiated by the Austrian engineer Eugen Wüster in the 1930s. Consequently, terminology is a relatively young discipline in comparison to lexicography, which brought forth thousands of specialised dictionaries before the first reference works based upon terminological principles came into being. Today terminology – or what is presented as “terminology” – is divided into a number of schools, tendencies and national variants, as already observed by Humbley (1997) and Kudashev (2007). As in lexicography, there is a continuous discussion of its basic principles, cf. Roche (2012). In chapter 7 and 8, we will have a closer look at some of these theoretical expressions and their practical outcomes in terms of reference works. Here we will limit ourselves to a brief discussion of the general relationship which exists – or is said to exist – between terminology and lexicography. In 1975, the ISO 1087 standard introduced the term terminography as a replacement of the term terminological lexicography, thus nurturing the differentiation between lexicography and terminography, as reflected in publications by Rey (1975, 1988) and Rondeau (1981), among others. The new term was obviously “coined on the analogy of lexicology:lexicography :: terminology:terminography”, cf. Hartmann & James (2001: 139). According to Igor Burkhanov, a Polish author of a dictionary devoted to the basic terminology of lexicography, this apparently innocent step seems to have had an ulterior motive: The primary aim of this term is to emphasize that lexicography should deal with the description of the general vocabulary only, whereas specialized terminology should be accounted for by another discipline. (Burkhanov 1998: 240)

The same idea is expressed in other words by the Spanish scholar María Chantal Pérez Hernández: When distinguishing between terminology and lexicography, most authors use to draw a parallel, on the one hand, between lexicology, a discipline concerned with the study and description of the lexicon of a language, and lexicography, understood as the applied branch of lexicology concerned with the compilation of dictionaries, and on the other hand, between terminology, an area of theoretical and methodological study, and terminography, the applied branch of terminology concerned with the compilation of specialised dictionaries. Hence, lexicology is to terminology what lexicography is to terminography. (Pérez Hernandez 2002: 3.3)

28 | Academic Status of Specialised Lexicography Consequently, an unholy alliance is established between terminologists and certain linguists engaged in lexicographical work. The false idea that lexicography is a branch of applied linguistics (lexicology) is used as a ram to breach new walls and claim that a similar relationship exists between terminology and terminography, thereby completely ignoring the branch of linguistics devoted to LSP research as well as rejecting the relevance, and sometimes even the existence, of specialised lexicography both as a theory and a practice. This strange approach to a century-old lexicographical reality has led to a relatively extensive literature on the relations between (specialised) lexicography and terminography and their possible similarities and differences, with contributions by scholars from both “camps” – see, for instance, Dubois (1979), Picht (1985), Rey (1988), Riggs (1989), Svensén (1992), Bergenholtz (1995), Humbley (1997), Bergenholtz & Kaufmann (1997), Toft (1998), Pérez Hernández (2002), Wiegand (2004), Kudashev (2007), Bevilacqua & Bocorny Finatto (2006), Bergenholtz & Nielsen (2006), and Bergenholtz & Tarp (2010). All the publications mentioned include the words lexicography and terminography (or terminology) in their titles but they only represent the top of the iceberg of the corresponding literature. Various journals like Hermes 18 (1997) and Terminology 12:2 (2006) have also dedicated special issues to this question; moreover, it should be added that search on the Internet produces a considerable number of references to web pages, books and articles where lexicography and terminography (or terminology) are contrasted. Scholars from the field of terminology in particular seem to have been very eager to point out the differences more than the similarities. In fact, Bergenholtz & Kaufmann (1997) note, not without some frustration: You may get the impression that there is an almost frantic search for differences, perhaps in the effort to present terminography as something real unique. (Bergenholtz & Kaufmann 1997: 93)

The two authors – one of them a lexicographer and the other a molecular biologist – present a number of “arguments put forward by terminologists/terminographers” in order to distinguish between the two traditions. In the same vein, Bergenholtz & Nielsen (2006) focus on five of the most important arguments, whereas Bergenholtz & Tarp (2010) list the following thirteen “statements”: – Lexicography deals with the description of general-language words, whereas terminography concentrates on the description of LSP terms. – As opposed to lexicographers, who work with an alphabetic macrostructure, terminologists prefer a systematic macrostructure. – Terminology is prescriptive, whereas lexicography is descriptive. – The target group of terminology is the expert, whereas in lexicography it is the layman. – While terminologists aim to help users encode texts, lexicographers aim to help users decode them.

Relation to terminology and terminography | 29

– –

– – – –

– –

In terminography they only use academic experts as informants, whereas lexicography can use any native speaker. Terminographers use computers and data bases, their results are mainly presented as electronical dictionaries also called terminological databases. However, lexicographers only present their results as printed dictionaries. Terminographers prepare tools for text production, lexicographers for text reception and translation. Terminographers only work with synchronic methods, lexicographers with diachronic. In terminography there is no polysemy, but sometimes more than one term to one concept. In lexicography, on the other hand, you find polysemy. In terminography they have a field systematic as starting point for the whole terminographical work, in lexicography they deal with each word unsystematically. In terminography concept relations are described, whereas lexicographers use linguistic methods. In terminology they work with concepts and terms and not with language signs. (Bergenholtz & Tarp 2010: 28)

All these statements, made with a view to opposing terminography to lexicography, have been discussed and dismissed in several publications by scholars doing research into specialised lexicography, for instance Bergenholtz & Tarp (1995: 10-11), Bergenholtz (1995), Bergenholtz & Kaufmann (1997), Bergenholtz (1998), Bergenholtz & Nielsen (2006), Kudashev (2007), Tarp (2008: 4-13), and Bergenholtz & Tarp (2010). Such statements may represent some tendencies which are more or less explicit in one or the other type of work, but they cannot be considered absolute and used as proofs for the classification of terminography and lexicography as two different disciplines. Having said this, it is nonetheless obvious that some differences in terms of underlying theory, methods applied and final presentation can be observed between, on the one hand, a group of terminographical products and, on the other, another group of specialised lexicographical ones. However, such differences also exist internally among reference works termed lexicographical (in the narrow sense of the word), as well as among those referred to as terminographical, as illustrated by the many schools, tendencies and national variants of terminology mentioned above. These differences may partially be explained by the distinct user needs which the respective works aim to meet (and in this case they are completely normal and justified). However, – as we shall see in chapter 8 – the dissimilarities may also involve theories, methods and forms of presentation, which prove to be more or less acceptable or problematic not only when the quality of the final product has to be evaluated in terms of its user-friendliness (its capacity to satisfy the foreseen user’s needs

30 | Academic Status of Specialised Lexicography in all aspects), but also when the production process itself is assessed in relation to the resources and time required to produce the respective works. From this perspective the various specific theories, methods and forms of presentation promoted by both lexicographers and terminographers should be subjected to scientific criticism and quality analysis. This might lead to certain ones being dismissed as inadequate, not because they belong to one or another school or tradition, but rather because they are not deemed capable of guiding the production of high-quality reference works within the shortest possible time and by means of a minimum of human and material resources. The reason why these theories, methods and forms of presentation can be subjected to the same criteria is that they are all developed to produce different types of reference work, regardless of their name. The artificial separation of these theories, methods and forms of presentation into lexicographical and terminographical (terminological) has no relevance or scientific basis when their final products are considered, from the point of view of the foreseen users, to be utility products designed to satisfy their different information needs. They all fall into the broad category of specialised dictionaries or lexicographical works, as they are defined by the function theory. Consequently, terminography – that is, the production of reference works based upon terminological principles – is here understood as an integrated part of specialised lexicography, whereas terminology – like LSP research – is taken as a discipline with which specialised lexicography interacts in the conception and production of a number of specific reference works.

3.4 Relation to information science Information science came into the world as an independent discipline in the first half of the twentieth century and has undergone an enormous upsurge during the last decades due, among other things, to the rapid development of the corresponding technologies. By comparison, lexicography, as a four thousand year-old cultural practice, has inevitably brought forth a large accompanying literature of academic reflections, but it is not until the twentieth century that the discipline has been subject to a systematic theory building, that is, more or less during the same period in which information science was founded and flourished. Until recently, the two disciplines have developed along parallel lines albeit with no or very few formal relation between them. However, with the growing focus on information in society and the incorporation of the new technologies into lexicographical practice, it has become increasingly clear that lexicography has a lot in common with information science, especially the part of it concerned with information organisation and retrieval. As a result, it is appropriate to emphasize that when an abstraction is made from the concrete and specific content of the needs lexicographical works have

Relation to information science | 31

attempted to meet during the millennia, what is left is their common nature of information needs in the true sense of the word. Essentially, the users of these works always look for and, if they are successful, retrieve information which they can subsequently use for countless purposes in accordance with their specific needs in each case. Consequently, lexicography and information science share a common field of interest, albeit at different levels. Specific sub-disciplines within information science – for instance, information organisation and retrieval – have until now mainly been dedicated to studying and developing systems, technologies and techniques that may provide access to already existing and relatively large documents (books, articles, etc.), as well as illustrations and other types of data sources, from which information can be retrieved. In this regard, information science operates almost exclusively with cognitive needs, as they are defined in the function theory (cf. section 5.4). Lexicography, on the other hand, also studies and develops systems and techniques that allow the users to access relevant data and retrieve the information required. However, lexicography differs from information science in at least three ways, cf. Bothma & Tarp (2012: 104): – the information needs covered by lexicographical works are not only cognitive but also communicative, operative and interpretive (cf. section 5.4); – the cognitive needs covered are in most cases (but not always) needs that may be met by relatively small sets of data; – the data to which access is provided are, as a rule, not already existing data but data selected, elaborated and prepared by the lexicographer. In spite of these obvious differences, history provides various overlapping examples where lexicographical works have been planned and produced according to principles similar to those of information science. The New Palgrave Dictionary of Economics Online (2012) is such an example of a lexicographical work which provides access to a huge number of scientific articles written by distinguished economists based upon principles that completely overlap with those of information science. Another even more impressive example is the Chinese Yongle Dadian (1408) which was produced in order to collect and provide access to all knowledge existing in China at that moment. Among its 11,095 volumes, this gigantic lexicographical work included both completely new data written by the authors and already existing books which were incorporated, i.e. rewritten, in their totality into the work and made accessible through a rhyming system for the characters as well as a complex system of indexes – thus creating an information tool very similar to the ones developed by modern information science. In view of these similarities, it is hardly surprising that during the last few years a number of researchers from both fields have started discussing various aspects of the relation between the two disciplines and have reached the conclusion that lexicography, although an independent discipline, should be reconsidered and viewed

32 | Academic Status of Specialised Lexicography as a part of information science, cf. Bergenholtz & Bothma (2011), Bothma (2011), Bothma & Tarp (2012), Fuertes-Olivera & Bergenholtz (2011b), Gouws (2011), Heid (2011), Leroyer (2011), Verlinde, Leroyer & Binon (2010), and Tarp (2007a, 2009a, 2010, 2011, 2012a). The general idea is that both fields, lexicography and information science, have certain important things to contribute and to learn from each other. For instance, when analysing the problems relating to innovation in lexicographical works on an electronic platform, Bergenholtz & Bothma (2011: 74) comment: One of the reasons for the lack of innovation in e-lexicography is that lexicography is usually treated as a part of linguistics and lexicographical tools are primarily compiled by specialists with linguistic background. Our main thesis, namely that lexicography is not a part of linguistics but a part of information science does not support this line of thought. (Bergenholtz & Bothma 2011: 74)

It is evident that this integration of two hitherto independent disciplines with long traditions of their own is not something which can be solved overnight, and neither can it be a unilateral process. Yet some progress has been made recently. Bothma (2011) as well as Bergenholtz & Bothma (2011) have, for instance, dealt with needsadapted data presentation in e-information tools, showing how information science may contribute to solving this complex problem which is shared by both disciplines. Conversely, Bergenholtz & Bothma (2013) have criticized the common informationscience idea of “information needs changing over time” and have, considering the complex function-theory definition of information needs, argued that no such change is taking place; this view, they maintain, actually confuses old needs occurring for a user in one particular situation with new needs arising for the same user in new and even different types of situations. Similarly, Bothma & Tarp (2012) have analysed the theory of relevance – mainly, but not exclusively, developed in the field of information science – in a lexicographical light, and have reached the conclusion that such an analysis is beneficial for both disciplines, as it indicates that “the theory of relevance in information science should be expanded” and that at the same time it “provides an elegant theory to understand relevance as a complex phenomenon that may have a profound influence on lexicographers’ analysis of users’ information needs”: Research in relevance theory […] enriches the theoretical underpinnings of both information science and lexicography, and has a practical implication for providing better information tools (lexicographic information tools as well as other information tools) to users of such tools. (Bothma & Tarp 2012: 106)

The redefinition of lexicography and its categorisation as an information science discipline does not imply that its independent character should be lost. On the contrary, it means that lexicography takes its place in a more natural environment and is close to a source from which, true to its traditional interdisciplinary vocation, it

Relation to information science | 33

can obtain inspiration and avail itself of useful theories, methods and techniques after adapting them to its specific nature; this has been emphasised in section 3.1 and exemplified above by means of the relevance theory. The intention is, therefore, not to dilute lexicography as a discipline with its own core characteristics, but rather to focus even more on these characteristics whilst adopting everything that proves to be valuable in this connection. Consequently, in chapter 6 we will discuss in further detail how various information science techniques can be very useful in the development of advanced specialised online dictionaries.

4 Concept of Lexicographical Theory A major problem which lexicography shares with many disciplines within social sciences and humanities is the indiscriminate use of the terms science and theory, with no real clarification of what exactly they refer to. This shortcoming is a frequent source of misunderstandings and unnecessary disputes in the accompanying literature. Apart from the careless use of the two terms, the problem is amplified by the fact that the philosophy of science knows various competing concepts of science and theory. The problem is similar to the frequent use of a non-systematic terminology discussed in chapter 2. In order to avoid discussing at cross-purposes, it is therefore expedient that all researchers explain the specific concepts of science and theory to which they are referring. With these introductory comments, we will now return to Henri Bejoint’s statement on the non-academic status of lexicography, already quoted in section 3.2: I simply do not believe that there exists a theory of lexicography, and I very much doubt that there can be one. Those who have proposed a general theory have not been found convincing by the community, and for good reasons. A theory is a system of ideas put forward to explain phenomena that are not otherwise explainable. A science has a theory, a craft does not. All natural phenomena need a theory, but how can there be a theory of the production of artefacts? There are theories of language, there may be theories of lexicology, but there is no theory of lexicography. Lexicography is about all a craft, the craft of preparing dictionaries, as well as an art, as Landau (2001) says. It may be becoming more scientific, but it has not become a science. (Bejoint 2010: 381)

In these few lines, the former President of the European Association for Lexicography and current Professor Emeritus at the University of Lyon presents three fundamental postulates of great relevance for the discussion in this book, all of them aimed at showing that there is no theory of lexicography: – Lexicography is above all a craft as well as an art. – Lexicography may become more scientific, but it is not a science. – A science has a theory, a craft does not. Upon this basis, Bejoint asks defiantly: “How can there be a theory of the production of artefacts?” Before we answer it is worth drawing attention to the clause immediately preceding this question, namely, that “all natural phenomena need a theory”. In this respect, he also states that “there are theories of language”, but “no theory of lexicography”. With these words he is actually saying that there is a big difference between language and lexicographical products, in the sense that language is a natural phenomenon, whereas lexicographical products are another thing, that is, artefacts produced by human beings. Probably without intending to, this statement supports the idea, defended in section 3.1, that lexicography has its own independent subject field composed of dictionaries, i.e. cultural artefacts; in

Is lexicography a science? | 35

contrast, the subject field of linguistics is language, a genetically determined and socially stimulated phenomenon inherent in modern Homo sapiens, without which we would not be human beings in a social and cultural sense.

4.1 Is lexicography a science? Bejoint is not the only scholar denying lexicography the status of science but he does not make explicit what he understands by science when making this claim. This is, in our view, an unfortunate shortcoming in his argumentation due to the existence of various concepts of science, some of which may even lead to the opposite conclusion. One such concept is the one formulated by Manfred Buhr and Georg Klaus in their Philosophisches Wörterbuch, and used as a basis for the reflections in this book. It can be summarised in the following way: A science is a system of knowledge growing out of social practice and developing on an on-going basis, comprising the acknowledgement of the most important properties, causal connections and legal considerations of nature, society and philosophy; rooted in the form of concepts, categories, defined goals, laws, theories and hypotheses, and constituting the basis of Man’s growing mastery of his natural and social environment. A science also consists of its own history, pre-theoretical ideas, contributions to methodology, instructions for practical action etc., cf. Buhr & Klaus (1971: 1083, 1169) If the existing literature on lexicography is related to the above concept of science, then it can be argued that this branch of human activity complies with all the requirements necessary to be treated and classified as a separate science or area of academic study, cf. Tarp (2008: 4f): – It constitutes a system of knowledge growing out of social practice; – it has its own subject field; – it is rooted in the form of concepts, categories, theories and hypotheses; – it comprises both the history of dictionaries and its own history, including pretheoretical ideas; – it contains independent contributions to methodology; – it includes instructions for practical action. This, of course, does not mean that everybody doing lexicography is engaged in science. It only suggests that it is possible to treat lexicography as such on the condition that the work is carried out with the necessary scientific rigor, especially in terms of theory building. From this perspective, the task of lexicographical science in society is partly to obtain a wider knowledge of dictionaries and their role as cultural products aimed at satisfying specific types of social need, and partly to help develop new and better dictionaries. It should, however, not be forgotten that the concept of science described is not the only one allowed for in the broad framework

36 | Concept of Lexicographical Theory of the philosophy of science; it is, for instance, different from the one generally accepted in the Anglo-Saxon tradition and espoused by many lexicographers with their roots in this tradition. It is noteworthy that even a well-established discipline like information science, which includes the term science in its very name, comprises an internal tendency belonging to the above-mentioned tradition which rejects or at least raises doubts over its status as a science, e.g. Machlup (1983), Brooks (1989) and Buckland (2012); this is a debate which other scholars regard as “not particularly rewarding” (Norton 2010: 8). If we leave aside this Anglo-Saxon tradition, which in the final analysis constitutes only a small – although vociferous – part of the world academic community, we will find that a number of disciplines which are not considered sciences in this particular tradition are actually regarded as such in others. In Danish universities, for instance, students can study cultural science, art science, literature science, religious science, social science, or information science, among many other disciplines. More or less the same disciplines can be studied as sciences with academic degrees at universities in Germany, South Africa, etc. Within lexicography it is not only the supporters of the function theory who consider their discipline to be a science which can, and ought to, be studied as such at university. The same view is held, for instance, by the Russian lexicographer V. Dorosevskij who, albeit with a different understanding of its relation to linguistics, considers lexicography to be “the science of the classification processes of word material and its presentation in dictionaries” (quoted in Sorokoletov 1978: 79). The Spanish scholar Martínez Motos writes that “it is widely acknowledged that Lexicography is a science concerned with compiling, writing and editing dictionaries, as well as developing the principles that govern the process of dictionary making of the general lexicon of a language” (Martínez Motos 2011: 5). Igor Kudashev defines lexicography as “a branch of science which deals with presentation and description of lexical units of natural languages in the dictionaries” (Kudashev 2007: 159). By comparison, H.E. Wiegand, who bases his views on a concept of science formulated by Posner (1988), regards lexicography as a scientific research area rather than a science, as it does not fulfil two of fifteen required criteria formulated according to this concept, the most important of them being that there is “no consistent metalexicographical theory available in a falsifiable form” (Wiegand 1989: 261). The way in which the various scholars understand the academic status of lexicography basically depends on the underlying concept of science accepted by each of them. We will later – in section 4.3 – return to the question of falsification, which is occasionally used to reject the function theory, but first we will briefly discuss the concept of theory, likewise a subject of much dispute.

A lexicographical theory is possible | 37

4.2 A lexicographical theory is possible It is impossible to read the academic literature on lexicography without encountering the word theory. It is used as reference both to general theories that cover the whole subject field of lexicography as a discipline, and to specific theories that cover any sub-area of this enormous discipline. Some of the general theories are wellknown among most members of the lexicographical community: – [Towards a] General Theory of Lexicography (Scerba 1940); – General Theory of Lexicography (Wiegand 1998); – Function Theory (Bergenholtz & Tarp 2002, 2003). Among the many specific theories formulated by a large number of researchers can be mentioned: – Theory of Bilingual Lexicography (Duda et al. 1986); – Theory of the Lexicographical Example (Hausmann 1985 and Zöfgen 1987); – Theory of Lexicographical Language Description (Wiegand 1983); – Theory of the Dictionary Form (Wiegand & Morán 2009), – etc. Against this background it is, therefore, surprising that a number of scholars – even excellent lexicographers such as Sue Atkins and Michael Rundell (2008) and Henri Bejoint (2010) – do not only deny the existence of a theory of lexicography but also the very possibility that it may exist. Even the late editor of International Journal of Lexicography, Paul Bogaards (2010), who confirmed that many authors refer to theory in their contributions to this journal, personally doubted that such a theory already existed. The only explanation for this contradictory phenomenon is that discussed in the previous section, namely, that different scholars refer to different concepts of theory and science. It goes without saying that the various general and specific lexicographical theories proposed vary a great deal, as some of them are strongly embedded in linguistics whereas others are based upon the idea that lexicography is an independent discipline (and even science) with a strong interdisciplinary vocation. Besides, it must be admitted that reference to theory is normally made without any clarification of what is meant by theory. As we saw with the concept of science, there are also various “competing” concepts of theory to be found in the academic and philosophical literature. In this regard, the function theory is explicitly based on a concept of theory, again taken from Philosophisches Wörterbuch, which can be summarised as follows: A theory is a systematically organised set of statements about an area of objective reality or consciousness, i.e. logical structures reflecting the fact that certain things have certain properties, or that certain relationships exist between these things, cf. Buhr & Klaus (1971: 155, 1083).

38 | Concept of Lexicographical Theory In this spirit, a general lexicographical theory can be defined as a systematically organised set of statements about dictionaries and other lexicographical works and their relationship with specific types of social need, cf. Tarp (2008: 9f). Furthermore, it is necessary to distinguish between different types of lexicographical theory: – general and specific theories, i.e. theories covering the whole discipline of lexicography and theories only covering a restricted subarea; – integrated and non-integrated theories, i.e. specific theories integrated or not in a general theory of lexicography; – contemplative and transformative theories, i.e. general or specific theories that merely describe existing lexicographical practice, and those which also provide guidance for future dictionaries. Let us once more return to Bejoint (2010: 381) who claims that “lexicography is above all a craft” and who, therefore, does not permit for any theory because there cannot be “a theory of the production of artefacts”. What should we think of these statements? First of all, it must be admitted that the practical production of dictionaries is neither a science nor a theory but a millenarian cultural practice that can reasonably be defined or viewed as a craft which, as all other crafts, has developed historically with a view to satisfying certain needs detected in society. However, this in no way excludes that this craft – as well as the needs giving rise to it, its practical products (dictionaries and other lexicographical tools) and the use made of these – can be the object of observation, empirical studies and theoretical generalizations. With the above definition of the concept of theory as our starting-point, it is perfectly possible: – to observe and study this craft in all its dimensions, – to isolate relevant phenomena with certain properties, – to establish the relations existing between them, – to make statements about these phenomena and relations, and – to systematize these statements. This is all that is required to formulate a coherent theory of lexicography if the underlying concept of theory is the one discussed above. It is also worth noting that this is actually what is encountered, at least in terms of the first four of these steps, in the bulk of the academic literature on lexicography; in contrast, it is still relatively uncommon to find the required theoretical generalizations and the systematization of the statements made in the form of logical structures. As a consequence of this, one cannot but agree with the Chinese scholars Heming Yong and Jing Peng, who, in a book on the history of Chinese lexicography, criticise previous studies in this area because, among other things, “dictionary compilation is viewed as a purely linguistic activity”, and also stress the need for a theory to support this type of research:

A lexicographical theory is possible | 39 It is frequently apparent in their research [the previous one] that more emphasis is laid on the parts than on the whole, that more attention is paid to the isolated analysis of cases than to theoretical generalizations, and that more consideration is given to accumulation of practical experiences than to formulation of lexicographical theories. (Yong and Peng 2008: 5)

Even with these limitations, the inevitable conclusion which can be drawn from the previous discussion is that a theory of lexicography is possible, and that various – frequently competing – general and specific theories of lexicography do exist, thus providing further arguments for the classification of lexicography as a science, as discussed in section 4.1. It was such a theory capable of explaining, guiding and even renovating the existing practice that Scerba (1940) intended to formulate in his ground-breaking contribution to lexicography; this was what Wiegand did with his general theory of lexicography, and what has been done with the function theory. It may be that these theories “have not been found convincing” (Bejoint 2010: 381) by the Anglo-Saxon lexicographical community, but this does not mean that they do not exist or are not possible, or that they are not needed by those trying to solve the complex problems within present-day lexicography. Whereas specific theories may deal with whichever area or subarea of the discipline, a general theory of lexicography should necessarily include all aspects of lexicography (including all types of dictionaries and other lexicographical works) and cannot be restricted only to a subset of dictionaries, e.g. dictionaries where special knowledge of linguistics has been required. A general theory of lexicography cannot ignore the point that lexicographical works are multi-faceted cultural artefacts and utility tools which, during the millennia, have met a wide range of different needs detected in society and covered almost all spheres of human activity and knowledge. Nor should it take its point of departure in the differences that separate all these works in terms of their specific content, structure etc., but in the aspects and elements that unite them and are common to all of them. According to the function theory (cf. chapter 5), these uniting aspects and elements which can be considered the core of lexicography are: – the design of utility tools – that can be quickly and easily consulted – with a view to meeting punctual information needs – occurring for specific types of users – in specific types of extra-lexicographic situations. If this core activity is studied and generalized in the form of a systematic set of statements, the theory thus formulated will be able to guide the design not only of an entirely new generation of lexicographic works of the kind we already know, but also of consultation tools covering areas that have still not been treated lexicographically. No theory is born fully armed like Athena from Zeus’ forehead, and

40 | Concept of Lexicographical Theory existing lexicographical theory, including the function theory, has to undergo a continuous process of validation and perfection as a result of the fruitful interaction with its alter ego, the lexicographical practice, which is also in a process of continuous progress.

4.3 Additional remarks on the concept of lexicographical theory In a recent publication in which he discusses the function theory, the Japanese researcher Yukio Tono (2010) asks whether “we really need a ‘theory’?” Tono’s final answer to his own question is affirmative. At a high level of abstraction, i.e. independent of the specific theory in question, this view is shared by members of the Russian tradition (e.g. Scerba 1940, Sorokoletov 1978), the German lexicographical tradition (e.g. Duda et al. 1986, Wiegand 1989, Zöfgen 1994, Weber 1996), Danish lexicographers (e.g. Kromann, Riiber & Rosbach 1984, Bergenholtz & Tarp 2003), Swedish lexicographers (e.g. Svensén 2005), Spanish lexicographers (e.g. FuertesOlivera & Arribas-Baño 2008), various South African lexicographers (e.g. Gouws & Prinsloo 2005), Brazilian lexicographers like Kilian et al. (2012), the Czech-American lexicographer Zgusta (1992), the Indonesian lexicographer Kwary (2012), and at least part of the Chinese tradition (e.g. Yong & Peng 2008), among many others. However, this appreciation of lexicography is, as we have seen in the previous sections, rejected by other lexicographers, especially some of the ones belonging to the AngloSaxon tradition. Unfortunately, neither Atkins & Rundell (2008), Bejoint (2010) nor Rundell (2012) provide any consistent argument to support their opinion that a theory of lexicography does not and cannot exist. Other scholars, however, relate the existence of theory to the problem of falsification. Paul Bogaards, for instance, does not deny that such a theory may exist but questions whether it has already been developed. Consequently, he does not accept the view that the function theory should be considered a theory, primarily because it “lacks any form of empirically verifiable or falsifiable hypotheses” (Bogaards 2010: 316). A similar comment was sent to one of the authors of this book: You finally define a theory (of lexicography) as ‘a systematically structured set of statements’. Without going into long epistemological debates, I would say that a theory should first of all consist of a number of hypotheses and axioms, as well as of a set of methods to falsify the hypotheses. (Bogaards: Personal communication)

Here it becomes clear that the discrepancy as to the validity of this specific theory is due to the underlying concept of theory used by each researcher. The function theory does not consider hypotheses to be part of the concept of theory, but as something pre-theoretical that may eventually lead to theory (or improved theory) and which belongs to the broader concept of science, cf. Buhr & Klaus (1971: 1169).

Additional remarks on the concept of lexicographical theory | 41

Falsification is a very useful method to validate theories within many disciplines but, as Bergenholtz, Bergenholtz & Tarp (2008: 157f) has argued, the idea that every theory has to be falsifiable is one which has been adopted from natural sciences and the Popperian tradition. As a rule, this idea is not valid in various social disciplines where a lot of empirically grounded theories – i.e. theories built on practice-based evidence – has been formulated that are absolutely necessary in order to understand practice and make progress within these disciplines, although they cannot be falsified in the same manner and with the same rigid criteria as theories relating to natural sciences. The same can be said of lexicographical theories. The idea that social sciences and humanities should be subjected to the same criteria as natural sciences is basically a reductionist one that does not take into account the special characteristics of these disciplines, reducing everything to natural phenomena. This reductionist view of science is in no way helpful for the development of lexicography. It is obvious that this rejection of falsification as a relevant method does not imply that the value of lexicographical theories cannot and should not be tested. The criterion for evaluating and validating a transformative lexicographical theory such as that of function is practice, which, from a lexicographical perspective, has three dimensions: 1) the product, 2) the production process, and 3) the quality of the final product: – Is it possible to produce lexicographical works when guided by the theory? This is very easy to test, and the answer constitutes the criterion to judge whether a theory is transformative or not. – Is it possible to produce lexicographical works in less time and employing fewer human and material resources when guided by the theory? This can easily be tested by simply measuring, and it makes it possible to compare various theories in terms of productivity. – Is it possible to produce lexicographical works of higher quality, i.e. those satisfying users’ needs in all aspects, when guided by the theory? This can also be tested by means of the most appropriate methods; however, it requires a much bigger effort as it involves time-consuming qualitative methods, cf. Bergenholtz, Bergenholtz & Tarp (2008) and Tarp (2009b). Basically, this is also a way to compare the value of various theories in terms of dictionary making. This kind of evaluation and validation is much more relevant to lexicography than the reductionist criterion of falsifiability. From this perspective, it is worth remembering that dictionaries have been tested by users on a daily basis for more than four thousand years; therefore, they have long stood the test as a type of tool capable of meeting man’s ever-increasing punctual information needs. Even when they are not the direct products of a theory, these works have always represented the materialisation of ideas previously taking shape in the mind of creative human beings. Modern lexicographical theory originates from this rich empirical basis, as well as from

42 | Concept of Lexicographical Theory pre-theoretical experience encouraged by the general social demand that lexicographical works should continuously be improved and designed to cover still new and more complex needs. However, even if the Popperian criterion is accepted it should not be forgotten that Popper himself regarded falsification as a method by which to judge whether or not a specific theory can be considered scientific, and in this way demarcate it from other theories deemed unscientific. He never claimed that unfalsifiable theories – e.g. theories relating to history and similar research areas – should be disregarded as theories and that they were necessarily incorrect, simply that, in his view, they could not be considered scientific: Thus the problem which I tried to solve by proposing the criterion of falsifiability was neither a problem of meaningfulness or significance, nor a problem of truth or acceptability. It was a problem of drawing the line (as well as it can be done) between statements, or systems of statements, of the empirical sciences, and all other statements... The criterion of falsifiability is a solution to this problem of demarcation because it says that statements or systems of statements, in order to be ranked as scientific, must be capable of conflicting with possible, or conceivable, observations. (Popper 1963: 38)

In this regard, Wiegand (1989) appears, as far as we are concerned, much more consistent than Bogaards (2010), since his acceptance of the fact that his own theories are not available in a falsifiable form does not prevent him from formulating new theories and presenting them as such. The same can be said of other lexicographers who do not even regard falsifiability as a relevant criterion within their discipline, and consequently continue their theoretical work largely unimpressed by the attempt to disregard their research. If scholars engaged in lexicography renounced theory-building for whatever reason, then practical lexicographical projects would resemble ships sailing the seven seas without a captain: it would take much longer for them to arrive at a safe port, assuming they could. This is not what lexicography needs, especially not at a moment when it happens to be in the middle of a challenging but also troublesome transition from printed to electronic dictionaries. In this respect, it is appropriate to quote the visionary words of the South African lexicographer Rufus H. Gouws: Looking back at the development of the theory and practice of lexicography it is clear that for too long the practice of printed dictionaries had to go without a sound theory, for too long lexicography did not establish itself as an independent discipline, for too long the pool of lexicographers had been restricted to experts from a single field, for too long innovation in the lexicographic practice was impeded by its theory being a follower and not a leader, for too long lexicographic theory was exclusively directed at being implemented in the production of dictionaries. Looking at the future, the planning and compilation of electronic dictionaries and the further development of a coherent and medium-unspecific theory we need to unlearn a lot, we need to learn a lot so that we can be innovative and produce better reference tools, including even dictionaries. (Gouws 2011: 29)

Additional remarks on the concept of lexicographical theory | 43

A theory such as the one sought by Gouws will be developed in the next chapter, with special focus on specialised lexicography.

5 General Theory of Specialised Dictionaries In this chapter we will discuss the function theory and some of its implications. The theory was originally inspired by ideas put forward by distinguished lexicographers such as Scerba (1940), Hausmann (1977), Kromann, Riiber & Rosbach (1984), and Wiegand (1977, 1989), among others, i.e. lexicographers from outside the AngloSaxon tradition. As explained by Tarp (2008: 33f), it developed through various phases and soon crystallised into an independent school of thought without abandoning its theoretical prehistory. From the outset it was conceived as a general theory of lexicography, albeit strongly influenced by specialised lexicography which provided unique material that helped overcome limitations observed in previous studies and theories. One of the things which characterise theories formulated within humanities and social sciences and distinguish them from most theories belonging to natural sciences, is that the subject field is undergoing constant – and frequently rapid – changes, not only quantitatively but also qualitatively. In lexicography this is reflected, for instance, in the means that have been used to embody its practical products over the millennia, from clay, papyrus and paper (handwritten or printed) to the various electronic platforms used today. A general theory will necessarily have to cover all aspects of the discipline including the works presented by means of the different media. Consequently, if such a theory had existed four thousand years ago there would be no reason to formulate or reformulate it in the present phase of gradual transition from printed to electronic dictionaries. However, two different phenomena interact and determine the need for constant improvement and reformulation of lexicographical theory. The first is common to all scientific work, namely, that human cognition by definition is a gradual and ever-going process characterised by a still deeper insight into and understanding of the subject studied, a process where theory engages in a fruitful interaction with practice. The second phenomenon appears interwoven with the first, and concerns the various qualitative changes in terms of the technologies employed to produce, present and make use of lexicographical products. In this respect, it is perfectly possible that elements which existed from the very outset of lexicography, yet only in embryonic form, were not sufficiently taken into account, or even ignored, until these previously “hidden” elements flourished at a later stage. This is, among other things, the case with elements linked to the concept of information, which did not manifest themselves in a fully developed form until the present “information society”, in which it has become increasingly clear that lexicography shares a lot of elements with information science and should, therefore, be reformulated as an independent discipline within this science (cf. section 3.4). Hence, the theory described in the following is an updated version of the function theory which will probably have to be refined even more in future works, when

Utility tools and user needs | 45

more experience has been gained in terms of practical products and confrontation of ideas.

5.1 Utility tools and user needs The function theory is based upon the fundamental postulate that lexicographical works are utility tools. This postulate is the result of observation of practice, incorporated in a body of dictionaries produced and consulted over more than 4.000 years. As such, there is no reason to “prove” its correctness by means of falsification or any other method; it is an axiomatic postulate similar to the ones to be found in theories within many other disciplines. In addition, it is widely accepted and frequently referred to in lexicographical circles. However, the challenge is not to pay lip service to an evident fact and accept it in general terms, but to accept the necessary consequences in all their dimensions. The first scholar to do this was Wiegand (1987), who was later followed by the advocates of the function theory. In fact, the attitude towards this fundamental postulate represents the first dividing line between those who defend the need for a lexicographical theory and those who deny its very existence. As utility tools, lexicographical works share the same essential characteristic of any other human-made tool, namely, that they are, or should be, designed with the purpose of satisfying specific types of human needs. Lexicographical works always represent a relation between at least two people, the user and the lexicographer; as such, lexicographical tools are basically social creations and lexicography is a social discipline. The main challenge is, therefore, to explore and describe the relation existing between lexicographically relevant needs detected in society – i.e. among individuals or groups of human beings – and the solutions or intended solutions provided by lexicographers in tools designed to meet these needs, as the latter cannot be understood outside the dialectical relation between specific social needs and their satisfaction. On the other hand, to take the starting point as that of lexicology (Pérez Hernandez 2002), terminology (Cabré 2003), pragmatics (Piotrowski 2009), Chomskyan linguistics (Ten Hacken 2009), or any other branch of general or specialised linguistics, does simply not make sense and runs counter to any serious scientific approach. Although the knowledge provided by these and other branches or sub-theories of linguistics may be absolutely necessary for the production of a number of specific dictionaries, they can never by themselves determine the concrete user needs to be solved by a particular dictionary, let alone constitute the basis for a general theory of lexicography, which takes into account all types of past, present and future dictionaries, be they general or specialised. Hence, the real point of departure for a theory of lexicography are the social needs giving rise to dictionaries and other lexicographical utility tools, regardless of their name. In this respect, it should not be forgotten that most scholars within the

46 | General Theory of Specialised Dictionaries field of lexicography refer in their writings to user needs, although the discussion, sadly enough, tends to stay at a general level without getting to the bottom of the issue; as a result, the necessary theoretical and practical measures are not taken. The function theory proceeds differently, as the main drive behind it was and still is to formulate a transformative theory, that is, one with the potential to orientate the development of an efficient methodology that may not only be applied in order to analyse and describe existing dictionaries but also, and more importantly, to furnish a more accurate picture of real user needs; in this way, it will be possible to determine the pertinent data to be included in a specific dictionary without the necessity of resorting to more complex, dubious and time-consuming methods. In fact, various methods to determine user needs may be observed within lexicography, among others: – Business as usual – Personal knowledge – User research – Functional approach The first type of “method” is probably the one most commonly employed in the compilation of dictionaries, especially in relation to specialised ones. It consists of doing business as usual by reusing or plagiarizing already existing dictionary concepts with no or few modifications. This “method” may lead to excellent dictionaries if the concept copied is of a high quality and suitable for solving the specific needs of the foreseen user group, but in most cases it results in dubious, low-quality products. As an example of this can be mentioned the many bilingual specialised dictionaries which offer only translational equivalents without meaning discrimination, definitions and relevant grammatical data, and which, therefore, are seldom helpful in the situation they claim to cover (mainly translation). Although it is always recommendable to study and learn from existing practice, this “method” of business as usual cannot, by definition, provide the innovation and creative solutions required in the present online environment. The second type of method is applied when the lexicographer has a close personal knowledge of the needs of the intended user group. One such example was the Danish priest Jens Høier Leth (1800), who personally observed the many comprehension problems among simple-minded children when he tried to teach the catechism whilst preparing their confirmation; he therefore decided to write a Dansk Glossarium (Danish Glossary), where he explained difficult words and expressions in a simple way (cf. Pálfi & Bergenholtz 2007). Another example is the British economist Malachy Postlethwayt (1774) who, according to his own words, had a rich communication with his users, and on this basis incorporated a lot of surprising data in the later editions of his Universal Dictionary of Trade and Commerce; this represented a lexicographical innovation from which we can learn even today (cf. Tarp 2013a). It is always an advantage to combine a creative mind with a profound

Utility tools and user needs | 47

knowledge of the real user, and many good dictionaries, mainly monofunctional ones, may still be produced in this way. However, the personal knowledge of the user is clearly not sufficient, and as a method appears a shade artisanal when the challenge is to solve the general and complex problems in the present transition to the electronic medium. A third way to provide information about users’ needs is represented by the various forms of user research, from the simple and easily handled – but strongly criticised – questionnaires to more complex and time-consuming methods such as interviews, protocols, tests, and observation of the consulting process, e.g. through eye tracking or the study of log files in online dictionaries. The relative usefulness of these methods is illustrated in a book authored by the German-Brazilian scholar Herbert Andreas Welker (2006), who provides the hitherto sole systematic overview of user studies published in European languages. These studies have employed various methods to provide data concerning user needs, the consultation process and the results of this process. Welker’s book shows that most – but not all – studies suffer from a very limited number of informants, in many cases less than twenty. If one accepts the principles of modern statistics, these studies are only representative for the informant group itself and cannot be considered so for a broader group of users; as a consequence, they are of little value when planning a specific dictionary. With this in view, Tarp (2009b: 293) concludes that most lexicographical user surveys represent “a waste of time and money”. Similarly, Bergenholtz & Bergenholtz (2011: 190) maintain that the majority of these studies have been carried out “in the most unscientific way imaginable, as they were conducted without any knowledge and without use of the methods of the social sciences”. It is interesting that Rundell (2012a: 50), in an attempt to provide arguments against the supporters of the function theory, explicitly disagrees with these statements, and writes that “this does not chime with my experience”. In doing so, he uses a few well-performed studies to justify the bulk of surveys, although he admits that “there are inevitably some unevenness in quality”. Thus, Rundell does not only deny the existence and relevance of lexicographical theory but also tacitly disregards the principles of modern statistics and social sciences. In this respect, it goes without saying that the function theory does not deny the value of user research conducted according to scientific standards, especially the research performed with qualitative methods. However, in most cases it considers this type of research to be too complex, too time-consuming and too costly to constitute a relevant method for providing the necessary knowledge of the intended users and their needs in terms of each and every new dictionary. The general information about user needs provided by a few serious studies, designed according to genuine scientific principles, is clearly not sufficient to guarantee that a particular dictionary turns out to be of a high quality as regards meeting the concrete needs of the foreseen user group.

48 | General Theory of Specialised Dictionaries

5.2 Concept of user needs From what we have seen above it is evident that lexicography needs an efficient methodology which, in a relatively easy, quick and cheap way, can be applied to develop specific concepts for specific dictionaries without compromising their quality. Such a methodology is provided by the function theory. It is based upon the encirclement and subsequent analysis of lexicographically relevant user needs, i.e. the dissection of these needs into their smallest meaningful parts. The first step in determining the concept of lexicographically relevant user needs involves delimiting and separating them from other types of human need. The fact is that human beings may experience countless needs, of which only a few can be met by lexicographical works. The very nature of these creations reduces the types of need they are able to satisfy to those of information. Besides, lexicographical works are designed to be consulted in one way or another. If one applies the terminology used by Hausmann (1977) and distinguishes between global and punctual user needs, then it is evident that lexicographical works are primarily conceived to meet the users’ punctual information needs in a linear perspective, where the binomial “global-punctual” should not be understood as a relation between “big and small”, but between “the whole and the part”. In this way, even a dictionary with fairly long articles, such as The New Palgrave Dictionary of Economics, provides punctual and concentrated vanguard information about a part of economics conceived as a whole. It goes without saying that lexicographical works are not the only tools designed to give punctual information on various subjects, as the same holds true for manuals, telephone guides, train and flight timetables etc. In this respect, it is interesting that Rodríguez (2013) shows how the methodology developed by the function theory may also be applied to describe and design other categories of utility tools, such as user guides and manuals of economics, which are not traditionally regarded as lexicographical works. This fact once more documents lexicography’s close affinity with information science, as discussed in section 3.4. However, it should not be ignored that information needs are not the only lexicographically relevant needs. They are the ones which initially lead to a lexicographical consultation, but this consultation itself may give rise to a new type of need different from the former and related to the correct handling and use of the lexicographical tool, with the specific purpose of retrieving the information required to satisfy the original needs. Once the concept of lexicographically relevant user needs in the broadest sense of the word has been clarified, the function theory proceeds to a more detailed description of these needs. Two fundamental factors interact in the formation of these needs which should never be considered abstract needs, that is, needs as such, because they always appear in a very concrete and specific form (although lexicographers, as a rule, have to take their point of departure in types of need; cf. section 6.2 for a discussion of the complex relation between concrete, individual needs and

Concept of user needs | 49

types of needs). The first of these factors is the characteristics of the concrete person experiencing the information needs, and the second is the social situation or context where these needs occur. As regards the first factor, it is evident that the precise content of the user’s needs depends on the specific characteristics of each and every person. In order to study these characteristics the function theory resorts to the relevance criterion originally developed in the framework of information science, cf. Saracevic (1996), Cosijn & Ingwersen (2000), Borlund (2003), Cosijn & Bothma (2005), and Bothma & Tarp (2012). Any human being has countless personal characteristics of which only a small minority are lexicographically relevant; for instance, from a lexicographical perspective whether a person is red-haired, left-handed, beautiful, physically strong or choleric, etc., is totally irrelevant. In the final analysis, it is the dual nature of dictionaries as both information tools and consultation tools which determines the user needs that may be relevant from a lexicographical point of view. Therefore, it is necessary to distinguish between the user’s characteristics relevant to the information needs originally giving rise to a lexicographical consultation, and those pertinent to the consultation itself. For instance, a Danish and a Namibian school child with the same level of proficiency in English as a foreign language (one specific type of user characteristics) may need the same kind of information when they have textproduction problems in English. But the lack of correlation in social and economic development – and the corresponding unfortunate reality that the former is more likely to be constantly online with a particular device, while the latter will very often not even have access to electricity – determines the type of vehicle for a lexicographical solution. This vehicle may, in the first case, be the Internet while in the second case it will no doubt be a printed book with letters which are not too tiny. As a consequence of these reflections, a list of lexicographically relevant user characteristics may be drawn up based on the following questions: – Function-relevant user characteristics: Which language is the user’s mother tongue or first language? What is the user’s proficiency level in the mother tongue? With which method is the user learning the mother tongue or first language? What is the user’s proficiency level in a second, third, etc., language? With which method is the user learning a second, third, etc., language? What is the user’s general cultural and encyclopaedic level? What is the user’s experience in translation between a specific set of languages? What is the user’s knowledge of a specific discipline (layman, semi-expert or expert)? What is the user’s comparative knowledge of a specific culture-bound discipline as expressed in two specific language areas? What is the user’s proficiency level in a specific specialised language?

50 | General Theory of Specialised Dictionaries What is the user’s experience in translation between a specific set of specialised languages? Etc., etc… –

Consultation-relevant user characteristics: What is the user’s experience of lexicographical consultations? Is the user blind, deaf, or suffers from any other handicap which may limit the use of specific types of lexicographical tools? Does the user have electricity and electric light? Does the user possess a device with access to the Internet? Does the user know to distinguish between right and left? Etc., etc…

The answers to the questions listed above may help to create a detailed user profile. A crucial question is, nevertheless, how to obtain the answers. Basically, there are at least three different ways to proceed, one of which is very time-consuming and complex: – One way is to approach an amorphous group of potential users and let them answer the questions by means of questionnaires, interviews, etc. As this method implies problems in terms of statistical representativeness and reliability, as well as the time and money needed to conduct this kind of user research, it should generally not be recommended when doing preparatory work for a specific dictionary. – Another and more efficient method is to take a reasonably well-defined group of potential users, e.g. school children of a specific grade, university students of a specific discipline, a specific group of specialists or professionals, etc., and then consult someone with an intimate knowledge of this group, for instance, a school teacher, a university professor, a professional translator, etc. – Finally, a very efficient method is that in which the lexicographers define the predicted user group and its characteristics on the basis of their general knowledge of the subject and the list of questions given above; in other words, they themselves simply decide to which specific user group they will dedicate the planned dictionary. The resulting user profile could, for instance, be a group composed of semi-experts and experts in microbiology, with Spanish as their first language and a low proficiency level in both general and specialised (microbiological) English, with access to Internet and experience in the use of dictionaries. This last method, which, so to speak, represents the opposite of normal commercial market analysis, has frequently been applied by supporters of the function theory when preparing concepts for specialised online dictionaries, and has generally

Concept of user needs | 51

yielded good results. When, therefore, a user profile is being designed for a specific specialised dictionary, it may be considered recommendable to use one of the two last methods, or a combination of them; for instance, the first version of the concept is prepared according to the third method, and then it is refined by means of the second. In comparison with the first of the three methods mentioned, the latter two are relatively quicker, easier and cheaper to apply, whilst being efficient as a solid basis for ensuring correspondence between the foreseen and real needs of the intended user group. This, or course, does not exclude feedback from a selected group of actual users of the alpha or beta version, i.e. before an online dictionary is made available to the public in general or even afterwards, but it should not be forgotten that conceptual changes at this stage may imply time-consuming and costly reprogramming of the dictionary. The above list of questions aimed at defining the characteristics of the foreseen user group of a specific dictionary is, obviously, an open list to which more lexicographically relevant questions may be added in each specific case. At the same time it is evident that not all the questions are relevant for all dictionaries. What determines the specific relevance of each question is the social situation in which the user needs originally occur. This means that it is not enough to do user research in order to produce a profile of someone with specific lexicographically relevant characteristics for a specific dictionary. It is also imperative to take into account the foreseen type of user situation in order to determine which characteristics are relevant to the profile required in each case. This situation represents the second of the two factors determining the specific content of the lexicographically pertinent needs in each case. When one observes and generalises social practice, that is, the immense empirical material continuously made available, or one’s own personal experience (introspection), at the highest level of abstraction, three main types of situation can be observed in which a person may experience precise information needs: – When a person needs information either to store in the memory as knowledge or as background knowledge in order to solve a specific task, for instance, the curriculum of a specific person when writing about him or her in an article. – When a person needs information in order to understand or interpret a specific phenomenon, sign, symbol, text, etc. – When a person needs information in order to perform an action of a physical, mental or linguistic nature. These three main types of lexicographically relevant situation may be called cognitive, interpretive and operative situations. However, as a large number of dictionaries have traditionally been conceived and consulted with a view to solving problems of text communication, the function theory has, for convenience reasons, operated from the very outset with a fourth fundamental situation. This is represented by everything that concerns the production, reception, translation, and revision of oral

52 | General Theory of Specialised Dictionaries or written texts, and which has been separated and merged into a joint overall category, the communicative situation. For this reason, the theory distinguishes between four fundamental and lexicographically relevant situations where concrete information requirements may occur: – Communicative situations – Cognitive situations – Operative situations – Interpretive situations These fundamental user situations can be further divided into a number of subsituations, as already indicated for the communicative situation (text production, reception, translation, etc.). In this respect, it should be stressed that the term knowledge in relation to a cognitive situation (cognition) is not used in its vulgar, daily meaning where it also very often includes skills. In the terminology applied by the function theory it is compared to and enters into a dialectical relation with skills, which are the user qualities required in the communicative, operative and interpretive situations mentioned. For instance, all normal people have mother tongue skills allowing them to communicate with other speakers of the same language, but it is only relatively few of them who, through their studies, have acquired a learned knowledge of this language. One may have skills without having any knowledge of a specific subject, in the same way that one may have knowledge without having any skills, as already shown more than 250 years ago by the German poet and philosopher Gotthold Ephraim Lessing, in the dialogue between the young, well-read Damis and his more prosaic servant Anton, cf. Lessing (1747: 8-9). A top professional journalist reporting on the famous Tour de France may, for instance, possess a profound knowledge of cycling, its history, great achievements, business model, social status, etc., without being able himself to climb Alpe d’Huez or any of the other legendary mountains on the route. In most situations a person will be better equipped when possessing both skills and knowledge, but this does not imply that the concepts should be confused. Through practice, knowledge can be transformed into skills, while a systematic observation and study of the latter can lead to knowledge. The information acquired through consultation of dictionaries can be stored as knowledge but it cannot be directly transformed into skills. This information can assist users in performing specific mental, physical and linguistic tasks related to interpretive, operative or communicative situations and, in this way, it may be gradually internalised and reappear in the form of skills. Of course, many of these tasks, e.g. specialised translation, cannot be performed without a certain amount of background knowledge; this, however, does not mean that skills and knowledge should be confused but rather that specialised lexicography should take cognisance of the fact that they are interrelated and even interdependent in specific situations; cf. Tarp (2008: 131-136) for a more detailed discussion of the concepts of information,

Concept of user needs | 53

knowledge and skills and the complex relation existing between them in a lexicographical environment. In each of the situations and sub-situations mentioned above specific types of punctual information need may occur. However, it is not the situation alone which determines the needs relevant to a specific dictionary. The analysis of the types of need occurring in each type of user situation may help to draw up of a list of needs, some of which may be relevant and others irrelevant, for a specific type of user with specific characteristics. In addition, it is also these characteristics that determine the specific content and shaping of each data category to be incorporated into a specific dictionary, as well as its overall design in terms of access possibilities etc. So, how are the needs determined once the type of user and the type of situation have been defined? In this respect, there are at least three different ways to proceed: –

By means of questionnaires, interviews or similar methods: A number of users of the specific type in question can be asked about their needs in their specific type of situation. By means of observation: A number of users belonging to the specific type can be observed in order to detect the problems and needs they may experience in the specific type of situation in question. By means of deduction: Lexicographers in collaboration with the subject-field experts can, based upon their experience and profound knowledge of the subject matter, deduce the relevant types of need which may occur for the specific type of user in the specific type of situation.

None of these methods are perfect: The questionnaires are reasonably easy to handle, also in terms of the needed amount in order to meet the requirements of modern statistics. However, they may at the same time be vitiated by the usual errors referred to by social science and certain lexicographers, e.g. Hatherall (1984) and Welker (2006). Besides, the following points should not be ignored regarding the users: 1) although fully aware that they experience a need, they may be unaware of its specific nature, or 2) they may not even be aware that they have any relevant needs, cf. Tarp (2009b). Wellprepared interviews may remedy some of these deficiencies. In this respect, a distinction should be made between interviews with predetermined questions, and the type of interview which Welker (2006: 26) calls “open interviews” and Zikmund (1997: 122) “in-depth interviews”; the latter author characterises these as “relatively unstructured, extensive interviews”, where the respondents may speak their mind and where the interviewers do not base themselves on predetermined questions. The first kind of interview represents quantitative research and in social science is frequently regarded as a special form of questionnaire suffering from many of the same problems as questionnaires (cf. Hansen & Andersen 2000: 98). The second kind represents qualitative research, and may as such furnish interesting hints and

54 | General Theory of Specialised Dictionaries indications of what potential users may need; however, should it be employed to offer statistically significant information on the relevant types of need to be covered by a specific dictionary, then it is a very time-consuming and costly method to be used in the preparation of such a dictionary. Observation may be an efficient method, at least in connection with certain types of user situation (communicative, operative, and interpretive, though probably not cognitive ones). If it is followed by an interview which takes an in-depth look at the action patterns observed, then it may certainly bring forth relevant information. If, on the other hand, it is expected to be statistically significant and provide a reliable basis for a concrete dictionary, then it must be regarded, once again, as a highly time-consuming and costly method. Consequently, although they may generate some interesting information, neither the first nor the second method can be recommended when preparing the concept for a concrete dictionary. This is the reason why supporters of the function theory generally use the third method mentioned, i.e. deduction. When this is employed by lexicographers in collaboration with subject-field experts, it may result in a relatively easier, quicker and cheaper way to detect most types of relevant need; hence, it can generally be recommended, especially in the preparation of specialised dictionaries. Deduction should not be confused with introspection, which is a completely different method. If, for instance, a person teaching specialised translation has, over the years, marked thousands of translated LSP texts and read students’ protocols relating to this activity, and if the same person has then discussed the corresponding problems, including those related to the use of reference works, with the students, then that teacher will surely possess a deep knowledge of the problems and needs of this particular group of potential users in terms of LSP translation. As a result, if this person works in collaboration with an expert trained in lexicographical theory and practice, then it will be perfectly possible for the two of them to typologise these needs and determine which of them are lexicographically relevant, i.e. which of them may be solved by consulting a specialised dictionary with the corresponding function. The thousands of marked LSP translations, protocols and subsequent discussions with students will in most cases constitute much better background material than the result of a few dozen observations or the dubious data provided by hundreds of questionnaires. This type of deduction is based upon real knowledge rather than introspection. However, it should not be forgotten that there will almost always exist some non-detected types of need, especially those which rarely appear among the predicted users. In this regard, deduction is not a one-hundred-per-cent-perfect method, but the possible lacunae or gaps may largely be filled by the various techniques developed to meet the individual needs foreseen in advanced online dictionaries (cf. chapter 6). In this connection, it is important to stress that it is not a user situation in general, but rather the specific situation that gives rise to a specific number of relevant types of need. For instance, there may not be the same types of text-production need

Concept of user needs | 55

in English, Spanish, Chinese, Japanese and African languages just as it is not the same types of cognitive need which exist in relation to the study of biology, chemistry, history, etc. In order to determine the needs based upon a specific combination of user type and situation, it is therefore necessary that the person (the lexicographer or a collaborator), who is in charge of this task, has a profound knowledge of the subject matter whether this is a specific language or a specific scientific discipline. The same kind of expert knowledge is required when the user situation(s) to be covered by a planned dictionary has to be defined. Here there are basically two ways to proceed: the lexicographers or their collaborators can either decide themselves which user situation(s) they want to cover (according to their general knowledge of the subject), or else they can do some research in order to find out which situations may be relevant to a specific user group. For instance, students of a university discipline may very often require assistance not only with cognitive needs directly related to their study, but also in foreign-language text reception, as many textbooks are written in a language different from their mother tongue. In this respect it is worth remembering that many specialised dictionaries, whether printed or online, are conceived to cover various user situations and, at the same time, to satisfy the various types of need which a heterogeneous group of users with different types of characteristics may have in these situations. This is a major reason why many printed dictionaries suffer from various degrees of information overload, a problem also repeated in many online dictionaries, especially those categorised in section 2.3 as Copy Cats and Faster Horses. The problem, however, can easily be avoided in online dictionaries if the underlying system is interactive and designed to calculate the exact amount and types of data to be presented to each type of user in each type of situation; but the precondition for this to function is that the necessary preparatory work has been carried out and that each data category has been attached to at least one specific combination of a user type and a situation type (Model T Ford) or, even better, to a specific type of user need (Rolls Royce). In chapter 9, we will provide some concrete examples of how the function theory and its methodology are used to conceive and plan a concrete lexicographical project based upon user’s situations, characteristics and user needs. In this respect, it is essential that these categories are treated with the necessary scientific rigor, and not simply mentioned for the sake of appearance, as it frequently happens. In his arguments against lexicographical theory in general – and the function theory in particular – Rundell (2012: 59f) argues that there is nothing new in referring to user characteristics and situations as text reception and production; to support this view he lists various authors who have touched this question, even back in the 17th Century. This is correct, Tarp (2008, 2013b) also mentions various scholars who have spoken about the two types of phenomenon:

56 | General Theory of Specialised Dictionaries Acknowledgement of this close causal connection between registered social needs and the production of dictionaries as cultural products, an acknowledgement which is the very basis for developing an independent scientific theory of lexicography, didn’t arise from one day to another but was the product of a century-long process that culminated, so far, in the last decades of the twentieth century and expressed itself in the growing focus on lexicographic functions in the theoretical discussion. Scerba (1940) is usually considered as the father or initiator of this lexicographic approach, but as Hausmann (1989) shows in his “Small World History of Metalexicography”, Scerba was only the culmination of a process that started several hundred years earlier within the European lexicographic tradition. (Tarp 2013b: 93)

It is in no way surprising that this happens as the basic categories of the function theory are easily deducible from existing lexicographical practice. However, the challenge is not to make general statements about user’s characteristics and situations but to accept the corresponding consequences, that is, to use this knowledge to develop a theory capable of orientating and continuously renovating practice at a still higher level. It is no secret that there is a rich prehistory from which a lot can be learned but until a few decades ago nobody has taken the full consequences and used the knowledge about users and user situations as a basis to formulate a transformative theory of lexicography, a theory whose practical products represent “new relations to the users” (Humbley 2002: 95). Lip service to evident facts cannot substitute genuine scientific work. An example of this is an article by Faber, León Araúz & Pérez Hernández (2010), in which the authors first discuss, with reference also to the function theory, user situations, characteristics and needs, and then go on to analyse 15 definitions of the winemaking term fining provided in various types of specialised lexicographical works. However, in their analysis they seem to forget the specific needs of the users of the respective lexicographical works, as they take as their real starting-point a complex linguistic-terminological concept of definition, a method which leads them to the conclusion that “none of these definitions is adequate for user needs” (p. 115), despite several of them, in fact, appearing to be excellent for specific types of users in specific types of situations. The Spanish authors’ methodological problem seems to be that they have as their basis a definition concept taken from outside the lexicographical world, without adapting it to the specific requirements of this particular world; such an adaption could actually have been made quite easily. Another example is an article by the Catalan scholar María Teresa Cabré, where she compares and discusses various theories of terminology, putting forward a number of basic “assumptions” for a “new theory” of terminology. In the second of these assumptions she writes the following about the relation between, on the one hand, “terminological applications” and, on the other, user types and situations: Terminology is a set of applications in as far as it allows the development of products specifically intended to satisfy needs. The most important characteristic of such products is their appropriateness. A terminological application must be oriented towards the solution of specific needs and therefore it must take into account its recipients and the activities they plan to carry

Data and information | 57 out by means of such a specific application. This leads us to think that, despite of what is usually said about standardised terminological glossaries, it is the circumstances of each situation which determine the type of application (glossary, lexicon, dictionary, software, text, poster, standard, etc. in one or several languages), the information they must contain (terminology, phraseology, definitions, variants, contexts, phonetic or phonological representation, foreign language equivalents, illustrations, etc.), their representation and even their means of dissemination. (Cabré 2003: 182f)

At first glance this may seem to be similar to the method used by the function theory and which is described above. However, if we take a closer look it appears to be the reverse of this method, as the real point of departure in this case is terminology (i.e. something already prepared), which can subsequently be “applied” in accordance with the type of user and situation. The function theory operates differently, its starting-point being in user situations, characteristics and needs, and then initiating the lexicographical process in terms of specific dictionary design and compilation. It frequently uses the material produced by terminology or linguistics in general, but only after subjecting it to a critical analysis and adapting it to the requirements of lexicography. This reverse method recommended by Cabré and others may be one of the reasons why several projects carried out according to the principles of terminology drag on for years and frequently use several times more human and financial resources per data category than projects guided by the function theory (cf. chapter 8). To summarise, in view of the function theory a complex dialectic relation exists between the type of user (characteristics), social situation and specific needs to be covered by a specific dictionary: – Any type of user has a number of lexicographically relevant characteristics. – The type of user situation to be covered by a specific dictionary determines which of the user characteristics are relevant when drawing a specific user profile, i.e. a profile of the foreseen type of user. – The type of user situation also gives rise to a number of possible types of user need relevant to this situation. – The specific user profile determines which of the possible types of user need are relevant when conceiving a concrete dictionary. – The types of user need determine the data categories to be included into a specific dictionary. – The user profile determines the specific form given to these data categories.

