• Author / Uploaded
• Claus

Citation preview

TEXT TYPOLOGY AND TRANSLATION

BENJAMINS TRANSLATION LIBRARY The Benjamins Translation Library aims to stimulate academic research and training in translation studies, lexicography and terminology. The Library provides a forum for a variety of approaches (which may sometimes be conflicting) in a historical, theoretical, applied and pedagogical context. The Library includes scholarly works, reference books, post-graduate text books and readers in the English language. ADVISORY BOARD Jens Allwood (Linguistics, University of Gothenburg) Morton Benson (Department of Slavic, University of Pennsylvania) Marilyn Gaddis Rose (CRIT, Binghamton University) Yves Gambier (Centre for Translation and Interpreting, Turku University) Daniel Gile (Université Lumière Lyon 2 and ISIT, Paris) Ulrich Heid (Computational Linguistics, University of Stuttgart) Eva Hung (Chinese University of Hong Kong) W. John Hutchins (Library, University of East Anglia) Werner Koller (Department of Germanic, Bergen University) José Lambert (Catholic University of Louvain) Willy Martin (Lexicography, Free University of Amsterdam) Alan Melby (Linguistics, Brigham Young University) Makoto Nagao (Electrical Engineering, Kyoto University) Roda Roberts (School of Translation and Interpreting, University of Ottawa) Juan C. Sager (Linguistics, Terminology, UMIST, Manchester) Maria Julia Sainz (Law School, Universidad de la República, Montevideo) Klaus Schubert (Technical Translation, Fachhochschule Flensburg) Mary Snell-Hornby (School of Translation & Interpreting, University of Vienna) Sonja Tirkkonen-Condit (Savonlinna School of Translation Studies, Univ. of Joensuu) Gideon Toury (M. Bernstein Chair of Translation Theory, Tel Aviv University) Wolfram Wilss (Linguistics, Translation and Interpreting, University of Saarland) Judith Woodsworth (FIT Committee for the History of Translation, Concordia University, Montreal) Sue Ellen Wright (Applied Linguistics, Kent State University)

Volume 26

Anna Trosborg (ed.) Text Typology and Translation

TEXT TYPOLOGY AND TRANSLATION

Edited by

ANNA TROSBORG The Aarhus School of Business

JOHN BENJAMINS PUBLISHING COMPANY AMSTERDAM/PHILADELPHIA

The paper used in this publication meets the minimum requirements of American National Standard for Information Sciences — Permanence of Paper for Printed Library Materials, ANSI Z39.48-1984.

Library of Congress Cataloging-in-Publication Data Text typology and translation / edited by Anna Trosborg. p. cm. -- (Benjamins translation library, ISSN 0929-7316 ; v. 26) Includes bibliographical references and index. 1. Translating and interpreting. 2. Discourse analysis. 3. Typology (Linguistics) I. Trosborg, Anna, 1937- . II. Series. P306.2.T49 1997 418' .02--dc21 97-38869 ISBN 90 272 1629 0 (Eur.) / 1-55619-710-1 (US) (alk. paper) CIP © Copyright 1997 - John Benjamins B.V. No part of this book may be reproduced in any form, by print, photoprint, microfilm, or any other means, without written permission from the publisher. John Benjamins Publishing Co. • P.O.Box 75577 • 1070 AN Amsterdam • The Netherlands John Benjamins North America • P.O.Box 27519 • Philadelphia PA 19118-0519 • USA

Contents Introduction

vii

Part I: Methodology Anna Trosborg Text Typology: Register, Genre and Text Type

3

J. C Sager Text Types and Translation

25

Christiane Nord A Functional Typology of Translations

43

Paul Kussmaul Text-Type Conventions and Translating: Some Methodological Issues

67

Approaches to Literary Genres Susan Bassnett Text Types and Power Relations

87

Viggo Hjørnager Description and Criticism: Some Approaches to the Translations of Hans Christian Andersen 99

Part II: Domain- and Genre-Specific Texts Christina Schäffner Strategies of Translating Political Texts

119

Anna Trosborg Translating Hybrid Political Texts

145

vi

CONTENTS

Morten Pilegaard Translation of Medical Research Articles

159

Jens Hare Hansen Translation of Technical Brochures

185

Vijay K. Bhatia Translating Legal Genres

203

Part III: Terminology and Lexicon Margaret Rogers Synonymy and Equivalence in Special-Language Texts

217

Eugenio Picchi and Carol Peters Reference Corpora and Lexicons for Translators and Translation Studies 247 Texts Types and Media Constraints Mary Snell-Hornby Written to be Spoken: The Audio-Medial Text in Translation

277

Thomas Herbst Dubbing and the Dubbed Text - Style and Cohesion

291

Henrik Gottlieb Quality Revisited: The Rendering of English Idioms in Danish Television Subtitles vs. Printed Translations 309

Index

339

Introduction Within translation theory and practice, there has been a shift from an overall concern with equivalence between source and target texts to a recognition of the need for adaptation to the target situation and purpose (cf. the skopos theory). In most cases, equivalence can hardly be obtained in translation across cultures and languages, and it may not even be a desirable goal. Therefore, other criteria for successful translation are needed. Translators have long been aware of the need for categorization in translation. Since Catford (1965: 83), the desire to have a framework of categories for the classification of varieties or "sub-languages" within a language has been acknowledged. Genre analysis has been concerned with establishing characteristics of particular types of text, but whereas the concepts of genre have a long tradition in literary studies, interest in the analysis of non-literary genres is of more recent date (e.g. Swales 1990, Bhatia 1993). It is only throughout the last decade or so that genre analysis has become popular in a variety of fields, e.g. rhetoric, discourse analysis, cognitive science, computational linguistics, business communication, a.o. Apart from literary studies, research on genres has been slow to penetrate the field of translation studies, and only little work has been done in this area. An urgent need for dictionaries in a universe with an intensively growing intercultural communication and a subsequent need for translation may have kept research at the level of register analysis with the main focus on the writing of dictionaries. However, recent developments in applied genre theory and its applications to Language for Special Purposes (LSP) (Swales 1990, Bhatia 1993) seem to have interesting implications for the theory and practice of translation and interpretation of professional discourse. Being prototypical rhetors, LSP specialists such as salesmen, lawyers, politicians, and even editorialists are far more constrained by situational factors than are poets, novelists, dramatists, and the like. Constraints and conventions of specific types of text deserve further attention, just as the value for translation studies needs to be investigated. The essays in this book explore the possibilities and the problems of generic scholarship with special reference to the translation of a wide range of professional discourse. The book attempts to demonstrate the value of text typology

viii

INTRODUCTION

for translation purposes, emphasizing the importance of genre analysis, analysis of communicative functions and text types in a broad sense as a means of studying spoken and written discourse. Sonnets, sagas, fairy tales, novels and feature films, sermons, political speeches, international treaties, instruction leaflets, business letters, academic lectures, academic articles, medical research articles, technical brochures and legal documents are but some of the texts treated in this volume. It is argued that text typology involving genre analysis can help the translator develop strategies that facilitate his/her work and provide awareness of various options as well as constraints. In this book, text type is used in a broad sense to refer to any distinct type of text and the notion includes genre. The central question of the book is: In what ways are translations affected by text types? The two main areas of investigation are: (A.) What are the advantages of focusing on text types when trying to understand the process of translation? How do translators tackle different text types in their daily practice? Is the translator specialization not only conditioned by subject matter, but also by text type? (B.) To what extent and in what areas are text types identical across languages? What similarities and dissimilarities can be observed in text types of original and translated texts? As the essays to follow clearly show, this is a wide-ranging task with many aspects to be explored. Texts may be classified in multiple ways. As put forward by Simons and Aghazarian (1986: 11), one might profitably view Martin Luther King's letter from Birmingham Jail as "black rhetoric, as southern rhetoric, as protest rhetoric, as ministerial, religious rhetoric, as a species of public letter-writing, and as an apologia of sorts". Likewise, there is no one way of grouping the essays in this volume, and I am aware that any single classification is bound to deflect attention from some of the distinctive characteristics of a particular text type. The essays in Part I of the book highlight methodological possibilities. For all the interest in genre analysis, there is still some bewilderment as to the extension of the term, just as the criteria to be used for classification and for identification of genre membership are still controversial issues which are the object of much ongoing discussion. For some systemicists, genre is sometimes used in a broad sense to refer to register variation, such as journalistic language, legal language, scientific discourse, etc. Other scholars mix genres with rhetorical types, naming expositions and argumentative texts as genres. How does the classification of genres work? How do genres relate to register and text types? How is the interrelation between the purpose of the communication and the rhetorical strategies determining the text type(s) employed to achieve the intended communicative goal? The essay by Trosborg sets out to explore these issues and presents a differentiated picture of the notions in question.

TEXT TYPOLOGY AND TRANSLATION

IX

What happens to a text when it is translated? And how about translation specific text types, are there text types which exist only as a product of translation? Traditional translation theory speaks of equivalence at the level of the content of the source and target text only. Sager's essay widens the concept of equivalence in translation by examining new translation strategies which change the content of a message, its intention and/or both. Viewed within the diverse communicative environments in which texts occur in practice, we are dealing with messages addressing primary readers, i.e. readers directly addressed by the originator, and messages addressing secondary readers, i.e. readers not addressed directly. Many translations are of the latter kind. However, in order to be communicatively effective, translations have to be modified at the level of the content or the intention so that the recipients of translated messages can be addressed as primary readers wherever appropriate. A new typology of translations, including the existence of translation-specific text types, is established based on these criteria. Communicative functions relate to the purpose of discourse, an issue which most authors in this volume are concerned with one way or another. These functions are the particular focus of attention for Nord in her framework classifying texts for translation purposes. After explaining her functionality-plusloyalty model of translation, which combines the criterion of target-text functionality with that of the translator's loyalty towards her/his partners in the translational interaction (source-text sender, client, target-text receivers), the author presents a functional typology of translations, which then serves as a basis for an application of the model to a sample text. She argues that the translator's basic option for either a documentary or an instrumental translation may serve as a guideline for all subsequent decisions as to whether pragmatic aspects, norms and conventions, etc. have to be reproduced from the source text or adapted to target-culture standards in the translation process. In translator training, the translation typology can be complemented by a typology of pragmatic, intercultural, interlingual and text-specific translation problems. Whereas Sager is concerned mainly with a typology of translations as product, Nord's classification focuses on the translation as process. Kussmaul begins by clarifying some basic concepts of text typology, namely convention, culture and text type. When people produce specific text genres, such as business letters, they can be expected to conform to specific regularities and rules. If they do not, communication may turn out to be difficult or even to break down. A further problem lies in the fact that conventions may differ not only between genres but also between 'identical' genres in different cultures. He observes two overriding aspects: Text-type conventions are interrelated with speech-act rules and with situational dimensions. A change of

X

INTRODUCTION

dimension may result in a change of text type. Kussmaul goes on to discuss macrostructures, mainly for the text type academic writing and looks at the situational dimensions they reflect. Then there is a discussion of various microstructures such as metacommunicative utterances, hedging and directives, again mainly with reference to academic writing but also to some other text types such as instruction leaflets, business letters and regulations. Here again the overall aspect is: Can situational dimensions and specific rules for speech acts be detected? Finally, there is a detailed discussion of the translation of macrostructures and microstructures from a functional point of view. Being concerned with aspects of text type, culture and convention in general, though with reference to specific genres for exemplification, Kussmaul's essay is a forerunner of a number of essays in Part II concerned with the translation problems in relation to particular genres in a cross-cultural perspective. Although genre has a prominent position in literary studies, the traditions of the old genres of the classics: ballads, odes, sonnets, tragedies, and comedies have been replaced by literary practice in the twentieth century, which seems to throw away convention. Authors break new ground. Howewer, a new genre is always the transformation of one or several old genres: by inversion, by replacement, by combination. A work that rebels against genre equally relies on the reader's recognition of the conventions being rejected. An appreciation of genre is therefore a necessary though not a sufficient condition for an appreciation of literature. But what happens to literary genres in translation? Why is it that some types of text resist translation while others come to enjoy immense popularity in the target culture? Susan Bassnett is concerned with text type and power. What is the reason that some genres enter a literary system through translation and become accepted, while others are rejected? Susan Bassnett examines the problem of the status of literary texts in the source and target systems. She argues that regardless of the genre, the needs and conventions of the target system will condition what ultimately happens both in the actual process of translation itself and in the fate of the translated texts. Using two examples, the Petrarchan sonnet and the Norse saga, which were received quite differently into the English literary system at different moments in time, she points out the difficulty of generalizing about patterns of literary translation. Both these text types may be seen as potentially innovative for English literature, but only the sonnet made the transition and was accepted into the English system, ultimately to become a canonical form. Although there are obviously many factors to be considered in an analysis of how and why a new text type enters a literary system, it does seem clear that the determining factors reside in the target system, and not in the source system, regardless of the importance of the text, author or text type in

TEXT TYPOLOGY AND TRANSLATION

XI

that system. The needs and constraints imposed by the receiving literature are paramount. Literary texts generally defy genre classification according to communicative purpose. Translations of literary texts must inevitably focus on form. The essay by Viggo Hjørnager discusses the relation between the description and criticism of translation, using the author's work on English translations of Hans Christian Andersen for purposes of illustration. The methods described include a discussion of translations as target language literature, a rank-bound linguistic comparison of an original and its translation(s), a study of abbreviated translations, and a contrastive study of vocabulary, phraseology, and idioms. A change of genre (into children's literature) is discussed within the manipulation approach to translation. The article concludes that, while any decent criticism must be based on a description of the translated text, such a description is only interesting for the scholar if it is followed by criticism. Cross-cultural aspects of text typology are of particular importance in the present context. While cross-cultural variation of spoken language has become a well-established area of discourse study (e.g. Wierzbicka 1985, 1991, BlumKulka et al. 1989), very little has been published in the case of written genres. Cultural taboos in the use of numbers, colours and shapes are well-known, high-context cultures and low-context cultures may differ in their approach, etc. It is also well-known that a message can be totally distorted if the implicit culture-specific information or a culture-specific word or an allusion is not grasped by the translator, but to what extent and in what ways are text types affected by culture? Cultural variation and local socio-cultural constraints seem to play a significant role in a number of genres representing specialized texts, as for example business letters, job applications, and some legal genres, while other genres are of the conformative type, in the sense that they are universally conventionalized and rarely show any variation in cross-cultural realization patterns. To gain more knowledge in this area is crucial to the translator. The essays in Part II constitute our probe of problems and possibilities of generic approaches to translation of professional genres in a practical perspective. The emphasis is on a combination of essential linguistic insights and sociocognitive and cultural explanations with a concern for the ways in which genre is embedded in sociologically determined communicative activities. The focus of attention is the specific linguistic realisations of genre conventions, which may be culture-specific rather than universal; the aim is a grounded description of language in use bringing in useful explanations of why a particular type of conventional realization of meaning is considered appropriate to a particular institutionalized socio-cultural setting. Of particular interest is: What are the

xii

INTRODUCTION

characteristics of particular text types? To what extent and in what areas are text types identical/different across languages? The essays in Part II are grouped according to field. This choice has been made for the convenience of organizing the material, in the knowledge that this is by no means a sufficient criterion for the classification of texts and that a given classification, as a metatextual construction, will always be one among many possible classifications of the text(s) in question. In addition to work on political texts, there is an essay on medical texts highlighting, in particular, features of the medical research, an essay on technical brochures, and one on legal documents. Schäffner offers a delimitation of political discourse into specific genres. With the growing internationalization of politics, translation too, becomes more and more important. Political genres fulfil different political functions, and each of them displays characteristic contextual, text-typological, pragmatic, etc. features. Therefore, they call for different translation strategies. Using multilingual treaties and documents, speeches and statements by politicians as examples, some translation problems encountered in these texts (German and English) and the strategies employed for their solution in authentic translations (target texts) are discussed. These translation problems concern formal constraints, text-typological conventions, terminological and syntactic conformity in a sequence of intertextually related documents, as well as ideological aspects of political concepts and culture-specific knowledge. The development of international standardized genres discussed by Schäffner points to hybridisation as the product of a compromise to achieve an internationally accepted standard. This aspect is further developed by Trosborg, who examines documents of the European Union (EU) within a sociocognitive approach to genre. In contrast to political speeches and statements by politicians, which reflect culture-specific conditions of their production, EU documents are a product of a multicultural discourse community; they come into existence as translation-specific hybrid text types, which are characterized by specific features (syntactical, lexical and textual) labelled "Eurojargon". There are no proper source texts and the translations are documents addressing primary readers. As these texts are no doubt marked as products of translation, the status of both the ST and the SL community would have to be reconsidered. In a world where the dissemination of knowledge and exchange of information is gaining ever greater pace and research has become truly international, medical translation is emerging as a rapidly growing field of specialized translation that presents a host of theoretical as well as practical challenges. The essay by Pilegaard presents a selective review of recent contrastive studies of the translation of medical research articles, and it suggests a number of translation strategies of particular relevance for the medical article, for example

TEXT TYPOLOGY AND TRANSLATION

xiii

adaptation to readership and attention to text focus, cultural presuppositions, and conversational rules at work in the TL, arguing that the principles and rules for the production of medical texts are highly genre-specific. For example, where the case report is mainly descriptive and expository, it is the nature of the research article to be more "experimental" and rhetorical: it raises questions and verifies hypotheses, gives answers and explains cause-effect relationships. In translation, pragmatic modifications must be guided by genre and culturespecific conventions, and a balance between the conventions of the international research paper and the idiosyncrasies of the ST author must be struck in close collaboration between the translator and the author. Communicative functions in a cross-cultural perspective are the focus of attention in Hare Hansen's article on technical brochures. Questioning the relevance of equivalence, he pays specific attention to culture specific problems in adjusting to the German market. His findings show that a semantic translation from German to Danish is too formal and too rigid and most often ineffective in achieving the intended sales purpose of technical brochures. Instead, the focus should be on pragmatic aspects. A shift from representative to directive/ persuasive function is required when translating from German into Danish. Furthermore, cultural information irrelevant to target culture addressees must be left out. In order to make the desired transformation, the translator must have in depth knowledge of text-type conventions in a cross-cultural perspective as well as marketing qualifications. The purpose is to further the sale of the products and the translation is to be adjusted accordingly. Bhatia stresses the importance of preserving text-type integrity in translation. Although, in the construction, interpretation and use of professional genres, it is often possible to take a rather liberal attitude towards lexicogrammatical innovations, he argues that it is crucial to maintain generic integrity of the intended genre. Taking examples from legislative contexts, his essay takes the position that in the teaching and learning of translation and interpretation, one needs to understand the rationale for the target genre as a prerequisite. This kind of understanding will go a long way in producing a pragmatically more sucessful translation of the original. Adaptation to readership in terms of "easification" is proposed. Part III is divided into two sections. The first set of issues to be addressed concerns terminology and lexicon, the second set is devoted to the influence of medium. The first two essays discuss tools assisting the translator in his/her tasks. Works on terminology and lexicon are crucial to the translator. The broadening perspective of language use described in the methodology section and elaborated on in essays devoted to particular genres in Part II is extended to terminological use.

xiv

INTRODUCTION

Translators deal with language in use. Yet, as Rogers points out, synonymy and equivalence - relations which are crucial to the translator's task - are represented in dictionaries as lexemes, i.e. at the level of system. In her essay, some implications of the use/system dichotomy are investigated in relation to terminology, the specialized vocabulary of an LSP, from a grammatical and a semantic perspective, focusing in particular on the grammatical category of number and the semantic relationship of collocability. Using an English and German bilingual corpus of texts from genetic engineering, it is shown that the relations of synonymy and equivalence, if seen as relations between word forms rather than as relations between lexemes, are more complex than either semasiological or onomasiological models generally predict. The discussion focuses on four synonyms in each language and their role in the formation of single-word and multiword compounds; potential equivalents are then considered. It is concluded that the probabilistic tendencies identified on the basis of the text corpus are not amenable to standardization, often seen as a goal of terminology work. Picchi and Peters present machine-readable corpora and specialized lexicons as the translator's indispensable workstation. The essay is divided into two parts. The first discusses some of the latest trends in the fields of computational lexicography and corpus linguistics that are of direct relevance to the translation of particular genres and studies of the translation process. The second part of the article concentrates on describing the two main components of a prototypical Translator's Workstation designed by the authors: the bilingual lexical database system and the bilingual text management system. Examples of how these systems can be queried and the type of results that can be obtained are given; their potential for all kinds of cross-language applications, e.g. bilingual lexicography, translating and language learning activities, is described. The constraints of mode and medium is the subject in the last three essays in Part III. A type of text may be principally characterized by its use of a particular medium or means of communication. Memos, telegrams, e-mail, etc. are not classified primarily by their communicative goals or rhetorical designs, rather, they stand out by their medium of communication, and dubbing and subtitling are translation methods resulting in translation specific text types. The audio-medial text type is characterized by its complex mode of expression. Translation problems in connection with mediation between the written and the spoken medium are discussed by Snell-Hornby. Her essay revives the term audio-medial introduced by Katharina Reiss in 1971, but it uses it in a strictly limited sense to refer to a text - such as a political speech, an academic lecture or a TV manuscript reporting the news - which has been written to be spoken aloud, and hence perceived audio-medially by its recipient. Rhetoric and speakability play an important part - examples analyzed are part of a

TEXT TYPOLOGY AND TRANSLATION

xv

speech by Churchill and a passage from the Authorized Version of the English Bible. For translation purposes academic lectures, particularly those intended for an international audience with English as a lingua franca and spoken by non-native speakers, provide a rich fund of material: with long and syntactically involved sentences speakability is often achieved in the English translation through devices such as postmodification and strong end-focus. By strategies aiming at reducing redundancy and condensing academic acoustic padding it is possible to produce a more speakable translation, hence facilitating understanding for the audience. The essay by Thomas Herbst is based on a corpus of television series and films dubbed from English into German. It is obvious that translation for dubbing is subject to a number of constraints due to equivalence criteria such as quantitative and qualitative lip sync. It is argued that, as far as their textual structure is concerned, the dubbed films analysed show a number of characteristics. The most important of these are (i) unmotivated style shifts, (ii) a certain tendency to use a formal style with elements typical of the written rather than the spoken language; (iii) a certain "lack" of cohesive ties. It is shown that these features can be traced back not so much to the specific demands on translation for dubbing as such but to a particular approach to translation. These features thus are not necessary features of dubbed text but the outcome of a misguided approach to dubbing. Type of text determined by medium is decisive for choice of translation strategies. In his empirical study on the translation of idioms, Henrik Gottlieb compares the strategies used by professional literary translators and television subtitlers. Although roughly the same strategies are found to be used by groups of translators, due to the different media-specific constraints of literary translation and subtitling, the relative frequences of these strategies vary significantly: books - monosemiotic by nature, communicating only through the abstract medium of writing - show a different distribution of idiom translation strategies than do subtitled television programs, polysemiotic per definition, where the written translation carries only part of the semantic content of the multi-channel discourse. Gottlieb sees idioms as extreme verbal items, yet units whose manifold translation options have major implications for our comprehension of translation strategies in general, and of the differences between books and television in particular. In this volume, contributors are concerned with general (maybe universal) characteristics relevant for a particular genre as well as with identifying crosscultural differences for particular genres. Translators will benefit from insights into the rationale underlying the selection and distribution of linguistic features characteristic of a specific genre, just as it is a growing concern to gain increas-

xvi

INTRODUCTION

ing knowledge of socio-cultural variation in (specific) features pertaining to a genre in a particular culture. Awareness of genre conventions is crucial, both in the understanding of the source text and the creation of the target text. Furthermore, categories of communicative functions and (media-specific) text types are significant to a theory of discourse for translation purposes; being few in number and universal in nature, though with culture-dependant realization patterns, they are more manageable than a whole list of academic disciplines and thus become a workable tool for the translator. The aim is not to establish rigid norms and simplification of text-type conventions; in order to achieve precision and variety in style, the translator must be aware of possible choices as well as obligatory conventions and sociocultural rules. In sum, this volume provides a theoretical (and to some extent historical) overview of major problems and possibilities as well as investigations into a variety of text types with practical suggestions that deserve to be weighed by anyone considering the relation between text typology and translation. It is my hope that this volume may assist the translator in his/her efforts to become a "competent text-aware professional". Anna Trosborg, The Aarhus School of Business, March 1997 References Bhatia, Vijay K. 1993. Analysing Genre. Language use in professional settings. London/New York: Longman. Blum-Kulka, Shoshana, House, Juliane and Kasper, Gabriele (eds). 1989. Cross-Cultural Pragmatics: Requests and apologies. Norwood: Ablex Publishing Corporation. Catford, J.S. 1965. A Linguistic Theory of Translation. Oxford: Oxford University Press. Reiss, K. 1971. Möglichkeiten und Grenzen der Übersetzungskritik. Munich: Max Heuber. Simons, Herbert W. and Aghazarian, Adam A. (eds). 1986. "Introduction". In Herbert W. Simons and Adam A. Aghazarian (eds), Form, Genre and the Study of Political Discourse. South Carolina: South Carolina Press. Swales, John M. 1990. Genre Analysis. English in academic and research settings. Cambridge, UK: Cambridge University Press. Wierzbicka, Anna. 1985. "Different cultures, different languages, different speech acts". Journal of Pragmatics 9: 145-178. Wierzbicka, Anna. 1991. Cross-Cultural Pragmatics. The semantics of human interaction. Berlin/New York: Mouton de Gruyter.

Part I Methodology

Text Typology: Register, Genre and Text Type Anna Trosborg The Aarhus School of Business

Introduction It is obvious that not all texts are of the same type. We may distinguish between political texts, legal texts and medical texts; fairy tales, novels and short stories differ from newspaper reports, essays, and scientific papers; food recipes, instructions booklets and advertisements may show similarities but they are not the same, expository texts differ from argumentative texts, etc. All these types of text differ in ways that are somewhat obvious, intuitively, but which nevertheless invite detailed analysis. The development in the fields of language and linguistics, communication and rhetoric, the ethnography of speaking, pragmatics and discourse, etc. have contributed to and influenced our view of text typology. Throughout the last decade, genre analysis, in particular, has enjoyed immense popularity. This field of study has attracted the attention of literary scholars, rhetoricians, sociologists, discourse analysts, cognitive scientists, machine translators, computational linguists, ESP specialists, business communication experts, language teachers a.o. (see Bhatia 1993: ix). This popularity does not mean, however, that there is a general consensus on the meaning of the term. A number of questions prevail. How do genres relate to register and text types? How is one genre to be identified and distinguished from other genres? Are the defining criteria text-internal, or is the classification based on text-external criteria, or both? Do we need uni-criterial or multi-criterial classification systems? What are the characteristics of specific genres? Do these characteristics differ cross-culturally and if so in what ways? Besides, our knowledge of specific genres still leaves much to be desired. The aim of this article is to point to a number of classificatory categories which each in their own right (and together) can be used to classify as well as explain ways in which types of discourse may usefully be categorized and ac-

4

REGISTER, GENRE AND TEXT TYPE

counted for. Terminological problems and considerations, comprising notions such as text and discourse, register, genre, and text type, discourse purpose, communicative purpose or communicative function, etc. are dealt with. No pretense to an exhaustive coverage is suggested in this brief outline, of course. A framework comprising a classification into registers and genres, with communicative function and text type as crucial categories within a discourse framework of field, tenor and mode will be suggested. This framework forms guidelines for identifying and generating conventions and functions of language and the implications for translating is discussed. Text and discourse For some scholars, text refers to written language and discourse to spoken language. For others, texts may be spoken or written, and they may involve one or more text-producers (cf. Virtanen 1990: 447). Halliday and Hasan (1976) and Quirk et al. (1985) talk about text, while e.g. Grimes (1975) and Sinclair and Coulthard (1975) and their followers deal with discourse. Do these scholars refer to the same thing, albeit the difference in their use of terminology? A study of the various uses of text and discourse in the literature during the past two decades (traced by Virtanen (1990)) highlights this problem.1 The two separate terms text and discourse have, in fact, been related to two different but complementary perspectives on language. A text may be viewed as structure and/or it may be regarded as a process. In line with these two approaches, text has often been used of a static concept - the product of a process - while discourse has been used to refer to a dynamic notion - the process of text production and text comprehension (Virtanen 1990: 453). However, the notion of text has expanded from a descriptive structural to a processual unit adopting situational factors into its scope. Seen within this development, it seems rather arbitrary today to maintain a strict boundary between text linguistics and discourse analysis. As a result, the two separate terms text and discourse may be used interchangeably - that is if no definition to the contrary has been proposed.2 Text and discourse can be directed to any aim of language or refer to any kind or reality; it can be a poem, a comedy, a sports commentary, a political speech, an interview, a sermon, a TV ad., etc. Register The concept of a "whole language" is so vast and heterogeneous that it is not operationally useful for many linguistic purposes, and the description of com-

ANNA TROSBORG

5

municative situations and events is now fairly widely recognized as a proper goal of linguistic analysis. Two sets of insights from anthropology and linguistics have been particularly influential, namely the work of Malinowsky (1923, 1935), and that of Firth (1935, 1951). Malinowsky's theory of context was originally developed with the translator in mind. Faced with the task of portraying remote cultures, he became increasingly concerned with the context of situation in order to truly convey cultural insights. Malinowsky believed that the cultural context, comprising a variety of factors ranging from the ritualistic to the more practical aspects of everyday life, was crucial in the interpretation of the message. The insights of Firth relate to culture as determining our world of language and cognition. Cultural factors influence and determine linguistic choices. This view of language was built on the views of Malinowsky and emphasized situation and culture. The contextual factors outlined were those components of speech events referred to in the ethnography of speaking research, i.e. setting, speaker-hearer role relationship, channel, genre, key. etc. (cf. Bauman and Scherzer 1975). The finding that language varies with its function led to descriptions of "varieties" of language use referred to as registers (Reid 1956, Halliday, Mcintosh and Strevens 1964). A framework devised by Halliday, Mcintosh and Strevens (1964) divided language into user-related varieties also termed dialects (Corder 1973), and use-related varieties known as registers. User-related varieties comprise geographical, temporal, social (non)standard dialects and idiolects, while registers comprise an open-ended set of varieties (or styles) of language typical of occupational fields, such as the language of religion, the language of legal documents, the language of newspaper reporting, medical language, technical language, etc. Register, as a functional language variation, is a "contextual category correlating groupings of linguistic features with recurrent situational features" (Gregory and Carroll 1978: 4). Sub-codes of a particular language were distinguished on the basis of the frequency of lexico-grammatical features of a particular text-variety (see, e.g. Crystal and Davy 1969, Gregory and Carroll 1978). Studies on the frequencies of syntactic properties (see, e.g. Barber 1962, Crystal and Davy 1969, Gustaffsson 1975) provide empirical evidence confirming intuitive and impressionistic statements about high/low frequencies of certain syntactic features in various varieties of language. Studies to investigate the relationship between grammatical choices and rhetorical functions (i.e. communicative functions) were carried out in written English for Science and Technology by, e.g. Lackstrom, Selinker and Trimble (1973), Swales (1981), Trimble (1985). An interesting finding was how specific linguistic features take on restricted values in the structuring of scientific communication.3 This line of research gave rise to the recognition of

6

REGISTER, GENRE AND TEXT TYPE

situation types. A recognition that it is often the collocation of two or more lexical items, rather than the occurrence of isolated items that determines the identity of a given register was another major finding. The account of language variation sheds light on the conscious stylistic choices made by language users. The factors which affect these choices became the focus of attention: The category of register is postulated to account for what people do with their language. When we observe language activity in the various contexts in which it takes place, we find differences in the type of language selected as appropriate to different types of situation. (Halliday et al. 1964: 87, quoted in Hatim and Mason 1990: 46). The question is what is meant by different types of situation.4 If the goal is that the user's awareness of conventional situation types is to facilitate effective and appropriate communication, register is too broad a notion. Focusing mainly on the language of a particular field (language of scientific reporting, language of newspaper reporting, bureaucratic language, legal language, etc.), register analysis disregards differences between various genres within a field. Even if there remains some shorthand convenience attached to retaining lables such as scientific, medical, legal or even newspaper English, in reality such terms can now be seen to be systematically misleading. They overprivilege a homogeneity of content at the expense of variation in communicative purpose, addresser-addressee relationships, and genre conventions (Swales 1990: 3). While it remains necessary to use texts in order to understand how texts organize themselves informationally, rhetorically and stylistically, textual knowledge remains generally insufficient for a full account of genre. To further confuse the matter, notice also usage like 'employer register' (Werlich 1976) focusing on tenor and 'written register' (Schleppegrell 1996) adjusted to mode. Genre Genres are the text categories readily distinguished by mature speakers of a language, and we may even talk about a "folk typology" of genres. Texts used in a particular situation for a particular purpose may be classified using everyday labels such as a guidebook, a nursery rhyme, a poem, a business letter, a newpaper article, a radio play, an advertisement, etc. Such categories are referred to as genres. Analysis of registers on their own reveal little about the nature of genres, so registers are divided into genres reflecting the way social purposes are accomplished in and through them in settings in which they are used. As pointed out

ANNA TROSBORG

7

by Bhatia (1993: 17), for example, a science research article is as legitimate an instance of scientific English as is an extract from a chemistry lab report. Aca­ demic conversation shows a variety of casual hallway chats, lectures, conversa­ tions between teachers and students in and out of class, e-mail, memos, schol­ arly papers, books (Bhatia 1993: 11). The legal register may comprise the language of the law in legal documents (legislative texts, contracts, deeds, wills), the language of the courtroom (e.g. the judge declaring the law, judge/ counsel interchanges, counsel/witness interchanges), the language of legal textbooks, and various types of lawyers' communication with other lawyers and with laymen (Trosborg 1991: 4). Only in the case of restricted registers is there a close relationship between register and genre (for example weather fore­ casts). By means of the concept of genre we can approach texts from the macrolevel as communicative acts within a discoursive network or system: Because it is impossible for us to dwell in the social world without repertoires of typified social responses in recurrent situations - from greetings to thank yous to acceptance speeches and full-blown, written expositions of scientific or scholarly investigations - we use genres to package our speech and make of it a recognizable response to the exigencies of the situation. (Berkenkotter and Huckin (1995: 7) A comprehensive study of genres by Swales (1990) analyses the development of the concept of genre in the fields of folklore studies, literature, linguistics and rhetoric (see Swales 1990: 34-45). Genre analysis has a long-established tradition in literary studies. It dates back to Aristotle, who distinguished genres as classes of texts, a view which still prevails. Today, the term genre, which was formerly used as "a distinctive type or category of literary composition" (Webster's Third Dictionary) is quite easily used to refer to a distinctive cate­ gory of discourse of any type, spoken or written, with or without literary aspi­ rations. However, within linguistics, few studies have distinguished register from genre. Swales points to register as a well-established and central concept in lin­ guistics, while genre is described as "a recent appendage found to be necessary as a result of important studies of text structure".5 It is only recently in the sys­ temic school that genre has become disentangled from register: Frow (1980: 78), for instance, refers to "discourse genre, or register". An unwillingness to demote register to a second position strengthened by large-scale investment in analysis of language varieties, for example for lexicographic purposes, may well be the reason for the relatively little interest in recognizing texts as genres, that is in seeing "how texts are perceived, categorized and used by members of a community" (Swales 1990: 42). 6

8

REGISTER, GENRE AND TEXT TYPE

Rhetorical scholars who have adopted a more inductive approach have tended to take context more into account and to give genre a more central place thus making a distinction between rhetorical situation and rhetorical genre, with emphasis on the recurrence of similar forms (together in constellation) in genre creation. Recently, rhetorical studies of genre have focused more on the social dynamics and social constitution of nonliterary forms of writing and speaking. With the work during the 1980s in the fields of Language for Specific Purposes and professional discourse, there was a shift of emphasis to a growing interest in the sociocultural functions of disciplinary genres, for example legal and scientific communication, and a number of surveys of key professional areas such as those by Maher (1986) on medical English and Bhatia (1987) on legal English have appeared. Recognizing the dynamic aspect of genres (amenable to changes), researchers now busy themselves with establishing genres of specialized language. Defining speech events in a community has become crucial; there is an interest in discovering in a community which communications are generically typed and what labels are used, in order to reveal elements of verbal behaviour which the community considers sociolinguistically salient. This has been the concern of ethnographers for more than a decade (cf. Saville-Troike 1982). It is basic to ethnography that the units used for segmenting, ordering and describing data should be the categories of the community and not a priori categories of the investigator (Saville-Troike 1982: 34). The procedure should be to develop sets of a posteriori categories based on empirical investigation and observation within which eliciting the community's category labels plays a central role. Swales's review includes statements by a number of researchers (e.g. Todorov (1976) and Fowler (1982)) to the effect that genres are not simply assemblies of more-or-less similar textual objects but, instead they are coded and keyed events set within social communicative processes. Recognizing those codes and keys can be a powerful facilitator of both comprehension, composition and translation. Similarly, Miller (1984: 151) argues that "a rhetorically sound definition of genre must be centred not on the substance or form of the discourse but on the action it is used to accomplish". Martin (1985: 250) considers genres to embrace each of the linguistically realized activity types which comprise so much of our culture. Genre is recognized as a system for accomplishing social purposes by verbal means. Genre "refers to the staged purposeful social processes through which a culture is realized in a language" (Martin and Rothery 1986: 243).

ANNA TROSBORG

9

Communicative purpose as the defining criterion of genre How then is genre to be identified, classified and described? How can one genre be distinguished from another? These and similar problems have been a scholarly concern for the last decade. For some scholars, genres are defined primarily on the basis of external criteria; newspaper articles are found in the news sections of newspapers, academic articles are found in academic journals (see Biber 1989: 6), while for other scholars, communicative purpose and/or linguistic content and form play a role. When accounting for the concept of genre, Swales emphasizes the sociorhetorical context of genre, the categories to be defined are those of the community, and communicative purpose is the decisive defining criterion. His analysis focuses on genre as a class of communicative events, and the principal criterial feature that turns a collection of communicative events into a genre is some shared set of communicative purposes. Exemplars or instances of genres vary in their prototypicality with the discourse community's nomenclature for genres as an important source of insight (see Swales 1990: 49-52). Crucial factors are discourse community, genre and task bound together by communicative purpose. It is communicative purpose that drives the language activities of the discourse community; it is communicative purpose that is the prototypical criterion for genre identity, and it is communicative purpose that operates as the primary determinant of task (Swales 1990: 10). Of recent studies, attention must be drawn to Bhatia (1993) and Berkenkotter and Huckin (1995). Bhatia (1993) (following Swales (1981, 1985, 1990)) takes genre to be primarily characterized by the communicative purpose(s) that it is intended to fulfil. It is this shared set of communicative purpose(s) which shapes the genre and gives it an internal structure, and a major change in the communicative purpose(s) is likely to result in a change of genre, while minor changes or modifications are likely to be distinctive of sub-genres, even though it is not possible to draw a fine distinction between genres and sub-genres (Bhatia 1993: 14).7 A further point to be stressed is that genres are meant not so much to classify but to clarify and explain the rationale of social behaviour (cf. also Fowler 1982: 286). The concept of genre as social action, one situated in a wider sociorhetorical context operates not only as a mechanism for reaching communicative goals but also as a means of clarifying what these goals might be (Swales 1990: 44). Aknowledging that there are a number of other factors, like content, form, intended audience, medium or channel, which influence the nature of the construction of a genre, Bhatia (1993: 13) also sees a close connection between the communicative purpose of a particular genre and its typical cognitive struc-

10

REGISTER, GENRE AND TEXT TYPE

turing. For exemplification, he points to a comparison of a typical news report and a feature article in a newspaper. Factors relating to mode (including channel and nature of participation) and tenor of discourse (including the status and the social distance between the participants) remaining the same, their communicative purposes change from an objective reporting in the news report to a balanced analysis of some interesting and controversial issue in the feature article. These differences in communicative goals require different strategies to be used in the two genres. In cases like these, where the communicative purposes of the genre-text are considerably different, requiring different cognitive structuring, the two texts are viewed as different genres (Bhatia 1993: 21-22). In agreement with the stand taken by Swales, Bhatia takes genre analysis from linguistic description to explanation. He has emphasized the importance of motive as an approach to linguistic analysis. His aim is to find answers to the question "Why do members of a specialist community write the way they do" (Bhatia 1993: 1). As such, genre analysis must attempt explanation and go beyond description to rationalize conventional aspects of genre construction and interpretation. Finally, Berkenkotter and Huckin (1995), who combine their expertise in the fields of discourse analysis and cognitively based rhetorical research, have developed what they call a sociocognitive theory of genre. Their theory is explained by Trosborg (this volume), who uses it as an explanatory approach when discussing the translation of documents of the European Union as hybrid political texts. A multi-dimensional approach to genre So far, genre has been established as a system underlying register.8 Taking register in its narrow sense of occupational field, contracts will always be part of the legal register, a sermon will involve the religious register, and so on, but a particular genre may cut across a number of registers. A study carried out by Swales (1981) has shown that a research article in chemistry may not be very different from a research article in, for example, sociology. This finding questions the assumption that genre is subordinated to register. Genres are subordinated to registers only in the sense that one register may be realized through various genres. Conversely, one genre may be realized through a number of registers just as a genre constrains the ways in which register variables of field, tenor and mode can be combined in a particular society. Some topics will be more suitable for lectures than others, while other topics are likely to be chosen for informal conversation between equals. Both concepts need to be considered. Registers impose constraints at the linguistic level of vocabulary and syntax, whereas genre constraints operate at

ANNA TROSBORG

11

the level of discourse structure. Furthermore, genre specifies conditions for beginning, structuring and ending a text, for which reason genres, unlike registers, can only be realized in completed texts (see Couture 1986: 82). Therefore, I do not see genre as subordinated to register or field.9 Instead, I see genres as having complementary registers, and communicative success with texts may require an appropriate relationship to systems of genre and register (cf. Couture 1986: 86). Acknowledging these points, a description of genre in its own right, independent of a subordination to a particular register, is needed. Furthermore, communicative purpose may be a dubious criterion for identification of genre. For Bhatia, the use of this criterion leads to the classification of advertisements and job applications as belonging to the same genre: to promote the value of something, be it an article or a person.10 There are genres for which purpose is unsuited as a primary criterion, for example poetic genres aimed at giving verbal pleasure defy ascription of communicative purpose. So do a number of texts types in which medium is a decisive criterion, for example memos, e-mails, faxes etc., which are characterized and influenced by their medium of communication. What we need is not a classification according to a narrow specification of field, neither is it a uni-criterial model focusing, for example on communicative purpose. Instead, we need a multicriterial model in which all relevant dimensions count. Recent approaches to language acknowledging language as text (beyond the sentence) and language as social action embedded in communicative situations led to the availability of just such a model. The approach developed by Michael Halliday and his colleagues in Britain in the 1960s and 1970s provided translation studies with an alternative view. They launched a functional approach to language, an approach "which attempts to explain linguistic structure, and linguistic phenomenon, by reference to the notion that language plays a certain part in our lives; that it is required to serve certain universal types of demand" (Halliday 1971: 331). This social theory of language, with its three-fold division in field, tenor and mode, and known as the systemic-functional model is now acknowledged world-wide in the fields of linguistics, sociolinguistics, communication studies, applied linguistics, etc. The use of this model for translation purposes is the great discovery of Vermeer and Nord (see also Hatim and Mason 1990; Baker 1992). Genres can be defined multicriterially through an extension of the variables field, tenor and mode, with a development of field in the ideational component covering linguistic content, of tenor in the interpersonal component covering communicative functions in relation to sender/receiver role relationships, and, finally, the development of mode in the textual component involving medium.

12

REGISTER, GENRE AND TEXT TYPE

While a genre can only be fully accounted for through a specification of field, tenor and mode and a description of linguistic features realized in the ideational, the interpersonal and the textual components of particular texts (see Eggins (1994) for further development), it may nevertheless be recognized by any (outstanding) feature which has been made the focus of attention. See also Kussmaul (this volume) for an extensive presentation of how change of a single parameter may result in a change of genre. Text types cutting across registers and genres Recent interest in the structure of discourse has brought attention to one of the oldest issues in the discipline of rhetoric. Two traditions of classifying texts run through the 2,400-year-old history of rhetoric, both deriving from Aristotle's Rhetoric.11 One tradition classifies texts according to purpose, the other by type (also called 'mode', see Kinneavy (1980); Faigley and Meyer (1983)). A discourse may be characterized in terms of its communicative function. Is the discourse intended to inform, to express an attitude, to persuade or create a debate, etc? Additionally, it may be classified according to text type into descriptive, narrative, expository, argumentative, instrumental etc. The focus is on functional categories, also termed rhetorical strategies, which is not normative, but abstract knowledge, fundamental in the creation of texts.12 Criticism has been launched on genre analysis to the effect that genre distinctions do not adequately represent the underlying text functions of English. Texts within particular genres can differ greatly in their linguistic characteristics; for example, newspaper articles can range from extremely narrative and colloquial in linguistic form to extremely informational and elaborated in form. On the other hand, different genres can be quite similar linguistically; for example, newspaper articles and popular magazine articles can be nearly identical in form. Whereas the notion of genre refers to completed texts, communicative function and text type, being properties of a text, cut across genres. Thus informative texts may comprise newspaper reports, TV news, textbooks, etc. and argumentative texts may comprise debates, political speeches, newspaper articles, etc. Genres and text types are clearly to be distinguished, as linguistically distinct texts within a genre may represent different text types, while linguistically similar texts from different genres may represent a single text type (cf. Biber 1989: 6).

ANNA TROSBORG

13

Communicative function, speech act and text act. In determining the purpose of discourse, many previous theories have taken the component of the language process into consideration, implicitly or explicitly. Four factors of the linguistic process are often listed: speaker, listener, thing referred to, and the linguistic material.13 According to a framework acknowledging Aristotle and Bühler as their sources, a text is classified into a particular type according to which component in the communication process receives the primary focus. If the main focus is on the sender, the discourse will be expressive; if on the receiver, it will be persuasive; if on the linguistic code, literary; and if the aim is to represent the realities of the world, it will be referential (see Jakobson 1960, Kinneavy 1971). In Linguistics and Poetics (1960), Roman Jakobson also emphasized language product as a pleasurable end in itself (cf. his poetic function) and added two other uses: the metalanguage, in which language is used to talk about language, and phatic communication, which is the use of language merely to keep the channel open, as in an introduction or in some seemingly trivial conversational crutches. Jakobsen's basic model has been adopted by some important anthropologists, for example Dell Hymes (1974). For a more recent account, see Kinneavy (1980: 65), who acknowledges the work of Aristotle and Aquinas, Cassier, Morris, Miller, Russell, Reichenbach, Richards, Bühler, Jakobson, when comparing their categories with his own. For a typology of texts based on communicative functions and used to form the basis of translator's decisions, see Reiss (1976). See also Nord (this volume) for whom communicative functions play a crucial role in establishing her functional typology of translations. Another body of research concerned with communicative purpose is speech act theory, which views language as action made up of communicative acts (cf. Austin 1962, Searle 1969, 1976, and their followers). Thus Searle (1976) distinguishes five major classes of speech acts: Representatives, directives, expressives (and evaluatives 14 ), commissives and declarations. Each class divides into a number of different speech acts; for example it makes a difference whether a speaker is begging, asking, ordering or threatening; the illocutionary point is the same, namely that of influencing the hearer, however, different illocutionary forces are expressed. Acknowledging Traugott and Pratt (1980) as their source, Hatim and Mason (1990) has adopted this framework for translation purposes. The illocutionary (or communicative) force, which is the dynamic element in communication, is not to be treated in isolation as the illocutionary force of each utterance. Rather, the interrelationship of speech acts within sequences leads to the notion of illocutionary structure of a text. The overall purpose may be that of achieving equivalence of illocutionary force at text level. For ex-

14

REGISTER, GENRE AND TEXT TYPE

ample, an advertisement may be predominantly referential in nature, consisting of informative (and expressive) statements, but still, as is well known, the aim is that of persuading the consumer to buy. Advertisements are difficult to classify at the level of typical formation patterns, but they all share the same function of promoting the sales of a product, i.e. they are directive at text level. Real text will display features of more than one type. As (this) multifunctionality is the rule rather than the exception, any useful typology of texts will have to be able to accommodate such diversity (Hatim and Mason 1990: 138). The expressed intent of the author may not be the real intent. In many cases in these overlaps, one of the aims is dominant and the other is a means. Information included in an advertisement is there to further the persuasion, so that persuasion is the primary aim. In some literature, it is obvious that persuasion is a secondary motif, etc. In expressing an opinion, factual knowledge as well as evaluative judgements may be brought as supportive statements. Studies of how entire sequences of speech acts are evaluated on the basis of higher order expectations about the structure of a text, and how these sequences of coherent microtexts contribute to the global coherence of a larger text have become the ultimate goal of text pragmatics (Ferrara 1985: 140). In translating, the aim is not necessarily matching speech act for speech act. The reader's (client or consumer, etc.) interest must be constantly matched against the communicative intent of the producer of the source text. For example, if the intention of the producer of the ST is to sell a product, any translation of the text as an advertisement must be evaluated in terms of how well it serves that purpose (i.e. the persuasive text act involved), rather than on the basis of a narrow linguistic comparison. If, on the other hand, a translation of advertising copy is required purely for information, the translator's product will be adjusted accordingly. The predominant illocutionary force of sequences of speech acts, i.e. the text act, must be recognized (see Hatim and Mason 1990: 78-82, cf. Horner 1975). Failure to recognize the text act can be a stumbling-block in conveying the communicative intention of a message and may easily lead to misunderstandings (see, e.g. Hatim and Mason 1990: 78-79). Recognizing the text act is therefore an important precondition in translation and interpreting. Text type and text type focus The varities of discourse have been established in register analysis and elaborated on in genre analysis. Particular genres, each with its own characteristics, within and across registers, have been discussed. Newspaper reporting imposes a certain demand that the main content be given in the opening sentences, and the details are to be given in successive sentences and paragraphs of the story

ANNA TROSBORG

15

(cf. the news pyramid), the validity of legal documents may depend on the expression of specific formulae, a business letter requires a high level of explicitness, while a private letter is the writer's own concern, etc. There are, on the other hand, similarities between certain types of discourses. The fairy tale, the novel, the short story, are all types of story telling; a first person novel may have much in common with a first person account of an informal sort. First person accounts, newspaper reporting, and historiography make claims to factuality. Essays and scientific papers have much in common as do sermons, pep talks and some political speeches. Food recipes have something in common with instruction booklets, etc. (cf. Longacre 1982). Certain discourse types are then somewhat similar to others. Our classification therefore needs to include both classification into registers and genres and specifications into modes of discourse made up by text types. While communicative purpose represents the overall aim of a text, rhetorical purpose is made up of the rhetorical strategies which constitute the mode of discourse realized through text types. Text types are identified as "a conceptual framework which enables us to classify texts in terms of communicative intentions serving an overall rhetorical purpose" (Hatim and Mason 1990: 140). While genres form an open-ended set (Schauber and Spolsky (1986), text types constitute a closed set with only a limited number of categories.15 Building on Aristotle, Kinneavy (1971, 1980) theorizes a classification of text types in terms of modes, which derive from philosophical concepts of how reality can be viewed. His primary distinction is between static and dynamic, between looking at something at a particular time and looking at how it changes over time, and he arrives at the four classes of narration, classification, description, and evaluation. If our static view of reality focuses on individual existences, we describe', if it focuses on groups, we classify. If our dynamic view of reality looks at change, we narrate; if it looks at the potential for reality to be different, we evaluate. As such, these text types are cognitive categories offering ways of conceptualizing, perceiving and portraying the world. Werlich's (1976) typology, which includes five idealized text types or modes: description, narration, exposition, argumentation, and instruction has later been adopted by Hatim and Mason (1990) for translation purposes with a divison of instruction into two subclasses: instruction with option (as advertisements, manuals, recepes, etc.) and instruction without option (e.g. legislative texts and contracts). The typology is based on cognitive properties of text types: Differentiation and interrelation of perceptions in space (description), differentiation and interrelation of perceptions in time (narration), comprehension of general concepts through differentiation by analysis and/or synthesis (exposition), judging, i.e. evaluation of relations between and among concepts through the extraction of similarities, contrasts, and transformations (argumen-

16

REGISTER, GENRE AND TEXT TYPE

tation), planning of future behaviour (instruction). For differentiation and further description of these categories, see Hatim and Mason (1990: 153-160), Albrecht (1995: 117). For example, exposition is to be distinguished from argumentation on the grounds of factuality established by means of a scene-setter, whereas argumentation is established through a tone setter as evaluative discourse. See also Biber (1989) for a classification based on formal linguistic criteria. Text types often cut across genre categorizations. The relationship between genres and text types is not straightforward. However, this finding does not invalidate genre analysis. Genres and text type categorizations have different theoretical bases, which are both valid as distinct text constracts. Genres correspond directly to the text distinctions recognized by mature adult speakers, reflecting differences in external format and situations of use. The theoretical basis of genres is independent from those for text types. Genres are defined and distinguished on the basis of systematic non-linguistic criteria, and they are valid in those terms. Text types may be defined on the basis of cognitive categories (as described above) or on the basis of strictly linguistic criteria (similarities in the use of cooccurring linguistic features, cf. Biber's (1989) typology). Contrary to most previous findings, Biber's types are valid in linguistic terms and captures the salient linguistic differences among texts in English. See also Longacre (1976, 1982) and Smith (1985). No theory of modes of discourse is rigid in its categorization. Most discourse employs multiple views of reality and is therefore multiple in type (cf. Kinneavy 1980: 37), and pure narration, description, exposition and argumentation hardly occur. Thus, a particular genre may make use of several modes of presentation, though typically with one of these as the dominant type. The idea of an overall function was recognized by what Morris (1946: 75) calls a "dominant" mode. Today, text type focus, or contextual focus, refers to text type at the macro level, i.e. the dominant function of a text type exhibited in or underlying a text (cf. Werlich 1976, Hatim and Mason 1990).16 If a text is incorporated into a larger text with a different overall purpose, the performative impact of the incorporated text may be changed by its incorporation. This explains why linguistic features at microlevel need not be isomorphic with the particular characteristics of the contextual focus. The same principle holds in uses of speech acts (see above for a dominant communicative function or text-act). Thus, we need a two-level typology for text types (as well as for communicative functions) rather than a single level of types only. At the macrolevel of discourse, text type may be assumed to precede the level of text-strategic choices, thus affecting the whole strategy of the text. The choice of microlevel text types on the other hand, has to do with the textualization process, which is determined by the text producer's text strategy. The text types employed in a

ANNA TROSBORG

17

particular text (or genre) need not agree with its contextual focus. An argumentative text-type focus may be realized through narration, instructions may take the form of description, and so forth. In the sense of various blends of different text types, a dominant text type is often recognizable. Hatim and Mason (1990: 146-148) account for types of "hybridisation", stressing the need for translators to become aware of these aspects. Of great interest is the interaction between communicative purpose and rhetorical purpose (text type), for example, in order to persuade one can narrate, describe, counterargue, etc. (cf. Hatim and Mason 1990: 145). The interrelation between the purpose of the communication and the rhetorical strategies determining the text type(s) employed to achieve the intended communicative goal is an object of study deserving further attention. Concluding comments The persistence of classifications of texts according to aims and modes in the 2,400-year-old history of rhetoric suggests that these classifications do reflect some fundamental properties of discourse. However, the discourse framework is decisive, so we are back to discourse situation and genre. The interaction between discourse situation and genre, on the one hand, and text type and communicative functions on the other, is an all important goal of future investigations within as well as across languages and cultures. Genre knowledge, knowledge of form-function relations of communicative functions and text types are important not only to scholars and researchers in the fields of communication, rhetoric, and sociology of science, to linguists who teach and conduct research in ESP and LSP, but also to practitioners who compose or translate in the disciplines. Today, there is a growing interest in assessing rhetorical purposes, in unpacking information structures and in accounting for syntactic and lexical choices. Moreover, the resulting findings are no longer viewed only in terms of stylistic appropriacy but, increasingly, in terms of the contributions they may or may not make to communicative effectiveness and appropriateness. We deal increasingly with the communicative character of discourse. Understanding the genres of written communication in one's field is essential to professional success. The function of the genre must be understood from the perspective af the composer/translator who must draw upon knowledge of register and genre to perform effectively. However, the acquisition of conceptual knowledge, like the learning of the use of tools, is "both situated and progressively developed through activity", and "to learn to use tools as practitioners use them, a student, like an apprentice, must enter the community and

18

REGISTER, GENRE AND TEXT TYPE

its culture" (Brown, Collins, and Duguid 1989: 33, reported in Berkenkotter and Huckin 1995: 12). Learning the genres of disciplinary and professional discourse requires immersion into the culture for a lengthy period of apprenticeship and enculturation (cf. Freedman 1993). The problem for translators operating in a different culture is that they are often asked to use the tools of a discipline without being able to adopt its culture. As compensation, knowledge of cross-cultural differences and similarities regarding text typology and conventions may be a useful source. A particular genre is often a highly structured and conventionalized communicative event. This specific structure and convention is of great importance to the translator. Likewise, the constraints on allowable contributions in terms of their intent, positioning, form and functional value are likely to constrain the translator who must conform to standard practices within the boundaries of a particular genre. Even if we grant that surface features and local decisions are highly contributory to the performance outcome, it is still very much the case that it is facilitative for a participant to have a sense of the underlying logic or rationale of text typologies in both reception and production. This is even more so in cases where no guiding linguistic features are identified. The translator may in some contexts have to pay particular attention to the way the "official" function of a text is being manipulated (Hatim and Mason 1990: 146). As mentioned, failure to recognize the illocutionary force of single utterances as well as the superordinate communicative intent of the text act can be a major stumbling-block in establishing the aims of discourse and may result in faulty translations. Similar problems hold for text types. Conducting a translation exercise, Hatim and Mason (1990: 149) found that the majority of twelve translator trainees produced translations which could be faulted on the grounds that they misinterpreted text-type focus. Recognizing text act and texttype focus are important goals in translation. In using the genres customarily employed by members of their discourse community, text producers help constitute the community and simultaneously reproduce it (Berkenkotter and Huckin (1995: 25). Similarly, translators aware of the conventions play the same role. Communicative functions and text types are universal, but subservient to cultural norms reflected in realisation strategies and the organisation of texts. Lack of relevant knowledge of genre, communicative functions, text types and culture may result in distorted translations. Text typology with genre conventions and knowledge of how communicative functions and text types are realized in different languages within and across genres are useful knowledge in translator training and in translation itself.

ANNA TROSBORG

19

Notes 1 For recent surveys on definitions of text and discourse, see also Vitacolonna (1988). 2 For example, Longacre (1982) has an article on "Discourse typology in relation to language typology", whereas Biber (1989) has one on "A typology of English texts". Both authors are concerned with the classical rhetorical classes of narration, description, exposition and argumentation. 3 For example, Swales (1990: 26) has shown how pre-modifying en-participles textualize two different aspects of chemistry text depending upon whether the author is exemplifying or generalizing. 4 For Hatim and Mason (1990: 48), "situation type" includes any number of similar situations (tokens) of the general type. For example, making your next appointment with the dentist's receptionist is a particular token of a recognized type of situation. 5 Linguistics as a whole (apart from Martin (1985), Martin and Rothery (1986) and Couture (1986) has tended to find genre indigestible (Swales 1990: 41). 6 In fact, it was mainly due to the influence of the register concept that recognition of differences between, say, medical journal editorials and articles (Adams Smith 1984) or between legislative prose, legal textbooks and legal case reports (Bhatia 1983) has developed rather slowly in the English for Specific Purposes field (cf. Swales 1990: 3). 7 The division of genres into sub-genres is dependent on the degree of specificity of the classification: letter/business letter/letter to The Bank of Westminster. 8 The relationship between genre and the longer established concept of register is not always very clear (see Ventola (1984) for a discussion of this uncertainty). 9 Hatim and Mason (1990: 75) see genre as a development of field. 10 Bhatia places advertising to sell goods and letter of job application as the same genre, neglecting the fact that the two kinds of text have different receivers: people (often in general) who may not be interested in the product, and a specific firm who have already formulated its needs and interest in the type of person being advertised for. 11 Rhetoric, traditionally (and for over a century in English departments), often refers to the whole field of the uses of language. A more specific meaning is the use of rhetoric to refer to modes of discourse realized through text types (narration, description, exposition, argumentation, etc), i.e. the classification of texts by type (see Kinneavy 1980: 3-4). To complicate the matter further, some scholars also refer to communicative functions as rhetorical strategies (cf. Trimble 1985). 12 Longacre (1976, 1982), Smith (1985) and Biber (1989) all refer to text types as "underlying shared communicative functions". Here we reserve communicative functions to a classification of speech acts in accordance with the typology suggested by Kinneavy and others, restricting text types to modes of discourse. For different authors in linguistics, "function" has referred to: the kind of reality referred to (Cassirer 1944: 171 ff, Urban 1939: 134 ff); the level of social formality of a given discourse (Kenyon 1952: 215 ff.) nonmorphological classes of words in grammar (Fries 1952). Clearly, none of these meanings are intended here. 13 In fact, Aristotle proposed what he called "a language concerned with things", and a "language directed to the hearer". A three-dimensional model of communication, resulting in the so-called communication triangle was originally proposed by Bühler (1933: 74). 14 Searle collapses expressives and evaluatives, originally distinguished by Austin as two separate classes, into one class, while Traugott and Pratt (1980) keep the six original classes. 15 See also Chafe (1982), who proposes a four-way classification of texts with respect to the parameters of 'involvement-detachment' and 'integration-fragmentation'.

20

REGISTER, GENRE AND TEXT TYPE

16 Note that Virtanen (1992) uses discourse type to refer to text type at the macrolevel reserving text type for microlevel analysis. However, as text and discourse have been defined to be used interchangeably (compare also the alternate use of discourse type and text type), it seems confusing to use discourse type at the macrolevel in contrast to using text type at the microlevel. For this reason, I do not use the term discourse type to represent text type at the macrolevel; instead, I use the term 'text-type focus' or 'contextual focus' employed by Werlich (1976) and adopted by Hatim and Mason (1990) to refer to the predominant text type. References Adams Smith, Diana E. 1984. "Medical discourse: Aspects of author's comment". The ESP Journal 3: 25-36. Aristotle. 1960. Rhetoric. (L. Cooper, trans.) New York: Appleton-Century-Crofts. Albrecht, Lone. 1995. Textual Analysis. Copenhagen: Samfundslitteratur. Austin, John L. 1962. How to Do Things with Words. New York: Oxford University Press. Baker, Mona 1992. In Other Words: A coursebook on translation. London and New York: Routledge. Barber, C.L. 1962. "Some measurable characteristics of modern scientific prose". In Contributions to English Syntax and Phonology, 1-23. Stockholm: Almquist & Wiksell. Bauman, R. and Sherzer, J. 1975. "The ethnography of speaking". In B. Siegel (ed), Annual Review of Anthropology 4: 95-120. Berkenkotter, Carol and Huckin, Thomas N. 1995. Genre Knowledge in Disciplinary Communication: Cognition/culture/power. Hillsdale, New Jersey: Lawrence Erlbaum Associates, Publishers. Bhatia, Vijay K. 1983. "Simplification vs. easification: The case of legal texts". Applied Linguistics 4(1): 42-54. Bhatia, Vijay K. 1987. "Textual-mapping in British legislative writing". World Englishes 6(1): 1-10. Bhatia, Vijay K. 1989. Nativization of job application - a microethnographic study. Paper presented at the International Conference of English in South Asia, Islamabad, Pakistan, 4-8 Jan., 1989. Bhatia, Vijay K. 1993. Analysing Genre. Language use in professional settings. London/New York: Longman. Biber, Douglas. 1989. A Typology of English Texts. Linguistics 27: 3-43. Brown, J.S., Collins, A., and Duguid, P. 1989. "Situated cognition and the culture of learning". Educational Researcher 18: 32-42. Bühler, Karl. 1933. "Die Axiomatik der Sprachwissenschaften". Berlin: Kant-Studien 38: 19-90. Cassirer, Ernst. 1944. An Essay on Man. New Haven: Yale University Press.

ANNA TROSBORG

21

Chafe, W.L. 1982. "Integration and involvement in speaking, writing, and oral literature". In D. Tannen (ed), Spoken and Written Language: Exploring orality and literacy. Norwood, NJ: Ablex, 35-53. Corder, S. Pit. 1973. Introducing Applied Linguistics. Harmondsworth: Penguin. Couture, Barbara (ed). 1986. Functional Approaches to Writing: Research perspectives. Norwood, NJ: Ablex. Crystal, David and Davy, Derek. 1969. Investigating English Style. London: Longman. Eggins, Suzanne. 1994. An Introduction to Systemic Functional Linguistics. London: Pinter. Faigley, Lester and Meyer, Paul. 1983. "Rhetorical theory and readers' classifications of text types". Text 3(4): 305-325. Ferrara, A. 1985. "An extended theory of speech acts: appropriateness conditions for subordinate acts in sequences". Journal of Pragmatics 4: 233-252. Firth, J.R. 1935. "The technique of semantics". Transactions of Philological Society, reprinted in Firth (1951), 7-33. Firth, J.R. 1951. Papers in Linguistics: 1934-1951. Oxford: Oxford University Press. Fowler, Alastair 1982. Kinds of Literature. Oxford: Oxford University Press. Freedman, A. 1993. "Show and tell? The role of explicit teaching in the learning of new genres". Research in the Teaching of English 27: 222-251. Fries, Charles Carpenter. 1952. The Structure of English. New York: Harcourt, Brace & World, Inc. Frow, John. 1980. "Discourse genres". The Journal of Literary Semantics 9: 73-79. Gregory, M. and Carroll, S. 1978. Language and Situation: Language varieties in their social contexts. London: Routledge & Kegan Paul. Grimes, J.E. 1975. The Thread of Discourse. The Hague/Paris: Mouton. Gustafsson, M. 1975. Some syntactic properties of English law publication (Publication no. 4). Turku, Finland: University of Turku, Dept. of English. Halliday, M.A.K. 1971. "Linguistic function and literary style: An inquiry into the language of William Golding's The Inheritors". In S. Chatman (ed), Literary Style: A Symposium. New York: Oxford University Press Halliday, M.A.K. and Hasan, R. 1976. Cohesion in English. London: Longman. Halliday, M.A.K., Mcintosh, A. and Strevens, P. 1964. The Linguistic Sciences and Language Teaching. London: Longman. Hatim, Basil and Mason, Ian. 1990. Discourse and the Translator. London/New York: Longman. Horner, W.B. 1975. Text act theory: A study of non-fiction texts. Unpublished PhD thesis, University of Michigan. Hymes, Dell. 1974. Foundations in Sociolinguistics: An ethnographic approach. Philadelphia: University of Pennsylvania Press. Jakobson, Roman. 1960. "Linguistics and poetics". In Thomas E. Sebeok (ed), Style in Language. M.I.T. Press: Cambridge, Massuchusetts, 350-378. Kenyon, John S. 1952. "Cultural levels and functional varieties of English". In Harold B. Allen, Reading in Applied Linguistics. Appleton-Century-Crofts, Inc.

22

REGISTER, GENRE AND TEXT TYPE

Kinneavy, James L. 1971. A Theory of Discourse: The aims of discourse. Englewood Cliffs, NJ: Prentice-Hall International. Kinneavy, James L. (1980) A Theory of Discourse. New York: Norton. Lackstrom. J. E., Selinker, L. and Trimble, L. 1973. "Technical rhetorical principles and grammatical choice". TESOL Quarterly, 1. Longacre, Robert E. 1976. An Anatomy of Speech Notions. Lisse: R de Ridder. Longacre, Robert E. 1982. "Discourse typology in relation to language typology". In S. Allén (ed), Text Processing. Text Analysis and Generation. Text Typology and Attribution. Stockholm: Almqvist & Wiksell. Maher, John. 1986. "English for medical purposes". Language Teaching 19: 112-45. Malinowsky, B. 1923. "The problem of meaning in primitive languages". Supplement 1 to C.K. Ogden and I.A. Richards, The Meaning of Meaning. London: Kegan Paul. Malinowsky, B. 1935. Coral Gardens and Their Magic, vol. 2. London: Allen & Unwin. Martin, James R. 1985. "Process and text: two aspects of human semiosis". In James D. Benson and Greaves, S. William (eds), Systemic Perspectives on Discourse, vol. 1. Norwood, NJ: Ablex, 248-274. Martin, James R. and Rothery, Joan. 1986. "What a functional approach to the writing task can show teachers about 'good writing'". In Barbara Couture (ed), Functional Approaches to Writing: Research perspectives. Norwood, NJ: Ablex. 241-265. Miller, Carolyn R. 1984. "Genre as social action". Quarterly Journal of Speech 70: 151-167. Morris, C.W. 1946. Signs, Language, and Behaviour. Englewood Cliffs, N.J.: Prentice-Hall, Inc. Quirk, R., Greenbaum, S., Leech, G., Svartvik, J. 1985. A Comprehensive Grammar of the English Language. London/New York: Longman. Reid, T.B.S. 1956. "Linguistics, structuralism and philology". Archivum Linguisticum 8. Reiss, K. 1976. Texttyp und Übersetzungsmethode. Der Operative Text. Kronberg: Scriptor. Saville-Troike, Muriel. 1982. The Ethnography of Communication. Oxford: Basil Blackwell. Schauber, Ellen and Spolsky, Ellen. 1986. The Bounds of Interpretation. Stanford, Cal: Stanford University Press. Schleppegrell, Mary J. 1996. "Conjunction in spoken English and ESL writing". Applied Linguistics 17(3): 271-285. Searle, John R. 1969. Speech Acts. Cambridge: Cambridge University Press. Searle, John R. 1976. "A classification of illocutionary acts". Language in Society 5: 1-23. Sinclair, John McH. and Coulthard, Malcolm. 1975. Towards an Analysis of Discourse: The English used by teachers and pupils. London: Oxford University Press.

ANNA TROSBORG

23

Smith, Edward L. 1985. "Text type and discourse framework". Text 5(3): 229-247. Swales, John M. 1981. Aspects of article introductions. Aston ESP Research Report No. 1, Language Studies Unit, University of Aston in Birmingham, Birmingham, UK. Swales, John M. 1985. "A genre-based approach to language across the curriculum". Paper presented at the RELC Seminar in Language Across the Curriculum, at SEAMEO Regional Centre, Singapore, April 1985. In M.L. Tickoo (ed) (1986), Language Across the Curriculum, Singapore, SEAMEO Regional Language Centre. Swales, John M. 1990. Genre Analysis. English in academic and research settings. Cambridge, UK: Cambridge University Press. Todorov, Tzvetan. 1976. "The origin of genres". New Literary History 8: 159-170. Tragott, E. C. and Pratt, M. L. 1980. Linguistics for Students of Literature. New York: Harcourt Brace Jovanovitch. Trimble, Louis. 1985. English for Science and Technology: A discourse approach. Cambridge: Cambridge University Press. Trosborg, Anna. 1991. "An analysis of legal speech acts in English Contract Law". Hermes 6: 1-25. Urban, Wilbur Marshall. 1939. Language & Reality. New York: The Macmillan Company. Ventola, Eija. 1984. "Orientation to social semiotics in foreign language teaching". Applied Linguistics 5: 275-286. Virtanen, Tuija. 1988. Discourse functions of adverbial placement in English: clauseinitial adverbials of time and place in narratives and procedural descriptions. Unpublished Licentiate thesis. Turku: University of Turku. Virtanen, Tuija. 1990. "On the definitions of text and discourse". Folia Linguistica XXIV/3-4: 447-455. Virtanen, Tuija. 1992. "Issues of text typology: Narrative - a 'basic' type of text?" Text 12(2): 293-310. Virtanen, Tuija and Wårvik, B. 1987. Observations sur les types de textes. In J. Härmä and I. Mäkinen-Schwanck (eds), Rencontre des professeurs de francais de l'enseignement supérieur: Communications. Publications du Département des Langues Romanes 6, Université de Helsinki, 91-114. Vitacolonna, L. 1988. "'Text'/'Discourse' definitions". I J.S. Petöfi (ed), Text and Discourse Constitution: Empirical aspects, theoretical approaches. Berlin/New York: de Gruyter, 421-439. Werlich, E. 1976. A Text Grammar of English. Heidelberg: Quelle & Meyer. Wierzbicka, Anna. 1985. "Different cultures, different languages, different speech acts". Journal of Pragmatics 9: 145-178. Wierzbicka, Anna. 1991. Cross-cultural Pragmatics. The semantics of human interaction. Berlin/New York: Mouton de Gruyter.

Text Types and Translation1 J. C. Sager University of Manchester

Introduction In recent years the concept "translation product" has been widened to include consideration of cultural differences between source and target languages and the purpose of translated messages. Since text types have been recognised as determiners of the global purpose of a text, recent discussions of translation have also included equivalence of text type as one of the major forms of equivalence to be aimed at. There are, however, well-established activities carried out by professional translators, such as bilingual abstracting, summarizing and gist translation, which do not readily fit into the paradigm of equivalence. In a translation theory which attempts to represent all current professional practice, the concept of equivalence has to be modified to accommodate these types of translation processes and products.

The interpretations of "equivalence" Hitherto translation theory has largely evolved around the concept of equivalence. The central position of this concept arises from a static view of translation which takes the source text as its starting point and sets up parameters for measuring the "interesting' differences between source and target text, i.e. those which are not simply the result of the linguistic structures of the languages involved. Historically, the concept of "equivalence" between source text and translation product has been assessed in terms of accuracy, fidelity, appropriateness, and other such poorly defined values. More recently, there have been various qualifications of this concept which we list here in chronological order. "Closest natural equivalent", "formal vs dynamic equivalence" (Nida 1964);

26

TEXT TYPES AND TRANSLATION

"retention of translation invariance at the content level" (Kade 1968); "communicative equivalence" (Jäger 1975, Reiss 1976); "pragmatic equivalence" (Wilss 1980); "adequacy" (Reiss 1984); "acceptability" (Toury 1980, Nord 1991); "faithfulness, fidelity, loyalty" (Nord 1991); and finally back again to "pragmatic equivalence" (Baker 1992, Koller 1992). Wilss (1995) recognises the vagueness of this concept when he states that translation aims at "some sort of equivalence". Equally vague is the now widely-used expression "functional equivalence" which implies consideration of the purpose of a translation and its original. So Reiss & Vermeer (1884: 101) "die Translate variieren in Abhängigkeit von den vorgegebenen Skopoi" (the translation products vary according to the specified "skopoi"). For the approach to translation pursued hitherto it was in principle enough to postulate that a text to be translated consisted of linguistic form and semantic content and that the operation of translation had to change the linguistic form while retaining the same content. Now that this static view is being challenged, discussions of translation take a dynamic approach, which means considering translation as one possible step in a communication process between two cultures. This widened scope includes the writer and the reader, between whom the translator has a mediating role. In this role the translator has a choice of types of equivalence, or non-equivalence at certain levels, in function of the situation created by the demand for a translation. This choice is chiefly exercised in terms of what is known about the requirements of the person who has requested or commissioned the translation. The commissioner of the translation may be a writer who wants to address one or several people in another culture, a publisher who wants to sell a novel in translation, or individual readers who need to know what is new in a scientific paper written in a language they do not understand. With the advent of machine translation, translation users, i.e. writers or readers who commission translations, now have the option of bypassing the human translator and getting a variety of quality levels of translation directly from a machine. Translators can also use machines to assist them in their work or provide edited machine products for their customers. The result of this diversification of products is that users have a much wider choice and that they can now, for the first time ever, have direct access to translation without the intervention of human translators. But in this changed situation, translators also need more specific instructions how to proceed with their work. In fact, it is now being recognised that translators need something akin to job specifications which do not only deal with the mundane matters of delivery dates and price, but also contain a characterisation of the future readers of the translation, their expectations of the translated text and the use they are likely to make of it.

J. C. SAGER

27

It is therefore useful to relativize the concept of equivalence much more than has been done hitherto and to develop new approaches to the relationship that exists between source and target text by taking into account both user needs and technical possibilities.

An approach to translation based on communication theory2 To redefine the relationship between source and target text, translation theory has to incorporate a number of useful concepts from communication theory. Messages - documents - texts Information scientists speak of documents as the units of text that have a function in communication. This concept is extremely valuable in translation if it is defined as a text, consisting of form and content, supplemented by the writer's intention which is integral to the document. Messages are documents in their pragmatic communicative situation between writer and reader. It is the writer's intention encoded in the document which distinguishes a document from a "text", i.e. a unit of form and content only, which can be given a new intention by an information mediator like an abstractor or indeed a translator when it is addressed to someone the writer did not have in mind. Intention is expressed both linguistically, through the choice of text type, the rhetorical structure of the text and the choice of words, and through the situation, i.e. the particular circumstances of production and expected reception of the document when it was first produced. For example, in messages that have served or lost their original raison d'être over time, and have subsequently been stored in some way, the original intention is also no longer relevant to later readers; they are then simple texts. Such texts can, however, be given a new intention by whoever extracts them from their resting place; for example: a researcher checking old parish records, a lawyer looking up legal precedents, a scholar elucidating an old manuscript. The writer's intention to influence the behaviour of a reader in a certain way is what motivates and directs the production of messages. In an optimally effective message the writer's intention coincides with the reader's expectation. This can, however, happen only when the writer knows the reader and vice versa. In translation this is seldom the case, translators becoming substitutes for writers. Translators, therefore mainly deal with documents, i.e. messages outside their original pragmatic communicative situation, which they have to reconstruct in order to fully understand the original message3.

28

TEXT TYPES AND TRANSLATION

When translators receive documents to translate they have to decide whether they are documents, i.e. texts incorporating a clear writer intention and hence a full specification of the addressees the writer has in mind, or simply texts in search of an application, a recipient, a sender that is to be specified by a commissioner of the translation. Because they are not given the history of a document, translators are often unaware of this non-linguistically expressed intention; without such external evidence they have to attempt to reconstruct the intention in order to gain a full understanding of the document. Faced with documents, translators have to decide whether the intention accompanying the original text is also the intention of the target text or whether, in the translation specification received from the commissioner of the translation, a new intention has been indicated. If the intention is specified as remaining the same — the case of most literary translations, for example — translators only have to assess the future readers' state of knowledge regarding the cultural associations and presuppositions of the source text. If the intention is specified as different, translators have to decide whether the source text lends itself to the proposed change of intention or not, and interact with the commissioner accordingly. Faced with texts, translators are always dependent on the commissioner to specify the intention of the target text, and cannot carry out their work without an indication of the expectations of the proposed readers. Secondary and primary readerships Another useful concept borrowed from information science is that of secondary reader as opposed to primary reader, a distinction which is of particular importance for translation as it is connected to the difference between message and text. A primary reader is the person a writer has in mind when producing a message. All readers not included in a writer's original scope of addressees are secondary readers. When there is a match between a writer's presuppositions about a reader's expectation and the reader's assumptions about the writer's intention, communication is maximally effective. Secondary readers have to evaluate the text they read as they think fit, because there is no relevant writer's intention to guide them and what linguistically expressed intention signals they encounter may be confusing rather than helpful. Primary readers receive messages; in the hands of a secondary reader such messages are downgraded to simple texts. By this definition most readers of translations are secondary readers. They are primary only if writers instruct translators to address them on their behalf, or if writers wrongly assume readers to understand their language and translators have to intervene, as sometimes happens with letters. If a reader asks for

J. C. SAGER

29

a translation of a message addressed to another reader or group of readers, the document immediately becomes a text which makes him by definition into a secondary reader. In daily practice, users of translations are at a serious disadvantage because - unable to know the content nor the original intention of the document - they cannot always judge the relevance of a document for their needs nor can they assess whether a particular translator can offer them suitable assistance in gaining access to the information they are seeking. In order to be able to decide whether a translation should be undertaken, they are therefore in the first instance dependent on external evidence surrounding the document, such as the place of an article in a serious journal, the name of the author, the date of appearance in relation to other recent information, and possibly key words and abstracts which they can sometimes find in translated form. As to the specification of the task, readers are at the mercy of specialist multilingal information scientists or translators who can act as information advisors before actually translating. It is at this level of work that human translation shows its absolute advantage over machine translation. While machines can only translate texts without adjustment of intention, translators have two options: they can translate the text with the original intention, thus putting the reader in the role of secondary reader; or they can modify the text so that it fits the reader as primary reader. The ability to transform the intention of the source document for either purpose depends partly on the translators' skill to interpret the intention of the source document and partly on their understanding of the readers's needs. In translations made without concern for the readers's specific needs, i.e. in cases where they are clearly secondary, readers may have to struggle with intention signals addressed to primary readers. For secondary readers the writer's contribution to the success of the communication diminishes, whereas the translator's skill in interpreting reader expectation and meeting it through a change of the linguistically expressed intention signals grows in importance. Translators as information mediators Extending the idea of primary and secondary readership to the functions of translators themselves, we can say that translators perform both functions simultaneously by adopting one or the other role as required by their dual responsibility to the document in hand and the document they are to produce. When they read a document in order to understand its content and intention they adopt the role of primary reader while in actual fact they are always and by nature secondary readers. When, during the actual process of translating, they

30

TEXT TYPES AND TRANSLATION

are writers of new documents, they will have assessed whether the original message is appropriate for the target language readership or whether it requires a change of intention. For example, instead of producing full-text translations, specialised translators working for specified clients regularly produce abstracts of scientific papers, précis of discussions, summary translation of business letters or relevant current awareness information extracted from documents in a source language. These examples also show that modification of content, e.g. through abstracting, must also be accounted for in a translation theory. In regular dealings with documents written in other languages, most professional readers have already developed habits of expectations which they are able to convey to translators through the job specifications. It is the interaction between habitual readers and translator in their role as information mediators which ensures the most pertinent information transfer. Text types In order to carry out the task of information mediators, translators need unambiguous means of identifying intention and, equally important, means of expressing it. The most visible means of expressing intention is through the choice of conventional text types. Text types have evolved as patterns of messages for specific communicative situations. When we write a message we first think of the text type that is suitable for the occasion and the content, and formulate our text accordingly. Regular repetitions of messages in particular circumstances have created expectations of recognisable structural and rhetorical features which condition our modes of reading a message. When we receive a message, we first identify the text type because it permits us to tune in to the appropriate mode of reception. For example, we know that manuals and grammars have a directive intention and are for repeated reading and for reference; application forms are pre-structured and writers only have to fill in slots and readers only have to read selectively. If a writer sends a document directly to a known reader, as, e.g. in hand-written correspondence, the original intention of the writer coincides with the expectation of the reader. In industrial environments where documents are frequently the result of multiple authorship, substantial editing and revision, text types are related through alternate uses in a variety of communicative situations and through modifications to make them suitable for different purposes. Shipping documents are a good example of the various purposes certain complex documents have to adjust to. While the lay person usually associates text type with content, e.g. medical report, police report, book review (a report on a book), cookery recipe (an instruction), in many cases the same content permits a variety of text types. There are, for example, a

J. C. SAGER

31

great variety of reports on different topics which have the same overall text structure.4 Knowledge content can also be expressed by non-linguistic means and many messages consist of combinations of such texts and graphics, tables, charts, illustrations etc. In summary we observe that categories of texts arise from conventionalised communicative situations. They can be highly effective in conveying information unambiguously because they result from common social and knowledge relationships between writer and reader. They are characterised further by topic, and mode of expression, as, for example, in the numerous subcategories of reports, many of which are written to more or less strict guidelines. Translation adds a further dimension of complexity to text types because of varying formats as required by different national authorities. Translation studies have accepted the need for analysing text types. Neubert (1985: 125) offers an interesting definition of text types by calling them "socially effective, efficient, and appropriate moulds into which the linguistic material available in the system of a language is recast". Wilss (1977: 135-9) surveyed the role that text types have increasingly been playing in translation theory, but also concluded that text types were mainly studied in order to determine translation methods or degrees of translatability. Reader expectation and text types Since the meaning of the text type of a document is the first impact of a message on a reader, the recognition of a particular text type conditions the reader's response to the message. The most common way of recognising text types is through the situation and the compositional features of the text. Readers recognise broad categories such as schedules, reports or essays or more specific ones, such as a railway timetable, a weather report, or a newspaper sketch, both through the textual structure and the situation. The text type tells readers who addresses them and how they are addressed, as an individual or as a group. Identifying a message as a schedule, readers know that the intention is purely informative and usually requires some prior familiarity with the conventions of expression of that broad text type. For example, airline timetables cannot be consulted by the general reader without previous study of the many symbols explained in the introduction. Reports, on the other hand, can be read, and hence translated, selectively, e.g. the conclusions and findings only, because the conventions of this text type require that the main parts of reports be written as self-contained units. Readers of memos know that the intention is directive or instructional or simply informative, according to its subtypes. In an essay, readers know that the intention is to develop an argument by which the writer hopes to persuade and influence the reader. Readers can, however, react to

32

TEXT TYPES AND TRANSLATION

documents they receive from two different positions. If they consider them as personal messages, they must receive them in relation to their individual background, involving the writer and possibly a separate sender, because a specific sender has attached a particular intention and/or the recipient has associated a text with a certain expectation in the pursuit of communication. If, on the other hand, readers are not directly addressed by a particular message, it becomes simply a text for them. When a text is divorced from writer and reader it finds itself outside its original pragmatic circumstances and it is then simply an item of writing which may be re-used with a different intention by another writer, by any reader, sender or recipient, or a translator, for whom it is then only a draft document. There are many historical documents in existence about whose origin and original intention we have little, if any, information and can only surmise their original intention, but our interest in them is always motivated by a different expectation, a fact which leads to misinterpretation of historical records. For us, such documents are deprived of their original message function and we can only study them as static objects, i.e. as texts. Since translators cannot always reconstruct the situation, a communicatively effective translation will ensure that readers readily identify the text type in order to adjust their expectations. Our preliminary conclusion of this section must be that the dynamic view of the translation process does not only require our acceptance of a change of linguistic code, but also the possible need for modifications of content (addition and reduction) and change of purpose (writer intention and reader expectation) in order to give the readers of translations the same advantage as that of a primary reader.

