TRANSLATION IN CONTEXT

btl.39.vw.p65 1 26/09/00, 2:43 PM

 BENJAMINS TRANSLATION LIBRARY

The Benjamins Translation Library aims to stimulate research and training in transla-
tion and interpreting studies. The Library provides a forum for a variety of approaches
(which may sometimes be conflicting) in a socio-cultural, historical, theoretical,
applied and pedagogical context. The Library includes scholarly works, reference
works, post-graduate text books and readers in the English language.

GENERAL EDITOR
Gideon Toury (Tel Aviv University)

ADVISORY BOARD
Marilyn Gaddis Rose (Binghamton University)
Yves Gambier (Turku University)
Daniel Gile (Université Lumière Lyon 2 and ISIT, Paris)
Ulrich Heid (University of Stuttgart)
Eva Hung (Chinese University of Hong Kong)
W. John Hutchins (University of East Anglia)
Zuzana Jettmarová (Charles University of Prague)
Werner Koller (Bergen University)
Alet Kruger (UNISA)
José Lambert (Catholic University of Leuven)
Franz Pöchhacker (University of Vienna)
Rosa Rabadán (University of León)
Roda Roberts (University of Ottawa)
Juan C. Sager (UMIST, Manchester)
Miriam Shlesinger (Bar-Ilan University, Israel)
Mary Snell-Hornby (University of Vienna)
Sonja Tirkkonen-Condit (University of Joensuu)
Lawrence Venuti (Temple University)
Wolfram Wilss (University of Saarbrücken)
Judith Woodsworth (Mt. Saint Vincent University, Halifax)
Sue Ellen Wright (Kent State University)

Volume 39

Andrew Chesterman, Natividad Gallardo San Salvador and Yves Gambier (eds.)

Translation in Context
Selected contributions from the EST Congress, Granada 1998

btl.39.vw.p65 2 26/09/00, 2:43 PM

 TRANSLATION
IN CONTEXT
SELECTED CONTRIBUTIONS FROM
THE EST CONGRESS, GRANADA 1998

Edited by

ANDREW CHESTERMAN
University of Helsinki
NATIVIDAD GALLARDO SAN SALVADOR
University of Granada
YVES GAMBIER
University of Turku

JOHN BENJAMINS PUBLISHING COMPANY

AMSTERDAM/PHILADELPHIA

btl.39.vw.p65 3 26/09/00, 2:43 PM

 TM The paper used in this publication meets the minimum requirements of Ameri-
8

can National Standard for Information Sciences — Permanence of Paper for

Printed Library Materials, ANSI Z39.48-1984.

Library of Congress Cataloging-in-Publication Data

EST Congress (2nd : Granada, Spain : 1998).
Translation in context : selected contribution from the EST Congress, Granada, 1998 /
edited by Andrew Chesterman, Natividad Gallardo San Salvador, Yves Gambier.
p. cm. -- (Benjamins translation library, ISSN 0929-7316 ; v. 39)
Includes bibliographical references and index.
1. Translating and interpreting--Congress. I. Chesterman, Andrew. II. Gallardo San
Salvador, Natividad. III. Gambier, Yves, 1949- IV. Title. V. Series.
P306.2.E76 1998
418’.02--dc21 00-056477
ISBN 90 272 1644 4 (Eur.) / 1 55619 986 4 (US) (alk. paper)
© 2000 - 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

btl.39.vw.p65 4 26/09/00, 2:43 PM

 Contents

Introduction ix
Part I: Conceptual analysis

Sandra Halverson
Prototype effects in the "translation" category 3

Cay Dollerup
"Relay" and "support" translations 17

Lourdes Pérez González

El concepto de numero difuso aplicado al perfil del traductor 27

Part II: Situational, sociological and political factors

Brian Mossop
The workplace procedures of professional translators 39
Hanna Risku & Roland Freihoff
Kooperative Textgestaltung im translatorischen Handlungsrahmen 49

Nike Kocijancic Pokorn

Translation into a non-mother tongue in translation theory 61

Anna Aguilar-Amat & Laura Santamaria

Terminology policies, diversity, and minoritised languages 73

Part III: Psychological/cognitive aspects

Candace Séguinot
Knowledge, expertise, and theory in translation 87
Astrid Jensen & Arnt Lykke Jakobsen
Translating under time pressure 105

Paul Kussmaul
Types of creative translating 117
M T. Bajo, F. Padilla & P. Padilla
Comprehension processes in simultaneous interpreting 127

Nedret Kuran-Burço lu
At the crossroads of translation studies andimagology 143

Part IV: Translation effects

Ritva Leppihalme
Foreignizing strategies in drama translation 153
Alexandra Lianeri
Translation and the shaping of modern democracy 163

Tiina Puurtinen
Translating linguistic markers of ideology 177

Kristiina Taivalkoski
Les simplifications narratives dans une traduction française
de Joseph Andrews 187

Part V: Computer aids in the translation process

Jeremy Munday
Seeking translation equivalents: a corpus-based approach 201

Anja Schwärzt
Machine translation for translators? 211

Martin Kaltenbacher

Aspects of universal grammar in human versus machine translation 221

Part VI: Studies of a text type

Eliana Franco
Documentary film translation: a specific practice? 233
K. Malmkjær, J. Milton & V. Smith
Translation and mass culture 243

Masa Nomura
Text, image and translation 261
Christina Valdés
Reception factors in translated advertisements 271

Susan Sar evi

Creativity in legal translation: how much is too much? 281

Rainer Kohlmayer

Frank Wedekind's sex tragedy Lulu in English and French versions 293

Part VII: Culture-bound concepts

Leticia Herrero
Sobre la traducibilidad de los marcadores culturales 307
Åse Johnsen
El mundo de [la filo]sofïa 317

Anne Martin
Institutional references from the English-speaking world
in the Spanish press 327

Stefania Taviano

British acculturation of Italian theatre 339

Part VIII: Translation history

Paul Bandia
Towards a history of translation in a (post)-colonial context: an African
perspective 353
Emine Bogenç Demirel & Hülya Yilmaz
La critique de la traduction littéraire en Turquie (1940-1992) 363

Outi Paloposki & Riitta Oittinen

The domesticated foreign 373

Index 391
 Introduction

The Granada context

There is currently a strong social demand for Translation Studies in

Spain, largely due to Spain's integration into the European Union. As part of
this growing interest, the University of Granada has become an important
Spanish forum for research and training in translation and interpreting. In
September 1998 it hosted the Second International Congress of the European
Society for Translation Studies. This is a selection of the contributions which
represent papers, interactive sessions and a plenary lecture (by C. Séguinot) of
the EST Granada Congress. Over three hundred people took part, coming from
more than twenty countries. Both the private and public sectors were well
represented, including scholars, professional translators and interpreters, and
postgraduate students. The Congress aimed at providing an international forum
for scholars, teachers and practitioners involved in all aspects of Translation
Studies.

The Translation Studies context

Translation Studies itself has been through many transformations since James
Holmes first approached the issue of naming the discipline and drew up a map
of Translation Studies which still seems surprisingly modern thirty years later.
Over the past two decades, the context of Translation Studies has been shifting
and expanding at an increasing rate.
Texts are now more specialised than in the past and translation is
inconceivable without documentation techniques, both factual and
terminological. Translation practices have undergone a radical change with the
incorporation of computer-based tools. Translation is no longer an isolated
activity; it has come into contact with, and often overlaps, other activities such
as documentation, publishing, technical and professional writing, and editing.
Analysis of the translation process has integrated new psycholinguistic and
cognitive aspects. Cultural, sociological, political, and situational factors have
been introduced that place translation within a broad social context. Translator
X Introduction

values have changed, too, in order to focus more on cultural and

communicative aspects rather than solely on language and text. Scholars
reflecting on the problems of theory and methodology in translation and
interpreting have needed to widen their scope more and more. There is an
increasing need to train researchers in the field of translation and interpreting,
and to establish stronger links between theoretical work and practical
applications such as translator and interpreter training. The effort to find
common criteria for research methodology and ways to systematise it is a
growing concern.
All in all, there are more meetings, communications and publications
focused on translation and interpreting than ever before. The Granada Congress
covered all of the themes mentioned above, and many of them are represented
in the contributions in this book.

Translation in Context

We have grouped the contributions into eight sections. Part I focuses on

conceptual analysis, refining some of the conceptual tools we can use in
Translation Studies. Part II contains four contributions dealing with the
situational, sociological and political context of translation: the translator's
working procedures, cooperation models, the native language issue, and
terminology policy. Part III then turns to the psychological and cognitive
context: issues of knowledge and expertise, problem-solving, creativity,
working memory, and the role of images. Part IV focuses on the study of
translation effects, at various levels: on theatre audiences, on cultural history,
on ideological and aesthetic reception. Part V is on computer aids in the
translation process. Part VI presents contributions dealing with translation
problems associated with different text types: film documentaries, advertising,
popular literature, legal texts, and an unusual literary text. Part VII centres on
the problem of translating culture-bound concepts and cultural stereotypes. The
final section focuses on translation in its historical context, in Africa, Turkey
and Finland.
Space limitations have kept the contributions fairly short, but we trust
that they give a fair picture of the wide range of research that was presented
and discussed in Granada. Our special thanks to the contributors for their
creativity and for their patience during the editing process; to the referees for
their critical comments; and to John Benjamins for including this volume in
their Translation Library series.

Andrew Chesterman
Natividad Gallardo
Yves Gambier
 PART I

Conceptual analysis
PROTOTYPE EFFECTS IN THE "TRANSLATION"
CATEGORY

Sandra Halverson
Department of English
University of Bergen, Norway

Zusammenfassung
Diese Untersuchung sucht empirisch zu bestätigen, daß die interlinguale Übersetzung
den zentralsten Bereich der Prototypenkategorie darstellt, die wiederum mit
verwandten Kategorien nach festgelegten Prinzipien verbunden ist. Studenten werden
um Abgabe von Werturteilen über eine Reihe von Übersetzungspaaren gebeten. Die
Zusammenstellung der Textreihe diente folgendem Ziel: Prüfung der Zentralität der
Sprache als ein System unter anderen und der Relevanz der Bewegung zwischen
semiotischen Systemen oder innerhalb derer. Die Haupthypothese beinhaltete die
Annahme über die Betrachtung der Beispiele mit intersystemischer Bewegung als das
beste/die besten Beispiel(e), insbesondere die Beispiele mit Bewegung zwischen zwei
deutlich distinkten Sprachen. Die Ergebnisse bestätigten die Hypothese.

Résumé
Dans cette contribution, nous avons cherché une base empirique à l'idée que c'est la
traduction interlinguistique qui est centrale dans une catégorisation prototypique. Nous
avons demandé à des étudiants d'évaluer des paires de traduction, notre objectif étant
de tester à la fois la place essentielle de la langue parmi les systèmes sémiotiques et
l'importance du mouvement entre ou à l'intérieur de ces systèmes. Notre hypothèse
principale a été que les meilleurs exemples seraient ceux portant sur le transfert
intersystémique, plus précisement entre deux systèmes linguistiques bien distincts. Les
résultats confirment l'hypothèse.

Resumen
Este estudio tiene como objetivo demostrar empiricamente que el concepto de
traducción interlingual es el elemento principal de una categoria prototipo que a su vez
esta muy ligada a distintas categorias relacionadas. Se pidió a un grupo de estudiantes
que seleccionaran los mejores ejemplos entre una serie de pares de traducción
disenados para comprobar la centralidad de los sistemas semióticos. La hipótesis
partia de la base de que los mejores ejemplos estarian relacionados con el movimiento
4 Sandra Halverson

intersistémico, especialmente los que suponian un movimiento entre dos lenguas

claramente diferenciadas. Los resultados confirmaron la hipótesis.

1. Introduction

In a set of previous papers (Halverson forthcoming; Halverson 1997, 1998,

1999), I outlined the philosophical and empirical problems associated with
attempts at defining the concept of "translation" in terms of a classical
category. The main thrust of the arguments presented there is that the classical
theory of concepts has been undermined philosophically, and that the current
drive in research technologies (corpus-based analysis, etc.) is revealing the
empirical inadequacies of the same approach. In the present paper I offer some
empirical evidence in support of the philosophical arguments. Here, I test my
hypothesis that the "translation" category will demonstrate prototype effects
(Rosch 1978, Gibbs 1994).
In what follows, I present an empirical investigation into subjects'
responses to the Norwegian concept of "oversettelse" ('translation'). The study
was carried out using a slightly adapted version of the Coleman-Kay (1981)
methodology. The initial aim was to establish whether the concept, in fact,
demonstrates prototype effects. An additional objective was to investigate the
pattern of prototype effects found (if any), to see what that pattern might
indicate with respect to the underlying cognitive structure that motivates them.
The fundamental question which underlies this investigation is actually
two-fold, and the conclusions I hope to draw will address two related issues,
i.e. the internal structure of the "translation" category, and the category's
structured relationships to other concepts. These are two issues which have
formed the crux of the move from classical to prototype categories. In brief,
prototype theory and research have shown that, instead of the uniform internal
structure and clear boundaries guaranteed by necessary and sufficient
conditions, virtually all natural language concepts show signs of having graded
membership (not all members are equal), and fuzzy boundaries (where one
concept stops and another starts is indeterminate). (For a survey of this subject,
see Lakoff 1987.)

2. Theoretical background

The questions of boundary drawing and category relationships (internal and

external) are not new to the field of Translation Studies. In fact, they represent
areas of considerable current tension (see e.g. Koller 1995, Toury 1995, Pym
1995, 1997). The discussion of translation types represents attempts at
 Prototype effects 5

sub categorization of the "translation" category, as well as attempts to link that

category to related ones, e.g. "paraphrase", "interpretation", "rendering", etc.
The well-known translation typology proposed by Jakobson (1959) -
intralingual, interlingual and intersemiotic - has been criticized by Derrida
(1985), and also by Toury (1986), who suggests a more refined classification of
translation types. In Toury's version, the main distinction is between
intrasemiotic and intersemiotic translation; intrasemiotic then divides into
intrasystemic (e.g. intralingual) and intersystemic (e.g. interlingual) translation.
However, both these classifications are based on classical categories, and there
is thus no privileging of one type of translation over another.
I have chosen to test the dimensions which figure in these two classic
accounts; I aim to show that none of them constitute necessary and sufficient
conditions for membership in the "translation" category, and that they play
different roles, i.e. demonstrate varying salience, in informants' evaluations.

3. Methodology

As mentioned in the introduction, the methodology used in this investigation is

that first introduced by Coleman and Kay (1981) in their study of the "lie"
concept. (For a review and critique of this methodology, see Myers 1994.)
In my study, the subjects were asked to respond to seven sets of
translation pairs using the same type of seven-point scale (though the order was
reversed, i.e. one through three were various degrees of "yes, it is a
translation", while five through seven were the corresponding "no, it is not a
translation" responses). The actual questionnaire is included in the appendix to
this paper. The seven translation pairs, like the Coleman and Kay "lie" stories,
were chosen to comprise various combinations of the dimensions mentioned in
Section 2: the type relationship (are the two systems the same kind, i.e. the
intra- versus intersemiotic dimension), the token relationship (are they the
same system, i.e. the intra- versus intersystemic dimension), and linguality (of
at least one of the pair members). Accordingly, on the type relationship
dimension, whether or not the pair were of the same semiotic type (i.e. both
linguistic, both numeric, etc) values of 0 (the same) and 1 (different) were
assigned. Similarly, 0 or 1 values were also assigned for the token relationship
dimension, whether the systems were the same (0) or different (1). Finally, on
the linguality dimension, a value of 0 was assigned if neither of the two texts in
a given pair was linguistic, and a value of 1 was assigned if at least one of the
texts was linguistic (by "text" I mean an internally coherent semiotic entity).
This assigning of values represents an adaptation of the original
methodology in that instead of giving +/- values, I have assigned values which
represent an opposition of same versus different, an orientation towards either
intra- or inter-, on the first two dimensions. This value assignment represents
6 Sandra Halverson

both a strengthening and a weakening of the methodology. First of all, it allows

for a more detailed analysis of the pattern of effects within the overall category,
i.e. the status of the various members relative to each other. On the other hand,
it makes testing for necessary or sufficient conditions slightly less
straightforward, at least with respect to these two dimensions. The 0/1 values
do not directly represent the presence or absence of something: they represent
two alternatives. However, the values do provide indirect evidence of the
presence or absence of something else, i.e. a boundary of some sort. In other
words, the 1 and 0 values indicate the presence or absence of a boundary,
respectively, and that boundary may exist at either the type or token level. But
this particular characteristic must be studied rather more indirectly than the
third, the presence or absence of language, which is actually directly reflected
in the 0 and 1 values. Nevertheless, I have opted to note whatever information
comes out of the test with regard to the roles of the dimensions as necessary or
sufficient, both direct and indirect, and aim for more detailed information about
the internal structure of the category.
The seven text pairs in the questionnaire also cover the six
subcategories outlined by Toury, and the various combinations of dimensions
may be shown in tabular form as follows, where the seven text pairs are
numbered and their designation in Toury's framework is also given:

Table 1. Translation pairs, dimensions and values

Text pair type relationship token relationship linguality

(boundary) (boundary)
1 - intrasystemic 0 0 0
2- intersystemic 0 1 0
3- intersemiotic 1 1 1
4- interlingual 0 1 1
5- interlingual? 0 0/1 1
6-intralingual 0 0 1
7-intersemiotic 1 1 0

It is important to note that, in the references to Toury's typology, not all of the
labels are at the same level of generality. Thus, while the seven text pairs on
the questionnaire are all at the lowest level of generality in Toury's typology,
the superordinate labels given to pairs one, two, three, and seven in Table 1 are
merely an artifact of the labeling system employed by Toury, and the lack of
conventional (or natural language?) designations for these particular forms.
A further note is necessary on text pair five, as it is an exception to the
overall system. The two texts in pair five are both linguistic, and thus are
assigned a value of 1 on that particular dimension. The two are the same type
 Prototype effects 7

of semiotic system; hence the value of 0 on that dimension. However, the

question of whether the two constitute two distinct systems or not (the token
relationship) was less clear. In text pair four, the translation was from English
to Norwegian; in other words, between two distinct languages. In text pair six,
on the other hand, the translation was within one system, i.e. one of Norway's
two official written languages (bokmål). In five, however, the translation was
from a recognized dialect (western Norwegian) into bokmål. Thus, the status of
the dialect as a distinct language was less clear. The reason for including this
particular example was precisely to study the significance of relative
indeterminacy on that particular dimension.
The questionnaire was administered to 103 undergraduate students at
the English Department at the University of Bergen, during the spring and fall
semesters of 1997. These students were all native speakers of Norwegian.
Two-thirds of the subjects had not taken the introductory course in practical
translation which is part of the Department's undergraduate program, while the
remaining one-third was approximately half-way through that course. The
representativity of the group is not statistically guaranteed, as sampling
techniques were not adopted. Nevertheless, the students were all considered to
be equally theoretically naive, and thus representative of the population in that
sense.

4. Results and discussion

It is important to keep in mind that the aim of this study is to determine

whether there is evidence of prototype effects (membership gradience and
fuzzy boundaries) in the "translation" category. If such evidence is found, then
the pattern of effects will also provide additional information regarding the
underlying cognitive structure which motivates them. In the following, the
results of the investigation are presented in Section 4.1, and the conclusions
that may be drawn with respect to these issues are presented in Section 4.2.
Certain implications of these findings for further work on the underlying
cognitive structure are outlined in Section 4.3.

4.1 Results

The mean scores and standard deviations for the seven translation pairs are
given in Table 2. Readers may recall that the seven-point scale ranged from 1
(yes, completely sure) to 7 (no, completely sure). A response of four indicated
"do not know".
8 Sandra Halverson

Table 2. Mean scores and standard deviation for each text pair

text pair mean score s

1 -intrasystemic 3.96 2.634
2-intersystemic 3.099 2.027
3-intersemiotic 1.752 1.615
4-interlingual 1.426 .864
5-interlingual? 1.663 1.458
6-intralingual 3.545 2.138
7-intersemiotic 5.04 1.849

It is clear from Table 2 that only one of the pairs was clearly excluded from the
category, i.e. text pair seven, intersemiotic (non-linguistic) translation
(translation from a visual (painting) system to an aural one (musical)). All
other text pairs were identified as translations, though the mean scores indicate
that the degree of certainty varied. Furthermore, the means clearly indicate
graded membership; text pair number four (interlingual translation) was found
to be the most central member, while text pair number one (intrasystemic) was
particularly close to the "do not know" score. The remaining pairs also showed
varying degrees of membership.
Interestingly, the most central member, text pair four, also showed the
lowest standard deviation, which indicates that the group of respondents was
most agreed on this particular pair. It is also interesting that as the mean
approaches 4.0 ("do not know"), the standard deviation also increases, and
there is a direct correlation. The same pattern is shown for both yes and no
responses; in other words, the standard deviation is lower for text pair number
seven, the only pair excluded from the category, than it is for text pair six, i.e.
intralingual translation, with respective means of 5.04 and 3.54. Thus, the
pattern is maintained. The conclusion that can be drawn is that there is greatest
intragroup certainty about the most central member, and that as average
uncertainty grows, so does the range of responses (a finding which is
intuitively appealing).
On the basis of the above, an additional test was carried out to test for
the overall significance in group responses. A one-way ANOVA (see Butler
1985: 132) indicated that both between-group and within-group variance was
significant (P = .0326 and P = .0001 respectively). A scheffé test was then used
to check the significance of differences between means, followed by a z-test.
(Full statistical details are available on request from the author.)
Results showed the score for text pair seven is significantly different
than those for all of the other pairs. In addition, the group of text pairs three,
four and five does not show significantly different ratings. Similarly, scores on
pairs one, two and six are not significantly different. Indeed, the six text pairs
 Prototype effects 9

which the respondents included in the category seem to show a rather ripple-
like pattern around the center, with the most central member most immediately
surrounded by the (not significantly different) text pairs three and five, and, in
the next "ring" the texts which were significantly different from three four and
five, i.e. text pairs one, two and six, which in turn were not significantly
different from each other.
This may be illustrated in Figure 1, below, in which the dotted circles
are intended to represent permeable boundaries:

Figure 1. The pattern of effects

4.2 Discussion

The results presented in Section 4.1 provide clear evidence of prototype effects
in the "translation" category. The most obvious and immediate evidence is that
indicating graded membership, as indicated by the means and corroborated in
the statistical analysis of subject ratings. As pointed out in the introduction,
graded membership is one of the characteristics of a non-classical category.
The second characteristic, the absence of a clear category boundary constituted
by necessary and sufficient conditions, is more difficult given the methodology
employed and requires a somewhat more detailed look at the results presented.
A clear category boundary is demonstrated if any of the dimensions
tested constitutes a necessary and sufficient condition by means of which
instances may be included or excluded. The only text pair that was excluded
from the category was number seven, which exemplified intersemiotic, non-
linguistic translation. It had values of 1 on type relationship, 1 on token
relationship, and 0 on linguality. In other words, there was a difference in
semiotic system at both type and token level and there was no language
10 Sandra Halverson

involved. On the language dimension, one might proceed on the assumption

that the lack of language in this example constituted the criterion for exclusion,
i.e. language is necessary. However, pairs one and two were included, and
neither of them are linguistic. Thus, language does not seem to be necessary,
though it seems to be sufficient, given the inclusion of pair number six
(intralingual).
Even though the values given on the other two dimensions preclude the
direct testing of the type and token relationships as necessary and sufficient
conditions, it is possible to determine whether the presence of a boundary (a
value of 1) at either level is necessary or sufficient. The exclusion of text pair
seven thus provides indirect evidence that a difference (boundary) between
systems at both type and token level is not sufficient for category membership.
Nor are distinctions at either level necessary, as demonstrated by the inclusion
of text pair one (which differs from seven in lacking a boundary at both levels),
or text pair two (which differs only in lacking a boundary at the type level).
The necessity of a token-level distinction (or boundary) is disconfirmed by the
inclusion of pairs one and six. Thus, indirectly, there is evidence that the
presence of a boundary at either type or token level in a pair of translation pairs
is neither necessary nor sufficient for category membership.
In sum, the conclusion that can be drawn from the results outlined here
is that "translation" demonstrates clear prototype effects of both types, i.e.
graded structure and fuzzy boundaries. The clustering demonstrated so clearly
in Figure 1 is a visualization of the pattern of effects. However, it must be kept
in mind that the effects are not the category's structure, and that the next task to
account for that pattern through the reconstruction of the cognitive structure
that has motivated them..

4.3 The role of the three dimensions

In Section 4.2, it was shown that none of the three dimensions constituted
necessary conditions for membership, though the involvement of language as
one of the semiotic systems seemed to be a sufficient condition. The presence
of a boundary at either type or token level was not found to be necessary or
sufficient. It seems obvious, however, that the various combinations on these
dimensions are playing a role in the judgments that the subjects made.
Figure 1 illustrated the fact that text pairs three, four and five were not
significantly different, though text pair four (interlingual translation) was
clearly the most central member (as indicated by the means). All of these pairs
involved linguistic systems, and all of them (with the possible exception of
number five) demonstrated a boundary at the token level, i.e. an intersystemic
relationship between the pairs. In addition, pair number three showed a
boundary at the type level. Thus, the most central members all seem to involve
 Prototype effects 11

at least one linguistic system and at least one boundary, more specifically, a
lower (token) level one.
In the next ring in Figure 1 are text pairs one, two, and six.
Interestingly, pair one does not demonstrate a boundary at any level; nor does it
involve language. This pair was probably close to being excluded from the
category, as it was closest to a score of 4 (don't know), and also showed the
largest standard deviation (2.634). Furthermore, two different statistical tests
gave different outcomes as to the significance of the difference between pairs
one and two. In other words, pair number one, at least relative to one other
member in the outer "ring", seems to be even more peripheral to the category.
At the very least, this suggests that the status of pair number one is less clear
than that of the other pairs within that group. It may provide further indication
of the overall borderline status of this particular pair with respect to the
category as a whole. Additional testing could illuminate this question further.
Text pairs two and six diverge from the central members along the language
and boundary dimensions, respectively. In other words, pair two has a
boundary at token level (but not at type level) and no language, while pair six
involves language, but no boundaries at either level. Thus, the category shows
typical radial structure, with extensions along the token-level boundary
dimension and the language dimension.
The results presented in Section 4.1 also provide information about the
relative salience of the main dimensions in various areas. In other words, the
involvement of a boundary and/or language is not equally important for each
pair judged. This is particularly clear with respect to the existence of a system-
level boundary. In order to investigate this issue in more detail, it is useful to
consider various combinations of text pairs which differ only along that
particular dimension. Text pairs one and two, for example, are similar on
dimensions one and three, while they differ only in that pair number two
involves two different systems, while pair number one does not. These two
pairs were not significantly different from one another according to the scheffé
test, though they were significantly different from each other according to the
z-test. This would suggest that the salience of a boundary in this area is
questionable.
If, on the other hand, one compares pairs four, five and six, the result is
slightly different. Also these three differed only along the token-level
dimension, and the presence of a clear boundary was decreasing from four to
six. Four had a clear boundary (between two distinct languages), five had a less
clear (or no) boundary (between a dialect and a standard language), while in six
there was no boundary. Pairs four and five, it will be remembered, were among
the most central members, while six was not. Pairs four and five were not
significantly different from each other, while both were significantly different
from pair six. Thus here, it seems that the presence of a boundary was
influential in the subjects' category judgments.
12 Sandra Halverson

The difference between text pairs one and two and the group of four,
five and six lies in the involvement of language. In pairs which are not
linguistic (the former), the involvement of a boundary seems to play less of a
role, if any. In pairs which are linguistic (the latter), on the other hand, the
clarity of the boundary does seem to be relevant. What this seems to indicate is
that the two dimensions, i.e. the presence of a boundary and the involvement of
language, are interacting, at least close to the center of the category.
In the interest of deriving the cognitive model, the relevant information
is that the involvement of language and the crossing of at least one, preferably
a token-level, semiotic boundary seem to be shared by all of the most central
category members. Furthermore, these dimensions, separately, seem to be used
productively in the process of extending category membership. Finally,
dimensions two and three seem to be interacting in a significant way, at least at
the center of the category. Thus, the cognitive model posited for this category
must account for the interplay of these two dimensions in the center, along the
periphery and outside the category (i.e. the exclusion of text pair seven).

5. Conclusion

It is my claim that "translation" is a non-classical category that demonstrates

prototype effects. Consequently, my view is diametrically opposed to the view
that categories are defined by necessary and sufficient conditions, a view which
entails that categories have clear boundaries and equal members. A further
characteristic of the classical view is that categories are structured
hierarchically through the addition of further specifying features.
Attempts to define translation have made use of such categories, while
at the same time (paradoxically) there has been wide recognition that
interlingual translation is in some way special (e.g. Jakobson's "translation
proper"). Furthermore, an extensional approach (i.e. "translation" is whatever
people recognize as such) has for some time been working under the
assumption that generalizations are possible without extensions being fixed in
any way. This is profoundly problematic.
The view argued in this and preceding papers (particularly Halverson
1998) is that extensions may be anchored in a cognitive structure, which is
more or less shared, and which may be investigated empirically. The effects
studied in this paper are surface phenomena which provide insight into the
cognitive structure which is used by subjects in the extensional task.
My results provide, to my mind, unequivocal evidence that the
"translation" category is indeed a non-classical one. The results of the test
demonstrate clear prototype effects, i.e. graded membership and fuzzy
boundaries. Furthermore, the intuition of translation theorists, i.e. that
 Prototype effects 13

interlingual translation is, in some way, special is also vindicated. This

particular type is found to be the most central member of the category.
The more particular findings indicate that the dimensions of language
and the presence of a system-level boundary are related in a complex way.
More specifically, these two dimensions provide a means of extending
category membership, while at the same time the two dimensions demonstrate
varying saliencies at central and peripheral areas of the category. These are
findings which must be taken into consideration when the cognitive structure is
posited at the next stage.
Another interesting aspect of the results of this test comes to light when
we consider the cross-linguistic picture. The subjects involved in this
investigation were all native speakers of Norwegian, and their judgments were
pertinent to the Norwegian concept of "oversettelse". It must be remembered
that the standard definition of this concept only overlaps with the situation
demonstrated by text pair number four. Consequently, it might be expected that
Norwegian subjects would restrict category membership to that particular pair.
The fact that they did not do so is highly interesting. It becomes even more
interesting in light of the fact that their category extensions are entirely in line
with the most common extensions (as reflected in additional senses) of the
English-language concept (cf. Oxford English Dictionary 1989:409). This
cross-linguistic similarity is highly interesting from the perspective of
developing a cognitive account.
It would be most interesting to see whether tests of this sort for related
"translation" concepts in other languages demonstrate the same pattern of
effects. It is my guess that at least languages that have derived their concept,
either directly or indirectly, from Latin will do so. Whether the same is true of
"translation" concepts which have developed from other sources is a question
of some import. If a different pattern of effects provides evidence of a different
type of cognitive structure, then extensional judgments may not coincide in
spite of an apparent "correspondence" of terms. Generalizations based on such
judgments will be ungrounded. Such concerns are of interest in numerous
respects, but when the issue at hand is the structuring of a category which
constitutes an object of study for an entire academic discipline, then the stakes
are considerably higher.
14 Sandra Halverson

APPENDIX - QUESTIONNAIRE

Vurdér hvorvidt du mener B er en oversettelse av A, og sett ring rundt det tallet

som best gir uttrykk for både svaret ditt og hvor sikker du er. Ved å sette ring
rundt ett av tallene fra 1-3 svarer du ja, at B er en oversettelse av A. Ved å sette
ring rundt ett av tallene fra 5-7 svarer du nei, at B ikke er en oversettelse av A.
Hvis du ikke kan svare hverken ja eller nei, setter du ring rund tallet 4.

Ja, det er en oversettelse Nej, det er ikke en

oversettelse
1 2 3 4 5 6 7
helt ganske noe yet Noe ganske helt
sikker sikker sikker ikke sikker sikker sikker
A B

1 1 2 3 4 5 6 7

2 $10 69,-kr. 1 2 3 4 5 6 7

3 vikeplikt 1 2 3 4 5 6 7

4 What day is Hvilken dag 1 2 3 4 5 6 7

it? er det?

5 Ka e' det? Hva er det? 1 2 3 4 5 6 7

6 Han er Han liker å ta 1 2 3 4 5 6 7
makelig ting med ro
anlagt.
7 * Mussorgskys 1 2 3 4 5 6 7
Hartmanns musikkstykke
malerier og 'Bilder på en
tegninger utstilling'
 Prototype effects 15

* Maleren Victor Hartmanns malerier og tegninger danner grunnlaget for komponisten

Mussorgskys musikkstykke 'Bilder på en utstilling', som er blitt beskrevet som
Mussorgskys gjenomgang av sine favorittbilder.

Translation/explanation of Questionnaire

1. Translation of instructions :
Please indicate whether you believe that B is a translation of A, and circle the
number which best represents both your response and how certain you are. By
circling one of the numbers from 1-3, you are answering yes, that B is a
translation of A. By circling one of the numbers from 5-7, you are answering no,
that B is not a translation of A. If you are not able to answer either yes or no,
circle the number 4.

2. Translation of response scale :

1 and 7: completely certain
2 and 6: quite certain
3 and 5 : somewhat certain
4: do not know

3. Translation/explanation of text pairs

Text pair no. 1 A: a red light

B:a stop sign
Both in current use in Norwegian road traffic system.
Text pair no. 2 A: specific amount in US dollars (10)
B: same amount (at current rate of exchange) in Norwegian kroner
Text pair no. 3 A: Norwegian yield sign
B: the word 'yield'
Text pair no. 4 A: 'What is it?' (in English)
B: 'What day is it?' (in standard Norwegian - structurally
isomorphic)
Text pair no. 5 A: 'What is it?' (in Norwegian dialect
B: 'What is it?' (in standard Norwegian
Text pair no. 6 A: 'He is rather laid-back by nature' (in standard Norwegian
B: 'He likes to take things easy.' (in standard Norwegian)
Text pair no. 7 A: * Hartmann's paintings and drawings
B: Mussorsky's musical piece 'Pictures at an Exhibition'

* The painter Victor Hartmann's paintings and drawings form the basis for the
composer Mussorsky's piece 'Pictures at an Exhibition', which has been described as
Mussorsky's review of his favorite pictures.
16 Sandra Halverson

References
BUTLER, C. 1985. Statistics in linguistics. Oxford: Basil Blackwell.

COLEMAN, L. & P. KAY. 1981. Prototype Semantics: The English word lie. Language 57(1),
26-44.

DERRIDA, J. 1985. Des Tours de Babel. Difference in Translation ed. by J. Graham, 165-207.
Ithaca and London: Cornell University Press.

GIBBS, R. 1994. The Poetics ofMind. Cambridge: Cambridge University Press.

HALVERSON, S. 1997. The concept of equivalence in translation studies: Much ado about
something. Target 9(2), 207-233.

HALVERSON, S. 1998. Translation studies and representative corpora: Establishing links

between translation corpora, theoretical/descriptive categories and a conception of the object of
study. Meta 43(4), 494-514.

HALVERSON, S. 1999. Conceptual work and the 'translation' concept. Target 11(1), 1-31.

HALVERSON, S. (forthcoming). Image schemas and metaphoric processes in the 'translate'

category. Resubmitted to Metaphor and Symbol.

JAKOBSON, R. 1959. On linguistic aspects of translation. On translation, ed. by R.A. Brower,

232-239. Cambridge, Mass.: Harvard University Press.

KOLLER, W. 1995. The concept of equivalence and the object of translation studies. Target
7(2), 191-222.

LAKOFF, G. 1987. Women, Fire and Dangerous Things. Chicago: University of Chicago
Press.

MYERS, D. 1994. The Chinese morpheme gong. Cognitive Linguistics 5(3), 261-280.

PYM, A. 1995. European translation studies, Une science qui dérange, and why equivalence
needn't be a dirty word. TTR - Traduction, Terminologie, Rédaction 8(1), 153-176.

PYM, A. 1997. Koller's 'Äquivalenz' revisited: Review of Werner Koller's Einfüring in die
bersetzungswissenschaft'. The Translator 3(1), 71-79.

ROSCH, E. 1978. Principles of categorization. Cognition and Categorization ed. by E.