5.3 Data and information In the previous chapters the terms data and information have been used with a very specific meaning which we will briefly explain here in order to avoid any confusion. In the theoretical literature on lexicography (and other cognate disciplines), one of

58 | General Theory of Specialised Dictionaries the many terminological inconsistencies that may lead to conceptual confusion and misunderstanding among those reading it are the various meanings attached to these two terms. In this respect, the function theory follows a well-defined terminology based upon the critical reflections by Wiegand (2000, 2002). In this terminology, the term data has two different meanings: it is used to denote both the raw material (raw data) which the lexicographers extract from various data sources such as corpora and the Internet, and the adapted, elaborated and prepared items which they finally incorporate in their dictionaries. Information, on the other hand, is what the users of the lexicographical products retrieve from these prepared data with a view to using it to their specific purpose, whether it is of a communicative, cognitive, operative, or interpretive nature. This terminology is different from the one generally applied within information science where the term information is also employed to denote the material included in information tools (among them dictionaries). Although the function theory has a close affinity with information science, there are two main reasons why it does not follow its terminology in this regard. The first is that the cognitive retrieval of information from data represents a highly complex mental process, the success of which requires that the data incorporated in dictionaries are adapted to the mental and cognitive characteristics of the user; in order to stress the difference between the input and output of this complex mental process, it makes sense to formulate a terminological distinction between the data incorporated in dictionaries and the information subsequently retrieved from such data by users. The second reason is that the exact meaning of information is still disputed within information science and frequently seems to embrace much more than the material included in information tools such as dictionaries (see, for instance, Buckland 1991, 1997, 2012, and Hjørland 1998). This also signifies that the use of the term information retrieval in the function theory is restricted to the mental process described, whereas it has a broader meaning in information science where, apart from the cognitive process, it also denotes what is normally called data selection within lexicography. In function-based lexicography, data selection presupposes the previous determination of the corresponding data categories. Once a foreseen user group’s punctual information needs in terms of the situations to be covered by a planned dictionary, have been clarified, then the next task in the conception of a dictionary – i.e. before its compilation in the strict sense of the word – is to decide which data categories to include in order to satisfying the detected needs, that is to say, not the concrete data which demands other methods, but the types of data. The crucial question is now how this should be done and by whom. Basically, in order to perform this task both specialised lexicographical and subject-field knowledge are required. Hence, although one and the same person may occasionally unite the various types of knowledge needed in relation to a specific dictionary, in most cases this task requires an interdisciplinary collaboration between an expert in lexicography and at

Data and information | 59

least one subject-field expert (in some multi-field dictionaries various types of nonlexicographical expert knowledge may be needed). The function theory, by means of the methodology described, is most useful above all when it comes to designing dictionary concepts and determining the specific data categories to be included in a specific dictionary project. However, it goes beyond abstract categorisation, being likewise concerned with giving content to these categories and providing methods and specific guidance to select the required data and prepare this for the users of specific dictionaries. It is, therefore, somewhat surprising that Rundell (2012: 60f), in a polemic article against this theory, claims that it says nothing about the origin of the selected lexicographical data, while De Schryver (2012: 496), in another article, even postulates that the data furnished by the supporters of the theory are “quite literally invented”. These postulates are, we consider, in flagrant contradiction to the facts, and it is indeed difficult to understand why such “arguments” are put forward by both lexicographers. It can hardly be due to ignorance, as the question has been dealt with in many readily-available articles and books, beginning with Bergenholtz and Tarp’s Manual of Specialised Lexicography; similarly, the methodology developed has been applied in a large number of general and specialised dictionaries designed with the guidance of this theory. Among the many contributions dealing with the selection of lemmata, equivalents, collocations and other types of data in specialised dictionaries, can also be mentioned those of Bergenholtz (1994), Bergenholtz, Tarp & Kaufmann (1994), Bergenholtz & Tarp (1994b) and, more recently, Nielsen & Almind (2011) and Fuertes-Olivera et al. (2013). In chapter 9, we will provide some concrete examples of how the lexicographical data has been selected and prepared in specialised dictionaries designed according to the principles of the function theory. In relation to this, a false and artificial opposition has been created in terms of the various attitudes to corpora. Supporters of the function theory have been criticized for not seeing the benefits of the electronic text corpora which have become available and fashionable during the last decades. The criticism which, for instance, has been raised by Kilgarriff (2012) and Rundell (2012) is, we believe, absurd. Advocates of this theory do not deny the value of corpora, defending, in addition, their relevance in the compilation of dictionaries; see, for example, Bergenholtz & Schaeder (1979), Bergenholtz & Mugdan (1989) and Bergenholtz & Pedersen (1994), where the composition of a text corpus for a specialised dictionary is undertaken, mainly with a view to extracting collocations. Nevertheless, it is true that the theory relativizes the usefulness of these corpora and does not engage in the chorus of blind and passionate love songs devoted to the “corpus revolution” (Hanks 2012). In some linguistic circles this “revolution” seems to have become a sort of miracle cure which may allegedly solve most lexicographical problems and free lexicographers from their individual responsibility towards the end users and their needs. The fact is that the use of corpora is absolutely necessary – or at least highly recommendable – in relation to specific data categories and dictionary types, whereas it is of little or

60 | General Theory of Specialised Dictionaries no use at all in the preparation of other data categories and dictionary types, cf. Fuertes-Olivera (2012). The position taken by the function theory on this question is probably due to the fact that the latter has its epistemological roots in the world of specialised dictionaries, where it is evident that text corpora have but a relatively limited value when data for a specific dictionary are being selected. Complex multi-word terms, for instance, cannot be detected by a concordance program or, at least hitherto, by the many prototypical tools for automatic term generation. In such cases it is necessary to employ other methods which can subsequently be complemented by corpus work, for instance, in the selection of relevant collocations once the terms have been established. In addition to this, supporters of the theory also visualize how selected corpora – and the Internet – can be integrated into advanced online dictionaries (Rolls Royces) and data from these sources can be taken in automatically in order to optimize the satisfying of user needs (see for instance Heid, Prinsloo & Bothma 2012 and Tarp 2012b). In this respect, the function theory also takes into account the time factor, that is, the time needed to compose a useful corpus when other easier ways to provide material for specialised dictionaries are within reach. It is difficult to agree with Rundell (2012), when he defines what he considers to be “the core task for lexicographers”: Function theory, it transpires, has little to say about what many of us see as the core task for lexicographers: analysing the evidence of language in use in order to identify what is likely to be relevant to dictionary users. (Rundell 2012: 60)

According to the function theory, the core task of lexicography is the conception and production of high-quality dictionaries, i.e. utility tools which can be quickly and easily consulted with a view to meeting specific types of punctual information needs, occurring for specific types of users in specific types of social situations. It may be that “analysing the evidence of language” is necessary when preparing quite a number of specific data categories and dictionaries, but it can never be the core task for lexicography as such because there are a big number of other data categories and dictionaries which do not presuppose any analysis of the evidence of language. This idea was already expressed in one of the earliest specialised dictionaries published in Britain, John Harris’ Lexicon Technicum with the eloquent subtitle “an Universal English Dictionary of Arts and Sciences: Explaining Not Only the Terms of Art, but the Arts Themselves”: That which I have aimed at, is to make it a Dictionary, not only of bare Words but Things; and that the Reader may not only find here an Explication of the Technical Words, or the Terms of Art made use of in all the Liberal Sciences, and such as border nearly upon them, but also those Arts themselves; and especially such, and such Parts of them as are most useful and advantageous to Mankind. (Harris 1704: Preface)

Data and information | 61

As a rule, the preparation and writing of definitions – or explications, as Harris calls them – for specialised dictionaries with cognitive functions require other types of analysis rather than the one mentioned above by Rundell, especially if the definitions should be understandable to the specific target user without being too boring and resembling “pickled frogs, outstretched upon a dissection board” (Van Loren 1962: 25). Although not extracted from any corpus, such definitions are in no way “literally invented”, but the result of a profound study and expert knowledge of the subject matter in question. Definitions may be found scattered throughout text corpora although they are always contextual, in other words, they may serve their purpose in a specific context, but in most cases they are more harmful than useful if they are incorporated uncritically as the only definitions in specialised dictionaries. (This does not exclude their use as a supplementary one to the main definition provided, cf. section 6.4). Many dictionaries compiled according to the principles of corpus-based definition “finding” are living proof of this (cf. chapter 8). A comment made by Adam Kilgarriff (2012) in his review of a book about elexicography, edited by Fuertes-Olivera & Bergenholtz (2011), also illustrates the somewhat exaggerated belief in the usefulness of data found in corpora, among them the Internet. Here, in his discussion of the contribution by Nielsen & Almind (2011), he criticises a definition of the accountancy term deemed cost used by the latter and taken from the Accounting Dictionaries. As his argument he opposes it with another definition which he has googled on the Internet (the second hit) and then comments: Accurate, extensive encyclopedic entries are very often already available, and very easily accessible via google, as here. A case has to be made for what value lexicographers are adding. (Kilgarriff 2012: 27)

Kilgarriff’s problem, as we see it, seems to be that he is a non-expert in accountancy and, therefore, not in a position to see that the definition found on the Internet might be correct in a specific context or for a specific purpose, but not as a definition of the term treated in the Accounting Dictionaries and here defined according to existing International Financial Reporting Standards (IFRS), and for the benefit of the specific target user in question. In addition, such terms may from time to time be redefined by the corresponding terminological language bodies (cf. section 5.6); for instance when changes are introduced in the relevant legislation. The definition googled by Kilgarriff is far too broad and covers much more than the IFRS standards relevant to the Accounting Dictionaries. In this respect, Kilgarriff is caught in the trap of “private property belonging to no one” (Harris & Hutton 2007: 212), i.e. “writing definitions for everybody and nobody” without targeting “specific user types” (Nielsen 2013: 57). Subject-field knowledge combined with that of the target users’ needs cannot and should not be replaced by blind faith in data extracted from the Internet or any other corpus.

62 | General Theory of Specialised Dictionaries

5.4 Lexicographical functions After the discussion in the previous sections it is now opportune to define the concept of lexicographical function from which the function theory is named. Such a definition should necessarily take its point of departure in the fact that a dictionary is a tool, and it should provide answers to the following questions: 1) What can a dictionary be used for? 2) Who can benefit from using it? 3) When can it provide assistance? This leads us to the following definition: A lexicographical function is the satisfaction of the specific types of punctual information need that may arise in a specific type of potential user in a specific type of extra-lexicographical situation. This definition answers the above-mentioned questions: – What can a dictionary be used for: to satisfy punctual information needs. In this way, the definition excludes both global information needs and other types of need such as reading glasses, printer cartridges for printing purposes, antivirus programs to ensure that the computer works etc., i.e. needs which must be covered in other ways. In addition, the formulation underlines that the relevant needs are not compulsory; they are possible needs that may arise in the potential user. – Who can benefit from using a dictionary: a specific type of potential user, i.e. users with specific characteristics. In this way, the definition underlines that punctual information needs are not abstract but associated with certain types of potential user. It also explains that dictionaries are not only viewed in relation to actual users, but in relation to potential users. In addition, it clarifies that these potential users have to be divided into types, because not everyone has the same types of need in the same types of situation. – When can a dictionary provide assistance: when punctual information needs occur in a specific type of extra-lexicographical situation. In this respect, the definition explains that lexicographically relevant needs are not only associated with a specific type of user, but that these users always find themselves in a specific type of situation when they experience needs of the relevant type. Functions are the heart and soul of lexicography. In the view of the function theory they determine – or ought to determine – everything that has to do with its practical products: the content and form, the data and their selection, preparation and accessibility. No decision about any aspect of specialised dictionaries should ever be taken without their respective function(s) being considered. They should, therefore, always be carefully defined and included as foundation stones in any lexicographical concept. In this regard, the above definition, together with the reflections in section 5.2, allows for a very precise determination of the function(s) to be covered by a specific specialised dictionary. A few examples of how this can be achieved are given here:

Lexicographical functions | 63

The function of the dictionary is to satisfy the communicative needs that may occur for students of specialised translation, with Danish as their mother tongue, a high proficiency level in general Spanish, a low proficiency level in the relevant specialised language and hardly any specialised knowledge of the subject, when translating texts about wind energy from Danish to Spanish. The function of the dictionary is to satisfy the punctual information needs that may occur for German students of economics with a medium proficiency level in general and specialised English, when they experiencing reception problem reading English texts about the discipline in question. The function of the dictionary is to satisfy the punctual information needs that may occur for German students of economics during their studies of this discipline. The function of the dictionary is to satisfy the punctual information needs that may occur for English-speaking experts of economics when doing advanced research related to this discipline.

Although they also depend on user characteristics, lexicographical functions are traditionally typologised and named in accordance with the basic user situation. The first two functions described above are, therefore, termed communicative, and the latter two cognitive. As already mentioned, one and the same dictionary may have either one or various functions. Many printed specialised dictionaries are multi-functional, especially when communicative functions are involved. There may be several reasons for this. The most relevant is probably of an economic nature, namely, that the publishing houses more often than not prefer to produce multifunctional dictionaries rather than to print a series of mono-functional dictionaries which may seldom be profitable, as the group of potential users of these publications is, in the majority of cases, relatively small. It is difficult to change this situation in the printed environment, but it should not be forgotten that this multifunctionality may lead to information overload – the famous concept originally popularised by Toffler (1970) – and create various unnecessary difficulties for users in terms of data access and information retrieval. Unfortunately, information overload is not restricted to printed information tools but has also been reproduced in the electronic environment. In recent years, under the aegis of the so-called information society, it has developed into an international plaque, especially when browsing and searching for information on the Internet, where the user is increasingly stressed with too many data and frequently placed in a situation similar to the search for the proverbial needle in a haystack. The relevance of the function theory in this context is precisely that it offers a vision of information tools and a corresponding methodology capable of reducing the problem, at least partially. For specialised lexicography, the main vision is to produce a new generation of online dictionaries which are either:

64 | General Theory of Specialised Dictionaries – – – –

purely mono-functional, multi-functional allowing for mono-functional data access, mono-functional allowing for individualised data access, or multi-functional allowing for individualised data access.

The two first types of specialised online dictionary represent the category of lexicographical Model T Fords, whereas the two latter are advanced lexicographical Rolls Royces (cf. chapter 6). The terminological discussion as to whether the second type should be considered one multi-functional dictionary or various mono-functional ones does not affect the basic idea which, in both cases, aims at function-based satisfaction of user needs, cf. Bergenholtz & Bergenholtz (2011) and Tarp (2011). In the light of the above reflections and with a view to improving user needs satisfaction, the function theory has formulated some fundamental principles in order to place specialised online – and other electronic – dictionaries on a more userfriendly path which takes advantage of the new techniques made available by the information and computer technologies: – Information overload should be avoided. – Users should be able to find the data required in each consultation as quickly as possible. These are the two fundamental principles which should guide the design of all dictionaries conceived to be used on electronic platforms. To these principles can be added four other basic ones: – A distinction should be made between the dictionary and the database which may even feed various dictionaries. Electronic dictionaries are not databases, but consultation tools based upon databases from which they take in the data required to meet their users’ information needs (see, for instance, Bergenholtz & Bergenholtz 2013). – The database should include as much data as possible, i.e. as much data as possible relevant to the type(s) of dictionary in question. – The specific dictionary should be able to present as much data as possible in terms of all possible consultations, i.e. the entire body of hypothetic articles resulting from these consultations. – The individual articles, namely, the dynamic data presented on the screen in each consultation, should include as little data as possible, i.e. exactly the types and amount of data needed by the user in each situation (Model T Fords) or each consultation (Rolls Royces), neither more nor less. In chapter 6, we will look at some of the techniques required to implement the above principles and little by little create new and more advanced – and more userfriendly – online dictionaries and information tools in general. Here it should only be stated that in order to make progress in this direction and choose the relevant

General methodology: an example | 65

techniques, an advanced theory is absolutely necessary. What the function theory provides in this respect is first of all a theoretical framework and methodology which makes it possible to determine the types and amount of data needed in relation to each function; consequently, this is the first step taken to avoid information overload and put “as little data as possible” at the disposal of the user and, simultaneously, to ensure that this “little data” is enough to effectively meet the user’s need in each situation. Upon this basis, and as a second step, the theory formulates guidelines in order to select the most relevant techniques to be used in each specific dictionary project. It goes without saying that this is not a question of uncritically incorporating techniques and creating eye-catching lexicographical devices with sophisticated high-tech names, as occasionally occurs in specialised lexicographical and terminological projects; cf. Fuertes-Olivera (2013a) and Tarp (2012c). Lexicography should not be replaced by lexicotainment, the pertinent term introduced by Almind, Bergenholtz & Vrang (2006). The real challenge is to design user-friendly products of increasingly higher quality in terms of the above principles. If this is to be successful, it should not be computer and information technology that takes charge of the process but rather lexicography, by means of an advanced theory capable of providing guidance and methodology. If such a theory was of great use in the printed environment, it is almost indispensable in the online environment if progress is to be made at the desired rate.

5.5 General methodology: an example In chapter 9 we will show, step by step, how a specialised online dictionary could be produced from scratch by following the methodology of the function theory. Here, though, we will anticipate this chapter by providing an example of how functional methodology can be applied to develop a general concept of a specific category of dictionary. Among the most frequent reasons why users consult specialised dictionaries are problems related to translation of specialised texts and the corresponding need to receive qualified assistance. Thousands upon thousands of people are engaged on a daily basis in specialised translation in one or another way, and almost all of them consult dictionaries as an integrated part of the activity. It is, therefore, surprising that the concept of a translation dictionary is among the most misunderstood and underdeveloped ideas within theoretical and practical lexicography; cf. Fuertes-Olivera (2013b) and Tarp (2004a, 2007b, 2013c). In the literature on lexicography there seems to be an ingrained habit of treating translation dictionaries and bilingual dictionaries as more or less synonymous terms or, at least, of considering the former as being bilingual per definition. Marello (2003: 325), for instance, defines “bilingual dictionaries only as those dictionaries which place the two languages in contact for purposes of translation”. Similar

66 | General Theory of Specialised Dictionaries ideas and approaches can be found in most of the existing literature on lexicography and translation. Here we will defend a much broader and complex concept of translation dictionary determined by the various types of lexicographically relevant need which may occur for the person engaged in translating texts from on language to another. The methodology used is that of deduction, as described in section 5.2, and combining knowledge of lexicography and specialised translation. There are basically three types of specialised translators requiring assistance from lexicography: 1) trained translators, 2) translation students, and 3) subjectfield experts performing translation. Each of these potential user types has their specific characteristics determining the particular kinds of needs they might experience in the various phases of the translation process and, consequently, the types of lexicographical data and access routes required to meet these needs. In this respect, there are several characteristics to be taken into account, of which the most important are: – Subject-field knowledge – Comparative subject-field knowledge (in culture-dependent disciplines) – Translation skills and experience – General skills in the source language – General skills in the target language – Terminological knowledge and skills in the source language – Terminological knowledge and skills in the target language – Knowledge and command of genre conventions in source language – Knowledge and command of genre conventions in target language Although there are no sharp dividing lines between them, all these characteristics can be graduated in low, intermediate and high (layman, semi-expert and expert) in accordance with the specific person engaged in specialised translation. Trained translators may be divided into two categories: those who have specialised in a certain field (e.g. accounting or legal translation), and those who are general (multi-field) translators. Both of them will, by definition, have highly developed translation skills and performance in the respective general languages, but they will differ considerably in terms of knowledge of the subject field, the relevant terminology and the corresponding genre conventions (cf. Nielsen 2010). In the first case, the translators are supposed to possess a fairly good knowledge of the terminology and genre conventions in question, and they may also have attained a subject-field knowledge which qualifies them as semi-experts. In the second case, the translators may be considered subject-field laymen, with a limited knowledge of the specific terminology and genre conventions in both languages. Students of specialised translation will probably have medium to high general proficiency levels in the respective languages, but apart from this they are basically characterised by insufficient (but increasing) translation skills combined with low knowledge of the subject field, corresponding terminology and genre conventions.

General methodology: an example | 67

Subject-field experts engaged in translation cover a broad spectrum of characteristics. They will always possess a great deal of knowledge of the terminology and genre conventions in question, though not necessarily in both languages. In addition, some of them will have developed high-level translation skills, whereas others will have skills similar to the translation students. As a rule, if a specific dictionary is designed to assist both trained translators and students, it will also cover the broad spectrum of needs occurring for the subject-field translators. As mentioned in section 5.2, the lexicographically relevant needs are determined not only by the potential user’s characteristics but also by the specific context in which they appear. As a result, the translation process has to be analysed in its entirety. Not surprisingly, modern translation science has shown translation to be a highly complex activity involving various phases and sub-phases. Lexicography, as an independent discipline, can be inspired by the theoretical and practical achievements of translation science but it cannot uncritically copy them. It is exclusively the phases and sub-phases where needs may occur that can be satisfied by means of dictionary consultation, which are of relevance to lexicography. Consequently, and although this may vary from translator to translator and from task to task, the overall translation process viewed from a lexicographical perspective comprises the following relevant phases and sub-phases: – a pre-translation phase where the translator a.

prepares the translation studying relevant background material in order to get a general idea of the subject field in question, b. reads the whole text in the source language; –

a translation phase (in the narrow sense of the word) where the translator a. reads specific text segments in the source language, b. transfers specific text segments from the source language to the target language, c. reproduces specific text segments in the target language;

a post-translation phase where the translator (or another person) a.

revises the translated text.

In all these phases and sub-phases, the translators may experience various types of need which require specific types of lexicographical data, as well as allowance for specific types of data access, in order to be satisfied. The needs may be either cognitive, when it is a question of background knowledge, or communicative when it is “merely” a question of reading, transferring, reproducing and revising the text. In this respect, it is worth emphasizing that cognitive needs may not only occur in the

68 | General Theory of Specialised Dictionaries pre-translation phase where general subject-field knowledge may be required, but also in the translation – and even the post-translation – phase, where lack of specific subject-field knowledge may derail or obstruct the process. In the pre-translation phase, the translators may need 1) a general and systematic introduction to the subject field or part of the subject field relevant to the task, 2) definitions of specific terms, and 3) background information about specific phenomena, processes and things related to the text. Here it should be noted that the corresponding lexicographical data do not require a bilingual solution to be duly accessed but can easily be provided in a monolingual source-language dictionary, a solution which is actually the best one when the source language is also the translators’ mother tongue. In the translation phase (in the narrow sense of the word), the translators may need – apart from 1) specific background information – 2) definitions of sourcelanguage terms, 3) equivalents of terms, collocations and fixed expression, as well as 4) information about orthography, gender, grammar (e.g. inflection and syntactic properties) and genre conventions in the target language. Here it is only the provision of equivalents which definitely requires a bilingual solution, whereas the latter is optional for the other data categories. Background material and definition of the source-language terms can also be furnished by a monolingual dictionary in this language, in the same way that information concerning the target language (grammar, genre conventions, etc.) can be provided either in a bilingual dictionary (from the source language to the target language), a monolingual dictionary (in the target language), or a bilingual one (based upon this language). Very often translators – and particularly experienced ones – do not experience any problems when reading and transferring the text, but they might when it comes to reproducing it in the target language (especially when this is not their mother tongue). In such cases, a bilingual dictionary from the source language to the target language would, in fact, be unable to provide assistance to the translator unless the whole translation process, starting with text reception in the source language, is reconstructed – a reconstruction which is illogical, time-consuming and counterproductive, and which lexicographers, therefore, cannot expect from their users. In the post-translation phase, text revision consists of a number of compulsory and optional sub-phases, where the translator 1) reads the source text, 2) reads the translated text, 3) compares the two texts, 4) considers whether or not there are problems in the target-language text, and 5) corrects or modifies the text, which is an activity very similar to text production or reproduction, cf. Tarp (2004b). Here the translator or person in charge of text revision may need 1) specific background information, 2) definitions in order to understand specific terms or ensure that these terms are correct, and 3) information about orthography, gender, grammar, collocations and genre conventions in the target language. Satisfying these complex needs requires a combination of lexicographical solutions. Reception-related data in the source language should be provided in either a monolingual solution in the source

General methodology: an example | 69

language (which is preferable when it is the translator’s mother tongue) or in a bilingual one based on this language (the best solution when it is not the translator’s mother tongue). A comparison of the two texts demands at least a bilingual solution from source to target language. Finally, it is a matter of course that the lexicographical data required to assist the evaluation and possible modification of the target language text must be furnished by either a monolingual or a bilingual solution, with this language as the starting-point. Before considering the radical consequences of the above reflections, we will take a look at the results of already-conducted user research in order to see whether such reflections are of substance or whether they should just be considered figments of the imagination. As mentioned in section 5.2, existing studies of user behaviour are above all suffering from a methodological debility, namely, that the lexicographically relevant user needs are not analysed directly as they occur in the translation process, but only indirectly as they are reflected in dictionary consultations or interpreted by the translators themselves. Besides, these studies are, as a rule, characterised by an excess of percentages and decimals showing how often the informants are using one dictionary or another, although such data do not have any statistical significance whatsoever due to the very small number of informants consulted in almost all existing studies. In this respect, it is worth reconsidering the relevance of quantitative research to lexicography: How useful exactly is the information that such and such a percentage of all dictionary users in such and such a percentage of their look-ups are searching for this or that? What relevance has the information that 80 percent of all look-ups have to do with orthography and semantics? Of course, commercial lexicography would be delighted and would hasten to produce dictionaries and publicity material focusing on these types of data. On the other hand, scientific lexicography would above all be interested in knowing in which situations – e.g. reception and production – these needs may occur. Then it would set itself the task of uncovering the needs users have in the last 20 percent of the look-ups, i.e. in one out of five consultations. And it would not stop here, but would try to go even deeper into the problem in order to discover the needs that only show up in one out of a hundred or one out of a thousand consultations, or, even more rarely, in order to conceive dictionaries capable of meeting all the users’ needs in specific types of situations. Whereto else, if not to dictionaries, should users direct themselves when they look for assistance to satisfy lexicographically relevant needs? (Tarp 2009b: 291-292)

These reflections do not necessarily imply that previously-conducted user studies should be disregarded wholesale. If the material provided by some of these studies is subjected to careful analysis, not from a quantitative but rather a qualitative perspective, then it is possible to confirm some of the observations made in the previous section concerning the phases and sub-phases where relevant needs may appear. We will now look briefly at a few pertinent studies carried out in relation to specialised translation. In a study undertaken more than 20 years ago, involving students translating economic texts from Danish into German, French and Spanish, respectively, Duvå,

70 | General Theory of Specialised Dictionaries Laursen & Maidahl (1992) applied the method of protocol-writing followed by interviews. The three Danish researchers concluded that lack of background knowledge of the discipline was one of the main reasons explaining the problems observed in the translation process. The informants’ difficulties in terms of equivalents were, among other things, due to four “uncertainties”, i.e. “uncertainty about the subject matter, uncertainty about the semantic content of the words, uncertainty about the words’ place in the universe, uncertainty about the usage of the words” (Duvå, Laursen & Maidahl 1992: 132). This conclusion comes as no surprise to teachers of specialised translation. Although it cannot be generalised due to the relatively low number of informants, it nevertheless confirms, as mentioned above, that at least some students need general and specific background information about the subject field when performing specialised translation. In a very interesting study carried out a few years earlier, the Polish lexicographer and translator Jerzy Tomaszczyk (1989) recorded all the consultations which he himself performed in connection with a translation from Polish into English of a book on the industrial use of diamonds. Tomaszczyk made a total of 691 consultations in 13 different monolingual and bilingual dictionaries, looking for various sorts of information. When he carried out this study, he had behind him 20 years of experience as a Polish-English translator; in other words, he had a very good command of English as well as considerable skills and experience as a translator. As a result, it is worthy of note that one third of all the consultations were made directly in an English-based dictionary. Tomaszczyk himself draws some important conclusions from his findings: Of the 691 items looked up, 373, or 54%, were more or less new to me, while in the remaining 318 cases (46%) a dictionary was consulted merely to confirm my own predictions about form, contextual appropriateness, spelling, etc. of various items. The proportion of items one would be able to retrieve from memory or simply guess successfully might be even higher if one tried harder but it is usually easier, and it takes less time, to consult a dictionary than to search one’s memory. […] This means that in a substantial number of cases one turns to a dictionary not to learn something one does not yet know, but either to make sure that what one already knows is in fact correct and appropriate in the context in which it is to appear or simply as a memory aid. […] Since what one looks up in such cases are L2 items (rather than L1 items and their L2 equivalents), it follows that in L1-L2 translations one can go a long way without a L1-L2 dictionary. This applies not only to general language problems but also to terminology, especially multiword combinations. Of the 318 predictions made, 303 (95.3% turned out to be correct. (Tomaszcyk 1989: 179)

In terms of statistical significance, the numbers and percentages obtained by Tomaszczyk are not relevant for anybody else, including maybe himself when he is translating texts outside the world of diamonds. But from a qualitative point of view, his findings confirm that translators sometimes (often?) do not experience any problems in the first two stages of the translation phase but only in the third one, and therefore “can go a long way without a L1-L2 dictionary”, i.e. a bilingual dic-

General methodology: an example | 71

tionary from the source language to the target language. This puts a dent in the side of the myth that translation into a foreign language always requires a bilingual dictionary into this language. Tomaszczyk’s experience – but with other figures – has also been confirmed by other studies. For instance, an in-depth study which the Finnish lexicographer Krista Varantola (1998) conducted in connection with four advanced translation students engaged in translating a semi-specialised text on fishing, also indicates the appropriateness of distinguishing between needs appearing in the transfer subphase and those occurring in the reproduction sub-phase. A comparison between the two studies shows that the experienced Polish translator goes straight into the foreign language more frequently than the translation students, even though the translated text on diamonds was no less difficult than the Finnish text on fishing. Similar conclusions can be drawn from a study performed by the German researcher Britta Nord, who examined the sources consulted by a group of professional translators of legal texts. Her study revealed that half of all these consultations were made in some form of bilingual dictionary, whereas in a quarter of all the cases the informants sought information in monolingual dictionaries and encyclopaedias, and in the rest in non-lexicographical sources (Nord 2002:175). What all the above studies show when they are analysed from a qualitative perspective is that both translation students and experienced translators – apart from equivalents furnished in a bilingual dictionary from the source language to the target language – at least sometimes need 1) cognitive background information and 2) terminological, linguistic and grammatical information provided by a dictionary based on the target language (whether monolingual or bilingual “in the other direction”). This supports some of the basic observations put forward above, and confirms the need for a broader concept of translation dictionary. To sum up, by applying a functional methodology we have recognised that, during the overall translation process, translators may experience cognitive and communicative needs when performing the following activities relating to the three main phases of this process: – In the pre-translation phase a. general study of the subject matter b. text reception in the source language c. specific study of a topic related to the subject matter –

In the translation phase a. text reception in the source language b. text transfer from source language to target language c. text reproduction in the target language i.

with problems in the previous sub-phases

72 | General Theory of Specialised Dictionaries ii. without problems in the previous sub-phases d. specific study of a topic related to the subject matter –

In the post-translation phase a.

text revision in the target language i. ii. iii. iv. v.

text reception in the source language text reception in the target language text comparison between source and target language text evaluation in the target language text reproduction in the target language

b. specific study of a topic related to the subject matter If a dictionary is supposed to give real assistance to translators of specialised texts, then it should be designed to meet all the lexicographically relevant needs appearing in the various phases and sub-phases of the overall translation process. This is not only a question of incorporating the right data into the dictionary but also of guaranteeing the easiest – and sometimes only possible – means of accessing these data. Here we will not go into detail regarding the relevant data types but will focus mainly on the last aspect, which has to do with the overall design of the dictionary. In this respect, there are many similarities but also some important differences in terms of the best way to satisfy the user’s needs in L1-L2 and L2-L1 translation, respectively. – Most translators, especially those who are not subject-field experts, would prefer that the systematic introduction to the subject field is written in their mother tongue although it could also be provided in both languages simultaneously. This is valid for both L1-L2 and L2-L1 translation, and it does not matter if the “outer” text is placed in connection with a monolingual or a bilingual solution. – The same holds true for the definitions of specific terms found in the source text as well as background explications of specific phenomena required in both the pre-translation and the translation phase. Here, the best option is to provide these data in a monolingual solution in relation to L1-L2 translation and in a bilingual one in relation to L2-L1 translation. The second best option would be a bilingual solution in the first case and a monolingual one in the second case, respectively. – Equivalents of terms, collocations and fixed expression should necessarily be furnished in a bilingual solution from source language to target language for translation in both directions.

General methodology: an example | 73

As it has been explained above, the need for information about orthography, gender, grammar and genre conventions in the target language may appear both when problems have occurred and not occurred in the previous transfer sub-phase. In the first case, the best option would be to place the corresponding lexicographical data in a bilingual solution from source language to target language, whereas it, in the second case, is indispensable to place these data in a monolingual (or bilingual) solution based on the target language. In order to be helpful, the lexicographical data required to meet the user’s needs in the post-translation phase should necessarily be furnished in various types of monolingual and bilingual solutions. Data related to text reception in the source language can be provided in the same solutions as mentioned above (point 2). By analogue, the data related to text comparison in the source language should also be provided in the same sort of bilingual solution as mentioned above (point 3). Finally, the data related to text reception, evaluation and reproduction in the target language should necessarily be provided in either a monolingual solution in the target language (the best option in L2-L1 translation) or a bilingual one taking the point of departure in this language (the best option in L1-L2 translation).

These reflections allow us to decide on the best overall design for specialised translation dictionaries from and into the translator’s mother tongue. Hence, the best overall design of a dictionary conceived to assist its users in L1-L2 translation of specialised texts consists of the following three components: – a monolingual L1 component – a bilingual L1-L2 component – a bilingual L2-L1 component Similarly, the best overall design of a dictionary conceived to assist its users in L2-L1 translation of specialised texts consists of the following two components: – a bilingual L2-L1 component – a monolingual L1 component In a printed environment, the ideal solution would be to print a series of three dictionaries for L1-L2 translation and another series of two dictionaries for L2-L1translation. However, this is seldom feasible. Due to the relatively small number of potential users of specialised translations dictionaries within most subject fields – and language pairs – this would not be economically attractive for any publishing house unless the project is carried out with public funding. In such cases, a pragmatic – but theory-based – solution would be to opt for the second best overall design as it has also been discussed above, namely, a combination of L1-L2 and L2-L1 components for L1-L2 translation and a combination of L2-L1 and L1-L2 components for L2-L1 translation.

74 | General Theory of Specialised Dictionaries With only a few minor lexicographical adjustments and additions this second best solution could be materialised in a combined L1-L2/L2-L1 dictionary conceived to assist users with both mother tongues performing translations in both language directions. The Hungarian-German, German-Hungarian Fachwörterbuch zur Rentenversicherung as well as the English-Spanish Encyclopedic Dictionary of Gene Technology and its Spanish-English counterpart Diccionario Enciclopédico de Ingeniería Genética, are all examples which show that such a solution may not only be viable but also of a high quality, cf. Bergenholtz, Tarp & Kaufmann (1994), Tarp (2005), and Fata (2009). However, in spite of the undeniable merits of these and similar dictionaries, it is no secret that they suffer from two main problems which cannot be solved satisfactorily in the printed environment. The first of these problems is the additional lookups which the users frequently have to make in other parts of the dictionary in order to find the required data. The second has to do with information overload, in the sense that the translators in most specific consultations may encounter a certain amount of superfluous data which is irrelevant to their specific purpose. In both cases this may delay the consultation process and obstruct the proper retrieval of the information needed in each specific consultation. This is an inevitable problem in printed dictionaries, which, however, can be easily avoided in the online environment, especially in the lexicographical Model T Fords and Rolls Royces. The Danish, Danish-English, English and English-Danish Accounting Dictionaries and the English-Spanish, Spanish and Spanish-English Diccionarios de Contabilidad are concrete examples showing how the Model T Ford-principles can be materialised (cf. section 9). In the latter (still under construction), the users are mainly native Spanish-speaking translators and other people writing and reading about accounting, cf. Fuertes-Olivera & Nielsen (2012). These users are initially given the option to access either a Spanish, a Spanish-English or an English-Spanish part, where they are offered the following “search methods” which almost instantaneously direct them to the data adapted to their specific activity:

General methodology: an example | 75




Illustration 1: Search methods offered in the Diccionarios de Contabilidad

In other Model T Fords it would be advantageous for the users to initially indicate their mother tongue, activity, and other relevant characteristics by means of interactive techniques (see chapter 6). This done, the above reflections indicate that the user should at least be offered the following activity-orientated search options when starting a consultation based on either L1 or L2 in relation to L1-L2 or L1-L2 translation, respectively:

76 | General Theory of Specialised Dictionaries Schema 1: Possible search options in a lexicographical Model T Ford


L1-L2 translation

L2-L1 translation


In L1

Cognition (L1)

Cognition (L1)

Cognition (L1)

Reception (L1)

Reception (L1)

Reception (L1)

Reproduction (L1)

Reproduction (L1)

In L2

Transfer (L1-L2)

Transfer (L1-L2)

Reproduction (L2-L1)

Reproduction (L2-L1)

Cognition (L2-L1)

Cognition (L2-L1)

Cognition (L2-L1)

Reception (L2-L1)

Reception (L2-L1)

Reception (L2-L1)

Transfer (L2-L1)

Transfer (L2-L1)

Schema 1 also illustrates how relatively easy it is to combine a dictionary for L1-L2 translation with one for L2-L1 translation as the two of them share most search options giving access to activity-related data types. If the lexicographers want to take further steps towards the lexicographical Rolls Royce, there are several available techniques which can be applied in order to further individualise the online dictionary, for instance mouse-sensitive areas, pop-up windows, hypertexts, annotations, links to Web pages with additional data, individual article modelling (see chapter 6). In this respect, it is important to stress that there does not and cannot exist any single detailed design of a specialised translation dictionary, or of any other general category of dictionary. Each one has to be designed according to the specific subject field, language pair, user group and user needs in question, and with consideration given to the medium in which it is to be published, cf. Fuertes-Olivera (2013b). When entering into detail to decide on the specific data categories to be included in the dictionary in terms of communicative and cognitive needs, it is necessary to possess

Prescription, description or proscription? | 77

not only expert knowledge of lexicography and specialised translation but also knowledge of the subject field to be covered; all of this will contribute to giving each dictionary its individual stamp with specific data types and access routes (see chapter 9). Even so, there are some common features regarding the overall design of a specialised dictionary for translation purposes, in the same way that there are common features for other general categories of dictionaries. In terms of translation dictionaries these have been discussed above, and they are valid for most dictionaries conceived with this aim.

5.6 Prescription, description or proscription? The selection and preparation of data to a dictionary is never performed in a vacuum. It always takes place in an environment where several official or semi-official bodies, in one way or another, pretend to regulate the language, whether specialised or general. As a result, when preparing the concept of a specialised online dictionary, the lexicographers in charge of this task have to decide the attitude to be taken towards the various language policies relevant to the specific subject field treated in the dictionary. Any lexicographical work is, by definition, an important player in the promulgation of a given language policy whichever body has laid it down. In fact, Bergenholtz & Gouws (2006) write: Every single lexicographical decision has a language policy relevance and therefore, in the end, a political dimension. (Bergenholtz & Gouws 2006: 14)

This statement may seem exaggerated as there are obviously several lexicographical decisions, e.g. about the structure and design of the data appearing on the screen, which do not have any direct relation to language policy. However, when it comes to the terms and other linguistic data included in the dictionary, these will always, to one degree or another, influence the linguistic behaviour of the factual users and therefore have “a language policy relevance” as stressed by the two lexicographers. This fact puts a great responsibility on the shoulders of the lexicographers engaged in both the conception and production of any specialised dictionary. There are several different definitions of language policy, cf. Bergenholtz & Tarp (2007). Here it is used as an umbrella term, referring to all the different levels of decisions and their implementation, aimed at regulating various aspects of language use, especially with regard to the formal recognition of words and terms as well as their spelling and inflection. Lexicographers have, on the one hand, to take cognisance of the different forms of language policy relevant to the dictionary being prepared and, on the other, to negotiate the best ways to ensure that the functions of this dictionary can be achieved. This implies that even if lexicographers obey the rules prescribed by formal bodies in the formulation of their language policies, the

78 | General Theory of Specialised Dictionaries particular user needs and functions of the dictionary may at times demand the inclusion of non-prescribed forms. There are a variety of language policies which, in one way or another, are relevant to specialised lexicography. Following the typology proposed by Gouws & Tarp (2008), they are here grouped into three main categories, i.e. national language policy, domain-specific language policy, and terminological language policy. The term national language policy refers to the intended regulation regarding use within a given speech community, be this community only one among others within a specific country (such as some of the national languages in Spain), the only community in a specific country (like Denmark), or a cross-border community like the Germanspeaking population in Germany, Austria and Switzerland. The national language policy is normally laid down by a certain official or semi-official body such as a national language board or an academy, and it generally prescribes correct spelling and inflection forms and occasionally pronunciation. In some language communities (like in Iceland and the Faroe Islands) it even prescribes the words to be used. The boards or academies responsible for the national language policy may wield different degrees of authority, and there are even speech communities with competing authorities in terms of language policy. The term domain-specific language policy refers here to the language policy laid down by “intermediate” entities such as companies, ministries, universities, local governments, and various kinds of organisations. This type of language policy may regulate the language, or languages, to be used within their sphere of influence, for example, inside a company or in external communication. Additionally, this particular language policy most often regulates the style to be used internally and externally in the body in question. It is normally subordinated to the national language policy, although it may prescribe specific terms, words and forms to be used, which in some cases even go against recommendations contained in the national language policy. Finally, the term terminological language policy refers to the regulation of terminology within one or several specialised subject fields. This policy may be decided by a national or regional terminological board or by separate organisations, companies or other entities. As a result, it is sometimes interwoven into the domainspecific language policy and even on occasions with the national language policy. The regulation of terminology normally embraces the selection of the recommended terms and their definition. As can be seen, the three types of language policy mentioned cover different areas and regulate different aspects of language use, although they may overlap to a certain extent. This has to be taken into account when planning and compiling different types of lexicographical work. Consequently, specialised lexicographers (or terminologists) dealing with dictionaries covering a specific subject field have to relate their work to the relevant terminological language policy. Similarly, those who are working with company, branch and similar dictionaries need to appreciate

Prescription, description or proscription? | 79

the domain-specific language policy wherever this is appropriate for their work. Likewise, all lexicographers, in one way or another, have to relate to the national language policy. Whilst it is important to relate to the language policy at a macro level, it is also important to determine the character of this relationship. However, before proceeding to this discussion, it is necessary to make a brief incursion into the field of methodology. Prescription, description and proscription represent different methods of collecting and utilising data from different sources, such as corpora, the Internet, linguistic surveys, text investigations, etc., cf. Bergenholtz (2003). In the interaction between language policy and lexicography, the principles of prescription, description and proscription play an important role. Regarding prescription, terminological language boards and other standardisation bodies are, by their very nature, prescriptive. They prescribe, for instance, which terms should be used within a given discipline, branch or at the national level, and how they should be spelled, inflected and defined. Prescription is not necessarily the acknowledgement of only a single form. Quite often a language body officially recognises different variants, e.g. orthographic variants of a single word. Prescription comes very much to the fore in dictionaries. When taking a prescriptive approach, lexicographers impose their point of view on the target users of the dictionary. Many users actually want this kind of guidance, especially when consulting a dictionary for text production, text revision or translation purposes. In these situations, and in order to make the consultation as quick and easy as possible, lexicographers should not let the users themselves choose among a large number of competing terms, words and forms, but should present them with a single correct term, word or form. According to Bergenholtz (2003), different forms and degrees of prescription can be distinguished. Here it suffices to say that, especially in the case of a strong prescriptive approach, it can be regarded as either presenting a single form (with regard to, for example, orthography, pronunciation, meaning or morphological possibilities), or more than one form, as the preferred form(s) of the dictionary without any reference to other words from non-standardised use. This hard-line prescription implies that only these forms and words should be allowed, as they are correct, whilst their variants or other words should be prohibited. When following a prescriptive approach, the lexicographer places himself in the position of having to judge the language and make a decision regarding the accepted terms, words and forms. In many instances a prescriptive dictionary will follow the rules laid down by a formal language body or entity. It will then function as an extension of and an instrument at the disposal of this standardisation authority, the prescription of the relevant body as a result being presented in the dictionary. The average dictionary user does not distinguish between different language bodies and academies, but regards the dictionary as the embodiment of authority. Moreover, just as domainspecific language policy sometimes runs counter to prescribing the national lan-

80 | General Theory of Specialised Dictionaries guage policy, lexicographic prescription does not necessarily always adhere to official rules. It also prevails when lexicographers, providing a single form and ignoring all other variants, give forms which deviate from the official language rules with regard to, for example, spelling, morphology, etc. In such a case, the dictionary is not an extension of the language body, but instead it follows its own prescriptive methods. Description does not only reflect the decisions of an official body or the implementation of an official language policy on any of the levels discussed above. Rather, it endeavours to give a comprehensive account of actual language use by presenting a variety of terms and forms, whether orthographic, morphological or pronunciation variants. Description avoids classifying occurring forms as either recommended or not permissible. It reflects the spectrum of actual language use. Description plays a significant role in some dictionaries. A descriptive approach in lexicography sees the lexicographer trying to reflect real language use, making provision for different variants but not indicating a recommended form or labelling a given form as not permissible. Bergenholtz (2003) refers to different forms and degrees of description. Depending on the functions of a dictionary, a descriptive approach can either frustrate or please the users. In a dictionary compiled for text reception, a presentation of all the different variants can assist users in an effective way. Yet those who consult a dictionary for text production or translation are often frustrated by a descriptive approach, because they do not find explicit guidance regarding the best, the appropriate, or the most correct form. Having to make choices and being confronted with variants is not what they expect from a dictionary. Owing to the fact that users see dictionaries as authoritative sources, even description can be interpreted by the average user as a presentation of the correct forms. In this respect, Wiegand (1986) refers to the normative force of descriptive dictionaries. Although many users do not distinguish between the various methods used to select data for their dictionaries, lexicographers need to take a firm decision regarding the approach to follow in their practical work. This decision must take cognisance of a range of implications that the method chosen may have on users. So far, much has been written about description and prescription in terms of lexicographic works. Different authors have argued in favour of one method or another as relevant in dictionary making. Some of these arguments are convincing but generally speaking the discussion has been based upon non-lexicographical principles and the authors’ own subjective preferences, with no visible relation to the users’ needs and the respective function(s) of the dictionaries. By contrast, the function theory always has as its basis the user’s needs and functions when determining the specific methods to be employed in concrete dictionary projects. With this in mind, and by applying the basic principle of methodology described at the end of section 3.1, this theory has critically subjected these two methods to the requirements of lexicography

Prescription, description or proscription? | 81

as an independent discipline, thereby developing a third method which is supplementary to them, namely, proscription. The term proscription, as it is introduced and used by Bergenholtz (2001, 2003), originates from the Latin word proscribere (to make public) and not from the modern English word proscribe (to forbid). In a proscriptive approach, the lexicographer wants to inform the user not only about language use but also about the form recommended by the lexicographer. In lexicography, a proscriptive approach sees the lexicographer deviating from the prescriptive way of saying “this should be done” and saying instead “this is recommended”. A proscriptive approach often recommends a single term or form, but it may also give different variants or include different terms, accompanied by a clear indication of the term or form the lexicographer recommends. In some specific cases, it could also lead to the recommendation of two or more terms or forms. This might, for instance, be the case when these terms and forms appear in a particular corpus with the same frequency, or when new terms are introduced into a given language and the future will decide whether, let us say, a simple loan, transliterated or coined term will prevail in the speech community. However, even if one or more recommended terms or forms are given, it does not imply that these are the only ones permissible. It merely represents the lexicographer’s recommendation and does not necessarily reflect the decisions of a formal language body. In his classification of different types of proscription, Bergenholtz (2003: 13) differentiates, among others, between exact proscription (where only one variant is recommended though other variants may be mentioned) and not exact proscription (where more than one variant is recommended but other variants may be mentioned). Although he refers to the possibility of other variants being mentioned, he does not distinguish a type of proscription based on the reference, or absence of it, to non-recommended forms. Such a distinction may be useful because it contributes to the contextualization, which often has an influence on the choices a user makes for a given utterance. In connection with this, Gouws & Tarp (2008:239) have proposed the dichotomy single versus complementary proscription when referring to this distinction. A procedure of single proscription sees only the recommended terms and form(s) included in the dictionary, whereas these are complemented by their non-recommended variants in complementary proscription. Here we will discuss three examples where there is disagreement either between the prescription laid down by different language bodies, or between the lexicographers and the relevant language policy. The first example is from Denmark, where the national Language Board – due to its general language policy regarding commonly used acronyms – accepts both dna and DNA, i.e. written with both small and capital letters, as the acronyms for deoxyribonucleic acid. The following two articles from Retskrivningsordbogen (Danish Orthographic Dictionary) – which, according to national law, is the official mouthpiece in which the Danish Language Board publishes its language policy decisions – show how this principle is implemented:

82 | General Theory of Specialised Dictionaries DNA (or dna) sb., DNA’en or DNA’et, in composita DNA-, e.g. DNA-molekyle, DNA-sekventering dna (or DNA) sb., dna’en or dna’et, in composita dna-, e.g. dna-molekyle, dna-sekventering Example 1: Two translated articles from Retskrivningsordbogen

The term dna spelled with small letters is, however, contrary to language use within scientific disciplines like molecular biology where DNA with capital letters is the internationally accepted and recommended acronym. This difference between two different language policies could obviously create confusion among dictionary users, especially those who have to choose the most appropriate variant when engaged in text production or translation. As can be seen, the dictionary also furnishes two different inflection paradigms (but without plural), thus adding fuel to the fire. This unnecessary and time-consuming confusion can easily be solved by a proscriptive approach, for instance the one chosen by Den Danske Netordbog (The Danish Online Dictionary) which recommends both a single spelling form (DNA) and a single inflection paradigm (incl. plural) while simultaneously mentioning the nonrecommended forms, especially for the benefit of users with text reception problems: DNA abbreviation also, but not recommended dna abbreviation also, but not recommended This spelling form is not recommended, use instead –>DNA Example 2: Two translated articles from Den Danske Netordbog

With a proscriptive approach as the one implemented above, the user does not need to waste time in choosing the appropriate spelling variant and inflection form. The following example has also to do with DNA but with a different problematic. As information tools specialised dictionaries need to convey the type of data determined by their functions and directed at meeting the real needs of the intended target user group. Consequently, they should not only be regarded as showcases of language prescription but as utility tools they should include all data necessary to fulfil their specific functions, even if this goes against or beyond the prevailing prescriptive vision. Although lexicographers have to take cognisance of the decisions of language bodies and, as a rule, should not endeavour to position their dictionaries in opposition to these decisions, the principles guiding these decisions may always be a dominating force when selecting data for inclusion in their dictionaries. In this respect, giving quick and easy access to relevant data should be regarded as one of the most important tasks of a lexicographer. Access, however, is not guaranteed if the lexicographer regards specific data in isolation. Contextualisation of data is paramount and in lexicographical works this contextualisation often implies that a word should not be isolated from its variants. Access to a given word can

Prescription, description or proscription? | 83

often best be achieved via a variant known to the users. In their report from a workshop where the concept for a dictionary of gene technology was decided, Bergenholtz, Tarp & Kaufmann (1994) refer to a discussion where a purely prescriptive approach would have impeded user access to specific entries. Let us take a look at the case. The Spanish equivalent for the English form DNA is ADN (ácido desoxirribonucleico). However, evidence from a Spanish corpus showed that in 40 per cent of the instances the English form DNA was used in Spanish scientific texts. Although the Spanish equivalent ADN may be both the recommended and the prescribed form, successful access to this form will often depend on the inclusion of the English loan form DNA as a separate lemma, reference to which can be restricted to a crossreference entry, guiding the user to ADN. A purely prescriptive approach would have led to the inclusion of only the prescribed form ADN. The omission of the loan form DNA would have prevented many users from accessing the prescribed form, especially when they had problems relating to text reception or translation of Spanish texts using this loan form. The importance of including both ADN and DNA in the dictionary could also be recommendable on practical grounds. If only the prescribed form were included, a user who was unfamiliar with the other form might encounter the omitted form (DNA) in the literature. With so many different acronyms are being used in the scientific field, this user might regard the form encountered as a wholly different acronym not related to the one included in the dictionary. Another example where a prescriptive approach clashes with user needs in terms of access, is taken from the South African wine industry. Here, Van der Merwe (2008) mentions a number of terms which are frequently used by ordinary people but not accepted as being correct according to the terminology laid down by the South African wine industry. One of these terms is the English drinking wine (drinkwyn in Afrikaans), which in the official wine terminology has been replaced by the term wine, despite the fact that many habitual wine drinkers still use drinking wine. Another example is tapvat, which is the official Afrikaans equivalent for the English term boxed wine. In this case, most ordinary people will use the colloquial bokswyn, which, however, also refers to a lesser quality wine and is therefore not allowed by the industry as an Afrikaans equivalent for boxed wine. A last example is the popular word champagne, which the European Union has now restricted only to products coming from the French district of Champagne. The official term used for wines produced by the same method, e.g. the Cap Classique, but not originating in the French district is sparkling wine, although champagne is still widely used in colloquial language. Thus, if drinkwyn, bokswyn and champagne are not included in a wine dictionary, at least as cross-reference lemmata, then many users of this dictionary will probably be unable to find the correct terms prescribed by the wine industry. In spite of this, in the final version of the South African Trilingual Wine Industry Dictionary one finds that only the term drinkwyn has been admitted, with the information that it is “superseded” and cross-referenced to wine, while bokswyn

84 | General Theory of Specialised Dictionaries is completely omitted and champagne only accepted as originally a French product; neither is there any reference to the prescribed South African term sparkling wine. Consequently, the dictionary is not especially helpful to its intended users, among them “wine lovers”, many of whom will probably use the colloquial terms without knowing that they are not accepted by the wine industry. The conclusion, therefore, is that lexicographers should look not only at the prescribed or recommended term but also at variants that are not allowed; in other words, a “descriptive” approach should prevail when they are selecting the data needed to ensure optimal access to the relevant dictionary article. In spite of prescription by terminological and other language bodies, lexicographers should realise that their users need access to relevant data. Furthermore, even if the lexicographer agrees with the decisions of a prescribing body, the lexicographic presentation should make allowance for different points of departure and different access routes for different users. Many users may try accessing the dictionary via a non-prescribed form which is possibly the only one with which they are familiar. In such cases, a strict adherence to the prescriptive principle would result in unsuccessful dictionary consultation. The inclusion of as many relevant forms as possible should ensure that access to the recommended form, which might or might not be the prescribed one, is successful, as more than one possible access route is employed. Relevant data presentation, the best possible access to the data, optimal retrieval of information, and compliance with the lexicographic functions of a dictionary, are paramount for ensuring a successful consultation process. To achieve this, lexicographers should not merely abide by the results of prescription. Specialised dictionaries could, we believe, do better by using the proscriptive method, thereby offering a wider selection of forms, complemented by well-motivated recommendations, in order to guarantee their success as consultation and information tools.

5.7 The lexicographical process: the lexicographer’s perspective As defined in section 3.1, the study object of lexicography does not only embrace dictionaries as such but also the processes related to their production, design and use. In this and the following section we will look more closely at these processes and some relevant aspects related to them. Basically, lexicographical processes can be seen from two different perspectives: that of the lexicographer and that of the user. The two types of process are interwoven, with various intersections and meeting points relevant both to the theory and practical work. In this section we will briefly discuss the process from the lexicographer’s point of view, i.e. the various phases and steps in the conception, production and publication of dictionaries. The following schema shows the three relevant phases and the main steps related to each phase:

The lexicographical process: the lexicographer’s perspective | 85

a pre-compilation phase where the lexicographers: a. observe one or more potential users in a specific context or situation, b. detect lexicographically relevant information needs among these people, c. and decide to start a lexicographical project in order to satisfy their needs;

compilation phase where the lexicographers: a. b. c. d. e. f. g. h. i.

design a dictionary concept (functions, data categories, access, etc.), compose a team of relevant subject-field collaborators, prepare the material needed for the dictionary project, prepare the computer program (fields, their number, content, relation, etc.), write lexicographical instructions, elaborate and evaluate a number of test articles, create a detailed plan with tasks and deadlines for the practical work, carry out the project in all its practical aspects, and prepare the dictionary for its publication on the chosen media;

a post-compilation phase where the lexicographers: a. put the dictionary at the disposal of the foreseen user group, b. observe how the users use of the dictionary, c. and check whether or not the users’ needs have been satisfied.

In the previous sections we have discussed some of the complex problems related to the pre- and post-compilation phases. In chapter 9, we will go into a much more detailed discussion of the various steps to be taken in the compilation phase in regard to specialised online dictionaries. Here we will just draw attention to three relevant questions: interdisciplinarity, production time, and production costs. Regarding the first question, the above schema clearly indicates that although lexicography may rightly be considered an independent discipline, the whole compilation, in all its aspects, of any specialised online dictionary is by its very nature an interdisciplinary endeavour. In the preparation of such a dictionary, none of the steps included in the compilation phase above should, as a rule, be carried out exclusively by experts in lexicography, but always in collaboration with the relevant subject-field experts, in terms of both the specific subject field dealt with in the dictionary as well as the necessary technical support provided by programmers, designers, etc. The second question is strongly defended by supporters of the function theory. It has to do with the overall production plan, where it is of the utmost necessity to establish realistic and compulsory deadlines for all the tasks to be accomplished, in order that these specialised lexicographical projects are completed within a reason-

86 | General Theory of Specialised Dictionaries able time limit, which, in most cases, should not exceed two years. This is particularly important for those projects which, in one way or another, cover one of the many disciplines being continuously developed and transformed due to new scientific discoveries, technological applications or laws or regulations, such as molecular biology, computer science, or EU legislation. The fact is that many specialised lexicographical and terminological projects drag on for years, frequently without ever being finished or, when they finally are finished, ending up with a lot of outdated and, therefore, more or less useless information. In this respect, FuertesOlivera (2013a) writes: The internet has allowed the compilation of new types of information tools, for example, the so-called terminological knowledge bases. These are proliferating around the world, especially because they obtain public money easily, although most of them do not deliver much. For instance, around 90% of the terminological dictionary projects funded by the Spanish R+D funding agency are still prototypes after several years of continuous and generous funding. (Fuertes-Olivera 2013a: 337)

The third question is related to the second, and involves the use of human, financial and other resources in the production of specialised online dictionaries. It is crucial that the projects and production plans should be meticulously prepared with a view to making full and rational use of the available resources. This requires an efficient methodology, aimed at compilation by means of as few resources as possible, for instance, one based on the function theory, as described in sections 5.2 and 5.3 in relation to the conception and preparation phase. This methodology should also include the important principle that, once the preparatory phase has been completed and compilation as such has begun, no modification of the chosen concept that implies changes of already-completed work should be allowed, even if there are good arguments for such changes. New data categories not affecting alreadyconducted work may, of course, be permitted, but any conceptual change with retrospective consequences should be postponed until a future edition. However, and as the above quotation also shows, many specialised lexicographical or terminological projects – regardless of their name – are, in our view, extravagant and waste a lot of resources in the production of “prototypes” and other poor results. We believe that this can never be the future of specialised online lexicography, where the number of information tools needed within the different branches and disciplines generally vary inversely in respect of the resources available. In this regard, the production costs in terms of the human, financial and other resources employed in lexicographical projects (referred to by Nielsen 2008 as “lexicographical information costs”) come to represent an important criterion for lexicographical success and quality.

The lexicographical process: the user’s perspective | 87

5.8 The lexicographical process: the user’s perspective The previous section dealt with the lexicographical process from the lexicographer’s point of view, i.e. the way dictionaries are conceived and produced. Yet, as the basis of all lexicographical work is ultimately the user’s needs, then the above process can only be fully understood if it is taken into account from the user’s perspective, as the two types of process are interwoven and share several intersection points. As has already been indicated, the function theory does not only study the process taking place from the moment the actual user starts consulting a lexicographical tool to the time when the required information is retrieved from the data contained in this tool, i.e. the intra-lexicographical consultation process. It is also interested in the extra-lexicographical processes taking place immediately before and after consultation. The reason for this approach is two-fold: on the one hand, it is necessary to know in which situation the lexicographically-relevant information needs occur in order to determine the nature of these needs, and, on the other, it is absolutely necessary to evaluate the post-lexicographical process to establish an objective criterion for success or failure rather than the subjective type used by questionnaires and the like. In this respect, and according to the function theory, a “normal” lexicographical process from the user’s point of view comprises the following phases and steps: – an extra-lexicographical pre-consultation phase where a user with specific characteristics finding him or herself in a specific context or situation: a. experiences an information need, b. becomes aware of the information need, c. and decides to start a lexicographical consultation; –

an intra-lexicographical consultation phase where the user: a. selects the relevant lexicographical information tool, b. accesses the relevant data, c. verifies that he or she has found the right data, i.e. relevant to the information need in question, d. and retrieves the needed information from the data;

an extra-lexicographical post-consultation phase where the user: a.

makes use of the retrieved information in order to solve a communicative or cognitive problem, to store it as knowledge, to perform a task or to interpret a sign, signal, symbol etc.