New types of translations and translation-related work These observations oblige us to enlarge the scope of the basis on which we construct a theory of translation. If we consider the source material for translation to consist of a content, a linguistic form and a purpose (a writer intention and a reader expectation) we now accept that, beside the obligatory change of linguistic form, the other two elements can also undergo modifications in translation. Applying the possibilities of changes in content and intention to the whole field of professional translation, it appears necessary to single out several new translation strategies which have hitherto not been fully recognised in the theoretical literature. The conventional concept of translation is represented by the strategy which aims at preserving both content and intention. This is applied when we trans-

J. C. SAGER

33

late a letter into a letter, a technical manual into another technical manual. This strategy suffices for such items as learned articles and other communications in the fields of science and technology, as well as for bilingual documents in countries with a bilingual regime. The success of the full transfer of content and intention in a translation can be measured by comparing the effect of the translated text on the new readership with the effect the source document had on the original readership because in such cases both linguistic communities will have a very similar culture. In this type of translation the fact that a text is translated may, but need not be acknowledged, because it is unimportant. Translation strategies changing content The content level of source texts is no longer constant if we accept such translation products as commented translation, translingual abstracts, and gist translations. The moment translators know the requirements of the new readers, they can judge which sections of the source document are relevant and which are conveniently summarised, abridged or omitted altogether. If we accept that the intergrity of the source text content is no longer the sole criterion of the quality of a translation, and at the same time give translators the dynamic role of mediators, we can also expect them to adjust their work more closely to the needs of secondary readers, or indeed, where suitable, upgrade the end users to the class of primary readers. There is obviously a danger in elevating omissions of text in translation to a general principle, and we certainly do not want to give the impression that we advocate the type of omissions that frequently occurs in translations because of translators' negligence, forgetfulness or deliberate avoidance of difficult passages. But, it must be recognised that in certain circumstances and for certain uses, simple omissions of parts of a source text can improve the quality of a document for the reader. It is therefore here proposed to recognise as a new type the translation strategy which aims at preserving the intention but accepts the need for reduction or addition of content, which sometimes also involves some modification of content. Typical forms resulting from this strategy are commented translations, annotated translations, or alternatively various forms of reduction or concentration of content. Theoretically this work involves two stages: reduction or expansion followed by translation. In practice, unless a source text reduction is available or usable, translators carry out both operations simultaneously because decisions about reduction directly involve major restructuring of sentences which is best done within the process of translation. A specific new form of content modification is the "gist translation" of, for example, letters in

34

TEXT TYPES AND TRANSLATION

which the translator only summarises the content. Gist can also be conveyed orally in the form of on-sight excerption of the essence of a message. An expanded translation, with footnotes or additions in the text itself, does not change in intention if it only aims at supplementing the new readerships' prior knowledge for the full understanding of the text. Strictly speaking, many translations of literature which contain elucidations of culture-specific items of the source language also fall into the category of expanded translation, but the necessity for this intervention is so widely recognised that it is not considered as representing a modification of content. Since many changes of content are accompanied by changes of intention, we shall exemplify content modification more fully below. Translation strategies changing intention Unless the intention of a document they receive for translation is self-evident, translators have to identify and define the new pragmatic communicative situation so that the appropriate text type can be selected. Among the text types available to translators we have to distinguish text types known in the target language, including types which have come into a language via translation, but are now fully accepted as genuine forms, and text types which are introduced into the target culture only through a new act of translation and which continue to have this status. It is the distance between the communicative situation of the source text and the situation of the target text which is mainly responsible for changes of intention and hence changes of text types. For example, oral or impromptu translation of selected passages of a scientific paper changes the intention just as much as the written translation of a transcript of a political speech destined to an audience at an open air rally. A change in the intention of a text can lead to the same content being expressed in two different text types. For example, an instruction in a source language may have to be accompanied by a lengthy introduction for a particular type of target language reader but not for others. The same content can be expressed in an informative leaflet and in rules, regulations or instructions. What is commonly understood as translation "for information only" -- i.e. the provision of a working document for internal use of a firm or personal use of a client, widely accepted in industrial and scientific circles -- clearly belongs into this category, but there are various interpretations of this loose label which is applied to a number of different text types and their translation. Another classic case of change of intention are the prose translations of poetry, which clearly destroy part of the aesthetic function of a text in order to preserve another part.

J. C. SAGER

35

Translators therefore need to know whether the target language message is to be given the same or a different intention than the source document. If the intention changes, there are certain limitations on what target language text types can be derived from source language text types. When, for example, the source text is highly situation-bound, its intention may not be transferable to the new communicative situation created through the translation. This then may be a reason for a text to be declared "untranslatable". The second new translation strategy which should be fully recognised is, therefore, that of translation involving a change of intention. A translated text may be intended to inform rather than to instruct. Such changes usually require a change of text type and with it a number of changes in the content and the structure in which the content is presented. For example, proceedings of a conference may be converted into a report, a report may become a summary, minutes of meetings may be conflated into a report. To the extent to which such activities are not first carried out on monolingual texts, we are also dealing with new translation-specific text types which are governed by their own conventions and expectations of quality. Translations involving a change in intention, proceed in two stages. Firstly, there has to be the selection of a suitable text type which may but need not be the same as that of the source text. Only then can translation proceed because, in case a non-matching text type is chosen, the content may also have to be restructured. Translation strategies changing content and intention The logically third new type of translation strategy involves both a change of intention and of content. It is most strikingly exemplified by target language abstracts of source language documents, adaptations and commentaries. This complex activity has become routine in the case of translingual abstracts whose intention is reflected in the many subcategories such as indicative, evaluative, slanted abstracts. The selection of content requires translators to bring specialist subject knowledge to bear on the process. Many such forms are conventionalised to the extent that they have prescribed reductions in relative or absolute length of text and other rules for the treatment of content. Finally a change in the content of a text can lead to a change in its intention. For example, a simple informative text may become imperative for a particular group of readers and may therefore need to be expressed in the form of an instruction. To this group also belong the translingual compte rendu, minutes and transcripts as they are practised at multilingual gatherings. The conversion of spoken source language messages into written form seeks to represent the original as closely as possible but within established conventions about the

36

TEXT TYPES AND TRANSLATION

conversion of spoken discourse into written sentence structures. Changes of both intention and content occur in the following situations: — When a source language document is steeped in its particular environment and specific to its culture it must undergo changes to become acceptable to a new readership. First, it must be given a new intention so that it becomes relevant. This may involve a change of text type. For example, a piece of legislation only applicable to the source language culture can be given a purely information intention and so become of interest to another culture. Secondly, the document must be annotated, edited, rewritten in translation in such a way that a target language reader finds his expectations satisfied. This can be done in two ways, similar to the traditional writer/reader orientation, by either bringing the reader close to the understanding of the original purpose of the source document, or by making the new document wholly relevant to the readers' current requirements. — When a source document has ceased to be a message to anyone, the text can be made available to anyone in another culture for any purpose. For example, historical records are now studied for what they can tell us about our present situation; ancient text fragments like the Bible in Gothic are now used to write a grammar of that language and no longer as a religious text. New text types as a result of computer involvement in translation

The advent of machine translation produces new situations for translations which also have an effect on text types because we cannot necessarily equate the language produced by computers with that produced by human translators. The product of an entirely human process of translation now has to be called "human translation" in order to differentiate it from the product entirely made by computers, which is called "machine translation". The language used by human translators is a "natural language", the language produced by a computer is a form of "artificial language". An artificial language is defined in the BSI Glossary of Documentation Terms (BS5408) as "a language whose rules are explicitly established prior to its use". Since these rules are usually conceived at the sentence level only, it follows that they are inefficient in making distinctions at the level of larger structures of text types. We can therefore identify the fundamental difference between human and machine translation as being the nature of the language in which they are written, irrespective of quality. The other major characteristic of machine translation, which makes all such products into text types of a different class, is the fact that machines can neither recognise nor express changes in the intention of a document. Only human translators can deal with the intention of documents. Computers can therefore be said to produce "texts" deprived of intention, and readers of translations

J. C. SAGER

37

produced by computers have to interpret such texts according to their own expectations. The comparison with machine translation also serves to show the full range of complex tasks translators perform which cannot be replaced by computer programs. In simplified form we can say that the human translator "understands" a document, whereas the computer simply processes linguistic information. In particular, the human translator has to deal with implicit information not contained in the document, whereas the computer can only deal with what is "explicit" in the input of linguistic substance. In addition, the human translator has knowledge which is not contained in dictionaries or grammars which, on the other hand, is the only information available for the computer because it is sufficiently codified to be programmed. ! In other words, the conversion from a natural language to the artificial language of the output of any machine translation system also implies a restriction in function and pragmatic scope. While it may be possible to fine-tune a machine translation system to be optimally efficient for certain text types, this is so far of limited scope, because our knowledge of characteristics of text types is not advanced enough to fully codify it in the form of computer programs. A translation produced by a computer is, therefore, another translation-specific text type. It is written in the code of an artificial language invented by the system designers on a restricted model of a natural language. Machine translation products also have their subtypes, but these are the result of differences in the artificial language codes which distinguish one machine translation system from another. In this largely theoretical discussion we exclude from consideration the many varieties of machine-assisted translation because in all these cases the human translator is in control and may ultimately, through editing, convert the machine output back into natural language.

A Typology of translations In order to accommodate the full range of translation products discussed here, we also need a wider definition of the concept of translation. Here is an attempt to offer a more comprehensive definition: Translation is the production of derived documents in a target language which are dependent for their content upon another text or message in a source language, with the restriction that the core knowledge contained in the derived target language document cannot in essence exceed the knowledge contained

38

TEXT TYPES AND TRANSLATION

in the source language document. Any document, even derived documents themselves, can become the source of another derived document, e.g. abstracts of translations and translated abstracts. This definition, which, admittedly, still leaves open what is to be understood as "core knowledge", suggests that it is preferable to treat translation as a generic concept, subdivided by a number of facets, such as - the mode of production (human/machine, with the concomitant conversion from natural to artificial language output); - change of content (full/modified by reduction or addition, with concomitant change of intention and possible text type); - change of intention (same/different, with concomitant change of text type). In all cases translators must consider the source texts they are modifying, i.e. transforming their language code, content and/or intention, as drafts for a new product destined for a new reader. This point of view eliminates the possibility of evaluating translations on the basis of such criteria as fidelity, accuracy, etc. which all involve some form of "equivalence" and replaces it by the single measurement of appropriateness for the intended use by the end reader. We have shown that text type is closely related to change of intention and choice of translation strategy. These relationships are most closely evident in strategies leading to translation-specific text types. On this basis we offer here a new major classification of translation products, the distinctions being based on production strategies. 1. Translations which convey the writer's intention unaltered and seek to widen the primary readership by extension to another language and culture. In this case we may assume that the translator either acts in agreement with the writer and functions as his mouthpiece, or he is contracted independently to correct what may be termed "a writer's misapprehension of the readers' ability to read his language". This translation strategy applies to most literary translation and also to instructional and directive texts. 2. Translations which are adjusted to satisfy reader expectation. In this case we may assume that the writer was or is unaware of the possibility of a secondary readership, and translation adapts the document to a particular group of secondary readers. The basic strategy within this major type is simply to change the linguistic code. However, where the secondary readership expectations differ from that of primary readers, translation then introduces other necessary modifications of the content and/or the text type of the document. Examples of such translations are gists, summaries, texts "for information only" etc.

J. C. SAGER

39

Another facet of classification, based on text types, which influences translation strategies, must differentiate between target language text types which are fully established in the target language culture and those which are specific to translation products and occur only as a result of translation. We therefore have two classes: 1. Existing target language text types Documents written in a particular text type and addressed to a specific primary readership can be translated into parallel text types known in the target culture and used for conveying the same intention. But they can also be translated into other text types known in the target culture if this is required by users. For example, an environmental health regulation which is directed at farmers in one culture cannot have the same impact in another culture where such a health problem is unknown and no such regulation exists; but it can be translated as an informative document, say, with suitable annotations so that readers perceive it as information rather than as an incomprehensible and inapplicable directive. 2. Translation-specific text types Different cultures may have different sets of text types because they have evolved different patterns of communication. Causes for such differences have to be sought in different degrees of literacy, poetic traditions, administrative arrangements, legal procedures etc. Translations strategies based on constant content, intention and text type assume that the respective target language has the equivalent or parallel text type. If this is not the case, the concept of equivalence between source and target language text can be usefully applied only to smaller textual units than the whole document. The finished translation is then no longer a text type inherent to the target language, but a product sui generis, a type of translation, expressed in a form that is novel to the target language community. This is, most obviously, the case for all unedited machine translation output5. Original new translationspecific text types can also be the result of reduction of longer source documents, as is the case with gists. We must finally acknowledge that many cultures have imported text types through translation, many of which have now become fully integrated into the target culture making it possible for future translations to be cast into preexisting types. The most obvious example is the introduction of written language forms through the language planning efforts carried out by missionaries. The need for translating the Bible required the codification of oral languages in different cultures, thus giving them the first written form of their language,

40

TEXT TYPES AND TRANSLATION

with dictionaries and grammars to regulate it. Subsequent commercial contacts with countries of written traditions then introduced many functional text types so that at certain levels of communication there is a fair chance of finding suitable parallel text types. This is, of course, not the case for innovative literature which differentiates cultures rather than uniting them. Notes 1 An earlier version of this paper appeared in: Sager 1996. pp. 43-53. 2 A full treatment of communication theory relevant to translation is given in Sager (1994) chapters 3-4. 3 By contrast, interpreters always deal with messages because, unlike translators whose work interrupts a communication process, they are direct participants in the communication process. 4 A full analysis of over 100 text types can be found in Sager, J.C. et al. (1980) chapters 6-7. 5 The growth and further development of machine translation may lead to yet further translation-specific text types. It is already common to identify translations produced by computers, so that readers know what they are getting and can adjust their expectations accordingly. References Baker, M. 1992. In Other Words: A Course Book onTranslation. London: Routledge. BSI 5408 1976. BSI Glossary of Documentation Terms. London: British Standards Institute. Jäger, G. 1975. Translation und Translationslinguistik. Halle: Niemeyer. Kade, O. 1968. Zufall und Gesetzmässigkeit in der Übersetzung. In Beiheft Fremdsprachen I. Leipzig: Verlag Enzyklopädie. Koller, W. 1992. Einführung in die Übersetzungswissenschaft. Heidelberg: Quelle & Meyer. Neubert, A. 1985. Text und Translation. Leipzig: Verlag Enzyklopädie. Nida, E. 1964. Toward a Science of Translation. Leiden: Brill. Nord, M. 1991. Text Analysis in Translation. Theory, methodology and didactic implications of a model for translation-oriented text analysis. Amsterdam: Rodopi. Reiss, K. 1976. Texttyp und Übersetzungsmethode. Kronberg: Skriptor. -- 1984. "Adäquatheit und Äquivalenz". In W. Wilss and G. Thome (eds), Translation Theory and its Implementation in the Teaching of Translation and Intepreting. Tübingen: Narr. pp.80-89. Reiss, K. and Vermeer, H.J. 1984. Grundlegung einer allgemeinen Translationstheorie. Tübingen: Niemeyer. Sager, J.C. 1996. "Stretching the limits of translation". In Übersetzungswissenschaft im Umbruch, Tübingen: Gunter Narr, pp. 43-53. -- 1994. Language Engineering and Translation. Amsterdam: John Benjamins.

J. C. SAGER

41

— et al. 1980. English Special Languages - Principles and Practice in Science and Technology. Wiesbaden: Brandstetter. Toury, G. 1980. In Search of a Theory on Translation. Tel Aviv: Tel Aviv University. Wilss, W. 1977. Übersetzungswissenschaft: Probleme und Methoden. Stuttgart: Klett. -- 1980. "Semiotik und Übersetzungswissenschaft". In W. Wilss (ed), Semiotik und Übersetzen. Tübingen: Narr, pp. 9-22. -- 1995. Knowledge and Skills in Translator Behaviour. Amsterdam: John Benjamins.

A Functional Typology of Translations Christiane Nord University of Hildesheim

Introduction: The crux of equivalence In her book Translation Studies: An integrated approach, Mary Snell-Hornby distinguished "two main schools of translation theory which now dominate the scene in Europe" (Snell-Hornby 1988: 14), one of which can be said to represent the linguistic approach, and the other the functional approach to translation. Although we may distinguish a considerable number of different approaches today - such as the descriptive approach, the feminist approach, the literary criticism approach or the translation-and-power approach, to name but a few - the fundamental opposition between linguistic and functional approaches does not seem to have been overcome so far. Linguistic translation theories hinge on the concept of equivalence, which has been one of the most ambiguous concepts in translation studies from the start. Consequently, it has been interpreted in very different ways, depending on which aspects of source and target text are required to remain invariant in the translation process: "linguistic elements" (Oettinger 1960: 110), "textual material" (Catford 1965: 21), the "message" (Nida and Taber 1969: 12), the "communicative value" (Neubert 1985: 138ff.), etc. As early as 1947, Eugene A. Nida coined the term "dynamic equivalence" (as opposed to "formal equivalence") in order to make clear that equivalence is not a static relationship between elements of language systems (cf. Nida 1964: 159). In 1979, Koller specified even five types of equivalence: denotative, connotative, text-normative, pragmatic and formal equivalence (Koller 1979: 187ff.), and some years later Neubert speaks of "text-bound equivalence" (Neubert 1984: 68 and 1986: 87ff.), which the translator constantly has to strive after and which may make up for non-equivalent translations on lower ranks (e.g. at the level of words or phrases). Snell-Hornby (1986: 13ff.) claims to have located 58 different types of "Äquivalenz" referred to in German trans-

44

A FUNCTIONAL TYPOLOGY OF TRANSLATIONS

lation studies which, at the same time, must not be equated with the use of the term "equivalence" in English translation studies (nor with "équivalence" or "equivalencia", for that matter). Under the heading "Equivalence could be all things to all theorists", Pym more recently (1992: 37) bases his notion of equivalence on the economic term "value" and claims that equivalence "is still what happens, on one level or another, whenever a translated text is received as if it were a merely transferred text", "given that what is exchanged, what the specific reader ideally wants and receives, is ultimately a representation of that part of Y [= the transferred text, C.N.] which is considered to be of value in the particular exchange situation concerned" (emphasis added, C.N.). In my experience, such redefinitions do not solve the problem of the inherent fuzziness of the concept, and to allow everybody to use their own concept of equivalence would not really help translation scholars understand each other either. The fundamental drawbacks of the equivalence model, as it is generally understood, lie in the following characteristics: 1. Although there are sporadic references to pragmatic aspects (such as function or communicative effect), the equivalence model focuses mainly on structural qualities of the source text, losing the intrinsic interrelationship between extratextual (i.e. situational) and intratextual (i.e. linguistic) factors of communicative interaction out of sight. 2. This is why cultural aspects do not come sufficiently into consideration, although language can be regarded as an intrinsic part of culture, and languageusers cannot but behave in a culture-specific way. 3. Considering the divergent definitions of its basic concept, the equivalence model lacks consistency. Some scholars praise "literalism" as the best way to secure equivalence (Newmark 1984/85: 16). Others, such as Koller, allow a certain amount of adaptive procedures, paraphrases or other non-literal procedures in specific, well defined cases, "where they are intended to convey implicit source-text values or to improve the comprehensibility of the text for the target readership" (Koller 1993: 53; my translation). These are two rather arbitrary criteria which do not take account of the fact that sometimes implicit values should remain implicit and that comprehensibility is not a general purpose common to all texts or text types. 4. It is interesting to note, moreover, that any non-literal translation procedures are more readily accepted in the translation of pragmatic texts than in literary translation. Thus, different or even contradictory standards for the selection of transfer procedures are set up for different genres or text types, which makes the model rather confusing.

CHRISTIANE NORD

45

5. Being based on an apparently "universal" concept (although this seems to be an illusion, too, as has been pointed out above), the equivalence model does not account for culture-specific differences in translational concepts. As a human activity which takes place within the boundaries of a particular culturecommunity, however, translation is bound to be guided by culture-specific norms and conventions, although the differences may not be striking between some culture-groups or even within larger culture-areas, such as the so-called "average Western culture". 6. The equivalence model excludes target-language texts which do not satisfy the criterion of equivalence, such as interlineal versions, philological translations, resumé translations or adaptations, from "translation proper" (cf. Neubert 1985: 162, and, more recently, Koller 1993), regarding them as "non-translations" or adaptations (cf. Jäger and Müller 1982: 55f), although it is a wellknown fact that such forms are asked for in professional translation practice. 7. In the equivalence model, the source text and its "value(s)" (?) are considered to be the one and only standard, to which the translator has to subordinate any decision in the translation process. Therefore, the model seems to perpetuate the low social prestige of the translator, whose activity is usually thought to be a kind of "nurturing profession" (Pym 1993: 55).

Functionalism + loyalty as an alternative model Since obviously the equivalence model is not adequate to the needs of professional translation in a modern society, let us look for an alternative. What we need is (1) a pragmatic model which takes account of the situational conditions of communicative interaction and, accordingly, of the needs and expectations of the addressees or (prospective) receivers of the target text, (2) a culture-oriented model giving consideration to the culture-specific forms of (verbal and nonverbal) behaviour involved in translation, (3) a consistent model able to establish a coherent theoretical and methodological framework which may serve as a guideline for an intersubjective justification of the translator's decisions in any type or form of translation task, (4) a comprehensive model which can be applied to all text types, both literary and non-literary, and to translations into and out of the foreign language, (5) an "anti-universalist" model which allows for culture-specific differences in translational concepts,

46

A FUNCTIONAL TYPOLOGY OF TRANSLATIONS

(6) a practical model capable of accounting for all forms of transcultural communication needed in professional translation practice, (7) an expert model which gives translators the prestige of being experts in their field, competent to make purpose-adequate decisions in full responsibility towards their partners. The skopos model In my opinion, Hans J. Vermeer's functional "skopos theory" is able to meet most of these requirements (cf. Vermeer 1978, Reiss and Vermeer 1984). It parts from the view that translation is a form of human interaction and, as such, determined by its purpose or skopos. One of the main factors in the skopos of a communicative activity is the (intended) receiver or addressee with their specific communicative needs (cf. Nord 1996). This applies both to the source text (ST) and the target text (TT). Being addressed to a specifically source-cultural audience, for whom it is intended to have a particular communicative function, the ST and its linguistic, paralinguistic and non-linguistic features are conditioned by the norms and conventions prevalent in the source culture (SC), whereas the TT, addressed to target-cultural readers/listeners and intended to be meaningful and functional for them, should naturally conform to the norms and conventions of the target culture (TC), taking into account what target-culture members can be expected to know or feel about the subject in question. Thus, the focus is shifted from the source text to the target text and its communicative function or functions. It is no longer the ST which sets the standards for the translator's decisions in the translation process, but the intended receiver of the translation, whose reception will be entirely guided by TC expectations, conventions, norms, models, real-world knowledge, perspective, etc. This model is pragmatic in that it takes target-orientation seriously and even makes the target receiver the most important determinant of translational decisions. It is culture-oriented because it considers translation as a "cross-cultural event" (Snell-Hornby 1987: 82). Translation is not the transcoding of words or sentences from one language to another, but a complex form of action, whereby someone provides information on a text (source language material) in a new situation and under changed functional, cultural and linguistic conditions, preserving formal aspects as closely as possible (Vermeer 1986: 33, transi. Snell-Hornby 1987: 82).

CHRISTIANE NORD

47

Moreover, the functional model can be regarded as consistent because it permits any translation procedures which will lead to a functional target text, and it is comprehensive because target function is considered to be the main standard for any translation process. It is practical because it can be applied to any assignment occurring in professional translation practice. Last, but not least, it is a model which gives due consideration to the role of the translator as an expert. This last aspect is particularly important for translator training. We do not want to train submissive "servants" who do what they are told by the source text (what is a text that it can "tell" people to do something), but fully responsible partners in a cooperative interaction between equals - no less nor more. Functionality plus loyalty There are only two reservations which I would like to make with regard to the skopos model in the form it has been presented here. One concerns the culturespecificity of translational models, and the other refers to the relationship between the translator and the ST author. Like the equivalence concept, the skopos theory claims to be a general model (see the title of Reiss and Vermeer 1984). Nevertheless, Vermeer mentions that "normally" we would expect a relationship of analogy or similarity to hold between ST and TT, which he calls "intertextual coherence" or "fidelity" (Vermeer 1978: 59); however, the demand for fidelity is subordinate to the skopos rule. If the skopos demands a change of function, then the required standard will no longer be intertextual coherence between ST and TT but adequacy or appropriateness with regard to the skopos (Reiss and Vermeer 1984: 139). The main idea of the skopos model could be paraphrased as "The translation skopos determines the translation procedures". Now, this seems to be acceptable whenever the translation skopos is in line with the communicative intentions of the original author. But what happens if the translation assignment requires a translation whose communicative aims are contrary to or incompatible with the author's opinion or intention? In this case, the paraphrase mentioned above could be interpreted as "The end justifies the means", and there would be no restriction to the range of possible "ends". In a "general theory", this could be accepted because we can always argue that general theories do not have to be applicable. But translation does not take place in "general", a-cultural or supra-cultural surroundings. Translators or interpreters act within a particular culture community, and therefore any practical application of the general theory must take this into consideration. According to the (usually implicit) culture-bound concept of what a translation is or should be, people (i.e. initiators, readers, authors of originals) in fact

48

A FUNCTIONAL TYPOLOGY OF TRANSLATIONS

expect particular relationships to hold between ST and TT, which often, vary according to the text type in question. They might expect, for example, that the TT gives exactly the author's opinion or that it is a "faithful" reproduction of the formal ST features or that it is precisely not a faithful reproduction but a comprehensible, readable text. These expectations must be taken into account by the translator. This does not mean that translators are always obliged to do exactly what the readers expect - but there is a moral responsibility not to deceive them (cf. Nord 1991: 94f.). This responsibility translators have towards their partners in the translational interaction is what I call "loyalty". Loyalty commits the translator bilaterally to the source and the target side. Loyalty must not be mixed up with "fidelity" or "faithfulness", which usually refer to a relationship of similarity between texts or even surface structures of texts, it is an interpersonal category referring to a social relationship between individuals. In the model, it is a sort of empty "slot" which, in a particular translation task, is filled up by the demands of the specific translational concepts of the cultures in question. If, for example, the target-culture expects a translation to be a literal reproduction of the original (although this might not be a functional translation from the source-culture's point of view), translators cannot simply translate in a non-literal way without telling the target audience what they have done and why. It is the translator's task to "mediate" between the two cultures, and mediation must not mean "cultural imperialism" (i.e., culture A pretending that their concept is superior and must, therefore, be adopted by culture B). Introducing the loyalty principle into the skopos model would also solve the second problem I see in "radical" functionalism. It concerns the relationship between the ST author and the translator. Normally, authors are not experts in translation, and therefore they are likely to insist in a "faithful" rendering of the ST surface structures. Only if they believe in the translator's loyalty, will they consent to changes or adaptations which may be necessary in order to make the target text "work" in the target-culture situation. And this confidence would again strengthen the translator's social prestige as a responsible and trustworthy partner. Therefore, the functional model presented here rests on two pillars: "functionality" and "loyalty" (see Nord 1991: 28ff. and 1993: 17ff.).

CHRISTIANE NORD

49

The notion of function General considerations Since functionality is one of the two central issues in this concept of translation, we have to consider what is meant by "function". A text is a "communicative occurrence" (de Beaugrande and Dressier 1981: 3). An essential prerequisite for such a communicative occurrence is, first of all, the existence of a "situation", fixed in time and space and comprising at least two participants who are able and willing to communicate with each other for a certain purpose and by means of a text (i.e. a combination of communicative signs). The text is transmitted over a suitable channel or medium, which we consider part of the situation, and (ideally) has the function of fulfilling the communicative purpose intended by the sender. If a text is regarded as a combination of communicative signs exchanged between sender and receiver, we can analyse text function from either the sender's or the receiver's point of view. The sender intends to achieve a certain purpose and therefore chooses certain strategies of text production considered appropriate for this purpose, using structural features in order to "signal" their intention to the addressee. But as we all know, the best of intentions does not guarantee a perfect result. It is the receiver who "completes" the communicative action by deciding to receive (i.e. to use) the text in a particular function. Whenever I refer to "function" in this context, I mean "function or set of functions" because texts are rarely intended for one function only. Various functions usually form a hierarchy of functions, subfunctions, etc. (cf. Nord 1990). The use the receivers make of the text depends on their individual interpretation of the situational markers and on their expectations towards the text. These are determined by general background knowledge, communicative needs, the communicative situation in which the text is received and - last, but not least - by the signals of the sender's intention found in the text. The sender's intention and the receiver's expectations may be identical, but they need not necessarily coincide nor even be compatible. Text function is, therefore, a pragmatic quality assigned to a text by the receiver in a particular situation and not something attached to, or inherent in, the text. Thus, it seems only logical that the function of the source text is specific to the original situation and cannot be left invariant or "preserved" through the translation process. The function of the target text, on the other hand, is specific to the target situation, and it is an illusion that a target text should have automatically the "same" function as the original.

50

A FUNCTIONAL TYPOLOGY OF TRANSLATIONS

A model of basic text functions There are several models of text functions which could be used as a frame of reference for functional analysis. To make things easier in translator training, I suggest a combination of the models elaborated by the German psychologist Karl Bühler (1934) and the Czech functionalist Roman Jakobson (1960), establishing four basic textual functions (including various subfunctions, among them Jakobson's metalinguistic and poetic functions). (1) Referential function (i.e. reference to objects and phenomena of the world). Some subfunctions: informative function (object: e.g. a traffic accident), metalinguistic function (object: e.g. a particular use of language), instructive function (object: e.g. the correct way of handling a washing-machine), teaching function (object: e.g. Geography) etc. Example 1: Directions for Bottling Fruit 1. Place the clean, warm jars in a large bowl of boiling water. 2. Pack the jars firmly with fruit to the very top, tapping jars on a folded cloth or the palm of the hand, to ensure a tight pack. 3. Fill the jars with boiling water or syrup to within1/4inch of the top. [...] (2) Expressive function (i.e. expression of the sender's attitude or feelings towards the objects and phenomena dealt with in the text). Some subfunctions: emotive function (expression of feelings, e.g. in interjections), evaluative function (expression of evaluation, e.g., in a political commentary). Example 2: Simone de Beauvoir: Une mort très douce. - Engl.: A Very Easy Death. Both versions of the title have an expressive function. The original French title expresses a personal emotion seen from "inside" the dying person ("douce"), whereas the English translation expresses an evaluation: whether a death is "easy" or not can be judged from outside, e.g. from the doctor's perspective. (3) Appellative function (i.e., appealing to the receiver's experience, feelings, knowledge, sensibility etc. in order to induce him/her to react in a specific way). Some subfunctions: illustrative function (intended reaction: recognition of something known), persuasive function (intended reaction: adopt the sender's viewpoint), imperative function (intended reaction: do what the sender is asking for), pedagogical function (intended reaction: learn certain forms of behaviour), advertising function (intended reaction: buy the product). Example 3: If you're an American living abroad and you need to keep track of your calls, you really ought to get the AT&T Card. First of all, you get a monthly itemized bill. A new option even lets you bill your AT&T Card calls

CHRISTIANE NORD

51

to your American Express Card account. [...] For an AT&T Card application, call us collect at 816 654-6004 Ext. 60, or write to AT&T Card Operations, P.O. Box 419395, Kansas City. (Newspaper announcement) (4)Phatic function (i.e., establishing, maintaining or finishing contact). Some subfunctions: salutational function, "small-talk" function, "peg" function (e.g. text introductions, such as the allusion to a proverb in the following example). Example 4: Hotel list HERZLICH WILLKOMMEN IN BREMEN! Wie man sich bettet, so schläft man, sagt ein Sprichwort. Dabei wollen wir Ihnen, lieber Gast, mit dieser Hotelliste behilflich sein. [...] WELCOME TO BREMEN! There is proverb [!] which says "As you make your bed so you must lie on it". That is why we hope that this Hotel List will be of service to you for your stay in Bremen. [...] The English translation cannot achieve the phatic function. The quoted phrase, which is not an English proverb, but a "pseudo-proverb" based on the idiomatic expression "you have made your bed and you must lie on it" (= "you must accept the bad results of your action", DCE 1978), seems to inform the readers that it was not a good idea to come to Bremen. Instead of the phatic function intended by the ST sender (and probably expected by the reader), the English TT has an informative function and a secondary appellative function (= "make the best of it!") which do not correspond to the sender's (conventional) intention. From a functional point of view, a semantic paraphrase like "After a long day of sight-seeing you need a nice place to rest" or any other proverb, quotation or introductional phrase would have done better than this "false friend" (not to mention the grammatical mistakes). A functional typology of translations While the function of the original is defined with regard to its own referent and situation, the function of the translation has to be first defined with regard to the source text. Here, we can distinguish two translation "types", which have split translation theorists into two camps since the days of Cicero (a detailed description of the typology is given in Nord 1989). The translation can be (1) a document of the situation in which an SC sender communicates with SC receivers via the source text, focussing on one or various aspects of the text, which are then reproduced in the target language, and

52

A FUNCTIONAL TYPOLOGY OF TRANSLATIONS

(2) an instrument in a new TC situation in which the ST sender communicates with TC receivers via the target text which has been produced under TT conditions using the source text as a kind of model. A "documentary translation" is a kind of metatext marked as a translation (e.g. by stating the source and/or the name of the translator), whereas "instrumental translations" are object texts which can serve any function a non-translated text can achieve. Documentary translation Interlineal version Literal translation Metatextual function Philological transi. Exoticising transi.

Instrumental translation Referential function equifunctional trl. Expressive function heterofunctional trl. Appellative function homologous transi. Phatic function

These general considerations on text function can be applied both to original texts and to instrumental translations, whereas documentary translations usually have a metatextual function, informing the target reader about certain aspects of the original text and its communicative situation. Let me now illustrate the functional translation typology with a few examples. In a documentary translation, certain aspects of the source text-in-situation are reproduced for the target receivers, who are conscious of "observing" a communicative interaction of which they are not a part. If the focus is on the morphological, lexical or syntactic features of the source language system as present in the source text, we speak of a word-for-word translation or interlineal version, as in the following example, taken from Wilhelm Busch's famous book "Max und Moritz". Example 5: Word-for-word translation Durch den Schornstein mit Vergnügen sehen sie die Hühner liegen, die schon ohne Kopf und Gurgeln lieblich in der Pfanne schmurgeln.

Through the chimney with pleasure see they the chickens lie which already without head and necks nicely in the pan spatter.

If a documentary translation is intended to reproduce the words of the original, adapting syntactic structures and idiomatic use of vocabulary to the target language norms, we speak of a literal or grammar translation, which is frequently used in foreign language teaching, but also in news texts when the utterance of a foreign politician is rendered literally. For the lines from "Max and Moritz" quoted above, a literal translation would look as follows: Example 6: Literal translation Durch den Schornstein mit Vergnügen sehen sie die Hühner liegen,

Peeping through the chimney, they notice with pleasure that the

CHRISTIANE NORD

die schon ohne Kopf und Gurgeln lieblich in der Pfanne schmurgeln.

53

chickens, without heads and necks, are spattering already nicely in the frying-pan.

If the target text reproduces the source text rather literally, adding, however, the necessary cultural or (meta)linguistic information in footnotes or glossaries, we speak of philological or learned translation. This form is used frequently with classical texts. For members of a target culture where people are accustomed to frying a chicken without chopping off head and neck beforehand, the translator therefore might have added: Example 7: Cultural explanation *In Germany, the head and neck of a chicken is usually chopped off before frying. Another form of documentary translation, which is the conventional form of literary translations in many cultures today, is called "exoticizing translation" because the "setting" of the story is left unchanged, producing an "exotic" effect on the target readers where the original readers found their own culture reflected in the text. Example 8: Proper names als culture markers The English translation of "Max und Moritz" leaves the names of the two boys in the German form (= exotizing translation), whereas in the Italian version they are called "Pippo e Peppo" (instrumental translation with cultural adaptation to produce equivalent effect). (See also Nord 1993a) There are also various forms of instrumental translation. If the function of the target text is the same as that of the original, we can speak of an "equifunctional translation". This form is normally used for technical texts, computer manuals, instructions for use, and the like. It is best represented by set phrases like the following: Example 9: Equifunctional translations Zutritt verboten! No entry. Défense d'entrer. Prohibido entrar. If the function or functions cannot be reproduced as a whole or in the same hierarchy of functions, we could speak of a "heterofunctional" translation. The following translation of the lines from "Max und Moritz" is instrumental in that it produces an amusing text in rhymes and with word plays specifically tailored to the new addressees' taste. We can see that the element "ohne Kopf und Gurgeln" ("without heads and necks"), whose main function was to rhyme with the onomatopoeic verb "schmurgeln", has not been translated semantically and that the translator added "reckless", which has no equivalent in the original text.

54

A FUNCTIONAL TYPOLOGY OF TRANSLATIONS

What has been translated, however, is the function of producing an original rhyme which fits into the verse pattern. Example 10: Heterofunctional translation Through the chimney, gay and reckless, They can see them, plump and neckless, Browning nicely in their batter, Grace the frying pan and spatter. (Arndt 1981) Thus, the referential function has been slightly changed, and the entertaining function - which obviously has been given priority - is more or less the same in the source and the target text. As the translation preserves the German setting by leaving the proper names unchanged, it does not reproduce the identificational (= appellative) function the original has for its readers. Thus, this function has been changed from "appealing to the readers' own cultural experience" to "informing about a strange culture with some resemblance to the readers' own culture". A third form of instrumental translation is what I would like to call a "homologous translation". It is intended to achieve a homologous effect by reproducing in the TC literary context the function the source text has in its own SC literary context (or polysystem). This form is often found in the translation of poetry. It is called "semiotic transformation" by Ludskanov or creative transposition in Jakobson's terms (see Bassnett 1991: 18). Example 11: Catullus, Poem 13, An invitation to dinner Cenabis bene, mi Fabulle, apud me paucis, si tibi di favent, diebus, si tecum attuleris bonam atque magnam cenam, non sine Candida puella et vino et sale et omnibus cachinnis. [...] To night, grave sir, both my poore house, and I Doe equally desire your companie: Not that we thinke us worthy such a ghest, But that your worth will dignifie our feast, With those that come; whose grace may make that seeme Something, which, else, could hope for no esteeme. (Ben Johnson, example borrowed from Bassnett 1991: 84ff.) In reading an instrumental translation, the reader is not supposed to become aware of reading a translation at all. Therefore, the form of the text is usually adapted to TC norms and conventions of text type, genre, register, tenor, etc.

CHRISTIANE NORD

55

The limits of translation In this typology, the limits of translation are marked by the amount and quality of adaptive procedures. Copying or merely transliterating a text or text part cannot be called "translating"; nor can we accept as a "translation" any rendering which does not respect the principle of loyalty. Example 12: In his book, "In Cuba", the Nicaraguan priest Ernesto Cardenal intended a subjective, politically committed éloge in honour of Fidel Castro's revolution. The German translation from 1972 ("In Kuba. Bericht einer Reise") gives the impression of a rather objective report of the author's journey to Cuba, in which the reader is constantly reminded that all that glitters is not gold. The German reader must come to the conclusion that the author has adopted a critical attitude toward Castro's Cuba. This conclusion does not conform to the author's intention. Therefore, it is not a loyal translation although it may be regarded as functional from the editor's viewpoint who in the Germany of the early seventies may not have dared to confront the readers with a "communist author". The translator should have refused to produce this "translation" on ethical grounds. Within the framework of the functionality-plus-loyalty model, an instrumental translation can be chosen only in those cases where the intention of the sender or author is not directed exclusively at SC receivers but can also be transferred to TC readers (which would have been the case with Ernesto Cardenal's book). If this is not the case, the translation must be realized in document function, informing about the source-text situation in the text environment (e.g. in a few introductory lines) and thus giving the target addressees an indication that they are reading a translated text.

Planning the translation process A text is "functional" when it serves the function or functions it is intended for, and text function is determined by the factors of the situation in which the text will have to serve as a communicative instrument (i.e. the time, place, and purpose of, and motive for reception, the medium by which it will be transmitted, and the audience it will be addressed to, see Nord 1991a). The process of producing a "functional" target text has to start, therefore, from an analysis of the target situation.

56

A FUNCTIONAL TYPOLOGY OF TRANSLATIONS

The skopos The TT function (skopos) is defined by the translation assignment given by the initiator or deduced by the (professional) translator from the conditions of the translation task, interpreted according to previous experience. Since trainee translators lack such previous experience, each translation task has to be accompanied by a "translation brief' (cf. Nord 1996), from which the students can draw the necessary information on the expected TT function(s). The translation brief should therefore contain (explicit or implicit) information about (1) the sender's intention(s), (2) the addressee(s), (3) the (prospective) time and place of text reception, (4) the medium over which the text will be transmitted, (5) the motive for text production or reception. Let us look at an example again: Example 13: JUBILÄUM AUS TRADITION IN DIE ZUKUNFT „Aus Tradition in die Zukunft". So lautet das Leitmotiv des Jubiläumsjahres 1986, in dem die Ruperto Carola 600 Jahre alt wird. Im Bewußtsein ihrer jahrhundertealten Tradition formt sich ihre künftige Funktion in Wissenschaft und Gesellschaft zum Auftrag von heute. Langfristige Jubiläumsprojekte sind das „Internationale Wissenschaftsforum Heidelberg", in dessen Rahmen Heidelberger Wissenschaftler mit auswärtigen Forschern zu Symposien zusammenkommen werden, ein Tiefmagazin für die wertvollen Bestände der Universitätsbibliothek und ein Rechnernetz zur intelligenten Informationsverarbeitung für alle Fakultäten. [...] (First paragraph of a folder published on the occasion of the sixth centenary of Heidelberg University)

SIX-CENTIEME ANNIVERSAIRE TRADITION ET MODERNISME „Tradition et modernisme": C'est sous ce double signe qu'est placée l'année 1986, année du six-centième anniversaire de la fondation de l'université Ruperto Carola. Forte de sa tradition séculaire, Heidelberg vit déjà à l'heure du future et a choisi d'anticiper sur les taches qui lui incomberont dans la science et la société de demain. Parmi les projets de longue haleine mis en oeuvre à l'occasion de cet anniversaire, citons le „Forum International des Sciences" qui fera de Heidelberg un lieu de rencontres et d'échanges entre scientifiques de toutes nationalités, la construction d'archives souterraines destinées à abriter les trésors de la Bibliothèque Universitaire et enfin l'installation d'un réseau informatique.

CHRISTIANE NORD

SIXTH CENTENARY FROM TRADITION INTO THE FUTURE „From tradition into the future" is the motto for 1986, the 600th anniversary of Heidelberg University. 1st present and future role, in academic and public life, is rooted in this tradition. Forward-looking projects to mark the occasion include the Heidelberg University International Forum (a conference centre for local and visiting scholars), the construction of underground archives for valuable University Library stacks and the establishment of a computer network available to all faculties.

57 VI CENTENARIO DESDE LA TRADICION HACIA EL FUTURO

„Desde la tradición hacia el futuro" es el lema bajo el que se conmemora en 1986 el VI Centenario. Se trata de resaltar la tradición secular de la Universidad Ruperto Carola. Su función actual y futura en la ciencia y en la sociedad surge como una misión que tiene su origen en esta tradición. Proyectos del VI Centenario a largo plazo son: el Foro Científico Internacional de Heidelberg, en el que se reunirân, en simposios, científicos de Heidelberg con investigadores de otras universidades; un almacén subterrâneo para los fondos valiosos de la Biblioteca Universitaria y una red de ordenadores para el procesamento inteligente de datos destinada a todas las facultades.

The German text is the original. The translation assignment could have been formulated as follows: Translate the text into (European/American) English/Spanish/French. There will be separate folders edited in each of these languages (plus Japanese) with the same arrangement of text and photographs. They will be distributed at Heidelberg University during the year 1986 in order to inform any visitors or interested persons about the anniversary celebrations and further projects and present a positive image of Heidelberg University. This assignment contains the following translation brief: (1) intended functions: referential (information), appellative (image), (2) addressees: Persons from Great Britain/the United States (Spain/Latin America, France) interested in academic institutions, (3 ) time of reception: 1986, (4) place of reception: (mainly) Heidelberg, (5) medium: monolingual folder with coloured photographs and short texts, (6) motive for text production/reception: 600th anniversary. In order to achieve the intended functions with the target audiences, the target text should conform to TC text-type and style conventions and a rather formal register in order to present the information in a way that ensures the presentation of a positive image. It should take account of the target-readers' culturespecific knowledge presuppositions. We may assume that the receivers are per-

58

A FUNCTIONAL TYPOLOGY OF TRANSLATIONS

sons with a specific interest in universities and academic life, since day tourists from abroad do not normally visit the University buildings where the folder is distributed. Source-text analysis and translation strategy The ST function, on the other hand, can be elicited by a similar analysis which considers both extratextual and intratextual factors of the original. In this case, the analysis shows that most extratextual factors are the same for ST and TT. What is different is the group of addressees, who, in the case of the German original, are not only German-speaking (i.e. German, Swiss, Austrian) visitors to Heidelberg interested in the University and academic life, but also potential sponsors or prospective students. Comparing the skopos with the ST functions, the translator will find out (a) whether the assignment can be fulfilled without violating the loyalty principle, and (b) which translation problems will arise in the process and what kind of transfer procedures will have to be used in order to solve them. Since both the original and the various translations have to function in exactly the same situation (as to time, place, medium and motive), we may assume that what is needed is an equifunctional instrumental translation. Any adaptations which may prove necessary are due to the difference in the addressed audience (cultural knowledge, socio-cultural background, not so specific interests in the case of the target receivers) and the differences in textual and stylistic conventions and language use. Since it is a common fact that translations tend to be longer than originals, the translator should be careful not to exceed the space limitations set by the layout. Categorizing translation problems The comparison of skopos and source text shows that, in the case of our example, the translation instructions are compatible with the loyalty principle. In view of the differences between ST and TT addressees and between SC and TC conventions and language norms, we can now categorize the translation problems which will arise in the process of translating the German original into one of the languages mentioned above. For didactic purposes, I have devised four categories of translation problems (see Nord 1987): pragmatic translation problems, intercultural translation problems, interlingual translation problems, text-specific translation problems. These categories, listed according to their degree of generalizability, have proved to be of valuable help in translation teaching. I will therefore illustrate them by means of some examples taken from the CENTENARY text (example 13).

CHRISTIANE NORD

59

(a) Pragmatic translation problems (PTP) Every source text can be translated into any target language for various purposes. And in each case, the translator is confronted with two communicative situations (ST situation and TT situation). The contrast between these situations gives rise to pragmatic translation problems, even in those cases where the individual factors do not differ (see also example 4). Example 14: Translation function The first decision the translator has to take is the one regarding the function of the translation. In our example, we have found out that the translations should be equifunctional and that the main function is the referential function. In the original and in the French translation, the connection between the university's future-orientation and its long-term projects becomes perfectly clear. In the Spanish and the English version, there is a lack of coherence between the first and the second part of the paragraph. Therefore, the slogan does not "work" either, which is also due to the fact that the preposition "aus" in the German text is ambiguous (meaning "from tradition" and "by tradition" at the same time), whereas the English and the Spanish slogan refer to a temporal extension only. Example 15: Culture-bound terms (receiver-orientation) The translation of culture-bound terms always poses pragmatic translation problems since the target readers often cannot be supposed to know the source culture. Therefore, the translation of "Ruperto Carola" by "University of Heidelberg" in the English version takes account of the fact that most Englishspeaking readers would not know the German habit of using the Latin names of the ancient universities. The French and the Spanish versions may seem slightly incoherent to those readers who are not familiar with this habit. Example 16: Space restrictions (medium) The space restrictions mentioned above pose another pragmatic translation problem. In this sense, the surprising shortness of the English version might even be "dysfunctional" because it does not "fill" the provided space. (b) Intercultural translation problems (CTP) Each culture has its own habits, norms and conventions of verbal and nonverbal behaviour. Intercultural translation problems arise from the differences in conventions between the two cultures involved, such as measuring conventions, formal conventions, text-type conventions, conventional forms of address and salutation formulae, etc. Since according to the translation brief the translation of the sample text should conform to TC conventions, I will just mention two examples of CTP.

60

A FUNCTIONAL TYPOLOGY OF TRANSLATIONS

Example 17: Text-type conventions Slogans can be regarded as a text types like titles and headings (cf. Nord 1993). Therefore, a translation of the slogan "Aus Tradition in die Zukunft", although it may correctly reproduce its semantic content (which is not the case in the Spanish version), will not be functional unless it "sounds" like a TC slogan, i.e. unless it conforms to the TC slogan conventions. The French version shows that this may require a complete restructuring of the ST formulation. Example 18: General conventions of style CTP may arise from the preference of certain stylistic patterns in SC and TC. If the TT receivers expect a translation to conform to the general conventions of "good" style of their culture, the translator has to adapt the ST structures in order to meet this requirement. The sample texts present some examples for the preference of nominal style in German as opposed to the preference of verbal style in the Romance languages: "ihre künftige Funktion in Wissenschaft und Gesellschaft" vs. "les tâches qui lui incomberont dans la science et la societé de demain" (the Spanish version does not conform to this principle although it also exists in Spanish stylistic conventions) or "ein Tiefmagazin für die wertvollen Bestände..." vs. "la construction d'archives souterraines destinées à abriter les trésors..."). (c) Interlingual translation problems (LTP) The structural differences in vocabulary, syntax and suprasegmental features of the two languages give rise to translation problems as well. Comparative grammar and some approaches to a didactic translation grammar (see Raabe 1979) provide valuable support for the solution of these problems. Just a few examples from the sample text: Example 19: Compounds A problem which often arises in translation where German is one of the languages involved is the translation of nominal compounds, such as "Jubiläumsjahr", "Jubiläumsprojekte", "Tiefmagazin", "Rechnernetz", "Informationsverarbeitung". For translation.teaching, it would be advisable to discuss the possible transfer procedures, such as modulations ("se conmemora en 1986 el VI Centenario"), transpositions ("1986, the 600th anniversary of Heidelberg University"), paraphrases ("projects to mark the occasion", "projets... mis en oeuvre à l'occasion de cet anniversaire"), or even reductions ("Rechnernetz zur intelligenten Informationsverarbeitung" vs. "réseau informatique" or "computer network") which may be more functional in this case than a long and complicated translation of all the details. Example 20: Adjective function The formulation "archives for valuable Library stacks" leads to the assumption that the archives are intended for valuable stacks only, whereas the Ger-

CHRISTIANE NORD

61

man adjective "wertvoll" is a mere epithet. Therefore, the French translation ("trésors") is functional, while the postposition of the adjective in the Spanish version ("fondos valiosos" instead of "valiosos fondos") wrongly points to a distinctive feature. Example 21: Focus In the German formulation "So lautet das Leitmotiv...", the anaphoric element "so" links the slogan with the following sentence, and "Leitmotiv" carries the sentence stress, which the French translator has rendered by the typical focus structure "C'est sous ce double signe que...". In the Spanish version, the use of the relative clause "el lema bajo el que se conmemora..." leads to the same effect, whereas in the English translation the intonation contour automatically has its culmination point in "1986". (d) Text-specific translation problems (TTP) TTP are those problems which arise in the translation of one specific text and whose solution cannot be generalized, although it is also based on functional criteria. In this category we find the translation of metaphors, similes, puns, rhetorical figures, etc. Since the sample text belongs to a rather conventional text type, it does not present any TTP. Therefore, I will quote a play on words from "Alice in Wonderland". Example 22: Play on words "How funny it'll seem to come out among the people that walk with their heads downward! The Antipathies, I think" (she was rather glad that there was no one listening this time, as it didn't sound at all the right word). (Carroll 1946: 4) "Das kann ja lustig werden, wenn ich bei den Menschen herauskomme, die mit dem Kopf nach unten laufen! Die "Antipathien" sagt man, glaube ich -" und diesmal war sie recht froh, daß ihr wirklich keiner zuhörte, denn das Wort klang ganz und gar nicht richtig. (Carroll 1973: 13) "iQue divertido será surgir de golpe por donde vive toda esa gente que anda sobre la cabeza! Los antipâticos, según me parece..." (Esta vez sí que se alegró de que no hubiera nadie escuchando, porque algo no le sonaba bien en lo que había dicho) (Carroll 1970: 34) The original pun is based on the phonetic similarity of "antipodes" and "antipathies", as is confirmed by the reference to the "sound" of the word, whereas the readers of the two translations must assume that Alice is just mixing up two difficult words. The specific point of the ST, its "function" to make the reader understand what is going on in Alice's head, is lost.

62

A FUNCTIONAL TYPOLOGY OF TRANSLATIONS

A functional hierarchy of translation problems When translating a given source text into a given target language, the "normal" procedure is to start from the source-language elements and transfer the text sentence by sentence or, more frequently, phrase by phrase or, if possible, even word by word in a kind of draft translation, whose quality may vary according to the translator's competence, and then try and polish it stylistically until it seems acceptable (judging from the translator's personal point of view) for the communicative situation it is needed for. Such a "bottom-up" process, which works from the linguistic text-surface structures (Stage 1) up to conventions (Stage 2) and finally pragmatics (Stage 3), giving much room to the translator's own stylistic preferences and the limitations of his/her linguistic and translational competence, has several drawbacks not only in translation practice, but particularly in translation teaching, which can be resumed as follows: (a) The students are tempted to keep as close to the ST structures as possible, which leads to linguistic interferences and mistakes even when translating into the native language. (b) Translating is seen as a code-switching operation, where lexical or syntactic equivalences play the most important part. (c) The focus of attention is directed to the smaller units of language, and the whole of the text in its communicative situation is often lost out of sight. (d) Decisions are often based on personal intuition, which makes it difficult to justify them by intersubjective reasons, e.g. from translator to customer or reviser in translation practice and from student to teacher in translator training, or vice versa. (e) A decision which has been taken on a "lower" stage, has often to be revised when reaching the next stage, as is shown in the following example. (f) Sometimes, the translation process is even blocked because of apparent "intranslatability" (as would have been the case in the English translation of example 4, if the translator had known the correct meaning of the idiom s/he had chosen). Example 23: Bottom-up decision process The translator has decided (stage 1) to render "Rechnernetz zur intelligenten Informationsverarbeitung für alle Fakultäten" by "computer network for intelligent data processing for all faculties". Checking on the text-type and style conventions (stage 2), she then finds out that, apart from being ambiguous, this phrase would give the whole sentence a clumsy nominal structure which is not acceptable in a conventional text of this type. Therefore, she goes back to stage 1, changing the end of the phrase into "which is available to all facul-

CHRISTIANE NORD

63

ties". Now, she looks at the communicative situation (stage 3), and finds that with this solution the text becomes too long for the available space and, moreover, the stress on this element at the end of the paragraph does not correspond to its subordinate function as one of several elements in an enumeration. Thus, the translator has to go down to stage 1 again, and to re-start the whole process, etc.etc. In functional translation, problems should therefore be dealt with in a top-down hierarchy:

This means that a functional translation process should start on the pragmatic level by deciding on the (intended, assigned) function of the translation (la: documentary vs. instrumental). This decision marks the way for the next stage (lb): which functional elements of the ST will have to be reproduced "as such" and which should be adapted to the addressee's background knowledge, expectations, communicative needs, to medium-restrictions, deixis requirements, etc. The chosen translation type then determines whether the translated text should conform to source-culture or target-culture conventions, both with regard to translation (for translational conventions see Nord 1991b) and style (stage 2). Only then, if at all necessary, will the differences in language system come into play (stage 3). If there is still more than one possible solution at this point, contextual aspects (stage 4) or even - in less conventionalized or unconventional, i.e. literary, texts - the translator's own personal preferences, always with

64

A FUNCTIONAL TYPOLOGY OF TRANSLATIONS

due respect to the function of the translation, will determine the ultimate decision (stage 5).

Conclusion The functional approach presented in this paper rests on two pillars: a communication-oriented concept of textuality, on the one hand, and a target-oriented concept of translation, based on the principle of loyalty to cooperation partners, on the other. The model is claimed to be valid for all fields of oral or written intercultural communication and for any culture and language pair. The application of the model to various text types (including literary texts) shows that there are a large number of translation problems which can, and should, be dealt with in a general manner . References Arndt, Walter W. 1981. The Genius of Wilhelm Busch. University of California Press. Translation published in Wilhem Busch. 1982. und Max und Moritz polyglott. Munich: dtv. Beaugrande, Robert A. de and Dressler, Ulrich W. 1981. Introduction to Text Linguistics. London-New York. Bassnett, Susan and Lefevere, André (eds). 1990. Translation, History and Culture. London/New York: Pinter. Bassnett, Susan. 1991. Translation Studies. Revised edition. London/New York: Routledge. Bühler, Karl. 1934. Sprachtheorie. Jena: Fischer. Carroll, Lewis. 1946. Alice in Wonderland and Through the Looking Glass, New York: Grosset and Dunlap. Carroll, Lewis.. 1970. Alicia en el Pais de las Maravillas. Transi. Jaime Ojeda. Madrid: Alianza. Carroll, Lewis. 1973. Alice im Wunderland. Transi. Christian Enzensberger. Frankfurt/M.: Insel. Catford, John C. 1965. A Linguistic Theory of Translation. An Essay in Applied Linguistics (Language and Language Learning). London: Oxford University Press. Dictionary of Contemporary English (DCE). 1978. London: Longman. Holz-Mänttäri, Justa and Nord, Christiane (eds). 1993. Traducere navem. Festschrift für Katharina Reiss zum 70. Geburtstag. Tampere: Universitätsverlag. Distributed by University of Tampere Sales Office. P.O. Box 617. FIN-33101 Tampere, Finnland. Jäger, Gerd and Neubert, Albrecht (eds). 1982. Übersetzungswissenschaftliche Beiträge 5. Leipzig: Enzyklopädie.

CHRISTIANE NORD

65

Jäger, Gerd and Müller, D. 1982. Kommunikative und maximale Äquivalenz von Texten. In Jäger and Neubert (eds). 43-57. Jakobson, Roman. 1960. "Linguistics and poetics". In Thomas A. Sebeok (ed). Style in Language. Cambridge/Mass, 350-377. Koller, Werner. 1993. Einführung in die Übersetzungswissenschaft. 4th ed. Heidel­ berg: Quelle & Meyer. Koller, Werner. 1993a. "Zum Begriff der 'eigentlichen' Übersetzung". In Holz-Mänttäri & Nord (eds). 49-63. Neubert, Albrecht. 1984. "Text-bound translation teaching". In Wilss & Thome (eds.). Die Theorie des Übersetzens und ihr Aufschlußwert für die Übersetzungs- und Dolmetschdidaktik. Tübingen: Narr, 61-70. Neubert, Albrecht. 1985. "Maximale Äquivalenz auf Textebene". Linguistische Arbeitsberichte 47 (1985), 12-24. Neubert, Albrecht. 1986. "Translatorische Relativität". In Snell-Hornby (ed). 85-105. Newmark, Peter. 1984/85. "Literal translation". Parallèles 7: 11-19. Nida, Eugene A. and Taber, Charles. 1969. The Theory and Practice of Translation. Leiden: Brill. Nida, Eugene A. 1964. Toward a Science of Translating. Leiden: Brill. Nord, Christiane. 1987. "Übersetzungsprobleme - Übersetzungsschwierigkeiten. Was in den Köpfen von Übersetzern vorgehen sollte..." Mitteilungsblatt für Dolmetscher und Übersetzer 2: 5-8. Nord, Christiane. 1989. "Loyalität statt Treue. Vorschläge zu einer funktionalen Übersetzungstypologie". Lebende Sprachen 34 (3): 100-105. Nord, Christiane. 1990. "Ausgangstextanalyse und Translatfunktion. Zur Rolle des Ausgangstexts in der funktionalen Translation". Fremdsprachen 3: 161-169. Nord, Christiane. 1991. Text Analysis in Translation. Theory, Methodology and Didactic Applications of a Model of Translation-Relevant Text Analysis. Amsterdam/Atlanta GA: Rodopi. Nord, Christiane. 1991a. "Scopos, loyalty and translational conventions". Target 3 (1): 91-109. Nord, Christiane. 1993. Einführung in das funktionale Übersetzen. Am Beispiel von Titeln und Überschriften. Tübingen: Francke. Nord, Christiane. 1993a. "Alice im Niemandsland". In Holz-Mänttäri and Nord (ed), 395-416. Nord, Christiane. 1996. Translating as a Purposeful Activity. Functionalist Approaches Explained, Manchester: St. Jerome (= Translation Theories Explained 1) (in press). Oettinger, Anthony G. 1960. Automatic Language Translation. Cambridge/Mass. (= Harvard Monographs in Applied Science, 8). Pym, Anthony. 1992. Translation and Text Transfer. An Essay on the Principles of Intercultural Communication (Publikationen des Fachbereichs Angewandte Sprach- und Kulturwissenschaft der Johannes Gutenberg Universität Mainz in Germersheim, 16). Frankfurt a.M. etc.: Peter Lang.

66

A FUNCTIONAL TYPOLOGY OF TRANSLATIONS

Pym, Anthony. 1993. Epistemological Problems in Translation and its Teaching. A Seminar for Thinking Students. Calaceite (Spain): Editions Caminade. Raabe, Horst. 1979. "Didaktische Translationsgrammatik". In Bausch, Karl-Richard (ed.). Beiträge zur didaktischen Grammatik. Probleme, Beispiele, Perspektiven. Königstein/Ts. 239-256. Reiss, Katharina and Vermeer, Hans J. 1984. Grundlegung einer allgemeinen Translationstheorie. Tübingen: Niemeyer. Snell-Hornby, Mary (ed). 1986. Übersetzungswissenschaft - eine Neuorientierung. Tübingen: Francke. Snell-Hornby, Mary. 1986. "Übersetzen, Sprache, Kultur". In Snell-Hornby (ed). 9-29. Snell-Hornby, Mary. 1987. "Translation as a cross-cultural event: Midnight's Children - Mitternachtskinder". Indian Journal of Applied Linguistics 2: 91 -105. Snell-Hornby, Mary. 1988. Translation Studies: An Integrated Approach. Amsterdam/Philadelphia: Benjamins. Snell-Hornby, Mary. 1990. "Linguistic transcoding or cultural transfer? A critique of translation theory in Germany". In Bassnett and Lefevere (eds), Translation, History and Culture. London/New York: Pinter, 79-86. Vermeer, Hans J. 1978. Ein rahmen für eine allgemeine translationstheorie. In Hans J. Vermeer (ed). 1983. Aufsätze zur Translationstheorie. Heidelberg: Vermeer, 48-61. Vermeer, Hans J. 1986. "Übersetzen als kultureller Transfer". In Snell-Hornby (ed). 30-53.

Text-Type Conventions and Translating: Some Methodological Issues Paul Kussmaul University of Mainz/Germersheim

Introduction When I was a young lecturer at a British university, I attended a discussion of a German Professor's paper that had just been delivered. My German colleague and I both found the paper very stimulating, and we took an active part in the discussion. The next day, my German colleague came to me and told me that an older English colleague had said to him "You were rather offensive last night". We were both rather puzzled and tried to figure out what had happened. As was normal in our own country we had both freely expressed our conflicting opinions and critical views of the German professor's paper. We tried to remember how the English colleagues had done it and in retrospect realised that they had acted in a completely different manner. They had never uttered straightforward criticism but had mainly asked questions such as "What do you think of the approach of colleague X?" "I would be interested to hear your opinion on the recent book by colleague Y?" To those familiar with the topic these questions were in fact rather critical in that they implied that the professor had not taken these approaches sufficiently into account, but he was given the chance to cover up his omissions or even his lack of knowledge and avoid losing face. There seemed to be two types of academic discourse, the German type and the English type. German academics seemed to be more direct in uttering their critical attitude and differing opinions, whereas English academics seemed to behave more indirectly, which did not, however, mean that they were not critical. Is this merely an anecdote or is it typical? Galtung (1985) has proposed cultural prototypes of intellectual discourse which are very much in line with this

68

TEXT-TYPE CONVENTIONS AND TRANSLATING

anecdote. In "Saxonic" discourse (in Great Britain and the USA) comments on a presentation or lecture usually begin with a few laudatory remarks, and then carry on with "however" as an introduction to some kindly worded criticism. In "Teutonic" discourse (in Germany), on the other hand, one immediately aims at the weakest point in the thesis being propounded and tries to shoot the proponent down in flames (Galtung 1985: 157f.). I have begun with this example because it can be used to illustrate and explain the notions of convention, culture, and text type. The type of discourse described in our example has a great affinity with the term text type as used in linguistics, as we shall see in a moment, and seems to be governed by certain rules or conventions about what is appropriate behaviour. These rules seem to vary between cultures.

Some basic concepts Conventions Let us look at the terms convention and conventional to begin with. They are used with a variety of meanings in linguistics. They can be used more or less synonymously with the term arbitrary (in contrast to the term motivated) and then refer to the relationship between concept and sound-image. Although we are not concerned with this type of relationship here, this use of conventional nevertheless implies notions which are also central to the description of text types. Conventions imply conformity and expectation (cf. Lewis 1969: 78), and when people use words they can normally be expected to use them in the same sense as other people do, i.e. to conform to generally agreed on regularities. In the same way, when people produce specific text types, such as business letters, they can be expected to conform to specific regularities and rules. If they do not, communication may turn out to be difficult or even break down. Within speech-act theory convention and conventional have come to be used for the conditions that need to be fulfilled for a felicitous performance of an illocutionary act (cf. Wunderlich 1972: 11ff.). Searle mentions two types of rules in this context: constitutive rules which create and define the speech act and take the form "X counts as Y in the context of C", and regulative rules which regulate independently existing behaviour and take the form "If Y, do X" (Searle 1969: 33ff.). Although both types of rules may be said to be subject to conventions, for the purpose of describing the nature of text types the second type of rules seems to be particularly useful. For instance, when describing the linguistic behaviour in the example mentioned above we may say that in the

PAUL KUSSMAUL

69

British culture there seems to be a rule which says "In academic discourse, if you want to criticise, make use of questions!" In the German culture there seems to be a rule like "In academic discourse, if you want to criticise, you may point out the weaknesses of your colleagues." Before, however, discussing regulative rules and speech acts within certain text types it may be useful for our purposes to look more closely at the terms culture and text type. Culture "Culture" is a very complex notion. In translator-training syllabuses "cultural studies" usually form a major part. Those who design them are faced with the notorious problem of deciding what topics should go into these studies. Usually included are political institutions, education, history and current affairs. But then, which periods of history should one concentrate on? Often included is the literature of a country, but then which books are part of the cultural heritage? What about other forms of art such as films and television programs most people in a country know and watch? Culture certainly has to do with common factual knowledge. However, there is also behavioural knowledge, and I agree with Vermeer who stresses the importance of just this type of knowledge for the purposes of training intercultural competence and indeed translation competence. He points out the importance of cultural norms, conventions and opinions which help us forming the guiding principles of our behaviour (Vermeer 1986: 11). It seems to me that behavioural knowledge is underrepresented in translator training, perhaps because it is more volatile than factual knowledge. For the purposes of handling texts this type of knowledge is of great importance. If we do not observe these cultural norms, we may have to face sanctions, and it is no excuse saying "but I am a foreigner". In the first example, it will be remembered, my German colleague was rebuked by an older English colleague for not complying with British norms of behaviour. Text type Text type is an ambiguous term. It may refer to what is called Texttyp in German, which is the more general notion which has been brought to our attention by Reiss (1971) who distinguishes between informative, expressive and appellative text types. For want of a better word text type may also serve as an equivalent for German Textsorte, which refers to entities such as manuals, instruction leaflets, business letters, weather reports, contracts, or also academic discourse, to give a name to our initial example. The very fact that there are names in a language for specific types of communications may be taken as a sign that these must have features in common.

70

TEXT-TYPE CONVENTIONS AND TRANSLATING

There is a close relationship between situation and text type. Crystal and Davy (1969) have put forward a model of situational dimensions and distinguish between INDIVIDUALITY, DIALECT, TIME, MEDIUM, PARTICIPATION, PROVINCE, STATUS, MODALITY and SINGULARITY. In their detailed analysis of texts they show how these dimensions have a bearing on how a text is written or, indeed, spoken. This model has been applied to translation and in some respects further specified by House (1977). If we apply the model to our initial example, which, it will be remembered, is a specimen of the text type "academic discourse", we shall see that text types reflect a pattern of situational dimensions. As far as MEDIUM is concerned our example is characterized by speech as opposed to writing. Within the PARTICIPATION dimension it is charaterized by dialogue as opposed to monologue. As to STATUS, the text is characterized by equal to equal SOCIAL ROLE RELATIONSHIP. The participants are all colleagues in the academic world; in a learned discussion their roles are of equal status. Within the SOCIAL ATTITUDE dimension (House's additional term 1977: 45ff.), which takes account of the degree of intimacy beween the participants, the situation of our example is marked by a moderate degree of intimacy, i.e. the participants of the discussion are not total strangers, they are colleagues but most likely not close friends. Finally, its PROVINCE dimension may be described as literary studies; the participants will for instance have used the specific vocabulary of this field. We can show that these dimensions are to some degree distinctive in the linguistic sense by changing them. For instance, if in our example we change the MEDIUM dimension from speech to writing, we get a different text type. It will no longer be an academic oral discussion but rather something like an exchange of letters, maybe by sending letters to the editor of a learned journal to comment on a colleague's paper. If we change the PARTICIPATION dimension from dialogue to monologue, we might get a paper read at a conference. If we change the STATUS dimension from equal-to-equal to higher-to-lower, the result might be a talk of a professor to his students giving them advice on their studies. Changing the SOCIAL ATTITUDE dimension might result in the discussion becoming an informal chat, and it would be interesting to see if the forms of indirectness would then still be necessary. If we make changes in the PROVINCE dimension, we will get a discussion within a different occupational field, which will entail a different style with regard to both vocabulary and syntax. Moreover, if we make changes in more than one dimension this will result in yet other text types. For instance, if we change both the MEDIUM to writing and PARTICIPATION to monologue, this might result in text types such as a book review or an assessment of a colleague's academic qualities. Or if we change both PARTICIPATION to monologue and STATUS to higher-to-lower the outcome may be a lecture to students. It might be argued that sometimes the names for these text types are rather long-winded and cumbersome, quite unlike "manual"

PAUL KUSSMAUL

71

or "weather report". We must bear in mind, however, that we are basically not concerned with names but with concepts. I think most people will have a pretty good idea of what the difference is in style, say, between a paper presented at a conference and a lecture to students. In the following sections, when analysing text types I shall try to show how situational dimensions are reflected in text-type conventions.

Macrostructures It is difficult to base macrostructures firmly on linguistic models. Efforts have been made to use theme-rheme progression for this purpose (cf. GerzymischArbogast 1986, 1987) but it seems difficult to describe large stretches of text in a way simple enough to be used as a guideline for the practical purpose of translating. When text linguists deal with macrostructures they usually examine the general line of thought of a text and the sequence of passages typical for the text type. Thus Göpferich (1995) has observed that in technical and science texts the flexibility of macrostructures varies according to the specific text type. For instance, patent specifications have a completely rigid macrostructure, which means that the individual passages always follow each other in the same sequence, whereas conference reports and articles in technical journals seem to have more flexible macrostructures. In general she found that the less technical a text is, the more flexible is its macrostructure, and there seems to be little difference between English and German texts in this respect (Göpferich 1995: 299ff.). The reason may be that science and technology are international fields. Technical products are sold all over the world and are normally not influenced by individual cultures. The situation is different for texts which are more firmly embedded in individual cultures. Clyne (1981,1987,1991) examined "academic texts" and found that they differ in the German and Anglo-Saxon cultures. English texts are marked by linearity, whereas German texts are marked by digressions. He sees a link between the structure of these types of texts and essay writing as taught in schools, where in Anglo-Saxon countries there are definite rules concerning a stringent and linear argument. As a result, in Anglo-Saxon academic writing there are less footnotes, and quotations and references are often integrated within the main text. Let us now return to our guiding principle and try to see which situational dimensions are implied in these macrostructural conventions. As far as PROVINCE is concerned, the similarity between English school essays and academic articles can be interpreted as a sign that they are less technical in style, whereas

72

TEXT-TYPE CONVENTIONS AND TRANSLATING

German academic writing definitely has a style of its own. With respect to English texts because they are less digressive are more oriented toward their readers, whereas the reader of German texts often must make a special effort to plough through all the digressions and footnotes, if he can be bothered at all. Clyne also found that in English texts there are more metacommunicative utterances, which refer to the structure and line of argument (Clyne 1991:279). What are the consequences of these types of observations for translating? Should we preserve the cultural macrostructures or are we allowed to change them? I mentioned above that we might be faced with sanctions if we do not comply with cultural norms, and if we do not comply with target cultural texttype conventions for academic texts, this might well result in the texts loss of acceptability. Clyne quotes a book review which might have been a reaction to a German academic text translated into English with all its digressions and footnotes: "For readers not extremely well-versed in German academic prose, the structure of the work limits its readability" (Clyne 1991: 376). On the other hand, when translating English academic texts into German, do we have to insert digressions and footnotes? I once translated a book on phonology by an English friend into German. When the publisher received the manuscript, he wrote back that there were not enough footnotes for a really academic text. So in order to fulfil the expectations of the German readers and to make the publisher happy we put in a number of additional footnotes. But should we also insert digressions, such as more detailed references to some theory or a brief historical overview? I think readability is the most important quality of a text in any culture, and all other considerations should be subject to this aim. PARTICIPATION

Microstructures Metacommunicative utterances Macrostructures deal with the elements in texts that are mainly informative and centred on content, but texts also contain elements which help to ensure that information is understood in the right way. These are usually called metacommunicative utterances (cf. Göpferich 1995: 381ff.) These types of microstructures may conveniently be analysed within the framework of the speech-act model. Some time ago I made a corpus-based analysis of metacommunicative utterances in German and English academic texts about literary studies and linguistics (Kussmaul 1978), the results of which may perhaps also hold for other subject areas. I shall quote a few typical examples:

PAUL KUSSMAUL

73

Announcements (la) I shall discuss this topic in the following chapter. (lb) We shall discuss this topic in the following chapter. (1c) The following chapter deals with... (1d) Das vorliegende Buch erörtert... (le) Diese Frage wird im folgenden Kapitel erörtert. Whereas in English announcements first person pronouns were used predominantly ( I (la) in 61% and we (lb) in 17% of the cases counted) the predominant construction in German was the mentioning of the book or chapter as a subject (1d) which was found in 55% of the cases, and the passive (le) occurred in 38% of the cases. Mentioning the book or chapter as a subject also occurred in the English announcements (lc), but only in 21% of the cases. English announcements thus seem to be more "personal" than German ones. In other words, they differ in the SOCIAL ATTITUDE dimension. There is a high degree of social distance in the German announcements and a medium degree in the English announcements. Obviously, SOCIAL ATTITUDE is a quantitative category, it varies by degrees. House makes use of Joos' degrees of formality: frozen, formal, consultative, casual and intimate style (House 1977: 45f.). Taking up these categories we might say, for instance, that example (la) is casual style, (le) is formal, and (lc) and (1d) are rather frozen. In addition, the first person plural form of the personal pronoun can also be interpreted as a sign of PARTICIPATION including the speaker and the addressee. There thus seems to be an affinity between the dimensions SOCIAL ATTITUDE and PARTICIPATION. It would be worth examining in general if texts containing forms suggesting a dialogue between author and reader also contain forms suggesting a. higher degree of intimacy. For formulating rules about conventional linguistic behaviour we can make use of Searle's pattern of regulative rules and suggest: "If you want to make an announcement, in an English academic text (with a literary or linguistic topic) use either the singular or the plural first person personal pronoun, in German texts mention the book or chapter as the subject of the sentence or use a passive construction!" Similar differences between English and German linguistic behaviour can be observed in other metacommunicative speech acts in academic texts. To quote a few more examples from my corpus: Back- references (2a) As we have seen in the first Chapter... (2b) Das erste Kapitel hat gezeigt...

74

TEXT-TYPE CONVENTIONS AND TRANSLATING

Requests (3a) Let us consider... (3b) Es ist dabei zu beachten... In these cases the first person plural pronoun is used in the English texts, and in the German texts we find a variety of "impersonal" constructions. The features of closeness in the English texts and distance in the German texts seem to be typical not only for metacommunicative utterances but for other speech acts as well, as the examples in my corpus suggest (cf. Kussmaul 1978). Göpferich made similar observations when analysing texts dealing with technology and science. She found that in general, i.e. not specifically in relation to metacommunicative utterances, the use of the pronoun we, which suggests that both the author and the reader are engaged in observing phenomena and solving problems, is much more common in English than in German texts of these types (Göpferich 1995: 374). Snell-Hornby made an observation which in a way is comparable to the ones made here when she analysed public directives. She found that English favours identification of the addressee (e.g. "Passengers entering or leaving the bus while it is in motion do so at their own risk"), whereas German favours abstract nouns (e.g. "Privatgrundstück. Benutzung auf eigene Gefahr") (SnellHornby 1988: 90). Identifying the addressee, i.e. actually addressing him or her, can be interpreted as a sign of PARTICIPATION, which again is in line with a lesser degree of distance as far as SOCIAL ATTITUDE is concerned. It would certainly be going too far if one claimed that English texts in general were more personal in tone than German texts, but it might be worthwhile to keep this aspect in mind when analysing other text types. Hedging Constructions with the pronoun we, passive and other impersonal constructions are interpreted by Schröder as a sign of hedging. He found that hedging is more common in German and Finnish texts about linguistics and philosophy than in English texts dealing with these subjects (Schröder 1987). What, in fact, is meant by hedging? It means hiding behind a hedge, as it were. The author of a text no longer takes all the responsibility for the truth of the propositional content of his utterance. The use of we may be seen to imply: "This is not only my opinion but also that of others." What the author says thus becomes more objective. The use of the passive voice can be understood to mean the same thing: "What I say are generally accepted facts." Whereas these types of hedges support the truth value of ones statements, other types of hedges can be used as "downtoners" weakening the truth value. These are phrases expressing one's

PAUL KUSSMAUL

75

opinion, such as English it seems to be, I find, I think, I suggest (in its meaning of putting forward a hypothesis), and such as German wie mir scheint, wenn ich recht sehe etc. What kinds of situational dimensions can be observed here? Those hedging features which strengthen the truth value of one's statements {we and the passive) can be interpreted as increasing the STATUS of the author. He is not alone, as it were, but is backed up by a number of other people, is the speaker of a group of experts. Neither do those hedging features that act as downtoners decrease the status of the author, rather they help him to maintain his status. Phrases such as I suggest, I believe imply that what one says is open to discussion, that there may also be other views. One cannot be attacked by someone when what one puts forward is presented as a contribution to a discussion. Thus, in both cases when increasing the truth value and when downtoning it, the author evades the danger of losing face. Keeping face and losing face is thus closely related with status. There seems to be a contradiction between Schröder's findings and my own. In my English examples quoted above there are hedges, such as we-forms and the passive, and these forms occur both in English and in German, whereas Schröder found that hedges are less common in the English texts he examined. Also, downtoners such as I suggest, seem to etc. in my experience are typical forms of introducing one's results in Anglo-Saxon academic writing, whereas donwtoners are less common in German academic writing. This, of course, would have to be demonstrated by a corpus-based study. The seeming contradiction can be explained when we observe the co-occurrence of hedging phenomena with particular speech acts, an aspect which Schröder did not take into account. Hedging, logically, can only help maintain the author's status when the speech acts he or she performs are such that they could actually decrease the author's status. This could happen in speech acts such as statements, presenting results, drawing conclusions etc. With metacommuncative utterances, which organize the line of thought of a text, status is not involved, and we-forms in these speech acts can therefore not be counted as hedges. Thus the interpretation of hedging phenomena depends on the type of speech act performed. For instance, the pronoun we in announcements and in back-references can be seen as a form of participation, typical of dialogue, where we includes you and I, whereas we in statements and presentations of results can be seen as a plural form including fellow researchers. Directives I shall use this speech act because it seems to me especially suited to illustrate text-type conventions. Instructions as a type of directive appear in leaflets and

76

TEXT-TYPE CONVENTIONS AND TRANSLATING

manuals. Some of the linguistic forms conventionally used for English and German instructions differ. In an English instruction leaflet for bath salts the following sentence appeared (cf. Kussmaul 1995: 76): (4a) Balneum Hernial is added to the bath-water (not vice versa) and is well mixed. It would be interesting to observe the reaction of English readers to this. On the basis of its syntax the sentence can be interpreted as a statement that the well mixed baths salts are already part of the bath water. This is nonsense, of course, and the only thing the reader can do is try to understand the sentence in a way which makes sense in an instruction leaflet. In the German original the sentence reads: (4b) Balneum Hermal wird in das Badewasser gegeben (nicht umgekehrt) und gut untergemischt. Although this sentence has the syntactic form of an affirmative clause normally used for statements it will be interpreted by the reader as an instruction. In German leaflets the passive voice, among other forms, is conventionally used as an illocutionary force indicator for instructions. In English leaflets, however, different forms are used. The most common one is the imperative. Thus an adequate translation would have been (4c) Add Balneum Hernial to the bath water and stir well until dissolved. A translator must know these conventions, unless he/she wants to risk the break-down of communication. In English the passive construction is used for instructions only in special syntactic contexts; it seems to be used before gerunds, e.g. (4d) The doors are locked by pushing in on the forward position of the rocker lever. (cf. Kussmaul 1995: 80) When translating instruction leaflets, teachers of translation often advise their students to use the imperative in English and the infinitive in German. In our example this variant would be (4e) Balneum Hernial in das Badewasser geben (nicht umgekehrt) und gut vermischen. This is a perfectly acceptable form, but when we look at instruction leaflets and manuals more closely, we find that there is quite a variety of forms which prevent these texts from becoming too monotonous in style. For instance, in a

PAUL KUSSMAUL

77

small corpus of texts (Kussmaul 1995: 75ff.) I found in the English texts the following: • • • • • • • • • •

the imperative please + imperative must + infinitive active must + infinitive passive have to it is advisable we recommend it is recommended it is important should

And in the German texts: • • • • • • • • • • • •

the infinitive the imperative "bitte" + imperative "bitte" + infinitive present tense passive present tense active müssen "ist zu/sind zu" "wir empfehlen" "es ist empfehlenswert" "es ist ratsam" "sollte"

I shall not go into these details here. I discussed this at some length in Kussmaul 1995: 77ff. (see also Göpferich's detailed account 1995: 308ff.). What I would like to point out, however, is that although there is a large variety of speech acts to choose from, not all forms of directive speech acts are conventional for all text types. Thus the German infinitive in its function as an illocutionary indicator for instructions is restricted to instruction leaflets, cookery books and manuals. The English imperative may be used in a large number of situations and text types but not, for instance, in regulations and rules. A sentence such as (5a) ? Make your application for a curriculum of advanced study to the Registrar not later than 1st May.

78

TEXT-TYPE CONVENTIONS AND TRANSLATING

seems to be rather unusual. Typical for regulations is (5b) Application for a curriculum of advanced study ... shall be made to the registrar not later than 1st May. (University of Bristol. Postgraduate Study) Similarly in German Verordnungen, neither the infinitive nor the imperative is used in these text types, nor indeed the modal verb sollen as a grammatical equivalent to English shall, but the form ist zu, and for a proper translation we would have to say (5c) Der Antrag für ein postgraduales Studium ist bis spätestens ersten Mai and den Registar zu richten. All other forms would be unconventional and would most likely not be taken seriously. If we want to put this into the form of a regulative rule, we might say: "When formulating orders within a legal context, In English use shall and in German use ist/sind zu as an illocutionary force indicator!" (see also Trosborg 1994: 312 who comes to the same conclusions on the basis of a corpus study of English contracts). German ist zu can be used in other text types as well, for instance in academic texts (cf. example 3b), and. I also found it in instruction leaflets. English shall, however, would look rather strange in instructions (4f) ? Balneum Hernial shall be added to the bath water. Nor does it, for instance, seem to be conventional in business letters when making reminders. A sentence such as (6a) ? You shall send us your cheque without further delay. seems strange. We would rather expect something like the following, which I found as a typical form in a corpus of commercial correspondence (Kussmaul 1977), (6b) We must ask you to send us your cheque without further delay. This latter sentence can be interpreted as a type of hedging (cf. Fraser 1975), because when taking this sentence literally the reader could understand that the writer was under some obligation by a third party to make this request, and the reader could in theory reply: "Must you really?" or "Don't worry, you don't have to." Thus here again, the dimension of STATUS is involved. The roles taken on here are of equal-to-equal. In German reminders a usual form used to be (6c) Wir bitten Sie, den Betrag innerhalb von 8 Tagen zu überweisen, (cf. Kussmaul 1977: 64).

PAUL KUSSMAUL

79

Nowadays reminders seem to be less standardised and there is a variety of highly individual and personal and polite forms. For instance, I recently received the following reminder (6d)Bei der Überprüfung unserer Konten-Listen haben wir festgestellt, daß für eine an Sie erfolgte Lieferung noch eine Rechnung zur Bezahlung offensteht. Da die Lieferung bereits längere Zeit zurückliegt und die Rechnung bei Ihnen vielleicht nicht mehr vorhanden ist, haben wir Ihnen eine neue Rechnung ausgedruckt. Bitte verwenden Sie zur Überweisung des offenstehenden Betrags die der Rechnung anhängenden und bereits ausgefüllten Überweisungsformulare. Sie können uns auch dafür eine Einzugsermächtigung erteilen. Der Betrag würde dann von Ihrem Konto abgebucht werden. Sollten Sie den Betrag aber bereits überwiesen haben, dann nehmen wir jetzt schon gedanklich Ihre Entrüstung für diese Erinnerung entgegen... Can situational dimensions be recognised in these examples? As far as the dimension is concerned the English imperative and the German infinitive, the most common forms in instruction leaflets and manuals, can be interpreted as a sign that there is little intimacy between writer and reader. However, it seems that there are subtypes of text types in which greater intimacy is favoured. In German software manuals there is a tendency to use the imperative in combination with Sie, the polite form of the second person pronoun, instead of the infinitive. Thus we usually find instructions like SOCIAL ATTITUDE

(7a) Ziehen Sie die Seitenrandbegrenzungen auf dem horizontalen und vertikalen Lineal jeweils an die gewünschte Stelle. (Microsoft Word. Textverarbeitungsprogramm 6.0, p.274) We do not find (7a) Die Seitenrandbegrenzungen auf dem horizontalen und vertikalen Lineal jeweils an die gewünschte Stelle ziehen. This would, however, be a typical form in hardware manuals. For grammatical reasons, in English texts of this sort the personal pronoun cannot be used in instructions. There is thus no choice between an impersonal and a more personal form, which means that the mere imperative does not carry stylistic meaning and cannot have the effect of formality or frozenness, to use Joss' terms. Still, in order to make "closeness" explicit, the pronoun you can be used in other places in manuals or instruction leaflets. As far as STATUS is concerned, it may look as if the imperative and the infinitive, which are not wrapped up in forms of politeness in manuals and

80

TEXT-TYPE CONVENTIONS AND TRANSLATING

instruction leaflets, show a higher-to-lower social-role-relationship. Seeing things in this way would be a fallacy though. The action to be performed is in the interest of the reader, and thus STATUS does not come into play here. In the same way, in everyday situations we just say "Have some more tea" or "Would you like some more tea?" and we do not have to say "Would you please have some more tea?" In fact, such an utterance would rather imply that the person being addressed would do us a favour by drinking more tea. STATUS, however, is implied in the examples quoted from regulations and business letters. English shall and German ist zu in regulations and orders imply higher-to-lower social-role-relationship, STATUS is here based on institutional power. Hedging forms like we must ask you to in business reminders imply a lower-to-higher social-role-relationship, but here, of course, this is a kind of role-playing, because in actual fact, if the customer delays paying the bill, legal action could eventually be taken. On the other hand, firms are interested in ensuring a friendly relationship with their customers. This becomes quite apparent in (6d) where reasons are given for sending the reminder and the firm hedges itself against possible indignation on the part of the client. There seems to be a regulative rule which is tacitly observed here and which runs: "If you want to send a reminder, use polite forms, i.e. play a role of lower-tohigher social-role-relationship !"