ROSCH and B.B. LLOYD, 27-48. HILLSDALE, NJ: Lawrence Erlbaum Associates.

TOURY, G. 1986. Translation : A cultural-semiotic perspective. Encyclopedic Dictionary of

Semiotics, ed. by T. SEBEOK, 2. 1111-1124. Berlin: Mouton de Gruyter.
"RELAY" AND "SUPPORT" TRANSLATIONS

Cay Dollerup
Centre for Translation Studies,
University of Copenhagen, Denmark

Zusammenfassung
Im Fokus dieses Artikels steht die Sachlage, daß viele Übersetzungen nicht die Sprache
des Originals als Ausgangssprache aufweisen, sondern Übersetzungen aus einer
intermediären Drittsprache (Viert- usw.) darstellen. Da die meisten intermediären
Realisationen für wirkliche Rezipienten gemacht worden sind, wird vorgeschlagen,
diese Realisationen aus Gründen der Exaktheit als 'relays' (Relais) für andere Texte im
Unterschied zu 'indirect translation' (indirekte Übersetzung) zu bezeichnen. Der Term
'indirect translation' bezieht sich auf intermediäre Realisationen mit original keinen
Konsumenten. Dem Verfahren der Übersetzer, manchmal Übersetzungen in andere
Sprachen als in die eigene Zielsprache zu konsultieren, wird ebenfalls im Artikel
Aufmerksamkeit geschenkt. Die Bezeichnung 'support translation' würde sich auf
diese Fälle beziehen.

Résumé
Les traductions ne sont pas toujours faites directement de la langue de l'original vers la
langue d'arrivée mais passent par une troisième (quatrième, etc.) langue intermédiaire.
Comme ces produits intermédiaires ont été réalisés pour des récepteurs bien précis, on
pourrait les appeler "relais", servant pour d'autres textes d'arrivée, tandis que l'expression
"traduction indirecte" pourrait désigner les cas où les textes intermédiaires n'ont pas
d'audience réelle. L'auteur parle aussi de "traduction-soutien", en constatant que les
traducteurs s'appuient parfois sur des traductions produites en d'autres langues que les leurs.

Resumen
Este articulo se centra en el hecho de que a menudo la traducción no se realiza
directamente entre la lengua origen y la lengua término, sino que pasa por una tercera
(o cuarta, etc.) lengua intermediaria. Puesto que la mayoria de las actuaciones
intermedias se han realizado pensando en un publico real, se argumenta que seria más
preciso llamar a estas actuaciones "relevos" para otros textos término y reservar el
término "traducción indirecta" para los casos en los cuales no existen unos receptores
18 Cay Dollerup

reales para las actuaciones intermedias. El articulo senala, además, que a veces los
traductores consultan traducciones a otras lenguas diferentes de su propia lengua
término, utilizando asi lo que se podria denominar una "traducción de apoyo".

1. Translation as communication

It is often argued that since translation is a type of communication, a model of

communication is a useful basis for discussion. In such a model, interlingual transfers
are characterised by the existence of a medium of translation (a translator, an
interpreter, a subtitler, or, for that matter, a machine). Consequently the communic-
ational chain looks like this:

Sender > message > (near) simultaneous reception/mediation/sending > message

> receiver.

Discussions focus on the relationship between the source-text original and the
target-language version. The terms employed about this relationship describe a
reality in which, even when there are intermediary realisations, the focus is on the
relationship between the original sender and the end consumer:

'Original sender ' > message > intermediary decoding and (near) simultaneous
encoding by transmitter > message > decoding by end receiver.

2. 'Indirect' translation

However, much interpreting, subtitling, and translation is not based on the text of
the original in the source language, but on realisations of the original in yet other
languages.
An interest confined exclusively to the relationship between the original
sender and the end consumer may occasionally be justified, provided the aim of the
communication is to convey the exact meaning between only two parties, two end
interlocutors. This was the case in a Danish murder trial, where a Thai witness made
a deposition in court as follows:

The woman's deposition reached the jury in circuitous ways, since it was first
rendered from Thai into English by one interpreter and then into Danish by yet
another interpreter. (Politiken. 12 September 1998. 1:4, cc. 5-6. My translation.)

In principle, then, a translation procedure might go via the following stages:

(1) New speech; (2) intermediary, ephemeral rendition (no consumers in this
target language); and (3) end which leads to response, that is another new speech.
 "Relay" and "support" translations 19

One imagines that this scenario of indirect translation, in which

intermediary realisations are merely a means for establishing communication
between only two parties, is rare. It would, presumably, be confined to liaison
interpreting or - as in the above case - court interpreting where there are well-
defined roles for the two parties at each end of the chain of communication (namely,
on the one hand, the judiciary establishment comprising the jury, the judge, as well
as the counsels for the defence and prosecution, and, on the other hand, the
witnesses and the defendant(s)).
In many types of interlingual transfer there may be translations into several
languages, one after the other. In these cases, a terminology which implies that these
realisations are of secondary importance, as does indirect translation with its
explicit focus on the end realisation, is misleading. This terminology does not really
allow for the fact that the vast majority of translational realisations of an original into
any language, are primarily intended for an audience. The term indirect translation
should be reserved for cases like the court proceedings cited, for situations where
two parties must communicate by means of a third intermediary realisation which
has no legitimate audience.

3. Relay

In order to focus sharply on the implications of these intermediary, non-original

realisations, I suggest we introduce relay, a term which is well-established in
interpreting, but the relevance of which seems to have been largely overlooked in
Translation Studies. Even in interpreting, relay is viewed mostly in relation to the
final users. (See the relevant entries in Shuttleworth and Cowie 1997, Snell-Hornby
et al. 1998; on intermediary realisations, see Toury 1995: 129-146.)
Within such a framework, relay can be defined as a mediation from source
to target language in which the translational product has been realised in another
language than that of the original; the defining feature is that the intermediary
translation has an audience, that is consumers, of its own. Unlike the renditions
rightfully termed indirect above, these intermediary realisations do not exist in a
vacuum or in an ephemeral interval between the original and the target language
version under discussion.
Roughly speaking the chain of communication is as follows:
1. Original -> audience + translator
2. Translation -> audience + translator
3. Translation -> audience + translator
4. Translation -> audience + translator, etc.

Relay often also involves a delay in the arrival of the message in interlingual
transmission as opposed to an instantaneous - but not necessarily simultaneous -
20 Cay Dollerup

rendition. Accordingly, it was no coincidence that the use of relay in interpreting

became obvious to non-professional outsiders when Danish became an international
language: relay is most apparent in language transfers involving minor languages,
and is most visible to public audiences in simultaneous conference interpreting. (See
further Dollerup 1996.)

3. J Conference interpreting

The Common Market (now the European Union) had four official languages,
namely Dutch, French, German, and Italian when originally founded. The entry of
the UK, Ireland, and Denmark in 1973, added two more languages, namely English,
covering both Ireland and the UK, and Danish. (For background information see
Dollerup and Ceelen 1995.) At the European institutions, it is (in principle) the
prerogative of national delegates to speak their own language. Danish members of
the European Parliament exercised this right, and also listened to Danish inter­
preters. Other nationalities noted that Danes were slow on the uptake, indeed
dimwitted. Clever journalists then noticed that there seemed to be degrees of dim-
wittedness among the members of parliament: if, for instance, a Dutch delegate
cracked a joke, the Dutch would laugh; after 5-10 seconds, the French and Germans
would get the point; after some more seconds, the English, the Irish, and the Italians
would catch on, and, then,finally,the Danes would join in the general merriment.
The figure of 5-10 seconds is traditionally cited in research on simultaneous
interpreting as the 'time lag' between the original sender's utterance and the inter­
preter's rendition. This is the explanation for the ripple response described above:
• The Dutch speaker is understood by his countrymen, who laugh right away.
• The German and the French interpreters render the joke in German and French,
respectively. Consequently, after 5-10 seconds, the German and the French
listeners appreciate the joke.
• The English and the Italian interpreters do not know Dutch, so, understanding
German and French, they use these languages as their source texts and their
delegates laugh after 10-20 seconds. In this example, the Danish interpreters use
English as their source-text: 'dim-wittedness' is due to relay in which all
listeners use interpreted renditions.
Bureaucrats, politicians, and interpreters who have lost touch with the
profession argue that relay leads to confusion, since there must be numerous errors
in every transfer from any one language into another. Working interpreters tend to
agree with my views which are based on fairly systematic listening to relays:
indisputable errors may arise, but there are few (Dollerup 1987).
It is true that there are errors in the product of relay: however, these have
rarely anything to do with the phenomenon of relay, but are due to the first in­
terpreter's misunderstanding of the original sender. Normally, these errors are
caused by the original speakers: delegates mumble, hit the microphone, turn their
 "Relay" and "support" translations 21

heads, speak dialect, use slang, quote figures in incomprehensible ways, and,
without warning, use prepared manuscripts, thus shifting delivery into a written
mode without informing anybody, least of all the interpreters (for an interpreter's
views on such occurrences, see Pearl 1995).
Similarly, the use of relay may lead to errors in the product. Let us assume
that Danish speeches are relayed by way of Dutch to English and German, from
these languages to French and Italian, and so on. This gives a time lag of some 30
seconds before the Danish speech reaches Portuguese. Let us then assume - and it is
a fair assumption - that the chair of the meeting understands Dutch. When the
Dutch interpreter concludes the brilliant summing-up of the last Danish speech, the
chair logically assumes that this is the end and gives the floor to the next delegate,
say, a Greek. The Greek delegate opens with a rousing statement. Here is the snag:
the Greek original begins before the entire original (Danish) message has made it
through all relays; at the same time, the interpreters have to reorganise their chains
of relay. This sometimes forces the interpreters either to cut short the previous
speech, or omit the beginning of the new speech. The delegates notice that
something has gone wrong and assume it to be the interpreters' fault. We see that it
is not due to relay as such, but to the management of relay: it is the fault of the chair.
Overall, then, at the institutions where relay is most often seen by the public
(the European institutions, the UN), the process of relay does not lead to many
misunderstandings at the linguistic level, for the interpreters are professionals who
are trained to deliver well-phrased and easily understood speech.
On the other hand, relay interpreting is undoubtedly a risky undertaking
when there are no professional middlemen around: in my experience, it usually
means that communication takes place by means of sign language and gestures.

3.2 Translation

Relay and concomitant delay are widespread in translation, indeed so common that,
in literary studies, for instance, it is hardly noted at all. We do not consider it strange
that literature takes time to cross cultural barriers. We accept that Shakespeare was
not translated into Spanish until the 18th century.
Sometimes the original no longer exits. The Jesus of the New Testament
must have spoken Aramaic, but no Aramaic text has been preserved. Modern Bibles
are made from relay editions using different languages. The English Bible overseen
by John Wyclif (c. 1385) used the Latin Vulgate as its source text. The Vulgate
derived from St. Jerome's Bible (c. 400), itself a Latin translation of Greek sources
which St. Jerome subsequently revised by means of Hebrew texts. William
Tyndale's English Bible of 1526 was based on Greek texts. Knowing no Greek,
Miles Coverdale (1535) used Tyndale's version of the New Testament and
translated the Old Testament from Hebrew and from Martin Luther's Bible, thus
also involving German. (See Delisle and Woodsworth 1995.) The main point is that
22 Cay Dollerup

some original source texts of the Bible no longer exist, so that translations are based
on relay translations in languages where, like the renditions of present-day
conference interpreters, they have had audiences of their own: in Greek, in Latin,
and in German.
Today, most translation activity is concerned with texts that are not intended
to last for generations, but are accepted as transitory and short-lived, such as letters,
instructions, manuals, used in industry, in trade, and in tourism. In these contexts,
translation is carried out soon after the appearance of the original, if at all: there is
no point in translating old manuals, instructions for obsolete tools, out-of-date
tourist brochures. Most translation in the modern world is therefore comparable to
consecutive interpreting: once the message is complete in the source language, it is
followed by a translation. In principle, it is immaterial whether an interval of five
minutes or one year elapses, for the point is that we are not talking about a century
or more between the appearance of the original and its translation(s).
Literature, however, offers many examples of relay in which the original is
still extant and where translation is both relayed and delayed.
The Tales of the brothers Grimm were initially published in 1812-1815 in
German. In 1823 an English translation by Edgar Taylor of selected tales achieved
success among English audiences. This collection - and its successors in English -
also provided source texts, relays, for numerous translations into other languages
(Sutton 1996, Dollerup 1999). Similarly German translations of fairytales first
published by the Dane Hans Christian Andersen, in 1835, were made for German
audiences as early as 1837 and enjoyed great success. Such German texts served as
the source texts for the first translations of Andersen into English and later into
central and southern European languages, for instance Slovene (personal
communication by Hans Vermeer and Silvana Orel). In subsequent periods direct
translations from Danish into English served as source texts for yet other target
translations into, for instance, Chinese (see e.g. Xu 1998).
I have a reason for singling out tales: they are popular and hence often
published. Accordingly, they offer a picture in miniature of relay, nowadays
especially so in European, indeed international, co-prints. In co-printing a publisher
in one country produces a book with illustrations and uses translations from national
publishers which fit these illustrations in terms of typography, that is, of length. The
point to note is that the text provided by the original publisher may already be a
translation and thus a relay of the original. This procedure is used extensively in the
publication and translation of children's literature and of illustrated textbooks. It is -
to the best of my knowledge - largely overlooked in comparative philological
translation studies.

3.3 Further aspects of relay

The term relay highlights the dynamics in the interlingual movements of translation.
 "Relay" and "support" translations 23

The end product of translation is static, but this static translation may be used for
other dynamic translation or interpreting processes, leading to yet other static
translations (Dollerup 1988). There is, in principle, no finality to this process,
translations of the same text can continue indefinitely, not only within the same
binary language pair, but also between languages and cultures. This is why relay is a
better term than indirect translation, which implies not only that intermediary
realisations are ephemeral, but also that the translation in hand is the only end
product imaginable.
It is, furthermore, a fact of life that literary translation is not highly
remunerated, and that translators commit errors, ranging from mistakes involving
false friends to complete distortions. Of course, most readers, including reviewers,
do not notice deviations because they do not collate the books they read for fun with
the originals; indeed there is no reason why they should read a translation at all if
they are well versed in the language of the original.
No matter how we define errors or deviations, it is clear that in written
public translations, each translator using relay will normally add new deviations to
those made by predecessors in the chain. In relation to the original, there is thus a
cumulation of deviations every time a work is relayed. In this respect, then, public
written translation differs radically from public professional simultaneous conference
interpreting, for thanks to the professionalism involved in delivery in relay
interpreting, nearly all errors occur between the original speaker and the first
interpreter and not in the chain of relay. Conversely, relay is indeed a major source
of deviations in written translation.
Cultural differences account for some deviations. Others are explicable in
terms of the presence or absence of the sender: in conference interpreting, the
presence of both senders and receivers in the same room, the body language of
senders, audiences, and other contextual features constitute strong corrective
elements. (See Gottlieb 1997 for errors in relay subtitling.) In literary translation,
relay translation (as well as delay) implies that the sender, the original author,
recedes into the background. The communicational chain is not complete. Fidelity
and loyalty to the author become weakened, not out of ill will, but for practical
reasons - the translator will not always be in a position to have the author elucidate
obscure points. In other words, the use of relay in translation shows that models of
the translation process which operate by means of unbroken chains from the original
sender to the final receiver are inapplicable to all the realities of translation work.

4. 'Support translation'

Relay translation should be distinguished from what I call support translation. This
is the strategy in which, translating a given source text, translators check translations
into languages other than their own target language in order to see whether
24 Cay Dollerup

colleagues have found satisfactory solutions to certain problems - usually only to

find that it is the same passages which prove problematic to translators in related
languages. Bush (1997) mentions one example. This strategy per se is rarely
mentioned in scholarly Translation Studies, possibly because, in literary translation,
there may be problems of droit moral. Historical examples such as Coverdale's
Bible exist. At the European Commission, translators are grouped according to
topics, not to languages, and in some of these groups it logically follows that other
translations are used as supports.
It will be appreciated that the relation between support and relay is
connected with the degree of dependence on other realisations: in pure relay, the
translator uses the totality of another translator's text, whereas in regular cases of
support translation we meet isolated fragments within these wholes. However, there
is obviously an enormous area in between and one which contains all sorts of
fascinating combinations. The German dramatist Bertold Brecht's translation of
Shakespeare's Coriolanus is a case in point: "he used two editions in English as well
as a modern [German] edition of Dorothea Tieck's translation, and the prompt book
from [a German 1936 staging] as well as [the Latin historians] Livy and Plutarch"
(Ewbank 1995: 8).

5. Concluding remarks

Finally, there is that interesting phenomenon: the multilingual original. Many

European Union documents can hardly be said to have a stable core original,
unambiguously placed in the context of one language only (Dollerup 1996). Kristine
Anderson (1997) discusses the dialectics in certain works by the Danish author Isaac
Dinesen, who wrote alternatively in Danish and in English, producing startlingly
different versions of works considered the same by her publishers and audiences. Xu
Yanhong (1998) calls attention to English versions which come into existence at the
same time as the Danish originals, and discusses examples of transfer routes from
Danish originals to Chinese realisations.
There are legitimate, even durable realisations made for consumers in target
languages; in dynamic processes, such translations may be used for other
translations in chains which are, in principle, infinite; these I suggest we term relay
translations in order to stress the multiplicity of audiences. On the other hand, we
find rare cases where ephemeral realisations in intermediary languages exist only as
subordinate prerequisites for establishing one clear chain of communication between
two (parties of) interlocutors. It is only in these last cases that we can truly speak of
indirect translation.
In actual translation work, there are more dimensions to be taken into
account: the presence and availability of the senders and receivers, the time gap, the
deviations introduced by translators, by accident or design, for any motive ranging
 "Relay" and "support" translations 25

from ignorance via censorship to deliberate exclusion in order to produce better

integration into target cultures.
The situation discussed above is, it is true, complex, and my terminology at
the secondary level is open to debate. Nevertheless, this article serves to underline
the point that in order to deal with these facets of the real world, terms used in
Translation Studies should try to describe the facts as precisely as possible. This is
all the more pertinent as the terms discussed also shed additional light on the vexed
relation between the authority of the original and the realisations commonly called
translations.

References

ANDERSON, Kristine. 1997. "Karen Blixen's bilingual oeuvre: the role of her English editors".
Perspectives: Studies in Translatogy 5(2). 171-189.

BUSH, Peter. 1997. "Strawberry Flowers in Realms of Chocolate: The Training of Literary
Translators". The Changing Scene in World Languages. Issues and Challenges (ATA Scholarly
Monograph Series IX) ed. by Marian B. LABRUM, 109-117. Amsterdam & Philadelphia: John
Benjamins..

DELISLE, Jean & Judith WOODSWORTH. 1995. Translators through history. Amsterdam &
Philadelphia: John Benjamins.

DOLLERUP, Cay. 1988. "An ontological approach to translation and intranslatability". Semantik,
Kognition und Äquivalenz ed. by Gert JÄGER & Albrecht NEUBERT, 138-147. Leipzig: VEB
Verlag Enzyklopädie.

DOLLERUP, Cay. 1987. "Control of interlingual mediation in practice: Denmark as a case study".
Multilingual Journal of cross-cultural and interlanguage communication. 6(2), 169-190.

DOLLERUP, Cay. 1996. "Language work at the European Union". Translation Horizons Beyond
the Boundaries of Translation Spectrum ed. by Marilyn Gaddis ROSE, 297-314. Binghamton:
Center for Research in Translation.

DOLLERUP, Cay & CEELEN, Leo. 1996. A corpus of consecutive interpreting comprising
Danish, Dutch, English, French, German and Italian. Copenhagen: Centre for Translation Studies.

DOLLERUP, Cay. 1999. Tales and Translation: the Grimm TalesfromPan-Germanic narratives
to shared international fairytales. Amsterdam & Philadelphia: John Benjamins.

EWBANK, Inga-Stina. 1995. "Shakespeare Translation as Cultural Exchange". Shakespeare

Survey. An Annual Survey of Shakespeare Studies and Production 48 ed. by Stanley WELLS, 1-12.
Cambridge University Press.

GOTTLIEB, Henrik. 1997. "Subtitling: diagonal translation". Subtitles, Translation and Idioms ed.
by Henrik GOTTLIEB, 107-134. Copenhagen: Centre for Translation Studies.
26 Cay Dollerup

PEARL, Stephen 1995. "Lacuna, myth and shibboleth in the teaching of simultaneous
interpreting". Perspectives: Studies in Translatology 3(2), 163-190.

SHUTTLEWORTH, Mark & COWIE, Moira. 1997. Dictionary of Translation Studies.

Manchester: St. Jerome.

SNELL-HORNBY, Mary, HÖNIG, Hans, KUSSMAUL, Paul & SCHMITT, Peter A. (eds). 1998.
Handbuch Translation. Tübingen: Stauffenburg Verlag.

SUTTON, Martin. 1996. The Sin-complex. A Critical Study of English Versions of the Grimms'
Kinder- und Hausmärchen in the Nineteenth Century. Kassel: Brüder Grimm Gesellschaft.

TOURY, Gideon. 1995. Descriptive Translation Studies and beyond. Amsterdam & Philadelphia:
Benjamins.

XU Yanhong. 1998. "The routes of translation". Perspectives: Studies in Translatology 6(1), 9-22.
EL CONCEPTO DE NUMERO DIFUSO APLICADO
AL PERFIL DEL TRADUCTOR

Lourdes Pérez Gonzalez

Universidad de Oviedo, Spain

Abstract
The concept of fuzzy number as applied to the translator's profile.
This paper uses the mathematical concept of fuzzy number to quantify the
communicabihty degree of a given message (nothing, little, much, all) according to
the gap, both in time and space, existing between the transmitter and the receiver, and
the degree to which they share the same code. If communicabihty is not complete, it
can be retrieved by introducing a new element, the translator himself, who will have to
intervene to a greater or lesser degree in order to ensure a degrate communication.

Résumé
La notion de nombre flou appliquée au profil du traducteur.
Notre travail utilise la notion mathématique de nombre flou pour mesurer la
communicabilité d'un message donné (rien, peu, beaucoup, tout), en fonction de la
distance espace-temps qui existe entre l'émetteur et le récepteur et le degré avec lequel
ils partagent un même code. Cette capacité de communiquer, au cas où elle ne serait
pas totale, peut être rétablie au moyen de l'introduction d'un élément nouveau, le
traducteur, qui devra intervenir plus ou moins afin de rééquilibrer l'éventuel
déséquilibre communicatif.

Zusammmenfassung
Der Begriff der diffusen Zahl mit Anwendung auf das Profil des Übersetzers
In diesem Artikel wird der mathematische Begriff der diffusen Zahl verwendet, um
den Kommunikationsgrad einer bestimmten Mitteilung (nichts/wenig/viel/alles) im
Verhältnis zur existierenden räumlichen und zeitlichen Entfernung zwischen Emittent
und Rezeptor sowie dem Grad, in dem beide die gleichen Kode teilen, zu
quantifizieren.
Im Falle eines Mangels kann der gewünschte Kommunikationsgrad durch die
Einführung eines neuen Elements erzielt werden: durch den Übersetzer, der in mehr
oder weniger hohem Maße eingreifen muss, um eine potentielle.
28 Lourdes Pérez González

Partimos de que la traduction forma parte del proceso comunicativo; es una

parte de la comunicación, forzosamente necesaria en muchos casos, ya que no
todos los mensajes comunican de igual modo; es decir, no todos los mensajes
son comprendidos por el receptor y/o no todos los mensajes, caso de ser
comprendidos, son comprendidos del mismo modo por distintos receptores.
La comunicación, pues, implica la transmisión de un mensaje entre un
emisor y un receptor que poseen en comün, al menos parcialmente, el código
necesario para la codificación y descodificación del mensaje. Esta transmisión
no es lineal, es un todo integrado, un proceso dinâmico que exige una actitud
activa por parte del emisor y del receptor y en la comunicación asi entendida
podriamos decir que todos los mensajes con vocación de comunicación, que
aspiran a ser eficaces, tienen que tener en cuenta el tipo de receptor al que se
dirige.
Por ello, toda comunicación, incluso la más sencilla, para paliar la
"desinformación" nunca comporta un ünico lenguaje -no podemos hablar, por
tanto, de un mensaje monofónico sino de un conjunto fluido y polifónico de
numerosos lenguajes- diversidad que sirve para matizar, contradecir o
reafirmar (traducir) el sentido de los otros. Este concepto de polifonia incluye
no sólo la existencia de personajes distintos en un ünico enunciado verbal -lo
que explica fenómenos como la ironia o la presuposición- sino que también
incluye la posibilidad de que lenguajes no verbales intervengan en el acto
comunicativo.
Es decir, para que ciertos mensajes cumplan la condición de
comunicabilidad en determinadas situaciones, es preciso que exista una
traducción - en el sentido de la acepción 2 de Maria Moliner (Moliner 1983):
"expresar en forma distinta algo ya expresado"- en el mismo código o en
otro, por parte del propio emisor, del receptor o por parte de una tercera
persona, el "traductor", que "adapte" dicho mensaje.
Asi pues, tres elementos fondamentales (emisor, receptor y código) han
de intervenir en todo acto comunicativo. En efecto, si alguno de estos
elementos foera nulo no se realizaria el acto comunicativo. Pero estos tres
elementos que han de estar presentes en todo acto comunicativo no participarân
en igual medida en todos los casos.
Asi, el receptor está más o menos en la medida en que su participación
es mâs o menos activa. Un receptor presente, que puede interpelar al emisor,
podria facilitar la comunicación mâs que un receptor ausente que tiene que
limitarse a un mensaje fijo. El emisor puede estar más o menos lejos del
receptor y conocer en mayor o menor medida su contexto, que puede ser
diferente o similar. Y el código, por su parte, puede ser más o menos
compartido y mâs o menos complejo en la medida en que tenga referentes
explicitos o implicitos, en que tenga un contexto mâs o menos actual, mâs o
menos compartido culturalmente por emisor y receptor. Se requiere, pues, un
cierto equilibrio entre los elementos que constituyen la comunicación para que
 El concepto de nómero difuso 29

esta no resuite perturbada. Ahora bien, lograr este equilibrio puede conseguirse
sin intervención o puede requerir la intervención de un elemento nuevo.
Y es este nuevo elemento, esta conexión a veces ajena pero necesaria
(el Traductor), lo que restablece el equilibrio necesario para que un mensaje
comunique. Pero la necesidad del traductor será mayor o menor en función del
desequilirio existente en el acto comunicativo. Y para poder determinar esa
mayor o menor necesidad de traduction utilizaremos el concepto matemâtico
de difusividad que permitirâ analizar la falta de precisián, la indefinición, la
vaguedad con la que se presentan estas situaciones.
Este concepto de difusividad fue propuesto por L.A. Zadeh en 1965
(Zadeh 1965:338-353) como nueva via para simplificar el estudio de una serie
de problemas nacidos de la creciente aplicación de la matemática a numerosos
campos de la ciencia y de la técnica.
En la teoria clâsica todos los resultados con los que se trabaja son
exactos. Sin embargo, la experiencia nos demuestra que esto no es siempre
posible en la realidad. Ni son tan claras nuestras preferencias personales a la
hora de tomar una decision, ni son tan precisos los resultados de un
experimento.
Supongamos que queremos considerar personas altas y bajas y, para
ello, dividiremos el conjunto de las estaturas en dos subconjuntos A y B.
Diremos que una persona es alta si su estatura está en el conjunto B y
diremos que es baja si está en el conjunto A. Estos dos conjuntos son conjuntos
ordinarios. Sin embargo la frontera entre estaturas bajas y altas no es tan clara
como parecen indicar los conjuntos anteriores, pues hay una serie de valores
que tanto podrian estar en un conjunto como en otro. No obstante, aunque no
nos decidamos claramente a introducirlos en uno de los dos conjuntos si nos
aventuramos a decir que una estatura es mâs bien alta que baja o viceversa. La
introducción de este juicio subjetivo en la construcción de los conjuntos se va a
reflejar en la función de pertenencia y darâ lugar a subconjuntos difusos.
La diferencia fundamental con los conjuntos clâsicos es que mientras en
éstos los elementos pertenecen o no a ese conjunto (si o no, 0, 1, blanco,
negro), en los conjuntos difusos los elementos pueden pertenecer en mayor o
menor medida al subconjunto difuso (entre 0 y 1 hay infinitas posibilidades de
pertenencia), pertenencia que se expresa mediante una función que permite
evaluar (cuantificar) la presencia de un elemento dado en dicho subconjunto, lo
que esta mâs de acuerdo con la mayoria de las situaciones de la vida real y de
las aplicaciones cientificas y técnicas, donde las clasificaciones excluyentes no
ofrecen un modelo adecuado de la mayor parte de los fenómenos en estudio.
El numero difuso permite una semântica del tipo "aproximadamente
entre a y b" por lo que va a permitir modelizar la capacidad comunicadora de
los mensajes, representada ésta por el numero difuso obtenido al multiplicar los
numeros difusos que describen, respectivamente, la posición relativa entre
emisor y receptor y el grado de compartición mutua del código utilizado. Como
30 Lourdes Pérez Gonzalez

una consecuencia del resultado obtenido para el grado de comunicabilidad,

podremos concluir también en términos de numeros difusos la necesidad o no
de la presencia del traductor.
La utilizatión del concepto de numero difuso permite infmitas
posibilidades pero, para simplificar, utilizaremos una cuâdrupla (a,b,a,ß )

traducible por nada, poco, mucho y todo

que se cuantificarian como sigue:

Nada Valores iguales a 0

Poco Valores menores de 0.5
Mucho Valores mayores de 0,5
Todo Valores iguales a 1

Asi pues, aplicando el concepto de dirusividad a los tres elementos:

emisor, receptor y código, obtenemos cuatro posibilidades, que en el caso del
emisor y del receptor se traducirian por presente, inexistente, lejano y cercano
(en el espacio y/o en el tiempo), y en el caso del código por totalmente
compartido, nada compartido, muy compartido y poco compartido.
La difusividad de los elementos subjetivos podria definirse como la
presencia real o virtual de los mismos. Esta presencia real o virtual no es sólo
fïsica, ya que al incorporar los factores espacio y/o tiempo, abarca un concepto
mucho más amplio, que es el de contexto. Asi un mensaje carente de
connotaciones tipo "hoy llueve", "yo soy una mujer", "tengo hambre" acercaria
mâs que un mensaje con carga contextual, tipo: "Pero el dos no ha sido nunca
un numero / porque es una angustia y su sombra, I porque es la guitarra donde
el amor se desespera / porque es la demostración de otro infinito que no es
suyo..." (Lorca 1971:530)
En definitiva, para modelizar los distintos tipos de mensaje contamos
con dos elementos:

1.- Emisor/receptor al que reconocemos cuatro posibles estados cuantificados

por los numeros difusos:
Inexistentes 0
Lejanos menores 0,5
Cercanos mayores 0,5
Presentes 1
 El concepto de numero difuso 31

2,- Código, al que también reconocemos cuatro posibles estados cuantificados

por los mismos numeros difusos:
Nada compartido 0
Poco compartido menores 0,5
Muy compartido mayores 0,5
Totalmente compartido 1

Las cuatro posibilidades del código: nada, poco, mucho, todo compartido, no
presentan dificultades de comprensión, ya que de lo que se trata es, en
definitiva, de los grados de "dominio" de una lengua o de un lenguaje, que un
receptor dado poseerá en mayor o menor medida.
Pero quizâ debamos insistir un poco más en las posibilidades de
emisor/receptor que cuantificâbamos como: presente, cercano, lejano e
inexistente, ya que incorpora las variables "en el tiempo y/o el espacio" y que
influirâ en el grado de comunicabilidad de los mensajes, similar a los niveles
de comprensión del mensaje definidos por Strawson (Strawson 1970) y que de
menor a mayor son: primer nivel (significacion lingüistica), segundo nivel
(significación referencial) y tercer nivel (significación completa).
Emisor/receptor presentes: se trata de presencia fïsica -que permite la
interpelación- por tanto de un código oral, siendo los contextos que movilizan
ambos netamente iguales. También podria tratarse de un texto escrito cuyo
contexto compartieran ambos.
Emisor y receptor cercanos: no existe presencia fïsica, ni posibilidad de
interpelación, se trata, pues, de un código escrito. Los contextos que movilizan
son diferentes - distancia espacio/temporal pequena - pero no sustancialmente
diferentes. Contextos culturales de paises cercanos - diferentes, pero similares-
o contextos históricos de momentos cercanos.
Emisor y receptor lejanos: no existe presencia fïsica ni posibilidad de
interpelación; se trata, también, de un código escrito. Los contextos que
movilizan son sustancialmente diferentes - distancia espacio/temporal grande.
Contextos culturales de paises remotos o contextos históricos de momentos
lejanos.
Emisor y receptor inexistentes. Se trata de casos inviables ya que al no
existir emisor/receptor no existe posibilidad de emisión de mensaje.
Las posibles combinaciones (16) nos va a permitir representar cuánto
comunica cada uno de los mensajes mediante el numero difuso producto de los
numeros que cuantifican los estados de las variables código y emisor/receptor
(Bonissone 1982:332). De modo que, en forma de cuadro, quedaria del
siguiente modo:
32 Lourdes Pérez González

Tabla Producto: Comunicabilidad De Los Mensajes

Emisor/ código
Receptor a b C d
A 1 toda <0,5 mucha >0,5 poca 0 nula
B <0,5 mucha <0,5 mucha >0,5 poca 0 nula
C >0,5 poca >0,5 poca >0,5 poca 0 nula
D 0 nula 0 nula 0 nula 0 nula

Antes de entrar en el detalle de este cuadro, hemos de recordar que hay una
serie de casos inviables: aquellos en los que la variable emisor/receptor tiene el
valor cero, es decir uno de ellos o ambos no existen, ya que en la medida en
que no hay quien emita es imposible que haya mensaje, pero a efectos de una
mejor "visualización" seguiremos incorporândolos a los cuadros. Como
podemos apreciar, solo en un caso (emisor y receptor presentes, código
totalmente compartido) el mensaje comunica totalmente, en el resto de las
situaciones hay un desequilibrio, falta "algo" en mayor o menor medida para
que se logre la comunicación. Y, para compensar ese desequilibrio hemos de
introducir el concepto matemâtico de complementario respecto de la unidad,
que se obtiene mediante la diferencia entre los numeros difusos (Bonissone
1982:332) y que, en forma de cuadro, quedaria del siguiente modo:

Complementario: Necesidad De Traductión

Emisor/ código
receptor a b C d
A 0 nula >0,5 poca <0,5 mucha 1 toda
B >0,5 poca >0,5 poca <0,5 mucha 1 toda
C <0,5 mucha <0,5 mucha <0,5 mucha 1 toda
D 1 toda 1 toda 1 toda 1 toda

Este complementario, que equivale a "lo que falta" para que exista un
equilibrio entre los elementos que constituyen el acto comunicativo, cuantifica
la necesidad (mayor, menor o total) de descodificación-recodificación y de
acercamiento entre emisor y receptor - de traducción - para cada tipo de
mensaje aqui analizado. Analizaremos los cuatro casos que se nos presentan.
En el primer caso los valores que cuantifican los conjuntos
emisor/receptor y código son mâximos, es decir se trataría de un código
totalmente compartido y de la presencia de emisor y receptor.
Seria una comunicación nitida, transparente.
Puede tratarse de mensajes elementales, esencialmente informativos,
que no requieren ninguna competencia especial por parte del receptor, porque
 El concepto de número difuso 33

los contextos del emisor y del receptor estân reducidos a cero. Imaginemos un
mensaje tipo: "esto es una mesa" o "son las tres y media".
Pero también puede tratarse de contextos activos pero iguales,
imaginemos una fórmula matemâtica comunicada entre dos expertos de igual
rango. En cualquiera de los casos, se trata de situaciones comunicativas nitidas,
que no requieren en absoluto traducción, ya que, como vimos, el
complementario es cero.
El segundo caso es la otra situación extrema, la contraria de la anterior,
es decir aquella en la que todos los elementos son cero. Situación en la que la
comunicación no es posible, ya que el grado de comunicabilidad es nulo, no
solo porque el código no es compartido en absoluto, sino porque emisor y
receptor son inexistentes.
En el tercer caso hay que incorporarle algo para que cumpla su función,
tiene que intervenir un elemento en principio ajeno pero absolutamente
necesario, (el Traductor), para que la comunicación exista. Este elemento
servirá para "compensar" la inicial no comunicabilidad, el inicial desequilibrio
existente entre los elementos de la comunicación. Existen varios tipos dentro
de este tercer caso según la mayor o menor necesidad de traducción requerida
por ciertos mensajes para que comuniquen. Y esta mayor o menor necesidad de
traductor, que estarâ más o menos cerca de los contextos del emisor y/o del
receptor y que tendrá que recodificar total o parcialmente es la que determinará
los distintos tipos de "ayuda" a la comunicación.