88 | General Theory of Specialised Dictionaries Lexicography, as the science of dictionaries and similar information tools, is supposed to study all these phases and steps. However, if we look at this overall lexicographical process and compare it with the user studies published, for instance, those analysed in Welker (2006), the hitherto sole systematic overview of such studies, we will find that most of them deal with the various steps in the intralexicographical consultation phase. As far as we know, there is no published user study dealing systematically with the extra-lexicographical pre- and post-consultation phases. There are, of course, retrospective studies where users, by means of questionnaires, interviews and similar methods, have been asked about their needs prior to consultation and whether they are satisfied with the results of this consultation. However, studies of users in the middle of one of these phases are, to our knowledge, non-existent. There could be several reasons for this gap in user research. An important explication may be that research into the pre-consultation phase is very complex and costly, especially if the results are expected to be statistically significant. This is the reason why the function theory recommends alternative methods – such as the ones discussed in section 5.2 – in order to establish the user needs upon which the theory, as well as its practical products, is built. As for research into the usefulness of the information retrieved and applied in the post-consultation phase, this would, although still time-consuming, probably be in the majority of cases less complex and expensive than research into the preconsultation phase. Of course, one research method would entail feedback from the users, but this, although useful in many aspects and relatively easy to handle in the online environment, is always biased to a certain degree, as it is based upon the users’ own subjective interpretation rather than on objective criteria. It should, therefore, be complemented with other methods, such as tests. It is well known that several such studies have been conducted, for instance, that of Caruso & De Meo (2013). Yet although this and other similar studies provide interesting and relevant information on problems experienced by users of specialised online dictionaries when translating specialised texts, the relatively small number of informants (in this particular case only 39) implies that they cannot be considered statistically significant and, hence, generalizable. Given the importance of this question, it is surprising that systematic research into the post-consultation phase has not yet been conducted; it is, after all, in this phase where we find the proof of the pudding, which, in the last resort, may prove the relevance of a lexicographical theory as well as the quality of specific dictionaries (cf. section 4.3). In spite of the many deficiencies characterising lexicographical user research in general, various examples of this research based upon scientific methods can be found in relation to the intra-lexicographical consultation phase, e.g. Nesi (2000), Tono (2001), Wingate (2002), Thumb (2004), and Lew (2004). By contrast, it seems as if most of the research conducted in relation to the pre- and post-consultation phases has been carried out according to the methods of commercial market analy-

The lexicographical process: the user’s perspective | 89

sis, based upon user expectations and even artificially created needs rather than upon knowledge of real needs; without consideration of such needs lexicography may still sell its practical products and gain profit but it will never become a scientific discipline, cf. Tarp (2009b). The different opinions regarding the academic status of lexicography may be influenced by this fact. When this is said, it is important to emphasize that the above ideal process presupposes that the user is actually aware of his or her information need and decides to take lexicographical action. In fact, when a lexicographically relevant information need occurs for an individual in any extra-lexicographical situation, this individual is automatically transformed into a potential user which may: – not be aware of the information need and therefore not proceed to any lexicographical consultation; – be aware of the information need but not proceed to any lexicographical consultation because he or she thinks – maybe based upon previous lexicographical experience – that the need cannot be solved by means of a lexicographical consultation or that this consultation may require too much time and trouble; – be aware of the information need and proceed to a lexicographical consultation but with a wrong idea and understanding of the real nature of this need; or – be aware of the real nature of the information need and proceed to a lexicographical consultation. In an ideal world, lexicographical tools should be able to meet the user’s information needs in all four cases, not only for the actual user as in case 4. In case 3, for instance, a lexicographical tool could, by means of various advanced interactive techniques and methods, guide the user in such a way that they will finally obtain the appropriate information required and not “the right answer to the wrong question” (cf. Zikmund 1997: 96). Nevertheless, even in cases 1 and 2, where the potential users do not, for whatever reason, make a consultation, there are lexicographical tools already available with solutions for the information needs occurring in specific types of situation in a digital environment. In this case, the above model will have to be transformed as follows: – an extra-lexicographical phase where the information need occurs but where the potential user for one reason or another does not take lexicographical action; – an intra-lexicographical phase where the lexicographical tool: a. detects the problem, b. and suggests a solution; –

an extra-lexicographical process where the user accepts the proposed solution and uses it in order to solve the hitherto unknown problem.

90 | General Theory of Specialised Dictionaries As we mentioned, such advanced lexicographical tools already exist, mainly in relation to text production and translation. Among them the most well-known is probably Word’s Spelling and Grammar Checker which, when activated, detects problems and suggests solutions regarding written text production in a digital environment. In such cases, the intra-lexicographical process takes only a few seconds or even less. This should be compared with the above-mentioned process, where the user himself has to take lexicographical action, which frequently may take several minutes. If, for instance, a professional translator needs to make 50 lexicographical consultations in a normal working day, and these take an average of 5 minutes, this works out at 250 minutes, or more than four hours, of consultation; for most people this would represent more than half a working day. This constitutes some rather expensive production costs for the translators and, as such, a terrible waste of time. In this respect, the time factor – translated into quick and easy data access and information retrieval – becomes another important criterion of lexicographical quality and relevance, as also happens in the production phase of specialised online dictionaries mentioned in section 5.7.

6 Special Problems Related to Online Dictionaries Since the year 2000, the Internet has increasingly consolidated itself as the dominant platform for electronic dictionaries. This development was initiated several years earlier with the introduction of computer and information technologies into the field of lexicography, a process which De Schryver (2003) dates back to the late 1960s. In these more than four decades lexicography has undergone a true revolution in most of its core expressions, from the selection process, through the production phase, to the publication and presentation of the final product to the users on some kind of electronic platform. But in a certain sense this revolution has stopped halfway. When a deep-rooted millenarian cultural practice like lexicography passes from one technology to another, one would expect such a gigantic step to be more than a mere change of production methods and medium; it would also be logical for it to involve improvements in terms of quality, expressed in a more accurate and personalized satisfying of user needs which has become possible thanks to the new technologies. As could be expected, the many variants of electronic dictionary have been accompanied by a considerable body of academic reflection. Among the large number of studies published during the past few years can be mentioned Haß & Schmitz (2010), Granger & Paquot (2010, 2012), Kosem & Kosem (2011), and Fuertes-Olivera & Bergenholtz (2011a). This literature undoubtedly contains many relevant ideas, yet it is disappointing to see how most authors focus on practical problems related to natural language processing and, to a lesser degree, on use of the available techniques to speed up access (Faster Horses), create fancy products (Stray Bullets), or conduct new types of user research (log files, eye tracking, online questionnaires, etc.). The new possibilities of combining higher lexicographical quality with lower production time and costs in terms of human, financial and technological resources allocated have been dealt with only peripherally. The same holds true for the more accurate and personalised user service, which has unfortunately been discussed by only a minority of authors, although these aspects are no less important to lexicography than, for example, language processing. As explained in section 5.4, advanced electronic dictionaries in general, and online dictionaries in particular, should avoid information overload and guarantee the quickest and easiest possible consultation process. This, among other things, could be achieved by offering the users only the minimum amount of data required to cover the needs related to each function (Model T Fords) or each individual consultation (Rolls Royces). There is no single model of any of these two dictionary types, but both of them are based on a series of principles (cf. section 2.3). These can only be applied in practice if the lexicographers – or their programmers – make use of special techniques, which mainly, but not exclusively, have been developed in

92 | Special Problems Related to Online Dictionaries the framework of information science and subsequently incorporated into lexicography, cf. Bothma (2011). In this chapter we will examine some of the relevant techniques. Before doing so, however, it is important to stress that the latter can merely hold the promise of more advanced lexicographical solutions; for this promise to become reality it should be embedded in a theory capable of providing the necessary guidance and visions. Technology should never be allowed to take command. From this perspective, the available techniques may be subdivided into two main groups. The first consists of the techniques that can be used to reduce the amount of data offered to the user in each consultation, while the second allows the user to access supplementary and additional data when needed. Although they may seem mutually exclusive and opposed to each other, these two types of technique are actually complementary. Both are needed in order to take the necessary steps towards a higher degree of user satisfaction.

6.1 Data filtering The first type of technique to be discussed here is data filtering. With this technique, only part of the data contained in the data base may be shown on the screen when performing a concrete consultation. The purpose is to adapt the data more precisely to the specific needs of the target user and avoid superfluous data (information overload). Data filtering can be based on various criteria and may be more or less advanced in terms of the individual user’s needs. In this respect, one relatively simple technique (from the point of view of the user) is to offer options for monofunctional data access. In section 5.5, one way of doing this was illustrated with examples from the Diccionarios de Contabilidad, i.e. using buttons which allowed for almost instantaneous access to exactly the types of data needed in relation to text reception, production, translation, etc. However, a lexicographical function is not only defined by the situation in which the user needs may appear, but also by the relevant user characteristics (cf. section 5.4). If users are to be offered data tailored to their specific needs, this demands that, in addition to data on the specific situation, the system also receives personal data characterising the user, such as data on mother tongue, foreignlanguage proficiency level, and subject-field knowledge. In this respect, the South African information scientist Theo J.D. Bothma writes: User profiling can be accomplished through the user supplying the system with specific data, by the system tracking user behaviour and thereby automatically constructing a profile of the user or a combination of the two. User profiles can be transitory (e.g., applied to a single search only) or persistent (e.g. be stored on the system and be used for future retrieval actions). (Bothma 2011: 84)

Data filtering | 93

The easiest way of constructing a user profile is by means of interactive fill-in options where the users are guided step by step when they supply the relevant data to the system. In accordance with the resulting user profile – and the specific user situation – the pre-programmed system will then automatically calculate the types and amount of lexicographical data required to fulfil the needs occurring for this specific user type in the situation in question. These data will be the only ones presented on the screen in the concrete consultation. This might relate to definitions and meaning discrimination in the user’s mother tongue, grammatical data in a foreign language (e.g. inflection of Spanish verbs), or explications adapted to the user’s level of subject-field knowledge, as already discussed and exemplified by Bergenholtz & Kaufmann (1997). While the user profile, as a rule, can be made once and for all and only needs to be refined when the user’s relevant characteristics change (for instance, in the case of students), the description of the situation has to be supplied to the system when starting each new task. Although it should always be possible to “re-saddle” in the middle of the process, in this way it is not necessary to go through all the timeconsuming steps for each consultation. Once a relevant characterization of the specific user has been submitted together with an indication of the user situation in question, the system will then automatically select, filter and present the specific data needed by the user. It goes without saying that this “requires a very sophisticated system both in terms of the variables that have to be made provision for and also in terms of the database structure and the way in which data is characterised” (Bothma 2011: 84). One of the requirements is that at a very early stage of a dictionary project the lexicographers decide which specific data fields should be included in the lexicographical data base, and provide a detailed description of these to the programmer. These fields should be planned and designed in such a way that each specific data type is attached to its own field in the data base. The techniques or methods mentioned above are common to both the Model T Fords and the Rolls Royces. If further steps have to be taken towards individualising the lexicographical tool, then various techniques are possible. A relatively simple one has already been applied in the Accounting Dictionaries and the Diccionarios de Contabilidad, where the possibility of searching for and gaining data access exclusively to collocations and their translation equivalents is an undeniable step towards a future lexicographical Rolls Royce (cf. section 5.5). Such data-typeorientated forms of access might also be provided for other relevant types of data in specific dictionaries. Another more time-demanding technique (from the point of view of the user) is article modelling. In this case, each individual user of an online dictionary will be given the option to design his or her own master article in terms of the types of data wanted and their arrangement on the screen. One way of doing this is by allowing the user to include or exclude data types from a general list in order to meet his or

94 | Special Problems Related to Online Dictionaries her personal needs in relation to a specific task, for instance, translation of specialised texts. Such options do already exist, for instance using a pop-up window, as can be found in the Swedish Lexin, which is a general learner’s dictionary for immigrants living in Sweden (see Illustration 2). This option would be much more relevant to some users of specialised dictionaries, for example, professional translators, who in most cases are very conscious of their own individual needs relating to their profession.

Illustration 2: Pop-up window from Lexin showing options allowing the users to design individual master articles by including and excluding data types.

As with the other techniques discussed above, the design of a master article may be accomplished when the user enters the online dictionary for the first time, but it can also be done when they begin a specific activity, start a specific consultation, and even when they are in the middle of a specific consultation. In this way, it is possible to re-saddle whenever necessary in order to individualize the final data presentation. Various scholars, partially based upon the experience with their own online dictionaries, have commented that users cannot be expected to engage in the necessary interactive process in advanced lexicographical tools (see, e.g. Trap-Jensen 2010). Such comments probably reflect the situation in many cases. However, among the users of specialised dictionaries there are at least some groups who are highly de-

The type and the individual | 95

pendent on these dictionaries in their daily work. For these users, higher quality combined with reduced consultation time would be an offer they cannot refuse. Translators of specialised texts undoubtedly belong to this group. They would find it very much in their interests if they could save just a few seconds in each consultation. This tendency among translators to dedicate more time to their dictionaries also seems to have been confirmed in an empirical study by Müller-Spitzer, Koplenig & Töpel (2011). In section 5.5, we saw how an experienced translator (Tomaszczyk 1989) made a total of 691 consultations during the translation of a specialised book on diamonds. If he could save, say, 30 seconds in each consultation after going through the interactive process, this would mean a total of 345 minutes, i.e. almost six hours during the whole translation process. Many professional translators would not hesitate to take the time to communicate the necessary data to the interactive system if they were afforded this possibility. Moreover, other frequent users of specialised dictionaries would probably do the same once they had become acquainted with the timesaving options offered them by the dictionary. Finally, the provision of “hidden data” which the users can access when needed (cf. section 6.3) should also be mentioned here. The corresponding techniques allow the presentation of less data on the screen “at first sight”. This is especially relevant when users are working with small screens like tablets, mobile phones or other hand-held devices.

6.2 The type and the individual Both the lexicographical Model T Fords and the more advanced Rolls Royces have left behind the static articles inherited from printed lexicography. These advanced information tools provide dynamic solutions designed either to meet the types of needs which a specific type of user may have in a specific type of situation or activity (Model T Fords), or the individual needs of an individual user in each consultation (Rolls Royces). This development raises the question of the relation between the type and the individual. The gradual introduction of the lexicographical Rolls Royces will eventually reconcile lexicography with the fact that no type of user has ever made a type of lexicographical consultation in order to access a type of data that may meet a type of information need occurring in a type of social situation. The only thing happening every day, hour and minute, is that an individual user with individual information needs occurring in an individual situation decides to make an individual lexicographical consultation in order to access the concrete and individual data that may satisfy his or her individual needs. Although each user, user situation, user need, data and consultation may be assigned to specific types, they are not in themselves types but individual and concrete phenomena.

96 | Special Problems Related to Online Dictionaries This evident fact poses some new and complex problems in the realm of theory, since no theory in the true sense of the word can be built directly upon individual phenomena that may differ from one another in many aspects. One example will be sufficient to illustrate the complexity of this problem. Let us consider a hypothetical and relatively simple online dictionary, containing 10,000 lemmata with 10 data items attached to each lemma. If such a dictionary was designed for fully individualized access and the user was given the possibility to choose the data types they wanted to visualize, then 1,023 (210-1) possible data combinations could be displayed on the screen for each lemma; this is a total of 10,230,000 combinations for the whole dictionary, if we take into account all the possible individual needs of a hypothetical group of users in relation to the data contained in the dictionary. Moreover, the total number of possible combinations in the individual articles and the dictionary as such would rise to the astronomical figures of 4,037,913 (1!+2!+3!+4!+5!+6!+7!+8!+9!+10!) and 40,379,130,000 (ca. 40 billions), respectively, were the users also allowed to arrange the data individually in the order required by each one. It goes without saying that it is impossible to develop any theory, let alone write any lexicographical instructions, based on such an enormous and unsystematic set of data. Scientific and theoretical work presupposes an abstraction from some of the less important characteristics of the individual phenomena, and the creation of concepts, categories and types which include phenomena with certain common characteristics deemed essential and relevant for the research field in question. At the level of theory it is, therefore, a sine qua non to continue the typologization of the phenomena observed and to work with the corresponding types of user, user situation, user need, data and consultation. This conclusion does not only hold true at the level of theory, but also at the level of dictionary design and production. No lexicographer, however skilled, is capable of dealing with each and every one of the infinite number of individual needs that an infinite number of individual users may have in an infinite number of situations. This is completely unthinkable and cannot be the vision for future lexicography. All this means that, just like theory is a precondition for a scientific practice, theory-based typologization is also a precondition for a well-conceived individualization of dictionaries. Lexicographers – in order to do both their practical and theoretical job well – will therefore still have to work with types of users, situations, needs, data, etc. When planning a new online dictionary, they still have to specify the exact data types and corresponding fields to be included in a data base and at the same time decide which techniques to be used in order to guarantee quality and the required individualisation. What is needed in this respect is full use of the available information science techniques that permit individualized access to the data prepared by the lexicographers and stored in a well-structured database, or made available on the Internet by means of links, as well as access to data recreated on the basis of already existing data.

Access to supplementary data | 97

6.3 Access to supplementary data Advanced specialised online dictionaries should, above all, be characterised by flexibility. Concrete user needs are not rigid categories, and lexicographical e-tools should reflect this reality. The principles of quick and easy access, on the one hand, and default presentation of as little data as possible, on the other, will inevitably lead occasionally to situations where supplementary and additional data is required to meet a specific user’s needs. There may be various reasons for this problem and, hence, various types of possible solution, such as: – The lexicographer has, on purpose, chosen to present less data than required in order to avoid an overloaded screen. The best solution here is to make allowance for data expansion by means of adaptive presentation or, alternatively, linking to internal or external sources. – The user has provided wrong information about his or her personal characteristics, for instance, exaggerating his or her knowledge and skills, and less data than required have therefore been presented. The best solution here is to resaddle in terms of user characteristics or, alternatively, to offer the same options as above. – The amount and types of data are the right ones but the user needs a more personalised approach. The best solution here is to use annotations. – The user simply discovers in the middle of the consultation process that he or she has supplementary or additional needs, for instance the need for supplementary text examples or for specialised background knowledge when translating specialised texts. In such cases the best solution would often be to link to internal or external sources or even to reuse the data from these sources. We will now briefly discuss some of the techniques or methods mentioned above: – Adaptive presentation. There are various techniques in terms of adaptive hypermedia but here we will mention only pop-up windows activated by either mouse-moving or clicking, cf. Prinsloo et al. (2012). Very often the lexicographer decides that not all data needed in a specific consultation should be presented at first sight. There could be various reasons for this; for instance, to prevent the user from having to scroll down so as to see all data (this depends on the size of the screen), or simply because the predicted type of user cannot be expected to overview too many data at the same time. Whatever the reason, the lexicographer can choose to add some “hidden” supplementary data, which can easily be accessed, for example, by moving the mouse over a specific word or area, or clicking on a link. Then a pop-up window will appear with supplementary text, tables, illustrations, photos, etc. It could, let us say, be a table with the inflected forms of a Spanish verb, which would occupy a large amount of screen space. Alternatively, it might be background information related to a specific term or phenomenon, etc. The content of the pop-up window could be specifically pre-

98 | Special Problems Related to Online Dictionaries

pared for this purpose, or it might consist of already-existing data taken from another dictionary article. Adaptive presentation is especially relevant when users are working with small screens like tablets, mobile phones or other handheld devices, cf. Kwary (2013). Indexing and abstracts. One of the on-going discussions among scholars in terms of e-lexicography is how much data should be presented at a given time to the dictionary user (see, e.g. Lew 2013). If the idea is to provide quick and easy data access, then the answer seems to be that the maximum amount of data is that which can be displayed on the screen without the necessity to scroll down. However, in many dictionaries, such as those with cognitive functions like The new Palgrave Dictionary of Economics, this represents an ideal which is impossible to achieve in each and every article displayed. One solution could, therefore, be the presentation of some sort of preliminary article index which can be further expanded by means of hypermedia (like in Wiktionary). Alternatively, a small abstract could be placed at the top of the article, as has been done in The New Palgrave Dictionary of Economics. The application of these and similar techniques immediately provides the user with an overview of the content. In addition, the possibility of expanding an index by means of hypermedia allows the user to go directly to the section of the article which is pertinent in each case. Annotation. In the Web 2.0 environment, users are allowed to add their own data and comments to already-existing documents without changing the original in any way, cf. Bothma (2011: 96). Applying this technique, users may become mini-lexicographers and “co-authors” of their own dictionary, for instance, when recommendation of certain terms (in case of synonymy) or addition of other terms and data is relevant to a particular company or branch. This technique can without any doubt contribute to the future individualization of lexicographical works. Moreover, although it can also be applied by individual users in order to personalize specific articles, it may above all be useful in the framework of predefined user groups such as companies, branches, public entities, research groups, etc. Reuse of data through linking. As stated above, an online dictionary may provide either static or dynamic articles in order to present the data stored in the connected database. However, in both cases, the dictionary contains a limited amount of data that may not be sufficient or accurate enough to fulfil the user’s information requirements in a specific consultation. In such cases, it could be advantageous to provide a link to additional data stored in external sources such as corpora and the Internet. The data in question could be additional collocations, example sentences showing the usage of collocations and other linguistic features, or so-called contextual definitions found in existing texts. In this respect, Rundell (2010) proposes:

Access to supplementary data | 99 Online dictionaries could link directly to general and specialized corpora, allowing users to search for examples of any word, pattern, or linguistic feature they are interested in. (Rundell 2010: 174).

In the same vein, Theo Bothma observes, from the point of view of information science: Much work has been done on the role of corpora in dictionaries; see, for example, Prinsloo 2009 and the opinion regarding the value of such corpora differ widely. Lexicographers should, however, consider exploring the vast stores of linked open data to provide access to additional information to satisfy especially cognitive needs of users. Currently e-dictionaries provide links only to so-called outer texts and to manually selected external examples; examples of both principles can be found in a number of the dictionaries by, as well as in many other free online dictionaries. This, however, does not make systematic and/or automatic use of reusable data and requires a tremendous input in terms of time and effort from the lexicographer. One example of a dictionary that does provide the option to link to a huge external database of examples is the Base lexicale du français (BLF). After getting examples of the meaning of a word the user has the option of linking to various corpora, including a set of documents of the European Parliament and Wikipedia [...] These examples are automatically searched by the BLF and the selection of the examples does not require any input from the lexicographer (except, obviously, specifying which corpora are to be searched). (Bothma 2011: 92)

Heid, Prinsloo & Bothma (2012) have dedicated an entire article to the integration of dictionaries and corpus data in web portals, and have put forward a number of interesting proposals in this regard. According to them: Electronic dictionaries should supply a natural bridge for the user from the dictionary article to a variety of internet resources to enhance access to potentially relevant information. In total, the user should be able to find a wealth of digestible information without being overloaded. (Heid, Prinsloo & Bothma 2012: 270)

When an individual experiences an information need, he or she may then access the data contained in a dictionary and thereby retrieve the required information. This is at least what occurred in lexicography until only a few years ago, with lexicographical works providing only direct access to lexicographically selected, elaborated and prepared data, and not to that prepared and made accessible elsewhere, for example, in books and archives. It is only recently that a few advanced lexicographical tools have tried to reuse already-existing data made available through a database or the Internet, and one of the visions today is not only to reuse such information but also to repackage it by adapting it to the specific information needs of the user in each situation, cf. Bothma (2011) and Tarp (2011).

100 | Special Problems Related to Online Dictionaries

6.4 The use of corpora and the Internet In the previous section it was argued that the reuse and repackage of data imported from external sources like corpora and, above all, the Internet, is bound to play an important role in future dictionaries in terms of individualization of needs satisfaction. Corpora and the Internet are already extensively used by lexicographers in order to select (mostly corpora) and validate (both) data during the compilation of dictionaries. However, what is at the stake here is the automatic incorporation of data into online dictionary articles and, in a far-reaching perspective, the automatic generation of such articles, that is, without the intervention of the human factor in the selection and validation process. During the past few years, various tools have been developed with a view to approaching this ideal. Some of them may make the job easier for lexicographers engaged in the production of various types of dictionaries, but they are still far from convincing if the vision is to completely supersede the lexicographer, i.e. the human factor. One such example is a bilingual term recognition system discussed by Vintar (2010). By means of this tool various English terms with their Slovene equivalents have been selected within the fields of tourism, accounting and the military. Unfortunately, Vintar does not reveal how many terms have been detected within each field, and it is therefore not possible to see to what extent terms are actually covered. However, at least one of the 10 “recognized” English accounting terms – disputable receivable – given as an example by Vintar (2010: 150) is not an accounting term but a collocation (the accounting term is receivable). This provision of a nonexisting term may not seem so important. Although it is not the case here, within accounting a “small” mistake of this sort could easily become a major one which may eventually lead to loss of millions of euros and compensation claims against the uncritical dictionary user, e.g. a translator. This is why it is recommended that all terms “recognized” are validated by a subject-field expert before becoming available to the end users (cf. section 5.7). In the modern world, especially with the advent of the Internet, there are a lot of sources by which individuals may satisfy their information needs. In this respect, it should never be forgotten that what distinguishes – or ought to distinguish – dictionaries and other lexicographical works is not only that they are, par excellence, consultation tools that provide quick and easy access to the data from which the required information can be retrieved, but also that such data are supposed to be reliable and trustworthy. Consequently, as long as artificial intelligence has yet to be developed up to a level at which it can supersede human intelligence – and this may never happen – the intervention of a lexicographer or subject-field expert, at least in terms of validating the data automatically selected or recognized, will continue to be a necessary precondition for reliability and trustworthiness. In section 5.4, we saw a negative example of what could happen when a non-expert (Kilgarriff 2012) tried

Online communication with the user | 101

to take in definitions by googling; it is easy to imagine the disastrous consequences should a non-human system be asked to do the same. Does this rather pessimistic vision invalidate the idea that lexicographical Rolls Royces should automatically recreate and re-present data taken in from corpora and the Internet in order to provide more dynamic and individualized solutions to the users? Not necessarily, but it puts it into perspective and gives rise to four principles or recommendations that could contribute to guaranteeing the quality of this advanced tool: – The present-day or near-future vision should not be a completely automatic, but rather a semi-automatic generation of dictionary articles, where central parts of the articles are compiled or validated by lexicographers or subject-field experts, and where the users in need of additional data are given the option to import these data from external sources. – Non-validated data should only be taken in and presented to individual users whose critical sense, subject-field knowledge, and lexicographical culture place them in a position to validate the respective data in a reasonable manner. – The user should, in one way or another, be clearly informed about the nonvalidated data and the sources from which these data have been imported. In this respect, it should be remembered that data taken from a well-composed and controlled corpus may have greater degree of reliability than data taken from the Internet. – The user should be given the option to determine the amount of additional data in order to meet his or her individual need and avoid information overload. With this in mind, the question remains as to which kind of data could or should be imported and re-presented in online dictionaries? In principle, such data might be any that could be addressed to the lemma or to already selected information, such as: – Collocations and other word combinations showing the syntactic properties of the lemma in question. – Example sentences illustrating the use of “any word, pattern, or linguistic feature”, including collocations and idioms. – Contextual definitions which could enrich the definition already included. – Background texts in order to meet additional cognitive needs related to the task performed.

6.5 Online communication with the user In the printed environment, communication between lexicographer and user was mainly one-way, where sporadic user feedback could only be taken into consideration in subsequent editions which frequently took several years to publish, if they

102 | Special Problems Related to Online Dictionaries were published at all. By contrast, in the online environment traditional communication has increasingly been replaced by the possibility of having two-way communication, where users can comment almost instantaneously on existing data, suggest modifications, or ask for additional data in order that their needs are met in a more complete way. This, of course, also modifies the traditional role of lexicographers, whose job no longer ends with the publication of a given dictionary; today they should be prepared to perform new tasks with regard to post-publication user service and continuous updating. In order to guarantee this partial incorporation of users into the production and perfection of the dictionary, it is important that the dictionary interface should in some way allow users to communicate directly with lexicographers, thereby benefiting the quality of the work and increasing its adaption to their needs. Apart from this open and direct communication, it is also extremely helpful for lexicographers if the underlying system allows them to receive continuous, indirect information about user behaviour, for instance about frustrated “look-ups” for specific terms or collocations. This type of information may not only assist the lexicographers in the perfection of the already finished product but may also interfere directly in their work during the phase where the dictionary is still under construction. The studying of log files reflecting user behaviour in some online dictionaries has shown that only a certain percentage of lemmata are ever consulted by the users. For example, only a third of the approximately 110,000 lemmata in Den danske netordbog (Danish Internet Dictionary) have ever been searched for, despite the fact that these have been available online for several years and that over 18 million consultations have been made, cf. Bergenholtz & Norddahl (2012: 220). The nonconsulted words may be old or new, long or short, difficult or easy, and belong to all parts of speech. So far no systematic explanation has been found for this strange phenomenon, which also seems to be repeated – with other figures – in specialised online dictionaries produced by the Aarhus-based Centre for Lexicography. If the above tendency were confirmed in other projects, editorial policy would have to be adapted correspondingly. Such a policy could, among other things, include the online publication of a new dictionary when, for example, only 30 per cent of all its expected lemmata, the “core” ones, have been completed. The remaining lemmata could then be incorporated gradually according to the pre-established plan, while the still unpublished terms and words subject to frustrated consultations – and relevant to the subject field and functions of the dictionary – could be finished and included as soon as possible after the frustrated consultation, e.g. in a period of 24 hours. The introduction of such working methods could benefit all parties: users could see their needs fulfilled more quickly and more accurately than before; publishing houses could start selling their products much sooner than otherwise; and lexicographers could receive feedback and assistance to improve the quality of their work and adapt it more successfully to the specific needs of their target users.

Shortening the access route | 103

6.6 Shortening the access route As has already been mentioned in previous chapters, the time factor in terms of a quick overall consultation process should also be regarded as an important criterion for quality in specialised online dictionaries. Saving time is important for many users of these works, especially those who have to consult dictionaries as part of their work. We have already discussed how a user-friendly article design with as few data as possible may contribute to reducing the time required to find the specific data and retrieve the information sought. But also the time spent from the moment the consultation starts until when the data appear on the screen is relevant here. If this can be reduced, various seconds can be saved in each consultation. In order to reduce additionally this part of the overall consultation time, various techniques have been developed in recent years. These techniques have, among other things, been embodied in applications connected to an e-dictionary. Once such an application has been downloaded and installed on the computer, a user who is reading, writing, translating or revising electronic texts has simply to point to the problematic word or word combination in order to get assistance. The application will then – in a pop-up window – provide a definition, a translation equivalent, the correct spelling or other relevant grammatical data. The content shown in the window can be adapted to the user’s specific activity, for instance, text production or translation. When this activity has been indicated, the user need only click on the required data in the pop-up window, for instance, data concerning an inflected form, and this will immediately be reproduced in the text with which they are working, thereby saving the time required to re-write it in the text. Bothma & Prinsloo (2013) have analysed one such application used to assist text reception in Kindle e-books. The two South African researchers show how this application can still be refined, and propose a number of solutions in this regards. If a user works with this or a similar application, then he or she does not any longer have to activate the online dictionary, write the relevant word or word combination in the search field and wait for the required data to appear on the screen. All these intermediate phases can now be dispensed with. It goes without saying that the technique can only be used in connection with electronic texts. When it is used in activities related to these, it represents a considerable reduction in overall lexicographical consultation time. For some categories of users, this will also mean a measurable reduction of work costs. Although the technique still has to be refined, as pointed out by Bothma & Prinsloo (2013), there is little doubt that it will gain a footing in online lexicography in the next few years. Designers of specialised online dictionaries should, therefore, seriously consider incorporating this technique and enabling applications such as those mentioned to be downloaded and used in connection with their products.

7 A Critical View of Terminography Designing, compiling and updating specialised dictionaries is also a key concern of terminology, the academic status of which received an impulse in the 1930s. By the 1840s terminology was mainly interested in documenting the use and meaning of specialised words with the aim of making scientific and technical communication precise and objective: Terminology signifies the collection of terms, or technical words, which belong to the science. But in fixing the meaning of the terms, at least the descriptive terms, we necessarily fix, at the same time, the perceptions and notions which the terms are to convey; and thus the Terminology of a classificatory science exhibits the elements of its substance as well as of its language. (Whewell 1847: 463-464)

Whewell uses substance to refer to the content of the concepts. His ideas, which were widely accepted in the mid-nineteenth century, were based on the conviction that common general language has a certain degree of looseness and vagueness, and is not well suited for intellectual discussion. Instead, scientific discussion and technical descriptions can be achieved only by agreeing on the meaning of terms: The meaning of technical terms can be fixed in the first instance only by convention, and can be made intelligible only by presenting to the senses that which the terms are to signify. The knowledge of a colour by its name can only be taught through the eye. (Whewell 1847: 464)

Some years later, especially following Wüster’s work in the 1930s, the two meanings of terminology envisaged in Whewell’s quotes became universally accepted. One of them is that terminology is the set of designations belonging to one special language. This definition is used when we refer, for instance, to the terminology of Economics or that of Physics. Another concerns the scientific discipline that studies the structure, formation, development, usage and management of specialised vocabularies in various subject fields (Schmitz 2006; ISO 704). This has resulted in the advent of “terminology science” (or terminology studies), which deals with concepts and their representations in special languages. In addition to the meanings above, some scholars, e.g. Rey (1975), and institutions such as the ISO, have added a new one, which is usually called terminography. This was coined on the analogy of the distinction lexicology/lexicography, and typically refers to the activity of processing terms in specialised dictionaries (cf. section 3.3.). This means that Rey and colleagues, e.g. Rondeau (1983), Riggs (1989), and Gaudin (1993), are reproducing the same pattern of work that is commonly defended by linguists working with general language: lexicology/terminology analyses common words or specialised terms by studying their nature, meanings, elements, and relations; lexicography/terminography is the practical counterpart, i.e. it is a practical activity that is concerned with the design and construction of information tools

Theories of terminology | 105

(cf. section 3.3). In the following section, we offer a critical view of the theories of terminology, that is, the theoretical constructions that terminologists use (or must use) for designing, making, and updating specialised dictionaries; these terminographical products include terminological knowledge bases, glossaries, term banks, terminological dictionaries, lexica, etc. Hence, we do not separate the three meanings of terminology, as this book defends the interconnection of theory or practice. This chapter offers a critical view of the theories of terminology in terms of their practical application for designing, making, and updating specialised dictionaries.

7.1 Theories of terminology Humbley (1997: 16) claims that terminology “is practiced differently, and thus viewed differently from place to place: some language communities have not much gone in for terminology (the USA is a case in point), others, the French-speaking ones in particular, have so much more, and this affects the perception of what terminology is and how it relates to lexicography”. To the French-speaking communities should be added several others, notably, Spanish-speaking, Catalan-speaking, German-speaking, and Dutch-speaking communities. Cabré (2003: 164), for example, mentions certain scholars who have been concerned with terminological theory and are members of one of the abovementioned language communities: French-speaking Gaudin (1993); Catalan and Spanish-speaking Cabré (1999, 2003); Dutch-speaking Temmerman (2000); Germanspeaking Wüster and colleagues, e.g. Budin (2001), etc. It might seem surprising that native English speakers have not been interested in terminological theory; perhaps this indicates the differences in conception of the term theory that most English-speaking scholars have compared to that of other scholars (cf. section 4.2.). The former are more interested in suggesting methodologies. For instance, Pearson (1998: 1) did not design a terminological theory but used Sinclair’s (1991) corpus paradigm with the aim of demonstrating that “corpora can be used for semi-automatic terminography.” This idea has been (and still is) very influential, and we will refer to it in several ways in the coming sections, especially to support the argument that the use of corpus in specialised lexicography must be limited to a number of tasks (cf. section 6.4; chapter 9; also Fuertes-Olivera 2012). Finally, we are witnesses to a plethora of recent publications which claim that specialised information tools can be generated (semi-)automatically by means of computer programs without the intervention of trained terminologists (or lexicographers) or a subject-field expert. Some of these publications are authored by the same scholars who espouse a particular theory of terminology. Consequently, this section and the following one can be seen as a quest for understanding the changes in terminology that have occurred since the 1930s. For operational purposes, we have analysed theories of terminology in three sub-groups, each of which is viewed

106 | A Critical View of Terminography critically so as to decide whether they are suitable for designing and producing (online) specialised dictionaries: (i) the General Theory of Terminology, (ii) Linguistics-based theories of Terminology, and Terminology and knowledge engineering.

7.2 Traditional terminology: The General Theory of Terminology Wüster initiated the discipline of Terminology in the 1930s by formulating a set of principles that aimed to eliminate ambiguity, enhance standardisation and establish terminology as the science of terms. He was following an academic tradition, as shown by the following quotes: The construction of a Classificatory Science includes Terminology, the formation of a descriptive language; (...). Terminology must be conventional, precise, constant; copious in words, and minute in distinctions, according to the needs of the science. The student must understand the terms, directly according to the convention, not through the medium of explanation or comparison. (Whewell 1847: xxxii) Terms must be constructed and appropriated as to be fitted to enunciate simply and clearly true general propositions. (...) This Aphorism may be considered as the fundamental principle and supreme rule of all scientific terminology. (...) The maxim is true of words appropriated as well as invented, and applies equally to the mathematical, chemical, and classificatory sciences. (Whewell 1847: lxxiii-lxiv) The new laws which the study of electro-chemistry brought into view, required a new terminology to express their conditions: and in this case, as we have observed in speaking of the Twelfth Maxim, arbitrary words are less suitable. Mr Faraday very properly borrowed from the Greek his terms Electrolyte, Electrode, Anode, Cathode, Anion, Cathion, Dilectric. In the mechanico-chemical and mechanical sciences, however, new terms are less copiously required than in the sciences of classification, and when they are needed, they are generally determined by analogy from existing terms. (...) In such cases, it is generally possible to construct terms both compendious and descriptive, without introducing any new radical words. (Whewell 1847: ciii)

This tradition permeates Wüster’s work, which was known by Wüster’s followers as the General Theory of Terminology, and it was made public some years later than originally described (Wüster 1979). At a very abstract level, Wüster initiated the development of an independent scientific discipline called terminology science or terminology studies, which has developed methods and procedures for managing terminology (Wright & Budin 1997, 2001). These methods and procedures focus mainly on analysing concepts and terms by making use of Odgen & Richards’ (1923) triangle, which illustrates the presence of a triadic relationship among concepts, terms, and objects that follows Peirce’s semiotic sign (1868) (we follow Schmitz (2006) in this description):

Traditional terminology: The General Theory of Terminology | 107

Objects: objects are concrete or abstract entities existing in the real world. Wüster (1979) clarified them by introducing the notion of individual object, which allows him to make a difference between the overall class or conceptual “type” of something, e.g. a table, and the individual instances or “tokens”, which are each individual manifestations included in the overall class, e.g. many (different) tables. Concepts: concepts are “cognitive representatives” (Felber & Budin 1989) or units of knowledge. They are created by a unique combination of characteristics that humans recognise and know that exist in a majority of individual objects of the same type and then “store these characteristics (e.g. remember them) and use them to impose order on the world of objects, in order to achieve mutual understanding when they communicate with other people.” (Schmitz 2006: 579) In other words, concepts are cognitive units that are formed in our minds and associated with verbal or non-verbal representations (terms, symbols, etc.). Terms: terms are verbal designations of concepts in a specific subject field. The term serves as the representation of the concept and therefore the term is used for communication. Some of the terms consist of more than one word. These terms are called multi-word terms or compounds. Terms always designate class concepts whereas names are representations of individual concepts. Phrases, phraseological units, acronyms, (chemical) symbols, formulas and images also function as terms.

Concepts are more important than objects and terms. Concepts can be individual, e.g. the Empire State Building, or general, e.g. building. They can be culturedependent or culture-independent, according to their possible modification to the cultural and social background of the human being who generates them and the environments in which they are used. Finally, concepts can be related to one another primarily in view of the characteristics with which they are associated. Concepts establish conceptual relations. These can be of a hierarchical or nonhierarchical nature. Hierarchical relations can be divided into generic (also, hyperonymic) and part-whole (also, meronymic) relations. Generic relations allow a relation between a broader (or superordinate) concept and a narrower (or subordinate) concept. The term designating a superordinate concept is hyperonym, whilst that referring to a subordinate concept is hyponym. Concepts that occur at the same level are called coordinate concepts, and they are termed co-hyponyms. Nonhierarchical relations are more difficult to define and include sequential as well as cause-and-effect relations. Conceptual relations are used in order to create a concept system, which is represented with these conceptual relations contained in it by means of graphical representations, numerical codes and ontologies (also called taxonomies). Nowadays, concept systems are being created by information specialists and knowledge engi-

108 | A Critical View of Terminography neers with the stated aim of showing all the major terms and data objects that are managed in a domain or sub-domain, etc. Furthermore, the merits of Odgen & Richards’ triangle for practical terminological work coincided with the view that terminologists must prepare information tools containing features of artificial languages, i.e. communication systems that create unique designations which are firmly fixed in place by definition and measurable distances to other designations (Sager 1984); these would eliminate the possibility of communication breakdown by emphasising that concepts are universal and independent of cultural differences. Taken together, these abstract and practical views resulted in a series of principles that Wüster and his followers - e.g. Felber (1979), Laurén, Myking & Picht (1998), and Budin (2001) – apply to the design and production of specialised information tools. The most important of these are summarised below: – Terminological information tools must deal solely with specialised language, emphasising the priority of concept over term. This means that terminologists must design information tools that show the precision of concepts (monosemy) and their univocity (i.e. absence of synonymy). As this situation is difficult to achieve – synonymy, homonymy, and lack of equivalence are widespread in all natural languages –, they have designed systems for representing these phenomena in a controlled way. For instance, they recommend using numbers such as 1, 2, 3, etc. to identify each meaning of a polysemous term. – Terminological information tools must use only written specialised texts as sources of information and experts as informants. More recently, the meaning of spoken forms has been recognised in, for example, contexts relating to language planning. – Terminological information tools must contain concept-related, term-related and administrative-related data categories (ISO 12620 (1999). Concept-related data categories are concerned with the concept underlying a terminological entry, and typically include the following: definition, subject field, illustration (e.g. a symbol or formula), classification, superordinate concept, subordinate concept, and co-ordinate concept. Term-related data categories contain data that refer to one particular term, such as term (including synonyms, abbreviated forms and orthographical variants), term type, context (usually with an example), grammar data, geographical restriction, linguistic restrictions, register, and project code. Finally, administrative categories refer to the entry as a whole and include identification number, date of creation (or adaptation), author, source, reliability, and comments. – Terminological information tools must be based on describing concept relations, which are based on systematic classifications. This has resulted in terminologists preparing information tools for describing terms synchronically. They can be arranged onomasiologically or semasiologically with systematic or alphabetical macrostructures, and may include as many phraseological units as

Terminology and linguistics | 109

are deemed necessary. Concept relations can be hierarchical or non-hierarchical. Terminological information tools must contribute to the general purpose of achieving international standardisation. This objective coincides with two approaches to terminology that share the general aims of terminology: (i) a translation-oriented approach and (ii) a language planning-oriented approach. The translation-oriented approach to terminology implies the design and construction of information tools whose genuine purpose is to facilitate translation and terminology processing, which is considered essential for supporting translation. The language planning-oriented approach emphasises official intervention in increasing the status of minority languages or languages which have, for some reason or another, been historically disadvantaged.

Wüster’s theories are still influential, especially as regards the interests terminologists still show in describing concepts and conceptual relations. We believe, however, that such an interest is erroneous and is based on a misunderstanding of the situation in which it originated. Wüster was an engineer, and believed that common language was ill-suited for describing scientific processes, theories and methodologies. In other words, his discipline was anchored on an objective, positivist view of science, and consequently he produced a theory that could be used as long as the sender and receiver shared the same level of understanding, i.e. a theory for communication among experts and for standardising terms. The reality, however, is that there are many more types of users of specialised texts, e.g. interested laypersons and translators. These are not in a position to understand specialised concepts and concept relations, no matter how these are worded. Hence, it seems unnecessary to offer conceptual relations to non-experts in, for instance, specialised dictionaries that mainly target non-experts and interested laypersons. This idea is supported experimentally and will be analysed fully in chapter 8, in which we will show that the specialised dictionaries defending the inclusion of conceptual relation in their articles can be separated into two groups. The first, e.g. IATE, includes them because they were included in its predecessor, EURODICAUTOM. The second, e.g. EcoLexicon and Genoma, defends the validity of Cognitive Linguistics tenets for making specialised dictionaries.

7.3 Terminology and linguistics Various scholars had started to criticise Wüster and his followers’ ideas by the 1990s. All of them agree that terms are more than linguistic representations of concepts. They are also instruments to communicate knowledge and are, therefore, subject to the same forces that permeate general language: indeterminacy, variation, multidimensionality, culture-dependency and context-dependency, etc. This

110 | A Critical View of Terminography prompted them to argue that terminology cannot be an independent science or discipline, but a dependent science that needs to be understood with contributions from cognitive science, language sciences and communication science (Cabré 2003). This argument supports the theoretical foundations on which different recent theories of terminology are based. As they all make use, in one way or another, of linguistic assumptions, concepts, and terms, we will group all of them in this section. For example, the communicative approach to terminology emphasises the communicative function of terminology, which is centred on facilitating effective communication and knowledge transfer within the professional community (Cabré 1999).

7.3.1 Cabré and the Communicative Theory of Terminology Maria Teresa Cabré (1999, 2003: 182) presents her theory by making two general assumptions. The first is that “terminology is simultaneously: a set of needs, a set of practices to resolve these needs, and a unified field of knowledge”. Her second assumption is that “the elements of terminology are the terminological units”. Regarding the first assumption, she enumerates a list of needs that are associated with the activities carried out in different fields: Firstly, terminology presupposes a need for all the activities related to the representation and transfer of specialised knowledge such as technical translation, the teaching of languages for specific purposes, technical writing, the teaching of special subjects, documentation, special language engineering, language planning, technical standardisation, etc. (Cabré 2003: 182)

Following on from this, she acknowledges that these needs must be addressed by developing “products” oriented towards the solution of specific requirements, whatever these might be. Regarding her second assumption, she describes the characteristics relating to the object of study of terminology: terminological units, which are at the same time “units of knowledge, units of language, and units of communication. Based on this approach, the description of a terminological unit must necessarily cover these three components: a cognitive component, a linguistic component and a sociocommunicative component”. (Cabré 2003: 183) Both assumptions imply that Cabré (2003: 187-192) is in favour of a terminological theory that is different from the one espoused by Wüster and his colleagues. Cabré equates terminological units and general words in their behaviour, and therefore supports the view that terminology is part of linguistics; she argues that the analysis of terminological units and the design and construction of terminological reference works can be carried out through linguistic theories and methodologies. For the purpose of this book, Cabré’s stance can be summarised as follows: – Terminological units are multidimensional and can be analysed and described from different positions and conceptions. This implies the necessity of using a

Terminology and linguistics | 111

– –

cognitive and functional linguistic theory, i.e. one that accommodates an analysis of the grammatical, pragmatic and semantic characteristics of terminological units describing their specific and general characteristics. Terminological units have to be studied within the framework of specialised communication. The latter shows a preference for certain text types and a strictly controlled knowledge structure that is manifested in exclusive lexical units whose meaning and use is restricted. This communicative framework is used for transmitting specialised knowledge among experts, experts and semiexperts, or experts and learners, and also for popularising science and technology. Terminological units are observed in many different oral and written situations in which experts address directly, or to specific recipients. Terminological units are recognised in an organised structure of knowledge. They are recognised “from their meaning in a subject field, their internal structure and their lexical meaning”. (p. 189) Terminological units are like general words. She postulates that lexical units are neither terminological nor general but that they are general by default, acquiring special or terminological meaning when “this is activated by the pragmatic characteristics of the discourse”. (190) The theory “is not directed towards any practical application”. (p. 190) The applied aspect develops “according to a methodology, based on the principles of the theory, which permits the projection of these principles in practical activities aimed at solving problems and developing resources for the needs of representing and communicating this special knowledge”. (p. 191) Terminological units are acquired through discourse, which means that terminological units are acquired through systematic learning in a professional environment.

We believe that Cabré’s previous views may be criticised on several accounts. At an abstract level, Cabré’s assumptions imply that the very concept of specialised language is a non-starter. For instance, the sets of conditions which distinguish terminological units from other similar but different units cannot be applied for differentiating between general and specialised units. For instance, terminological units “depend on a thematic context”, “can have lexical and syntactic structures”, “exploit all the devices of word formation and the processes of acquiring new units”, etc., are characteristics of all types of lexical units, be they specialised or general. Furthermore, the author enumerates a list of conditions that terminological units must fulfil, posing many questions regarding their applicability. For instance, the conditions associated with the cognitive component run counter to the true nature of language units, whose meanings are in a constant flux and cannot be located exactly within a conceptual structure. How do we place the Spanish synonyms amortización and depreciación when translating them into English, whose

112 | A Critical View of Terminography equivalents (amortisation and depreciation) refer to different concepts? Moreover, it is easy to question claims such as the following: terminological units do not occur in general texts, or only experts fix their meanings. For instance, translators of specialised texts change and fix specialised meanings, as seen in the case of deterioro, a Spanish term used in accounting that acquired its current meaning through a translation of the English impairment. This translated meaning is so widespread that it has almost eliminated the traditional Spanish term for referring to the concept denoted by the English impairment (provisión por depreciación). At a more practical level, Cabré’s theory makes few inroads. She does not mention the needs of her envisaged specific users when carrying out, say, the abovementioned activities (cf. section 5.2.). Moreover, she uses “application” as a synonym of “information product”, and enumerates a list of applications that are ontologically different and, therefore, cannot be included in the same list for use with the same purpose. For instance, a dictionary is for specific consultation, whereas a text is not: This leads us to think that, despite of what is usually said about standardised terminological glossaries, it is the circumstances of each situation which determines the type of application (glossary, lexicon, dictionary, software, text, poster, standard, etc. in one or several languages), the information they must contain (terminology, phraseology, definitions, variants, contexts, phonetic or phonological representation, foreign language equivalents, illustrations, etc.), their representation and even their means of dissemination. (Cabré 2003: 183)

Finally, she believes that her theory does not need to be translated into practical tools, thus questioning the purpose of her work: We therefore do not share the idea expressed by Budin (2001: 15, cited above) that alternative theories to the traditional one are explained solely by their different practical orientation. (Cabré 2003: 190)

To sum up, Cabré’s theory has not resulted in the design, production or updating of high quality specialised dictionaries, as our analysis of Genoma will show (cf. section 8.14.1).

7.3.2 Temmerman and the Sociocognitive Approach By the mid 1990s, scholars such as Gaudin (1993), Boulanger (1995), and Temmerman (1997) had emphasised that Wüster’s ideas did not take into consideration real language usage. These authors adopted a more restricted focus than Cabré. They criticised in particular the concepts of special field and expert used by Wüster, by indicating that experts do not form homogeneous groups and that special fields tend to be interdisciplinary and dynamic, as they must adapt to an ever-changing reality. Such arguments laid the foundation for new approaches to terminology that

Terminology and linguistics | 113

were based on Sociolinguistics (Gaudin 1993), Functionalism and the Cognitive Sciences (Temmerman 2000). The former emphasised the influence of situational factors, as well as social and ethnic criteria, on terminology. By way of example, gender expectations are responsible for the introduction of synonymous terms such as chair and chairperson. To the best of our knowledge, the sociolinguistic approach, usually called socioterminology, has paid more attention to signalling Wüster’s rather weak points than defending a theory of terminology that could be used for designing and compiling specialised dictionaries. Temmerman, however, has proposed a terminological theory known as sociocognitive terminology. This also recommends the term as the starting point and an analysis of real usage for formulating the theory: The reason why we claim that our first line of thought links up with hermeneutics is that we concentrate on textual information in order to find out about categorisation and its intricate relation with language. The textimony of how language exists (parole) is in the textual archives. (Temmerman 2000: 54)

Sociocognitive terminology is grounded in notions proposed by Lakoff (1987) and Johnson (1987), arguing that our knowledge with regard to the world of science and technology is experiential, and embodied in, in other words, based on, our sensory perceptions and the result of our reasoning capacity, which allows humans to create categories in the mind, usually in the form of prototypes (Rosch 1978). Consequently, sociocognitive terminology considers that language plays a part in the understanding of the world, that the world is (partly) in the human mind, and that an understanding of language cannot be separated from the understanding of the world. Within the above general framework, Temmerman (2000: 219-232) proposes new forms of terminology description that includes a list of theoretical principles and their application to the design and construction of terminographical products. These are the principles on which sociocognitive terminology is based: – Terminology does not deal with concepts but with units of understanding. A unit of understanding typically has a prototype structure and is described by combining three perspectives: nominalistic, mentalistic and realistic. Terminology should try and describe the relationship between these three perspectives. This signifies that the terminological work is concerned with relating the sense of the word with the idea people have of it in their mind and its presence in the external world. – Understanding is a structured event. In other words, there are no clear-cut concepts but units of understanding with intracategorial and intercategorial structures that function in cognitive models. – Meaning description varies in order to adapt to the level and type of specialisation of the sender and the receiver. Thus, there are different types of units of understanding and as a result it is possible to imagine a template of understand-

114 | A Critical View of Terminography

ing. This consists of different information modules which contain more or less essential information. An analysis of real texts shows there is little arbitrariness in categorisation and lexicalisation. This means that synonymy and polysemy are functional, that there are different ways of referring to prototypically structured categories, and that polysemisation occurs because of changes in the world due to new technology or social change, changes in the understanding of the category, and language constraints. Units of understanding are constantly evolving and terms are not arbitrary but motivated. This means that they can be understood in terms of cognitive models such as metaphorical idealised cognitive models (ICMs); in other words, figurative expressions are also part of the terminological description.

In addition, Temmerman claims that terminology also needs tools, i.e. new methods and techniques for analysing terminological data: Sociocognitive Terminology is a paradigm involving a number of principles which differ from the paradigm of traditional Terminology. In order to be a fully-fledged discipline Terminology needs more than principles. It needs tools, i.e., methods and techniques for the analysis of terminological data which can serve a double purpose. On the one hand these methods and techniques allow the researchers to test the validity of the principles and on the other hand they provide the terminographer with data for his descriptive practice. (Temmerman 2000: 229)

The terminological tool envisaged by Temmerman and her colleagues is termontography (Temmerman & Kerremans 2003; Kerremans 2004; Kerremans et al. 2005a; Kerremans et al. 2005b; De Baer et al. 2006; Kerremans et al. 2010). Kerremans et al. (2005: 93) indicate that termontography is a multidisciplinary and functional approach for terminology description “in which theories and methods in terminography are combined with methods and principles in ontology engineering”. Termontography is based on the insight gained from sociocognitive terminology management, which takes into consideration linguistic information, cultural data and cognitive perspectives during terminology description. It is presented as a categorisation framework, that is, a methodology whose goal is “to build comprehensive lexical resources that may help to overcome certain communication problems”. (De Baer et al. 2009: 535) This methodology consists of several phases, all of which are described below with examples referred to the FF POIROT, a 2002-2005 EU funded project that explores the use of ontology technology for fraud prevention and detection. This project aims at compiling “a multilingual computationally tractable and shareable knowledge repository for the financial forensic domain”. (Homepage) The first stage is the analysing phase. It comprises two steps that are needed to integrate the analysis of users, applications and purposes of terminology bases. The first step requires writing a user requirement report, whilst the second stage in-

Terminology and linguistics | 115

volves the development of a categorisation framework in the knowledge specification phase. Kerremans et al. (2005: 95) describe the requirement specifications for the FF POIROT project. These deal with macrostructural and microstructural aspects. The macrostructural requirements are the general framework that determines the texts from which suitable terminology can be extracted. In this project, this comprises legal rules, descriptions of product features, an understanding of commercial transactions, categories related to companies and their structure, and data concerned with investment websites. The microstructural requirements consist of the specific elements that offer the context meaning which must be present in every terminological record: definitions of the terms retrieved, co-texts, references to the textual sources from which descriptions and co-texts have been retrieved, and relational markers that link two terms in a given context. The above requirements were the result of applying a knowledge specification phase with which terminographers can determine which “linguistic words/patterns are considered terms given the applications, users and purposes of the terminology base”. (Kerremans et al. 2005: 95) They also indicate that in a multilingual project, the framework for specifying knowledge must be language-independent knowledge; this signifies that it is based on experts’ feedback for informing a terminographer of the necessity of adding categories to the framework and terms to the terminology database so as to fill a gap within the scope of the purpose defined. At a later stage, the categorisation framework can be enhanced with information that is culturespecific. The second stage is the information gathering phase, which is carried out under expert guidance. Experts in the field point terminographers to relevant domainspecific textual material or websites where they can search with a view to compiling and managing a corpus of domain-specific texts. This search can be performed on the Internet or made available in electronic format through Optical Character Recognition (OCR). In the FF POIROT project, relevant texts were regulations, domainspecific texts, and websites. Thirdly, there is the search phase. Once the corpus has been established, Kerremans et al. (2005: 99) indicate that “the terminographer will be able to extract from the domain-specific texts, terms denoting categories in the categorisation framework as well as words or linguistic patterns indicating the intercategorial relations”. During the search phase the compilers of the FF POIROT project found that the concept of term has to be adjusted to the characteristics of the domain. For instance: In the case of law texts, we considered a term to be any word or expression used to express a category in the legal rule that needs to be represented in the ontology of financial forensics. (...) This means e.g., that from the first rule in article 41 of the Legislative Decree 58 (Section 3.3) – i-e- “[a]sset management companies may market investment fund units and client-by-client portfolio management services abroad” – the following terms were extracted: ‘asset manage-

116 | A Critical View of Terminography ment companies’, ‘investment fund units’ and ‘client-by-client portfolio management services’. This term extraction method sometimes obliged us to extract terms which one would never encounter in a traditional terminological dictionary. (Kerremans et al. 2005: 100)

The fourth phase is the refinement, verification and validation phase. The purpose of the first action is to “further complete the terminological database by for instance: aligning those terms that are equivalent, specifying the co-texts or concordances in which terms occur as well as the reference to the source from which each co-text was extracted”. (Kerremans et al. 2005: 101) Then comes the verification phase, which refers to the process by which the terminology base is checked for consistency, i.e. a final check “to see whether the content of the terminology base really meets the requirements specified in the ‘analysis phase’”. (Kerremans et al. 2005: 101) We believe that Temmerman and her colleagues’ sociocognitive approach to terminology is to some extent out of touch with real lexicographical problems and potential solutions to these, perhaps because they are not considering that reference works are tools. For instance, most terminology users, such as translators and students, cannot put the first principle into practice: how can translators relate the sense of a term to the idea they have in their mind when, while translating a specialised text, they may have little or no proper knowledge of the ideas being discussed? How can they relate the sense of the term with the object in the real world when perhaps they do not even know that such an object (or function, relation, etc.) exists? Similarly, the remaining principles are taken as gospel truths and rest on generalizations that can be easily disproved. For instance, there are clear-cut concepts and categories for some users, even if they do not resemble prototypes. By way of example, Eskimos can easily classify the different types of snow, describing them in detail and with certainty without previously “categorising” a snow prototype. This unrealistic situation may explain a surprising fact: termontography is being used in several EU-funded projects, e.g. FF POIROT, PoCeHRMOM, and PROLIX; however, neither of these projects have produced real specialised information tools (at least, we have not been able to discover any), although generous funding has been received, e.g. around 1.6 million euros, for FF POIROT.