By way of consclusion: Consequences for translation When translating these microstructures we are faced with the same decisions as when translating macrostructures. Shall we preserve the source-text-type structures and thus create a kind of alienation effect, or shall we conform to the target-text-type conventions and thus create a text which looks perfectly normal? First of all we must take care that the speech act will be understood. Sometimes we cannot stick to the source-text forms of the illocutionary indicator simply because they would not be idiomatic in the target language; a case in point is example (4a). Secondly, there are cases where the illocutionary indicator is determined (i. e. conventional) by a specific text type. Here again, it is doubtful if the speechact would be felicitous if we do not stick to the target-text-type conventions; cases in point are examples (5a) to (5c) and (6a) and (6b). Finally, there are the cases where the illocutionary force indicator as such is not directly concerned but where an influence of STATUS, SOCIAL ATTITUDE and PARTICIPATION on the form of the speech act can be observed. These cases can overlap with the ones just mentioned. Within the STATUS dimension this happens when politeness is required, such as in reminders (examples 6b - 6d), but

PAUL KUSSMAUL

81

also when politeness is ruled out either by institutional power as in the regulation (5b) and (5c), or because the action to be performed is in the reader's interest as in most instructions. Since politeness is a very important factor in making a text acceptable to its readers, translators will be well advised to make sure they know the politeness conventions in the target culture in order to be able to observe them. Within the dimension of STATUS as apparent in the hedging forms of nondirective speech acts the situation seems to be somewhat different. When an author of an English academic text presents his or her results, it will be remembered, there seems to be a tendency toward hedging (I suggest, it seems to be, I think etc.). In German texts of this type there seems to be less hedging. When translating such texts do we have to conform to the target-culture conventions or are we allowed to introduce source-culture conventions? One may hold the opinion that the Anglo-Saxon forms sound less authoritarian and in an egalitarian academic society should also be used in German texts. Should translators change the style of academic writing then? I think it all depends on STATUSagain, in this case, however, on the status of the translators. If they have a position strong enough to do so, they may introduce new conventions. At present, however, the professional status of translators does not seem to be particularly high. Similarly, the higher degree of intimacy observed in metacommunicative utterances in English academic texts might be regarded as an asset that should be transferred into German academic culture. Here again the status of the translator is involved, but in this case one might also consider the status of the source-text author. A very intimate style may be a typical feature of his or her writing. I remember when translating some of John Searle's work into German that he used the personal pronoun I very often in metacommunicative utterances. Now Searle has a rather high status among linguists, and I therefore preserved these stylistic features in the translation. As translators we will have to weigh the status of the source-text authors, our own status and the status of the readership carefully against each other when making these kinds of decisions. Bibliography Austin, John L. 1962. How to Do Things with Words. Oxford: Clarendon Press. Clyne, Michael. 1981. "Culture and Discourse Structure". Journal of Pragmatics 5: 61-66. Clyne, Michael. 1987. "Cultural differences in the organization of academic texts". Journal of Pragmatics 11: 201-238. Clyne, Michael. 1991. "Zu kulturellen Unterschieden in der Produktion und Wahrnehmung englischer und deutscher wissenschaftlicher Texte". Info DaF 18(4): 376383.

82

TEXT-TYPE CONVENTIONS AND TRANSLATING

Crystal David and Davy, Derek. 1969. Investigating English Style. London: Longman. Fraser, Bruce. 1975. "Hedged performatives." In P. Cole and J.L. Morgan (eds), Syntax and Semantics. Vol. 3: Speech Acts. New York: Academic Press, 187-210. Galtung, Johan. 1985. "Struktur, Kultur und intellektueller Stil. Ein vergleichender Essay über sachsonische, teutonische, gallische und nipponische Wissenschaft'Mn Wierlacher, Aloi (ed), Das Fremde und das Eigene: Prolegomena zu einer interkulturellen Germanistik. München: Iudicium Verlag, 151-193. Gerzymisch-Arbogast, Heidrun. 1986. "Zur Relevanz der Thema-Rhema-Gliederung für den Übersetzungsprozeß". In Mary Snell-Hornby (ed), Übersetzungswissenschaft - eine Neuorientierung. Tübingen: Francke, 160-183. Gerzymisch-Arbogast, Heidrun. 1987. Zur Thema-Rhema-Gliederung in amerikanischen Wirtschaftstexten. Tübingen: Narr. Göpferich, Susanne. 1993. "Die translatorische Behandlung von Textsortenkonventionen in technischen Texten". Lebende Sprachen 28(2): 49-53. Göpferich, Susanne. 1995. Textsorten in Naturwissenschaften und Technik. Pragmatische Typologie - Kontrastierung - Translation. Tübingen: Narr. House, Juliane. 1977. A Model for Translation Quality Assessment. Tübingen: Narr. Kussmaul, Paul. 1977. "Auffordern im Englischen". Kongreßberichte der 8. Jahrestagung der Gesellschaft für Angewandte Linguistik GAL e.V. Mainz 1977. Stuttgart. Hochschulverlag, 57-66. Kussmaul, Paul. 1978. "Kommunikationskonventionen in Textsorten am Beispiel deutscher und englischer geisteswissenschaftlicher Abhandlungen". Lebende Sprachen 2/1978, 54-58. Kussmaul, Paul. 1980 (ed) Sprechakttheorie. Ein Reader. Wiesbaden: Athenaion. Kussmaul, Paul. 1990. "Die Übersetzung von Sprechakten in Textsorten". Der Deutschunterricht, Jg.42, Heft 1.17-22. Kussmaul, Paul. 1995. Training the Translator. Amsterdam/Philadelphia: Benjamins. Lewis, K. David. 1969. Convention. A Philosophical Study. Cambridge (Mass.): Harvard University Press. Reiss, Katharina. 1971. Möglichkeiten und Grenzen der Übersetzungskritik. München: Hueber. Schröder, Hartmut. 1987. "Hedging and its linguistic realizations in German, English, and Finnish philosophical texts: a case study". Erikoiskielet ja käännösteoria. VAKKI-seminaari VII. Vaasa 1987, 47-57. Searle, John R. 1969. Speech Acts. An Essay in the Philosophy of Language. Cambridge: Cambridge University Press. Searle, John R. 1976. "A classification of illocutionary acts". Language in Society 5: 1-23. Snell-Hornby, Mary. 1988. Translation Studies. An Integrated Approach. Amsterdam: Benjamins. Trosborg, Anna. 1994. "'Acts' in contracts: Some guidelines for translation". In Mary Snell-Hornby, Franz Pöchhacker and Klaus Kaindl (eds), Translation Studies. An Interdiscipline. Amsterdam/Philadelphia: Benjamins, 309-318.

PAUL KUSSMAUL

83

Vermeer, Hans J. 1986. Voraussetzungen für eine Translationstheorie. Einige Kapitel Kultur- und Sprachtheorie. Heidelberg: th - Translatorisches Handeln. Universität Heidelberg. Wunderlich, Dieter. 1972. "Zur Konventionalität von Sprechhandlungen". In Dieter Wunderlich (ed), Linguistische Pragmatik. Frankfurt am Main: Athenäum, 11-58.

Approaches to Literary Genres

Text Types and Power Relations Susan Bassnett University of Warwick

The power of the source It is often said that literary translation causes more problems than technical translation. But this commonly held myth is only true in one sense - the power and authority of the source text is more likely to be threatening to the translator where literary texts are involved, particularly if they are regarded as canonical. Otherwise, the processes of decision taking and of linguistic transfer are the same. Arguably, the technical translator has more freedom, since he or she makes choices based principally on the function of the text itself - what it is meant to be doing, what the stylistic conventions in both source and target languages are, what the context is in which the translation is taking place. In a text that gives safety instructions for leaving a ship, for example, or a legal document, stylistic criteria will depend solely on the target system, not on the source system at all. But the literary translator is constrained by the source text in a different way, and if it is a famous text or a canonical text then problems arise for the translator if he or she feels bound by the place that source text occupies in its native system. Translators may, of course, refuse to be bound by the source text, aiming their translations at a specific readership and tailoring their work in accordance with the horizon of expectations of that readership. The early translations of Zola into English, for example, are masterpieces of censorship, as the translators fashioned their work to suit Victorian moral scruples. So are many 19thcentury translations of Sappho, where translators sought to remove any sense of lesbian love from the texts. A recent Polish edition of the poetry of Alan Ginzburg removed all traces of homoeroticism. History abounds with examples of such textual manipulation. It is what happens when the translator does not feel intimidated by the source text, and seeks to bring it into the receiving culture adapted for mass acceptability, rather than taking his or her readers to

88

TEXT TYPES AND POWER RELATIONS

the text in another culture. This fundamental distinction lies at the heart of literary translation. As Nietzsche put it: One can gauge the degree of the historical sensibility an age possesses by the manner in which it translates texts and by the manner in which it seeks to incorporate past epochs and books into its own being. Corneille's Frenchmen and even those of the Revolution - took hold of Roman antiquity in a manner that we- thanks to our more refined sense of history - would no longer have the courage to employ. And then Roman antiquity itself ... the Romans translated ... to suit their own age and ... intentionally as well as carelessly, they swished into oblivion the dust from a butterfly's wing Nietzsche was referring to the French translation practice of the 17th-Century that was dominated by techniques of acculturation. Suggesting that translation is an indicator of the cultural sensitivity of an age or of a society, Nietzsche evaluates the French practice negatively, since translators sought primarily to incorporate other literatures into their own. A more refined sense of history, he argues, encounters texts from other cultures in less predatory manner. His view is representative of what might be loosely termed the German tradition, whereas Voltaire represents the opposing, French perspective when he writes: I am convinced that we have two or three poets in France who would be able to translate Homer very well; but I am equally convinced that nobody will read them unless they soften and embellish almost everything because, Madame, you have to write for your own time, not for the past. The debates about translation between French and German translators in the 17th-century and 18th-Century have, in one respect, continued to the present day, though no longer characterized by national difference. And in analysing translation, different approaches need to be examined from various perspectives. What is clear, however, is that any discussion needs to take into account the status of the source text in its context, the expectations of the designated readership and questions of typology.

Translation in context Translation is a very complex activity, and anyone engaged in it knows full well that there is no such thing as equivalence conceived of as sameness across languages. The translated text will never be the same as the source text. Moreover, there is always a context in which translation takes place which influ-

SUSAN BASSNETT

89

ences the decisions that the translators have to take. Henry Rider, the English translator of Horace in the 17th-Century commented that Translations of Authors from one language to another, are like old garments turn'd into new fashions; in which though the stuffe still be the same, yet the die and trimming are altered, and in the making, here something added, there something cut away. Using the image of tailoring to describe translation, Rider suggests that the translator makes something new and contemporary by changing the shape, style and colours of the original, cutting away and adding new components. He is very much aware of the role played by the translator in this shaping process, for the translator's subjectivity is a significant factor, and one that has tended to be marginalized by translation theorists. The translator not only selects the text for translation, he or she then reshapes it in the target language. In that process, the question of typological substitution is therefore very much a matter of individual choice, though obviously within the constraints of context and convention. As James Holmes points out, in his essay on the translating of poetry, a translator may opt to recreate the formal properties of the source text, for example, terza rima into terza rima, or may foreground the function of the form in the source system and use an analogical form, such as blank verse for English translations of classical French alexandrines (Holmes 1970). Other translators may, of course, eschew any attempt at retaining formal properties of texts, and focus instead on the content. E.V. Rieu's famous 1946 version of The Odyssey rendered Homer into prose, on the grounds that the novel was the functional equivalent of the ancient Greek epic poem. In his introduction, Rieu describes Homer as a "novelist": In form they (the Iliad and the Odyssey) are epic poems: but it will perhaps make their content clearer to the modern reader if I describe the Iliad as a tragedy and the Odyssey as a novel. ... the Odyssey, with its well-knit plot, its psychological interest and its interplay of character, is the true ancestor of the long line of novels that have followed it. Polysystems theory expanded the horizons of translation scholarship by urging further consideration of cultural history in any examination of translation, and also by insisting upon the significance of norms in translation. As Even-Zohar pointed out as far back as the 1970s, patterns can be traced of ways in which texts for translation are selected by the target culture and of ways in which texts adopt specific norms, behaviours and policies that are a result of the interaction between systems. Following Tynjanov, he suggests that there is a constant struggle for a central position in the target system, a struggle in which translated works and other literary texts compete (Even-Zohar 1978). The

90

TEXT TYPES AND POWER RELATIONS

status of translated texts is therefore intimately linked to the role translated texts play in a given literary system, and the actual practices of translation are conditioned by that role. But at the same time, the subjectivity of the translator needs to be taken into account as well. By focussing on typological questions in the translation of literary texts, we are compelled to look at both the microstructures and macrostructures of the translation process, without placing undue emphasis on either source or target. The work of a radical translator like Rieu, therefore, needs to be seen from the dual perspective of his own idiosyncratic decision-taking and the context within which he was operating as a translator. Let us take as examples, two types of text that have had very different histories. The sonnet, which developed in Italy in the late Middle Ages has had an astonishing rate of success in a range of different languages. As a verse form, it has travelled across cultures, undergoing all kinds of metamorphoses but never losing its fundamental characteristic of 14 lines.1 In contrast, the Norse sagas have not been successfully received in other literary systems, and remain outside the canon of European mainstream texts. What we have, then, is one type of text that has successfully moved across literary systems, and another that has resisted transfer. Endeavouring to map out some of the criteria that may be relevant in understanding these different histories we come immediately to the role played by translation.

The case of the sonnet The sonnet first appeared in southern Italy in the early thirteenth century and reached a peak of aesthetic excellence in the Rime of Petrarch (1304-74). Sonnets in the vernacular developed in Spain in the fifteenth century, in France, England and Scotland in the sixteenth century and shortly thereafter in German. In his study of the development of the sonnet in Renaissance Britain, Michael Spiller suggests that the constraints of the form provided at once a security and a challenge: The sonnet pre-emptively solves two problems: proportion and extension; and, while this is a challenge, it is also a security, a kind of metrical extension of feudalism, a definite service required and requited. The "proportioned mental space" which the sonneteers so consistently chose to inhabit emerges, right at the start, as the familiar fourteen-line sonnet, with eleven syllables (or ten, depending on the vernacular) to a line, dividing into eight or six, and using in the octave two rhymes arranged either ABAB ABAB or ABBA ABBA; and two or three rhymes rhyming CDCDCD or CDECDE or almost any possible arrangement of these in the sestet. (Spiller 1992)

SUSAN BASSNETT

91

This is the basic structure of the Italian sonnet; a fourteen line schema, with definite possibilities of rhyming patterns allowing for a certain amount of variation. The skill of the poet thus depended on being able to manipulate the tightly determined form to his or her own ends, with the aim of producing an innovative text within an agreed framework. Moreover, as Spiller points out, the sonnet was essentially a courtly form, and its principle theme was love. Originally developed in one kind of southern European court, as feudalism and the courtly love ethos was replaced by the more urbane Humanistic culture of the Renaissance court, so the sonnet continued its ascent to ever-increasing popularity, becoming the dominant poetic form in the courts of Urbino, Mantua or Ferrara. Just as courtiers displayed their wealth and social status ostentatiously for others to envy and admire, so the practitioners of the sonnet under the powerful influence of the sonnet master Pietro Bembo (1470-1547) increasingly demonstrated their skills in handling the poetic form. The French historian of the sonnet, Hugues Vaganay, has declared that the sixteenth century was the century of the sonnet. Certainly it is in this period that the sonnet began to be translated extensively into other European vernaculars. These processes of transfer have been well-documented from the perspective of the literary historian, but from the translation studies perspective what needs to be considered is why this form should have suddenly become so popular several centuries after it had first emerged in Italy, and whether the process of translation brought about any significant changes to the basic formula. Such large questions are, of course, necessarily constrained by the space of this essay, but what we can see with some clarity is that the sonnet did undergo several major alterations through translation. One of the most significant is the adaptation of the sestet to result in a rhyme scheme of four lines, followed by a final couplet. This variation first appeared in the sonnets of Sir Thomas Wyatt (1503?-42) and was developed later by Sidney and Shakespeare. The effect of this was, of course, to alter the tonal range of the sonnet completely; a form that had been used as a vehicle for expressing sentiments of courtly love could now be used as a vehicle for irony. The final couplet could be used to reverse the sentiments expressed in the previous lines, or to comment or to return the reader to the point of departure in a witty circular motion. So if we consider Wyatt's version of Petrarch's Rime 189 "Passa la nave mia colma d'oblio" we can see how the rhyme scheme of ABBA ABBA CDE CDE has been altered to ABBA ABBA CDDC EE. This has the effect of focussing attention onto the final couplet, where Wyatt significantly changes the sentiments expressed in the Italian source. Petrarch compares himself to a ship that is tempest-tossed and unable to find a safe haven. The ship's captain is Love, and the poet cannot see the bright eyes of his beloved who might guide him back to safety. His last three lines read as follows:

92

TEXT TYPES AND POWER RELATIONS

Celansi i duo mei dolci usati segni; morta fra l'onde e la ragion e l'arte, tal ch'i' incomincio a desperar del porto. (Hidden are my two sweet familiar signs; reason is dead amid the waves, and art, so that I begin to despair of reaching harbour.) In Wyatt's version, "My galley charged with forgetfulness", the speaker makes no reference to a lady. "The stars are hid that led me to this pain", he states bluntly in line 12, and the final couplet is a rueful comment on the speaker's inability to seek consolation in reason: Drowned is the reason that should me comfort And I remain despairing of the port. In his edition of Wyatt's poetry, Rebholz significantly uses different phrases to describe the Tudor poet's translations. At times he uses the phrase "this sonnet translates Petrarch's Rimé", but in the case of sonnet xviii he writes that "this sonnet freely imitates Petrarch's Rime 98", whilst sonnet xiii "closely translates Petrarch's Rime 224", and sonnet xi "is probably a free imitation of Petrarch's Rime 190, and sonnet xxv "imitates very freely Petrarch's Rime 258". Such distinctions suggest that Rebholz as editor was not comfortable with Wyatt's translations of Petrarch, and had some notion of what a translation ought to be, as opposed to a free imitation. However, when we look at Wyatt's texts, what becomes apparent is that he had a very clear idea of what he was doing with Petrarch's sonnet form. He selected some specific texts for translation, as opposed to others, made significant modifications and tailored his own work for his English readership. Wyatt was, after all, a man of his time, a Renaissance courtier whose love affairs and political allegiances in the conspiracy-ridden court of Henry VIII almost cost him his life on more than one occasion. The sonnet, originally composed by the idealistic scholar-philosopher Petrarch and adapted as a form for expressing the torments and joys of unattainable love became, in Wyatt's restructured version, a vehicle for ironic self-scrutiny and for complaints against the tyranny of love. The case of the sonnet's transfer into English is a particularly good example of a text type entering a literary system at a moment when that system appeared to need it. The Renaissance was, of course, characterised by great debates about the status and authority of vernacular languages, and literary systems such as English translated avidly from modern and classical languages in an attempt to extend the stylistic boundaries of the native literature. The Englishing of the Bible in the sixteenth and early seventeenth centuries that resulted in

SUSAN BASSNETT

93

the canonical King James version of 1611 was accompanied by extended and often bitter debate, not only about issues of doctrine but also about style and language. In France, the execution of Etienne Dolet in 1546 appears to have been linked to his determination to exalt the vernacular to the status of canonical (classical) languages.2 The introduction of a new type of text, such as the sonnet, by an individual seeking to establish himself both as a poet and as a courtier in the world of the new Tudor dynasty can be said to have happened at exactly the right time in English literature for that form to develop, because both the language and the literary system were ready to absorb it.

The Norse saga In complete contrast is the case of the translation of Norse sagas into nineteenth century England. Since the eighteenth century there had been a vogue for what might loosely be termed anthropological translations, that is, translations of texts from distant cultures accompanied by detailed footnotes that sought to provide insights into the culture. This fashion was linked to developments in tourism and also to the extension of imperial interests around the globe. But within Europe there was a commensurate interest in early Germanic texts, linked to the rediscovery of Anglo-Saxon, which was introduced into universities with the gradual development of English literary and philological studies, and this interest was part of a general movement in the nineteenth century to establish Englishness as the basis of national pride and identity. Several renowned scholars and translators visited Iceland, among them William Morris whose Volsunga Saga appeared in 1875, entitled The Story of Sigurd and the Fall of the Niblungs and Sir Richard Burton, best known as the Orientalist who translated The Kamasutra and The Perfumed Garden. Burton's study of Iceland, Ultima Thule, or a Summer in Iceland also appeared in 1875. The tone is quasi-scientific, with opinion offered as anthropological data. Here, for example, are some of Burton's remarks on Icelandic women: The women at first sight appear tall compared with the men, but not so notably as in the case of the little Welshman and his large wife. They are, as they should be, better looking than their mates, whilst the chubby and rosy children are better looking than their mothers. The expression of countenance is hard and uncompromising. We involuntarily think of "those chilly women of the north who live only by the head"; and they gorgonize us into stony statues. Regularity of features is hardly to be expected so near the Pole.

94

TEXT TYPES AND POWER RELATIONS

The vogue for sagas was connected to the vogue for an heroic proto-Germanic past, and translations begin to trickle into English through the second half of the nineteenth century. Most of those translations are unreadable today, having been constructed in heavy pseudo-medieval language or extensively censored, despite the feelings of some critics that Icelandic texts were somehow "masculine".3 The violence of the sagas, particularly where women were concerned, was clearly problematic for Victorian readers, whilst the translators were concerned with creating a glorified impression of an idealized heroic past that all too frequently clashed with the content of the sagas themselves, composed as they were in a much bloodier age. The sagas entered English literature at a point when the canon was already formed. It is one of the ironies of literary history that these great texts should have somehow managed to remain outside the European mainstream. They are great classical works and yet they are marginalized, and one explanation for this lies in the fact that they only began to be translated at a point in time when literary canons were already well-established. Moreoever, the epic form had ceased to have any real significance in English literature by the nineteenth century. The sagas, for all their reputed greatness, remained curiosities. And if we look at the actual translations, it is possible to discern the sense of uneasiness that the translators must have felt, as they struggled to recreate texts that could have no proper function in the target system, despite the best of intentions. In 1861, the Icelandophile George Webbe Dasent published his translation of Njal's Saga, under the title The Story of Burnt Njal or Life in Iceland at the end of the Tenth Century. The title indicates that he had a non-literary agenda, in that like Morris and Burton, he was interested in conveying a sense of both Iceland's geography and history to English readers. His lengthy introduction and extended appendix contained details of Icelandic history and culture, for the text was perceived primarily as an historical document. The task of the translator was therefore not only to render the text into English, but to explain its historical credibility. Webbe Dasent's was to remain the standard translation for decades, and was republished in a revised form in 1912 in the Everyman Library series as The Story of Burnt Njal. Bayerschmidt and Hollander's 1956 version and Magnusson and Pahlsson's 1960 version both similarly contain a lengthy introduction, setting the saga in its historical perspective. The Bayerschmidt and Hollander version refers to Dasent's translation and complains about its Victorian prudery. Magnusson and Palsson complain that Dasent's version had a "deliberately archaic flavour", and was too literal in rendering Icelandic style and syntax, making the translation "unnecessarily alien to the modern reader." Both editions also provide the

SUSAN BASSNETT

95

reader with a breakdown of the storyline, though Magnusson and Pahlsson state that "it is impossible to summarize briefly the "plot" of Njal's saga", before going on to attempt just such a summary, to little avail. The provision of this kind of additional information is very interesting, as are all translator's notes and prefaces. What both these texts seem to be saying is: we are translating this Icelandic text which you English readers will not understand, you will not be able to follow the plot, the whole text is confusing, we have not translated the verse forms or the poetic diction but doubtless you will be better people if you manage to reach the end.' This is simplistic, of course, but it indicates that the translators are impelled principally by the status of the source text, and less by the needs of the readers. Interesting though the introductions are, they do not bode well for the pleasure of reading that is meant to follow. For Dasent appears to have set a model for later translators to follow, in that he ranked the pleasure of the story being told less than the historical authenticity of the tale. What this means is that the sagas were introduced into English literature not so much as literary texts, but rather as historical documents and this convention has continued into the twentieth century. Let us take at random one episode. In chapter 6 the noble Hrut is about to leave Norway and return to his bride Unn. Bayerschmidt and Hollander summarize the plot as follows: In Norway, Queen Gunnhild singles him out for her paramour and helps him regain his inheritance. But in her jealousy at his leaving her to return to his bride in Iceland she lays a curse on him which prevents him from cohabiting with Unn so that, on her father's advice, she divorces him. When sued by Mord for not returning her dowry, Hrut challenges him to single combat instead, which the old man, of course, refuses. Magnusson and Palsson, however, offer a different version of the same event: For Hrut's marriage to Unn is blighted by the witchcraft of the nymphomaniac Queen Mother of Norway; and when Unn later divorces Hrut, his pride is so affronted that he refuses to return her dowry. Now Hallgerd reappears, strikingly beautiful, arrogant, volatile, passionate and ruthless. She marries twice (956 and 959); both her husbands have occasion to slap her face and promptly die for it at the hands of her jealous and unruly foster-father, Thjostolf. The information selected by the two sets of translators is strikingly different, and we may well wonder how we are to sort our way through this narrative when even the summaries are so widely discrepant. It seems clear that any sense of "plot summary" is widely divergent, and the value judgements of the

96

TEXT TYPES AND POWER RELATIONS

two versions are also very different. The Magnusson version offers "historical fact" in the form of dates, but far more subjective, impressionistic views on the situation. The Queen of Norway is a "nymphomaniac", Hallgerd is described with no less than five adjectives and an adverb. The information provided in the two passages is radically different, and is due to the interpretation given to the text through extratextual material selected by the translators.

Difference and translation Of course we shall always find differences in translations. It is common knowledge that if twenty translators take the same poem, we end up with 20 poems, which seems like a very good reason for rejecting any notion of the absolute authority of the source text. It is also common knowledge that conventions vary and are not consistent: the language and style that is acceptable at one moment in time may be antiquated and rendundant within a few years. There is such a thing as the life span of a text, and here we encounter the great paradox of translation because, as Walter Benjamin reminds us, it is so often translation that ensures that the life of a text continues after its allotted span. Translation can ensure the survival of a text, can even bring about its regeneration or its resurrection. But the translation of the sagas does not appear to have prolonged the life of these texts, rather it appears to have preserved them in a metaphorical specimen jar. The sagas were introduced into English literature at a point when the canon was already so well-established that there was little room for innovation. What there was, however, was an anthropological interest in "primitive" cultures, where "primitive" was used to define cultures outside Europe or distanced in time, as was the case with Iceland. Presented to the reader as curiosities, as texts that should not be seen as epic poems but rather as "authentic" histories, even though the level of violence in the texts was softened in accordance with Victorian taste, the sagas failed to impact upon the English literary system. That the early translations of Dasent and Morris were written in an archaic English also lessened the possibility of the sagas being seen as anything other than exotic texts, with no real significance for English readers. The case of the translation history of the sonnet form and the saga illustrate the difficulty of generalizing about patterns of literary translation. Both types of text may be seen as potentially innovative to the target system, but only one entered into that system. The sonnet was transferred into English, with modifications to its basic formal structure and acquired a status within the target system that has continued through several centuries. The saga, in contrast, remined outside the English literary system and successive translators have endeav-

SUSAN BASSNETT

97

oured to compensate for the failure of the text type to acquire any literary significance by providing readers with additional extra-textual information. That this is also linked to the status of the Icelandic language may also be posited, for translations of Greek and Latin epic poetry, not to mention translations of Italian epic were undertaken with greater success throughout the nineteenth century, presumably because there was already an established tradition into which new translations could be slotted. Icelandic remained a distinct minority taste, despite the interest shown in the language and culture by certain individuals. What seems clear is that the needs and conventions of the target system will condition what happens in the translation process and the eventual fate of the translation. This is not an historical constant, but it is certainly the case in English literature of the nineteenth and twentieth centuries. It is not that the English, like Nietzsche's assessment of the Romans, or the French in the seventeenth century, were bold and foolhardy translators, who cared little for the context and status of the texts they translated, but rather that there was resistance to foreign literary imports. This meant that it was extremely difficult for translated texts to have any innovatory impact upon the target system. Regardless of the type of text, or of its status in the source system, the destiny of the translated text was determined by the constraints of the target system. Notes 1 The rogue sonnet of 16 lines, employed by George Meredith in his Modern Love may appear to contradict this statement. However, some critics consider this form to be a verse monologue, rather than a variant on the sonnet, and in any case, the impact of this text depends on the reader's recognition that the 16 lines are violating an accepted formal code, just as the subject matter of the sequence of poems violates Victorian moral and social norms. 2 I am grateful to Theo Hermans for his brilliant interpretation of the case of Dolet in a hitherto unpublished lecture, given at the University of Warwick, Spring, 1993. 3 For a discussion of the "masculinity" of William Morris' translations of the sagas, see Bassnett (1996). References Bassnett, Susan 1996. "Engendering Anew: Translation and Cultural Politics". In Malcolm Coulthard (ed), The Knowledges of the Translator: from Literary Interpretation to Machine Translation Lampeter: The Edwin Mellon Press pp. 53-66. Bayerschmidt, Carl F. and Lee M. Hollander (eds and transi.). 1955. Njal's saga London: Allen and Unwin. Burton, Richard F. 1875. Ultima Thule; or a Summer in Iceland Vol. I. London and Edinburgh: William P. Nimmo.

98

T E X T TYPES AND POWER RELATIONS

Dasent, George Webbe. 1861. The Story of Burnt Njal or Life in Iceland at the end of the Tenth Century 2 vols. Edinburgh: Edmonston and Douglas. Even-Zohar, Itamar. 1978. "The position of translated literature within the literary polysystem". In James Holmes, Jose Lambert and Raymond Van den Broeck (eds), Literature and Translation: New Perspectives in Literary Studies (Leuven:ACCO). Helgasson, Jon Karl. 1995. The Saga of Njal's Saga: A Study in Rewriting. Unpublished doctoral dissertation, University of Massachussetts, Amhersts. Holmes, James. 1970. "Forms of verse translation and the translation of verse form". In James Holmes, The Nature of Translation. The Hague and Paris: Mouton. Magnusson, Magnus and Palsson, Hermann (eds and transi.). 1960. Njal's Saga, Harmondsworth: Penguin. Nietzsche, Friedrich. 1992. "On the problem of translation", transi. Peter Mollenhauer. In Rainer Schulte and John Biguenet (eds), Theories of Translation: An Anthology of Essays from Dryden to Derrida. Chicago: University of Chicago Press pp. 6871. Petrarca, Francesco. 1954. II Canzoniere, (Milan: Rizzoli). Poem no. 189. Rider, H. 1638. All the Odes and Epodes of Horace: Translated into English Verse by Henry Rider, Master of Arts of Emanuel College in Cambridge. London: R. Rider. Rieu, E.V. 1946. Introduction to The Odyssey. Harmondsworth: Penguin. Spiller, Micahel R.G. 1992. The Development of the Sonnet: An Introduction. London: Routledge. pp.2-3. Vaganay, Hugues. 1899. Le sonnet en Italie et en France au XVI ième siècle. Louvain. Voltaire, extract from a letter to Anne Dacier, 1720, reprinted in André Lefevere (ed), Translation/History/Culture:A Sourcebook. London: Routledge, p. 30. Wyatt, Sir Thomas. 1978. The Complete Poems. R.A. Renholtz (ed). Harmondsworth: Penguin.

Description and Criticism: Some Approaches to the English Translations of Hans Christian Andersen Viggo Hjørnager Pedersen The University of Copenhagen The English translations of Hans Christian Andersen form a vast subject, and the study of them is complicated by the fact that for many Danish scholars and most Andersen specialists are Danes - Hans Christian Andersen has the status of a cult figure, whose sacred words, in an ideal world, no translator should be allowed to tamper with. Discussions of translations of Andersen therefore usually take the form of criticism of the failure of a given translator to render adequately some detail(s) of the original. Unlike some modern translation critics, I would not condemn such an approach out of hand; but it is clearly inadequate unless supplemented by other points of view. Particularly important is the fact that translation debunking is unlikely to appeal to native speakers of English, who are less interested in the subtleties of the Danish language than in understanding the texts that they are actually reading. In addition, they may perhaps also be interested to hear about the reasons for the differences between English versions of the 'same' text. On the other hand the dissatisfaction with translated texts is too general to be ignored. But rather than raging about the loss of individual beauties or the distortion of details, it is more valuable to look for general tendencies that may help to explain this dissatisfaction. The aim of the following paper, therefore, is twofold: first, a number of different approaches to the study of the English translations of Hans Christian Andersen will be examined; and, secondly, a special area, that of the translation of idioms and semi-idioms, will be explored in more detail in the hope that this will lead to the discovery of grounds for the dissatisfaction mentioned above. In conclusion, some general observations will be made concerning the status of literary translations vis à vis their originals, and concerning the respective roles of description and criticism in literary translation studies.

100

DESCRIPTION AND CRITICISM

The problems In her Translation Studies Snell Hornby (1988: 1-31) contrasts two schools of translation scholars: 1. The linguists, who, in her opinion, tend to be equivalence-addicts; which is bad, for, as classical equivalence is a chimera, the 'Quest for 'Equivalence" is hardly a promising research strategy, since it is bound to end in despair1. 2. The new 'manipulators', who concentrate on the target text and on literary approaches, but in order to describe, not to criticize literary translations. A couple of quotations may be illuminating here. In the introduction to his book The Manipulation of Literature, Hermans (1985: 8) criticizes The strongly evaluative orientation of literary criticism, ... [which] has meant that translation has found itself consistently relegated to the periphery ... And concerning the new group, introduced in the book, he says (1985: 7) that it has: An approach to literary translation which is descriptive, target-oriented, functional, and systemic; and an interest in the norms and constraints that govern the production of translations ... Lambert and Gorp (in Hermans: 1985) have systematized the analytical approach of this group, suggesting an analytical model that rather resembles a bill of fare, offering a little of everything. In any case, they are mainly interested in the contexts in which translations are made, and actual comparison of linguistic details is placed rather low on their agenda. Still, an analysis of this kind might perhaps be attempted with Andersen in English - one might, so to speak, do a Lambert and Gorp on him. But such a project is obviously beyond the scope of an article. Instead, I have decided to use the change of approach to the study of literary translation that is taking place at the moment as an opportunity to take stock of my work to date, by describing some of the work I have done over the last 10 years as illustrating a number of different approaches hopefully of interest to scholars working on similar projects.

The translations First, however, a word about the material to be studied: the English translations of Andersen. There are many. Nobody knows how many, but there are more

VIGGO HJØRNAGER

101

than 30 sizable editions of the tales, including about half a dozen more or less complete ones.2 However, these editions are not completely independent of each other; borrowing from older translations seems to be the rule rather than the exception. Some are by Danes, others by Englishmen, others still by Germans. It would be an oversimplification automatically and generally to prefer one group to the others. Native speakers of English tend to be inaccurate - through insufficient knowledge of the source language, or simply because they cannot be bothered as when, in "The Nisse at the Grocer's", Caroline Peachey reduces the objects in the shop specified by Andersen to 'sundry other articles'. Danes, on the other hand, tend not to be free enough, and not good enough at English - though, who is? One reason why translations are so often seen as inadequate is that as a rule they are not made by the best people - the best people are busy producing their own stuff. Germans, generally, have been better at handling the Danish than British or American translators, but also somewhat wooden and unidiomatic, frequently with interference from German in their work. This applies, as we are going to see, to the (in)famous Dr. Dulcken, the first to bring out a translation comprising the majority of the tales. In addition to the complete translations, there are many shortened and adapted ones, particularly of the 20 - 30 most popular tales, and these abbreviations obviously belong in the total picture. In the case of the most popular tales of them all, such as "The Ugly Duckling" there are certainly several hundred versions, and the number may well run into four figures.

Approaches Now, what can one do with a material like this? In the following, I shall describe some of my own approaches (cf also Pedersen 1991a). Hans Andersen as an English writer The article "Hans Andersen as an English Writer" (1984) stressed the place of the translation in the target culture - something which, in spite of the efforts of Hermans and other 'manipulators', is not readily recognized even today. A translation belongs both in the source language and in the target language literary system. And if, as in the case of Andersen, the author takes an active interest in the translation, this is one reason more for regarding it as a version of the 'work' as understood by Wellek and Warren, i.e. an ideal 'structure of norms' beneath the surface manifestations of the text (cf Pedersen: 1990c). In

102

DESCRIPTION AND CRITICISM

other words, the translation is one among several manifestations of the work, and a translation may in fact come to influence the way the original is understood and interpreted: Kierkegaard would not have the status and the following that he has in Denmark today if translations of his work had not appealed to French existentialists. In Andersen's case, it is worth noting that he enters fully into the English literary system in that he is frequently quoted and imitated, even today, and that the genre of Children's Literature in English would have been very different but for his presence. Rankbound comparison of target language with source language text To facilitate contrastive studies of source and target texts, I have worked out an analytical model which is based on the Firth/Halliday tradition, and which ensures that no level of analysis is forgotten. The idea is to analyse the text level for level and element for element, beginning with larger units like paragraphs, and working downwards towards individual words (cf appendix 1). Of course this method works best if the translation does follow the original to a reasonable extent, as in my analysis of four English versions of "The Little Mermaid" (Pedersen 1990b). But even with abbreviated texts it can be used up to a point, as in Pedersen (1990a and 1991b); and, in any case, analysis will not be equally detailed at all levels in each case: the idea is to concentrate on those levels and phenomena that seem most interesting in each individual situation. With Andersen, I have found the level of paragraph relatively uninteresting, whereas the levels of sentence, phrase, and word are about equally interesting: all translations of Andersen introduce syntactic changes, and Andersen's phraseology as well as his vocabulary presents many difficulties. Abbreviated texts The majority of modern English translations of Andersen's tales are shortened versions. Therefore, I have found it interesting to analyse the ways in which a shortened version of a tale differs from a complete text. This work has led to some unexpected results, for example that an abbreviated tale tends to approach the folktale norm: happy endings are introduced (several times in "The Little Mermaid", cf the Walt Disney film version), and social satire and person-centred sentimentality tend to be left out, together with 'literary', romantic features like descriptions of scenery. Thus several descriptive scenes in "The Ugly Duckling" go by the board, or are drastically reduced (cf Pedersen 1990a). The fact that abbreviated versions are aimed specifically at children also has consequences. Often the translations concentrate on the surface action rather

VIGGO HJØRNAGER

103

than on the deeper message, and elements that are considered unsuitable for children, such as violence or references to sexuality, tend to be toned down or left out: in Peachey's "Tinderbox", the soldier is not allowed to kiss the princess like the true soldier that he is in Andersen's text, but only to kneel down and kiss her hand. And in several versions the scene from "The Ugly Duckling" in which the two young ganders ate shot so that the water is coloured red by their blood is either left out or made less sanguinary. However, normally the core of a tale remains even in drastically reduced versions. For instance, "The Ugly Duckling" is a fairy tale to the extent that it is about a hero of rather unpromising background, who sets out to find a place in the world, and succeeds - though the duckling is rather more passive than a traditional fairy tale hero. This success aspect is always present, and almost invariably linked to the idea of growing up. A critical edition of Andersen in English Pedersen (1987b) discusses an untraditional way of presenting the vast number of data relevant for the discussion of a literary translation. A plea for the preparation of a critical edition of Andersen is presented. Such an edition would have Danish text on the left-hand pages, and English on the right-hand ones. The translation could either be a new text especially prepared for this edition, or an existing translation, possibly with some emendations. Under the Danish, there would be a commentary (in English) on the text, whereas important variants to the translation chosen would be registered in an apparatus under the English text. A sample page from the opening of "The Nisse at the Grocer's", which presents a number of difficulties and exists in several rather different versions, is shown in Appendix 2. A contrastive study of vocabulary, including phraseology The Center for Translation Studies and Lexicography, Copenhagen University, and the Hans Christian Andersen Center at Odense University are cooperating with the Royal Library, Copenhagen, in establishing a machine readable corpus of Andersen texts. So far the Danish standard edition of the tales by Erik Dal has been copied and stored in Macintosh format at the Hans Christian Andersen Center, and the Jean Hersholt translation at the Center for Translation Studies. The texts are also available in IBM compatible format. Initially, the texts were typed in, a slow and laborious process, but scanning began in 1993. Pedersen (1993a) describes a number of ways of analysing this material: by means of a simple WP macro it is possible to search for and copy individual words in micro context. Words subjected to this form of analysis include high-

104

DESCRIPTION AND CRITICISM

frequency, emotionally loaded adjectives. Meanwhile the corpus + 5 of Hans Christian Andersen's novels (in Danish) have become available on CD-ROM with a WORDCRUNCHER programme. This programme has been used to establish word frequency lists. What is still lacking is a format that will allow rapid comparison of large quantities of Danish and English material. A better knowledge of Andersen's vocabulary is obviously an advantage for translators and scholars, even without the contrastive element. A natural next step would be to prepare an Andersen concordance, and, on the basis of that, a monolingual Andersen dictionary. This, in turn, might build on the linguistic analysis already present in the ODS - the Danish equivalent of the OED - and the Holberg dictionary. A bilingual Andersen dictionary is hardly a commercial proposition. But in a database version, typical English (German, etc.) equivalents of the Danish items might be listed. Special problems of equivalence A special problem is the patterning of Andersen's work due to the frequent recurrence of certain words and phrases. Pedersen (1993a) discusses among other things the difficulty of preserving this pattern in translation, and the problems that occur when typically Danish phenomena without any equivalent in English are to be rendered3. These include Danish institutions, place-names, and historical data whose associations are obvious for Danes, but not for anybody else. Thus in "The Galoshes of Fortune" reference is made to Kong Hans' Tid - the time of King Hans, i.e. 1481-1513 - an allusion that will be meaningless to the vast majority of English readers. Individual translators I am conducting a series of studies of the work of individual translators, in particular early translators, such as Mary Howitt, Caroline Peachey, and Clara de Chatelain (cf Pedersen 1993b). In these studies I try to take into account the background of the various translators, including their other literary activities, and the situation in which they produced their translation. Thus it is significant that the three ladies mentioned above intended to produce suitable reading for children according to the norms applying to children's literature at the time. My procedure is obviously close to that of the Dutch-Israeli manipulators. But in my view such studies should not confine themselves to description, but should include interpretation and criticism, cf p. 111 below.

VIGGO HJØRNAGER

105

Idiomaticity: Det er ganske vist/It's Quite True (Dulcken)/It is Perfectly True (Haugaard) The point of the heading of this section, devoted to work in progress, is that something which is proverbial in the Source language is not necessarily so in the target language. The Danish phrase just quoted has a resonance which the English translations do not. In the following, I shall apply this observation to an examination of some text examples, taking "Nattergalen" (The Nightingale) as my point of departure. About the artificial as opposed to the real nightingale the poor fishermen in H.C. Andersen's tale "Nattergalen" have the following to say: "det klinger smukt nok, det ligner ogsaa, men der mangler noget, jeg ved ikke hvad!" (A4: 23)4 "it sounds beautiful, and like the bird's song, but something is missing, though I don't know what it is." (Haug: 208) When considering the English translations of Hans Christian Andersen's tales this may be a reasonable starting-point: if for a moment we regard the translation as artificial and the original as real, what is lacking in the TL version is idiomaticity. As a rule, literary translations contain fewer idioms than originals do; and individual words tend to be less justes than those of the original - in particular if the translator tries to render as much as possible of the semantic content of the original. However, it should be stressed at the outset that of course Hans Christian Andersen can be translated into English. He is neither more nor less difficult to translate than other great writers and poets. It is just as easy - or difficult - to translate Andersen into English as it is to translate Shakespeare into Danish. In both cases, even a poor translation will preserve so much of the original work of art that the TL audience will feel the presence of something great and unusual. But at the same time this audience will often wonder about oddities in the text which are due to the translator's inability to bridge the gap between the two languages. Or, if the audience does not feel the presence of such a gap, the translation critic will most likely be able to prove that the translator has left out or added details, or reformulated the original text. Which is regrettable, but often necessary in order to obtain a decent TL work. Therefore, the following is not primarily to be regarded as a criticism of translation 'mistakes' but rather as a demonstration of the presence of a number of problems which, difficult and often insoluble as they are, make it impossible to reach the perfect translation. Part of the value of such an exercise, therefore,

106

DESCRIPTION AND CRITICISM

lies in the fact that it forces scholars to focus on a number of textual details which only attract attention when they are looked at from a translation point of view.

Reading the text Reading the text is the first step in any analysis. Like so many of Hans Christian Andersen's tales, The Nightingale' is at one and the same time easy and difficult. The story about the natural as opposed to the artificial, and about the differences between the Emperor of China and the poor fisherman, can be understood in any society with some considerable distance between the top and the bottom. Nor is the setting unduly difficult: at any rate in Andersen's own day, interest in things Chinese was general in Europe and thus ought not to cause any particular difficulties in translation. But, in spite of that, there are special Danish and special Andersen nuances. In Hans Christian Andersen's time, Copenhagen was in every respect a small city with less distance between the various sections of society than was the norm in Europe. It was a patriarchal society, one in which the King Frederik VI - received the keys of the city every evening when the gates were locked, put them under his pillow and kept them there till the next morning. This is part of the background of Andersen's feeling on intimate terms with kings and emperors - besides the fact that he had known Frederik VII as a boy, and played with him, and it is an attitude which has often proved difficult for English people to grasp; and even if this emperor does not, like the one in "The Swineherd" run down to the pigsty in his slippers, it is still essential to preserve an intimate, informal tone, as when the Emperor is juxtaposed with the street arabs (A: 24): Gadedrengene sang "zizizi! klukklukkluk!" og Keiseren sang det -! jo det var bestemt deiligt! The street arabs sang "zizizi! klukklukkluk!" and the Emperor sang it -! it was indeed wonderful! which is not quite adequately rendered by Peachey's '... and the Emperor himself sang too' (my italics): Peachey expresses wonder at the fact that proletarians and emperors may react in similar ways; Andersen takes it for granted.

VIGGO HJØRNAGER

107

Pragmatics This seems a suitable point to embark on the question of pragmatics - phenomena which Andersen's original audience were likely to understand differently from a modern English-speaking one. Although the distinction between pragmatic and linguistic difficulties is in one sense irrelevant in translation, since both have to be overcome by linguistic means, it is perhaps useful to preserve it all the same. Certainly a pragmatic problem, once it has been grasped, may usually be overcome by circumlocution, substitution of a target language phenomenon for a source language one, or simply by adding a few words of explanation. This is not always the case with linguistic problems. The point of treating the matter here is that part of the idiomatic nature of a text is that it draws freely on the audience's knowledge of a shared linguistic and cultural background. A couple of examples will have to suffice here. Andersen loved things Chinese, and used them frequently. However, in this story one detects a mild irony which is not necessarily grasped by an innocent translator. A phrase like det var nu saa ganske chinesisk (A: 23) ('this was indeed very Chinese') is probably partly, at least, to be understood in the light of meanings listed in ODS such as mandarinagtigt, gammeldags, omstændeligt, uforståeligt ('mandarin, old-fashioned, longwinded, incomprehensible'), as exemplified in the following marvellous quotation from P.O Brønsted: Vi talte om denne erbarmelige Nation (i.e. the French nation), dens chinesiske Væsen og onde Noder. We talked of this pitiful nation, its chinesisk nature and evil ways. Similar senses are not listed in OED, so perhaps the proper translation of chinesisk is 'mandarin' rather than 'Chinese'. Another problematic word is Spekhøkerbørn {elleve Spekhøkerbørn bleve opkaldte efter den, men ikke een af dem havde en Tone i Livet. (A: 22 - for translations, see below)). This is clearly to be understood ironically. But many English translators fail to register the fact that to Andersen's contemporaries, the word Spekhøker was in itself comical, denoting a philistine (and not a creature from the lower orders, like the various variants of English 'shopkeeper'). Cf the poet Heiberg: "Hvems er det deilige Barn?" "Det tilhører en Spekhøker i Adelgaden. " "En Spekhøker! Det er aldrig muligt! Hvor Fanden skulde en Spekhøker komme til et saadant Barn! " "Whose is that lovely child?" "It belongs to a Spekhøker in Adelgade." "A Spekhøker! impossible! How the devil should a Spekhøker come by such a child!"

108

DESCRIPTION AND CRITICISM

Hersholt translates the word as 'pork-butcher', Dulcken as 'pedler'. Other solutions are 'provision-dealer', 'huckster', and 'grocer' (Peachey). The modern equivalent is probably 'delicatessen shop owner' (Haugaard) - but then the story is set in the 19th century, not in the 20th. Theevand as in drukket sig lystige i Thevand (A: 23) is also a word whose understanding presupposes knowledge of the source culture. Theevand (literally 'tea-water') was formerly synonymous with tea, however, with a suggestion of it being weak (the traditional Danish way of serving this beverage, which differs markedly from the English). A proper translation would be 'weak tea'; but most translators miss this nuance: D: tipsy upon tea, for that is quite the Chinese fashion; Haug: got drunk on too much tea; H: gotten tipsy on tea, Chinese fashion. Though it also fails to render this nuance, there is a lot to be said for Peachey's free version: as if they had all been drinking tea; for it is tea that makes the Chinese merry. Loanwords, though common to SL and TL, need not have the same connotations in the two languages. Couleur (A: 20) is often the opposite of black and white; the word was apparently common in the 19th century. However, a phrase like kostbare Steene af adskillige differente Couleurer ('costly stones of several different colours') suggests that the word is felt to be a loanword to a greater extent than in modern Danish, since different is a rare loanword in Danish; couleurer might perhaps be rendered by 'hues'. Succès (den vil gjøre en stor succès ved Hoffet A: 21) also seems relatively rare, although common today. There is just the one example in the MAGNUScorpus, which makes one suspect that at the time the word was perceived more as a loanword than it is today. Andersen here as in "The Swineherd" seems to be using affected French to characterize courtiers of whom he disapproves. It is probably impossible to render this nuance in English. But the fact that it is necessary to 'undertranslate' the word should be kept in mind with a view to compensating for the loss elsewhere in the text. Presenter (A: 23) in the sense of 'gifts', modern Danish gaver, is a period marker without any English equivalent; but of course the loss could and should be compensated for elsewhere in the text.

Puns To exemplify the difficulty of overcoming linguistic problems let us take the punning on Nattergal:

VIGGO HJØRNAGER

109

Hele Byen talte om den mærkværdige Fugl, og mødte to hinanden, saa sagde den Ene ikke andet end: "Nat-! " og den Anden sagde "gal! " og saa sukkede de og forstode hinanden ... (A: 22) The whole town talked about the marvellous bird. Whenever two people met in the street they would sigh; one would say, "night", and the other "gale"; and then they would understand each other perfectly. (Haug: 207). Dulcken's 'Nightin ... gale' is rather less amusing than Andersen, since gal in Danish means 'mad' or 'besotted', and other translations are hardly better, if attempted: P: -; H: 'one could scarcely say "night" before the other said "gale"'. The point is, however, that there is no satisfactory solution. The translator's choice is to under-translate or to compensate elsewhere.

Idioms The translation of idioms is as difficult as it is central. There is no general agreement about the definition of idiom. To some linguists, all stock phrases or habitual collocations are idioms. Others require that they should not be transparent, so "once in a blue moon" would be an idiom, whereas "that's the way it is" would not; However, idiom is here also used in the sense of 'characteristic way of expressing oneself'. Some idioms have only one fixed form, whereas in the case of others a number of variant forms may be found. There is also a grey area where it is debatable whether we have idioms or not - but then again we certainly do not have free variation. For example, det gik ret til Hjertet ('it went straight to the heart') (A: 21), where it is not easy to think of an acceptable substitute for the noun Hjertet, if the rest of the string is preserved, even though native speakers would hardly think of the phrase as an idiom. The literature on the subject is limited: there is a survey in Fernando and Flavel (1981), and a discussion of some of the problems involved in Pedersen (1986). Some points may also be gathered from Kjær (1990). Idioms give flavour to a text, and their absence consequently impoverishes it. That an idiom sometimes has no equivalent in the target language poses no immediate problem; it can be rendered by an unidiomatic expression, or by a single word. But if this solution is embraced regularly in a given text, the translation becomes flatter than the original. A few examples from "The Nightingale": saa at det var en Lyst (A: 20) // D: 'most delightfully'; H: 'burst into song'; Haug: 'as beautifully as he could'.

110

DESCRIPTION AND CRITICISM

man kunde ikke høre Ørelyd (A: 21) // D: 'one could not hear oneself speak'; H: 'you could scarcely hear yourself talk'; Haug: 'you couldn't hear what anyone said'. Land og Rige (A: 23 & 25) (literally 'land and realm') is one of those tautologies Andersen was so fond of. It is not easy to imitate: D: 'country and empire'; H: 'land'; Haug: 'empire'. ikke een af dem havde en Tone i Livet (A: 22) // D: 'not one of them could sing a note' ; H: 'not one could sing' ; Haug: 'not one of them could sing'.

Pseudo-idioms If the literature on idioms is thin, that on what have here been called pseudoidioms is thinner. I have set out the idea in Pedersen 1986, viz. that idioms constitute one end of scale whose other end is free variation; "once in a blue moon" is an idiom, "twice in a bad year" is free variation, because you might just as well say "three times in the peak season". But in between we have not only a number of idioms with latitude, like "when in Rome, do as (or like) the Romans / a Roman / (do)", but a number of patterns which aspire to idiom status. Thus the phrase from "Kjærestefolkene" ("The Sweethearts") ligget i Skuffe sammen (literally 'lain in drawer together' (A: 29)) presupposes a pattern of vb + a prepositional phrase with i denoting locality + sammen: sove i seng sammen, komme i by sammen ('sleep in one bed, meet in the same town'), etc., which it is difficult to imitate in translation. The implications of ligge i skuffe sammen are clearly sexual, which need not be true about translations; cf Peachey: '... since we are thrown so much together'. Here are two more examples: It is said about the professor in "The Flea and the Professor" after the departure of his wife that han tabte sit gode Humeur (literally 'he lost his good humour'). As we know, Andersen set great store by et godt Humeur, the title of one of his delightful little stories, but the point here is the tone, which is jocular, but sad. If we look for parallel examples in the Vinterberg and Bodelsen Danish/English dictionary we are struck by the fact that 'gode + possesive pronoun + noun' often collocates with something negative: gå ud af sit gode skind, miste sit gode navn (literally 'go out of one's good skin' (i.e. be very upset), 'lose one's good name'). A parallel example from the same story is en kold Fornøielse ('a cold pleasure'). It is of course a case of contradictio in adjecto, and there is no obvious English equivalent. But in fact there is a sub-pattern where fornøjelse - a

VIGGO HJØRNAGER

111

word that is normally positive - collocates With negative adjectives. The result is a certain jocularity of tone: en dyr/kort (i.e. 'expensive/short') ƒ Most translations fail to satisfy here, as they are either inaccurate or non-idiomatic: D: -; H: 'a chilly sort of pleasure'; Haug: 'no fun'.

Conclusion: Description versus criticism It will be obvious that the above consists of hints and guesses - notes towards a description of the 19th-century English translations of Andersen. However, at least two conclusions may be drawn: 1. Translations do tend to be inferior to originals from an artistic point of view, unless the translator is prepared to be very free, skipping bits that cannot be translated adequately, and possibly adding embellishments of his own devising. This is not a new observation. Cowley, du Bellay and other 16th and 17th century scholars said the same (cf Pedersen (1987a: 13-26). However, it is a fact which has been so completely ignored since the beginning of the 19th century that it is necessary to restate it with some emphasis. 2. Translations must be evaluated both in relation to target language norms and to the work they purport to render; and, as far as I am concerned, the objective description of a translation should be followed by a critical evaluation of it: indeed I regard it as an important responsibility for a translation scholar to furnish such a criticism, as nobody else is in a position to do so. Both these statements, I know, will seem controversial to a number of colleagues. But it seems to me an illusion that any student of the arts can limit himself to objective statements about his object of study; and I believe that the time has come for a return to translation criticism, provided that criticism is based on careful analysis of the texts concerned, and provided that it does not forget the importance of the TL context. It seems to me an important, if subjective, statement that Hans Christian Andersen is not as good in English as he is in Danish, whereas with Shakespeare, the reverse is the case. But, even with the inevitable losses or changes sustained in translation, both are important and welcome additions to the literature of any target language, and without their presence in translation, both Danish and English literature would be very different from what they are.

112

DESCRIPTION AND CRITICISM

Notes 1 Cf Jennifer Draskau (1987), which explores the possibility of establishing equivalence between Francois Villon's original poems and English translations and concludes that in spite of the ingenuity of many translators, it is often impossible to achieve such equivalence. 2 Examples are the translations by Peachey, Dulcken, Hersholt, and Haugaard. 3 For a similar approach, see Knowles and Malmkjær (1991). 4 A = Andersen (1964-91); D = Dulcken; H = Hersholt; Haug = Haugaard; P = Peachey; - = not translated; ODS = Ordbog over det danske sprog (A Dictionary of the Danish Language). References Andersen Editions Andersen, H.C. Eventyr og Historier. 1964-91. Ed. by Erik Dal and Erling Nielsen. Copenhagen. Andersen, H.C. (1889) 1983. The Complete Illustrated Stories, (transi. Dulcken). London. Andersen, H.C. 1942-47. The Complete Andersen (transi. Jean Hersholt). New York. Andersen, H.C. 1985. The Complete Fairy Tales and Stories (transi. E.C. Haugaard, 1974) Harmondsworth: Penguin. Andersen, H.C. 1852. Danish Fairy Legends and Tales (transi. Peachey). 2nd enlarged ed. London. Other Sources Draskau, J. 1987. The Quest for Equivalence: On Translating Villon. Copenhagen. Fernando, C. and Flavel, R. 1981. On Idiom. Critical Views and Perspectives. Exeter. Hermans, Th. 1985. The Manipulation of Literature. London: Croom Helm. Kjær, Anne Lise. 1990. "Phraseology research - State-of-the-art" and "Context-conditioned word combinations in legal language". Terminology Science and Research 1(1-2): 3-32. Vienna. Knowles, M. and Malmkjær K. 1991. "Key terms in H.C. Andersen's fairytales and their translations into English". Babel 37(4): 203-12. MAGNUS. 1992. A CD-ROM issued by CD-Danmark, Palægade 4. Copenhagen. Pedersen, V. H. 1984. "Hans Andersen as an English writer". In H. Ringbom and M. Rissanen (eds), Proceedings from the 2nd Nordic Conference for English Studies. Åbo, 95-109. Pedersen, V. H. 1986. "The Translation of Collocations and Idioms". In L. Wollin and H. Lindquist (eds), Translation Studies in Scandinavia. Lund: CWK Gleerup, 12631. Reprinted in Pedersen (1988). Pedersen, V. H. (1st ed. 1973) 1987a. Oversættelsesteori. 3rd revised ed. Copenhagen: Nyt Nordisk Forlag. Pedersen, V. H 1987b. "A critical edition of Hans Andersen?". In I. Lindblad and M. Ljung (eds), Proceedings from the Third Nordic Conference for English Studies (Acta Universitatis Stockholmiensis LXXIII). Stockholm, 291-99.

VIGGO HJØRNAGER

113

Pedersen, V. H. 1988. Essays on Translation. Copenhagen: Samfundslitteratur. Pedersen, V. H. 1990a. "Ugly ducklings? Reflections on some English versions of Hans Andersen's "Den grimme Ælling"". In P.N. Chaffey et. al. (eds), Translation Theory in Scandinavia. Proceedings from SSOTT HI. Oslo, 229-42. Pedersen, V. H. 1990b. "A Mermaid translated: An analysis of some English versions of Hans Andersen's "Den lille Havfrue"". In Shirley Larsen (ed), The Dolphin 18. Århus: Aarhus University Press, 7-20. Pedersen, V. H. 1990c. "The mode of existence of a literary translation". In Eric Jacobsen et. al. (eds), Essays Presented to Bent Nordhjem (PDE vol. 18). Copenhagen: Department of English, University of Copenhagen, 141-152. Pedersen, V. H. 1991a. Oversættelsesteoretiske Problemstillinger/Aspects of Translation Theory (DAO 1). Center for Translation Studies, University of Copenhagen, Copenhagen. Pedersen, V. H. 1991b. Oversættelse eller parafrase?/Translation or Paraphrase? The Hans Christian Andersen Center, Odense University, Odense. Pedersen, V. H. 1993a. "A wonderful story of a true soldier and a real princess. Problems in connection with the rendition of Hans Andersen's vocabulary in English." In de Mylius et al. (eds), Andersen and the World. Odense: Odense University Press, 197-209. Pedersen, V. H. 1993b. "Mary Howitt's translations of Hans Christian Andersen". In Julian Meldon D'Arcy (ed), Proceedings of the Fifth Nordic Conference for English Studies. (1992). Reykjavik, 286-94. Ordbog over det danske sprog. 1919. (ed. V. Dahlerup et al.). Copenhagen. Snell-Hornby, M. 1988. Translation Studies. An Integral Approach. Amsterdam: John Benjamins. Vinterberg, H. and Bodelsen C.A. 1990. Dansk/Engelsk Ordbog. (1st ed. 1954-56; 3rd edition, ed. V. Hjørnager Pedersen, 1990). Copenhagen: Gyldendal.

114

DESCRIPTION AND CRITICISM

Appendix 1 Analytical figure from Pedersen (1988: 110f) showing at what levels various sorts of modifications of a text may occur in translation. "+" indicates that modification is possible. The model is based on 5 different levels of a Firthian kind, and on 5 different types of operation. The 5th of these, "change" may well be redundant, since it may be redefined as deletion + addition; but it is a practical sort of shorthand to use in the analytical procedure. Obviously more or less obligatory changes do not count - only optional changes are of interest. However, delimitation of this category is far from easy.

Text

Paragraph

Sentence

Phrase

Word

+

+

+

+

+

Compression

+

+

+

+

+

Deletion

+

+

+

+

+

Addition

+

+

+

+

+

Expansion

+

+

+

+

+

Change

115

VIGGO HJØRNAGER

Appendix 2 First draft of a page of a critical edition of Andersen in English. The text is slightly modified Hersholt. Source: Pedersen (1987b: 299). There was a s t u d e n t , who l i v e d i n t h e a t t i c and d i d n ' t own a n y t h i n g . There was a l s o a g r o c e r , who l i v e d on t h e ground floor

and owned t h e w h o l e

the grocer, could

house,

and t h e

nisse

stuck

to

f o r e v e r y C h r i s t m a s Eve i t was t h e g r o c e r who

afford

him a bowl

of

porridge

with

a big

pat

of

butter in i t .

So t h e n i s s e s t a y e d i n t h e g r o c e r y s h o p , and

t h a t was v e r y

educational.

One e v e n i n g t h e s t u d e n t came i n by t h e back door t o buy some c a n d l e s and c h e e s e . He had no one t o s e n d , and t h a t ' s why he came h i m s e l f .

He g o t what he came f o r , p a i d f o r

it,

and t h e g r o c e r and h i s w i f e nodded "Good e v e n i n g . " She was a woman who c o u l d do more t h a n j u s t nod, s h e had t h e

gift

of

was

the

gab.

reading

The

student

nodded,

s o m e t h i n g on t h e p i e c e o f

too,

but

while

he

p a p e r w h i c h was wrapped

around h i s c h e e s e , he s u d d e n l y s t o p p e d . I t was a p a g e t o r n out of

an o l d book t h a t o u g h t n e v e r t o h a v e b e e n t o r n u p ,

an o l d book f u l l of "There's woman

a

more o f

few

eightpence,

coffee

poetry. it,"

said

beans

for

the it.

S i r , you s h a l l have t h e

grocer. If

you

"I g a v e an o l d will

give

me

rest."

1. The Misse at the Grocer's The Goblin and the Grocer H.; The Brownie at the Butterman's Brækstad; The Goblin and the Huckster Dulcken, Sievers; The Goblin at the Provision-Dealer's Kingsland. 2. student H.; real student Peachey; proper student Kingsland; (He was) a student of the good old sort Breakstad. 2. a t t i c Peachey; garret H. 2. owned H.; possessed Brakstad; 6. porridge H.; plum porridge Dulcken; gruel Peachey; jam Mrs. Paul. 6. pat H.; lump Brakstad; piece Mrs. Paul 7. stayed H.; settled down in Brakstad. 7. educational H.; where there was much to learn Brakstad; and was right comfortable there Peachey; which was very cunning of him Mrs. Paul. 11. and that i s why H., Dulcken; so Brakstad. 11. got what he came for H.; procured what he wanted Dulcken. 12. Good evening H.; Good night Brakstad. 13. g i f t of the gab Kings land; she was gifted with a g l i b tongue H.; she had an unusual g i f t of speech Brakstad. 17. torn up Dulcken; put to this purpose H.; torn to pieces Brakstad. 21. eightpence Peachey; eight pennies H.; two groschen Dulcken; sixpence Kingsland.

Part II Domain- and Genre-Specific Texts

Strategies of Translating Political Texts Christina Schäffner Aston University

Politics and translation Research into politics has mainly been concerned with the consequences of political decisions and actions for (the history of) a society. However, politics is constituted to a great extent by text and talk. Not only, of course, since physical and economic coercion are involved, but these too depend on or utilise discourse. In the analysis of political discourse and political texts, the broader societal and political framework in which such discourse is embedded has to be taken into consideration. Van Dijk has recently argued, that "despite some studies on 'political language', discourse and conversation analysis has thus far had little to offer to political science", and he has called for discourse analysis to be a "genuine social, political or cultural analysis" (van Dijk 1994: 164)."These aspects are of equal relevance when we look at political texts from the point of view of translations. Political text itself, however, is a vague term. It is an umbrella term covering a variety of text types1, or genres. Political discourse includes both innerstate and inter-state discourse, and it may take various forms. Examples are bilateral or multilateral treaties, speeches made during an electioneering campaign or at a congress of a political party, a contribution of a member of parliament to a parliamentary debate, editorials or commentaries in newspapers, a press conference with a politician, or a politician's memoirs. The characterisation of a text as political can best be based on functional and thematic criteria. Political texts are a part and/or the result of politics, they are historically and culturally determined (see the contributions in Bochmann ed. 1986). They fulfil different functions due to different political activities. Their topics are primarily related to politics, i.e., political activities, political ideas, political relations, etc. Another characteristic feature is that - in the

120

STRATEGIES OF TRANSLATING POLITICAL TEXTS

majority of cases - they are meant for a wider public. Each individual text is embedded in a wider political discourse, the texts thus showing a high degree of intertextuality (de Beaugrande and Dressier 1981). Political discourse is often of relevance not only for the specific culture of the text producer(s) but may be of interest for a wider audience. Since politics is more and more internationalised, translation, too, is becoming more and more important. Translation is first of all communication, to be precise, interlingual and intercultural, or cross-cultural, communication. It is a source text induced target text production (see Neubert 1985: 18, Neubert and Shreve 1992: 7). It is usually the case that the source text (ST) itself fulfils a particular function in the source language (SL) community, at a particular time, at a particular place, addressed to a more or less specific audience with knowledge about the subject of the text and probable text-typological conventions. When the text is translated, it will address a new audience in the new target language (TL) community, at a different time and place (Neubert 1985: 71 speaks of displaced situationality). The subject knowledge of the target text (TT) addressees may be more or less the same as that of the ST addressees. Depending on the particular text type, there may be more or less conventionalised text-structures (in the sense of van Dijk's 1980 superstructures) or preferences for specific syntactic structures or lexical choices. The function of the TT may be the same as the one of the ST or it may be different {Funktionskonstanz vs. Funktionsveränderung, Hönig and Kussmaul 1982). Based on this general description of translation as mediated cross-cultural communication, it will be illustrated that the various factors that influence the TT (addressees, situation, function of TT in TL community, text type) are not of equal relevance, and that it is above all the functions of the ST and the TT in their respective cultures that determine the translation strategies. Translation strategy is defined by Lörscher (1991) as "a potentially conscious procedure for the solution of a problem which an individual is faced with when translating a text segment from one language into another" (cited in Chesterman 1993: 13). In discussing examples from the sample texts we will not only ask "what has been done?" but also try to answer the question "why has a specific strategy been applied?". Each of the various political (or politically relevant) text types has its own contextual, text-typological, pragmatic, etc. conventions and calls for different translation strategies. In this chapter, some observations of translation problems and solutions are presented. The focus is on English and German (and partly French) texts and their authentic translations. The discussion is mainly descriptive, i.e. individual translation products are studied in order to describe and

CHRISTINA SCHÄFFNER

121

explain the specific problems encountered and the strategies employed for their (more or less appropriate) solution. The main aim is to develop an awareness for some phenomena typical for political texts. It is not intended, however, to say what 'should' be done, let alone to formulate translation rules (see Toury's 1992 arguments against a "normative conditioning in translator training"). The sample texts chosen for illustration can all be characterised as political, but they fulfil different political functions, from laying down rules or recommendations for international political behaviour to individual reflections about political events. They also display different degrees of culture-boundedness. On the one hand, there are texts that reflect in a specific way the social context and the historic period of their production (e.g. policy statements delivered on particular occasions). On the other hand, there are multilingual but equally authoritative texts (e.g. documents of the Conference on Security and Cooperation in Europe) which reflect specific production conditions, and which may provide evidence of (overt or hidden) internationalisation processes.

Diplomatic discourse in multinational institutions Quite a lot of translation is being done in the diplomatic sphere. For example, bilateral and multilateral treaties as instances of legal texts are agreed and translated. The research on legal texts, for example by Thiel and Thome (1987) and Rothkegel (1984), has shown that contracts and treaties as text types display special conventions. Especially in the preamble, there are standardised phrases in each language translators are expected to be aware of. Such phrases, which fulfil identical functions within the specific texts, are for example: English

French

Spanish

German

conscious

conscients

conscientes

desiring; desirous

désirant; désireux

con el deseo; deseosos

in dem Bewußtsein; eingedenk in dem Wunsch; von dem Wunsch geleitet

There are also syntactic and lexical conventions for expressing speech acts in legal documents (see Trosborg 1994). Another formal feature characteristic of contracts and treaties is that one sentence in the ST must correspond to one sentence in the TT. This is necessary to ensure mutual understanding: in oral or written negotiations that are based on a treaty the negotiators must be able to refer, for example, to "Article 1, paragraph 2, sentence 3" without causing any misunderstandings for their counterparts.2

122

STRATEGIES OF TRANSLATING POLITICAL TEXTS

"Violations" of syntactic rules are tolerated, condoned, due to the more important function of referring practices during negotiation. This is clear evidence of the fact that linguistic aspects of translation are a function of overarching functional and socio-cultural strategies of cross-cultural communication. Examples for "violations" of syntactic rules can be found in the Helsinki Final Act, although in example (1) the German text reads "clumsily", with three verbs put one after the other at the end of the sentence (italics, bold print, and underlining added for easier identification): (1) Declare that they are resolved to respect and carry out, in their relations with one another, inter alia, the following provisions which are in conformity with the Declaration on Principles Guiding Relations between Participating States. Werden in ihren Beziehungen zueinander unter anderem die folgenden Bestimmungen, die in Übereinstimmung mit der Erklärung über die Prinzipien, die die Beziehungen der Teilnehmerstaaten leiten, stehen, achten und anwenden, The results is an extremely impaired readability and comprehensibility. Example (2) shows that in the case of a German prefix verb, anerkennen, the prefix and the verb stem have not been split, resulting in a structure that is not typical for the majority of text types (where we would find erkennen ... an), but commonly used in legal text types (see Thiel and Thome 1987). (2) The participating States recognize the universal significance of human rights and fundamental freedoms, respect for which ... Die Teilnehmerstaaten anerkennen die universelle Bedeutung der Menschenrechte und Grundfreiheiten, deren Achtung ... The CSCE discourse is of interest from the point of view of translation mainly in another respect. Major representative examples of this discourse are the Final Act of the Conference on Security and Cooperation in Europe (CSCE) signed in Helsinki in 1975 by the high representatives of all, then 35, European countries (but with the exception of Albania) as well as of the United States of America and of Canada, and the documents of follow-up conferences (Belgrade 1978, Madrid 1983, Vienna 1989, Paris 1990, Helsinki 1992). These texts have an obvious political role to play in that they outline regulations for political behaviour - even if these regulations were not binding for the governments (for a more detailed analysis of textlinguistic aspects of the Final Act see Schäffner 1986, and for political interpretations based on the German version see Schäffner in print).

CHRISTINA SCHÄFFNER

123

All these documents exist in six official languages: English, French, German, Italian, Russian, and Spanish. All six texts are considered to be equally authentic. Although the Helsinki Final Act was drawn up by negotiation, the six texts are nevertheless the product of translations. That is, the original version was the English text. It was translated into Russian, French, and German. In the process of drafting the Final Act, however, English, French, and Russian mutually influenced each other, meaning also, that the translatability affected the creation of the text. The French text was mainly the basis for the translations into Italian and Spanish. The German text of the Final Act was the only text which had been agreed upon by "Western", "Eastern" and neutral states. That is to say, the five national teams of the then Federal Republic of Germany and German Democratic Republic, of Austria, Switzerland and Liechtenstein worked together and consulted each other for the translation of the Final Act into German. The Final Act was revised whenever language problems occurred. It was linguistically and politically negotiated. The result was a kind of hybrid text[type], reflecting some fairly specific syntactic and lexical features. Syntactic aspects concern mainly the extremely close similarity of the surface structures of all the six versions of the CSCE documents, and for which examples (1) and (2) above are evidence. Another interesting point is the specific CSCE terminology. One procedure may actually be called "creation of specific terms in the CSCE discourse". This can be seen in example (3), referring to the phrase Refraining from the threat or use of force in the Final Act (italics in the original): (3) II. Refraining from the threat or use of force The participating States will refrain in their mutual relations, as well as in their international relations in general, from the threat or use of force against the territorial integrity or political independence of any State,... Reaffirming that they will respect and give effect to refraining from the threat or use of force ... II Non-recours à la menace ou à l'emploi de la force Les Etats participants s'abstiennent dans leurs relations mutuelles, ainsi que dans leurs relations internationales en général, de recourir à la menace ou à l'emploi de la force, soit contre l'intégrité territoriale ou l'indépendance politique d'un Etat,... Réaffirmant qu'ils respecteront et rendront effectif le non-recours à la menace ou à l'emploi de la force ... II Enthaltung von der Androhung oder Anwendung von Gewalt

124

STRATEGIES OF TRANSLATING POLITICAL TEXTS

Die Teilnehmerstaaten werden sich in ihren gegenseitigen Beziehungen sowie in ihren internationalen Beziehungen im allgemeinen der Androhung oder Anwendung von Gewalt, die gegen die territoriale Integrität oder politische Unabhängigkeit irgendeines Staates gerichtet.... ist, enthalten. Unter Bekräftigung, dass sie die Enthaltung von der Androhung oder Anwendung von Gewalt achten werden und ihr Wirkung verleihen werden,... There is continuity in the English and the French texts as regards the keyphrase of international law refraining from the threat or use of force and nonrecours à la menace ou à l'emploi de la force, respectively. The fairly clumsy Enthaltung von der Androhung oder Anwendung von Gewalt is found only within the CSCE discourse, although not exclusively. The more common [Gewalt]Verzicht, which is typical for political discourse outside the CSCE framework is used in the Madrid document. The reason for this phrase is that the translation into an "agreed" German resulted in a specific "CSCE-German", i.e. a certain jargon, known to the CSCE negotiators and CSCE staff, including translators and interpreters. Dollerup (1993) reports similar observations with reference to EC documents. He speaks of the "language of the Eurocrats" which "is not only a sociolect, but also a language for special purposes, namely Community negotiations". The Helsinki Final Act set the example for the wording of the aims of the CSCE process, and due to the functional and thematic continuity specific formulations were taken up again, often in exactly the same wording. We thus find quite a number of identical formulations in the Final Act and the documents of follow-up conferences. This was taken into account for the translations as well, as illustrated in example (4). The TT production in this special case can be characterised as induced both by the ST and by a previous TT which functions as a TL reference point:3 (4) Convinced of the need to exert efforts to make détente both a continuing and an increasingly viable and comprehensive process, universal in scope, ... (Helsinki 1975) The participating States express their determination - to exert new efforts to make detente an effective, as well as continuing increasingly viable and comprehensive process, universal in scope, as undertaken under the Final Act; (Madrid 1983) Überzeugt von der Notwendigkeit, Anstrengungen zu unternehmen, um die Entspannung im universellen Sinne sowohl zu einem dauerhaften als auch zu einem immer lebensfähigeren und umfassenderen Prozeß zu machen,... (Helsinki 1975)

CHRISTINA SCHÄFFNER

125

Die Teilnehmerstaaten bringen ihre Entschlossenheit zum Ausdruck, - neue Anstrengungen zu unternehmen um die Entspannung im universellen Sinne zu einem wirksamen wie auch dauerhaften, immer lebensfähigeren und umfassenderen Prozeß zu machen, gemäß den in der Schlußakte eingegangenen Verpflichtungen; (Madrid 1983) The continuity in the texts of the CSCE documents is to a large extent determined by intertextuality relations to the UN Charter. The authentic texts of the UN Charter are Chinese, English, French, Russian, Spanish, but not German. That means, that in drafting the Helsinki Final Act, reference could be made to the wording of the UN Charter. Example (5) shows the authentic English and French texts of the UN Charter referring to the principle of refraining from the threat or use of force. The formulations in the Final Act, given in example (3) above, reflect the close link to the wording of the UN Charter: (5) All Members shall refrain in their international relations from the threat or use of force against the territorial integrity or political independence of any state, or in any other manner inconsistent with the Purposes of the United Nations. (UN Charter, ch. I, Art. 2,4) Les Membres de l'Organisation s'abstiennent, dans leurs relations internationales, de recourir à la menace ou à l'emploi de la force, soit contre l'intégrité territoriale ou l'indépendance politique de tout Etat, soit de toute autre manière incompatible avec les Buts des Nations Unies. For establishing a CSCE discourse in German, there was no universally valid German copy of the UN Charter, although several translations existed (with slight variations, however). Alle Mitglieder unterlassen in ihren internationalen Beziehungen jede gegen die territoriale Unversehrtheit oder die politische Unabhängigkeit eines Staates gerichtete oder sonst mit den Zielen der Vereinten Nationen unvereinbare Androhung oder Anwendung von Gewalt. (Beck'sche Textausgaben, München 1979) Alle Mitglieder [der Organisation] enthalten sich in ihren internationalen Beziehungen der Gewaltandrohung oder Gewaltanwendung, die gegen die territoriale Unverletzlichkeit oder politische Unabhängigkeit irgendeines Staates gerichtet oder in irgendeiner anderen Weise mit den Zielen der Vereinten Nationen unvereinbar ist. (Staatsverlag, [Ost]Berlin 1974 [1961]) There is more variation between the German texts of the UN Charter and the Final Act than is in the English and French texts. The fact that there was an East German and a West German version of the UN Charter did - to the best of my knowledge - not result in heated political

126

STRATEGIES OF TRANSLATING POLITICAL TEXTS

arguments based on different, politically motivated, interpretations of the text. A case in point, however, where political interests did in fact result in two different German translations was the Quadripartite Agreement on West Berlin signed in 1971.4 The authentic texts were in English, French and Russian. The paragraph dealing with the relations between West Berlin and the Federal Republic of Germany says that the ties will be developed. Ties (French liens) had been translated as Verbindungen in the East German and as Bindungen in the West German version, with Verbindungen denoting relations that are not so tight as those denoted by Bindungen (see Kade 1980: 57ff.). To conclude the discussion of multilingual and equally authentic texts: One could of course ask whether this is in effect translation. Do these texts fit at all into a framework that defines translation as ST induced TT production? With the CSCE discourse we have a specific case of mediated communication: There is no primary communicative situation in which a ST has fulfilled its communicative function for its SL addressees and where subsequently a TT is to fulfil or has fulfilled its function for TL addressees in a secondary communicative situation. In fact, there is no proper ST displaying characteristic features or specific conventions of the SL text type, and there are also no proper parallel texts in the TL culture to serve as a model for the TT. All six versions were meant to fulfil an identical function, their main addressees disposing of (relatively) identical background knowledge, pursuing fairly identical political aims, and fulfilling practically the same political roles in their respective societies. There will not be universal agreement that this is in fact translation and that these texts are functionally equivalent. But such diplomatic multilingual texts definitely do have some bearing and call for a specific treatment within translation studies. Due to increasing internationalisation processes, more texts of this kind will pose challenges for translators, for example texts of the European Community, European Parliament, United Nations. Translation and interpreting services in the various institutions of the European Union (EU), for example, have expanded massively, due to the EU language policy with the right for everybody to use their own national language (on language and translation policy in the EU see, for example, Arthern 1994 and Volz 1993). In the case of such international bodies we would have to differentiate between texts for "internal" use (i.e., addressed to the politicians and negotiators, for example legal documents, constitutional texts) and texts for "external" use (i.e., addressed to the broad public, for example manifestos of the party groupings in the European Parliament). The texts that are produced (through intercultural negotiation as well as through translations) may display features which are a reflection of specific conventions in the two (or more) cultures. The resulting

CHRISTINA SCHÄFFNER

127

text might be a "hybrid", incorporating different norms from the cultures involved (see Schütte's 1993 illustration of how legislative "Eurotexts" are interactively created during multilingual negotiations and the effects on their linguistic structure, often labelled "Eurospeak"). Are such "hybrids" new supra-cultural (or real intercultural) text types? Are specific textual features (vocabulary, syntax, style, etc.) compared to language-specific features a loss or a gain?5 Whatever the answer, it will definitely influence the conditions of text production and reception, as well as aspects of translation training. For translation studies this means that the status of both the ST and the SL community would have to be reconsidered. Texts produced within multi- or supranational institutions, regardless whether they are intended for internal or external use, are clear evidence that nations, cultures and languages do not totally coincide (see also Lambert 1994).

Speeches and statements by politicians In contrast to such texts that are interactively negotiated in a supranational setting with the overall aim of achieving and reflecting consensus, there is another group of text (types) that do indeed reflect culture-specific conditions of their production. The sample texts chosen are speeches or statements by politicians, mainly speeches by German and American politicians delivered within the context of German unification (1989/90) (see appendix). These texts deal with political events in one particular country, Germany, although the events definitely were of relevance for international politics as well. They are examples of political documents that have fulfilled a specific purpose in a particular communicative situation in a SL community. By virtue of their political function, however, they are not exclusively addressed to an audience in this source community but they have a wider political audience in mind. Their translations, therefore, inform a target audience about a communicative act that had already been fulfilled in the source community, with the content of the texts being of equal relevance for a target audience (or an audience in various target communities). Within this group of texts we can differentiate two sub-groups due to their function in the political communication. The speeches by Bush and Baker were delivered to a political body, a small group of other politicians. These texts are examples for internal political communication (politicians speaking to politicians), and they fulfil a function in policy-making. The speeches by Kohl and Weizsäcker, on the other hand, were first of all addressed to the German people. They are examples for external political communication (politicians

128

STRATEGIES OF TRANSLATING POLITICAL TEXTS

speaking to the public on a special occasion, with these speeches having been subsequently published in mass media), and their function is to explain and justify political decisions. The STs were produced in a very specific SL context, addressed to an SL audience, and according to their main purpose they are persuasive rather than informative. The TTs, however, do not attempt to persuade the audience in the TL culture but to inform them about events in the SL culture (although cultures are not homogeneous, "culture" in the contexts to be discussed can be said to coincide with "nation"). These changes in the text function determine the translation strategies.6 A few observations will be presented, based on an evaluative comparison of the source language texts and their translations. There are of course always translation problems that are caused by differences in the language systems, but since they are not specific for political texts they will not be discussed here. The translation strategies pursued will be discussed on the basis of a textlinguistic-discursive and functional approach to translation (based on Neubert 1985, Snell-Hornby 1988, Nord 1988), i.e. an approach that treats the text in its situation and with a specific function. The illustrations are product-oriented. A widely used translation strategy, namely omitting or deleting information, is found in speeches by politicians as well, but not very often. This again can be explained by reference to the function of the texts in internal or external political communication. Here sentences or passages were deleted which point to the specific situation of the original text production or which are of relevance for the textual meaning only at a microstructural level, as in example (6), at the end of President Bush's remarks before signing a letter transmitting the Treaty on German Unification to the U.S. Senate for its advice and consent. In example (7), a standard phrase frequently used by Chancellor Kohl to end. a speech has been omitted: (6) And now for the signing. (Bush) (7) Gott segne unser deutsches Vaterland! (Kohl) Another phenomenon concerns specific naming conventions, which often are related to diplomatic and/or legal discourse. The more explicit version in the TT in example (8) is due to such differences in the conventionally preferred terminology (italics in all of the following examples are mine): (8) President Bush announced ... a major new Helsinki initiative ... (Baker) hat Präsident Bush ... eine neue Initiative im Rahmen des KSZE-Prozesses angekündigt,...