PRIMER TIPO
Que identificamos con los casos en que la necesidad de traducción es poca,
representados cuantitativamente por >0,5

Emisor/ código
receptor a b c d
A >0,5 poca
B >0,5 poca >0,5 poca
C
D

En los dos primeros casos se trata de mensajes emitidos en códigos muy

compartidos, aunque la presencia de emisor/receptor es variable.
Imaginemos una situación para cada uno de ellos:
- Emisor y receptor presentes, código muy compartido.
La presencia de emisor y receptor hace posible no sólo la interpelación
sino la movilización de los mismos recursos contextuales. Se trata del caso-tipo
y único de la auto-traducción. El emisor, que comparte contexto con el
receptor, emite su mensaje y simultâneamente, porque la presencia permite la
34 Lourdes Pérez Gonzalez

interpelación, traduce aquellas partes que "sabe" pueden ser malinterpretadas o

incomprendidas o sobre las que el receptor pregunta para mayor claridad.
En este caso, es el propio emisor el que ejerce de traductor.
- Emisor y receptor cercanos código muy compartido.
En este caso existe una pequeña distancia (en el espacio y/o en el
tiempo) entre emisor y receptor, que no permite la interpelación. Sus contextos
no son idénticos, aunque comparten el código en alto grado. Imaginemos un
mensaje emitido en una lengua compartida pero en un lenguaje de otra época
(no muy lejana, por ejemplo argot juvenil de hace diez años) o cuyo contenido
se refiera a situaciones "parcialmente" desconocidas (por ejemplo editorial de
actualidad de periódico sudamericano).
Traductor, el propio receptor, mediante analogias.
- Emisor y receptor cercanos, código totalmente compartido.
Este caso que, como se puede apreciar por el cuadro, pertenece a una
columna distinta que los anteriores, tiene la peculiaridad de que se trata de un
código totalmente compartido, a pesar de que emisor y receptor no estân
presentes. Imaginemos el ejemplo anterior pero recibido por un experto en ese
lenguaje o en esa realidad. El compartir el código totalmente permite que el
receptor - que posee las claves suficientes para comprender el texto en su
totalidad - supla al traductor.
Traductor, el propio receptor, con sus propios recursos.

SEGUNDO TIPO
Que identificamos con las situaciones en las que el grado de comunicabilidad
es pequeno por tanto la necesidad de traducción es grande

Emisor/ código
receptor a b c d
A <0,5 mucha
B <0,5 mucha
C <0,5 mucha <0,5 mucha <0,5 mucha
D

En los tres primeros casos se trata de las mismas situaciones que en el epigrafe
anterior, incluyendo el caso de emisor/receptor lejanos, pero el código está
poco compartido, por tanto la necesidad de traducción es mayor. Imaginemos
las situaciones anteriormente mencionadas pero con mensajes emitidos en una
lengua solo parcialmente conocida por el receptor. Evidentemente se impone la
necesidad de traducción aunque, al no ser esta necesidad total, la traducción
puede ser ejercida por un traductor no experto. Imaginemos un mensaje
informativo emitido en un idioma que el destinatario sólo conoce parcialmente,
éste no captarâ el mensaje en su totalidad, pero la intervención de algün
espontáneo - en el caso de emisor/receptor presentes - que conozca la lengua
 El concepto de número difuso 35

del receptor aunque no la domine, puede lograr la comunicabilidad del

mensaje.
Traductor: tercera persona no experta
- Emisor y receptor lejanos código muy compartido.
En este caso, la distancia que separa el emisor y el receptor es grande.
Esto significa que sus contextos son diferentes. Se trataria del caso de un
mensaje emitido en una lengua común pero en un lenguaje no compartido, por
ejemplo un lenguaje técnico. En este caso la traduction podria proceder bien de
una tercera persona o de un código anejo, por ejemplo un código grâfico.
Imaginemos un manual de funcionamiento de un aparato (solo texto) y un
receptor profano... los esquemas y dibujos le serân absolutamente necesarios a
la hora de hacer las posibles conexiones.
Traductor: tercera persona o el propio receptor mediante códigos
anejos.
- Emisor y receptor lejanos, código totalmente compartido.
En cuanto a este ultimo caso, sucede lo mismo que con el último caso
del epigrafe anterior, pero aqui la distancia es mayor, los contextos estân más
alejados; el receptor, dada la dificultad del texto tendrá que acudir a alguna
fuente o consulta para autotraducir integramente el mensaje.
Traductor: tercera persona o propio receptor con recursos externos.

TERCER TIPO

Emisor/ Código
receptor a b c d
A 1 toda
B 1 toda
C 1 toda
D

En este caso, la existencia de un código nada compartido exige la presencia

nitida de un traductor - que ha de ser experto - para que se cumpla el objetivo
de la comunicación. Aqui no hay autotraducción posible ni posibilidad de
comunicabilidad solo mediante códigos anejos, ya que el mensaje requiere una
recodificación total y no parcial como en los casos anteriores.
Traductor: tercera persona experta.
Esta clasificación de los mensajes en fimción de su mayor o menor
grado de comunicabilidad y el complemento que, en cada caso, ha de aportar el
traductor determina las competencias que tiene que poner en funcionamiento el
traductor en cada desequilibrio comunicativo. Por otra parte, esta clasificación
le será util al traductor para valorar la cercania/lejania entre el emisor y el
receptor y poder elegir "su" lugar de acción y "situarse" en el punto que
36 Lourdes Pérez Gonzalez

considere mâs adecuado a efectos de una correcta comunicabilidad del

mensaje.
En conclusión, la traducción es una actividad comunicativa en la que
intervienen factores linguisticos y extralingüisticos cuyo objetivo es la
comunicabilidad de un mensaje dado. La mayor o menor necesidad de
traducción condiciona que el traductor pueda ser el propio emisor, el receptor,
un traductor no experto o un traductor experto. Esto genera una distinción entre
traducciones propiamente dichas (en las que el traductor aporta el grado de
traducción necesario para que se logre la comunicabilidad) y
pseudotraducciones (en las que el grado de traducción aportado es inferior al
necesario). Y la noción de intraducible pierde consistencia en la medida en que
consideramos la relación emisor - traductor - receptor como dinâmica y
reciproca.
La cuantificación de la necesidad de traducción mediante el concepto de
numero difuso permite delimitar los "bordes" de la actividad traductora, situar
las competencias que se requieren para cada tipo de mensaje y establecer las
fronteras entre traducciones y pseudo-traducciones.

Referendas bibliograficas

BELLMAN, R & ZADEH, L.A. 1970 "Decision making in a fuzzy environment".

Management Science 17. 141-164

BONISSONE, P.P. 1982 "A fuzzy sets based linguistic approach: Theory and applications"
Approximate Reasoning in Decisions Analysis. 329-339. Amsterdam, New York, Oxford:
North-Holland Publishing Company

GARCIA LORCA, F. 1971 Obras completas. Madrid: Aguilar.

MOLINER, M. 1983. Diccionario de uso del español Madrid : Gredos.

STRAWSON, P. 1970 "Phrase et acte de parole". Langages 17.

ZADEH, L.A. 1965 "Fuzzy sets Informa". Control 8.338-353

 PART II

Situational, sociological and political factors

THE WORKPLACE PROCEDURES OF
PROFESSIONAL TRANSLATORS

Brian Mossop
Government of Canada Translation Bureau and York University
School of Translation, Canada

Zusammenfassung
Eine Untersuchung zur Klärung der von Übersetzen bei ihrer Arbeit verwendeten
Prozeduren, wird vorgeschlagen. Wie ist die Abfolge der unterschiedlichen Aufgaben
beim Übersetzen (Interpretieren, Konstruieren, Recherchieren, Kontrollieren)?
Welche Kontroll- und Recherchiermethoden finden Anwendung? Sind die Prozeduren
anders bei einem kurzem, gutgeschriebenen Text als bei einem langen, mangelhaft
verfaßten Text? In diesem Artikel werden Schwächen des ÜberSetzungsunterrichts
wegen Nicht-Vorhandenseins einer Untersuchung dieser Art angesprochen. Behandelt
werden auch die möglichen praktischen und theoretischen Vorteile der Untersuchung
neben der interessanten Frage, warum keine Untersuchungen über diese Prozeduren
vorliegen.

Résumé
Nous proposons une enquête destinée à mettre à jour la méthode de travail employée
par différents traducteurs en milieu de travail. Dans quel ordre les différentes tâches
(interprétation, rédaction, recherche, vérification) sont-elles executées? Quelles sont
les méthodes de recherche et de vérification employées? La méthode varie-t-elle selon
que le texte à traduire est court ou long, bien ou mal rédigé? Nous indiquons les
faiblesses de tout conseil prodigué aux apprenants en l'absence d'une telle enquête,
les avantages pratiques et théoriques qu'on pourrait tirer de l'enquête, et les raisons
pour lesquelles la méthode de travail n'a pas été étudiée de façon systématique.

Resumen
Se propone la realización de un estudio para revelar los procedimientos seguidos por
varios traductores en el lugar de trabajo. Se investiga el orden seguido en la
realización de las distintas etapas (interpretación, composición, investigación,
revisión), cuáles son los métodos utilizados en la revisión e investigación y si se sigue
el mismo proceso para un texto corto bien escrito que para un texto largo mal escrito.
40 Brian Mossop

Se estudian los fallos que la ausencia de un estudio semejante puede producir en los
consejos ofrecidos a los principiantes. También se investigan los beneficios teóricos y
prâcticos que este trabajo puede aportar a la teoria y la práctica de la traducción, asi
como las razones por las que nunca se han estudiado estos procedimientos.

1. A Research Proposal

There is a curious gap in our knowledge about translation. We have no

systematic observations, or even self-descriptions, of how professional
translators proceed when they translate, that is, the order in which they perform
the various operations involved in translating. What we do have, in
pedagogical writings, are all sorts of procedural prescriptions. But in the
absence of large-scale observations, what justifies these prescriptions?
The production of a translation can be described in terms of three
phases and five tasks:
Three phases of translation production
(1) pre-drafting (before sentence-by-sentence drafting begins)
(2) drafting
(3) post-drafting (after sentence-by-sentence drafting is complete)

Five tasks to be performed

(1) Interpret the source text.
(2) Compose the translation.
(3) Conduct the research needed for tasks 1 and 2.
(4) Check the draft translation for errors and correct if necessary.
(5) Decide the implications of the commission: how do the intended
users and uses of the finished product affect tasks 1 to 4?

What we do not know, and need to know, is how translators distribute the
tasks over the phases. Each task can be performed during any phase (with the
exception of task 4, which obviously cannot occur during phase 1). One can try
to interpret a passage (task 1) either before or during drafting, or one can leave
a blank in the draft and interpret during the post-drafting phase. Similarly with
composing (task 2): one can jot down possible translations during pre-drafting;
or one can leave blanks during drafting and fill them in during the
post-drafting phase. Thus phases and tasks are not to be confused: checking
(task 4) is not a synonym of post-drafting (phase 3); composing (task 2) is not
a synonym of drafting (phase 2). 1
The purpose of this paper is to suggest why it would be worth carrying
out a large-scale study of work procedures in terms of tasks and phases. The
proposed study would consist of:
(1) systematically observing the procedures used by a large number of
 The workplace procedures 41

translators in the workplace; and perhaps:

(2) interviewing the translators or having them fill out questionnaires;
and also:
(3) determining the quality of the translators' output; and
(4) timing the three phases, in order to correlate quality with speed.

A provisional definition of quality would be 'whatever constitutes acceptable

quality in the market for which a particular translation is prepared'.
Appendix 1 gives an idea of what I mean by observations. The two
examples are purely illustrative, not actual observations of individual
translators. They are based partly on my own experience and partly on
information I have gathered during professional development workshops. The
words in parentheses refer to unobservables; this type of information could be
gathered during interviews, though it must be borne in mind that self-
descriptions and explanations of one's own behaviour are not always reliable.
The two sets of observations in Appendix 1 reveal different distributions
of the research task over the three phases. In Case A, there is no pre-drafting
research, most of the research is done during drafting, and then there is a small
amount of post-drafting research. In Case B, there is considerable research
during all three phases.
With regard to the checking/correcting task, there is a small amount of
correcting during the drafting phase in Case A, and a considerable amount in
Case B, but the main difference has to do with the particular way of
accomplishing this task, rather than its timing. In A, there is a single
read-through to check for errors, plus a paragraph count. In B, there are three
read-throughs, each having a different purpose. A more ambitious study would
gather this sort of information in addition to information on the temporal
distribution of the tasks.
Appendix 1 should make it clear that the proposed study is very different
from think-aloud protocol (TAP) studies. First, TAP researchers use as data
their subjects' verbalizations during translation production, along with their
own observations of the subject; in the proposed study, the researchers use as
data their own observations, and then may or may not supplement these with
the post-production comments of subjects. Second, the immediate objective of
TAP researchers is to identify translators' thought processes (for example:
their decision-making criteria; their self-concept of what they are doing as
translators); the objective of the proposed study is to correlate the sequence of
workplace actions with quality and speed. Third, TAP researchers often report
microlevel details (for example: the subject consulted a unilingual TL
dictionary on arriving at the 4th word of the 3 rd sentence of the 2 nd paragraph);
the proposed study would concern itself solely with the macrolevel (for
example: the translator consulted a unilingual TL dictionary only during the
post-drafting phase). Finally, in TAP research, the subjects are not always
42 Brian Mossop

experienced professionals engaged in translation projects for clients; in the

proposed study, all subjects would be professionals so engaged.

2. Practical benefits of the study

The study might reveal - though I think this highly unlikely - that one
particular way of distributing the tasks over the phases, or one way of
organizing the research or revision work, makes it possible to produce higher
quality work within a given amount of time, or to complete jobs faster at a
given quality level, or even to achieve both higher quality and greater speed.
Of course, the procedure in question would not necessarily be adoptable by all
translators, since it may not be compatible with their personal psychologies.
A more likely outcome of the study would be that there are two or three
'best' procedures, all of which lead to similar results with respect to quality
and speed. And even more likely, I would guess, is that the best procedure will
vary with the type of project: short or long text, urgent request or not, familiar
or unfamiliar topic, well written or poorly written source text, translation for
publication or for information, and so on. It may become apparent early on in
the study that valid comparisons among translators can only be made for texts
having a given set of features (for example: short, familiar topic, poorly
written, urgent request).
It is of course also possible that almost any work procedure can yield
acceptable results. In other words, the secret of producing good translations
quickly may lie elsewhere - not in procedure at all.
It is unclear whether there would be any practical benefits for translation
students since the procedures used by experienced professionals may not be
conducive to learning translational skills. A simple example: professionals will
use either a unilingual or a bilingual dictionary to obtain information,
whichever is fastest, whereas students should be encouraged to use a unilingual
dictionary simply because they tend to be overreliant on bilingual ones.

3. Weakness of advice in the absence of the proposed study

In the absence of an empirical basis for advising professionals that this or that
procedure is better, advice given during professional development workshops,
or in the workplace to junior translators, will inevitably be based on some
combination of received prescriptions, experience and logic.
Professional development trainers use 'received prescriptions' when they
pass on concepts of how one ought to proceed - concepts remembered from
translation school or gleaned from books they have read or lectures they have
heard (for example: "start by reading the source text in its entirety, underlining
 The workplace procedures 43

problem expressions"). The problem with prescriptions is that their

relationship to anyone's actual practice is unclear, or they may be suitable for
students but not for professionals.
Trainers use 'experience' when they advise others to use whatever
procedures they themselves use. Or, if they are aware of the procedures used
by their colleagues, they may recommend trying out a range of possibilities.
Experience, while superior to prescriptions, is problematic because, in the
absence of a study of the kind proposed here, there is no way of knowing
whether the prescribed procedure actually yields better or faster translations.
One may have an intuition that a certain procedure is better, but notoriously,
intuitions can be completely wrong.
A step above prescription-based and experience-based advice is advice
based on logic, but it too has weaknesses. Here is an example of logic-based
advice on how to check the draft for errors:

(1) The reader of the translation will be getting the meaning from the
translation, not from the source text. Therefore, during comparative checking,
read a sentence of the translation first, then the corresponding passage of the
source text in order to emulate the reader's experience.

The idea here is that if you read the source text first, you will already have the
intended meaning in your mind, and you may then erroneously read this
meaning into your translation. This certainly seems logical, but a study of the
kind advocated here might well reveal that the order of reading translation and
source text is completely unrelated to final quality.
Here are some more examples of logical advice about checking:

(2) It is hard to focus simultaneously on small units (words, phrases) and on

large units (the flow of an argument or the flow of language in a paragraph).
Therefore read the translation once for detail (comparison of source and
translation) and once for flow (translation only, without reference to source).

(3) Changes to the draft take time and there is a risk of introducing error while
making the changes. Therefore minimize changes.

(4) Making corrections on a printout and then inputting them is

time-consuming. Therefore check and correct on screen.

The results of the advocated study may well prove all this advice wholly or
partly invalid. With regard to (2), it may turn out that with, say, half the
translators observed, one read-through was just as good as two: the extra time
required for two read-throughs did not lead to higher quality. With regard to
(3), it could turn out that this advice is valid, but only for a quarter of the
observed subjects, who do tend to introduce errors and would be best off
44 Brian Mossop

staying with their original drafts; others improve their translations significantly
by making large numbers of changes. With regard to (4), the study might show
that the extra time spent checking a printout rather than checking on-screen is
well worth the while because far more errors are caught; the best procedure
might prove to be checking on a printout but then saving time by inputting any
changes directly on screen rather than handwriting them first on the printout.
Of course, time permitting, an individual translator could conduct an
empirical study on herself in order to validate such logical advice. She could
save a batch of translations checked on screen and another batch checked on a
printout. If possible, she would allow for comparability between the batches in
terms of topic, urgency, length and writing quality. Then, after a few months,
she or a colleague would go over a sample of each of the two batches of
translations to see whether more errors were missed when checking was done
on screen. Such individual diagnoses would probably be necessary even if the
advocated large-scale study were to be carried out. For it is unlikely that, on all
matters for which advice is required, the study would point to one particular
approach (say, checking on a printout) as the best one for all translators.

4. Theoretical benefits of the study

An important goal of translation theory is to explain why translations are the

way they are. Presumably, one determining factor is the production procedure.
For example, one set of translations might be more idiomatic than another, in
part because the translator did a separate check for this feature, rather than a
combined check for idiomaticity and accuracy.
The results of the study could also be used to check a variety of
hypotheses.
Hypothesis 1 A translator's procedure reflects his or her general
writing habits. Daniel Chandler (1993) reports four basic writing strategies
differentiated by the approach taken to planning and revising. Table 2 is my
summary of his findings.

Table 2: Four Writing Strategies (based on Chandler 1993)

task/phase: planning/pre-drafting revising/drafting or

\ writing strategy post-drafting
architect major minimal, after drafting
bricklayer major major, during drafting
watercolourist minimal minimal, during drafting
[ oil painter minimal major, during & after drafting j
 The workplace procedures 45

A corollary would be that a translator's procedure reflects the writing methods

taught in the particular national school system in which he was educated.
Hypothesis 2: Translators with several years' experience are methodical in the
sense that they follow a certain procedure for a certain sort of text: one
procedure for texts on a familiar subject and another for those on an unfamiliar
subject; one procedure for long texts and one for short; one for urgent and
another for non-urgent translation requests; one for poorly written and one for
well written source texts. The familiar/unfamiliar distinction might prove to be
of special interest because with unfamiliar subject matter, the translator is
more likely to be operating in a highly conscious, problem-solving mode, and
this may make the procedure quite different from what will be found when
familiar subject matter is being translated with little conscious effort. It would
be interesting to discover inconsistency; for example, long and well-written,
non-urgent texts on unfamiliar subjects sometimes dealt with one way,
sometimes another.
Hypothesis 3: Translators' procedures are logical in that they correspond to the
sort of logic-based advice discussed in the previous section.
Hypothesis 4: Organizational environment determines procedure with respect
to research and checking/correcting, in the sense that translators who work in a
large bureaucracy (or perhaps, began their careers in a large bureaucracy)
follow different procedures from those who work alone. Because of the
division of labour in large bureaucracies, it is hypothesized, a translator's
procedure will be affected by the need to coordinate with in-house
terminologists and librarians for research, and with quality controllers and
secretarial support staff for checking and correcting. It would be interesting if
the study showed that during the time when the text is in the hands of the
translator (rather than another type of employee), research and checking
procedure differs little from that of a freelance working alone.

5. Why have translators' work procedures not been investigated in

Translation Studies?

A study of the sort advocated here would be very expensive. It would take
many researchers and much time to agree on the study methods, find subjects,
make the observations and analyse the results. For a study aimed at universal
validity, there would also be problems of international coordination. The best
starting point would perhaps be a small-scale study of work procedures in a
single large translating organization. This would make it possible to verify the
research methodology and it would eliminate the problem of trying to compare
subjects who are working to different quality standards.
However I do not think the main obstacle to the proposed study is
financial or logistical. Rather the main obstacle lies in the discipline of
46 Brian Mossop

Translation Studies (TS) itself. There are at least four features of TS which
discourage observation of translators' work procedures.
First, with the exception of think-aloud studies, TS is still strongly
oriented to the study of product rather than process. The translator is still not at
the centre of TS, even though some progress is being made in this regard.
Second, work in the area of translation pedagogy is almost entirely
oriented toward the teaching of students. Professional development - a regular
theme at congresses of practising translators - is practically never discussed at
TS conferences. As a result, the idea of observing professionals at work in
order to help them improve has not arisen within TS.
Third, the practitioners of TS are almost all university-based, but the
idea of studying workplace procedures would most likely seem attractive to a
practising professional. Unfortunately, few of us are active in TS.
Finally, TS remains strongly oriented to its source disciplines -
linguistics and literary/philosophical studies. The proposed study does not fall
within the scope of these fields. As a study of workplace procedures, it is most
closely related to the field variously known as methods engineering, work
design, and work measurement. Approaches that have been developed in these
areas might well be worth considering when designing the proposed study.

6. Conclusion: A possible misuse of the study

My reference to methods engineering might make some readers worry about

the uses to which the proposed study would be put. Methods engineering is an
applied field heavily oriented toward improving productivity (and profit!) in
offices and factories through standardization. It has not always been marked by
the greatest regard for workers' autonomy, dignity and job satisfaction. The
proposed study might well find itself being funded by agencies hoping to
increase productivity/profit through standardized procedures. Such a hope
would I think be misguided. For translation would seem to be an occupation
where standardization is not applicable to products or to work procedures but
rather to skills (Mintzberg 1979:5-6).
Translators are obviously not engaged in producing standard products:
each text is unique, and each project is unique because one of a range of
possible commissions is given to the translator, which he or she applies to that
unique text. What can be standardized are the skills translators apply, such as
the skill of quickly extracting pertinent information from a dictionary or the
skill of using Internet search engines to find documentation on a subject. These
skills (for carrying out the five tasks of interpreting, composing, researching,
checking and applying commissions) are either acquired on the job or taught at
translation schools. As contacts among instructors and schools grow, the skills
are slowly being standardized.
 The workplace procedures 47

But what about the procedures, in the sense discussed in this article?
The idea of a standardized translation procedure is not nonsensical. However a
highly likely outcome of the proposed study is that while certain procedures
are indeed inefficient in terms of quality/speed, and should perhaps be
discouraged, the range of efficient procedures is quite large, so that there
would be no point in proclaiming one of them as a standard.2 Better to leave
the choice to individual translators.

Notes

1. Sager (1994:151-242) presents four phases (he treats interpretation of the commission as a
phase preceding the pre-drafting phase), each phase being divided into several stages,
such as the 'reading comprehension' stage and the 'dictionary look-up' stage (15 stages in
all). However he makes no distinction between a temporal phase and a task.

2. The German Standards Institute's proposed standard for the conduct of translation projects
(Deutsches Institut für Normung 1996) contains no specifications for procedures in the
sense discussed in this article. Most of the 12-page document is devoted to matters like
preparing the contract and billing. There is a very brief discussion of checking, but only to
list the things that must be checked (completeness etc.), not the procedure for doing so.

Appendix 1: Sample Procedures

Note: The italicized phrases started reading and when draft completed mark the beginning of
the drafting and post-drafting phases respectively. Words in parentheses refer to
unobservables, which might be elicited through interviews.

Case A
1 Glanced very briefly through the ST (found topic was familiar, writing quality poor).
2 Looked at commission: 1000 words by 4 hours from now, for one person's information.
3 Started reading the ST and composing sentence-by-sentence, fixing up poor writing
while translating.
4 Did only a little research (when the concept was unclear). Instead left blanks, or
guessed and inserted ?-marks.
5 (Where the concept was clear but no TL word came to mind), sometimes briefly
consulted a bilingual wordbook or thesaurus.
6 Consulted an electronic database for abbreviations, as encountered.
7 (When a passage clarified a previous passage), went back to fill in blank or correct.
8 Rarely edited a sentence after composing.
9 When draft completed, consulted an SL colleague about some of the ?-marks and made
appropriate changes.
10 Called the author (about the meaning of remaining ?-marked passages) - author did not
call back - left ?-marks in the translation.
11 Did a full proofreading of the TT on screen for gross language errors but made no style
corrections. Looked at the ST on two occasions (when the draft seemed nonsensical)
and reworked the relevant passage of the translation.
12 Ran electronic spell check.
48 Brian Mossop

13 Made a printout and checked it for formatting errors. Counted paragraphs in ST and TT
(to make sure none omitted).

Case B
1 Paged through ST (found topic unfamiliar, but text well written).
2 Looked at commission: 9,000 words for 1 week from now; translation to be published
and used by specialized readership for many years; author unavailable.
3 Scanned the text (looking for keywords) and searched an electronic term bank for them.
4 Searched the Internet for parallel TL texts.
5 Searched an in-house library for parallel TL texts.
6 Read passages of the parallel texts and wrote a few notes in the margins of the ST.
7 Phoned to clarify the commission: there would be a subject-matter editor.
8 Started reading the ST and composing sentence-by-sentence.
9 (Where concept unclear), stopped to do research, but occasionally aborted the research
effort and inserted a ?-mark.
10 (Where concept clear but no TL word came to mind), often consulted a thesaurus or
collocation dictionary or bilingual wordbook.
11 Phoned for a copy of the intertext (because part of the ST proved to be a detailed
response to another text).
12 (When a passage clarified a previous passage), went back to fill in blank or correct.
13 Frequently edited a sentence after composing.
14 When draft completed, went to a university library to read some items in the article's
bibliography, and made changes in the translation accordingly.
15 Telephoned a TL subject-matter expert with questions and made appropriate changes.
16 Ran electronic spell check (to eliminate distracting typographical errors).
17 Printed out a copy, and did a full comparative re-reading for accuracy and
completeness, handwriting changes on the printout.
18 Input all changes, made a fresh printout and did a full language & style editing.
19 Separately checked transcription of numbers (because important to message).
20 Revised the title (in light of the actual content of the text).
21 Created the table of contents.
22 Ran a final spell check.
23 Made a printout and checked it for formatting errors.

References

CHANDLER, Daniel. 1993. "Writing Strategies and Writers' Tools". English Today 9(2), 32-
38.

Deutsches Institut für Normung. 1996. DIN 2345 - Übersetzungsvorhaben.

MINTZBERG, Henry. 1979. The Structuring of Organizations. Englewood Cliffs, NJ:

Prentice Hall.

SAGER, Juan. 1994. Language Engineering and Translation: consequences of automation.

Amsterdam & Philadelphia: Benjamins.
KOOPERATIVE TEXTGESTALTUNG IM
TRANSLATORISCHEN HANDLUNGSRAHMEN

Hanna Risku & Roland Freihoff

Danube University of Krems and University of Vienna, Austria &
University of Tampere, Finland

Abstract
Today's complex translation projects require the co-operation of experts from
different fields. In this paper, we use the approaches of constructivism and situated-
ness to develop a theoretical basis for understanding co-operation in translation. We
apply this framework to authentic examples of co-operation between a freelance
translator and his co-operation partners. The two focal points of the analysis are as
follows:
1. Co-operation partners among such parties as the initiator of the process, the
commissioner, designers and developers of graphics or text, the external providers of
information, and the user or recipient, and
2. Fields of co-operation such as co-ordination, planning, developing, and evaluating.

Résumé
La complexité des projets de traduction actuels exige une coopération entre
spécialistes de différents domaines. Nous référerons au constructivisme et à l'analyse
situationnelle pour développer une base théorique utile à la compréhension de cette
coopération. Un tel cadre sera appliqué ensuite à des exemples réels de coopération
entre un traducteur indépendant et ses partenaires experts. Deux points surtout vont
retenir notre attention:
1. Les partenaires de la coopération, comme par exemple l'initiateur de la traduction,
le donneur d'ouvrage, les éditeurs chargés de la mise en page, les fournisseurs
d'information, le récepteur ou usager du document traduit;
2. Les moments de coopération, comme la coordination, la planification, le
développement et l'évaluation.

Resumen
En la actualidad, los proyectos de traducción complejos requieren la cooperación de
expertos en diversas areas. En este trabajo utilizamos el enfoque del constructivismo y
del análisis situacional para desarrollar una base teórica para estudiar el fenómeno de
la cooperación en la traducción. Aplicamos este marco a ejemplos reales de
50 Hanna Risku & Roland Freihoff

cooperación entre un traductor autónomo y los socios con los que coopera. Los dos
puntos esenciales del análisis son los siguientes:
1. La cooperación entre socios tales como el iniciador del proceso, el que encarga la
traducción, los disenadores gráficos o del texto, los proveedores externos de la
información y el usuario o destinatario y
2. Las areas de cooperación tales como la coordinación, planificación, desarrollo y
evaluación.

Es geht in diesem Beitrag um einen theoretischen Rahmen der Translation:

Translation als Kooperative Textgestaltung. Wir wollen zeigen, wie dieser
Ansatz zum besseren Verständnis und zur Bewältigung der Komplexität des
Berufsalltags beitragen kann. Die Notwendigkeit des kooperativen Ansatzes
wird begründet. Danach werden praktische Erfahrungen anhand eines Rollen­
modells besprochen.
Das Übersetzen ist eine höchst kreative, situationsabhängige und damit
flexible Tätigkeit, bei der kommunikative Hilfsmittel für Zielsituationen
erstellt werden. Darum sind die jeweiligen Zielgruppen und Anwendungs-
zwecke für jeden einzelnen Fall/Auftrag genau zu spezifizieren und dabei die
Rolle des Ausgangstextes (AT) immer wieder neu zu durchdenken und zu
gewichten. Der Auftraggeber liefert dem Translator Informationen über seine
Vorstellungen und Bedürfnisse, aber letztlich ist die Rollenverteilung jeweils
von den Übersetzenden zu erarbeiten und zu vertreten - das ist eine ihrer
Hauptaufgaben, wie die Theorie vom Translatorischen Handeln (Holz-Mänttäri
1984) postuliert. Ziel dieser Theorie ist, die Ausarbeitung fallspezifischer
Strategien zu ermöglichen. Es geht dabei also nicht um fertige Rezepte.

1. Von erlernten Verhaltensabfolgen zur situativen Handlung

Als das menschliche Handeln und so auch das Übersetzen noch vor allem als
Befolgen gelernter Schemata und Umsetzung vorhandenen Wissens betrachtet
wurde, war die Sache einfach: die Beschreibungsformeln lieferten eine Theorie
und damit auch eine Methode. Konventionen und Methoden dienten als fertige
Handlungsanleitungen. Abweichungen von der üblichen Praxis und Sprach­
verwendung wurden als situationsbedingte Sonderfälle behandelt. Diese Sicht
ist typisch für Sprachtheorien, die sich in der Auflistung von Merkmalen,
Scripts und anderer Wissenseinheiten erschöpfen (propositionale Theorien,
Symbolmanipulation, semantische Netze, frühe konnektionistische Ansätze)
und genauso typisch für Translationstheorien, die sich auf solche Sprach­
theorien stützen bzw. Translation als die Anwendung erlernter Handlungs-
abfolgen beschreiben (z.B. äquivalenztheoretische und fertigkeitsbasierte
Translationstheorien).
 Kooperative Textgestaltung 51

Die konkrete jeweilige Situation ist jedoch nicht etwas, womit wir
unsere vorgefassten Schemata auffüllen, sondern gerade sie ist die fundamen­
tale Quelle und unverzichtbare Ressource für die Sinnbildung und damit für die
Handlungsfähigkeit. Schemata und Konzepte sind der Handlung vor- und nach­
gelagert, sie bestimmen nicht die Handlung selbst, sondern dienen lediglich als
vage Pläne und nachträglich erstellte Erklärungen. Handlungen und Bedeu­
tungen dagegen bauen sich immer im Hier und Jetzt auf. Die Aufgabe der
Schemata ist nur, uns in eine Position zu bringen, aus der heraus wir die
momentanen Ressourcen nutzen und eine sinnvolle Interpretation der Situation
samt ihrer verbalisierten Teile (der darin situierten Texte) entwickeln können.
Deshalb sind Situationen der Schlüssel zum Erlernen komplexer Aufgaben. Die
Expertenkompetenz liegt nicht im Vollziehen eines Plans, sondern im kontinu­
ierlichen Erstellen neuer Hypothesen (Suchman 1987:188; Hendriks-Jansen
1996).

2. Strategien zum Umgang mit Unsicherheit

Wenn wir also jetzt das Konstruktive und Situative, das Kreative, an der Trans­
lation zutage fördern, fragen sich viele, wie die daraus resultierende Unsicher­
heit und Unvorhersehbarkeit zu bewältigen ist. Werden hier nicht Willkür und
Unkontrollierbarkeit gezüchtet? Schon bahnt sich eine Gegenreaktion an: die
Überlegung, ob wir nicht durch eine graduelle Vorfixierung der Texte, der
Übersetzungsprozesse und -aufgaben eine Stabilität erzwingen könnten. Dies
wäre jedoch eine totale Verkennung des Grundproblems: derartige Fixierungen
führen in eine erstarrte Scheinwelt, in der wir mit quasi-objektiven, letztlich
aber willkürlich festgelegten Prozeduren und zwischenkulturellen Äquivalenz­
setzungen arbeiteten.
Der vorliegende Beitrag zeigt eine andere Lösung: die kooperative
Textgestaltung. Wenn alle Beteiligten in den Übersetzungsprozess eingebun­
den werden, lassen sich Flexibilität, Kreativität und Verantwortung voll aus­
schöpfen - ohne Realitätsverlust. Das ist durch die kognitiven und kommuni­
kativen Grundlagen der Translation zu erklären: erfolgreiche Kommunikation
ist nicht auf Erlernung und Beherrschung vorgefasster Bedeutungen und
Konventionen zurückzuführen, sondern im Gegenteil auf unsere Fähigkeit, im
Interpretationsprozess ständig unsere Hypothesen zu prüfen, Missverständnisse
aufzudecken und zu bewältigen. Bei diesem Prozess der Kohärenzbildung grei­
fen wir auf vielfältige situative Hinweise zurück: auf den Anlass, den Raum,
das bisher Gesagte, die beteiligten Personen. Wir leben nicht nur mit und in der
Unsicherheit, sondern wir benötigen sie sogar für komplexe intelligente Tätig­
keiten. Darin besteht die eigentliche Leistung, und nicht etwa in der Beseiti­
gung der Ambivalenz. Wir sollten nicht versuchen, diese fundamentale Un­
sicherheit und Unvorhersehbarkeit wegzuerklären oder wegzuregulieren, son-
52 Hanna Risku & Roland Freihoff

dem uns auf die Ressourcen konzentrieren, die uns die Bewältigung der Fall­
spezifik eröffnen (Suchman 1987: 69). Wesentlich sind also die Strategien des
Krisenmanagements. - Die durch Verallgemeinerungen und Vereinfachungen
gewonnenen, oft nur vermeintlichen „Standardfälle", in denen sich erlernte
Routineprozeduren oder Automatismen zur Texterstellung anbieten, werden
immer mehr mit Hilfe von Textbaustein- und Übersetzungsspeichersystemen
am Computer erledigt (Freihoff 1998).