7.3.3 Martin, Faber and frame-based terminology Humbley (2008: 261) claims that the lexicographer Willy Martin is the instigator of the approach to terminology using frame semantics. Martin (1998, 2001, and 2006) proposed frames as a model to define terms at the Professional Communication and Knowledge Transfer conference held in 1998, in which he argued that frame-based definitions offer the following advantages:

Terminology and linguistics | 117

  

greater consistency (terms belonging to the same concept type can be defined in the same way); more flexibility (the same or similar terms can be deliberately defined in a different way, depending on the different knowledge of the intended user); greater completeness at the level of representation (concept tokens can inherit from different types and so generate maximal frames). (Martin & Van der Vliet 2003: 347)

Martin’s approach basically rests on the conception of frame semantics espoused in Fillmore (1976, 1982 and 1985) and Fillmore & Atkins (1992). This approach maintains that the meaning of a term derives from context, which is presented as a dynamic notion containing both external factors (situational and cultural) and internal cognitive ones, which offer an encyclopaedic view of each language element. Hence, each term must be defined in relation to the whole conceptual scenario or semantic frame. This definition is shaped through information extracted from real data, typically that of a corpus, and is worded in a way that makes “visible” the conceptual scenario underpinning a particular term as well as its related senses and semantically related terms. Discussion on the merits of the frame-based terminology approach gained momentum at the beginning of the 2000, e.g. at the conference on terminology, computing, and translation held in Wales in 2004 (Ten Hacken 2006). The basic idea of this approach is that frame semantics offers a solution to practical work on terminology, especially when problems of polysemy and lexical representation are dealt with. Furthermore, the methodology developed in projects such as FrameNet can be replicated to describe the meaning of terms in a network of concepts (Faber et al. 2005, 2006 and 2007; Van der Vliet 2006). Terminologists should study terms (and words) through a detailed analysis of corpus data, describe the frames or conceptual structures which underlie them, examine sentences extracted from a large corpus of specialised texts that contain these terms and words, and record the way in which information from associated frames are expressed in the sentence. Van der Vliet shows an example for terminology management. The above steps must be used when constructing a conceptual system or frame (3): Gastritis ISA disease M-AFFECTS stomach F-AFFECTS metabolism ORGANISM patient ETIOLOGY infectious Example 3: The frame for gastritis. Source: Van der Vliet 2006: 61

Van der Vliet claims that using frame-based concept systems as a basis for the description of terms has advantages in terminology management. He mentions advan-

118 | A Critical View of Terminography tages concerning the acquisition of terms, the representation and application of domain knowledge and the associated terms: Concerning the representation, using frames leads to an explicit link between the (language based) terms and the (relatively language independent) domain knowledge. (...) The acquisition can profit from frames because of the fact that the frames contain coherent domain knowledge. This knowledge can guide the acquisition of terms and collocations, since it more or less predicts the terms in the field; (...) At the level of application a database with frame-based description of domain knowledge offers better retrieval possibilities. The database can still be entered using the form of terms, but one can also use the slots and the fillers to search for individual terms or sets of terms. Frame-based knowledge representation can also be used in various applications of the computational processing of natural language. (Van der Vliet 2006: 61-62)

For instance, Van der Vliet mentions an advantage of representation: frames like example (3) provide the material for one or more definitions: Gastritis is a disease which effects (sic) the stomach and the system of metabolism of a patient and is caused by an infection. The frame provides the possibility to base a definition on a selection out of the represented knowledge. This leads to more flexible, tailor-made definitions. (Van de Vliet 2006: 61)

In addition, frame-based terminologists argue that a conceptual network for terminology should also contain explicit information on combinatorics, i.e. construction patterns visible as multiword terms (Van der Vliet 2006: 66). They are seen “as a combination of slots waiting for their appropriate fillers” (Humbley 2008: 261), as evidenced in medical language, where these patterns are very productive. For instance, the pattern “DRUG + therapy” may become slightly more complex: “DRUG + [multi/combination/mono]therapy” (Van der Vliek 2006: 67). More recently, some scholars have included Cabré’s multidimensionality as a fundamental principle of frame-based terminology (Faber, 2012). Faber (2009: 123), for example, indicates that frame-based terminology “focuses on (1) conceptual organization; (2) the multidimensional nature of terminological units; and (3) the extraction of semantic and syntactic information through the use of multilingual corpora”. (Faber 2009: 123) Faber and her colleagues (e.g. Prieto Velasco & LópezRodríguez 2009; Reimerink et al. 2010) illustrate this new addition by indicating that the design and construction of specialised information tools must adopt a multimodal approach. The above-mentioned approach defends the argument that dictionary entries must contain more than frames and textual material. They may, for example, contain visual information:

Terminology and linguistics | 119 Frame information helps users to better understand the conceptual structure of the general event, and therefore, the specialized domain. Textual information, selected according to knowledge richness, shows the syntactic behaviour of the LUs [lexical units], and provides the user with the necessary information on how they interrelate within the same frame as well as with other frames. Finally, visual information, also selected according to criteria of knowledge richness, complements the textual information regarding the relation between processes and their linguistic designations. (Reimerink et al. 2010: 1948)

Visual material consists of (a) visual object(s) that the terminographer associates with the concept/term in the information tool “in order to activate previous knowledge in the user’s mind, as well as the new knowledge potentially transmitted by the lexis of the definition”. (Prieto-Velasco & López-Rodríguez 2009: 181) This means that the visual material should be coherently linked with other data, focus on significant aspects of definitions, highlight the most relevant conceptual relations, and reflect the ontological organisation of the domain structure as well as adding substance to vague contexts. To sum up, images can be selected to: a. encode necessary parts of the definitions which can only be described graphically; b. provide supplementary data corresponding to a certain definition; c. represent similar terms that non-experts understand as synonyms because they are conceptually related; d. depict the different facets of a term belonging to one more than one subject area; e. indicate the level of expertise of the intended users; f. restrict the meaning of a term or to disambiguate meaning, in the same way as the linguistic contexts do. (Prieto Velasco & López-Rodríguez 2009: 193)

We believe that the tenets of frame-based terminology rest on very unstable foundations. In addition to the comments relating to Cabré and Temmerman's views (cf. sections 7.3.1. and 7.3.2.), we add two more in this section. The first is that framebased terminologists are not offering a solution to practical terminological work. For instance, in the MacMillan English Dictionary for Advanced Learners (2nd edition), water is a noun with five different definitions and a verb with three different definitions, whereas in EcoLexicon (cf. section 8.14.2), water is only a noun with a general definition (example 4). water: colourless, tasteless, and odourless substance that is essential to all known forms of life and is the most universal solvent. Example 4: Definition of water in EcoLexicon.

The EcoLexicon definition does not mention that water is a chemical substance, nor does it give its formula or composition. It does not say that it is a liquid at temperature above 0ºC at sea level, but a solid or a gas at temperature below 0ºC, etc. Briefly, the definition of water is conceptually poor, although definitions are regarded as mini-knowledge representations or frames in frame-based terminology.

120 | A Critical View of Terminography The second is a critique of the notion of corpus used in Corpus Linguistics. This notion rests on assumptions that are irrelevant in specialised texts, e.g. the notions of frequency and representativeness (Fuertes-Olivera 2012). Linguistics-based terminologists have yet to answer why a frequent term rather than a relevant one should be included in a specialised dictionary (cf. sections 3.4 and 6.4.). Furthermore, corpus data are mostly used for selecting one- or two-word lemmas, which are rare in specialised languages. For example, almost 70% of Spanish accounting terms are of the multi-word type, many of them being composed of three or four orthographic words (cf. section 9.3.3.). Such terms will not be easily identified through corpus methodology, nor could the data be used for describing the grammatical behaviour of a large body of them; e.g. many accounting terms with method are only found in the singular and either with or without the definite article (cf. section 9.3.3.). To sum up, the frequently quoted idea that linguistically-conceived corpora offer the best methodology for producing a specialised dictionary is questioned by the function theory, which reinforces its position by indicating that corpus data cannot be used with confidence without actual work by experts on the dictionary project, i.e. without experts working side by side with specialised lexicographers (cf. chapter 9). In conclusion, the linguistics-based approach to terminology agrees on several arguments in the design and preparation of specialised dictionaries. These arguments share a basic criticism of Wüster’s approach to terminology: – Concepts are no longer considered universal and independent of cultural differences. This means that terminologists must no longer prioritise concepts over terms. This principle runs counter to the basic idea of traditional terminology, i.e. making scientific and technical communication exact hinges on representing and describing culture independent concepts. If terms are to be on a par with concepts, terminologists must incorporate a description of culture-dependent concepts and their terms. – It is almost impossible to draw a line between general knowledge and specialised knowledge. Hence, there is no consensus on where the concept selection process should begin or end. In fact, the decision on whether or not a particular term should be included in a terminological information tool is almost a question of faith. Terminologists must deal with terms, semi-terms, i.e. “non-subject specific specialised terms, as they are used in more than one domain and constitute most of the vocabulary of a specialised language in a discipline” (Gómez González-Jover 2006: 221), and general words. – Both concepts and terms occur in context. They are no longer in a vacuum; rather, they are part of a culture and a specific text or discourse. Consequently, concepts and terms are to be selected from real data, e.g. from corpora, and analysed with empirical data. The analysis of real data will show that terms are a means of communication and expression “and according to these two variables the discourse will be marked by redundancy, conceptual and synonymic varia-

Terminology and knowledge engineering | 121

tion and, in addition, permit the observation that there is not always a perfect equivalence between languages.” (Cabré 2003: 178) Communication has diversified scenarios, and Wüster and his followers have misunderstood most of them. For example, Cabré (2003: 177) claims that [A] theory can never be prescriptive because a theory is a unit of coherently integrated axioms or essentials which permits the description of an object, its properties, its relations and operations within a specific framework. The theory developed by Wüster is not prescriptive but rather descriptive, even though the data he described were not representative of the full range of terminology and hence the theory he inferred from these data was biased.

However, the specialised dictionaries designed and compiled by linguistics-based terminologists are not different from traditionally-conceived ones, as shown in chapter 8 (cf. sections, 8. 10; 8. 11; 8.12; 8.13; and 8.14). What makes them different is that they, e.g. Cabré's Genoma and Faber's EcoLexicon, have embraced data modelling, usually termed knowledge engineering, the new buzzword in the field of terminology.

7.4 Terminology and knowledge engineering For the purpose of this book, knowledge engineering is the set of activities that are mainly concerned with building, maintaining and developing knowledge-based management systems, usually through (semi-) automated methods. Clark et al. (2012: 555-558) present an overview of these, and indicate that methods for building domain knowledge structures from text collections are undertaken in two ways, known as data-driven and knowledge-driven approaches. Data-driven models are statistical models that are based on extracting key words or phrases which capture concepts, and employ three main data-driven algorithms: unsupervised learning, weakly supervised learning, and supervised learning (Fig. 7. 1.).

Figure 7. 1.: Data-driven domain modelling. Source: Clark et al. 2012: 558

122 | A Critical View of Terminography Unsupervised learning takes raw data with which to learn a model. This is a very active research area that tends to be based on frequency analysis, formal concept analysis, mining query logs, clustering text, and adaptive filtering. Weakly supervised learning requires some annotation in order to learn a model. Phrase and/or syntactic patterns are typically identified empirically and used on large corpora “to harvest entities satisfying the pattern.” (Clark et al. 2012: 559). The algorithm is generally based on clustering concepts, relation extraction, and ontology concepts enrichment. Supervised learning methods require pre-labelled training, e.g. lexical and syntactic association patterns (cf. Clark et al. 2012), for a review of these approaches as well as their main advantages and disadvantages). Knowledge-driven domain modelling focuses on building domain models from knowledge sources that can be fully structured (e.g. WordNet) or semi-structured (e.g. Wikipedia). Clark et al (2012: 561-563) claim that there are two main lines of research in this field. One of them is the use of explicit knowledge sources “to general applications (e.g. word sense disambiguation), the creation of large scale knowledge bases (e.g. YAGO), and the extension of general lexicons (e.g. WordNet) with semantic relations (e.g. Navigli 2009a; Pennacchiotti & Pantel 2006)”. (Clark et al. 2012: 561) The other line of research consists of enriching existing domain models, especially domain ontologies, and creating terminological knowledge bases. Regarding terminology, a review of the literature indicates that the relationship between terminology and knowledge engineering has comprised two main sets of activities, although the line of demarcation between them is blurred. Firstly, researchers have developed methods and created software tools, some of which can be downloaded from the Internet freely, for extracting candidate terms, either from ad hoc specialised corpora or from the Internet as a corpus. The main objective of this line of work is automated term-extraction and structuring, which is presented as an initial and necessary step for designing and constructing (semi-) automatic dictionary articles. (Kageura & L’Homme 2008) This line of research usually uses parallel and/or comparable linguistic corpus, that is, corpus compiled according to the criteria described in Corpus Linguistics, e.g. Sinclair (1991) and Bowker & Pearson (2002), for term extraction. Most ad hoc corpora include research and review articles from specialised journals, texts from manuals and handbooks, and popularised texts, typically from newspapers, all of which are usually constrained in terms of authorship, geographical variety, location of publication, etc. Subsequently, the corpus is processed by means of software tools and the researchers choose several extracted candidate terms and extracts concordances of each one, for instance, by using random sampling in the concordancing feature of WordSmith Tools (Scott 2008). After this, they evaluate the results of their work by comparing what they have extracted (semi-) automatically with results obtained manually or those already encountered in existing repositories.

Terminology and knowledge engineering | 123

Regarding parallel corpora, i.e. corpora from previously translated texts, there are several tool suites for extracting multilingual terms, e.g. the GIZA++ statistical machine translation toolkit (Och a& Ney 2003) and the iTools suite which performs single- and multi-word alignment, and includes graphical and interactive tools (Merkel & Foo 2007). Rocheteau & Daille (2011) claim that these tool suites have several drawbacks, insofar as parallel corpora are in short supply, are available only for certain major languages, and are non-existent in several domains. Rocheteau & Daille (2011: 9) argue in favour of using comparable corpora, namely, “corpora that are formed by texts in different languages that are not translations of each other, but deal with the same or similar issues”. They illustrate their proposal with the TTC TermSuite, which is (p. 9) “the first tool suite for the multilingual extraction of terminology from comparable corpora”. They also describe the functional architecture of TTC TermSuite, which is divided into four steps. Step 1 requires text pre-processing, performed by a set of tools, e.g. web crawlers, text extractors, etc., which provide a clean textual content without any linguistic information. Step 2 is carried out by tools such as word tokenizers and part-of-speech taggers, which are used for linguistic analysis. Step 3 requires processes of “singleword term (SWT), multi-word term (MWT) and morphological compound detection” (p. 10), as well as term variant processing such as acronym detection. Finally, step 4 aligns bilingual terms. Up the time of writing this book, a demonstration of the workings of the TTC TermSuite was planned for two pairs of language: FrenchEnglish and English-Chinese. There are also several methods for extracting candidate terms from corpora. Vintar (2010: 142-144) offers an overview of term recognition methods, which she identifies as traditional and innovative. Traditional approaches to automatic term recognition (ATR) exploit the distributional features of terms, employ morphological patterns, or combine both approaches. Their aim is to obtain a list of candidate lexical units and weight their termhood, i.e. the condition of being a real term in a specific domain. Analysing the behaviour of lexical language patterns in knowledgerich contexts, i.e. ones in which there are segments of text that can be useful for achieving conceptual information (Meyer 2001), is an example of a traditional approach to term recognition. Lexical language patterns are knowledge patterns that “prototypically take the form X + MARKER + Y, in which X and Y are the expressions of related concepts, and the MARKER indicates the relation that holds between them”. (Marshman 2008: 125127) In other words, such patterns contain a lexical unit or a series of lexical units that will serve as a marker of a relationship between two concepts, as seen in the following lexical language patterns: (Marshman 2008: 125) - GENERIC-SPECIFIC relations codified as X is a Y (e.g. lung cancer is a disease); - CAUSE-EFFECT relations represented as X causes Y (e.g. radiotherapy causes side effects ); - PART-WHOLE relations indicated as X is part of Y (e.g. cost accounting is part of accounting);

124 | A Critical View of Terminography - ASSOCIATION represented as an observed co-occurrence of two factors (e.g. managerial accounting and financial accounting).

Innovative term recognition methods have extended the orthodox combination of statistics and linguistic term properties, and have included semantic properties of term and term variation as novel methods with which terms can be recognised and extracted automatically. As an innovative approach, Vintar (2010: 145) classifies her bilingual term extractor for English-Slovene, LUIZ, which is presented as “a hybrid term extractor that uses parallel or comparable corpora as input and outputs monoand bilingual lists of term candidates”. LUIZ is especially innovative in its use of the so-called bag-of-equivalents approach, i.e. the source word can potentially be assigned several translation equivalents, each of which is assigned a probabilistic score. She indicates that this approach has advantages regarding the handling of term variations and multiple translations, which makes it more precise than other approaches. This line of research is in its infancy and, as a result, will not be looked at in much detail in this book. Most of these projects are only pilot studies. Only a handful of them have been employed when making specialised dictionaries and still have a very limited use, as shown in Vintar’s quote relating to the use of LUIZ for making a printed specialised dictionary in the military domain. She indicates that the users of LUIZ were a team of professional terminographers who first compiled parallel, comparable and monolingual corpora of military terms, and then actively cooperated in the fine-tuning of the system. They found that the number of term candidates extracted was small, and that ATR systems are more useful as auxiliary tools, e.g. for finding collocations, than as primary tools for term extraction. She also mentions that low recall is still commonplace and that these systems need a lot of funding, which casts doubts on its use in most projects, which are typically underfunded: The number of term candidates that were initially selected for inclusion in the dictionary was — unsurprisingly — relatively small compared to all terms extracted, however the feedback of the dictionary makers was very positive. Instead of having to scan through huge amounts of data manually or using concordancing software, they were rather able to invest more time into the conceptual representation of the field and the descriptions of dictionary entries. Also, their impression was that by using ATR better domain coverage could be achieved in the dictionary. It seems that ATR can serve as an effective support tool for terminographers, whereby precision – within reasonable limits – does not play a pivotal role. Low recall on the other hand is a problem, because if the terminographer is nevertheless forced to manually scan the corpora for less frequent that the system had missed, there is little to be gained from ATR in the first place. It should however be noted that although in this particular ATR task the provision of specialized corpora had been tackled beforehand by the dictionary-making team and did not in itself pose a problem, in part certainly due to sufficient funding, in other domains this may not be the case. In fact many terminographic projects possess neither the resources nor the technical

Terminology and knowledge engineering | 125 skills to compile a representative specialized corpus, much less a bilingual one. (Vintar 2010: 152-153)

The second line of research deals with modelling knowledge based on semantic ordering principles, with the aim of translating it into the construction of ontologies and other information tools, typically referred to as lexicons, terminological knowledge bases, taxonomies, concept systems, thesauri, and glossaries. (Wright 2007) This line of work was initially proposed in the early 1990s, e.g. in the work of Meyer et al. (1992). At that time, terminologists started to claim that the volume of texts that may be used in terminology and terminography work is so vast that computer tools are required for carrying out knowledge acquisition, conceptual description, creation of concept systems, formulation of definitions, and establishment of equivalence between terms. (Marshman 2008) This line of research is broader than the first, as the results reported are not primarily concerned with terminology, but with domain modelling, i.e. the process of capturing and structuring knowledge embedded within a selected domain (for example, a collection of documents) in several disciplines, of particularly relevance being the research results reported in three fields: artificial intelligence (AI) (especially the Semantic Web (SW)), natural language processing (NLP) and information retrieval (IR) (Fig. 7. 2.).

Figure 7.2.: Examples of domain modelling in various disciplines. Source: Clark et al. 2012: 553

Work in this field of research can be split into several areas, depending on the level and degree of automatic methods used for constructing knowledge representation resources, i.e. tools that contain sets of concepts which are usually arranged in a way that reflects some kind of internal structure. For the purpose of this book, this section aims only to present results that can be applied to the design and construction of specialised dictionaries. Consequently, bibliographic retrieval projects (e.g. Medline) will not be analysed in this book, although this project is a well-known example of a knowledge representation resource. Instead, we will focus on know-

126 | A Critical View of Terminography ledge representation resources that are presented as terminological tools. They are usually constructed through semantic relations, the so-called “pattern-based approach” (Hearst 1992) being particularly well known. Auger & Barrière (2008: 2-3) define this approach as a “bottom-up” methodology which is characterised by two assumptions: – the target relation is a specific (named) relation; – this relation is explicitly expressed in text between words or lexical units. They also summarize the methodology underlying this approach, which comprises four main steps: – defining the semantic relation of interest, – discovering the actual patterns which explicitly express this relation in text as well as the syntactic conditions by means of which the meaning of the targeted relation is conveyed, – searching for instances of the relation using the pattern, – structuring the new instances as part of a new or existing ontology (or terminology database). Within this field of research, there is a vast array of projects, two of which will not, however, be discussed in this book: – Projects aiming at constructing ontologies for NLP (these are typically designed for being used by machines); for instance, we will not analyse mined dictionaries, i.e. those that are constructed automatically by mining the web: The recent development of the Internet has attracted the interest of a number of researchers in studying mining bilingual dictionaries from the Web, because more and more multilingual resources with high quality can be freely obtained (e.g. Wikipedia), and new words in different languages usually emerge first on the Web. (Ye et al. 2012: 2474) An empirical observation one would make about the Wikipedia corpus is that the related words in different languages are likely to co-occur with the same concepts or highly related concepts. (...) The above observation motivates us to devise a graph-based approach to mining a highquality cross-language association dictionary by considering the co-occurrences of related words in the description of the same concept in different languages. (Ye et. al. 2012: 2477)

Projects aiming at constructing teaching collections in which the constructed ontology does not aim to assist users with concrete information needs, but with systematic ones (i.e. the ontology is similar to a teaching package). For instance, Boyatt & Joy (2010: 6) report on the EU funded MALog project. This project is building an ontology to support the development of high-quality learning materials in mathematical logic: One of the primary aims of the MALog project is to produce pedagogically high quality learning material units to help develop mathematical logic knowledge and skills. The learning material

Terminology and knowledge engineering | 127 will consist of all necessary theory, example problems, exercise sheets and, where appropriate, exam questions. Also included within the learning material will be appropriate metadata including learning goals, prerequisite information and ontology data.

Instead, this book will analyse ontologies that are part of so-called terminological knowledge bases (TKBs) (cf. section 8.14.). Such ontologies are systems that aim to store and retrieve data which are used for creating and displaying relationships between units of information, data accessed through a web page, usually in the form of graphic user interface (GUI). They are commonly referred to as terminological knowledge bases, and it appears that most of them are constructed by the use of three types of relationship: generic-specific; whole-to-part, and associative relations, such as concept, text, and text or source relations. Google searches and queries in several academic databases, e.g. the ISI Web of Science, MLA and Scopus, show two main findings. The first is that there are a large number of projects aiming at constructing terminological knowledge bases, usually EU-funded projects. The second is that many of these projects have resulted only in prototypes; in other words, we have not been able to find the tool these projects aimed to build, e.g. De Baer et al. (2006). An analysis of the work of the above-mentioned authors is sufficient to signal the main characteristics of this line of work. Firstly, researchers typically use generic-specific, whole-to-part, and associative relations for designing and constructing their TKBs. De Baer et al. (2006: 814), for example, claim that the genericspecific and whole-part relations “may be used to group related information together and to identify and retrieve this information by means of an abstract group name”. In addition, they claim that the concept relations are adequate for creating a formal representation of the domain “by means of the domain concepts and their inter-categorical relationships”, that term-relations are useful for relating terms within a concept, those belonging to different concepts, or ones which belong to different categories, and that text or source relations are necessary for creating hyper-links that “refer to other texts”. Secondly, these systems are typically equipped with descriptions of search engines that are not operative. For example, De Baer et al. (2006: 817) mention that their system is equipped with the KBExplorer. This is presented as a knowledge discovery tool, which “could be used to browse through the resulting knowledge base in a flexible and powerful way”. To the best of our knowledge, the KBExplorer is inoperative; that is to say that we were unable to find it through Google searches nor queries in academic databases (e.g. ISI Web of Science, Translation Studies Bibliography, MLA, and Scopus). Knowledge engineering is often referred to when proclaiming the often-quoted association between terminology and technology. Bowker & Marshman (2009: 62), for instance, claim that the primary purpose of terminological knowledge bases (they refer to them as term banks) “was to help store and disseminate terminological

128 | A Critical View of Terminography data”, and that these “are still in active development and remain valuable tools for language professionals”, especially because they are on the internet and “electronic texts can be consulted more quickly and easily than printed texts”. These quotes take for granted the association between easily accessible and quality, a relationship that does not always hold true, as will be shown in the next chapter. Kageura & L’Homme (2008) seem to warn against this view in their editorial report on 15 years of research published in Terminology, in which they affirm what they consider the raison d’être of the journal: 

Connecting terminology to its related domains: The status of terminology as a field of study is twofold. On the one hand, as was said in the previous editorial statement: “terminology is a field of study with its own applications and a theory to accommodate them.” On the other hand, terms and terminology are at the center of many applicational, practical and theoretical issues in a variety of domains, ranging from translation, computational linguistics, language teaching, and information retrieval to business management, pharmaceutics, medicine, etc. The communication between these different groups is of the utmost importance. Bridging theoretical and applied issues of terminology: One of the unique features of the study of terminology is that its theory is often intimately related to its applications, perhaps much more so than in linguistics. Now that the field has matured, there is, however, a danger of researchers engaging in theory for the sake of theory per se, which could lead to a decline in research dynamics. The significance of constant interaction between theory and practice as well as the feedback from applications to theory and vice versa cannot be overstated. Putting local issues in a global context: Though terms are basically language- dependent, given the rapid growth of international communications, multilingual issues are becoming increasingly important and unavoidable. In this context, it is important to put local and language-dependent questions within a global context and essential to understand and share local and language- dependent experiences globally. (Kageura & L’Homme 2008: 153154)

7.5 Summary Following the line of Kageura & L’Homme, we also believe that terminology theories are totally justified if they can be put into practice, i.e. if they result in terminographical products for meeting users' needs in potential situations. To the best of our knowledge, this connection, namely, the translation of terminological theories into real and working terminographical products, has so far left a lot to be desired. Our next chapter will investigate this impression. It will analyse some examples of specialised dictionaries with two main objectives. The first is to discover the theoretical basis on which the tool was created, and the second is to evaluate its quality in terms of the criteria we have been defending in chapters 2 to 6: specialised tools must be based on a sound theory that takes into consideration both the nature of information tools and information costs, that is, the time and money needed for designing, creating and updating them.

8 An Analysis of Specialised Online Dictionaries As shown in section 2. 2., the production of specialised information tools is an activity with a very long history. This has resulted in different terms being applied (e.g. dictionary, encyclopaedia, encyclopaedic dictionary, lexicon, vocabulary, glossary, term bank, thesaurus, terminological knowledge base, etc.) and different means of constructing them, e.g. there are both printed and electronic dictionaries. In this chapter, we will review several reference works that share two characteristics. The first is that their compilers present them as specialised reference works which have been designed for consultation purposes. The second characteristic is that they are all accessed through the Internet. We acknowledge that the Internet has such great potential that the future of lexicography – and this book aims to offer a theory and practice of specialised online dictionaries – rests on both maintaining the core nature of lexicography (cf. section 4.2.) and making use of the technologies the Internet has brought into our lives (cf. chapter 6). Briefly, this chapter aims to review the current online editions of several specialised dictionaries: – Dictionary of Business and Management (Law 2012). – New Palgrave Dictionary of Economics Online (New Palgrave Online 2012). – Musikordbogen (Bergenholz 2011). – Glossary of Mortgage and Home Equity Terms (Wells Fargo 2012). – Cambridge Business English Dictionary (Cambridge Business 2012). – TermFinder (Peters et al. 2011). – Business English-Spanish Glossary (Miles 2012). – Glossary of FAO Database and Information System (FAO Glossary 2012). – Kicktionary (Schmidt 2009). – InterActive Terminology for Europe (IATE 2012). – CercaTerm (Termcat 2012). – UNTERM (United Nations 2012). – Multilingual Glossary of Technical and Popular Medical Terms in Nine European Languages (Multilingual Glossary 2010). – European Dictionary of Skills and Competences (Disco 2012). – Banc de Coneixement sobre el Genoma Humà (Genoma 2003). – Terminological Knowledge Base on the Environment (EcoLexicon 2012). The dictionaries were selected for exemplifying current practices in this field of research; the following characteristics have been observed: –

There are those which, for one reason or another, are worthy of imitation: the Musikordbogen and the New Palgrave Online. These two dictionaries are intended to meet the needs of well-identified users, employ existing Internet technologies favouring individualisation, are regularly updated, and pay attention to costs and means for achieving their basic objectives.

130 | An Analysis of Specialised Online Dictionaries –

Some have been designed and constructed by scholars who defend particular theories of terminology (cf. chapter 7). Hence, they are interesting for studying the relationship between theory and practice: the Genoma and EcoLexicon. There are dictionaries that are sequels to well-known printed dictionaries: The Dictionary of Business and Management, The Cambridge Business English Dictionary, and The Glossary of Mortgage and Home Equity Terms. These are suitable for studying the possible influence of the Internet on adapting printed dictionaries to the new media. Certain dictionaries are funded as national or international projects. These are run by national and/or international organizations and receive a considerable amount of funding: IATE, CercaTerm, UNTERM, The Multilingual Glossary of Technical and Popular Medical Terms in Nine European Languages, and The European Dictionary of Skills and Competence. These are suitable for observing whether the results merit the amount of money received, which, as we have said, is usually large and comes from taxpayers. There are those that are designed and produced as “innovative projects” and run by individuals or groups of individuals, usually with little funding. They merit our attention as there are many scholars who aim to construct these tools and discuss their “prototypes” in international conferences, journals and symposia: Kicktionary, Business English-Spanish Glossary and TermFinder.

These dictionaries are studied under a common framework of analysis, which is explained below.

8.1 A functional framework to dictionary criticism: A list of criteria for reviewing specialised online dictionaries Following Wiegand (1984), some scholars regard dictionary criticism as an integral part of a theory of lexicography (Gouws 2012). This means that for scholars such as Gouws, dictionary criticism is the part of metalexicography that is concerned with evaluating the quality of one or more dictionaries, typically by reviewing them or by studying their use on the basis of one or more lexicographically relevant criteria (our emphasis). As we also espouse this view, we claim that dictionary criticism will benefit if it is undertaken within a theory of lexicography, a thing which does not often happen in most reviews of dictionaries and studies in dictionary use published so far. In these, we have observed the following main deficiencies, which have been extensively discussed in recent lexicographical publications, e.g. Bergenholtz & Johnsen 2007; Chan & Taylor 2001; Nielsen 2008, 2009a and b; Swanepoel 2008; Tarp 2009b; Welker 2006 [2010]:

A functional framework to dictionary criticism | 131

– –

Dictionary criticism is mostly undertaken by linguists/teachers who do not target the majority end-users of dictionaries, i.e. the users themselves; instead, these authors write for other linguists/teachers. Dictionary criticism does not address lexicographical problems and possible solutions. For instance, it does not usually discuss the updating process (if there is one), which is crucial in specialised lexicography. Dictionary criticism is mostly done without any explicit statement of its purpose. The evaluation process is confusing and typically lacks a consistent list of criteria or guidelines, e.g. evaluation criteria that are based on a theory of lexicography and are applied to the book(s) in a systematic and consistent way. This gives rise to the publication of reviews and/or studies in dictionary use, most of which are methodologically deficient and lack academic consistency (cf. section 5.1). Similarly, an analysis of most dictionary reviews shows that for many scholars “incidental sniping (takes) the place of any real exploration of the intentions with which the works being criticized had been set up.” (Osselton 1989: 229; cited in Hartmann 2000: 194). Hartmann’s (2000: 194-196) rejoinder to Urdang’s (2000) review of the Dictionary of Lexicography is in line with Osselton's quote and may explain why the academic activity of reviewing dictionaries is losing ground to other academic activities. Hartmann enumerates three main problems in Urdang's review: 1. 2.


90% of the review only covers 2% of the entries included in the book. Urdang uses evaluative adjectives, e.g. “pedestrian”, “amateurish”, and “inaccurate”, for describing the definitions used in the Dictionary of Lexicography. These evaluations are neither explained nor exemplified at all. Urdang’s review is not very professional: it typically uses circular arguments and does not focus on central issues in lexicography, for instance, on finding a right balance between terms from lexicography proper and those from neighbouring disciplines such as linguistics and computing.

Authors of dictionary criticism tend to highlight certain dictionary components, whereas they do not usually pay attention to others. For example, reviewers tend to concentrate on studying language matters - typically, morphological, phonetic, phonological, syntactic, pragmatic and semantic information - and do not consider other key lexicographical components, e.g. the user's guide or introduction. Authors of dictionary criticism are unsure about drawing conclusions, especially negative ones. This might be the result of dictionary criticism being primarily factual and descriptive rather than evaluative, focusing on and describing the corpus on which the words are based (i.e. where they originate), the presence or absence of specific

132 | An Analysis of Specialised Online Dictionaries kinds of information (i.e. what is included), and the structure and presentation style of the entries given (i.e. how they are treated). (Chan & Taylor, 2001: 174)

In order to avoid the above deficiencies, we support the view that dictionary criticism must be based on a theory of lexicography, be mostly evaluative, focus on users' needs in user situations, pay attention to the dictionary in its totality and complexity, and make use of a list of criteria that guide the process of dictionary criticism; the latter must have two main interrelated and reinforcing goals, i.e. one of the objectives gives feedback to the other and vice versa. The first goal is to help users understand the merits of the book(s) for satisfying their needs in one or more user situations. The second is to support lexicographers in the design and compilation of reference works aiming to meet previously identified user needs. Gouws (2012: 460) partly concurs with our view when he mentions that dictionary use is a recurrent topic in International Journal of Lexicography (e.g. McCreary & Dolezal 1999; Scholfield 1999; Rundell 1999, Al-Ajmi 2002; Lew 2011), adding that this "is in line with one of the most prominent features of modern-day lexicography, that is the user-perspective that puts the user, the needs and the reference skills of the user at the centre of all decision regarding the planning and compilation of dictionaries". Gouws' quote seems optimistic, although we have observed a change of attitude in the field of dictionary criticism published in mainstream lexicographic journals such as International Journal of Lexicography. For instance, Tono (2011) uses a qualitative methodology, enumerates caveats, and draws conclusions which seem to be in line with the view that dictionary criticism must be undertaken within a theory of lexicography that aims at a better understanding of dictionary reference skills: Dictionary reference skills are complex, integrated skills which involve language skills, knowledge about dictionary conventions, problem-solving skills, information processing skills among others. Users have to identify encoding/decoding problems in context, and make a decision about problem-solving strategies. (...) At every step, users have certain expectations about what to see or get. All the query results will be judged against the criteria of whether it will meet the expectations or not. Speculating on the process of dictionary look-up in this way might be useful to build up a theoretical model of dictionary reference behaviour, but actual look-up processes are far more complicated and unpredictable than this. (Tono 2011: 149) While the eye mark recorder is a powerful tool, the setting inevitably becomes artificial. In order to calculate gaze points accurately, it was necessary to fix the subjects' head onto the chinrest and ask them to look at the PC monitor, instead of real dictionaries. This could elicit some unnatural behaviour from the subjects. Also they worked on twenty-four different versions of the two entries, MAKE and FAST. Although the tasks were counter-balanced across subjects, there were bound to be some carry-over effects. The combination of target lexical information and the entries which contain the information needs to be carefully designed. Individual dif-

A functional framework to dictionary criticism | 133 ferences are limited, individual differences in response patterns are hard to control. Overall, the approach itself is quite promising, but a more solid design and its replications with different subjects or exploitation of a single-subject design would definitely be needed. (Tono 2011: 151)

Tono (2011) has applied eye-tracking technologies to analyse the process of dictionary consultation by learners of English as a foreign language. His work observes the behaviour of eight subjects who are subjected to an experiment that makes use of independent variables, moderator variables and dependent variables. Tono found that around one third of look-ups did not succeed, that menus and signposts do not help users as much as expected, that the location of the information influences the rate of success of the look-up, and that it is "sometimes misleading to compare groups with or without particular dictionary information". (ibid, p. 151) Briefly, Tono has found that the look-up process is complex and we believe that complex processes cannot be dealt with properly without their being given sufficient theoretical consideration. Our proposal regarding dictionary criticism is, therefore, framed within the tenets of the function theory explained in previous chapters (cf. chapter 5). It consists of an open list of criteria that constitute initial steps in the method used to determine user needs (cf. section 5.1). These criteria are based on the core of lexicography defended in this book (cf. section 4.2), the technological possibilities of the Internet (cf. section 2.3. and chapter 6), and the time and money needed to design and update working and running reference works for consultation purposes. In other words, these are standards that aim to help users and lexicographers understand the main characteristics of the work reviewed so as to guide them, either when they are consulting (as users), or when they are designing and compiling them (as lexicographers): – Author’s view: Does the dictionary include outer texts that inform on relevant lexicographical characteristics, e.g. the targeted user(s), and expected user situations the dictionary aims to cover? Similarly, have the compilers discussed the main characteristics of the dictionary in the lexicographical literature? With these questions, we emphasise how important the authors’ opinion is for building a theory of lexicography. – Function(s): Does an independent analysis of the dictionary match the information given in outer texts and lexicographic literature on the function(s) the dictionary aims to cover? In other words, do the data presented in the dictionary support the function(s) identified? The focus is on whether the data match the needs of the target group(s). – Access routes: Does the dictionary contain access routes that favour the process of consultation? Here we are considering whether the data are presented so that users can process them to get the information they need to solve their problems in a simple and easy way.

134 | An Analysis of Specialised Online Dictionaries –

– –

Internet technologies: Is the dictionary using existing Internet technologies? If so, is it using them adequately? Here we are focusing on whether the dictionary offers options that favour consultation, e.g. dynamic articles with dynamic data; this means that in the overall architecture (cf. sections 6.1, 9.2, and 9.3), we have a data collection (“database”) as well as selection, ordering and presentation filters towards user-/function-specific “dictionaries”. So-called “dynamic articles” result from the application of different filters, according to different user needs, applied to the data from the data collection. Lexicography: Does the dictionary make use of lexicographical theories and methodologies or is it based on other types of theories, e.g. Cognitive Linguistics, as suggested recently (e.g. Faber 2012)? Production costs: Are lexicographers using time and money in a sensible way? For instance, why do some dictionaries, e.g. DANTERM, have so much funding at their disposal when their results can be achieved with less? Information costs: Are lexicographers paying attention to the amount of time and effort users may need to look up, understand and interpret their findings? For instance, why do some dictionaries force users to search two or more times in order to retrieve data? In other words, are lexicographers paying attention to the distinction between comprehension-related and search-related information costs discussed in the literature, e.g. Nielsen (2008)? Updating: Is the dictionary being updated regularly, a must in specialised lexicography? Experts: Are real experts in the field included in the production team of the dictionary? This should be an obligation in specialised lexicography: it will increase the quality of the data included, for we believe that other methods of extracting specialised knowledge, e.g. non-experts working with corpus data, cannot be used for most lexicographical data; one example here is the definition given of motivation (Illustration 4, below): how can someone with corpus data craft such definition? Data selection: Does the dictionary use reliable sources for selecting and dealing with the lexicographic data included in the dictionary article? This criterion aims to assess the reliabilism of both the raw material included and its lexicographical treatment. The focus is on considering whether or not the data included and the way in which it is handled lexicographically is the result of a process that is documented and knowledge-based, i.e. data that have been validated by expert knowledge. To sum up, investigating the reliabilism of the data selected and its lexicographical handling is necessary and can be accomplished by, say, performing several random analyses in this regard in the dictionaries reviewed.

The Dictionary of Business and Management | 135

8.2 The Dictionary of Business and Management In the homepage of The Dictionary of Business and Management we access the main characteristics of the dictionary, which has been designed to provide quick information on traditional and recent topics in the domain of Business and Management: This wide-ranging and authoritative dictionary contains 7,000 entries covering all areas of business and management, including marketing, organizational behaviour, business strategy, law, and taxation. Written by a team of experts, it features the very latest terminology, for example, the recent vocabulary associated with structured finance and the associated subprime lending crisis, including collaterized debt obligation and special purpose vehicle. The new edition of this established bestseller dispels modern financial and management jargon, defining entries in a clear, concise, and accessible manner. It contains US business terms, general management concepts (e.g. competence, knowledge management), named theories (e.g. Tannenbaum and Schmidt, Blake and Mouton) as well as expanded coverage of the contemporary theory of the firm and human resources. New terms are included from the fastmoving areas of current affairs (e.g. MiFID), Internet business and information technology and there is full coverage of the new Companies Act. With recommended web links for many entries, accessible and kept up to date via the Dictionary of Business and Management companion website, this edition is more informative than ever. This A-Z reference work is essential for business students, teachers and professionals, and useful for anyone needing a guide to business terminology. Readership: Students on business and management courses at all levels; business professionals including lawyers, bankers, accountants, advertising agents, and insurance brokers; the general reader looking for clarification of everyday business terms (encountered, for example, in house-buying, taxreturns, or share investment. (Homepage)

An analysis of some of its dictionary articles, e.g. illustrations (3) and (4), as well as the information quoted in the homepage, e.g. the previous quote, and in Oxford Reference shows that the dictionary identifies its targeted users - experts, semiexperts and interested laypersons – and explains its main function: it is a cognitiveoriented dictionary that aims to assist users who wish to gain some knowledge about business and management, understood here in a very general way due to its covering management proper as well as law and taxation in a business and management context. The dictionary achieves its purpose: it has long and rather complete definitions (e.g. the definition of motivation, Illustration 4), which are separated by Arabic numbers when appropriate. The lemma list is arranged alphabetically, contains synonyms (e.g. ESOP) and English and American variants (e.g. “In the USA they are known as employee stock option plans.”). It also has internal (e.g. see) and external cross-references (i.e. links), and devices for making the search quicker (e.g. the link “Back to top”).

136 | An Analysis of Specialised Online Dictionaries

Illustration 3: Dictionary articles in The Dictionary of Business and Management

The Dictionary of Business and Management | 137

Illustration 4: Definitions in The Dictionary of Business and Management

In terms of the above functional framework, this dictionary seems to be based on lexicographical practices, makes use of experts, is updated from time to time, and assists any user who is in a cognitive situation and wishes to gain some basic knowledge about business and management. The data matches and supports the function identified and is adequate for the intended users and for the intended purpose of offering quick information about a specific subject field. However, the dictionary does not contain hyperlinks nor does it make good use of Internet technologies. It employs a kind of scroll-up menu, e.g. “Back to top”, and contains external

138 | An Analysis of Specialised Online Dictionaries hyperlinks, but its design, conception and technical characteristics are similar to those of printed dictionaries. For instance, synonyms and similar structures are not hyperlinked. This suggests that the dictionary is a sequel to a previous printed edition, i.e. a copycat or Faster Horse. The access route confirms this fact: it displays a menu containing the letters of the alphabet. When clicking on one of them, users retrieve all the terms that start with that letter. This primitive access system does not invalidate the value of this dictionary for gaining basic knowledge about a subject field, especially since the lexicographical cost of the dictionary is low, because the fact that it is well marketed signifies that little funding is required for designing, producing, and updating.

8.3 The New Palgrave Dictionary of Economics Online The authors’ view on The New Palgrave Dictionary of Economics Online is present in the dictionary and in economic literature: R. H. Inglis Palgrave's original Dictionary of Political Economy (1894-9) was a landmark in both publishing and economics: a liberal and scholarly overview of the whole sphere of economic thought in its day. Henry Higgs's revised edition, Palgrave's Dictionary of Political Economy (1923-6), retained the spirit of the original publication while embracing new concepts in the development of economics as a discipline. The four-volume The New Palgrave: A Dictionary of Economics, edited by John Eatwell, Murray Milgate and Peter Newman, was published in 1987 to international acclaim. Its scope had expanded and evolved almost beyond what Palgrave himself would have recognized, but the tradition of drawing together eminent contributors from across the spectrum of methodological and ideological schools produced not only an unsurpassed work of reference on the grand scale, but also many individual classic essays of enduring importance. Well into the 21st century it remained a standard work for economists in all spheres of the discipline and, as Palgrave described his original work, 'an almost unique example of economic cooperation'. In 2008, Palgrave Macmillan published The New Palgrave Dictionary of Economics, 2nd edition, edited by Steven N. Durlauf and Lawrence E. Blume. While some classic articles from the 1987 were retained, around 80% of the text was either entirely new or substantially rewritten to reflect the depth of change within the discipline between the editions. The new edition retains the inspiring tradition of bringing together the world's most influential economists writing in their own voice in their areas of expertise. Released simultaneously online The New Palgrave Dictionary of Economics is a dynamic, updated resource serving the information needs of a new generation of economists. (Homepage) The latest edition of this landmark work (...) has more than 1,750 articles, 1,500 distinguished contributors, and almost 6 million words. The subscription-based online edition incorporates quarterly additions, corrections, and updates. The list of contributors includes a “Who’s Who in Economics,” featuring Nobel Prize winners and leading authorities in their fields.

The New Palgrave Dictionary of Economics Online | 139 The articles cover every facet of economics from early philosophers such as Aristotle to classical thinkers like Adam Smith and Karl Marx to the modern masters of Keynes and Milton Friedman. Virtually every one of the many subfields of economics is fully treated, including experimental and behavioral economics, game theory, international economics, and financial economics. The specialized methodologies used by economists such as econometrics, linear programming, and Monte Carlo methods are also explored. Significant historical economic events are extensively covered in lengthy essays on the Price Revolution, the South Sea Bubble, the Great Depression, and others. The dictionary features many fine tools that facilitate the exploration of any topic: references to the most important scholarly literature, lists of related articles in the dictionary, an abstract with keywords at the head of each article, a table of contents outlining extended articles, and Journal of Economic Literature Classification Codes. For the serious study of economics, the New Palgrave Dictionary of Economics is without peer. Although its primary target audience is economists and advanced economic students, many of the articles are clearly written at a level that will inform the general reader and serious student. The pricing model of the online edition requires substantial yearly payments, making it unaffordable for most libraries. (McKay 2008: 139-140).

Our analysis of The New Palgrave Online shows that this is a cognitive-oriented online dictionary that permits a profound study of one or more subject fields, depending on the conception of Economics as single or multi-field. The data is reliable and matches the functions and needs of its targeted users, experts and would-be experts, e.g. students in the field. As indicated, the online version is a sequel to well-known and professionally acclaimed printed specialised dictionaries. This online version has kept pace with Internet technologies. For instance, each article has a Table of Contents that facilitates consultation and is a kind of scroll menu for moving up and down the dictionary article (illustration 5). This is a kind of Fast Horse dictionary that incorporates several search options, and shows some traces of Model T Ford dictionaries: there is a kind of customization device, “My Dictionary”, which allows users to save searches, bookmark articles and make notes for future reference.

140 | An Analysis of Specialised Online Dictionaries

Illustration 5: The “Table of Contents” and other data types in The Palgrave

The search system is located in a horizontal menu that consists of three options: “quick search”, “browse” and “advanced search”. The “quick search” option allows users to search by typing words or parts of them in the search engine (e.g. searching for crisis retrieves 210 articles), whereas the “advanced search” option allows users to retrieve in several sections, namely, “full text”, “bibliographies”, “article titles”, “contributors”, “abstracts”, “keywords”, “edition”, and “topic area”, i.e. the approximately 40 sub-fields into which Economics is currently divided. For instance, searching crisis in “bibliographies” and in “abstracts” retrieves 130 and 16 articles, respectively. The Palgrav Onlinee is similar to a peer-review article published in a journal, is accessed via subscription, and has been created by well-known scholars in the field, e.g. Nobel laureates. The editor(s) select(s) the scholars. The current online Palgrave version contains the full text of the 8-volume 2008 print edition, as well as upgraded access routes, interactive mechanisms and more relevant data. This dictionary remains a standard work for economists in all spheres of disciplines, is written by experts in their fields of expertise, and is updated quite regularly. It has a great deal of influence in its discipline. For instance, it is highly regarded as a means of preparing lectures, exploring lines of research, and understanding general developments in Economics. Each entry includes the author’s name, a long scholarly-written article, and an access structure that favours cross-references and interactivity. It also makes use of hyperlinks for internal and

The Musikordbogen | 141

external cross-reference. Furthermore, it contains several mechanisms for upgrading its scientific credibility: the authors of the articles are usually identified by name; users are informed of the date of each article and, if applicable, when this was modified; the entries are grouped by means of conceptual categories; an abstract precedes each article; finally, functions such as ‘how to cite the article’ and ‘download printed versions in friendly ways’ are provided. To sum up, this dictionary gives information concerning its characteristics, is based on sound lexicographical practices, and fulfils its main purpose, which is to assist experts and students in a cognitive situation where they need a profound study of one or more concepts in a subject field. The data matches and supports the function identified and is sufficient for the intended users and the intended purpose of offering in-depth information on Economics. It uses modern and well-devised search systems, each of which is adequate for quick and specific consultation. This project is very expensive and can be carried out only by well-established publishing houses, e.g. Macmillan in the case of the Palgrave, which must provide considerable funding and have a professional management running the project on a daily basis. Its adaptation process is continuous and it is the most cited and acclaimed reference work in Economics for punctual consultation, as revealed in several review articles on the Palgrave (e.g. in the Journal of Economic Literature and other well-known journals).

8.4 The Musikordbogen The authors’ view is described in several outer texts and authors’ publications: The music dictionaries intend to be tools for music students in universities and music schools, for both amateurs and professional musicians and for every interested person who wants aid by reading texts on music or who wishes to get further information on musical terms and topics. (Bergenholtz & Bergenholtz, 2013: 119)

Our analysis of this dictionary shows that the Internet has allowed individual scholars to design and compile multifunctional specialised online dictionaries that can be used in cognitive and communicative situations. For instance, the Musikordbogen is suitable in cognitive situations for both providing quick information about a specific subject and introducing its intended users to in-depth data concerning several relevant issues. It is also adequate in a communicative situation, especially for text reception and text translation, as shown in the previous quote. We believe that the Musikordbogen stands out on account of three main lexicographical concepts. The first is that it contains dictionary articles and a systematic introduction, i.e. a separate dictionary component that aims at providing help in cognitive-oriented and communicative-oriented use situations (Bergenholtz & Nielsen 2006). In the former situations, the systematic introduction provides an intro-

142 | An Analysis of Specialised Online Dictionaries duction to or a systematic, detailed presentation of the subject field covered by the dictionary, thereby fulfilling its cognitive function. In communicative-related situations, the component supplements the encyclopaedic information with language information in the form of collocations, auxiliary words and standard LSP usage. The Musikordbogen has a systematic introduction of several key music concepts, e.g. notation, illustration (6). This “simply offers a condensed description in an overview with many music examples. These music examples serve as illustrations linked to in many dictionary entries”. (Bergenholtz & Bergenholtz 2011: 196).

Illustration 6: Systematic introduction in the Musikordbogen (excerpts)

The second important concept is that the dictionary is the product of a significant change in lexicographical theory: the lexicographical database and the dictionary are separate concepts (cf. sections 9.1.) This means that the dictionary articles are different depending on the user situation in which the user is searching. In the Musikordbogen, users can retrieve data to assist them in four situations: reception, knowledge, production, and translation (Bergenholtz & Bergenholtz 2013). All these data are stored in a lexicographical database, comprising 4,015 cards with definitions, historical background, synonyms, references and links, pictures etc. (Bergenholtz & Bergenholtz 2013). The third key factor is that it is conceived as a Model T Ford dictionary: it provides quick access and creates dynamic articles with dynamic data (cf. section 2.3.). Furthermore, the access routes incorporate several options that offer monofunctional solutions in cognitive and communicative use situations (Bergenholtz 2011). In practical terms, this means that users access dynamic data in dynamic articles. For instance, if they are in a reception situation, such as a user who has bought a CD with Spanish music and wants an explanation of, for example, zarzuela, they re-

The Glossary of Mortgage and Home Equity Terms | 143

ceive a short definition or explanation (e.g. zarzuela is a “Spanish form of ballad opera with spoken dialogue”). Then again, the same user may be in a different user situation, for instance, one involving knowledge, in which they also need data on musical topics, genres, instruments, etc. Example (5) shows what a user receives when searching for zarzuela in this situation: Zarzuela Short explanation A Spanish form of ballad opera with spoken dialogue Elaboration Zarzuela was practised in the seventeenth and eighteenth centuries in connection with the fiestas or festivals of the Spanish Court. The ballad opera is named after one of the hunting lodges of King Philip IV of Spain. After 1850 it had a renaissance which was in style nearer to the more popular género chico. See also género chico More information Example 5: Zarzuela in the knowledge dictionary. Source: Bergenholtz & Bergenholtz 2011: 201.

In terms of the above functional framework, this dictionary provides information on its characteristics, focuses on the types of help its intended users need in the two situations targeted by the dictionary: cognitive and communicative. The data presented is reliable and matches the needs of the target users, and the dictionary uses current Internet technology. For instance, its team – this dictionary is designed, produced and updated by an expert in the field, a lexicographer and an IT expert can make use of log files and active feedback for updating and enlarging the lemma list. It is marketed by an IT company, does not require a great deal of funding, and shows a means of working in the field of specialised online lexicography in which three types of experts are needed, once again, a lexicographer, an expert in the field, and an IT expert.

8.5 The Glossary of Mortgage and Home Equity Terms The author’s view of The Glossary of Mortgage and Home Equity Terms is not found in outer texts or publications. On reading its homepage, we can infer that it is part of the promotional homepage of the bank, Wells Fargo. Our analysis of the dictionary reveals that it is a monolingual (English) communicative-oriented specialised dictionary which is designed and produced for assisting users in several situations, particularly text reception. Its main purpose seems to be to explain the meanings of the 230 mortgage and home equity terms that are in the publications sponsored by

144 | An Analysis of Specialised Online Dictionaries the bank in relation with mortgages. The data offered is reliable for the intended user but is handled inadequately as shown below. Dictionaries such as The Glossary of Mortgage and Home Equity Terms abound, and may be considered examples of specialised online monolingual promotional dictionaries, i.e. ones which are designed by organisations and/or companies – be they private or public, national or supranational, etc.- with the aim of explaining the meaning of the terms they use for referring to the basic characteristics of the products and/or services these organisations “sell” or deal in. Basically, the genuine purpose of this Glossary is to describe the mortgage contracts Wells Fargo sells. This dictionary lacks lexicographical know-how: it contains only short definitions of the terms. These are arranged alphabetically and target potential customers of the organisation that has designed and created the dictionary. The latter contains few lexicographical components and has primitive search systems: a horizontal menu with the letters of the alphabet. Clicking on one of them retrieves all the terms beginning with the same letter and there are no hyperlinks among them. In short, this dictionary is a copycat, i.e. it was created as a printed dictionary and uploaded as a pdf file (cf. section 2.3.). Illustration (7) shows that it contains only short definitions and synonyms, e.g. BPMI and “also known as “the mortgagor” in borrower.

Illustration 7: Screenshot for The Glossary of Mortgage and Home Equity Terms

In terms of the above functional framework, this dictionary does not describe its characteristics nor does it take into consideration lexicographical theories. It stands out in that it has three defining factors. The first is that this type of dictionary type is quite abundant and has low lexicographical information costs: most public or pri-

The Cambridge Business English Dictionary | 145

vate organisations design and produce these tools, most of which are regularly updated, easily accessible, and can be put together with little expenditure. The second factor is that these dictionaries do not benefit from Internet technologies: they are typically printed dictionaries that are uploaded as pdf files. Thirdly, they contain few dictionary components and are, therefore, only adequate for disambiguating the meaning of the terms used in the publications of the organisation sponsoring the dictionary; this is a modern-day trend in a world where technology is no longer selfexplanatory. Instead, technology [N]eeds instructions, leaflets for describing the product, the installation procedures, etc. In sum, modern terminology needs improving communication strategies. Communication between developer and user can only work if the text author and the text recipient share the same terminology, i.e., if a given term denotes exactly the same concept for author and reader. Language, and primarily written language, is the prerequisite for our modern technology. (Teubert 2005: 98)

8.6 The Cambridge Business English Dictionary An analysis of the data posted on the homepage of The Cambridge Business English Dictionary shows that this monolingual (English) communicative-oriented dictionary covers a whole domain, mainly targets students and professionals in the working place, focuses on written and spoken English, is informed by the Cambridge Nottingham Business English Corpus, uses illustrations and contains “Topic Boxes”, i.e. dictionary structures that give extra information about particular business topics: The Cambridge Business English Dictionary is a brand new dictionary of over 35,000 businessrelated words, phrases and meanings used in business and the world of work. Including the most up-to-date vocabulary from the rapidly evolving world of business and business English, the Cambridge Business English Dictionary is ideal for anyone studying business-related subjects and for anyone using English for their work. The dictionary gives thousands of examples from real business texts, helpfully presented information about grammar, and there is a strong emphasis on collocation. Informed by the unique Cambridge Business Corpus, the dictionary includes the very latest business-specific vocabulary. Most of the words in the dictionary have a business subject label, such as Marketing, Finance, or Computing. (...) (Homepage)

This dictionary is part of a teaching package: it is uploaded on the homepage of a publishing house that sells teaching materials and aims to meet the reference needs

146 | An Analysis of Specialised Online Dictionaries of those learning an LSP, e.g. Business English in illustrations (8) and (9). This means that The Cambridge Business is accessed in Cambridge Dictionaries Online, a homepage which is “the most popular online dictionary and thesaurus for learners of English”. Hence, it follows the lexicographical tradition of English learner dictionaries and includes the same dictionary data for describing terms (illustration 8) and words (illustration 9): the lemma, its (American and British English) pronunciation, part of speech, grammar sub-categorisation, phonetic transcription, short definitions, several collocations, some of which are exemplified, subject field labels, examples, hyperlinks to related terms, visual pictures, links to other dictionary articles and other dictionaries on the website.

Illustration 8: Screenshot for Cambridge Dictionaries Online

The Cambridge Business English Dictionary | 147

Illustration 9: Screenshot for Cambridge Dictionaries Online

Our analysis of The Cambridge Business English Dictionary shows that it describes its main characteristics and follows the lexicographical practices common to the design and construction of English learner dictionaries. For instance, it contains data and dictionary components that are suitable for assisting its primary users. The search engine is also well-conceived, although it does not allow for individualisation. We believe that the dictionary has a few drawbacks, perhaps because it does not pay due attention to a theory of lexicography. The first shortcoming is that it does not make use of Internet technologies for favouring individualisation, thus continuing the tradition of multi-functional dictionaries that are not equipped with technological solutions for offering mono-functional solutions. This means that the dictionary assumes all potential dictionary users have the same requirements, which is doubtful since learners vary in several key factors: mother tongue, command of mother tongue and foreign language, cultural background, knowledge of the domain, etc. Another problem is that the search engine has several, let us say, “fancy functions”. For instance, “My Site Dictionary”, which allows users to insert the name of their organisation in the search results. We believe that these functions cost time and money and should be abandoned as they can confuse users. Finally, the dictionary does not make much use of experts, e.g. its graphical interface somewhat unpolished.

148 | An Analysis of Specialised Online Dictionaries

8.7 TermFinder The merits and characteristics of TermFinder are to be found on its homepage and in several publications (e.g. Peters et al. 2006). Our analysis shows that TermFinder is a monolingual (English) communicative-oriented dictionary with a bilingual dimension (Chinese) (Bergenholtz & Tarp 1995: 49). TermFinder is “a collaborative project between Macquarie University teaching staff and dictionary experts, designed to help beginning students understand technical terminology.” (Homepage). All the dictionary structures are in English with the exception of a Chinese equivalent, which is included in the upper part of the dictionary article. This equivalent accompanies short definitions of the English lemma as well as data on its grammar, alternative forms, and usages, e.g. examples. It also contains cross-references, usage notes, an audio file, a subject field label and, in some entries, graphical representations, e.g. balance sheet (illustration 10).

Illustration 10: Screenshot for TermFinder

TermFinder is a Faster Horse dictionary that uses an access system with several options, although neither of them allows individualisation and mono-functional solutions (cf. section 2.3.): – A search limited to a discipline by selecting the appropriate thematic filter.

The Business English-Spanish Glossary by A.D. Miles | 149

– –

Wildcards such as *, and - can be used for retrieving headwords whose spelling is unknown or incorrect. TermFinder retrieves alternatives to the term the user has typed in, and explains the reasons for having retrieved alternatives.

Peters et al. (2006) explain that the underlying reason for designing TermFinder is that students, particularly novice students, are faced with the need to acquire large amounts of specialised academic and professional vocabulary, which is necessary for recognition as members of discourse communities (Swales 1990), or communities of practice (Wenger 1998). In order to achieve these aims, they rely basically on corpus frequencies, which have resulted in a small lemma list (around 600 terms in biology, statistics, accountancy, geology, neuropsychology, and a sampling of general academic words) with few multi-word terms. The project seems to have been terminated in 2008, and its dictionary entries have not been updated since then. This can be especially dangerous in culture-dependent subject fields. By way of example, in accrual accounting there is a usage note that refers to the diverse possibilities Australian companies have for preparing their accounts. The possibilities are based on annual figures, which are subject to constant change and must, therefore, be handled very carefully in this type of reference work. In terms of the functional framework given previously, this dictionary gives information regarding its main characteristics, uses experts, and contains adequate data for helping the targeted user in a particular user situation. In other words, it presents data in a way that supports the communicative function identified, which is to assist users in understanding the meaning of certain terms as they are employed in academic institutions. However, the dictionary also has its disadvantages: it is not updated regularly, contains very few lemmas for its stated purpose, and does not rely on a theory of lexicography. It claims, for example, that its data is mostly corpus-based, an aspect which is problematic in specialised lexicography, especially in relation to crafting definitions and describing grammar. This might explain why the dictionary contains very few lemmata and why the project was abandoned; extracting terms from corpus data must be complementary in specialised lexicography (cf. section 9.3.1).

8.8 The Business English-Spanish Glossary by A.D. Miles The author’s view of The Business English-Spanish Glossary is absent. Furthermore, the information posted in its homepage is scanty and irrelevant. It only indicates that it is a bilingual biscopal (English-Spanish/Spanish-English) communicativeoriented dictionary being compiled by A. D. Miles (a Business English teacher). Miles started this project in the 1980s and is still engaged in it:

150 | An Analysis of Specialised Online Dictionaries Expressions included in the A. D. Miles Business English to Spanish Glossary were compiled between 1980 and today by Andrew D. Miles. The glossary mostly comprises business terminology that cannot be found in general dictionaries. This online version has over 10,000 entries and is offered for free. (

Our analysis of its components shows that it contains a word list of lemmas and their equivalents, and practically nothing else, that is, it lacks definitions and the grammatical – especially syntactic – data needed to provide qualified assistance in translation as well as foreign text production, which are the most relevant functions in relation to bilingual dictionaries. It does not identify targeted users and uses two very basic search systems. One of them is a search engine that allows fuzzy searches and has two options: words beginning with a letter from English to Spanish or from Spanish to English. This search retrieves the term and a list of equivalents, whose separate senses are differentiated by Arabic numbers (illustration 11). The second search system is a horizontal menu containing the letters of the alphabet in both languages (it is accessed by clicking on “About the glossary"). For instance, clicking on “b” retrieves four options: (i) B2B – banking; (ii) bankrupt-berth; (iii) bid-bribery; (iv) bridge-byte. In “bid-bribery”, for instance, we obtain 146 alphabetically arranged English lemmas and their Spanish equivalents, together with hyperlinks on the right to advance in the search, e.g. “go to: bill, billion, etc.”. The one-word terms are accompanied by a grammar code indicating their part of speech, e.g. “n” for noun (illustration 12).

Illustration 11: Screenshot for The Business English-Spanish Glossary

The Glossary of FAO Database and Information Systems | 151

Illustration 12: Screenshot for The Business English-Spanish Glossary

In terms of the above functional framework, the dictionary does not inform its users of its main characteristics. Our analysis has shown that it does not identify users, and that the data presented and the dictionary components used are rather unsatisfactory. Furthermore, the access system is somewhat primitive, and experts do not work on the dictionary. It appears that it is updated quite regularly.