CHRISTINA SCHÄFFNER

129

Since the STs were meant for an audience of the SL community, some of the information could be left implicit by the speakers because they could assume mutual knowledge (Clark and Marshall 1981). Since the TT audience cannot automatically be assumed to share specific background knowledge about the SL culture, the translator, as the knowledgeable mediator, would have to decide if lacking background knowledge would have to be accounted for, and if yes, how this should most appropriately be done. Mutual knowledge due to community co-membership (Clark and Marshall 1981: 37) is in most cases activated by lexical items, but it may also be activated by syntactic structures, for example definite articles. In example (9), Weizsäcker refers to ethnic Germans who, after the end of the second World War in 1945, had to leave former German territories in Poland and Czechoslovakia (which had mostly been occupied by Germany during the war). In order to come up with this interpretation, the TT reader would need quite a lot of knowledge about Germany and German history: (9) Unsäglich hart war der erzwungene Heimatverlust (Weizsäcker) Those who were forced to leave their homeland suffered immeasurably. Similar in example (10): Kohl thanks Eastern European politicians who initiated reform processes in their countries. His use of the definite article die Flüchtlinge is motivated by the fact that a frame GDR people protesting against their political system had been activated before during text comprehension, die Flüchtlinge referring to GDR citizens who had fled to Hungary (some of them staying there in the embassy of the Federal Republic of Germany, FRG) in order to go (via Austria) to the Federal Republic as their ultimate aim. (10) Dank schulden wir auch den Reformbewegungen in Mittel-, Ost- und Südosteuropa. Vor gut einem Jahr ließ Ungarn die Flüchtlinge ausreisen. (Kohl) Our thanks are also due to the reformist movements in Central, Eastern and South Eastern Europe. Just over a year ago, Hungary permitted the refugees to leave. Leaving implicit information in the ST also implicit or possibly vague in the TT may mean risking miscomprehension of the TT or preventing comprehension altogether. However, differences in the encyclopaedic background knowledge of the respective addressees have hardly been a problem in these speeches by politicians. For example, in the sample texts analysed there are no cases where information had been added to provide background knowledge about the source culture for the TT addressees (see Schäffner 1992a). This seems to be due, first of all, to the function of these texts in the internal politi-

130

STRATEGIES OF TRANSLATING POLITICAL TEXTS

cal communication. The speeches by Kohl and Weizsäcker too, although examples of external political communication in the source culture, changed into examples of internal communication after they had been translated: The TTs were produced by translation offices of the German government and distributed in limited editions to the governments and official institutions in the target culture(s). Another reason for the extremely rare cases of omissions and additions is the topic of the texts: German unification was widely reported about in 1989/90, thus providing the necessary background information. And a final point: such speeches (for instance, televised addresses to the nation) seem to be more explicit than speeches addressed to a smaller or more specific audience, but more evidence for this would have to be provided. Closely related to the problem of (the lack of) sufficient background knowledge for text comprehension is the aspect of political concepts and their (ideologically determined) meanings. A concept is part of a coherently organised conceptual structure and usually linked to a linguistic label. A concept represents a particular amount of stored knowledge within such a structure. In cognitive linguistics, these conceptual structures are referred to as frames, scripts, schemata (see, for example, Fillmore 1976, Lakoff 1986). They are shared by a group of people or by a whole speech community, depending on their social experience, including experience with texts. Concepts in politics, as well as in the humanities in general, can be seen as representing a "whole theory" (Wallerstein 1991). Concepts have histories, they not only evolve historically but they cannot be understood without linking them to the total historical process. Political debates have often been about (the correct interpretation of) concepts, or political keywords. In the words of Robertson (1988: viii), "Politics as an art, and political studies as a science are overwhelmingly about words, shades of meaning, ideological linkages often neither grammatically nor logically determined." Understanding the concepts and knowing (about) the frames in which they are embedded is of high relevance for translation as well, because, as Wallerstein (1991: 15) says: "If an individual reader misreads, he suffers the consequences individually. If a translator misreads, he leads innumerable others astray, all of whom pay the consequences as well."7 Examples (9) and (10) above included, albeit not very obviously, concepts of a frame that could be labelled German nationality. More widely known examples are such political key concepts as democracy, or human rights, which during the Cold War had different, ideologically motivated meanings in western and eastern Europe (and there still are different interpretations in Europe and the United States on the one hand, compared to Asian countries on the other). Sometimes, political keywords that are commonly used within one

CHRISTINA SCHÄFFNER

131

country or a group of ideologically united countries, are not taken up in international documents because another country or group of countries objects to them. For example, the Warsaw Pact countries did not succeed in having the term peaceful co-existence included in the Helsinki Final Act, and due to British resentments, the word federalism was not used in the Maastricht Treaty to denote the aim of European integration. But not only the avoidance of certain terms is politically relevant, but also, and even more so, the choice of particular words and phrases. Such words (as well as the political texts themselves) are rooted in a particular ideology, and the translator has to be aware of those "cues and symbolic conventions that create frames of interpretation" (Neubert 1985: 120). With reference to political texts of the former Soviet Union, Markstein (1994: 105) even speaks of a "propagandistic linguistic nomenclature": words whose meanings have been ideologically determined and which are a "code for insighters".8 For example, it was an established practice in political texts of the former GDR to refer to the border between the GDR and the FRG as Trennlinie (dividing line), to the other Warsaw Pact countries as sozialistische Bruderländer {fraternal socialist states), and to the western democracies as nichtsozialistisches Wirtschaftsgebiet (non-socialist economic area). Subtle changes in more or less established naming conventions may also be of significant political relevance. Example (11), a quotation from Weizsäcker's speech, is an illustration of this: (11) Ich möchte der verantwortlichen politischen Führung in beiden bisherigen deutschen Staaten, den gesetzgebenden Körperschaften und ... danken. (Weizsäcker) I wish to thank the political leaders on both sides, their parliaments, and ... It had not been typical of the West German government before unification to speak of two German states, due to the official policy of regarding the division of Germany as unnatural. One of the alternative wordings was zwei Staaten in Deutschland (two states in Germany). In the GDR political discourse, however, it was highly common to speak of two German states, thus putting emphasis on the sovereignty of the GDR. Weizsäcker's formulation, at first glance, seems to be in contradiction to the established norms. However, looking at this sentence from a macro-propositional perspective, we see that it is embedded in the context of President Weizsäcker thanking individual people, political organisations, and states for their part in making German unification possible. With politische Führung he is actually referring to the CDU-led coalition government in the FRG and to the GDR government under Prime Minister de Maizière, also a CDU-led coalition government that was formed after the first (and

132

STRATEGIES OF TRANSLATING POLITICAL TEXTS

last) free democratic elections in the GDR in March 1990. All these subtle changes are not reflected in the English translation. Examples (12) and (13) show different readings of the ST and the TT with respect to the concept of Europe, which at least have the potential for political arguments: (12) Die Sowjetunion ... Sie will die alte Distanz zu Europa überwinden. (Weizsäcker) ... The Soviet Union wishes to close the traditional gap between itself and the rest of Europe. (13) Wenn wir Deutschen solche Signale in Richtung auf das ganze Europa setzen, dann... (Weizsäcker) When we Germans send such messages to the whole of Europe we do so as ... Example (12) concerns the political discussions on the concept of Europe and on the Soviet Union being part of Europe or not. In (13), the German text makes use of the movement schema, with das ganze Europe functioning as the destination, the end point on the path; in the English text, however, the whole of Europe is the recipient in a sending schema.

Politically relevant texts by non-politicians There is a third group of texts for the translation of which culture-specific background knowledge is highly relevant. These are speeches or essays and articles by writers and intellectuals. At a first glance, these texts may not seem to belong to the category of political texts at all. They probably are not prototypical exemplars of political texts, but they are nevertheless politically relevant. The sample texts here are three essays on the political developments in Eastern European countries in 1989/1990, written by Günter Grass, Christa Wolf, and Timothy Garton Ash, i.e., intellectuals who had frequently given (more or less) critical comments on these events. The texts by Christa Wolf and Günter Grass were actually speeches that had been delivered on a particular occasion, but subsequently published for wider circulation (see appendix). Günter Grass, a well-known German writer, spoke to the Parliamentary representatives of the Green Party and the Bündnis 90 Alliance on the eve of German unification (October 2, 1990). This was a highly topical occasion, and Grass' intention was to express a warning. The text by Christa Wolf is a speech she gave at a mass demonstration for democracy and freedom of speech in East Berlin in November 1989 (transmitted live by GDR

CHRISTINA SCHÄFFNER

133

radio and TV). Although Grass' speech was originally an example of internal political communication (addressed to members of parliament), both speeches were ultimately of relevance for the nation as a whole, which was the motivation for their subsequent publication in mass media. The sample text by Timothy Garton Ash, an historian and publicist who repeatedly reported on Eastern Europe - before and after the peaceful revolutions there - , is a political essay, published in January 1990 in The New York Times Review of Books. It is a personal account of the political events in Czechoslovakia in November 1989. That all these three texts were considered to be of political relevance for an audience outside their countries of origin is reflected in the fact that they were translated and published in mass media of the target culture: the text by Grass in the Guardian Weekly, the one by Wolf in New Statesman & Society, and the one by Garton Ash at first in Lettre Internationale and later in a volume of essays by him. These texts are highly culture-bound. However, the essay by Garton Ash is slightly exceptional in this respect: The source culture of text production and reception is not identical with the culture reported about. Author and addressees belong to the same culture in a wider sense (i.e. Anglo-American). The text was written by an outsider looking at another culture, namely Czechoslovakia (for a comparison of the German and French translations of this text see Schäffner and Herting 1994). Highly culture-bound texts, i.e. texts with references to a wide range of cultural patterns of the society in question, including aspects of its economic, political and legal life, require a lot of background knowledge for a coherent interpretation. Since the SL author usually shares knowledge about socio-political and socio-cultural aspects of life in the SL community with his or her ST addressees, the texts are often characterised by a high degree of implicit information and present, thus, particular problems for translators. Only a few examples will be discussed which show very drastically how a message may be totally distorted if the implicit culture-specific information or a culture-specific word meaning or an allusion is not grasped by the translator. Culture-specific aspects become most obvious in mistranslations. It is mostly a concept, a word or a phrase, which calls for the activation of (culture-specific) background knowledge needed for a coherent text interpretation. If misread, the political perspective of the text or the author's attitude may not come across to the TT audience. Of special interest in this respect are once more politically relevant keywords (or: concepts) whose meaning is determined by the culture in which they are used. One such phenomenon is information which is related to the mechanisms of Communist Parties and to the theory of Marxism-Leninism.

134

STRATEGIES OF TRANSLATING POLITICAL TEXTS

The translators' strategies in these cases are evidence of their consideration of their readers as well as of their own (mis)understandings of the ST information. Examples (14) to (17) are cases in point, albeit cases for obvious misunderstandings or misinterpretations. The context of example (14) is Garton Ash's report about meetings of the Czech Civic Forum to discuss the formulation of a communiqué stating the aim of a planned general strike. (14) The general strike is here described as "an informal, nationwide referendum on whether or not they should go on humiliating us, and whether this country should continue to be ruined by the leaders of one political party, permanently abrogating to itself the leading role." (Garton Ash) (a) Nun wird der Generalstreik so deklariert: "... ein informelles, nationenweites Referendumdarüber,ob sie uns einfach immer weiter erniedrigen dürfen und ob das Land weiterhin von den Führern einer politischen Partei ruiniert werden darf, die permanent selbst auf die Ablösung ihres Führungsanspruchs hinarbeitet."(LettreInternational, italics in the original, underlining is mine) (b) ... ob sie uns einfach immer weiter beleidigen können und ob das Land weiterhin von den Führern einer politischen Partei ruiniert werden sollte, die permanent selbst ihren Führungsanspruch ad absurdum führen". (book) A reader of version (a) who had had an education in the former GDR, and thus familiar with the ideology of Marxism-Leninism, will definitely be struck by the information given: that the Communist Party itself was orienting its work towards abolishing its leading role. This information is totally contradictory to the teachings of Marxism-Leninism, namely that the Communist Party has the leading role in the society and that it will strengthen this leading role in order to make better and better the life for everybody in a socialist country. Moreover, the dictionary meaning of abrogate (abolish, do away with, put an end to) is in contrast to the facts referred to in the message. The information centres around the leading role of the Communist Party (which had been mentioned before in the text - so there is internal coherence). The propositions of example (14), specifying the aim of the general strike, refer to negative effects due to the policy of the leading Communist Party. This is reflected in the lexical items (humiliating, ruined). So the last proposition should also reflect a negative fact, namely that the Communist Party assumed a power to which it had not been entitled by free elections. In the book, version (b), the translator had obviously noticed her mistake. One possible explanation for the incoherent initial version is lacking background knowledge of the translator. There is, however, another problem, a problem that actually lies in the source text, in Garton Ash's (mis)use of the word abrogate. Abrogate (with the obsolescent sense 'to do

CHRISTINA SCHÄFFNER

135

away with, put an end to') and arrogate (i.e., arrogate to oneself as to claim or assume that to which one is not entitled, or to claim or assume unreasonably or arrogantly, as explained in Partridge 1980) are often mixed up. How to deal with more or less obvious mistakes in a source text? This is a question which Newmark (1991) wants to solve by differentiating between authoritative and informative texts. In this case, however, it can be suspected that the translator's main strategy of staying as close as possible to the source text came into conflict with her insufficient background knowledge - a conflict she was unable to solve in her initial translation. From a functional perspective, the question whether or not the translator is allowed to correct errors in the ST is but a pseudo-problem. The TT is produced in order to fulfil a communicative function for TT addressees. These functional aspects outweigh any loyalty to the ST - which Newmark acknowledges himself when he says "...a translation has to be understood even if the original isn't, or if the original can only be guessed" (Newmark 1994: 214). Example (15) reflects a particular meaning and connotation of the word Staatsvolk. (15) Unglaubliche Wandlungen, das Staatsvolk der DDR geht auf die Straße, um sich als Volk zu erkennen. (Wolf) Incredible changes. The GDR's head of state goes on to the streets in order to recognise himself as one of the people. Staatsvolk is not a specific word out of Communist terminology. Christa Wolf contrasts it to the word Volk (people), which had been used in the slogan Wir sind das Volk (We are the people) with which in the autumn of 1989 the GDRpeople had openly demonstrated that they were not subject to the wishes of the ruling party, but that they themselves wanted to decide on what their life should look like. With the people becoming the head of state in the target text, however, the whole idea is distorted, and also the head of state (Honecker at that time) is evaluated positively, which was not at all the case (for a more detailed analysis of this text see Schäffner 1992b). In example (16) we have an allusion to Chancellor Kohl who is said to have a habit of not reacting immediately to certain problems but waiting instead until they have either been solved by themselves or have disappeared. In the case of German unification, however, he quickly took the initiative for fear of coming too late, of "missing the train", and of being unable to influence the way of unification. In the translation, however, there is no reference to Kohl, but a generalisation instead. (16) Als der Weisheit letzter Schluß wurde eine Bahnhofsdurchsage wiedergekäut: "Der Zug ist abgefahren!" Und jemand, der sonst Probleme auszusitzen

136

STRATEGIES OF TRANSLATING POLITICAL TEXTS

pflegt, glaubte, den Mantel der Geschichte rauschen zu hören, sprang auf und griff zu. (Grass) Wisdom's last message came in the form of a repeated station announcement: "The train has departed!" And people otherwise used to riding out problems thought they heard the swish of history's cloak, so they jumped up and grapped at it. (16) and also (17) below are good examples to illustrate that a functionally-oriented textlinguistic-discursive analysis can actually help the translator to achieve a coherent text comprehension. The indefinite pronoun jemand is a signal which calls for cognitive reasoning and inferencing. Missing information has to be searched for, and if it is not found in the text itself recourse to stored knowledge, i.e. to frames or scripts, is necessary (coherence being achieved by the interaction of text-presented knowledge and stored world knowledge, see de Beaugrande and Dressier 1981: 6). The ST reader would then come up with a proposition which is something like The 'somebody' refers to Chancellor Kohl because the knowledge that he sometimes hestitates to take decisions is part of a Kohl frame. Looking at the ST from a micro-structural perspective alone, jemand would allow for a more generalised reading, due to its meaning potential and the syntactic structure. This potential other version has been made obvious in the TT, with the consequence, however, that the English readers of the TT may interpret this passage as telling them something about typical reactions and attitudes of all West Germans. In example (17) Kohl and Genscher were mixed up. (17) Als sich vor einigen Monaten die Außenminister der Sowjetunion und der Bundesrepublik Deutschland in der einst polnischen, seit 1939 weißrussischen Stadt Brest trafen, war das ein Treffen neben vielen. Dieser Ort kümmerte die Welt wenig. Nur die Polen reagierten verschreckt, weil Schewardnadse und Genscher taktlos genug waren, einander dort zu begegnen, wo im Herbst 1939 die Wehrmacht und die Rote Armee ihre Waffenbrüderschaft gefeiert und ihre Siegesparade abgehalten hatten. EineSchande, zumal nicht, wie gewohnt, der Kanzler Kohl Polen gegenüber den Rüpel herauskehrte, sondern ein Politiker Rücksichtslosigkeit bewies, der allgemein als schlau, vorsichtig, ja, behutsam gilt. (Grass) When some months ago the Foreign Ministers of the Soviet Union and of West Germany met in the formerly Polish but, since 1945, Byelorussian city of Brest, that was one of many meetings. The world was not greatly concerned about the location. Only the Poles reacted with fear, because Shevardnadze and Genscher were tactless enough to meet in the place where, in the autumn of 1939, the Wehrmacht and the Red Army had celebrated their comradeship in arms and had held their victory parade. It is particularly shameful

CHRISTINA SCHÄFFNER

137

that Chancellor Kohl did not, as he usually does, play the lout towards the Poles, but instead demonstrated the thoughtlessness of a politician who is generally reckoned to be shrewd, cautious, indeed wary.

It is not Kohl who was generally reckoned to be a cautious politician but Genscher, the former Minister of Foreign Affairs. Kohl had been known for his often rude words and behaviour in relation to Poland. Grass finds fault with Genscher who, although usually a cautious politician, demonstrated thoughtlessness towards Polish sentiment in one particular instance. In the German text, the propositions in the last sentence are connected by a semantic relation of opposition (nicht Kohl, sondern ein Politiker - not Kohl, but a politician who - see Mann and Thompson 1986), and there is coherence to the previous information, exemplified particularly clearly by the semantic relation between taktlos and Rücksichtslosigkeit (tactless, thoughtlessness). These two words are textually synonymous, which implies that they denote an identical referent. Therefore we have co-reference between Genscher and ein Politiker. There is a relatively high number of mistranslations in the English version of the Grass text, most of which were obviously caused by the translator's lack of background knowledge about the SL culture (for a more detailed analysis see Schäffner 1993). The TT readers get different, often totally wrong, information. This is particularly worrying, since as a motive for publishing this text it is said in the Guardian Weekly: ... For Grass, the two Germanies' rush to unification is not a time for jubilation. Here he explains why. Germany, Britain, and Europe as a whole cannot afford to ignore his impassioned argument.

Conclusion One could argue that some of the mistranslations illustrated were actually caused by insufficient linguistic competence or by an insufficient analysis of the ST. This might have been the case as well. But the discussion should also have made it clear that linguistic knowledge cannot be seen as an autonomous system, largely independent of socio-cultural knowledge. The lexical items or syntactic structures that either were misinterpreted or whose meanings and/or connotations were more or less distorted in the TT are evidence that concepts have meanings only by virtue of being embedded in socio-culturally determined frames (which are more or less culture-specific). Or in the words of Agar "the separation between 'language' and 'culture' makes no sense ... it makes more sense to speak of "languaculture" (Agar 1992: 13).

138

STRATEGIES OF TRANSLATING POLITICAL TEXTS

Political texts come in the form of various text types, only a few of which have been taken into consideration in this chapter. Almost every text (type) has the potential to become politically relevant. Kohlmayer (1994), for example, discusses ideologically determined translations of Oscar Wilde's plays for the German stage during the Nazi period. All these different text types are characterised by different features and, thus, pose different problems for translation. And, of course, the products of translation have different depths of impact in the TL communities, depending most of all on the respective text type and its function in the internal or external political communication. Text types in the field of politics only rarely display highly conventionalised features both at macro- and micro-levels. Those that do, for example treaties, require of the translator, as the expert in interlingual and intercultural communication, to take such text-typological conventions into account and produce a TT that conforms to the TL conventions. In a large number of cases, however, political texts are primarily meant to function within their source culture, and when they are translated, their function in and for the target culture will change, and a need for conforming to TL text-typological conventions does not arise. Thus, the most important factor determining the particular textual makeup of the TT is its function in the TL community - taking into account the necessity to reflect anew about the status of SL culture and TL culture as well as of ST and TT in view of internationalisation processes. Further analysis of texts such as international or multilateral treaties, governmental documents, writings or speeches by politicians, writers, or other "representative voices", will have to test the translation strategies discussed, and will, highly probably, reveal some more translation problems. Notes 1 Text type is understood with Suter (1993: 48): "A traditional text type is what a given speech community, at a given time and over a considerable period of time, accepts as a tra­ ditional, conventional and in some specific way linguistically standardised textual model to be constantly re-used for specific communicative purposes. " 2 See the explanation in a German booklet laying down guidelines for standardised formula­ tions of treaties issued in 1962 by the language department of the Foreign Office: Jedem fremdsprachigen Satz muß auch ein deutscher Satz entsprechen; ... Diese Regel ist auch gegenüber etwaigen stilistischen Bedenken im Einzelfall unbedingt einzuhalten; denn bei mündlichen oder schriftlichen Verhandlungen, denen eine Übereinkunft zugrunde liegt, müssen die Unterhändler jederzeit z.B. auf "Artikel 1 Absatz 2 Satz 3" u.ä. Bezug nehmen können, ohne daß dies bei der Gegenpartei, die natürlich die Fassung in ihrer eigenen Spra­ che benutzt, zu Mißverständnissen führt. (Paragraph 16 - Der Satz als unantastbare Einheit. In Standardformulierungen für deutsche Vertragstexte).

CHRISTINA SCHÄFFNER

139

3 In 1987 the German Foreign Office published a book on the CSCE terminology in the six official CSCE languages. It says in the introduction: "Inaccuracy of quotation and reference is apt to cause confusion in public debate. In order, therefore, to eliminate this danger as far as possible, the following practical terminology booklet has been published for the benefit of all who are required to draft texts on subjects concerning the CSCE/CDE, or who wish to quote, reproduce or comment on passages from CSCE/CDE documents." 4 There was also a politically motivated difference as to the object of this agreement: whether it was an agreement on Berlin as a whole, or just on the western sectors of Berlin. Gowers speaks of a "threat" to the English language: "The English in the Community's publications, which has been derisively called Euro-English, is distorted under the influence of other European languages and displays a bureaucratic jargon of its own. The impositions of Euro-jargon should be firmly opposed and the Europeanisms resisted." (Gowers 1986: 186f, cited in Schütte 1993: 108) 6 In the former German Democratic Republic quite a number of political texts were translated, usually by the state-owned Intertext translation agency. Although texts such as reports to Communist Party Congresses had as a main function to persuade and mobilise ST addressees, this persuasive function was - more or less overtly - also ascribed to the TT, justified by the educational function of the Marxist-Leninist ideology. Translators, too, were regarded as mediators of politics, and the aim of propaganda literature sent abroad was to rouse interest and respect for the GDR's policy (see Ringleb 1983). The TT were usually very close to the surface structure of the ST. Understanding these texts required a familiarity with the style of Marxist-Leninist discourse and the specific terminology, as can be seen in an example from the report to the 1986 Congress of the Socialist Unity Party (SED) of the GDR: Auch künftig haben wir die sozialistischen Produktionsverhältnisse so auszugestalten, daß sie die dynamische Entwicklung der Produktivkräfte im Interesse der Stärkung des Sozialismus fördern. With our sights set on further strengthening socialism, we will have to continue to manage our socialist relations of production in a manner conducive to the dynamic development of the productive forces. 7 Wallerstein (1991: 6) links original text production and translation: "An author who seeks to communicate well normally has a target audience in mind. The author is likely to make some assumptions about the existing knowledge of the putative group of readers. When the author employs a concept, the reader is presumed to know the implicit theory of history (even if the reader does not share it). Or if the author does not think the reader knows what the concept implies, a sensible author will seek to explain it." 8 "[...] propagandistische Sprachnomenklatur: Wörter mit terminologischem Charakter [...], mit denen die Bezeichnung bestimmter Sachverhalte ideologisch festgelegt wird, gleichsam als Code für Eingeweihte." (Markstein 1994: 105) References Agar, M. 1992. "The intercultural frame". Paper delivered at the LAUD Symposium, Duisburg 1992.

140

STRATEGIES OF TRANSLATING POLITICAL TEXTS

Arthern, P. J. 1994. "European Community translation in Belgium". Meta xxxix(l): 150-158. de Beaugrande, R. and Dressler, W. 1981. Introduction to Text Linguistics. London: Longman. Bochmann, K. (ed.). 1986. Eigenschaften und linguistische Analyse politischer Texte (Linguistische Studien LS/ZISW/A 152). Berlin: Zentralinstitut für Sprachwissenschaft. Chesterman, A. 1993. "From 'Is' to 'Ought': Laws, Norms and Strategies in Translation Studies". Target 5(1): 1-20 Clark, H. and Marshall, C.R. 1981. "Definite reference and mutual knowledge". In A. Joshi, B. Webber and I. Sag (eds), Elements of Discourse Understanding. Cambridge: Cambridge University Press, 10-63. Dollerup, C. 1993. "Translation, interpreting and negotiations as forces in language change in member states of the EC". Paper delivered at the AILA Congress, Amsterdam 1993. Fillmore, C. J. 1976. "Frame semantics and the nature of understanding". Annals of the New York Academy of Science, 280: 20-32. Gowers, E. et al. 1986. The Complete Plain Words. London: HMSO. Hönig, H.G. and Kussmaul, P. 1982. Strategie der Übersetzung. Tübingen: Narr. Kade, O. 1980. Die Sprachmittlung als gesellschaftliche Erscheinung und Gegenstand wissenschaftlicher Untersuchung (Übersetzungswissenschaftliche Beiträge 3). Leipzig: Enzyklopädie. Kohlmayer, R. 1994. "Übersetzung als ideologische Anpassung: Oscar Wildes Gesellschaftskomödien mit nationalsozialistischer Botschaft". In M. Snell-Hornby, F. Pöchhacker and K. Kaindl (eds), Translation Studies: An interdiscipline. Philadelphia/New York: Benjamins, 91-101. Lakoff, G. 1986.Cognitive Semantics. (Berkeley Cognitive Science Report 36). Berkeley: Berkeley University. Lambert, J. 1994. "The cultural component reconsidered". In M. Snell-Hornby, F. Pöchhacker and K. Kaindl (eds), Translation Studies: An interdiscipline. Philadelphia/New York: Benjamins, 17-26. Lörscher, W. 1991. Translation Performance, Translation Process, and Translation Strategies. Tübingen: Narr. Mann, W.C. and Thompson, S.A. 1986. "Relational propositions in discourse". Discourse Processes 9(1): 57-90. Markstein, E. 1994. "Sprache als Realie: Intertextualität und Übersetzung. Am Beispiel totalitärer Sprachen". In M. Snell-Hornby, F. Pöchhacker and K. Kaindl (eds), Translation Studies: An Interdiscipline. Philadelphia/New York: Benjamins, 103111. Neubert, A. 1985. Text and Translation (Übersetzungswissenschaftliche Beiträge 8). Leipzig: Enzyklopädie. Neubert, A. and Shreve, G.M. 1992. Translation as Text. Kent and London: Kent State University Press. Newmark, P. 1991. About Translation. Clevedon: Multilingual Matters.

CHRISTINA SCHÄFFNER

141

Newmark, P. 1994. "Paragraphs on translation - 34". The Linguist 33(6): 214-218. Nord, C. 1988. Textanalyse und Übersetzen. Heidelberg: Groos. Partridge, E. 1980. Usage and Abusage. Harmondsworth: Penguin. Ringleb, K. 1983. "Voraussetzungen für die Wirksamkeit übersetzter Texte der politischen Auslandsinformation". In Probleme der Übersetzung gesellschaftswissenschaftlicher Texte. Berlin: Hauptvorstand der Vereinigung der Sprachmittler der DDR, 69-71. Robertson, D. 1988. The Penguin Dictionary of Politics. London: Penguin. Rothkegel, A. 1984. "Sprachhandlungstypen in interaktionsregelnden Texten". In I. Rosengren (ed.), Sprache und Pragmatik (Lunder Symposium 1984). Lund: Almqvist & Wiksell, 255-277. Schäffner, C. 1986. "Analysis of interrelationships between keywords as a means of determining their meaning in the Final Act". In L'Acte Final d'Helsinki (1975): L'analyse semiotique dans la recherche comparative - Le vocabulaire des relations internationales. (Occasional Paper 3). Vienna: Vienna Centre, 95-104. Schäffner, C. 1992a. "Translating Governmental Documents". In C. Mair and M. Markus (eds), New Departures in Contrastive Linguistics / Neue Ansätze in der Kontrastiven Linguistik. Proceedings of the Conference held at the Leopold-FranzensUniversity of Innsbruck, Austria, 10-12 May 1991 (Innsbrucker Beiträge zur Kulturwissenschaft, Anglistische Reihe Band 5). Innsbruck: University Press, vol. II, 143-154. Schäffner, C. 1992b. "Sprache des Umbruchs und ihre Übersetzung". In A. Burkhardt and K.P. Fritzsche (eds), Sprache im Umbruch. Politischer Sprachwandel im Zeichen von "Wende" und "Vereinigung". Berlin: de Gruyter, 135-153. Schäffner C. 1993. "Meaning and Knowledge in Translation". In Y. Gambier and J. Tommola (eds), Translation and Knowledge (SSOTT IV, Scandinavian Symposium on Translation Theory. Turku 4-6 June 1992). Turku: University Centre for Translation and Interpreting, 155-166. Schäffner, C. in print. "CSCE documents from the point of view of translation". In A. Neubert, G.M. Shreve and K. Gommlich (eds), Basic Issues in Translation Studies, Proceedings of the Fifth International Conference (Kent Forum on Translation Studies, vol. II). Kent: Kent State University Press. Schäffner, C. and Herting, B. 1994. "'The Revolution of the Magic Lantern': A crosscultural comparison of translation strategies." In M. Snell-Hornby, F. Pöchhacker and K. Kaindl (eds), Translation Studies: An Interdiscipline. Philadelphia/New York: Benjamins, 27-37. Schütte, W. 1993. "'Eurotexte' - Zur Entstehung von Rechtstexten unter den Mehrsprachigkeitsbedingungen der Brüsseler EG-Institutionen". In J. Born and G. Stickel (eds), Deutsch als Verkehrssprache in Europa (Institut für deutsche Sprache, Jahrbuch 1992). Berlin/New York: de Gruyter, 88-113. Snell-Hornby, M. 1988. Translation Studies: An Integrated Approach. Amsterdam/ Philadelphia: Benjamins. Suter, H.-J. 1993. The Wedding Report: A Prototypical Approach to the Study of Traditional Text Types. Amsterdam/Philadelphia: Benjamins.

142

STRATEGIES OF TRANSLATING POLITICAL TEXTS

Thiel, G. and Thome, G. 1987. Resolutionen. Ein empirisch entwickelter Beitrag zur Textanalyse.Tübingen: Narr. Toury, G. 1992. "'Everything has its price': An alternative to normative conditioning in translator training". Tijdschrift voor toegepaste linguïstiek / Journal of Applied Linguistics 6(2): 60-72. Trosborg, A. 1994. "'Acts' in contracts: Some guidelines for translation". In M. SnellHornby, F. Pöchhacker and K. Kaindl (eds), Translation Studies: An interdiscipline. Philadelphia/New York: Benjamins, 309-318. van Dijk, T. A. 1980. Macrostructures. An Interdisciplinary Study of Global Structures in Discourse, Interaction, and Cognition. Hillsdale, NJ: Erlbaum. van Dijk, TA. 1994. "Editorial: Discourse analysis as social analysis". Discourse & Society 5(2): 163-164. Volz, W. 1993. "Deutsch im Übersetzeralltag der EG-Kommission". In J. Born and G. Stickel (eds), Deutsch als Verkehrssprache in Europa (Institut für deutsche Sprache, Jahrbuch 1992). Berlin/New York: de Gruyter, 64-76. Wallerstein, I. 1991. "Scholarly concepts: Translation or interpretation?" Paper delivered at the conference Humanistic Dilemmas: Translation in the Humanities and Social Sciences, Binghamton, 26-28 September 1991. Appendix: Analysed texts (1) Konferenz über Sicherheit und Zusammenarbeit in Europa - Schlußakte (3-71) Conference on Security and Cooperation in Europe - Final Act (73-135) Conférence sur la sécurité et la coopération en Europe - Acte Final (201-263) In L'Acte Final d'Helsinki: texte et analyse. - European Coordination Centre for Research and Documentation in Social Sciences; International Social Science Council. Wilhelmsfeld: Egert, 1990 (2) "A New Europe, a New Atlanticism: Architecture for a New Era", Address by Secretary of State James A. Baker, III, to the Berlin Press Club (12-12-1989) "Ein neues Europa, ein neuer Atlantizismus: Architektur für ein neues Zeitalter". Rede von Außenminister Baker vor dem Berliner Presseclub (13-12-1989 / Amerika Dienst 45) (3) Bush Sends German Unification Treaty to Senate (25-9-1990) Bush leitet Vertrag über Vereinigung Deutschlands dem Senat zu (3-10-1990 / Amerika Dienst 39) (4) "Aus eigener Kraft allein hätten wir es nicht geschafft", Fernsehansprache des Bundeskanzlers Helmut Kohl am 2. Oktober 1990 {Der Spiegel, Dokument, no. 5, 1990,10-11) Television Address by Dr Helmut Kohl, Chancellor of the Federal Republic of Germany, on the Eve of the Day of German Unity. (5) "Sich zu vereinen, heißt teilen lernen", Ansprache des Bundespräsidenten Richard von Weizsäcker zur Deutschen Einheit, 3. Oktober 1990 (Süddeutsche Zeitung 11-101990)

CHRISTINA SCHÄFFNER

143

'The Day of German Unity", Speech by Richard von Weizsäcker, President of the Federal Republic of Germany at a State Ceremony in the Philharmonie, Berlin on October 3, 1990 {Statements & Speeches, 14-11-1990, Vol. XIII, no. 20) (6) Günter Grass "Ein Schnäppchen namens DDR". In G. Grass, Ein Schnäppchen namens DDR. Letzte Reden vorm Glockengeläut. Frankfurt/Main: Sammlung Luchterhand, 1990, 39-60. Günter Grass "The West German business Blitzkrieg". Guardian Weekly 11-11-1990, 22-23. (7) Christa Wolf "Rede auf der Demonstration für Pressefreiheit in Berlin am 4. November 1989". Neue Deutsche Literatur, 3/1990, 173-175. Christa Wolf "Was tun? Germany's greatest living writer, celebrates a revolutionary moment of renewal and asks what happens next". New Statesman & Society, 17-111989,11. (8) Timothy Garton Ash "The Revolution of the Magic Lantern". The New York Review of Books, 18-1-1990,42-51. Timothy Garton Ash "Prag. Laterna Magica. Aus dem Herzen der Revolution". Lettre Internationale, Frühjahr '90, 34-43 (translated by Yvonne Badal) "Die Revolution der Laterna Magica". In T. Garton Ash Ein Jahrhundert wird abgewählt. Aus den Zentren Mitteleuropas 1980-1990. (Aus dem Englischen von Yvonne Badal), München: Carl Hanser Verlag, 1990, 401-450.

Translating Hybrid Political Texts Anna Trosborg The Aarhus School of Business

Introduction Political discourse is an umbrella term covering a variety of text types, or genres. It comprises inner-state and inter-state discourse and ranges from highly culture-bound texts, which reflect culture-specific conditions of their production, at one end of the scale, to diplomatic discourse in multinational institutions and standardised formulations of treaties at the other end (see Schäffner, this volume). Highly culture-bound political texts, for example speeches and statements made by politicians, refer to a wide range of cultural patterns of the society in question, including aspects of its economic, political and legal life. In contrast to such texts, there are political texts interactively negotiated in a supranational setting, for the overall purpose of achieving and reflecting consensus. Such texts include standardised formulations of treaties, documents of NATO, documents of the European Union, the European Parliament, etc. Contracts and treaties as genres display special conventions. The translations into different languages represent different versions meant to fulfil an identical function with their main addressees disposing of relatively identical background knowledge pursuing fairly identical political aims, and fulfilling practically the same roles in their respective societies.1 Due to increasing internationalisation processes, texts of this kind will pose challenges for translators. In today's world, the translation of documents is linked to cultural, social and political realities, a point which is often overlooked in translation theory (Dollerup 1996: 312). What happens when cultures want to or have to communicate by means of translations? What consequences will this have for the specific target text to be produced or for the target text type or genre? The concern of this paper is the way interlingual transfer work operates in the European Communities. Documents produced in a suprana-

146

TRANSLATING HYBRID POLITICAL TEXTS

tional multicultural discourse community where there is no linguistically neutral ground are referred to as "hybrid texts".2 Hybrid texts Hybrid texts are a feature of contemporary intercultural communication. They result from cultures and languages being in contact. Such contacts are initiated for differing communicative purposes, e.g. informative, propagandist, educational, entertaining, thought-provoking, or as here for political and commercial reasons. A provisional definition of a hybrid text is "a text that results from a translation process and shows features that somehow seem 'out of place'/'strange'/ 'unusual' for the receiving culture, i.e. the target culture (TC)". These features are not the result of a lack of translational competence or examples of "translationese", but evidence of conscious and deliberate decisions by the translator. Although the hybrid text is not yet fully established in the TC (because it does not conform to established norms and conventions), it is accepted in its TC because it fulfils its intended purpose in the communicative situation (at least for a certain time). Hybrid texts of the kind to be discussed in this article are arrived at as an outcome of negotiations between cultures and the norms and conventions involved as well as through translation.3 Such texts may display features which are a reflection of specific conventions in two (or more) cultures. Translation as intercultural communication Hybrid texts have been defined as being derived from a translation process. If we define translation as ST induced production, and as an instance of displaced situationality, translation into a foreign language will be an instance of intercultural communication. Culture is understood here, not in the narrower sense of man's advanced intellectual development as reflected in the arts, but in the broader anthropological sense of all socially conditioned aspects of human life. Salient aspects are: (a) The concept of culture as a totality of knowledge, proficiency and perception; (b) its immediate connection with behaviour (or action) and events; (c) its dependence on norms, whether those of social behaviour or those accepted in language usage (Vermeer 1986: 33). Thus culture has to do with common factual knowledge usually including political institutions, education, history, and current affairs. The concept of culture as a totality of knowledge, proficiency and perception is fundamental in the approach to translation. Translation will have to bridge the gap, small or large, between two different languages and cultures, and will hopefully result in contact, and not in conflict.

ANNA TROSBORG

147

Cultural strategies in translation The problem for the translator is how to comply with cultural norms, i.e. which norms are to be decisive: the cultural norms of the ST community as reflected in genre conventions, the cultural norms of the target text (TT) community, or maybe a combination of the two cultures, or a compromise between two or more cultures? Choice of cultural strategy may result in source culture(SC)bound translation (the translation stays within the SC), target culture(TC)-oriented translation (the text is translated into the TC) or hybrids (the translation is a product of/a compromise between two or more cultures. Finally, international texts are a product of a dominant culture. 1. SC-bound translation. Through translation, a new texttype or genre is introduced into culture B which builds on the model of culture A, either by culture A imposing its rules, text types, etc. on culture B, or by culture B importing the textual conventions of culture A. The political/ideological dimension may influence the decisions as to imitation or adoption. A distinction can be made between text types for which models already exist in the TC, and text types which are introduced into the TC only through translation. 2. TC-oriented translations. Due to the impact of translations, culture B modifies an existing text type in analogy to the text-typological conventions in culture B, doing away with the original culture-specific text-typological conventions of culture A. 3. Hybrid texts. The translation is a product of two or more cultures, or a compromise between a number of cultures. There are situations where the impetus to contact is mutual, thus resulting in jointly produced translations. Cultures A and B (and possibly more cultures and languages) come into contact due to increasing internationalisation processes. As already mentioned, examples include texts that are produced through collaboration in the European Community, the European Parliament, the United Nations, etc. 4. International texts. If a text type or genre resists changes across cultures, we can talk about international texts. Genres which remain more or less the same when translated across cultures are often a product of a dominant culture. This goes for types of advertising which help to create a new global culture (cf. Cook 1992), and for texts within international fields, such as science, technology and international politics. Factors influencing the choice of cultural strategy comprise:

148

TRANSLATING HYBRID POLITICAL TEXTS

- Type of text: Some texts are likely to be culture-bound, e.g. political speeches, legal documents, while other text genres are more likely to develop international norms. As mentioned above some technical and scientific texts are representative of science and technology as international fields. See also Göpferich (1995) reported in Kussmaul, this volume. - Purpose of translation: Texts may be translated for different purposes (functions). The function may be metatextual with the purpose of reporting exactly what is conveyed in a particular text (e.g. a political speech, a legal document), or the translation may function as a new text in its own right, for example as an advertisement in the TC. One and the same text may also be translated for different purposes. Thus, an announcement of death may be translated as a documentary translation to convey to relatives abroad what was in the announcement, or as an instrumental translation as an announcement adapted to TC conventions in order to appear in a paper in the TL as an announcement in its own right.4 - Status/dominance of the ST or ST-author: Some texts are representative of a dominant culture to be conveyed in the SC, or composed by a prestigious author, whose idiolect has to be kept in the translation. There is an increasing Americanisation, so that, for example, McDonalds would market their product almost the same all over the world. By contrast, Danish products sold on the German market will have to adapt to German taste (see Hare Hansen, this volume). Compare also English/American influence on the development of the research article.5 Countries with a digressive style of writing may have to adapt to more linear norms. - Readability'. A distinction has been made between reader versus writer responsibility. Thus English uses a writer-responsible rhetoric: it is the responsibility of the writer to make the text clear to the reader. In contrast, in readerresponsible rhetoric it is the responsibility of the reader to understand what the writer has intended to say. There is a tendency to encourage readability and hence writer-responsible rhetoric (see Kussmaul, this volume). - Loyalty: The question of loyalty is a tricky one: loyalty to whom? to the ST writer or to the TT reader? In principle, loyalty goes both ways - to the ST writer as well as to the TT reader (see Nord 1991), but as we have seen status of ST writer and cultural dominance may play a role. A possibility is that a compromise has to be reached. New text types or genres may result from the process of translation. The formulation of, for example, standardised treaties may involve the creation of

ANNA TROSBORG

149

hybrid texts. Texts that are produced through intercultural negotiation (as well as through translation) display features which are often a reflection of specific conventions in two or more cultures. The resulting text is a hybrid incorporating different norms from the cultures involved. Problems connected with this type are: Why do hybrid political texts occur? Who is responsible for their creation? How can we define the process of hybridisation? What are the identifying characteristics of hybrid political texts as product? I will now discuss these questions within a sociocognitive approach to genre analysis and with documents within the European Union (EU) as the particular object of investigation. A sociocognitive approach to hybrid texts A genre comes into existence as a recognizable response to an exigent state of affairs. In the words of Berkenkotter and Huckin (1995: 7), "We use genres to package our speech and make of it a recognizable response to the exigencies of the situation". Combining their expertise in the fields of discourse analysis and cognitively based rhetorical research, Berkenkotter and Huckin have developed what they call a sociocognitive theory of genre comprising five principles that constitute their theoretical framework: Dynamism, situatedness, form and context, duality of structure and community ownership (see Berkenkotter and Huckin 1995: 4-24). Berkenkotter and Huckin point out that genres are dynamic and change over time in response to their users' sociocognitive needs. Genre knowledge is a form of "situated cognition", as our knowledge of genres derive from our participation in the communicative activities of daily and professional life. It embraces both form and content, including a sense of what is appropriate usage in a particular context. This knowledge, rather than being explicitly taught, is transmitted through enculturation as apprentices become socialized to the ways of speaking in particular disciplinary communities. Duality of structure refers to the process of writing, in which the use of rhetorical genres is both constitutive of social structure (as it is instantiated through our observing a genre's rules-for-use or conventions) and generative: "As we draw on genre and genre rules to engage in professional activities, we constitute social structures (in professional, institutional, and organizational contexts) and simultaneously reproduce these structures" (Berkenkotter and Huckin 1995: 17). The fifth aspect community ownership refers to genre conventions as signalling "a discourse community's norms, epistemology, ideology, and social ontology" (Berkenkotter and Huckin 1995: 4).

150

TRANSLATING HYBRID POLITICAL TEXTS

EU documents In the following, EU documents are discussed within a sociocognitive approach to genre. Situatedness Sociopolitical changes in a given culture create the need for new or modified text types, just as the increasing internationalisation of communication processes breaks down text type areas. There is an increasing convergence between cultures, a partly inevitable tendency. Also institutional patterns of behaviour may occur in similar forms in different cultures as a result of international strategies (in multinational companies, international alliances for defence, trade, education, etc.). Translators at the Commission of the European Communities work with different types of source texts. Hybrids may occur with common communications (e.g. circulars to staff) and also with legal documents. Cultural boundaries change due to social and political integration and internationalisation processes. EU documents are evidence of such (overt or hidden) internationalisation processes. Community ownership The stimulus may come from the society as a whole, a specific subgroup, an individual, or a multi- or supra-cultural environment. The European Market (now The European Union) rose in response to a demand for cooperation on matters of shared interest, in particular to regulate production and trade. The Common Market was formed in 1952 and consisted originally of six member states: France, Germany, the Netherlands, Belgium and Italy (see Dollerup 1995: 28; Dollerup 1996: 297ff for historical accounts). In 1973, three more nations became members: the United Kingdom, Ireland and Denmark. Today the EU has 15 nations as members and no fewer than 11 official languages, and the areas of cooperation have been extended to include common regulations for agriculture, fishing, trade and industry, economical and environmental aspects. Officially, all eleven languages are equally valid. According to current language policy in the EU, all eleven languages are spoken in the European Union, and all documents are translated into each of the official languages. It is a matter of language policy in the individual countries, though, whether all non-native speakers have a legal right to interpreting and translation in court. In principle, there may be the same degree of involvement of any official language, but in practice, this is not the case. Most documents are drafted in one of the dominant languages (French, English, German and Spanish), even

ANNA TROSBORG

151

by non-native speakers (Karker 1993: 37). A draft in one language (most often French) is then translated into the 10 other official languages. At national meetings, national delegates use national language documents, which are revised until there is agreement about the final document. The final version is collated by legal linguists; they check that the wording is the same in all versions of the document, which in turn are equally binding in all official languages (cf. Dollerup 1996: 300). Form and content Hybrids reflect specific textual features (vocabulary, syntax, style, etc.) which may clash with target language conventions. In the process of establishing political unity within the Union, linguistic expressions are levelled to a common, (low) denominator. EU documents have developed a specific language involving coinage of new concepts as well as new terms for the drafting of documents and for use in Community negotiations. As a result, eurotexts are characterized by reduced vocabulary, meanings that tend to be universal, and reduced inventory of grammatical forms (cf. Schütte 1993, Pym 1993). Standardisation The aim of the language work in the Union is standardisation of texts permitting for multilingual formulas and nomenclatures, which would make for simplification and rationalisation of translation work. New socio-cognitive realities require new terms. Where established expressions and technical terms with equivalent terms in the TL are not available, new terms have to be coined. These terms are often unfamiliar and strange to the TLs into which the documents are translated. As an example can be mentioned the use of subsidiaritetsprincippet in Danish translations. This term is a derivation of the German word Subsidiarität, which refers to the principle that a higher institution in society (especially the State) can only take on tasks and burdens which lower authorities, for example the family, cannot handle alone. While the term is well-established in German usage, the translation sounds foreign to the Danish language and there is a preference for the more tangible term nærhedsprincippet Similarly, the French expression l'acquis communitaire has been adopted in English as well as in Danish; the word acquis appears in italics and quotation marks, respectively, in English and Danish texts: "the Community's acquis, as it is known"; "også kaldet Fællesskabets 'acquis'" (see Karker 1993: 38-39). In written translation, there is a demand for uniformity and the organisation of the material must be the same in order to ensure mutual understanding. The

152

TRANSLATING HYBRID POLITICAL TEXTS

keyword is "sameness" between texts in all 11 languages and the result is standardisation. The documents are typed out in much the same way in terms of sentences (and largely layout). In oral and written negotiations that are based on a document, the negotiators must be able to refer to a particular article, paragraph, and sentence, etc., which may result in specific syntactic and lexical features. Furthermore, there are constraints on sentence length: one sentence in the ST must correspond to one sentence in the TT. Another rule which determines the structure of EU documents is the fullstop rule, which prescribes an equal number of full stops in ST and translations. The constraints on sentence length lead to long and complicated sentences, which is a violation of, for example, the Danish recommendations for short and clear sentences. Where the translator into Danish would like to split up a complicated sentence structure, s/he is forced to keep a structure which is often more difficult to comprehend. Karker (1993: 60) mentions an example with no less than 57 words between the subject and the corresponding verb. All that Danish translators can do to facilitate comprehension is to insert semicolons. The punctuation constraint is highly visible in the opening statement of preambles for all legally binding documents from the EU. Union legalese The demand for "sameness" is a must in legal acts. The result is a kind of Union legalese, which is established in an intricate interplay between all EU government officials and the permanent EU staff. Each member nation has its own network between delegates, interpreting and translation services. In the common Union cooperation, common directives and laws which are legally binding for all nations are worked out. They create a new reality, which, because it is new, needs a new language and terminology for being described. It comes as laws, directly so in the form of regulations which are linguistically the same in the EU member states, and indirectly in directives which are used for national legislation. This kind of legalese is centrally bound and heavily biased towards the "sameness" format of all 11 EU languages (see Dollerup 1996: 306-308). A striking example is the so-called preamble which inevitably precedes all legal documents in the Union. This preamble is remarkable for its mere existence. In Danish, for example, it is only treaties with foreign powers which are preceded by a preamble. Furthermore, the preamble is remarkable for its characteristic grammatical structure. It follows a predetermined pattern prescribed by the Council's formulas. It has the following formulation in Danish, English, and French, respectively:

ANNA TROSBORG

153

Rådet for de europæiske fællesskaber har under henvisning til Traktaten...., under henvisning til forslag fra Kommissionen, under henvisning til udtalelse fra Europa-Parlamentet, og ud fra følgende betragtninger: I....omhandles...., i....er det bestemt, at...., fremgangsmåden er...., anvendelsen af foranstående regler medfører, at...., som følge heraf bør....udstedt følgende forordning: The Council of the European Communities, Having regard to the Treaty...., Whereas...., has adopted this regulation Le Conseil des Communautés européennes, vu le traité que...., a arreté...., considerant que...., a arreté le présent règlement. The number of parts in the preamble varies a great deal, and the specific references and considerations may vary in quantity. In Danish, this procedure results in the separation of auxilary and main verb, which is contradictory to recommended Danish sentence formation and sounds very queer in Danish. The more references and considerations the preamble comprises, the stranger the document appears. The language of the Eurocrats As we have seen, EU documents have developed a specific language of the Eurocrats for the purpose of document drafting and Community negotiations. The result is a certain jargon, with which the EU negotiators and staff, including translators and interpreters, are familiar. This language, which is often blurred, complicated and hard to understand, has been labelled eurojargon. The new reality spans the spectrum of the EU and also, at another remove, the reality of each of the member nations, where it will hardly sound "native" in all nations. This EU jargon is most outspoken in Union legalese, but there is also a certain jargon, a special EU-dialect which is slowly manifesting in the spoken languages in the EU. It is known to delegates and EU staff, including translators and interpreters. This language is not only a sociolect, but often becomes a language for special purposes, namely EU negotiations which are often inpenetrable to outsiders. Union legalese may lead to lack of comprehension. For example, in the early days of Danish participation, there were cases where a number of factors,

154

TRANSLATING HYBRID POLITICAL TEXTS

including lack of interpreting and deficient feedback from Denmark led to legal documents which were incomprehensible in Danish, so much so that the ministry involved had to publish an instruction in which it explained the new law to Danes. It is reasonable to expect that a translation into Danish or any of the EU languages can be read as original prose like any other text. However, the official guidelines for written Danish, emphasizing the use of short, clear sentences, are often violated in many EU translations. For example, Europe and the challenge of enlargement becomes Europa og Udfordringen Forbundet med Udviklingen, instead of Europa og den udfordring, der er forbundet med udviklingen, which sounds more Danish. In addition to complex sentence structure, complex NPs, etc., overuse of abstract nouns and extensive nominalisation are typical features diminishing the readability of EU documents and texts.6 The translators rely on their general expertise, on their access to previous documents pertinent to the task, and on specialist terminologies (Dollerup 1996: 300). In addition, translators may call the "originators" of the documents as well as other experts for elucidation. Subsequently, their translations may even be checked by "revisers". To assist EU translators and interpreters in their work, a common terminology database called Eurodicautom with half a million items in two languages and up to nine languages has been worked out. This database has to be used with care, though, as a term may have entered the base from a subject matter totally different from the one the translator is working with. As a reaction to the often incomprehensible and stilted language in EU documents, some reformation has found its way into directives. Furthermore, Danes are now allowed to conclude letters the Danish way: "Med venlig hilsen", instead of the French "Veuillez agréer, monsieur le professeur, l'assurance de mes sentiments distingués" (Dollerup 1996: 306). Dynamism and duality of structure The EU is a dynamic union in which proposals for extended cooperation, such as monetary union and military cooperation, are continually negotiated. Besides, the linguistic management of the EU institutions has a noticeable effect on member state societies including their languages. As EU documents become part of the national legislation in the individual member states, they are likely to have a strong regulative role on the language in these communities and, consequently, contribute to language change. The impact is felt most strongly in documents which are the outcome of an interplay between interpreting and translation at meetings, whereas the linguistic

ANNA TROSBORG

155

traces of the EU origin will be less obvious in a directive worked out in national legislation (Dollerup 1996: 308). Even if an effort is made to find equivalent terms in the respective target languages, interference from the ST is hard to avoid, and the translated products are said to be in EU Danish, EU Dutch, etc. They appear as translations, not as originals. As the end products of international negotiations are agreements with a legally binding power, translation and interpreting are formative elements in the target societies, and thereby formative elements in language change in the national languages of the member states (Dollerup 1996: 308).

Hybrid texts as primary communication Generally, translated texts are instances of secondary communication, i.e. the receivers are not the receivers for which the text was originally created but a secondary target group intended as receivers of the translation. A text which in itself does not fulfil a communicative function, but serves as a draft for translation may be referred to as a "pseudo-text". There is no primary communicative situation in which the ST has fulfilled its communicative function for its addressees and where, subsequently, the TT is to fulfil its function for TL addressees in a secondary communication situation. Furthermore, there is often no proper ST displaying characteristic features or specific conventions of the SL text type, and no proper parallel texts in the TL culture to serve as a model for the TT. From this point of view, translations of pseudo-texts come close to primary texts addressing a primary target group, even though they are nevertheless derived as an outcome of translation. Examples would be documents to fulfil the same function but within different societies (e.g. brochures and manuals) or texts in which uniformity is the guiding principle (e.g. legal documents and treaties). Accepting the rendering into different languages and cultures of pseudotexts as translations means that the status of both the ST and the SL community would have to be reconsidered, a step which some scholars are not ready to take. Compare the standpoint taken by Gutt (1990) that such translations are not translations at all as there are no proper STs. The question arises whether hybrid texts of the kind discussed in this article are translations at all? In this connection, I would like to draw attention to Sager's (this volume) important distinction between "documents" and "texts". "Documents" are primary texts which are constructed by the author with a specific intended audience in mind. In contrast, he refers to "texts" as being produced with no specific audience in mind; texts are secondary communication and often a product

156

TRANSLATING HYBRID POLITICAL TEXTS

of translation. Within this terminology, most hybrid texts of the kind treated in this article are "documents", even though they are translations. The hybrid "source text" is a pseudo-text arrived at as a compromise between several languages and cultures, for which the specific translations have a defined intended readership, and as such they function as documents in their own right. Considering the formation process, hybrid texts evolve as a product of various languages and cultures and their respective translations serve as instances of primary communication. The fact that these documents in many cases are linguistically marked as translations, written as they are in EU Danish, EU Dutch, etc., confirms their status as translations, although they function as documents in their own right within a supranational discourse community. Once a text is accepted in the target culture, is it then still a hybrid? Or are hybrid texts a transitional and historical phenomenon? What has been claimed here is that EU documents are hybrid texts with specific genre characteristics. Like genres, hybrid texts may change over time in response to their users' sociocognitive needs, they are a form of "situated cognition", and they derive from participation in the communicative activities of professional life in a particular setting. Specifically, EU documents are a product of a multicultural discourse community arrived through a process of negotiation and compromise. Both form and content are adjusted to what is appropriate to this particular purpose in this particular situation at a particular point in time. Through the process of writing/ translation, hybrid texts are both constitutive of social structure (as negotiated, for example, by EU societies) and generative, as these texts are reproduced in professional, institutional, and organisational contexts. Although such texts are a product of a community ownership representing several languages and cultures, they may eventually become established text types with their own norms and conventions. Finally, we may ask whether specific textual features (vocabulary, syntax, type, etc.) compared to language-specific features are a loss or a gain? And what factors influence the reception of hybrids in the target society? When EU texts function within the Community within which they are created (e.g. for the staff, or for meetings of the respective bodies), there are clearly defined user needs. These texts are accepted due to the limited communicative functions of the texts. When EU texts are translated for use in the member states, different criteria for acceptance may apply. The receiving culture may adopt - a defective attitude (enriching its own culture with foreign elements), - a transcultural attitude (a kind of carelessness in adopting L2-elements in the L1),

ANNA TROSBORG

157

- a defensive attitude (rejecting the "otherness" and doing everything to stop it) (cf. Robyns 1994). There may be a hidden political or ideological agenda governing acceptance and institutionalisation depending on the attitude of political parties and individual members. Notes 1 See also Schäffner, this volume, for representative examples of CSCE discourse, i.e. representative of the Final Act of the Conference on Security and Cooperation in Europe (CSCE), which is remarkable for its special terminology with creation of specific terms in the CSCE discourse resulting in CSCE-specific expressions. 2 The author acknowledges the insight gained in a workshop on hybrid texts arranged by Christina Schäffner and Beverly Adab at EST Translation Congress, Prague, September 1995. A preliminary version of this article was presented at a symposium organised by John M. Swales, Thicker Description of Genre Analysis, Aila '96, Yväskylä, Finland. 3 Thus hybrid texts are not hybrids in the sense that they are a mix of various text types in the rhetorical sense of narration, description, exposition, argumentation, etc., or as a mix of representative, expressive or directive (appellative) communicative functions. Neither are they of a kind like hybrid Hamlets. See Kennedy (1993) for an account of cultural adaptation in translations of Shakespeare's Hamlet. 4 For an account of documentary vs. instrumental translation, see Nord (1991), Nord this volume. 5 For example, in Czech culture, like many Eastern European cultures, people are reluctant to reveal their points; the article should read like a detective story. This is quite contradictory to Western norms, where progression is preferred to digression, and the style of Czech articles often needs to be adjusted for publication in international journals (see Cmejrková 1996). 6 Examples from Karker (1993: 53) are havnemiljøbeskyttelsesforanstaltninger, markedsføringstilladelsessag, personaleressourceplanlægning, transportinfrastrukturomkostningerne. References Berkenkotter, Carol and Huckin, Thomas N. 1995. Genre Knowledge in Disciplinary Communication: Cognition/culture/power. Hillsdale, New Jersey: Lawrence Erlbaum Associates, Publishers. Cmejrková, Svetla. 1996. Translation, intellectual style, cross-cultural communication. Paper presented at EST Translation Congress, Prague, September. Cook, Guy. 1992. The Discourse of Advertising. Routledge: London. Dollerup, Cay. 1995. "English in the European Union". In Reinhard Hartmann (ed), The English Language in Europe, EUROPA 2.3. Oxford, 24-36. Dollerup, Cay. 1996. "Language work at the European Union". In Marilyn Gaddis Rose, Translation Horizons Beyond the Boundaries of Translation Spectrum. State University of New York at Binghamton: Center for Research in Translation, 297314.

158

TRANSLATING HYBRID POLITICAL TEXTS

Gutt, Ernst-August. 1990. "A theoretical account of translation - Without a translation theory". Target 2.2: 135-164. Göpferich, Susanne. 1995. Textsorten in Naturwissenschaften und Technik. Pragmatische Typologie - Kontrastierung - Translation. Tübingen: Narr. Karker, Allan. 1993. Dansk i EF - en situationsrapport om sproget. København: Gads Forlag. Kennedy, Dennis. 1993. Foreign Shakespeare. Cambridge University Press: Cambridge. Nord, Christiane. 1991. Text Analysis in Translation. Amsterdam/Atlanta: Rodopi. Pym, Anthony. 1993. Epistemological Problems in Translation and its Teaching. A seminar for thinking students. Caminade: Calaceit. Robyns, Clem. 1994. "Translation and discursive identity". In Clem Robyns (ed), Translation and the (Re)Production of Culture. Leuven, 57-81. Schütte, Wilfried. 1993. "Eurotexte" - Zur Entstehung von Rechtstexten unter den Mehrsprachligkeitsbedingungen der Brüsseler EG-Institutionen. In Joachim Born und Gerhard Stickel (eds), Deutsch als Verkerhsspräche in Europa (Jahrbuch/ Institut für Deutsche Sprache; 1992) Berlin/New York: de Gruyter, 88-113. Swales, John M. 1990. Genre Analysis. English in academic and research settings. Cambridge: Cambridge University Press. Vermeer, Hans J. 1986. "Übersetzen als kulturelle Transfer." In Mary Snell-Hornby (ed), Übersetzungswissenschaft - Eine Neuorientierung. Tübingen: Francke, 30-53.

Translation of Medical Research Articles Morten Pilegaard The Aarhus School of Business

Introduction Over the past two decades, much attention has been given to the phenomenon of linguistic variation by various linguistic schools, especially British neoFirthian contextualism, Czech functional stylistics and American discourse analysis. However, current literature on language for special purposes (LSP) is fairly unanimous in adopting a mainly sociolinguistic rather than a functional approach to medical language. The prevailing sociolinguistic approach to medical language describes and analyses it in terms of speakers and communicative situations. A notable exception is Salager (1983), who argues that the peculiarities of LSP are first and foremost of quantitative nature. Medical language may be described as a type of register, i.e. a variety of language appropriate to different occasions and situations of use. Ranges of variation of this register have been described in terms of technicality, formality, and channel of communication. As a highly technical, formal written text, medical texts have, among others, been assigned to the fields of discourse called 'technical English' or 'scientific English' (Brunt 1987: 448). Defined in terms of register, a medical text must be described by means of its contents, i.e. medicine in general or one of the various specialist fields, its mode (e.g. oral or written, written medical language for research reports or written medical language for equipment manuals, etc.), its tenor, its communicative function and national language. The 'tenor' approach to the study of medical language allowed Lankamp (1989: 21) to distinguish between major variants of medical language: (1) language of medical education (e.g. textbooks), (2) language of medical occupation (e.g. journal articles), (3) language of medical journalism (popular medicine), (4) doctor/patient language, and (5) medical technical language (e.g. manuals). The communicative purpose of medical language is first and foremost to provide a nonambiguous and preferably nonsynonymous language by

160

TRANSLATION OF MEDICAL RESEARCH ARTICLES

means of terminologies in order to express relevant concepts, especially in the expert-to-expert tenor. Medical English has also been studied under a variety of other headings depending on the focus and scope of the study: 'English for Occupational Purposes' (EOP), 'English for Academic Purposes' (EAP), and 'English for Special (Specific) Purposes' (ESP). However, this article will not be concerned with whether medical English is one or the other of these varieties. It suffices to note that there is an isolatable variety or subregister called 'medical English' which varies according to the roles and status of the participants, their degree of specialised training, and the setting in which the communication takes place. Medical English is here seen to comprise both highly technical jargon in the expert-to-expert tenor and the colloquial vocabulary of the layman. The history of medical translation is long. Quoting Fishback (1986: 16), medical translation along with religious translation "is the most universal and oldest field of scientific translation because of the homogeneous ubiquity of the human body". This rather common perception of medical translation is often accompanied by claims that medical translation poses fewer difficulties than many other types of scientific translation (Fishbach 1986: 19) because the medical vocabulary is almost universally based on Greek and Latin roots. In support of this claim, it should be mentioned that physicians are still using Latin and Greek for records and prescriptions. Although tradition places the birth of modern Western medicine in Greece, it must be recognized that Greek medical scholars and translators had access to earlier writings in Sanskrit, Egyptian and other languages. The purpose here, though, is not to give an account of the preHippocratic history of medical translation (see Fishbach 1986, Hoof 1993), but merely very briefly to discuss certain aspects of translation of medieval medicine relevant to today's medical translation. In the 15th century, accurate translations of Greek writers began to appear in English. The Liber Uricrisiarum, a 15th century treatise on uroscopy, provides scholars with the first major text in Middle English on this subject. The Liber Uricrisiarum was translated from the treatise De Urinis by Isac Judaeus, a Hebrew physician of the late 9th and early 10th centuries. The existence of this and other similar texts in Middle English demonstrates the early emergence of vernacular texts, usually translations from Latin or Greek in the late 14th and early 15th centuries. Thus, the gradual development of English as the international language of medicine only began some 500 years ago. Another major novelty of that time was that a major work on new drugs discovered due to the widespread colonisation of the New World was not written in Latin but in Italian. This work was translated into English in 1577 (Fishbach 1986: 19). For a brief history of medical and pharmacological nomenclature, see Gabrieli (1986). English medical translators of the 15th and 16th centuries were basically facing the dual challenge of translating from the

MORTEN PILEGAARD

161

classical languages or contemporary vernaculars into English on the one hand, and accepting the use of borrowed or exotic terms from non-classical languages, on the other hand. Still, at the time, progress in the medical sciences was comparatively slow and the translator's task was simplified by the fact that the basic anatomical and physiological elements of medical communication were largely the same all over the world. This sameness of object or concept to be translated "gives the medical translator an edge over his colleagues in other fields" (Fishbach 1986: 19), so from the point of view of some, medical translation was a comparatively easy task. It may, indeed, previously have been a relatively simple task, but with the current pace and proliferation of medical research, this is hardly so any longer, and convincing evidence to the contrary has been given, among others by Newmark (1976a, 1976b). The relevance of medical translation may be discussed both from a theoretical and a practical point of view. Theoretically, a closer look at this specific LSP register may reveal interesting aspects of its dynamic interactions. Firstly, aspects of language typology (English, Spanish, French, German, Chinese, Danish, etc.) may give rise to different ways of semantic conceptualisation and selection of lexical and syntactic patterns in medical texts. Secondly, aspects of text typology (scientific, semi-scientific, layman, etc.) may account for differences between the source language (SL) and the target language (TL) at the level of the discourse structure in a given register. Thirdly, incorporation of reader/text focus, i.e. observance of sociolinguistic norms (e.g. Newmark 1976: 42), cultural presuppositions, and conversational rules at work in SL as opposed to TL, may reveal interesting cultural differences between individual languages. For example, the translation of a Nordic language medical interview into most other languages, e.g. French and German, would have to account for the fact that in the Nordic countries the 2nd person singular pronoun du as both formal and informal form of address corresponds to the courteous form of the 2nd person plural pronouns vouz in French or Sie in German. Plenty of other examples can be found in multiethnic societies such as Australia, where the choice of lexemes reflects variations in terminology according to country of origin and terminology may have to be adapted to the age of the intended reader as discussed by Chesher (1988). So the 'taste' of the language in the context of its culture must be catered for, also in a technical register such as medicine. The mode of presentation may also undergo changes during translation. For example the typical Nordic informal style may not work in the TL and translators must abandon SL style where it is likely to confuse the reader or give the wrong impression. From a practical point of view, the study of medical translation is very relevant. Firstly, reality is far more complex than suggested by Fishbach (1985). Newmark (1976: 43), for example, demonstrates that in medicine the translator

162

TRANSLATION OF MEDICAL RESEARCH ARTICLES

must, in fact, often deal with a lack of physical correspondence between concepts in different languages. Hence, French and German have no terms for knuckle, French none for shin and in Russian there is no distinction between hand and arm. Secondly, and more importantly, the relevance of medical translation springs mainly from the fact that English has today replaced Latin as the language of international medical communication. English is the international language of science and technology per se, and as technology is a growing subset of medicine, any medical translator hence increasingly relies on his mastery of the English language for the acquisition and dissemination of knowledge about the medical subject field, which, to a large extent, has come to include chemistry, electronics and statistics, among others. Thirdly, owing to rapid progress in medicine, the need for exchange of information is great. Moreover, new names are being constructed daily by researchers, with little coordination (Gabrieli 1986: 22). It may, indeed, be claimed that medical translators may run into the problem that there is no such thing as a systematised knowledge of the syntactic, semantic and pragmatic limitations of medical terminology because specialised registers in general and medical terminology in particular undergo constant innovation, adaptation and change. Medical terminology provides an exemplary illustration of such rapid language change. WHO has estimated that several thousand new terms are created annually. Many of these neologisms are products of new discoveries in the biomedical and related sciences. Others are due to theoretical reorientations, as in the reclassification and renaming of viruses. Many others are duplications or synonyms of existing medical concepts (Barkman 1974: 28). For a single medical entity, there may be as many as 30 (near) synonyms, each with slightly different shades of meaning. Gould (1990), for example, discusses how the meaning of AIDS is derived from medical facts about AIDS and, importantly, from their connotations as they combine with deeply rooted symbolism and cultural attitudes about medicine, religion, and society in social discourse. The proliferation of terms that have sprung up in the wake of the AIDS epidemic since 1980 represents a very illustrative example of how rapidly new terms are being made. From 1979 to 1980, when the first reports on the epidemic nature of AIDS in Western societies were being published by researchers at the National Institute of Health, Bethesda, U.S.A., it took less than two years before a whole new set of specialised English loan words had found their way into Danish medical language. The adaptation and translation of these terms from English into Danish by researchers in the laboratories of the Danish Cancer Society in Aarhus and the State University Hospital in Copenhagen were almost instantaneous, yet highly culture-specific. A more serious problem than neologisms, the proliferation of synonyms and the hasty appearance of new specialised terms is medical homonymy, i.e. the

MORTEN PILEGAARD

163

use of a single term to describe different medical concepts. In this context it is interesting to observe that in his search for features common to ancient Latin and modern English medical terminology, Langslow (1989: 49) finds that the incidence of synonym pairs and translation pairs between Latin and English does not decrease, quite the opposite. The problems that may arise from the use of a single term to describe different medical situations may be exemplified by the French medical term petit mal, which may refer to either 3% or 80% of all cases of epilepsy, depending on how the specialist chooses to group the different types of seizures involved. This type of homonymy often causes confusion across language boundaries as well as within the same language. Other examples are schizophrenia, chronic bronchitis and peptic ulcer which have different meanings in German, French and English, formal linguistic similarities across the three languages notwithstanding. In an effort to solve this problem, various national and international groups have attempted to standardise medical terminology. However, their efforts have failed to reach the vast majority of people communicating about medical topics, and many special fields of biomedicine have not yet even confronted the major problems involved in the development of a standardised, systematic terminology. Medical language is not only a technical vocabulary, it is also fraught with professional jargon (Hoof 1959: 419). The user's choice of TL lexeme often varies according to the genre, which could be a case report (narratives of single cases of disease or injury), a disease review (summations of knowledge concerning specific diseases or conditions), a treatment-focused report (reports focusing on how to treat specific diseases/conditions), an experimental report (reports aimed at providing potentially applicable, experimentally derived medical/scientific information), and a speech (printed versions of speeches) made to professional medical groups on medical topics (Atkinson 1992: 337374). It should be noted that many, particularly non-native physician-writers also lack English writing skills and training in writing in English. When they write, their conscious rhetorical focus is on the subject, to the apparent exclusion of persona and audience as also noted by Kaufmann (1989). The result is that use of technical jargon often looms high. In medical translations across genres, this is certainly an extra challenge for the translator. The relevance of medical translation can hardly be questioned. It is currently one of the most pertinent areas of LSP translation, the more so as focused training in medical translation is sparse, or even non-existent at traditional translation schools. This paper will focus on the medical research article as one of the more relevant and challenging medical text genres from a translation point of view. Apart from the general problems in medical translation, some of which were introduced above, the translator of a medical research article faces a range of other challenges. For example, in medical research papers the au-

164

TRANSLATION OF MEDICAL RESEARCH ARTICLES

thor's communicative strategy is traditionally concerned with indexing of modesty and caution in the narration of experiments. Lexemes are usually chosen with the explicit aim of encoding this strategy. Hedges are introduced to show an appropriately cautious attitude towards the interpretation of the empirical evidence presented. Research articles are hence not merely straightforward accounts of laboratory findings: they can be seen as complex conventionalised structures whose main rhetorical function is to appear trustworthy to readers who have no direct access to the recorded empirical events (Atkinson 1992: 340). Linguistically, certain features accordingly tend to occur differently across different sections of a research article. This does not imply that convention, whether on the preference for a specific variant lexeme in a speciality or on the appropriate narrative strategy, is entirely resistant to change. On the contrary, the formal aspects of scientific writing are inextricably bound up with the epistemology of science itself and changes in scientific "thought-styles" are amply reflected in altered styles of writing, as shown by Atkinson (1992). Moreover, culture-specific differences affect the pragmatics of scientific discourse and divergences between scientific discourse styles are often striking. The translator of medical texts in general and scientific articles in particular is therefore up against a set of fascinating challenges. It is the purpose of this paper to make a selective review of recent contrastive studies of the translation of medical research articles and, by addressing some of these challenges, to suggest a number of translation strategies of particular relevance in the medical research article. Hence, this paper ignores a large number of other issues in medical language, e.g. sociological studies of doctor-patient interaction (Pendleton 1983), aspects of machine translation and construction of automated medical lexicons (Barkman 1974, Graitson 1975, Garciahidalgo and Dunham 1981), and translation of middle (medieval) English into today's English (e.g. McConchie 1988) or Arabic medical terms into modern French (Tropeau 1986).

Medical translation: Contrastive studies Most studies of medical translation focus on grammar rather than style and few direct their attention to the special constraints and challenges of the medical research article. Hie (1988) is a notable exception. She discusses the possibility of developing a translation paradigm for social-medical texts by identifying the structural interlingual equivalences between Swedish and Romanian. Although most studies are contented to state that translation problems stem from differences between language systems, it must be noted that translation problems also stem from conceptual differences, as will be shown below.

MORTEN PILEGAARD

165

The approach to the study of medical translation has mainly been contrastive and it is almost exclusively focused on translation into English. A selective review will be given of studies of the translation of major world languages (Spanish, French and German) into English. Reference to more exotic languages (Danish, Dutch and Chinese) will be made partly to illustrate points not captured by the previous studies. Major world languages In an interdisciplinary contrastive French-English study of gynaecology and obstetrics, Alexander (1987) repudiates the common fallacy that scientific discourse in general and medical discourse in particular is subject to straightforward objectivity. The author presents data that undermines the view that medical conditions can be defined in objective scientific terms. Alexander suggests that the Anglo-French difference in approach to preterm labour is reinforced, and perhaps even caused by the terminology used to describe its symptoms. English medical terminology is bestowed with more than half a dozen terms to describe uterine activity, but French only has one currently used term, viz contractions. It is also worth noticing that even though the English term impending preterm labour is used in centres in the U.S.A. where active management of preterm labour is sought, no such term exists in the U.K. In Britain there is simply no word for a situation in which the clinician observes clinical signs which may foretell the onset of labour, but in all likelihood do not (see also Atkinson 1987: 183). From a translational point of view, the main thrust of Alexander's paper is that, in practice, translators should always check the comparability of the concepts at issue "by looking closely at the words people use" (Atkinson 1987: 182). As demonstrated by Atkinson, problems at the lexical level of translation are culture-bound and stem partly from the simple fact that apart from bilingual medical dictionaries between major world languages, which often have a most rudimentary microstructure, very few specialist lexicographic tools are, indeed, available to the medical translator, particularly so if the translator is working from/to a more exotic language (Pilegaard and Baden 1994b). Examples of specialised English-French lexicographic works are, firstly, a bilingual glossary on immunology compiled in connection with a diachronic/etymological study of the development of immunology as an autonomous science and its relationship to other medical and biological sciences (see Jammal 1993). Secondly, the specialised terminology of neurosurgery in English and French is presented in a lexicon covering 203 concepts presented and organised into five areas: generalities, anatomy of the nervous system, radiological examination, surgical instruments and surgical techniques (Valiquette 1991). According to a comprehensive review of monolingual and bilingual

166

TRANSLATION OF MEDICAL RESEARCH ARTICLES

medical dictionaries by Dressier and Schaeder (1994), very few lexicographers have endeavoured to make comprehensive bilingual dictionaries on medicine. In the context of bilingual societies based on the Canadian (e.g. Rouleau 1993a, Rouleau 1993b; Jacopoosie 1993) and the American experience (Weaver 1982), problems of translating at the level of syntax have been given some attention. The purpose of such studies has mainly been to help French translators of English medical texts produce idiomatically adequate translations in recognition of the fact that English in recent decades has rapidly been gaining ground as the international language of scientific exchange. Formerly an important international language in medicine, French is becoming increasingly constricted in its ability to function in a technological setting (Decennie 1974). Despite efforts in France and Canada to reestablish French as a prominent medical language, English remains thoroughly monopolistic. Most contrastive studies of English and French are content today to point to specific translation problems in translating from French into English and not vice versa. A point of particular difficulty for French translators of English medical texts is the translation of compounds, especially sequences of nouns and/or adjectives. A thorough account of such problems is offered by Martin-Valiquette (1991), who demonstrates how modified heavy noun phrases in English correspond to prepositional phrases in French. Thus, the French prepositional phrase épreuve volumétrique de lyse par le glycérol must be rendered by an adjective plus a noun group in English, viz volumetric glycerol lysis test. The particular problem the translator meets here is that he will not arrive at the correct rendition of the concept in English by simply translating word for word. Furthermore, it is traditionally assumed that the English propensity for passives in scientific writing is also found in medical research articles and that this stands in contrast to the predominant use of active sentences in French scientific writing. However, in a quantitative study, Rouleau (1993a) studied the occurrence of the passive voice in French and English medical and paramedical texts and found that the traditionally assumed preference in French for active forms in comparison to English passive forms did not hold true: the passive was used with the same frequency in both English and French medical and paramedical texts (42%). However, the passive was used more extensively in French medical research articles than in other technical texts. Rouleau (1993a) presents one of the few contrastive studies of the translation of medical texts and specifically argues that the idea that the structure 'one/they + active verb' in French medical texts, if translated into an English passive, will result in an unidiomatic translation. Style is also traditionally an area of concern for the translator writing between English and French. In a comparative study of English and French medical articles and case reports, Newmark (1976a) describes the main gram-

MORTEN PILEGAARD

167

matical and stylistic differences between medical texts in the two languages: Where the English papers are structured carefully and the language is mainly concrete and unpretentious, French papers are claimed to show much greater variety of style. French medical articles often have longer clausés and use more subordinate clauses than similar English texts. This makes the French style more emotive and sometimes even climatic according to Newmark (1976a: 41-46). A translator translating medical texts into English would therefore have to tone the phrases down to more factual style. This view is shared among others by Kourilovâ (1993: 15), who states that English research reports are target-oriented and inductive, whereas the French are deductive. French discourse is also described as being more fragmented than English. Somewhat ethnocentristically, Newmark (1976a: 43) even claims that French style is "showy" and "stylistically most incongruent". According to Kourilová (1993), this difference is also reflected in the theme-rheme projection of the French report as opposed to the rheme-theme structure of the English report. Although Spanish is a major world language, very few studies have dealt with Spanish-English medical translation. The studies that do exist have solely been preoccupied with translation problems at the lexical level. Navarro (1994), a typical example of one of the studies that do exist, points to English as the language of international communication in the medical sciences and reports on the impact of English medical terminology on Spanish. Navarro provides a short glossary of false friends, unacceptable Anglicisms and pharmacological terms not conforming to the International Common Denomination, but the study is generally of little practical use to translators. Major studies of relevance to the translation of medical research articles from Spanish into English are not available. As with Spanish-English, studies of German-English translation of medical texts have mainly been concerned with translation problems at the lexical level, in particular the esoteric terminologies and different designations of diseases, treatment methods and instruments that result from a high degree of medical specialisation (e.g. Grave's disease in English versus Morbus Basedow in German). In a contrastive study, Hilary (1992) discusses the widespread resort to proper names (personal names, toponyms) to designate diseases and illnesses and the difficulties encountered in the translation of such designations from German into American English and British English. The English and German language systems differ significantly. It may be relevant to point to translation differences at the syntactical level. Where the traditional German preponderance for exhaustive treatment of circumstances, considerations, deliberations, afterthoughts and objections is reflected in a highly complex syntax, not least in scientific texts, the English research article is characterised by a simple, fairly straightforward syntax where each sentence

168

TRANSLATION OF MEDICAL RESEARCH ARTICLES

usually carries a single meaning. Translation of German medical research articles into English hence usually requires extensive modification of syntax, as extensively discussed by Newmark (1976b). Other languages A number of other studies discuss medical translation in languages other than English, French, Spanish and German extending beyond lexicon, syntax and style. Thus, a somewhat more exotic, though certainly no less relevant issue in medical translation is raised by Bendix (1988). Although this work has a practical focus on the cross-cultural (English-Chinese) practitioner-patient interaction in the therapeutic situation, it touches upon the important question of cognitive models in medical discourse, more specifically the translatability of metaphor in cross-cultural medical communication from high to low context based individuals (Hall 1959). This may, indeed, be relevant to translation of medical research articles, too. Bendix lends direct support to Benjamin Lee Whorf and Edward Sapir's famous hypothesis that people speaking different languages may be discussing quite different things when they are allegedly talking about the same subject. Bendix thereby also provides support for the Glossematics school (Louis Hjelmslev, Hans Jørgen Uldall and others), who maintained that linguistics may be logically prior to science and that science is merely a branch of linguistics dealing with nature according to linguistically predetermined prejustices (see also Celt 1987). Metaphor is only one of several intricate aspects of translation from Chinese into English, which is also fraught with a whole set of other problems, for example the lack of understanding of the thousands of confusing, unusual technical terms used in Chinese medicine. Problems in determining word meanings and establishing equivalence have for example been discussed in connection with the translation of a major Qung dynasty work on the Chinese materia medica into English, the Bencao Beiyao, first published in 1683 (Zmiewski 1987). A similar though less pervasive lack of understanding may, by analogy, also exist among less disparate cultures, as witnessed for example by the above mention of terminological confusion over preterm labour vocabulary in France, the U.S.A. and the U.K. The above review of some essential studies of translation from major world languages into English invites a number of conclusions. Firstly, only a few studies have examined medical research papers from a translational perspective. Secondly, translation problems are encountered on all levels. Thirdly, there is an apparent lack of theoretical and practical tools available for translators in the field of medicine.

MORTEN PILEGAARD

169

The medical research paper The principles and rules for the production of medical texts are highly genrespecific. For each of the classical medical text genres - case report, treatmentfocused report, research paper (original contribution), disease review, dissertation, medical textbook and speech - we may identify specific characteristics at the lexical, syntactical, stylistic, pragmatic and cognitive levels. Moreover, the interaction between content, form, structure, and medium is close and culturebound. For example, the use of modality varies much from genre to genre. In medical text books, whose purpose is to present facts, (epistemic) modality is not frequently employed. In research articles, however, modality is mostly employed in the Introduction and the Discussion where the author downtones propositions, possible interpretations and conclusions. Moreover, even in a narrowly defined genre such as the research article, the elaborateness of the modality employed varies as a function of cultural value orientations. Hence, there is apparently a propensity in British English medical journal editorials to use modality more extensively than in American journals (Pilegaard, unpublished). Another example of genre specificity is voice: it is often argued that the use of passives in the descriptive sections of the research article is predominant. However, it is rarely mentioned that the use of passive depends on content or topic. Thus, a paper focusing on methodological issues in a particular clinical situation will have many sentences in the active voice because this is better suited for describing procedures than the passive. Such discrepancies are evidently even more pronounced across medical genres. A speech will for example have more sentences in the active voice than a research paper, typically personifying a non-human subject (cells) as in the individual cells do not show the morphologic picture of malignancy or (skeleton and elements) in the following sentence this intermediate skeleton forms a rigid network upon which the contractile elements can exert their focuses. There is a variable degree of genre-specific stereotypicality and the research paper is clearly one of the more uniform medical text genres. Comparing the case report and the research paper, for example, we see that the situational context influences the grammatical and textual structures in different ways. The case report is mainly descriptive and expository, whereas it is the nature of the original contribution to be more "experimental" and rhetorical: it raises questions, verifies hypotheses, gives answers and explains cause-effect relationships. The constitutive features of each genre are unique. The present section of this paper will focus on the research article. Its organisation will be largely hierarchical: it begins at the simplest level, morphology and words, and grows in complexity by widening the focus to syntax and text.