Die Strategien des Krisenmanagements sind zum großen Teil kooperativer

Natur. Weil Übersetzende kaum direkten Zugriff auf die Verwendungssitua­
tionen von AT und ZT haben, sind sie meist auf die Beteiligten als Informanten
angewiesen. Wichtigster Kooperationspartner ist der Besteller (Auftraggeber).

Wenn die Forschung die Kontakte der Übersetzenden zu den anderen Beteilig­
ten ignoriert, so verzichtet sie auf wesentliche Teile, die dieser Arbeit als
Hilfeleistung, als Ermöglichung von Verständigung für andere, erst einen Sinn
geben. Denn letztlich sind es die gegenseitigen Erwartungen der Beteiligten,
die die jeweiligen Ziele und den Handlungsrahmen für die translatorische
Produktentwicklung bestimmen. - Wir wollen die Diskussion auf den Boden
der Tatsachen holen und verwenden dazu das Rollenmodell von Holz-Mänttäri
(1984:106). Anhand von praktischen Erfahrungen erläutern wir die impliziten
und expliziten Erwartungen der Beteiligten.

3. Situative Einbettung der translatorischen Handlungen

Im folgenden werden hier die Rahmenbedingungen der Kooperation in einem

authentischen Handlungskontext besprochen. Dabei konzentrieren wir uns auf
die Gesichtspunkte WER (Rollen) und WORIN (Kooperationsbereiche).

EIN HANDLUNGSRAHMEN

Ein auf Haftlaminate für selbstklebende Etiketten spezialisiertes Unternehmen belie­

fert Druckereien, Verarbeiter und Anwender mit Rollenmaterial aus Papier und Kunst­
stoff. Die Kategorien der benötigten Translate umfassen Produktbeschreibungen,
Lieferbedingungen, Produktdatenblätter, Broschüren und Faltblätter, Anzeigen, Kun­
denmagazin, Kundeninformationen (Begleitbriefe, Einladungen zu Messen und
Schulungsveranstaltungen), Schulungsmaterial, Texte für mündliche Präsentationen,
Videofilme, Pressemeldungen sowie Fachartikel. Hauptsprachen sind Englisch und
Finnisch. Ausgangssprache ist meist Englisch, doch auch Finnisch, d.h. EN- oder FI­
DE, -FR, -IT, -SP, -FLA etc.
 Kooperative Textgestaltung 53

3.1 Rollen (WER)

Die Kooperationspartner des Translators im angesprochenen Handlungsrah­

men werden in Abbildung 1 dargestellt. Rollen als Aufgabenbündel betrachtet
erlauben verschiedenen Personen sich eine Rolle (d.h. Aufgaben) zu teilen,
oder eine Person kann mehrere Rollen übernehmen (z.B. Initiator + Besteller).

Abb. 1. Translatorischer Handlungsrahmen, Rollen bzw. Kooperationspartner

AUSGANGSKULTUR AUSGANGS- ODER ZIELKULTUR ZIELKULTUR

Rollen und Kooperationspartner:

1. Der Initiator /Bedarfsträger braucht einen Zieltext (ZT)

• Das Unternehmen ist Initiator und Bedarfsträger. Es initiiert die Translation, hat
einen Kommunikationsbedarf und braucht Hilfe bei der interkulturellen Texterstel­
lung. Seine Ziele sind der Anlass für die translatorische Handlung und für die Bildung
eines Kooperationsnetzwerkes zur Herstellung des Translats.
• Aufgaben: Bedarf verbalisieren, Feedback geben und einholen.

2. Der Besteller bestellt einen ZT

Die Kommunikationsabteilung ist zuständig für Medien, Werbung, PR, Informationen
/ Mitteilungen extern und intern.
• Die Kommunikationsabteilung erteilt den Auftrag, ist also Besteller und zugleich
Koordinator. Auch die Überprüfung der Zieltexte erfolgt durch den Besteller über die
künftigen Applikatoren vor Ort - meist auf Wunsch des Translators.
• Aufgaben: Bedarf (und Bedarf des Initiators) verbalisieren, Feedback geben und
einholen, Translator und Textgestalter für Kooperation kontaktieren.

3. Der AT-Gestalter erstelltundgestaltet einen Ausgangstext (AT)

Die Ausgangstexte bzw. -vorlagen werden von einem englisch-amerikanischen
Autorenteam nach Vorgaben produziert, redigiert, übersetzt (Fi-En).
• Der AT wird als Ausgangsmaterial für die Translation verwendet.
• Aufgaben: Bedarf verbalisieren und Feedback geben und einholen.
54 Hanna Risku & Roland Freihoff

Textge stalter, Textproduzent und Textdesigner sind Oberbegriffe für Verfasser und
Gestalter. Verfasser/innen schreiben Texte im Hinblick auf die situative Einbettung.
Gestalter/innen entwerfen (und verwirklichen) die drucktechnische Gestaltung
(Layout, Typographie).

4. Der Applikator arbeitet mit dem ZT

Die Zieltexte werden in den Auslandsniederlassungen u.a. für Werbung, Verkauf und
interne Schulung verwendet. Verkaufsabteilungen, Händler und Kursleiter sind somit
Zieltextappli katoren. Ihre Aufgabe ist auch das Testen der Translate, die auf Wunsch
des Übersetzers vom Besteller an die Applikatoren geschickt werden.
• Der Applikator übermittelt den ZT an die Rezipienten.
• Aufgaben: Bedarf verbalisieren, Feedback geben

5. Der Rezipient rezipiert den ZT

• Intendierte Zielgruppe des Translats sind die Rezipienten und Endbenutzer. Papier­
händler, Druckereien, Verarbeiter und Anwender der Haftlaminate. Die Beziehungen
zu den Rezipienten sind nur „virtuell" (s. auch Holz-Mänttäri 1984: 69, „unscharfe
Einschätzung der Situation").
• Aufgaben / Reaktionen: Bedarf verbalisieren, Feedback geben

6. Der Translator erstellt einen ZT

Mit der Erstellung der zielsprachlichen Texte beauftrage Freelancer und Überset­
zungsbüros sind in diesem Fall die Translatoren und zugleich Rezipienten der AT und
des Auftrags sowie Gestalter der ZT. Sie klären in der Kooperation ggf. den Besteller
über die Vorbedingungen einer qualifizierten Arbeit auf.
• Der Translator produziert als Textgestalter im Auftrag des Bestellers einen ZT.
• Aufgaben: Verständnis für die Bedürfnisse der Kooperationspartner entwickeln,
die Rolle des Ausgangsmaterials durchdenken und gewichten, Ziel, Methode und
Vorgangsweise bei der ZT-Produktion bestimmen. Bedarf verbalisieren, Feedback
geben und einholen.

3.2 Vertextung

Die Vertextung der Botschaft lässt sich in Komponenten aufteilen (Schröder

1993: 192). Die Botschaft geht aus einer komplexen pragmatisch-rhetorischen
Konstellation hervor, in die der jeweilige Text auf vielen Ebenen integriert ist.
Funktion, Ort und Zeit, Medien, Sender- und Empfängerpragmatik sind Rand­
bedingungen, die als Faktoren in die Aufgaben(bündel) und damit in die Rollen
der Personen im jeweiligen Kooperationsrahmen eingehen. „Kontextfreie"
sprachliche Äquivalenzen beim Übersetzen zu fordern erweist sich damit als
absurd. Translation verlangt die Einbettung der Vertextung in Handlungs- und
Kooperationsebenen.
 Kooperative Textgestaltung 55

3.3 Handlungs- und Kooperationsebenen

Nun gilt es prospektiv zu denken und zu zeigen, wie für den Texter die Rah­
menbedingungen sozusagen das Wurzelgeflecht bilden, auf dem die Texte
wachsen. Eine Grafik mit den verschiedenen Ebenen von Zielbildung und
Maßnahmen kann uns dabei helfen (s. Abb. 2).

Die Handlungs ebenen sind:

(1) Interpretation der Gesamtsituation durch den Translator (Intentionen der
Beteiligten, Rahmenbedingungen). Sie ergibt, worauf es ankommt:
(2) den (präzisierten) Skopos: das fallspezifische Handlungsziel des Translators als
individuelle Gesamtperspektive. Der Skopos legt (3) fallspezifische Vorgehensweisen
nahe: die Entscheidung über den Umgang mit dem AT und die Rolle des AT sowie die
Wahl der Mittel. (4) Die adäquaten/persönlichen MikroStrategien und konkreten Maß­
nahmen bzw. Werkzeuge bewirken letztlich die endgültige Form des Translats (4).

Abb. 2. Handlungsebenen der Vertextung

1. Interpretation der Gesamtsituation

2. Translator-Skopos

3. Vorgehensweise

4. Form

In der Kooperation werden also alle Ziel- und Entscheidungsbereiche berührt

und bei Rückfragen wieder durchlaufen: Die Kooperationspartner geben ihre
Interpretationen der Gesamtsituation, entwickeln Skopoi, Vorgehens­
weisen und MikroStrategien zur Erstellung einer sinnvollen Form (s. Modell in
Risku 1998: 125): Initiator/Bedarfsträger und Translator einigen sich über den
Bedarf (1) oder auch über allgemeine Ziele der Textproduktion (2), Gestal-
tungsentwürfe (3) bzw. über die Wahl bestimmter Bilder, Stile, Termini und
Materialien (4). Auch die übrigen Kooperationspartner können Meinungen
und Kritik äußern.

In unserem Beispielkontext sind die Kooperationspartner des Übersetzers (1) Fach­

leute der Bereiche Kommunikation, Marketing, Produktion, Forschung und Entwick­
lung sowie (2) Druckereien und/oder Werbeagenturen, Webmasterin, ein englisch-
amerikanisches Autorenteam (AT-Gestalter), die Niederlassungen im Ausland (Ziel­
text-Applikatoren) und Informanten für sporadische Einzelinformationen. Mit dem
Initiator und mit den Endrezipienten gibt es keine direkte Kooperation. Der Überset-
56 Hanna Risku & Roland Freihoff

zer muss sich von der Welt der Endrezipienten durch Zeitschriften und Rückfragen
beim Bedarfsträger/Besteller ein Bild machen.

Als „Stamm-Freelancer" kennt der Übersetzer „Tradition" und Usancen der Firma und
der Branche. Für Recherchen nutzt er sein Paralleltextarchiv, Übersetzungsarchiv
(teils elektronisch), gedruckte Übersetzungen (Prospekte, Broschüren etc.), Glossare
(offizielle und eigene), Internet-Quellen und Ausgaben aktueller Fachzeitschriften, die
das Unternehmen ausleiht.

3.4 Medien (WODURCH)

Interaktive Kooperationsmedien verdrängen den traditionellen Postweg. In

unserem Fall sind die verwendeten Medien in der Reihenfolge ihrer Nutzungs-
frequenz: Telefon (in Finnland kulturspezifisch Nr. 1), E-Mail für Lieferung
von AT und ZT sowie Klärung von Fragen, Telefax für Texte auf Papier sowie
kleine Druckfahnen, Post für Disketten und Texte und Boten für Druckfahnen­
pakete.

3.5. Kooperationsbereiche (WORIN)

Der Translator arbeitet in einzelnen Bereichen mit den Partnern zusammen.

Der Besteller, die Kommunikationsabteilung, ist Partner für Auftrag, Überprüfung,

Koordination der Übersetzung sowie auch für Druckereien und Werbeagenturen,
Webmasterin, was Text und Bild, Layout etc. betrifft.

Experten (Marketing, Produktion, Forschung und Entwicklung) sind Partner für Sach­
wissen und Terminologie. Sie bestätigen die Vorschläge, geben selbst den Terminus
vor, den sie unbedingt im Text haben wollen, oder sie erläutern das „Konzept", damit
der Übersetzer selbst weitersuchen kann. Oft vermittelt die Kommunikationsabteilung
den Kontakt und leitet Fragen des Translators weiter.

Sonstige Informanten sind Partner z.B. für Formulierungsfragen oder Sachwissen. Sie
haben mit dem Auftrag weder direkt noch indirekt zu tun. Es sind z.B. Bekannte mit
Spezialkenntnissen und Muttersprachler, die eine Formulierung beurteilen können.

AT-Gestalter als Partner sind Experten, die einen Fachartikel schreiben: Produkt­
entwickler, ein Werbebüro bzw. dessen Texter, zuweilen auch ein Regisseur für Ton-
aufhahmen, wenn der Übersetzer Texte für Präsentationen selbst übersetzt, über­
arbeitet und auch auf Band spricht. Die Gestaltung wird auch vom AT-Autorenteam
beeinflusst bei Termindruck und Absprachen. Auch die Niederlassungen als ZT-Appli­
katoren sind Partner für Überprüfung und Feedback. -
 Kooperative Textgestaltung 57

Typische Kooperationsbereiche sind (a) Auftragserteilung und -besprechung,

(b) Terminologie, Werbesprache, (c) Defekte im AT, (d) Korrekturlesen, (e)
Nachträgliche Textänderungen und (f) Formulierungsfragen und Layout:

(a) Auftragserteilung und -besprechung:

Kooperation mit neuen Mitarbeiter/innen verlangt die Artikulation von Bedürfnissen
und Eingehen auf bestehende Unsicherheiten. Em TRANSLATION SERVICE GUIDE
wäre nützlich: Um sinnvoll arbeiten zu können, brauchen Übersetzer hinreichende
Informationen über die text-externe Situation, die Zielgruppe und ihre Umweh, und
darüber, was mit dem Text bezweckt wird, was gerade in diesem Fall erwartet wird: für
welches Druckerzeugnis ist der Text gedacht? Bilder? Andere Texte, z.B. frühere
Übersetzungen? Terminologieanweisungen? Gestaltung: Wo stehen die Titel? Auch
Schriftgröße und -font können die Übersetzung beeinflussen.

(b) Terminologie und Werbesprache:

Kooperationspartner (Produktion, Verarbeitung, Marketing, Werbung, Endanwender)
entwickeln jeweils einen eigenen Sprachgebrauch, der ihrer „Perspektive" entspricht.
Deshalb ist Abstimmung erforderlich. Beispiel: Trägerpapier, Abdeckpapier, Trenn­
papier, Silikonpapier, In allen Fällen handelt es sich hier um das abtrennbare silikoni-
sierte Papier, von dem sich das Haftetikett ablösen lässt. Finnisch: taustapaperi,
irrokepaperi, Englisch: backing (paper), release paper, liner.

(c) Defekte im AT-Material

In Krisenfallen tragen die Kooperationspartner gemeinsame Verantwortung. Sach­
fehler in Druckerzeugnissen müssen in allen Sprachen verhindert werden. Häufig
treten Fehler bei Zahlen, Maßeinheiten und Namen auf. Ursache ist z.B. gedanken­
loses Übernehmen oder Kopieren alter Textelemente. - Sprachlich inkompetente Mit­
arbeiter/innen schreiben in der Fremdsprache. Kommt in einem unidiomatischen AT
der Gedanke trotzdem klar heraus, ist das für die Übersetzung irrelevant. - Der
Translator wird durch Kooperation zum Gesprächspartner bei (interkulturellen) Sach-
und Sprachproblemen.

(d) Korrekturlesen: Sinnvolle Textgestaltung, letzte Unsicherheiten

Ewig wiederkehrende Fehler im Deutschen sind Silbentrennung, Schreibung von ß/ss;
Groß- und Kleinbuchstaben. Finnische Druckereien und Werbebüros haben kaum
deutschkundiges Personal. Fehler fallen immer auf den Übersetzer zurück. Deshalb ist
Korrekturlesen im Übersetzungspreis inbegriffen. Bei der Korrektur stößt man auch
auf „eigenmächtige" Ergänzungen, nachträgliche Änderungen, aber vor allem auf
Mängel, die eigentlich in die Verantwortung der Druckerei gehörten: sprachfremde
Typographie, finnische oder englische statt deutscher Anführungsstriche und Gedan­
kenstriche, fehlende Zwischenräume bei Wörtern und Buchstaben (Ausgleich, Ker­
ning), fehlende Wörter und Sätze, vertauschte Absätze usw. Auch das Layout spielt
eine Rolle, etwa in bezug auf sinnvolle(re) Trennung von Wörtern und sinnvolle
(sprechbare) Verteilung zentrierter Texte auf eine bestimmte Zahl von Zeilen. All das
fuhrt dazu, dass der Translator letztlich selbst für die Korrektur zuständig ist.
58 Hanna Risku & Roland Freihoff

(e) Textänderungen: Unabgeschlossene Kooperation, Offenheit der Texte

Die Texte sind so lange „offen", bis sie in Druck gehen. Texte, die in den Niederlas­
sungen des Unternehmens im In- und Ausland begutachtet werden, erfahren Erwei­
terungen, Kürzungen oder auch starke Veränderungen. Das ist vielleicht frustrierend,
gehört aber zum Charakter der redaktionellen Arbeit.

(f) Formulierungsfragen und Layout

Wenn Ausgangstexte für Drucksachen zunächst nur als Brottext geliefert werden, gibt
es oft Schwierigkeiten mit Überschriften und Bildern: Worauf kommt es eigentlich
an? Was ist im Bild zu sehen? Wieviel Platz ist für Überschrift oder Bildtext vorgese­
hen? All das beeinflusst den Formulierungsspielraum des Übersetzers. Wird die Über­
setzung zu lang, muss umformuliert werden. Reduziert man nur Laufweite und Punkt-
größe der Schrift, so wird der Text unleserlich, und das Layout leidet insgesamt.

4. Ausblick

Wer in welchem Zusammenhang mitzureden hat oder entscheiden muss und

die Hauptverantwortung trägt, hängt vom sozialen und fachlichen Status, d. h.
vom Expertenstatus ab. Um Verantwortung übernehmen zu können, müssen
Übersetzende Experten werden - und sich auch ständig weiterbilden, um
Experten zu bleiben. Es geht um die translatorische Kompetenz, um die
Grundlagen des Übersetzens als Expertentätigkeit (s. Risku 1998). - Darin liegt
auch die Lösung der zur Zeit in der Translatologie diskutierten „Machtfrage".
Sie ist mit der möglichen Verantwortung verbunden. Wie alle anderen Berufe
auch, müssen sich Übersetzende ihr Ansehen immer wieder verdienen, indem
sie ihre Arbeit voll verantworten. Es gibt in allen Bereichen leitende, assistie­
rende, dienende und ausführende Aufgaben bis hin zum Routinejob.

Verbraucherinnen und Verbraucher als Bedarfsträger, Besteller, Applikatoren

und Rezipienten wissen oft nicht, was sie eigentlich bestellen und geliefert be­
kommen, und Übersetzerinnen und Übersetzer lassen sie oft im Stich. Daher
müssen Interkulturelle Kommunikation und Translation
• Schulfach werden (in allgemeinbildenden und in Fachschulen),
• in Lexika praxisgerecht beschrieben werden und
• in der Fachausbildung in den Bereichen Publizistik, Handel, Diplo­
matie, Wissenschaft etc. einen höheren Stellenwert erhalten.

Um das Niveau anzuheben und sinnvolle Arbeit entsprechend honoriert zu

bekommen, müssen wir die Bedarfsträger von der Qualität unserer Arbeit
überzeugen und die allgemeine Ignoranz beseitigen. Für das interkulturelle
Gespräch über den Status der Übersetzer in der Gesellschaft dürfte im Deu­
tschen das Wort „Autorität" besser geeignet sein als „Macht" (für engl. power),
 Kooperative Textgestaltung 59

weil letzteres auch Aggressivität, Willkür und Durchsetzung persönlicher Vor­

stellungen gegen die Interessen anderer anklingen lässt. „Macht" passt nicht
zum Übersetzen als Dienstleistung, da gerade der Translator den ganzen
kooperativen Handlungsrahmen, die sachlichen Belange und die Interessen
anderer berücksichtigen muss. Der Translator handelt im Auftrag, er ist dazu
„befugt" und im jeweiligen Rahmen verantwortlich.

Um also Abwehrhaltungen gegen eine neue, vermeintliche Selbstherrlichkeit

eines sehr braven Berufsstandes zu vermeiden, sollte man die Kompetenz und
das Verantwortungsbewusstsein des Translators als Kooperationspartner stär­
ker in den Vordergrund rücken, als dies bisher geschehen ist.

Literatur

CLARK, Andy 1997. Being there. Putting Brain, Body, and World Together Again.
Cambridge: The MIT Press.

FREIHOFF, Roland 1999. „Textproduktion und Übersetzen als Wiederverwertungsprozeß -

Kommunikation und Sprachtechnologie". Detlef Wilske (Hrsg.). Fachsprachen und Über­
setzungstheorie. VAKKI-Symposium XVIII/1998, 70-81. Vaasa: Vaasan yliopisto.

HENDRIKS-JANSEN, Horst 1996. Catching Ourselves in the Act. Situated Activity, Inter­
active Emergence, Evolution, and Human Thought. Cambridge: The MIT Press.

HOLZ-MÄNTTÄRI, Justa 1993. „Textdesign - verantwortlich und gehirngerecht". Traducere

navem. Festschrift für Katharina Reiß zum 70. Geburtstag. Justa HOLZ-MÄNTTÄRI &
Christiane NORD (Hrsg.) ,301-320. Tampere: Tampereen yliopisto, Studia translatologica a 3.

HOLZ-MÄNTTÄRI, Justa 1984. Translatorisches Handeln. Theorie und Methode Helsinki:

Suomalainen Tiedeakatemia, Annales Academiae Scientiarum Fennicae B 226.

RISKU, Hanna 1998. Translatorische Kompetenz. Kognitive Grundlagen des Übersetzens als
Expertentätigkeit. Tübingen: Stauffenburg.

SCHRÖDER, Hartmut 1993. „Semiotische Aspekte multimedialer Texte". Fachtextpragmatik.

Hartmut SCHRÖDER (Hrsg.), 189-213. Tübingen: Lang

SUCHMAN, Lucy 1987. Plans and Situated Actions. Cambridge University Press.
TRANSLATION INTO A NON-MOTHER TONGUE
IN TRANSLATION THEORY: DECONSTRUCTION
OF THE TRADITIONAL

Nike Kocijancic Pokorn

University of Ljubljana, Slovenia

Zusammenfassung
Nur wenige Übersetzungswissenschaftler haben ernsthafte Überlegungen über die
Unterschiede zwischen Übersetzungen in die Muttersprache und aus der
Muttersprache des Übersetzers angestellt. Trotzdem wird oft die Annahme vertreten,
daß die Übersetzer nur in ihre eigene Muttersprache übersetzen sollten, um sprachlich
und kulturell akzeptable Übersetzungen zu produzieren. Auf Grund einer genauen
Lektüre von einigen grundlegenden Werken zur Übersetzung läßt sich diese
traditionelle Annahme jedoch auf aprioristische Auffassungen zurückführen. Es fehlen
wissenschaftliche Beweise für die tatsächliche sprachliche und kulturelle
Überlegenheit der Übersetzungen in die Muttersprache. Die Annahme ist entgegen
weitverbreitetem Glauben nicht traditionell oder universal akzeptiert - weder in der
Praxis noch in der Theorie des Übersetzens.

Résumé
Peu de chercheurs en traductologie ont analysé sérieusement les différences entre les
traductions faites vers la langue maternelle du traducteur et celles faites à partir
d'elles. Malgré tout, on affirme souvent que les traducteurs devraient travailler
exclusivement vers une langue, de façon à produire des versions acceptables, à la fois
linguistiquement et culturellement. Un regard attentif sur des ouvrages fondamentaux
en traduction révèle néanmoins que cette affirmation "traditionnelle" s'appuie plus sur
des conditions a priori que sur des preuves scientifiques. En dépit de l'opinion
courante, l'affirmation est loin d'être universellement acceptée, que ce soit du côté des
praticiens ou du côté des théoriciens.

Resumen
En los Estudios de Traducción, pocos estudiosos han considerado seriamente la
diferencia entre las traducciones directas e inversas realizadas por los traductores. A
pesar de esto, suelen pensar que los traductores sólo deberian traducir hacia su propia
62 Nike Kocijancic Pokorn

lengua para garantizar asi traducciones lingüistica y culturalmente aceptables. Sin

embargo, un estudio detallado de algunas obras importantes sobre la Traducción
revela que esta premisa "tradicional" tiene su origen en opiniones realizadas a priori
sin un apoyo cientifico que demuestre la superioridad lingüistica y cultural de la
traducción directa. Además, esta premisa, al contrario de la opinión general, no se
acepta de manera tradicional y universal ni en la teoria ni en la práctica de la
traducción.

1. Mystification of the native speaker - the translator as owner of the TL

The assumption that translators can master only their mother tongue and must
therefore translate only in that direction developed rather late in the Western
world. In fact, it seems to have been Martin Luther who for the first time
assumed that one translated satisfactorily only into one's own language (Kelly
1979:110). His conviction was taken over and strengthened by the first and the
second, nationalist generation of Romantic authors, who also made a great
contribution to the rise of national philologies, especially because of the almost
obsessional identification of language and nation (Lefevere 1982:49). It was
perhaps Wilhelm von Humboldt who expressed most powerfully the
mystification of the native speaker and the special status the mother tongue had
at that period. Humboldt claimed that the nation is deeply connected with its
language, in fact that the nation's language is the spirit of that nation,1 which
means that only those who speak the language of a particular community can
access the hidden essence of the nation. That is also why the translator can
never write the way the author of the original would have written in the
language of the translator (Humboldt 1977:42), because the complete passage
from one language to another is impossible. Every language has its own way of
expression, which remains inaccessible to everyone who does not speak that
particular language from birth. Translation should therefore always proceed
from foreign languages to one's mother tongue and never vice-versa, since the
hidden essence of the target language is not attainable by any foreign speaker.
The roots of the conviction that we can grasp the ungraspable only in our
mother tongue, and consequently create a convincing translation only in our
native language, thus stem from this Romantic identification of the
transcendental nature of the nation and its language.
The claim that one should always translate into one's mother tongue
can also be found in contemporary writings on translation. Thus for example
Peter Newmark is short but blatantly direct in regard to this same problem; he
too states that translators cannot translate into a non-mother tongue and so
repeats and strengthens the "traditional" position.
 Translation into a non-mother tongue 63

(... ) A foreigner appears to go on making collocational mistakes however long

he lives in his adopted country, possibly because he has never distinguished
between grammar and lexicology. An educated native will also make mistakes
in collocation, particularly if he is under the influence of interference, but he
will correct himself intuitively. Sprachgefühl means awareness of
collocations. For the above reasons, translators rightly translate into their own
language, and a fortiori, foreign teachers and students are normally unsuitable
in a translation course (Newmark 1981:189).

Translators should thus translate only into their mother tongues; even if a
person lives in a TL culture for years, his or her writing will be, according to
Newmark, "unnatural and non-native", full of "unacceptable or improbable
collocations" (ibid.). In spite of these statements, however this "indisputable"
stand reveals some considerable cracks. In examining Newmark's assertions
more closely, it seems difficult to avoid the following questions: Who is the
native speaker Newmark is talking about? The speaker of English, Scottish,
Canadian, American English or also the speaker of Singapore, Indian,
Pakistani, Nigerian English? Are the children of immigrants who are born in an
English-speaking country also native speakers of English or not? Is a person
who moved to an English-speaking environment in his/her childhood a native
speaker of English, and if so, when does childhood stop? What about pairs of
translators consisting of a native and a non-native speaker of English - which
language is their mother tongue?
Newmark does not answer these questions, and in fact, even if he had
wanted to, he would not have been able to provide final answers. The notions
of'native speaker" and "mother tongue" are far from being well defined in
linguistics; in fact, there is a considerable variation in the connotations
attributed to those terms, which seem to depend on the ideological position of
the person providing the definition, or at least on the motives hidden behind
his/her need to determine them (see Skutnabb-Kangas & Phillipson 1989:450-
477; Phillipson 1992:39).
Moreover, linguists have no answer to the question of when the critical
age for acquiring a particular language occurs (see Krashen 1981:72-76; Stern
1983:362-63; Littlewood 1984:66-67; Singleton 1992:47; Cook 1996:108); nor
do they, in spite of some attempts (e.g. Coppieters 1987, Davies 1991), have
any acid tests which would once and for all determine the difference between
native speakers and highly proficient non-native speakers of a particular
language (Paikeday 1985). Although we must agree that the majority of foreign
learners never achieve native-like competence, it remains unproven whether
some individuals nevertheless eventually succeed in being accepted by a
particular foreign community as native speakers of that language.
Newmark's categorical claims thus seem more than suspect. He seems
to ethnocentrically defend the notion of the superiority of the "natural native
64 Nike Kocijancic Pokorn

speaker", the innate state that can never be acquired, and thus rejects the
marginal and peripheral (i.e. translators from immigrant communities and the
practice of team translation) as necessarily inferior.2

2. The idealisation of the translator- a perfect bilingual translator

It is the vagueness of the basic terms "mother tongue" and "native speaker"
that has most probably led many translation theoreticians to avoid the question
whether translators should be native speakers of the source or of the target
language; instead, they have idealised the subject involved in the process of
translation. In other words, translators are assumed by some scholars to be
perfect bilingual speakers of the source and target languages, translating from
one mother tongue to another.
We find the claim that translators should be bilingual for instance in J.
C. Catford (1965:27): "The discovery of textual equivalents is based on the
competent bilingual informant or translator", where the vagueness and
evasiveness of the notion "equivalence" is coupled with a similarly vague and
abstract notion of "a bilingual informant or translator". The notion of
translators as perfect bilingual or even bicultural persons can also be found in
theoretical works by Mary Snell-Hornby, e.g. where she writes that the aim of
all translation teaching is to create "not only a bilingual but also a bicultural (if
not multicultural) specialist working with and within an infinite variety of areas
of technical expertise" (Snell-Hornby 1992:11). Linguists argue that "real", i.e.
perfect, bilinguals are rare, if they exist at all:

The notion of proficiency raises some very complex issues. Again, the
"obvious" answer is to say that people are bilingual when they achieve native*-
like fluency in each language. But this criterion is far too strong. People who
have "perfect" fluency in two languages do exist, but they are the exception,
not the rule (Crystal 1987:362).

Nevertheless, the notion of a perfect bilingual translator - this idealisation

which seems to have no tangible reflex in real life, this quintessential state we
all aspire to and never really attain - is often found in linguistic as well as in
"cultural" translation theories (see for example Gutt 19903:143; Bell 1991:15,
38, 40, 424; and Lefevere & Bassnett 1990:115).

3. A hidden traditional conviction

Theoreticians rarely discuss openly the possibility of choosing one's TL in

translation; however, they do covertly express their conviction that only
 Translation into a non-mother tongue 65

translation into one's mother tongue guarantees a good translation. This opaque
discourse on the translators' mother tongue reflecting the "traditional" view
can thus be found in Victor Hugo's writing on translation (Hugo 1992:18)6
through Friedrich Schleiermacher's "Über die verschiedenen Methoden des
Übersetzens" (Schleiermacher 1985:322)7 to Walter Benjamin (1982), where
he writes:

It is the task of the translator to release in his own language that pure language
which is under the spell of another, to liberate the language imprisoned in a
work in his re-creation of that work (Benjamin 1982:80).

Benjamin demands that a translator should not only master his own mother
tongue but also allow the foreign language to transform the target language by
means of translation and thus liberate the power of the pure, original language
- and this extremely important and difficult task seems to be only possible if
one is translating into one's mother tongue. By accepting the assumption that
translators always translate, or at least should translate, into their mother
tongue, Benjamin's work seems to reveal one of its sources - German
Romantic thought.
In addition to traditional writings, the hidden assumption that one
always translates into one's mother tongue can also be found in contemporary
theories - thus we can find it in Barbara Johnson's deconstructive thoughts
(Johnson 1985:142)8, as well as in Sherry Simon's feminist work (Simon 1996
94)9. But this hidden discourse is especially interesting and contradictory in
George Steiner's hermeneutic work on translation theory.
Steiner never openly discusses the translators' choice of the target
language, despite the fact that he repeatedly fails to determine his own native
language, supposedly possessing equal fluency in English, French, and
German. In fact, he even claims that he experiences his first three tongues as
perfectly equivalent centres of himself (Steiner 1992:120). Although he
recognises the difficulties in defining the notion of the mother tongue itself, he
indicates repeatedly that the TL of the translator should also be his mother
tongue. For example: "The translator labours to secure a natural habitat for the
alien presence which he has imported into his own tongue and natural cultural
setting" (ibid.:365). Or: "He [the translator] will import from abroad
conventions, models of sensibility, expressive genres which his own language
and culture have not yet reached" (ibid.:370). And finally: "[...] it is logically
conceivable that the translator, having gained great mastery over a source-
language, will conclude 'I understand this text but find no way of restating it in
my own native tongue "'(ibid. :372).
Although he cannot define his native language, and although he has
done some translation himself, Steiner never opens the question of choosing
one's target language in translation. The principle that the translator is allowed
66 Nike Kocijancic Pokorn

to translate only into his native language seems so deeply rooted in his thought
that he never challenges it. He also never disputes another principle, connected
to the first one and also typical of the canonised translational norms in the
English-speaking world - the principle of fluency and naturalness. And indeed,
the norm of fluency and naturalness has prevailed over other translational
strategies in English-speaking cultures and shaped the canon of foreign
literatures in English (see Venuti 1995). And since it was also agreed that
"perfect" fluency in the TL and the mastery of its different styles could only be
achieved in one's mother tongue, the norm that the translator (of at least
literary texts) should be a native speaker of the TL became widely accepted
too; indeed, it seems even more deeply grounded in English-speaking cultures
than the fluency principle.
But norms can change. Thus Lawrence Venuti challenges the absolute
validity of the norm of fluency and tries to get his readers to reflect on the
ethnocentric violence of a transparent, fluent translation, i.e. of a translation
that does not appear to be a translation but imposes itself as the "original"
(Venuti 1995:41) - he even pleads for the production of translations that reveal
"the linguistic and cultural difference of foreign texts" (ibid). But in spite of
his openness towards the foreign, and sensitivity to ethnocentric violence,
Venuti never touches the problem of the translator's TL, and thus accepts,
perhaps not consciously though, the prevailing and ethnocentric norm that
proclaims the superiority of TL translators.
Venuti seems to completely ignore the prevailing practice in peripheral
language communities, where many translators work into a language that is
non-native to them, and the deplorable fact in Western societies that many
translations are praised despite the fact that their "translators" did not
understand the SL10. In fact, he does not find this to be an issue of discussion;
thus he quotes Goethe translated by André Lefevere, a Belgian translating from
German into English (ibid. :99), and he explains the translational practice of
Ezra Pound, without mentioning his "Cathay" (1915), despite the fact that this,
probably the most praised of Pound's translations of Chinese poems, is also
famous for the fact that Pound did not understand Chinese when he translated
from E. F. Fenollosa's transcription of and commentary on the source text.
Moreover, Venuti emphasises that the TL should be the translators'
mother tongue, since he qualifies the language and culture the translator is
supposed to translate into as "domestic", and the SL culture as "foreign": "[...]
the translator's interpretive choices answer to a domestic cultural situation and
so always exceed the foreign text" (ibid.:37; emphasis added). Or: "Because
translation is a double writing, a rewriting of the foreign text according to
domestic cultural values, any translation requires a double reading - as both
communication and inscription" (ibid. :312). According to Venuti then,
translators choose a "foreign" text and translate it according to a "domestic"
 Translation into a non-mother tongue 67

cultural situation, which implies that they never work away from their native
language but always into their mother tongue.
By not discussing the possibility of translations into a non-native
language, by undertheorizing the basic notions of "foreign" and "domestic",
Venuti thus accepts the prevailing practice in major-language communities,
where, indeed, translation is usually done into the translator's mother tongue.
But since he not only accepts this practice but also generalises it, he thus
imposes yet another ethnocentric norm on the rest of the peripheral world by
suggesting the necessity and thus superiority of translations into the translator's
mother tongue, i.e. he covertly imposes the superiority of the TL culture.