8.9 The Glossary of FAO Database and Information Systems The author’s view is not found on the homepage of the Glossary of FAO Database and Information Systems (FAO Glossary) or in publications. Our analysis reveals it to be a multilingual communicative-oriented dictionary that is mainly designed for assisting translators. These are usually members of supranational organisations and are required to translate documents from one official language into another. For instance, the United Nations and its different agencies design and produce specialised online dictionaries in Arabic, Chinese, English, French, Russian and Spanish. Our analysis of its components shows that it includes a lemma list and a short definition in each language, without any equivalents and cross-references. This means that users have to match the lemma in language "a" and its equivalent in language "b" depending on their own knowledge (in other words, they can do this only if they already know the meaning of the concept). By way of example, the Spanish perfiles de pastos de los países is found under the letter "p", whereas its English equivalent country pasture profiles is located under the letter "c" (illustration 13); they are, then, not cross-referred. Furthermore, this dictionary type does not include grammar data or contextual clues, e.g. examples, which makes it un-

152 | An Analysis of Specialised Online Dictionaries suitable for translation purposes, although its compilers maintain that it was and is designed for assisting translators.

Illustration 13: English Lemma country pasture profiles and Spanish equivalent perfiles de pastos de los países in the FAO Glossary.

This dictionary, which is a typical copycat (cf. section 2.3.), contains around 75 terms in each language, which can be accessed by scrolling up and down its homepage. It is characteristically integrated in dictionary aggregators and international databases such as UNTERM (cf. section 8.11. 3, below), and uses some hyperlinks to texts of the same organisation. For instance, by clicking on ASFA users are crossreferred to Aquatic Sciences and Fisheries Abstracts, which is [A]n International Cooperative Information System which comprises an abstracting and indexing service covering the world's literature on the science, technology, management, and conservation of marine, brackish water, and freshwater resources and environments, including their socio-economic and legal aspects. (homepage)

In terms of the above functional framework, this dictionary does not inform on its main characteristics. Furthermore, our analysis has indicated that the FAO Glossary is lacking in quality, is not based on lexicographical theories or practices, and has high information costs (Nielsen, 2008). For example, it does not contain dictionary structures or data that support the translation function. In addition, its access route is very complicated and cannot be used by inexperienced translators, who must “figure out” the equivalent, as each dictionary follows an alphabetic ordering in its own language without any cross-referencing among these. It does not provide information on the role of experts, is updated sporadically, makes no use whatsoever of Internet technologies, and does not explain data selection.

Kicktionary | 153

8.10 Kicktionary Schmidt (2009) and the homepage of Kicktionary indicate that this is an online dictionary of the language of football (soccer) containing around 800 German, 600 English, and 530 French terms. It was developed between September 2005 and July 2006 during Schmidt’s stay as a visiting researcher for the FrameNet project. Therefore, Kicktionary demonstrates that we are also witnessing a high level of interest in designing and producing multilingual specialised online dictionaries as individual projects. A review of recent publications in this field of research shows the existence of many scholars who describe their projects and on-going work, e.g. Schmidt’s Kicktionary (2009). Schmidt claims that the main purpose of the project “is to explore how linguistic theories about lexical semantics, methods from corpus linguistics, technologies for hypertext and hypermedia and techniques from computer language processing can help to make lexical resources that are better than (or: good in a manner different from) traditional paper dictionaries”. (Homepage) An analysis of illustration (14) shows that Kicktionary is based on frame-based terminology (cf. section 7.3.3.). The dictionary contains a term and its part of speech in one of the languages, followed by its “scene” and “frame”. Users then access a list of examples, each of which contains frame elements and codes as well as semantic relations, e.g. synonyms and hypernyms, and a list of “frame elements”. Some entries also contain a short definition.

Illustration 14: Dictionary article in Kicktionary

154 | An Analysis of Specialised Online Dictionaries On the left, users have a search system that allows them to investigate lexical units (LU), scenes, conceptual hierarchies, spoken examples, and parallel texts. Searching through “Lexical Units” retrieves an alphabetically organised list of all the terms included in the dictionary. These are, in their entirety, in the same list and, therefore, users can access German abblocken, and English appeal under “a” (Illustration 15):

Illustration 15: Searching for Lexical Units in Kicktionary (excerpts)

Searching through “scenes” and “conceptual hierarchies” retrieves a list of 16 scenes, i.e. prototypical events, and around 40 conceptual hierarchies, each of which organises a set of synsets, i.e. group of words with identical or similar meanings, “into a tree via lexical relations such as hypernymy/hyponymy (X is-a-kind-of Y), holonymy/meronymy (X is-a-part of-Y), and troponymy (to X is to Y in some way)”. (Homepage) Finally, searching through “spoken examples” and “parallel texts” retrieves chunks of aligned corpus texts that contain lexical units, some of which are links to the dictionary article. For instance, clicking on goal (illustration 16) retrieves the dictionary article of goal as well as excerpts of texts.

Illustration 16: Retrieval of parallel texts in Kicktionary

In terms of the above functional framework, we believe that Schmidt’s stated aim falls a long way short of being fulfilled. Users will find its dictionary articles very difficult to understand. They will also find that the annotation used (e.g. SCORER after Mario Frick; Illustration 14) does not help in any user situation. Furthermore, they will be surprised that the compiler has spent his time offering “frames” when much needed data, e.g. definitions and equivalents, are sometimes present but other times not. Briefly, Kicktionary is an example of lexicotainment, i.e. an informa-

Kicktionary | 155

tion tool designed and constructed for fun (Almind, Bergenholtz & Vrang 2006). Litkowski (2010: 105), for example, concurs with our view, when he maintains that the theoretical descriptions of frame semantics given in Boas (2009) can be viewed only as “a set of further theoretical developments in frame semantics, rather than a guide to be used in current lexicography”. Although Atkins & Rundell (2008: 144149) maintain that frame semantics is an appropriate theoretical construction for describing the valencies of lexical items in dictionaries, we have not been able to support this view in existing dictionaries. How can a typical user, e.g. one with no knowledge of theoretical linguistics, consulting illustration (17) obtain the information they might need in a user situation? This is a good example of comprehensionrelated information costs, which have not been taken into consideration when designing the dictionary.

Illustration 17: A Dictionary article in Kicktionary

156 | An Analysis of Specialised Online Dictionaries

8.11 IATE, CercaTerm and UNTERM The homepages of IATE, CercaTerm, and UNTERM present these as term banks or terminological databases. These are the names used in terminological circles for referring to specialised dictionaries whose design and elaboration started in the 1960s in large language services belonging to governmental organisations and large enterprises, in standards bodies and in language planning institutions (Schmitz, 2006). The above reference work typically shares a common structure that consists of a number of elements (Table 8. 1.) and values, as well as an interest in describing concepts and their relations within a particular domain or sub-domain (ISO 12620). Table 8. 1.: List of elements and their meaning. Source: TermSciences (adaptation) Element Terminological entry Concept identifier Language section Language identifier Term section Administrative status

Remarks Terminological entry (concept). Identifier for the entry. Language section made up of all the terms in a given language. Identification for a language, e.g. “en” for English and “es” for Spanish. One section per term. Status of a term in its original reference work (e.g. preferred term or deprecated term in MeSH). Used to specify the update of term (e.g. modification, deletion, input). Used to specify the date a term was updated. Origin of an element, generally the name of an institution. The name of the terminological resource the term came from.

Transaction Date Original institution Originating database name Broader concept generic Broader concept; used here to describe the “is a” type of relation but sometimes used for the “part of” type relations. Certain broader concepts are contextual meaning specific to a user in a given field (see domain Expert). Related concept Related concept. Domain Expert Defines a specific field of knowledge in which the concept is used. Subject Field The field a terminology entry comes from originally; one or more scientific fields to which the concept is related. Application Subset Field of application of a concept. Associated concept Non-hierarchical theme-based relation between two concepts. Here it corresponds to the semantic category a concept belongs to (e.g. illness). Definition Describes a concept to differentiate it from other concepts in the same scientific field. Valid for all terms in a given terminological entry. Source Bibliographic data on the origin of a data item such as a definition, etc. Part of speech Grammatical category of a given term. Grammatical gender Gender of the term, e.g. masculine. Grammatical number Number of the term, e.g. singular. Lexeme term type Type attributed to a lexeme or term. Can include several different values: abbreviation, full form, formula, etc. Context Extract from a text which shows the term in context.

IATE, CercaTerm and UNTERM | 157 Note Element working status

Application note. Gives other information on any part of a lexical resource. Can complete a definition. Status of a terminological entry currently being ratified for scientific approval.

The number of terminological databases is enormous, and all of them contain specialised terms and their definitions as well as an array of data that aim to situate the term in a conceptual map and assist in its standardisation. These systems are mainly designed for helping in three situations: translation, language planning, and standardisation.

8.11.1 Inter-Active Terminology for Europe (IATE) In the homepage of the Inter-Active Terminology for Europe (IATE), we are informed that IATE is the “EU inter-institutional terminology database”. An analysis of publications and Internet documents on IATE shows that it has been used in EU institutions and agencies since 2004 for the collection, dissemination and management of EU-specific terminology. It is the successor of several EU databases: EURODICAUTOM, TIS, EUTERPE, EUROTERMS and CDCTERM. IATE was launched “with the objective of providing a web-based infrastructure for all EU terminology resources, enhancing the availability and standardisation of the information”. (Homepage) In March 2007, it provided the general public free access to multilingual terminology in the fields of activity of the European Union. For the purpose of this book, IATE is an example of a plurilingual translationoriented specialised online dictionary, whose main characteristics are summarised below (our analysis): – IATE covers a broad spectrum of domains: politics; international relations; European Communities; law; economics; trade; finance; social questions; education and communications; science; business and competition; employment and working conditions; transport; environment; agriculture; forestry and fisheries; agri-foodstuffs; production and technology research; energy; industry; geography; and international organisations. Each record typically stores the information supplied by EU sources, which has resulted in an information system that contains more than 1.6 million records, around 7 million terminological entries, and 8.7 million terms, “including approximately 540,000 abbreviations and 130 phrases”. (Homepage) – Although it covers a broad spectrum of disciplines in all the official EU languages, its consideration is restricted to EU policy (our emphasis), i.e. it reflects only the meaning accorded within the European Union. In other words, both the description of the concepts and terms covered, as well as the names given to the domains included in IATE are idiosyncratic, and can or cannot also be used in other parts of the world. For instance, the domain social questions is only a do-

158 | An Analysis of Specialised Online Dictionaries

main in IATE. Similarly, comercio, which is the Spanish equivalent of trade, is omitted in the domain of Politics. In addition, the Spanish equivalents of Southern Cone Free Trade Zone are Mercosur, Mercado Común del Sur, Mercado Común del Cono Sur and Mercado Común de América del Sur. IATE covers all official languages, English, French, and German being those with more terms. There are about 1.6 million English terms and around only 35,000 in Hungarian. In addition, it includes Latin for botanical and zoological names. IATE is a prototypical example of a translation-oriented specialised dictionary. This means that its genuine purpose is to facilitate translation and terminology processing, which is considered essential for supporting translation, e.g. in multilingual organizations (UNTERM) and multilingual countries (Canada). Hence, IATE mainly targets translators, interpreters and writers of EU texts who access data aimed at increasing the reliability of the solution proposed in all (or many) of the EU languages: Its added value is the result of the terminological processing of information (for instance, a document search, the addition of a definition or reference, the designation of the preferred term for consistency reasons, the endorsement of a solution suggested by a translator, etc.) in order to increase the reliability of the proposed solution. Multilingualism, and consistency between languages in particular, is also an added value, since it enhances the usefulness of the database. (IATE Best Practice. Quoted from de Vicente 2007)

IATE records have three levels. At the top, users access language-independent data. The second level contains language data and the third level includes term data. This means that IATE is based on concepts, not only terms, and offers more definitions, relevant references, fewer duplicates and multi-languages entries. The selection of terms has three approaches: – A thematic approach, which selects EU core terminology and terms documented in existing EU collections/glossaries, in addition to domains assigned to IATE partners (these are: the European Parliament, Council, Commission, Court of Justice, Court of Auditors, European Economic and Social Committee, Committee of the Regions, European Investment Bank, European Central Bank, and Translation Centre). – A frequency/machine-based approach: this selects monolingual entries and duplicates according to consultation frequency. – An ad-hoc approach: this consists of proposals by translators, terminologists or other institutions, etc. In terms of the functional approach presented above (cf. section 8.1.), we believe that IATE has several important drawbacks due to a lack of adequate lexicographical planning. These deficiencies are to be found in the fact that there is no matching of data, functions and users' needs (Illustrations 18 and 19), as well as in the defi-

IATE, CercaTerm and UNTERM | 159

cient use of Internet technologies; nevertheless, the cost of IATE’s undertakings runs into millions of euros. Finally, this database is not updated regularly, and seems to use experts on a sporadic basis. These drawbacks are explained below in more detail, in view of IATE’s being very influential and, as we have said, costing European taxpayers a great deal of expense. Illustration (18) shows “global results”, i.e. the retrieval of all the data included in IATE when the user investigates via the search engine. Illustration (19) shows the results per dictionary entry, which are obtained when clicking on “Full entry”, a search system located on the right side of the records.

Illustration 18: Screenshot for trade in IATE

160 | An Analysis of Specialised Online Dictionaries

Illustration 19: Screenshot for the entry Anti-dumping code in IATE

We consider the main inconsistencies or drawbacks of IATE to be: – The system uses a reliability rating of the terms included that is founded on unstable and untested bases. In other words, it includes data with no stated purpose or interest for the user. As an illustration of this, what can a user do if the system indicates that a term has “minimum reliability”, or “reliability not

IATE, CercaTerm and UNTERM | 161

– –

verified”? Furthermore, these two ratings are very common. For instance, an analysis of 30 English terms with accounting (English – Spanish search) resulted in: 8 “reliable terms”, 10 “minimum reliable” terms, and 12 “reliability not verified” terms. The analysis was carried out with terms from accounting to special allowance for accounting officers. The system does not use labels for helping users. For instance, ADC and Antidumping code and Acuerdo Antidumping and Código antidumping are synonyms, but in fact no indication of this is given. In other words, users have to decide by themselves whether or not these are synonyms. Antidumping (with an initial capital letter) and antidumping (with a small letter) are part of the equivalent of Anti-dumping Code. The reason for using either a capital or a small letter is not mentioned, which will make users hesitate about the correct way to write the term. The search system retrieves trade in several dictionary components without specifying the information retrieved. Although such information is useful for experienced translators, it needs to be systematic in order to be helpful. For example, trade is in trade description (the term or dictionary entry) as well as in several collocations, e.g. “effect on trade between Member States”, and examples, e.g. “Customs Cooperation Council’s Recommendation of 1 January 1975 designed to enable Statistical Data on International Trade collected on the basis of the Brussels Nomenclature to be expressed in terms of the second revision of the Standard International Trade Classification”. The term trade description is in the domain “Community law, agriculture, forestry and fisheries, Tariff policy, Trade policy, agri-foodstuffs (Council)”, whereas the collocation is in that of “Competition law, European Union law, Trade”, and the example is in the domain “Community act, international agreement, Tariff policy, Community law (Council)”. This way of presenting the data will demand a great deal of processing time, as users have to find out, say, the relationship between a term in domain “a” and contextual clues in domains “b” and “c”, if, that is, one actually exists. The search system retrieves data that are not part of the item investigated. Why, for example, does it retrieve Anti-dumping Code and ADC when searching for trade? The domain is only in English, not in Spanish. The definition is only in English, not in Spanish. Furthermore, there is an English definition for the term Anti-dumping code but not for the term Agreement on implementation of Article VI of the General Agreement on Tariffs and Trade. Both terms are in the same record, and the definition is at the top without any link to the term defined. How is a user to know to which term the definition refers? The note in the English entry is easy to understand, whereas the Spanish note is less clear. We think that Spanish translators are more familiar with “GATT” than with “MISC”. In fact, we were unable to find the meaning of “MISC”.

162 | An Analysis of Specialised Online Dictionaries –

The Spanish entry contains hyperlinks to external sources, whereas these are absent in the English entry.

Furthermore, IATE is a very expensive project - Tarp (2012c), for instance, claims that it is probably the world’s most expensive lexicographical work ever –; it does not deliver much if we consider the expense and human resources employed in its design, preparation, and adaptation, containing data, e.g. reliability, Term Ref., and Term Note, which are useless for its intended users, and its completion requires a large investment in time and funds. Furthermore, IATE is not updated on a regular basis, which is a very important drawback in such a system, especially when errors and mistakes are easily recognised. For instance, searching accounting in any domain retrieves 859 hits. The first of these is accounts and accounting in Medical science and Community finance. Clicking on “full entry” retrieves illustration (20), in which we find data difficult to understand. For instance, the definition is “charge”, and accounts and accounting are abbreviations:

Illustration 20: Screenshot for IATE

8.11.2 CercaTerm The homepage of CercaTerm indicates that it is a multilingual specialised online dictionary developed by TERMCAT, the Catalan Centre for Terminology. Our analysis of CercaTerm shows that it covers a broad spectrum of domains, and is a prototypical example of a language-planning oriented specialised online dictionary. This

IATE, CercaTerm and UNTERM | 163

means that its main purpose or function is to help users increase the status of Catalan, a minority language which has, for one reason or another, been historically disadvantaged. Hence, the rationale for CercaTerm is that Catalan must have up-todate coherent terminology to ensure professional communication in all fields. Users can access CercaTerm and other language planning services on the homepage. CercaTerm includes human languages, typically Catalan, Spanish, English, French and German terms (also languages such as Guaraní), in addition to other communication systems, e.g. numbers. An analysis of this dictionary in terms of the above functional framework (cf. section 8.1) reveals the following characteristics: – CercaTerm targets every possible user requiring Catalan terminology. In short, its main objective is to replace terminology imported from the language of technologically dominant countries, thereby fostering term formation in Catalan. – CercaTerm has two search systems. The basic search organises the findings by thematic areas and language, which function as thematic area and language filters (Pastor & Alcina 2010: 328), in that they “add a search restriction to the query”; these may assume various forms, such as a restriction in terms of part of speech, thematic field or language. A search for flujo de caja (cash flow) with the thematic area filter Economía. Empresa (‘Economics. Business’), finds 19 results that are further subdivided into sub-domains such as ‘accounting’, ‘management’, ‘audit’, and ‘financial markets’. By clicking on each result, users retrieve a list of equivalents with grammar labels and a brief definition in Catalan. On the right side of the screen, users can click on two menus, abreviacions utilizades (‘abbreviations’), and indicadors de llengua (‘languages of the translation equivalents’). These display a vertical menu explaining the abbreviations used (illustration 21).

164 | An Analysis of Specialised Online Dictionaries

Illustration 21: CercaTerm results for flujo de caja (basic search)

The advanced search system allows more focused searches (e.g. “acabar en”; En: “ends with”, “empezar por”, En: “begins with”), maintains the thematic and language filters and includes others: definition, denomination, note, lexical category, and hierarchy. This allows different types of searches, with the important aspect that both searches offer the same data when a user clicks on one of the retrieved terms (illustration 22). Again, on the right side, abbreviations are displayed, e.g. codes for the languages included in the system (illustration 22). The system does not retrieve dynamic data in dynamic articles.

IATE, CercaTerm and UNTERM | 165

Illustration 22: CercaTerm results for flujo de caja (advanced search)

Finally, the data included for each record are represented by a Catalan definition, equivalents and several codes, basically language and part of speech codes. It does not mention the role experts play in its design and compilation.

To sum up, CercaTerm is basically a Faster Horse dictionary whose main advantages and disadvantages may be listed as follows (cf. section 2.3.): Its main advantages are: (i)

it adopts the one-to-one relationship between concept and term that characterizes specialised discourse;

(ii) the definitions are suitable for the primary user groups, who can disambiguate meanings easily; (iii) the grammar data concerning the Catalan term (either as lemma or as equivalent) support translation. The main drawbacks are conceptual, and show that this dictionary was based on linguistic, not lexicographical, traditions and practices. Firstly, most of the data included in the system, for example, in Terminologia Oberta and Biblioteca en Línia, are only useful for experts who happen to know which document they have to download and search for. In other words, they are

166 | An Analysis of Specialised Online Dictionaries ‘raw data’ which were not primarily designed to provide assistance to translators with specific and situation-dependent needs. Secondly, the dictionary does not include collocations, phrases and examples, nor genre conventions and necessary words. Thirdly, the terms have not been lemmatized according to lexicographical practices appropriate for the compilation of specialized dictionaries. For instance, translators of specialized texts need complex terms to be lemmatized, especially when they are highly relevant, and hence highly important, in the subject field. In accounting, concepts such as ‘income from shares in group undertakings’, ‘income from interests in associated undertakings’, and many more, represent items in profit and loss accounts and refer to conceptual realities, most of which are culture dependent (in Spanish, the former is the equivalent of ingresos por participaciones en empresas del grupo, whereas the latter is ingresos por intereses en asociadas). The Spanish equivalents are terms extracted from corresponding Spanish profit and loss accounts, cuenta de pérdidas y ganancias, and illustrate two related but different concepts for English undertakings: ‘empresas del grupo’, and ‘asociadas’). Cercaterm does not lemmatize these complex terms, which can result in important difficulties when translating, say, the Spanish profit and loss account of Banco de Santander into English, as this must be done according to current legislation on the disclosing of financial information by multinational companies. Finally, the online dictionary does not use the technological possibilities of the Web and follows the tradition of constructing polyfunctional dictionaries that do not pay attention to search customization and user profiling. (Fuertes-Olivera & Nielsen 2012: 199- 200)

8.11.3 The United Nations Multilingual Terminology Database (UNTERM) The era of the Internet has also favoured the design and production of specialised online dictionaries which put several objectives on an equal footing. For instance, the United Nations Multilingual Terminology Database (UNTERM) has been made accessible on the Internet with the stated aim of helping any possible user, not only those working within the institution: This database was compiled over the years in response to diverse and wide-ranging demands of United Nations language staff for terminology and nomenclature. It is being put on the Internet to facilitate the efforts of people around the world who participate in the work of the United Nations but do not have access to the Secretariat’s intranet. (Homepage)

UNTERM is updated daily and contained 85,000 entries in six official languages at the time of writing this chapter (May, 2013). It is internally generated and aims to offer accurate and consistent usage of terminology in UN documents: UNTERM is a multilingual terminology database which provides United Nations nomenclature, technical or specialized terms and common phrases in all six official languages – English, French, Spanish, Russian, Chinese and Arabic. The database is mainly intended for use by the language staff of the United Nations to ensure accurate and consistent usage in documents published by the Organization.

IATE, CercaTerm and UNTERM | 167 UNTERM content is generated by the Terminology Team of the Terminology and Reference Section, Documentation Division, DGACM. Technical support is provided by ICTS and ITSD. (About UNTERM, Homepage)

UNTERM uses a search engine that allows users several options. They can select from one to six official languages, use wild cards, use keywords with Boolean operators, limit the search and also employ fuzzy searches (illustration 23). In our view, the search system is very complicated and difficult to put into practice, i.e. performing a search takes longer than it should. This is revealed in two examples that are commented on below.

Illustration 23: Search tips in UNTERM

The first example deals with the use of quotation marks. It is claimed that there is no need to use quotation marks around the search string, “unless it includes the words field or sentence.” The reason is that these two words are special, the so-called “reserved” words (Illustration 23 above). The second example consists of a system of codes that limit the search for the keyword being investigated. These codes are displayed on the homepage and restrict the search to main fields, acronyms or notes (Illustration 24):

Illustration 24: UNTERM field names for restricting searches

168 | An Analysis of Specialised Online Dictionaries When searching in the search engine, users retrieve up to 250 documents that comprise the following: a card, subject(s) in English, and terms in the 6 official languages, including variants (Illustration 25).

Illustration 25: Screenshot for UNTERM (excerpts).

Clicking on “card” retrieves more data: a hyperlink to a related term and an English definition, English synonyms and spelling variants, and the subject field; there are also several notes that are, in our opinion, difficult to understand (for example, we cannot be sure of the exact meaning of the data included in “Note Chn”). All the terms are “validated” (Illustration 26).

Illustration 26: Card for checking account in UNTERM

The Multilingual Glossary of Technical and Popular Medical Terms | 169

In general, we believe that terminological databases such as IATE, CercaTerm and UNTERM do not actually fulfil the purposes they claim to. Although they receive a great deal of funding and resources, our view is that they do not really deliver the goods, especially in terms of translation purposes. These terminological databases provide a better service for documentation purposes than for punctual consultation. Furthermore, they do not make adequate use of the Internet, as a result of which they are for us partly examples of Copycats and partly examples of Faster Horses (cf. section 2.3.). Let us take, for instance, Illustration (26), which seems to contain the same data as a printed version of the database. Finally, we do not know the role, if any, experts play in compiling them.

8.12 The Multilingual Glossary of Technical and Popular Medical Terms in Nine European Languages IATE and similar reference works, e.g. TERMIUM PLUS®, are influencing the allocation of funds distributed by several research financing agencies. Here, there are several EU-funded projects that are conceptually similar to IATE. Two of these projects have been – and still are – working in the design and construction of specialised online dictionaries, such as The Multilingual Glossary of Technical and Popular Medical Terms in Nine European Languages (Medical Terms Glossary) and The European Dictionary of Skills and Competences (Disco) (cf. section 8.13.). In the Introduction of the Medical Terms Glossary, users are informed that this is a collection of 1,830 technical and popular medical terms in eight official European languages (not nine, as its title says): English, Dutch, French, German, Italian, Spanish, Portuguese and Danish. (Homepage) The Introduction also informs us that the dictionary aims to assist technical writers and translators of medical texts to write medical leaflets in plain language (EEC Directive 92/27/EEC). Its authors claim that the rationale for constructing this dictionary is that the writers and translators of medical texts must have at their disposal an information tool that describes the meaning of medical concepts, in addition to the two terms used for referring to them: the technical and the popular. This is achieved through an English definition of the technical term – the same definition is included in the eight dictionaries - as well as the technical and popular term in each of the European languages covered. The English definition, together with the number code used for identifying each term, help users to disambiguate the meaning of polysemous entries (Illustration 27).

170 | An Analysis of Specialised Online Dictionaries

Illustration 27: abdomen and its popular synonym gut in the Medical Terms Glossary

In the Introduction, we are also informed that the lemma selection process is based on an English word list stemming from a Dutch frequency search. The process started with a selection of Dutch words that were included in a CD-ROM Dutch compendium of medical products. A computer program searched the compendium and generated an alphabetical list of approximately 45,000 Dutch word forms. This initial list was then reduced to around 11,000 words through the elimination of function words (e.g. articles, prepositions, etc.), words with a frequency count lower than five, and those with different spellings for the same concept (e.g. ph/f). This list of 11,000 words was further reduced through eliminating terms that did not pertain to the domain, e.g. chemical structure terminology, compounds, and "terms with a frequency of less than 10". This resulted in a lemma list of around 2,200 Dutch terms that was then submitted to a third reduction by the elimination of synonyms and a regrouping of related medical terms. This Dutch word list was translated into English, the lingua franca of the medical field. This resulted in 1,830 English technical terms that were subsequently translated into the rest of the languages included in the dictionary. Finally, the technical lemma list in each language was augmented via the inclusion of a popular term (e.g. gut is the popular term for technical abdomen (Illustration 27, above). In short, the selection process and methodology are not based on lexicographical theories.

The Multilingual Glossary of Technical and Popular Medical Terms | 171

An analysis of its data and elements shows that the former does not match the intended functions, and that the user interface of the online version has three different approaches: (i) language-specific, whereby users choose the working language and browse through the list of technical and popular terms as well as the so-called "glossary", which includes "the comments and critics of translators and users"; (ii) concept-oriented, by means of which the user selects from the 1,830 English terms (which are alphabetically indexed) and retrieves all the translations in the different languages; (iii) the so-called "speed search" for specific words, which allows users to enter a word or string of words in the search engine and retrieves "all occurrences of that word". At the moment of writing this book (April, 2013), the “speed search" did not work. The search process was slow, with users having to carry out several searches in order to be sure they had found what they were looking for. For instance, a search for the Spanish word abdomen retrieved an alphabetical list of the technical term in bold followed by grammar codes “nm” (noun masculine) and the popular term in italics (vientre), also followed by the same grammar codes (Illustration 28).

Illustration 28: Spanish abdomen in the Spanish Medical Glossary.

The most positive aspect of this project is that experts in several field (46 scholars are mentioned on the homepage) worked on the project. Although they had substantial financing at their disposal, the end result is, we consider, rather poor, perhaps because they did not apply a lexicographical approach. Some of its main deficiencies are as follows: – the number of terms covered is rather small; – the data are not updated regularly;

172 | An Analysis of Specialised Online Dictionaries –

the data included for each entry is of little assistance (for instance, users are not provided with collocations, examples or spelling variants, aspects that should always be incorporated in a reference work that aims to help users in a production situation, which is the one the dictionary aims to cover; the lemma selection process is somewhat misguided and, as far as we are concerned, painfully slow (this may explain that it contains a lemma list of only 1,800 terms). By way of example, although multi-word terms are very common in Medicine they have been left out. Similarly, the dictionary only contains frequent words, namely, ones which do not pose a real problem for experienced technical writers and translators. Such users actually need terms that are not frequent, as these are the problematic ones; the search process is slow and requires a good knowledge of English for disambiguating the meaning of the terms in the other languages.

8.13 The European Dictionary of Skills and Competences (Disco) The European Dictionary of Skills and Competences (Disco) illustrates a far-reaching situation within Europe. This is an Austrian and EU-funded project begun in 2004, which provides access to a thesaurus containing more than 90,000 skill and competence terms and about 8,000 example phrases in ten European languages: Czech, English, French, German, Hungarian, Lithuanian, Swedish, Spanish, Italian and Slovakian. It is currently in its second stage (November 2010 – October 2012), and aims to design a “thesaurus explorer” and a “start document profiler” (Illustration 29).

The European Dictionary of Skills and Competences (Disco) | 173

Illustration 29: Homepage of Disco

Our analysis of Disco shows that this is an example of a mixed package, i.e. a EUfunded project that aims to design and produce two or more related systems. The first is a reference work that is typically designed under the tenets of traditional or linguistics-based theories of terminology. The second system usually deals with the Internet as a technology for automating language processes, e.g. the automatic creation of academic and/or professional profiles (e.g. the Disco project), the creation of ontologies (e.g. the PROLIX project, De Baer et al. 2009), automating term extraction (e.g. the TTC project, e.g. Heid & Gojun 2012), or developing commercial contacts between small and medium-size enterprises (e.g. the ESPA Dictionary project). Below, we will restrict our analysis to the Disco project: the PROLIX and the TTC projects have yet to come to fruition, whilst the ESPA dictionary is similar to Disco.

174 | An Analysis of Specialised Online Dictionaries Our analysis of the thesaurus and its technical report shows that the data does not match the functions identified. For instance, it contains a structured vocabulary for describing skills and competence in many working environment contexts. However, this vocabulary only contains frequent terms and patterns from corpora, in other words, not what typical users of specialised dictionaries really need. The search system consists of an engine that allows general or restricted searches. For instance, if business is specified without restriction, then 25 results are retrieved in the Show Term Browser, each with its corresponding equivalents in the other languages (Illustration 30). However, the same search string, e.g. business restricted to "education" retrieves only four results, each of which has its corresponding equivalents when clicked on in the "Show Term Browser" (Illustration 31).

Illustration 30: Screenshot for Show Term Browser in Disco. (General Results)

The European Dictionary of Skills and Competences (Disco) | 175

Illustration 31: Screenshot for Show Term Browser in Disco. (Restricted Results).

We believe that the funding provided for the Disco project is proving rather unproductive. As with many similar projects, the concept of user need is employed as a catch-all phrase. For instance, users searching for the English business retrieve 131 terms and 28 phrases. When clicking on one of these, they can retrieve the equivalent in one or more of the languages of the project. The Spanish equivalent and the English lemma are lemma lists with no further data; in other words, they are not well-conceived information tools (Illustration 32). This has two dangerous consequences. The first is that the EU research agency is wasting money, and the second is that users are not being given access to quality reference works despite the considerable cost.

176 | An Analysis of Specialised Online Dictionaries

Illustration 32: Results for business in Disco.

Furthermore, we have followed the instructions given under “Europass CV” (the second system) and have obtained a list that, for us, is neither informative nor helpful, especially if we take into consideration the time spent on discovering how to convert the data given into information (Illustration 33).

Genoma and EcoLexicon | 177

Illustration 33: Europass CV in Disco

Finally, we have been unable to find the role of experts in the design and compilation of Disco.

8.14 Genoma and EcoLexicon Since the 1990s, a lot of research has been devoted to designing and constructing ontologies: In the context of computer and information sciences, an ontology defines a set of representational primitives with which to model a domain of knowledge or discourse. The representational primitives are typically classes (or sets), attributes (or properties), and relationships (or relations among class members). The definitions of the representational primitives include information about their meaning and constraints on their logically consistent application. In the context of database systems, ontology can be viewed as a level of abstraction of data models, analogous to hierarchical and relational models, but intended for modeling knowledge about individuals, their attributes, and their relationships to other individuals. Ontologies are typically specified in languages that allow abstraction away from data structures and implementation strategies; in practice, the languages of ontologies are closer in expressive power to firstorder logic than languages used to model databases. For this reason, ontologies are said to be at the "semantic" level, whereas database schema are models of data at the "logical" or "physical" level. Due to their independence from lower level data models, ontologies are used for integrating heterogeneous databases, enabling interoperability among disparate systems, and specifying interfaces to independent, knowledge-based services. In the technology stack of the Semantic Web standards [Berners-Lee, Hendler & Lassila, 2001], ontologies are called out as an explicit layer. There are now standard languages and a variety of commercial and open source tools for creating and working with ontologies. (Gruber 2009)

178 | An Analysis of Specialised Online Dictionaries Software tools, e.g. Protégé Ontology Editor, are used for creating and handling ontologies. Boyatt & Joy (2010: 6) illustrate the process of constructing the ontology used in the MALog project, which is an EU-funded project primarily designed to produce “pedagogically high quality learning material units to help develop mathematical logic knowledge and skills”. This process comprises six steps: – Identifying the domain of the scope of the ontology to be constructed. Therefore, in the early stages of the project a “needs-analysis survey was conducted to investigate the areas of mathematical logic already taught in schools and universities, and identify topics where more teaching materials were needed”. (Boyatt & Joy 2010: 9) – Conducting research on existing resources in order to “ensure that nothing is available that could be adapted or extended”. (Boyatt & Joy 2010: 9) – Collecting relevant information to identify many different terms. The resources used “include mathematical text books, mathematical logic teaching resources, websites and published mathematical literature”. (Boyatt & Joy 2010: 9) – Constructing the class hierarchy, which provides the very basic structure composed of classes and their relationships. – Starting to construct individual instances of every class, all of which “should fit within the definition of that class”. (Boyatt & Joy 2010: 10) – Evaluating the ontology alongside “the construction of the mathematical logic learning materials”. (Boyatt & Joy, 2010: 10) By the early 2000s, terminologists such as Temmerman and her colleagues (cf. section 7. 3. 2) began to focus on ontologies as a more viable means of putting terminological work into practice, thus paving the way for the construction of terminological knowledge bases (cf. section 7.4.); this is the name used in linguistics-based terminological circles for addressing the creation of information tools that are associated with rules and/or axioms that are anchored at certain nodes or relations in the conceptual framework, with the result that ontologies can be used to generate automatic, rulebased inferences, such as answers to random queries”. Wright (2007: 166)

Genoma and EcoLexicon are two examples of the type of information tool linguisticsbased terminologists aim to design and construct for assisting users to understand the meaning of concepts and their use in several cognitive-oriented and communicative-oriented situations (cf. sections 7.3.1 and 7.3.3.). In addition to traditional generic-specific and part-whole relationships, advocates of these information systems propose other relations, typically those of cause-effect, object-function and ones which are domain-specific (Bowker 2008). In general, scholars working in this field claim that these systems are being designed and produced with the final aim of satisfying “not only the communicative expectations of the user, but also their cognitive expectations”. (Reimerink et al. 2010: 1929) As our analysis will reveal, this

Genoma and EcoLexicon | 179

sentence is also a “catch-phrase” that is not supported when real data are examined.

8.14.1 Genoma The homepage of the Genoma project and several other publications reveal that this project was started in 2002, and has both general and specific objectives. There are two general objectives. The first is connected with providing a sound theoretical basis for computer-aided unit detection, semi-automatic mapping of cognitive nodes and conceptual relations (Feliú et al. 2002: 1885). The second objective is to construct an information retrieval system, capable of improving current systems using terminological control. It is claimed that this second objective can be achieved by taking advantage of the grammatical, semantic and pragmatic information associated with the units that convey specialised knowledge (Feliú et al. 2002: 1885). The specific objectives can be summarised in a simple one: the construction of an information tool equipped with a terminological extraction system based on formal patterns and lexical ontologies that meets the needs of several users: translators, terminologists, lexicographers, information science experts, specialized writers and journalists, researchers and scholars, and, finally, linguists. (Hospedales & Rodríguez: 2004: 1305) Our analysis of this dictionary shows that the project is no longer under construction. At the time of writing this chapter (May, 2013), the information posted on the homepage of Institut Universitari de Lingüística Aplicada (IULA), which is the research group responsible for designing and constructing Genoma, was simply a summary of that published between 2002 and 2004 (Feliú et al. 2002; Cabré et al. 2004): – the project bank is designed as a modular structure with four modules: (i) a textual database; (ii) a document database; (iii) a terminological database; and (iv) an ontology; – the ontology is equipped with OntoTerm™, an ontology management tool (Moreno Ortiz 2000). This allows researchers to build an ontology integrating, at the same time, the ontology and a terminological database; – the ontology is built on four different but interrelated modules: the ontology, term- base, corpus, and entities module (Fig. 8. 1);

180 | An Analysis of Specialised Online Dictionaries

Figure 8. 1: Architecture of the Genome-KB project. Source: Feliú et al., 2002: 1886

the system works like this: 1.



Terminological units have clear-cut borders in each specialised domain and correspond to a concept to be found in an ontology. This approach led them to use an ontology management tool that permits the building of an ontology integrating both this and the terminological database. With the aid of a domain expert they created an ontology of 100 basic concepts. In the Ontology Editor, the user can introduce a brief description (concept definition) and the conceptual hyponymy-hyperonymy relation. The system “organizes the information on the basis of the concepts, their attributes and the conceptual relations among them”. (Cabré et al. 2004: 88) The ontology module is closely linked with the term-base module. This means that “no term entry is possible if its corresponding concept has not been previously introduced in the ontology”. (Cabré et al. 2004: 88) This module allows users to retrieve the following: (i) the term itself in Catalan, Spanish and English; (ii) the part of speech; (iii) the number and gender assignation; (iv) the usage contexts and their sources; (v) the lemmatised form and some administrative data. They add that the “term definition and its source and some usage notes are optional”. (Cabré et al. 2004: 88) All data are extracted from a corpus that has been linguistically processed as in the usual way for NLP applications. In 2004 the genomic sub-corpus contained texts in Catalan, Spanish and English. “Near one million tokens per language have been collected”. (Cabré et al. 2004: 89)

Genoma and EcoLexicon | 181

4. The document module complements terminological and textual corpora information. This module offers bibliographic and factographic data, i.e. references as well as the names and addresses of institutions working in the field of the genome. 5. The homepage is equipped with a query-system that “foresees three different query levels: simple, complex and combined”. (Cabré et al. 2004: 89) For each of the terms found, the system proposes five icons giving access to the following information: term base content; ontology information; term variants and equivalents in other languages; corpus contexts; and corpus frequencies. We have analysed Genoma in terms of the functional framework described in section 8.1. Our analysis reveals that access is restricted to registered users. As registration is free, we are unsure as to why it is necessary to register so as to access it. This process takes time and obliges users to remember the user name and password. Once a potential user has registered they access a homepage that contains four “tools”: (i) “banc terminològic/ontològic” (BT), (ii) “corpus textual” (Corpus), (iii) “banc documental” (BD), and (iv) banc factogràfic” (BF). For the purpose of this book, we will search only in BT, as the other tools have no functions connected with punctual consultation. For instance, “Corpus” retrieves corpus concordances and corpus frequencies, “BT” bibliographic data, and “BF” the names and addresses of institutions operating in the genome area. BT is a specialised trilingual online dictionary that can be used in cognitive and communicative situations; users are not identified. It retrieves data in Catalan, Spanish and English, although the data included may differ substantially, as we see in example (6), which shows definitions and contexts of DNA in English, Catalan and Spanish: DNA (English) Definició: The molecule that encodes genetic information. DNA is a double-stranded molecule held together by weak bonds between base pairs of nucleotides. The four nucleotides in DNA contain the bases adenine (A), guanine (G), cytosine (C), and thymine (T). In nature, base pairs form only between A and T and between G and C; thus the base sequence of each single strand can be deduced from that of its partner. Ref. 1 Contextos: Once the DNA sequence of the human genome is known, scientists will be able to compare the information to that produced by efforts to sequence the genomes of other species, yielding a fuller understanding of how life on the earth evolved. Ref. 2 Genes are made of DNA, which contains the instructions for making proteins. DNA never leaves the nucleus of the cell; its molecular recipes are read out in the form of messenger RNA, which leaves the nucleus and enters the cytoplasm, where proteins are made. Ref. 3

182 | An Analysis of Specialised Online Dictionaries Genes are carried in the DNA molecules of the chromosomes in the cell nucleus. Ref. 4 DNA (Catalan) Contextos: El DNA és una doble hèlix, semblant a una escala de cargol, formada per quatre tipus elementals de graons o nucleòtids, designats correntment amb les inicials A, C, G i T, que es van repetint i alternant fins assolir un nombre d'aproximadament tres mil milions en la dotació haploide de la nostra espècie. Ref. 1 La mutació és un procés poc usual que es produeix bàsicament en el moment de fer la còpia del DNA per ésser transmès a la descendència. Ref. 2 El DNA s'empaqueta en unes estructures anomenades cromosomes, que tenen forma de bastonet (com una I) amb un estrenyiment, el centròmer. Ref. 3 DNA (Spanish) Definició: Abreviatura de ácido desoxirribonucleico (en inglés deoxyribonucleic acid o DNA). Es la molécula que contiene y transmite la información genética de los organismos excepto en algunos tipos de virus (retrovirus). Está formada por dos cadenas complementarias de nucleótidos que se enrollan entre sí formando una doble hélice que se mantiene unida por enlaces de hidrógeno entre bases complementarias. Los cuatro nucleótidos que forman el ADN contienen las bases adenina (A), guanina (G), citosina (C) y timina (T). Dado que en el ADN la adenina se empareja sólo con la timina y la citosina sólo con la guanina, cada cadena del ADN puede ser empleada como molde para fabricar su complementaria. Ref. 1 Contextos: La reparación de las lesiones inducidas en el DNA es un proceso complejo que generalmente requiere la participación de una DNA polimerasa. Ref. 2 El Genoma Humano está compuesto por 23 pares de moléculas de DNA contenidas en estructuras llamadas cromosomas, que se localizan en el núcleo de la célula (genoma nuclear), y por una pequeña molécula de DNA circular contenida en un órgano de la célula denominado mitocondria (genoma mitocondrial). Ref. 3 Una enzima denominada polimerasa de DNA es responsable de la construcción de las nuevas cadenas del DNA, emparejando cada base de la nueva cadena con la complementaria de la cadena «vieja», que actúa de molde. Ref. 4 Example 6: Definitions and Contexts of DNA in Genoma.

The differences affect the conceptual description of the term; perhaps because these definitions were worded with corpus data may explain the important conceptual differences observed between the different entries: – In Spanish, users are told that DNA is an abbreviation for a molecule that encodes genetic information of organisms except some types of retrovirus. This information is not given in English or Catalan. – In Spanish and Catalan, users are informed that DNA is a double helix. This information is absent in English.

Genoma and EcoLexicon | 183

– – – –

In Spanish, users are told that the double helix is bonded together by hydrogen. This information is absent in English and Catalan. In Spanish, users are told that the bases are complementary. This information is absent in English. In Spanish, users are informed that the Human Genome is composed of 23 pairs of DNA molecules. This information is absent in English and Catalan. In Spanish, users are informed that each strand of DNA can be used as pattern in order to produce its complimentary molecule. This information is absent in Catalan and may be implied from the English definition, although this will take time and a good knowledge of the area on the part of the user.

We have not been able to discover the number of lemmas included in this dictionary. Our analysis reveals that most of them are multi-word terms that are accessed by using the search button “Que contingui” (contains). The system possesses another three search engines for “exact lemma”, “starting with”, and “finishing with”, which offer poorer results. For instance, when searching for sequence, molecular biology, hereditary, genetics, entirety, organism, encode, DNA, virus, and non-coding, by means of the four buttons, we found the following: – organism, and DNA (exact lemma); – DNA sequence; hereditary disease; hereditary elliptocytosis; hereditary ovalocytosis; eucaryotic organism; eukaryote organism; eukaryotic organism; multicellular organism; one cell organism; one-cell organisme; organism; pluricellular organism; procaryotic organism; prokaryote organism; prokaryotic organism; DNA; DNA replication; chloroplast DNA; complementary DNA; copy DNA; mitochondrial DNA; and pox virus; (“starting with”, “contains”, and “finishing with”). The above findings show that there are only two search systems: “exact lemma” and the rest of the search buttons. This activity takes a long time, with users needing to perform several searches before finding what they might be looking for. For instance, searching hereditary disease retrieves five icons in each language: terminological information, variants and equivalents, ontology, corpus: contexts, and corpus: frequencies. This means that users have to perform several searches before they can find, say, the equivalent needed in a translation situation. Terminological information contains a lemma, concept, language, definition, grammatical category, and contexts. Definitions and contexts are followed by codes such as, e.g. Ref.1, Ref. 2, etc. Clicking on them retrieves documentation sources, e.g. the text which was used for defining the term, also a good example of search-related information costs that has not been taken into consideration (Illustration 34).

184 | An Analysis of Specialised Online Dictionaries

Illustration 34: Terminological information for hereditary disease in Genoma (English)

Supposing a user is in a translation situation, then they have to perform another search by writing in the search engine the equivalent in the language they need. This means that if a user does not know that the Spanish equivalent of hereditary disease is enfermedad hereditaria, they cannot find what is required in the translation situation. In order to carry out such a search, they are obliged to search via another icon which retrieves variants and equivalents. On this screen, our user will retrieve the Spanish equivalent and by clicking on the icon for terminological information they will retrieve the Spanish term, its concept (the same as in English), language, definition (in Spanish), grammatical category (in Catalan), gender (in Catalan), contexts (in Spanish) and more terms that are supposed to be synonyms, although this information is not mentioned (Illustration 35).

Illustration 35: Spanish equivalent enfermedad hereditaria in Genoma

Genoma and EcoLexicon | 185

To sum up, Genoma is a long way from fulfilling the claim made by Feliú et al. (2002) concerning its purpose and objectives: The main goal of this first project is to provide a sound theoretical basis for computer-aided unit detection and semi-automatic mapping of cognitive nodes and conceptual relations. It is foreseen that our working methodology —oriented to improve information retrieval (IR) systems— would combine strategies both from the cognitive sciences and linguistics. We will also resort to indexation strategies and thesaurus building standards, coming from information science, and some other linguistic engineering working lines, such as natural language processing and statistical analysis. (Feliú et al. 2002: 1885)

8.14.2 EcoLexicon EcoLexicon deals with the environment, and organises its content in such a way that this goes from very general to more specific levels. On its Homepage it is claimed that the most generic level is the Environmental Event (EE), which provides a frame for organising the concepts in the information system (Fig. 8. 2).

Figure 8. 2.: The environmental event in EcoLexicon. Source: Reimerink et al. et al. 2010: 1932

Its authors have explained its main characteristics in several publications (e.g. Faber 2012; Faber et al. 2006, 2007; Prieto-Velasco & López-Rodríguez 2009; Reimerink et al. 2010): Thus, the EE has two types of agents that can initiate processes. Such agents can be inanimate (natural forces) or animate (human beings). Natural agents, such as water movement (e.g.

186 | An Analysis of Specialised Online Dictionaries waves, tides, and currents) and atmospheric phenomena (e.g. winds and storms) cause natural processes such as littoral drift and erosion in a geographical area such as the coast. These processes affect other entities or patients (e.g. beaches, sea ports, and seabed) which, as a result, may suffer changes (e.g. loss/creation of beaches, and modifications in seabed composition). Human agents can also implement artificial processes (e.g. constructions), which can generate or prevent effects normally caused by natural processes. For instance, a coastal construction can be created with the purpose of controlling erosion. With the help of instruments, man can build structures such as groynes, whose function is to protect a shore area, retard littoral drift, reduce longshore transport, and prevent beach erosion. (Reimerink et al. 2010: 1932)

These publications also reveal the main characteristics of EcoLexicon, which is in turn designed and constructed under the tenets of Frame-Based Terminology (cf. sections 7.3.3 and 8. 10.): – EcoLexicon offers access to a definition that describes each term conceptually, lexically and pragmatically. This definition goes hand in hand with images of the concept, where appropriate. These graphical resources and combined contexts are accessed through the tag “Resources” (left side in Illustration 36). – EcoLexicon also offers conceptual knowledge, which “is displayed in a 3D visual thesaurus, where concepts are placed in terms of their semantic relationships with other similar concepts”. (Reimerink et al. 2010: 1932) – Users have access to a list of specific frames, “under the tag ‘Domains’, with hyperlinks to the parent frames of the concept, i.e. MOVEMENT, NATURAL PROCESS and PROCESS.” (Reimerink et al. 2010: 1933) – Users have access to frame relations, “displayed in a dynamic network of related concepts such as FLOW and WATER-CYCLE, related to the concept INFILTRATION through an IS-A and PART-OF relation, respectively”. (Reimerink et al. 2010: 1933) – Users have access to the terminological units designating the concepts in several languages. – Users have access to data regarding selection patterns, as well as the collocational and grammatical tendencies of each terminological unit. These are listed in the box “Terms” (left side in Illustration 36). – Users have access to contexts that offer syntactic, conceptual and pragmatic data.

Genoma and EcoLexicon | 187

Illustration 36: Screenshot for EcoLexicon

We have analysed EcoLexicon in terms of the above functional framework (cf. section 8.1.) and have found that the dictionary is an example of a Stray Bullet (cf. section 3.2). Hence, it contains several limitations. The first is that experts in the field do not make their contributions on a regular basis, which may explain why the definitions are rather poor. For instance, the definition of water (example 4; c.f. section 7.3.3.) does not include basic information that may to some extent affect the environment, which is the subject field covered by the dictionary: – water is a chemical substance; – H2O is its chemical symbol; – water is a liquid at temperature above 0 C at sea level, but a solid at temperature below 0 C; – water is also found as a gas; – water also exists in a liquid crystal state; – water covers around 71% of the Earth surface; – the role of water in erosion, etc. The second main limitation is the access system. We found that the “3D visual thesaurus” takes several minutes to retrieve the data searched, and that during this time the screen is moving up and down, which is, in our view, an important drawback of the system. The third principal shortcoming is that it does not target specific users and, therefore, the data included and the manner in which it is dealt with lexicographically is not very helpful in communicative or cognitive usage situations.

188 | An Analysis of Specialised Online Dictionaries For example, when searching for water in the search engine, located in a horizontal menu at the top of the screen, users have to select the language (it can be English or Spanish). Then, we obtained a kind of cloud with more than 70 terms in several languages, and water appeared on the left of the screen. Clicking on Water retrieved a cloud of over 200 terms in several languages, each of which was related by generic-specific, part-whole, and non-hierarchical relations, represented by green, blue and red lines (Illustration 37). In short, this is a good example of how searchrelated information costs have not been considered in the design of the dictionary.

Illustration 37: Dictionary article for water in EcoLexicon

To recapitulate, then, Genoma and EcoLexicon fall short of their stated objectives. They are not updated regularly, do not usually employ experts in the handling of data, do not target specific users in specific user situations, do not appear to have many lemmas, and employ rather primitive and time-consuming search systems. We believe their main shortcoming is that they are using incorrect methodologies

Summary and The Work Ahead: Working with the Function Theory of Lexicography | 189

and, consequently, do not offer data for real situations involving real users. Neither system can be used in a translation situation: a translator will need around five minutes per consultation and will have to “imagine” a large amount of data that might be included but which is not presented in a simple way. Furthermore, these two systems are not conceptually different, despite the fact that their designers espouse different theories of terminology (cf. sections 7.3.1 and 7.3.3.). Moreover, what is the purpose in a translation context of the relations shown at the bottom right part of the screen and the hierarchy of categories? A translator would appreciate that, e.g. water is an uncountable noun, that it can also be a verb, etc. However, the absence of such data demonstrates the flawed conception of both dictionaries.

8.15 Summary and The Work Ahead: Working with the Function Theory of Lexicography Table 8.2 summarises the findings of the 16 dictionaries reviewed in terms of the functional framework presented in 8. 1. Table 8. 2: Summary of the Functional Framework applied to 16 dictionaries

Business & Manage. Palgrave Musikord. Mortgage Cambridge TermFin. A.D.Miles FAO Kicktiona. IATE Cercaterm UNTERM Medical Terms Disco Genoma EcoLex.

Author’s Funcview tion. (1) (2)

Access Internet Lexicog- Produc. Inform. Updat- Experts Data tech. raphy costs cost ing Select. (3) (4) (5) (6) (7) (8) (9) (10)

Yes Yes Yes No Yes Yes No No Yes Yes Yes Yes Yes

Yes Yes Yes Yes Yes Yes No No No No Yes No No

No Yes Yes No No No No No No No No No No

No Yes Yes No No No No No No No No No No

Yes Yes Yes No Partly No No No No No Partly No No

Yes Yes Yes Na Yes Yes Yes Yes Yes No No No No

No Yes Yes Yes No Yes No No No No Yes No No

No Yes Yes No No No Yes No No No Yes Yes No

Yes Yes Yes Yes No Yes No Yes No Na Na Na Yes

Yes Yes Yes Na Yes Yes Na Yes Yes Yes Yes Yes Yes

Yes Yes Yes

No No No

No No No

No No No

No No No

No No No

No No No

Yes No No

No No No

Yes Yes Yes

Table 8.2 shows several findings. The first is that there is solid justification for proposing a theory and practice of specialised online dictionaries. This book supports the argument that such endeavours must be based on a sound lexicographical the-

190 | An Analysis of Specialised Online Dictionaries ory, e.g. the Function Theory of Lexicography, which is presented as the bedrock for the design and production of online specialised dictionaries. In the same vein, the functional framework presented in this chapter and used for analysing an array of specialised online dictionaries allows us to identify several practices that can be improved or totally discarded. These are as follows: – An analysis of access routes and information costs shows that most dictionaries are not using various existing Internet technologies for presenting dictionary articles. Putting it another way, most dictionaries do not contain dynamic articles with dynamic data, i.e. data adapted to a specific user situation (cf. 8.1 “Internet technologies”). Instead, these dictionaries prefer accumulating data rather than offering ways to individualise searches. – An analysis of functions shows that some dictionaries take for granted the concept of user needs. In other words, although they might refer to and present the concept, they do not include data that support the function(s) and user’s needs in specific user contexts. This means that most of the data presented in the dictionary do not support any specific function nor contain dictionary components which are needed to sustain the functions envisaged. – An analysis of lexicography shows that some dictionaries, especially those designed and compiled by linguists who have suddenly become terminologists, are using incorrect methodologies for applying these two activities to specialised online dictionaries. A case in point is the use of linguistic corpora, e.g. corpora designed and compiled in the manner defended by Sinclair and his followers, for selecting lemmas, preparing definitions and describing the grammar of the particular term. We believe that these three areas – lemma selection, definitions of terms and grammar descriptions- should use corpus data only as auxiliary tools and that this should be employed in very restricted tasks, e.g. finding examples or language patterns (cf. chapter 9). – An analysis of updating shows that some dictionaries are not updated regularly or are updated only sporadically. This practice is especially dangerous when dealing with specialised topics and languages. A principal characteristic of many domains is that they are in a constant process of adaptation to new regulations, rules, standards, etc. Hence, considering lexicographical costs and devising ways of updating specialised online dictionaries is a crucial aspect in any practice concerned with these. – An analysis of authors’ view reveals that authors and dictionaries tend to include outer texts explaining key aspects of the reference work in question. However, there is still plenty of room for upgrading this variable: some dictionaries, e.g. Genoma, do not include relevant data such as number of lemmata and targeted users. – An analysis of production costs shows that some dictionary projects are availing themselves of considerable amount of funding, although the end result leaves something to be desired. This is especially visible in many EU-funded dictionary

Summary and The Work Ahead: Working with the Function Theory of Lexicography | 191

projects, which aim to offer dictionaries and other tools at the same time. The results of these projects do not justify their financing. An analysis of experts reveals that some dictionaries do not employ experts in their production process. This is quite problematic. We believe that Internet experts are fundamental in the production of online dictionaries, and that experts in the field cannot be replaced in specialised lexicography. Without these two types of experts specialised online dictionaries cannot be constructed, and as a consequence they are absolutely necessary when it comes to producing highquality specialised online dictionaries.

To sum up, our analysis indicates that we need a theory and practice of specialised online dictionaries to help lexicographers design and produce quality information tools. In this book, we have defended the Function Theory of Lexicography as a way to direct the practice of designing, constructing and updating specialised online dictionaries. Our next chapter shows how this theory has guided our practice.

9 Designing, Making and Updating Specialised Online Dictionaries In previous chapters we have argued that lexicography constitutes a system of knowledge arising from social practice, and have emphasised that any lexicographical theory must include instructions for practical action (cf. section 4.1). We have also explained our conviction that one of the things characterising theories formulated within the humanities and social sciences is that the subject field, i.e. dictionaries in lexicography, undergoes constant changes. These ideas constitute our case for advocating transformative theories, that is, those with a potential to orientate the development of an efficient methodology that may be applied not only to analysing and describing existing dictionaries, but also to furnishing a more accurate picture of the real user needs; this would mean determining the pertinent data to be included in a specific dictionary without having to resort to more complex, dubious and time-consuming methods. Now we will explain the practice of designing, compiling and updating specialised online dictionaries. This practice is based on the tenets of the function theory described so far, and is illustrated with data from the Accounting Dictionaries and/or the Diccionarios de Contabilidad. The set of dictionaries referred to as Accounting Dictionaries or Diccionarios de Contabilidad are the result of a joint project involving teams from the Centre for Lexicography at Aarhus University in Denmark, and the International Centre for Lexicography at the University of Valladolid in Spain (Bergenholtz 2012; Bergenholtz & Nielsen 2006; Fuertes-Olivera 2009, 2011, 2012; Fuertes-Olivera & Nielsen 2011, 2012; Fuertes-Olivera & Niño Amo 2011; Fuertes-Olivera & Tarp 2011; Nielsen 2002a and b; Nielsen & Almind 2011; Nielsen & Fuertes-Olivera 2013; Nielsen & Mourier 2005 and 2007; Tarp 2012c). Each team consists of lexicographers and accounting experts. In addition, programmers and experts in Information Science are working on the project, together with experts in the marketing and selling of online dictionaries. These experts are employed by, Denmark's biggest provider of Web-based dictionaries and/or, its German-based subsidiary.

9.1 A Multidisciplinary Activity Designing, compiling, and updating a specialised online dictionary is a multidisciplinary activity, usually headed by a main editor or main lexicographer. The advent of the Internet has modified the timing of traditional lexicographical tasks, such as selecting lemmas, equivalents, crafting examples, etc. These activities, which are described and illustrated in sections 9.2 to 9.4, must be performed once IT experts and the main lexicographer have prepared storage systems that allow the scope of the lexicographical work to become concrete.

A Multidisciplinary Activity | 193

The making of specialised online dictionaries starts with IT experts and lexicographers working together with two main objectives. The first is to translate the lexicographical vision of the planned work into a storage system, e.g. a lexicographical database, whose design and technical characteristics must facilitate this vision. In the Accounting Dictionaries and the Diccionarios de Contablidad, Professor Bergenholtz’s work with Richard Almind (IT expert) resulted in the design of the accounting lexicographical database. This allows lexicographers, experts and translators, who are based in different countries, to work on the project together and communicate easily. For instance, the “memo” file in the accounting database is reserved for internal communication among the members of the project. The design of the accounting database is founded on a key idea: there is a practical distinction between dictionaries and databases, as shown in Bergenholtz (2011 and 2012), and Nielsen & Almind (2011: 142-148) and summarised below: – Databases are storage systems and nothing else; they have no functionality per se. They contain discrete fields that store specific types of data. For instance, discrete fields such as collocation, example or synonym are used in the accounting lexicographical database, i.e. the storage system used for storing the data types (Table 9. 2., below) prepared for compiling the Accounting Dictionaries and the Diccionarios de Contabilidad. – When designing a lexicographical database, it is very important to avoid redundancy. For instance, examples and collocations can be reused and addressed to several specialised lemmas: In a database, a specific example only exists virtually on the computer and can be related to many terms. Changing it in one place changes it everywhere it is being used. Relations can under certain circumstances be made automatically, but this is not recommended. (Nielsen & Almind 2011: 142)

Interdependence of data is very important when designing a lexicographical database. For instance, grammar cannot exist without a lemma: In order to create such dependencies, data elements are placed into tables, for instance a table for collocations, one for examples and so on. These fields are related hierarchically to each other and form paths along which queries can travel to retrieve data. Data retrieval then becomes a matter of querying the database in a predetermined order and collecting data as the query moves along the relational paths. (Nielsen & Almind 2011: 142-143)

Data elements that are not related cannot be retrieved and can put at risk the coherence of the database: Either the lexicographical element is coded into the relational hierarchy or it is omitted entirely. Usually, the process of creating a relational hierarchy reveals serious flaws in lexicographical instructions and, although painful, correction of those flaws always leads to more coherent dictionaries. (Nielsen & Almind 2011: 143)

194 | Designing, Making and Updating Specialised Online Dictionaries –

The data displayed when sought are called a data set. This set depends on users’ needs as the set is formatted and displayed to them. “At this point, when the result is formatted and displayed to users, the data ceases to be a part of the database and becomes the dictionary”. (Nielsen & Almind 2011: 143) The structure of the database is Anglo-centric and has formed a hub-and-wheel structure that allows monolingual use, as shown in Figure 9. 1. With this structure, consolidation will take place at the level of definitions, as shown in Figure 9. 2.