170

TRANSLATION OF MEDICAL RESEARCH ARTICLES

Equivalence at word level Translators are concerned with communicating the overall meaning of stretches of language, and start by decoding the units and structures which carry that meaning: the words. However, in medicine perhaps more than in any other LSP, the smallest element most important to meaning is not the word but the morpheme. In discussing meaning, the paper will centre on 'propositional meaning' and will largely ignore aspects of 'expressive meaning', 'presupposed meaning' and 'evoked meaning' (Cruse 1986). Anyone who comes across compound medical terms for the first time is likely to run into what we may call a 'lexical barrier': the opacity of semantically complex medical words. This problem is exacerbated by the pace of change in the medical sciences and the haste with which the medical vocabulary is being expanded and neologisms proposed by scholars, scientists, technologists, educators and others in the medical community. Still, virtually all the complex words of Latin-Greek origin are built up from simpler, smaller parts. If the meanings of the smaller parts are known, it is possible to deduce the meaning of the complex word itself; for example the word otorhinolaryngology can be broken down as follows: oto means ear, rhino means nose, laryngo means larynx and logo means study. Therefore, otorhinolaryngology is the study of the ears, nose and larynx, which is perfectly understandable to most native speakers of English, because all these words are part of the non-specialised English vocabulary. This also includes larynx which is originally a Greek word meaning the upper part of the windpipe. However, the translation of the English word otorhinolaryngology into another language is not just a simple procedure of finding the meaning of the root and the meaning of the modifying prefixes and suffixes and simply translating these word parts into the other language. For a translation of this complex word from English into Danish to succeed, the word larynx must be translated into the general language of Danish, of which it is not a part. So, Danes meet a 'double barrier': they also have to 'translate' the Greek loan-word larynx into non-LSP Danish, i.e. hals. Another example will show that the lexical barrier may be raised to a level where it is not easily surpassed, particularly not if the task involves a change of tenor, say from a specialist text to a non-specialist text: the compound word pseudohypoparathyroidism is made up of the words pseudo, which means false in the sense a resemblance which is only apparent, not real; hypo which means under, para which means beside', thyroid which refers to the thyroid gland; and ism which means condition or state. Together, para and thyroid refer to the parathyroid glands beside the thyroid gland. So pseudohypoparathyroidism is a condition resembling underactivity of the parathyroid glands. In Danish expert tenor this word only has its Latinate form, pseudohypoparathyreoidismus, and it simply

MORTEN PILEGAARD

171

cannot be translated into a noun or noun phrase in the non-specialist tenor. Its translation would require a full sentence. We may argue that some languages have a double-layered medical vocabulary, i.e. most scientific words and expressions have a popular counterpart (Ilie 1988); 'popular' in the sense that the word has a popular origin or is adopted in medical discourse when the medical professional wants to make himself comprehensible to the layman. When one language can deploy a doublet to cover a medical concept, while the other language only has one lexical item to cover the same concept, usually a cognate of the Latinate half of the doublet, we have what Lankamp (1989: 155) calls a 'register mismatch'. Particularly pervasive examples of problems of 'register mismatch' are described by Chesher (1988), who reports on the achievements of the Health Translations Service of New South Wales, Australia. Another example is Danish and Dutch which can use the doublet blindtarmsbetændelse/ appendicitis (in Danish) and blindedarmontsteking/appendicitis (in Dutch), whereas English only has appendicitis. This mismatch occurs because the English word appendicitis is taken to be both a common and a specialised medical term, whereas appendicitis in Danish and Dutch is used as a specialist medical term. The Danish layman's medical vocabulary, on the other hand, is generally less opaque than that of the English layman. For example the Danish lay expressions for stofskifte (in English metabolism) and gulsot (in English jaundice) are intuitively more readily understandable than the corresponding English layterms. The medical professional is in a diametrically opposite situation: the Danish physician uses far more Greek and Latin terms than does his British colleague. English doctors simply inflexionally "anglify" Greek and Latin roots and fixes. Hence, the distance between the layman's medical vocabulary and that of the medical professional is greater in Denmark than in Britain. Translating from Danish into English thus often forces the Danish translator either to split words into their meaning segments, reduce two or more words to one, or change from one part of speech to another. For example the Danish nouns hepatoblastoma, pneumocyt, spinalanæstesi, and lymforeticulosarkomer become noun groups liver blastoma, lung cell, spinal anaesthesia, and lymph sarcomas and reticulosarcomas when translated into English. The Danish adjective palpatorisk becomes an English prepositional phrase: when palpated. In absence of a 1:1 equivalence between SL and TL, we will often have to analyse the semantic contents of the SL word. Thus, for example the Danish adverbs brutalt (in a negative context) and mager must be translated into two parallel adverbs in English suddenly and sharply and thin and wasted, respectively, and the Danish verb at styrte must be rendered by a verb and a qualifying adverb, to fall violently, in English. It should also be noted that the differences between general and specific language go beyond lexis where the jargon of a particular speciality may have a meaning that escapes the layman and even influences

172

TRANSLATION OF MEDICAL RESEARCH ARTICLES

grammatical usage. Statistical analyses of the patients' understanding of medical terms and diseases have concentrated more on patients' knowledge of the aetiology, treatment, and prognosis of certain diseases than on their ability to agree with the majority of doctors over the more basic issues of definition. Boyle (1970) reports significant differences between patients' and doctors' interpretation of some common medical terms. He also rightly states that the educational level of the patient must be considered, as relationships between vocabulary performance and educational background must be expected. The implication of this and other studies is that translation across levels of technicality must cater for the large areas of misunderstanding between conventional medical opinion and the vagaries of the lay mind. This applies across cultures but also within cultures where patients are exposed to mental models based on abstraction processes that emphasise the patterns or configurations coded within the semantic domain of the physician. The patient finds it difficult to follow the physician's language and often tacitly or even explicitly specified symbolic recording. "The patient's literacy or rationality, even if he or she is highly educated, is no match for the physician's language and external memory system, a system that is in constant flux if the doctor is to be able to keep up with new developments in medicine" (Cicourel 1981: 84). Cicourel, as one of the leading American ethnomethodologists, has done extensive studies of doctor-patient discourse with particular emphasis on speech acts, but this data is only of indirect relevance to the present paper. The transparency/opacity of medical words, i.e. how much we understand of the meaning and structure of our own language, largely depends on the language family. In some Romance languages and Latin, what is abstract conceptualisation in English and more so in Danish is understandable to the layman in that language (e.g. the Latin ad renem is closer to Spanish rinòn and French rein than English adrenal gland, Danish nyre and Chinese shenshangzian (kidney-atop-gland). Translation of medical research papers rarely involves a change of tenor and problems with the 'register mismatch' and the 'lexical barrier' are accordingly not particularly pertinent in this context. A more intriguing kind of non-equivalence at word level arises where concepts are culture-specific or where the SL is not lexicalised in the TL. Thus, as mentioned above, Alexander (1987) notices that even if the English term impending preterm labour is used in centres in the U.S.A. where active management of the preterm labour is sought, no such term exists in the U.K. Hence, we meet equivalence problems even among countries which are, culturally speaking, very close. This problem is evidently greatly exacerbated when the language pairs in question are further apart. Again, we may use translation of English into Chinese as an illustrative example. Celt (1987: 20) says about translation from English into Chinese that "what it loses in terms of pure

MORTEN PILEGAARD

173

abstract conceptualising it gains in everyday understanding". According to Celt, the major difficulty in translating into and from Chinese lies in our own ethnocentricity and our expectations. We expect medical language to sound Greco-Latinate and medical. But Chinese medical language and thought, beyond the problems posed by the Chinese four-character summaries (or aphorisms) often used to describe medical conditions, are much more geometrical, mathematical and even legalistic in feel. Celt gives, as an example of the "everyday understanding" of these terms, the words for the two bones we have in our lower arm. The only names we have for them in English today are ulna and radius. These are the scientific Latinate names. These two bones are important to us, but we have no everyday terms for referring to them in English, although there is clear evidence from historical linguistics that these bones once had other names. The ulna was once called the el, the radius either spike or spoke. It is interesting to note, though, that modern German does have names for these two bones: die Elle and die Speiche. In modern English the place where the "el" makes a bend or "bow" is called the elbow (Ellenbogen in German, albue in Danish). In Chinese, these words translate as "foot-measure bone" (close to the meaning of "el") and "rowing bone". It is interesting that all bones and all locations in the body have similar down-to-earth names in Chinese, whereas in Western societies many terms are linguistically off-limits, except to doctors. In medical translation from one language into another as disparate as English and Chinese, we may, indeed, have to account for the fact that a system is not a system and a word is not a word. Following Benjamin Whorf quoted in Celt (1987): "there is no Chinese word for word' and in Chinese a "character can be anything from a piece of a word, sometimes similar to a prefix or suffix, to an unbound particle, to a fully unmistakable word on its own, to a free-standing abbreviation for a two-character phrase, which might be either one word or two words in English, or even an abbreviation for a fourcharacter construction" (Celt 1987: 20). We are thus able to firmly repudiate Fischbach's claim, quoted above, that the medical translator has "an edge over his colleagues in other fields" (Fischbach 1986: 19). The lexicon of English presents special problems to non-native speakers due to the vast quantity of Latin and Greek loan words which are, especially in the case of post-Renaissance borrowings, not thoroughly assimilated into the fabric of the language. Students or learners of other languages meet the 'lexical barrier' both in respect of conventional Latin-Greek terms and conventional intraEnglish or native metaphors. For Romance-speaking learners, conventional English native metaphor readily constitutes a lexical barrier. However, nonRomance-speaking learners would seem to face a double lexical barrier owing to the opacity, for them, of the multitude of conventional Latin-Greek and native conventional metaphors. This problem is greatly exacerbated when we

174

TRANSLATION OF MEDICAL RESEARCH ARTICLES

deal with medical texts because something like three-quarters of the terms used in Anglo-Saxon medical language and its associated disciplines are of Greek and Latin origin. As these languages are no longer widely studied, newcomers to medical and paramedical subjects often find that they have to cope with a whole new vocabulary as well as learn the unfamiliar concepts of their subjects. The ubiquity of metaphor in medical texts makes this problem no less serious: an example is the metaphorical description of the body as a container (Lakoff 1987: 383). Furthermore, as the meaning we attribute to medical terms and our use of metaphorical versus literal terms is closely tied to our cultural tradition (Bendix 1988), translation of medical texts across widely different cultures may, indeed, be a particularly complicated task. Medical terminology owes another part of its opacity to the widespread use of eponyms. Following Hoof (1959: 421), any translator of medical texts into English should avoid using unfamiliar eponyms because, in medicine, eponyms play a particularly tricky role. For example, the translator will usually not meet any difficulties translating Parkinson's disease into French la maladie de Parkinson or Hodgkin's disease into la maladie de Hodgkin. But, although Down s disease is currently the preferred term for mongolism in the U.K., the French are not familiar with la maladie de Down but prefer the term le mongolisme (Covington 1983). Moreover, some eponyms bear features lacking in others. One such feature is that over time some eponyms have lost their possessive form whereas others have not, as discussed by Mühlhaus (1995: 51). So where the translation of eponyms is sometimes straightforward, this is certainly not the rule. Medical language probably offers the richest proliferation of synonyms in any technical language. Innumerable concepts are synonyms or near synonyms but differ according to level, be it anatomic, clinical, pathogenetic, historic or geographic. In the translation process, it is important to note that synonyms cannot always be translated into other languages. Thus in French polyarthrite déformante, which by some is still being referred to as rheumatisme chronique déformant, is, today, more properly termed polyarthrite chronique évolutive. Its equivalent in English is rheumatoid arthritis, which in turn has the following synonyms: atrophic arthritis, arthritis deformans, chronic infectious arthritis and proliferative arthritis. The delicate task for the translator, then, is to pick the synonym preferred by the likely readers of his translation. This choice can only be made if the translator regularly consults medical journals in both languages. Where the Anglo-Saxons mainly use rheumatic fever and not its synonyms acute articular rheumatism and polyarthritis rheumatica, the French seem to prefer the term rheumatisme articulaire aigu and perhaps even more frequently maladie de Bouillaud, although they do have the terms fièvre rheumatismale and polyarthrite aiguë fébrile. A more detailed explanation of

MORTEN PILEGAARD

175

the proliferation of synonyms and the difficulties in translating these from French into English is given by Hoof (1959) and Newmark (1976a, 1976b). It should also be borne in mind that finer shades of meaning are not necessarily expressed in the same way in the SL and the TL, a fact which is particularly important in those sections of a research paper where the author evaluates his conclusions. For example, in Danish the word let is an adjective that indicates smallness or insignificance in physical contexts. It may be used to modify both smerte (in English pain) and skade (in English [physical) damage). Thus, the Danish en let smerte should correspond to a mild pain in English. However, in English physical damage is usually not qualified by the adjective mild (corresponding to Danish let) but by negated severe. Translation of the Danish absolute comparative en lettere skade into English a milder damage would be unidiomatic (and ungrammatical). The correct translation is a less severe damage. Another similar example is that imperfectly understood in English is equivalent to forholdsvis uerkendt in Danish: forholdsvis (+) uerkendt (_) becomes imperfectly (_) understood (+). In other words, collocational restrictions may involve a shift during translation between positive (+) and negative (_) meaning. The pursuit of equivalence at the word level must hence be most cautious. For all practical purposes translation of specialist terms and phrases unfamiliar to the translator should be conducted within the framework of what Gile calls "équipes bidisciplinaires" (i.e. interdisciplinary cooperation) (Gile 1986) between professional translators and specialist physicians. This advice finds strong support among LSP metalexicographers (Schaeder and Bergenholtz 1994) and has led to recent practices in bilingual lexicography as discussed in Pilegaard (1994a). Pragmatic equivalence A few preliminary observations are needed before discussing translation above the word level. Firstly, any translator is judged by his ability to produce clear, comprehensible translations and should take an active role in trying to ensure that the text is clear, readable English. Often, the first step in a translation process is to "translate" the SL text into a clear, coherent and cohesive SL text. The importance of this is underlined by the fact that health-care professionals often write about their speciality in a language which is filled with their own specialist terminology and do not always realise that it is, in fact, jargon which may not be understood by those who are not working in the speciality in question. Secondly, health-care professionals who have developed expertise in a medical speciality rarely have a parallel expertise in the art of writing. Moreover, in their effort to be scientifically accurate, they often make convoluted and almost incomprehensible sentences (Chesher 1988: 3). This is, of course, a

176

TRANSLATION OF MEDICAL RESEARCH ARTICLES

generalisation that cannot be expected always to hold. The German rhetorical style in research papers, for example, is more convoluted than the English style, as discussed above. In the case of language revision, i.e. where texts are written in the TL by non-native health-care specialists who may have some foreign language competence at the lexical level in their particular speciality but rarely master the syntax or the pragmatics of the TL, it is even more important that the translator or language revisor points out ambiguities and devotes his first effort to improving the comprehensibility of the text. In practice, the borderline between translation and language revision is not always entirely clear. The question then is, when should a SL text be revised before translation is undertaken? As a rule of thumb we may apply the "native speaker reaction" (Chesher 1988: 41) which implies that if the translator as a competent ignorant is having difficulty understanding the information in his own first language, then it is reasonable to assume that he/she will also run into serious problems while translating the text. This problem may partly be overcome by recourse to encyclopedia and dictionaries, but there is a serious risk that these works may lead the translator off on a false trail, particularly in highly specialised domains where proper lexicographic tools are rare. Secondly, styles of argumentation are, among others, reflected at the syntactical level of the research article, viz. the above discussion of differences between English on the one hand and French and German on the other hand. At the argumentative deep level, the persuasive power of each sentence depends, for example, on elements such as the use of passive or active verb forms, the positioning of adverbs and adjectives, the choice of first person or third person narrative, and reference patterns. Convincing arguments for the introduction of a clinical hermeneutic approach in peer reviews are given by Horton (1995). The question of striking a balance between the maximally objective presentation of the real world and the subjective evaluation of the facts and findings presented is a delicate one. The use of epistemic modality, for example, is influenced by the nature of the topic, the writer's intentions, the degree of subjective certainty and the prospective reader (Kourilová 1993: 5). Linguists whose main interest is LSP unfortunately tend to concentrate on the commonest errors of lexis and syntax (Webber 1993). There is an evident scarcity of studies providing insight into the problems of conveying signals of claims and mitigated claims, denials of claims and many other pragmatic facets of scientific communication and interactions between writers and readers typically demonstrated by non-native writers of scientific articles. This is a serious problem, as rightly pointed out by Kourilovâ (1993). Thirdly, the recurrent stumbling blocks for (competent) translators are the differences between the languages: that whereas both anaphoric and cataphoric patterns occur in Danish and English and are equally acceptable in English, the

MORTEN PILEGAARD

177

latter is only marginally possible in Danish; or that time indications often fill the subject slot in English, which they rarely do in Danish, or that English makes greater use of non-finite forms of the verb than does Danish, etc. etc. A more interesting problem is culturally determined usage of modality in the text genre in question. This issue has recently received some attention (Myers 1989, Swales 1990, Prince 1992, Kourilová 1993). In general, it has been established that modality, broadly conceived, varies with genre. In textbooks, which supposedly present established facts and findings, the level of modality is lower than in research articles, in which it is most apparent in the Introduction and Discussion sections, where the author identifies with or distances himself from certain positions, interpretations or conclusions. In the classical IMRAD (Introduction, Material and Method, Results and Discussion) structure of the research article, each section has a distinct communicative function whose communicative purpose is served by different linguistic means, including modality. Thus, the Methods and Results sections are highly factual and modality is low, a finding supported by Kourilová (1993: 8-9), who reports a highly significant difference between the use of epistemic modality in the reporting and commenting sections from a 50,000-word corpus of medical research articles. However, whereas some attention has been drawn to the fact that modality is a basic feature of the research article in general and that its use varies with the macrostructure, the transfer of SL modality norms into TL modality during translation remains a virtually unexplored domain. It appears, though, as if traditional, culture-bound conventions for scientific reporting are often transferred from the SL into the TL by non-native speaker science writers and physicians writing in the TL. Even highly educated Danish physicians rarely use the great repertoire of modality forms and structures available in the English language when reporting to international journals, and they often run into serious problems when trying to project subtle shades of truth value in English. A related problem is hedging. Hedging is one of the areas where the pursuit of dynamic equivalence between the SL and the TL makes it necessary for the Danish translator translating into English to be slightly "illoyal" to the SL text. The Danish physician generally uses fewer hedges than his English colleague. Part of the translator's adaptation of the text to TL norms consists in adding hedges because appropriate hedging is important for the way in which the message of the research article is received. Failure to use conventional hedging sends the signal that what is being communicated is, indeed, not the author's opinion of a relevant topic, his presentation of a new discovery, etc. Even Nobel Prize winners hedge their statements. Watson and Crick, who discovered the DNA structure, introduced the paragraph describing their new discovery with the humble words: We wish to suggest. In the Discussion section of a research paper, claims and conclusions are al-

178

TRANSLATION OF MEDICAL RESEARCH ARTICLES

ways presented as preliminary. This is done, for example, by means of impersonal constructions, the results suggest, which are often hedged into the results seem to suggest. Proper translation of an unhedged Danish sentence like Undersøgelser blandt læger viser, at der er op mod 22.600 alkoholiserede læger i U.S.A. (unhedged Danish verb viser, in English shows) requires that the corresponding English verb be hedged Surveys of physicians suggest that there exists up to 22,600 alcoholic physicians in the U.S.A. (hedged verb suggest). The addition of hedging in the presentation of results in the TL will often interact with other syntactical shifts in the translation process as for example in Tidligere undersøgelser viser, at indvirkningen på cellens cyklus er ... which in English becomes Earlier studies indicate a cell cycle dependent effect of ... In this translation the Danish unhedged verb (viser, in English shows) and infinitive sentence (viser, at indvirkningen er ...) is translated into a hedged noun phrase in English (indicate a .... effect). Addition of hedging is even more important where the SL culture differs from the TL culture in terms of conventional speech act realisation. The translator working from Danish into English will generally have to modalise critique to match the TL form to its readers' culturally determined expectations, e.g. by using impersonal constructions and passives as in The increase was claimed to be induced by the release of... (Smith et al. 1993). Another case in point is that Danish physicians rarely master the combination of declarative verbs indicating probability with modal verbs in English; a problem which is particularly pertinent given the importance of hedging and modality for proper TL decoding as just mentioned. A rough outline of the most frequent declarative and modal verbs in reporting sections of English research papers may look as follows: Illocutionary force

Weak

Strong

Declarative verb

Modal verb

suggest indicate show demonstrate

might may can

Danish physicians almost consistently use declarative verbs from the 'strong' end of the continuum and occasionally combine these verbs with modal verbs also from the 'strong' end. Inversely, English science writers never choose declarative and modal verbs that lie at the same end of the continuum. They in-

MORTEN PILEGAARD

179

variably use combinations, e.g. may/might show/demonstrate and can/may suggest/indicate. Translation of unmitigated Danish verbs or verb groups into the TL would then not only lead the reader off on a false interpretation track, but would also result in an incorrect or infelicitous translation. The translator of medical research articles must therefore obey the norms of the genre and must tailor the TL to target culture expectations of form and usage. Medical English belongs to the fields of discourse called 'technical English' or 'scientific English' and we may accordingly be inclined to think of medical English as a highly factual and objective kind of writing. However, its presentation conforms to a large extent to the classical ideas of rhetorical presentation, and in the words of Horton (1995: 987) "the text describes a specific path, carefully carved by the authors, through a complex undergrowth of competing arguments". The IMRAD structure of the research paper is in close agreement with Aristotle's four elements of successful oratory: introduction, narration, proof, and epilogue. The research paper is usually a most carefully structured and highly idiomatic text that serves a rhetorical purpose at several levels. The rhetorical purpose is served at the deep level by lexical and morpho-syntactical structures, i.e. the choice of active or passive forms, the positioning of adverbs and adjectives, the choice of first person or third person narrative, and the referencing pattern, among others. At the surface level, argumentation is concerned with the limitations and quality of the methods and the material, the theoretical and/or clinical importance of the work, reflections on the meaning of the results, claims about the success of the original purpose of the study, and, usually, directions for future research (Horton 1995). The individual bits of text are normally linked together in various more or less effective and explicit ways, and before the text is translated, it should be subjected to a close reading which should bring about a full realisation of the way the text has been put together. In the process of translation, the translator must attempt to strike a balance between the highly conventionalised demands of the genre and the particular style demands of the scientific journal in question. It is evident that there are many situations where rearrangement or substitution is required if the result is to be idiomatic and pragmatic equivalence is to be reached. However, it goes beyond the scope of the present paper to discuss thoroughly how such rearrangements and/or substitutions can be made. It suffices to note that in any medical translation the process of establishing proper pragmatic equivalence should be performed in a close cooperation with the ST author in order not to upset or twist his argument.

180

TRANSLATION OF MEDICAL RESEARCH ARTICLES

Conclusion It is concluded, firstly, that in medical translation equivalence must be sought at both lexico-syntactic and textual levels. Secondly, lexico-syntactic and pragmatic modifications must be guided by genre and culture specific conventions at the levels of lexicon, word classes, verbal categories, syntactic functions, modality, relationships between sentences, and polarisation. Thirdly, the balance between the conventions of the international research paper and the idiosyncrasies of the ST author must be struck in a close collaboration between the translator and the author. Finally, there is a great need for specialised translation tools and for further studies of translation of medical texts. References Alexander, Sophie and Slater Catherine. 1987. "Labouring under linguistic delusions: The impact of linguistic factors in international studies of preterm labour". Language & Communication: An Interdisciplinary Journal 7(2): 179-185. Applewhite, L. 1979. "Examination of the medical/scientific manuscript". Journal of Technical Writing and Communication 9: 17-25. Atkinson, Dwight. 1992. "The evolution of medical research writing from 1735 to 1985: The case of the Edinburgh Medical Journal". Applied Linguistics 13, 4 Dec: 337-374. Barkman, Bruce. 1974. "The translation of SNOP: A first step toward the construction of an automated medical lexicon". Meta 19: 28-42. Bendix, Edward H. 1988. "Metaphorical and literal interpretations: Cross-Cultural communication in medical settings". CUNYForum: Papers in Linguistics 13: 1-16. Bork, Albert G. 1992. "Translating medical records from Spanish". InFrontiers: Proceedings of the 33rd Annual Conference of the American Translators Association, Medford, NJ: vii. Boyle, CM. 1970. "Differences between patients' and doctors' interpretations of some common medical terms". Brit. Med. Journal 2: 286-290. Brown, Charles Barret. 1942. The Contribution of Greek to English with special attention to medical and other scientific terms. Nashville. XIII: vii-ix, 1-5, 56-57, 144-145, 204-205, 254-255. Brunt, R.J. 1987. "My waterworks are playing me up something chronic". In Lörscher, Wolfgang and Schulze, Rainer (Hrsg.), Perspectives on Language in Performance. Studies in Linguistics. Literary Criticism, and Language Teaching and Learning. To Honour Werner Hüllen on his Sixtieth Birthday. Tübingen. 447-469. Celt, Sandra. 1987. "The challenge of translating Chinese medicine". Language Monthly: The International Journal for Language and Translation. April: 43: 1921.

MORTEN PILEGAARD

181

Chesher, Theresa G. 1988. "How to keep healthy in 17 languages: Translating and interpreting in New South Wales Health Care Delivery". Aral 11:1: 34-46. Cicourel, A.W. 1981. "Language and the structure of belief in medical communication". Studia Linguistica 35/1-2: 71-85. Coleman, Hilary. 1992. "Probleme und Freuden der medizinischen Fachübersetzung". Hieronymus 3: 7-12. Cruse, D.A. 1986. "Lexical semantics". Cambridge: Cambridge University Press. Dressler, Stephan and Schaeder, Burkhard. 1994. "Wörterbücher der Medizin. Beiträge zur Fachlexikographie". Lexicographica. Series Maior 55, Niemeyer. Eyraud, Daniel. 1974. "Bilan d'une Decennie". Meta 19: 13-27. Fischbach, Henry. 1986. "Some anatomical and physiological aspects of medical translation, lexical equivalence, ubiquitous references and universality of subject minimise misunderstanding and maximise transfer of meaning". Meta 31: 16-21. Fischbach, Henry. 1993. "Translation, the great pollinator of science: A brief flashback on medical translation". In S.E. Wright and L. Wright (eds), Scientific and Technical Translation. Amsterdam: Benjamins: 89-100. Gabrieli, E.R. 1986. "Construction of a biomedical nomenclature". Meta 31: 22. Gambier, Y. 1989. "Introduire ... ou il n'y a que le premier pas qui coute!". In C. Lauren and M. Nordman (eds), From Office to School: Special Language Internationalisation. Clevedon: Multiling. Matters: viii, 169pp. Garciahidalgo I. and Dunham, G. 1981. "An experiment in English-Spanish Automated translation of medical language data". Methods of Information in Medicine 20:1:38-46. Giebson, Edith. 1990. "Zur sprachlichen Realiserung der Handlungsstruktur ausgewahlter medizinischer Fachtexte: Deutsch-russischer Vergleich ..." In Gunter Frohne and Olga Pavlovna Ermakova (eds), Ausgewahlte Beitrage zur Kommunikationslinguistik. Potsdam: 195-203. Gile, D. 1986. "La traduction médicale doit-elle être réservée aux seuls traducteursmédicins? Quelques réflexions". Meta 31: 26-30. Gould, Clinton. 1990. "Beyond disease: A conceptual analysis of the language and origins of the meaning of AIDS in a sociohistoric and medical context". Dissertation Abstracts International Nov:51(5): 1537A. Graitson, M. 1975. Identification et transformation automatiques des morphèmes terminaux dans le lexique médical français. (Cahiers de Lexicologie 26):1: 85-109. Hall, Edward T. 1959. "The Silent Language". Doubleday. Hankins, Freda Richards. 1992. "Leechbook". In Dissertation Abstracts International, vol 53/07/-A: 2363. Hinkkanen, S. 1989. "Translation and author's editing of biomedical communications written by Finnish authors". In C. Laurén and M. Nordman (eds), Special Language: From Humans Thinking to Thinking Machines. Clevedon, Philadelphia: Multilingua Matters Ltd: 225-233. Holden, S. (Hrsg). 1977. "English for Specific Purposes". In J. Allwright and R. Allwright (eds), An Approach to the Teaching of Medical English. London. 56-62.

182

TRANSLATION OF MEDICAL RESEARCH ARTICLES

Hoof, H. van. 1959. "Reflexions sur le langage médical. Le point de vue du traducteur scientifique". La Presse Médicale 67: 419-421. Hoof, H. van. 1970. "La traduction medico-pharmaceutique". Langage et l'omme 12. 65-77. Hoof, H. van. 1974. "Bibliographie de la traduction médicale". Meta 19:l:Mar: 43-66. Hoof, H. van. 1993. "Historie de la traduction medicale en occident". Cashiers de l'nstitut de Linguistique de Louvain v 19(1-2): 75-125. Horton, Richard. 1995. 'The rhetoric of research". British Medical Journal 310: 985986 Ilie, Cornelia. 1988. "Syntactic and semantic equivalence in translating Swedish medical texts into Romanian". In N.C. Chaffey, N, A.F. Ryding and S.S. Ulriksen (eds), Translation Theory in Scandinavia: Stockholm. Jacopoosie, Peter. 1993. "Medical interpreting in Canada's North". Meta 38: 42-44. Jammal, Amal, Leblanc, Louise, Proulx, Mylene, Rinfret, Aline. 1993. "Si l'immunité m'était contée... Introduction a une étude des termes de l'immunologie". Meta 38: 502-516. Jasin, Joanne. 1983. "A critical edition of the middle English liber uricrisiarum in welcome MS 225". In Dissertation Abstracts International, Vol. 44/09-A: 2761. Kaufmann, James Michael. 1989. "The rhetoric of medical writing: Case studies of physicians writing for journal publication". Dissertation Abstracts International Aug: 50(2): 427A. Kourilová, Magda. 1993. "Epistemic modality in written scientific discourse". Meta 38:4-18. Langslow, David. 1989. "Latin technical language: Synonyms and Greek words in Latin medical terminology". Transactions of the Philological Society 87(1): 33-53. Lankamp, Robert Eduard. 1989. A Study of the Effect of Terminology on L2 Reading Comprehension. Amsterdam: Rodopi. Lurquin, Georges. 1985. "La Page du terminologue". La Langage et l'Homme 20:1(57):46-49. M. Covington. 1983. "The translation of eponyms". The incorporated Linguist 22: 198-199. Maher, J. 1986. "The development of English as an international language of medicine". Applied Linguistics 7: 206-218. Manuila, A. (ed). 1981. Progress in Medical Terminology. Switzerland: S. Karger. McConchie, R.W. 1988. It Hurteth Memorie and Hindreth Learning: Attitudes to the use of the vernacular in sixteenth century English medical writings (Studia Anglica Posnaniensia): 21, 53-67. Moreau, A. 1986. "Traduction et terminologie médicales". META. Journal des Traducteurs. Organe d'Information et de Recherche dans les Domaines de la Traduction et de l'Interpretation 1: 98-105. Mühlhaus, Susanne. 1995. "Describing medical eponyms". English Today 42:11,2: 4853. Navarro, FA., Hernandez, F. 1994. "Glosario de "falsos amigos" y palabras de traducción enganosa en el inglés de los textos médicos". Lebende Sprachen 39.1: 24-28.

MORTEN PILEGAARD

183

Naylor, RB. 1989. "The Message and the medium: Exploring their relationship in oral medical presentations". In C. Lauren and M. Nordman (eds), Special Language: From Humans Thinking to Thinking Machines. Clevedon, Philadelphia: Multilingua Matters Ltd: 190-201. Newmark, P. 1976a. "A layman's approach to medical translation. Part 1". The Incorporated Linguist 15: 41-47. Newmark, P. 1976b. "A layman's approach to medical translation. Part 2". The Incorporated Linguist 15: 63-68. Newmark, Peter. 1992. "Paragraphs on translation". The Linguist 31.4: 114-119. Nwogu, K.N. 1990. "Discourse variation in medical texts: Schema, theme and cohesion in professional and journalistic accounts". Dissertation Abstracts International. Winter. v51(4): item2060C. Pendleton, D. 1982. "Doctor-patient communication: A review". In D. Pendleton and J. Hasler (eds.), Doctor-Patient Communication. London, 5-53. Pilegaard, Morten. 1994. "Bilingual LSP dictionaries. User benefit correlates with elaborateness of 'explanation'". In B. Schaeder and H. Bergenholtz (eds), Fachlexikographie. Fachwissen und seine Repräsentation in Wörterbüchern pp 211228. Pilegaard, Morten and Baden, Helge. 1994. Dansk-engelsk, engelsk-dansk medicinsk ordbog. Gad, København. Reeves, C A . 1990. "The characterization of a medical problem: An analysis of the writing on AIDS in medical science". Dissertation Abstracts International March: 50(9): 2875A-2876A. Regent, O. 1985. "A comparative approach to the learning of specialized written discourse". In Philip Rile, and Christopher N. Candlin (eds), Discourse and Learning. Papers in Applied Linguistics and Language Learning. London. Rouleau, Maurice. 1993a. "La voix passive dans les textes médicaux et para-médicaux". Meta 38: 440-448. Rouleau, Maurice. 1993b. "Des traquenards de la version médicale. I. Action, effect, potency et effectiveness". Meta 38: 268-274. Salager-Meyer, F. 1986. "Infinitive clauses in medical English literature: A rhetoricogrammatical approach". Estudios de Linguistica Aplicada. Jan. 4(5): 66-86. Salager-Meyer, F. Defives, G. Jensen C.and M. de Filipis 1989. "Communicative function and grammatical variations in medical English scholary papers: A genre analysis study". In C. Lauren and M. Nordman (eds), Special Language: From Humans Thinking to Thinking Machines. Clevedon, Philadelphia: Multilingua Matters Ltd: 151-160. Salager-Meyer, Francoise. 1991. "A genre-based and text-type analysis of hedging in written medical English discourse". Interface: Journal of applied linguistics 6:1: 33-54. Salager-Meyer, Francoise. 1994. "Hedges and textual communicative function in medical English written discourse". English for Specific Purposes 13.2: 149-70. Sammons, Susan. 1993. "Challenges in minority language programming in Canada's Eastern Arctic: The training of Aboriginal language interpreter-translators". Meta XXXVIII: 1:45-50.

184

TRANSLATION OF MEDICAL RESEARCH ARTICLES

Schaeder, Burkhard and Bergenholtz, Henning. 1994. Fachlexicographie. Fachwissen und seine Repräsentation in Wörterbüchern. Gunter Narr Verlag Tübingen. Sliosberg, A. "A propos de la traduction médicale et pharmaceutique". Babel 23: 107115. Tropeau, Gerard. 1986. "Les problèmes posés par la traduction de l'arabe médical ancien en français moderne". Meta 31:1: 11-15. Valiquette, Michele. 1991. "Lexique anglais-français de neurochirurgie". Meta 36:4: 633-644. Weaver, Charlotte Ann. 1982. "Role evolution of language translators in a major medical center". Dissertation Abstracts International vol. 43/02-A: 494. Yanoff, K.L. 1989. "The rhetoric of medical discourse: An analysis of the major genres". Dissertation Abstracts International April 49(10): 3011 A. Ylönen, S., and Neuendorff, D. and Effe, G. 1989. "Zur kontrastiven Analyse von medizinischen Fachtexten. Eine diachrone Studie". In C. Laurén and M. Nordman (eds), Special Language: From Humans Thinking to Thinking Machines. Clevedon, Philadelphia: Multilingua Matters Ltd: 203. Zmiewski, Paul David. 1987. "The concise pharmacopeia: An abridged translation with a systematic method for rendering Chinese medical terms". Dissertation Abstracts International, vol. 49/06-B: 111

Translation of Technical Brochures Jens Hare Hansen The University of Aalborg

Introduction Through the last few decades translation studies have developed considerably. The theory of translation has, like the research of language for specific purposes,1 obtained rich inspiration from many disciplines, not least linguistics. For a number of years attention was specifically concentrated on questions concerning structural linguistics. This has had great impact on the conception of equivalence which has played a prominent part in many works about a theory of translation.2 Later linguistics and theory of translation have developed into a direction in which one works with language in a social context. The most recent development of translation studies is turning to the cognitive processes which take place in the act of translating. In the field of translation, new scientific advances have received mixed reviews concerning their importance and relevance. Among theorists this development is considered as a positive trend, while practising translators have a pronounced scepticism concerning the use of these theories. Trosborg stresses that practising translators "usually tend to be sceptical of any kind of theorizing" (1994: 9). The general scepticism of theory of translation among practising translators is shared by many teachers at universities and other educational establishments.3 Translation studies thus needs some sort of justification. In the following I shall show that translation studies can definitely contribute to the improvement of concrete translations. This concerns the translation of the LSP text type technical brochures.4 In opposition to Werner Koller (1992: 21ff.) I state that there is reason to assert that practicians need translation studies, and that you can actually improve translations if you command a methodological concept.

186

TRANSLATION OF TECHNICAL BROCHURES

This article is based on Danish and German brochures and on the problems that arise when you translate between these languages. However, the results are also very relevant to translations between other languages. In this article attention is concentrated on pragmatic aspects, while semantic and syntactical questions should be referred to the very detailed literature on the subject.

LSP and translation In LSP-research it has long been an acknowledged point of view that the purpose of LSP was to describe reality, and that LSP has an informative language function (cf. Hoffmann 1985). In accordance with this, the point of view was reached that LSP-translation primarily is a question of creating a denotative equivalence. The above mentioned view of LSP is not least due to the fact that the work primarily concerned technical-scientific texts, while other forms of LSP were largely ignored. Until the beginning of the eighties the concentration was at the same time primarily concerned with technical terms and, to a lesser degree, with the syntax of LSP, while pragmatic and epistemic conditions were generally not examined (cf. Göpferich 1995: 1). Since then pragmatics and text linguistics have gained ground in the research of LSP, which has had a considerable influence on translation studies. This has resulted in the fact that LSP-translation is not now just considered a matter of achieving denotative equivalence. This new insight into LSP research implies that scientific attention has now also been stressed on the functional aspects of text. There is a pronounced increase in awareness of the fact that translation of different LSP text types demand different kinds of translation strategies. It is still a widespread view that technical translation concerns a precise description of the objective world in the target-language.5 This has caused technical language to often be described as a form of communication where the technical content is most important, and where the sender has merely a limited concern for the receiver (cf. Fluck 1984). Stolze, for example, quotes Zeumer's impression that the sender does not think of the receiver of the message (1992: 79). In my opinion, however, such a view does not stand a closer examination, especially in the case of the text type of technical brochures, where the consideration for the receiver has a very important role.

JENS HARE HANSEN

187

Technical brochures seen from linguistics and marketing theory As already mentioned, technical brochures are among the least examined text types. This is the case from a linguistical as well as from a marketing-related view. The German marketing theorist Axel Bänsch writes about informative advertising - of which the brochures are a part - that it seeks to win the customer for the product by help of factual information. According to Bänsch informative advertising is often just giving pure facts. And yet such form of advertising can also hold suggestive elements (1991: 215). Berhard Sowinski, the linguist, arrives at a similar result from another basis, and writes that informative advertising contains a technical description of the product and an account of its factual qualities. This makes the message of the advertising seem reliable, and the producer more trustworthy. Sowinski writes that often the advertisement only has its objective character on the surface. He does not explain more concerning such advertisements' language. According to Katharina Reiss, certain types of advertising can be so factual that they, in her classification, must be characterized as informative. This applies to the sort of advertising that does not appeal emotionally, but factually to the receiver (1983: 57f.). In contrast to Reiss, Göpferich (1992) suggests that all advertising is operative, because it serves an economic purpose. It is typical of technical advertisements, Göpferich asserts, that they contain "operative elements". She omits, however, to explain more precisely what she understands by this concept. When it comes to translation it would certainly be an advantage if it was possible to identify the elements that Göpferich and others mention, whether they use the term operative elements or something else.6 There is a general agreement among translation scholars that it is necessary for translation to analyse the source language text (cf. Nord 1991). However, there is disagreement about what is relevant in that sort of analysis. In the following chapter I shall concentrate on selected aspects which are especially important to the translation of technical brochures. Speech acts and text functions in technical brochures It is to be supposed that texts are produced with the purpose of achieving a communicative purpose (cf. Hatim and Mason 1993). To use language, in other words, is to act. The question is: How can these aspects best be described and analysed? Here there are clear divergencies in the different concepts. For lack of space I shall refrain from discussing the subject in detail, instead I refer to Heinemann and Viehweger (1991: 50ff).

188

TRANSLATION OF TECHNICAL BROCHURES

Speech act theory seems to me to be suited to explain important aspects of action, but because of the enormous complexity of the subject the theory is not unproblematic, and it would be necessary first to discuss parts of that. This article is far too short to hope for such comprehensive discussion. From the above, it follows that representatives and directives are especially of interest in an analysis of technical brochures. Various attempts have been made to classify speech acts. The most famous classification - and according to Brinker (1992: 100) also the best - has been proposed by Searle. His classification is here to be used as the starting point of a discussion of illocutionary aspects in technical brochures. Searle defines representatives as follows: The point or purpose of the members of the representative class is to commit the speaker (in varying degrees) to something's being the case, to the truth of the expressed proposition. All of the members of the representative class are assessable on the dimension of assessment which includes true or false (1976: 10). In technical brochures there are many examples of this type of speech act. An example from a brochure: Die 500er verfügen über 4 Zapfwellengeschwindigkeiten mit 540/750/1000/1400 U/min (= The 500 C range tractor is fitted with a 4-speed PTO of 540/750/1000 and 1400 rpm). This statement can easily be proved as regards truth or falseness. The same thing goes for the following sentence: Ein Druckbehälter mit 15 Liter Inhalt speichert die Druckluft. (= Air is stored in a 15-litre capacity tank). Statements as the ones here quoted correspond to what has traditionally been understood by LSP. Technical language especially has often been described as a usage characterized by extensive usage of representative speech acts. Add to this a pronounced use of terminology such as Zapfwellengeschwindigkeit, Druckbehälter and Druckluft. Furthermore, 3rd person is used.7 It turns out, however, that it is not always possible to discern representatives from other types of speech acts. I shall return to this fact later. Directives are defined by Searle in the following manner: The illocutionary point of these consists in the fact that they are attempts (of varying degrees, and hence, more precisely they are determinates of the determinable which includes attempting) by the speaker to get the hearer to do something (1976: 11). In German technical brochures, speech acts which can be classified as obvious directives are pretty rare. Directives as: Es lohnt sich, Fendt zu wählen (=It pays to choose Fendt) or Wählen Sie X (=Choose X) may be found, but normal-

JENS HARE HANSEN

189

ly it is preferred to let the directive aspect appear indirectly.8 In German brochures great value is put on making the text seem most informative and factual.9 Under the informative surface, however, is hidden an extensive implicit content. An example from a brochure about a feeding machine is: Nach beendeter Futterfahrt kehrt der Wagen in seine Ausgangsposition zurück und wird automatisch wieder aufgefüllt. The example describes how a feeding machine after carrying out its job returns to its starting position, and is automatically refilled. At first glance this sentence is obviously representative, as it is possible to verify it. At the same time, it has a directive aspect, because of an implicit content. Indirectly it is said that the farmer's work is made easier if he chooses the machine referred to in the brochure. The machine performs a certain function automatically, so that he does not have to worry about that job. Brandt et. al. (1983) distinguish in their analytical model between the expressed propositional structure (= APS) and the total propositional structure (= PS). Because of the inferencing it is claimed that it is profitable to choose the machine in question. This implies that even though the sentence superficially is a representative speech act, a directive speech act will prevail when you take the implicit contents into account. You may say that APS to a pronounced degree is representative, while the directive aspect is mainly expressed through PS. An example of referential speech use which becomes directive when inferencing, is the information that the brakes of the machine has asbestos-free linings. The receiver himself can infer that by choosing the machine in question he can avoid exposing himself to the health damage that asbestos presents. At the same time it is an appeal to the farmers' wish to spare the environment. Statements, which at first seem to be purely referential, may also have another function by serving as subsidiary illocutions in relation to other illocutions.10 An example of this from a Fendt-brochure: Die Axialkolbenpumpe mit gleichzeitiger Druck- und Mengensteuerung arbeitet mit einem Mindestdruck von nur 20 bar. Resultat: Kraftstoffeinsparung im Vergleich zu herkömmlichen Systemen (= The axial piston pump with simultaneous pressure and volume control operates at a minimum pressure of 20 bar. Fuel savings are much greater than with other, conventional systems). If the author had only written that the Fendt-machine was fuel-saving compared to the tractors of the competitors it would, to a marked degree, have had the character of a postulate, to which the receiver would have been sceptical. The technical character of the subsidiary illocution contributes to making the assertion far more trustworthy.11 That is to say that the apparently purely refer-

190

TRANSLATION OF TECHNICAL BROCHURES

ential illocution also has a direct aspect, as it serves as a sort of documentation of the fact that it is indeed possible to save fuel by choosing Fendt.12 Atelsek (1981) suggests another speech act type, namely opinions. Her suggestion is based on a type of speech act related to Searle's representatives. Opinions are, however, characterized by not being verifiable or falsifiable like representatives, rather, they are expression of an evaluation. It must be asserted that the use of opinions is a dominant constituent characteristic of technical brochures. In many cases they express an evaluation, and not a fact. This is true of sentences like Die verschleißfreie Turbokupplung bietet höchsten Fahrkomfort (=The wear-resistant turbo clutch offers maximum driving comfort). This is clearly a subjective estimation, for what does maximum driving comfort mean? Also the following example expresses an opinion Nach Abnehmen des Frontgrills läßt sich der großdimensionierte Trockenluftfilter werkzeuglos, schnell und leicht überprüfen (=After removing the front grille, the large dryair filter is checked quickly and easily, and without need of tools). It is not possible to verify whether this statement is correct. For what does it mean that you can do something quickly and easily? Certainly this can only be illustrated through a comparison, and that is not shown. According to the theory of illocutionary structure it may be said that the total text function is operative, as it is possible to infer a dominant directive illocution in the brochures, which says It pays to choose this machine. The other illocutions are subsidiary compared to the dominating illocution. This means that apparently representative speech acts acquire a directive aspect at the same time as they support the dominating illocution. It follows that from a total evaluation it is hardly possible to distinguish between operative elements and non-operative elements. Based on the above considerations, it may be said that brochures are argumentative texts, for they belong to the category of texts that are "utilized to promote the acceptance of evaluation of certain beliefs or ideas as true vs. false, or positive vs. negative" (Beaugrande and Dressier 1992: 184). This is true, even if the texts superficially seem to be descriptive.

Actional aspects in translation Until now I have stressed the actional aspects of source-language texts (ST). It is however, not sufficient to only take the results of this analyses into consideration. In translations the purpose of the target-language text (TT) should be the dominant criterion.13 The following considerations are based on the condition that the texts should fill the same function in ST as in TT.

JENS HARE HANSEN

191

Most of the time brochures are translated on the assumption that the important thing is reproducing the denotative contents. The translation of technical brochures, however, demands far more from the translator than is usually supposed. I aim to show that a long series of conditions must be taken into consideration. The main thing is that the source text is tied to a certain culture or socioculture, and that, consequently, one must take into account that the cultural conditions and relevant marketing conditions are different in the targetlanguage culture. a) Even though the Danish and German languages are closely related, and even if the Danish and German cultures are equally closely related, there are differences in the way that brochures are linguistically formulated. There is a tendency to express illocutions differently in the two languages. Things are often said more directly in Danish brochures. This means, for example, that the directive aspect is clearer to be seen. This can be illustrated by the following examples: Det er klogt at vælge Agrometer fra starten (= It is wise to choose Agrometer from the beginning) Den næste roeoptager bør også være en Tim (= The next beet digger should also be a Tim) Derfor er Hydrema det sikre valg (= That's why Hydrema is the right choice) Tim er de bedste (= Tim machines are best) Køb dansk, når det er bedst! (= Buy Danish - when it is best) That sort of very direct speech acts can also be found in German brochures, but here they are much rarer. The German producers are to a certain degree more modest, as they normally do not explicitly assert that they produce the best machines, neither do they talk directly about the right choice. When German concerns want to throw themselves into relief compared to the competitors, this is normally done in relation to certain technical conditions/problems, which may be seen in the following example: Heute wissen wir, daß die Bodenbearbeitung häufig übermechanisiert ist. Pflügen oder Grubbern, Bearbeiten und Säen verlangen sowohl bodenbiologisch wie betriebswertschaftlich einen hohen, oft zu hohen Preis. Die vernünftige Alternative ist die Komplett-Bodenbearbeitung mit der DUTZI. In the English translation of the brochure the quotation is translated into: Today, agricultural specialists are aware of the overmechanisation of soil cultivation. Ploughing or conventional cultivation, tillage and sowing have their price, in many cases a high one. Not least as far as soil biology and economy

192

TRANSLATION OF TECHNICAL BROCHURES

are concerned. All-in-one soil cultivation by using the DUTZI is the sensible alternative. DUTZI describes their own product as a machine which is well suited to solve the problems that might arise. As it is clear from the texts, DUTZI also says that it is wise to choose especially their machine, and that DUTZI machines are best, but the message is less directly formulated than in the Danish examples, which have been quoted above. Danish brochures have more direct statements than the German ones.14 In Danish sales brochures you will often find directive speech acts like Save manpower, Save fodder and Try Us. Obviously German brochures want to say the same things, but here they normally do so in a sort of language which is typical for technical language. In German, for example, they mention that the machine implies Verminderungen des personellen Aufwandes, which means that it reduces man-power requirements. Instead of writing Save fodder the German brochures typically have Es wird Kraftfutter gespart (= Fodder is saved), Das Kraftfutter wird optimal ausgenutzt (= Concentrates are utilized to an optimal degree), or Die Futterkosten werden gesenkt (= Fodder costs are reduced).The superficially representative German sentences will, as is shown above, become directive through inferencing. Danish brochures tend to use more informal language, including many colloquial expressions and puns.15 Direct address is more often used in Danish brochures than in German ones. They say, e.g. You will not have to leave your driver's seat; The well-being of your cows is important, if...; You will reduce your fuel consumption to about 1/3 and Did you know that... ? German brochures are normally kept in a more impersonal style, where the third person singular or plural is used. Therefore they mostly describe technical aspects, and hereby they may point to the fact that certain functions can be performed directly from the driver's seat. Instead of writing The well-being of your cows is important, if..., the German brochures tend to stress that a certain technique will lead to a greater output. While the Danish texts can quite well use You save fuel, the German expression will more often be that the use of the machine will result in the saving offuel, or that it has a low energy dissipation. In Danish brochures it will be all right to write Did you know that... Such a use would be uncommon in German, where a technical thing is described without questioning whether the receiver knew of it. Both in Danish and German brochures a series of positive words and expressions are used. Generally the Danish brochures use a more superlative style than the German ones. In Danish brochures, for example, it is often mentioned that the products are ideal or optimal, that they are very sturdy, very easy to

JENS HARE HANSEN

193

service, or that they are extremely professional. You may also get the information that it is sheer joy to work with the machines. In German brochures you may find similar expressions, but here they are far less frequent. In German, for example, they relatively seldom use adverbs such as sehr (= very). They would rather choose the somewhat weaker besonders. Instead of writing very easy to service, the German brochures will more often use wartungsarm (low-maintenance) or wartungsfreundlich (= easy to service). In Danish professional is used far more often than professionell in German. In German brochures you do not actually find sentences that mention something as being sheer joy. The style of the Danish brochures is much more talkative than that of the German ones. Soellner, a German translator, writes that the realisation of technical language depends on the mentality of a people (1980: 176). Even though it may be problematic to talk about the mentality of a people, it must be said that the language of brochures seems to affirm the fact that Danes have a tendency to be rather informal, while Germans are more formal and thorough.16 According to the science of marketing it is very important that the language of advertising is well adapted to the target group you address. This means that Danish brochures, with their more informal style, should not be directly translated into German. Because, if you do not choose a (superficially) more business-like style, the text will not seem trustworthy.17 Therefore, it will be advisable to adapt to the linguistic conventions of German brochures. This often demands considerable changes compared to the source text.18 Generally brochures are very literally translated. They are often translated sentence by sentence, and the information is very similar in source text and target text. This means that interferences are definitely a risk. Both in texts translated into Danish, and in texts translated into German, the influence of the source text can normally be clearly felt. In other words, it is easily recognizable as a translation. The language of brochures translated from German into Danish will often turn out to be formal and too rigid. Certainly, Danish words are used, but this is done in a way that would not have been the case if a Dane had worked out the text without looking at the German source. For example, there will be too many substantives, and the language will be very rigid, compared to normal Danish language. Examples of this: Generende griben efter flere andre gearstænger, som afleder koncentrationen, bortfalder and Sparekraftudtagenes udveksling er gearet til 75% motoromdrejninger, her i konstantydelsesmotorens maximale drejningsmoment står faktisk den fulde motoreffekt til rådighed ved lavere brændstofforbrug. This form of language seems unnatural to a Dane. The first sentence would in a Danish text very often be changed into direct address, or: Føreren behøver

194

TRANSLATION OF TECHNICAL BROCHURES

ikke at... or Føreren er fri for at ... The sentence with Sparekraftudtagene should contain several subordinate clauses or several main clauses. In order to improve translation it is necessary to translate more freely. The risk of an artificial language is much greater, when you try to translate sentence by sentence more or less slavishly. The question for the translator is to regard the text as an instrument which can be used to obtain certain aims. From this point of view it will be possible to make the prevailing actional aspects the basis of translation. That is, instead of translating sentences, you must translate a message or a text. The translator can benefit by first noting the central information as key-words and then start to formulate the target text. This implies that he is far more independent of the surface structure of the source text, so the risk of interference errors will be diminished.19 In this way it will be possible to obtain translations that function.20 b) Many facts have different meanings in different countries. Most Danes regard Danish products as quality products. This fact is used by Danish machine producers when they often stress that the machines have been produced in Denmark. They speak of a Danish quality product or mention that the firm is owned totally by Danish investors. In some cases they write a Danish product and a Danish reader will infer that this is a good product. This positive evaluation of Danish products is not to the same degree shared by Germans, when it concerns technical products. In the view of the Germans, German firms produce the technically best products, followed by the United States and Japan (Schwenzer and Schwenzer 1984). They do not have the same confidence in the ability of the Danish industry to produce technical quality products (List and Wagner 1992: 214). This implies that Danish firms will generally find it inexpedient to emphasize the Danish origin,21 though this is extensively done on the home market. An example of the fact that the quality of products are valued differently in Denmark and Germany may be found in combine harvesters. Danish farmers regard the Dronningborg combine which is produced in Denmark, to be the best combine, while Germans prefer CLAAS, which is produced in Germany. A Danish producer informs in a brochure that the firm is placed at Lem near Ringkøbing in the west of Jutland. Such a piece of information would be of no use to a German. Equally sentences like Denmark's first producer of manure spreaders would hardly impress the German addressee. Some brochures point out that it is possible to obtain tax benefits and finances benefits if you choose certain types of machines. As the same conditions are not the case in the German market, it would be meaningless to translate this information. Some firms seek to obtain greater trustworthiness by writing that their products have been tested at the Danish Technological Institute, which does not, however,

JENS HARE HANSEN

195

give the same impression to a German reader, for whom a Danish institute does not have the same credibility as it has to a Dane. For minor firms in Germany it can be a selling point in their brochures that they are suppliers to major firms that are well-known for their good products, e.g. the tractor factory of Fendt. The reader will then draw the conclusion that when the market leader (Fendt) wants to make use of the products of the firm, it must be because the products are good. To the Danish reader the argument that the firm is a supplier of Fendt's will not have the same effect, as Fendt has not nearly the same reputation in Denmark. Even though the socio-cultural conditions are different, texts are often translated as if they were just presenting information. An example of this is found in a Danish brochure where it is said that the dimensions of the machine are made according to Scandinavian directives. Even though it is to be doubted that this information can be used by a German reader, it is directly translated into German. This sentence will hardly impress a German reader, while it will be a guarantee of quality to a Danish recipient. c) In many cases it will not be sufficient just to make use of parallel texts to become familiar with target language conventions. It can be problematic to imitate the style of a certain text. This is due to the fact that brochures present considerable variations. In Germany, great and well-known firms, e.g. CLAAS and Fendt, use a more informal style with many positive adjectives and a lot of colloquial language. In some brochures direct speech is strongly to be seen; likewise they have a great deal of colloquial features. The following sentence from a CLAAS brochure is an example of this: Wenn Sie das Schneidwerk mal zugefahren haben, dann werden Sie froh sein, daß CLAAS an eine Reversiereinrichtung gedacht hat.

When great firms like CLAAS can use such a colloquial style, it is due to the fact that the firm is among the greatest in the world. Any German farmer knows that the CLAAS machines are among the very best. Therefore the credibility of the firm is not weakened by the use of an informal style. If a Danish firm, which is unknown in Germany, wants to penetrate into the German market, it should choose a much more neutral style. Otherwise the brochure would seem untrustworthy to a German reader. d) Frequently the source texts are defective. There can be language mistakes and clumsy formulations.22 Another, and much greater problem is the fact that the source texts often leaves a great deal to be desired in pragmatic respect. This subject is dealt with by Davidow (1987). Here the author writes that good

196

TRANSLATION OF TECHNICAL BROCHURES

marketing campaigns start by the marketing people clearly stressing the advantages that the customer will get by choosing the product in question. In practice, this mainly happens in the consumer market. The same thing is certainly not always the case in the producer market. According to Davidow this is because people in the high-tech industry are often interested in the product itself rather than in the benefit that the customer derives from it.23 It is advisable, in connection with such texts, of which there are some inside the sector of agricultural machines, to make considerable text improvements of the source texts.24 When you translate texts that are pragmatically deficient, the translator faces considerable problems. First of all he must make himself familiar with the technical aspects of the machine, and find out what characteristics can be emphasized as advantages. In order best to obtain this it is also necessary for the translator to know the products of the competitors, so that he can stress the advantages of his "own" machine.25 In this respect, the same requirements principally applies to the presentation of the target text as to the preparation of source texts. In addition to this, the translator has to get familiar with the mentality of the market segment in question. Then the text must be formulated in accordance with the way that this group regards the world, that is to say, the values, norms and expectations of the target group. The demands that are made on the translator in connection with defect source texts, are considerable.26 In order better to make impact it can be expedient for the sender to use what Große (1976) calls "gruppenindizierende Sprache". That is to say a kind of language use in which you emphasize, for example, common qualities or aims in life. If the receiver regards the sender as belonging to the same group, the possibility for influencing the receiver is increased. As farmers regard themselves as practical men, and as they are very sceptical to theorists, many machine producers stress that they, too, are practical men. Also the use of colloquial language in many brochures can be regarded as "gruppenindizierende Sprache". e) Brochures are only part of the total communication-mix. Probably some firms will stress the use of brochures on the export market more than they will on the home market. This can imply that it is necessary to include more information in the translation than in the source text. It can, for example, be advantageous to present the firm in the translated brochure, even though such a presentation is not to be found in the source edition.27

JENS HARE HANSEN

197

Conclusion It follows from the above mentioned that the division into specialized and nonspecialized texts does not show very much about how the text type of technical brochures should be translated. Though technical brochures are LSP-texts it is very clear that the denotative equivalence cannot be regarded as the dominant criterion. Technical brochures demand a certain technical knowledge of the translator. Furthermore, the translator must have a certain knowledge of text type conventions. It is also necessary that he has marketing qualifications and intercultural competence. Different cultures and different marketing aspects between language areas play a greater part than generally supposed. In many brochures there is information, which is completely or partly irrelevant in the target language area; this information should at least be adjusted to other target language conditions.28 It is decisive for the quality of the translation that the translator takes into consideration that people express themselves differently in Danish and German. German brochures emphasize a factual and technical style, while Danes express things in a less "technical" manner, e.g. with fewer substantives. Furthermore Danish brochures also use far more subordinate clauses than German ones. There are many inadequate translations of brochures where it is immediately realized that "this is a translation".29 Business concerns are usually very cost-conscious when it comes to translations. This is strange when you take into consideration that brochures constitute an important part of the external communication of the firm, and that it is often the first contact with the customer. And of course a bad impression can have unfortunate consequences.30 In order to improve bad translations, one must abandon the strategy of translating sentence by sentence. If one chooses a very direct form of translation it is actually only suitable as a starting point for an advertising agency, which will then change the text in order to adapt it to the target language style. When translating the text type of technical brochures it is necessary to take into account the receiver much more than is generally done today. Under present conditions too much importance is attached to the source text. The semantic aspect is stressed too much, at the expense of pragmatic consideration. Heinemann and Viehweger write about the sender of source texts that he makes a series of linguistic decisions considering the superior communicative targets in order to acquire the maximum effect on the receiver. This concerns the choice of propositions, for example. When it comes to brochures, the pur-

198

TRANSLATION OF TECHNICAL BROCHURES

pose is to further the sale of the products; this is the main purpose for the texts. There is every reason for the translation of brochures to use the same criteria. The choice of linguistic means in translations should also be made so that the message can obtain the goal that is desired. This means that the importance of the source texts is considerably decreased. The dominant criterion of the target text is, in other words, to consider the aim of the brochure. It is clear that this implies a very free translation which is not least determined by sales psychology, though of course also by purely technical aspects. Notes 1 Cf. Göpferich (1995) and Bungarten (1993). 2 Cf. e.g. Koller (1992: 90ff. and 214ff). 3 This scepticism of theory is surprising, as educations that only stress practical skills cannot be called academic. - But universities now also need to provide professional training. The problem is that theory and practice need to be linked. 4 This article is mainly based on the examination of technical brochures concerning agricultural machines, but the predominant results should also imply when it comes to similar texts from other technical sub-languages. 5 Cf. for example Stolze (1992: 97) "Der Übersetzer im Dienste der Technik ... hat keine welterschließende Aufgabe wie im Bereich der Dichtung, sondern die demonstrative Funktion der präzisen Darstellung der Objektwelt". 6 Römer (1976) for example, speaks of semantische Aufwertung, and Möhn and Pelka (1984) of aufwertende Adjektive, but they do not define the conception so that it can be made operative. 7 Further features typical for LSP-texts are dealt with by Fluck (1984: 73ff). 8 Obviously directive speech acts seem to be more frequent in technical advertisements than in brochures. Gläser (1990: 252) briefly deals with directives in English advertisements of technical products. 9 In the theory of persuasive communication it is often stressed that it can be advantageous to hide the aim of influencing the receiver (cf. e.g. Breuer 1974: 56ff.). 10 The question of the relation of sentences to illocutions is not yet finally decided. Here, I regard sentences as illocutions, even though this approach may be far from unproblematical. 11 The author of the text has clearly tried to anticipate possible objections from the receiver. Anticipation of reaction from the receiver, as a fundamental element, is extensively dealt with in Zimmermann (1984). 12 Cf. Stolze (1992: 199), who writes that the description of technical details can give the impression of precision, efficiency, and competence. 13 This is stressed by Hönig and Kußmaul (1982), Reiss and Vermeer (1991) and Trosborg (1994). 14 The corpus consists of fifty Danish and fifty German brochures. 15 Examples are: Få det perfekte indeklima med lune besparelser, Os imellem er der kun varmt vand (about a heat pump); Agrometer gøder jorden for et effektivt landbrug. 16 This is dealt with by List and Wagner (1992).

JENS HARE HANSEN

199

17 Credibility is one of the most important concepts in marketing theory (cf. Schifko 1982). 18 As mentioned above, the greater objectivity in German brochures is a surface phenomenon. As is the case of the Danish brochures, the German ones are also clearly evaluative. 19 The problem of interference is dealt with in Stolze (1992: 76ff.). 20 Also Soellner (1980) comes to the conclusion that direct translations are artificial. 21 On the other hand, the marketing of Danish foodstuffs benefit by stressing that the products are Danish, as such products are regarded as high quality products (cf. Bänsch 1991). 22 Göpferich mentions that in her five years as a translator, she has not come upon texts that had not at least small mistakes. 23 Pomplitz (1990) draws the attention to the fact that many firms are unwilling to invest much money in advertising for industrial products, and that there is no special prestige for copy writers in making such texts. 24 Also Kühn (1990) mentions the problem of bad industrial advertising, which is not based on the knowledge and interests of the receiver, but on purely technical facts. 25 Cf. Bänsch (1991: 26) about USP. 26 Actually, there can be many things that prevent the translator from living up to the demands that have been mentioned in this article (cf. Sager 1986), but this does not change the fact that good translations must take these into account. 27 Pomplitz (1990a) deals with relations between brochures and other forms of market communications. 28 Therefore it can be said about the translation of technical brochures that it is far from being "a mere transcoding process, but ... a cross-cultural event" (Trosborg 1994: 25). 29 Lykke Jacobsen states that industrial translators must produce texts that seem natural. "So natural, in fact, that translations are indistinguishable from originals" (1993: 67). Relatively few translations of brochures meet this demand. 30 There are, however, some translations which are so good that it is only possible to realize that it is a translation, because you know that this is the case. References Atelsek, Jean. 1981. "An anatomy of opinions". Language in Society 10: 217-225. Bänsch, Axel. 1991. Einführung in die Marketing-Lehre. 3., wesentliche erneuerte und erweiterte Auflage. München: Vahlen. Baumann, Klaus-Dieter and Kalverkämper, Hartwig (eds). 1992. Kontrastive Fachsprachenforschung. Tübingen: Narr. Beaugrande, Robert-Alain de and Dressler, Wolfgang Ulrich. 1992. Introduction to Text Linguistics. London/New York: Longman. Brandt, Margareta, Koch, Wolfgang, Motsch, Wolfgang, Rosengren, Inger and Viehweger, Dieter. 1983. "Der Einfluß der kommunikativen Strategie auf die Textstruktur - dargestellt am Beispiel des Geschäftsbriefes". In Inger Rosengren (ed), Sprache und Pragmatik. Lunder Symposium 1982. Lund: Almqvist and Wiksell International, 105-135. Breuer, Dieter. 1974. Einführung in die pragmatische Texttheorie. München: Fink. Brinker, Klaus. 1992. Linguistische Textanalyse. Eine Einführung in Grundbegriffe und Methoden. 3., durchgesehene und erweiterte Auflage. Berlin: Erich Schmidt Verlag.

200

TRANSLATION OF TECHNICAL BROCHURES

Bungarten, Theo. 1993. "Hinsichten zu einer Theorie der Fachsprachen. Zur Einführung". In Theo Bungarten (ed), Fachsprachentheorie. Tostedt: Attikon. Davidow, William H. 1987. High Tech Marketing. Der Kampf um den Kunden Erfahrungen und Rezepte eines Insiders. Frankfurt: Campus. Fluck, Hans-Rüdiger. 1984. Fachdeutsch in Naturwissenschaft und Technik. Heidelberg: Groos. Gläser, Rosemarie. 1990. Fachtextsorten im Englischen. Tübingen: Narr. Große, E. U. 1976. Text und Kommunikation. Eine linguistische Einführung in die Funktionen der Texte. Stuttgart: Kohlhammer. Göpferich, Susanne. 1992. "Eine pragmatische Typologie von Fachtextsorten der Naturwissenschaften und der Technik". In K.-D. Baumann and H. Kalverkämper (eds), 190-210. Göpferich, Susanne. 1993. "Die translatorische Behandlung von Textsortenkonventionen in technischen Texten". Lebende Sprachen 38(2): 49-53. Göpferich, Susanne. 1995. Textsorten in Naturwissenschaften und Technik. Pragmatische Typologie - Kontrastierung - Translation. Tübingen: Narr. Hatim, Basil and Mason, Ian. 1993. Discourse and the Translator. London/New York: Longman. Heinemann, Wolfgang and Viehweger, Dieter. 1991. Textlinguistik. Eine Einführung. Tübingen: Niemeyer. Hoffmann, Lothar. 1985. Kommunikationsmittel Fachsprache. Eine Einführung. Tübingen: Narr. Holz-Mänttäri, Justa. 1984. Translatorisches Handeln. Theorie und Methode. Helsinki: Suomalainen Tiedeakatemia. Holz-Mänttäri, Justa and Nord, Christiane (eds). 1993. Traducere Navem. Festschrift für Katharina Reiß. Tampere: Universitätsbibliothek Tampere. Hönig, Hans G. and Kußmaul, Paul. 1982. Strategie der Übersetzung. Ein Lehr- und Arbeitsbuch. Tübingen: Narr. Jakobsen, Arnt Lykke. 1993. "Translation as textual (Re)production". In Justa HolzMänttäri and Christiane Nord (eds), Traducere Navem. Festschrift für Katharina Reiß zum 70. Geburtstag. Tampere: Universitätsbibliothek . Jakobsen, Arnt Lykke (ed). 1994. Translating LSP Texts: Some Theoretical Considerations. Copenhagen: Samfundslitteratur. Koller, Werner. 1992. Einführung in die Übersetzungswissenschaft. Heidelberg: Quelle und Meyer. Kroeber-Riel, W. 1991. Strategie und Technik der Werbung. Stuttgart: Kohlhammer. Kußmaul, Paul. 1990. "Die Übersetzung von Sprechakten in Textsorten". Der Deutschunterricht 1: 17-22. Kühn, Jürgen M. 1990. "Mausgrau oder schmetterlingsbunt? Oder Der Investtexter zwischen Menschen und Maschinen". In Ernst Högn and Hans-Jürgen Pomplitz (eds), Der erfolgreiche Werbetexter. Landsberg am Lech: Moderne Industrie. List, Pia and Wagner, Johannes. 1992. "Nationale Stereotype im internationalen beruflichen Alltag: Überlegungen anhand eines Fallbeispiels". In A. Grindsted and

JENS HARE HANSEN

201

J. Wagner (eds), Communication for Specific Purposes. Fachsprachliche Kommunikation. Tübingen: Narr, 210-226. Möhn, Dieter and Pelka, Roland. 1984. Fachsprachen. Eine Einführung. Tübingen: Niemeyer. Niss, Hanne. 1994. Made in Denmark. Nationalitetens betydning i international markedsføring. Aalborg: Aalborg Universitetsforlag. Nord, Christiane. 1991. 2nd edition. Textanalyse und Übersetzen. Heidelberg: Groos. Pomplitz, H. J. 1990. "'Viel Arbeit - wenig Ehr'. Oder: Das harte Brot der Investtexter". In Ernst Högn and Hans-Jürgen Pomplitz (eds), Der erfolgreiche Werbetexter. Landsberg am Lech: Moderne Industrie. Pomplitz, H. J. 1990a. "Wie textet man einen Prospekt?" In Ernst Högn and HansJürgen Pomplitz (eds), Der erfolgreiche Werbetexter. Landsberg am Lech: Moderne Industrie. Poulsen, Sven-Olaf. 1990. "Zur Problematik des textsortenbezogenen Übersetzens". Der Deutschunterricht 1: 29-35. Reinhardt, W. et al. 1992. Deutsche Fachsprache der Technik. Hildesheim: Olms. Reiss, Katharina. 1983. Texttyp und ÜberSetzungsmethode. Der operative Text. Heidelberg: Groos. Reiss, Katharina and Vermeer, Hans J. 1991. Grundlegung einer allgemeinen Translationstheorie. Tübingen: Niemeyer. Römer, Ruth. 1976. Die Sprache der Anzeigenwerbung. Düsseldorf. Sager, Juan D. 1986. "Die Übersetzung im Kommunikationsprozeß: der Übersetzer in der Industrie". In Snell-Hornby (ed), 331-347. Schwenzer, J. E. and Schwenzer J. M. 1984. "Völkerimages. Selbstbild und Bilder anderer Völker bei den Deutschen". Marktforschung 2, 58-64. Searle, John R. 1976. "A classification of illocutionary acts". Language in Society 5: 1-23. Snell-Hornby, Mary (ed). 1986. Übersetzungswissenschaft. Eine Neuorientierung. Tübingen: Francke. Soellner, Rolf. 1980. Form und Inhalt. "Betrachtungen zur literarischen und technischen Übersetzung". W. Wilss and S. O. Poulsen (eds), Angewandte Übersetzungswissenschaft. Internationales übersetzungswissenschaftliches Kolloquium an der Wirtschaftsuniversität Århus/Dänemark, 19-21 Juni 1980. Århus, 165-179. Sowinski, Bernhard. 1979. Werbeanzeigen und Werbesendungen. München: Oldenburg. Splinter, R. and Hartig, P. 1980. "Fachsprachliche Besonderheiten der Werbesprache, untersucht am Beispiel der Anzeigenwerbung". In Textgattungen der Technik. Edited by Gunter Neubert. Berlin. Stolze, Radegundis. 1992. Hermeneutisches Übersetzen. Linguistische Kategorien des Verstehens und Formulierens beim Übersetzen. Tübingen: Narr. Stolze, Radegundis. 1994. Übersetzungstheorien. Eine Einführung. Tübingen: Narr. Sørensen, Henrik Selsøe. 1994. "Knowledge and LSP translation. When does a translator have to be unfaithful? Some Cases of LSP Translation". In A. Jakobsen 13-31.

202

TRANSLATION OF TECHNICAL BROCHURES

Trosborg, Anna. 1994. "Translation studies: Some recent developments." Hermes. Journal of Linguistics.12.1994. 9-28. Vestergaard, Torben and Schröder, Kim. 1985. The Language of Advertising. Oxford: Blackwell. Wilss, Wolfram. 1979. "Fachsprache und Übersetzen". In Terminologie als angewandte Sprachwissenschaft. Gedenkschrift für Univ.-Prof. Dr. Eugen Wüster, edited by H. Felber u. a. München/New York/London/Paris, 177-191. Zimmermann, K. 1984. "Die Antizipation möglicher Rezipientenreaktionen als Prinzip der Kommunikation". In I. Rosengren (ed), Sprache und Pragmatik. Lunder Symposium 1984. Malmö: Almqvist and Wiksell International, 131-158.

Translating Legal Genres Vijay K. Bhatia City University of Hong Kong

Recent work in applied genre analysis (Swales 1981, 1990, Bhatia 1983a, 1993, Dudley-Evans 1986, Berkenkotter and Huckin 1995) has reiterated the importance of linguistic analysis in the teaching and learning of translation. It does not look at discourse analysis simply as an act of linguistic description but more as linguistic explanation, attempting to answer the question, why do members of specific discourse communities use the language the way they do? The answer requires input not from linguistics alone, but equally important, from sociolinguistic and ethnographic studies, psycholinguistic and cognitive psychology, communication research, studies of disciplinary cultures and insights from members of such discourse communities, to name only a few crucial sources of information. Taking communicative purpose as the key characteristic feature of a genre, the analysis attempts to unravel mysteries of the artifact in question. Genre analysis thus, has become one of the major.influences on the current practices in the teaching and learning of languages. In this paper I would like to extend these areas of application to demonstrate the use of genre theory to the teaching and learning of translation. I will do this in the following four sections. 1. The first section will highlight some of the complexities of translation process, especially in legal contexts. 2. The second section will characterize some of the major characteristic features of genre theory which have serious implications for the theory and practice of translation. 3. The third section will discuss some of the characteristic features of legislative writing which need special treatment in translation work. 4. The final section will then illustrate how a genre-based framework can be utilized to maintain generic integrity in translation work, taking examples from intra-lingual legislative translations.

204

TRANSLATING LEGAL GENRES

Complexities of translation process Translation is not simply a matter of linguistic transference alone, but a genuine act of communication in its own right, perhaps more complex than communication in a singular semiotic system. It is an attempt to communicate someone else's message through another language. It is complex because it takes place through different semiotic systems in the context of diverse socio-cognitive and disciplinary cultures in response to a variety of motivations. In short, it is an attempt to communicate one world in terms of another. In order to accomplish it successfully and effectively one requires an understanding of the two semiotic systems at the same time. This, among a number of other things, means awareness not only of the source text, the writer and the readers in the source language context, lexico-grammatical resources in the source language, generic conventions in the source disciplinary culture, but also awareness of the target text (translation), the writer (translator) and the readers in the target language context as well as awareness of lexicogrammatical resources in the target language and the generic conventions associated with the relevant disciplinary cultures in the target language. Translation thus requires exploration of a socio-cultural, as well as semiotic systems of two languages to make appropriate linguistic and hence communicative choices depending upon individual motivations. In addition to this, the translator is also a reader and is involved in cognitive work as well. This, in all probability, will require one to take into account some or all of the following considerations. • •

Interpretation (intended use of the translation): the question of generic integrity Easification or facilitation of the original text (especially in the context of legal texts)

These two related concerns give rise to several important issues for translation theory, including the following: Do we convey or recreate other people's messages? Do we create and clarify context in which one wishes to communicate? Candlin (1990) discussing the complexities involved in the processes of translation aptly points out, Here the somewhat sterile debates about translation as process or translation as product give way to fresh opportunities to cohere the semiotic, the linguistic, the social, the cultural and the psychological perspectives on communicating. In short, it offers a broader conception of what it means to understand. Although Malinowski's notion of context of situation itself was originally developed in the context of translation, much of translation has concentrated on

VlJAY K. B H A T I A

205

the source text in order to produce a target text, to the neglect of a number of other contextual determinants of such texts, whether in the source language or in the target language, some of which being • rhetorical context • communicative purpose • textual organization • generic knowledge • generic dynamism Many of these factors have been brought into prominence in the recent developments within a socio-cognitive genre theory as discussed in Swales 1990, Bhatia 1993, and Berkenkotter and Huckin 1995, especially the role of communicative purpose and the associated concern with generic integrity in text construction, interpretation and use in professional contexts. Genre analysis Genre analysis is the study of situated linguistic behaviour in institutionalized academic or professional settings. Some of its crucial characteristics for our purpose are the following. • Genre analysis shows a genuine interest in the use of language to achieve communicative goals rather than a detailed extension, validation or otherwise of one linguistic framework or the other. In this sense, it is not an extension of linguistic formalism. •

Genre theory exploits all aspects of socio-cognitive knowledge situated in disciplinary cultures in order to analyze construction, interpretation and use of linguistic communication to achieve non-linguistic goals. A large part of this kind of generic knowledge is embedded in the discoursal practices of the professional community which habitually engages in such generic forms.

Although these generic forms are highly conventionalized they are not static. The very fact that they do account for the way expert users of language manipulate generic conventions to achieve a variety of complex goals, indicates that they are highly dynamic and clarificatory. The most important contribution of genre theory towards the understanding of professional discourse is that it allows one to understand how a professional writer is able to maintain generic integrity on the one hand, and exploit generic convention, on the other hand, to achieve private intentions within the constraints of socially recognized communicative purposes.

206

TRANSLATING LEGAL GENRES

A genre-based approach to translation One of the most crucial aspects of translation, which has been neglected in recent literature on the subject, is the concern to maintain the generic identity of the target text. In other words, a newspaper report in English, when translated into Danish, should preserve all or at least the representative generic features of newsreports in the Danish version. Sometimes such a decision may become rather complex if generic conventions and their realizations in the two languages show a considerable degree of variation, in which case one may need to take a further decision as to which generic pattern one must represent. Either way, one cannot afford to ignore the generic realities of the text genre one is translating. Once it has been decided to maintain generic identity, one needs to determine the approach to achieve this most effectively and economically in the translation classroom. In my view, there are at least four distinct, though systematically related, areas of competence that the learner needs to develop in order to handle the question of generic integrity in the translation work, especially in professional discourse. Assuming that most of these learners already possess a reasonably adequate competence in the use of the two languages for general every day functions, they will still need to develop understanding of the specialist codes, familiarity with the dynamics of specialist genres, which includes the rhetorical forms and content, specific contexts they respond to and the conventions they tend to use in their responses, and finally, a proficiency in the manipulation of specialist genres to respond to the exigencies of unfamiliar and novel situations. In other words they need to develop the understanding of the two codes, the acquisition of genre knowledge associated with the specialist cultures, sensitivity to cognitive structuring of specialist genres in the two specialist areas and then, and only then, they can hope to exploit generic knowledge of a repertoire of specialist genres by becoming effective translators of specialist discourse in a particular chosen field. Let me elaborate on these four stages I have identified by taking them up one at a time. Knowledge of the codes The knowledge of the codes, of course, is the pre-requisite for developing communicative expertise in any specialist or even everyday discourse. Most of the language programmes all-over the world aim to achieve this to varying degrees of success. However, let me point out that an almost perfect knowledge of the codes is neither necessary nor sufficient to achieve success in translation

VIJAY K. BHATIA

207

instruction, though it does seem to be a popular myth many of us often tend to believe in. On the one hand, it is often possible to develop translation expertise in specialized areas of discourse in a relatively short period of time; on the other it is naive to imagine that all good bilinguals will necessarily make good translators. Acquisition of genre knowledge Translating a particular professional genre presupposes an acquaintance not only with the communicative goals of that particular discourse community, but also with the communicative purposes associated with specific use of genres to achieve those goals. Therefore, before the learner is directed to undertake any communicative activity, he or she needs to become aware of appropriate rhetorical procedures and conventions which are typically associated with the specialist discourse community. Genre knowledge of this kind is a form of 'situated cognition', which seems inextricable from professional writers' procedural and social knowledge (Berkenkotter and Huckin 1995: 13). Learners need to acquire genre knowledge, procedural knowledge (which includes a knowledge of tools and their uses as well as their discipline's methods and interpretive framework) and social knowledge (in the sense of familiarity with the rhetorical and conceptual context) in order to become better informed apprentices. As Fairclough (1992:126) points out, "... a genre implies not only a particular text type, but also particular processes of producing, distributing and consuming texts." Sensitivity to cognitive structures Having understood the goals of the specialist community and to some extent internalized some of the conventions associated with specialist genres used by them, the learner will then require to develop acquaintance with the way language is typically used to achieve these goals and communicative purposes, on the one hand, and to exploit these conventions sometimes to respond to changing socio-cognitive demands in specific professional contexts or certain novel situations in the target culture and sometimes even to achieve private intentions within the socially recognized purposes. This can be developed by sensitizing apprentice learners not simply to the generic forms and content in genre-specific texts but also to their emerging responses to changes in social practices in the two semiotic systems.

208

TRANSLATING LEGAL GENRES

Exploitation of generic knowledge It is only after the learner has developed some acquaintance or, better still, expertise at levels discussed above, that he or she can confidently handle specialist discourse; interpret it, translate it, and exploit it or may be, take liberties with it. Let me point out here that the first three stages, to a large extent, involve understanding conventions, whereas the last stage includes exploiting and taking liberties with conventions to achieve pragmatic success through translated texts suitable for specified professional contexts.

Legislative discourse Typical characteristics of legal discourse, especially legislative provisions, have been adequately documented in several studies (see Bhatia 1982, 1983a, 1984, 1987, 1993, 1994b, Gustafsson 1975, 1984, Gunnarsson 1984, Maley 1994, White 1982). As White (1982:423) points out one of the most problematic features of legal discourse is that it is 'invisible'. He claims, the most serious obstacles to comprehensibility are not the vocabulary and sentence structure employed in law, but the unstated conventions by which language operates. What he means here is that there are expectations about the way in which language operates in legal contexts, but such expectations are never explicitly stated anywhere but in legal culture and which are simply assumed by the linguistic surface in such contexts. At the surface level, in addition to the use of a high degree of nominalisation, one confronts a typical use of qualifications to express complex contingencies. In order to make legislative statements not only simple, clear and unambiguous but all-inclusive also, these qualifications are inserted at various points in the syntax of legislative sentences where they introduce syntactic discontinuities which become formidable obstacles to an effective processing of legislative statements. They also tend to introduce excessive information load at various points in the syntax of such statements thereby creating barriers to effective understanding of cognitive structuring in such statements (see Bhatia 1993: 101-117). In order to be able to understand and, to some extent, translate legislative provisions, whether from one language to another, or from one audience to another, one is inevitably required to take into account the typical difficulties imposed by some of these factors. Translation process, whether intra-lingual or inter-lingual, must take into account the relative accessibility of the target text for a specific audience, which makes the notion of easification particularly relevant to the process of

VIJAY K. BHATIA

209

translation. Easification, as introduced in Bhatia (1983b) and subsequently developed in Bhatia (1993), is a process of making a text-genre more accessible to an intended readership without sacrificing its generic integrity. One of the most important easification procedures, particularly relevant to legislative statements is what Bhatia (1993: 209) calls clarification of cognitive structuring, which may include the use of textual-mapping devices (Bhatia 1987), often making it possible for the writer to reduce information load at various points in the syntax of legislative statements. Let me now illustrate how this can be effectively used for intra-lingual translation purposes by taking up an extract from a publisher's contract. The author hereby warrants to the Publishers that the author has the right and power to make this Agreement and that the Work is the Author's own original work, except for material in the public domain and such excerpts from other works as may be included with the written permission of the copyright owners, and will in no way whatever give rise to a violation of any existing Copyright, or a breach of any existing agreement, and that the Work contains nothing defamatory or libelous and that all statements contained therein purporting to be facts are true and that nothing in the Work is liable to give rise to a criminal prosecution or to a civil action for damages or any other remedy and the author will indemnify the Publishers against any loss, injury or expense arising out of any breach or alleged breach of this warranty. The Publishers reserve the right to alter or to insist the Author alter the text of the Work in such a way as may appear to them appropriate for the purpose of removing or amending any passage which on the advice of the Publishers' legal advisers may be considered objectionable or likely to be actionable at law without affecting the Author's liability under this Clause in respect of any passage not so removed or amended. The foregoing warranties and indemnities shall survive the termination of this agreement. A provision like this could be relevant to a number of professionals in industrial, management and bureaucratic institutions, who though not legal experts, are nevertheless required to be able to read, understand, interpret and translate legal documents as part of their daily professional activities. To be able to handle such statements they may need to be trained in the use of the language in typical legal settings. They may have the ability to use languages in their every day life when they find it necessary; however, when it comes to the question of handling legal language, especially legal rules and regulation, legal agreements and contracts, it is likely that many of them will quite understandably feel nervous about it. As part of their daily routine, most of them are often required, among other things, to read, interpret, translate and explain ordinances, legislative acts, contracts and agreements to their superiors and quite often to members of the general public.

210

TRANSLATING LEGAL GENRES

In order to become proficient in the use of language(s) in legislative contexts, in general, and for translation purposes, in particular, they need to develop some or all of the following skills. • •

The ability to understand why legal documents are written the way they are The ability to understand how these documents are constructed, interpreted and used • The ability to read and clarify these legal documents for the benefit of lay audience And, most of all, • to develop increased self-confidence, and sensitivity to the use of legal genres by acquiring genre skills, including those of rhetorical consciousness

Easification as intra-lingual translation Assuming that the learners are adequately equipped with linguistic competence in everyday use of language, they still need to be given enough background information about the contexts in which legislative rules are drafted, interpreted and used. Particular attention needs to be given to the dual characteristics of legal rules, i.e., clarity, precision and unambiguity, on the one hand, and allinclusiveness, on the other. The learners also need to be given sufficient practice in analysis of legislative sentences, especially focusing on the use of lexico-grammatical devices which are typically used to make their interpretation and use certain as well as flexible. Particular attention must be paid to the identification and use of complex-prepositional phrases and qualificational insertions to make rules clear, precise and unambiguous and to binomial expressions to make them all-inclusive. A lot of attention can also be paid to cognitive structuring typically associated with legislative sentences. In addition, the learners will require some practice in identifying psycholinguistic problems resulting from discontinuities in syntax as a result of qualificational insertions. The learners can then be introduced to the notion of easification f or the specialist audience and simplification for the lay audience, which are introduced as two different genres, because they serve two different communicative purposes, and are meant for two different audiences (see Bhatia 1994a for a detailed discussion of this issue). One may be required to decide whether one is translating for lay or specialist audience. Depending upon their decision, they may be asked to undertake one of the following tasks.