4. Translation into a non-mother tongue and team translation as a part of

translation practice

Contrary to common belief, the principle that translation should always be

done into one's mother tongue does not have a long history. On the contrary,
translation into a non-mother tongue can also be found at the dawn of Western
history: in Ancient Greece, for example, the native language of the translator
was not an issue, or at least not one of the criteria according to which the
quality of the translation was assessed. Thus, for example, the praised Jewish
translators of the Old Testament had either Greek or Hebrew as their native
language, which means that at least some of them were translating out of their
mother tongue. In classical Rome the great and famous translators were native
speakers of Latin; however, the first Christian Latin translators seem to have
been Greeks. After a short period when Latin speakers like St. Jerome
dominated the field (and also Jerome, according to traditional accounts, worked
with a group of helpers whom he used as walking dictionaries), we enter a
period when nobody translating into Latin spoke it natively (Kelly 1979:109).
At the end of the Middle Ages, when the most heated debates about
translation were usually connected with the translation of the Bible, there were
few who found it objectionable that both Reformers and counter-Reformers
translated it into Latin. After that period translation into a non-mother tongue
still remained alive in science, where Latin was used as an international lingua
franca until the end of the eighteenth century.
After the Reformation period we find few translations out of the mother
tongue outside the scientific context and language teaching; however, in the
seventeenth century Thomas Power translated Milton's Paradise Lost into
Latin epic hexameter and in the eighteenth century a number of Greek patristic
works were translated into Latin.
Also in the twentieth century translation out of one's mother tongue is
not such a rare occurrence; it may be seldom found in major-language
societies, but it is a common translation praxis in minor-language communities,
68 Nike Kocijancic Pokorn

or to use the current euphemism, in communities which use "a language of

restricted distribution or limited diffusion" and which are forced to translate
into foreign languages if they want their works to be translated at all (see also
McAlester 1992:292-296). Translation into non-mother tongues is also
common in other large but peripheral language communities, for example in
China, where Chinese translators are trying to change, according to them, the
distorted image of Chinese poetry which earlier translations created (Lefevere
1995a; 1995b).
This means that translation into a non-mother tongue is common in
small as well as in large language communities; however, it is undoubtedly
more common in cultures which do not have a central status and are forced to
the global periphery. Very seldom, though, do translators from peripheral
cultures work alone - the common practice adopted in those cultures is co-
operation between a native and a non-native translator, or a translator who is a
native SL speaker and a TL stylist. This co-operation on an equal basis has
been theoretically described in the theory of Bible translation.
Eugene A. Nida's Toward a Science of Translating is dedicated to such
pairs of translators; in fact, Nida's work goes even further since it is primarily
destined to help English-speaking Bible translators (missionaries) who translate
the Greek and Hebrew originals into one of the non-Indo-European languages
(Nida 1964:147) - i.e. translators who translate from one foreign language into
another, often with the help of secondary source languages (French, English),
which are used as substitute bases for translation (Nida & Taber 1982:6). At
first, Nida insists that ideally the translator should be bilingual in both the
source and the target languages (1964:149). And we have to agree with him
that "this ideal is rarely realized"; indeed, it seems obvious that "this ideal" is
in fact never realized at all.
That is why Nida soon leaves behind the realm of the ideal and focuses
on the real problems his group of translator-missionaries is facing. Since he is
aware of the fact that an ideal set of abilities in one person cannot be found, he
distributes the essential elements in the role of translator among several
persons in various ways (ibid. :153). According to him, in a translation team,
roles should be distributed among three persons: one person should interpret
the meaning of the source-language message, the second should suggest the
equivalent rendering in the receptor language and the third should be
responsible for style (ibid. :154). The basic principles of team translation
described in Nida's theoretical work could also be applied to the co-operation
between translators in peripheral communities, which provides this
translational praxis with a possible theoretical basis. Translation out of one's
mother tongue thus becomes not only a common fact in the contemporary
world but also a theoretically grounded action with fixed rules of conduct.
 Translation into a non-mother tongue 69

5 Conclusion

Translation into a non-mother tongue is known in Western history from

antiquity onwards, and can find one of its possible theoretical groundings in
Nida's work. This translational practice is especially common in languages
with restricted distribution and in larger linguistic communities which are
pushed into a peripheral position because of the global distribution of power.
Western translation theory in general ignores this practice, and accepts the
"traditional" assumption that translators should work only into their own
language (when translating all types of text, but especially when translating
literature) if they want to create linguistically and culturally acceptable
translations - an assumption which most probably stems from the Romantic
identification of the nation with the transcendental essence of language.
However, the direction of translation, i.e. whether translators work into or out
of their mother tongue, is not a reliable criterion of the linguistic and cultural
quality of their translations. Moreover, the fundamental notions used by the
majority of translation scholars (e.g. "native speaker", "mother tongue",
differences in proficiency between native and non-native speakers) are blurred
and elusive and cannot be objectively defined. Therefore, any axiomatic
assumption that translators should always translate into their mother tongues is
prejudiced and has no universal applicability. In fact, it appears high time for
translation scholars to reconsider the notion of the linguistic and cultural
inferiority of translations out of a mother tongue, since it seems that if this
conviction persists in translation theory, its only purpose will be to
ethnocentrically defend the superiority of post-Romantic West-European
concepts concerning translation and translational practice, and thus
consequently the a priori superiority of the translators and translational
practice of major-language communities.

Notes
1. "Die Sprache ist gleichsam die äußerliche Erscheinung des Geistes der Völker; ihre
Sprache ist ihr Geist und ihr Geist ihre Sprache, man kann sich beide nicht identisch genug
denken" (Humboldt in Stolze 1994:24).

2. Newmark's influence is particularly felt in books on translation teaching. See also Alan
Duff (1981:111, 125 and 1989:11) and Geoffrey Samuelsson-Brown (1995:16).

3. "[ ... ] the only complication is that the source language communicator does not master the
receptor language. Therefore he needs the help of a bilingual person to produce a receptor
language stimulus that will communicate his informative intention" (Gutt 1990:143).

4. "[...] the translator must possess linguistic competence in both languages and
communicative competence in both cultures [...]" (Bell 1991:42).
 Nike Kocijancic Pokorn
70

5. Lefevere and Bassnett are even more demanding: "Since languages express cultures,
translators should be bicultural, not bilingual" (Lefevere & Bassnett 1990:11).

6. "To translate a foreign writer is to add to your national poetry [...]" (Hugo 1992:18;
emphasis added).

7. "Seine [a translator's] erste Regel muß sein, sich wegen des Verhältnisses, in dem seine
Arbeit zu einer fremden Sprache steht, nichts zu erlauben was nicht auch jeder
ursprünglichen Schrift gleicher Gattung in der heimlichen Sprache erlaubt wird"
(Schleiermacher 1985:322; emphasis added)

8. "Through the foreign language we renew our love-hate intimacy with our mother tongue"
(Johnson 1985:142).

9. "Translation plays out in the open the 'everyday frustrations' of writing, projecting them
into an external form. We transfer our frustrations to the mother tongue [...]" (Simon
1996:94).

10. "Some of the most persuasive translations in the history of the metier have been made by
writers ignorant of the language from which they were translating (this would be so
notably where rare, "exotic" languages are involved)" (Steiner 1992:375).

References
BENJAMIN, W. 1982. "The Task of the Translator". Walter Benjamin, Illuminations ed. by
H. ARENDT, 69-82. London: Fontana.

BELL, R. T. 1991. Translation and Translating. London: Longman.

CATFORD, J. C. 1965.,4 Linguistic Theory of Translation. Oxford: Oxford University Press.

COOK, V. 1996. Second Language Learning and Language Teaching. London, New York &
Sydney, Aukland: Arnold.

COPPIETERS, R. 1987. "Competence Differences Between Native and Near-Native

Speakers". Language 63(3). 544-573.

CRYSTAL, D. 1987. The Cambridge Encyclopedia of Language. Cambridge: Cambridge

University Press.

DAVIES, A. 1991. The Native Speaker in Applied Linguistics. Edinburgh: Edinburgh

University Press.

DUFF, A. 1981. The Third Language: Recurrent Problems of Translation into English.
Oxford: Pergamon Press.

DUFF, A. 1989. Translation. Oxford: Oxford University Press.

 Translation into a non-mother tongue 71

GUTT, E.-A. 1990. "A Theoretical Account of Translation - Without a Translation Theory".
Target 2(2), 135-164.

HUGO, V. 1992. "Introduction to the translation of Shakespeare". Translation, History &

Culture: A Sourcebook ed. by A. LEFEVERE, 18. London & New York: Routledge.

HUMBOLDT, W. von. 1977. "Einleitung to Aeschylos' Agamemnon metrisch übersetzt von

Wilhelm von Humboldt, Translating literature: The German Tradition: From Luther to
Rosenzweig (Approaches to Translation Studies 4) ed. by A. LEFEVERE, 40-45. Assen,
Amsterdam: Van Gorcum.

JOHNSON, B. 1985. "Taking Fidelity Philosophically". Difference in Translation ed. by J.

GRAHAM, 142-148. New York: Cornell University Press.

KELLY, L. G. 1979. The True Interpreter: A History of Translation Theory and Practice in
the West. Oxford: Basil Blackwell.

KRASHEN, S. D. 1981. Second Language Acquisition and Second Language Learning.

Oxford: Pergamon Press.

LEFEVERE, A. 1982. "Translated Literature in the Study of Literature". Translation in the

Literary Process. Publikacie Vedeckovyskumne pracovisko literarney a experimentalnej
motodiky pedagogikej fakulty v Nitre. Vyskume materialy, studie, bibliographie 60.
Department of literary communication and experimental methodology at pedagogical faculty
at Nitra. (Dedicated to the Congress of ICLA, 10th, New York, August 22-29, 1982). 41-62.
Nitra: Nitrianske tlaciarne.

LEFEVERE, A. 1990. "Translation: Its Genealogy in the West". A. LEFEVERE & S.

BASSNETT (eds) 14-28.

LEFEVERE, A. 1995(a). "Chinese and Western Thinking on Translation". Unpublished

lecture at Cetra translation seminar in Leuven.

LEFEVERE, A. 1995(b). "Is this the same text?: A Consumer's Approach to Translation".
Unpublished lecture at Cetra translation seminar in Leuven.

LEFEVERE, A. & S. BASSNETT (eds) 1990. Translation, History and Culture. London:
Pinter.

LITTLEWOOD, W. 1984. Foreign and Second Language Learning. Cambridge: Cambridge

University Press.

McALESTER, G. 1992. "Teaching Translation into a Foreign Language: Status, Scope and
Aims". Teaching Translation and Interpreting: Training, Talent and Experience ed. by C.
DOLLERUP & A. LODDEGAARD, 291-297. Amsterdam & Philadelphia: John Benjamins.

NEWMARK, P. 1981. Approaches to Translation. Oxford: Pergamon Press.

NEWPORT, E. L. 1990. "Maturational Constraints on Language Learning". Cognitive

Science 14(1). 11-28.
72 N i k e Kocijancic Pokorn

NIDA, E. 1964. Toward a Science of Translating: With Special Reference to Principles and
Procedures Involved in Bible Translating. Leiden: E. J. Brill.

NIDA, E. & TABER C. R. 1982. The Theory and Practice of Translation. Leiden: The
United Bible Society, E. J. Brill.

PAIKEDAY, T. M. 1985. The Native Speaker is Dead!: An informal discussion of a

linguistic myth with Noam Chomsky and other linguists, philosophers, psychologists, and
lexicographers. Toronto & New York: Paikeday Publishing Inc.

PHILLIPSON, R. 1992. Linguistic Imperialism. Oxford: Oxford University Press.

QUIRK, R. 1990. "Language Varieties and Standard Language". English Today 21(6/1).3-10.

SAMUELS SON-BROWN, G. 1995. A Practical Guide for Translators. Clevedon:

Multilingual Matters.

SCHLEIERMACHER, F. 1985. "Über die verschiedenen Methoden des Übersetzens". Les

Tours de Babel: essais sur la traduction, 279-347. Mauvezin: Editions Trans-Europ-Repress.

SIMON, S. 1996. Gender in Translation: Cultural Identity and the Politics of Transmission.
London & New York: Routledge.

SINGLETON, D. 1992. "Second Language Instruction: the When and How". In Language
Teaching in the Twenty-first Century: Problems and Prospects. AILA Review (9).46-54.

SKUTNABB-KANGAS, T. & PHILLIPSON, R. 1989. "Mother Tongue: The Theoretical

and Sociopolotical Construction of a Concept". Status and Function of Languages and
Language Varieties ed. by Ulrich AMMON. Berlin: Walter Gruyter.

SNELL-HORNBY, M. 1992. "The Professional Translator of Tomorrow: Language

Specialist or All-round Expert". Teaching Translation and Interpreting: Training, Talent and
Experience ed. by C. DOLLERUP and A. LODDEGAARD, 9-22. Amsterdam &
Philadelphia: John Benjamins.

STEINER, G. [1975] 1992. After Babel: Aspects of Language and Translation. New York,
Oxford: Oxford University Press.

STERN, H. H. 1983. Fundamental Concepts of Language Teaching. Oxford'. Oxford

University Press.

STOLZE, R. 1994. ÜberSetzungstheorien: Eine Einfuhrung. Tübingen: Gunter Narr Verlag.

VENUTI, L. 1992. "Introduction". Rethinking Translation: Discourse, Subjectivity, Ideology

ed. by L. VENUTI, 1-17. New York & London: Routledge.

VENUTI, L. 1995. The Translator's Invisibility: A History of Translation. London & New
York: Routledge.
TERMINOLOGY POLICIES, DIVERSITY, AND
MINORITISED LANGUAGES*

Anna Aguilar-Amat & Laura Santamaria

Department of Translation and Interpreting (FTI)
Autonomous University of Barcelona, Spain

Zusammenfassung
Nach der Beschreibung der Unterschiede zwischen Minoritätensprache und
minorisierter Sprache, skizzieren die Verfasser die heutige Situation der Terminologie
in Katalonien, und zwar angesichts des Sprachenkontakts, der bis hin zur
Vermischung besteht. Diese Situation macht eine Erörterung der Funktion der
Terminologie notwendig: Ist sie auf die Standardisierung der Verwendung
ausgerichtet Anwendungen oder auf die Entwicklung von Besonderheiten der
jeweiligen linguistischen Gemeinschaft? Verschiedene Argumente sind vorgebracht
worden, aber offensichtlich lenkt der Status der betreffenden Sprachen die
Stellungnahmen. Einige Beispiele schliessen die Studie ab.

Résumé
Après avoir différencié langue minoritaire et langue minorée, les deux auteurs
esquissent la place de la terminologie aujourd'hui en Catalogne, quand il y a contact
de langues, parfois jusqu'au mixage. Cette situation oblige alors à s'interroger sur les
fonctions de la terminologie: vise-t-elle à standardiser les usages ou à développer les
spécificités de chaque communauté linguistique? Divers arguments sont avancés mais
d'évidence le statut des langues concernées infléchit les prises de position. Quelques
principes concluent la démonstration.

Resumen
Las autoras analizan la situación del catalan, lengua que hasta hace poco se
encuadraba dentro del grupo de lenguas minoritarias y que en la actualidad se
denominan lenguas menos favorecidas; las politicas terminológicas; la normalización;
la discriminación frente a otras lenguas en la investigación terminológica y la
necesidad de que las lenguas minoritarias del Estado se incluyan en los planes de
estudio de las Facultades de Traducción e Interpretación.
74 Aguilar-Amat & Santamaria

1. Minority languages and minoritised languages

The term minority language has proved to be problematic in recent years.

Some of the languages included in this category could be referred to as lesser-
used languages, such as Norwegian in the political context of Europe as a
whole. However, this article does not set out to deal with these languages.
Rather, we shall be considering the case of languages that are sometimes
referred to as less favoured languages, and which, until recently, were
invariably included within the group of minority languages.
In effect, the word "minority" implies "a small number". Minority
languages are languages spoken by a relatively small number of people. Thus,
the category is based on a statistical notion of the number of speakers. As
Pueyo (1992) points out however, the label tends to be applied to languages
which suffer from functional difficulties, and these difficulties are not always
linked to the demographic weight of the relevant speech communities but
rather to historical and socio-economic conditions and psycho-social
mechanisms. Catalan, for example, is seen as a minority language, whereas
Danish, which has a similar number of speakers, is not. Furthermore, we must
bear in mind that groups are not minorities, rather they become minorities. As
Michael Cronin (1995) points out:

The concept of "minority" with respect to language is dynamic rather than static.
"Minority" is the expression of a relation not an essence. The relations can
assume two forms: diachronic and spatial. (p. 86)

A language which is minoritised diachronically is one whose lack of symmetry

with respect to the predominant or "majoritised" language has come about as
the result of an historical event (for example, in the case of Irish, the famine
which took place in the 19th century). A language which becomes minoritised
spatially is one which loses its dominant position as a result of a reorganisation
of territorial borders (the case of the Russian language, for example, in the
Baltic Republics after the break-up of the former Soviet Union). However,
there are further cases in which the speakers of what is clearly a majority
language live quite evidently in a minoritised situation, for instance under
colonialised rule (see Miquel Pueyo 1992:21). In this light, we must bear in
mind that although many minority languages are or have been minoritised
languages, not all minoritised languages are minority languages. A number of
African and Amerindian languages are the language of the majority in their
territory, but they are dysfunctional as a result of the conceptual colonialism
which prevails there; in effect, they are minoritised languages.
Returning now to the point made by Pueyo, we shall draw a distinction
between the terms minority language and minoritised language (the latter term, in
any case, has now acquired something of a tradition at least in some Romance
 Terminology policies 75

languages); and we shall also distinguish between the terms majority language
and majoritised language. Viewed from this perspective, "all languages are
potentially minority languages" (Cronin, 1995:87-88). These concepts are also
relative, since a language can be majoritised in one context, for example,
Castilian in Spain, and minoritised in another, such as the position of Castilian in
Europe - only around 1% of texts produced in the European Union are written in
Castilian. In comparison with other labels such as expanding languages and
receding languages, the terms we propose here have the advantage of not
implying any given "direction".

2. Terminology in the Catalan translation tradition

Terminology is a discipline which has developed from a number of different

perspectives. The approach we take has stemmed from Translation Studies, and in
the contextual setting of our country (Catalonia), it defends multilateral action.
The process of linguistic standardisation, as has taken place with Catalan, aims to
give the linguistic system back its vitality via its own etymological,
morphological, syntactic and semantic systems. When a language's development
has been stunted by repressive historical conditions it is necessary to take steps to
enrich and lend cohesion to the system, and especially crucial is the
standardisation of vocabulary and terminology: a language without a scientific,
administrative and technological vocabulary has little or no chance of surviving
or competing in the modern world.
According to Auger (1987) and Cabré (1992:33), the three main trends in
terminology are as follows: (a) terminology oriented towards the linguistic system
(focusing on concepts and the standardisation of terms and notions); (b)
terminology oriented towards translation (database creation); and (c) terminology
oriented towards linguistic planning (systematic intervention with a view to
changing the status of a rather unstable language). Current (a) is represented by
three schools: Vienna, Prague and Moscow. The Vienna School is perhaps the
most widely influential since it has developed its own theoretical system of
principles and foundations for terminology based on the work of E. Wüster and
disseminated by means of the ISO standards.
Currents (b) and (c) have developed mainly in the Francophone world
(Quebec-Canada, Belgium) and in other countries where Romance languages
are spoken, including Catalonia. However, the work undertaken by the
Francophone countries has been monodirectional, in that the criteria drawn up
for the defence and expansion of French have not been applied to the
indigenous languages of their former colonies (minoritised majority
languages). It is our view that language rights can and should be defended
more democratically.
16 Aguilar-Amat & Santamaria

Apart from the work done in Catalonia by the various standardisation

bodies and the input of experts from a range of specialised scientific fields, an
important role has been played by the so-called Serveis Lingüístics and the
university Faculties of Translation and Interpreting, in which terminology is a
compulsory subject. However, it is impossible in the case of Catalonia to
separate the translation-linked current of terminology from the political and
social dimension, that is, (b) from (c). This is because the translation-linked
approach to terminology leads to a defence of linguistic, conceptual and
cultural pluralism as opposed to the range of globalising alternatives which
currently exist, whether artificial, as in Esperanto or Neokoiné, or natural as in
English. This implies a defence of the idea that every language must seek its
own terminological resources and these must be in keeping with the specific
nature of the linguistic system. The importance of lexical and morphosyntactic
coherence should not be affected by the fact that the language is either widely
spoken or minoritised, or whether it has a strong economic support base or not.
We cannot defend diversity and at the same time forget about the most
vulnerable languages, and their speakers, whose unique values rooted in the
past of their collective psychology are now so seriously threatened.
Although what we call the coercive dimension of terminology is
unusual in Europe outside Catalonia (with the historical exception of France)
and in America outside Quebec, it is nevertheless a worthwhile practice in all
areas in which there are languages in contact. The existence of such
pseudolanguages as pidgin in south-east Asia and the "Spanglish" found in
Puerto Rico is due to the total absence of the coercive dimension, or what Jean-
Claude Boulanger (1991), highlighting the conditioning or promotional
function performed by the organisms which apply it, calls l'axe
aménagemental. Linguistic dependence is the first and the last symptom of
conceptual, cultural and economic dependence. Some might well hold that
pseudolanguages, being the outcome of a need to communicate, are not in any
way objectionable, and indeed this is a reasonable claim. However, the
problem is not the acquisition of a new language as such; rather, it lies in the
discarding of a language that has underlain an enormous cultural tradition in
favour of a set of arbitrary signs. If, for example, we take the word rufo (from
the English word "roof') which is employed among the inhabitants of Puerto
Rico instead of its Spanish equivalent tejado, we can see how the logical
connection relating the Spanish tejado with teja (roof-tile) has been lost, not to
mention the links with the other members of this lexical group, such as
tejadillo (a small single-piece or lean-to roof section, as over a porch,
outbuilding or extension), tejar (a tile factory), tejuelo, and also such phrases
as jugar a los tejos, tirar/echar los tejos, a teja vana, a toca teja, de tejas abajo
and de tejas arriba.1 Nor does the fact that Puerto Ricans say rufo mean that
their "Spanglish" is necessarily enriched by the corresponding English lexical
family made up of such words and compounds as "roof-garden", "roof of the
 Terminology policies 77

mouth", "roof-rack", "roof-top restaurant", "roof of heaven", or such phrases

as "to hit the roof', "to raise the roof', and so on.
This regression of Spanish in relation to English is in broad terms
similar to the regression suffered by Catalan in relation to Spanish or Castilian.
In Catalonia and writing in Catalan, we tend to send letters which include an
acuse de recibo (Spanish for "acknowledgement of receipt") as opposed to an
avis de recepció (Catalan). Similarly, we tend to refer to long-term planning as
planificació a llarg plaç (directly calqued from the Spanish plazo) as opposed
to planificació a llarg termini - the autochthonous Catalan term; we sell to the
highest bidder, whom we call the millor postor (from Spanish postor: 'bidder')
instead of the Catalan major oferent. Bankrupt companies undergo suspensió
de pagos instead of suspensió de pagaments; gushing taps release a chorro
d'aigua instead of a raig d'aigua or a doll d'aigua, and we say that it is
forbidden to enter public parks with dogs which are sueltos (not on a lead or
loose - in Spanish) instead of gossos aviats.
Of course, the Catalans who speak "Catanyol" cannot be held
responsible for this state of affairs. We simply wish to point out one of the
reasons why we support the idea of speaking and writing as far as possible in
accordance with the intrinsic structure of the language: we believe that it is
much more natural for speakers to learn and use a lexeme, a number of rules
for word formation and a system of analogies than to memorise vocabulary and
concepts which do not have these natural lexical links. Direct borrowing from
other languages impoverishes the expressive capacity of the individual speaker
and the speech community in general. It is the coherence of the linguistic
system, with the links which it gives rise to, that shapes us as individuals,
regardless of whether or not it could be said to actually determine us; and
mastery of this system is the first step towards attaining a certain level of
freedom of thought, word and deed.

3. Homogeneity versus diversity: international standardisation

A defence of minoritised languages and their associated conceptual systems

comes into conflict with linguistic globalisation and the attempts to create a
single universal scientific language. Consideration of linguistic diversity and
the matter of one language's predominance or subordination with regard to
another brings us to the following dilemma which has repercussions on a
world-wide scale: should terminology aim to attain maximum homogeneity in
each specialised language and across different languages so as to facilitate
international communication? Or should it adapt to each individual language
and thus aim to guarantee maximum accuracy, utility and satisfaction in the
monolingual community even if this implies increased distancing from the
equivalent foreign terminology?
78 Aguilar-Amat & Santamaria

In favour of the first option, Arntz & Picht (1989:171) defend the
standardising role of terminology as a means of ensuring unequivocal
communication among specialists. Efforts to clarify the linguistic resources of
specialised fields are part of a long tradition. Versalius (1514-1564), the
founder of modern anatomy, can be considered one of the precursors.
However, it was not until the second half of the 19th century that
terminological work became systematic. This was mainly due to the
technological revolution and the disordered coexistence of a whole range of
technological discoveries. At first, in Germany there were 15 different units of
measurement for electrical resistance, and eight different units for motor
power. Inevitably, this situation gave rise to expensive mistakes. As a result,
the industrialised states began to see the need to standardise objects, procedures
and their associated terminology.
However, although these authors also use the word "standardisation",
the objectives of this process on an international scale are totally opposed to
those of the standardisation of a minoritised language. In the former case,
certain lexical options are imposed on other languages, whereas in the latter
case, the aim is to preserve and strengthen the autochthonous terms. Therefore,
we shall refer to this international level as "homogenisation" or "international
standardisation", as opposed to simple "standardisation", as it was translated by
Amelia de Irazazâbal.
In the area of homogenisation, the International Electrotechnical
Commission (IEC) was founded in 1906 and set about the task of clarifying
electronic terminology. The DNA (Deutscher Normenausschluss - German
Standardisation Commission) appeared in 1919, followed by the ISA
(International Standards Association). Wüster (1898-1977) published his work
Internationale Sprachnormung in der Technik, besonders in Elecktrotechnik
(International Linguistic Standardisation in Technology, especially in
Electronics) in 1931, which was based on de Saussure's structuralism. The
main objections raised against this work were the decisive role assigned to the
idea of system and the belief that language cannot be standardised. Wüster's
reply was that he was not referring to everyday language, but only to
specialised language. Wüster assumed that international standardisation cannot
be conceived as static, that it is not equally meaningful in all specialised areas
and that it cannot be applied with equal intensity in all cases.
The controversy aroused by Wüster was interrupted by the outbreak of
the Second World War. In 1946 the ISO (International Standards Organisation)
was founded to replace the ISA. In 1950 a technical committee was formed to
deal with terminological matters (ISO/TC 37 Terminology - principles and co-
ordination). The committee completed the drafting of the basic standards in
1973, and these standards have been undergoing a process of revision since
1974. Other bodies such as the IEC (International Electric Committee) and
Technical Committee 7, also work in conjunction with the ISO.
 Terminology policies 79

As regards the second option, projects to develop artificial lingua franca

were enlightened efforts to facilitate international communication. (See e.g.
Álvarez & Candel 1994.) At present however, projects such as Esperanto and
Neokoiné would seem to be Utopian and even counter-productive, given that
they are an attempt to introduce standardised elements and geometric
rationality into the realm of culture.
From our standpoint, respect for diversity and the individual
characteristics of different languages would seem to be the only universal
guideline which is valid today. Alvarez and Candel claim that this is the
consequence of earlier homogenisation campaigns which gave rise to a
reaction. "In general, the universalisation of cultural forms has consisted of the
simple destruction of traditional forms without any parallel creation of new
forms. As a result, those who see the survival of their deeply held values
threatened, react aggressively to the efforts to impose a single, standard
culture" (id. :417). Furthermore, we can justly ask how decisions of such global
consequence can be taken on the basis of the consideration of a diminutive
group of languages rather polemically referred to as the languages of culture.
This debate between the globalising and the diversifying approach has a
profound bearing on the standardisation and dissemination of terminology. At
present, the countries which produce technology also produce terminology. In
terms of market, these countries stand to benefit from a homogenising
dissemination of their vocabulary. In other words, scientific and technical
evolution is placed in the hands of the countries which control production and
terminology and which, in the long term, end up imposing their conceptual and
cultural systems on other countries. Whether this is a natural process of human
evolution or whether, on the contrary, it is simply a case of the hegemony of
certain cultures which are economically powerful over others which can be
seen as minorities, a hegemony which will lead to a general impoverishment in
the long run, is a question which should be pondered by participants in any
terminology project, with respect to translation or to language planning.
On page 1 of DIN 820 {Deutsches Institut für Normung - German
Standards Institute - founded in 1917 and based in Berlin), "standardisation" is
defined as follows:

Standardisation is the planned unification of material and immaterial objects

which the interested parties carry out for the benefit of the public. In no
circumstances should it be in response to a desire for individual profit.
Standardisation promotes rationalisation and the guarantee of quality in the
economy, science and administration. It is at the service of human and
material safety and the improvement of quality in all aspects of life. It also
serves to establish a reasonable order, to facilitate information in the
respective areas subject to standardisation.
80 Aguilar-Amat & Santamaria

However, in this definition, unquestionably drawn up with the best of

intentions by the German standards committee, there are a number of issues
which could be the subject of a wide-ranging debate. For example, despite the
fact that it is explicitly stated that standardisation should not be carried out for
profit, in practice however, terminological homogenisation often serves the
interests of certain manufacturers who, under the guise of a drive for
"rationalisation", "quality", "safety", "improvement" or "order" (all of these
highly subjective words appear in the paragraph quoted above), merely seek to
expand their markets. It is not clear just how many countries are included in
what is referred to as "the benefit of the public", though it would seem that the
non-industrialised countries in Africa, Asia and Latin America are excluded
from the benefits of terminological homogenisation affecting articles which
they cannot hope to be able to afford.
This issue is currently the subject of debate within international bodies.
The countries traditionally linked to such pioneering organisations in
terminographical and terminological research as Infoterm (Austria and
Germany), countries which are also producers and disseminators of technical
terminology, believe that the preservation of terminological homogeneity is a
central part of their international activity. Other countries, mainly those which
speak Romance languages2, observe this interest with suspicion in the
aftermath of a period in which the promotion of homogeneous terminology was
often interpreted as just another form of cultural hegemony. These countries,
normally receivers of technology and its accompanying terminology and
almost always poorer in resources, campaign for international terminological
work which respects linguistic and conceptual pluralism and which strives to
seek the most appropriate solution for each situation.
All standards bodies, both the international homogenisers and those
trying to halt the decline of a minoritised language, are concerned about the
extent of acceptance or otherwise of the proposed terminological standards and
lexical items. Bodies such as DIN have the legal status of registered
associations and cannot, therefore, impose their standards by "decree". And
despite the fact that organisations in other European states such as France have
wider powers, standardisation normally takes the form of "recommendations
from the experts". In the business world, non-compliance with certain
standards normally involves financial loss, which means that the standards are
generally accepted and acted upon.
In contrast, in terminological standardisation things are not so easy.
This is because language use, even in specialised areas, is a complex process
which is marked by psychological and sociological factors that cannot be
simply regulated at will. The fact that it is so difficult to accept proposed
terminology when it differs from established use suggests that there is a need
for a change of perspective in terminological discourse. In general, users of a
language are highly reluctant to use an established term with a new meaning or
 Terminology policies 81

to replace an already established term with a new one. In many cases, the new
term and the old term continue to be used alongside each other as synonyms,
and this of course can lead to ambiguity.
Neologism is a valid resource for certain new or emerging fields of
specialisation, but it stands little chance of success if it used in an effort to
replace the traditional conceptual and lexical system. This point must be firmly
borne in mind when dealing with minoritised languages.

4. Terminology and development

We now turn to languages which cannot resort to neologisms based on Greek

and Latin roots. Many countries with languages which use non-Latin alphabets
or whose linguistic tradition is oral, place their hopes in terminology as one of
their most important links to the international community, and at the same time
they see it as a means of achieving independence from English by creating
their own native terminology. This is the motivation which underlies the recent
efforts made in this area by such countries as South Africa, India and Israel.
However, in many countries which are lacking in resources, investment
in terminology is aided from the outside. The case of Cameroon is an
interesting example. In the first Terminology Summer School organised by the
Pompeu Fabra University in 1997, Marcel Diki-Kidiri (from CNRS-LLACAN,
Centre National de la Recherche Scientifique, Langage, Langues et Cultures
d'Afrique Noire) defended the use of "cultural metaphors" to reproduce certain
concepts in the oral Sango language spoken in the north of Cameroon. For
example, in order to translate a term such as "human rights", it is possible to
refer to the African custom whereby after hunting and killing an animal, certain
people have rights to certain parts of the dead animal (the head, for example,
goes to the chief of the village or to the oldest person, the left leg goes to the
owner of the weapon that killed the prey, etc.). Each of these portions of the
carcass is known as ngura. The term "human rights" is then translated as ngura
ti zo (// zo = of the person). This term is not entirely felicitous, however, since
the idea of people being equal in terms of their rights does not underlie the
ngura concept, which includes a clear hierarchy and differentiation. Similarly,
it is almost impossible to translate the expression "right to equality of men and
women". For a Cameroonian, and for Africans in general, the difference
between men and women and between people in general is a basic fact, one on
which their interpretation of the world is based. Without going into the social
consequences of this fact, it is nevertheless clear that in this context conceptual
homogenisation is not feasible. Should we not therefore include other rights in
our list of human rights, such as the right to be different, to think differently
and to hold different values?
82 Aguilar-Amat & Santamaria

Such problems are of concern to the approach known as "development

terminology", whose influence seems to be becoming increasingly powerful in
recent times. What worries practitioners here is the possibility that these efforts
to homogenise on an international level are solely aimed at achieving market
expansion through the creation of technological needs in developing countries,
and, consciously or unconsciously, at rejecting a role for the native languages
as equally useful and enriching creators and conveyors of concepts. For
example, returning once again to the point made by Marcel Diki-Kidiri, one of
the specific tasks facing African terminologists is the provision of accurate
information to cotton pickers on the application of certain pesticides in the
fight against predatory insects (cotton is the main source of wealth in the north
of Cameroon). It is necessary to create African neologisms to denote the
insects in accordance with the classifications of western entomology, since the
local farmers only have words for two types: those that fly ("fliers"), and those
that walk ("walkers"). There can be no doubt as to the importance of the cotton
harvest for the local economy, but nevertheless one cannot help wondering
whether it would not be possible and even more effective - though perhaps not
so lucrative - to manufacture and sell an insecticide for "fliers" and another for
"walkers" to these people. Is not rejection of their conceptual system perhaps
the most fundamental element of linguistic and cultural discrimination?
Each language is unique and unrepeatable, not only in its linguistic
aspects but also because it shapes the perception of the world, and the
individual and collective ways of being of its speakers. Just as biological
species disappear off the face of the earth, a language which dies out takes with
it valuable information which could contribute to explaining human culture and
evolution.
The defenders of cultural homogeneity reason that a few languages
more or less is not a matter of great importance. They consider it possible, even
desirable, to achieve a monolingual world. Ever since the European expeditions
which conquered the world at the end of the 15th century, multilingualism has
been perceived by some as a threat to the state, a recalcitrant voice raised
against the national destiny imposed by the dominant minority. Nevertheless,
multilingualism is part of the normal way of life of hundreds of millions of
people all over the world. However, fewer than a quarter of the world's states
officially recognise more than one language and very few indeed recognise
three or more official languages. There exists a widespread prejudice,
erroneously based on the theory of the evolution of species, according to which
the most widely spoken languages are in some way stronger and will therefore
survive and eliminate the weaker languages as part of a natural process. Before
going along with this prejudice, we should study the real causes which lead to
one language being imposed on another, and determine whether efforts to wipe
out languages and impose authority are in fact "natural".
 Terminology policies 83

Such is our lack of knowledge of humanity's linguistic heritage that we

do not even know exactly how many living languages there actually are in the
world. "The monolingual tends sometimes to believe that reality and its verbal
expression are one and the same thing, and this can therefore lead to a
dogmatic and absolutist way of thought, while those who are accustomed to
living through a number of languages and cultures find it easier to accept
divergence and ambiguity" (Miquel Siguán 1996:17). Multilingualism and
multiculturalism are rich resources and should be fostered by our education
systems, the media and society in general. Strengthening our cultural pluralism
is a fundamental first step towards increasing our tolerance and helping us to
advance towards "the culture of peace", in accordance with the resolution
passed at Leioa. By fostering linguistic diversity and protecting minoritised
languages, we can make real progress in the area of international understanding
and can avoid the development of a single, standard way of thinking.