Figure 9. 1.: Relational database structure for monolingual use. Source: Nielsen & Almind, 2011: 145.

A Multidisciplinary Activity | 195

Figure 9. 2.: Relational database structure for bilingual use. Source: Nielsen & Almind, 2011: 146.

Figure 2 shows, for instance, that for the purposes of the bilingual dictionaries, “translation is defined as the process of finding the data related to the definition of any given L2-accounting term, which in turn is related to the definition of any given L1-accounting term as a result of user queries”. (Nielsen & Almind 2011: 146) These dictionaries have a triadic structure: (a) a lexicographical database; (b) a user interface where one or more dictionaries are placed; (c) a search engine that mediates between the database and the user interface: This search engine allows users to search for data in the database and from there it retrieves the relevant data for each of the dictionaries and presents the results of searches to users according to their requests. In this case, there are several dictionaries and one database, and the relationship between database and dictionary is, thus, a one-to-many relationship. (Nielsen & Almind, 2011: 147)

The second objective concerns the formal properties of online dictionaries. These must facilitate consultation and be in line with the ultimate goal of the project. For instance, if the dictionary is going to be sold through subscription, its formal properties must comply with the marketing of the company. The formal properties of the Accounting Dictionaries and the Diccionarios de Contabilidad, which have been extensively discussed, e.g. Almind (2005); Bergenholtz & Johnsen (2005); Fuertes-Olivera & Niño Amo 2011; Nielsen & Mourier (2005 and 2007), are currently visible on an Internet homepage, run and maintained by IT

196 | Designing, Making and Updating Specialised Online Dictionaries experts from and This homepage (Illustration 38, below) facilitates consultation and is the result of the work of Professor Bergenholtz, Richard Almind and several IT experts enrolled in or

Illustration 38: Homepage of the Accounting Dictionaries (excerpts).

An analysis of Illustration 38 shows that the homepage of these dictionaries contains the solutions prepared for coping with the formal properties of online dictionaries (Table 9. 1.). Table 9. 1.: Demands and Solutions related to the formal properties of the Accounting Dictionaries and the Diccionarios de Contabilidad. Source: Adaptation from Almind (2005). Demand An Internet dictionary must be easy to find It must have smart searches It must strengthen interactivity

It can be part of a reference portal It must use an elegant and pleasant layout

Solution It has a simple Internet-address, which is maintained in operation 24-hourss a day. It has smart searches (cf. 9. 4, below). Editors can be consulted, and it contains features like “tools” and “chat”, and other functionalities that facilitate interaction. It is in a dictionary portal: and It uses user-friendly colours, fonts, font-sizes, etc. It also divides the page static parts (for example the part for header, footer, and navigational menu), and a dynamic data part.

Pre-compilation phase | 197 It must use a familiar and reassuring virtual environment It must allow users to retrieve data in different formats It must be provided with extra software The search field must be the centre of attention It must have readable articles

It must be regularly updated and corrected It must offer instant, simple, and a reduced number of results It must use advanced searches It must display the results logically

The screen does not contain ‘extra elements’ such as advertisements. Users can access internal and external texts through hyperlinks. It contains links to other functionalities. For example, it allows the copy-and-paste facility. The search field is visually centred near the top of the homepage. It uses “good” colours for important data that are easy to see and read. For instance, black on white is good, but red on blue is much less readable. Contrast is also used in the right way, and each piece of information is put on its own line and uses headings prudently. It is updated every three months. It offers instant and simple results. Depending on the function, the amount of results can be simpler or longer. It uses all alternatives to an alphabetically sorted list. The results are mostly sorted alphabetically.

What follows describes the work of lexicographers, accounting experts and translators who store the data in the accounting database. This work is continuous and has become a very successful business operation. By way of example, the Danish and English combinations of the Accounting Dictionaries, which are currently sold by, have been instrumental in this Danish company being granted the Gazelle award for five consecutive years. This has meant that Ordbogen has been the company with the highest growth rate in Denmark over the last five years. To sum up, the process of designing, compiling and updating specialised online dictionaries is a multidisciplinary activity that demands skills and knowledge of Lexicography, Information Science, and one or more subject field(s). In addition, it also needs a business model if the dictionary is intended to be sold. These skills and knowledge can be applied in order to prepare storage systems designed for allowing lexicographers to see the vision of the projected lexicographical work materialise.

9.2 Pre-compilation phase The pre-compilation phase contains two main and closely interconnected tasks; these are concerned with (a) deciding on potential users in a specific context or situation, (b) detecting lexicographically relevant information needs among potential users, and (c) deciding to set up a lexicographical project in order to satisfy their needs (cf. section 5.6). This phase was accomplished with the methodology advocated in the function theory, which is basically a combination of experts’ knowledge on well-defined group(s) of potential users and deduction (cf. section 5.2). In practical terms, lexicographers must (a) decide on users’ needs in potential user situa-

198 | Designing, Making and Updating Specialised Online Dictionaries tions, and (b) evaluate human, financial and other resources that could be employed in order to have working dictionaries within two/three years (Nielsen 2002a and b). Regarding user needs in potential user situations, the Accounting Dictionaries and the Diccionarios de Contabilidad are directed at three Danish and Spanish user groups: (1) translators and language staff; (2) accounting experts and semi-experts; and (3) students and laypersons interested in Danish, English and Spanish accounting topics. These may consult lexicographical data when they are reading, writing and translating accounting texts (i.e. in communicative situations) and/or gain some knowledge about accounting (i.e. in a cognitive situation). The above analysis conditioned the evaluation of human and material resources: the project needed experts in lexicography, accounting, and information science, all of whom started working at the Centre for Lexicography, University of Aarhus, in 2003. Four years later (in 2007) experts in lexicography, accounting, and translators from the University of Valladolid joined the project. Since then, both groups have continued to work together, and the project has resulted in several combinations of Danish, English and Spanish accounting dictionaries. Both groups have used private and public resources, totalling around 5,000 Euros per language per year. Briefly, our experience shows that a project such as the Accounting Dictionaries and the Diccionarios de Contabilidad needs the guidance of one main editor, who must be a lexicographer, two or three experts per language (at least one lexicographer and one expert in accounting), and one or two programmers along with one Internet expert, that is, experts in designing tools for the Internet.

9.3 Compilation Phase The compilation phase runs almost parallel to the pre-compilation phase. During this period several decisions must be taken. These are basically concerned with dictionary concepts, preparing the material needed for the dictionary project, writing the lexicographical instructions, assignment of specific tasks to every member of the team, and preparation for the book’s publication. The first task is formulating a lexicographical database that could be used for storing data with which our envisaged users could solve potential needs in communicative and cognitive situations. Table 9. 2. shows the data types of the accounting database, i.e. our lexicographical database, and the rationale for including such types, which is based on a user needs analysis in communicative and cognitive situations:

Compilation Phase | 199 Table 9.2: Data Types in the Lexicographical Accounting Database Data Type Lemma Homonymy index (superscript in the dictionary)

Rationale Self-evident: all dictionaries describe lemmas Included when homographs have different inflectional paradigms or are countable or uncountable, respectively. It assists all user types in all situations: it restricts the meaning of lemmas. Polysemy index (Arab numbers in Included when homographs have different definitions. It assists the dictionary all user types in all situations. It restricts the meanings of the same lemma. Language code to lemma Included for indicating language and language variant of the lemma. It assists all user types in all situations. Grammatical data addressed to It offers inflections, countability, and active and passive forms of lemma verbs. It assists all user types in two communicative situations: production and translation. It offers linguistic profiles of lemmas. Equivalent One Danish, English or Spanish equivalent in several dictionary combinations: Danish-English (Danish Lemma; Danish definition; English equivalent); English-Danish (English lemma; English definition; Danish equivalent); English-Spanish (English lemma; English definition; Spanish definition; Spanish equivalent); Spanish-English (Spanish lemma; Spanish definition; Spanish equivalent). It assists all user types in communicative situations. Language code to equivalent Included for indicating language and language variant of the equivalent. It assists all user types in communicative situations. Grammatical data addressed to It offers inflections, countability, and active and passive forms of equivalent verbs. It assists all user types in two communicative situations: production and translation. It offers linguistic profiles of equivalents. Definition of lemma It explains the meaning(s) of lemmas. Always one definition per sense. Definitions are crafted in the language of lemma. There is one exception: in the English-Spanish dictionary, definitions are in English and Spanish. It assists all user types in all situations. Collocations They are short and long phrases but not full sentences. They are crafted in the language of the lemma. They mainly assist users in two communicative situations: production and translation. In addition, the Spanish dictionary also includes collocations that assist users in cognitive situations (this is a useful method for dealing with culture-bound subject fields). Translation of Collocations In some dictionary combinations collocations are translated: English collocations are translated into Danish and Spanish, and Danish and Spanish collocations are translated into English. They assist all user types in two communicative situations: production and translation. Language code to translation of Included for indicating language and language variants of the collocations lemma and equivalent. It assists all user types in two communicative situations: production and translation. Examples They are full sentences showing lemmas in use. They assist all user types in cognitive situations and two communicative situations: production and translation.

200 | Designing, Making and Updating Specialised Online Dictionaries Translation of Examples

They are full sentences showing equivalents in use. They are in some dictionary combinations: English examples are translated into Danish and Spanish, and Danish and Spanish examples are translated into English. They assist all user types in cognitive situations and two communicative situations: production and translation. Synonyms and antonyms adThey assist all user types in two communicative situations: prodressed to lemmas or equivalents duction and translation. Language code to synonyms and Included for indicating language and language variant of the antonyms synonym and antonym. It assists all user types in two communicative situations: production and translation. Source Hyperlinks to external texts. It mainly assists experts in cognitive situations. It can also assist translators and semi-experts in two communicative situations: production and translation. Lexicographical notes Usage, and/or contrastive notes that are addressed to lemma or equivalent, especially for indicating factual differences between lemma and/or equivalent, particular usages, cultural details, etc. They mainly assist all user types in cognitive situations. Grammar notes attached to lemma They indicate language profiles of some lemmas and/or equivaand/or equivalent lents. They mainly assist users in two communicative situations: production and translation. Proscriptive notes They are used for recommending lemmas. They mainly assist all user types in communicative situations. Cross-references Hyperlinks to internal texts. They assist all user types in cognitive situations and two communicative situations: translation and production.

Table 9. 2. shows that each data type may have one or more specific goal(s) within the general framework of the user's needs, functions, and resources agreed on at the start of the project. Next, we will comment on several aspects of the data types, more specifically the methodology used for selecting them and their main function(s). For operational reasons, they are separated into subsections, although these tasks are usually undertaken simultaneously and not necessarily in the order described here.

9.3.1 Lemma Selection Lemma selection is a continuous process, and must be performed by lexicographers and subject field experts working side by side. It starts by delimiting the subject field, which in the case of accounting comprises decisions concerned with "macroaccounting", typically the treatment of national accounts, and "microaccounting", which deals with internal reporting, i.e. management accounting, and external reporting or financial accounting. Although the three branches of accounting are included in the attention paid to accounting terms and concepts in the Accounting Dictionaries, there is a preponderance of financial accounting with respect to the other two branches: around 85% of the terms come from financial accounting,

Compilation Phase | 201

whereas approximately 15% deal with national accounts and management accounting, e.g. cost accounting (2% and 13%, respectively). The above imbalance is in accordance with users' needs and user situations: most published accounting texts are concerned with financial accounting, e.g. the following text types: (a) national and international standards, codes and regulations such as European Union legislative texts, International Accounting Standards (IASs) and International Financial Reporting Standards (IFRSs); (b) a company's financial statements; (c) auditing; (d) manuals, handbooks, interpretations and practical examples of accounting problems and solutions. National accounts are peripheral areas of accounting and are mostly concerned with balance of payments data and similar comparable statistics. Management accounting is basically an internal activity, usually related to cost accounting. As a result, there are not many texts available, and this activity is not very systematised. For instance, English loans, calques and borrowings crop up regularly in Spanish cost accounting texts, that is, management accounting. This has a lexicographical solution in the Diccionarios de Contabilidad, which is the inclusion of these English terms either as Spanish ones, equivalents or synonyms of a Spanish lemma. For instance, ABC (an English lemma) and método ABC (a combination of Spanish and English) (example 7) were included as synonyms of the Spanish lemma systema ABC, also a hybrid of Spanish and English. We believe that this solution helps potential Spanish users to understand terms whose Spanish counterpart, should there be one, is not particularly stable. The solution adopted is documented, usually by performing Google searches such as "el método ABC" (one of our synonyms) + "contabilidad". This retrieved around 40,000 hits, which shows that the solution adopted in the dictionary is in line with potential Spanish users’ expectations. sistema ABC sinónimo ABC método ABC sistema de coste basado en la actividad Example 7: Synonyms of sistema ABC in the Diccionarios de Contabilidad

A word of warning is necessary before continuing. We believe that using the web as a lexicographical corpus makes sense provided that “safe practices” are adopted. In other words, all the data taken from Google searches must be analysed by the corresponding expert working on the dictionary project; it is this expert who, in the end, decides on their inclusion or rejection. In this regard, we used Google searches to complete the profile of the Spanish lemma señal: googling “señal” retrieves almost 69 million hits. Such a number of hits show that señal is polysemous: it is used in general Spanish, Information Technology, Medicine or Accounting, etc. In order to

202 | Designing, Making and Updating Specialised Online Dictionaries obtain its language profile in accounting, we used the accounting definition crafted by our expert in this area as a restriction element for the several Google searches undertaken: (a) we searched “señal” + “arras” (En: pledge) + “contabilidad” (En: accounting), retrieving hits such as “contrato de arras”; (b) we then searched “contrato de arras” + “señal” + “contabilidad”, and obtained 11, 000 hits; (c) an analysis of the first 500 hits showed that “contrato de arras”, “señal”, and “contabilidad” coincided with “entrega a cuenta” (En: down-payment); (d) this prompted us to search “señal” + “contabilidad” + “down-payment”, for which we obtained around 400 hits. Its analysis allowed us to complete several data types: – the English equivalent of señal is pledge; – the language profile of señal in accounting is ; i.e. it is always singular; – the accounting definition of señal must indicate that this term is like a kind of down-payment, which is used in mortgage agreements with the aim of reserving property. The next step is to select lemmas out of texts that are representative of the subject field, e.g. the three branches of accounting. We started with well-known general accounting texts, i.e. texts that do not deal with specific companies, problems, situations, etc., but with general aspects of accounting. For instance, the Spanish team started the selection process with existing accounting dictionaries, e.g. the equivalents of our English-Spanish accounting dictionary, as well as institutionalised accounting texts, that is, standards, codes, legislative texts (e.g. the Plan General Contable, i.e. the current Spanish accounting code), manuals and handbooks. Almost 60% of the lemma list of the Diccionarios de Contabilidad, around 4,000 terms, come from these representative and general sources. For instance, there are around 645 Spanish IAS/IFRS terms that are lemmatised in the Spanish accounting dictionary. Most of these have equivalents in Danish and English that also occur in institutionalised texts downloaded from Internet sites, such as "The EU Single Market" and "IFRS". Private texts are our second main source of lemmas: for instance, there are around 1,500 Spanish terms (around 22% of the lemmata) in the Diccionarios de Contabilidad that have been selected from private texts. Private texts deal with specific situations, problems, and practices, such as the financial reports, financial statements and auditing reports of a specific company, e.g. those commented on and analysed in Memento Contable, which is a book of observations and explanations of Spanish accounting rules, standards, practices, etc. One key element of these texts is that they are uniform in that they refer to the same concepts. This means that a representative text of a company, for example, its income statement, is sufficient for working with accounting concepts that are used by all companies in the same sector or industrial activity. In other words, representativeness in accounting is not associated with frequency but with relevance, both

Compilation Phase | 203

function-relevant user characteristics and consultation-relevant user characteristics (cf. section 5.2). For instance, non-profit organisations use different accounting texts from those of quoted entities. As a result, the lemma selection team chose two relevant text(s) from both types of organisations rather than selecting many text types of the same kind of organisation. The lemma selection team worked with the Profit and Loss Account of Banco Santander, which contains the same concepts as Banco Bilbao Vizcaya (both Spanish multinational banks), and its equivalent in a nongovernment organisation, which is known as Income and Expenditure Account. Although these two text types are very much related, they also contain key differences that are based on the different goals these two types of entities are trying to achieve. Private texts also contain culture specific elements, i.e. elements that describe accounting traditions. Only experts in accounting can easily spot and illustrate these elements, especially similarities and differences between variants, accounting operations, etc. By way of example, the Spanish accounting concept “amortización” corresponds to two English lemmas: amortisation and depreciation, which refer to two different English concepts. Due to English influence, Spanish accountants are also using depreciación as a synonym of amortización, which was not originally the case in Spanish accounting. Hence, we decided to include depreciación and amortización as Spanish lemmas, and indicate that they are treated as synonyms in Spanish although they are conceptually different in English. An in-house corpus of around 3 million words was also used to select lemmas. This corpus basically contains research and popularised online retrieved accounting texts, i.e. those extracted from accounting journals such as Partida Doble and accounting articles published in dailies, bank leaflets, prospectuses and so on. Most of the approximately 1,000 lemmas selected from the corpus have two main characteristics: (a) they consist of one or two orthographic words; (b) they are either semiterms, i.e. terms that are used in several subject fields with different meanings, or words, that is, general words that may also have a specific meaning in accounting. For instance, the lemma accumulate, i.e. "to increase or grow over a period of time because of periodical additions" (definition in the Accounting Dictionaries) is lemmatised because it forms lexical profiles with key accounting terms such as cost, cash, financial instruments, etc. Similarly, the term representation (a polysemous and homonymous term) is lemmatised because its three senses show nuances of meaning that are relevant in accounting (example 8). 1. representation1 (a representation, the representation, representations) 1. A representation is a formal or factual statement made by e.g. management, an auditor or a party to a contract. 2. A representation constitutes a particular format, layout or way in which something is presented, e.g. a plan, data or financial statements.

204 | Designing, Making and Updating Specialised Online Dictionaries 2. representation2 (no indefinite article, the representation, no plural) Representation exists when e.g. a person or an enterprise is present or acts officially on behalf of another person or enterprise, i.e. represents that person or enterprise for a particular purpose or in a particular place. Example 8: Representation in the Accounting Dictionaries

Finally, another source of lemmas must be the use of a systematic compilation process. This means that during the process of compiling dictionaries the lexicographic team is aware of synonyms, antonyms and new or changed terms. This process is very productive and easy to implement when working with experts in the field. These readily produce synonyms, antonyms and other related terms when a definition is crafted. For instance, around 500 new terms have been added in the Diccionarios de Contabilidad during the 8-month period spent in making definitions of the previously selected terms. One example is sistema institucional de protección (abbreviated as SIP) and its synonym fusión fría, a colloquial term introduced in 2010 in Spanish public discourse to refer to fake mergers of Spanish saving banks. Our experts easily found that both terms referred to the same concept, although sistema institucional de protección is an institutionalised term and fusión fría is the ad-hoc formation coined by newspapers. Briefly, the process of lemma selection is based on the concept of relevance (Tarp, 2008; Fuertes-Olivera & Nielsen 2011 and 2012; Bothma & Tarp 2012) (cf. also section 3.4). We mainly selected lemmas by relying on expert knowledge, both lexicographic and that of accounting. Hence, this process was totally focused on the potential usefulness of the selected lemmas in several user situations rather than depending on frequent terms, as described in many other dictionary projects. In reality, frequent lemmas are not very useful in specialised lexicography, which may explain why so many projects fail, insofar as they aim to design specialised dictionaries that start with frequency and continue for a long time without reaching completion: The internet has allowed the compilation of new types of information tools, e.g. the so-called terminological knowledge bases. These proliferate around the world, especially because they obtain public money easily, although most of them do not deliver much. For instance, around 90% of the terminological dictionary projects funded by Spanish R + D funding agency are still "prototypes" after several years of continuous and generous funding. (Fuertes-Olivera, 2013a: 337)

Relevance is also useful in this process as it allows experts to take informed decisions in some situations, e.g. when Spanish equivalents of English lemmas cannot be found or when English lemmas appear routinely in Spanish accounting texts. In such situations, we adopted three decisions:

Compilation Phase | 205

– –

Some loans, calques and borrowings that exist side by side with their Spanish counterparts were also included as Spanish lemmas (and equivalents). For instance, a Google search of "el stress test" + "banca" or + "banco" retrieved around 30,000 hits each. Similarly, a Google search of "los stress test" + "banco" retrieved more than 130,000 hits. The Spanish counterpart (“prueba de resistencia” + “banca” or “banco”) retrieved more than 30 million hits. Certain proper names, such as FASB and similar accounting bodies, which are named in English, were also lemmatised as Spanish lemmas. Occasionally, Spanish lemmas were created from scratch, of which there are around 100; e.g. enfoque estándar del coste por absorción is our translation of English absorption standard costing. This situation was practically restricted to management accounting, and especially to cost accounting: sistema de coste retroactivo (En: backflush costing); gestión de costes basada en la actividad (En: activity-based management), etc.

9.3.2 Explanation of meaning Explanation of meaning is a key lexicographical activity whose purpose is to offer potential users data types that facilitate the understanding of each term. This is basically concerned with crafting definitions and attaching equivalents where necessary (Nielsen & Mourier, 2007). These types of data are found in the accounting database (Table 9.2., above) and their inclusion was decided once the user profile in several user situations was decided. Lexicographers and subject experts must perform this activity jointly. The subject expert describes/explains the concept; the lexicographer decides on the definitional style and writes the definition. In the Accounting Dictionaries and the Diccionarios de Contabilidad, definitions are crafted as complete sentences, usually by relating the definiendum to the definiens using a word or phrase which expresses a clear relationship of identity between them, e.g. “is”, “es”, etc. There are some exceptions to this definition type, e.g. function words such as less, and cum, but these exceptions are irrelevant in the dictionaries (for instance, there are only 31 function words in the Diccionarios de Contabilidad). In terms of user needs, the definitions can be divided into two main types: (a) around 50% of these are general definitions, that is, they are valid for assisting all potential users of the dictionary; (b) around 50% are special definitions, in that they have undergone lexicographical treatment with the aim of assisting specific types of potential users: – The first type of special definition occurs in articles describing institutionalised terms, e.g. many IAS/IFRS terms. In this regard, the definition of impracticable in IAS 8 is intended to furnish experts’ expectations; it is a rather long and very detailed definition that offers prolific details (example 9). This definition was

206 | Designing, Making and Updating Specialised Online Dictionaries simplified in the Accounting Dictionaries (example 10), which also uses hyperlinks (source in example 10; cf. section 9.3.5) to the official text: Impracticable Applying a requirement is impracticable when the entity cannot apply it after making every reasonable effort to do so. For a particular prior period, it is impracticable to apply a change in an accounting policy retrospectively or to make a retrospective restatement to correct an error if: - the effects of the retrospective application or retrospective restatement are not determinable; - the retrospective application or retrospective restatement requires assumptions about what management’s intent would have been in that period; or - the retrospective application or retrospective restatement requires significant estimates of amounts and it is impossible to distinguish objectively information about those estimates that: (i) provides evidence of circumstances that existed on the date(s) as at which those amounts are to be recognised, measured or disclosed; and (ii) +would have been available when the financial statements for that prior period were authorised for issue from other information. Example 9: Definition of impracticable in IAS 8. impracticable When something is impracticable it is impossible to do or achieve it, even if all practical things or matters reasonably expected to solve a task or problem have been done or dealt with. Source: IAS 8. Paragraph 5 Example 10: Definition of impracticable in the Accounting Dictionaries and hyperlink to IAS 8.

The second type of special definition consisted of including data that makes the cultural specificity of one particular term very clear and precise. This solution assists potential users with different backgrounds. For instance, the English definition of building society indicates that it is a financial institution located in the UK that works as a bank or savings bank and offers mortgages and other loans for the purchase, construction or improvement of dwellings. The Spanish definition also informs us that a building society is similar to Spanish cooperatives or mutual savings banks, two organisations that are easily identified in Spain and that have a similar function to the building society described in the dictionary article (example 11):

Compilation Phase | 207 building society Building societies in the UK are financial institutions whose objectives include accepting deposits for saving, granting mortgage loans as well as unsecured loans to the depositors for the purchase, construction or improvement of owner-occupied dwellings as well as offering various financial services. Las sociedades de crédito hipotecario o inmobiliario son instituciones financieras que funcionan en el Reino Unido como cooperativa o mutua constructora. Sus objetivos incluyen el depósito de cuentas viviendas, la concesión de préstamos hipotecarios con avales y sin avales utilizados para la compra, construcción, o mejora de viviendas así como la oferta de varios tipos de servicios financieros. Example 11: English and Spanish Definitions of building society in the Diccionarios de Contabilidad.

The third type of special definition was used in combination with other data types so as to disambiguate the meaning of polysemous lemmas, i.e. ones which are preceded by Arabic numbers, each corresponding to one specific sense (cf. section 9.3.3). This has occurred in three situations. Firstly, the crafting of the definition and the polysemous indices informs us that the concept varies due to legal requirements (example 12). Secondly, the crafting of the definition, the polysemous index and a source hyperlink indicates that the concept varies depending on the institutionalised text consulted (example 13). Thirdly, in addition to the polysemous index, the definition uses different superordinates in order to situate the term in a particular conceptual scenario (example 14): CFC 1. UK definition: A CFC is a controlled foreign company in which a UK company has a 10% stake or more. 2. US definition: A CFC is a controlled foreign corporation whose voting stock is more than 50% owned by US stockholders, each of whom owns at least 10% of the voting power.

Example 12: UK and US definition of CFC in the Accounting Dictionaries current price 1. The current price is the latest price quoted on the stock exchange for a security. Source: IAS 39, paragraph AG 72; IFRS 2, paragraph B6 (c)

208 | Designing, Making and Updating Specialised Online Dictionaries 2. The current price is today’s price in an active market for an asset, e.g. properties, commodities or other products or services. Source: IAS 40, paragraphs 45 and 46 Example 13: Definitions of current price and hyperlinks in the Accounting Dictionaries notice 1. A notice is an announcement about something, often written or formal, made to one or more persons. 2. A notice is a warning or notification that a certain event will take place. Example 14: Two definitions of notice in the Accounting Dictionaries

Selecting equivalents for the bilingual dictionaries (English-Danish/Danish-English; English-Spanish/Spanish-English) was also carried out by lexicographers and accounting experts. They go with the definitions, with the result that these dictionaries prove innovative in many ways: they are L1 (lemma) + L1 (definition crafted in the language of the lemma) + L2 (equivalent). In addition, the English-Spanish combination contains a Spanish definition, making it an L1 + L1 + L2 + L2 dictionary (i.e. a lemma, a definition in the language of the lemma, an equivalent and a definition in the language of the equivalent). Most of the candidates for becoming equivalents were selected through existing lexicographical resources, parallel texts and comparable texts. All were validated by experts in accounting or through Google searches, with consideration being given to the caveats mentioned in section 9. 3. 1. Google searches were useful when we selected equivalents that we could not document in the three main sources employed: (i) existing dictionaries and institutionalised texts; (ii) private texts; and (iii) an inhouse accounting corpus. For example, in the English-Spanish combination, there are 54 English lemmas and 49 Spanish equivalents with ratio; an analysis illustrates five specific procedures for selecting the Spanish equivalents: – The equivalent of English acid test ratio is índice de liquidez. Here, as we could not find a Spanish equivalent in the texts we used for documenting them, we performed a Google search: we quoted the English lemma and a Spanish term connected with its meaning, i.e. “acid test ratio” + “liquidez”. This resulted in 4,100 hits and analysis showed that índice de liquidez could be used as an equivalent. Next, we googled “acid test ratio” + “índice de liquidez”, and obtained 306 hits. This examination convinced us that acid test ratio and índice de liquidez referred to the same concept. A similar procedure was used for ratio de capital de máxima calidad (En: core capital ratio), ratio de endeudamiento (En: debt-equity ratio), etc.

Compilation Phase | 209

The equivalent of equity ratio is ratio de solvencia, although there are accounting dictionaries and texts that include different equivalents. For instance, The Diccionario Técnico Económico Financiero Actuarial includes two different equivalents for the same English lemma: índice de capitalización; coeficiente de los fondos propios. Our accounting experts decided that coeficiente de los fondos propios was unsuitable (it is not used in Spanish, as can be seen by a Google search of this term retrieving only 1 hit), and that índice de capitaliación was less informative than ratio de solvencia. Our decision was to use ratio de solvencia as an equivalent and índice de capitalización as a synonym. Similar decisions were taken for índice de beneficio bruto (En: gross margin ratio), ratio préstamo-valor (En: loan-to-value ratio), ratio precio-valor contable (En: market to book ratio), precio/flujo de efectivo por acción (En: P/CF ratio), precio/beneficio (En: P/E ratio), etc. The equivalent of financial ratio is ratio financiero. The Spanish equivalent is a loan translation of the English lemma and is easily documented in all the accounting dictionaries and Spanish accounting text consulted. This solution applies to many equivalents, e.g. cálculo de ratios financieros (En: calculation of financial ratios), ratio de contribución (En: ratio of contribution), ratio de apalancamiento (En: gearing ratio), ratio precio beneficio (En: price earnings ratio), ratio precio/valor en libros (En: price/book ratio), etc. The equivalent of dividend ratio is ratio de dividendos. Both ratio de dividendo (singular dividend) and ratio de dividendos (plural dividend) could be used as an equivalent: both were easily documented; we decided on "plural dividend", on the assumption that plural variants represent characteristics of many accounting texts. Similar procedures, that is, using one language variant instead of another, were also used for ratio de costes (En: cost ratio), análisis de ratios (En: ratio analysis), etc. The equivalent of payout ratio is ratio de distribución de dividendos. We found several Spanish translations of this English term, e.g. porcentaje que representan los dividendos respecto a las utilidades de la compañía), but these are idiosyncratic, unstable, and inadequate for most potential users, such as translators. In these situations, we “created” an illustrative translatable equivalent, i.e. one that could be used in a translation situation and could represent the concept in a way which is easy to understand. Similar procedures are used in ratio de capital de máxima calidad (En: tier 1 capital ratio), razón ventas a capital invertido (En: sales to capital employed ratio), etc.

9.3.3 Restricting Meaning Unlike that of general language words, the meaning of most specialised terms is restricted; in other words, it is specific to a particular process, function, person,

210 | Designing, Making and Updating Specialised Online Dictionaries situation, fact, etc. Including homonymous and polysemous indices, as well as the addition of modifiers and lexicographic notes, restricts the meaning of the lemma. There are three main dictionary data types for restricting meaning: (a) polysemous and homonymous indices; (b) grammatical data; (c) lexicographic notes, e.g. grammatical, contrastive, and usage notes. Polysemy and homonymy must be scarce in specialised fields, as the core nature of specialised language is the description of monosemous entities. This must result in the construction of specialised dictionaries that prefer monosemy to polysemy, and exact meaning to the vague kind. This idea agrees with findings from the Accounting Dictionaries and the Diccionarios de Contabilidad. For instance, some 5% of the lemmas are polysemous (e.g. there are 337 and 357 polysemous English and Spanish terms). An analysis of these polysemous lemmas also reveals two interesting findings. The first is that around 85% of polysemous lemmas have two meanings, typically one very specific and used only in accounting, and one more general that can also be used in general language (a word) and/or in another subject field (a semi-term). For instance, profit has an accounting meaning, namely, the earnings of an enterprise in an accounting period, and a general sense representing a gain obtained when buying and selling, two daily activities all human beings perform (example 15). Similarly, amortisation has an accounting meaning, i.e. the gradual reduction in the cost of an intangible asset, and one in terms of banking, that is, the regular repayment of instalments and interest on a loan (example 16): profit 1. For an accounting period, profit constitutes the earnings of an enterprise, i.e. the amount by which revenue exceeds expenses, resulting in increased total assets. 2. A profit is the amount by which the sales price exceeds the costs of products or services sold or the gain obtained on the sale of an asset. Example 15: Two definitions of profit in the Accounting Dictionaries amortisation 1. Amortisation is the gradual reduction of an intangible asset's cost by systematically allocating the asset's amortisable amount over its definite economic life. 2. Amortisation means the regular repayment of instalments and interest on a loan. Example 16: Two definitions of amortisation in the Accounting Dictionaries

The second finding is that around 15% of polysemous lemmas have three or more senses. Most of them include a restricted meaning, one that is either institutional-

Compilation Phase | 211

ised, e.g. crafted in an IAS/IFRS that is typically used within the context of the IAS/IFRS texts, or used in a particular country. For instance, guarantee as a noun has three meanings in the Accounting Dictionaries. One of them is UK specific and corresponds to US guaranty, which is included as synonym (example 17): guarantee 1. A guarantee is a promise made by a guarantor to repay a loan and accept liability on behalf of a borrower in the event of the latter failing to meet the repayment obligations. A guarantee typically involves providing security in the form of particular assets to the lender. 2. A guarantee is a contract whereby a person or an enterprise, the guarantor, agrees to become liable for the debt of another, the principal debtor, in case the principal debtor defaults on the obligations he owes to the creditor. 3. A guarantee is a promise made by a supplier, manufacturer or seller to give compensation for, or to repair or replace any defective or malfunctioning goods or services delivered subject to defined conditions concerning extent and guarantee period. Example 17: Three definitions of guarantee in the Accounting Dictionaries

Homonymy indices, which are indicated as superscripts in the Accounting Dictionaries and Diccionarios de Contabilidad, are attached to lemmas and equivalents where relevant. There are not a large number of homonyms (for instance, around 2% of English lemmas and equivalents and 0.7% of Spanish lemmas and equivalents are homonyms); such figures are in line with our conviction that terms tend to be monosemous, perhaps because most of them are multi-word terms that are coined with the basic aim of restricting the meaning of more general terms. The imbalance between English and Spanish is due to the presence of “zero derivation”, e.g. cost as noun or verb in English and its absence in Spanish. This also explains the fact that approximately 95% of the homonymous English lemmas consist of one orthographic word. Homonymy indices are typically identified during the process of compilation, usually when definitions are being crafted. English homonyms are typically noun and verb used for conceptualising accounting concepts, such as cost, balance, aggregate, guarantee, etc., most of which are also polysemous (example 18): balance1 noun 1. A balance is the net amount or amount remaining after part has been taken away, i.e. the difference between debits and credits in an account. 2. A balance is an amount owed, i.e. an outstanding debt payable.

212 | Designing, Making and Updating Specialised Online Dictionaries balance2 verb passive To balance means to reconcile the debit and the credit side of an account to determine the balance. Example 18: Homonymous lemmas in the Accounting Dictionaries

A second type of homonymy is typically identified during the process of contextualising meaning (cf. section 9.3.4.). It usually consists of specific accounting terms and general words such as polysemous account (a term in accounting and banking) and the word account “consideration” in different accounting contexts, e.g. “take risks and uncertainties into account when preparing the financial statements”. As this can assist our potential users, such as translators, both lemmas were included in the accounting dictionaries, as shown in example (19); similar processes are also seen with act, authority, etc. account1 noun 1. An account is a record in monetary terms of accounting transactions listing items of a similar type on a debit or credit basis. Accounts are part of an accounting system and usually classified according to a specific category, e.g. cash account, ledger account, nominal account, contra account, deposit account etc. 2. An account is an arrangement with a firm allowing credit and deferring payments to a later date, usually the end of the month, or a statement of money paid or due for goods or services. 3. An account is an expression for a regular client or customer who does a large amount of business and has an account with a particular supplier or other enterprise. account2 noun Account means consideration, e.g. to give consideration to something when you plan. Example 19: Homonymous lemmas in the Accounting Dictionaries

The inclusion of grammatical data restricts the meaning of all terms, especially those that are typically formed by the addition of modifiers to a basic node; this system characterises the process of term formation in specialised languages. Consequently, the process of term formation usually contemplates the aggregation of nouns, verbs, articles, adjectives, prepositions, etc. to a basic node, which typically results in an abundance of multi-word terms (MWT), for example, in languages such as specialised English and Spanish. In this context, around 75% of Spanish accounting terms are multi-word expressions (Table 9.3.; figures have been rounded, as the process of adding lemmas is continuous):

Compilation Phase | 213 Table 9.3: Grammar categories in the Diccionarios de Contabilidad: Number and Percentage Grammar Category Noun Verb Adjective Adverb/ Preposition/Conjunction Abbreviation Multi-word Terms Total

Number 921 254 219 31 150 4,925 6,505

% 14 4 3 0.4 2.3 75 100

Regarding grammatical data, the lexicographers working on a specialised dictionary project such as the Accounting Dictionaries and the Diccionarios de Contabilidad have attached inflections, articles with nouns, a description of countability, and active and passive verb forms to all the lemmas and equivalents present in the project. These dictionaries offer language profiles of every lemma and equivalent with the objective of assisting users, especially in communicative situations where this profile situates the term in a linguistic paradigm, e.g. a countable noun, and makes the meaning precise, favouring monosemy and restricting its polysemous interpretations. Around 99% of the grammar data included was documented with Google searches (cf. section 9.3.1). As was previously indicated, the accounting team created around 1% of equivalents and lemmas, which means that we also created the grammar data attached to these “lemmas and equivalents". The process of attaching grammar data starts with an analysis of each definition. For instance, the definitions of nouns and multi-word expressions are usually crafted in order to indicate the following, where relevant: (a) number; (b) gender (e.g. in Spanish); (c) definiteness, i.e. it can be used with a definite or indefinite article or both. The second phase consists of performing Google searches of the different grammar possibilities, analysing the hits and storing the information retrieved in the accounting database. An analysis of method (106 lemmas) and método (104 lemmas) illustrates the working methodology used: – 82 of the 106 English lemmas (77%) and 76 of the 104 Spanish lemmas (73%) are used only in the singular with a definite article. This signifies that the meanings of these terms are restricted to a specific accounting method, namely, one used in accounting to perform an accounting task: for example, the pooling-ofinterest method (also pooling of interest method) is used in connection with business combinations with the aim of recognising all assets and liabilities in the acquired company at carrying value at the time of combination instead of at fair value and without recording goodwill. This information, which is included in the accounting database, is presented to users as shown in example (20).

214 | Designing, Making and Updating Specialised Online Dictionaries pooling-of-interest method

Example 20: Grammar data attached to the lemma in the Accounting Dictionaries

24 of the 106 English lemmas (23%) and 28 of the 104 Spanish lemmas (27%) are used in singular and plural with both definite and indefinite articles. As a result, the lemma is not necessarily accounting specific but a kind of superordinate term which refers to a general class of methods. For instance, an allocation method (also method of allocation) is a general method for distributing profits, costs, taxes, etc., on accounting elements or separate entities: there are specific allocation methods, that is, methods for distributing profits, costs, etc., in a specific way, such as the incremental cost-allocation method, which ranks users of a cost object in a prioritised order and allocates costs to users in accordance with this ranking. The grammar information on these “superordinate” methods, which is also included in the accounting database, is presented to users as shown in example (21): allocation method

Example 21: Grammar data attached to the lemma in the Accounting Dictionaries

There may be grammar differences among languages. For instance, there are English lemmas that accept only singular definite articles (e.g. the account analysis method), whereas their Spanish equivalents accept singular definite and indefinite articles (e.g. un método de análisis contable and el método de análisis contable) (example 22): account analysis method

método de análisis contable

Example 22: Grammar data attached to the lemma and equivalent in the Diccionarios de Contabilidad

The method described in the above paragraphs was employed for every lemma and equivalent, and the results indicate that attaching grammar to specialised terms is

Compilation Phase | 215

necessary and very useful for all types of users, as shown in the following findings referring to Spanish nominal, that is, lemmas that behave syntactically as nouns: – Around 14% of Spanish lemmas take a plural definite article (los, las), but do not take plural indefinite articles (unos, unas). – Around 2% of Spanish lemmas and equivalents do not take indefinite articles (un/una; unos/unas). – Around 5 % of Spanish lemmas and equivalents take only a singular definite article (el, la). – Around 5% of Spanish lemmas and equivalents take only plural definite and/or indefinite articles (unos/unas, los/las). – Around 2% of the Spanish lemmas and equivalents do not take inflections or articles. – Around 40% of Spanish lemmas and equivalents take feminine articles. – Around 55% of Spanish lemmas and equivalents take masculine articles. – Around 3% of Spanish lemmas take both masculine and feminine articles. Lexicographic notes restrict the meaning of a term in a very specific sense, generally explaining the following: (a) a grammar peculiarity of the lemma and equivalent, where relevant; (b) factual differences between lemma and/or equivalent(s), where relevant; and (c) usage information on lemma and equivalent, where relevant. As the Accounting Dictionaries and the Diccionarios de Contabilidad include grammatical data addressed to lemmas and equivalents, the inclusion of grammar notes in the accounting database is negligible. Instead, it contains between 3% (English lemmas) and 1% (Spanish lemmas) of dictionary articles with lexicographic notes that are prepared by lexicographers and experts together and are worded in the language of the lemma. The imbalance between the two languages is due to the presence of Spanish as well as English definitions of English lemmas in some dictionaries of the Diccionarios de Contabilidad (e.g. the set describing English terms for Spanish users). As shown when comparing examples (23 and 24), the Spanish definition of an English lemma may contain data that situate cultural differences between English lemmas and Spanish equivalents, i.e. balance sheet accounts are included only in Groups, 1, 2, 3, 4, and 5 of the Cuadro de Cuentas, a specific Spanish accounting document described in the Plan General Contable: balance sheet account A balance sheet account records the entries concerning the enterprise’s assets and liabilities with resulting balances and forms the basis of the balance sheet. Example 23: Entry for balance sheet account in the Accounting Dictionaries

216 | Designing, Making and Updating Specialised Online Dictionaries balance sheet account A balance sheet account records the entries concerning the enterprise’s assets and liabilities with resulting balances and forms the basis of the balance sheet. Una cuenta de balance registra las entradas relacionadas con los activos, pasivos y patrimonio neto de una empresa con sus saldos correspondiente y forma la base del Cuadro de Cuentas de los grupos 1, 2, 3, 4 y 5 del Plan General Contable. cuenta de balance Example 24: Entry for balance sheet account in the Diccionarios de Contabilidad

The lexicographic notes, which are worded in the language of the lemma, typically offer contrastive information, e.g. factual differences between culture-bound concepts, and usage information; in many cases they indicate that a particular concept of the term has been recently modified or is used in specific contexts. From the user’s point of view this difference is irrelevant: what is important for them is the restriction of meaning they introduce, as example (25) shows: A share In some countries, A shares may be non-voting ordinary shares. Whether A shares are voting or non-voting shares will appear from a company's articles of association. asset The term asset is also used metaphorically to describe things that cannot actually be owned, e.g. the skills, loyalty and commitment of employees. audit opinion The expression “audit opinion” may also refer to the whole audit report. long-term provisions Long-term provisions are recognised as an item under non-current liabilities in the balance sheet. management report The management report typically appears in tandem with the auditors' report and makes explicit the distribution of responsibilities between the company's management and the independent auditors. Example 25: Lexicographic notes in the Accounting Dictionaries

Compilation Phase | 217

9.3.4 Contextualising Meaning Contextualising meaning implies the inclusion of data types that aim to present the meaning of lemmas and their equivalents in one or more relevant context(s). There are three main types of relevant contexts in specialised lexicography: the discursive, factual and sociolinguistic context. During the process of definition, lexicographers and experts must work together in order to prepare dictionary types of data that contextualise the meaning of lemma. The discursive context is the language context of the term, typically the collocational context in which the term occurs. In the Accounting Dictionaries and the Diccionarios de Contabilidad, the inclusion of some collocations and examples primarily aim to assist in communicative situations by including data types in Danish, English and Spanish (and translations between these languages, where relevant) that present the discursive context of the lemma and its equivalent. The factual context is the inclusion of facts, figures, procedures and culture data that primarily are intended to help users in cognitive situations. In the Accounting Dictionaries and the Diccionarios de Contabilidad, most examples and some collocations contain accounting data that contextualise the factual meaning of the lemma and its equivalent (where relevant). The sociolinguistic context is the language code(s) attached to lemmas, equivalents, synonyms, antonyms, collocations and examples, where relevant. These refer to a particular language or to a language variant. In the Accounting Dictionaries and the Diccionarios de Contabilidad, two types of codes are used: (a) language codes such as “DK”, “UK”, “US”, and “E”, for identifying the languages of the particular dictionaries: Danish (DK), British English (UK), American English (US) or Spanish (E) lemmas; (b) dialect codes such as “IAS/IFRS” and “NIC/NIIF” for referring to international accounting variants of English and Spanish, i.e. language variants which are introduced by accounting committees and organisations such as the Financial Accounting Standard Board (FASB) or the International Accounting Standard Board (IASB), etc. The former type of language code is necessary in dictionary projects such as the Accounting Dictionaries and the Diccionarios de Contabilidad, where the search system offers different solutions in different languages. For instance, users retrieve Danish and English lemmas when initiating a search of “balance” (Illustration 39):

218 | Designing, Making and Updating Specialised Online Dictionaries

Illustration 39: Screenshot from the Accounting Dictionaries

The second type of language code is also useful: it openly indicates that the element concerned, e.g. the definition of a lemma, is an important type of institutionalised concept, one crafted by a normalising accounting board whose decisions result in the creation of harmonised concepts; the latter refer to the same accounting reality everywhere in the world where the International Accounting Standards and the International Financial Reporting Standards are in operation. We use lexicographical concepts of collocation and example as defined in the Function Theory of Lexicography (cf. chapter 5; also Bergenholtz, 2009; FuertesOlivera et al. 2013). A collocation is an umbrella term for referring to combinations of signs, typically words, which contextualise the meaning of lemmas and equivalents. They are composed of two or more orthographic signs, do not constitute a full sentence, but offer potential users the possibility of obtaining relevant information on discursive and encyclopaedic contexts. Examples perform the same function but are different from collocations, in that they are crafted as full sentences, usually unedited. Collocations and examples are extracted from accounting texts (almost all of which can also be documented by googling) under one or all of the following criteria: (a) the variant included is more frequent than the other variant; e.g. a plural form instead of a singular one; (b) they offer a language profile of term and equiva-

Compilation Phase | 219

lent (e.g. they are found only in the plural, or they attach definite articles but not indefinite ones; (c) they offer factual information if googled; (d) they instance real accounting language in use, especially relevant for disambiguating polysemous lemmas and equivalents; (e) they are also lemmas and/or equivalents in the Accounting Dictionaries and the Diccionarios de Contabilidad. The following collocations (example 26) and examples (example 27) illustrate the process of selecting the collocations and examples used: - a depreciable asset: un active amortizable Rationale: depreciable asset and activo amortizable are English and Spanish lemmas and equivalents; they also offer factual knowledge: Spanish accountants do not differentiate between “depreciation” and “amortisation”, therefore the Spanish equivalent is “amortizable” and not “depreciable”). - assets adjusted upward: activos ajustados al alza Rationale: they show the language profile of term and equivalent; there are many instances of upward and downward movements in accounting; for instance, in the Diccionarios de Contabilidad, around 20% of the lemma contain collocations that conceptualise movements, e.g. alcanzar su objetivo de ratio de coste, conseguir un objetivo de coste, igualan los niveles de tipos de interés actuales, etc.) - assets leased under finance leases: activos en arrendamiento financiero Rationale: these are instances of real accounting language; differences among English (financial leases is plural) and Spanish (arrendamiento financiero is singular). Spanish accountants do not use plural arrendamientos financieros with activos en) - assets, total: activos, total Rationale: these are instance of real accounting language. Important where the comma and the two language elements are; different from total assets, which is a different term; “assets, total” is an expression in a balance sheet, i.e. in a typical accounting text type). - 100. Capital social Rationale: if googled, users access Internet sites that define the concept, explain how the concept works in accounting, related accounts, offers examples of entries, and its equivalent concept and indicate its synonymous term (where relevant) in the Plan General Contable of 1990 (the previous official accounting code), as shown in the following Internet site when googling "100. Capital social" (Illustration 40)

220 | Designing, Making and Updating Specialised Online Dictionaries

Illustration 40: Internet site when googling the collocation 100. Capital social - entradas contables Rationale: the plural is more relevant than singular entrada contable. Example 26: List of collocations in the Diccionarios de Contabilidad The ongoing performance audit was carried out in conjunction with the compliance audit. = La auditoría de resultados en marcha se llevó a cabo conjuntamente con la auditoría reglamentaria. Rationale: language profile of term and equivalent: the passive voice, especially the socalled Spanish se pasiva, is more relevant in accounting texts. Any material uncertainties related to events or circumstances that may affect the enterprise’s ability to continue as a going concern should be disclosed. = Deben publicarse cualesquiera incertidumbres materiales relacionadas con sucesos o circunstancias que pudieran afectar a la capacidad de la empresa para continuar su actividad. Rationale: they are instances of real accounting language.

Compilation Phase | 221 The standards meet the criteria of understandability, relevance, reliability and comparability required of the financial information needed for making economic decisions and assessing the stewardship of management. = Las normas cumplen los criterios de comprensibilidad, relevancia, fiabilidad y comparabilidad exigidos a la información financiera necesaria para llevar a cabo la toma de decisiones económicas y la valoración de la actuación de la dirección. Rationale: they offer factual information. Example 27: List of examples in the Diccionarios de Contabilidad

The accounting database contains around 30,000 collocations and 2,000 examples in each language, many of which are translated into the other languages covered in the dictionary by the joint work of lexicographers, experts in accounting and experienced translators. This makes these dictionaries very much appreciated and fulfils one of their most important genuine purposes: to assist users in a translation situation (cf. section 9.4). Language codes such as “IAS/IFRS” and “NIC/NIIF” illustrate a fundamental characteristic of certain specialised languages: there are specific creations of concepts and terms. This gives rise to four different situations, all of which are considered lexicographically in the Accounting Dictionaries and the Diccionarios de Contabilidad, as is revealed in the following analysis. The first situation occurs when the created concept varies in the different IAS/IFRS terms published so far, as shown in example (14, above). The second situation relates to the created concept differing from already existing ones. For instance, Spanish cedente (English equivalent: cedant) is defined in the NIIF 4, which applies to insurance contracts (including both insurance and reinsurance contracts) but does not address other aspects of accounting by insurers, such as accounting for financial assets. This cedente has an institutionalised meaning, i.e. one crafted in the NIIF 4: a cedente is “the policyholder in a reinsurance contract”. In addition, Spanish cedente has another accounting meaning, namely, “any person who transfers assets to another person”. This situation resulted in a dictionary article with two senses in Spanish and one meaning in English. The third situation occurs when the created concept introduces a new term that takes over from accounting terms which have traditionally been employed. Finally, competing IAS/IFRS terms exist. Situations three and four are very relevant in culture-bound subject fields in which a dominant language, typically English, influences the coining of terms to describe the newly created concepts (Fuertes-Olivera & Nielsen 2011 and 2014). For example, Spanish deterioro (En: impairment), and deterioro del valor de los activos (En: impairment of assets) refer to concepts analysed in IAS 36 (Impairment of Assets). This text studies accounting situations that have been studied in accounting texts since this subject field was created in ancient Egypt. Consequently, languages such as Spanish had been using their own terms for describing the impairment of

222 | Designing, Making and Updating Specialised Online Dictionaries assets. When the IAS 36 text was translated into Spanish, translators did not use the Spanish terms for these concepts, which were (and are) provisión por insolvencias, depreciación, and provision por depreciación, despite the fact that these are frequently used in Spanish accounting texts; for instance, a Google search of provisión por insolvencia retrieved around 80,000 hits. We dealt with this situation in the following lexicographic way: deterioro and deterioro de valor de los activos are the equivalents of impairment and impairment of assets, and provisión por insolvencias, depreciación, and provisión por depreciación are synonyms. In addition, the dictionary articles of these Spanish terms included extra data, and the definitions indicate that they are the traditional Spanish accounting terms; we recommend the use of deterioro and deterioro del valor de los activos, as these two terms are also the ones used in the official Spanish accounting code: The Plan General Contable (example 28): provisión por insolvencias Una provisión por insolvencias es el término tradicional español para "deterioro", es decir, para referirse a una disminución del valor de un activo registrado en el balance debido a un daño sufrido o a una reducción en el precio de mercado del mismo. Esta reducción se presenta como un valor a recuperar inferior al valor contable del activo. Not recommended, use instead deterioro Example 28: Provisión por inslovencia taken over by deterioro in Spanish accounting.

Competing IAS/IFRS terms are the result of translations of IAS/IFRS texts. Each new version of one such text produces several modifications of concepts and terms, or even the introduction of nonsensical terms. We have adopted three lexicographical decisions for dealing with this situation. The first is to differentiate between two related codes: (a) “IAS/IFRS + E” and (b) “IAS/IFRS”. The former code indicates that the concept and term are extracted from the institutionalised texts that are currently recommended on the official Spanish accounting Internet site: the ICAC or “Instituto de Contabilidad y Auditoria de Cuentas”. The second code indicates that the term and concept have been used in previous versions of the IAS/IFRS texts, usually ones that were translated in South America, and some of which may be still in use in one or more South American countries. In the Diccionarios de Contabilidad, there are 404 “IAS/IFRS + E” terms and 241 “IAS/IFRS” terms. For instance, actividad interrumpida, operación abandonada, and operación discontinua refer to the same concept: actividad interrumpida is an “IAS/IFRS + E” term, whereas operación abandonada and operación discontinua are “IAS/IFRS” terms. The second decision refers to the introduction of recommendations, i.e. proscription (Bergenholtz, 2003; cf. sections 5.6 and 9.3.5). Generally speaking, we have recommended the “IAS/IFRS + E” terms for two reasons. Firstly, they are more eas-

Compilation Phase | 223

ily cross-referred (e.g. access to the ICAC Internet site is free), and secondly their translations seem better than previous ones. The third means of dealing with this situation is to introduce lexicographic notes that explain a particular point relevant to these codes. These notes typically indicate that the “IAS/IFRS” term is not precise, unnatural or even nonsensical (example 29): coste por jubilación Spanish accountants prefer ‘coste por jubilación’ or ‘coste por retiro’ to the “IAS/IFRS” term ‘coste por prestaciones definidas post-empleo’. This term is unnatural Spanish. costes de producción sin variación de efectivo Spanish accountants use ‘costes de producción sin variación de efectivo’ instead of the “IAS/IFRS” term ‘coste imputado’. This term is not precise and can lead to misunderstandings. exposición al riesgo en moneda extranjera The “IAS/IFRS” term ‘exposición a moneda extranjera’ is nonsensical in Spanish. Example 29: Use of lexicographic notes for contextualising the meaning of “IAS/IFRS” terms in the Diccionarios de Contabilidad

9.3.5 Offering More Choices Offering more choices concerns the inclusion of dictionary data types which are intended to increase users’ options in some or all of the user situations envisaged during the pre-compilation phase, e.g. synonyms, antonyms, internal and external cross-references, and proscription notes in the Accounting Dictionaries and the Diccionarios de Contabilidad. In online dictionaries, these data types are hyperlinks, occasionally preceded by symbols or expressions. For instance, the expressions Synonyms, Antonyms, See also, Sources, Do not use, use instead are included in the Accounting Dictionaries and Diccionarios de Contabilidad for introducing synonyms, antonyms, internal and external cross-references, and proscription, respectively (Illustration 41).

224 | Designing, Making and Updating Specialised Online Dictionaries

Illustration 41 a: Screenshot of the Diccionarios de Contabilidad

Illustration 41 b: Screenshot of the Diccionarios de Contabilidad

Synonyms and antonyms are incorporated mainly to provide users with different stylistic options when producing and translating specialised texts. Cross-reference directs users to additional information and is, therefore, chiefly employed in cognitive situations. Synonyms, i.e. terms that can usually replace the lemma and/or equivalent where relevant, should not proliferate in specialised dictionaries, as terminological variation must be avoided in specialised language. This situation does not occur in the Accounting Dictionaries nor in the Diccionarios de Contabilidad: many lemmas and equivalents have synonyms (around 35% of the English lemmas and 55% of the Spanish ones). The above numbers can be explained lexicographically. Firstly, the dictionary articles include only one equivalent per sense. This increases the options of including synonyms. Secondly, we believe that proscription leads to the inclusion of more synonyms: there are often synonymous variants of the recommended term. Finally,

Compilation Phase | 225

lexicographers and experts in the field must work together, which favours the identification of synonyms. By way of example, the following list contains terms that are synonyms only and exclusively in accounting: non-current and long-term; current and short-term; cuenta de reservas por ganancias acumuladas and beneficio pendiente de asignación; gasto asignable and coste asignable (cost and expense are different concepts in accounting), patrimonio neto and fondos propios, ajustes por cambio de valor y subvenciones, etc. It would have been difficult for non-experts to envisage, say, estado de situación patrimonial as synonymous with balance. In addition, the Diccionarios de Contabilidad contain 20% more dictionary articles with synonyms than English and Danish dictionary combinations. This imbalance is mainly due to three factors. The first is that many institutionalised Spanish accounting terms have synonymous variants, as explained in the inclusion of both Spanish “IAS/IFRS + E” and “IAS/IFRS” terms (cf. section 9.3.4). The second factor is a lexicographical decision: several English terms are also treated as Spanish synonyms, as we consider this to be necessary when the term is not very well systematised (cf. section 5.6.). For instance, stress test is a synonym of prueba de resistencia; this, we believe, favours meaning extraction and is in line with users’ expectations. Hence, a Google search of "el stress test" + "banca" or + "banco" retrieved around 30,000 hits each, and similarly a Google search of "los stress test" + "banco" retrieved more than 130.000. The third factor is the influence of the linguistic characteristics of English and Spanish on term formation. Many Spanish accounting terms are English translations and these show two main variations: – Many variations are the result of either using or not using function words in the Spanish term. For example, the synonym of análisis de estados financieros is análisis de los estados financieros; in other words, there is a definite article (los) before the nominal expression estados financieros. – Many variations are the result of maintaining Spanish traditional accounting terms and new translated terms. For instance, póliza de seguro is the synonym of contrato de seguro, i.e. the Spanish traditional accounting term (póliza de seguro) and the Spanish translation of English insurance contract (contrato de seguro). Antonyms, or terms that refer to opposite concepts to those of lemmas and/or equivalents where relevant, are not very common in specialised dictionaries nor in the Accounting Dictionaries and the Diccionarios de Contabilidad. For instance, around only 7% of English and Spanish dictionary articles include antonyms. We believe this is due to operational factors: experts do not usually explain concepts by referring to opposites, but rather to similar ones. Hence, it is easier to identify synonyms rather than antonyms when working with specialised texts. In terms of users’ needs, the synonyms and antonyms contained in these dictionaries, which are always documented in the texts used and always accepted by

226 | Designing, Making and Updating Specialised Online Dictionaries lexicographers and accounting experts, aim to help users in communicative situations, typically when writing or translating. In addition, they can be used to assist users in disambiguating polysemous terms. For example, the presence of synonyms and/or antonyms in approximately 2% and 4% polysemous English and Spanish lemmas respectively, can help users in cognitive situations, especially when meaning differences are very subtle, as these three prototypical situations show: – There are two specific accounting concepts. In the case, let us say, of ingreso diferido, this has two senses: one is very specific and has no synonym or antonym (ingreso diferido as deferred income or deferred revenue: “this is an item in the balance sheet which is recognised under the accruals concept”); yet the same term can be also a more general accounting concept and, therefore, may have synonyms and antonyms, e.g. the synonym cobro anticipado and the antonym gasto diferido help users to disambiguate ingreso diferido as deferred revenue, that is, “cash payments from customers prior to the enterprise’s provision of the good and services to which the payments relate”. – There is one specific accounting concept and one general sense. For instance, incremento is an accounting concept for referring to resources acquired by an enterprise within a specific accounting period. This concept has synonyms and antonyms: adiciones and sumas (synonyms) and restas (antonym). Incremento is also a general word for referring to an extension or expansion of something, e.g. a market share. As a general word, it can also have synonyms, such as the English expansion and the Spanish aumento and antonyms, for example, the Spanish reducción. – There is one specific accounting concept and one specialised concept, i.e. one that is not accounting specific. In the case of cargar and charge, these refer to the act of recognising an amount in the profit and loss account. As accounting specific, it has synonyms (adeudar and debitar; En: debit) and antonyms (abonar, En: debit). Cargar also refers to the act of imposing a levy (En: tax) and has synonyms (imponer, gravar) in the Diccionarios de Contabilidad. Cross-references are typically abundant in specialised online dictionaries, usually with the purpose of assisting users in cognitive situations. Approximately 35% of the articles in the Accounting Dictionaries and the Diccionarios de Contabilidad contain cross-references, most of which can be used in a similar way to encyclopaedic notes and sections. This means that these, namely, hyperlinks in online dictionaries, are basically included in specialised dictionaries not only to inform users about related concepts but also to offer options for increasing the level of specificity and details of the search term, as shown in the external cross-reference to IAS 8 of example (10). In the Accounting Dictionaries and the Diccionarios de Contabilidad, the inclusion of so many internal and external cross-references as well as the use of certain Internet technologies has eased the process of converting the original polyfunctional dictionaries into Model T Ford dictionaries, or ones in which the articles

Compilation Phase | 227

and the visualised lexicographical data are adapted to the various functions displayed by the dictionary (cf. section 9.4.). For instance, clicking on partes vinculadas in the hyperlink See also (illustration 41) also retrieves eight dictionary articles in which partes vinculadas is present: deudas a corto plazo con partes vinculadas; deudas a largo plazo con partes vinculadas; inversions financieras a corto plazo en partes vinculadas; inversiones financieras a largo plazo en partes vinculadas; operaciones con partes vinculadas; participaciones en partes vinculadas; rela-ción entre partes vinculadas; transacción con partes vinculadas (Illustration 42):

Illustration 42 a: Internal and external cross-references in the Diccionarios Contabilidad (excerpts) I.