VlJAY K. BHATIA

211

Tasks 1. Considering the communicative purpose of this text to be to regulate the future legal relationship between the Publishers and the Author, how would you write an easier version of this section meant for the specialist audience? 2. How would you simplify the content of this section for a non-specialist audience who simply would like to be informed about the content of the clause? A typical learner response to task (1) of easification for specialist audience will be something like the following. An easified version for specialist audience (1) The author hereby warrants to the Publishers that (a) the author has therightand power to make this Agreement, and (b) the Work is the Author's own original work, except for material in the public domain and such excerpts from other works as may be included with the written permission of the copyright owners, and will in no way whatever give rise to a violation of any existing Copyright, or a breach of any existing agreement, and (c) the Work contains nothing defamatory or libellous, and (d) all statements contained therein purporting to be facts are true and, (e) nothing in the Work is liable to give rise to a criminal prosecution or to a civil action for damages or any other remedy, and (f) the author will indemnify the Publishers against any loss, injury or expense arising out of any breach or alleged breach of this warranty. (2) The Publishers reserve the right to alter or to insist the Author alter the text of the Work in such a way as may appear to them appropriate for the purpose of removing or amending any passage which on the advice of the Publishers' legal advisers may be considered objectionable or likely to be actionable at law without affecting the Author's liability under this Clause in respect of any passage not so removed or amended. (3) The foregoing warranties and indemnities shall survive the termination of this agreement. A typical response to Task 2, on the other hand, will be something like the following:

212

TRANSLATING LEGAL GENRES

A simplified version for lay audience • The author declares that the work is author's own original work. • Wherever the author has included extracts from other sources with the permission of the copyright owners, they do not violate any existing agreement. • The work does not contain any thing defamatory, libellous or factually untrue. • The work will not give rise to any criminal prosecution or civil action for damages. • If it does any of these things, the author will be responsible for making good any losses to the publisher. • The publisher can remove or amend any parts of the text if he finds it objectionable. A quick comparison of the three versions demonstrates the self-confidence the learner seems to have gained in handling complexities of legal discourse and the extent to which he has become sensitive to the specific demands imposed on informed readership of such documents. It also indicates the internalization and use, among other things, of the following abilities and strategies typically used in the construction, and interpretation of specialized genres associated with the legal culture, which include an ability • • • •

to cope with the complexities of legal syntax in legislative contexts, to handle the use of excessive information load in the legislative sentence, to use textual-mapping devices in the expression of complex contingencies, and to distinguish communicative purposes in the two versions used above.

An adequate understanding of the easification procedures establishes the importance of generic integrity in translation work. The approach establishes the importance of interpretation or facilitation in translation work, on the one hand, and of maintaining generic integrity in the translated work, on the other. The main advantage of such a genre-based approach to the teaching and learning of translation is that the learner does not learn language in isolation from specialist contexts, but is encouraged to make relevant connection between the use of language on the one hand and the purpose of communication on the other, always being aware of the question, why do members of the specialist discourse community use the language in this way? This develops in the learner an explicit desire to participate consciously in the professional community, rather than just being able to translate legal texts as a computer does, without being a participant in the communicative event. This awareness of participation in the ownership of the genres of legal culture is what Swales (1990) calls raising rhetorical consciousness in the learner.

VIJAY K. BHATIA

213

References Berkenkotter, C , and Huckin, Thomas N. 1995. Genre Knowledge in Disciplinary Communication — Cognition/Culture/Power, New Jersey: Lawrence Erlbaum Associates, Publishers. Bhatia, V. K. 1982. An investigation into the formal and functional characteristics of qualifications in legislative writing and its application to English for academic legal purposes, a Ph.D. thesis, University of Aston in Birmingham, UK. Bhatia, V. K. 1983a. An Applied Discourse Analysis of English Legislative Writing. A language Studies Unit research report, University of Aston in Birmingham, UK. Bhatia, V. K. 1983b. "Simplification v. easification: The Case of Legal Texts". Applied Linguistics 4 (1): 42-54. Bhatia, V. K. 1984. "Syntactic discontinuity in legislative writing and simplications for academic legal purposes". In A. K. Pugh and J. M. Ulijn, (eds), Reading for Professional Purposes - Studies and Practices in Native and Foreign Languages. London: Heinemann Educational Books, 90-96. Bhatia, V. K. 1987. "Textual-mapping in British Legislative Writing". World Englishes 6(1): 1-10. Bhatia, V. K. 1993. Analysing Genre — Language Use in Professional Settings. London: Longman, Applied Linguistics and Language Study Series. Bhatia, V. K. 1994a. "Generic integrity in professional discourse". In Britt-Louise Gunnarsson, Per Linell and Bengt Nordberg (eds), Text and Talk in Professional Contexts, ASLA:s skriftserie 6, Uppsala, Sweden. Bhatia, V. K. 1994b. "Cognitive structuring in legislative provisions". In John Gibbons (ed), Language and the Law, London, Longman, 136-55. Candlin, C. N. 1990. General Editor's Preface to Basil Hatim and Ian Mason's Discourse and the Translator. London: Longman. Dudley-Evans, Tony. 1986. "Genre analysis: an investigation of the introduction and discussion sections of M.Sc. dissertations". In M. Coulthard, (ed), Talking about Text. Birmingham: English language Research, University of Birmingham, U.K. Fairclough, N. 1992. Discourse and Social Change. London: Polity. Gunnarsson, Britt-Louise. 1984. "Functional comprehensibility of legislative texts: Experiments with a Swedish act of parliament". Text 4/(1-3): 71-105. Gustafsson, Marita. 1975. Some Syntactic Properties of English Law Language. Turku, Department of English , University of Turku. Gustafsson, Marita. 1984. "The syntactic features of binomial expressions in legal English". Text 4(1-3): 123-141. Maley, Yon. 1994. "The language of the law". In John Gibbons (ed), Language and the Law. London: Longman, 11-50. Swales, J.M. 1981. Aspects of Article Introductions. Aston ESP Research Report No.l, Language Studies Unit, Universy of Aston in Birmingham, Birmingham, U.K. Swales, J.M. 1990. Genre Analysis - English in Academic and Research Settings. Cambridge: Cambridge University Press.

214

TRANSLATING LEGAL GENRES

White, J. B. 1982. "The invisible discourse of law: reflections on legal literacy and general education". Michigan QuarterlyReview,420-38.

Part III Terminology and Lexicon

Synonymy and Equivalence in Special-language Texts A Case Study in German and English Texts on Genetic Engineering1 Margaret Rogers University of Surrey, UK

Introduction The process of translation focusses attention on the problems of establishing terminological equivalence between languages, or more precisely, between terms in a particular source text (ST) and those in its emerging translation, the target text (TT). It also highlights problems of synonymy as the translator attempts to establish which terms are synonymous and what the conditions are for their use. Indeed, the translator must regularly make decisions not only concerning synonymy within the ST and within the TT, but also concerning the crosslinguistic relations between these synonyms. Although translators must deal with synonymy and equivalence in practice on a daily basis, these are theoretically problematic notions, each in its own right. So what is the relationship between synonymy and equivalence?

Synonymy and Equivalence In terminology science, the 'meaning' of the term is said to be the concept, which is the level at which any terminological analysis must start (Felber 1984: 98), i.e. the approach is onomasiological. Form and meaning are not regarded as two aspects of the same unit, as in the semasiological approach, but as independent entities. Synonyms are defined as two or more terms from the same language representing the same concept. Equivalents can be defined as two or more terms from different languages representing the same concept. In the terminological view, there is no direct meaning relation between words; relation-

218

SYNONYMY AND EQUIVALENCE

ships of meaning are mediated through the concept. We can therefore arrive at a diagrammatic representation of synonymy and equivalence from a terminological perspective as in Figure 1. On the basis of such a model we can, however, make no predictions concerning the equivalence relations between particular synonyms in the source language (SL) and particular synonyms in the target language (TL) except to assume that they are fully substitutable within each language and that all are equally possible as equivalents across languages. Experience tells us, however, that this is unlikely. While a concept-based approach may help to establish the fact of synonymy (often as a prelude to its elimination in the interests of clear LSP communication), it is not in itself able to shed any light on the criteria for choosing particular synonyms (or equivalents) for a particular concept.

Figure 1: Synonymy and equivalence from a terminological perspective

A semasiological approach presents us with a model as in Figure 2, in which the mediating role of the concept is absent. Here, we can regard regard synonymy synonymy and and equivalence as relations between lexemes which share the same denotative meaning; terms are said said not notto tohave have connotative connotativemeaning. meaning.

Figure 2: Synonymy and equivalence from a word-based perspective

MARGARET ROGERS

219

Figures 1 and 2 are both, however, idealised pictures. It is well-known that absolute synonymy, i.e. substitutability in all contexts, is an elusive notion, even in LSPs. The notion of substitutablity introduces another aspect of meaning beyond that of denotative meaning, namely that of syntagmatic relations, or collocability. While evidence on denotative - or conceptual - meaning may be derived from knowledge of extra-linguistic entities, processes, states, etc., evidence concerning collocability is to be found in linguistic artefacts, i.e. texts. In general language, the role of syntagmatic relations in constraining synonymy is well recognised. It is possible, for instance, to 'get a letter' and to 'receive a letter', but while it is possible to 'get a cold', it is not possible to *'receive a cold' (cf. Bußman 1990: 763 for German examples). So 'get' and 'receive' are synonymous when collocating with 'letter' but not with 'cold'. But in LSP communication, the desire to eliminate or reduce ambiguity may impose different constraints, since the use of synonyms is judged unhelpful and even misleading rather than simply an aspect of language use. One example is the alternation of synonyms within a single special-language text: here there is normally no question of text-typological variation2 and one might therefore expect to find consistency of terminological choice in the interests of unambiguous communication. However, it is well known that synonymic variation is common in special-language texts despite the best efforts of standardising bodies. It is often assumed in such cases that the author's use of synonyms rather than of a single preferred term is arbitrary and by implication sloppy. But this is an assumption which requires further investigation. The occurrence of apparently arbitrary alternative forms has, for instance, been studied extensively in sociolinguistics. By investigating the relationship between various social factors and alternating forms, it has been possible in many cases to establish variational patterns, indicating the non-arbitrary nature of such behaviour. As Hudson has pointed out, variability in speech cannot be explained "[...] in terms of some kind of general 'performance' factor, such as laziness or limitations of memory [...]" since the same performance factors do not necessarily lead to the same outcomes (Hudson 1980: 181). Instead, systematic patterns of variation need to be explained. The arbitrary or non-arbitrary nature of synonym variation is of particular importance for translators, who have to decide whether lexical variation in the ST should be retained in the TT. This brings us back to the question of synonymy and equivalence: even if the synonymy in a ST can be shown to be well motivated in some way, e.g. pragmatically or syntagmatically, what does this indicate about translation decisions for the TT? Will the pattern of variation in the TT match that of the ST in some way? Can the translator reckon with oneto-one correspondences? Or will the variational pattern in the TT, although

220

SYNONYMY AND EQUIVALENCE

equally well motivated, be a different one? In this chapter, I would like to examine how synonymic variation within one broadly-defined text type may be compared crosslinguistically with a view to identifying whether any regular patterns of equivalence may be established. The chosen domain is that of genetic engineering, a relatively new and still emergent field3; the exemplar languages are English and German. The text type is journal (not expert-to-expert); the method is corpus-based. Before moving on to the detailed study of the corpus, I would like to consider some of the implications of studying terms as word forms in text rather than as lexemes.

Terms and text Synonyms and equivalents are recorded in dictionary entries as a kind of lexeme (or lemma), i.e. a citation form which is an abstract representation of all possible word forms connected with that lexeme (lexemes will be indicated through the use of capital letters, e.g. DISEASE; word forms are shown in single inverted commas, e.g. 'disease', 'diseases', etc.). Synonymy and equivalence are defined in both terminology science and linguistics at the abstract level of the lexeme, as we have seen. Regardless of the theoretical framework, however, what is available for analysis in texts is not lexemes, but word forms which enter into textual relations as well as sense relations. It is well known, for instance, that synonymy between items in a text contributes to the cohesion of that text (cf. for instance, Halliday and Hasan 1976: 278-9; Hoey 1992: 6), one of the seven standards of textuality identified by de Beaugrande and Dressier (1981: 3). In special-language texts we can frequently observe the progressive 'stripping' of compound nouns, as in the following example from Automotive Engineering: 'closed loop control system', 'closed loop control', 'control'. There is a kind of hierarchy here which is discernible in the chosen ordering. While such hierarchies of synonyms are difficult to represent in specialist dictionaries or terminologies, I do not think that they normally present significant difficulties in text creation, since the systematicity of such variation is relatively transparent and seems non-arbitrary. Other types of synonym variation are, however, generally less amenable to our intuitions about what makes a text a text. As we have seen, a common and non-trivial problem encountered by translators is the use by ST authors of synonyms in an apparently arbitrary fashion. Dictionary sources often compound the problem by recycling the apparently synonymous terms in definitions.

MARGARET ROGERS

221

Take, for instance, the terms DISEASE and DISORDER which Levy notes (1995: 51) are apparently used randomly in English scientific writing on genetic engineering. disease a specific disorder that features a recognizable complex of physical signs, symptoms and effects Figure 3: Definition of 'disease' in Collins Medical Dictionary, 1992 The definition in Figure 3 suggests that DISEASE is actually a kind of DISORDER, i.e. a hyponym. Yet evidence from elsewhere (cf. Figure 4) confusingly suggests that a specific kind of DISORDER may be a kind of DISEASE. Clotting disorders and other diseases involving circulating gene products Figure 4: A sub-heading from a journai article on gene therapy (Mulligan 1993) For the individual translator the variable use of such terms, as also the German KRANKHEIT and ERKRANKUNG (similarly proposed by Levy as potential synonyms, Levy 1995) and the matching of these synonyms across languages may indeed seem arbitrary. So on what basis can translation decisions be made? Reference to the concept, assuming that equivalence is established, will not provide any solutions, as we have seen. Neither will reference to the sense relations between word forms in the text as paradigmatic relations. An alternative possibility is to look to the linguistic context of the terms, i.e. the syntagmatic relations between the variant forms and their linguistic context in LSP texts. Since translation decisions must take into account the behaviour of potential equivalents in the TT context, as well as their relationship with terms in the ST, it seems reasonable to assume that one TL synonym may be preferred over another TL synonym according to the linguistic context in the TT. In other words, equivalence needs to be established not only in terms of ST-TT relations, but also in terms of TT-TT relations. Moreover, the constraints imposed on lexical choice in special-language texts are likely to be tighter than those which operate in general-language texts. The kinds of linguistic factor which can be considered to potentially influence the choice of a translation equivalent from a number of possible variant forms include the following:

222

SYNONYMY AND EQUIVALENCE

morphological factors: • the ability of the variant to form compounds • the ability of the variant to form derivatives • the number (e.g. singular versus plural) of nominal variants • the conjugability of verbal variants semantic factors: • the collocability of the nominal variants with in particular governing verbs, dependent adjectives and dependent attributes syntactic factors: • the ability of the variant to change word class and to relate transparently to terms in other word classes • the syntactic function of nominal variants (eg Subject, Object) and their relation to LSP verbs In the next section, an empirical study is described which aims to investigate some of the questions raised so far concerning the factors which may influence the choice of synonyms and equivalents where such variation is not obviously motivated by pragmatic or denotational factors, but rather by immediate linguistic context.

Method If factors such as those outlined in the previous section do indeed influence patterns of variability in the choice of synonyms and hence of equivalents in the translation of special-language texts, these patterns are more likely to be detected in a corpus of texts rather then in an individual text, and they are more likely to be detected from an automatic analysis of the corpus than by manual methods, particularly if the choices are probabilistic rather than absolute. For this reason, a bilingual German and English corpus of texts in genetic engineering4 was compiled, consisting of articles from scientific publications all written between 1993-19945 (following Levy 1995). The German corpus contains 17,057 words (7 articles), the English corpus 33,872 (6 articles), of which the majority (5/6) are American English. In other words, the texts chosen are not of the highest level of abstraction in LSP such as an academic journal article written by an academic expert in genetic engineering for his or her counterpart, but are intended for a scientifically-informed but not expert readership. The English corpus covers a wider range than the German. While terminologists stress the importance of selecting expert-to-expert texts as the basis of high-quality terminology work, it was felt that the choice of less spe-

MARGARET ROGERS

223

cialised texts had certain advantages for the present analysis. Firstly, it could be expected that such texts would have a higher level of explicitness, assuming less shared knowlege between author and reader. And secondly, the use of terms could be expected to show greater variation, potentially yielding more data on synonyms and hence also on equivalence relations. From a number of synonym sets proposed by Levy (1995) and following a preliminary analysis of the corpus, it was decided to focus the analysis on the two synonym sets in Table 1. English DISEASE DISORDER ILLNESS SYNDROME

German KRANKHEIT ERKRANKUNG MORBUS SYNDROM

Table 1: Sets of synonyms in English and German from the genetic engineering corpus (after Levy 1995) The reason for selecting these sets in preference to others was simply the amount of data which they provided for analysis. The corpus was analysed using the University of Surrey's text-processing tool KonText, part of the lexical workstation System Quirk (cf. Ahmad 1995). It was decided to focus in the first instance on one morphological factor, namely, the ability of the variant to form compounds, since the majority of terms in special-language texts are normally nominal compounds, and compounds may be regarded as a kind of collocation, although this is not generally the view taken in terminology science. Number was found early on in the analysis to play a potentially significant role in the distribution of the chosen synonyms in each language and was therefore taken into account in all subsequent analyses. These aimed to identify the occurrences of the 8 selected terms in Table 1 in compounds as well as standing alone as terms in their own right. Two operations were performed: concordance and index. As a first step, the frequency distribution of each of the terms in Table 1 was established, as an initial indication of possible equivalence. The concordance showing each search term in its context was then used to identify multiword compound forms (e.g. 'mutationsbedingte Erkrankungen' in German; 'coronary artery disease' in English). Inflected forms of the terms were included in the searches. The concordance output was also used to identify the occurrence of the 8 terms as single-word terms in their own right (e.g. 'Erkrankung'; 'diseases'). Single-

224

SYNONYMY AND EQUIVALENCE

word compound terms in German were identified using an index with wild card option placed to the right or left of the search term (e.g. krankheit* to give terms such as 'Krankheitsprozeß'; *krankheiten to give terms such as 'Autoimmunkrankheiten'). Having established the compounding patterns of the selected terms, proposals were made for synonyms in each language. The issue of equivalence was then reconsidered by comparing the distribution of these compound terms crosslinguistically.

Results and Analysis Lexemes and word forms The frequency distributions of the 4 terms in each sub-corpus were calculated using the concordance outputs. The results are shown in Tables 2a and 2b, giving both absolute and relative frequencies based on the lemmatised term in the first instance. term (lemmatised) DISEASE DISORDER ILLNESS SYNDROME Table 2a: Frequency distribution of (Corpus: N=33,872) term (lemmatised) ERKRANKUNG SYNDROM KRANKHEIT MORBUS Table 2b: Frequency distribution of (Corpus: N=17,057)

ƒ 143 12 8 3

f/N (%) 0.4222 0.0354 0.0236 0.0089

DISEASE, DISORDER, ILLNESS

ƒ 81 31 18 2

and

SYNDROME

f/N (%) 0.4749 0.1817 0.1056 0.0117

ERKRANKUNG, SYNDROM, KRANKHEIT

and

MORBUS

The frequency distributions suggest that DISEASE and ERKRANKUNG are the preferred terms in English and German respectively. Indeed, their relative frequencies are close (DISEASE: 0.4222; ERKRANKUNG : 0.4749), suggesting that

225

MARGARET ROGERS

they are potential candidates for equivalence. Tables 3a and 3b show a breakdown of the word forms associated with the lemmatised terms in Tables 2a and 2b. term disease diseases

number singular plural

75 68

singular plural

2 10

singular plural

3 5

singular plural

3 0

ƒ

total disorder disorders

143

12

total illness illnesses total | syndrome | syndromes

8

3

total

f/N (%) 0.2214 0.2008 0.4222 0.0059 0.0295 0.0345 0.0089 0.0148 0.0236 0.0089 infinity 0.0089\

Table 3a: Frequency distributions of 'disease', 'diseases', 'disorder', 'illness', 'illnesses', 'syndrome' and 'syndromes'

term number (case) Erkrankung singular Erkrankungen plural total Syndrom singular (nom; acc; dat) Syndroms singular (genitive) Syndrome plural (nom; acc; gen) Syndromen plural (dative) total Krankheit singular Krankheiten plural total Morbus total

ƒ 28 53 14

'disorders',

f/N (%) 0.1642 0.3107 0.4749 81 0.0821

8 7

0.0469 0.0410

2

0.0117 31

0.1817 0.0352 0.0704

6 12 18 2

0.1056 0.0117

2

0.0117

Table 3b: Frequency distributions of 'Erkrankung', 'Erkrankungen', 'Syndrom', 'Syndroms', 'Syndrome', 'Syndromen', 'Krankheit', 'Krankheiten' and 'Morbus'

226

SYNONYMY AND EQUIVALENCE

Looking at Table 3b, it can be seen that the plural forms 'Erkrankungen' and 'Krankheiten' occur approximately twice as often as their related singular forms. By contrast, the plural forms 'Syndrome'/'Syndromen' occur in aggregate less than half as often as the singular forms 'Syndrom'/'Syndroms'. The distribution of word forms in the English synonym set is rather different, as Table 3a shows. The plural form 'disorders' is 5 times as frequent as 'disorder', whereas 'illnesses' occurs only marginally more frequently than 'illness' (although absolute frequency is low here). The singular form 'disease' is, on the other hand, marginally more frequent than its plural counterpart. In fact, if we look at the data in terms of word forms rather than lexemes, a more complex picture of equivalence emerges than that suggested by Tables 2a and 2b. So while DISEASE and ERKRANKUNG are the most frequent lexemes in each of the sub-corpora, the distribution of the word forms belonging to each of these lexemes is rather different: while just under half of the occurrences of DISEASE 47.55% (68/143) - are in the plural form, two thirds of the occurrences of ERKRANKUNG - 66.67% (64/96) - are in the plural. Of the English terms, it is DISORDER and ILLNESS which occur more frequently in the plural form (respectively: 10/12, i.e. 83.33%; 5/8, i.e. 62.5%). These figures suggest that equivalence may usefully be considered at the level of the word form rather than at the lexeme level if patterns of use in original texts are to be reflected. Single-word terms Before we turn to compounding patterns for the chosen synonyms in each language, let us first consider their occurrence as single-word terms in their own right. The frequency distributions in Tables 3a and 3b show all occurrences of the chosen synonyms as single-word terms in their own right or as components of multiword terms, as identified in the concordance outputs. Tables 4a and 4b focus on the relative frequencies of single-word terms in their own right as a proportion of all occurrences of the term: in the case of German, it was necessary to include in the data presented in Table 4b occurrences of the 4 synonyms in single-word compounds such as 'Grunderkrankung', 'Grundkrankheit' and 'Krankheitsbild' (also considered separately in Tables 9a and 9b below). These forms were not detectable from the concordance output and were identified using indexes. The results of the concordances and the indexes are merged in Table 4b. The frequency of each term is shown as a percentage of the total number of occurrences of each term in each sub-corpus.6

227

MARGARET ROGERS

term 36 24

48.00 35.29

75 68

totals

60

disorder disorders

143 2 10

1 1 1 5

12

3 0

33.3 infinity

1 69

totals grand totals

75.00

8

6 1 0

16.67 33.33 100.00

3 5

totals syndrome syndromes

41.96 50.00 10.00

2

totals illness illnesses

f/N(%)

N

ƒ

disease diseases

33.33 47.57

3 166

Table 4a: The English synonyms as single-word terms in their own right as a percentage of all occurrences (N)a \

term

f

| 10

32 a

Erkrankungen

18

64 b 28

totals Syndrom Syndroms Syndrome Syndromen

0 3 3 1

96 14 8 7 2

7

totals

f/N(%) 31.25 28.13

N

Erkrankung

29.17 infinity 37.5 42.86 50.00

31

22.58

Krankheit

3

24 c

12.50

Krankheiten

2

24 d

8.33

totals Morbus

5

total grand totals

0 40 \

10.42

48 2

0

infinity 2 177

| infinity 22.60

a. includes 4 occurrences as head of single-word compound b. includes 11 occurrences as head of single-word compound c. includes 2 occurrences as head of single-word compound and 16 as modifier of single-word compound d. includes 12 occurrences as head of single-word compound Table 4b: The German synonyms as single-word terms in their own right as a percentage of all occurrences (N)

228

SYNONYMY AND EQUIVALENCE

By looking at those forms which are most likely to appear as single-word terms in their own right, we can begin to draw some preliminary conclusions about the propensity of particular word forms to act as components of compounds. Overall, there is a much clearer tendency for the chosen lexemes to occur as terms in their own right in the English data (41.57%) than in the German (22.60%). In other words, there is more compounding (in both multi- and single-word compounds) going on in German than in English. Table 5 shows the English and German word forms in descending rank order of their observed tendency to occur as single-word terms in their own right. Those which are ranked lowest can be inferred to have a greater tendency to enter into multiword compounds in the case of English, or, in the case of German, mulitiword and single-word compounds. RANK

1 2 3 4 5 5 7

English

%

illnesses disorder disease diseases illness syndrome disorders

occurrences as single-word terms in own right 100.00 50.00 48.00 35.29 33.33 33.33 10.00

RANK

1 2 3 4 5 6 7 8 8

German

%

occurrences as single-word terms in own right Syndromen 50.00 42.86 Syndrome 37.50 Syndroms Erkrankung 31.25 Erkrankungen 28.13 12.50 Krankheit Krankheiten 8.33 Syndrom infinity Morbus infinity ]

Table 5: English and German word forms in descending rank order of their observed tendency to occur as single-word terms in their own right Table 5 highlights some further differences arising from the behaviour of word forms rather than lexemes and showing the interaction of number and compound formation. So, for instance, the plural form 'illnesses' always occurs in the present corpus as a single-word term, whereas the plural form 'disorders' is very likely to occur as part of a multiword term. In the German data, the word forms associated with the lexeme SYNDROM have the greatest tendency to appear as single-word terms in their own right with the exception of the uninflected 'Syndrom' which only occurs in compounds. Of the remaining German lexemes, KRANKHEIT is more likely to form compounds than ERKRANKUNG. If we return to our original comparison of ERKRANKUNG and DISEASE as poten-

229

MARGARET ROGERS

tial equivalents, we must note here yet a further difference (in addition to the distributional differences between singular and plural forms), namely, the propensity to form compounds. For DISEASE, just over 40% of occurrences can be accounted for by single-word terms in their own right (cf. Table 4a) whereas only c. 30% of occurrences of ERKRANKUNG can be accounted for in this way (cf. Table 4b). In other words, the German term is more likely to be part of a compound than the English term. Another point to note is that some single-word terms, e.g. 'Erkrankung', 'Erkrankungen', 'disease', 'diseases' and 'illnesses', show a tendency to appear in attributive phrases with 'von', 'der' (genitive) and 'of'. Semantically these single-word terms acting in their own right are hyperonyms of the compound terms in which the term acts as head (e.g. 'Herzerkrankung' is a kind of 'Erkrankung'; 'inherited disease' is a kind of 'disease'). Some examples are shown in Table 6. English word form

N

disease

36

diseases

24

illnesses

5

| German word form Erkrankung Erkrankungen

10 18

Examples of attributive phrases the manifestation of disease 5 different versions of disease 2 this family of diseases the treatment of serious diseases an ever-broadening range of other illnesses 2 inherited forms of illnesses sekundäre Folgen der Erkrankung 6 als Ursache einer Erkrankung Diagnostik, Therapie und Prävention von 7 Erkrankungen eine Vielzahl von Erkrankungen

ƒ

f/N 13.89 8.33 40.00

60.00 38.89

Table 6: Single-word terms in their own right occurring in attributive phrases (f) shown as percentage of occurrences of single-word terms in their own right (N)

Since the frequencies are low, further data are needed, although there seems to be a more marked tendency in the German to use the generic term in such phrases. Let us now go on to look at the actual compounds which our chosen synonyms enter into. We first look at multiword compounds in English (Tables 7a and 7b) as well as possible synonymic relations between those compounds (Table 8). We then move on to single-word and hyphenated compounds in German (Tables 9a, 9b and 9c). Finally, multiword compounds in German are

230

SYNONYMY AND EQUIVALENCE

investigated (Table 10), as well as possible synonymic relations between those compounds (Table 11). The question of equivalence is then reconsidered in the light of the newly-proposed synonym sets consisting of compound terms. Compounding patterns in English Tables 7a and 7b below show the collocations identified from the English subcorpus based on each of the synonyms in the English set. The candidate multiword compounds terms identified have the search terms functioning as the head of the multiword term (Table 7a see next page) or as its modifier (Table 7b) respectively. The frequencies of each compound are shown for both singular and plural forms. According to the evidence of the corpus, not all synonyms are freely substitutable in all combinations. For instance, in the present English sub-corpus there is no evidence for the compounds 'hereditary syndrome' or 'hereditary disease', 'immune system illness', or 'residual disease'. However, a larger corpus may indicate more possibilities. While the evidence of the English sub-corpus is that all 4 synonyms enter into multiword compounds in which they serve as the head, only DISEASE seems to function as the modifier in multiword compounds (Table 7b). candidate multiword compound term DISEASE (types: N=5) disease progression disease profile disease process disease transmission disease strains total DISORDER | none observed total ILLNESS | none observed total SYNDROME | none observed total Table 7b:

tokens (ƒ)

5 0 0 0

o\ ■o

o\

Candidate multiword terms with DISEASE, DISORDER, ILLNESS and SYNDROME

as modifier of the compound

231

MARGARET ROGERS tokens (ƒ) plural

candidate multiword compound term singular DISEASE types: N=29) ACQUIRED GENETIC DISEASE ACQUIRED DISEASE ALZHE MER'S DISEASE ANIMAL DISEASE BRAIN DISEASE CORONARY ARTERY DISEASE HUMAN BRAIN DISEASE BULK DISEASE CARDIOVASCULAR DISEASE CREUTZFELDT-JAKOB DISEASE DEGEN ERATIVE DISEASE FATAL DISEASE GENET] C DISEASE HEART DISEASE HUMAN DISEASE HUMAN GENETIC DISEASE INFECTIOUS DISEASE INHERITED DISEASE INHERITED HUMAN DISEASE LIFE-THREATENING DISEASE LIVER-BASED GENETIC DISEASE LIVER DISEASE LUNG DISEASE MAD COW DISEASE LESCH-NYHAN DISEASE PRION DISEASE RESIDUAL DISEASE SCRAPIE-LIKE DISEASE VIRAL DISEASE I

totals

34 0 0 1 0 0 0 0

totals ILLNESS (types: N=2) HEREDITARY ILLNESS VIRAL ILLNESS

44

1

totals

totals

Table 7a: Candidate multiword terms with DISEASE, as the head of the compound

2 1 3 1 1 1 1

2

0 1

0 0 2

10 1 1

0 0

1 1

78

9

2

|

1 4 1 1 2 1 1 1 6 2 2 1 20 3 7 1 1 2 1 1 1 1 1 1 1 4 2 6 2

2 1 2 1 1 1 1

1 1

SYNDROME (TYPES: N=2) DI GEORGE SYNDROME GERSTMANN-STRÄUSSLER-SCHEINKER SYNDROME |

1 4 0 1 2 0 0 0 1 0 1 0 15 0 2 1 1 2 0 0 1 0 1 0 0 4 0 6 1

0 0 1 0 0 1 1 1 5 2 1 1 5 3 5 0 0 0 1 1 0 1 0 1 1 0 2 0 1

DISORDER (types: N=7) BLOOD DISORDER CLOTTING DISORDER GENETIC DISORDER HEREDITARY DISORDER HUMAN DISORDER IMMUNE SYSTEM DISORDER IMMUNOLOGIC DISORDER

| total

0

DISORDER, ILLNESS

..

2

and SYNDROME

232

SYNONYMY AND EQUIVALENCE

Tables 7a and 7b show that the selected terms more easily assume the head role in English multiword compounds than the modifier role. It is notable that the heads of the compounds in Table 7b - 'progression', 'profile', 'process', 'transmission' - tend to fall into the category of what might be called subtechnical terms, after Trimble (1985: 29), since they have a wide distribution through many domains but a relatively low frequency in any particular text. The exception here is the more domain-specific 'strains'. Having established an inventory of candidate multiword terms in English based on the original synonym set, we are now in a position to investigate the extent to which these terms may themselves be thought of as synonyms. Table 8 lists those found in the English sub-corpus (based on Table 7a): potential synonym sets variation in modifier of compound CARDIOVASCULAR DISEASE HEART DISEASE CORONARY ARTERY DISEASE CREUTZFELDT-JAKOB DISEASE MAD COW DISEASE IMMUNE SYSTEM DISORDER IMMUNOLOGIC DISORDER | variation in head of compound HUMAN DISEASE HUMAN DISORDER HEREDITARY DISORDER HEREDITARY ILLNESS GENETIC DISEASE GENETIC DISORDER VIRAL DISEASE VIRAL ILLNESS totals

ƒ (sing.)

%

ƒ (plu.)

%

5 3 1 2 1 0 0

90.00

1 0 0 100.0 0 0 0.00 1 1

0.00

5 0 0 1 5 1 1 1

62.50

37.50

50.00 26.10 66.67 26

10.00

100.0

2 1 1 0 15 2 1 0

50.00 73.90 33.33

25

total 1 (100%) 6 3 1 2 1 1 1 7 1 1 1 20 3 2 1 51

Table 8: Potential synonym sets for multiword compounds in English showing frequencies of singular and plural f orms

The synonymy proposed for the various sets shown in Table 8 arises in 3 cases from the variation of the modifier (e.g. CARDIOVASCULAR/HEART) and in 4 cases from the variation of the head (e.g. DISEASE/DISORDER). Whilst the most common head variation is between DISEASE and DISORDER, both may also alternate with ILLNESS. Where there is variation in the modifier rather than the head, the patterns of distribution between singular and plural forms is

MARGARET ROGERS

233

relatively consistent between synonyms. This suggests that, in determining the interchangeability of apparently synonymous compound terms in translation, formal aspects such as head or modifier variation need to be taken into account. The present data suggest that interchangeability is greater for those compounds in which the modifier is variable and the head constant. Let us now turn to the German sub-corpus, looking first at single-word compounds. Compounding patterns in German The results of analysing the compounding patterns of the German search items are shown in Tables 9a (head) and 9b (modifier) for single-word compounds. Hyphenated compounds - an increasing trend in modern German word formation - in which the search term is the head are shown separately in Table 9c. tokens

candidate single-word compound term

ERKRANKUNG (types: N=9) ADENOSINDEAMINASEMANGELERKRANKUNG AUTOIMMUNERKRANKUNG GRUNDERKRANKUNG HERZERKRANKUNG HIRNERKRANKUNG INFEKTIONSERKRANKUNG LUNGENERKRANKUNG PRIONERKRANKUNG TUMORERKRANKUNG totals SYNDROM (types: N=0) none observed totals KRANKHEIT (types: N=6) AUTOIMMUNKRANKHEIT AUTOIMMUNOKRANKHEIT ERBKRANKHEIT GRUNDKRANKHEIT IMMUNSCHWÄCHEKRANKHEIT INFEKTIONSKRANKHEIT totals MORBUS (TYPES: N=0) none observed totals

(f)

singular

plural

total

1 0 1 1 1 0 0 0 0

0 1 0 0 0 2 1 1 6

1 1 1 1 1 2 1 1 6

4

11

0

0

0 0 0 1 1 0

7 1 1 0 0 3

1 1 1 1 1 3 14 0

0 0

°\

12

2 0

15

0

o\

Table 9a: Candidate single-word compound terms with ERKRANKUNG, SYNDROM, KRANKHEIT and MORBUS as the head of the compound

234

SYNONYMY AND EQUIVALENCE

Table 9a shows that both ERKRANKUNG and KRANKHEIT tend to occur in their plural forms (11/15, i.e. 75.33% and 12/14, i.e. 85.71% respectively) as heads of the single-word compounds shown. When functioning as modifier in singleword compounds, however, it is KRANKHEIT which is the most productive component, as shown in Table 9b. candidate single-word compound term ERKRANKUNG (types: N=0) | none observed | total SYNDROM (types: N=0) | none observed | total KRANKHEIT (types: N=6) | Krankheitsbild | Krankheitsempfindlichkeit Krankheitsentstehung Krankheitsprozeß Krankheitsverständnis | Krankheitszustand | total MORBUS (types: N=0) | none observed total

tokens (ƒ)

|

0

o\ 0

o\ 3 1 2 4 5 1

16\ 0

o]

Table 9b: Candidate single-word compound terms with ERKRANKUNG, SYNDROM, KRANKHEIT and MORBUS as modifier of the compound

As with the English compounds in which the search term functions as modifer (cf. Table 7b above) the head of the compound in German may also be described as a subtechnical term: 'Bild', 'Empfindlichkeit', 'Entstehung', 'Prozeß', 'Verständnis' and 'Zustand'. In Table 9c we return to the search term as head of the compound, but this time in compounds in which the components are orthographically linked by hyphens.

MARGARET ROGERS

235

candidate hyphenated compound term

tokens

ERKRANKUNG (types: N=6) ALZHEIMER-ERKRANKUNG GSS-ERKRANKUNG HERZ-KREISLAUF-ERKRANKUNG JACOB-CREUTZFELDT-ERKRANKUNG JC-ERKRANKUNG PARKINSON-ERKRANKUNG WILSON-ERKRANKUNG

plural

total

1 1 0 2 1 1 1

0 0 2 0 0 0 0

1 1 2 2 1 1 1

totals SYNDROM (types: N=10) DELETIONS-SYNDROM DOWN-SYNDROM ?HYPER-IGM-SYNDROM LESH-NYHAN-SYNDROM MARFAN-SYNDROM MIKRODELETIONS-SYNDROM PRADER-WILLI-SYNDROM TURNER-SYNDROM WAGR-SYNDROM X-SYNDROM |

(f)

singular

2

7 0 5a 1 3a 1

totals KRANKHEIT (types: N=l) PELIZAEUS-MERZBACHER-KRANKHEIT 1 | totals MORBUS (TYPES: N=0) | none observed 0 totals

1 0

1 5

0 0

1

5 1 1 1 1

0

1 0

0 0

20

4

1

]

1

0 0 0 0 16

|

3

0 3b

2a 1 1 1 1

9

1 0

0

o\

Notes: a. includes genitive form/s b. includes dative form Table 9c: Candidate hyphenated compound terms with ERKRANKUNG, SYNDROM, KRANKHEIT and MORBUS as head of the compound

The data in Table 9c show that when part of hyphenated compounds ERKRANKUNG and SYNDROM occur primarily as names of particular diseases, and in this role in their singular forms. This is in contrast to the pattern established for

236

SYNONYMY AND EQUIVALENCE

as a single-word term and, as we shall see, multiword terms, both of which appear more frequently in their plural forms. The multiword compound terms identified in the German sub-corpus are shown in Table 10. The pattern which seems to emerge is that both ERKRANKUNG and KRANKHEIT tend to appear in their plural forms in multiword compounds in which they function as the head (33/44, i.e. 75.00% and 10/12, i.e. 83.33% respectively). In this respect, the German KRANKHEIT has a similar distribution pattern within multiword compounds to that of the English DISORDER (cf. Table 7a). We can recall that 90.00% (9/10) of multiword compounds with DISORDER as head occur in the plural form.

ERKRANKUNG

The data presented in Tables 9a, 9b, 9c and 10 suggest the sets of synonyms as shown in Table 11. As with the English data on possible synonym sets for compound terms (cf. Table 8), variation occurs in both head and modifier roles. The commonest alternation in variant forms for the head role in German is that between ERKRANKUNG and KRANKHEIT. If we focus on those occurrences where the variation occurs in the modifier, then there is a notable tendency - as in the English - for the distributional pattern between singular and plural to be maintained across variants, although absolute frequencies are low. However, in contrast to the English data, there is also a fairly consistent pattern of singularplural distribution between some proposed synonym sets in which the head varies. In other words, the German synonyms ERKRANKUNG, KRANKHEIT and SYNDROM seem, on the basis of the limited data available here, to be more interchangeable than the English DISEASE, DISORDER and ILLNESS. What emerges from the sets of synonyms for compounds proposed both for English and for German is that synonymy is not a case of synonymy between the lexemes DISEASE, DISORDER, ILLNESS and SYNDROME or between the lexemes ERKRANKUNG, KRANKHEIT, SYNDROM and MORBUS, but rather between these terms operating in conjunction with particular modifiers in compounds in the singular or plural form. Summary At the beginning of this analysis, it was suggested that, on the basis of the relative frequency of lexemes, ERKRANKUNG and DISEASE were potential candidates for equivalence. However, we have seen that these lexemes, when analysed from the point of view of word forms, show different distributional and compounding tendencies. Tables 12a and 12b summarise these data for English and German respectively.

237

MARGARET ROGERS

Candidate multiword compound term singular ERKRANKUNG (types: N=24) 3-CHROMOSOMENLOCUS GEBUNDENE ERKRANKUNG 1 BAKTERIELLE ERKRANKUNG 0 DEGENERATIVE ERKRANKUNG 1 ERBLICHE ERK RANKUNG 4 ERBLICHE NEUROLOGISCHE ERKRANKUNG 0 ERWORBENE GENETISCHE ERKRANKUNG 0 GENETISCH BEDINGTE ERKRANKUNG 1 0 GENETISCHE ERKRANKUNG GERSTMANN-STRÄUSSLER ERKRANKUNG 1 HLA-DO-ASSOZIIERTE ERKRANKUNG 0 INFEKTIÖSE ZNS-ERKRANKUNG 0 KARDIOVASKULÄRE ERKRANKUNG 0 MENSCHLICHE ERKRANKUNG 0 MONOGENE ER KRANKUNG 1 MULTIFAKTORIELL BEDINGTE ERKRANKUNG 0 MUTATIONSBEDINGTE ERKRANKUNG 0 NEURODEGENERATIVE ERKRANKUNG 0 NEUROLOGISCHE ERKRANKUNG 0 ONKOLOGISCHE ERKRANKUNG 0 0 PARASITÄRE ERKRANKUNG PSYCHIATRISCHE ERKRANKUNG 0 0 VERERBTE ERKRANKUNG VIRALE ERKRANKUNG 0 WILSON'SCHE ERKRANKUNG 2 totals SYNDROM (types: N=4) GENETISCHES SYNDROM 0 0 NEUROLOGISCHES SYNDROM TRICHORHINOPHALANGEAL-SYNDROM TYP I 1 TRICHORHINOPHALANGEAL-SYNDROM TYP II 1 totals KRANKHEIT (types: N=7) FAMILIÄRE ALZHEIMER KRANKHEIT 1 GENETISCH BEDINGTE KRANKHEIT 0 GENETISCHE KRANKHEIT 0 HETEROGENE KRANKHEIT 1 LETALE KRANKHEIT 0 NEURODEGENERATIVE KRANKHEIT lo NEUROLOGISCHE KRANKHEIT 0 totals MORBUS (types: N=2) MORBUS BECHTEREW MORBUS ALZHEIMER

totals

tokens (f) plural total 1 1 2 5 2 1 3 5 1 1 1 1 1 1 2 1 2 6 1 1 1 1 1 2

0 1 1 1 2 1 2 5' 0 1 1 1 1 0 2 1 2 6 1 1 1 1 1 0 11 1 1 0 0

1 1 1 1

2

2 0 4 2 0 1 1 2

2

4 i 4 2 1 1 1 2

2 1 1

44

33

10

o o

12\ 1 1

o1

2\

Table 10: Candidate multiword terms with ERKRANKUNG, SYNDROM, KRANKHEIT and MORBUS as head of the compound

238

SYNONYMY AND EQUIVALENCE

potential synonym sets | variation in modifier of compound GERSTMANN-STRÄUSSLER ERKRANKUNG GSS-ERKRANKUNG HERZERKRANKUNG HERZ-KREISLAUF-ERKRANKUNG KARDIOVASKULÄRE ERKRANKUNG JACOB-CREUTZFELDT-ERKRANKUNG JC-ERKRANKUNG WILSON'SCHE ERKRANKUNG WILSON-ERKRANKUNG ONKOLOGISCHE ERKRANKUNG TUMORERKRANKUNG variation in head of compound INFEKTIONSERKRANKUNG INFEKTIONSKRANKHEIT GRUNDERKRANKUNG GRUNDKRANKHEIT GENETISCHE ERKRANKUNG GENETISCHE KRANKHEIT GENETISCHES SYNDROM MORBUS ALZHEIMER ALZHEIMER-ERKRANKUNG ?FAMILIÄRE ALZHEIMER KRANKHEIT NEUROLOGISCHE ERKRANKUNG NEUROLOGISCHE KRANKHEIT NEUROLOGISCHES SYNDROM NEURODEGENERATIVE ERKRANKUNG NEURODEGENERATIVE KRANKHEIT ?DEGENERATIVE ERKRANKUNG GENETISCH BEDINGTE ERKRANKUNG GENETISCH BEDINGTE KRANKHEIT VERERBTE ERKRANKUNG ERBKRANKHEIT ERBLICHE ERKRANKUNG AUTOIMMUNERKRANKUNG AUTOIMMUNKRANKHEIT AUTOIMMUNOKRANKHEIT ?IMMUNSCHWÄCHEKRANKHEIT [ totals

ƒ (sing.)

%

ƒ (plu.)

%

total (100%)

1 1 1 0 0 2 1 2 1 0 0

100.0

o

0.00

1 1 1 2 1 2 1 2 1 1 6

0 25.00

100.0

o

75.00

2 1 0 0

0.00

100.0

0.00

o

0 0 1 1 0 0 0 1 1 1 0 0 0 0 0 1 1 0 0 0 4 0 0 0 1

0.00

1 6

100.0

0.00

2 3

100.0

100.0

o

0.00

0.00

100.0

0.00

20.00

14.29 57.14

10.00

21

0 5 2 1 0 0 0 6 2 1 2 1 1 2 4 1 1 1 1 7 1 0

100.0

0.00

100.0

80.00

85.71 42.86

90.00

54

2 3 1 1 5 2 1 1 1 1 6 . 2 1 . 2 1 2 3 4 1 1 5 1 7 1 1 75\

Table 11 : Potential synonym sets for single-word, hyphenated and multiword compounds in German showing frequencies of singular and plural f orms

MARGARET ROGERS

239

Tables 13a and 13b show the environments in which each of our original synonyms is most and least likely to appear. Of the 4 lexemes in Table 13a, ILLNESS is the most likely to occur as a singleword term in its own right, DISEASE to a lesser extent. These lexemes behave differently, however, with regard to their distribution between singular and plural forms (cf. Table 12a). ILLNESS tends to occur more frequently in the plural form (5/6, i.e. 83.33%) whereas DISEASE occurs more frequently in the singular form (36/60, i.e. 60.00%). As noted earlier, DISORDER is the most likely to form multiword compounds in which it functions as the head; the majority of these compounds appear in the plural form (9/10, i.e. 90.00%). DISEASE also appears frequently in this environment, but relatively less so than DISORDER. However, when it does, it tends to occur in the plural form (44/78, i.e. 56.41%) in contrast to its behaviour as a single-word term where it appears more frequently in the singular (36/60, i.e. 60.00%). DISEASE is the only lexeme which functions as the modifier of a multiword term. Of the 4 lexemes in Table 13b, ERKRANKUNG is the most likely to appear as the head of a multiword term and as a single-word term in its own right. In both these environments, it tends to occur more frequently in the plural form (respectively: 33/44, i.e. 75%; 18/28, i.e. 64.29%) (cf. Table 12b). KRANKHEIT is the most likely to occur in single-word compound terms, both as head and as modifier. When functioning as the head of the term, it tends to occur in the plural form (12/14, i.e. 85.71%) as also ERKRANKUNG. SYNDROM is the lexeme which is most likely to occur as the head of a hyphenated term, but in this case, the singular form is more common (16/20, i.e. 80.00%), since, as we can recall, these terms are mostly the names of particular diseases. KRANKHEIT is less likely than either ERKRANKUNG or SYNDROM to appear as a term in its own right, and may therefore be regarded as a 'carrier' in this LSP, as also DISORDER in the corresponding English LSP. Equivalence What implications does such a complex picture have for equivalence? One way of looking at the problem is to compare the tendencies of the different lexemes to participate in compounding, as already indicated in Table 5. Comparisons along these lines would suggest that the roles performed by the lexemes can be compared as in Table 14.

240

SYNONYMY AND EQUIVALENCE

singleword term in own right ƒ DISEASE singular plural totals DISORDER singular plural totals ILLNESS singular plural totals SYNDROME singular | plural totals

multi­ word term (head) ƒ

f/N %

60

33 100

1 5

33 0

1 0 1

0 0

3 5

0 0

67

8

0

0 0

2

12

0

0

67 0

33

2 10

0

25

2 0

143\

3

0 0

2

75 68

0 0

83 67 0

(N) 100%

%

0 0

10

75

6

f/N

5

50 90

2 0

total

1 0

55

1 9 17

2

5 0

78

42 50 10

1 1

45 65

f/N

34 44

48 35

36 24

%

multi­ word term (mod) ƒ

3 0

0

3

0

Table 12a: Summary of English data singleword term in own f/S right ƒ %

singleword com­ pound term (head) f/N ƒ %

singleword com­ pound term (mod) f/N

f

ERKRANKUNG 13 singular 10 31 4 0 17 28 11 plural 18 0 16 28 29 15 totals KRANKHEIT singular 13 2 3 8 16a 12 50 2 8 0 plural 14 29 5 10 totals SYNDROM | singular 14 0 3 0 0 4 40 | plural 0 0 0 23 0 0 totals 7 MORBUS singular 0 0 0 0 0 0 0 0 plural 0 0 0 0 totals 0 0

hyph­ enated term f/N ƒ

%

ƒ

0 0

7 2

22 3

11 33 9

67

1

4

2

0

0

0

10

16

a. The singular has been treated as the default torm. Table 12b: Summary of German data

9

0

53 0 0

0

16 4 0

0 0

2 76 2 40 65 20

0 0 0

3

1

2 0

0 0 0

total

f/N

%

0

0

multi­ word term (head)

0

(N) 100

% %

32 34 64 52 44 46 96 8

24

24 42 12 25 48 21 10 20 10 4 13 3l 100 2 0 0 2 100 2

241

MARGARET ROGERS

DISEASE

% single-word term in 42 own right multiword term 55 (head) multiword term 3 (modifier) 100

DISORDER

ILLNESS

rank 2

% 17

rank 2

% 75

rank 1

% 33 a

rank | 2

1

83

1

25

2

67^

1

3

0

3

0

3

0

3

100

100

SYNDROME

100

Note:

a. represents an absolute frequency of only 3 in aggregate. Table 13a: 'Preferred' environments for DISEASE, DISORDER, ILLNESS and SYNDROME

The parallels in Table 14 suggest that, according to their role within the bounds of a term, ILLNESS and ERKRANKUNG may be in some sense regarded as equivalents (single-word terms in their own right), as may also DISORDER and ERKRANKUNG (head of multiword term) and arguably DISORDER and KRANKHEIT (head of single-word compound term) since the English data has no evidence of single-word compounds. Finally, DISEASE and KRANKHEIT could also be considered equivalent since they both function as modifiers in compounds (multiword and single-word respectively). This indicates that the one-to-one relations which normally characterise entries in bilingual dictionaries do not adequately describe the complex syntagmatic relations which words enter into. To conclude, let us look more specifically at the possible equivalence relations for the compounds identified in our analysis, including the synonym sets proposed for each language (cf. Tables 8 and 11). The comparison is set out in Table 15. The proposed equivalences shown in Table 15 indicate a relatively high degree of synonymic variation in both languages, giving some insight into the problems faced by the translator when attempting to decide on equivalents in this subject field, even at comparable levels of communication. While the distributional data could form the basis of standardisation decisions concerning preferred terms and their equivalents, probabilistic trends such as the tendency of the term to appear in the singular or plural form play no role in the standardisation of concepts and their designations, i.e. terms.

242

SYNONYMY AND EQUIVALENCE

ERKRANKUNG rank % 2 single-word term in 29 own right single-word term 3 16 (head) single-word term 0 5 (modifier) 4 hyphenated term 9 1 multiword term 46 (head) 100

KRANKHEIT rank 4 10

%

SYNDROM rank % 23 2

MORBUS % rank 0 2

29

2

0

4

0

2

33

1

0

4

0

2

3 25

5 3

65 13

1 3

0 100 a

2 1

100

100

100

Note: a. represents an absolute frequency of only 2. Table 13b: 'Preferred' environments for ERKRANKUNG, KRANKHEIT, SYNDROM and MORBUS

preference in English (% occurrences in plural) single-word term in own ILLNESS right (83.33%) multiword term (head) DISORDER (90.00%) single-word term (head) no occurrences environment

| multiword term (mod.) single-word term (mod.) hyphenated term

DISEASE no occurrences no occurrences

preference in German (% occurrences in plural) | ERKRANKUNG (64.29%) ERKRANKUNG (75.00%) KRANKHEIT (85.71%) no occurrences | KRANKHEIT SYNDROM (20.00%)

Table 14: Parallel functions of the selected lexemes in English and German

Conclusion In this essay, I have considered the linguistic behaviour of two sets of potential synonyms in English and German texts from the domain of genetic engineeering, based on a corpus of texts aimed at a scientific but not necessarily expert readership. The analysis has highlighted a number of constraints which are of relevance to translators as text creators.

243

MARGARET ROGERS English GENETIC DISEASE GENETIC DISORDER DEGENERATIVE DISEASE

CREUTZFELDT-JACOB DISEASE MAD COW DISEASE HUMAN DISEASE HUMAN DISORDER VIRAL DISEASE VIRAL ILLNESS HEREDITARY DISORDER HEREDITARY ILLNESS

German ƒ sing plu 5 [ 15 GENETISCHE ERKRANKUNG GENETISCHE KRANKHEIT 1 2 GENETISCHES SYNDROM 1 1 ?DEGENERATIVE ERKRANKUNG NEURODEGENERATIVE ERKRANKUNG NEURODEGENERATIVE KRANKHEIT

2 1 5 0 1 1

0 1

0 0 2 1 1 0 !1 0

GERSTMANN-STRÄUSSLERSCHEINKER SYNDROME

l

0

CARDIOVASCULAR DISEASE HEART DISEASE

6 3

1 0

BRAIN DISEASE

1

2

LUNG DISEASE PRION DISEASE IMMUNE SYSTEM DISORDER IMMUNOLOGIC DISORDER

0 0 0 0

1 4 1 1

ALZHEIMER'S DISEASE

1

0

LESCH-NYHAN DISEASE INFECTIOUS DISEASE

1

0 1

0

ƒ sing plu 0 !5 2 0 1 0 1

1

0

2

0

1

2 | 1 0 MENSCHLICHE ERKRANKUNG

0 0 1

VIRALE ERKRANKUNG

0

1

ERBLICHE ERKRANKUNG VERERBTE ERKRANKUNG ERBKRANKHEIT GERSTMANN-STRÄUSSLER ERKRANKUNG GSS-ERKRANKUNG HERZ-KREISLAUFERKRANKUNG KARDIOVASKULÄRE ERKRANKUNG HERZERKRANKUNG

4 0

o 1

1 1 1

1 1

0 0

0

2

0 1 1

1 0 0

0 0 0 0 0 1 0 0 0 1 1 1

6 2 1 1 1 0 1 7 1 0 0

JACOB-CREUTZFELDTERKRANKUNG JC-ERKRANKUNG

HIRNERKRANKUNG NEUROLOGISCHE ERKRANKUNG NEUROLOGISCHE KRANKHEIT NEUROLOGISCHES SYNDROM LUNGENERKRANKUNG PRIONERKRANKUNG IMMUNSCHWÄCHEKRANKHEIT AUTOIMMUNERKRANKUNG AUTOIMMUNKRANKHEIT IMMUNSCHWÄCHEKRANKHEIT MORBUS ALZHEIMER ALZHEIMER-ERKRANKUNG ?FAMILIÄRE ALZHEIMER KRANKHEIT LESH-NYHAN-SYNDROM INFEKTIONSKRANKHEIT INFEKTIONSERKRANKUNG

3

p

[0 Table 15: Candidate equivalent compounds based on corpus citations

0 p

3 [2

244

SYNONYMY AND EQUIVALENCE

The data which has been analysed suggests that the notion of synonymy, and by extension equivalence, needs to take into account a number of factors at the linguistic level which are not usually considered in terminology work or indeed in lexicographical work. These factors include the role of the grammatical category number and the role of the combining elements in compound formation as head or modifier. The notion of synonymy, if seen as a relation between word forms rather than as a relation between decontextualised lexemes, is more highly constrained, revealing many complex relations of overlap and exclusion which are neither logically predictable in all cases nor, it would seem, amenable to standardisation procedures, which deal in absolutes rather than tendencies, and, of course, with lexemes rather then word forms. Notes 1 This is a revised version of a paper originally given at the 10th European Symposium on Language for Special Purposes, Multilingualism in Specialist Communication, Vienna, 29 August - 1 September 1995. 2 It should be noted, however, that some texts do 'shift' type. For example, customer information accompanying an electrical product may contain legal information regarding the conditions of guarantee as well as persuasive text on the benefits of the product. 3 Genetic engineering is a relatively new science which developed fast during the 1980s and continues to find new applications. As a new subject field in the biological sciences, its terminology is still likely to be developing, drawing on related fields, and evolving its conceptual structures. A phenomenon commonly associated with the emergence of new ideas is that of synonymy. In any developing subject field, there is likely to be a creative tension between incremental shifts in the defining characteristics of concepts on the one hand and the requirement for normalisation on the other hand (cf. for instance, de Beaugrande 1989 on the subject field of linguistics). 4 I would like to thank Iris Levy for providing me with the corpus, as well as for posing the original question about synonymic variation in relation to scientific papers in genetic engineering (cf. Levy 1995). 5 The texts are taken from the following publications: English - Scientific American, Science, Annu. Rev. Biochem.,New Scientist; German - Internist, Deutsches Ärzteblatt. 6 Except for 5 occurrences of 'disease' which may be accounted for by adjectival compounds (e.g. 'disease-causing', 'disease-specific') and 2 occurrences of 'diseases' in what might be called 'pre-terms', e.g. 'diseases of the central nervous system'. References Ahmad, Khurshid. 1995. "Pragmatics of specialist communications: Workbenches and the engineering of special languages". In H. Picht and G. Budin (eds), Biannual Conference on Terminology and (Micro) Computer Applications. Vienna: INFOTERM/UNESCO, 5-52. Bußman, Hadumod. 1990. Lexikon der Sprachwissenschaft. Stuttgart: Kröner. 2te Auflage.

MARGARET ROGERS

245

de Beaugrande, Robert. 1989. "Special purpose language as a complex system: The case of linguistics". In C. Laurén and M. Nordman (eds), Special Language: From Humans Thinking to Thinking Machines. Clevedon: Multilingual Matters, 3-29. de Beaugrande, Robert and Dressier, Wolfgang.U. 1981. Introduction to Text Linguistics. London: Longman. Felber, Helmut. 1984. Terminology Manual. Paris: UNESCO and INFOTERM. Halliday, Michael and Hasan, Ruqaiya. 1976. Cohesion in English. London: Longman. Hoey, Michael. 1991. Patterns of Lexis in Text. Oxford: OUR Hudson, Richard A. 1980. Sociolinguistics. Cambridge: CUR Levy, Iris. 1995. "Das Prinzip der Gentherapie" and "Molekularbiologie in der Neorologie". A Translation with Commentary. MA Dissertation: University of Surrey, Guildford. Mulligan, R.C. 1993. "The basic science of gene therapy". Science 260: 926-31. Trimble, Louis. 1985. English for Science and Technology - A Discourse Approach. Cambridge: CUP.

Reference Corpora and Lexicons for Translators and Translation Studies Carol Peters1 and Eugenio Picchi2 1 Istituto di Elaborazione della Informazione, CNR, Pisa, Italy 2 Istituto di Linguistica Computazionale, CNR, Pisa, Italy

Introduction Translators, especially professional translators, need fast and flexible tools to assist them in the task of rendering a text written in one language as skilfully as possible in another. They require rapid and reliable access to authoritative sources of reference (both dictionaries and texts) that can provide them with guidance in selecting the most appropriate translation equivalent for a given word or expression used in a given context, to suit the particular text type being treated. The requirements of teachers or scholars of translation skills and techniques are similar: they need procedures which permit them to access documented evidence on the possible realisation of a concept in different languages, according to a number of contextual factors, e.g. usage, style, register, domain, etc., in order to be able to perform contrastive text analyses. The difference between the two groups is mainly one of emphasis: whereas speed and precision of retrieval are of major importance in the first case, in the second priority is generally given to the possibility of obtaining responses which are as complete and exhaustive as possible. The aim of this paper is to give an idea of the latest results in the field of Computational Linguistics which directly interest both translation studies and the work of the translator. Our attention is concentrated on recent developments in the areas of computational lexicography and corpus linguistics, i.e. on the creation of electronic lexicons and language reference corpora and the design and implementation of procedures to analyse and query them. Although such tools can provide important data for natural language processing (NLP) and machine translation (MT), we are here primarily concerned with machine-aided

248

REFERENCE CORPORA AND LEXICONS

human translation activities. The following two sections thus briefly present the current state of the art in these areas and, as the topic of this book regards approaches to text typology in translation, in the second of the two sections, we concentrate on one of the latest trends in corpus linguistics: the construction of bilingual and multilingual reference corpora. The implications to both theoretical studies and practical applications in the near future of the availability of enormous quantities of different types of contrastive text data are discussed. For this reason, after presenting our integrated mono- and bilingual lexical database system in the fourth section, we describe our bilingual text management system in the fifth section and discuss how the system can be used to help process and analyse different types of texts and extract useful cross-linguistic data from them. This system, together with a system for monolingual corpus management and interrogation, is one of the two main components (the other is the lexical database system) of the prototype Translators' Workstation, which has been implemented at the Istituto di Linguistica Computazionale (ILCCNR), Pisa, and is outlined in the final section.

Electronic dictionaries Until very recently the main tool available to help the translator has been the printed dictionary. The first search for a suitable word or phrase to express a concept in another language is normally made in the bilingual dictionary which provides direct access to the target language by supplying lists of possible translation equivalents for a given word or phrase. The monolingual dictionary is then often used to verify the accuracy and appropriateness of the translations proposed by the bilingual. The disadvantages are evident. The times needed for this kind of manual dictionary look-up are considerable, especially when a multiple search is involved, and imply a break in concentration which often has negative effects on the quality of the translation being produced. Even more important, the static, alphabetical ordering of the printed volume means that much of the information actually contained in mono- and bilingual dictionaries cannot be found because there is no valid access key. The user simply does not know how to get at it. In addition, the amount of information that can be contained in the printed dictionary is clearly limited by factors of both cost and space. For these reasons (among others), much research in computational lexicography in the late 80's focused on the construction of mono- and bilingual lexical databases which offer dynamic, hypertextual access to dictionary data, making it possible to retrieve all kinds of lexical information (phonologic, mor-

CAROL PETERS AND EUGENIO PICCHI

249

phologic, syntactic, semantic) wherever it is stored in the lexicon. In particular, much work was done on the implementation of procedures to analyse and represent computationally the lexical data contained in machine readable versions of published dictionaries. As examples of first results in this field we can cite the pioneer work by Boguraev (1986), Byrd et al. (1987), and also Picchi et al. (1990). The starting point for all studies was the definition of a computational model of the lexical entry (see, for example, the computational model of the dictionary entry defined for the ACQUILEX project in Calzolari et al. 1990). This led to much discussion on the formalisms to be adopted, the types of information to be included, and the most suitable structure to represent the complex network of relationships between the different elements in the lexicon. The aim was to provide a tool that would permit the construction of different paths for navigation through the lexical data. The dictionary publishing world followed these studies with interest and, in addition to providing data for research purposes, was at times directly involved in them. Most dictionary publishers are now aware that the computer should not be limited to the final stage of the publishing process but should be involved in all stages of production, from the initial definition of the entry structure and acquisition of citation material from machine readable text archives to the printing or producing in electronic form of the finished product. The benefits are manifold: a computational model permits a tight control over the work of the lexicographer, ensuring far greater consistency and coherency in the data, and facilitating revisions and updates. But perhaps the main advantage for the commercial world is the reusability of the computerised lexicon. This is the reason why major dictionary publishers are now building up large, exhaustive lexical databases from which to generate their dictionaries. Not only can many different types of traditional printed dictionaries be printed from a single comprehensive database, i.e. learners' dictionaries, lexicons for special languages, dictionaries for synonyms and antonyms, etc., but from the same source it is also possible to produce electronic dictionaries of various types and dimensions. Computerised dictionaries are thus becoming increasingly common on the market and are rapidly acquiring popular consensus. The advantages of fast, flexible access to the data far outweighing the pleasure of a leisurely but limited consultation. This expansion in the diffusion of the electronic dictionary has been greatly encouraged by the impressive technological advances over the last few years, especially in the hardware area. Prices have fallen and it is now possible to compress and store enormous volumes of data on lightweight disks or CD-ROM's which can be loaded on inexpensive personal or portable PC's. This means that such resources are easily accessible not only for the profes-

250

REFERENCE CORPORA AND LEXICONS

sional translator employed in an office environment but also for the free lancer working at home. A number of manufacturers and publishing houses have recently released systems which combine text processors with access to monolingual and bilingual dictionaries and spelling checkers. The user, writing or editing his/her document, can invoke the system dictionary(ies), usually in separate windows, and can move back and forth between the text in construction and the reference sources, as and when he/she needs them. Although these systems are certainly of great utility to translators, providing them with a more flexible tool for dictionary look-up and helping to speed up production, they still present considerable limits. The dictionaries maintained by such systems tend to be relatively small and, despite the provision of some functions for dynamic access, the lexical information provided is generally restricted to that traditionally supplied by the printed dictionary. Further on in this paper, we will describe the latest developments in the prototype mono- and bilingual lexical database system we have installed in Pisa, in order to give an idea of the type of advanced dictionary interrogation tools which should be available to the public in the not too distant future. However, recently a second important reference source is being proposed for the translator: the bilingual text corpus. The main difference is that, whereas the bilingual dictionary provides translation equivalents for lexical items considered in isolation as headwords (with at most a few associated examples generally of particular cases), the bilingual corpus provides translations for words or expressions according to the particular context. Therefore, while the answers in the first case are given at the formal level of the lemma and can thus be considered as lexical hypotheses, in the second case the results represent real world documented evidence of how utterances or samples of text in one language can be rendered in another, according to the particular type of text being treated. In the next section, we will describe recent research in this area.

Corpus linguistics and the development of reference corpora Corpus linguistics is that branch of linguistics which deals with the empirical study of large bodies of language. Over the last decade there has been considerable growth of interest in corpus construction as language reference corpora, i.e. representative collections of texts in machine readable form, have become widely recognised as being important sources of information, not only for lexicographic purposes - in dictionary construction - but also in many other types of linguistic studies including the acquisition of knowledge for natural language processing systems1. There is general agreement with Sinclair (1987b)

CAROL PETERS AND EUGENIO PICCHI

251

that "the ability to examine large text corpora in a systematic manner allows access to a quality of evidence that has not been available before". Indeed it is felt that a sufficiently large or truly representative corpus will make it possible to document the whole range of linguistic structures for a given language or sublanguage. For this reason the dimensions of text collections have increased enormously over the last few years. This construction of very large corpora has been encouraged by the recent technological evolution in optical character readers and data storage devices, already mentioned above, which means that the rapid acquisition, maintenance and management of huge volumes of data at relatively low costs is becoming increasingly feasible. In addition, the adoption of computerised procedures in the production of most written material nowadays (both for publication and in the office environment), means that a vast amount of all types of texts in machine readable form are potentially available for corpus purposes. Monolingual corpora At the beginning of the eighties, the Brown Corpus of contemporary American English (Francis and Kucera 1982) was considered large at one million words; nowadays, many text collection initiatives have samples of text running into hundreds of millions of words. The underlying assumption motivating this interest, already summarised by Firth in the fifties: "You shall know a word by the company it keeps" (Firth 1957), was that a word acquires sense from its context. Consequently, when empirically based studies regained popularity in the eighties, it was felt that to study language in depth it was necessary to have access to large and representative samples of written and spoken text data. As was evidenced, for example by the lexicographers working on the Cobuild corpus (the basic corpus of general contemporary English used to produce the Cobuild dictionary was around 20 million words, backed up by a range of more specialised texts coming to a total of about another 20 million, see Sinclair 1987a), a word used in a given sense very often reveals not only typical syntax but typical patterns of lexical cooccurrence. This has considerable implications for many language studies and is of course of great importance for translators attempting to generate in a second language. They need as much information as possible not only on the semantics of a given word, which can presumably be found in the dictionary, but on its combinatorial properties, i.e. its preferred syntax or structural patterns and collocations. To some extent, the Cobuild dictionaries have met this need with their introduction of a new kind of definition statement or "equation" in which, while the right part gives an explanation of the meaning - similar to that of traditional dictionaries, the left part presents the headword in its typical context. However, clearly no printed dictionary can do

252

REFERENCE CORPORA AND LEXICONS

much more than supplying just one or two proto-typical examples of the lexical item in use, selected in the Cobuild case from the wide range of concordances found for that particular item in the corpus, whereas ideally the language learner would benefit from free access to the entire, or a significant part of, the corpus itself. This is now becoming possible as will be described below in the section on our bilingual text system. An important activity now under way in most of the initiatives aimed at the creation of large language corpora is the development of software which will automatically annotate large bodies of textual material with various kinds of information: in the first place part-of-speech taggers (usually operating on a probabilistic basis), and syntactic parsers (more often rule-based), which give a grammatical analysis of sentences in a text, i.e. identifying the different components - noun phrase, verb phrase, etc. - and how they relate to each other. Many studies are also now in course aimed at word sense disambiguation. The availability of syntactically and semantically tagged corpora will of course increase their potential value for all kinds of linguistic studies. Bilingual and multilingual corpora After the first work on monolingual corpora, at the beginning of the nineties, researchers in machine translation (e.g. Brown et al. 1993), and bilingual lexicography (e.g. Klavans and Tzoukermann 1990) began to turn their attention to bilingual text archives. The interest was motivated by a growing awareness of the value of bilingual corpora as in-depth sources of documented evidence of how texts written in one language can be rendered in another, according to various contextual factors, such as style, register, domain, etc., and therefore of their potential for exploitation in many types of cross-language comparisons and investigations. Such corpora can be considered as important tools in bilingual dictionary compilation, translating, and language learning activities. For example, bilingual lexicographers can use them in much the same way as the monolingual lexicographer refers to a monolingual language corpus: as a test-bed for their intuitions; to provide reliable material for examples. While a monolingual dictionary must provide an adequate representation of the lexical system of a given language, in the bilingual the focus is on representing the complex network of relationships between the lexical system of two languages. The necessity to represent a given headword on the basis of the different ways it can be translated in the target language (TL), and also to specify the factors which affect the choice of the TL equivalent, will influence the formulation of the source language (SL) entry and, in particular, can modify its breakdown into senses when compared to the equivalent monolingual entry. A bilingual refer-

CAROL PETERS AND EUGENIO PICCHI

253

ence corpus may well supply evidence to support the adjustment of a first SL analysis of an entry to meet the demands of the TL; it can even provide a starting point for new hypotheses on the relationships between the two languages in question. Access to bilingual text data should thus help the lexicographer to provide better bilingual dictionaries for translation purposes in the future. However, a bilingual corpus also represents an important direct fount of inspiration for the translator, permitting him/her to find translations for words or expressions which are not listed in any dictionary. Hartmann (1994: 292) points out the limits for the translator of the bilingual dictionary2, stressing that "the semantic abstraction that is built into the lexical inventory of the dictionary has deprived each of these words of their natural context, and the translator must compensate for the lack of contextual information from his/her own bilingual discourse competence, particularly in that most intractable area of 'culturespecific' vocabulary", adding that it is from this point of view that the contrastive textologist will want to consult bilingual text archives, to go "beyond the mere comparison of given parallel texts as translation products and search instead for the actual code-switching operations that allowed the competent translator to 'find' a suitable target-language equivalent in the first place". In fact, bilingual corpora should be considered as important repositories of data which make it possible to study many of the processes involved in transferring information, ideas and concepts from one language to another. As mentioned above, there is much research activity at the moment focused on the development of different kinds of procedures for syntactic and semantic tagging of texts. Once annotated bilingual or multilingual texts become more easily available, methods could be developed to evaluate data extracted from them at different levels: for example, to examine changes in style and register or to investigate correspondences and differences in sentence structure and discourse organization between a target language version and its source text. Such studies, useful in themselves in helping us to understand and improve the work of the translator, could also permit the acquisition of data that reflects important distinctions between the two languages being considered and could help in the formulation of generalisations on the relationships between them. Data derived from analyses on bilingual corpora should also provide valuable input for MT systems. In particular, for example-based systems Nagao (forthcoming) has stressed the importance of including detailed collocational information in the transfer dictionaries of such systems: there are many specific expressions which must be translated in a specific way in a given TL and knowledge of this sort improves the quality of an MT system greatly. To acquire such information, many collocational expressions with their translations must be accumulated and bilingual texts are important sources of such data.

254

REFERENCE CORPORA AND LEXICONS

From parallel to comparable corpora So far most of the studies on bilingual (or multilingual) text archives have been on parallel texts. These are sets of translationally equivalent texts, in which generally one text is the source text and the other(s) are translations, although at times all the texts may be translated from an original not included in the archives. One major criticism that has been made with respect to the results of analyses based on translationally equivalent texts is that target texts are not true examples of natural language: a translation is always influenced by the source text and therefore, while parallel text corpora are important sources for studying the translation process, the target language or L2 text can never be considered in itself as a true representation of that language in usage. As stated by Hartmann (1994) "the translated text(s) cannot by definition share the full range of linguistic features of genuine texts produced in the respective target language". For this reason, many researchers have suggested that a more reliable source for certain types of studies, i.e. how a particular concept is rendered independently in different languages (the starting point being the idea and not an already existing text in a given language), should be a bilingual or multilingual comparable3 text corpus. Comparable text archives are sets of texts from pairs or multiples of languages which can be contrasted and compared because of their common features. They have been described as collections of "texts which, though composed independently in the respective language communities, have the same communicative function" (Laffling 1992). Thus, they are not translation equivalent but are supposed to share certain basic features of different types, such as, period of time, author, topic, functionality, style, genre, register, sublanguage, domain, etc. However, as Corazzari and Picchi (1994) state, the sharing of these features is not sufficient in itself to guarantee the comparability of different texts; comparability also implies the existence to a certain extent of a common vocabulary. These researchers, who are working on the Multext project, suggest a set of selection criteria which should be followed when attempting to construct truly comparable corpora. They propose a combination of external criteria, such as domain, topic, time period, and internal criteria, such as the selection of articles containing a given number of key words taken from a pre-defined vocabulary of a restricted sub-language. Thus, whereas parallel corpora can contain texts from any language types, comparable corpora always concern a restricted sub-language. Consequently, parallel corpora are composed by real-world texts from which data on in-context translation equivalents can be extracted while comparable corpora can be processed to obtain cross-language lexically equivalent contexts within a given domain.

CAROL PETERS AND EUGENIO PICCHI

255

While the potential of parallel text systems is now widely acknowledged, the concept of comparable corpora is relatively new; researchers are thus just beginning to explore their possibilities. Zanettin (1994) discusses an application specifically designed for the trainee translator. He describes the construction of comparable subcorpora from commercially available newspaper corpora on CD-ROM (issues of The Independent and Il Sole 24 Ore) in order to provide the translator with a range of texts on a specific topic. He asserts that "being topic-related, the subcorpora tend to express and report similar facts: being instances of attested behaviour, they represent models against which to check hypotheses on language use". Again, the sets of corpora are constructed on the basis of a choice of semantically corresponding keywords in the two languages. His students are trained to create ad-hoc subcorpora of texts on a given topic, and explore and analyse them in order to be able to use them as resources for translation. They use commercially available monolingual concordancing software to do so. In order to discover how a given word or expression found in texts in one language (L1) could be rendered in the other, his students are taught navigational strategies; they search for clues and significant keywords in the L1 texts in order to build access keys which will help them find equivalent expressions in the L2 corpus. Zanettin reports good results in his class-room tests. As this method is interactive, it is only suitable for relatively small corpora; however, it has the strong advantage in that the students are stimulated to browse the second language texts in depth and given the chance to compare and study examples of usage over languages, allowing them to extract useful translation data and thus improve their performance. John Laffling, perhaps one of the first advocates of comparable corpora, has used them in the construction of the transfer lexicon of an MT system (Laffling actually uses the term parallel to refer to this kind of bilingual text (see note 3)). Laffling has built a system for the automatic translation of German/English political texts. For this purpose, he has built up a test corpus of German/ English political manifestos. Laffling claims that such special domain texts "... provide the computer with a knowledge base that is comparable to the enriched linguistic repertoire the translator has at his disposal". They provide "selection pointers which function in a much more refined way than general equivalence relations found in a bilingual dictionary". Laffling recognises the importance of lexical collocates for sense disambiguation (in text analysis) and in the construction of natural language sentences (text generation). Thus, for translation (both human and machine guided), it is essential to have full information on the most significant collocates for a lexical item in each of its possible senses. This type of information is not to be found in a dictionary but can be extracted from a corpus. The tendency to co-occur may be preconditioned by a particular text-

256

REFERENCE CORPORA AND LEXICONS

type; thus for each entry in his transfer dictionary, Laffling includes information that he has extracted from the corpus on its preferred collocates for each sense. However, senses that do not appear in his corpus are eliminated from his lexicon as being irrelevant for that particular domain. His thesis is that appropriate word selection requires "abandoning procedures for comparison of the potentialities of two linguistic codes in favour of examination and coordination of the actualities of textual usage". He insists that the "coordination of parallel texts offers MT system advantages which ... can be individualised as follows: it explodes the myth of post-disambiguation one-to-one equivalence; it provides the computer with a wide reading so necessary to the human translator; it facilitates the production of stylistically cohesive TL texts" (see Laffling 1991). We have studied a way to process comparable corpora automatically in order to identify cross-language lexically equivalent contexts. In a later section, we describe our approach to this problem. Corpus design and accessibility. From the above discussion it appears that the utility of bilingual and multilingual corpora consisting of large sets of translation or domain equivalent texts for many kinds of cross-language studies has received wide recognition. The main problem now being faced by the research community is the creation and management of such archives. In fact, the construction of a large-scale bilingual text corpus is not a task to be undertaken lightly as it implies a considerable investment of time and resources. The goal must be a high quality corpus, sufficiently representative of the object it aims at modelling (whether the "entire" lexicon or particular sub-sets of it), and sufficiently large to provide valid data for a wide range of linguistic studies4. A motivated and systematic selection and collection of the texts to be included is essential. User needs must be considered at all times and exchangeability and reusability must be considered as priorities. Before any decisions are taken, the criteria to be adopted when assembling corpus material must be carefully evaluated. Unfortunately, no hard and fast guidelines are yet available which can be used to define the "correct" design criteria, not even for monolingual corpora. Points to be considered include the text acquisition methods to be adopted, the final size of the corpus, the definition of the sampling units, the identification of the different text types which constitute the "language" to be represented by the corpus, the proportions of each to be included (depending on the level of representativeness to be guaranteed), the time period to be covered, for parallel corpora the labelling of samples as source or target texts (and for target texts, some indication of trans-

CAROL PETERS AND EUGENIO PICCHI

257

lation status so that the users can judge the value of the material they are handling). Consultations with language specialists and potential users of the corpus are essential during this stage in order to evaluate the correctness of the approach adopted. Another problem which has been much discussed in recent years but still remains a major difficulty is that of copyright. Although publishing companies are often ready to enter into cooperation agreements and generous in providing research institutions with access to textual material in machine-readable form (MRF), it is not always clear what is the legal status of products derived from analyses made on such data nor whether the results can be made freely available to other members of the research community. The question of accessibility is of crucial importance. Not only should access to a corpus constructed for research purposes be granted to other members of the scientific community but such resources should be made generally available to the wider community of people working in some way with language. This is no longer a dream but is rapidly becoming a reality. The Bank of English at Birmingham University, for example, already provides a 20 million word sample of contemporary English for remote consultation. A number of CD-ROM's have recently been published containing collections of literary texts or language reference corpus data, together with procedures for dynamic access and retrieval. To give just one example of literary text collections now available commercially in this way, we can cite LIZ (Stoppelli and Picchi 1995). The second version of LIZ was released in spring 1995 and contains a complete set of the best known Italian literary texts, from Dante and Boccaccio to Pirandello and D'Annunzio. Similar initiatives exist for many other literary collections. Important sources of computerised contemporary text are newspapers, many of which are now available on Internet and/or sold on CD-ROM. Examples that can be cited are II Sole - 24 Ore for Italian and The Independent for English, but many other newspapers are beginning to appear in these two electronic forms. Indeed, with the rapid diffusion of networking facilities, e.g. Internet, it will soon be easy for the user working at home to access, query and acquire in real time data archives held anywhere in the world. With reference to the availability of bilingual and multilingual data, the European Corpus Initiative has already released a CD - Multilingual Corpus 1 (ECI/MCI). This CD is based on a corpus of almost 100 million words of text in at least 23 languages, mostly European languages but also some Chinese and Japanese, including 13 sets of bilingual or multilingual parallel archives, and covering a wide variety of text types. The ECI5 collects machine readable text for the purpose of scientific and humanistic research, and distributes it at cost and without royalties. The Multext project6 is now working on producing the first large-scale, annotated multilingual corpus covering six European languages (English, French, German, Italian, Spanish, Dutch). The corpus will be

258

REFERENCE CORPORA AND LEXICONS

of 3 parts: a parallel corpus, consisting of 2 million words per language - 50% will be marked and hand-validated for sentence alignment, 10% will be tagged for part-of-speech; a comparable corpus, again of 2 million words per language, composed of comparable texts from 2 or 3 different domains - 10% will be POS tagged; a small speech corpus. The intention is to serve as a multilingual reference corpus for research purposes just as the Brown Corpus has done for English and the Canadian Hansard Corpus for English/French. Corpus and tools will be distributed at cost and without royalties, and will be directly usable by researchers in a wide range of disciplines, including language understanding and generation, translation, etc. More information on international language data collection efforts and their availability can be found in (Church and Mercer 1993). It is evident that the definition of standards for corpus design, management and analysis is now a priority. We expect that the work of EAGLES7 (Expert Advisory Group on Language Engineering Standards) will be important in this area, helping also to avoid much of the actual dispersion of resources and duplication of efforts in fragmented uncoordinated initiatives. So far, we have attempted to give an idea of the current state of the art in Computational Linguistics with respect to computerised lexicons and reference corpora. In the second part of the paper, we will describe our current work at the Istituto di Linguistica Computazionale, Pisa, in this area: the implementation of a complex system for mono- and bilingual dictionary and text processing, querying and analyses.

The bilingual lexical database system Our bilingual lexical database system was first described in Picchi et al. (1990); it now forms part of the MLDB, a multilingual integrated lexical database system described in detail in Marinai et al. (1990). The lexical components of the MLDB include the Italian Machine Dictionary - mainly based on the Zingarelli Italian Dictionary -, and LDB's derived from the Garzanti Nuovo Dizionario Italiano, and the Collins Concise Italian-English, EnglishItalian Dictionary; an English LDB is now being included. In this section, we attempt to give an idea of the potential of the system for the user. For full details on how a lexical database can be designed and implemented on the basis of the dictionary data extracted by procedures which parse and analyse publishers' computer type-setting tapes, the reader is referred to the publications mentioned above.

CAROL PETERS AND EUGENIO PICCHI

259

Querying the bilingual LDB The translator will primarily be interested in the bilingual dictionary data. Using the bilingual LDB system, much valuable information can be retrieved for a given lexical item at all levels (e.g. translation equivalents, examples of usage, syntactic information, etc.) which is inaccessible using traditional dictionary look up. The LDB query system offers dynamic search procedures that permit the user to navigate through the dictionary data and within the different fields of the entry in order to access and retrieve information in whatever part of the dictionary it is stored, specifying the language on which the query is to operate. Any lexical item or combination of items entered as a value is searched in the database with reference to its particular function in the entry and the results (i.e. number of occurrences of the item) are displayed field by field. The user can then select, view and print those results that interest him/ her. It is important to stress that the system response times are virtually immediate. For a full description of the LDB query language and a list of the main functions implemented, see Marinai et al. (1990). The system is implemented on IBM-compatible systems and the most recent version of the system runs under Windows. In the following figures we give some examples of how the user can query the bilingual LDB to retrieve information that is not easily accessible in the printed dictionary. Let us suppose that a suitable translation for the Italian adjective imprudente is being searched. Figure 1 shows how the results are displayed on the screen as a result of a look-up of this word: we can see that imprudente appears just once as a headword, once in the Semantic Indicator field, once in the Example Translation field, and six times as a translation. It can be seen that the object of a search is not the single word but the word+data field pair. We refer to this as the Search Unit (SU). At this point the user can request the system to display the contexts for the particular SU's that interest him.

Figure 1 : Querying the Bilingual Lexical Database:Results of a Search for "imprudente "

260

REFERENCE CORPORA AND LEXICONS

Figure 2 shows a screen in which three windows have been opened. At the top left hand corner of the figure, we see the two SU's selected: imprudente as Headword and in the Translation data field. On the right side we see the display of the LDB entry for the headword; it can be seen that three translations are suggested, qualified simply by a Semantic Indicator denoting general usage: careless, foolish, imprudent', another translation, unwise, is proposed for when the word is used in the semantic context qualified by osservazione. At the bottom of the screen, we see the six entries on the English side of the dictionary in which imprudente appears in the Translation field. It can be seen that foolhardy, improvident, rash, reckless and unguarded, are additional possible translations for this word, which were not listed on the Italian-English side of the dictionary and thus could not be found easily in the printed dictionary.