5. Conclusions

We therefore propose the following principles for academic planning as

concerns minoritised languages. (1) There should be no discrimination against
languages in systematic terminological research. (2) The State's minority
languages should be included in the course syllabuses of the various Faculties
of Translation and Interpretation. (3) Seminars on minoritised majority
languages should be held, for example the lingua francas in Africa such as
Fulani, Wolof, Lingala, Bambara and Swahili, and in Latin America, including
Mayan, Quiche, Cakchiquel, Mam, Kekchi and others. (4) There should be
democratic interchange of knowledge, i.e. a two-way exchange of concepts. (5)
We must reject the patronising attitude implied by the new forms of cultural
colonialism. (6) Value judgements must be excluded from our exchanges of
concepts.

Notes

* Paper translated from Catalan by Carl Mac Gabhann

1. Tejo: a piece of roof-tile which is used as a toy. Tejuelo: a piece of metal which serves as a
support for a machine, a metal blank which will be made into a coin, or a piece of paper
attached to a book's spine for cataloguing purposes. Tirar/echar los tejos: to make verbal
attempts to seduce someone. A teja vana: covered in roof-tiles but without any interior
covering or ceiling. A toca teja: to pay in cash. De tejas abajo or de tejas arriba: of this
world or of a supernatural world.
84 Aguilar-Amat & Santamaria

2. United into networks and forums like Unión Latina, RITERM (Red Iberoamericana de
Terminologia), REALITER (Red Panlatina de Terminologia).

References

AGUILAR-AMAT, A., SIDIBÉ, P. &VILARNAU, J. 1997. "La terminologia del parentesco en

lengua fulfulde: un intento de sistematización". Nazioarteko Terminologia Biltzarra (Congreso
Internacional de Terminologia), edited by IVAP/UZEI, San Sebastián, 415-426.

ALVAREZ, P. & CANDEL, M. 1994. "Neokoiné: una alternativa a Babel". Terminologie et

Traduction 1/1994. 417-426.

ARNTZ, R. & PICHT, H. 1989. Introducción a la terminologia, translated by A. de Irazazábal et

al. Fundación Germán Sánchez Ruipérez. Madrid: Pirámide.

AUGER, P. & PARADIS, C. 1987, "La terminotique ou la terminologie à l'ère de rinformatique".

M e t a 32(2), 102-110.

BOULANGER, J.C. 1991. "Quelques composantes lingüistiques dans l'enseignement de la

terminologie". Actas del Coloquio Iberoamericano sobre Enseñanza de la Terminologia,
Granada. 61-74.

CABRÉ, M.T. 1992. La terminologia. La teoria, els mètodes, les aplicacions. Barcelon: Empúries.

CRONIN, M. 1995. "Altered States: Translation and Minority Languages". Traduction,

Terminologie, Rédaction 3(1), 85-103.

DIKI-KIDIRI, M. 1997. La diversité dans l'observation de la realité, la Escola d'Estiu de

Terminologia. UPF, Barcelona.

DUARTE, C, ALSINA, A. & SIBINA, S. 1991. Manual de llenguatge administratiu. Barcelona:

Escola d'administració publica de Catalunya, Generalitat de Catalunya.

ISO/R 860 Unification internationale de notions et de termes.

PUEYO, M. 1992. Tres escoles per al català. Lleida: Pagès, coll. Argent Viu.

ROJAS LIMA, Flavio. 1992. Los indios de Guatemala. Madrid: Mapfre.

SIGUAN, M. 1996. Aspectos sociopoliticos de la riqueza lingüistica europea. Madrid: Alianza

Editorial.

TERMCAT. 1990. La terminologia en el seu entorn, (series of lectures), Dep. de Cultura de la

Generalitat.
 PART III

Psychological/cognitive aspects
KNOWLEDGE, EXPERTISE, AND T H E O R Y IN
TRANSLATION

Candace Séguinot
York University, Toronto, Canada

Zusammenfassung
Der Artikel beschäftigt sich mit der Übersetzung im Licht der Erforschung von
Entwicklung, Anwendung und Institutionalisierung von Wissen. Es wird hier dem
Verständnis von der Natur der Probleme von der kognitiven zu einer mehr ökologischen
Sichtweise nachgegangen und vorgeschlagen, daß die Unterschiede zwischen
Übersetzungstheorien von einer prozeßorientierten versus wissensorienterten
Herangehensweise an Probleme herrühren. Die Termini expert und professional sind
manchmal synonym verwendet worden, aber die Unterschiede zwischen diesen beiden
Termini sprechen für den Bedarf einer eher soziologisch orientierten Untersuchung.

Résumé
Le point de vue sur la traduction adopté ici est celui du développement, de l'utilisation
et de l'institutionnalisation des savoirs. Le traitement cognitif de la compréhension des
problèmes cède alors le pas à une approche plus écologique. Une ligne de partage est
proposée entre les différentes théories de la traduction, selon que la résolution des
problèmes y est perçue comme processus ou comme savoir. Parfois considérées comme
négligeables, les différences entre savoir-expert et savoir professionnel incitent à
donner plus d'assise sociologique à la recherche.

Resumen
Este trabajo enfoca la traducción desde el punto de vista del desarrollo, uso e
institucionalización del conocimiento. Parte de una comprensión cognitiva de los
problemas hasta el desarrollo de una perspectiva más ecológica. Concluye que las
diferencias entre las distintas teorias de la traducción radican en si se aproximan a los
problemas basándose en el proceso o en el conocimiento. En ocasiones se han utilizado
los términos experto y profesional como sinónimos pero existe una diferencia entre
ambos que apunta a la necesidad de una investigación basada en un enfoque más
sociológico.
88 Candace Séguinot

1. Introduction

The impulse behind this paper, written and delivered as a talk, was to look at
translation in the light of research into institutionalized forms of knowledge.
How do people learn to translate? How do people learn to translate well? You
might think that we could find the answer by studying the people who have
made a career of translating, but it is difficult to use professionals as subjects.
Because so many of their processes have become automatic, they lose
themselves in their work, forget to verbalize their thoughts, and their expertise
is recorded only in the text they have created. In a testimony to the subliminal
flow state described by Douglas Robinson (1997b:95), a professional translator
once told me that he is completely unaware of what he will say until he sees his
own words on the page. One kind of information that is available for the study
of the translation process is what happens when things break down, when there
are difficulties that interrupt the process. It is then that we learn about
knowledge and can identify signs of expertise.

2. The nature of problems

To understand the development of procedural knowledge, I would like to step

outside the world of translation for a moment and suppose that I am in my
kitchen. And let's suppose that I'm helping my niece. She has a problem: she
needs to bring dessert to the school picnic tomorrow. That's what I'm doing
here, helping her to bake cookies. We've already got out the ingredients for the
first batch, and we're just. . .

What's that honey? It says to cut the butter into the flour? You need two knives
to do that . . . wait a minute, I think I've got a special tool...where did I put
it...it looks like a little rocker with blades...actually, we could have done this
part in the Cuisinart. No, you can't use a blender. No matter, we can just use
the knives, see...like this. That's right. Now blend in the rest of the ingredients,
and . . .

And the first recipe is done, and we're going to start on a second. This is my
grandmother's recipe for Coconut Peaks. You won't believe the secret
ingredient...it's made from mashed potatoes. Uh oh . . . I don't have any
leftover potatoes. What do you want to do, sweetheart, we can make something
else or. . You really want to use Nana's recipe? O.K. We'll just go to the store
and get some potatoes. Hey, while we're here, let's pick up some ice cream.
I've got a 20% off coupon so we might as well use it. Wait a minute, that can't
be right. You only charged me 50 cents for the ice cream...it should be at least
a couple of dollars. Well yes, I realize there's a long line in back of me. No, I
 Knowledge, expertise, and theory in translation 89

don't want to cause trouble. Well, you're the one who's going to be in trouble
because you're losing money on that ice cream . . . All right, all right.

So I pay the cashier, and we can leave, and return to the reality of the written
page. No problem. Now what do I mean when I say that? That in all likelihood,
you already know that you are sitting in front of a written page. You do not
have difficulty imagining yourself in that situation.
But now, let's return to my niece and her problem. The difference
between the problem in 'no problem' and my niece's situation in the kitchen is
that her 'problem' was a task to perform. She had to bring a dessert to school,
and the culture dictates, home made. Had she been older, and in her own home,
and I in hers, there would still have been a task, but no problem. However in
this version, she was not in her home, she was in mine: the problem was that she
needed my permission, perhaps even my supervision and certainly my guidance,
to use my kitchen and work with my recipes. 'The' problem turned out to have
several different component problems, each of which either inhibited or
interrupted the task.
There are two routes to understanding the nature of task-oriented
behaviour, what strategies lead to success or failure, what kinds of experience
best contribute to the development of expertise, and so on. The first is to start
with an understanding of particular situations, such as this one, and generalize
from there. The second is to look at what has been written about problems and
tasks.

2.1 Problems and their relationship to knowledge

Let's start with the second approach and look at interrupted tasks from the
perspective of a cognitive psychologist. Richard Mayer (1989:39-41), for
example, wrote:

A problem solver has a problem when a situation is in one state, the problem
solver wants the situation to be in a different state, and the problem solver does
not know an obvious way to eliminate obstacles between the two states. In
short, a problem consists of three components: the given state, the goal state,
and obstacles that block movement from the given to the goal state . . .

From the perspective of translation, problems have been defined with much the
same flavour in the empirical literature. Wolfgang Lörscher (1991:79-80), for
example, cited a different source defining problems in essentially the same way.
The programming model of human thought from which it stems is obvious;
obvious, too, its influence in shaping process models of translation and
interpretation. And not surprisingly, as the behavioural sciences have been
greatly influenced by the claim that all human thinking is information processing.
90 Candace Séguinot

For many cognitive psychologists a decade ago, the actual processes needed in
artificial intelligence were a source of information about human intelligence.
Artificial intelligence does not have an intelligensor, so information
processing could be studied separate from the processor taken as a whole. In
this view of the world, hypotheses can be generated about any processes, the
unobservable as well as the observable. This meant that the mind was once
again allowed to be an object of speculation, and from there it became central.
Knowledge became key. In fact, from the point of view of meme theory, people
have been described as incidental to the ideas that are housed in them, mere
libraries. The ideas live on long after their unsuspecting hosts pass on. A
problem, from the cognitive perspective, equates to a need for knowledge. Not
knowing how to get from where you are to where you want to be explains why
processes are difficult or unsuccessful.
Not knowing implies a view from the inside, speculating on what goes on
in my niece's or the translator' s head. What is outside is modeled in our minds
in a variety of ways, through networks or hierarchies of associations, frames,
schema, etc. Problems, for the cognitive psychologist, do not actually exist "out
there". It is our perception that identifies something as a problem. In other
words, it is the construct of an individual.
In contrast to this psycholinguistic approach, there are a number of
schools of Translation Studies that can be categorized as focussing on the
knowledge side of the equation, i.e. that try to identify what a competent
translator needs to know. They look at problems in a different way. The
understanding of what constitutes a translation problem seems to take one of
three forms. The first, for linguistically-oriented scholars like Vinay and
Darbelnet (1958) and Catford (1965), is in the structural differences between
languages. The second is an intellectual construct: you compare a source to a
target text, you match up bits that appear to be related, and you label what
appeared in the source text a problem and in the target text a solution
(Chesterman 1997). The third knowledge-based approach is the one exemplified
by Mary Snell-Hornby (1988:52 and 96), who spoke of problematic situations,
as when words have different meanings in context than in isolation or when a
linguistic expression is multidimensional. Christiane Nord (1997:64) has most
recently articulated the difference in perspective between the internal and
external views. Outlining the benefits of source-text analysis, she rejects the use
of the word 'problem' for what she calls the subjective difficulties encountered
during a translation session.
So on the one hand we have the process-oriented researchers, like
Lörscher, Tirkkonen-Condit, Jääskeläinen, and several others, including myself,
who study the individual and work toward finding objectively verifiable ways of
identifying problems that are realized in the course of translating sessions; and
on the other, linguistic, text, and culturally oriented studies whose authors'
 Knowledge, expertise, and theory in translation 91

interests are in the shared habits and norms of a translating population.

3. Translation theory and the notion of progress

One explanation for the differences is, as Chesterman (1997:42-46) points out,
our different starting points. Looking at a particular question will lead to one
answer, starting with a different question will lead to another. For as Karl
Popper describes the world of scientific enquiry (1989:47),

Observation is always selective. It needs a chosen object, a definite task, an

interest, a point of view, a problem. And its description presupposes similarity
and classification, which in their turn presuppose interests, points of view, and
problems. 'A hungry animal', writes Katz, 'divides the environment into edible
and inedible things. An animal in flight sees roads to escape and hiding places .
. . Generally speaking, objects change . . . according to the needs of the
animal'. We may add that objects can be classified, and can become similar
and dissimilar, only in this way by being related to needs and interests. This
rule applies not only to animals, but also to scientists.

Yet to point out the lack of coherence is to miss the progress on the descriptive
front where it was formerly assumed that linguistic difficulties could predict
difficulties in the translation process. The pivotal change from the view of the
translation problem as external to the individual to the data-driven re-orientation
of that understanding can be seen in a study published by Hans Krings in 1987.
In the section on methodology Krings began (1987:162),

In addition to 'ordinary' grammatical, semantic and stylistic translation

problems found in almost every type of text, these texts included puns,
metaphorical expressions and other instances of literary finesse. I wanted these
to be included because I assumed that the structure of the translation process
would depend on the type of translation problem . . .

On analyzing the data, however, he realized that only one-fifth of the problems
were common to all his subjects, and over half the problems were idiosyncratic
(1987:167-169). Furthermore, the fact that 90% of the problems were lexical
spurred other researchers to ask whether that figure wasn't more a function of
the subject population (he was studying second language learners) than an
insight into what was truly problematic for translators. This was the beginning
of an empirical focus on the question of whether translation as practised by
professionals is different from the translating of the more inexperienced, which
was an important step in understanding the nature of translation competence.
For as Gideon Toury (1995:238) has said,
92 Candace Séguinot

It is . . vital to give some thought to the process whereby one type of

'translation competence' evolves into, or is perhaps replaced by the other. What
we need to know, in other words, is not only what it takes to perform
translation, but also what it takes to become a translator.

This latter point, the question of whether socio-cultural factors contribute to the
development of translators who will be successful as professionals, leads to two
questions which will be raised again later in the paper. The first is whether there
is a clear understanding of the difference between translator competence, or
expertise, and professional competence, meaning the ability to assume a position
in the translation marketplace. The second is whether current empirical research
methods which depend on psycholinguistic methods can be enhanced or
complemented by other avenues of research to explore these socio-cultural
factors.

4. Developing translator competence

4.1 Going beyond the given

We have begun to tackle the notions of competence and professionalism by

looking at the nature of problems, how they have been viewed in the
psychological and in the translation literature. We started with one definition, a
'not knowing'. But now it is time to return to the kitchen experience to see if
knowledge explains all the problems we encountered. And of course it does not.
You will remember that when I started the second recipe, I noticed that one of
the ingredients was missing. If I had checked beforehand, or if you had looked
at the contents of my refrigerator with a copy of my recipe in hand, you would
have been able to predict that I was going to have a problem. To produce an
ideal translation, translators need knowledge, yes, but they also need access to
information. I'll return to this distinction a little later. Documentation is the
most frequent source of information, but not the only one. Looking at an object
or an image of that object, manipulating software, in other words coming into
contact with a product that is being described, can be even more important. A
translator needs to be able to verify the meaning of a text and at times look for
meaning elsewhere. To a certain extent it is the profession itself that creates the
expectation that translation can be carried out with only the translator's
knowledge and a text. It does so through a particular role, a role referred to by
Douglas Robinson (1997a) as "the invisible hands that control translation".
What the status of the "invisible medium" does is contribute to what Daniel
Simeoni (1998:27) has referred to as "imposed submissiveness".
The missing element, then, is one of the problems that the professional
 Knowledge, expertise, and theory in translation 93

translator has to face on a regular basis. How someone comes to learn to cope
in the face of the unexpected is one of the hallmarks of expertise. The paradox
is that in order to help someone develop their expertise, we need to make clear
the opposite, the expected. Problems come in two types: the routine, and the
unfamiliar. Routine problems are those for which the solutions are ready-made.
That does not mean that the novice necessarily knows the solution, but that the
procedure for arriving at a solution is well-known.
What role does academic instruction play in establishing those routine
connections? We know we contribute to the development of good translating
habits by showing students how to use documentation correctly, we know we
encourage automatic correspondences through exercises in comparative
stylistics. These are activities that have built-in repetitions. What is less clear is
how material we present in the form of facts, for example types of approaches
for different text types, gets transferred into procedural knowledge.

4.2 Differences in levels of expertise

To begin with, there is necessarily a form of learning before a habit is

established. A beginner needs to recognize that an instance is similar to or
different from an instance that was previously encountered. A fact must appear
to be in the same form, a situation have the same characteristics, a rule the same
form and domain of application. To foster that development in class, we focus
on particular issues and we simplify what might be a workplace-encountered
experience by providing examples that reinforce recognition.
What the research on experts and novices (Dreyfus and Dreyfus: 1986)
suggests is that there are several stages in the acquisition of expertise, and that
from this first matching approach (same or different) which is conscious, a
learner develops the ability to store some of the information in order to free up
conscious memory for new learning. As competence develops, the learner is
able to choose between alternatives and do so unconsciously. The greater the
development of expertise, the more the choice of procedure becomes intuitive,
and from the situation-bound understanding of the novice we have a form of
understanding that is context-free.
Now anyone who has taught translation knows how hard it is to see the
progression. What masks the development of progress is the inevitability of
error. People make mistakes in areas of language use for three reasons: (1) They
have not learned a regularity or rule. (2) They have not recognized the domain
of application of a regularity or rule. They need more practice. (3) They are not
paying attention.
94 Candace Séguinot

4.3 What professionals have to say...

Whenever we focus attention on one aspect of translation, we can expect some

backsliding in others (Séguinot: 1990). It is not the case, as Brian Harris once
tried to argue with me a few years ago, that professional translators never make
mistakes. What happens, at least from my own research, is that professionals
attend to aspects of their production as they are translating and are able to catch
and correct the effects of language interference. This attending aspect is
different from the matching of the beginner. It is a monitoring for the
unexpected that sends a signal when things are somehow not right. The classic
example from everyday experience is the grandfather clock. It ticks out the
minutes day by day, and you are so used to it that you no longer hear it until it
stops ticking, and then you say, "Did you hear that?"
The situation with my niece in the kitchen is analagous to the structuring
that goes on, consciously or unconsciously, in more formal teaching situations.
If I were baking on my own, I would not organize the session the way I did with
my niece. Normally when I cook I take out the ingredients as I need them, but I
had her take out all the ingredients and do all the measuring before we began
the steps of combining and mixing. In courses where students are expected to
write papers, it is not unusual for some instructors to ask students to write an
outline before they start their essay. In translation courses, it is not unusual for
some instructors to ask students to analyze their source text in specific ways
before they begin their translation, read the whole text through before they
begin to translate, etc. Is there any evidence that these preparatory exercises
reflect the way we work ourselves?
There are, of course, three answers: the instructor who feels guilty
because he or she feels this is what they should be doing; the instructor who
believes that certain approaches help intervene in the student's process of
development to further the process along, regardless of whether they will later
be integrated or abandoned; and the instructor who truly believes that he or she
understands his or her own translation strategies and can pass this knowledge
on to students.
Donald Schön (1983:51), an influential author of books on the nature of
professional practice, has said that this last position, the professional teacher
believing they know what they do when they perform, is not necessarily the
case. Professionals may think they know how it is they perform, but what they
are doing is reflecting on their performance. That reflection is not in itself part
of the procedure. As Gilbert Ryle (1949:32) has put it,

What distinguishes sensible from silly operations is not their parentage but their
procedure, and this holds no less for intellectual than for practical
performances. "Intelligent" cannot be defined in terms of "intellectual" or
"knowing how" in terms of "knowing that"; "thinking what I am doing" does
 Knowledge, expertise, and theory in translation 95

not connote "both thinking what to do and doing it."

You might think the answer would be to study professionals in action to see
their strategies first-hand. Yet even as I carry out this kind of research, I know
that you cannot have the desired effect - the integration of a description of
practices into the practice itself - because of the developmental stages involved
in acquiring expertise. For reasons that will become clear later, the nature of the
expert's knowledge, as well as the processes themselves, is different from that
of less experienced learners.

5. Expertise = multiple goals and alternative strategies

The point of looking at the problem of the unexpected is to see that experts
handle such situations in ways that distinguish them from the inexperienced, and
they do so in ways that may have little to do with what they are capable of
explaining. There is a second developmental issue in my kitchen example that I
would like to turn to now. As an adult, I feel the pressure of time. It would have
been more productive to switch to another recipe than to go to the store if my
only goal had been to make cookies. But real life, as you know, is messy. As
adults, we have to learn to deal with multiple goals. In this case, I was making
cookies, but I saw this as a way to spend quality time with my niece. That, for
me, was the more important of the goals, and so I let my niece decide what we
would do. Productivity is not a child's goal, and children do not readily
compromise, so we stuck with our original plan and went to buy the missing
ingredient.
Productivity is not necessarily a byproduct of expertise, though in the
field of translation we tend to conflate the characteristics of the professional
with those of expertise. Anthony Pym (1994:2), for example, includes a
reference to expert status in his explanations of the use of the word 'expert' in
discourse about translation. What does characterize the behaviour of the expert?
To return to my everyday world example, my expertise in the kitchen, at least
relative to my niece, showed itself in two ways. I had an arsenal of recipes at my
disposal; my niece did not. I also had more than one strategy for accomplishing
a part of the procedure. If we go back to the bit where she questioned me about
how one went about cutting butter into flour, my answer contained more
information than a simple instruction. I told her indirectly that there were three
ways we could go about cutting in butter: using two knives, using a special tool,
or using a kitchen machine. So in addition to knowledge, my experience has
taught me that there are often alternatives.
In the literature on procedures, a distinction is made between what are
called toolbox skills and algorithmic skills. In the case of a toolbox skill, you
96 Candace Séguinot

have a problem so you reach into your toolbox and pull out a tool, one of many.
In the case of an algorithmic skill, there is only one formula: learn it, and you
can arrive at your solution. Translation is a toolbox skill, and the learner needs
to be at a given level to take advantage of the offering of alternatives. Because
the beginner tends to focus on the lower level problems, however, there is an
opening up that comes from seeing how others have handled the same material.
In an article on creativity in translation, Paul Kußmaul (1991:98)
stresses the need to ensure that students always preserve a critical attitude
toward the ideas that come into their heads. During the incubation and
illumination phases "they must be able to stand back, as it were, and observe
what they have been doing." My store of experience leads to my facing novel
experiences with a set of expectations. When I went to the grocery store, for
example, I had an idea of how much my groceries should cost. No matter that
the total was significantly less than I expected, that it was to my advantage from
a purely monetary perspective. I felt moved to say something. I put aside the
task of buying groceries for a moment, and in so doing I created more problems
for myself by opening myself up to hostility on the part of the cashier and the
people in the line behind me. One of the ways that we differ from computer
programs is that our monitoring includes a scanning for ethical and aesthetic
implications. To return to the definition of a problem, it seems unlikely that we
store our responses to moral dilemmas as individual instances that we recognize
on subsequent encounters. How, then, do we account for the ways in which
ethical and aesthetic norms affect the performance of a task?

5.1 The search for ecological validity

One answer lies in the theories of behaviour that have developed from a
recognition of the inadequacies of the cognitive view. No one would deny if we
have a task to perform, whether to imagine ourselves here or elsewhere, to bake
cookies or to translate a text, the 'how to' takes place in our head. We rely on
our previous knowledge, we think about alternatives, and we come to decisions.
This is the arena that Mayer (1989:42-43) is addressing when he explains the
nature of problem solving :

Problem solving can be analyzed into two major phases: representation and
solution. Representation involves moving from a statement or presentation of
the problem in the world to an internal encoding of the problem in memory.
More specifically, representation involves mentally encoding the given state,
goal state, and legal operators for a problem. Solution involves filling in the
gap between the given and goal states. More specifically, solution involves
devising and carrying out a plan for operating on the representation of the
problem...The representation and solution processes may be analyzed,
respectively, as building a problem space and using procedures for searching
 Knowledge, expertise, and theory in translation 97

for a path through the problem space.The input to the solution process is a
cognitive description of the problem (including the given state, goal state,
intermediate states, legal operators) and the output is a search path (i.e.,
application of a series of operators) between the given and goal states. The
search through the problem space (i.e., the connections among the given, goal,
and all intervening states) is controlled by a set of procedures, such as a means-
ends analysis strategy.

But neither the assignment of the task nor the attempts to deal with it take place
in a vacuum of mental processes. To carry out her task, my niece looked to me
for help; I asked for her opinion. In the first instance, there was an obstacle, and
it was a lack of knowledge. My niece looked outside herself, to another person,
for the information that she could not understand. In taking her wishes into
account I was acknowledging our relationship and the importance of that
relationship in the determination of how I was going to act.
One could, of course, build a model that allows for an interpersonal
component. As Donald Norman points out (1985:310), however, there is a
fundamental objection to the mind-as-machine approach. This is its "lack of
consideration of other aspects of human behaviour, of interaction with other
people and the environment, of the influence of the history of the person, or
even the culture". The search for models that more accurately conform to our
common sense understanding of how people operate, referred to as the search
for ecological validity, has meant a movement away from mechanistic
explanations involving chains of cause and effect and a move toward more
complex, decentralized explanations (Resnick:1994). Mitchell Resnick offers the
analogy of a flock of birds. They fly in formation, changing direction as if to a
hidden signal, but there is no leader to give the signal.

6. Knowledge as a social construct

We know that a successful translator is concious of the receptivity and reactions

of the target audience. Some of that recognition may be stored in the form of
knowledge, but each new client and every new assignment brings with it the
potential for interaction and novelty. What is stored, then, is the knowledge of
what clients in the past have been like, of successful strategies for interacting
with clients and with resource people. We do store information about people
and events. However, it cannot be that we comprehend the world around us
solely by matching the known to the unknown, because there is a wealth of
evidence that we act not on the basis of an objective reality but on our
expectations of that reality. In addition, our perceptions of the world we come
into contact with are coloured by our attitudes and beliefs. An example that
comes to mind is Janet Fraser's (1996:92) study of two groups of professional
98 Candace Séguinot

translators. She looked at what the translators said about how they were
translating the source text and concluded that the differences could be attributed
in part to their degree of involvement with the target culture: one translator was
involved in the legal system, another a volunteer with a refugee association.

6.1 Professional status and expertise are not the same

How does my experience in the kitchen relate to the behaviour of professional

translators? I am not a professional baker, but I have more expertise than my
niece in the kitchen. Why do I say 'but I have expertise', as though my
competence as a baker depends on my professional status? If we substitute
'translator' for 'baker', we return to the confusion between competence or
expertise and profesional status alluded to before. As an illustration of the use
of these words as alternatives, here is a recent statement by Chesterman
(1997:147): 'The contemporary conception of the translator's role is that of an
expert, someone with expertise, with professional translational competence."
On the one hand, we seem to see the professional as the consumate
expert. Yet empirical studies of translation have shown that some professionals
do not show greater competence than some students (for example, Krings 1988,
Tirkkonen-Condit 1989, Jääskeläinen 1989, Lörscher 1991). This has led
researchers to speak of successful and unsuccessful strategies rather than
assuming that the status of the subject, that is the degree of experience, will
always correlate directly with the quality of the translation. If we look again at
the psychological research on experts and novices, this may not be so surprising.
That research shows (Hoffman 1992) that in experiments, students in the
professions sometimes did better than their teachers on written tests, but rarely
performed to the level of experts in actual job situations. Why this should be so
helps us understand what we can expect to be able to pass on to our students in
the classroom and what has to come with experience. We expect and are
expected to be able to pass on knowledge. But knowledge is not out there, like
a fruit to be harvested. What is out there is life: people, objects, and events. Our
previous experience helps us organize the stream of stimuli around us into
recognizable patterns. Whether we see particular patterns in terms of
information depends both on our experience and our current needs. How many
of us have heard speeches in a language we translate from or into and been
struck by a particular expression. "So that's how you say that in Japanese", you
think to yourself, and the moment passes. And then that term shows up in a text
to be translated, and it becomes a piece of information.
 Knowledge, expertise, and theory in translation 99

6.2 Knowledge as organization

Knowledge is not simply an accumulation of facts; it is information that has

been socially or biologically integrated. It is constructed by the individual
interacting with a given context and his or her own history of experience.
Knowledge is not necessarily retained in its original form. It is in the
nature of experiential learning that we reconstruct our knowledge in the light of
our current understanding. As we gain more experience, two things happen.
What once required the recalling of certain procedures becomes automatic as
we establish routines. And with more experience we can see patterns emerge
that allow us to make use of the principles of our profession, the academic kind
of learning, in a new way, recognizing what is recognizable in each new
situation. What marks the expert in a particular field, then, is more than the
acquisition of a body of declarative knowledge, and more than a body of
procedural knowledge. Expertise stems from the integration of experiential
knowledge with the principles of the discipline.

7. The nature of professional knowledge: A historical explanation

But what about the difference between expert knowledge and professional
knowledge? For that we need to change from a psychological to a sociological
perspective; we need to look at the development of the professions. Expertise in
the form of a commodity is a relatively recent notion. Prior to the industrial
revolution, people relied on authority, and their own common sense. The
division of labour and the densification of population with the growth of cities
made it possible for people to sell their advice and services, and from there it
became possible for people to support themselves through this form of
"commodification". The expectation of payment became combined with a
degree of autonomy. There was a recognition within society that a certain form
of learning was necessary for the successful performance of services, and
mechanisms were developed to ensure exclusivity to rights of practice, notably
through some form of certification. The combination of these factors served to
define a group of practitioners as professionals. Today when we think of
professional translators we tend to think of these outward aspects still: people
who support themselves by translating.
Meanwhile the learning that was necessary for the exercise of the
professions had to be passed on, and in this century it became the job of the
universities. In return for the rewards of being absorbed into institutions of
higher learning, the professions had to accept the positivist ideology that is the
hallmark of the modern university: science leads to progress, and a better world.
But there was a problem when it came to practical knowledge, because
100 Candace Séguinot

it could not be investigated by either of the two standard scientific methods,

direct experience through the senses or logical deduction. The solution was to
look at the professions in instrumental terms. What the expert practitioner
needed to know was the relationship of means to ends. Science could help in the
examination of causal relationships, which would provide the means for
evaluating the best means to specific goals. As the professions were accepted
into the university sphere, then, they adopted the epistemology of practice
which Donald Schön refers to as the model of Technical Rationality.
The participation of professionals in the university led to a hierarchy in
which the scientist-scholars were expected to produce the basic research which
would then be used in the applied forms of the discipline. The clearer the goals
of the discipline, the more likely that research could contribute to an
understanding of the best means to achieve them. For example, in the case of
medicine, the goal is understood to be good health. The professions are
supported by what are seen as more fundamental disciplines providing a base to
prepare the would-be practitioner, before he/she undertakes courses in the more
practical procedures. The professional part per se comes in the form of
knowledge that is based on the day-to-day activities in the practice of the
profession. And as that is passed on overtly, so covertly are the beliefs and
attitudes defining the profession's view of the proper relationship to the public it
serves, and in some cases the nature of the role of the professional in
marketplace relations.

8. Research and training, expertise and professional performance

It becomes clearer why there should be so many approaches to translation

theory when we look at translation in the light of this explanation of the
ideology of research in the professions. What we have in the case of translation
are a multitude of ends, some linked to the notion of expertise, and others more
obviously tied to the pragmatics of the workplace. By way of illustration: one of
my professional colleagues, himself a teacher and a researcher, was asked to
take part in a study of translation. When asked how he woud translate a given
metaphor he asked in return, "How much time do I have?" Obviously for him
the constraints of the workplace have led him to prioritize his use of strategies
with productivity in mind, a hallmark of the successful professional.
To return to Gideon Toury's question about whether there are two kinds
of competences, what it takes to perform translation and what it takes to be a
translator, the answer, I think, lies here, in the social, as opposed to the
intellectual side of the performance. If we look at classroom practices more
closely, not just the types of material we cover, literary or pragmatic, but at the
structuring of the learning experience, we may be able to separate out some of
 Knowledge, expertise, and theory in translation 101

the factors that contribute to the development of the professional as opposed

now to the expert skills of the professional translator. The factors we might
want to study include the following.

(1) The types of deadlines that are associated with translation assignments and
translation tests.
(2) The types of workplace simulations that are built into the programme.
(3) The weight given to the surface manifestations of professionalism:
requirements that assignments be word processed, be in on time, comply with
particular standards, etc.
(4) The reinforcement of routines through relevant forms of practice (e.g.
building and managing a personal term file).
(5) The regulating of curriculum offerings on academic or professional grounds.
(6) Whether Schools of Translation and Interpretation regulate themselves with
a governing body or charter that imposes particular kinds of courses.

These are just some of the ways that we define the academic and the
professional in setting out our interpretation of translation goals. This says
nothing about the preparation of students for their role as translators, though
classes play a part in this aspect as well. We could say that translation as a
discipline is, to the extent that it is directed toward the professional, concerned
with the transmittal rather than the construction of knowledge. To that extent it
reflects the reality of the institutionalized practice. The individual translator
working in a commercial setting has very little power (see Pym 1997:100 and
Robinson 1997b). The norms with which today's translators operate emanate
from the institutions in which or with which they work, and not only do those
institutions differ in their expectations, their expectations are changing to
incorporate the less obviously translation-specific activities required to produce
camera-ready copy and text suited to both print and screen. The importance of
contextualizing translation in order to predict the norms that will surface makes
it difficult for empirical studies to isolate particular variables for study. We
know instinctively that experimenting with short, artificial corpuses as is done in
psychology would not reflect the reality of the translation workplace (Séguinot
1996b). So while process studies appear to be a promising window into the
nature of professional expertise, they are constrained both by the rigor of
scientific methodology and by the notion of ecological validity.

9. From problem-solving to problem-setting

So much the better, Jean-René Ladmiral (1987:193-194) has argued, if

translation were to abandon the search for a coherent theory that takes all
102 Candace Séguinot

aspects of translation into account. Bound by epistemological and rhetorical

constraints, the systematic descriptive approach produces a discourse oriented
to explaining existing rather than future translations, and therefore, he argues, it
ill serves the needs of practitioners for whom conflict and ambiguity are
everyday realities.
But there is another way out, and that is the type of research that reflects
the reality of the practising professional. This is essentially the position of
Donald Schön, who rejects the conclusion from which we started, that end-
means relationships are what drives the success that we equate with expertise in
the professions. In Schön's words (1983:39-40),

Increasingly we have become aware of the importance to actual practice of

phenomena - complexity, uncertainty, instability, uniqueness, and value-
conflict - which do not fit the model of Technical Rationality. . . From the
perspective of Technical Rationality, professional practice is a process of
problem solving. Problems of choice or decision are solved through the
selection, from available means, of the one best suited to established ends. But
with this emphasis on problem solving, we ignore problem setting, the process
by which we define the decision to be made, the ends to be achieved, the means
which may be chosen. In real-world practice, problems do not present
themselves to the practitioner as givens. They must be constructed from the
materials of problematic situations which are puzzling, troubling, and
uncertain.

The question of how we can help our students become problem setters, as well
as problem solvers, brings us back to the topic with which I began, namely the
problem of problems. The more we structure classroom practices around what
commercial interests see as problems - the need for specialized vocabulary,
familiarity with technology, and productivity - the more we need to introduce
and therefore reward activities that tend to conform. With this in mind I would
like to close with some words from David Bohm, a physicist who wrote about
the relationship between art and science, and who thought a great deal about the
distinction between the mechanical and the creative. He wrote that (Bohm
1998:4)

. . . fear of making a mistake is added to one's habits of mechanical perception

in terms of preconceived ideas and learning for specific utilitarian purposes. All
of these combine to make a person who cannot perceive what is new and who is
therefore mediocre rather than original.