228 | Designing, Making and Updating Specialised Online Dictionaries

Illustration 42 b: Internal and external cross-references in the Diccionarios Contabilidad (excerpts) II.

An analysis of the retrieved definitions of the eight dictionary entries can help users to understand the gist of the concept, thereby representing a means of assistance for interested laypersons, students, and translators in a cognitive situation. In addition, working with external sources, such as clicking on NIC 24.9, can be of aid to users for acquiring a more accurate understanding of the concept, in other words, helping experts in a cognitive situation. Proscriptive notes inform the user not only about language use but also about the form recommended by lexicographers and experts working together. In the Accounting Dictionaries and the Diccionarios de Contabilidad, around 5% of the articles contain proscriptive notes. An analysis shows the principal characteristics: – There are proscriptive notes for recommending IAS/IFRS variants. For instance, we recommend precio vendedor in preference to precio actual demandado por el vendedor. This recommendation is based on an examination of current IAS/IFRS texts: we decided to recommend the variant downloaded from the official internet homepage of ICAC (The Spanish official accounting body). Similarly, we recommend activos comunes de la entidad as against activos comunes de la

Compilation Phase | 229

compañía (entidad (En: entity) is the term currently in use in accounting institutionalised texts). There are proscriptive notes for recommending orthographic and grammar variants. For instance, we recommend diminishing-balance method as opposed to diminishing balance method. The recommended term is more frequent in the accounting texts used for documenting lexicographical handling of accounting concepts and terms. Similarly, we recommend actividad de ventas instead of actividad de venta. There are proscriptive notes for recommending very specific terms. For example, we recommend compound financial instrument as against compound instrument. The recommended term is more descriptive and offers users a more precise description of its conceptual characteristics, e.g. the instrument is “financial”. There are proscriptive notes for recommending US terms instead of UK terms. For example, we recommend realization value and not realisation value. The rationale is that both IAS/IFRS bodies tend to prefer American variants and that there are many more American accounting texts published than British ones (most multinationals being American). There are proscriptive notes for recommending certain variants as opposed to others. For instance, we recommend weighted average cost formula rather than average cost formula. The recommended term is not only more descriptive but also more frequent in Google searches. There are proscriptive notes for recommending particular ways of wording the same concept. For instance, we recommend income earning and not earning of income. The rationale here is their frequency in Google: “income earning” retrieves around 800,000 hits, whereas “earning of income” retrieves around 100,000. Similarly, we recommend acciones ordinarias con potencial dilutivo in preference to acciones ordinarias con potencial de dilución. There are proscriptive notes for recommending specific accounting terms rather than more general ones. For instance, we recommend construction contract as opposed to contract. The recommended term is accounting specific, e.g. construction contracts are analysed in institutionalised accounting texts such as “IFRIC 15” and “IAS 11”. Similarly, we have recommended lease payment in preference to rent, etc. There are proscriptive notes for recommending well-established accounting terms and not ad hoc translations. For instance, we recommend acciones consideradas como pasivos financieros rather than acciones no incluidas en patrimonio neto. Similarly, we recommend autocartera as against acciones propias en cartera.

To sum up, the compilation phase comprises key lexicographical activities; these are basically concerned with the selection of lexicographical data and how this is

230 | Designing, Making and Updating Specialised Online Dictionaries handled so as to satisfy users’ needs in potential user situations. Our previous description of these activities has shown that each dictionary is fundamentally a specific project, i.e. it contains general features (basically deriving from the core nature of lexicography) and specific features which demand specific skills and knowledge. For instance, subject fields such as accounting mix specific cultural traditions with general ones. Hence, selecting and handling accounting data must always balance the working of the two opposing forces. Certain subject fields, let us say, chemistry, are not constrained by cultural forces. In other words, selecting and dealing with lexicographical data requires that we should take into account general concepts (i.e. they are valid for all reference works) and specific concepts. These are based on the nature and characteristics of the subject field in question. For instance, when preparing a dictionary on law or accounting, the easiest way to start the process of selecting and treating lexicographical data is by analysing institutionalised texts, e.g. statutes, standards, acts, etc. Once the process has started, the team will easily discover the way ahead, for example, using the Internet as a corpus, working with other reference works, consulting appendices and glossaries in Manuals and Handbooks, constructing their own in-house corpus, etc. What the team must not do is follow erroneous methodologies, such as starting the selection and treatment of the data by using that taken from Linguistics-conceived corpora, as most items of relevant data cannot be extracted from these. This may explain why there are many dictionary projects that take a very long time and are never, in fact, operational.

9.4 The Post-Compilation Phase: A Process of Continuous Updating During the post-compilation phase, lexicographers must put the dictionary at the users’ disposal, observe how it works, and check whether or not those consulting it are satisfied. For specialised lexicography, these tasks are inseparable from the subjecting of the dictionary to a process of regular and continuous updating. Constant modification of language and facts characterise the ontological nature of subject fields, requiring the use of theoretical assumptions, methodologies and technologies to facilitate the process of constant updating, which must be emblematic of the design and compilation of specialised online dictionaries. Below, we illustrate how the post-compilation phase operates in the Accounting Dictionaries and Diccionarios de Contabilidad. The Danish, English and Spanish dictionaries were initially published as printed and online polyfunctional dictionaries (Fuertes-Olivera 2010; Nielsen 2007; Nielsen & Mourier 2007; Nielsen et al. 2009; Nielsen & Fuertes-Olivera 2013). Consequently, the original dictionaries were equipped with static articles together with static data, as Illustration (43) shows:

The Post-Compilation Phase: A Process of Continuous Updating | 231

232 | Designing, Making and Updating Specialised Online Dictionaries

Illustration 43: Screenshot from El Diccionario Inglés-Español de Contabilidad (Nielsen et al. 2009).

Illustration (43) shows which types of data users retrieved when searching for accounting in El Diccionario Inglés-Español de Contabilidad (Nielsen et al. 2009), published in 2009. This search retrieved the following: – Several definitions and equivalents of accounting, i.e. the homonymous and the polysemous meanings of accounting1; both the English definitions and the Spanish equivalents (contabilidad and contable) make the meaning very precise. Arabic numbers and superscripts are used for indicating polysemy and homonymy, respectively. – Grammar data attached to each homonymous meaning of accounting, e.g. . – Language code attached to accounting1 and synonym accountancy, e.g. IAS/IFRS/US and UK.

The Post-Compilation Phase: A Process of Continuous Updating | 233

– – –

Several English collocations and examples together with their Spanish translations. English and Spanish synonyms, e.g. contabilización and accountancy. Cross-references, e.g. financial accounting.

By 2010, the original polyfunctional dictionaries were changed into a set of Model T Ford online dictionaries, each using the available technology with a view not only to providing quicker access, but also to creating dynamic articles with dynamic data, i.e. articles that vary from one consultation to the next in terms of their lexicographical content. This variation resulted from our conviction that the future of online dictionaries resided in providing articles and lexicographical data adapted to the various functions displayed by the dictionary. This change not only demonstrated the theoretical validity and practical suitability of the Function Theory of Lexicography for designing and producing genuine online dictionaries; it also showed that proponents of this theoretical construction can easily put into practice new conceptions; for instance, the database and the dictionary are different concepts (Bergenholtz 2011 and 2012; Nielsen & Almind 2011), and use existing technologies with the final aim of offering users real solutions at affordable time and money costs. This resulted in the conception of 27 dictionaries (Bergenholtz 2012; Nielsen & Fuertes-Olivera 2013), most of which are already on the market. The new dictionaries not only use more data, e.g. Spanish and English definitions of the same English accounting term in the English-Spanish combinations, but they also use the type of data, and no other type, needed in prototypical user situations. Within this new conception, dictionary data such as that shown above (Illustration 43) is no longer retrieved; instead, users retrieve data adapted to a reception situation (Illustration 44), a translation situation (Illustration 45), a knowledge situation (Illustration 46), and a specific translation situation (Illustration 47) (cf. section 5.5.). Illustration (44) shows what we believe Spanish users need in a reception situation: definitions of the English lemma; one Spanish equivalent per English and Spanish definition of the lemma; a part of speech of the English lemma; the language code "IAS/IFRS + US”, and polysemous and homonymous indices. Illustration (44) also shows an important difference between these dictionaries and other specialised dictionaries. We discovered that the proficiency level of Danish and Spanish users is different, the Danish being much more proficient in English than the Spanish. This resulted in the addition of a Spanish definition of the English lemma:

234 | Designing, Making and Updating Specialised Online Dictionaries

Illustration 44: Accounting in El Diccionario Inglés-Español de Contabilidad: Recepción. (Diccionarios de Contabilidad)

Illustration (45) shows what, in our view, users need in a translation situation: the types of data given in illustration (44), plus the following: inflections of Spanish equivalents; grammar restrictions of Spanish equivalents; synonyms of Spanish

The Post-Compilation Phase: A Process of Continuous Updating | 235

equivalents; English collocations and examples alongside their Spanish translations:

Illustration 45: Accounting in El Diccionario Inglés-Español de Contabilidad: Traducción. (Diccionario de Contabilidad)

Illustration (46) shows what we consider users require in a cognitive situation: the data types commented on in illustrations (44) and (45), plus the following: inflection of English lemmas; grammar restrictions of English lemmas; English synonyms; internal cross-references:

236 | Designing, Making and Updating Specialised Online Dictionaries

Illustration 46: Accounting in El Diccionario Inglés-Español de Contabilidad: Conocimiento. (Diccionarios de Contabilidad)

llustration (47) shows what we believe some very experienced translators need: a list of collocations and examples with the term being searched:

The Post-Compilation Phase: A Process of Continuous Updating | 237

Illustration 47: Accounting in El Diccionario Inglés-Español de Contabilidad: Frases y Expresiones. (Diccionarios de Contabilidad)

Clicking on each of the "frases y expresiones" of Illustration (47) retrieves the dictionary article(s) in which the collocation or example in question may be encountered: this is highlighted, favouring consultation and saving time. For example, clicking on the collocation "the transaction-based accounting model" retrieves two dictionary articles, one for accounting model and another for transaction-based: users, then, are provided with definitions and Spanish equivalents, inflections and grammar restrictions of Spanish equivalents, English collocations (and examples, where relevant) and their Spanish translations. In Illustration (48) there are two different translations of the same English expression. This makes this dictionary very much appreciated by users, especially professional translators (cf. section 5.5.). These two different translations show the importance of cultural data in certain culture-bound subject fields: one of the translations is "el modelo contable del pre-

238 | Designing, Making and Updating Specialised Online Dictionaries cio de adquisición"; this refers to cultural information, which may be accomplished due to the joint work of experts, translators and lexicographers:

Illustration 48 a: The transaction-based accounting model in El Diccionario Inglés-Español de Contabilidad: Frases y Expresiones. (Diccionarios de Contabilidad) I.

Summary and Conclusion | 239

Illustration 48 b: The transaction-based accounting model in El Diccionario Inglés-Español de Contabilidad: Frases y Expresiones. (Diccionarios de Contabilidad) II.

9.5 Summary and Conclusion This chapter has put into practice a working methodology for compiling specialised online dictionaries. It has shown how the Function Theory of Lexicography allows lexicographers to design and produce dictionaries that favour customisation, do not need a large number of resources for compiling and updating, and can cope with the characteristics of specialised subject fields. Briefly, we are now able to draw three important conclusions in this field of research. The first is that the process of compiling specialised dictionaries is always a continuous one; hence, the Accounting Dictionaries and the Diccionarios de Contabilidad are, on average, updated every three months. We are working towards updating them every week. This brings us to the second main conclusion. The use of Internet technologies requires the acknowledgement that lexicographical databases and dictionaries are different concepts (Bergenholtz 2011; Nielsen & Almind 2011). In other words, the future of online dictionaries does not reside in "thinking quantitatively" but on "thinking qualitatively". This means that the construction of Model T Ford and Rolls Royces needs an advanced theory capable of guiding lexicographers and other specialists, who must do whatever is necessary to help users and not bombard them

240 | Designing, Making and Updating Specialised Online Dictionaries with much more data than is necessary. Users do not need "fancy gadgets", which consume money and time; instead, they need simple solutions that save time and resources. The third conclusion is that the Function Theory of Lexicography also makes provision for dealing with the specific characteristics of the dictionary being planned. Culture-bound subject fields such as accounting are different from other others. Consequently, lexicographers and experts working on accounting dictionaries should always take into consideration the three main characteristics of accounting: – Some of the concepts are institutionalised, i.e. they are defined and framed in restricted legislative contexts. For instance, the so-called IAS/IFRS terms, i.e. terms and concepts that are dealt with in the International Accounting Standards and International Financial Reporting Standards, describe the same reality in all languages. Hence, they should be handled with data taken from the same text types. The best place, then, to search for institutionalised concepts and terms is in institutionalised texts, and the best way to find them is by extensive reading of these texts. – Several of the concepts are popularised on a daily basis, i.e. they are discussed in newspapers in view of the fact that they affect our lives. By way of example here, Fuertes-Olivera & Nielsen (2011, 2014) have mentioned that when the global financial crisis was in its infancy, experts developed new terms such as sub-prime debt to explain what was happening; this term did not alarm nonexperts. When, however, people gradually began to realise what in fact was occurring and what ‘sub-prime debt’ actually meant, a new term, toxic assets, was coined and frequently used; everyone understood why this concept, whose definition is identically worded for sub-prime debt and toxic assets, was so important. – Certain concepts are culture specific and need specific lexicographical handling. In addition to usage and lexical notes, the Diccionarios de Contabilidad also use collocations (Fuertes-Olivera et al. 2013), and definitions for explaining cultural differences.

10 Conclusions This book defends the argument that lexicography in general and specialised lexicography in particular is both a millenarian cultural practice and an independent academic discipline with its own system of scientific theories. Being independent does not mean that it is placed in a walled room away from the rest of scientific disciplines in which human knowledge has been partitioned; instead, (specialised) lexicography is related to many other disciplines, above all information science, linguistics, terminology, and specific subject fields. For instance, a relationship between lexicographers and subject field experts is completely necessary in specialised lexicography, as we believe – and have made our point quite clearly in this book – that high-quality specialised dictionaries cannot be compiled without a sound knowledge of the facts, that is, concepts that permeate a particular domain. Understanding the facts is, therefore, a necessary condition for constructing specialised dictionaries; this signifies that describing the language profile of a particular term must be dependent on the meaning of the term in the subject field in question. To put it in a more straightforward way, if someone with a deep knowledge of a domain is not actively working on a specialised dictionary project, the chances of designing, compiling, and updating high-quality specialised dictionaries are greatly reduced. This simple fact may explain a surprising finding: we have come across many different dictionary projects that are well-funded, e.g. certain projects funded by EU research agencies, but which have produced poor results or barely none at all. As lexicography is an independent academic discipline, this also implies that there must be a theory and methodology for designing, constructing and updating high quality specialised dictionaries. This book has made our case for defending the theoretical tenets and practical applications of the Function Theory of Lexicography, i.e. the academic construction originally initiated in the Centre for Lexicography at the Aarhus School of Business (Bergenholtz & Tarp 2002, 2003 and 2004; also Tarp 2008); this theory, as this book shows, has since been evolving continuously. This has prompted us to follow two main directions. The first is our defence of the concept of theory applied to lexicography. Our stance is based on accredited traditions. Lexicography fulfils all the requirements necessary for it to be treated and classified as a separate science or area of academic study: – Lexicography constitutes a system of knowledge growing out of social practice. – Lexicography has its own subject field. – Lexicography is rooted in the form of concepts, categories, theories and hypotheses. – Lexicography comprises both the history of dictionaries and its own history, including pre-theoretical ideas.

242 | Conclusions – –

Lexicography contains independent contributions to methodology. Lexicography includes instructions for practical action.

The above claims have allowed us to observe and study lexicography in all its dimensions, to isolate relevant phenomena with certain properties, to establish the relations existing between them, to express opinions concerning such phenomena and relations, and to systematize these opinions. This has resulted in the formulation of several principles that form the core of lexicography defended in this book: The core of lexicography is the design of utility tools that can be quickly and easily consulted with a view to meeting precise information needs occurring for specific types of users in specific types of extra-lexicographical situations.

The second direction stems from the above concept on the core of lexicography. A theory exists because it can guide practice. As we believe that the future of lexicography is very much dependent on the coming of age of the Internet, we have focused on translating the above principles into the construction of high-quality specialised online dictionaries. In other words, we have attempted to demonstrate that the theoretical tenets of the function theory of lexicography are solid, because they can be used to guide the design, construction, and updating of such dictionaries. This decision has had two main implications. One of them is concerned with analysing existing specialised online dictionaries, with the aim of discovering whether or not they satisfy the tenets of the functional approach to lexicography supported in this book. To perform this analysis, we refined Tarp’s (2011) previous classification of online dictionaries and built a template of 10 criteria, which was applied consistently to a list of 16 specialised online dictionaries that were considered examples of the different orientations and practical applications currently operating in this field of work. The main conclusion obtained is that there is still some way to go in the design, compilation, and updating of high-quality specialised online dictionaries. For instance, it was surprising to ascertain that the conception of many online dictionaries follows ideas used for constructing paper dictionaries, thereby not using Internet technologies for solving some of the problems associated with the Internet, such as the so-called Google effect that occurs when users retrieve many more data than needed; added to this is the absence of dynamic articles with dynamic data, i.e. that adapted to a user’s situation and needs. We also found surprising the concept of user needs espoused in many quarters, especially in terminology circles. Designers of terminological information tools typically use the concept of user needs as a catch-all phrase, thereby failing to target specific users and continuing to incorporate types of data that could have been used for documentation but not for meeting other requirements, e.g. ones related to translation, reception, production, etc. For instance, terminological information tools generally include data on conceptual relations among terms and concepts.

Conclusions | 243

Even supposing that such data could be discovered consistently and handled in a straightforward manner, its inclusion is, in our view, unimportant, or occasionally even rather hazardous, for most potential users of specialised information tools. The second implication is that we have developed and explained a methodology for putting into practice our idea of lexicography, namely, one for making, designing and updating high-quality specialised dictionaries. To this end, we have devoted several sections (c.f. section 5.5) and a full chapter (c.f. chapter 9); these explain the decisions taken during the different phases into which we have divided the process of designing, compiling, and updating high-quality online dictionaries. This practice has allowed us to identify the kind and type of knowledge needed in order to produce these dictionaries. We require the following: a) lexicographers, and in particular one or more with knowledge of the main principles and developments in the subject field in question; b) subject field experts; c) experts in information science, especially those who have a knowledge of the design of Internet databases that can be fed from far-away sources and who can offer technical possibilities for providing search systems which allow users to retrieve dynamic articles with dynamic data; d) experts in marketing and selling via Internet. All these experts are working on several lexicographical projects that are currently in operation in the Centre for Lexicography (University of Aarhus) and the International Centre for Lexicography (University of Valladolid). For the purpose of this book, we have explained decisions concerned with the Accounting Dictionaries and the Diccionarios de Contabilidad. These are a set of Danish, English and Spanish dictionaries that are oriented at three Danish and Spanish user groups: 1) translators and language staff; 2) accounting experts and semi-experts; and 3) students and laypersons interested in Danish, English and Spanish accounting matters. These users can look up lexicographical data when they are reading, writing and translating accounting texts (i.e. in communicative situations) and/or gaining some knowledge of accounting (i.e. in a cognitive situation). Our description of these lexicographical works illustrates that the proponents of the function theory work with real data, although not necessarily with the kind taken from corpora in the sense Corpus Linguistics attributes to this concept. We believe that the lexicographical data to be included in a specialised dictionary must be relevant data, no matter whether this is frequent or not. Relevant data is that needed to meet a user’s needs in the situation envisaged, and such data is always the result of a deductive process that experts in lexicography and a particular subject field perform so as to ascertain the type of information which must be prepared with the above aim in mind. Working with the Accounting Dictionaries and the Diccionarios de Contabilidad has likewise allowed us to apply another important tenet of the function theory of lexicography: dictionaries are not static but dynamic works, which means that they have general characteristics usually derived from the core nature of lexicography, and specific characteristics related to a specific and particular project.

244 | Conclusions To sum up, this book has offered a proposal for avoiding the schism in which lexicography has been mired for the last two hundred years. It has defined specialised lexicography as the branch of lexicography concerned with the theory and practice of specialised dictionaries, i.e. dictionaries, encyclopaedias, lexica, glossaries, vocabularies, terminological knowledge bases, and other information tools covering areas outside general cultural knowledge and the corresponding Language for General Purposes (LGP); mainly, but not exclusively, the reference is to disciplines related to trade, industry, technology, natural and social sciences, and humanities. In addition, this book has emphasised that specialised lexicography must deal with terms and concepts on an equal footing, and hence has opposed artificial divisions by approaching the theory and practice of designing, compiling and updating specialised dictionaries. These are set in a continuum that contains a series of lexicographical works with different characteristics, sizes, and names. All of them are specialised dictionaries, or tools that are designed for consultation purposes in a situation where users may need information on the language and facts in a particular domain. We believe that this book is of special significance at this particular time, when globalisation and internationalisation are dominant forces. High-quality information tools are a necessity in such an environment, where the Internet is the dominant factor. We have argued in favour of using Google searches to complete the data types included in specialised dictionaries. This is a safe procedure provided that some cautionary measures are taken into account. The most important measure is that any data which is to be included in the dictionary should be deemed suitable by the experts working on the dictionary project. In the final analysis, designing highquality specialised online dictionaries is a human activity, and this is what this book is about: it supports a theory and explains a practice for designing, compiling, and updating man-made, high-quality online dictionaries.

References Lexicographical works Accounting dictionaries = Sandro Nielsen, Lisa Mourier and Henning Bergenholtz 2012: Accounting dictionaries. (A series of 13 interconnected Danish, Danish-English, English-Danish and Danish dictionaries). Odense: American farmer’s encyclopedia and dictionary of rural affairs = Cuthbert W. Johnson and Governeur G. Emerson 1844: The American farmer’s encyclopedia and dictionary of rural affairs. Third Edition. Philadelphia: Carey and Hart. Base lexicale du français = Serge Verlinde et al. 2010: Base lexicale du français (BLF). Katholieke Universiteit Leuven. [] Bilteknisk Ordbog = Ebbe Skjerk 2007: Bilteknisk Ordbog Engelsk-Dansk. Jyllinge: Forlaget Veterania. [Online version:] Business-English Spanish Glossary = A. D. Mills 2012: A-D. Miles’ Business English to Spanish Glossary.[] Cambridge Business English Dictionary = Cambridge Dictionaries Online 2013: []. CercaTerm = Termcat 2013: CercaTerm.]. Concise Oxford Dictionary of Linguistics = Peter H. Matthews 2005: Concise Oxford Dictionary of Linguistics. Oxford: Oxford University Press. Cyclopædia = Ephraïm Cambers (Ed.) 1728: Cyclopædia; or, an Universal dictionary of arts and sciences. London: James and John Knapton. Dansk Glossarium = Jens Høier Leth 1800: Dansk Glossarium. En ordbog til Forklaring over det Danske Sprogs gamle, nye og fremme Ord og Talemaader for unge Mennesker og for Ustuderede. Et Forsøg. Med en Fortale af Professor Rasmus Nyerup. Kiøbenhavn: S. Poulsen. Dantern = DANTERMcentret 2013: Copenhague Business School. [] Den danske netordbog = Henning Bergenholtz and Heidi Agerbo Pedersen in cooperation with Filip Bodilsen, Kathrine Brosbøl Eriksen, Helene R. Gudmannn, Stine Busk Hedegaard, Helene H. Jensen, Christian Kjølhede, Emilie Dittmer Laursen, Jane Nguyen, Ricard Almind and Rasmus Theodor Styrk 2011: Den danske netordbog. 6th edition. Odense: [] Diccionario español de contabilidad = Pedro A. Fuertes-Olivera/Henning Bergenholtz/Sandro Nielsen, Pablo Gordo Gómez/ Marta Niño Amo/Ángel de los Ríos Rodicio/ Ángeles Sastre Ruano/ Sven Tarp/ Marisol Velasco Sacristán 2013: Diccionario español de contabilidad. (A series of interconnected Spanish accounting dictionaries). Hamburg: [] Diccionario de la Lengua Española = Diccionario de la Lengua Española 2005: Madrid: Real Academia Española. [Online version:] Diccionario Enciclopédico de Ingeniería Genética = Uwe Kaufmann and Henning Bergenholtz (Eds.) 1998: Diccionario enciclopédico de ingeniería genética español-inglés. Toronto: Lugus. Diccionario inglés-español de contabilidad = Sandro Nielsen, Lise Mourier, Henning Bergenholtz y Pedro Fuertes Olivera, Pablo Gordo Gómez, Marta Niño Amo, Angel de los Rios Rodicio, Angeles Sastre Ruano, Sven Tarp, Marisol Velasco Sacristán 2009: El Diccionario InglésEspañol de Contabilidad. Aarhus: Centre for Lexicography. Diccionario de lexicografía práctica = José Martínez de Sousa 1995: Diccionario de lexicografía práctica. Barcelona: Biblograf.

246 | References Diccionarios de contabilidad = Pedro A. Fuertes-Olivera, Henning Bergenholtz, Sandro Nielsen, Pablo Gordo Gómez, Lise Mourier, Marta Niño Amo, Ángel de los Ríos Rodicio, Ángeles Sastre Ruano, Sven Tarp and Marisol Velasco Sacristán 2012: Diccionarios de contabilidad. (A series of interconnected English-Spanish and Spanish-English dictionaries). Hamburg: [] Dictionary of Business and Management = Jonathan Law 2012: A Dictionary of Business and Management. 5th edition. Oxford: Oxford University Press. [] Diccionario Técnico Económico Financiero Actuarial = Julio García Villalón and Josefina Martínez Barbeito 2008: Diccionario Técnico Económico Financiero Actuarial Inglés-Español. Valladolid: own edition. Dictionary of the English Language = Samuel Johnson 1755: Dictionary of the English Language. London: J. & P. Knapton. Dictionary of Lexicography = Reinhard Hartmann and Gregory James 2001: Dictionary of Lexicography. London, New York: Routledge. Dictionary of Linguistics and Phonetics = David Crystal 1991: A Dictionary of Linguistics and Phonetics. Cambridge: Basil Blackwell. Dictionnaire analytique d’économie politique = Charles Ganilh 1826: Dictionnaire analytique d’économie politique. Paris, Brussels: Ladvocat. Dictionnaire universel de commerce = Jacques Savary des Bruslons 1723: Dictionnaire universel de commerce. Paris: Chez J. Etienne. Disco = The European Dictionary of Skills and Competences 2012: The European Dictionary of Skills and Competences [] Dizionario della economia politica e del commercio = Gerolamo Boccard 1857: Dizionario della economia politica e del commercio. Torino: Sebastiano Franco e figli e comp. EcoLexicon = Pamela Faber et al. 2013: Terminological Knowledge Base on the Environment. Granada: University of Granda. [] Encyclopedic Dictionary of Gene Technology = Uwe Kaufmann and Henning Bergenholtz (Eds.) 1998: Encyclopedic Dictionary of Gene Technology. English-Spanish. Toronto: Lugus. Encyclopédie = Denis Diderot and Jean le Rond D’Alembert (Eds.) 1751-1780: Encyclopédie, ou Dictionnaire raisonné des sciences, des arts et des métiers. Tome 1-35. Paris: Briasson. Encyclopædia Britannica = William Smellie (Ed.) 1768-1771: Encyclopædia Britannica; or a Dictionary of arts and science. Edinburgh: A. Bell and C. Macfarquhar. Fachwörterbuch zur Rentenversicherung = Idilkó Fata 2005: Ungarisch-Deutsches, DeutschUngarisches Fachwörterbuch zur Rentenversicherung. Szeged: Grimm Kiadó. FF POIROT = Gang Zhao et al. 2005: [] FrameNet = International Computer Science Institute, Berkley. [] Genoma = Mª Teresa Cabré et al. 2003: Banc de Coneixement sobre el Genoma Humano. Barcelona: IULA (Universitat Pompeu Fabra) []. Glossary of Archival and Records Terminology = Richard Pearce-Moses 2005: A Glossary of Archival and Records Terminology. Chicago: Society of American Archivists. [Online version:] Glossary of FAO Databases and Information Systems = Food and Agriculture Organization of the United Nations 2012: Glossary of FAO Databases and Information Systems []. Glossary of Mortgage and Home Equity Terms = Wells Fargo 2012: Glossary of Mortgage and Home Equity Terms. Wells Fargo.[]

Lexicographical works | 247 Glossary of Photographic Terms = Kodak 2010: A Glossary of Photographic Terms. Kodak. [] IATE = Inter-Active Terminology for Europe 2013: Inter-Active Terminology for Europe. Brussels: European Union [] Illustrated Glossary for Transport Statistics = Illustrated Glossary for Transport Statistics. 4th Edition 2009: Luxembourg, Paris, Geneva: EUROSTAT, ITF, UNECE. [] Illustrated Glossary on Stone Deterioration Patterns = V. Vergès-Belmin 2008: Illustrated glossary on stone deterioration patterns. Paris: ICOMOS International Scientific Committee for Stone. [ _Sites_15_ISCS_Glossary_Stone.pdf] International Encyclopedia of Lexicography = Franz Josef Hausmann, Oskar Reichmann, Herbert Ernst Wiegand and Ladislav Zgusta (Eds.) 1989-1991: Wörterbücher, Dictionaries, Dictionnaires. An International Encyclopedia of Lexicography. Volumes I, II and III. Berlin, New York: Walter de Gruyter. International Encyclopaedia of Social Sciences = William A. Darity (Ed.) 2008: International encyclopaedia of social sciences. Detroit: Macmillan Reference USA. Kemisk Ordbog = Sune Damhus, Søren Møller and Alexander Senning 2009: Kemisk Ordbog. Tredje Udgave. Copenhagen: Nyt Teknisk Forlag. [Online version:] Kicktionary = Schmidt, Thomas 2009: Kictionary. The Multilingual Electronic Dictionary of Football League. [] Lexicography: A Dictionary of Basic Terminology = Igor Burkhanov 1998: Lexicography: A Dictionary of Basic Terminology. Rzeszów: Wydawnictwo Wyzszej Zskoły Pedagogicznej. Lexicon Technicum = John Harris 1704: Lexicon Technicum: or, an Universal English Dictionary of Arts and Sciences: Explaining Not Only the Terms of Art, but the Arts Themselves. London: D. Brown, J. Walthoe, J. Knapton, B. and S. Tooke, D. Midwinter, B. Cowse, T. Ward, E. Symon, E. Valentine, and J. Clark. Lexin = Sven-Göran Malmgren (Ed): Lexin. 2012. Stockholm: Språkrådet. [Online version:] Linguistisches Wörterbuch = Theodor Lewandowski 1984: Linguistisches Wörterbuch. Heidelberg: Quelle & Meyer. MacMillan English Dictionary for Advanced Learners = Michael Rundell (Ed.) MacMillan English Dictionary for Advanced Learners. 2nd edition 2007: London: MacMillan. McGraw-Hill Dictionary of Scientific and Technical Terms = McGraw-Hill Dictionary of Scientific and Technical Terms 1974: New York: McGraw-Hill Book Company. Multilingual Glossary = Heymans Institute of Pharmacology and Mercator School, Department of Applied Linguistics 2010: Multilingual Glossary of Technical and Popular Medical Terms in Nine European languages. []. Musikordbogen = Inger Bergenholtz in cooperation with Henning Bergenholtz and Richard Almind. 2011: The Musikordbogen (The Danish Musik Dictionary). Odense: Ordbogen. [] New Palgrave Dictionary of Economics = Steven N. Durlauf and Lawrence E. Blume (Eds.) 2012: The New Palgrave Dictionary of Economics. Basingstoke, New York: Palgrave Macmillan. [] Nordisk leksikografisk ordbok = Henning Bergenholtz, Ilse Cantell, Ruth Vatvedt Fjeld, Dag Gundersen, Jón Hilmar Jónsson and Bo Svensén 1997: Nordisk leksikografisk ordbok. Oslo: Universitetsforlaget. Oxford English Dictionary = John Simpson (Eds.) 2012: Oxford English Dictionary. Oxford: Oxford University Press. [Online version:]

248 | References Philosophisches Wörterbuch = Manfred Buhr and Georg Klaus (Eds.) 1971: Philosophisches Wörterbuch. Band 1 & 2. Berlin: Das europäische Buch. PoCeHRMOM = PoCeHRMOM 2007: []. Popular Encyclopedia = A. Whitelaw (Ed.) 1846: The popular encyclopedia: or, ‘Conversations lexicon’: being a general dictionary of arts, sciences, literature, biography, history, ethics, and political economy. Volume 1-7. Glasgow: Blackie & Sons. PROLIX = Prolix 2010: []. Retskrivningsordbogen = Dansk Sprognævn 2012: Retskrivningsordbogen. Copenhagen: Dansk Sprognævn. [Online version:] South African Trilingual Wine Industry Dictionary = South African Trilingual Wine Industry Dictionary. Pearl: Winetech and SAWIS 2012. [] TermFinder = Peters, Pam in collaboration with Claudia Oliveira, Alan Jones, Jan Tent, Adam Smith, Yasmin Funk, Theresa Winchester-Seeto, Peter Petocz, Jenny Middledorp, Kehui Luo, Alan Kilgore, Fred Wang, and David Raftos 2011: Term Finder. Dictionary Research Centre/Centre for Language Sciences. Macquarie University. [] TERMIUM PLUS® = Termium Plus® 2013: [] TermSciences = Portail Terminologique Multidiciplinaire 2006: [] Universal Dictionary of Trade and Commerce = Malachy Postlethwayt 1774: The Universal Dictionary of Trade and Commerce. The Fourth Edition. London: W. Strahan, J. and F. Rivington, J. Hinton. Unterm = United Nations Multilingual Terminology Database, 2013: United Nations Multilingual Terminology Database. New York: United Nations. [] Wikipedia = Wikipedia. [] Wiktionary = Wiktionary. The free dictionary 2012: Wikimedia. [] Wordnet = Wordnet. A lexical database for English. [] Yongle Dadian = Jin Xie (Ed.) 1408: Yongle Dadian. Volume 1–11,095. China. Other Literature D’ Alembert 1754 = Jean le Rond D’Alembert 1754: Dictionnaire. In Denis Diderot and Jean le Rond D’Alembert (Eds.): Encyclopédie, ou Dictionnaire raisonné des sciences, des arts et des métiers. Tome IV. Paris: Briasson, 958-969. Al-Ajmi 2002 = Hashan Al-Ajmi 2002: Which microstructural features of bilingual dictionaries affect users’ look-up performance. International Journal of Lexicography 15(2), 119-131. Almind 2005 = Richard Almind 2005: Designing internet dictionaries. Hermes. Journal of Linguistics 34, 37-54. Almind/Bergenholtz/Vrang 2006 = Richard Almind, Henning Bergenholtz and Vibeke Vrang 2006: Theoretical and computational solutions for phraseological lexicography. Linguistik online 27(2). [ Auger/Barrière 2008 = Alain Auger and Caroline Barrière 2008: Pattern-based approaches to semantic relation extraction: A state-of-the-art. Special Issue on Pattern-Based Approaches to Semantic Relation Extraction. Terminology 14(1), 1–19 Atkins/Rundell 2008 = B.T. Sue Atkins and Michael Rundell 2008: The Oxford Guide to Practical Lexicography. Oxford: Oxford University Press. Bejoint 2010 = Henri Bejoint 2010: The lexicography of English. Oxford: Oxford University Press. Bergenholtz 1994 = Henning Bergenholtz 1994: Fachsprache und Gemeinsprache. Lemmaselektion in Fachwörterbuch. In Henning Bergenholtz and Burkhard Schaeder (Eds.): Fachlexikographie. Fachwissen und seine Repräsentation in Wörterbüchern. Tübingen: Günter Narr Verlag, 285304.

Other Literature | 249 Bergenholtz 1995 = Henning Bergenholtz 1995: Wodurch unterscheidet sich Fachlexikographie von Terminographie? Lexicographica 11, 37-46. Bergenholtz 1998 = Henning Bergenholtz 1998: Das Schlaue Buch. Vermittlung von Informationen für textbezogene und textunabhängige Fragestellungen. In Arne Zettersten, Viggo Hjørnager Pedersen and Jens Erik Mogensen (Eds.): Symposium on Lexicography VIII. Proceedings of the Eighth International Symposium on Lexicography May 2-5, 1996 at the University of Copenhagen. Tübingen: Niemeyer, 93-110. Bergenholtz 2001 = Henning Bergenholtz 2001: Proskription, oder: So kann man dem Wörterbuchbenutzer bei Textproduktionsschwierigkeiten am ehesten helfen. In A. Lehr et al. (Eds.): Sprache im Alltag. Berlin, New York: De Gruyter, 499-520. Bergenholtz 2003 = Henning Bergenholtz 2003: User-oriented understanding of descriptive, proscriptive and prescriptive lexicography. Lexikos 13, 65-80. Bergenholtz 2009 = Henning Bergenholtz 2009: Klassifikationen in der Linguistik und in der Lexikographie: Wortarten und Wortverbindungen. Zeitschrift für Anglistik und Amerikanistik. A Quartery of Language, Literature and Culture 57(3): 275-288. Issue edited by Henning Bergenholtz, Katrin Götz-Votteler and Thomas Herbst. Bergenholtz 2011 = Bergenholtz 2011 [2013]: Henning Bergenholtz: Access to and presentation of needs-adapted data in monofunctional internet dictionaries. In Pedro A. Fuertes-Olivera and Henning Bergenholtz (Eds.): e-Lexicography: The Internet, Digital Initiatives and Lexicography. London & New York: Continuum, 2011, 30-53. Also in Pedro A. Fuertes-Olivera and Henning Bergenholtz (Eds.): e-Lexicography: The Internet, Digital Initiatives and Lexicography. London/New Delhi/New York/Sydney: Bloomsbury Academic, 2013, 30-53. Bergenholtz 2012 = Henning Bergenholtz 2012: Concepts for monofuctional accounting dictionaries. Terminology 18 (2), 243-263. Bergenholtz/Bergenholtz 2011 [2013] = Henning Bergenholtz and Inger Bergenholtz 2011: A dictionary is a tool, a good dictionary is a monofunctional tool. In Pedro A. Fuertes-Olivera and Henning Bergenholtz (Eds.): e-Lexicography: The Internet, Digital Initiatives and Lexicography. London & New York: Continuum, 2011, 187-207. Also in Pedro A. Fuertes-Olivera and Henning Bergenholtz (Eds.): e-Lexicography: The Internet, Digital Initiatives and Lexicography. London/New Delhi/New York/Sydney: Bloomsbury Academic, 2013, 187-207. Bergenholtz/Bergenholtz 2013 = Henning Bergenholtz and Inger Bergenholtz 2013: One database, four monofunctional dictionaries. Hermes, Journal of Language and Communication in Business 50, 187-207. Bergenholtz/Bergenholtz/Tarp 2008 = Carsten Bergenholtz, Henning Bergenholtz and Sven Tarp 2008: Leksikografi i videnskabsteoretisk perspektiv: sand, falsk eller irrelevant. LexicoNordica 15, 155-168. Bergenholtz/Bothma 2011 = Henning Bergenholtz and Theo J.D. Bothma 2011: Needs-adapted data presentation in e-information tools. Lexikos 21, 53-77. Bergenholtz/Bothma 2013 = Henning Bergenholtz and Theo J.D. Bothma 2013: Information needs changing over time: a critical discussion. South African Journal of Library and Information Science 29, 22-34. Bergenholtz/Gouws 2006 = Henning Bergenholtz and Rufus H. Gouws 2006: How to do policy with dictionaries. Lexikos 16, 13-45. Bergenholtz/Johnsen 2005 = Henning Bergenholtz and Mia Johnsen 2005, Log files as a tool for improving internet dictionaries. Hermes. Journal of Linguistics 34, 117-141. Bergenholtz/Johnsen 2007 = Henning Bergenholtz and Mia Johnsen 2007: Log files can and should be prepared for a functionalistic approach. Lexikos 17, 1-20.

250 | References Bergenholtz/Kaufmann 1997 = Henning Bergenholtz and Uwe Kaufmann 1997: Terminography and lexicography. A critical survey of dictionaries from a single specialised field. Hermes, Journal of Linguistics 18, 91-125. Bergenholtz/Mugdan 1989 = Henning Bergenholtz and Joachim Mugdan 1989: Korpusproblematik in der Computerlinguistik: Konstruktionsprinzipien und Repräsentativität. In Istvàn S. Bàtori, Winfried Lenders/Wolfgang Putschke (Eds.): Computional Linguistics. An International Handbook on Computer Oriented Language Research and Applications. Berlin, New York: de Gruyter, 141-149. Bergenholtz/Nielsen 2006 = Henning Bergenholtz and Sandro Nielsen 2006: Subject–field components as integrated parts of LSP dictionaries. Terminology 12:(2), 281-303. Bergenholtz/Norddahl 2012 = Henning Bergenholtz and Bjarni Norddahl 2012: Ordbogsartikler, som ingen læser. LexicoNordica 19, 206-222. Bergenholtz/Pedersen 1994 = Henning Bergenholtz and Jette Pedersen 1994: Zusammensetzung von Textkorpora für die Fachlexikographie. In Burkhard Schaeder/Henning Bergenholtz (Eds.): Fachlexikographie. Fachwissen und seine Repräsentation in Wörterbüchern. Tübingen: Narr, 161-176. Bergenholtz/Schaeder 1979 = Henning Bergenholtz and Burkhard Schaeder (Eds.) 1979: Empirische Textwissenschaft. Aufbau und Auswertung von Text-Corpora. Königstein: Scriptor. Bergenholtz/Tarp 1994a = Henning Bergenholtz and Sven Tarp (Eds.) 1994a: Manual i fagleksikografi. Udarbejdelse af fagordbøger – problemer og løsningsforslag. Herning: Systime. Bergenholtz/Tarp 1994b = Henning Bergenholtz and Sven Tarp 1994b: Mehrworttermini und Kollokationen in Fachwörterbüchern. In H. Bergenholtz and B. Schaeder (Eds.): Fachlexikographie. Fachwissen und seine Repräsentation in Wörterbüchern. Tübingen: Gunter Narr Verlag, 385419. Bergenholtz/Tarp 1995 = Henning Bergenholtz and Sven Tarp (Eds.) 1995: Manual of Specialised Lexicography. The Preparation of Specialised Dictionaries. Amsterdam: John Benjamins. Bergenholtz/Tarp 2002 = Henning Bergenholtz and Sven Tarp 2002: Die moderne lexikographische Funktionslehre. Diskussionsbeitrag zu neuen und alten Paradigmen, die Wörterbücher als Gebrauchsgegenstände verstehen. Lexicographica 18, 253-263. Bergenholtz/Tarp 2003 = Henning Bergenholtz and Sven Tarp 2003: Two opposing theories: On H.E. Wiegand’s recent discovery of lexicographic functions. Hermes, Journal of Linguistics 31, 171196. Bergenholtz/Tarp 2004 = Henning Bergenholtz and Sven Tarp 2004: The concept of ‘dictionary usage’. In Cay Dollerup (Ed.): Worlds of Words. A tribute to Arne Zettersten. Special Issue, Nordic Journal of English Studies 3:1, 23-36. Bergenholtz/Tarp 2007 = Henning Bergenholtz and Sven Tarp 2007: Politik und Sprachpolitik in der Lexikographie. In H. Gottlieb and J.E. Mogensen (Eds.): Dictionary Visions, Research and Practice. Amsterdam: John Benjamins, 217-240. Bergenholtz/Tarp 2010 = Henning Bergenholtz and Sven Tarp 2010: LSP lexicography or terminography? The lexicographer’s point of view. In Pedro A. Fuertes-Olivera (Ed.): Specialised Dictionaries for Learners. Berlin, New York: Walter de Gruyter, 27-37. Bergenholtz/Tarp/Kaufmann 1994 = Henning Bergenholtz, Sven Tarp and Uwe Kaufmann 1994: Vore mænd i Havanna: Udarbejdelse af konception til en spansk-engelsk genteknologisk ordbog. Hermes, Journal of Linguistics 13, 291-304. Bergenholtz et al. 1997 = Henning Bergenholtz, Ilse Cantell, Ruth Vatvedt Fjeld, Dag Gundersen, Jón Hilmar Jónsson and Bo Svensén 1997: Nordisk leksikografisk ordbok. Oslo: Universitetsforlaget.

Other Literature | 251 Berners-Lee/Hendler/Lassila 2001 = Tim Berners-Lee, James Hendler and Ora Lassila 2001: The Semantic Web. Scientific American. Available at: [] Besomi 2011 = Daniele Besomi (Ed.) 2011: Crises and Cycles in Economic Dictionaries and Encyclopaedias. London, New York: Routledge. Bevilacqua/Bocorny Finatto 2006 = Cleci Regina Bevilacqua and Maria José Bocorny Finatto 2006: Lexicografia e Terminografia: alguns contrapontos fundamentais. Alfa. Revista de Lingüística 50:2, 43-54. Boas 2009 = Hans C. Boas (Ed.) 2009: Multilingual Framenets in Computational Lexicography: Methods and Applications. Mouton de Gruyter. Berlin. Bogaards 2010 = Paul Bogaards 2010: Lexicography: science without theory? In Gilles-Maurice de Schryver (Ed.): A Way with Words: Recent Advances in Lexical Theory and Analysis. A Festschrift for Patrick Hanks. Kampala: Menha Publishers, 313-322. Borlund 2003 = Pia Borlund 2003: The Concept of relevance in IR. Journal of the American Society for Information Science and Technology 54:10, 913-925. Boulanger 1995 = Jean-Claude Boulanger 1995: Présentation: images et parcours de la socioterminologie. Meta 40(2), 194-205. Bothma 2011 [2013] = Theo J.D. Bothma 2011: Filtering and adapting data and information in an online environment in response to user needs. In Pedro A. Fuertes-Olivera and Henning Bergenholtz (Eds.) 2011: e-Lexicography: The Internet, Digital Initiatives and Lexicography. London & New York: Continuum, 2011, 71-102. Also in Pedro A. Fuertes-Olivera and Henning Bergenholtz (Eds.) 2013: e-Lexicography: The Internet, Digital Initiatives and Lexicography. London/New Delhi/New York/Sydney: Bloomsbury Academic, 71-102. Bothma/Prinsloo 2013 = Theo J.D. Bothma and D.J. Prinsloo 2013: Automated dictionary consultation for text reception: a critical evaluation of lexicographic guidance in linked Kindle edictionaries. Lexicographica 29, 165-198. Bothma/Tarp 2012 = Theo J.D. Bothma and Sven Tarp 2012: Lexicography and the relevance criterion. Lexikos 22, 86-108. Bowker 2008 = Lynne Bowker 2008: Terminology. In Mona Baker and G. Saldanha (Eds.): Encyclopedia of Translation Studies (2nd edition). London, New York: Routledge, 286-290. Bowker/Pearson 2002 = Lynne Bowker and Jennifer Pearson 2002: Working with Specialized Language. A Practical Guide to Using corpora. London, New York: Routledge. Bowker/Marshman 2009 = Lynne Bowker and Elizabeth Marshman 2009: Better integration for better preparation. Bringing terminology and technology more fully into translator training using the CERTT approach. Terminology 15(1), 60-87. Boyatt/Joy 2010 = Russel C. Boyatt and Mike S Joy 2010: Developing an ontology of Mathematical Logic. Scientific Journal of the Technical University of Civil Engineering Bucharest 6-11. Available at: [] Brooks 1989 = Terrence A. Brooks 1989: The model of science and scientific models in librarianship. Library Trends 38(2), 237-249. Buckland 1991 = Michael Buckland 1991: Information as thing. Journal of the American Society for Information Science and Technology 42:5, 351-360. Buckland 1997 = Michael Buckland 1997: What is a “document”? Journal of the American Society for Information Science and Technology 48:9, 804-809. Buckland 2012 = Michael Buckland 2012: What kind of science can information science be? Journal of the American Society for Information Science and Technology 63:1, 1-7. Budin 2001 = Gerhard Budin 2001: A critical evaluation of the state-of-the-art of terminology theory. ITTF Journal 12(1-2), 7-23.

252 | References Burkhanov 1998 = Igor Burkhanov 1998: Lexicography: A Dictionary of Basic Terminology. Rzeszów: Wydawnictwo Wyzszej Zskoły Pedagogicznej. Cabré 1999 = María Teresa Cabré 1999: Terminology. Theory, methods and applications. Amsterdam/Philadelphia: John Benjamins. Cabré 2003 = María Teresa Cabré 2003: Theories of terminology. Their description, prescription and explanation. Terminology 9:2, 163-199. Cabré et al. 2004 = María Teresa Cabré, Carme Bach, Judit Feliú, Rosa Estopá, Gemma Martínez and Jorge Vivaldi 2004: The GENOMA-KB project: towards the integration of concepts, terms, textual corpora and entities. LREC 2004 Fourth International Conference on Language Resources and Evaluation. Lisbon: European Languages Resources Association, 87-90. Available at:: [] Caruso 2011 = Valeria Caruso 2011: Online specialized dictionaries: a critical survey. In Iztok Kosem and Karmen Kosem (Eds.): Electronic Lexicography in the 21st Century: New Applications for New Users. Ljubljana: Institute for Applied Slovene Studies, 66-75. Caruso/De Meo 2013 = Valeria Caruso and Anna De Meo 2013: Comunicare saperi sul web: il caso dei dizionari specialistici. In Christina Bosisio and Stafania Cavagnoli (Eds.): Comunicare le discipline attraverso le lingue: prospettive traduttiva, didattica, socioculturale. Atti del 12° Congresso dell’Associazione Italiana di Linguistica Applicata, Macerata 23-24 febbraio 2012. Perugia: Guerra, 63-83. Chan/Taylor 2011 = Alice Yin Wa Chan and Andrew Taylor 2001: Evaluating learner dictionaries: What the reviewers say. International Journal of Lexicography 14(3), 163-180. Clark et al. 2012 = Malcom Clark, Yunhyong Kim, Udo Kruschwitz, Dawei Song, Dyaa Albakour, Stephen Dignum, Ulises Cerviño Baresi, María Fasli and Anne De Roeck 2012: Automatically structuring domain knowledge from text: An overview of current research. Information Processing and Management 48, 552-568. Cosijn/Bothma 2005 = Erica Cosijn and Theo J.D. Bothma 2005: Contexts of Relevance for Information Retrieval System Design. In Fabio Crestani and Ian Ruthven (Eds.): Context: Nature, Impact, and Role. 5th International Conference on Conceptions of Library and Information Sciences, CoLIS 2005. Glasgow, UK, June 4-8, 2005. Berlin, Heidelberg: Springer, 47-58. Cosijn/Ingwersen 2000 = Erica Cosijn and Peter Ingwersen 2000: Dimension of relevance. Information Processing and Management 36: 533-550. Cowie 2009 = A.P. Cowie (Ed.) 2009: The Oxford History of English Lexicography. Volume II: Specialized Dictionaries. Oxford: Oxford University Press. De Baer/Kerremans/Temmerman 2006 = Peter De Baer, Koen Kerremans and Rita Temmerman 2006: Bridging communication gaps between legal experts in multilingual Europe: Discussion of a tool for exploring terminological and legal knowledge resources. In E. Corino, C. Marello and C. Onesti (Eds.): Proceedings of the XII Euralex International Congress, 813-818. De Baer/Meersman/Temmerman 2009 = Peter De Baer, Robert Meersman and Rita Temmerman 2009: Termontography and DOGMA for knowledge engineering within PROLIX. In R. Meerman, P. Herrero and T. Dillon (Eds.): On the Move to Meaningful Internet Systems: OMT 2009 Workshops. Berlin: Springer, 534-543. De Schryver 2003 = Gilles-Maurice De Schryver 2003: Lexicographer’s dreams in the electronicdictionary age. International Journal of Lexicography 16:2, 143-199. De Schryver 2012 = Gilles-Maurice De Schryver 2012: Trends in twenty-five years of academic lexicography. International Journal of Lexicography 25:4, 464-506. Diderot 1755 = Denis Diderot 1755: Encyclopédie. In Denis Diderot and Jean le Rond D’Alembert (Eds.): Encyclopédie, ou Dictionnaire raisonné des sciences, des arts et des métiers. Tome V. Paris: Briasson, 635-648.

Other Literature | 253 Dubois 1979 = Claude Dubois 1979: La spécificité de la définition en terminologie (par comparaison avec la définition en lexicographie). In Actes du sexième colloque international de terminologie. Québec: Office de la Langue Française, 45-59. Duda et al. 1986 = W. Duda, M. Frenzel, E. Wöller and T. Zimmermann 1986: Zu einer Theorie der zweisprachigen Lexikographie. Überlegungen zu einen neuen russisch-deutschen Wörterbuch. Berlin: Akademieverlag. Duvå/Laursen/Maidahl 1992 = Grete Duvå, Anna-Lise Laursen and Lisbet Maidahl 1992: Brugerundersøgelse vedrørende oversættelse af fagtekst. In Ruth Vatvedt Fjeld (Ed.): Nordiske studier i leksikografi. Rapport fra Konferanse om leksikografi i Norden 28.-31. mai 1991. Oslo: Nordisk forening for leksikografi, 105-133. Faber 2009 = Pamela Faber 2009: The cognitive shift in terminology and specialized translation. In A. Vidal and J. Franco (Eds.): A (Self-) Critical Perspective of Translation Theories Una visión (auto)crítica de los estudios de traducción. (= Monografías de Traducción e Interpretación 1). Universidad de Alicante, Universitat Jaime I and Universitat de Valencia, 107-134. Faber 2012 = Pamela Faber (Ed.) 2012: A Cognitive Linguistics View of Terminology and Specialized Language. Berlin, Boston: De Gruyter Mouton. Faber/León Arauz/Pérez Hernández 2010 = Pamela Faber, Pilar León Araúz and Chantal Pérez Hernández 2010: A Wine Dictionary for Non-Drinkers: LSP Dictionary Functions. In Miguel Ibáñez Rodríguez (Ed.): Vino, lengua y traducción. Volume 2. Valladolid: Universidad de Valladolid, Secretariado de Publicaciones e Intercambio Científico, 109-120. Faber/Márquez Linares/Vega Expósito 2005 = Pamela Faber, Carlos Márquez Linares and Miguel Vega Expósito. 2005: Framing terminology: A process-oriented approach. Meta 50 (4) (CDROM). Faber et al. 2006 = Pamela Faber, Silvia Montero Martínez, Maria Rosa Castro Prieto, José Senso Ruiz, Juan Antonio Prieto Velasco, Pilar León Arauz, Carlos Márquez Linares and Miguel Vega Expósito 2006: Process-oriented terminology management in the domain of coastal engineering. Terminology 12(2), 189-213. Faber et al. 2007 = Pamela Faber, Pilar León Araúz, Juan Anatonio Prieto Velasco and Arianne Reimerink. 2007: Linking Images and Words: the Description of Specialized Concepts. International Journal of Lexicography 20(1), 39-65. Fata 2009 = Idilkó Fata 2009: Das zweisprachige Translationswörterbuch für Fachsprachen in der wissenschaftlichen Theorie und Praxis. Budapest: Tinta Könyvkiadó. Felber 1979 = Helmut Felber (Ed.) 1979: Theories of Terminology and Terminological Lexicography. Vienna, New York: Springer. Felber/Budin 1989 = Helmut Felber and Gerhard Budin 1989: Terminologie in theorie und praxis. Tübingen: Gunter Narr Verlag. Feliú/Vivaldi/Cabré 2002 = Judit Feliú, Jorge Vivaldi and María Teresa Cabré 2002: Towards an ontology for a human genome knowledge base. Proceedings of the Third International Conference on Language RTesources and Evaluation, LERC 2002, May 29-31, 1885-1890. 2002, Las Palmas, Spain. Available at: []. Fillmore 1976 = Charles J. Fillmore 1976: Frame Semantics and the nature of language. In Steven R. Harnad, Horst D. Steklis and Jane Lancaster (Eds.), Origins and Evolution of Language and Speech. Annals of the NY Academy of Sciences, 280: 20-32. Fillmore 1982 = Charles J. Fillmore 1982: Frame Semantics. In The Linguistic Society of Korea (Ed.) Linguistics in the Morning Calm. Seoul: Hanshin, 111-137. Fillmore 1985 = Charles J. Fillmore 1985: Frames and the semantics of understanding. Quaderni di Semantica 6(2), 222-254. Fillmore/Atkins 1992 = Charles J. Fillmore and B. T. Sue Atkins 1992: Towards a frame-based organization of the lexicon: The semantics of RISK and its neighbors. In Adrienne Lehrer and Eva Kit-

254 | References tay (Eds.), Frames, Fields, and Contrast: New Essays in Semantics and Lexical Organization. Hillsdale, N.J: Lawrence Erlbaum, 75-102. Frawley 1988 = William Frawley 1988: New forms of specialized dictionaries. International Journal of Lexicography 1:3, 189-213. Fuertes-Olivera 2009 = Pedro A. Fuertes-Olivera 2009: Systematic introductions in specialised dictionaries: Some proposals in relation with the Accounting Dictionaries. In Sandro Nielsen and Sven Tarp (Eds.): Lexicography in the 21st Century. In Honour of Henning Bergenholtz. Amsterdam, Philadelphia: John Benjamins, 161-178. Fuertes-Olivera 2010 = Pedro A. Fuertes-Olivera (Ed.) 2010: Specialized Dictionaries for Learners. Berlin, New York: De Gruyter. Fuertes-Olivera 2011 = Pedro A. Fuertes-Olivera 2011: Equivalent selection in specialized elexicography: A case study with Spanish accounting terms. Lexikos 21, 95-119. Fuertes-Olivera 2012 = Pedro A. Fuertes-Olivera 2012: Lexicography and the interned as a (re-) source. Lexicographica 28, 49-70. Fuertes-Olivera 2013a = Pedro A. Fuertes-Olivera 2013a: Elexicography: The continuing challenge of applying new technology to dictionary making. In H. Jackson (Ed.): Bloomsbury Companion to Lexicography. London, New Delhi, New York, Sydney: Bloomsbury Academic, 323-340. Fuertes-Olivera 2013b = Pedro A. Fuertes-Olivera 2013b: The theory and practice of specialised online dictionaries for translation. Lexicographica 29, 69-91. Fuertes-Olivera/Arribas-Baño 2008 = Pedro A. Fuertes-Olivera and Ascensión Arribas-Baño 2008: Pedagogical Specialised Lexicography. Amsterdam, Philadelphia: John Benjamins. Fuertes-Olivera/Bergenholtz 2011a [2013a] = Pedro A. Fuertes-Olivera and Henning Bergenholtz (Eds.) 2011a: e-Lexicography: The Internet, Digital Initiatives and Lexicography. London, New York: Continuum. Also in Pedro A. Fuertes-Olivera and Henning Bergenholtz (Eds.) 2013: eLexicography: The Internet, Digital Initiatives and Lexicography. London, New Delhi, New York, Sydney: Bloomsbury Academic. Fuertes-Olivera/Bergenholtz 2011b [2013b] = Pedro A. Fuertes-Olivera and Henning Bergenholtz 2011b: Introduction: The construction of internet dictionaries. In Pedro A. Fuertes-Olivera and Henning Bergenholtz (Eds.): e-Lexicography: The Internet, Digital Initiatives and Lexicography. London, New York: Continuum, 1-16. Also in Pedro A. Fuertes-Olivera and Henning Bergenholtz (Eds.) 2013: e-Lexicography: The Internet, Digital Initiatives and Lexicography. London, New Delhi, New York, Sydney: Bloomsbury Academic, 1-16. Fuertes-Olivera/Nielsen 2011 = Pedro A. Fuertes-Olivera and Sandro Nielsen 2011: The Dynamics of terms in accounting: What the construction of the Accounting Dictionaries reveals about metaphorical terms in culture-bound subject fields. Terminology 17(1), 157-180. Fuertes-Olivera/Nielsen 2012 = Pedro A. Fuertes-Olivera and Sandro Nielsen 2012: Online dictionaries for assisting translators of LSP texts: The Accounting Dictionaries. International Journal of Lexicography 25(2), 191-215. Fuertes-Olivera/Nielsen 2014 = Pedro A. Fuertes-Olivera and Sandro Nielsen 2013: The Dynamics of accounting terms in a globalized environment: The role of English as Lingua Franca. In Rita Temmerman and Marc van Campenhoudt (Eds.): The Dynamics of Culture-bound Terminology in Monolingual and Multilingual Communications: An Interdisciplinary Perspective. Amsterdam, Philadelphia: John Benjamins (To appear) Fuertes-Olivera/Niño Amo 2011 [2013] = Pedro A. Fuertes-Olivera and Marta Niño Amo 2011 [2013]: Internet dictionaries for communicative and cognitive Functions: El Diccionario Inglés-Español de Contabilidad. In Pedro A. Fuertes-Olivera and Henning Bergenholtz (Eds.): e-Lexicography: The Internet, Digital Initiatives and Lexicography. London, New York: Continuum, 168-186. Also in Pedro A. Fuertes-Olivera and Henning Bergenholtz (Eds.) 2013: e-Lexicography: The Internet,

Other Literature | 255 Digital Initiatives and Lexicography. London, New Delhi, New York and Sydney: Bloomsbury Academic, 168-186. Fuertes-Olivera/Tarp 2011 = Pedro A. Fuertes-Olivera and Sven Tarp 2011: Lexicography for the third millennium: Cognitive-oriented specialized dictionaries for learners. Ibérica 21, 141-162. Available at: []. Fuertes-Olivera et al. 2012 = Pedro A. Fuertes-Olivera. Henning Bergenholtz, Sando Nielsen and Marta Niño Amo 2012: Classification in lexicography: The concept of collocation in the Accounting Dictionaries. Lexicographica 28, 291-305. Fuertes-Olivera et al. 2013 = Pedro A. Fuertes-Olivera, Henning Bergenholtz, Pablo Gordo Gómez, Sandro Nielsen, Marta Niño Amo, Ángel de los Rios Rodicio, Ángeles Sastre Ruano and Marisol Velasco Sacristán 2013: The construction of online specialized dictionaries: theory and practice. Fachsprache. International Journal of Specialized Communication XXXV 1-2, 25-42. Fuertes-Olivera et al. 2010 = Pedro Fuertes Olivera, Pablo Gordo Gómez, Marta Niño Amo, Angel de los Rios Rodicio, Angeles Sastre Ruano, Sven Tarp, Marisol Velasco Sacristán and Sandro Nielsen, Lise Mourier, Henning Bergenholtz 2010: Diccionario de Contabilidad Inglés-Español. Navarra: Thomson Reuters-Aranzadi. Ganilh 1826 = Charles Ganilh 1826: Préface. Dictionnaire analytique d’économie politique. Paris/Brussels: Ladvocat. Gaudin 1993 = François Gaudin 1993: Socioterminologie: propos et propositions épistémologiques. Le Language et l’Homme 28(4), 247-258. Gómez González-Jover 2006 = Adelina Gómez González-Jover 2006: Meaning and animorphism in modern lexicography. Terminology 12 (2), 215-234. Gouws 2011 [2013] = Rufus H. Gouws 2011: Learning, Unlearning and Innovation in the Planning of Electronic Dictionaries. In Pedro A. Fuertes-Olivera and Henning Bergenholtz (Eds.) 2011: eLexicography: The Internet, Digital Initiatives and Lexicography. London, New York: Continuum, 17-29. Also in Pedro A. Fuertes-Olivera and Henning Bergenholtz (Eds.) 2013: e-Lexicography: The Internet, Digital Initiatives and Lexicography. London, New Delhi, New York, Sydney: Bloomsbury Academic, 17-29. Gouws 2012 = Rufus H. Gouws 2012a: Theoretical lexicography and the International Journal of Lexicography. International Journal of Lexicography 25(4), 450-463. Gouws/Prinsloo 2005 = Rufus H. Gouws and D.J. Prinsloo 2005: Principles and Practice of South African Lexicography. Stellenbosch: SUN Press. Gouws/Tarp 2008 = Rufus H. Gouws and Sven Tarp 2008: A lexicographic approach to language policy and recommendations for future dictionaries. Lexikos 18, 232-255. Granger/Paquot 2010 = Sylviane Granger and Magali Paquot (Eds.) 2010: eLexicography in the 21st Century: New Challenges, New Applications. Louvain-la-Neuve: Presses Universitaires de Louvain. Cahiers du Cental 7. Granger/Paquot 2012 = Sylviane Granger and Magali Paquot (Eds.) 2012: Electronic Lexicography. Oxford: Oxford University Press. Gruber 2009 = Tom Gruber 2009: Ontology. In Ling Liu and M. Tamer Öszu (Eds.): The Encyclopedia of Database Systems. Berlin: Springer. Available at: [] Haensch et al. 1982 = Günther Haensch, Lothar Wolf, Stefan Ettinger and Reinhold Werner 1982: La lexicografía. De la linqüística teórica a la lexicografía práctica. Madrid: Gredos. Haiman 1980 = John Haiman 1980: Dictionaries and encyclopedias. Lingua 50, 329-357. Hanks 2012 = Patrick Hanks 2012: The corpus revolution in lexicography. International Journal of Lexicography 25:4, 398-436. Hansen/Andersen 2000 = Erik Jørgen Hansen and Bjarne Hjort Andersen 2000: Et sociologisk værktøj. Introduktion til den kvantitative metode. Copenhagen: Hans Reitzels Forlag.