Figure 2: Display of Contexts for "imprudente" in the Bilingual LDB In Figure 3 we give an example of an LDB look-up for a combination of words. In this case, we search for cooccurrences of the terms "rolling" and "money". It can be seen that the expression "rolling in money" appears seven times, each time in the Example Translation field and therefore not retrievable in the printed volume, and that six different translations are suggested.

CAROL PETERS AND EUGENIO PICCHI

261

Figure 3: Display of Results of a Search for Cooccurrences of "rolling" and "money" in the Bilingual LDB Of course, the translator can also access and query the monolingual dictionaries maintained by the system. The different perspective on the data provided by a monolingual entry often gives a more complete view of a given lexical item and its usage than is provided by the bilingual entry alone. A procedure has thus been implemented to permit semi-automatic mapping between bilingual and monolingual LDBs. Equivalent entries from separate dictionaries can be combined and links are created between them semi-automatically at the sense level, mainly on the basis of information that can be extracted from definitions, examples and semantic labels. In this way, we create a more complete composite entry which represents the sum of the information contained in the individual dictionaries (see Marinai et al. 1994). The translator can use this procedure to access, compare and scan rapidly the lexical information given for the same item in different source dictionaries. Specialising the bilingual LDB In the version of the bilingual LDB that we are implementing in the Translator's Workstation, the users will also have functions available so that they can add their own information to the bilingual entry. This will be particularly useful for translators working in a specific domain who may well accumulate information on the usage of particular terms and expressions within this discipline which is not registered in any dictionary. They can call the User Update Pro-

262

REFERENCE CORPORA AND LEXICONS

cedure which permits them to add to the data in the lexical entries as they wish, as long as the data representation schema is respected. The procedure will work in interactive mode. The user calls the lexical entry to which he/she wishes to add information by entering the headword on the keyboard. The structured and tagged entry is displayed on the screen. The user then invokes a Help function to display the different functions that can be used to intervene on the entry. All the information added by the user is recorded in a special User Memo Section. Within this section, he/she is given a choice of fields in which to enter the data. These fields are similar to those used in the rest of the Entry schema, and consist of fields for translations, examples, translations of examples, semantic indicators, and various kinds of semantic labels: subject, usage, geographic and register codes (for a description of the data representation schema we use, see Marinai et al. 1990). With the exception of a User Note field used for free comments by the translator, purpose-written, dynamic indexing procedures will then be executed on this new data so that it becomes directly accessible for subsequent querying. In this way, the translator is able to exploit and reuse information acquired as a result of his/her own experience and activity.

The bilingual text management and query system In a previous section, we attempted to give an idea of the current state of the art concerning the development, availability and utility of language reference corpora and especially of bilingual and/or multilingual reference corpora. In this section we describe our experience in the actual development and implementation of a comprehensive bilingual corpus system. The system initially consisted of a set of procedures designed to create and interrogate parallel text archives. The parallel text system is now in an advanced stage of development and is described in some detail here below. We are now also working on procedures to process comparable domain-specific textual material. Our comparable system is still in a preliminary stage - in the last part of this section, we will describe the strategy that we intend to implement to process and query this type of bilingual texts. The parallel corpus system So far most of the systems studied to manage bilingual and multilingual corpora work on parallel text archives and use statistically based procedures to align the texts at the sentence level. The results of a query are given as parallel sentences. Church and Gale (1991) presented a system of this type and also de-

CAROL PETERS AND EUGENIO PICCHI

263

scribed a word-based concordance tool in which the possible translations for a given word are discovered from the corpus on the basis of a pre-computed index indicating which words in one language correspond to which words in the other. Other papers have suggested different ways for isolating and identifying translation equivalents, within previously matched sections (Brown et al. 1993), and without pre-alignment (Fung and Church 1994, Jones and Somers 1995). All these studies use text-internal evidence. Our approach to the problem is quite different. We use external data provided by a bilingual LDB to create links between pairs of texts, where one is the translation of the other, on the basis of SL/TL8 translation equivalents. These links are then used by the bilingual text query system to construct parallel contexts for any form or cooccurrence of forms found in either of the two sets of texts; the term searched and the position of its translation equivalent are indicated to the user. A preliminary version of this system is described in Marinai et al. (1991). At the moment, the system runs on a sample set of Italian/English texts, chosen to cover a number of different language varieties, and including extracts from scientific articles, university and school text books, magazine articles, short stories, novels, and poetry. This set of texts was collected in the first place in order to provide a test-bed for the system, but we hope that it will provide useful empirical data to assist us in the definition of valid, more objective design criteria, which can then be used in a subsequent extension of these archives. In the version of the system which has been implemented in the Translator's Workstation, the translator has the possibility of creating a reference corpus from his/her own material and adding new texts to it as they become available. An easy-to-use interface has been prepared to guide the translator step-by-step as he/she inputs pairs of texts to the system. Creating the parallel archives Given a new pair of "bilingual" or translation equivalent texts, the first stage is to structure them in text database format using the DBT procedures (see the final section for information on the DBT system). The texts are scanned to recognise and identify the different elements composing them. For example, word forms are distinguished from the other tokens, such as punctuation marks, numbers, line and paragraph breaks; codes are added to distinguish between full stops and abbreviation marks, between dashes and hyphens, between the different use of the apostrophe in Italian and in English, etc. This stage is simple, rapid, and once a few preliminary instructions have been given, automatic. Once a pair of texts has been stored in DBT format, it must be input to the text "synchronisation" procedure which establishes as many links as possible

264

REFERENCE CORPORA AND LEXICONS

between translation equivalents in the two texts. This procedure is totally automatic and operates as follows. Each word form in the text selected as the source text is input to the morphological analyser for that language in order to identify its base lemma which is then searched in the bilingual LDB. All translations given for this lemma are read and input to the morphological generator for the TL; all the forms generated are then searched over the relevant zone in the target text. If the procedure finds more than one possible base lemma for a given form, the translations for each will be read as, in the case of grammatical homography, it is quite possible that the translation equivalent does not respect the category of the source language and, in the case of lexical homography, it is presumed unlikely that the translations of the "wrong" lemma will find a correspondence in the target text. Articles, pronouns, prepositions and a small list of stop words are excluded from this search as of little significance to the matching procedure and liable to create noise. When one of the translation equivalent forms is found in the searched section of the L2 text, a link - consisting of a physical address which locates the equivalent word in the L1 text - will be created. When no entry for a word in the L1 text is found in the dictionary, it may be that the form being examined is either a proper noun or a word from a highly specialised vocabulary not included in our bilingual LDB. An attempt is thus made to match such forms against any equivalent character strings in the relevant zone of the L2 text, ignoring the last characters to allow for morphological variations as, in the two languages in question, proper nouns and scientific terms frequently resemble each other. The matching procedure continues, word by word, to the end of the source language text. The execution of the "synchronisation" procedure is rapid and totally transparent. When it has terminated, the results are presented to the user in terms of the number of successful "matches" of translation equivalents between the Source and Target texts. The procedure will be considered to have "failed" if the number of matches is less than a given percentage of the total text. This procedure must be executed just once for each pair of bilingual texts, when they are added to the archives. Querying the parallel corpus When the parallel texts have been processed by the synchronisation procedure, all the links obtained are memorised in the text archives so that they can be used by the parallel query system. For each source text word or combination of words searched by the user, the parallel concordances for the target text are constructed in real time and displayed on the screen. The source context is constructed with the searched word at the centre; this word is highlighted and, dur-

CAROL PETERS AND EUGENIO PICCHI

265

ing the creation of its context, any translation links to the target text associated with the other words in the context are read. If the searched word itself has an associated link, this is used to identify directly the corresponding word in the target text, which will also be highlighted and used as the central point for the construction of the TL context; other words that have been linked in the paired contexts can be optionally evidenced in a different colour. When there is no directly linked TL form for the SL word being searched, then all the links for words found in the source context are used to calculate an "average" value which identifies the central point around which the relative TL context will be constructed. The calculation of this "average" value allows for the possibility of uneven concentrations of matched words in the contexts. The two linked forms which are closest to the point calculated as the middle of the target text context are evidenced in a different colour, as indicators of the likely position of the translation equivalent. "Wrong" links between potentially falsely recognised translation equivalents which disturb context calculation are identified and eliminated by the query system. For each direct link created, the procedure also estimates the probable value for the TL equivalent of that SL form on the basis of the other links created in the surrounding piece of text. If the estimated value differs considerably from the actual value, then the query procedure discards this particular direct link as being unreliable and recalculates the parallel contexts on the basis of those links recognised as valid. The user can either search for single word forms or, using the morphological generator, for all the forms of a given lemma. Parallel concordances of interest can be printed out or saved in a separate file for future reference. Figure 4 gives an example of the results of a search for parallel concordances of any occurrence of the lemma "glance" (the forms glance, glances, glances, and glancing were found) in a test corpus based on James Joyce's "The Dubliners". In the figure, the words being searched and the words linked directly with them in the parallel contexts are shown in bold, whereas the indicators of the position of the translation equivalents are underlined.

266

REFERENCE CORPORA AND LEXICONS

One of the most important uses for the translator of a system of this type is the extraction of information on the translation of bound or semi-bound expressions, idioms, collocations. A dictionary cannot contain an extensive range of such information, owing to space factors. The corpus can be important source. Figure 5 shows parallel concordances for the results of a search for the English expression "time being", i.e. the query specified cooccurrences of "time" adjacent to "being".

Figure 5: Examples of Parallel Concordances for "time" and "being"

CAROL PETERS AND EUGENIO PICCHI

267

In a new version of the system, to be released shortly, we have inserted a function that will provide a preliminary statistical analysis of the results. For any SL word queried, the TL equivalents found in the parallel texts will be listed in descending order of occurrence for those words for which direct links have been made, followed by the number of parallel contexts in which no direct link has been found for the searched word in the TL text, and the users can then select the particular contexts they wish to view without necessarily having to scan through them all. An example of how the results will be displayed is shown in Figure 6 for occurrences of "time" in our bilingual corpus

Figure 6: Frequencies for the Results of a Search for Parallel Concordances of "time'

Applications

in

lexicography

So far, we have described how our parallel corpus system can provide a useful source of reference for translators or language learners. Another potentially important area for the application of such a system is in bilingual lexicography. Often corpus data based on real world evidence shows up discrepancies or gaps in dictionary coverage and can thus provide an important supplement to the lexicographers' intuition. For example, the translations given in our bilingual lexical database (see previous section) and also in printed dictionaries we have consulted for the Italian lemmas animo and anima can be summarised as follows: animo, sm (a) (mente) mind; (b) (coraggio) courage anima, sf (a) (gen) soul; (b) (persona) soul, (abitante) inhabitant.

268

REFERENCE CORPORA AND LEXICONS

Here we have a clear distinction between animo shown as meaning simply mind or courage, and anima meaning the non-material part of a human being and, by extension, the human being itself. If, however, we examine our monolingual LDB we find that the third sense given for anima (sede degli affetti e dei sentimenti) coincides with the first sense of animo {sede e principio degli affetti e della volonta) (cfr soul 1.2 in Cobuild "... where the person's true nature and deepest thoughts and feelings are believed to be"). If we then examine the evidence provided by our bilingual text archives we find confirmation of this overlapping of meaning between the senses of the two lemmas. Whereas anima is indeed translated by "soul" in most of the examples and has been given no direct translation in a few cases, animo is also translated in many contexts as soul. Thus the evidence found in the bilingual corpus contradicts the clear (but misleading) distinction made in the bilingual dictionary between animo and anima and confirms the analysis of the monolingual dictionary. Another example of incomplete information can be found if we considered the data currently contained in our LDB under the entry for the English noun "air". This information can be summarised as follows: air n (a) aria (with examples to make it clear that this is air in the "mixture of gases" sense) (b) (Radio, TV) with no direct translation but examples of usage (c) (appearance) aria, aspetto In our corpus three distinct senses also clearly emerge (however, the Radio, TV sense is not registered, probably due to the small dimensions of our archives): (i) "air" in the sense given by (a) above in approximately 60% of cases, always translated by aria (where a direct translation appears); (ii) "air" in the sense of appearance (see (c) above) in about 25% of occurrences, always translated by aria. (iii) "air" in the musical sense in roughly 15% of cases, translated by motivo, motivetto, melodia and canzone. This suggests that a relatively important sense of "air" has been neglected in the bilingual LDB (and in the dictionary on which it was based). It is difficult to understand whether this omission in the dictionary data is due to an oversight or is deliberate, the sense not being considered sufficiently frequent to warrant recording. The translator or language learner attempting to translate "air" in this sense would find a number of possible answers in the corpus which are not provided by the dictionary. The bilingual lexicographer considering the corpus evidence would probably revise the existing bilingual entry for "air" or, if writing a new one, would take care to include "air" in the musical sense.

CAROL PETERS AND EUGENIO PICCHI

269

Building a comparable corpus system We are now studying procedures for constructing and querying comparable corpora. The comparable corpus system also makes use of the bilingual LDB and morphology but the approach is quite different as we do not have direct translation equivalents on which to work. Instead, we are using lexical and linguistic knowledge extracted from a specific domain corpus in one language and projecting it onto a comparable corpus in the other, i.e. given a particular term or set of terms in the texts in one language (L1), the aim is to be able to identify contexts which contain equivalent or related expressions in the texts of the other (L2). To do this, we attempt to isolate the vocabulary related to that term in the L1 corpus - hypothesising that lexically equivalent terms will be associated with a similar vocabulary in L2. We apply the Mutual Information Index (Church and Hanks 1989) to compute the most significant collocates of the L1 term and, using the bilingual LDB to find translation equivalents, search the L2 corpus attempting to identify and retrieve equivalent L2 contexts. These procedures are still in an experimental phase, but the first results on a set of Italian/ English parliamentary debates are very encouraging, i.e. we can retrieve contexts which refer to a particular argument represented in L1 by a given expression (term or set of terms), without the necessity for a known translation equivalent for that expression being present. A preliminary version of this system has been described in (Peters and Picchi 1995). We are now working on a revised version and intend to implement it with a user-friendly interface. This kind of data should be very interesting for an advanced language learner as it will provide him/her with second language natural language data (rather than translation equivalents) for a given L1 expression in a particular sublanguage. The parallel system can thus be used extract information from all types of translation equivalent texts whereas, when it is fully implemented, the comparable system should serve the needs of the user wanting information on cross-linguistic equivalents for special domains (i.e. in the field of terminology). Again, as with the parallel system, an important application for a system of this type should be in bilingual lexicography and especially in the construction of lexicons and thesauri for special languages.

The workstation The lexical and text management systems described above in the previous two sections have been integrated into a prototype Workstation, shown as in Figure 7. All the components included in the workstation form part of the PiSystem,

270

REFERENCE CORPORA AND LEXICONS

an open-ended modular set of tools, designed and developed to meet the various needs of literary and linguistic text processing and analyses. The core component of the system is the DBT, a textual database system that has been implemented in different configurations to perform specific text and dictionary processing tasks. Other components can be integrated with this system kernel as required, depending on the needs of a particular application. (For a detailed description of the DBT, see Picchi 1991.) In the application we have described in this paper, specific modules which permit the translator to create his/her own text archives and add his/her own lexical data to the dictionary databases have been included.

Figure 7: The Prototype Workstation The user can move freely and rapidly from one component to another, using the information extracted from one to query or supplement information contained in another. Links between the textual databases and the LDBs will permit look-up on detailed morphological, syntactic and semantic information in the dictionaries for any lexical item found in the texts; in addition semantic information, such as taxonomic data, derived from dictionary definition parsing

CAROL PETERS AND EUGENIO PICCHI

271

operations can also be invoked to search the corpus data (see Picchi and Calzolari 1985, for examples of this kind of text querying). The workstation is implemented on IBM-compatible personal computers and runs under Windows; context sensitive Helps can be invoked to explain the functionality of each command. This workstation can be employed in many types of activities: by the translator as we have described, and also by the lexicographer, by the language learner, or indeed by any user interested in exploiting to the full the possibility of being able to dynamically access, browse, and extract the different kinds of linguistic information contained in dictionary and text databases. Notes 1 See, for example, the success of important corpus-based dictionary projects such as COBUILD or the Trésor de la Langue Française, and the decision to set up a Network of European Corpora (NERC Consortium, 1991) to serve the various European language engineering needs under the auspices of the European Union. For information on corpus resources and their general availability, see RELATOR, a project in the Linguistic Research and Engineering programme of the CEU coordinated in Pisa (see the RELATOR Web site at http://guagua.echo.lu/langeng/en/lre2/relator.html). 2 Recently, proposals have been made for a new type of dictionary: the bilingualised dictionary, which should to some extent "bridge the gap" between the traditional bilingual and monolingual dictionaries. These dictionaries are different from monolinguals because they supply translation equivalents and also from bilinguals because they provide semantic equivalents (definitions) of the original text. Those available so far are generally "bilingualised" versions of English monolingual learners' dictionaries; the translation equivalents provide the key to help the user interpret what is for him/her an L2 definition. 3 We here use the terms 'parallel' and 'comparable' commonly used by computational linguists to distinguish between the two kinds of bilingual or multilingual texts. However, other researchers have proposed a distinction between bi-texts, for translationally linked texts, and parallel texts for texts that are functionally similar in situational motivation and rhetorical structure (see Hartmann, 1995). 4 Although corpus representativeness is a clear goal for the future, important results can still be obtained in the meantime from material which is not balanced, as long as the user is aware of the type of data he/she is analysing. For instance, the availability of texts such as the Hansards transcripts of the Canadian parliamentary debates has permitted much experimentation on the statistical processing of French/English bilingual texts which would have been impossible without the availability of such enormous volumes of data (see, for example, Church and Gale 1991). 5 The ECI is a volunteer effort, sponsored by the Association for Computational Linguistics (European Chapter), carried out at the Human Communication Research Centre, University of Edinburgh (HCRC) and Institut Dalle Molle pour les etudes semantiques et cognitives, University of Geneva (ISSCO), with additional support from the European Network in Language and Speech (ELSNET) and the Network for European Reference Corpora (NERC). 6 The MULTEXT project is part of the LRE Programme of the European Union. The project has developed a set of generally usable software tools for text corpus manipulation and

272

REFERENCE CORPORA AND LEXICONS

analysis, together with lexicons and multilingual corpora in several European languages. It has established conventions for the encoding of corpora and harmonised specifications for computational lexicons, building on and contributing to the preliminary recommendations of the relevant international and European standardisation initiatives. All project results are freely and publicly available (see the MULTEXT Web site http://guagua.echo.lu/langeng /en/lre2/multext.html). EAGLES is an EU initiative in the framework of the Linguistic Research and Engineering (LRE) programme, begun in January 1993 and coordinated at ILC-CNR, Pisa. The aim of EAGLES is to accelerate the provision of standards for: (i) very-large language resources; (ii) means to manipulate such knowledge, via computational linguistic formalisms, mark-up languages and various software tools; (iii) means to assess and evaluate resources, tools and products. The EAGLES Guidelines can be accessed at the EAGLES Web site at http://www.ilc.pi.cnr.it/EAGLES/home.html. For each pair of texts, we use the term Source to indicate that text, independently of the language in which it is written or whether it is an original or translated version, which is taken as input by the bilingual text "synchronisation" and query procedures and on which they operate to identify or construct, respectively, for each form, the corresponding translation equivalents or contexts in the other text, denoted as Target text. The procedures can treat either of the two paired texts as Source or Target, indifferently. References Boguraev, B.K. 1986. "Machine-readable dictionaries and research in computational linguistics". In Proceedings of the Workshop "On Automating the Lexicon". Grosseto, Italy, 1986. Brown, P., Delia Pietra, S., Delia Pietra, V. and Mercer, R. 1993. "The mathematics of machine translation". Computational Linguistics 19(2). Byrd, R.J., Calzolari, N., Chodorow, M.S., Klavans, J.L., Neff, M.S., Rizk, O.A. 1987. "Tools and methods for computational lexicography". Computational Linguistics. 13(3-4), 219-240. Calzolari, N., Peters, C. and Roventini, A. 1990. "Computational model of the dictionary entry: Preliminary report". ACQUILEX, Esprit BRA 3030, Six Month Deliverable, ILC-ACQ-1-90, Pisa, 90p. Church, K.W. and Gale, W. 1991. "Concordances for parallel text". In Using Corpora, Proc. 7th Annual Conference of the UW Centre for the New OED and Text Research. Oxford: OUP, 40-62. Church, K.W. and Hanks, P. 1989. "Word association norms, mutual information and lexicography". In Proceedings 27th Annual Meeting of ACL, Vancouver, B.C., 7683. Church, K.W. and Mercer, R.L. 1993. "Introduction to the special issue on computational linguistics using large corpora". Computational Linguistics 19(1), 1-24. Corazzari O. and Picchi, E. 1994. "A Proposal for the Construction of Comparable Multilingual Corpora". Working paper for the Multext Project, ILC-CNR, Pisa. Firth, J. 1957. "A synopsis of linguistic theory 1930-1955". In Studies in Linguistic Analysis, Philological Society, Oxford; reprinted in F.Palmer (ed.). Selected Papers of J.R.Firth. Longman. 1968.

CAROL PETERS AND EUGENIO PICCHI

273

Francis, W. and Kucera, H. 1982. Frequency Analysis of English Usage. Houghton Mifflin. Fung, P. and Church, K.W. 1994. "K-vec: a new approach for aligning parallel texts". In COLING 94 - The 15th International Conference on Computational Linguistics, Kyoto, Japan, 1096-1101. Hartmann, R.R.K. 1994. "The use of parallel text corpora in the generation of translation equivalents for bilingual lexicography". In Euralex 1994 Proceedings. Amsterdam, 291-297. Hartmann, R.R.K. 1995. "From contrastive textology to parallel text corpora: Theory and Applications", paper for Jacek Fisiak Festschrift. Jones, D.B. and Somers, H.1995. "Bilingual vocabulary estimation from noisy parallel corpora using variable bag estimation". In JADT95. Proceedings of the Third International Conference on Statistical Analysis of Textual Data, CISU, Rome, 255-262. Klavans, J., and Tzoukermann, E. 1990. "The BI-CORD system". In COLING-90 Proc, 15th International Conference on Computational Linguistics, 174-179. Laffling, J. 1991. "Towards high-precision machine translation. Based on contrastive textology". Distributed Language Translation, 1. Berlin: Foris. Laffling, J. 1992. "On constructing a transfer dictionary for man and machine". Target 4:1,17-31. Marinai, E., Peters, C. and Picchi, E. 1990. "The Pisa multilingual lexical database system". Esprit BRA 3030, Twelve Month Deliverable, ILC-ACQ-2-90, Pisa, 61p. Marinai, E., Peters, C. and Picchi, E. 1991. "Bilingual Reference Corpora: A System for Parallel Text Retrieval". In Using Corpora, Proc. of 7th Annual Conference of the UW Centre for the New OED and Text Research. Oxford: OUP, 63-70. Marinai, E., Peters, C. and Picchi, E. 1994. "A prototype system for the semi-automatic sense linking and merging of mono- and bilingual LDBs". In N. Ide and S. Hockey (eds.), Research in Humanities Computing. Oxford: OUP, 88-108. Nagao, M., Dictionaries for Machine Translation, Linguistica Computazionale, 8, forthcoming. Peters, C. and Picchi, E.1995. Capturing the comparable: A system for querying comparable text Corpora, in S.Bolasco, L.Lebart, A.Salem (eds), JADT'95 - 3rd International Conference on Statistical Analysis of Textual Data, CISU, Rome, 255262. Picchi, E. 1991. "D.B.T.: A textual data base system". In L. Cignoni and C. Peters (eds), Computational Lexicology and Lexicography. Special Issue dedicated to Bernard Quemada. II, Linguistica Computazionale, 7, 177-205. Picchi, E. and Calzolari, N. 1985. "Textual perspectives through an automatized lexicon". In Proceedings of the XII International ALLC Conference, Slatkine, Geneva. Picchi, E., Peters, C. and Calzolari, N. 1990. "Implementing a bilingual lexical database system". In T. Magay and J.Zigány (eds), BUDALEX '88 Proceedings, Budapest, 317-329.

274

REFERENCE CORPORA AND LEXICONS

Sinclair, J. 1987a. "Introduction". In J.M. Sinclair (ed), Looking Up: An Account of the COBUILD Project in Lexical Computing. Birmingham: Collins COBUILD, vii-ix. Sinclair, J. 1987b. "Grammar in the dictionary". In J.M. Sinclair (ed), Looking Up: An Account of the COBUILD Project in Lexical Computing. Birmingham: Collins COBUILD, 104-115. Stoppelli, P., Picchi, E. (eds) 1995. LIZ. Letteratura Italiana Zanichelli, CD-ROM, Zanichelli Editore, Bologna. Zanettin, F. 1994. "Parallel words: Designing a bilingual database for translation activities". In Andrew Wilson and Tony McEnery (eds), Corpora in Language Education and Research: A Selection of Papers from TALC 94. Lancaster University, UK, 99-111. Dictionaries Collins Concise English-Italian, Italian-English Dictionary. 1985. Collins, London and Glasgow. Il Nuovo Dizionario Italiano Garzanti. 1984. Garzanti, Milano. Sinclair, J., Hanks, P., Fox, G., Moon, R. and Stock, P. (eds). 1987. Collins COBUILD English Language Dictionary, Collins, London and Glasgow, 1987. Zingarelli N. 1970. Vocabolario della Lingua Italiana, Zanichelli, Bologna.

Texts Types and Medium

Written to be Spoken: The Audio-Medial Text in Translation Mary Snell-Hornby University of Vienna 'I wish to speak to you today about the tragedy of Europe. This noble continent, comprising on the whole the fairest and the most cultivated regions of the earth, enjoying a temperate and equable climate, is the home of all the great parent races of the western world. It is the fountain of Christian faith and Christian ethics. It is the origin of most of the culture, arts, philosophy and science both of ancient and of modern times. If Europe were once united in the sharing of its common inheritance, there would be no limit to the happiness, prosperity and glory which its three or four hundred million people would enjoy. Yet it is from Europe that have sprung that series of frightful nationalistic quarrels, originated by the teutonic nations which we have seen even in this twentieth century and in our own lifetime wreck the peace and mar the prospects of mankind.' These were the opening words of the famous address to the 'academic youth of the world' given by Winston Churchill on 19 September 1946 in the great hall of the University of Zürich. The event is commemorated by a plaque bearing an inscription with the rousing closing appeal of that address: 'Therefore I beg you, let Europe arise!'. Europe was then in ruins, and the speech is in essence a visionary description of Churchill's dream of a 'United States of Europe' which he saw as the remedy for the ills of the time - and which bears an extraordinary resemblance to the European Union that is the reality of today. It is however not the purpose of this essay to discuss the political developments over the last fifty years. What is of interest to us here is the language that Churchill - an acknowledged master of English rhetoric - used to get his message across. It is a kind of prose which today strikes us as being forceful but singularly old-fashioned, with structures ('if Europe were once united in the sharing', 'let Europe arise') and lexical items ('fairest', 'mar') which would not fit in with the text-type conventions of politicians' speeches in the 1990s. And

278

WRITTEN TO BE SPOKEN

yet even today the language seems to possess an unusual immediacy and vitality. Its power lies not only in its lexical density and precision, but in its rhythm and the ebb and flow of its marked syntactic patterns - in its speakability, in fact. This is not merely rhetorical prose, but essentially language that was written to be spoken - and as such, it is subjected to rules of its own.

Text-type and translation method It was Katharina Reiss (1971) who first investigated the intricate relationships between text-type and translation. The English word text-type covers both Texttyp and Textsorte in German. Textsorten are defined by their specific formal and linguistic features and by the particular situation in which they have to function, examples being cooking recipes, instructions for use, business letters or publicity leaflets. Texttypen are distinguished by their underlying function as based on Karl Bühler's model of 1934 (1965: 28), where the three functions of language are shown to lie in Darstellung (representation), Ausdruck (expression) and Appell (appeal). It is this sense that Reiss used to develop her theoretical frame of reference for methods of translation (1971, 1976). From Bühler's threefold division she derived three corresponding 'dimensions of language' (logical, aesthetic and dialogic) and three corresponding text-types {Texttypen), viz. informative (such as a scientific report), expressive (such as a lyric poem) and operative (such as an advertisement). She then developed criteria for translation according to the respective text-type. However, even in 1971 Reiss identified a fourth text-type which she defined as audio-medial: such texts have been written to be spoken or sung (1971: 34) and are hence dependent on a non-linguistic (technical) medium or on other audio-visual forms of expression for their full realization; language is only part of a broad complex of elements (1971: 49, 1976: 15). These observations gave rise to a heated debate among translation scholars in Germany, and after some time Reiss modified her position. Following a suggestion by Bernd Spillner (1980: 75) she changed the term audio-medial to multi-medial to include texts (such as comics) which have visual but not acoustic elements (1984: 211), and in 1990 she conceded that such texts do not form a fourth text-type in her original sense, but contain elements belonging to any of her other three text-types. Examples of multi-medial texts would be songs, stage-plays, film-scripts and operatic texts, as well as comics and advertising material that includes audiovisual elements. Meanwhile Reiss' original term audio-medial has been replaced completely by multi-medial l and has disappeared from the discussion. I would here like to

MARY SNELL-HORNBY

279

make a case for reviving it in the strictly limited sense of describing texts that have been written - or selected - for the express purpose of being spoken, which are therefore intended to reach their ultimate recipient by means of the human voice and not directly via the printed word. Examples would be formal speeches such as Churchill's address of 19 September 1946, academic lectures read from the manuscript, also passages from the Bible and other elements of liturgy read during religious services. The term multi-medial would apply for texts dependent on more than this one acoustic medium for their full realization, hence as soon as visual elements, music, screen or stage properties are included. The borderline between audio-medial and multi-medial is fluid, and much depends on how texts are actually used. Biblical texts have been used in oratorios and in drama, and are then multi-medial, and a case could be made for a radio-play being an audio-medial text even if background music and other sound-effects are included. Similarly, TV newsscripts are audio-medial in that they too have been written to be spoken, even if a visual medium is then added. There would be little point in splitting hairs on borderline cases: the important issue here, particularly where translation is concerned, is the type of language, and the nature of the text-type conventions, that are involved where a text is intended to be spoken aloud for an audience and not read silently. On the other hand, it lacks the added dimensions of dramatic dialogue (see Snell-Hornby 1993) and is not subjected to the constraints imposed on other multi-medial texts as by the music in opera (Kaindl 1995) or the interplay of varying codes in film-dubbing (Manhart 1995). Two basic factors are of course rhetoric and speakability. It is a feature of the audio-medial text, as against spontaneous spoken discourse or casual conversation, that such elements are planned and contrived. This means that the audio-medial text can be analysed with the same means and devices that characterize literary texts. The rhetorical structure of the opening paragraph of Churchill's speech quoted above is for example unusually vivid and transparent. Simple sentences alternate with complex syntactic structures, and these are characterized by a build-up of parallel constructions leading up to a climax of end-weight and end-focus (This noble continent, comprising..., enjoying..., is the home of all... of the western world'). Other parallel constructions are contradicted as it were by an antithesis ('It is the fountain... It is the origin.If Europe were once united...Yet it is from Europe...') and this again is resolved in another resounding climax ('wreck the peace and mar the prospects of mankind'). Particularly striking, as this last example shows, is the interplay of abstract terms and evocative, resounding monosyllables. Interplay, rhythm, antithesis, climax - these are the rhetorical devices that give life to the audiomedial text.

280

WRITTEN TO BE SPOKEN

At the heart of Katharina Reiss' theoretical approach is the hypothesis that it is the text-type that determines the method of translation (1976:1). As a general principle this may be debatable, as most texts are in fact hybrid forms, multidimensional structures with a blend of sometimes seemingly conflicting features (cf. Snell-Hornby 1988: 31), but in the case of the audio-medial text it is certainly arguable that the translation strategy should depend on the possibilities of expression inherent in the human voice. This can develop into an issue where the principles of rhetoric and speakability conflict with culturespecific expectations and language-specific text-type conventions, as the following examples from German and English may show.

The poetic text: the radio play and the Bible as liturgy Dylan Thomas called his Under Milk Wood, first broadcast as a radio play on 25 January 1954, a 'play for voices', and as such it is an borderline example of the audio-medial text.2 For our topic it is however interesting to note the artistic potential of the acoustic medium. In his Second Preface of 1974 the editor Daniel Jones - significantly a phonetician - declares: '...the play, whether heard in the mind or from a stage, is meant for the ear, which, unlike the eye, imposes no limits upon the imagination.' (1975: xii) Not unrelated to that are his remarks on translation: The success of Under Milk Wood in translation is significant. Verbal subtlety, word-play, the rhythm and the essence of the poetry, the strongly individual style of the 'prose' - if that is what we can call it - all these are lost in the process of translation. What is left? Humour, characterization, images, structure, the forward movement of the play (the word 'plot' can hardly be used), and robustness. If I had to choose one word to describe the quality that has ensured the survival of the play after the ordeal of translation, I would choose 'vitality'. (1975: xii) Jones does not mention which translation(s) he is referring to, but for our topic it is significant that precisely the acoustic qualities are those he finds missing, whereas an undefined quality of 'robustness' or 'vitality' is what survives. This might well be an indication that the translators have attempted to be faithful to the content of the text, but have overlooked the fact that it is actually intended for the voice and the ear. This is not however a generalization that can be made for all translation, not even of Dylan Thomas's other audio-medial literature. Looking at another of his texts A Child's Christmas in Wales (also derived from a radio script broadcast in BBC Wales Children's Hour on 20 December 1945) in its German translation by Erich Fried, we can see that the marked rhythm

MARY SNELL-HORNBY

281

and the many sound effects of the prose are precisely the elements that have survived in the translation. In my discussion of the opening paragraphs (SnellHornby 1988: 128) I emphasized that 'the syntactic pattern has been convincingly recreated in the German sentence-structure with all its cumulative effect of rhythm, piling up of phrases and convoluted but never confusing complexes of dependent clauses.' It would certainly be a worthwhile task to assess audio-medial texts - and their translations - which have been written to be read aloud and listened to (a clear skopos indeed) from the view point of their speakability and their message as a pattern of sense through sound. One English translation which has survived the centuries as a revered masterpiece of language (actually quite apart from its main skopos) is the King James - or Authorized - Version of the English Bible. In the Preface to the Revised Version (1952: v) we can read as follows: Its revisers in 1881 expressed admiration for 'its simplicity, its dignity, its power, its happy turns of expression... the music of its cadences, and the felicities of its rhythm'. It entered, as no other book has, into the making of the personal character and the public institutions of the English-speaking peoples. It is this poetic and acoustic quality - 'the music of its cadences, and the felicities of its rhythm' - which has enabled the Authorized Version of the English Bible to retain its position as an outstanding liturgical text, despite the claims of newer translations to be more easily comprehensible and more transparent for the churchgoer of today.3 In the text-typology of Katharina Reiss, the Bible could be described as a dominantly operative text, exhorting the faithful to certain modes of belief and behaviour, based on lengthy historical and narrative accounts, hence informative texts. Much is however intensely expressive, and some excellent examples can be found in the visions and exhortations attributed to the prophet Isaiah, such as this famous passage (from Chapter 40): 9. O Zion, that bringest good tidings, get thee up into the high mountain, O Jerusalem, that bringest good tidings, lift up thy voice with strength; lift it up, be not afraid; say unto the cities of Judah, Behold your God! 10. Behold the Lord God will come with strong hand, and his arm shall rule for him, behold, his reward is with him, and his work before him. 11. He shall feed his flock like a shepherd; he shall gather the lambs with his arm, and carry them in his bosom, and shall gently lead those that are with young.

282

WRITTEN TO BE SPOKEN

The rhetorical appeal of this passage is based on what are principally quite simple devices. The most obvious of these are rhythm and repetition, giving the language its speakability and musical potential (verses 9 and 11 are recognizable as the - in part modified - text for two Airs for the mezzosporano soloist in Handel's 'Messiah'). The syntactic structure of verse 9 falls into three parts; the first two have an identical triple structure, a vocative, an identical relative clause, and an exhortation; the imperative ('lift up thy voice') is taken up, extended by further imperatives to culminate in the forceful phrase 'Behold your God!' A similar principle is observed in verse 11, which is however a gentler paratactic construction based on short, simple verbs ('he shall feed', 'he shall gather', 'and carry', 'and shall gently lead'), in each case with an object and prepositional phrase of similar length, giving the effect of a simple stanza in a poem. Let us now look at the version in the New English Bible of 1961 : You who bring Zion good news, up with you to the mountain-top: lift up your voice and shout, you who bring good news to Jerusalem lift it up fearlessly; cry to the cities of Judah, Your God is here'. Here is the Lord God coming in might, coming to rule with his right arm. His recompense comes with him, he carries his reward before him. He will tend his flock like a shepherd and gather them together with his arm; he will carry the lambs in his bosom and lead the ewes to water.' Particularly in the first of these verses (9) it is striking that the rhetoric, the rhythm and indeed the musicality are greatly reduced. This is due mainly to the fact that the syntactic structures are uneven, both in form and length, hence weakening the force of the rhythm; in particular the final climatic phrase 'Behold your God', with strong end-focus on God, in the New English version sputters out in the weak statement (without end-focus) 'Your God is here'. Secondly, while the Authorized Version is stylistically unified and harmonious, the style of the New Version is again uneven, combining antiquated structures like 'you who bring' with colloquial phrases like 'up with you'. And thirdly, there is a weakness in the cohesion: the referent of it in 1.4 is not clear (strictly, it should refer to the last-mentioned noun, hence Jerusalem, and not voice back in 1.2.).

MARY SNELL-HORNBY

283

It is a matter for theologians and Classical philologists to debate on the most accurate turns of phrase to render the meaning of the Bible, and it is a matter for educationalists to decide on the most transparent wording for the growing generation of today; as part of the liturgy of Divine Service - as epistle, gospel or lesson - the language of the Authorized Version still remains for many unsurpassed as a masterpiece of literary translation and 'speakability'. The question remains whether a modern version of the divine truths it enshrines could be made to equal it.

The academic lecture One of my most frequent translation assignments over the years has been to provide English versions of academic papers and lectures by German-speaking colleagues for presentation at international conferences with English as lingua franca. This means that a speakable text had to be provided suitable to be read aloud by a non-native speaker (who was often no expert in English) and understood by a largely non-native audience. As those familiar with German conventions in scholarly language will know these are hardly conducive to such an aim (cf. Clyne 19914): scholarly standards are all too often linked with highly abstract language, complicated sentences full of long premodifying phrases and involved clusters of dependent clauses. This is a standard feature of scholarly language in German, but it is certainly not completely foreign to English, as can be seen from the following passage chosen at random from a classic in linguistics: Given alternative formulations of a theory of grammar, one must first seek to determine how they differ in empirical consequences, and then try to find ways to compare them in the area of difference. It is easy to be misled into assuming that differently formulated theories actually do differ in empirical consequences, when in fact they are intertranslatable - in a sense mere notational variants. Suppose, for example, that one were to modify the standard theory, replacing condition (3) by the condition that lexical items are inserted just prior to a transformation affecting the configuration in which they appear. Making this notion precise, we could devise an apparent alternative to the standard theory which, however, does not differ at all in empirical consequences, although the notion 'deep structure' is not defined, at least in anything like the sense above. Given the central character of this notion in the standard theory, the alternative would appear to be significantly different, though in fact it would be only a notational variant. (Chomsky in Steinberg and Jakobovits 1971: 187)

284

WRITTEN TO BE SPOKEN

It must be emphasized that this is not an audio-medial text, but part of a linguistic essay intended for silent scholarly perusal - but it is still the type of language not infrequently used (particularly in the German-speaking area) for conference papers. The high level of abstraction, the convoluted syntax and the density of Hard Words with their clusters of unstressed syllables make language of this kind difficult to articulate - as anyone who attempts to read the above passage aloud will confirm. It is the task of the audio-medial translator to make out of such language rhythmical prose with patterns of stress and focus that facilitate articulation and hence understanding. The example I am using to illustrate what is involved here is not however a piece of abstract jargon, but a passage combining special language with description, hence an informative text in Reiss' typology. It was part of an authentic translation assignment given to me in 1989 by a colleague in Zürich, Professor Rudolf Schnyder, Curator of the Swiss National Museum and President of the International Academy of Ceramics. The subject was Tendencies in modern European ceramics', and the text was to constitute his opening keynote lecture at the Academy's World Congress in Sydney, Australia. The audience would therefore consist of specialists - using English as a lingua franca. The lecture was illustrated by a number of slides (all of them made available to the translator), was therefore in part a multi-medial text dependent on visual material. The passage chosen here however formed part of the introduction and is purely audio-medial. The problem we are going to concentrate on is one of focus and stress-patterns that can be created by the translator through special strategies of word-order and syntax. Soviel zur Zeit vor 200 Jahren, zum Jahr 1788, dem Vorjahr zur französischen Revolution. Die französische Revolution aber wird dann das grosse Ereignis, in dem die sozialen Probleme der Zeit in einer gewaltigen Massenbewegung zum Ausbruch kamen. Sie brachte mit den für alle geforderten persönlichen Freiheitsrechten auf dem Kontinent auch die Einführung der Handels- und Gewerbefreiheit, brachte damit letztlich dem Unternehmertum mächtigen Auftrieb und leistete einer Entwicklung Vorschub, die auch auf dem Gebiet der Keramik rasch zu einer weiteren Technisierung der Arbeitsabläufe, zum Einsatz von Maschinen an Stelle der Hand, kurz, zu weiterem Fortschritt in der Industrialisierung führte. Diese Entwicklung wurde durch einen mit grossem Einsatz betriebenen Ausbau der Verkehrswege und der Transportmittel unterstützt und vorangetrieben; vor allem der Bau der Eisenbahn zeitigte in der Folge wirtschaftliche Auswirkungen von unabsehbarem Ausmass. Auf dem billigen Schienenweg erreichten die preisgünstigen Produkte von immer leistungsfähigeren Grossunternehmen bald die entlegensten Absatzmärkte. Damit schlug auch die Stunde

MARY SNELL-HORNBY

285

für viele jener in lokaler Tradition verwurzelte Kleinbetriebe, die sich bis dahin noch recht und schlecht durchgebracht hatten; manch einem gelang es nun nicht mehr, unter den veränderten Konkurrenzverhältnissen von seiner Hände Werk zu leben. Wer in seinem gelernten Beruf kein Auskommen mehr fand, musste nach andern Möglichkeiten suchen, sich durchzubringen. Viele suchten ihr Glück in der Ferne, wanderten aus. Difficulties are presented particularly by the long and syntactically involved sentence in the first paragraph (1. 4-11) and by repeated premodifying adjectival phrases (e.g. 1. 5, 11, 18-19), as well as a number of dependent clauses and phrases that lengthen the sentences and cause difficulties for the speaker. Concentrating particularly on these points, I arrived at the following version: So much for the time 200 years ago, so much for 1788, the year before the French Revolution, that great turning-point which was to make the social problems of the time erupt in a gigantic mass revolt. It led, not only to the civil rights and liberties demanded for all citizens on the Continent, but also to freedom of trade and competition; thus it boosted free enterprise and encouraged a development which even in the field of ceramics meant further mechanization of production, the use of machines instead of manual labour - in short, increased industrialization. Further impetus was provided by a massive expansion in transport and communications; the railways in particular had an immeasurable effect on the economic development. Cheap rail transport meant that the inexpensive products of ever-expanding commercial enterprises soon reached the remotest markets. That was the death-knell for many of those small local businesses which up to then had just managed to keep going; under the changed conditions many people could not get a living from the work of their hands and had to look elsewhere. Some tried their luck in distant lands - and emigrated. What is striking is that the translation is a good deal shorter than the German original (19 lines instead of 24), although no information has been omitted. The technique was firstly to avoid unecessary repetition (e.g.'French Revolution in 1. 2., see too 1. 19-20 of German text) and tautology (e.g. 1. 14 'unterstützt und vorangetrieben') and secondly to replace phrasal expressions in German by succinct lexical items, where possible emphatic monosyllables (e.g. 'erupt' for 'zum Ausbruch kamen", 'boosted' for 'brachte ... mächtigen Auftrieb', 'immeasurable' for 'von unabsehbarem Ausmaß'). In this way acoustic 'padding' was avoided, leading to a text that was more manageable for the non-native speaker to read out. In particular the long sentence (1. 4-11) in the first paragraph of the German would in comparable English syntax be an unnecessary strain on both breathing and intonation - even in the somewhat shorter English

286

WRITTEN TO BE SPOKEN

sentence both are alleviated by the semi-colon and the slight pause that goes with it. A special problem arises with the premodifying adjectival phrases that are a common feature of German for special purposes ('mit den für alle geforderten persönlichen Freiheitsrechten', 'durch einen mit grossem Einsatz betriebenen Ausbau', 'für viele jener in lokaler Tradition verwurzelte Kleinbetriebe'); through the clusters of unstressed words and syllables these syntactic structures are especially 'unspeakable'. Strategies here include a shifting of stress through syntactic changes, particularly postmodification ('the civil rights and liberties demanded for all citizens on the Continent') and telescoping into a more succinct phrase where only subsidiary information is reduced ('a massive expansion', 'small local businesses'). In the basic sentence patterns rhythm and stress were heightened by intensifying repetition ('So much for the time 200 years ago, so much for 1788'), by strong end-focus (e.g. 'gigantic mass revolt') and by exploiting patterns of intensification ('inexpensive products...everexpanding commercial enterprises...remotest markets'). Here too speakability is aided by the interplay of abstract terms and resounding monosyllables, and by a forward movement produced by patterns of stress and focus based on succinct and precise expression.

The audio-medial text and simultaneous interpreting Audio-medial translation as described above is of course a conscious, carefully planned process based on deliberate strategies. As such it cannot be compared with the more spontaneous task of interpreting, although this too is an audiomedial activity5. However, if we study the transcripts of interpreted material, we can confirm that professional interpreters, because they are bound to the spoken medium and have to work under extreme time-constraints, naturally tend to apply some of the strategies (e.g. reduction of complicated syntactic expression and redundancy, cf. Strolz 1992: 155) which we have been discussing above. One interesting example can illustrate this. During the 'Central European Symposium' held in Vienna in 1991 for colleagues from former Eastern bloc countries, a paper on opera translation (Kaindl 1992) was interpreted simultaneously for those present who could not understand German. Due to its density of specialized expression it was a special challenge for the interpreter. This passage may serve as an example: Wenn wir nun wie Fritz Paepcke davon ausgehen, daß nicht Wörter oder Sprachen, sondern Texte übersetzt werden, dann müssen wir zunächst die Oper als Ganzheit untersuchen. Die Tatsache allein, daß an der Textkon-

MARY SNELL-HORNBY

287

stitution mehrere Medien beteiligt sind, hat wenig Erkenntniswert für die Erfassung der eigentlichen Textgestalt, die sich erst aus den Relationen der einzelnen Teile zueinander ergibt. (...) Wesentlich für unsere Zwecke ist daher die Feststellung, daß die am Gesamttext beteiligten Medien nicht nebeneinander ablaufen, sondern sich durch simultane Wahrnehmung zu einer Ganzheit fügen. Durch das Aufeinandertreffen von musikalischen, sprachlichen und szenischen Zeichengestalten entsteht etwas qualitativ Neues, das nicht mit der Summe seiner Teile ident ist. Wenn sich die eigentliche Textgestalt erst aus der Wechselwirkung der in unterschiedlicher Gewichtung am Gesamttext beteiligten Medien ergibt, so kann das Schwergewicht der Analyse nicht auf den einzelnen Teilen liegen, sondern muß das Relationsgefüge, die Beziehungen der einzelnen Elemente zum Gegenstand haben. (Kaindl 1992: 60) Let us compare that with the transcript of the simultaneous interpretation: If we agree with Fritz Paepcke that we do not translate words or languages but texts, then I think we should view opera as a holistic entity. The fact that opera as text is constituted by various media tells us little about the actual gestalt qualities of the text, which are derived from the relationships between the various elements. What is important for us here is the realization that the media involved in the text do not run parallel but rather are perceived simultaneously and do constitute a coherent whole. The combination of music, language and scenic action leads to the emergence of something that has a different quality and is more than just the sum of its parts. If the gestaltlike nature of the text is derived from the interaction among the components and their varying emphasis, then our analysis must not focus on isolated parts but rather must look at the network of relations among the constituent parts. An analysis of these two versions leads to the conclusion that similar strategies have been used as was the case with the English translation of the passage from the lecture on ceramics: telescoping into succinct phrases ('tells us little' for 'hat wenig Erkenntniswert für die Erfassung'; 'music, language and scenic action' for 'musikalischen, sprachlichen und szenischen Zeichengestalten'), the reduction of redundancy, which in the German version serves as explanation ('the network of relations' for 'das Relationsgefüge, die Beziehungen der einzelnen Elemente'), plus a noticeable tendency towards strong end-focus. And these are features which, at least for the English audio-medial text, heighten speakability and facilitate understanding. In recent years Translation Studies has been devoting more attention than was previously the case to the characteristics and conventions of texts, not as static specimens of language, but as they function within their individual extralinguistic situation and for the specific purpose they have been intended

288

WRITTEN TO BE SPOKEN

for. Within such a broad constellation of texts and text-types, the audio-medial text - or the text used for audio-medial purposes - deserves scholarly attention. The deficit that exists at present often becomes glaringly obvious at international conferences, as the number of only partially intelligible papers given unwittingly in English by non-native speakers (cf. Pöchhacker 1994) may show. The crucial element is here the potential - and the limitations - brought about by the medium of the human voice, in this case not in combination with other audiovisual elements, but as the vehicle of direct communication, whether the text is expressive, operative or informative. We have been concentrating here on strategies for heightening the immediacy and clarity of expression in English - for other languages other strategies and conventions will apply, and for other cultures there will be other norms governing the interaction between speaker and hearer or speaker and audience. In a time of ever-increasing communication by means of the spoken word, the audio-medial text might well prove to be a topic worth resurrecting. Notes 1 To avoid confusion with the term multimedia as used in a computer context, the German term multi-medial has been taken over unchanged into English. 2 It borders of course on dramatic dialogue. It was indeed presented on the stage in the Edinburgh Festival in 1956, of which extracts were shown on television. The film of the play opened the Venice Festival in 1971. Here however we are concerned only with the acoustic dimension as present in the text for voices. 3 It is also favoured for dramatic or ceremonious situations or to elicit an emotional response, and it is significant that during the ceremony marking the 50th anniversary of the end of World War II in Hyde Park on May 7 1995 - actually an emphatically 'popular'event in trendy modern style - the passage chosen to be read out from the Bible was from the Authorized Version and not from a modern translation. 4 I am indebted to the University of Zürich for placing the unpublished manuscript at my disposal. 5 The basic difference emerges very clearly from Otto Kade's definition of Übersetzen as against Dolmetschen (1968: 35), the one concerning texts which are permanent and hence open to control and repeated correction, the other concerning discourse or text-material (usually presented orally) which for practical reasons (e.g. lack of time) can hardly be controlled or corrected at all. References Source texts Bible Translations: The Holy Bible containing the Old and New Testaments, translated out of the original tongues; and with the former Translations diligently compared and revised, by His

MARY SNELL-HORNBY

289

Majesty's special command, appointed to be read in churches, Cambridge: CUP (Authorized Version). The Holy Bible, Revised Standard Version containing the Old and New Testaments, translated from the original tongues, being the version set forth A.D. 1611, revised A.D. 1881-1885 and A.D. 1901, compared with the most ancient authorities and revised A.D. 1952, London: Nelson (Revised Version). The New English Bible with the Apocrypha, Oxford: Cambridge UP, 1970. Churchill, Winston, Ansprache an die akademische Jugend der Welt, Zürich, 19. September 1946.5 Kaindl, Klaus, '...weil ich innig liebe'. Stimme und Gestalt in der Oper 'Carmen' von G. Bizet, in: M.Snell-Hornby (ed.) Translation in Mitteleuropa, Prague: Charles University 1992, p.60. Schnyder, Rudolf, 'Tendenzen in der europäischen Keramik von heute', unpubl.ms. Secondary literature Bühler, Karl. 21965. Sprachtheorie. Stuttgart: Fischer. Clyne, Michael. 1991. "The sociocultural dimension: The dilemma for the German scholar". In Hartmut Schröder (ed), Subject-oriented Texts. Berlin: de Gruyter, 4968. Jones, Daniel. 1975. Second Preface to Dylan Thomas, Under Milk Wood. London: Everyman. Kade, Otto. 1968. Zufall und Gesetzmäßigkeit in der Übersetzung.Beihefl zur Zeitschrift Fremdsprachen 1. Leipzig. Kaindl, Klaus. 1995. Die Oper als Textgestalt. Perspektiven einer interdisziplinären Übersetzungswissenschaft. Tübingen: Stauffenburg. Manhart, Sibylle. 1995. Zum übersetzungswissenschaftlichen Aspekt der Filmsynchronisation in Theorie und Praxis: Eine interdisziplinäre Betrachtung. Vienna, unpubl. diss. Pöchhacker, Franz. 1994. Simultandolmetschen als komplexes Handeln. Tübingen: Narr. Reiss, Katharina. 1971. Möglichkeiten und Grenzen der Übersetzungskritik. Kategorien und Kriterien für eine sachgerechte Beurteilung von Übersetzungen. Mün­ chen: Hueber. Reiss, Katharina. 1976. Texttyp und Übersetzungsmethode. Der operative Text. Kron­ berg: Scriptor. Reiss, Katharina, 1990. "Brief an den Herausgeber". Lebende Sprachen 4, 185. Reiss, Katharina and Vermer Hans J. 1984. Grundlegung einer allgemeinen Translationstheorie. Tübingen: Niemeyer. Snell-Hornby, Mary. 1988. Translation Studies. An Integrated Approach. Amsterdam: Benjamins. Snell-Hornby, Mary. 1993. "Der Text als Partitur: Möglichkeiten und Grenzen der multimedialen Übersetzung". In Justa Holz-Mänttäri and Christiane Nord (eds), Traducere Navem. Festschrift für Katharina Reiss zum 70. Geburtstag, Tampere: Univ. Press.

290

WRITTEN TO BE SPOKEN

Spillner, Bernd. 1980. "Semiotische Aspekte der Übersetzung von Comics-Texten". In Wolfram Wilss (ed), Semiotik und Übersetzen. Tübingen: Narr, 73-86. Strolz, Birgit. 1992. Theorie und Praxis des Simultandolmetschens. Argumente für einen kontextuellen Top-down-Ansatz der Verarbeitung und Produktion von Sprache. Vienna, unpubl. diss.

Dubbing and the Dubbed Text - Style and Cohesion: Textual Characteristics of a Special Form of Translation Thomas Herbst Universität Erlangen-Nürnberg

Introduction Although the two most widely established forms of film translation, dubbing and subtitling, differ greatly in character, it is by no means obvious which of the two is superior. Research suggests that although there are clearly established preferences amongst viewers in favour of the one or the other, these vary between different countries and the tendency seems to be that viewers in socalled subtitling countries such as the Netherlands or the Scandinavian countries state subtitling as their preferred form of film translation, whereas viewers in the so-called dubbing countries such as France, Italy or Germany clearly prefer dubbing - with social class, educational background and the individual's own command of a foreign language proving to be important factors influencing the strength of the preference stated.1 It is obvious that for viewers with some command of the original language of a film subtitling has obvious advantages. Also, the practice of subtitling probably has a very positive effect on the overall standard of foreign languages in the subtitling countries. Nevertheless, it cannot be denied that as far as any notion of translational equivalence is concerned, subtitling is subject to the inbuilt problem of a shift of medium in that it provides a written translation of a spoken text. Dubbing is subject to constraints of a different type. One major problem, of course, is presented by finding suitable voices to match the voices of the characters of the original film. In practice, this is not as difficult a task as it may first seem since research has shown that there is absolutely no need for the voices of the original actor and the dubbing actor to be similar. Viewers tend to accept a particular voice as that of a given person, and apparently it is only in cases where one actor dubbing a particular character in a tv series has to be re-

292

DUBBING AND THE DUBBED TEXT

placed by another for one reason or another that viewers complain that the new voice is unnatural or artificial. On the whole, if the casting of dubbing actors is done well, the dubbing voice of a particular character seems to appear as natural to a tv or cinema audience as the voice of the original actor is to the audience seeing the film in the original. Dubbing voices only seem unnatural to people who are familiar with the real voices of a character (but, equally, the real voice seems unnatural to people used to the dubbing voice.). For this reason, it is important that well-known actors should always be dubbed by the same dubbing actor. Apart from the fact that due to a lack of dubbing actors, the same dubbing actors tend to be used for a number of foreign actors, which may lead to some estrangement in the viewing audience, voice casting in itself does not seem to present any great problems negatively affecting the quality of the translated film.

Lip sync and nucleus sync as constraints in dubbing A much more important problem than that of voice casting is presented by the fact that - with the exception of song translation - dubbing is the only form of translation in which the length of the translated text has to be identical with the length of the original text. Furthermore, the dubbed text has to match certain visual features of the original film, which imposes far more serious constraints on translation for dubbing than in any other form of translation. With respect to translational equivalence, two central criteria have to be added to the established list of factors to be considered:2 1. lip sync, i.e. the correlation of the lip movements of the original actors with the sounds perceived in the dubbed version, where a distinction can be made between quantitative lip sync (referring to the correlation of beginning and end of visible lip movements and sound) and qualitative lip sync (referring to the correlation of actual lip movements or shape of the mouth and the quality of the dubbed sound), 2. nucleus sync, i.e. the correlation of gestures or particular movements of the head (such as the raising of the eyebrows) with nuclei, i.e. syllables carrying accentual prominence in the dubbed version. Although these two types of sync impose serious constraints on the creation of dubbed dialogue, their importance should not be overrated: Considerable room for manoeuvre is gained because the degree to which these factors actually influence the translation varies considerably between films or within a film depending on the number of off-scenes, close-ups, etc. For instance, a rough

THOMAS HERBST

293

estimate for one episode of the British tv series Yes Minister calculates that lip and nucleus sync would have to receive considerable attention for about one quarter of the text and could almost be neglected for 33% of the text. Furthermore, the standard of sync required in a normal production is actually limited: It is a fact that not only would it be impossible to achieve absolute sync, it would also be completely unnecessary. Research shows quite clearly that viewers do not find even relatively crass violations of sync at all disturbing. This applies to quantitative as well as qualitive lip sync. Instances of dubbed text setting in 0.56 secs (14 pictures) before the lip movements or finishing up to 0.84 sees (21 pictures) later do not seem to be noticed by viewers; furthermore such translations as ich weiß, wovon ich rede for and I have with bilabial consonants in the place of an open vowel seem to pass unnoticed. Hence, within certain limits, there seems to be a considerable latitude as far as the requirements of sync are concerned and thus room for textual manipulation in the translation process.

The scope of the analysis The aim of this analysis is to reveal to what extent the contextual restraints on the translation and other factors inherent to dubbing result in specific features that could be regarded as characteristic of dubbed text. Dubbing practices seem to vary in different countries - with Germany apparently generally putting more emphasis on lip sync and Italy perhaps more emphasis on natural dialogue - so no general conclusions about the nature of dubbed language will be aimed at here and only German will be considered as a target language. Furthermore, it has to be said that since there are considerable differences in the overall quality of the approaches taken within dubbing into German, it will not be argued that the features identified below can be found in any film dubbed into German. For instance, the dubbing of the BBC-Shakespeare-productions into German is exceptional in that it poses problems of quite a different nature.3 However, it will be claimed that the characteristics listed are typical of dubbed material of a particular genre that makes up a considerable proportion of the programmes shown on German television and which also obtains peak viewing rates.4

Film language as spoken language: text typological considerations Dubbed texts, like other forms of film dialogue or drama, take up an intermediate position between the written and the spoken language. On the basis of the distinctions made by Gregory (1967), Elam (1980) and Soil and Hausmann

294

DUBBING AND THE DUBBED TEXT

(1985), three criteria can be used to distinguish the language of film from other texts: 1. Intended channel of perception: They are intended to be listened to not read, which distinguishes them from novels or newspaper articles (which are intended to be read) or poems and perhaps plays (which can be both read or listened to). 2. Degree of preparation: They have been carefully prepared, which does not apply to spontaneous conversation or some types of informal talk. 3. Genuineness of the communication: The communication taking place between the partners in the conversation is not genuine, unlike in spontaneous conversation, since the real level of communication is that between the author of the film and the viewers.5 For these reasons, the language of film (and drama) differs from that of spontaneous conversation in a number of ways. Most typically, all those features that, in spontaneous speech, result from performance-related factors (such as pauses or hesitation phenomena6) or overlaps are typically absent from film dialogue.

The unnaturalness of dubbed text With respect to the parameters identified above, there are no typological differences between the original film dialogue and the dubbed text. However, if one follows general observations, original film dialogue must be different from dubbed dialogue in a few striking ways. One indication of this is that just as most people - without seeing the pictures - would be able to tell on the basis of a tape recording whether they are listening to a recording of spontaneous conversation or the sound track of a film (which can be explained in terms of the absence of performance factors) they are also able to tell quite easily whether they are listening to the sound track of an original or a dubbed film. One factor7 that may account for this is the intonation and speech style of many dubbed programmes: There seems to be a tendency for dubbed films to lack the full range of accentual contrasts and pitch movements. A certain monotony of delivery can easily be accounted for in view of a recording technique that cuts up the film into a large number of so-called takes, which are all dubbed separately so that it becomes very difficult for the dubbing actors to achieve natural intonation and exploit the pitch range to the full. A further factor that contributes to a certain unnaturalness of dubbed films and distinguishes them from original programmes in many cases is the extensive use of the standard language: In Germany, at least, all films are dubbed in-

THOMAS HERBST

295

to the standard language. This is the inevitable result of the translation process because - with very few exceptions - no regional or social accent (apart from the standard language) can be used as a target variety in translation without producing a comical or strange effect. Even if accent contrast (between British and American English or particular social varieties of British English) form an important element of the original film, they cannot be expressed by a similar contrast of accents in the target language. Although certain social connotations of accents may be similar or the idea of regional contrast might be signalled in this way, all varieties other than the standard language are regionally marked within the target culture and thus cannot be used. The use of the standard language in dubbing has two important consequences: (i) the loss of those elements of meaning that would fall under Leech's (1981) category of social meaning, (ii) a certain lack of distinction between various speech styles in the dubbed version. This is a considerable disadvantage with respect to the differentiation of various characters in a film; furthermore, lack of linguistic contrast inevitably results in a certain monotony of expression. In the case of a language such as German, the use of the standard poses one further problem, which is caused by the fact that in the German-speaking world the standard accent Hochdeutsch is hardly ever used in natural speech; at least not in informal conversation. Unlike Received Pronunciation in Britain, Hochdeutsch (in the sense of an accent that betrays no regional influence whatsoever) is not so much a social accent as a variety of German whose use is mainly determined by situational factors. Although the increased use of the standard language could generally be seen as a feature that distinguishes natural conversation from film language, comparable German productions tend to use (modified) regional and social accents so that the use of a regionally neutral pronunciation (in situations where it would not normally be expected) must be seen as a key characteristic of dubbed films.8

Stylistic features of dubbed text This latter aspect of the use of the standard language has important consequences with respect to the way the text is perceived stylistically. Since, in the German-speaking world, the absence of regional forms of pronunciation tends to coincide with rather formal occasions and speech styles (such as reading the news on television or most forms of drama on the stage), the use of the standard as the only variety by characters in dubbed films in which greater regional variation would be expected in corresponding German films may be reallocated in terms of style.

296

DUBBING AND THE DUBBED TEXT

In fact, one of the main characteristics of films dubbed from English into German is that the dialogue displays a considerable number of formal elements. Although it has been pointed out above that with the exception of performance-related features film language otherwise tends to have the characteristics of the spoken language, this could not be said of dubbed programmes. On the contrary, many features typical of the written language can be found in German dubbed programmes: (i)

insertions that are not typical of the spoken language:

(1) Ich möchte keine Wunden aufreißen, aber du hast selbst, wenn auch nur für kurze Zeit, erfahren, was es heißt, ein Kind zu haben. Du kannst dir sicher vorstellen, was ich empfinde.

I don't want to bring up sad memories ... but you were a father ... even for only a few minutes of that poor baby's life ... you can understand how I feel, can't you?

(2) Wie du weißt, wurde ich aber nicht von ihm gerettet, weil so ein Mörder nicht existiert.

Except he didn't get me out, did he?

(ii)

use of the subjunctive:

(3) Chippie sagte mir, er hätte dir erzählt. daß es mir nicht besonders gut ging.

Chippie said he'd seen you, and told you I had muddled through.

(4) Ich hörte, du hättest einen Unfall gehabt ... du bist vom Pferd gefallen.

I ... oh I heard you had an accident, thrown from a horse. Are you all right?

(5) Und ich wünschte, ich könnte dabeisein.

I just hope I'm around to see it.

(iii) subordinate clauses typical of the written language: (6) Sollte das die Wahrheit sein....

If that's it, Krystle ...

(7) Die Polizei sagt, wäre ich in der Zeit zu Hause gewesen, hätte man mich getötet.

The police told me that had I been there during that robbery, I could have been killed!

(iv) use of participle constructions: (8) Deinem Vorschlag folgend hatte ich Fallon gebeten, das Kind zu mir zu bringen, aber das wollte sie nicht.

So I did what you suggested. I went to Fallon and I asked her to bring the child to me. But she wouldn't agree.

THOMAS HERBST

297

(v) heavy premodification in noun phrases: (9) Ich habe die bisher noch bruchstückhaften Fakten dieses Falles überprüft.

I've reviewed the basics of this case sketchy as they are for the moment.

(vi) nominal style: (10) Ich habe nämlich das Gefühl, daß Ihr Ärger über den Verlust von Stevens Zuneigung nur die Spitze eines Eisberges ist.

Well I happen to think that your loss of Steven's affection is only the tip of the iceberg here.

(vii) frequent use of the past tense instead of the present perfect: A further element contributing to the formality of dubbed texts is the frequent use of the past tense to refer to past time, where many dialects of German would prefer present perfect forms. Admittedly, regional differences within Germany are considerable in this respect and to a certain extent the stylistic associations connected with the use of the past tense may vary within the German speech community, but a statistical comparison of the use of past tense and present perfect forms in original German productions and programmes dubbed into German shows a higher proportion of present perfect forms in the original productions (69% versus 57% in dubbed films).9 Since in many varieties of German the past tense is typically used in the written language, its frequent use in dubbed dialogue constitutes another formal element of such texts: 10 (11) Und an dem Abend, in dem Restaurant, ... ist da noch irgend etwas, äh, besonderes passiert... das Sie ärgerte?

Now on the night in question, did anything else in particular happen to upset you?

(12) Vor einer halben Stunde rief ich an» wie lange dauert es, bis Sie ein Pferd satteln?

I phoned down here one hour ago. How long does it take to get one horse ready?

(13) Als du mich gestern anriefst. Adam, und sagtest, du hättest einige Papiere, die ich unterzeichnen müßte ... dachte ich, du bringst sie mir ins Büro.

When you phoned me last night, Adam, and told me that you had some routine papers for me to sign, I assumed that you'd bring them to the office.

Apart from grammatical features, lexis plays an important part in establishing the formal style typical of dubbed texts. Thus, some of the vocabulary that is used in dubbing is most unlikely to occur in informal conversation:

298

DUBBING AND THE DUBBED TEXT

(14) Ich hab' gesehen, wie mein Vater Sie angeblickt hat, Sie haben ihm Angst eingejagt.

I saw how my father looked at you. You frightened him.

(15) Und selbst wenn ich mit irgendwelchen Menschen Probleme habe, verglichen mit deinen, sind sie geringfügig.

And, if I have any problems in my life, believe me, they're minor compared to yours.

(16) In meinem Bestreben, das Beste für dich zu wollen ... habe ich wohl eure Gefühle füreinander falsch eingeschätzt.

Well, in my desire to want the very best for you ... I think I've ... a ... misjudged your feelings for each other.

(17) Nur keine Sorge,... ich versichere Ihnen, nichts wird Blake Carrington widerfahren. für das ich nicht persönlich verantwortlich bin.

Don't worry. I'll make certain nothing happens to Blake Carrington that I'm not personally responsible for.

Such lexical choices further emphasize the impression of a formal style in dubbed films although the words employed need not always be considered to be formal as such. What is very important in this context is the phenomenon that has been described as a shift in style value by Söil and Hausmann (1985) for French, an approach which has been applied to English by Barnickel (1980: 146), who explains the principle as follows: Danach gehört ein bestimmtes Wort im code parlé einer höheren Stilebene an als im code écrit und umgekehrt, d.h. beim code-Wechsel verschiebt sich die Stilzugehörigkeit eines Wortes um eine Stufe. The principle of a medium-dependent style value may not apply to all lexical items of a language in the same way, but it is extremely important to be aware of the fact that words that would not be considered marked stylistically if used in the written medium might appear formal in the spoken language. Although the stylistic classification of individual words may present certain problems it is clear that words such as Bestreben, anblicken or widerfahren, which are styl­ istically unmarked (or at least not very formal) in written German, appear much more formal in the type of spoken language that dubbed dialogue typic­ ally represents. Thus the use of vocabulary that - for spoken conversation appears inappropriate in this respect adds to the written stylistic dimension of dubbed texts. Some of the stylistic inadequacies can be seen irrespective of the written/ spoken dichotomy, however. This applies, for instance, to the following collo­ cations and phrases, which are by far too formal for any kind of natural dia­ logue:

THOMAS HERBST

299

(18) Ich glaube, daß Sie mich angreifen ... weil Ihnen der Fall auch noch Krystles Zuneigung entzogen hat.

I think you're attacking me because this suit has obviously cost you her affection as well.

(19) Nachdem ich Dir meine heimlichen Ab­ sichten enthüllt habe, enthüll' mir deine.

I've told you my ulterior motives, now let's hear yours.

(20) Du bist nicht nur clever ... und sexy sondern auch noch im Begriff, mich zu einem sehr glücklichen Mann zu ma­ chen.

Not only are you smart... and beautiful... and sexy ... but you are also about to make me a very happy man.

(21) Ich bin nur um meines Kindes willen hergekommen und wechsle nicht die Seiten wie du.

I'm here for my baby's sake. I don't keep changing sides the way you did.

Thus there can be no doubt that one of the key features of the language of films dubbed into German is a tendency to aim for a stylistic level which is too formal to be appropriate to the spoken language of film dialogue (even if this differs in many ways from spontaneous conversation, as pointed out above). What is even more typical is that this formal style level is usually not a consistent feature, rather formal elements occur in a random way in the text. Hence, dubbed language is typically characterized by a large number of unmotivated style shifts: (22) Okay, okay, Fallon, würdest du bitte mal einen Moment die Luft anhalten? Bevor wir jetzt ins Haus kommen, wo dein Bruder ist und Krystle und all das Durcheinander, sollten wir die Gelegen­ heit zu einem Gespräch nutzen.

La, la, la ... Fallon, would you wind down for a minute, please? We go back to the house, with your brother there and Krystle and all that confusion and may not get another chance to talk.

Cohesion in dubbed texts The stylistic characteristics of dubbed language are also relevant to another parameter in the analysis, namely textual cohesion. In this context, it is useful to consider Enkvist's (1964: 34-35) notion of style markers: We may ... define style markers as those linguistic items that only appear, or are most or least frequent in, one group of contexts. ... Elements that are not style markers are not stylistically neutral. This may be rephrased: style markers are mutually exclusive with other items which only appear in

300

DUBBING AND THE DUBBED TEXT

different contexts, or with zero; or have frequencies markedly different from those of such items. If one accepts this view, one might argue that a word such as Haus might be seen as having a neutral style value because there are no stylistic choices, but that words such as anblicken or widerfahren are stylistically marked in the spoken language and that it is these words that contribute to the cohesion of the text if style is seen as a relevant factor in this context (cf. Brown and Yule 1983: 194). The cooccurrence of elements that are marked as formal (such as die Gelegenheit zu einem Gespräch nutzen in the above example) with elements that are marked as informal {wo dein ... ist, und all das ...) results in a type of stylistic inconsistency that impinges on textual cohesion. Similar discrepancies arise with respect to other factors that play a part in establishing textual cohesion. For instance, inappropriate use of pro-forms is equally typical: (23) A: Du hast doch wohl nicht etwa Angst? B: Angst? ... Nein die habe ich natürlich nicht. (24) A: Kein Kommunique entspricht jemals dem, was Sie wirklich sagen! B: Und warum gibt's die?

Well, you're not afraid, are you? Afraid? No. Of course, I'm not afraid.

Well no communique ever bears any rela­ tion to what you actually say. Then why have one?

This not only applies to pro-forms referring to noun phrase antecedents but also in cases of clausal antecedents: (25) A: Gäbe es Jeff nicht, wären wir vielleicht noch Freunde, so wie früher, aber ich ... B: Fallon, ich habe jetzt wirklich keine Lust, mit dir darüber zu sprechen. A: Aber ich möchte es.

I mean, maybe if Jeff weren't here, we could still be friends like before. But — I don't want to have this conversation, Fal­ lon. Well, I do.

THOMAS HERBST

301

Inappropriate pro-forms are very common with questions: (26) A: Ihn zu verstecken, ist keine Lösung. B: Welche ist es dann? Sie weitermachen zu lassen, so wie Sie ihn verteidigt haben? (27) A: Hast du wirklich keine Beschwerden? B: Nein, keine.

Look, hiding him is no answer. Well then, what is? Letting you handle it, like you handled his defense?

Are you sure there's nothing wrong? No. Nothing.

Similarly deviant uses can also be found with another characteristic means of creating cohesion identified by Halliday and Hasan (1976: 143), namely ellipsis. Ellipsis is used in an untypical way in dubbed films because, for instance, elliptical questions such as the following would seem rather strange in German: (28) Jeannette, irgendeine Nachricht von meinem Vater, während wir weg waren?

Jeannette, has there been any word from my father while we were out?

(29) Es hat sich nichts geändert. Schon eine Entscheidung?

Nothing's changed. Any verdict yet?

Within lexical cohesion the original and the dubbed text sometimes employ different types of reiteration (cf. Halliday and Hasan 1976: 278). If, for instance, repetition in the original is replaced by the use of a near-synonym, cohesion as such may not be affected; since what CGEL (19.23) refers to as the "confirmation of a phenomenon" is missing, there may be negative effects with respect to the rhetorical impact: (30) A: Soll ich dir was verraten, Steven ... Wann ich das erste Mal gemerkt habe, daß ich dich liebe? Als du in meiner Küche einen Teller zerschmissen hast... weißt du noch? B: Ich erinnere mich.

You know what, Steven, you know the first time that I knew I loved you? It was when you dropped that dish on my kitchen floor. Do you remember?

I remember.

302

DUBBING AND THE DUBBED TEXT

One further aspect that deserves attention are those relations within a text that might not be easily describable in terms of formal cohesion in the sense of Halliday and Hasan (1976) - unless one wished to expand their rather vague notion of collocation further - but which are covered by concepts such as "pragmatic or semantic implication" in CGEL (19.20) or a theory of frames (van Dijk 1977: 99): An important COGNITIVE condition of semantic coherence is the ASSUMED of the worlds involved. That is, our expectations about the semantic structures of discourse are determined by our KNOWLEDGE about the structure of worlds in general and of particular states of affairs or courses of events. NORMALITY

The textual character of any such real world link can only be realized, however, if it is apparent in the text in some way, which is not always the case in dubbed text: (31) A: Sieht so aus, als hättest du deine Chance verpaßt. B: Mich interessiert keine Wette. (32) A: Oh, oh! Wie heißt es so schön in Romanen? Wir dürfen uns so nicht mehr treffen. B: Dramatisch klingt das.

Looks you've missed your chance.

Look, I ... I'm not interested in placing a bet. Oh, oh. What is that saying? We've got to stop meeting this way.

Something like that.

An utterance such as (32) is meaningless to a German viewer since Wir dürfen uns so nicht mehr treffen can in no way be related as a quotation or standard phrase from novels. On a different level, the referent of the pro-form das cannot be identified in: (33) A: Mach mir doch gleich was mit ... ich esse alles, was du ißt. B: Ich bin nicht richtig hungrig.... Ich weiß gar nicht, warum ich hier bin. A: Nein, nicht, Fallon, ich finde, wir sollten das klären.

Cold ham on mine - I'll take whatever you're making. I'm not really hungry. I don't even know why I came down. Come on, Fallon. Let's clear the air, okay?

THOMAS HERBST

303

The conclusion to be drawn from such examples is that, in German dubbed material, the use of cohesive elements such as those identified by Halliday and Hasan (1976) or Brown and Yule (1983 ) has revealed two characteristics: 1. violations of the linguistic norms with respect to the realization of cohesion in cases where ellipsis or pro-forms are used inappropriately, 2. an overall reduction of the cohesive ties in cases where the point of reference of a pro-form is unclear or not retrievable from the text at all. Even if the degree of formal cohesion of text may not be measurable, it seems to be a characteristic of German dubbed material that certain cohesive ties which are explicit in the original are lacking. This allows interesting conclusions as to the role of explicit cohesive ties with respect to the textuality of a text because there are no indications whatsoever that dubbed films are more difficult for viewers to follow than original productions, which definitely points to the importance of semantic (or content-driven) strategies of interpretation rather than dependence on formal realizations, as is pointed out, for instance, by de Beaugrande and Dressier (1981: 88) and by Brown and Yule (1983: 196): It seems to be the case then that 'texture', in the sense of explicit realisation of semantic relations, is not criterial to the identification and co-interpretation of texts.

Anglicisms One further characteristic of texts dubbed from English into German, which cannot be discussed here in any detail (cf. Herbst 1994, 1995), is the high occurrence of particular types of anglicism. It is not the use of English words in German text, which is a main source of anglicisms in many spheres of German such as advertising, pop music programmes on the radio etc., that is particularly important in this context. Dubbed films, however, contain a high proportion of loan meanings where the German word takes over meaning components from an etymologically related word in English (such as kontrollieren - control, lernen - learn or sehen - see), loan translations (Küstenlinie for coastline), loan idioms (der frühe Vogel fängt den Wurm - the early bird gets the worm) or anglicisms at the level of grammar (reden über used in situations where reden von would be appropriate) or at the level of pragmatics, which is particularly noticeable with greetings or phrases of the type Mir geht es gut, which are used in inappropriate contexts.11

304

DUBBING AND THE DUBBED TEXT

Textual characteristics of films dubbed into German Although, as emphasised above, the analysis of the material investigated may not allow any general conclusions to be drawn that would necessarily apply to all films dubbed from English into German - the type of film and the quality of the translation being very important variables -, it seems possible to conclude that certain features can be identified that characterise a considerable proportion of the programmes dubbed from English shown on German and Austrian television. The most important of these features are • a very high occurrence of anglicisms (of the type indicated in the section above) • stylistic elements that are inappropriate to the language of (film) dialogue and are often more typical of written than spoken German • inappropriate or reduced realization of cohesive elements.

The characteristics of dubbed text and the translation process It must be understood that the fact that the films analysed share certain characteristics should not be taken to mean that we can establish a text type dubbed film on these grounds because these characteristics cannot be regarded as consciously chosen by the author or as conventionally established but as the (unintended) result of a translation process that is subject to more severe constraints than other forms of translation. The question is, of course, to what extent these features are the inevitable result of dubbing. For some of the features outlined above, it is certainly true that they derive from the fact that dubbed texts are translated texts. Especially the use of the standard language is a necessity in any kind of translation into German, although it has the negative effects of using a rather formal stylistic connotation and of levelling out linguistic differences between characters. However, it would be a mistake to view the other main characteristics of dubbed texts - anglicisms, lexical and grammatical items typical of formal style and reduced cohesion - as automatic and unavoidable consequences of translation for dubbing, at least to the extent to which they actually occur in many dubbed films. Rather, they must be seen as the result of a particular approach to dubbing. It seems to be common practice in Germany to have two stages of translation for films (cf. Whitman-Linsen 1992: 104-22): first, a kind of rough translation is made on the basis of the script of the original film (but often without seeing the original film). This rough translation is used as the basis for the writing of the dubbing script, which in very many cases is carried out by the dub-

THOMAS HERBST

305

bing directors themselves. This dubbing script is based on the takes into which the film is divided for the purposes of the actual dubbing in the studio. The dubbing script is often modified during the process of dubbing to achieve better lip sync etc. This means that at no stage a translation is produced that would be in line with central demands of any theory of translation, i.e. the text is not seen as a text and the corresponding equivalence criteria are not considered. Since the main function of the rough translation seems to be to serve as a kind of semantic quarry for the dubbing script, no attempt is made at this stage to produce natural dialogue or idiomatic German. Given the apparent function of this rough translation (and the relatively poor pay), such an approach may perhaps seem justified; however, since a considerable amount of the actual phrasing of the rough translation seems to find its way into the final televised version, its dangers become fully apparent. The script writers, for their part, are mainly concerned with producing dialogue that corresponds to the needs of sync. Thus it is not at all surprising that the attention at this stage is focussed on the individual sentence or the individual take, which usually does not comprise more than one or two sentences. As a result, the textual level of translation, which is the only level at which equivalence can sensibly be achieved, is not really considered in the dubbing process at any stage. If equivalence is seen at a sentence-for-sentence (or a take-fortake) level, it is fairly obvious that the lexical and grammatical choices in the target language are too limited and that hence words and constructions have to be used that match the conceptual or denotative meaning (Leech 1981) of the source language without, however, being appropriate to film dialogue as far as stylistic meaning is concerned. Similarly, shortcomings with respect to cohesion - i.e. inappropriate realizations of pro-forms, ellipsis etc. - can easily be accounted for in view of an approach in which the text never seems to be considered as a textual whole. It has to be said, however, that the quality of dubbed texts could be improved considerably by taking a pragmatic approach to dubbing (cf. Herbst 1987, 1994, 1995), in which the sense of a scene (especially with respect to its plot-carrying elements) and the naturalness and appropriateness of the translated dialogue - together with central requirements of sync - are rated quite highly within the equivalence criteria. Such an approach to dubbing would neither be more expensive nor more time-consuming. It is important to emphasize that lip and nucleus sync, although they may present enormous difficulties in individual instances in a film, cannot generally serve as an excuse for the textual shortcomings of much of the dubbing into German. It is also to be emphasized that there are dubbed productions which are very much in line with the

306

DUBBING AND THE DUBBED TEXT

principles of a pragmatic approach and achieve a very high standard of both sync and translation. At the same time, the German viewing public (and film reviewers) do not seem to be very critical of the quality of the dubbing. If any shortcomings of dubbed dialogue are noticed at all, they are probably seen as the inevitable outcome of the constraints of lip sync. Thus, in commercial terms, there is no need to change current practice. And this probably means that the characteristics outlined above will remain valid for a considerable proportion of films dubbed from English into German. Whether this means that these features serve to establish a particular text type in the sense that viewers would expect a dubbed film to contain these characteristics and be dispappointed if it contained natural film dialogue without anglicisms, unmotivated style shifts or loose cohesive ties is very much to be doubted, however. Notes 1 Cf. Luyken (1991: 111-128) 2 For a theory of translation for dubbing see Herbst (1994: 219-275). 3 Cf. Rabanaus (1982). For an analysis of the dubbing of Shakespeare's plays see Herbst (1994: 263-273). 4 The corpus from which the following conclusions are drawn and from which the examples are taken consists of some 50 hours of American and British television series (soap operas, detective series and satirical comedies). Cf. Herbst (1994: 4-6). 5 Cf. the model by Elam (1980: 39). 6 Cf. Brown and Yule (1983: 4-19) or Barnickel (1980: 84-151). 7 One factor that might be mentioned in this context is the recurrent use of the same dubbing actors, which is common practice in Germany. This, however, seems to be more disturbing to foreigners than to the German audience, who are used to this and who would not normal­ ly be familiar with the voices of the original actors. Interestingly, people tend to generally accept voices for particular characters. 8 This only applies to situations in which regional accents would normally be used and not to other texts such as classical drama, of course. For a more detailed discussion of the problem of standard language see Herbst (1991 and 1995). 9 Cf. Herbst (1994: 139). Note, however, that the 69% of past tense forms excludes one episode of a German Derrick-production with very formal language and thus an exception­ ally high number of preterite forms. 10 To a certain extent, the frequent use of the past tense is also a consequence of the use of the standard language discussed above. 11 Cf. also Gellerstam's (1996: 56) examples of translationese.