The irony is that research and education that are tied to producing
specific kinds of knowledge, to meet needs specified by the profession, may not
serve the profession as well in the long run as less instrumental approaches that
bring both the profession and the discipline back to question the nature of
 Knowledge, expertise, and theory in translation 103

translation expertise.

References

BOHM, D., ed. by NICHOL, L. 1998. On Creativity. London & New York: Routledge.

CATFORD, T.C. 1965. A Linguistic Theory of Translation. London: Oxford University Press.

CHESTERMAN, A. 1997. Memes of Translation. Amsterdam & Philadelphia: John

Benjamins.

DREYFUS, H. L. & DREYFUS, S.E. 1986. Mind over Machine. Oxford: Blackwell & New
York: The Free Press.

FRASER, J. 1996. "Mapping the process of translation ". Meta 41(1), 84-96.

HOFFMAN, R. R. (ed). 1992. The Psychology of Expertise. New York: Springer-Verlag.

JÄÄSKELÄINEN, R. 1989. "Translation assignment in professional vs. Non-professional

translation: A think-aloud protocol study ". The Translation Process ed. by C. SEGUINOT.
87-98. Toronto: H.G. Publications,.

JÄÄSKELÄINEN, R. 1990. Features of Successful Translation Processes: A Think-Aloud

Protocol Study. Unpublished Licentiate Thesis. Savonlinna School of Translation Studies,
University of Joensuu.

KRINGS, H. 1987. "The use of introspective data in translation". Introspection in Second

Language Research ed. by C. FAERCH & G. KASPER, 159-176. Clevedon & Philadelphia:
Multilingual Matters.

KUßMAUL, P. 1991. "Creativity in the translation process: empirical approaches".

Translation Studies: The State of the Art ed. by K. M. VAN LEUVEN-ZWART & T.
NAAIJKENS, 91-101. Amsterdam & Atlanta, GA: Rodopi.

LADMIRAL, J.-R. 1987. "Technique et esthétique de la traduction". Encrages 17.190-197,

Actes des Journées européennes de la traduction professionnelle. Paris: Hachette.

LÖRSCHER, W. 1991. Translation Performance, Translation Process, and Translation

Strategies. Tübingen: Narr.

MAYER, R. E. 1989 "Human nonadversary problem solving". Human and Machine Problem
Solving ed. by K. J. GILHOOLY, 39-56. New York and London: Plenum Press.

NORD, C. 1997. Translating as a Purposeful Activity. Manchester: St. Jerome Publishing.

NORMAN, D. A. 1985. "Twelve issues for cognitive science". Issues in Cognitive Modeling
ed. by A. M. AITKENHEAD & J. M. SLACK, 309-336. Hillsdale, N. J.: Lawrence Erlbaum
104 Candace Ségumot

POPPER, K. 1989. Conjectures and Refutations: the Growth of Scientific Knowledge. 5 rev.
ed. London: Routledge.

PYM, A. 1994. "Ideologies of the expert in discourses on translator training". Koiné 4.139-
149.

RESNICK, M. 1994. Turtles, Termites, and Traffic Jams. Cambridge, Mass.: The MIT Press.

ROBINSON, D. 1997a. "The invisible hands that control translation". Paper posted in the
On-Line Colloquium on translation hosted over the internet by the Facultat de Traducció i
d'lnterpretació of the Universitat Autônoma de Barcelona.

ROBINSON, D. 1997b. Becoming a Translator, London: Routledge.

RYLE, G. 1949. The Concept of Mind, London: Hutcheson.

SCHÖN, D. A. 1983. The Reflective Practitioner. New York: Basic Books.

SEGUINOT, C. 1990. "A study of student translation strategies". Empirical Research in

Translation and Intercultural Studies ed. by S. TIRKKONEN-CONDIT, 79-88. Tübingen:
Narr.

SEGUINOT, C. 1996b. "Accounting for variability in translation". Cognitive Processes in

Translation and Interpretation ed. by J. DANKS, G. SHREVE, S. FOUNTAIN & M.
MCBEATH, 104-119. London : Sage.

SIMEONI, D. 1998. "The pivotal status of the translator's habitus". Target 10 (1), 1-39.

SNELL-HORNBY, M. 1988. Translation Studies. An Integrated Approach. Amsterdam &

Philadelphia: John Benjamins.

TIRKKONEN-CONDIT, S. 1989. "Professional vs. non-professional translation: A think-

aloud protocol study". The Translation Process ed. by C SEGUINOT, 73-85. Toronto: H.G.
Publications.

TOURY, G. 1995. Descriptive Translation Studies and Beyond. Amsterdam & Philadelphia:
John Benjamins.

VINAY, J-P. & DARBELNET, J. 1958. Stylistique comparée du français et de l'anglais.

Paris: Didier.
TRANSLATING UNDER TIME PRESSURE
AN EMPIRICAL INVESTIGATION OF PROBLEM-
SOLVING ACTIVITY AND TRANSLATION
STRATEGIES BY NON-PROFESSIONAL AND
PROFESSIONAL TRANSLATORS

Astrid Jensen & Arnt Lykke Jakobsen

Copenhagen Business School, Denmark

Zusammenfassung
Dieser Artikel berichtet von einem Übersetzungsexperiment, bei dem die Methode des
Computerlogging von real-time writing-Daten in Kombination mit der Methode des
Lauten Denkens angewendet wurde. Drei verschiedene Texte wurden von sechs
Personen unter drei unterschiedlichen zeitlichen Bedingungen übersetzt. Alle
Tastatureingaben, d.h. alle Cursorbewegungen und Korrekturen bei der Produktion
und Revision des Zieltextes, wurden mit Hilfe der Software Translog registriert, und
die Daten des Lauten Denkens wurden auf Tonband festgehalten. Obwohl nicht alle
anfänglichen Hypothesen durch das Experiment bestätigt wurden, half die
Kombination quantitativer und qualitativer Daten bei der Ermittlung wichtiger
Informationen über den Einfluss des Zeitdrucks auf die Revision des Zieltextes, über
die Benutzung von Wörterbüchern und die Anwendung von wörtlicher Übersetzung
oder Paraphrase.

Résumé
Dans l'expérience rapportée, on a eu recours à la fois à l'écriture saisie
informatiquement en temps réel et à la verbalisation à haute voix. On a demandé à six
sujets de traduire trois textes de presse différents, selon trois délais différents. Toutes
les frappes utilisées pour écrire et éditer les traductions ont été saisies grâce au
programme Translog; tout ce qui a été verbalisé a été enregistré. Toutes les hypothèses
initiales n'ont pu être étayées par notre expérience mais on a pu souligner fortement,
grâce aux données quantitatives et qualitatives, combien la finalisation des textes
d'arrivée est dépendante des délais, de l'utilisation de dictionnaires et de la stratégie
de traduction (littérale ou paraphrastique) choisie par chacun des groupes testés.
106 Jensen & Jakobsen

Resumen
El experimento de traducción objeto de este reporte tue llevado a cabo combinando la
escritura informatica en tiempo real y el razonamiento en voz alta. El experimento
consistió en la traducción realizada por seis sujetos de tres textos diferentes extraidos
de la prensa bajo tres restricciones de tiempo distintas. Todas las pulsaciones
utilizadas al escribir y editar los textos de llegada se registraron por medio del
programa Translog, mientras que la producción verbalizada fue grabada en cinta
magnética. A pesar de no poder corroborar mediante el experimento todas las
hipótesis iniciales, merced a la combinación de datos cuantitativos y cualitativos se ha
podido poner claramente de manifiesto hasta qué punto la calidad de los textos de
llegada está en relación con el periodo de tiempo disponible, la posibilidad o no de
utilizar diccionarios, y el tipo de estrategia traslativa elegida por cada uno de los
grupos sometidos al experimento.

1. Introduction

Since the middle of the 1980s, research into the process of translation has
tended to focus on methods of eliciting and analysing qualitative data. In
particular, the method of thinking aloud has been widely used. In the present
paper, we report on an experiment in which quantitative data logged by a
computer program were used in combination with qualitative think-aloud data.
We believe a quantitative process-oriented approach is an interesting
alternative or supplement to the various approaches hitherto imported from
psycholinguistics. Logging keystrokes in real time is a method which allows us
to track and study the writing process involved in translation across time and
through all revisions (Jakobsen 1998; 1999). Furthermore, this can be done in a
controlled but ecologically valid environment while escaping some of the
criticism that has been levelled against qualitative approaches. It is also
possible to combine a quantitative approach with more invasive, qualitative
methods such as observation, interviewing, or thinking-aloud protocolling.
Combining a quantitative and a qualitative approach makes it possible to
synchronise and compare data elicited by different methods but pertaining to
the same phenomenon. Finally, the logging tool that we use for collecting data
also permits us to work with retrospection in a new way that seems to enhance
both the accuracy and the quantity of retrospective information obtained from
subjects.

2. The Translog program

The key-log program used in the experiment was Translog, which was
originally designed by Arnt Lykke Jakobsen in 1995. The program has a
facility for editing and displaying source text in various ways and a facility for
 Translating under time pressure 107

editing and accessing a document dictionary. This allows the experimenter to

control dictionary information and to control source text display in various
ways. The source text is displayed in the top half of the screen and can be made
to appear either in full or in successive units. There is an automatic facility for
presenting source texts paragraph by paragraph, sentence by sentence or in
predefined units. The length of time a text unit is displayed can be anything
between 5 and 600 seconds. Alternatively, the source text can remain
accessible until the program is terminated. (For full documentation of the
Translog program, see Hansen 1999.)
As the source text appears in the top half of the screen, a translator can
begin to write a translation of it in the bottom half of the screen, where
ordinary text-editing facilities are available. During the entire writing event, all
keyboard activity, including dictionary lookups, cursor movements, and
deletions, is logged. At the end of a session, this information is saved in a
logfile. The information that is logged is essentially information about what
keys were pressed, and when they were pressed. Every keystroke is recorded
with the accuracy permitted by the DOS-clock.
A logfile management component makes it possible to represent logfile
information in two different ways. The Replay facility uses logfile information
to replay writing events as in a video recording. The top window again displays
the successive source-text units that were available to the translator, and in the
bottom half of the screen the translator's writing of the target text is reproduced
with real-time delay, text revision, and so forth. The replay function can be
paused and resumed at any point, and any time the replay function is paused,
the current text version can be saved in a file and printed. From the point of
view of the program, the final text version is just the last of a very large
number of current text versions. The chief experimental functions of the replay
function (and the pause option) are (1) to reinforce a translator's memory of
what s/he was thinking of at a given moment and thereby to enrich
retrospective data, and (2) to synchronise audio- or videotape information (or
other observations) with logged data.
The other way in which the logfile management component can represent
logfile information is with the View facility. The way in which this is done may
best be illustrated by showing two different representations of the same target
text segment, a translation into Danish of the first 32 words of one of the
source texts we used, the A text, which ran as follows: "Unit trust advertising:
Wrapped in weasel words. It is possible that no document on earth has
committed as many sins against clear language. The prose trips off the tongue
like peanut butter... "
108 Jensen & Jakobsen

Log file 02A.LOG

Source text Financial Times
Keystrokes total 2821
Dictionary lookups 0
Time lapse 00.18.08,29
Time unit 4.00 sec.

[*:56.41][↓][↓]*[↓][↑][↑][*:106.88]investeringsforening- sreklame: pakket ind-

i-weasel-words 5**Det ermuligt atikkenogetandetdokument på i-verdenhar-
begået-så-mange-synder-mod-klart- sprog.*- Sproget-flyder- *fr *r som-*** smør
[*:59.70]

Figure 1. Translog View representation with time unit = 4.00 seconds

In addition to supplying information about the file and text used, the total
number of keystrokes, (electronic) dictionary lookups and the total duration of
a writing event, the View facility represents a writing event as a linear
succession of keystrokes (characters, cursor movements, deletions, dictionary
lookups, etc.) with an indication of time in relevant places. Time delay is
indicated by asterisks. In the first representation, the time unit has been set at
four seconds. This means that for each time there was a delay lasting four
seconds an asterisk is printed. However, if a delay occurred during the writing
event which would require ten or more asterisks, the View facility writes the
absolute time delay instead, e.g. [*:56.41].
A time-unit value of four (or more) seconds gives a very rough
representation of time delay. What can be seen in this first representation of the
writing event is that, initially, there was a long orientation pause lasting 56.41
seconds, followed by scrolling of the source text [i] and [↓] to bring the entire
source text into view. This was followed by an even longer pause lasting
106.88 seconds. If the Translog data were our only source of information, we
would assume that this pause was spent processing the headline of the source
text. Synchronised audiotape data might supply additional information, e.g.
that the subject was looking up a word in a manual dictionary, or the subject
might supply think-aloud data at this point. At the end of this brief extract,
there is another long pause lasting almost a minute (59.70 seconds). This could
be interpreted either as a monitoring pause during which the target text so far
produced (or a part of it) was monitored, or as a forward-looking processing
pause in which subsequent target text was being processed.
A representation with a time-unit value of 4 seconds (or more) allows
us to observe only the major delays, which tend to appear at text, paragraph or
sentence boundaries. If the main focus of interest is on such major pauses, or if
it is on, say, text revision rather than time delay, a representation with a high
time-unit value probably brings out the interesting data most clearly. A high
 Translating under time pressure 109

time-unit value has the advantage of representing the target text in a way that is
usually easy to read. The target text appears to have been typed very regularly
with occasional minor delays appearing between sentences (after the headline
line break, after the sentence ending with 'mod klart sprog'), in connection
with the (proposed but abandoned) revision of 'flyder', or between words
(som*** smør). (The centred dot represents a touch of the space bar.)
Translog's default time-unit value (i.e. the value of one asterisk) is 0.50
seconds. With this value, the identical text segment will be represented in the
following manner:

Log file 02A.LOG

Source text Financial Times
Keystrokes total 2821
Dictionary lookup 0
Time lapse 00.18.08,29
Time unit 0.50 sec.

[*:56.41][l]*****[^]*********[^]*******[t]**[t][*:106.88]investeringsforening*
**.*** sreklame****:*-*pakket-*ind-i-*weasel-**words** 5[*:8.24]Det-ermulig
t- **at- **ikke-noget-ande*t-dokument-på* ****i verden- * **har-beg*ået-så-inange-
syndermod*klar*t**sprog.[*:7.96],**Sproget,****flyder*********fi:******
********r*som-[*:12.20]smør[*:59.70]

Figure 2. Translog View representation with time unit = 0.50 seconds

As can be seen, the only difference in the way the View facility displays logfile
information is in the way time delays are represented. With a time-unit value of
half a second, more detailed (and more accurate) information about time delay
is made to appear. We can see that time elapsed between all of the initial scroll
movements of the source text, and that sometimes there were delays not just
between words and sentences, but also within words (ande*t, beg*ået, klar*t).
However, delays appear on the whole very clearly to coincide with
grammatical (word and sentence) boundaries. Finally, we notice that all delays
lasting 5 seconds or more are now represented with the absolute delay time,
e.g. [*:8.24]. If we are interested in studying time delay in even greater detail,
we can ask the program to recalculate the logfile information with a smaller
time-unit value.

3. Aim and design of the experiment

For translation within the news media, time and deadlines are key issues.
Another serious challenge in newspaper translation is the multitude of topics
dealt with., In order to cope with these two main constraints, the translator of
110 Jensen & Jakobsen

texts for the news media must have efficient methods of information retrieval,
must be able to solve comprehension and translation problems quickly, must be
capable of making fast decisions, and must have efficient translation strategies.
Some problems are no doubt solved in a flash and many decisions made in an
instant. However, a lot of problem solving takes measurable time and therefore
can be more easily investigated.
One purpose of the experiment carried out by Astrid Jensen was to see
what problem-solving activity could be identified by comparing think-aloud
and logfile data and to see how such activity was distributed across eighteen
translation events. The effect of time pressure (if any) on the subjects' use of
translation strategies was also measured. Translog was used to help identify
problem-solving activity and to ascertain both what translation strategies were
used by non-translators and professional translators and what changes in the
use of translation strategies would correlate with different time conditions. It
was assumed that non-translators and professional translators would behave
quite differently. This assumption was based on a number of studies within the
past ten years comparing non-translators with professional translators using
introspective methods (Börsch 1986; Gerloff 1986; Krings 1986a and 1986b;
Jääskeläinen 1989a and 1989b; Lörscher 1991; Jääskeläinen & Tirkkonen-
Condit 1991; Fraser 1996). Translation strategies were operationally restricted
to direct translation and paraphrase. The extent to which dictionaries were
resorted to was also observed. (On time pressure, see also Jensen 1999.)
Six translators, two non-translators, two recently graduated
professionals, and two experienced professional translators were asked to
translate three different journalistic texts under three different time constraints
using Translog. The dictionaries consulted were traditional (non-electronic)
dictionaries. While doing the translation, subjects were asked to think aloud.
What they said was audiotaped, transcribed and synchronised with the process
information recorded by Translog. The two non-translators, or 'educated
laymen' (Jääskeläinen 1993: 99) had an M.Sc. degree in engineering, and were
accustomed to using English as a working language but had no experience with
translating. The time conditions set were 15, 20 and 30 minutes (with an initial
warm-up exercise). Each subject translated texts under all three time
conditions. Thus a total of 18 translations were produced.
The texts chosen comprised three topics from different sources: the
Economist, the Financial Times, and the Observer. The texts were translated
from English into Danish, the mother tongue of all the subjects, and subjects
were instructed to translate as if the articles were to appear in the Danish
newspaper Berlingske Tidende, a well-known quality national newspaper. For
the professional translators, this set-up was very similar to a routine translation
task. The average length of the texts was 120 words. Subjects were allowed to
use dictionaries of their own choice, and the experiment took place in their own
office, uninterrupted by the researcher.
 Translating under time pressure 111

4. Data representation and transcription

For the purpose of analysing the translation, the source text was segmented into
portions consisting of individual sentences (S1, S2, etc.) and the relevant
verbalisations from the subjects' protocols were combined with the data from
the Translog print-outs in a table (see Table 1). In that table, the duration of
time delays was represented in absolute figures, and relevant editing was
inserted in the text segments where it belonged. Further, a distinction was made
according to whether the editing took place during translation or during
revision.

Table 1: Source language text units synchronised with logfile and think-aloud
data

Text segment Words Letters Translation Revision Total

time time time
Source language text: 3 21 2.56- 12 sec
S1 Unit trust advertising 3.08
Translog:
(4*) Investeringsforeningsreklame
Think-aloud protocol
Unit Trust - er det ikke en
investeringsforeningsreklame
Source language text 4 20 3.08- 14.42- 53 sec
S2 Wrapped in weasel words 3.20 15.23
Translog:
Pakket ind i weasel words
Tvetydigheder
Think-aloud protocol
Pakket ind i weasel words,
Jeg ved ikke hvordan man skal
oversætte det lige nu, men det
kommer jeg tilbage til
Revision:
Weasel word - tvetydige ord -
tvetydigheder

5. Problem-solving activity

An initial, quantitative survey was made from the Translog records of pauses
that could be associated with problem solving. By problem-solving activity we
understand information processing aimed at solving a specific and identifiable
translation problem. This may be reflected in verbalisation, but not necessarily.
In the experiment, there were very considerable differences in verbalisation.
Some subjects produced very few comments, while others, perhaps more eager
112 Jensen & Jakobsen

to please the researcher, spoke volumes. There was no correlation between the
amount of verbalisation and the speed (or quality) of target-text production.
Therefore, no correlation was found between the quantity of verbalisation and
the amount or frequency of problem-solving activity.
On the other hand, verbalisation was clearly linked to problem solving.
After replaying the recorded translation event in Translog and listening to the
synchronised audio-tape recording several times and thus comparing logged
text-production data with think-aloud data, it was intuitively clear what
problems were being processed in what delays. Some delays were caused by
dictionary search, others by meta-comments, memory search, or word
association. A coding scheme for six identifiable problem-solving activities
was applied, and though not all delays could be linked to a specific problem-
solving activity, a correlation emerged between delays lasting four seconds (or
more) and think-aloud evidence of problem-solving activity.
A subsequent count of all delays lasting more than four seconds did not
reveal significant differences across the six subjects. Nor were there any clear
differences in the number of delays devoted to each of the three texts used. All
subjects had an average of very close to 2 problem-solving delays per minute,
regardless of text and time condition. There was a clear increase in the number
of problem-solving delays as more time was granted for a task (32% more with
20 minutes than with 15; 57% more with 30 minutes), the increase
corresponding to the increase in translation time available.
However, by differentiating between delays occurring during the initial
editing phase and during revision, two significant tendencies emerged. With
more time, the number of problem-solving delays during the initial editing of
the target text increased much less (20% and 32%) than the number of
problem-solving delays during revision (83% and 137%). Also, there was a
markedly higher incidence of problem-solving delays during revision in the
translations by the two most experienced professionals than in the other
translations.
These findings were interpreted as suggesting that raw target-text
production has its own rhythm and speed, which responds to the time factor but
is fairly independent of it (within the time ranges tested in the experiment). The
increase in the incidence of problem-solving delays during revision suggests
that for the majority of subjects in the experiment 15 minutes did not allow
much time for revision. In other words, the number of translation problems
negotiated was relatively independent of the three time conditions in the
experiment. What happened when more time was available was that time was
sometimes spent in two (or more) separate rounds of attempting to solve
problems to which problem-solving time had already been devoted.
 Translating under time pressure 113

6. Dictionary lookups

There was no significant variation in the number of dictionary lookups per time
condition or per text. The most striking differences in the use of dictionaries
were (again) between three groups of subjects. Non-translators used
dictionaries far more frequently (45 times in all) than the young professionals
(with 1-3 years of professional experience), who in their turn used dictionaries
far more often (32 times) than the experienced professionals (with more than 8
years of professional experience; 18 dictionary look-ups).

7. Translation strategies

In addition to examining the occurrence of pauses that could be associated with

problem solving and the use of dictionaries, an investigation was made of the
translation strategies employed in all eighteen translations. Translation
strategies were distinguished as being either direct, as in literal (or semantic)
translation, or indirect, as in free (paraphrastic or communicative) translation
(cf. Bell 1991:70). Direct strategies included borrowing, loan translation and
literal translation. (No loan translations were recorded here.) Indirect strategies
were categorised as instances of either paraphrase or adaptation. Once again,
the most striking differences were between three groups of translators:

Table 2. Occurrence of direct translation strategies by group

Direct translation Non-translators Young profs Exp. profs

strategies
Borrowing 17 7 3
Literal translation 32 9 6
Total direct strategies 49 16 9

Table 3. Occurrence of indirect translation strategies by group

Indirect translation Non-translators Young profs Exp. profs

strategies
Paraphrase 4 35 44
Adaptation 0 3 0
Total indirect strategies 4 38 44

No significant correlation was found between the three time conditions or the
three texts and the use of translation strategies.
The preferred translation strategy by the group of non-professional
translators was literal translation, whereas professional translators all preferred
paraphrase. There were only half as many instances of direct translation in the
114 Jensen & Jakobsen

group of expert translators as in the group of young professionals and also

fewer instances of borrowing.
Whereas non-professional translators appear to rely both on extensive
parallelism between form and meaning across English and Danish and on one-
to-one relationships at word level, professional translators rarely find it
possible to establish adequate equivalence of meaning through formal
equivalence or at word level. They are far more aware of the need to create a
comprehensible, idiomatic and readable target text. Expert translators rely
almost exclusively on paraphrase. Indeed, they seem to rely on indirect
translation strategies to such an extent that one might suspect them of
deliberately avoiding direct translation strategies, possibly in order to escape
the tyranny of the source-text form.

8. Discussion and conclusion

The only significant effect of time pressure that was identified in the
experiment was the decrease in problem-solving activity during revision as less
time was given for a translation task. However, it appears that only the non­
professionals came under serious time pressure to the extent that in at least one
case they were unable to complete a first draft of the translation. Therefore, the
assumption that different translation strategies would be used, given sufficient
time pressure, could neither be confirmed nor dismissed. In retrospect, it can be
seen that there ought to have been a shorter time condition than 15 minutes so
that all subjects would have come under more acutely felt time pressure.
The main findings of the experiment were in unexpected areas, in the
correlation of dictionary use and professional translational experience and in
the correlation of paraphrase and professional translational experience. Both
findings confirmed familiar assumptions. The results corresponded rather
closely with observations on 'developmental stages' made e.g. by Jääskeläinen
(1996:67) but seemed to suggest that the translational behaviour of young
professionals may be closer to that of semi-professionals (graduate students)
than to that of professionals with eight or more years of experience.
An initial assumption that time pressure would force all translators to
use some of the coping tactics observed in interpreting (cf. Gile 1995:207),
because of limited processing capacity or because of gaps in the subjects'
knowledge base, was only partially supported by the data. The findings
confirmed that all groups made use of coping tactics, in fact almost a third of
all translation strategies employed could be analysed as coping tactics.
However, at this point it would be premature to conclude that the use of coping
tactics was caused by time pressure only. It seems more probable that the
combination of (some) time pressure and certain types of translation difficulty
motivated the use of coping tactics.
 Translating under time pressure 115

Methodologically, we found it very helpful in tracing what translation

strategies were used to be able to draw on a complete, logged record of the
entire editing process. More importantly, we found that the possibility of
contrasting and comparing parallel data from different sources offers new
research opportunities. The request for subjects to think aloud while writing
their translation may have slowed down their translation process somewhat and
thus introduced an undesirable laboratory distortion of the data, but there
appear to us to be very considerable advantages in being able to make
observations in several parallel sets of data, in our case one qualitative set and
one quantitative. Our operational definition of what constitutes a problem-
solving pause was arrived at intuitively by synchronising and comparing data
from both sources and, when applied to one set of data (the automatically
logged real-time data), helped identify significant differences in translational
behaviour during revision.
We believe such an approach opens up new lines of inquiry, data
interpretation, and hypothesis testing. Real-time logging of all keyboard
activity allows us to study the process and the details of target-text production
with much greater accuracy than has been achieved by other methods. The
possibility of generating parallel sets of translation process data allows us to
interpret one set in the light of another and to look for converging
interpretations. The fact that by this method it will be possible to test
hypotheses formulated on the basis of qualitative data on a parallel quantitative
data set and vice versa could give future hypotheses about translation greater
strength.

References

BELL, R.T. 1991. Translation and Translating, Theory and Practice. London: Longman.

BÖRSCH, S. 1986. "Introspective methods in research on interlingual and intercultural

communication". HOUSE, J. & S. BLUM-KULKA (eds), 195-209.

FRASER, J. 1996. "Mapping the process of translation". Meta 41(1), 84-96.

GERLOFF, P. 1986. "Second language learners' reports on the interpretive process: talk-aloud
protocols of translation". HOUSE, J. & S. BLUM-KULKA (eds), 243-261.

GILE, D. 1995. Basic Concepts and Models for Interpreter and Translator Training.
Amsterdam & Philadelphia: Benjamins.

HANSEN, G. (ed.) 1999. Probing the Process in Translation: Methods and Results.
Copenhagen Studies in Language 24. Copenhagen: Samfundslitteratur.

HOUSE, J. & S. BLUM-KULKA (eds). 1986. Interlingual and Intercultural Communication.

Tübingen: Narr.
116 Jensen & Jakobsen

JAKOBSEN, A.L. 1998. "Logging time delay in translation". Copenhagen Working Papers in
LSP 7, 73-101. Copenhagen: Copenhagen Business School.

JAKOBSEN, A.L. 1999. "Logging target text production with Translog". HANSEN, G. (ed.),
9-20.

JAKOBSEN, A.L. & L. SCHOU. 1999. "Translog documentation". HANSEN, G. (ed.),

Appendix, 1-36.

JENSEN, A. 1999. "Time pressure in translation". HANSEN, G. (ed.), 103-119.

JÄÄSKELÄINEN, R. 1989a. "The role of reference material in professional vs. non­

professional translation: A think-aloud protocol study". Empirical Studies in Translation and
Linguistics ed. by S. TIRKKONEN-CONDIT, 175-200. Studies in Languages 17. Joensuu:
University of Joensuu.

JÄÄSKELÄINEN, R. 1989b. "Translation assignment in professional vs. non-professional

translation". The Translation Process ed. by C. SEGUINOT , 87-89. Toronto: York University.

JÄÄSKELÄINEN, R. 1993. "Investigating translation strategies". Recent Trends in Empirical

Translation Research ed. by S. TIRKKONEN-CONDIT & J. LAFFLING, 99-120. Joensuu:
University of Joensuu.

JÄÄSKELÄINEN, R. 1996. "Hard work will bear beautiful fruit? A comparison of two think-
aloud protocol studies". Meta 41(1), 60-74.

JÄÄSKELÄINEN, R. & S. TIRKKONEN-CONDIT. 1991. "Automatised processes in

professional vs. non-professional translation: A think-aloud protocol study". Empirical
Research in Translation and Inter cultural Studies. ed. by S. TIRKKONEN-CONDIT, 89-110.
Tübingen: Narr.

KRINGS H.P. 1986a. Was in den Köpfen von Übersetzern vorgeht. Tübingen: Narr.

KRINGS, H. P. 1986b. "Translation problems and translation strategies of advanced German

learners of French (L2)". HOUSE, J. & S. BLUM-KULKA, 263-276.

LÖRSCHER, W. 1991. Translation Performance, Translation Process, and Translation

Strategies'. A Psycholinguistic Investigation. Tübingen: Narr.
TYPES OF CREATIVE TRANSLATING

Paul Kussmaul
University ofMainz, Germersheim, Germany

Zusammenfassung
Um etwas über kreative Prozesse bei englisch-deutschen Übersetzungen (die meisten
stammen aus T S . Eliot's Old Possum 's Book of Practical Cats) zu erfahren, wird von
den Begriffen Perspektive, Fokus, Figur/Grund Relation, Szene, Rahmen, Kern,
unscharfe Ränder und Verkettungen als heuristische Mittel Gebrauch gemacht.
Dadurch lassen sich eine Reihe kreativer Techniken unterscheiden: 1. Verkettung von
Kategorien 2. Auswahl von Szenenelementen 3. Szenenerweiterung 4. Einrahmung
einer Szene 5. Neurahmung einer Szene.

Résumé
Jusqu'où y a-t-il créativité dans les traductions anglais-allemand, notamment dans les
traductions de T.S. Eliot? Pour l'évaluer, nous avons fait usage heuristique des notions
comme celles de point de perspective, focus (ou saillance), figure, fond, scénario, frame
(ou cadre cognitif), éléments saillants, frontières floues et liant ou enchainement. On a
distingué finalement 5 types de modes créatifs: l'enchainement de catégories, la
sélection d'éléments du scénario (ou script), l'élargissement du scénario, la mise en
cadre d'un scénario et la mise en place d'un nouveau frame.

Resumen
Con el fin de estudiar la creatividad en las traducciones del inglés al alemân (la mayoria
tomadas del Old Possum's Book of Practical Cats de T.S. Eliot), se han utilizado como
medio heuristico los conceptos perspectiva, enfoque, alineación flgura/suelo, escena,
marco, nücleo, bordes difusos y encadenamiento. Se pueden distinguir varias técnicas
creativas: 1) Encadenar categorias 2) Seleccionar elementos de la escena 3) Ampliar la
escena 4) Enmarcar la escena 5) Imaginar un marco nuevo.
118 Paul Kussmaul

1. Heuristic models

"He can't think straight for a minute!" This usually is not exactly a laudatory
comment. Creative people, however, very often do exactly this; they jump from
one idea to another. This does not mean they cannot think straight, but when it
comes to having a bright new idea, thinking straight is replaced by thinking
around the corner, or, in the terms of creativity research, by thinking laterally
(de Bono 1970) or divergently (Guilford 1975). Lateral thinking, a term coined
by Edward de Bono, has indeed become one of the basic notions in creativity
research. Lateral thinking leads to new ideas, and so it does in translation.
As to the products that result from new ideas, it is generally recognised
that they should be new in one way or other and also that they should be
adequate to the task (cf. Preiser 1976:6ff). These two features should be kept
in mind when evaluating creative translations. I am not interested here in
translations that are not good enough or even bad. I am specifically studying
good translations. They can be taken as models of how to translate successfully,
especially if the cognitive categories behind them can be revealed.
There are two further notions closely connected with lateral thinking:
perspective and focus (cf. Brodbeck 1995:40-41, de Bono 1990:154-168 uses
the synonym "attention area"). Changing the perspective and changing the focus
are ways of seeing familiar things in a new light and thus of coming up with new
ideas. When we deal with linguistic categories and with the use of language
these notions are, of course, metaphorical. It may be helpful, therefore, to use
pictures as a starting point in order to illustrate what goes on in our minds when
we are creative.

Figure 1

When we look at the cube in Figure 1 from the usual point of view, its front side
is on the lower left and we look upward to the right. But we can change the
point of view and look at it so that the front side is on the upper right and we
look down to the left.
When we make a drawing of a man on a bike we usually look at him
from the side, but we may change the perspective and look at him from the top,
and when he is a Mexican he will look like Figure 2:
 Types of creative translating 119

Figure 2

These are amusing games, but they also reveal some basic features or creative
thinking. Creative thinking makes us see details we may not have noticed
before: in the drawing of the Mexican on a bike the unusual size and the
sunshade function of his hat.
It is interesting to see that the notions of perspective and focus are also
used in cognitive semantics. Ronald W. Langacker uses the notion of
figure/ground alignment to explain perspective and focus (1987:120ff.). Here
again, a picture may help to illustrate what he means. (Langacker only gives a
verbal description.) In Figure 3 we usually see a white dot on a black field. The
focus is on the white dot.

Figure 3

In Figure 4 we usually see a black frame or "round window".

Figure 4

The focus in Figure 4 is on the window. Langacker's explanation is that the

larger area is seen as the ground and the smaller area as the figure. Thus the
white dot is smaller than the black ground, and when it becomes bigger and
bigger until it reaches the size of a disc, it becomes the ground for the figure,
which then takes the form of a frame for looking through. Langacker makes use
120 Paul Kussmaul

here of the notion "canonical viewpoint", which is the usual and natural way of
looking at things (Langacker 1987:123). Thus, when looking at the cube we
may find it hard to preserve the new point of view; after a few seconds we may
relapse into the old and familiar perspective.
Although we may find it difficult to do so we can structure things in a
different way. The first figure we can see as a black piece of cardboard with a
little hole in it, and the second figure we can see as a white plate on a black
table mat. This is a more unusual focus, and it is very similar to the change of
focus used as a starting point for divergent thinking. In creative translating, as
we shall see, changing the perspective and focus plays a very important part.
Being guided by the principle of fidelity, we will try to preserve the notions and
concepts of the source text in our translation, but we may change the point of
view from which we see a concept and/or we may focus on different elements of
a concept.
There is another frequently used term closely linked with prototypes.
When referring to larger notions, use is made of scene or even scenario (cf.
Fillmore 1977). Scenes may consist of a large number of elements. It will be
seen that the notion of a scene and focusing on scene elements are helpful tools
for explaining what happens in creative translating.
Focus and perspective, it seems to me, are also at work in prototype
semantics and in what George Lakoff calls chaining, a notion based on
prototype semantics. Chaining takes place via core elements (which are
focused) of scenarios (Lakoff 1987:95). When translating creatively we
sometimes do this. If we do not want to throw the principle of fidelity
overboard altogether there will, after all, have to be some link between the
source and target texts.
Scenes are linguistically represented by frames (cf. Fillmore 1977). In
creative translating, it will be seen, scenic details mentioned in the source text
are sometimes replaced by a frame in the target text. For practical purposes we
can distinguish the frame from the scene because usually a frame is more
abstract and consists of fewer words than a description of a scene.