256 | References Harris 1704 = John Harris 1704: Preface. Lexicon Technicum: or, an Universal English Dictionary of Arts and Sciences: Explaining Not Only the Terms of Art, but the Arts Themselves. London: D. Brown, J. Walthoe, J. Knapton, B. and S. Tooke, D. Midwinter, B. Cowse, T. Ward, E. Symon, E. Valentine, and J. Clark. Harris/Hutton 2007 = Roy Harris and Christopher Hutton 2007: Definition in Theory and Practice. Language, Lexicography and the Law. London: Continuum. Hartmann 2000 = Reinhard Hartmann 2000: A rejoinder to Laurence Urdang. International Journal of Lexicography 13(3), 194-196. Hartmann 2009 = Reinhard Hartmann 2009: [Review of] B.T. Sue Atkins and Michael Rundell. The Oxford Guide to Practical Lexicography. International Journal of Lexicography 22(1), 89-94. Hartmann/Gregory 2001 = Reinhard Hartmann and James Gregory 2001: Dictionary of Lexicography. London, New York: Routledge. Haß/Schmitz 2010 = Ulrike Haß and Ulrich Schmitz (Eds.) 2010: Lexikographie im Internet. Thematic section. Lexicographica 26. Hatherall 1984 = Glyn Hatherall 1984: Studying dictionary use: Some findings and proposals. In R.R.K. Hartmann (Ed.): LEXeter '83 Proceedings. Papers from the International Conference on Lexicography at Exeter, 9–12 September 1983. Lexicographica. Series Maior 1. Tübingen: Max Niemeyer, 183-189. Hausmann 1977 = Franz Josef Hausmann 1977: Einführung in die Benutzung der neufranzösischen Wörterbücher. Tübingen: Niemeyer. Hausmann 1985 = Franz Josef Hausmann 1985: Kollokationen im deutschen Wörterbuch. Ein Beitrag zur Theorie des lexikographischen Beispiels. In Henning Bergenholtz and Joachim Mugdan (Eds.): Lexikographie und Grammatik. Akten des Essener Kolloquiums zur Grammatik im Wörterbuch 28.-30.6.1984. Tübingen: Niemeyer, 118-129. Hausmann 1989 = Franz Josef Hausmann 1989: Kleine Weltgeschichte der Metalexikographie. In Herbert Ernst Wiegand (Ed.): Wörterbücher in der Diskussion. Vorträge aus dem Heidelberger Lexikographischen Kolloquium. Tübingen: Niemeyer, 75-109. Hearst 1992 = Marti A. Hearst 1992: Automatic acquisition of hyponyms from large text corpora. Proceedings of the 14th international conference on computational linguistics (COLING 1992) (Vol 2). ACL, 539-545. Available at: [] Heid 2011 [2013] = Ulrich Heid 2011: Electronic Dictionaries as Tools: Towards an Assessment of Usability. In Pedro A. Fuertes-Olivera and Henning Bergenholtz (Eds.): e-Lexicography: The Internet, Digital Initiatives and Lexicography. London & New York: Continuum, 287-304. Also in Fuertes-Olivera, Pedro A/ Bergenholtz, Henning (Eds.) 2013: e-Lexicography: The Internet, Digital Initiatives and Lexicography. London, New Delhi, New York and Sydney: Bloomsbury Academic, 287-304. Heid/Gojun 2012 = Ulrich Heid and Anita Gojun 2012: Term candidate extraction for terminography and CAT. In Ruth E. Vatvedt Fjeld and Julie Matilde Torjusen (Eds.): Proceedings of the 15th EURALEX International Congress. Oslo: Department of Linguistics and Scandinavian Studies, 585594. Available at: [] Heid/Prinsloo/Bothma 2012 = Ulrich Heid, D.J. Prinsloo and Theo J.D. Bothma 2012: Dictionary and corpus data in a common portal: state of the art and requirements for the future. Lexicographica 28, 269-291. Henriksen 1992 = Petter Henriksen 1992: Forholdet mellem encyklopedologi og leksikografi. In Ruth Vatvedt Fjeld (Ed.): Nordiske studier i leksikografi. Rapport fra Konferanse om leksikografi i Norden 28.-31.mai 1991. Oslo: Nordisk Forening for Leksikografi, 219-227. Hjørland 1998 = Birger Hjørland 1998: Theory and metatheory of Information Science: A new interpretation. Journal of Documentation 54:5, 606-621.

Other Literature | 257 Hoare 2009 = Michael Rand Hoare 2009: Scientific and technical dictionaries. In Anthony P. Cowie (Ed.): The Oxford History of English Lexicography. Volume II: Specialized Dictionaries. Oxford: Oxford University Press, 47-93. Hospedales/Rodríguez 2004 = Margarita Hospedales and Manel Rodríguez 2004: The GENOMA-KB Platform: Queries over Integrated Linguistic Resources. LERC 4, 1305-1308. Available at: [] Humbley 1997 = John Humbley 1997: Is terminology specialized lexicography? The experience of French-speaking countries. Hermes, Journal of Linguistics 18, 13-31. Humbley 2002 = John Humbley 2002: Nouveaux dictionnaires, nouveaux rapports avec les utilisateurs. Meta, 47:1, 95-104. Humbley 2008 = John Humbley. 2008. Review of Ten Hacken, Pius (Ed.). 2006. Terminology. Computing and Translation. Terminology 14(2): 261-265. ISO 12620 (1999). Computers applications in terminology - Data categories. ISO 12620 (2009). Terminology and other languages and content sources - Specification of data categories and management of a Data Category Registry for language resources. ISO 704 (2009). Terminology work. Principles and Methods. Johnson 1755 = Samuel Johnson 1755: Dictionary of the English Language. London: J. & P. Knapton. Johnson 1987 = Mark Johnson 1987: The Body in the Mind. Chicago: The University of Chicago Press. Kageura/L’Homme 2008 = Kyo Kageura and Marie-Claude L’Homme 2008: Reflections on fifteen years of research and development in terminology. Terminology 14(2): 153-157 Kalverkämper 1989 = Hartwig Kalverkämper 1989: Gemeinsprache und Fachsprachen - Plädoyer für eine integrierende Sichtwiese. In Gerhard Stickel (Ed.): Deutsche Gegenwartsprache. Tendenzen und Perspektiven. Berlin, New York: de Gruyter, 88-133. Kalverkämper 1990a = Hartwig Kalverkämper 1990a: Das Fachwörterbuch für den Laien. In Franz Josef Hausmann, Oskar Reichmann, Herbert Ernst Wiegand and Ladislav Zgusta (Eds.): Wörterbücher, Dictionaries, Dictionnaires. An International Encyclopedia of Lexicography. Second Volume. Berlin, New York: Walter de Gruyter, 1512-1523. Kalverkämper 1990b = Hartwig Kalverkämper 1990b: Der Begriff der “Fachlichkeit” im modernen Italienischen. Lexikalische Systematik und textuelle Integration. Quardini di Semantica. Rivista internazionale di semantica teorica e applicata 11, 79-115. Kerremans 2004 = Koen Kerremans 2004: Categorisation frameworks in termontography. Linguistica Antverpiensia New Series 3: 262-277. Kerremans et al. 2005a = Koen Kerremans, Isabelle Desmeytere, Rita Temmerman and Patrick Wille 2005a: Application-oriented terminography in financial forensics. Terminology 11(1), 83-106. Kerremans et al. 2005b = Koen Kerremans, Yan Tang, Rita Temmerman and Gang Zhao 2005b: Towards ontology-based e-mail fraud detection. In G. Bento, A. Cardoso/G. Dias (Eds.), Proceedings of EPIA 2005 BAOSW workshop of 12th Portuguese conference on AI, 106-111. Available:[ 0.pdf] Kerremans/Temmerman/DE Baer 2010 = Koen Kerremans, Rita Temmerman and Peter De Baer 2010: Competency-based job descriptions and termontography. The case of terminological variation. In M. Thelen and F. Steurs (Eds.): Terminology in everyday life. Amsterdam, Philadelphia: John Benjamins, 179-191. Kilgarriff 2012 = Adam Kilgarriff 2012: [Review of] Pedro A. Fuertes-Olivera/Henning Bergenholtz (Eds.). e-Lexicography: The Internet, Digital Initiatives and Lexicography. Kernerman Dictionary News, July 2012: 26-29. Kilian et al. 2012 = Christiane Krause Kilian, Ana Eliza Pereira Bocorny, Rodrigo Wilkens and Aline Villavicencio 2012: Critérios de seleção de termos utilizados na construção de um glossário pedagógico online baseado em corpus especializado. Revista Entrelinhas 7:2, 277-292.

258 | References Kosem/Kosem 2011 = Iztok Kosem and Karmen Kosem (Eds.) 2011: Electronic Lexicography in the 21st Century: New Applications for New Users. Ljubljana: Institute for Applied Slovene Studies. Kromann/Riiber/Rosbach 1984 = Hans-Peder Kromann. Theis Riiber and Poul Rosbach 1984. Überlegungen zu Grundfragen der zweisprachigen Lexikographie. In Herbert Ernst Wiegand (Eds.): Studien zur neuhochdeutschen Lexikographie V. Hildesheim, New York: Olms, 159-238. Kudashev 2007 = Igor Kudashev 2007: Terminography vs. lexicography opposition revisited. In Nina Niemelä and Esa Lehtinen (Eds.): Theory of Translation, LSP and Multilingualism: 27th VAKKI Symposium, Vaasa 9.-10.2.2007. Vasa: Vaasan yliopisto, 157-166. Kwary 2012 = Deny Arnos Kwary 2012: Adaptive hypermedia and user-oriented data for online dictionaries: A case study on an English dictionary of finance for Indonesian students. International Journal of Lexicography 25:1, 30-49. Kwary 2013 = Deny Arnos Kwary 2013: Principles for the Design of Business Dictionaries on Mobile Applications. Hermes, Journal of Language and Communication in Business 50, 69-82. Lakoff 1987 = George Lakoff 1987: Women, fire and dangerous things. Chicago: University of Chicago Press. Landau 2001 = Sidney Landau 2001: Dictionaries: The Art and Craft of Lexicography. Cambridge: Cambridge University Press. Laurén/Myking/Picht 1998 = Christer Laurén, Johan Myking and Heribert Picht 1998: Terminologie unter der Lupe. Vienna: TermNet. Law, Jonathan 2012. A Dictionary of Busienss and Management. Oxford: OUP. Layton 1965 = David Layton 1965: Diction and dictionaries in the diffusion of scientific knowledge: an aspect of the history of the popularization of science in Great Britain. The British Journal for the History of Science 2(3), 221-234. Leroyer 2011 [2013] = Patrick Leroyer 2011: Change of Paradigm: From Linguistics to Information Science and from Dictionaries to Lexicographical Information Tools. In Pedro A. Fuertes-Olivera and Henning Bergenholtz (Eds.): e-Lexicography: The Internet, Digital Initiatives and Lexicography. London & New York: Continuum, 121-140. Also Pedro A. Fuertes-Olivera and Henning Bergenholtz (Eds.) 2013: e-Lexicography: The Internet, Digital Initiatives and Lexicography. London, New Delhi, New York and Sydney: Bloomsbury Academic, 121-140. Lessing 1747 = Gotthold Ephraim Lessing 1747: Der Junge Gelehrte. Stuttgart: Reclam 1965. Lew 2004 = Robert Lew 2004: Which Dictionary for Whom? Receptive Use of Bilingual, Monolingual and Semi-bilingual Dictionaries by Polish Learners of English. Poznań: Motivex. Lew 2011 = Robert Lew (Ed.) (2011): Studies in dictionary use: Recent developments. Special Issue of International Journal of Lexicography 24(1), 1-153. Lew 2013 = Robert Lew 2013: Space restrictions in paper and electronic dictionaries and their implications for the design of production dictionaries. In Piotr Bański and Beata Wójtowicz (Eds.): Issues in Modern Lexicography. München: Lincom Europa. (to appear) Litkowski 2010 = Kenneth C. Litkowski 2010: [Review of] Hans C. Boas (Ed.) Multilingual Framenets in Computational Lexicography: Methods and Applications. Mouton de Gruyter. Berlin. 2009. X + 350 pages. ISBN 978-3-11-021296-9. List price € 99.95, US$ 140.00. International Journal of Lexicography 23:1, 105-109 Machlup 1983 = F. Machlup 1983: Semantic quirks in studies of information. In F. Machlup/ U. Mansfield (Eds.): The Study of Information: Interdisciplinary Messages. New York: John Wiley & Sons, 641-671. Marello 2003 = Carla Marello 2003: The Bilingual Dictionary. In R.R.K. Hartmann (Ed.): Lexicography. Critical Concepts II. London, New York: Routledge, 325-342. Marshman 2008 = Elizabeth Marshman 2008: Expressions of uncertainty in candidate knowledgerich contexts. Terminology 14(1), 124-151.

Other Literature | 259 Martin 1998 = Willy J.R. Martin 1998: Frames as definition models for terms. In A. Munteanu (Ed.), Proceedings of the International Conference on Professional Communication and Knowledge Transfer 2. Vienna: TermNet, 189-221. Martin 2001 = Willy J.R. Martin 2001: Frame-based definitions and the selection of multiword term candidates in DOT. In Actes des quatrièmes rencontres 'Terminologie et Intelligence Artificielle'. Fr. Vandoeuvre: Unité de Recherche et Innovation, INIST-CNRS, 55-65. Martin 2006 = Willy J.R. Martin 2006: In quest of a profile: Portrait of a terminologist as a young sublanguage expert. In Pius Ten Hacken (Ed.): Terminology, Computing and Translation. Tübingen: Gunter Narr Verlag, 73-94. Martin/van der Vliet 2003 = Willy J.R. Martin and Hennie van der Vliet 2003: Design and production of terminological dictionaries. In Peter van Sterkenburg (Ed.): A practical guide to lexicography. Amsterdam, Philadelphia: John Benjamins, 333-349. Martínez de Sousa 1995 = José Martínez de Sousa 1995: Diccionario de lexicografía práctica. Barcelona: Biblograf. Martínez Motos 2011 = Raquel Martínez Motos 2011: The role of interdisciplinarity in lexicography and lexicology. In Isabel Balteiro (Ed.): New Approaches to Specialized English Lexicology and Lexicography. Newcastle upon Tyne: Cambridge Scholars Publishing, 3–13. McCreary/Dolezal 1999 = Don R. McCreary and Frederic T. Dolezal 1999: A study of dictionary use by ESC students in an American university. International Journal of Lexicography 12(2), 107-145. McKay 2008 = Peter McKay 2008: [Review] of The New Palgrave Dictionary of Economics. 2nd ed. Ed. By Steven N. Durlauf and Lawrence E. Blume. New York: Palgrave Macmillan, 2008. 8 vols. Cybdragon2, 139-140. Meier 2003 = Hans H. Meier 2003: Lexicography as Applied Linguistics. In Reinhard Hartmann (Ed.): Lexicography. Critical Concepts III. London, New York: Routledge, 307-318. Menagarishvili 2003 = Olga Menagarishvili 2012: Dictionaries of Science as Participants in the Scientific Knowledge Economy. Minnesota: University of Minnesota. Merkel/Foo 2007 = Magnus Merkel and Jody Foo 2007: Terminology extraction and term ranking for standardizing term banks. In Proceedings of the 16th Nordic Conference of Computational Linguistics (NODALIDA-07). Tartu, 349-354. Meyer 2001 = Ingrid Meyer 2001: Extracting knowledge-rich contexts for terminography. In Didier Bourigault, Christian Jacquemin and Marie-Claude L'Homme (Eds.): Recent Advances in Computational Terminology. Amsterdam/Philadelphia: John Benjamins, 279-302. Meyer et al. 1992 = Ingrid Meyer, Douglas Skuce, Lynne Bowker and Karen Eck 1992: Towards a new generation of terminological resources: An experiment in building a terminological knowledge base. In Proceedings of COLING-92, Nantes, August 23-28, 1992, 956-960. Mikkelsen 1994 = Hans Kristian Mikkelsen 1994: Danske fagordbøger. In Henning Bergenholtz and Sven Tarp (Eds.): Manual i fagleksikografi. Aarhus: Systime, 258-305. Miles 2012 = A. D. Miles 2012: A-D. Miles’ Business English to Spanish Glossary.[] Moreno Ortiz 2000 = Antonio Moreno Ortiz 2000: OntoTerm™: un sistema abierto de representación conceptual. In Actas del XVI Congreso de la Sociedad Española para el Procesamiento del Lenguaje Artificial /(SEPLN),Vigo: España, 37-46. Available: [] Müller-Spitzer/Koplenig/Töpel 2011 = Carolin Müller-Spitzer, Alexander Koplenig and Antje Töpel 2011: What makes a good online dictionary? – Empirical insights from an interdisciplinary research project. In Iztok Kosem and Karmen Kosem (Eds.): Electronic Lexicography in the 21st Century: New Applications for New Users. Ljubljana: Institute for Applied Slovene Studies, 203–208.

260 | References Navigli 2009 = Navigli, Roberto 2009: Using cycles and quasi-cycles to disambiguate dictionary glosses. Proceedings of the 12th conference of the European chapter of the association for computational linguistics (EACL 2009). Athens: Greece, 594-602. Available at: [] Nesi 2000 = Nesi, Hilary 2000: The Use and Abuse of EFL Dictionaries. Tübingen: Max Niemeyer. Nielsen 2002a = Sandro Nielsen 2002a: Overvejelser angående det leksikografiske grundlag for en elektronisk bilingval regnskabsordbog. In LexicoNordica 9, 173-194. Nielsen 2002b = Sandro Nielsen 2002b: Lexicographical Basis for an Electronic Bilingual Accounting Dictionary: Theoretical Considerations. 2002. Available at: [] Nielsen 2007 = Sandro Nielsen 2007: Dictionary networking in an LSP learning context. In Maria Kuteeva/Helder Fanha Martins (Eds.) Teaching and Learning LSP: Blurring Boundaries. Lisbon: ISCAL/AELFE 2007, 367-373. Nielsen 2008 = Sandro Nielsen 2008: The effect of lexicographical information costs on dictionary making and use. Lexikos 18, 170-189. Nielsen 2009a = Sandro Nielsen 2009a: Reviewing printed and electronic dictionaries. A theoretical and practical framework. In Sandro Nielsen and Sven Tarp (Eds.), Lexicography in the 21st Century. In Honour of Henning Bergenholtz. Amsterdam, Philadelphia: John Benjamins, 23-42. Nielsen 2009b = Sandro Nielsen 2009b. The evaluation of the outside matter in dictionary reviews. Lexikos 19, 207-224. Nielsen 2010 = Sandro Nielsen 2010: Specialized Translation Dictionaries for Learners. In Pedro A. Fuertes-Olivera (Ed.): Specialized Dictionaries for Learners. Berlin, New York: De Gruyter, 69– 82. Nielsen 2013 = Sandro Nielsen 2013: Domain-specific knowledge in lexicography: how it helps lexicographers and users of accounting dictionaries intended for communicative usage situations. Hermes, Journal of Language and Communication in Business 50, 51-60. Nielsen 2014 = Sandro Nielsen 2014: LSP lexicography and typology of specialized dictionaries. In Gerhard Budin, Christer Laurén and John Humbley (Eds.): Language for Special Purposes. An International Handbook. Berlin, New York: De Gruyter, (to appear) Nielsen/Almind 2011 [2013] = Sandro Nielsen and Richard Almind 2011: From data to dictionary. In Pedro A. Fuertes-Olivera and Henning Bergenholtz (Eds.): e-Lexicography: The Internet, Digital Initiatives and Lexicography. London, New York: Continuum, 141-167. Also in Pedro A. FuertesOlivera and Henning Bergenholtz (Eds.) 2013: e-Lexicography: The Internet, Digital Initiatives and Lexicography. London, New Delhi, New York, Sydney: Bloomsbury Academic, 141-167. Nielsen/Fuertes-Olivera 2013 = Sandro Nielsen and Pedro A. Fuertes-Olivera 2013: Development in lexicography: From polyfunctional to monofunctional accounting dictionaries. Lexikos 23, 323.347. Nielsen/Mourier 2005 = Sandro Nielsen and Lise Mourier 2005: Internet accounting dictionaries: Present solutions and future opportunities. Hermes. Journal of Linguistics 34: 83-116. Nielsen/Mourier 2007 = Sandro Nielsen and Lise Mourier 2007: Design of a function-based internet accounting dictionary. In Henrik Gottlieb and Jens Erik Mogensen (Eds.): Dictionary Visions, Research and Practice. Amsterdam, Philadelphia: John Benjamins, 2007, 119-135. Nielsen et al. 2009 = Sandro Nielsen, Lise Mourier, Henning Bergenholtz and Pedro A. Fuertes Olivera, Pablo Gordo Gómez, Marta Niño Amo, Ángel de los Rios Rodicio, Ángeles Sastre Ruano, Sven Tarp, Marisol Velasco Sacristán 2009: El Diccionario Inglés-Español de Contabilidad. Aarhus: Centre for Lexicography. Nord 2002 = Britta Nord 2002: Hilfsmittel beim Übersetzen. Eine empirische Studie zum Rechercheverhalten professioneller Übersetzer. Frankfurt am Main: Peter Lang.

Other Literature | 261 Norton 2010 = Melanie J. Norton 2010: Information and Information Science. In Melanie J. Norton (Ed.): Introductory Concepts in Information Science. New Jersey: Information Today Inc., 3-18. Och/Ney 2003 = Franz Josef Och and Hermann Ney 2003: A systematic comparison of various statistical alignment models. Computational Linguistics 29(1), 19-51. Odgen/Richards 1923 = Charles K. Odgen and Ivor Armstrong Richards 1923: The Meaning of Meaning: A Study of the Influence of Language upon Thought and the Science of Symbolism. London: Routledge & Kegan Paul. Opitz 1990 = Kurt Opitz 1990: The technical dictionary for the expert. In Franz Josef Hausmann. Oskar Reichmann, Herbert Ernst Wiegand and Ladislav Zgusta (Eds.): Wörterbücher, Dictionaries, Dictionnaires. An International Encyclopedia of Lexicography. Second Volume. Berlin, New York: Walter de Gruyter, 1505-1512. Osselton 1989 = Noel Edward Osselton 1989: The history of academic dictionary criticism. In Franz Josef Hausmann. Oskar Reichmann, Herbert Ernst Wiegand and Ladislav Zgusta (Eds.): Wörterbücher, Dictionaries, Dictionnaires. An International Encyclopedia of Lexicography. Second Volume. Berlin, New York: Walter de Gruyter, 225-230. Pálfi/Bergenholtz 2007 = Loránd-Levente Pálfi and Henning Bergenholtz 2007: Dansk Glossarium til at hjelpe de Lægfolk tilrette, som gjerne gad læse en dansk Bog. Nomos 5:1, 55-91. Pastor/Alcina 2010 = Verónica Pastor and Amparo Alcina 2010: Search techniques in electronic dictionaries: A classification for translators. International Journal of Lexicography 23(3): 307354. Pearson 1998 = Jennifer Pearson 1998: Terms in Context. Amsterdam, Philadelphia: John Benjamins. Pérez Hernández 2002 = M. Chantal Pérez Hernández 2002: Explotación de los córpora textuales informatizados para la creación de bases de datos terminológicas basadas en el conocimiento. Capítulo 3.3: Terminografía y lexicografía. Estudios de Lingüística Española 18. [http: //] Peter et al. 2006 = Pam Peters, Alan Jones, Adam Smith, Theresa Winchester-Seeto, Jenny Middledorp and Peter Petocz 2006: TermFinder: creating online termbanks of technical terms for early university study. Journal of Applied Linguistics 3(2): 219-248. Peirce 1868 = Charles S. Peirce 1868. Questions concerning certain Faculties claimed by Man. Journal of Speculative Psychology 2: 102-114. Pennacchiotti/Pantel 2006 = Marco Pennacchiotti and Patrick Pantel 2006: Ontologizing semantic relations. Proceedings of the 21st international conference on computational linguistics and the 44th annual meeting of the association for computational linguistics (COLING-ACL 2006). ACL, 793-800. Available at: [ 33] Picht 1985 = Heribert Picht 1985: Leksikografi – Terminografi. Nordisk terminologikursus 2. Bind 1, 351-372. Piotrowski 2009 = Tadeusz Piotrowski 2009: [Review of] Sven Tarp. Lexicography in the Borderland Between Knowledge and Non-knowledge. General Lexicographical Theory with Particular Focus on Learner's Lexicography. International Journal of Lexicography 22(4), 480-486. Popper 1963 = Karl Raimund Popper 1963: Conjectures and Refutations. London: Routledge & Kegan Paul. Postlethwayt 1749 = Malachy Postlethwayt 1749: A Dissertation on the Plan, Use, and Importance, of the Universal Dictionary of Trade and Commerce. London: John and Paul Knapton. Posner 1988 = Roland Posner 1988: What is an academic discipline? In Regina Claussen/ Roland Daube-Schackat (Eds.): Gedankenzeichen. Festschrift für Klaus Oehler zum 60. Geburtstag. Tübingen: Stauffenburg Verlag, 165-185.

262 | References Prieto Velasco/López Rodríguez 2009 = Juan Antonio Prieto Velasco and Clara Inés López Rodríguez 2009: Managing graphic information in terminological knowledge bases. Terminology 15(2), 179-213. Prinsloo 2009 = D.J. Prinsloo 2009: The role of corpora in future dictionaries. In Sandro Nielsen and Sven Tarp (Eds.): Lexicography in the 21st century. In Honour of Henning Bergenholtz. Amsterdam: John Benjamins, 181-206. Prinsloo et al. 2012 = D.J. Prinsloo, Ulrich Heid, Theo J.D. Bothma and Gertrud Faaß 2012: Devices for Information Presentation in Electronic Dictionaries. Lexikos 22, 290-320. Reimerink/García de Quesada/Montero Martínez 2010 = Arianne Reimerink, Mercedes García de Quesada and Silvia Montero Martínez 2010: Contextual information in terminological knowledge bases. Journal of Pragmatics 42(7), 1928-1950. Rei 1975 = Alain Rey 1975: Terminologies et ‘terminographie’. Banque des Mots 10, 145-154. Rey 1988 = Alain Rey 1988: Terminologie et lexicographie. Parallèles 10, 27-35. Riggs 1989 = Fred W. Riggs 1989: Terminology and lexicography: Their complementarity. International Journal of Lexicography 2:2, 89-110. Roche 2012 = Christophe Roche 2012: Should terminology principles be re-examined? In Guadalupe Aguado de Cea, Mari Carmen Suárez-Figueroa, Raúl García-Castro and Elena Montiel-Ponsoda (Eds.): Proceedings of the 10th Terminology and Knowledge Engineering Conference: New frontiers in the constructive symbiosis of terminology and knowledge engineering. Madrid: Universidad Politécnica de Madrid, 17-32. Rocheteau/Daille 2011 = Jérôme Rocheteau and Béatrice Daille 2011: TTC TermSuite: A UIMA application for multilingual terminology extraction from comparable corpora. IJCNLP, Chiang Mai, Thailand. Available at: [] Rodríguez 2013 = Ángel Rodríguez Gallardo 2013: The function theory and its application on manuals of economics. Hermes, Journal of Language and Communication in Business 50, 83-95. Rosch 1978 = Eleanor Rosch 1978: Principles of categorization. In E. Rosch and B. Lloyd (Eds.), Cognition and categorization. Hillsdale: Lawrence Erlbaum, 27-48. Rossenbeck 1994 = Klaus Rossenbeck 1994: Enzyklopädische Information im zweisprachigen Fachwörterbuch. In Henning Bergenholtz and Burkhard Schaeder (Eds.): Fachlexikographie. Fachwissen und seine Repräsentation in Wörterbüchern. Tübingen: Günter Narr Verlag, 133159. Rondau 1981 = Guy Rondeau 1981: Introduction à la terminologie. Montréal: Centre éducatif et culturel. Rundell 1999 = Michael Rundell 1999: Dictionary use in production. International Journal of Lexicography 12(1), 35-53. Rundell 2010 = Michael Rundell 2010: What future for the learner’s dictionary? In Ilan J. Kernerman and Paul Bogaards (Eds.): English Learners’ Dictionaries at the DSNA 2009. Jerusalem: Kdictionaries, 169-175. Rundell 2012 = Michael Rundell 2012: ‘It works in practice but will it work in theory?’ The uneasy relationship between lexicography and matters theoretical. In Ruth E. Vatvedt Fjeld and Julie Matilde Torjusen (Eds.) (2012): Proceedings of the 15th EURALEX International Congress. Oslo: Institutt for lingvistiske og nordiske studier, 47-92. Sager 1984 = Juan C. Sager 1984: Terminology and the technical dictionary. In R.R.K. Hartmann (Ed.): LEXeter’83 proceedings. Papers from the International Conference on Lexicography at Exeter, 9-12 September 1983. Niemeyer: Tübingen, 315-326. Saracevic 1996 = Tefko Saracevic 1996: Relevance reconsidered ’96. In P. Ingwersen and N.O. Pors (Eds.): Proceedings of CoLIS 2: Second International Conference on Conceptions of Library and Information Science: Integration in Perspective. 13-16 October 1996. Copenhagen: The Royal School of Librarianship, 201-218.

Other Literature | 263 Scerba 1940 [1995] = Lev V. Scerba 1940: Towards a general theory of lexicography. International Journal of Lexicography 8:4 1995, 315-350. Schmidt 2009 = Thomas Schmidt 2009: The Kicktionary – A multilexical lexical resource of football language. In Hans C. Boas (Ed.) Multilingual Framenets in Computational Lexicography. Berlin & New York: De Gruyter, 101-132. Schmitz 2006 = Klaus-Dirk Schmitz 2006: Terminology and terminological databases. In Keith Brown (Ed.), Encyclopedia of Language and Linguistics, 2nd edition. Amsterdam: Elsevier, 578587. Online version: [] Scholfield 1999 = Phil Scholfield 1999: Dictionary use in reception. International Journal of Lexicography 12(1): 13-34. Scott 2008 = Michael Scott 2008: WordSmith Tools. Oxford: Oxford University Press. Sinclair 1987 = John Sinclair (Ed.) (1987): Looking Up: An Account of the COBUILD project in lexical computing and the development of the Collins COBUILD English language dictionary. London: Collins. Sinclair 1991 = John Sinclair 1991: Corpus, Concordance, Collocation. Oxford: Oxford University Press. Sorokoletov 1982 = Fedor Petrovich 1982: Traditionen der sowjetrussischen Lexikographie. In Werner Wolski (Eds.): Aspekte der sowjetrussischen Lexikcographie. Übersetzungen, Abstracts, bibliographische Angaben. Tübingen: Niemeyer, 63-88. Svensén 1992 = Bo Svensén 1992: Terminologi och lexikografi. In Terminologiläran och dess relationer till andra områden. Stockholm: NordTerm, 97-122. Svensén 2009 = Bo Svensén 2009: A Handbook of Lexicography. The Theory and Practice of Dictionary-Making. Cambridge: Cambridge University Press. Swales 1990 = John Swales 1990: Genere Analysis. Cambridge: Cambridge University Press. Swanepoel 2008 = Piet Swanepoel 2008: Towards a framework for the description and evaluation of dictionary evaluation criteria. Lexikos 18, 207-231. Tarp 2004a = Sven Tarp 2004a: How can dictionaries assist translators? In Chan Sin-wai (Ed.): Translation and Bilingual Dictionaries. Tübingen: Max Niemeyer Verlag, 23-39. Tarp 2004b = Sven Tarp 2004b: Korrektur og retning som leksikografiske funktioner. Hermes, Journal of Linguistics 33, 117-147. Tarp 2005 = Sven Tarp 2005: The pedagogical dimension of the well-conceived specialised dictionary. Ibérica 10, 7-21. Tarp 2007a = Sven Tarp 2007a: Lexicography in the Information Age. Lexikos 17, 170-179. Tarp 2007b = Sven Tarp 2007b: ¿Qué requisitos debe cumplir un diccionario de traducción en el siglo 21? In Pedro A. Fuertes Olivera (Ed.): Problemas lingüísticos en la traducción especializada. Valladolid: Universidad de Valladolid, 227-256. Tarp 2008 = Sven Tarp 2008: Lexicography in the Borderland Between Knowledge and Nonknowledge. Tübingen: Niemeyer. Tarp 2009a = Sven Tarp 2009a: Beyond lexicography: New visions and challenges in the Information Age. In Henning Bergenholtz, Sandro Nielsen and Sven Tarp (Eds.): Lexicography at a Crossroads: Dictionaries and Encyclopedias today, Lexicographical Tools tomorrow. Bern: Peter Lang, 17-32. Tarp 2009b = Sven Tarp 2009b: Reflections on lexicographic user research. Lexikos 19, 275-296. Tarp 2010 = Sven Tarp 2010: Reflections on the academic status of lexicography. Lexikos 20, 450465. Tarp 2011 [2013] = Sven Tarp 2011: Lexicographical and other e-tools for consultation purposes: Towards the individualization of needs satisfaction. In Pedro A. Fuertes-Olivera and Henning Bergenholtz (Eds.) 2011: e-Lexicography: The Internet, Digital Initiatives and Lexicography.

264 | References London, New York: Continuum, 54-70. Also Pedro A. Fuertes-Olivera and Henning Bergenholtz (Eds.) 2013: e-Lexicography: The Internet, Digital Initiatives and Lexicography. London, New Delhi, New York and Sydney: Bloomsbury Academic, 54-70. Tarp 2012a = Sven Tarp 2012a: Do we need a (new) theory of lexicography? Lexikos 22, 345-356. Tarp 2012b = Sven Tarp 2012b: Online dictionaries: today and tomorrow. Lexicographica 28, 253267. Tarp 2012c = Sven Tarp 2012c: Specialised lexicography: 20 years in slow motion. Ibérica 24, 117128. Tarp 2012d = Sven Tarp 2012d: Three hundred years of crisis in economic dictionaries and encyclopaedias. Hermes, Journal of Language and Communication in Business 49, 67-77. Tarp 2013a = Sven Tarp 2013a: Old wisdom: The highly relevant lexicographical knowledge obtainable from a specialized dictionary from 1774. Lexikos 23, 394-413. Tarp 2013b = Sven Tarp 2013b: Lexicographical functions. In Rufus H. Gouws, Ulrich Heid, Wolfgang Schweickard and Herbert Ernst Wiegand (Eds.): Dictionaries. An International Encyclopedia of Lexicography. Supplementary volume: Recent Developments with Special Focus on Computational Lexicography. Berlin, New York: Walter de Gruyter, 93-100. Tarp 2013c = Sven Tarp 2013c: What should we demand from an online dictionary for specialized translation? Lexicographica 29, 146-162. Temmerman 1997 = Rita Temmerman 1997: Questioning the univocity ideal. The difference between sociocognitive terminology and traditional terminology. Hermes. Journal of Linguistics 18, 5190. Temmerman 2000 = Rita Temmerman 2000: Towards New Ways of Terminology Description. The Sociocognitive Approach. Amsterdam, Philadelphia: John Benjamins. Temmerman/Kerremans 2003 = Rita Temmerman and Koen Kerremans 2003: Termontography: Ontology building and the sociocognitive approach to terminology description. In E. Hajicová, A. Kotesovcova and J. Mirovsky (Eds.): Proceedings of CIL 17. Prague: Matfyzpress. (CD-ROM) Ten Hacken 2006 = Pius Ten Hacken (Ed.) 2006: Terminology, Computing and Translation. Tübingen: Gunter Narr Verlag. Ten Hacken 2009 = Pius Ten Hacken 2009: What is a dictionary. A view from Chomskyan Linguistics. International Journal of Lexicography 22:4, 399-421. Teubert 2005 = Wolfgang Teubert 2005: Language as an economic factor: The importance of terminology. In Geoff Barnbrook, Pernilla Danielsson and Michaela Mahlberg (Eds.): Meaningful Texts. The Extraction of Semantic Information from Monolingual and Multilingual Corpora. London/New York: Continuum, 96-106. Thumb 2004 = Jenny, Thumb 2004: Dictionary Look-up Strategies and the Bilingualised Learner's Dictionary. Tübingen: Max Niemeyer. Toffler 1970 = Alvin Toffler 1970: Future Shock. New York: Random House. Toft 1998 = Bertha Toft 1998: Terminologi og leksikografi: nye synsvinkler på fagene. LexicoNordica 5, 91–105. Tomaszczyk 1989 = Jerzy Tomaszczyk 1989: L1–L2 technical translation and dictionaries. In E. Pöhl and M. Snell-Hornby (Eds.): Translation and Lexicography. Papers Read at the Euralex Colloquium Held at Innsbruck, 2–5 July 1987. Amsterdam: John Benjamins, 177-186. Tono 2001 = Yukio Tono 2001: Research on Dictionary Use in the Context of Foreign Language Learning. Focus on Reading Comprehension. Tübingen: Max Niemeyer. Tono 2010 = Yukio Tono 2010: A critical review of the theory of lexicographical functions. Lexicon 40, 1-16. Tono 2011 = Yukio Tono 2011: Application of eye-tracking in EFL learners’ dictionary look-up process research. International Journal of Lexicography 24(1), 124-153.

Other Literature | 265 Trap-Jensen 2010 = Lars Trap-Jensen 2010: One, Two, Many: Customization and User Profiles in Internet Dictionaries. In Anne Dykstra and Tanneke Schoonheim (Eds.): Proceedings of the XIV Euralex International Congress, Leeuwarden, 6-10 July 2010. Ljouwert: Fryske Akademy, Afuk, 1133‒1143. Urdang 2000 = Laurence Urdang 2000: [Review of] Hartmann and James (1998). International Journal of Lexicography 13(1), 35-42. Van der Merwe 2008 = Michele F. Van der Merwe 2008: Wine and words: A trilingual wine dictionary for South Africa. Lexikos 18, 337-348. Van der Vliet 2006 = Hennie Van der Vliet 2006: Combinatorics for special purposes. In P. Ten Hacken (Ed.): Terminology, computing and translation. Tübingen: Gunter Narr Verlag, 57-72. Van Doren 1962 = Charles Van Doren 1962: The idea of an encyclopedia. American Behavioral Scientist 6:1, 23-26. Varantola 1998 = Krista Varantola 1998: Translators and their Use of Dictionaries. In B.T. Sue Atkins (Ed.): Using Dictionaries: Studies of Dictionary Use by Language Learners and Translators. Tübingen: Max Niemeyer, 179-192. Verlinde/Leroyer/Binon 2010 = Serge Verlinde, Patrick Leroyer and Jean Binon 2010: Search and you will find. From stand-alone lexicographic tools to user driven task and problem-oriented multifunctional leximats. International Journal of Lexicography 23(1), 1-17. Vicente 2007 = Francisco de Vicente 2007: IATE: La base de datos terminológicos de las instituciones europeas. II Jornadas de Terminología y Traducción Institucional, Madrid, 25 de mayo de 2007. Available at: [] Vintar 2010 = Špela Vintar 2010: Bilingual term recognition revisited. The bag-of-equivalents term alignment approach and its evaluation. Terminology 16(2), 141-158. Weber 1996 = Nico Weber (Ed.) 1996: Theorie der Semantik und Theorie der Lexikographie. Tübingen: Niemeyer. Welker 2006 [2010] = Herbert Andreas Welker 2006: O Uso de dicionários: Panorama geral das pesquisas empíricas. Brasília: Thesaurus. Also in English: Herbert Andreas Welker [2010]): Dictionary Use: A General Survey of Empirical Studies. Brasília: Author’s Edition. Wenger 1998 = Etienne Wenger 1998. Communities of Practice: Learning, Meaning and Identity. Cambridge: Cambridge University Press. Whewell 1847 = William Whewell 1847: The Philosophy of the Inductive Sciences, founded upon their history. London: John W. Parker, West Strand. Wiegand 1977 = Herbert Ernst Wiegand 1977: Nachdenken über Wörterbücher. Aktuelle Probleme. In Günther Drosdowski, Helmut Henne, Herbert E. Wiegand: Nachdenken über Wörterbücher. Mannheim, Vienna, Zürich: Bibliographisches Institut, 51-102. Wiegand 1983 = Herbert Ernst Wiegand 1983: Was ist eigenlich ein Lemma? Ein Beitrag zur Theorie der lexikographischen Sprachbeschreibung. In Herbert Ernst Wiegand (Eds.): Studien zur neuhochdeutschen Lexicographie III. Hildesheim, New York: Olms, 401-474. Wiegand 1984 = Herbert Ernst Wiegand 1984: On the structure and contents of a general theory of lexicography. In R.R.K. Hartmann (Ed.), LEXeter ‘83 Proceedings. Tübingen: Max Niemeyer, 1330. Wiegand 1986 = Herbert Ernst Wiegand 1986: Von der Normativität deskriptiver Wörterbücher. Zugleich ein Versuch zur Unterscheidung von Normen und Regeln. In H. Henne (Ed.): Sprachnormen in der Diskussion. Beitrage vorgelegt van Sprachfreunden. Berlin/New York: De Gruyter, 72-101. Wiegand 1987 = Herbert Ernst Wiegand 1987: Zur handlungstheoretischen Grundlegung der Wörterbenutzungsforschung. Lexicographica 3, 178-227.

266 | References Wiegand 1988 = Herbert Ernst Wiegand 1988: Was eigentlich ist Fachlexikographie? Mit Hinweisen zum Verhältnis von sprachlichem und enzyklopädischem Wissen. In Horst Haider Munske, Peter von Polenz, Oskar Reichmann and Reiner Hildebrandt (Eds.): Deutscher Wortschatz. Lexikologische Studien. Ludwig Erich Schmitt zum 80. Geburtstag von seinen Marburger Schülern. Berlin, New York: Walter de Gruyter, 729-790. Wiegand 1989 = Herbert Ernst Wiegand 1989: Der gegenwärtige Status des Lexikographie und ihr Verhältnis zu anderen Disziplinen. In Franz Josef Hausmann, Oskar Reichmann, Herbert Ernst Wiegand and Ladislav Zgusta (Eds.): Wörterbücher, Dictionaries, Dictionnaires. An International Encyclopedia of Lexicography, First Volume. Berlin, New York: Walter de Gruyter, 246280. Wiegand 1994 = Herbert Ernst Wiegand 1994: Zur Unterscheidung von semantischen und enzyklopädischen Daten in Fachwörterbüchern. In Henning Bergenholtz and Burkhard Schaeder (Eds.): Fachlexikographie. Fachwissen und seine Repräsentation in Wörterbüchern. Tübingen: Günter Narr Verlag, 103-132. Wiegand 1998 = Herbert Ernst Wiegand 1998: Wörterbuchforschung. Untersuchungen zur Wörterbuchbenutzung, zur Theorie, Geschichte, Kritik und Automatisierung der Lexikographie. 1. Teilband. Berlin, New York: de Gruyter. Wiegand 2000 = Herbert Ernst Wiegand 2000: Wissen, Wissenrepräsentationen und Printwörterbücher. In Ulrich Heid, Stefan Evert, Egbert Lehmann and Christian Rohrer (Eds.): Proceedings of the Ninth EURALEX International Congress, EURALEX 2000, Stuttgart, Germany, August 8th– 12th, 2000. Stuttgart: Institut für Maschinelle Sprachverarbeitung, Universität Stuttgart, 15-38. Wiegand 2002 = Herbert Ernst Wiegand 2002: Wissen in der Sprachlexikografie. Ein Plädoyer für einige immer noch notwendige Differenzierungen. In Kennosuke Ezawa, Wilfried Kürschner, Karl H. Rensch and Manfred Ringmacher (Eds.): Linguistik jenseits des Strukturalismus. Akten des II. Ost-West-Kolloquiums Berlin 1998. Tübingen: Gunter Narr Verlag, 265-281. Wiegand 2004 = Herbert Ernst Wiegand 2004: Über die Unterschiede von Fachlexikographie und Terminographie. Am Beispiel des Wörterbuchs zur Lexikographie und Wörterbuchforschung. Germanistische Linguistik 178, 135-152. Wiegand 2013 = Herbert Ernst Wiegand 2013: Lexikographie und Angewandte Linguistik. Zeitschrift für angewandte Linguistik, 13-39. Wiegad/Fuentes Morán 2009 = Herbert Ernst Wiegand and María Teresa Fuentes Morán 2009: Estructuras lexicográficas. Aspectos centrales de una teoría de la forma del diccionario. Granada: Tragacanto. Wierzbiecka 1985 = Anna Wierzbiecka 1985: Lexicography and Conceptual Analysis. Ann Arbor: Karoma. Wingate 2002 = Ursula Wingate 2002: The Effectiveness of Different Learners Dictionaries: An Investigation into the Use of Dictionaries for Reading Comprehension by Intermediate Learners of German. Tübingen: Max Niemeyer. Wright 2002 = Sue Ellen Wright 2007: Coping with indeterminacy. Terminology and knowledge representation resources in digital environments. In Edem Bassey Antia (Ed.): Indeterminacy in Terminology and LSP: Studies in Honour of Heribert Picht. Amsterdam/Philadelphia: John Benjamins, 157-179. Wright/Budin 1997 = Sue Ellen Wright and Gerhard Budin (Eds.) 1997: Handbook of Terminology Management. Volume 1: Basic Aspects of Terminology Management. Amsterdam/Philadelphia: John Benjamins. Wright/Budin 2001 = Sue Ellen Wright and Gerhard Budin (Eds.) 2001: Handbook of Terminology Management. Volume 2: Application-oriented Terminology Management. Amsterdam/Philadelphia: John Benjamins.

Other Literature | 267 Wüster 1979 = Eugene Wüster 1979: Introduction to the General Theory of Terminology and Terminological Lexicography (English translation of Einführung in die allgemeine Terminologielehre und terminologische Lexikographie). Wien: Springer. Ye et al. 2012 = Zheng Ye, Jimmy Xiiangji, Ben He and Hongfei Lin 2012: Mining a Multilingual Association Dictionary from Wikipedia for Cross-language Information Retrieval. Journal of the American Society for Information Science and Technology 63(12), 2474-2487. Yong/Peng 2008 = Heming Yong and Jing Peng 2008: Chinese Lexicography. A History from 1046 BC to AD 1911. Oxford: Oxford University Press. Zgusta 1971 = Ladislav Zgusta 1971: Manual of Lexicography. The Hague: Mouton. Zgusta 1992 = Ladislav Zgusta 1992: The Czech-Chinese dictionary and the theory of lexicography. International Journal of Lexicography 8:2, 85-128. Zikmund 1997 = William G. Zikmund 1997: Business Research Methods. Fort Worth: The Dryden Press/London: Harcourt Brace College Publishers. Zöfgen 1987 = Ekkehard Zöfgen 1987: Kollokation – Kontextualisierung – (Beleg-)Satz. Anmerkungen zur Theorie und Praxis des lexikographischen Beispiels. In Alberto Barrera-Vidal, Hartmut Kleineidam and Manfred Raupach (Eds.): Französische Sprachlehre und bon usage. Festschrift für Hans Wilhelm Klein zum 75. Geburststag. München: Hueber, 219-238. Zöfgen 1994 = Ekkehard Zöfgen 1994: Lernerwörterbuch in Theorie und Praxis. Ein Beitrag zur Metalexikographie mit besonderer Berücksichtung des Französischen. Tübingen: Niemeyer.

Index A Glossary of Archival and Records Terminology 15 Accounting Dictionaries 61, 74, 93, 192f., 195ff., 200, 203ff., 210ff., 221, 223ff., 228, 230, 239, 243 Al-Ajmi, H. 132 Almind, R. 59, 61, 65, 155, 192ff., 233, 239 Andersen 53 Arribas-Baño, A. 40 Atkins, S. 24, 37, 40, 117, 155 Auger, A. 126 Barrière, C. 126 Bejoint, H. 24, 34f., 37ff. Bergenholtz, H. 5, 8, 10f., 28f., 32, 37, 40f., 46f., 59, 61, 64f., 74, 77, 79ff., 83, 91, 93, 102, 130, 141ff., 148, 155, 192f., 195, 218, 223, 233, 239, 241 Bergenholtz, I. 47, 64, 141ff. Berners-Lee, T. 177 Besomi, D. 9, 11f., 20 Bevilacqua, C.R. 28 Bilteknisk Ordbog 15 Binon, J. 32 Boccardo 14 Bocorny Finatto, M. J. 28 Bogaards, P. 37, 40, 42 Borlund, P. 49 Bothma, TJD 31f., 49, 60, 92f., 98f., 103, 204 Boulanger, JC 112 Bowker, L. 122, 127, 178 Boyatt, RC. 178 Boyatt, Russel C. 126 Brooks, T. 36 Buckland, M. 36, 58 Budin, G. 105ff., 112 Buhr, M. 35, 37, 40 Burkhanov, I. 27 Business English-Spanish Glossary 129f., 149ff. Cabré, T. 45, 56f., 105, 110ff., 118f., 121, 179ff. Cambridge Business English Dictionary 129f., 145, 147 Caruso, V. 15, 88 Centre for Lexicography 2, 16, 22, 102, 192, 198, 241, 243 CercaTerm 129f., 156, 162ff., 169, 189 Chamber 8

Chan, AY. 130, 132 Clark, M. 121f., 125 Concise Oxford Dictionary of Linguistics 26 Cosijn, E. 49 Cowie, A.P. 5, 25 Crystal, D. 26 Cyclopædia 8 d’Alembert, Jean le Rond 1, 8 Daille, B. 123 Dansk Glossarium 46 DANTERM 134 De Baer, P. 114, 127, 173 De Meo, A. 88 De Schryver, G-M. 59, 91 Den Danske Netordbog 82, 102 Diccionario de la Lengua Española 14 Diccionario de lexicografía práctica 4 Diccionario Enciclopédico de Ingeniería Genética 74 Diccionario Inglés-Español de Contabilidad 232, 234ff. Diccionarios de Contabilidad 16, 74f., 92f., 192f., 195f., 198, 201f., 204f., 207, 210f., 213ff., 219ff., 228, 230, 234, 236ff., 243 Dictionary of Business and Management 129f., 135ff. Dictionary of Lexicography 5f., 131 Dictionary of Linguistics and Phonetics 26 Dictionary of the English Language 1 Dictionnaire analytique d’économie politique 14, 20 Dictionnaire universel de commerce 14, 25 Diderot 1, 8, 11 Disco – European Dictionary of Skills and Competences Véase Dizionario della economia politica e del commercio 14 Dolezal, FT. 132 Dorosevskij, V. 36 Dubois, C. 28 Duda 37, 40 Duvå, G. 69f. EcoLexicon 16, 109, 119, 121, 129f., 177f., 185ff. Encyclopædia Britannica 8 Encyclopedic Dictionary of Gene Technology 74

270 | Index Encyclopédie 1, 8, 11 English Dictionary of Arts and Sciences 1, 60 European Dictionary of Skills and Competences 129, 169, 172 Faber, P. 56, 116ff., 121, 134, 185 Fachwörterbuch zur Rentenversicherung 74 Fata, I. 74 Felber, H. 107f. Feliú, J. 179f., 185 FF POIROT 114ff. Fillmore, C. 117 Foo, J. 123 FrameNet 117, 153 Frawley, W. 26 Fuertes-Olivera, Pedro A. 32, 40, 59ff., 65, 74, 76, 86, 91, 105, 120, 166, 192, 195, 204, 218, 221, 230, 233, 240 Ganilh 14, 20 Gaudin, F. 104f., 112f. Genoma 109, 112, 121, 129f., 177ff., 181f., 184f., 188ff. Glossary of FAO Database and Information System 129, 151 Glossary of Mortgage and Home Equity Terms 15, 129f., 143f. Glossary of Photographic Terms 15 Gouws, RHG 32, 40, 42f., 77f., 81, 130, 132 Granger, S. 91 Gruber, T. 177 Haensch 24 Haiman, J. 9f. Hanks, P. 59 Hansen 53 Harris, J. 1, 14, 60f. Harris, R. 61 Hartmann, RKK 5ff., 23, 27, 131 Haß, U. 91 Hatherall, G. 53 Hausmann, FJ. 37, 44, 48, 56 Hearst, MA. 126 Heid, U. 32, 60, 99, 173 Hendler, J. 177 Henriksen 8 Hjørland 58 Hoare 6 Høier Leth, J. 46 Hospedales, M. 179 Humbley, J. 27f., 56, 105, 116, 118 Hutton, C. 61

IATE 156ff., 162, 169, 189 Illustrated Glossary for Transport Statistics 14 Illustrated glossary on stone deterioration patterns 14 Ingwersen, P. 49 Inter-Active Terminology for Europe 15, 129, 157 International Centre for Lexicography 2, 16, 192, 243 International encyclopaedia of social sciences 12 International Encyclopedia of Lexicography 6 ISO 12620 108, 156 James, G. 5ff., 27 Johnsen, M. 130, 195 Johnson, M. 113 Johnson, S. 1, 14f. Joy, Mike S. 126, 178 Kageura, K. 122, 128 Kalverkämper 5, 8 Kaufmann 28f., 59, 74, 83, 93 Kemisk Ordbog 15 Kerremans, K. 114ff. Kicktionary 129f., 153ff. Kilgarriff, A. 2, 59, 61, 100 Kilian 40 Klaus, G. 35, 37, 40 Kosem, I. 91 Kosen, K. 91 Kromann 40, 44 Kudashev, I. 27ff., 36 Kwary 40, 98 L’Homme, Marie-Claude 122, 128 Lakoff, G. 113 Landau, S. 9, 23f., 34 Lassila, O. 177 Laurén, C. 108 Laursen, AL. 70 Layton 19 León Araúz, P. 56 Leroyer, P. 2, 32 Lew, R. 88, 98, 132 Lewandowski, H. 26 Lexicon Technicum 14, 20, 60 Lexin 94 Linguistisches Wörterbuch 26 Litkowski, KC. 155 López Rodríguez, Clara I. 118, 119, 185 Machlup, F. 36

Index | 271 MacMillan English Dictionary for Advanced Learners 119 Maidahl, L. 70 Manual of Lexicography 9 Manual of Specialised Lexicography 2, 5, 59 Marello, C. 65 Marshman, E. 123, 125, 127 Martin, W. 116f. Martínez de Souza 4 Martínez Motos 36 Matthews, Peter H. 26 McCreary, Don R. 132 McGraw-Hill Dictionary of Scientific and Technical Terms 20 McKay, P. 139 Meier, Hans H. 23 Meier, Hans, H. 23, 25 Menagarishvili, O. 20 Merkel, M. 123 Meyer, I. 123, 125 Mikkelsen, H.K. 19 Miles, A.D. 129 Morán, T. 37 Moreno Ortiz, A. 179 Mourier, L. 192, 195, 205, 230 Mugdan 59 Multilingual Glossary of Technical and Popular Medical Terms in Nine European Languages 129f., 169 Musikordbogen 129, 141f. Myking, J. 108 Navigli, R. 122 Nesi, H. 88 New Palgrave Dictionary of Economics 12, 31, 48, 98, 129, 138 Ney, H. 123 Nielsen, S. 7, 28f., 59, 61, 66, 74, 86, 130, 134, 141, 152, 166, 192ff., 198, 204f., 221, 230, 232f., 239f. Niño Amo, M. 192, 195 Nord, B. 71 Norddahl, B. 102 Nordisk leksikografisk ordbok 4f. Norton, M. 36 Och, FJ. 123 Odgen, C.K. 106, 108 Opitz, K. 6 Osselton, NE. 131 Oxford English Dictionary 26

Oxford History of English Lexicography 5, 25 Pálfi 46 Pantel, P. 122 Paquot, M. 91 Pearson, J. 105, 122 Pedersen 59 Peng, J. 38ff. Pennacchiotti, M. 122 Pérez Hernandez 27, 28, 45, 56 Peters, P. 129, 148f. Philosophisches Wörterbuch 35, 37 Picht, H. 28, 108 Piotrowski, T. 23f., 45 PoCeHRMOM 116 Popper 42 Postlethwayt, M. 1, 14, 24f., 46 Prieto Velasco, Juan A. 118f., 185 Prinsloo, D. 40, 60, 97, 99, 103 PROLIX 116, 173 Reimerink, A. 118f., 178, 185f. Retskrivningsordbogen 81f. Rey, A. 27f., 104 Richards, I.A. 106, 108 Riggs, Fred W. 26, 28, 104 Riiber 40, 44 Roche, C. 27 Rocheteau, J. 123 Rodríguez, A. 48 Rodríguez, M. 179 Rondeau, G. 27, 104 Rosbach 40, 44 Rosch, E. 113 Rossenbeck, K. 10 Rundell, M. 24, 37, 40, 47, 55, 59ff., 98f., 132, 155 Saracevic, T. 49 Savary des Bruslons 14, 25 Scerba 37, 39f., 44, 56 Schaeder 59 Schmidt, T. 129, 135, 153f. Schmitz, KD. 106f. Schmitz, U. 91, 104, 156 Scholfield, P. 132 Scott, M. 122 Sinclair, J. 105, 122, 190 Sorokoletov 21, 23, 36, 40 South African Trilingual Wine Industry Dictionary 83 Svensén, Bo 21, 28, 40

272 | Index Swales, J. 149 Swanepoel, P. 130 Tarp 5, 11, 21f., 28f., 31f., 35, 37f., 40f., 44, 46f., 49, 52f., 55f., 59f., 64f., 68f., 74, 77f., 81, 83, 89, 99, 130, 148, 162, 192, 204, 241f. Taylor, A. 130, 132 Temmerman, R. 105, 112ff., 116, 119, 178 Ten Hacken, P. 24, 45, 117 TermFinder 129f., 148f. Teubert, W. 145 The American farmer’s encyclopedia and dictionary of rural affairs 14, 20 Thumb, J. 88 Toffler, A. 63 Tomaszczyk, J. 70f., 95 Tono, Y. 40, 88, 132f. Universal Dictionary of Trade and Commerce 1, 14, 24f., 46 UNTERM 129, 152, 156, 158, 166ff., 189 Urdang, L. 131 Van der Merwe, M. 83 Van der Vliet, H. 117f. Van Loren 61

Varantola, K. 71 Verlinde, S. 32 Vintar 100, 123ff. Vrang, V. 65, 155 Weber 40 Welker, Herbert A. 47, 53, 88, 130 Wenger, E. 149 Whewell, W. 27, 104, 106 Whitelaw 8 Wiegand, H.E. 5, 9f., 21, 28, 36f., 39f., 42, 44f., 58, 80, 130 Wierzbiecka, A. 24 Wiktionary 98 Wingate, U. 88 WordNet 122 Wright, Sue Ellen 106, 125, 178 Wüster 27, 104ff., 112, 120f. Ye, Zheng 126 Yong, H. 38f., 40 Yongle Dadian 31 Zgusta, L. 9, 40 Zikmund, William G. 53, 89 Zöfgen, E. 37, 40