THOMAS HERBST

307

References Barnickel, Klaus-Dieter. 1980. Sprachliche Varianten des Englischen. Band 2: Register und Stile (Hueber Hochschulreihe 45/1). München: Hueber Verlag. de Beaugrande, Robert-Alain and Dressler, Wolfgang Ulrich. 1981. Einführung in die Textlinguistik (Konzepte der Sprach- und Litraturwissenschaft). Tübingen: Niemeyer. Brown, Gillian and Yule, George. 1983. Discourse Analysis (Cambridge Textbooks in Linguistics). Cambridge: Cambridge University Press. van Dijk, Teun A. 1977. Text and Context. Explorations in the semantics and pragmatics of discourse (Longman Linguistics Library). London/ New York: Longman. Elam, Keir. 1980. The Semiotics of Theatre and Drama. London/ New York: Methuen. Enkvist, Nils Erik. 1964. "On defining style. An essay in applied linguistics". In N.E. Enkvist, J. Spencer and M.J. Gregory (eds), Linguistics and Style. London: Oxford University Press, 1-56. Gellerstam Martin. 1996. "Translations as a source for cross-linguistic studies". In K. Aijmer, B. Altenberg and M. Johansson (eds), Languages in Contrast. Papers from a Synposium on Text-based Cross-linguistic Studies. Lund 4-5 March 1994. Lund: Lund University Press, 53-62. Gregory, Michael. 1967. "Aspects of variety differentiation". Journal of Linguistics 3: 177-198. Halliday, M.A.K. and Hasan, Ruqaiya. 1976. Cohesion in English (English Language Series). London: Longman. Herbst, Thomas. 1987. "A pragmatic translation approach to dubbing". ebu-review XXXVIII/6: 21-23. Herbst, Thomas. 1991. "Neutralisation als Kriterium zur Bestimmung von Standardsprache". ZAA 39 (3/4): 201-208. Herbst, Thomas. 1994. Linguistische Aspekte der Synchronisation von Fernsehserien. Phonetik, Textlinguistik, Übersetzungstheorie (Linguistische Arbeiten). Tübingen: Niemeyer. Herbst, Thomas. 1995. "People do not talk in sentences. Dubbing and the idiom principle". Translatio - Nouvelles de la FIT - FIT Newsletter. Nouvelle série XIV (1995) 3-4. Communication Audiovisuelle et Transferts Linguistiques. Audiovisual Communication and Language Transfer. International Forum Strasbourg, 251-271. Herbst, Thomas. 1996. "Why dubbing is impossible" (forthcoming). Leech, Geoffrey. 1981. Semantics. The study of meaning. Harmondsworth: Penguin. (2nd edition) Luyken, Georg-Michael (with Herbst, Thomas, Langham-Brown, Jo, Reid, Helen and Spinhof, Hermann). 1991. Overcoming Language Barriers in Television. Dubbing and subtitling for the European audience. Media Monograph 13, The European Institute for the Media. Quirk, Randolph, Greenbaum, Sidney, Leech, Geoffrey and Svartvik, Jan. 1985. A Comprehensive Grammar of the English Language. London: Longman. [CGEL]

308

DUBBING AND THE DUBBED TEXT

Rabanus, Gert. 1982. "Shakespeare in deutscher Fassung. Zur Synchronisation der Inszenierungen für das Fernsehen". Jahrbuch Deutsche Shakespeare-Gesellschaft West, 63-78. Söll, Ludwig and Hausmann, Franz Josef. 1985. Geschriebenes und gesprochenes Französisch. Berlin: Erich Schmidt Verlag. (3rd edition) Whitman-Linsen, Candice. 1992. Through the Dubbing Glass. The Synchronization of American Motion Pictures into German, French and Spanish. (European University Studies) Frankfurt: Peter Lang.

Quality Revisited: The Rendering of English Idioms in Danish Television Subtitles vs. Printed Translations Henrik Gottlieb University of Copenhagen In a collection of articles emphasizing the text type-specific nature of (any) translation, an initial presentation of the subject matter in question is a sine qua non. In this study, the basic distinction in operation is, naturally, the one between • discourse expressed through a static, solely verbal medium: the printed text, whether fiction or non-fiction, and • discourse expressed through a fleeting medium using a multitude of communicative channels simultaneously: the audiovisual text, ranging from last-minute interviews on TV to well-rehearsed dramas and feature films. So, as text types are concerned, the distinction has to do with the number of interacting communicative channels present: Books are monosemiotic, putting all their faith in the printed word,1 whereas films and TV programs are polysemiotic; several channels, i.e. picture + dialog + music & effects, contribute to the total communicative effect. Adding to this text type-defined distinction, which could be labeled semiotic complexity, is another distinction which arises in the moment texts are translated: physical status. While translations of monosemiotic originals appear as physically independent works - together with translated versions of live polysemiotic entities like drama and songs - audiovisual polysemiotic texts are dependent works; in retaining all non-verbal elements of the original, they are modifications of the original and, in a physical sense, include this. Thus they can be labeled inclusive translations, as opposed to exclusive translations. In one case, even the original verbal message is retained: subtitling is the extreme form of inclusive translation. In adding subtitles to a film or a TV program, the translator's work is more like frosting a wedding cake than baking a bread of his own.

310

IDIOMS IN SUBTITLES VS. PRINTED TRANSLATIONS

Subtitling in a historical perspective Subtitling as a means of film translation dates back to 1929, when the new sound films reached an international audience. The method of adding strips of translated dialog to moving pictures was invented as an inexpensive alternative to postsynchronization, or dubbing. In subtitling, the original dialog remains an integral part of the film, while in dubbing, it is replaced by more or less lipsynchronous dialog in the target language. When in the 1930s dubbing became the preferred mode of film translation in the world's big-market speech communities, subtitling proved to be the only viable technique in several minor European speech communities, such as Finland, Norway, Sweden, Denmark, the Netherlands, and Flemish-speaking Belgium. Small countries without a national language of their own - for example, Switzerland and Austria - tended to share their neighbors' pro-dubbing habits. This division between larger, dubbing-minded speech communities and smaller ones favoring subtitling is found even today, after more than four decades of television. Thus, while Italy, Germany, Spain, and Russia still prefer voice replacement techniques, the Nordic countries, Holland, Belgium, Portugal, Slovenia, Croatia, and Greece subtitle foreign-language cinema films and TV programs. Outside Europe, Francophone and Spanish-speaking countries tend to favor dubbing, while subtitling is making headway elsewhere. Straddling the fence, Britain has formed a tradition of dubbing foreign-language films and television programs aimed at a mass audience, whereas 'art movies' are shown with English subtitles. To a lesser extent, this tendency is found in the US and in France, where intellectual audiences tend to prefer subtitling to dubbing. Today, the balance between the two rivals of screen translation is slowly shifting toward subtitling, partly for economic reasons: the increasing production and exchange of films and TV programs, combined with the fact that the number of TV channels is outgrowing the number of TV households, calls for a translation method quicker and cheaper than dubbing. But to viewers in subtitling countries, the economic advantages of subtitling are secondary; retaining the authenticity of the original production is paramount. Once considered a nuisance or a necessary evil, subtitling is now becoming the preferred mode of language conversion for literate film and television audiences worldwide. However, subtitling does not by definition operate between languages. In countries like the US, Japan and Germany, the most common type of subtitling operates within one language, typically the dominant national language.

HENRIK

GOTTLIEB

311

Thus, the two main types of subtitling are: 1. Intralingual subtitling (in the original language): a) Subtitling of domestic programs for the deaf and hard of hearing. b) Subtitling of foreign-language programs for language learners. 2. Interlingual subtitling (from the original language): This type has a diagonal quality: in interlingual subtitling, the subtitler 'crosses over' from interpreting the spoken foreign-language dialog to presenting a written domestic-language translation on the screen. In the following, the term 'subtitling' refers to interlingual subtitling, only. Subtitling defined Subtitling can be defined as a (1) written, (2) additive, (3) synchronous type of translation of a (4) fleeting and (5) polysemiotic text type. 1. Being of a written - as opposed to spoken - nature, subtitling differs from all other types of audiovisual translation. 2. The label additive indicates that verbal material is added to the original, retaining the source-language discourse. 3. The label synchronous reflects the fact that the original work (with or without the original dialog) and the translation are presented simultaneously unlike 'simultaneous' interpreting, where the interpreter always lags behind. 4. The term fleeting refers to the fact that, in filmic media, all discourse is presented in a flowing manner, beyond the control of the receptor. 5. As we already know, the term polysemiotic states the fact that the targetlanguage rendering only covers one of several interacting channels of communication in the translated version.

312

IDIOMS IN SUBTITLES VS. PRINTED TRANSLATIONS

In the table below, six major types of translation are distinguished, using the five defining qualities of subtitling as parameters: Table 1 Subtitling

Form Written +

Role Additive +

Dubbing

-

-

+

+

+

Simultaneous interpreting

-

-

-

+

-

Translated drama

-

-

-

+

+

Comic book translation

+

-

+

+

+

Literary translation

+

-

Presentation Synchronous +

Reception Fleeting +

Composition Polysemiotic +

-

-

-

Constraints, strengths, and intra-textuality In many respects, the above-mentioned features characterizing subtitling must be seen as constraints limiting the linguistic freedom of the subtitler. And certainly, the subtitler of a film or television episode cannot take the same liberties as the ones enjoyed by the translator of monosemiotic works. Apart from having to face the time-and-space constraints of the audiovisual media, the subtitler never escapes the fact that there is always someone, in most cases at least a large minority of his viewers, who understands very well what is said by the people on the screen. So, with many language combinations, as for instance English-Danish, audience reception of a subtitled TV program - which is always based on the interaction subtitles-picture anyway - is strongly influenced by a further factor:feedback from the original dialog. In a number of cases, this tell-tale effect is felt by both subtitler and viewer as negative, for instance when satirical programs present puns referring, for example, to source-language specific homonyms or idioms, or to jokes presupposing a detailed knowledge of people, places and events in the source culture. However, in other situations the feedback from the original version is of a positive nature. As polysemiotic media - where the communicative burden is shared by several channels - film and television present the translator with many deictic solutions to potentially ambiguous utterances: the subtitler can

HENRIK GOTTLIEB

313

actually see or hear how a certain passage should be interpreted, and this privilege is passed on to the audience. And, of course, this interactive intra-textuality has to be taken into account by critics and scholars, as well. In this study, I will try to do this, but, alas, the medium of this article is strictly monosemiotic, blotting out sound altogether and limiting vision to written signs. Therefore, trying to compare TV translation and book translation through the medium of a book demands a certain power of abstraction, from the author as well as the reader. I hope both parties succeed.

Idioms at large Throughout the centuries, the concept of 'idiom' has had a range of meanings, the common denominator of which could be labeled "way of expression", covering phenomena as disparate as for instance 'music', 'artistic style', and 'language dialect'. In 1925, one of the first "collectors of curiosities" of English, Logan Pearsall Smith,2 used the term 'idiom' "in its narrower sense, meaning the idiosynchrasies of our language, and, above all, those phrases which are verbal anomalies, which transgress, that is to say, either the laws of grammar or the laws of logic." (Smith 1928: 168) Today, an idiom is often seen as "a fixed group of words with a special meaning that cannot be guessed from the combination of the actual words used." (Longman Dictionary of Contemporary English, 1981.) According to a similar definition, an idiom is "a phraseological unit whose meaning cannot be arrived at from the separate meanings of the constituents of the unit." (Balint 1969: 3.) Thus, idioms are normally considered 'fixed multi-word units with non-transparent meaning'. But how fixed is 'fixed', and how opaque are idioms? In the following, I will comment on these cornerstones of the modern idiom concept.

How fixed are idioms? In his monumental Manual of Lexicography, Ladislav Zgusta mentions that, in the case of multi-word lexical units, "it is the impossibility to substitute a part without a total change in the over-all meaning which seems to be the basic criterion." (Zgusta 1971: 145.) However, some elements in idiomatic multi-word units may very well be substituted, as they belong to sets of functional synonyms: A frequent idiom, with the following matrix, '(drive*/hammer*/etc) a(nother) nail in(to) ...'s coffin / the coffin of...'

314

IDIOMS IN SUBTITLES VS. PRINTED TRANSLATIONS

consists of up to six units.3 Only two of these are nuclear elements4 ('nail' and 'coffin') while one acts as an optional auxiliary (the verb), and two are toggles, alternating between a limited number of potential elements. The final unit is the joker, represented by the three dots in the matrix. In this position, any animate - or personified - subject can be inserted. While the joker is an open-class element, the auxiliary and the toggles are closed-class elements. In schematic form, our idiom looks like this: Status Auxiliary: Toggle: Nucleus 1: Toggle: Nucleus 2: Joker:

1. 2. 3. 4. 5. 6.

Usage example (B) knocked the last nail into the coffin of the Keynesian 130om

Function / Identification verb of motion, open to conjugation article, open to modification killer-metaphor preposition of inward motion death-metaphor subject to death

The authentic examples below show the versatility of our idiom: A. B. C. D. E. F.

" we drive another nail into our own coffin." "it knocked the last nail into the coffin of the Keynesian boom." " the nail in the coffin of the (second) international, " " the first nail in the coffin of doctrinaire Marxist thinking." "The final nail was hammered into the coffin of psychiatry " " another nail will be driven into the coffin of colonialism."

These examples were all found in the COBUILD corpora - of then 44.3 million words - when I investigated frequencies and variant forms of English idioms at Birmingham University in August 1991. In this case, by asking for KWIC concordances of sentences with the strings 'nail(s)' and 'coffin(s)', the computer came up with these 6 examples of idiomatic usage, plus one example with the original, literal meaning intact, all presented in contexts of one hundred words.5 In trying to answer the question of (lacking) idiom flexibility, our 'nail & coffin'-idiom seems to prove that the usual labeling of (entire) idioms as 'fixed units' is, to say the least, misleading. In our documented usage examples, we see flexibility at many levels: Adding to all the paradigmatic liberties allowed for in the idiom matrix, syntagmatic variation is also found in the corpus, in the form of sentence passivization (examples E & F). 6 Thus, instead of using the term 'fixed' when identifying idioms I would suggest the word 'marked' to indicate the recognizable, yet flexible, whole of the idiom. Only the nucleus or

HENRIK GOTTLIEB

315

nuclei of the idiom may be designated as 'fixed'; the idiom matrix is very productive, indeed.

How opaque are idioms? As illustrated by the 'nail & coffin' example, the idiom may be read as a normal, literal expression; only the nuclear elements act as metaphors. Very few idioms in English present syntactic anomalies, as found in 'needs must/ must needs' or 'dog eat dog'. A more substantial number of idioms comprise unical elements, i.e. verbal or morphological fossils previously in free distribution, now only used in isolated contexts (cf. Fleischer 1982: 42-53), such as 'spick and span' or 'up hill and down dale'. The majority of idioms simply consist of a number of well-known words, arranged in a normal way, but found in a seemingly abnormal context. These full-fledged idioms look "innocent", but exactly because a literal interpretation of them makes sense, they may present problems to the "naive" reader/listener, and indeed, to translators. Such problems of decoding are usually said to distinguish idioms from (non-metaphorical) collocations, while as far as encoding is concerned, neither idioms nor collocations (e.g. 'take a walk' or 'a single mother'), are constructed in a self-evident way. Indeed, according to modern definitions, an idiom is difficult - if not impossible - to decode correctly for someone who only knows the normal meanings of its constituent elements: "An idiom is a phrase which you cannot understand by putting together the meanings of the words in it." A sweeping and admirably brief definition, I must admit.7 But judging idioms totally opaque, or impossible to guess, seems to be hitting above the mark. Naturally, some idioms can be difficult to guess, but most idioms are shaped like our 'nail & coffin' model: as a metaphorical phrase with unmistakable connotations. Idioms are figurative, not abstract artefacts. The exact encoding may escape the reader/listener, but the textual and pragmatic context explains a lot, even in monosemiotic text types. It is through contexts that we learn the meanings of new words, and new meanings of old words and phrases. So, while understanding the meaning of certain idioms may not be all that problematic in real life, even simple "transparent" collocations may cause difficulties: Is a 'single mother', for example, a) the only mother among a larger group of people, b) just this one mother,

316

IDIOMS IN SUBTITLES VS. PRINTED TRANSLATIONS

c) a divorced woman without her children, or d) a divorced woman with her children? These examples show that there is no clear-cut borderline between collocations and idioms, thus no absolute difference in terms of 'possibility' vs. 'impossibility' of decoding. Rather, we find a gradual transition from total or partial transparency via established metaphorical status to (near-)obscurity. And not only is it a case of intersubjective relativity, it is even a case of individual subjectivity: "not all speakers make the same sense of all idioms." This statement by George Lakoff (Lakoff 1987: 451) is followed by another testimony of the relativity of idiom decoding: Just as there are considerable speaker-to-speaker differences in the details of rules of grammar, and very great differences in vocabulary, so there are differences in the images associated with idioms. For most of the imageable idioms I have studied, there seem to be between one and three prevalent associated images, though in some cases there may be between a half-dozen and a dozen. This is by no means an unseemly amount of lexical variation. In fact, since associated images are hardly ever consciously taught or consciously learned, it is remarkable that there is any uniformity at all." (ibid.) Returning to our 'corpus idioms', "naïve" speakers of English, confronted with a 'nail & coffin' expression in actual usage, might not know whether the phrase was coined by the speaker/author or already established as an English idiom. But they would surely understand the literal meaning and transfer this into the (social) context referred to in the examples (A) through (F).8

Translating idioms From a contrastive point of view, it would be tempting to define idioms, traditionally considered language-peculiar, as 'expressions which cannot be translated word by word'. But just by looking at a language pair as similar as English and Danish - two Germanic languages - this lay definition must be discarded: A number of expressions with a shared heritage, e.g. in Latin, can indeed be translated "naively". A representative of this group of idioms is 'look* for a needle in a haystack', which translates - morpheme-by-morpheme, even - into 'lede* efter en nål i en høstak'. Furthermore, quite a few Danish idioms are calques, based on English models, e.g. 'ikke min kop te' (not my cup of tea) or the lazy journalese newcomer 'skeletter i skabet' (skeletons in the closet/ cupboard).9

HENRIK GOTTLIEB

317

Apart from having different etymologies, these three examples illustrate how idiomatic expressions may act as verbal, adjectival, or nominal phrases, respectively. Other idioms obtain an adverbial function, e.g. 'like a snowball in hell' or act as expressive markers - often elliptical - as in "No way!" But idioms may also act as full sentences, e.g. "Kysten er klar". This statement is sometimes humorously phrased in English: "The coast is clear!" by Danes who are convinced that no native English speaker would say so. But as we all know, they may! This is an example of a surprisingly frequent interlingual phenomenon : false false friends.

A great divide: three modes of translation As languages are not equivalent, a fundamental decision must be made by every translator - and every translation scholar - very early in the process: "Do I consider the surface structure & elements of the original relevant for the wording of the translation?" In case of a 'yes', you want to "bring the reader to the text". That means that you will have to transplant all culture-specific references and languagespecific elements - idioms, for instance - found in the original. This formal loyalty may easily result in an outlandish, awkward translation. On the other hand, its provenance is clear: The translator has no illusions of adapting the original text to domestic taste; the formal translation reads as a translation. In creating it, you don't decompose the text in order to recompose it, you simply transfer it. An advocate of this (presently) controversial strategy, Brian Mossop, Canadian Secretary-of-state translator, argues that when translating a minority speech into a majority language, "the whole point of such a translation is precisely to make readers aware of its 'outside' origin." (Mossop 1989: 14) However, if you say 'no' to the question in hand, you want to "bring the text to the reader". Following this approach, a certain amount of linguistic modification will be necessary, and - of special interest here - idioms should not be emulated, if at all rendered as idioms. In deciding not to stick to the letter of the original, there are two alternative strategies: Focusing on the meaning of the original, or on the intended effect on readers. In the first case, what you opt for is stylistic loyalty, recreating the verbal atmosphere of the original. In the second case, you go for conceptual loyalty, aiming at adapting the original text to a new communicative situation. Below, some basic distinctions are outlined:

318

IDIOMS IN SUBTITLES VS. PRINTED TRANSLATIONS

Mode: Ideal: Loyalty: Unit in focus: Beneficiary:

APPROACHES OF TRANSLATION TRANSFER10 TRANSMUTATION Authenticity Idiomaticity Formal Stylistic Words Text Author Message

ADAPTATION Viability Conceptual Effect Reader

Again, we are not dealing with absolutes, but - as always within the Humanities - with fuzzy-edged concepts and delimitations. However, the above taxonomy will serve as an overview, not of a world divided, but rather of a series of options available to the translator. As Mossop puts it: "As for the wording of the translation, there is no generally valid approach for all texts, and no one correct approach for the whole of any particular text. Within one short passage, it may be appropriate to use both a recent gallicism and an ancient idiom, to render one thought through a close approximation to each lexical item of the original and then let another be understood purely by implication, even though it is explicated at considerable length in the original. The translator must always be ready to switch from one mode of translation to another, and a proper doctrine of translation should allow for this." (Mossop 1989: 19)

Typologies and strategies In this discussion of strategies for idiom translation, we have encountered a conceptual paradox: Idiomaticity in translation sometimes means that idioms should not be rendered with idioms. The compulsory idiom-to-idiom strategy aiming at formal correspondence - is found in transfer, not in idiomatic transmutation. Sometimes, the search for a target-language idiom may be all in vain: "The fact that in English a certain concept has or has no idiomatic expression [...] by no means implies that anything similar is likely to be found in the target language in question." (Makkai 1972: 180) Or, if a target-language idiom does exist, the policy of using this in a translation - or in a bilingual dictionary - "for the sake of rather illusory style equivalence [...] sometimes results in glosses that are either obsolescent or just not semantically equivalent." (Healey 1968: 96) What are the options, then, for translating idioms? Operating within the framework of contrastive linguistics, Eckhard Roos, one of the first West Europeans to discuss this issue, sets up this typology (see Roos 1981):

HENRIK GOTTLIEB

319

1. SL idioms11 matched by TL idioms: a) Congruence (identical at word level) b) Equivalence (differing at word level) 2. SL idioms matched by other TL lexemes: c) Single-word matches d) Formula (i.e. non-metaphorical multi-word matches) e) Free form (i.e. encyclopedic renderings) In his Handbok i lexikografi, aimed at bilingual dictionary makers, Bo Svensén suggests the following four categories (see Svensén 1987): 1. 2. 3. 4.

Idioms with identical metaphors in SL and TL Idioms with related metaphors Idioms with different metaphors SL idioms with no metaphorical counterparts

Finally, in her coursebook on translation (1992), Mona Baker distinguishes between the strategies of a. b. c. d.

Using an idiom of similar meaning and form Using an idiom of similar meaning but dissimilar form Translation by paraphrase Translation by omission

Discussing this latter category, Baker touches upon the strategy of compensation: "Briefly, this means that one may either omit or play down a feature such as idiomaticity at the point where it occurs in the source text and introduce it elsewhere in the target text." (Baker 1992: 78) Synthesizing these three sets of strategies, the following typology can be set up: Strategy 1) Congruence 2) Equivalence 3) Correspondence 4) Reduction 5) Paraphrase 6) Expansion 7) Omission 8) Compensation

Process SL idiom > identical TL idiom SL idiom > similar TL idiom SL idiom > different TL idiom SL idiom > TL word SL idiom > TL phrase SL idiom > TL circumlocution SL idiom > 0 (void) SL non-idiom > TL idiom

These eight strategies fall into four distinct categories:

320

Adherence Literalization Deletion Idiomatization

IDIOMS IN SUBTITLES VS. PRINTED TRANSLATIONS

(1,2, and 3 : (4, 5, and 6 : (7 : (8 :

idioms rendered metaphorically) idioms rendered non-metaphorically) idioms being omitted) non-idioms rendered metaphorically)

The pilot study Now, having looked at how idioms may be translated, the logical thing to ask is: "How are they translated? How do professional translators deal with idioms, and how successful are they?" In the following, I will try to answer these questions by using a modified version of the typology established in the previous section. The object of study here is a small bilingual text corpus12, comprising 10 translations of contemporary books and films in English, and the 9 corresponding originals: 2 British novels and their translations 1 North American novel, in two different translations 2 British TV episodes, with Danish subtitles 2 American TV episodes, with Danish subtitles 1 American feature film, with Danish (TV) subtitles 1 American documentary, with Danish subtitles I use the rendering of idioms as a key to establishing a comparison between the registers of linguistic options available to the literary translator and to the television subtitler, respectively. Or, rather, the strategies used within these two dominant types13 of interlingual transmission. The question is: To what extent do media-specific features determine the strategies chosen in static texts (books) as opposed to dynamic texts (TV programs)? However, before looking at possible answers to this question, we need some clarification of objects and methods: Delimitations, considerations and hypotheses The material, idioms in translation, covers: 1) Translations from English into Danish, i.e. from a dominant language partially understood by most members of the target culture. This direction of translation is chosen because of its implications for the minority language, and because it is much more common than minority-into-majority language transmission.

HENRIK GOTTLIEB

321

2) Idiomatic phrases with at least one nominal element, e.g. 'head over heels' and 'put one's foot down', thus excluding a) compound nouns where one or more elements have gained metaphorical meaning, such as 'head hunter' or 'jet lag'. b) simple phrasal verbs with metaphorical meaning, e.g. 'put down', 'put up with', etc. 3) All occurring idiom14 tokens in all text pairs, i.e. established idioms as well as spin-offs and neologisms, English as well as Danish, no matter their stylistic value. Novels and translations were read as separate texts, without immediate collationing, whereas subtitles were read on the screen while watching the film and listening to the original dialog. Only then were all idiom tokens paired with their SL or TL counterparts. No computer scanning or other electronic shortcuts were tried: In searching for all types and tokens, semantic judgments had to be made throughout, so the scholar-asreader/viewer approach seemed to be the only viable method. In this pilot study, I focus on: 4) Differences and similarities between literary translation and subtitling of TV fiction, with special emphasis on "similar" genres in the printed media and television, respectively. Concerning the latter media, a fair amount of Anglicisms could be expected to appear (in the subtitles) due to the audiovisual feedback effect15 and the structural similarities between English and Danish. Due to the time-and-space constraints of TV subtitling, deletive strategies must be expected to be more frequent in subtitling than in literary translation: Idioms may be seen as space-demanding "luxury items" dealt away with in the condensation of dialog. Following from points (1) and (4), dealing with the extra-textual and the intratextual communicative situation, respectively, 5) my evaluation of the translation data in the study is based on the ideal of idiomaticity, thus seeing transmutation as an adequate 'default mode' of translation within the text types and the direction of transmission studied. Method and terminology Returning to our structural typology of strategies for idiom translation, we will now expand this with a semantic dimension. We start by setting up a matrix involving the four structural relations between idioms in source and target texts. In this matrix, an idiom is represented by a plus (+), a non-idiom by a minus (-), and a "missing" expression by a void

(0):

322

IDIOMS IN SUBTITLES VS. PRINTED TRANSLATIONS RELATION

A) B) C) D)

SOURCE TEXT

TARGET TEXT

+ + + -

+

Adherence Literalization Deletion Idiomatization

0 +

Any idiom found in any of the 10 + 9 texts investigated will necessarily belong to one of these four structural categories. But this classification of the idiom inventory in the text corpus only takes us halfway. For judgments of semantic and stylistic quality, the matrix below is needed:

1) 2) 3)

EFFECT Correspondent Insufficient Defective

QUALITY OF RENDERING + 0

Of course, it takes some nerve to allocate these three "grades" to the idiom data in the study. Out of the cool shadow of structural objectivity, things may get hot and subjective. And rightly so: "Quality" is not easily quantifiable. No matter how many factors or dimensions translation scholars or critics insert between the phrasing of the texts analyzed and their individual judgments on quality, their own (hopefully fine-tuned) intuition will always be the final judge. 16 Combining the two matrices, representing the dimensions of "how" ("What relation?") and "how successfully" ("What effect?"), the following taxonomy of strategies17 can now be established: Relation Adherence

Original

Translation

+

+

Quality

0 Literalization

+

Deletion

+

+

0 0

+

0 Idiomatization

+

0

Strategy Transposition Emulation Falsification Paraphrase Reduction Lesion Elimination Lacuna Amputation Elaboration Complication Alienation

HENRIK GOTTLIEB

323

In my study, this typology of twelve potential strategies is developed to describe types and degrees of adequacy in the translation of idioms. However, such a typology may also serve for contrastive analysis of other phenomena, extra-linguistic as well as intra-linguistic, e.g. culture-specific elements. A reason for setting up this typology for the analysis of my 'pilot corpus' was to be able to quantify the corpus data, thus ensuring a method of inter-textual comparison. As mentioned, a central feature of the corpus is the juxtaposition of static and dynamic texts. Hence, before discussing corpus data in general, I will dwell on the idiom similarities and differences between two comparable pairs of texts in the corpus studied: 1. "The novel": an English detective novel & its Danish translation, and 2. "The film": the TV adaptation of the novel & the Danish, subtitled version.

"Last Bus to Woodstock": One title, four texts In this double couple, we find one general author, one main story, two media (books and television), two (script) writers, two translators, and two directions of transmission: horizontal (SL writing into TL writing) and diagonal (SL speak into TL writing).18 The novel The author of this (very) English detective story, Colin Dexter, was born in 1930 and had a rather late debut as a writer. His first novel, from 1975, titled "Last Bus to Woodstock", was to become the first of eleven Chief Inspector Morse Mysteries. The Danish translation of "Last Bus to Woodstock" appeared in 1991, at which time the Inspector Morse character was well established with the Danish TV audience. The TV film In 1988, BBC ran their first series of "Chief Inspector Morse" episodes. These detective films received top ratings among British viewers, and by 1993, more than 30 episodes had been made. The most recent TV episodes were not screen adaptations, but merely productions "based on the characters by Colin Dexter". The episode discussed here is a genuine Dexter adaptation, however, with Michael Wilcox credited as screenplay writer. (A newer, non-Dexter episode is also found in the pilot corpus.)

324

IDIOMS IN SUBTITLES VS. PRINTED TRANSLATIONS

All episodes have been broadcast in Denmark, with subtitles. The duration of each episode is 105 minutes, twice the length of the usual TV fiction episode, thus providing "Chief Inspector Morse" with the feature film duration necessary for developing the plot of these intricate television narratives. The idioms found in novel and TV film By comparing the translated novel with Dexter's original, it turns out that in the translation of idioms all strategies, except one, are represented. Juxtaposed, the two books muster 227 idiom units, i.e. cases where an idiom is found in one or both texts. The original text contains 176 idioms, while the translation reaches a figure of 165. But, as can be seen from the following table, this net difference in number covers far more than just a structural loss of 11 idioms in the Danish translation. As for the rendering of the film dialog, "only" 9 strategies are found in the material. With 53 idioms in the film dialog, and only 28 in the subtitles, the net loss is more dramatic, but again, the idiom traffic is quite complex. The overall structural-stylistic picture looks as follows: Table 2

THE NOVEL Translation vs. original

1. Correspondent renderings Transposition 84 cases Paraphrase 46 cases Elimination 1 case Elaboration 42 cases 173 cases 2. Insufficient renderings Emulation 23 cases Reduction 10 cases Lacuna 0 cases Complication 4 cases 37 cases 3. Defective renderings Falsification 7 cases Lesion 3 cases Amputation 2 cases Alienation 5 cases 17 cases Total 227 cases

THE FILM Subtitles vs . dialog

37.0% 20.3% 0.4% 18.5% 76.2%

15 cases 25 cases 1 case 11 cases 52 cases

23.4% 39.1% 1.6% 17.2% 81.3%

10.1% 4.4% 0.0% 1.8% 16.3%

2 cases 3 cases 2 cases 0 cases 7 cases

3.1% 4.7% 3.1% 0.0% 10.9%

3.1% 1.3% 0.9% 2.2% 7.5% 100.0%

0 cases 1 case 4 cases 0 cases 5 cases 64 cases

0.0% 1.6% 6.3% 0.0% 7.9% 100.1%

HENRIK GOTTLIEB

325

Judging from these figures, the TV subtitles seem to render idioms more suc­ cessfully than does the printed translation. The percentage of defective render­ ings, however, is roughly the same. But, returning to the (four) correspondent strategies found in three quarters of the Novel data, and in four fifths of the Film data, some significant differences are easily detected: A) The ratio between transposition and paraphrase is reversed, going from the novel to the film. The literary translator - devoid of the time-and-space con­ straints of subtitling - has rendered more idioms as idioms. But this does not leave the book reader with a more idiomatic text than the one received by the TV viewer, because B) In this novel, the translator uses structural calques three times as often as the subtitler of the film: The strategy of emulation is the fourth most frequent in the novel, whereas in the subtitles it is only used marginally. This runs con­ trary to our hypothesis that the dialog, still fully audible, would influence the wording of the subtitles. It suggests (at least) one of three explanations: 1) The subtitler is more talented than the novel translator. 2) The novel translator aims at authenticity, not idiomaticity. 3) The constraints of TV forces the subtitler to recompose rather than simply transfer elements. In this specific case, I think factors (3) and (1) are at play. At all events, talented translators working in the electronic media get better because of the constraints present, not in spite of them. Creativity does not arise in a vacuum. C) Due to the media-specific constraints, the drastic strategies of lacuna and amputation are found ten times as often in (this) subtitling than in (the similar) printed translation. Summarizing these three differences: As idioms often take up more space and need more time to be read than 'literal' constructions, the subtitler has, as ex­ pected, reached the high 'correspondence score' at the cost of idioms. But a remarkable similarity must be noted: D) Contrary to expectations, in "Last Bus to Woodstock" the strategy of elabo­ ration is just as frequent on the TV side as in the literary department: a major part of the idiom inventory in the Danish subtitles have non-idiom counterparts in the film dialog. As illustrated below, compensatory strategies (idiomatizations) are more frequent than deletions:

326

IDIOMS IN SUBTITLES VS. PRINTED TRANSLATIONS

Table 3

THE NOVEL Translation vs. original 50.2% 114 cases 59 cases 26.0% 3 cases 1.3% 51 cases 22.3% 227 cases 99.8%

Adherence Literalization Deletion Idiomatization Total

THE FILM Subtitles vs. dialog 17 26.6% 29 45.3% 7 10.9% 11 17.2% 64 100.0%

Examples of strategies found in the novel translation and the TV subtitles In the following, the quality of my quality judgments can be judged, as well as my criteria for allocating idiom units to the different strategies. For each strategy found in the novel data and in the film data, respectively, I give one example. However, the category 'transposition' are granted two examples: one showing lexical similarity within the unit, the other illustrating lexical substitution. For those not familiar with Scandinavian languages, I have back-translated all selected passages "naively", but with a modicum of English syntax. I am aware, however, that in some cases these back-translated constructs may seem so queer that the reader might be taken aback by their "unlikely" phrasings. Yet, that is just how idiom-sensitive groups19 react to real idioms the first time they meet them. In order not to alienate the reader unnecessarily, though, I have refrained from "funny" back-translations of the elements in the Danish subtitles which are not part of the idiom. The printed sources are rendered verbatim. All subtitles are rendered as broadcast, as their line breaks is an integral part of their "visual phrasing". The film dialog is transcribed as heard, as I have had no access to the film script.20 "LAST BUS TO WOODSTOCK": THE NOVEL ENGLISH ORIGINAL (Colin Dexter 1975)

DANISH TRANSLATION (Gerd Have 1991)

1.

CORRESPONDENT RENDERINGS

It's high time I had a try.'

Transposition Det er på høje tid, jeg gør et forsøg."

, and I saw Palmer myself on Friday Friday night when I had it all out with him.

, og jeg mødte selv Palmer fredag aften og fik kortene pa bordet."

BACK-TRANSLATION

It is on high time I make a try. and I met Palmer myself Friday night and got the cards on the table.

327

HENRIK GOTTLIEB

'Do you usually drink

"Er det almindeligt,

Is it common

on your own?'

at De drikker alene?"

that you drink alone?

'I've often heard, sir, that when

"Jeg har tit nok hørt, at når det

I have often enough heard that when

a couple are hitching the girl

er et par, der tomler, lader de

it is a couple who hitch, they let

shows a leg, as it were, and the

pigen stå og vise ben,

the girl stand and show legs, while

man keeps out of the way.

mens fyren holder sig af vejen."

the guy keeps out of the way.

Why were both these people

Hvorfor var det dem begge livet

Why was it to them both the life

anxious to keep quiet?

om at gøre at holde mund?

about to do to shut their mouths?

Elimination

Elaboration

INSUFFICIENT RENDERINGS Emulation Morse reluctantly reined back

Det lykkedes med besvær Morse

Morse had trouble succeeding to

the w i l d horses.

at få styr over sine vilde heste.

gain steering over his wild horses.

'Yes. She wrote it.'

"Jo. Hun skrev det selv."

"Yes. She wrote it herself."

'The children have been waiting

"Børnene har stået og ventet

The children have stood and waited

no end of a time.'

meget længe."

very long.

Reduction

lacuna (No occurrences) Complication ; and I reckon it's odds-on

Og som udgangspunkt mener jeg,

And as a starting point I think that

she got out in the front

det nok vil have en del for sig at

it probably will have a lot for i t

nearside and into the back

gå ud fra, at hun stod ud foran

to go out from that she stood out

nearside and that he did the same

til venstre og ind bagi til venstre,

in the front leftside and into the

on his side.

og at han gjorde det samme på

back leftside, and that he did the

højre side.

same on the right side.

DEFECTIVE RENDERINGS Falsification For the next hour the two officers

De to politimænd sad i de næste

The two policemen sat for the next

exchanged notes on the after­

par timer og udvekslede noter

couple of hours and exchanged notes

noon's interrogation,

vedrørende eftermiddagens forhør,

on the afternoon's interrogation,

The confident head of the murder

Den selvtillidsfyldte leder af mord­

The confident leader of the

inquiry, if ever invited to take

undersøgelsen ville, hvis han nogen­

murder inquiry would, if he

his eight discs to a desert island

­­inde blev anmodet om at vælge,

ever was asked to choose what

would have answered 'Committees'

hvad han ville tage med t i l en øde

he wanted to take with him to

to the inevitable question about

øf have svaret 'udvalgsmøder' på

a desert island, have answered

what he would be most glad

det uundgåelige spørgsmål om,

'committee meetings' to the

to have got away from.

hvad han helst ville slippe væk fra.

inevitable question about what

Lesion

he most wanted to get away from.

328

IDIOMS IN SUBTITLES VS. PRINTED TRANSLATIONS

Amputation For standing no more than four

For dér, i en afstand af ikke over

For there, at a distance of

feet away from him, ostensibly

en meter og femogtyve stod den

no more than one metre and twenty-five stood that detective

checking the times of the next

kriminalassistent, der samarbejdede

programme, but keeping himself

med kriminalkommissær Morse om

assistant who collaborated with

carefully and unobtrusively out

efterforskningen af mordet på

detective commissioner Morse on

of the limelight, was the sergeant

Sylvia Kaye.

the inquiry into the murder of Sylvia Kaye.

seconded to Detective Chief In­ spector Morse for the inquiry into the murder of Sylvia Kaye.

Alienation In the BBC she might have been

I BBC ville man måske have tildelt

In BBC they would maybe have

accorded the distinguished title

hende den fornemme stillingsbeteg-

accorded her the distinguished

of 'continuity girl'; but she was

tegnelse "script girl"; men hun var

title "script girl"; but she

in a dead-end job with the local

fanget i en blindgyde på lokal­

was caught in a dead-end alley

radio station.

radioen.

at the local radio.

"LAST BUS TO WOODSTOCK": THE TV FILM ENGLISH DIALOG

DANISH SUBTITLE

(Michael Wilcox 1988}

(Jane Bierbum 1991)

BACK-TRANSLATION

CORRESPONDENT RENDERINGS Transposition Coded messages, murder...

Kodebreve og mord ...

Code-letters and murder...

Right up my street It's not

Det er lige mig! Dagen starter godt.

That's just mel The day starts well.

a bad way to start the day. Looks like we're

Nu er vi i samme båd.

Now we're in the same boat

in the same boat now. Paraphrase • If it's the wretched professional chair, it's in the bag.

• Dit professorat er jo hjemme. - Er det?

• Your professorship is indeed home. - Is it?

- Is it? Elimination You must know something about

Du må vide noget om brevet.

You must know something about

the letter. Tell me what the hell

Hvad fanden er det?

the letter. What the devil is it?

is going on, for God's sake. Elaboration • Max? What are you doing here?

- Max... hvad laver du her?

• Max ... what do you do here?

-I could ask the same of you,

• I lige måde, Morse.

• In equal fashion, Morse.

Morse.

HENRIK GOTTTLIEB

2.

329

INSUFFICIENT RENDERINGS Emulation

Hang on ...

Vent... Nyt fra min front

Wait... News from my front

More news from the home front. Reduction ■ Anyway, I've bent over backwards to keep you out of this.

Men jeg har gjort ait

But I've done everything

for at holde dig udenfor.

to keep you out.

- Thanks. Lacuna ■ We should know one way or another, in a week or two.

Vi får besked senest om 14 da ■ Vil du ikke godt skifte emne?

■ We get a message at the latest in 14 days. ■ Won't you please change

■ Oh, talk about something else

the subject?

for God's sake, will you? Complication (no occurrences)

DEFECTIVE RENDERINGS Falsification (no occurrences) Lesion You'll do well out of me, dammit!

Du tjener på det

You make money on i t

That's what you wanted, wasn't it?

Var det ikke det, du ville?

Wasn't that what you wanted?

- Cunning old buzzard ...

Den snu gamle grib.

Amputation That cunning old buzzard.

- What's up, sir? Alienation (no occurrences)

I hope the examples speak for themselves, so one brief comment will suffice: The fact that the two lists of "idioms in context" are quite different is not intentional: In Last Bus to Woodstock, I only found one idiom unit common to both novel and film. The differences between the book and the TV adaptation go beyond the media-determined alterations of the novel, which leave only (part of) the dialog within the realm of verbal language. In the TV film, the plot has been modified considerably, yielding stretches of dialog not found in the novel. Not only as a subtitler21 have I experienced the whip of time-and-space constraints; also as a writer of articles one must accept certain limits. Therefore, I trust readers to manage without text examples in the presentation of the remaining data of this pilot study.

330

IDIOMS IN SUBTITLES VS. PRINTED TRANSLATIONS

Table 4

RENDERING OF IDIOMS IN TRANSLATED NOVELS AND TV PROGRAMS STRUCTURAL AND STYLISTIC TYPOLOGY: STRATEGIES USED FILMS

BOOKS 1) Writer 2) Translator 3) Title 4) ST/TT 5) Size, ST 6) Idiom frequency 7) Idiom units

Gibson Herløv

Dexter Have

Wilcox Cullen Bierbum Bierbum

Towne Dannow

Lynch Voxpop HPM Eskelund

Star Tonboe

Fischer Neuroma Neuroma 1980/80 1984/89 1984/92

Woodst 1975/91

Woodst Serpent 1988/91 1990/91

ChtownTwPeaks Deceive 1974/90 1989/90 1991/92

Beverly 1990/92

Greene Vesterholt

Gibson HPM

132 p. 0.4 i/p 78

173 p. 0.4 i/p 115

173 p. 0.4 i/p 103

200 p. 0.9 i/p 227

8) Correspondence 9) Transposition 10) Paraphrase 11) Elimination 12) Elaboration

6887.1 90*78.2 80*77.7 173*76.2 23*29.4 22*19.1 21*20.4 84*37.0 20*25.6 35*30.4 36*35.0 46*20.3 1* 0.4 25*32.1 33*28.7 23*22.3 42*18.5

13) Insufficience 14) Emulation 15) Reduction

10*12.9 17*14.8 18*17.5 37*16.3 8*10.3 1* 0.9 4* 3.9 23*10.1 - 11*9.6 12*11.7 10*4.4

16) Lacuna 17) Complication

2* 2.6

18) 19) 20) 21) 22)

Defection Falsification Lesion Amputation Alienation

0* 0.0 8* 7.0 - 1* 0.9 • 5* 4.3 - 1 0.9 -1*0.9

23) 24) 25) 26)

Adherence Literalization Deletion Idiomatization

27) 28) 29) 30)

Words, ST Words, TT Volume change Average change

31) 32) 33) 34)

Idioms, ST Idioms, TT Density change Average change

35) Idiom pairs 36) Aver, frequency

1* 0.9 4* 3.5

2* 1.9

31*39.7 24*20.2 26*25.2 114*50.2 20*25.6 51*44.7 52*50.5 59*26.0 0* 0.0 2* 1.8 0* 0.0 3* 1.3 27*34.7 38*33.3 25*24.3 51*22.5 78,000 87,000 +12%

51* 1.3 77* 1.5 78* 1.6 176* 2.3 58* 1.4 62* 1.2 51* 0.9 165* 1.9 + 11% -25% 40% -16% BOOKS: 18% 2.0 BOOKS: •

2.3 2.3

2.1 —

105 m. 0.6 i/m 64

125 m. 0.5 i/m 71

90 m. 0.5 i/m 51

57 m. 0.2 i/m 26

94 m. 0.6 i/m 61

52*81.3 48*75.0 56*78.9 35*68.6 19*73.1 50*82.0 15*23.4 12*18.8 21*29.6 12*23.5 5*19.2 8*13.1 25*39.1 31*48.4 22*31.0 13*25.5 3*11.6 35*57.4 1* 1.6 • 1* 2.0 11*17.2 5* 7.8 13*18.3 9*17.6 11*42.3 7*11.5 7*10.9 14*21.9 13*18.3 10*19.6 2* 3.1 3* 4.2 3* 4.7 7*10.9 7* 9.9 8*15.7 2* 3.1 7*10.9 3* 4.2 2* 3.9

5*19.2

7*11.5

5*19.2

6* 9.8 1* 1.6

5* 7.9

2* 3.1

2* 2.8

6*11.8

2* 7.7

4* 6.6

1.6 6.3

2* 3.1

2* 2.8

6*11.8

1* 3.8 1* 3.8

4*6.6

4* 1.8

5* 4.9 17* 7.5 1* 1.0 7* 3.1 4* 3.9 3* 1.3 - 2 * 0.9 5*2.2

39,200 50,000 50,000 40,100 53,700 54,200 +2% +7% +8% BOOKS:+7%

105 m. 0.5 i/m 64

2.9

17*26.6 12*18.8 24*33.8 12*23.5 5*19.2 8*13.1 29*45.3 38*59.4 29*40.8 21*41.2 8*30.8 41*67.2 7*10.9 9*14.1 5* 7.0 9*17.6 1 * 3.8 5* 8.2 11*17.2 5*7.8 13*18.3 9*17.6 12*46.1 7*11.5 c.9,500 c.9,500 c.7,000 c.7,000 -26% -26% FILMS:

5,300 4,725 -11%

9,850 7,000 -29%

53* 5.6 59* 6.2 58* 5.6 42* 4.7 14* 2.6 28*4.0 17*2.4 37*4.7 21*3.5 17*3.6 -28% -61% -16% -25% +36% FILMS: 26%

54* 5.5 15*2.1 -61%

6.7 FILMS: •

10,400 c.9,000 7,850 c.6,000 -25% -33% 25%

6.7

6.8 ..

5.7 6.2

4.9

6.2

HENRIK GOTTLIEB

331

On the previous page I have compressed the findings concerning all 10 text pairs, including the two "Chief Inspector Morse" versions discussed so far. Following the table is a list explaining the data line for line. For easy reference, all figures are supplied with percentages giving relative frequencies of idioms and strategies found in the material. Explanations to Table 4 1)

2) 3)

4)

5)

6) 7) 8-26)

27-30)

31-32) 33-34) 35-36)

Authors / screenplay writers are: Graham Greene, William Gibson, Colin Dexter, Michael Wilcox, Alma Cullen, Robert Towne, Mark Frost & David Lynch, people of Reno (Nevada), and Darren Star. Translators / subtitlers are: Ole Vesterholt, Hans Palle Mortensen, Arne Herløv Peter­ sen, Gerd Have, Jane Bierbum, Jesper Dannow, Jacob Eskelund, and Dorte Tonboe. Titles are: "Doctor Fischer of Geneva or The Bomb Party" / "Doktor Fischer i Ge­ nève; "Neuromancer" I Neuromantiker"; "Last Bus to Woodstock" / "Sidste bus til Woodstock"; "The Infernal Serpent" / "Den djævelske slange"; "Chinatown"; "Dream Deceivers" / "Sangen til døden"; "Beverly Hills, 90210". ST and TT are short for Source Text and Target Text, respectively. Years of book pub­ lications are self-explanatory. Dates for Danish TV broadcasts, necessary for identifi­ cation of individual programs, are: 'Woodst' September 9. 1991 (DR-TV); 'Serpent' November 1. 1991 (DR-TV); 'Chtown' September 1. 1990 (TV2); TwPeaks' (Epi­ sode 1) rerun January 2. 1991 (DR-TV); 'Deceive' March 16. 1992 (TV2); 'Beverly' (Episode 1) March 23. 1992 (TV2). Source text sizes are given in printed pages / TV program durations, in minutes. All works have been studied in their entire length, except 'Neuroma', where the three first sections have been analyzed. Frequencies refer to 'number of idioms per page / per minute'. Idiom units: Cases where idioms are found in one or both texts. Distribution of idiom units in the 3 times 4 taxonomy of strategies. Left figures in each column: number of idiom units. Right figures: percentages of total number. Fig­ ures in boldface are accumulated, so that line 8 accumulates lines 9-12, line 13 accu­ mulates lines 14-17, line 18 accumulates lines 19-22. Accordingly, line 23: 9+14+19, line 24: 10+15+20, line 25: 11+16+21, and line 26: 12+17+22. Figures have been computed via samples, due to the impossibility of doing electronic counts. The English definite article 'the' counts as a word - as opposed to the Danish definite nominal suffix '-en' or '-et'. On the other hand, contractions of the type 'she'll' or 'don't' - not found in Danish - are counted as one word, whenever they oc­ cur in spelling or pronunciation. These two trivial, but statistically important factors, form a pretty good balance, so that the registrered changes in text volume should give a valid picture of expansion / condensation. Left figures in each column: number of idiom tokens in each text. Right figures: idiom density (numbers of idioms per 1,000 running words). Idiom density changes in translations / subtitlings. Notice the plus/minus distribution. Numbers of idiom units per 1,000 running words. The dialog-only TV discourse shows an idiom density almost three times as high as the multi-purpose text in novels.

332

IDIOMS IN SUBTITLES VS. PRINTED TRANSLATIONS

Facts and interpretations: individual texts Looking at the ten text pairs contrasted in table 4, some interesting differences and similarities strike the eye: Similarities 1) Correspondence figures show only moderate variation: Books range from 76.2 to 87.1 percent; films from 68.6 to 82.0 percent. 2) Two of the books display similar profiles: Mortensen and Herløv are close in the strategies transposition, paraphrase, reduction, and lesion. And at the structural level, these two novel translations are atypical in that they show roughly the same adherence/literalization ratio as the films, i.e. they both tend to render idioms as non-idioms. Presumably, the idioms of their "shared" (cyberpunk) novel are difficult to render in Danish. 3) All subtitled fiction films "prefer" paraphrase to transposition, and - as mentioned above - all films favor literalization. 4) The non-correspondent strategies of complication and falsification are not found in any of the subtitled films. Differences 1) Insufficience figures vary for films; defection figures vary throughout the material. 2) As the only one, the Vesterholt translation is free of defective solutions. 3) The TV documentary ('Deceive') displays a unique profile: The strategy of elaboration comes in as number one, with paraphrase ranging third, unlike any other film. 4) The two "Chief Inspector Morse" films, with different screenplay writers, but the same subtitler, show quite different profiles: 'Woodst' uses significantly fewer paraphrases and fewer insufficient strategies than found in 'Serpent'.

HENRIK GOTTLIEB

333

Facts: books vs. films The central data in table 4 can be summarized this way: Table 5 STRATEGIES Correspondence Transposition Paraphrase Elimination Elaboration Insufficience Emulation Reduction Lacuna Complication Defection Falsification Lesion Amputation Alienation

BOOKS 79.8 26.5 27.8 0.1 25.4 15.4 6.3 6.4 0.2 2.5 4.9 1.3 2.4 0.5 0.8

FILMS 76.5 21.3 35.5 0.6 19.1 16.9 1.2 11.7 4.0 0.0 6.7 0.0 0.3 5.7 0.6

Adherence Literalization Deletion Idiomatization

33.8 36.7 0.8 28.7

22.5 47.5 10.3 19.8

RELATIVE DIFFERENCE -4% -20% + 28% + 500 % -25% + 10% -82% + 83% +1900 % - 100 % + 37% - 100 % -88% +1040 % -25% -33% + 29% + 1188% -31%

General conclusion In terms of quality, the rendering of idioms in films and books show the same statistical pattern: In both media, more than three quarters of the solutions are correspondent, with renderings of mediocre and poor quality covering approximately 16 and 6 percent of the material, respectively. The relatively high quality of the strategies found in subtitling is achieved via time-and-space saving literalizations, notably paraphrase and reduction. Fewer idioms are rendered as idioms in subtitles, but on the other hand, the non-idiomatic strategy of emulation is found five times as often in the novel translations. No surprise, deletive strategies are found 11 times more often in the subtitles than in the printed translations, but unexpectedly, one out of five idioms

334

IDIOMS IN SUBTITLES VS. PRINTED TRANSLATIONS

found in the Danish subtitles has no idiom counterpart in the English dialog 22 . Thus, whenever time, space and Danish grammar permits it, subtitlers seem willing to compensate for the indisputable loss of idioms in the diagonal process from spoken English to written Danish. Judging the subtitles more or less idiomatic than the English film dialog, however, is beyond the scope of the present study.

Idioms: Coming in from the cold In this study, I have focused on idioms as the tip of the phraseological iceberg and "abused" them to shed light on translation strategies. Although I have tried to soften the traditional view of idioms as rigid and obscure expressions, I have remained faithful to the fundamental notion that idioms are exceptions rather than extreme examples of general phenomena. However, modern text corpus research shows that we have to abandon the image of idioms as the "freaks of language". Rather than being bizarre creatures in the circus of linguistics, placed at a safe distance from normal grammar and lexis, idioms - as the ultimate multi-word units - are reminders that there are no such things as isolated words. Context is all there is, and meaning is always context-derived: "Most everyday words do not have an independent meaning, or meanings, but are components of a rich repertoire of multi-word patterns that make up text." (Sinclair 1991: 108). As in the natural sciences, the objects studied never cease to interact, and the units in focus are interdependent, not atomistic entities living in splendid isolation. John Sinclair, one of the prominent figures in the field of corpus linguistics, labels this interdependence in language 'the idiom principle', a concept which he presents in this way: "One of the main principles of the organization of language is that the choice of one word affects the choice of others in its vicinity." (Sinclair 1991: 173). So, far from "collecting curiosities" to keep up our belief in "normal" gram mar and lexis, future idiom research may, by studying idioms in their natural habitat, contribute to our understanding of language as an intricate web of recurring, flexible multi-word matrices, with an occasional single-word surprise here and there. Notes 1 In this context, we refer to unillustrated books, typically aimed at an adult audience. Books for children - like a number of non-fiction genres for all ages - are normally what could be labeled as bisemiotic: the expressive force of the verbal information is enhanced by pictures. 2 In his preface to "Words and Idioms", Pearsall Smith concludes: "... I have yielded most unreservedly to the temptation to make a collection of curiosities which should be as complete

HENRIK GOTTLIEB

335

as possible. If I may be accused of encouraging or inventing a new vice - the mania, or "idiomania", I may perhaps call it - of collecting what Pater called the "gypsy phrases" of our language, I have at least been punished by becoming one of its most cureless and incorrigible victims." 3 In this case, I consider the 'genitive + coffin' elements as one unit. A normal word count of this idiom would reach 7 or 8, and in example B below, the idiom comprises no less than 11 words. The asterisk (*) indicates that the element in question is open to grammatical transformation. 4 My categorization of idiom elements is based on Gottlieb 1992a. 5 For further discussion of actual idiom types and frequencies, see Gottlieb 1994a. An exhaustive description of the compilation of the COBUILD corpora is found in Sinclair (ed.) 1987. 6 A number of studies have revealed that idioms - and the psychological processing of them range on a cline from frozen expressions to loose matrices accepting a varying number of grammatical transformations. (See Fraser 1970; Swinney & Cutler 1979; and Gibbs & Gonzales 1985.) 7 Found in the introduction to A Learner's Dictionary of English Idioms (McCaig & Manser 1986: iii) 8 Ironically, in some of these examples, readers may need to look up the non-metaphorical joker, instead: What is a 'Keynesian boom', for instance? 9 The established expression 'lig i lasten' (corpses in the cargo) is becoming less frequent. The main reason is probably that Danish journalists use so many English-language sources that their Danish vocabulary and idiom inventory are affected. 10 As for the terminology of this tripartite division, the Collins COBUILD English Language Dictionary (1987) states: "If you transfer something, you move it to a different place or position." ('transfer', definition 1). Also, "When information is transferred, it is copied from one type of storage to another." ('transfer', definition 3). As for the second term, "If something is transmuted into a different form, it is changed into that form; a formal word." Finally, "Adaptation is the changing of something so that it becomes suitable for a new purpose or situation." ('adaptation', definition 2). 11 In the following, SL stands for Source Language; TL means Target Language. 12 A larger, more balanced corpus, comprising ten best-selling books and their Danish translations together with two weeks of subtitled Danish television, was compiled in 1994. This corpus is discussed in Gottlieb 1996 b. 13 In the minor speech communities of Western Europe, subtitling has a larger readership than literary translation, and people spend more time reading subtitles than books, see Gottlieb 1992b, 1994b, 1994c:137-157 and 1996 a. 14 In the following, I will use the term 'idiom' instead of 'idiomatic phrase' as defined above. Hence, the 'idioms' treated in this study all conform to the above-mentioned definition. 15 The impact from the soundtrack and the image may lead the television subtitler to come up with a formally correspondent rendering, modeled on the SL dialog, rather than a more idiomatic solution, cf. Gottlieb 1994b. 16 An elaborate scheme for translation quality assessment was set up by Juliane House (House 1977/1981): Via analyzing the source text according to three language user and five language use dimensions, and establishing syntactic, lexical, and textual correlates, a textual profile is established. The same procedure is then implemented on the target text. The final verdict is reached this way: "The degree to which TT's profile and function match or do not match ST's, is the degree to which TT is more or less adequate in quality." (ibid: 245). In her preface to the second edition, however, the author admits that "the model's unwieldiness as a working tool

336

IDIOMS IN SUBTITLES VS. PRINTED TRANSLATIONS

for the practising translator and translation evaluator and its relatively technical complexity for use in teaching have been mooted ...". 17 The term 'strategy' rarely implies a conscious decision in the mind of the translator. Most often, idioms - as other elements of language - are rendered without translators worrying about the structural status or the pragmatic effects of the choices they have made; they may "sound" right, they may be makeshift solutions, or considered the lesser of evils. 18 The categories horizontal, diagonal, and vertical translation (SL speak into SL writing) are discussed in Gottlieb 1994b. 19 The three major such groups - the core clientele of idiom dictionaries - are (native-speaking) children, the prelingually deaf, and (adult) learners of the foreign language in question. 20 However, as scripts are often lacking in reliability when it comes to auditively difficult passages, no script is better than an imprecise rendering of what was actually said. 21 In the years 1981-1992 I worked as a professional freelance subtitler, with DR-TV and Danish TV2. 22 One of the few major empirical translation studies (Kjär 1988) found that, surprisingly, no less than c. 15% of the verb metaphors in Swedish novel translations from German had a nonmetaphorical counterpart in the original (Kjär 1992: 200). Of the original metaphors, 67 percent were rendered as metaphors, 30 percent were "neutralized", and 3 percent were deleted (Kjär 1988: 120-121). References Baker, Mona. 1992. "Idioms and other fixed expressions". In In Other Words: A Coursebook on Translation. London: Routledge, 63-81. Balint, Andras. 1969. "Sector analysis and idioms". Kivung (Journal of the Linguistic Society of the University of Papua New Guinea) 2(1): 2-12. Collins COBUILD English Language Dictionary. 1987. John Sinclair (Editor in Chief). London and Glasgow: Collins. Fleischer, Wolfgang. 1982. Phraseologie der deutschen Gegenwartssprache. VEB Leipzig. Fraser, Bruce. 1970. "Idioms within a transformational grammar". Foundations of Language 6(1): 22-42. Gibbs, R.W. and Gonzales, G.P. 1985. "Syntactic frozenness in processing and remembering idioms". Cognition 20: 243-259. Gottlieb, Henrik. 1992a. "Idioms into Danish. Principles and design of a bilingual dictionary of current idiomatic English usage: The English-Danish idiomatic dictionary". In J.E. Nielsen (ed) Words that Teem with Meaning. Copenhagen Views on Lexicography. Copenhagen: Museum Tusculanum Press, 56-81. Gottlieb, Henrik. 1992b. "Subtitling - A new university discipline". In Cay Dollerup and Anne Loddegaard (eds), Teaching Translation and Interpreting. Amsterdam/ Philadelphia: John Benjamins Publishing Company, 161-170. Gottlieb, Henrik. 1994a. "Idioms in corpora: Types, tokens, frequencies, and lexicographical implications". In K. Hyldgaard-Jensen and V. Hjørnager Pedersen (eds), Symposium on Lexicography VI. Proceedings of the Sixth International Symposium on Lexicography May 7-9, 1992, at the University of Copenhagen (Lexicographica. Series Maior, Band 57). Tübingen: Max Niemeyer Verlag, 85-91.

HENRIK GOTTLIEB

337

Gottlieb, Henrik. 1994b. "Subtitling: People translating people". In Cay Dollerup and Annette Lindegaard (eds) Teaching Translation and Interpreting 2. Amsterdam/ Philadelphia: John Benjamins Publishing Company, 261-274. Gottlieb, Henrik. 1994c. Tekstning - Synkron billedmedieoversættelse (Danske Afhandlinger om Oversættelse 5). University of Copenhagen, Center for Translation Studies. Gottlieb, Henrik. 1996a. "Tekstning, et polysemiotisk puslespil". In Finn Frandsen (ed) Medierne og sproget. Aalborg Universitetsforlag, 151-171. Gottlieb, Henrik. 1996b. "Idiomer: Oversatte eller oversete? Et punktnedslag i oversættelsen af litterære tekster". In Subtitles, Translation and Idioms. Ph.D. thesis, University of Copenhagen, Center for Translation Studies. Healey, Alan. 1968. "English idioms". Kivung (Journal of the Linguistic Society of the University of Papua New Guinea) 1(2): 71-108. House, Juliane. 1977/1981. A Model for Translation Quality Assessment. Tübingen: Gunter Narr Verlag. Kjär, Uwe. 1988: "Der Schrank seufzt". Metaphern im Bereich des Verbs und ihre Übersetzung. Göteborg: Acta Universitatis Gothoburgensis. Kjär, Uwe. 1992. "Die Übersetzung von Verbalmetaphern". Babel 37(4): 193-202. Lakoff, George. 1987. Women, Fire, and Dangerous Things: What Categories Reveal about the Mind. Chicago: University of Chicago Press. Longman Dictionary of Contemporary English. 1981. Paul Procter (Editor in Chief) and Robert I1son (Managing Editor), London: Longman. Makkai, Adam. 1972. Idiom Structure in English. The Hague & Paris: Mouton. McCaig, Isabel and Manser, Martin H. 1986. A Learner's Dictionary of English Idioms. Oxford: Oxford University Press. Mossop, Brian. 1989. "'Write idiomatically and translate ideas not words:' Three defects of the prevailing doctrine of translation". In Candace Séguinot (ed), The Translation Process. School of Translation, York University (Toronto): H.G. Publications, 7-20. Roos, Eckhard. 1981. "Contrastive analysis and the translation of idioms: Some remarks on contrasting idioms". In Wolfgang Kühlwein, Gisela Thome and Wolfram Wilss, (eds) Kontrastive Linguistik und Übersetzungswissenschaft. Akten des internationalen Kolloquiums Trier/Saarbrücken 25-30.9.1978. München: Fink, 230-238. Sinclair, John (ed). 1987. Looking up. An account of the COBUILD Project in lexical computing. London and Glasgow: Collins ELT. Sinclair, John. 1991. Corpus, Concordance, Collocation. Oxford: Oxford University Press. Smith, Logan Pearsall. 1928. "English idioms". In Words and Idioms. Studies in the English Language. London: Constable's Miscellany, 167-278. Svensén, Bo. 1987. Handbok i lexikografi - Principer och metoder i ordboks-arbetet. Stockholm: Esselte Studium. (English edition: Practical Lexicography. Principles and Methods of Dictionary Making. Oxford University Press 1993).

338

IDIOMS IN SUBTITLES VS. PRINTED TRANSLATIONS

Swinney, D.A. and Cutler, A. 1979. "The access and processing of idiomatic expressions". Journal of Verbal Learning and Verbal Behavior 18, 523-534. Zgusta, Ladislav. 1971. "Combinations of words". In Manual of Lexicography. Prague: Academia / The Hague, Mouton, 138-165.

339

Index abstracts 35 accessibility 257 adaptation 318 adaptations 35 additions 34 ambiguity 219 Andersen, Hans Christian 99 anglicisms 303 annotated translations 33 anthropological 96 anthropological translations 93 appellative function 50 argumentation 15, 176 artificial language 36 Atelsek 190 Atkinson 163 audio-medial 278 audiovisual text 309 back-translated 326 Baker, Mona 319 Bank of English 257 Bayerschmidt and Hollander 94, 95 behavioural knowledge 69 Benjamin, Walter 96 Berkenkotter and Huckin 149, 207 Bhatia 7, 209 Biber 12, 16 Bible 36, 39 bilingual and multilingual reference corpora 248 bilingual corpora 252

bilingual dictionary 248, 250 bilingual lexical database 267 bilingual lexical database system 258 bilingual lexicography 252, 267 bilingual text 248, 262 bilingual text corpus 250 bilingual the 252 Brown Corpus 251, 258 Bühler 278 Burton, Sir Richard 93, 94 Celt 172 Center for Translation Studies and Lexicography, Copenhagen University 103 change of content 38 change of intention 38 Chief Inspector Morse 323 Chinese 172 Clyne71 Cobuild 251,252 Cobuild corpora 314 codes 206 cognitive condition 302 cognitive structures 207 cohesion 220 cohesion in dubbed texts 299 collocability 219 collocation 223, 251, 266, 315 commentaries 35 commented translation 33 commissioner 28

340

TEXT TYPOLOGY AND TRANSLATION

communicative purpose 9, 205 communicative situations 31 community ownership 149, 150 comparable corpora 254 comparable corpus system 269 compound formation 228, 244 compounds 222, 223, 228 computational lexicography 247, 248 Computational Linguistics 247, 258 computational model of the lexical entry 249 concept 130, 133, 217 concepts in politics 130 conceptual loyalty 317 concordance 223, 252 constitutive rules 68 constraints 312 content 32 contextual focus 16 contrastive text analyses 247 contrastive text data 248 contrastive textologist 253 convention 68 copyright 257 corpus 222 corpus accessibility 256 corpus design 256 corpus linguistics 247, 248, 250 credibility 195 criticism 104 cross-language studies 256 Crystal 70 culture 68, 69, 161, 172, 191 culture-bound 133 culture-boundedness 121 Dal 103 Danish 162, 170, 178 Dasent, George Webbe 94, 95, 96 Davy 70 DBT 270

DBT system 263 de Chatelain, Clara 104 decoding 315 description 15, 104 diagonal translation 336 dictionaries 220, 241 directive 31, 76 directive intention 30 discourse 4 documentary translation 52 Dolet, Etienne 93, 97 Dollerup 150, 152, 155 Dr. Dulcken 101 Dressier 166 duality of structure 149 dubbed dialogue 292 dubbing 310 dynamism 149 EAGLES 258 easification 204, 210 easified version 211 electronic dictionaries 248 electronic lexicons 247 encoding 315 end-focus 279 end-weight 279 English 162 eponyms 174 equifunctional translation 53 equivalence 25, 43, 104, 217, 225, 226,236, 239 essay 31 Eurocrats 153 eurojargon 153 European Corpus Initiative 257 evaluative abstracts 35 Even-Zohar 89 exclusive translations 309 existing target language text types 39 exoticizing translation 53

341

INDEX

expanded translation 34 Expert Advisory Group on Language Engineering Standards 258 exposition 15 expressive function 50 Fairclough 207 false friends 317 feedback 312 footnotes 34 for information only 34 formal loyalty 317 French-English 165 function 49 functionalism + loyalty 45 fuzzy-edged concepts 318 Galtung 67, 68 generic knowledge 208 genre 6, 7, 8, 169 genre analysis 205 German 217 German-English 167 Gerzymisch-Arbogast 71 gist translations 33 Göpferich71,74, 186 Gutt 155 Hansard Corpus 258 Hatim and Mason 14, 16 hedging 74 Hersholt 103 heterofunctional 53 hierarchy of translation problems 61 Hoffmann 186 Holmes, James 89 Homer 88, 89 homologous translation 54 homonymy 162 Hoof 174 Horace 89

Horton 179 House 70 Howitt 104 human guided 255 hybrid texts 146, 147 hybrid text[type] 123 hyphenated compounds 235 hyphenated term 239 hyponym221 idiom 109,313 idiomaticity 105 idiom principle 334 idiom units 324 Ilie 164 illocutionary act 68 illocutionary indicator 80 illocutionary structure 190 impromptu translation 34 IMRAD 177 inclusive translations 309 index 224 indicative abstracts 35 information mediators 29 informative 31 instruction 15, 34 instructional 31 instrumental translations 52 intention 31 intercultural translation problems 58, 59 interference 194 interlingual subtitling 311 interlingual translation problems 58, 60 Internet 257 intralingual subtitling 311 Italian Machine Dictionary 258 jargon 163 job specifications 26

342

TEXT TYPOLOGY AND TRANSLATION

Karker 151, 152 Kinneavy 12, 13 knowledge content 31 Koller 185 Kourilovâ 176 Kussmaul 73 Lackstrom, Selinker and Trimble 5 Ladislav Zgusta 313 Lakoff, George 316 language family 172 language learner 252, 268, 269 language learning activities 252 language reference corpora 247, 250, 262 Lankamp 171 leaflet 34 lexeme (or lemma) 220, 224, 236, 244 lexical barrier 170 lexical database 249, 258 lexical database system 258 lexical entry 262 linguistic form 32 lip sync 292 literal translation 52 LIZ 257 loan words 173 Logan Pearsall Smith 313 loyalty 45,48 LSP 185 LSP translation 163 machine-aided human translation 247 machine guided 255 machine translation 26, 36, 247, 252 macrostructures 71 Magnusson and Pahlsson 94,95, 96 marketing 195 Martin 8 meaning 175

medical English 160 medical language 159 medical terminology 162 medical translation 160 medium 70 medium-dependent style value 298 memos 31 messages - documents - texts 27 metacommunicative utterances 72 microstructures 72 MLDB 258 modality 177 mode of discourse 15 mode of expression 31 mode of reception 30 modification of content 33 mono- and bilingual lexical databases 248 mono- and bilingual lexical database system 248 monolingual corpora 251 monolingual dictionary 248, 252 monolingual LDBs 261 monosemiotic 309 Morris, William 93, 94, 96, 97 Mossop, Brian 317 MT 253, 255, 256 Multext 254, 257 multilingual 258 multilingual reference corpus 248, 258 multi-medial 278 multiword compounds 228, 239 multiword compound terms 236 multiword lexical units 313 multiword term 239 Mutual Information Index 269 narration 15 natural language 36

343

INDEX

natural language processing (NLP) 247, 250 Newmark 161 Nietzsche 87, 88, 97 Norse sagas 90, 93 nucleus sync 292 number 222, 223, 228, 244

professional translators 320 prose translations of poetry 34 province 70 pseudo-idiom 110 purpose 32

onomasiological 217

reader expectation 27, 31 reader's assumptions 28 Rebholz 92 reduction or addition of content 33 reference corpora 250 referential function 50 register 4, 5, 159 register mismatch 171 regulative rules 68 Reiss 69, 187,278 religious text 36 reports 31 research article 163 research paper 169 rhetoric 277, 279 rhetorical consciousness 212 Rider, Henry 89 Rieu, E. V. 89, 90 Roos, Eckhard 318 rules, regulations or instructions 34

parallel concordances 264, 265, 266 parallel corpora 254 parallel corpus system 262 parallel texts 254 participation 70 part-of-speech taggers 252 passive 166 Peachey 101, 104 Petrarch 90, 91, 92 Petrarch's sonnet 92 phatic function 51 philological translation 53 PiSystem 269 politeness 81 political concepts 130 political discourse 119, 120, 131 political key concepts 130 political keywords 130 politically relevant keywords 133 political text 199, 128, 131, 132, 138 polysemiotic 309 postmodification 286 postsynchronization 310 pragmatic communicative situation 34 pragmatic equivalence 175 pragmatics 107 pragmatic translation problems 58 primary readers 28 primary readerships 28 printed text 309

quality 322

sagas 94,95,96 Salager 159 Sapir 168 schedule 31 Schröder 74 scientific paper 34 Searle 68,188 secondary and primary readerships 28 secondary readers 28 semasiological 217, 218 semiotic complexity 309 sense disambiguation 255

344

TEXT TYPOLOGY AND TRANSLATION

simplified version 212 Sinclair, John 334 single-word compounds 228, 233 single-word terms 226, 239 situatedness 149, 150 situation 70 situational dimensions 70, 71 skopos theory 46 slanted abstracts 35 Snell-Hornby 74 social attitude 70 social role relationship 70 sociocognitive theory of genre 149 socioculture 191 sonnet 90, 91, 92, 93, 96, 97 source text (ST) 120, 134, 135 Spanish-English 167 speakability 279 speech-act 72, 188 speech-act theory 13, 68 Spiller, Michael 90, 91 Spillner 278 ST 120, 121, 124, 126, 128, 129, 132, 133, 134, 135, 136, 137, 138 standardisation 151, 241, 244 status 70, 81, 90, 92, 95, 96, 97 Stolze 186 strategies 318 style 166 stylistic features of dubbed text 295 stylistic loyalty 317 subcorpora 255 sub-language 254 subtechnical terms 232, 234 subtitling 309 Svensén,Bo 319 Swales 7 synonyms 174 synonymy 217,219 syntactic parsers 252

syntagmatic relations 219 system 248, 262 system for monolingual corpus 248 System Quirk 223 target text (TT) 120, 135 technical brochures 185 technical language 186 tenor 159 terminology 269 terminology science 223 text 4, 36 text corpus 320 text creation 220 Textsorte 278 text-specific translation problems 58, 61 Texttyp 278 text-type 15, 25, 30, 68, 69 text-type conventions 279 text-type focus 16 text-typological considerations 293 text-typological conventions 120, 138 textual characteristics of films 304 textual database system 270 theme-rheme 71 the mode of production 38 tools 247, 250 topic 31 transcript 34 transfer 318 translation 217, 221,222, 233, 255 translation assignment 47 translation brief 56 translation problems 58 translation-specific text types 35, 39 translation strategy 32, 33, 58 translator 219, 220, 221, 241 translator as an expert 47 Translators' Workstation 248, 261, 263

INDEX

translingual abstracts 33, 35 translingual compte rendu, minutes and transcripts 35 transmission: diagonal 323 transmission: horizontal 323 transmutation 318 Trosborg 78, 185 TT 120, 121, 124, 126, 128, 129, 130, 132, 133, 135, 136, 137, 138 types of translations 32 typology of translations 37, 51 unical elements 315 union legalese 152 unnaturalness of dubbed text 294 untranslatable 35 variant forms 221 variation 219, 232, 236 Vermeer 69 vertical translation 336 Virtanen 4 voice 169 Voltaire 88 Webbe Dasent, George 94 Wellek and Warren 101 Whorf 168 word-for-word translation 52 workstation 269 writer's intention 27 writer's presuppositions 28 Wunderlich 68 Wyatt, Sir Thomas 91, 92

Benjamins Translation Library A complete list of titles in this series can be found on www.benjamins.com 74 Wolf, Michaela and Alexandra Fukari (eds.): Constructing a Sociology of Translation. 2007. vi, 226 pp. 73 Gouadec, Daniel: Translation as a Profession. 2007. xvi, 396 pp. 72 Gambier, Yves, Miriam Shlesinger and Radegundis Stolze (eds.): Doubts and Directions in Translation Studies. Selected contributions from the EST Congress, Lisbon 2004. 2007. xii, 362 pp. [EST Subseries 4] 71 St-Pierre, Paul and Prafulla C. Kar (eds.): In Translation – Reflections, Refractions, Transformations. 2007. xvi, 313 pp. 70 Wadensjö, Cecilia, Birgitta Englund Dimitrova and Anna-Lena Nilsson (eds.): The Critical Link 4. Professionalisation of interpreting in the community. Selected papers from the 4th International Conference on Interpreting in Legal, Health and Social Service Settings, Stockholm, Sweden, 20-23 May 2004. 2007. x, 314 pp. 69 Delabastita, Dirk, Lieven D’hulst and Reine Meylaerts (eds.): Functional Approaches to Culture and Translation. Selected papers by José Lambert. 2006. xxviii, 226 pp. 68 Duarte, João Ferreira, Alexandra Assis Rosa and Teresa Seruya (eds.): Translation Studies at the Interface of Disciplines. 2006. vi, 207 pp. 67 Pym, Anthony, Miriam Shlesinger and Zuzana Jettmarová (eds.): Sociocultural Aspects of Translating and Interpreting. 2006. viii, 255 pp. 66 Snell-Hornby, Mary: The Turns of Translation Studies. New paradigms or shifting viewpoints? 2006. xi, 205 pp. 65 Doherty, Monika: Structural Propensities. Translating nominal word groups from English into German. 2006. xxii, 196 pp. 64 Englund Dimitrova, Birgitta: Expertise and Explicitation in the Translation Process. 2005. xx, 295 pp. 63 Janzen, Terry (ed.): Topics in Signed Language Interpreting. Theory and practice. 2005. xii, 362 pp. 62 Pokorn, Nike K.: Challenging the Traditional Axioms. Translation into a non-mother tongue. 2005. xii, 166 pp. [EST Subseries 3] 61 Hung, Eva (ed.): Translation and Cultural Change. Studies in history, norms and image-projection. 2005. xvi, 195 pp. 60 Tennent, Martha (ed.): Training for the New Millennium. Pedagogies for translation and interpreting. 2005. xxvi, 276 pp. 59 Malmkjær, Kirsten (ed.): Translation in Undergraduate Degree Programmes. 2004. vi, 202 pp. 58 Branchadell, Albert and Lovell Margaret West (eds.): Less Translated Languages. 2005. viii, 416 pp. 57 Chernov, Ghelly V.: Inference and Anticipation in Simultaneous Interpreting. A probability-prediction model. Edited with a critical foreword by Robin Setton and Adelina Hild. 2004. xxx, 268 pp. [EST Subseries 2] 56 Orero, Pilar (ed.): Topics in Audiovisual Translation. 2004. xiv, 227 pp. 55 Angelelli, Claudia V.: Revisiting the Interpreter’s Role. A study of conference, court, and medical interpreters in Canada, Mexico, and the United States. 2004. xvi, 127 pp. 54 González Davies, Maria: Multiple Voices in the Translation Classroom. Activities, tasks and projects. 2004. x, 262 pp. 53 Diriker, Ebru: De-/Re-Contextualizing Conference Interpreting. Interpreters in the Ivory Tower? 2004. x, 223 pp. 52 Hale, Sandra: The Discourse of Court Interpreting. Discourse practices of the law, the witness and the interpreter. 2004. xviii, 267 pp. 51 Chan, Leo Tak-hung: Twentieth-Century Chinese Translation Theory. Modes, issues and debates. 2004. xvi, 277 pp. 50 Hansen, Gyde, Kirsten Malmkjær and Daniel Gile (eds.): Claims, Changes and Challenges in Translation Studies. Selected contributions from the EST Congress, Copenhagen 2001. 2004. xiv, 320 pp. [EST Subseries 1] 49 Pym, Anthony: The Moving Text. Localization, translation, and distribution. 2004. xviii, 223 pp. 48 Mauranen, Anna and Pekka Kujamäki (eds.): Translation Universals. Do they exist? 2004. vi, 224 pp.

47 Sawyer, David B.: Fundamental Aspects of Interpreter Education. Curriculum and Assessment. 2004. xviii, 312 pp. 46 Brunette, Louise, Georges Bastin, Isabelle Hemlin and Heather Clarke (eds.): The Critical Link 3. Interpreters in the Community. Selected papers from the Third International Conference on Interpreting in Legal, Health and Social Service Settings, Montréal, Quebec, Canada 22–26 May 2001. 2003. xii, 359 pp. 45 Alves, Fabio (ed.): Triangulating Translation. Perspectives in process oriented research. 2003. x, 165 pp. 44 Singerman, Robert: Jewish Translation History. A bibliography of bibliographies and studies. With an introductory essay by Gideon Toury. 2002. xxxvi, 420 pp. 43 Garzone, Giuliana and Maurizio Viezzi (eds.): Interpreting in the 21st Century. Challenges and opportunities. 2002. x, 337 pp. 42 Hung, Eva (ed.): Teaching Translation and Interpreting 4. Building bridges. 2002. xii, 243 pp. 41 Nida, Eugene A.: Contexts in Translating. 2002. x, 127 pp. 40 Englund Dimitrova, Birgitta and Kenneth Hyltenstam (eds.): Language Processing and Simultaneous Interpreting. Interdisciplinary perspectives. 2000. xvi, 164 pp. 39 Chesterman, Andrew, Natividad Gallardo San Salvador and Yves Gambier (eds.): Translation in Context. Selected papers from the EST Congress, Granada 1998. 2000. x, 393 pp. 38 Schäffner, Christina and Beverly Adab (eds.): Developing Translation Competence. 2000. xvi, 244 pp. 37 Tirkkonen-Condit, Sonja and Riitta Jääskeläinen (eds.): Tapping and Mapping the Processes of Translation and Interpreting. Outlooks on empirical research. 2000. x, 176 pp. 36 Schmid, Monika S.: Translating the Elusive. Marked word order and subjectivity in English-German translation. 1999. xii, 174 pp. 35 Somers, Harold (ed.): Computers and Translation. A translator's guide. 2003. xvi, 351 pp. 34 Gambier, Yves and Henrik Gottlieb (eds.): (Multi) Media Translation. Concepts, practices, and research. 2001. xx, 300 pp. 33 Gile, Daniel, Helle V. Dam, Friedel Dubslaff, Bodil Martinsen and Anne Schjoldager (eds.): Getting Started in Interpreting Research. Methodological reflections, personal accounts and advice for beginners. 2001. xiv, 255 pp. 32 Beeby, Allison, Doris Ensinger and Marisa Presas (eds.): Investigating Translation. Selected papers from the 4th International Congress on Translation, Barcelona, 1998. 2000. xiv, 296 pp. 31 Roberts, Roda P., Silvana E. Carr, Diana Abraham and Aideen Dufour (eds.): The Critical Link 2: Interpreters in the Community. Selected papers from the Second International Conference on Interpreting in legal, health and social service settings, Vancouver, BC, Canada, 19–23 May 1998. 2000. vii, 316 pp. 30 Dollerup, Cay: Tales and Translation. The Grimm Tales from Pan-Germanic narratives to shared international fairytales. 1999. xiv, 384 pp. 29 Wilss, Wolfram: Translation and Interpreting in the 20th Century. Focus on German. 1999. xiii, 256 pp. 28 Setton, Robin: Simultaneous Interpretation. A cognitive-pragmatic analysis. 1999. xvi, 397 pp. 27 Beylard-Ozeroff, Ann, Jana Králová and Barbara Moser-Mercer (eds.): Translators' Strategies and Creativity. Selected Papers from the 9th International Conference on Translation and Interpreting, Prague, September 1995. In honor of Jiří Levý and Anton Popovič. 1998. xiv, 230 pp. 26 Trosborg, Anna (ed.): Text Typology and Translation. 1997. xvi, 342 pp. 25 Pollard, David E. (ed.): Translation and Creation. Readings of Western Literature in Early Modern China, 1840–1918. 1998. vi, 336 pp. 24 Orero, Pilar and Juan C. Sager (eds.): The Translator's Dialogue. Giovanni Pontiero. 1997. xiv, 252 pp. 23 Gambier, Yves, Daniel Gile and Christopher Taylor (eds.): Conference Interpreting: Current Trends in Research. Proceedings of the International Conference on Interpreting: What do we know and how? 1997. iv, 246 pp. 22 Chesterman, Andrew: Memes of Translation. The spread of ideas in translation theory. 1997. vii, 219 pp. 21 Bush, Peter and Kirsten Malmkjær (eds.): Rimbaud's Rainbow. Literary translation in higher education. 1998. x, 200 pp. 20 Snell-Hornby, Mary, Zuzana Jettmarová and Klaus Kaindl (eds.): Translation as Intercultural Communication. Selected papers from the EST Congress, Prague 1995. 1997. x, 354 pp.

19 Carr, Silvana E., Roda P. Roberts, Aideen Dufour and Dini Steyn (eds.): The Critical Link: Interpreters in the Community. Papers from the 1st international conference on interpreting in legal, health and social service settings, Geneva Park, Canada, 1–4 June 1995. 1997. viii, 322 pp. 18 Somers, Harold (ed.): Terminology, LSP and Translation. Studies in language engineering in honour of Juan C. Sager. 1996. xii, 250 pp. 17 Poyatos, Fernando (ed.): Nonverbal Communication and Translation. New perspectives and challenges in literature, interpretation and the media. 1997. xii, 361 pp. 16 Dollerup, Cay and Vibeke Appel (eds.): Teaching Translation and Interpreting 3. New Horizons. Papers from the Third Language International Conference, Elsinore, Denmark, 1995. 1996. viii, 338 pp. 15 Wilss, Wolfram: Knowledge and Skills in Translator Behavior. 1996. xiii, 259 pp. 14 Melby, Alan K. and Terry Warner: The Possibility of Language. A discussion of the nature of language, with implications for human and machine translation. 1995. xxvi, 276 pp. 13 Delisle, Jean and Judith Woodsworth (eds.): Translators through History. 1995. xvi, 346 pp. 12 Bergenholtz, Henning and Sven Tarp (eds.): Manual of Specialised Lexicography. The preparation of specialised dictionaries. 1995. 256 pp. 11 Vinay, Jean-Paul and Jean Darbelnet: Comparative Stylistics of French and English. A methodology for translation. Translated and edited by Juan C. Sager, M.-J. Hamel. 1995. xx, 359 pp. 10 Kussmaul, Paul: Training the Translator. 1995. x, 178 pp. 9 Rey, Alain: Essays on Terminology. Translated by Juan C. Sager. With an introduction by Bruno de Bessé. 1995. xiv, 223 pp. 8 Gile, Daniel: Basic Concepts and Models for Interpreter and Translator Training. 1995. xvi, 278 pp. 7 Beaugrande, Robert de, Abdullah Shunnaq and Mohamed Helmy Heliel (eds.): Language, Discourse and Translation in the West and Middle East. 1994. xii, 256 pp. 6 Edwards, Alicia B.: The Practice of Court Interpreting. 1995. xiii, 192 pp. 5 Dollerup, Cay and Annette Lindegaard (eds.): Teaching Translation and Interpreting 2. Insights, aims and visions. Papers from the Second Language International Conference Elsinore, 1993. 1994. viii, 358 pp. 4 Toury, Gideon: Descriptive Translation Studies – and beyond. 1995. viii, 312 pp. 3 Lambert, Sylvie and Barbara Moser-Mercer (eds.): Bridging the Gap. Empirical research in simultaneous interpretation. 1994. 362 pp. 2 Snell-Hornby, Mary, Franz Pöchhacker and Klaus Kaindl (eds.): Translation Studies: An Interdiscipline. Selected papers from the Translation Studies Congress, Vienna, 1992. 1994. xii, 438 pp. 1 Sager, Juan C.: Language Engineering and Translation. Consequences of automation. 1994. xx, 345 pp.