2. Types of creative translating

2. 1 Chaining of categories

In the Asterix comics the Druid's name in the French original is Panoramix. The
name refers to the far-sightedness and visionary powers of the wise old man. In
the English version he is called Get-a-fix, an allusion to the distribution of the
magic potion, and in the German version his name is Miraculix which refers to
the miraculous effects of the potion. These translations can be regarded as
 Types of creative translating 121

creative because they are new in the sense that they differ from the source text.
If, following Lakoff, we think of the names as categories, we might find it
difficult to see the chaining or linking element between the original name and its
translations: what have visionary powers to do with the magic potion? Are these
not two completely different categories? The connection becomes more
obvious, though, if we see the categories within an overall scenario: the druid
with all his various functions and capabilities. They are the linking elements
between the various names for him. We may regard the different names as a
result of a different perspective. For someone interested in planning and looking
ahead a proper name for the Druid may well be Panoramix, for someone
interested in the effects of a special type of booze he may adequately be called
Get-a-fix, and for those who take him along to fight against the Romans he is a
proper Miraculix. The different names can also be explained by focus; the
different names in the French, English and German version put the emphasis on
different features of the character. In Langacker's terms we may call these
features the figure against the ground of the overall scenario, and what was part
of the ground in the French source text (remember the restructuring of
Langacker's figure-ground alignment!) becomes the figure in the translations.
The fact that categories are parts of larger scenarios can be used as a
hypothesis to explain creative translations. These embeddings are not dissimilar
to the hierarchical structure of semantic fields, but in contrast to them they are
not caused by the language system but by our comprehension of the text and by
our experiences. This is the psychological aspect of the model, and this is what
enables us to chain categories.

2.2 Picking out scene elements

In translations we can often observe that a relatively abstract word of the source
text is replaced by a more detailed and concrete expression. This can be
analysed from a cognitive point of view.
In T. S. Eliot's Old Possum's Book of Practical Cats (which was used as
the libretto for the musical Cats) one of the poems is about Gus, the Theatre
Cat, who as an old cat tells everyone how famous he was in the old days:
For he isn't the cat that he was in his prime;
Though his name was very famous, he says, in its time.
(T.S.Eliot 1961:72)

The words name, famous, in its time are the frames which are filled in, as it
were, by the reader's mental pictures. These pictures are part of the
comprehension process, which is closely linked, as I have been able to observe
in many instances, with the actual translation process. Translators can be
122 Paul Kussmaul

creative at the comprehension stage, and when they verbalise what they have
comprehended they produce a creative translation. Thus, imaginative reading,
one might say, is be a prerequisite for creative translating. Carl Zuckmayer, who
translated this poem, appears to have been an imaginative reader. He wrote:

Nein, er ist nicht der Kater mehr, der er gewesen,

Als man täglich von ihm in der Zeitung gelesen.
(Eliot 1961:74)
'He is no longer the cat that he was before
when one could read about him in the papers every day'

The abstract frame famous gives rise to a scenario which potentially consists of
many elements, and one of them is picked out and verbalised by Zuckmayer.
Being based on our experiences, scenarios are historically determined. Today
the celebrity scenario would include the fact that we know famous actors from
talk shows and television serials and commercials, and if Carl Zuckmayer still
lived and had to translate the poem today, he would perhaps make use of one of
these elements.
I propose to call the creative strategy underlying this translation: picking
out scene elements within a frame. Here again we can observe the phenomenon
of'focusing. One element is mentioned in the translation, and we thus view one
particular element of the scene, which may be said to stick out as the figure
against the ground of the rest of the scene. It might have been possible to view
a different element, for instance the well-known theatres or the famous plays in
which he acted or the high fees he earned.
We may ask if reading about an actor in the newspapers is a core element
of the notion fame. I think it is. With fame we associate publicity and with
publicity we associate the media, and newspapers are part of them. Is the
translation adequate? It seems we are on the safe side if we pick out core
elements as translations, because they are prototypical. Television commercials,
although perhaps part of the celebrity scenario, are not core elements really but
rather positioned at the fuzzy edges. (See Rosch 1973.) Thus a translation such
as the following would not be so good:

Nein, er ist nicht der Kater mehr, der er mal war,

Als man ihn täglich in Werbespots sah.
'He isn't the cat he used to be
when you could watch him daily in commercials'
 Types of creative translating 123

2.3 Enlarging a scene

Enlargement, by definition, is a quantitative notion. In translation the question

arises: how many and which kinds of elements can be added to a scene without
getting into another scene that is no longer adequate as a translation, i.e. no
longer related in some degree to the source text? If we accept that mental
categories are characterised by fuzziness, there cannot be a definite answer to
this question. Another example, again taken from T.S. Eliot's poem about Gus
the theatre cat, may serve to illustrate this point:

His coat's very shabby, he's thin as a rake

And he suffers from palsy that makes his paw shake. (Eliot 1961:72)

Carl Zuckmayer translated:

Sein Flausch ist recht schäbig, er selbst ein Gerippe
Mit zittrigen Pfoten schon halb auf der Schippe. (Eliot 1961:74)
'His fur's very shabby, he's merely skin and bones
with trembling paws he's at death's door'

The central elements of the source-text scene are "old", "weak", "invalid".
These are the elements which link (or "chain") source and target scenes. The
target scene, however, is more intensified and thus adds a new aspect to the
source scene. The "old-invalid-man scene" is enlarged to the "old-man-close-to-
death scene". Is this enlargement adequate? One could turn things round and
say that the new scene includes the old one. Death implies - as prototypical
notions - sickness and old age. In other words, the larger frame would include
the smaller one. Logically, the translation would thus be adequate, but we have
only looked at two lines of the poem. We would, of course, have to make a
detailed analysis of the complete poem in order to find out if the scene
enlargement, with the mentioning of its details, is justified. There is no room to
do this here: we have just pointed out some aspects of evaluation.

2.4 Framing a scene

In the example mentioned above about Gus, the theatre cat, as a famous actor, a
frame was replaced by an element of the corresponding scene. The converse
method is the replacement of a scene element by a frame. Mr. Mistoffelees, the
conjuring cat, knows many tricks, for instance
He can pick any card from a pack. (Eliot 1961:59)

For the musical Cats this was translated by

124 Paul Kussmaul

Seine Kartentricks sind sehr gefragt.

(Deutsche Originalaufhahme. Deutsch von Michael Kunze, aufgenommen im
Motiva Studio, Wien, November 1983).
'His card tricks are very popular.'

Card tricks is a phrase that covers many things. It is a frame for a scene that
consists of many elements. One of them is the element chosen by T.S. Eliot. But
other parts of the scene are also the person who shows the tricks and the
audience who is baffled by them. There can be different kinds of tricks, too. The
magician may guess which card a person from the audience has chosen, or a
person may tell the magician - as in our example - which card to draw, and
there are many more varieties. In Kartentricks the focus is not on a particular
one of these elements, but the word includes them all. We zoom out, as it were,
until we reach the frame so that the whole scene is in the picture.
In the translation the source text has been altered, that is, to some extent
it fulfils the criterion of novelty so essential for creative products. But is it really
creative? It was mentioned earlier that a creative product must not only be new
but also adequate. Semantically speaking the translation is adequate insofar as a
frame always includes the elements of the corresponding scene. But is it also
adequate from a poetical point of view? One might argue that card tricks is less
vivid than picking any card from a pack, and that vividness, after a careful
analysis of T.S. Eliot's Book of Practical Cats, might turn out to be a central
feature of the poems. Thus the translation by Friedrich Podszus in the English-
German version of the Suhrkamp edition, although more literal, would be more
creative because he preserved the vividness of the original by writing:

Er zieht auf Wunsch jede Karte aus dem Päckchen heraus. (Eliot 1961.61)
'At the audience's wish he draws any card from a pack.'

We are faced here with a specific feature of creativity in translating, which at

first sight seems to be a paradox. On the one hand translators can fulfil the
requirement of novelty only if they move away from the source text; on the
other hand it may be more adequate for the overall purpose not to move very
far from the source text and thus be less creative. It should be mentioned here
that, although he did not change the main phrase, Podszus added the phrase auf
Wunsch to the source text, and thus he at least changed the source text a little
bit. (I am using words like less or more creative and a little bit because
creativity is a matter of degree.) The paradox can be solved if we consider the
purpose or skopos, i.e. the pragmatic embedding, of the translation. The
restrictions of the musical (melody, rhythm etc.) may have made it necessary to
use the more abstract form. In other words, as a translation of a poem Podszus'
 Types of creative translating 125

version is more adequate, although less creative, and as a libretto for the
musical the translation by Michael Kunze is an adequate and more creative
version.

2.5 Thinking up a new frame

Kartentrick is an already existing frame for the scene suggested by the source
text. This is the frame people use for the kinds of mental pictures suggested by
the source text. A new frame, however, would be a tailor-made frame, as it
were; in lexical terms, a neologism.
I asked my students to translate a satirical text about British motorway
restaurants (cf. Kussmaul 1995:27). The assignment was: "Preserve the
humorous tone in order to add to the enjoyment of reading the text! The text
will appear in a booklet on British everyday culture." The second part of the
sentence "Starch is cheap and filling, so sausages composed chiefly of bread
seem somehow right" was translated by a student by "...Hot dogs mit viel Brot
und Würstchen mit Baguette" ('hot dogs with a lot of bread and sausages with
baguette'). In a seminar we discussed this translation, and said that top-down
processes had here probably blotted out the bottom-up material. We tried to
remember what English sausages looked and tasted like, namely that they
consisted of about 50% bread, 40 % fat and 10% meat. One student remarked
that in the German translation the bread was round the sausage and in the
English text the bread was in the sausage. After having agreed that with regard
to the assignment the translation was not adequate one of the students
suggested: "Let us use Brotwurst ('bread sausage')!"
This is a creative translation. It is adequate because it refers to the
British culture and it is new, because the word does not exist in German. From a
cognitive point of view it comprises, i.e. frames, by implication, the elements
mentioned in the source text: cheap, filling, composed chiefly of bread, but it
focuses on the last one. In addition, it is funny, because it is a play on words
{Bratwurst 'sausage for frying' - Brotwurst) and thus preserves the humour of
the source text.

3. Conclusion

The five types of creative translation I have presented are chiefly based on four
types of mental processes: (a) moving from frame to frame (Asterix examples)
(b) moving from frame to scene (the theatre cat) and (c) moving from scene to
scene (again the theatre cat) and (d) moving from scene to frame (the conjuring
cat and the British motorway text). They all involve some degree of focusing,
perspective change or zooming.
126 Paul Kussmaul

No matter what we do, the processes which inspire creative solutions

are always linked with scenic visualisations. This opens up the way for
changing the focus in order to pick out elements in the scene or including them
all Scenic visualisations thus contribute to the novelty of a translation and help
make it a creative product.

References

DE BONO, Edward 1970. Lateral Thinking. A Textbook of Creativity. London: Ward Lock
Educational.

BRODBECK, Karl-Heinz 1995. Entscheidung zur Kreativität. Darmstadt: Wissenschaftliche

Buchgesellschaft.

ELIOT, T. S. 1961. Old Possums Katzenbuch. Englisch und Deutsch. Frankfurt am Main:
Suhrkamp.

FILLMORE, Charles J. 1977. "Scenes-and-Frames Semantics". Linguistic Structures

Processing ed. by ZAMPOLLI, Antonio, 55-88. Amsterdam: North Holland,

GUILFORD, Joy Peter 1975. "Creativity: A Quarter Century of Progress". Perspectives in

Creativity ed. by I. A. TAYLOR. & J. W. GETZELS, 37-59. Chicago: Aldine.

KUSSMAUL, Paul 1995. Training the Translator. Amsterdam & Philadelphia: John
Benjamins.

LAKOFF, George 1987. Women, Fire and Dangerous Things. What Categories Reveal about
the Mind. Chicago: University of Chicago Press.

LANGACKER, Ronald W. 1987. Foundations of Cognitive grammar. Stanford: Stanford

University Press.

PREISER, Siegfried 1976. Kreativitätsforschung. Darmstadt: Wissenschaftliche

Buchgesellschaft.

ROSCH, Eleanor 1973. "Natural categories". Cognitive Psychology 4, 328-350.

COMPREHENSION PROCESSES IN
SIMULTANEOUS INTERPRETING 1

Maria Teresa Bajo, Francisca Padilla & Presentacion Padilla

Departamento de Psicologia Experimental
Departamento de Traduction e Interpretation
Universidad de Granada, Spain

Zusammenfassung
In zwei Versuchsreihen haben qualifizierte Dolmetscher, Dolmetschstudenten,
zweisprachige Versuchspersonen und Fachkräfte aus anderen Berufszweigen
Aufgaben gelöst im Zusammenhang mit dem Arbeitsgedächtnis, dem Textverständnis
und dem Zugang zur Semantik und dem Lexikon. In der ersten Serie haben wir
versucht, das Textverständnis unserer Versuchspersonen zu definieren, sowie auch
andere Aspekte dieser Kompetenz (Aktivierung des Lexikons und der Semantik). In
der zweiten Versuchsreihe haben wir versucht die Fähigkeit und Effizienz des
Arbeitsgedächtnisses zu überprüfen. Die Ergebnisse erweisen eine relative
Überlegenheit der Dolmetscher im Bereich der linguistischen Fähigkeiten und in der
Benutzung des Arbeitsgedächtnisses und sind ein Hinweis dafür, dass diese besondere
Strategien benutzen zur Erweiterung ihrer Fähigkeiten und Effizienz und damit ihr
Verständnis verbessern.

Résumé
Au cours de deux séries d'expérience, des interprètes professionnels, des étudiants en
interprétation, des sujets bilingues et des praticiens provenant d'autres domaines du
savoir ont réalisé des tâches incluant la mémoire opérationnelle, la compréhension et
l'accessibilité lexicale et sémantique. Lors de la première série, nous avons cherché à
établir la capacité de compréhension de nos sujets, mais aussi à cerner certains aspects
relatifs à cette activité (activation lexicale et sémantique). La deuxième série de tests
visait à établir la capacité et l'efficacité de la mémoire opérationnelle des interprètes.
Les résultats ont montré la supériorité des interprètes quant aux compétences
linguistiques et l'usage de leur mémoire opérationnelle, car il a été prouvé que ceux-ci
acquièrent des stratégies spécifiques dans le but d'améliorer leurs capacités et leur
efficacité, renforçant ainsi leurs processus de compréhension.
128 Bajo & Padilla

Resumen
En dos series experimentales, intérpretes profesionales, estudiantes de interpretation,
bilingües y profesionales de otros campos del saber realizaron tareas relacionadas con
la memoria de trabajo, comprensión, acceso léxico y semántico. En la primera,
intentamos definir la habilidad de comprensión de nuestros sujetos, asi como aspectos
diferentes relacionados con esta habilidad (activación léxica y semântica). En la
segunda serie, intentamos définir la capacidad y eficacia de la memoria de trabajo de
los intérpretes. Los resultados mostraron la superioridad de los intérpretes en destrezas
lingüisticas y en el uso de su memoria de trabajo, al indicar que desarrollan estrategias
peculiares para ampliar su capacidad y eficacia, y asi potenciar sus procesos de
comprensión.

1. Introduction

It is only recently that cognitive psychologists have shown interest in the

processes involved in translation and interpretation (Danks, Shreve, Fontan &
McBeath 1997). In many cases this interest arises from the special difficulty
that these tasks impose upon the processes of comprehension and speech
production. For example, in simultaneous interpreting, these processes have to
be performed concurrently, in different linguistic codes and under strong
temporal pressure.
The interpreter must simultaneously concentrate on and understand a
unit of meaning or chunk of discourse in a given language (L1) while
translating and producing a previous unit of meaning or chunk of that discourse
in a second language (L2). In order to implement this the interpreter must be
able to maintain the new unit in his/her working memory, access the meaning
of the words and phrases involved, connect the information received to
previous information, and translate this unit into a new linguistic code while
producing the translated version of a previous unit.
Obviously, if the task is to be adequately accomplished, all these mental
operations must be performed efficiently. The theories proposed in interpreting
studies (Gile 1995) emphasize the role of working memory in producing
quality interpretation. The efficient use of working memory is especially
significant for the comprehension processes (Gile 1995). In fact, in interpreting
it is estimated that at least 80% of the effort or cognitive resources is devoted
to listening and understanding the discourse and only 20% to speech
production (Padilla 1995). The research presented here focuses on the
processes of comprehension in simultaneous interpreting.
The methodological approach used in most of the experiments in this
paper is that of cognitive-correlates introduced by Hunt, Frost and Lunnenborg
(1973) in their studies on human cognitive abilities. This methodological
approach starts with the selection of groups of subjects who differ in the
cognitive process or skill that we are interested in. Subjects in these groups are
 Comprehension processes 129

then compared in their performance in a series of simple tasks which are

assumed to be involved in this skill.
When applied to translation or interpreting, the method consists of
selecting subjects with different levels of skill in both tasks (practitioners,
trainee interpreters/translators, bilingual subjects etc.) and then comparing their
performance in tasks involving processes such as memory or comprehension
that are assumed to play a role in the skill. The superiority of the highly-skilled
subjects in one of these simpler tasks would suggest that its underlying process
is also present in the more complex translating/interpreting task.

2. Hypothesis

The first question that we explored was whether the comprehension processes
involved in interpreting are different in nature from those involved in normal
reading or listening and if so, how. According to psychohnguistic theories of
discourse comprehension (Kintsch 1988; Gernsbacher 1990) two groups of
processes play an essential role in comprehension, especially when it takes
place in conditions of simultaneity (Gile 1997). First, linguistic processes such
as lexical/semantic access, propositional analysis, syntactic processing etc.
need to be performed efficiently to construct a mental representation of the
discourse. Second, working memory capacity and resources need to be
distributed among the different processes involved. In our experiments we have
focused our attention on these linguistic and memory processes, and we have
attempted to determine whether they are performed differently in interpreting.
Dillinger (1994) has recently suggested that interpreters do not carry out
comprehension processes in ways that differ qualitatively from the ways in
which other people carry them out. The ability to interpret, that is, to
simultaneously understand, translate and produce discourse, is only due to the
greater linguistic skills that these people possess because of their degree of
bilingualism. That is, knowledge and use of two or more languages enhance the
linguistic skills of the interpreters and these enhanced abilities make it possible
for them to simultaneously understand and produce language. Alternatively,
other proposals (Daro & Fabro 1994; Padilla, Bajo, Carlas & Padilla 1995)
emphasise the role of working memory and linguistic training as determinants
of the interpreting skill. This debate has important pedagogical implications,
because if the first hypothesis is proved, much of the training given to potential
interpreters in cognitive skills would be irrelevant, and it should be oriented
only to increase knowledge of their second or third language. Therefore, an
important question in our experiments was to determine whether the
interpreters' training and experience had an effect over and above the possible
effect of knowledge and use of two languages.
130 Bajo & Padilla

In our first experiments we compared the performance of professional

interpreters, bilingual subjects, interpreting students and professionals (from
other fields) in cognitive tasks emphasising linguistic abilities (reading time
and accuracy, lexical access, semantic processing etc.). We expected that the
interpreters would perform better than the groups of students and professionals
from other fields. Our hypothesis was that this superiority would be due to their
training and practice of the linguistic skills involved in comprehension and not
only to their biligualism. The comparison of interpreters and bilinguals would
provide some insight into the nature of this superiority. If, as Dillinger (1994)
suggests, knowledge of two languages is the essential factor in
translation/interpretation, interpreters and bilinguals should show equivalent
performances. The comparison of bilinguals and interpreters involves
evaluation of the performance of two groups with superior linguistic skills due
to their knowledge of two languages. Hence, if there are no significant
differences in their performance, and they are both superior to the students and
other professionals, we would have an indication that the quality of interpreting
depends much more on these skills than on factors such as memory or
attention. On the other hand, if the interpreters are superior to the bilingual
subjects we would assume that the linguistic skills acquired through second
language learning are not the essential determiners of good performance in
interpreting.

3. Subjects

The interpreter group was composed of 10 practitioners, trained at the School

of Translators and Interpreters of the University of Granada (former degree).
Half of them had just passed their final interpretation exam with excellent
results; the other half had been working as interpreters for about 5 years.
The bilingual group was composed of 8 bilingual (English, Spanish,
German or Arab) subjects; 6 of them had been living in the country of their
second language for an average of 13 years, two of them regularly spoke one of
the languages at home and the other at school. None of them were translators
or interpreters, but they were students at the University of Granada.
The student group was composed of 10 students from the second year
of the interpreting program. They had not yet had any training in simultaneous
interpreting when they performed the experimental tasks (at the beginning of
the academic year).
Finally, a group of 10 professionals from other fields made up the
control group. The persons selected for this group had finished their
undergraduate studies in fields such as Linguistics, Arts, Philology and Law.
All of them had an excellent academic record.
 Comprehension processes 131

4. Comprehension task: method and results

Every subject was asked to perform a moving window comprehension task

adapted to Spanish by de Vega (1987). This moving window method allows
the progressive presentation of the material by the subject him/herself At the
beginning of the trial, the subjects viewed a computer screen filled with
patterns of spaces in the place where text would usually be found; dashes
corresponded to letters, and spaces corresponded to spaces between words.
Successive words were revealed by pressing a key; except for the current word
the screen remained masked with dashes. With each subsequent key press, the
previous word was masked and a new word appeared. Words appeared to
march across the screen at a pace set by the readers' key presses. The interval
between key presses defined reading times. In this way this method allows the
measurement of reading speed. In addition, after presentation of the text,
subjects had to answer a set of questions that appeared on the screen. The
purpose of these was to evaluate understanding and recall of different aspects
of the text. The moving window task thus permits an evaluation of the speed
and accuracy of comprehension processes.
Figure 1 shows the proportion of correct recall for each of the groups.
As can be seen, the bilinguals tend to be slightly worse than the rest of the
groups, but this tendency did not reach significance (p > 0.05). Thus all the
subjects showed an approximately similar understanding and recall of the
stories.

Figure 1: Comprehension: % correct recall

However, there were significant differences in reading times among the groups
(F(3,34) = 2.77, p = 0.05), in the analysis of variance. Comparisons between
the different groups showed significant differences between the interpreters
132 Bajo & Padilla

and non-interpreters (p = 0.01) and between the interpreters and the bilinguals
(p = 0.05). Figure 2 shows mean reading times obtained for each group.
Results seem to indicate that the interpreters show greater efficiency in
processing the information present in the text. They were capable of reading at
greater speed with no decrease in accuracy in their understanding and recall of
the text. Our first experiment thus indicates that the superiority shown by
interpreters is not only due to the linguistic skills derived from their knowledge
of two languages, but possibly to the training in some other skills involved in
comprehension. In our second and third experiments we attempt to clarify
whether other linguistic processes normally trained in interpretation schools are
also enhanced in the interpreters.

Figure 2: Comprehension: Reading times

5. Lexical decision and categorization: method and results

Part of the interpreters' training concentrates on reinforcing their skill in

accessing semantic information. For instance, some exercises consist of
quickly verbalizing a synonym or a term related to a word given by the
lecturer, clozing, etc. This training concentrates on immediate understanding of
the meaning, but also on accessing the form of the word so as to produce it
quickly. To explore the possible superiority of the interpreters in these
linguistic processes we selected two tasks: one emphasizing lexical access and
another emphasizing semantic access.
Our subjects participated in a lexical decision and a categorization task.
In lexical decision, a word or a group of pronounceable letters (a non-word)
were presented on the computer screen. Subjects had to decide as quickly as
 Comprehension processes 133

possible if the presented stimulus was a word or a non-word by pressing the

corresponding key. In categorization a word and the name of a category was
presented and the subject had to decide if the concept denoted by the word was
a member of the category. In lexical decision we manipulated the difficulty of
the task by manipulating the frequency of the words involved. Lexical
decisions for less frequent words require more processing demands and
therefore we hypothesized that the possible differences between the groups
might then be larger.
Figure 3 shows mean response times in the lexical decision task for
words and non-words as a function of the groups. The effect of group was not
significant when we analysed the word data (p > 0.05); however, there were
significant group differences in the analysis performed on the response times to
non-word F(3,34) = 3.01, p = 0.04. As can be observed in figure 3 the
interpreters tend to be better than the rest of the groups. The differences are
especially significant if we compare the interpreters to the bilinguals (p =
0.005) in the non-word condition. Contrary to what we expected, although
word frequency had an overall effect (F(l,34) = 95.88, p < 0.01), the
interaction between frequency and group was not significant (p > 0.05).

Figure 3: Lexical decision to word and non-word

The categorization task also showed the superiority of the interpreters. In this
task we presented pairs of words and the subjects had to decide if the concept
denoted by the second word was a member of the category represented by the
first. An important variable in categorization is the typicality of the exemplars
of the category (Rosch 1975). Many categorization experiments have shown
that response times depend on typicality, with more typical exemplars being
responded to faster than less typical ones (Rosch & Mervis 1975). For this
134 Bajo & Padilla

reason we manipulated this variable in an attempt to manipulate the difficulty

of semantic access. We hypothesized that the possible superiority of
interpreters might be more evident when more difficult categorical relations
were involved.
Figure 4 shows the results of the categorization task. Both typicality
(F(l,34) = 75.14, p < 0.01) and the interaction between typicality and group
(F(3,34) = 4.402, p = 0.01) were significant. When non-typical exemplars were
involved, the interpreters' responses were faster than those of the bilinguals (p
< 0.05), the professionals control (p < 0.05) and the interpreting students' (<
0.05). Figure 4 shows that reaction times of the bilinguals, students and
controls were very similar and slower than the times of the interpreters. This
difference is more significant for the less typical exemplars (136 ms difference
with respect to the nearest group). Hence, the interpreters seem to have more
rapid access to semantic information, especially when more difficult relations
are involved.

Interp Biling Students Non-interp

Groups

Figure 4: Categorization to typical and non-typical examples

These results seem to indicate again that the linguistic skills of the interpreters
are superior to those of the bilinguals, and that, therefore, different processes
from those involved in second language competence may be responsible for
this superiority.

6. The role of training in the interpreters' superiority

We also wanted to explore whether the differential effects in linguistic

operations that we found in interpreters were really due to training. This is
important for teaching interpreting because we would be able to infer that,
 Comprehension processes 135

apart from knowledge of the second language (as Dillinger suggests, 1994),
specific training in other linguistic skills leads to better use of them.
With this purpose in mind, at the end of the academic year we asked
subjects in the student group to take part in the second phase of the study. Our
aim was to compare their performance in the second phase with that of the first
phase. Thus, the students performed the comprehension, categorization and
lexical decision tasks once at the beginning of the academic year, before
receiving training in linguistic skills, and then at the end of the year, after the
intensive training received during this period (8 subjects participated in both
phases). To take into account the possible effect of practice on the tasks, a
group of 10 university students (non-interpreters) also participated in the
experiment as controls. They participated at the beginning of the academic year
(phase 1) and at the end of it (phase 2). If, in our previous experiments, training
was the cause of the interpreters' superior performance, we expected that
interpreting students would show improved performances in the second phase.

Students Controls

Figure 5: Comprehension (phase I and II)

Figure 5 shows mean reading times for interpreting students and controls as a
function of phase. As can be seen, only the group of interpreting students
experienced a reduction in the response times from the first phase to the second
(t(15) = 2.08, p = 0.06). The slight reduction in time from phase 1 to phase 2
experienced by the control group was not significant (t(16) = 0.89, p = 0.38).
The categorization task showed the same pattern of results (Figure 6).
The control group showed no significant improvements in any of the three
conditions of the task. Thus both typical and non-typical members were
categorized equally fast in the first and second phase of the experiment, t(16) =
- 0.63, p =,0.53 for typical exemplars and t(16) = 0.61, p = 0.55 for non-typical
136 Bajo & Padilla

exemplars. On the other hand, the group of interpreting students showed

significant improvements for both typical (t(15) = 2.12, p = 0.05) and non-
typical (t(15) = 2.40, p = 0.02) exemplars. The training received by the students
of interpretation does seem to have helped them to have faster access to their
semantic knowledge.

Figure 6: Categorization to typical and non-typical examples (phase I and II)

Finally, Figure 7 shows the results of the lexical decision task. As can be seen,
there is a tendency for the interpreting students to respond faster to words in
the second phase (t(15) = 1.86, p = 0.09). This improvement was significant in
the non-word condition (t(15) = 3.37, p = 0.04). In contrast, response times for
the control group remained identical in both phases and conditions (t(15) = -
0.17, p = 0.86 and t(16) = 0.82, p = 0.42).
Our results seem to indicate that linguistic skills such as access to
semantic and lexical information that are particularly developed in interpreters
are not only due to the greater linguistic skills that knowledge of two languages
gives them. This superiority did not show in the bilingual group, which also
had linguistic competence in two languages. Our data show that these skills are
susceptible to training, as our interpreting students performed the tasks much
more effectively after training.
 Comprehension processes 137

Students Students Controls Controls

word non-word word non-word

Figure 7: Lexical decision to word and non-word (phase I and II)

7. Comprehension and working memory

A second component involved in comprehension is short-term or working

memory (Baddeley, 1990). As we know, the task of interpreting imposes a
great burden on working memory, as interpreters need to simultaneously store
information and perform a whole series of mental operations. Because of this,
interpreters seem to show a large memory capacity (Padilla, Bajo, Cañas &
Padilla 1995), and this capacity may also be the basis for their greater
comprehension skills. Psycholinguistic theories like that of Kintsch and van
Dijk, (1978) Kintsch (1988) and Gernsbacher (1990), assign to working
memory a central role in the construction of an adequate representation of the
text. Psychological research has shown that the capacity of this memory is to a
great extent responsible for individual differences in the skills of reading and
understanding sentences.
An important component of good comprehension skills is the capacity
to compute semantic and syntactic relations between words and successive
phrases so as to construct a coherent representation of the discourse. To
integrate the new information with previously processed information it is
necessary to have access to the results of previous processes. If temporal
storage of all this information is crucial to understanding, those people who
have less memory capacity will be less capable of maintaining this information
in memory and therefore will be less capable of carrying out the processes
necessary for comprehension (tracking references, making inferences,
resolving ambiguities etc.). In the same way, people with greater memory
138 Bajo & Padilla

capacity will be more capable of carrying out these processes and will show
greater reading skills.
Daró and Fabbro's research (1994) has also shown the importance of
this memory when the task is performed in the conditions of simultaneity
imposed by interpreting. For this reason we thought that the greater skill of
interpreters in the comprehension task could be due to their greater short-term
memory capacity. Therefore our next objective was to obtain data on the
memory capacity of our interpreter subjects. In this second series of
experiments, we omitted the group of bilinguals, since there was no particular
reason to suppose that they might be superior and our previous experiments
had showed that their linguistic skills were the same as those of the students.
In order to measure working memory capacity we used two tasks that
have been widely used for this purpose: the digit span task and the phrase span
task. In the digit span task, three sequences of four digits were presented first.
Then the number of digits was gradually increased until it was impossible for
the subjects to recall them in order. In the phrase span task, subjects were
presented with a group of phrases and asked to recall the last word of each
phrase. The number of phrases in the group was gradually increased. The size
of the largest group of phrases in which all the words were recalled represented
the memory span of that subject. This task was also included because it
simultaneously measures the processing and storage capacities of working
memory, while the digit span test is more related to the storage capacity.
Daneman and Carpenter (1980) have shown that this task can predict
individual differences in reading skills.
Figure 8 shows the memory span obtained by each of the groups. As
can be seen, the interpreters' memory span as measured by the two tasks was
higher than that of the rest. The analysis of variance computed on the digit span
data showed the differences to be significant (F(2, 27) = 3.26, p = 0.05). The
analysis on the phrase span data also showed the differences to be significant
(F(2,27) = 5.32, p < 0.01).
In addition we calculated the correlation between the scores obtained by
our subjects in these memory tasks and the reading times obtained in the
previous experiments. The Kendall correlation between phrase span and
reading times was significant (K(25) = 0.24, p = 0.05). These results suggest
that the memory capacity of the interpreters is larger than that of non-
interpreters or interpreting students, and that this larger capacity may be the
basis for their efficient reading skills.
Daneman and Carpenter (1980) suggest that the greatest source of
differences in reading skills is not in the passive storage of the information but
in the most active part of the processing of the working memory (the central
executive, in Baddeley's term). People differ in their functional capacity, that is
to say, in the processes they perform in order to use the limited resources they
 Comprehension processes 139

D Digits
□ Phrase

Interp Students Non-interp

Figure 8: Memory span tasks: digits and phrases

possess in an optimum way. Working memory is assumed to have processing

and storage
capacities. Hence, individuals may differ in their skill of dividing their
resources between the two functions. Daneman and Carpenter suggest that a
functionally smaller storage capacity is linked to deficits of comprehension. In
the same way a functionally greater storage capacity will correlate with greater
comprehension skill.
To explore this hypothesis we carried out another experiment in which
we attempted to occupy the greatest storage component that the working
memory has: the phonetic component (articulatory loop, in Baddeley's term).
At the same time our subjects were required to process a list of words that they
would later have to recall. The usual way of blocking the phonetic component
is by articulatory suppression, which consists of asking the subjects to
verbalize a syllable while they process other material. It is obvious that
interpreters have to learn to carry out a number of processes (understanding,
recalling or translating) in conditions where the phonetic component is
occupied transmitting the result of these mental processes. Therefore, if their
training includes other tasks, we think that they will be capable of dividing
their resources of storage and processing more efficiently, so they will be less
affected by articulatory suppression.
All the subjects performed a memory task (3 lists of 16 words 3 seconds
each) with or without articulatory suppression. In the condition of articulatory
suppression the subjects had to verbalize the syllable "bla" as they tried to
memorize the words.
140 Bajo & Padilla

As illustrated in Figure 9, the interaction between group and

articulatory suppression was significant (F(2,27) = 4.58, p = 0.01). Without
articulatory suppression all the groups recalled approximately the same number
of words (F(2,27) = 1.91, p > 0.05). However, with articulatory suppression the
interpreters' performance was better than that of the other groups (F(2,27) =
4.55, p < 0.01). The students and non-interpreters performed worse when their
phonetic component was occupied by verbalization of the syllable (p < 0.01),
whereas the interpreters maintained their performance in these conditions (p >
0.05).

Interp Students Non-interp

Figure 9: Free recall with and without articlatory suppression

It thus seems that the interpreters' superior comprehension is not only due to
their greater speed in accessing semantic information, but also to their greater
short-term memory capacity and to their efficient use of this memory.

8. Conclusions

The results of our experiments suggest that training and experience in

interpreting develop a set of cognitive skills involved in comprehension. Thus,
interpreters show fast and accurate reading abilities, faster access to lexical and
semantic information, larger working memory capacity and a more efficient
use of this capacity, so that the ability to understand and memorize a list of
words is not impaired by suppression of the phonological component.
From a theoretical point of view, these data suggest that interpreters
develop specific strategies to extend their memory capacity and efficiency.
Similarly, they enhance their comprehension processes. First, they seem
 Comprehension processes 141

quicker when accessing lexical and semantic information stored in their long-
term memory. Second, they are able to perform a series of mental tasks without
the support of a rehearsal component in their working memory (the articulatory
loop). Contrary to Dillinger's results, interpreters' superiority in
comprehension processes is not due to the linguistic abilities acquired by
knowledge of a second language, as our bilingual subjects did not show that
superiority.
In agreement with other theoretical approaches in interpreting (Daró &
Fabbro, 1994), our data support the key role of working memory in
simultaneous interpreting. Efficiency in comprehension seem to be due to
better management of working memory resources. Our data show that better
reading abilities go together with better performance in tasks implying the use
of working memory. An important component in improving comprehension is
the capacity to compute syntactic and semantic relationships between
successive words and phrases in order to build a coherent representation of the
discourse. For new information to be integrated with previously processed
information one must have access to the results of previous processes. Storing
this information temporarily is of crucial importance to understanding.
Therefore, people with a larger memory capacity will be more capable of
storing information and, consequently, will be better able to execute a series of
sub-processes needed for comprehension (looking for referents, making
inferences, solving ambiguities etc.).
According to our conclusions, the training of future interpreters must be
focused, to a large extent, on an efficient use of their memory, in such a way
that intensive practice in related tasks has immediate repercussions on the
acquisition of the technique of interpreting.

1. Preparation of this paper was supported by Grant PB95-1180 of the DGICYT (Ministerio
de Educación y Cultura) of the Spanish Government to the first author. The authors are
indebted to José Canas for his help in preparing and discussing many of the experiments
presented in the paper.

References

BADDELEY, A. 1990 Human Memory: Theory and Practice. Hove: Lawrence Erlbaum
Assoc.

DANEMAN, M & CARPENTER, P.A. 1980. "Individual differences in working memory and
reading''. Journal of Verbal Learning and Verbal Behavior 19, 450-466.

DANKS, J.H.; SHREVE, G.M.; FONTAN, S.B. & MCBEATH, M.K., eds. 1997. Cognitive
Processes in Translation and Interpreting. London: Sage Publications.
142 Bajo & Padilla

DARÓ , V. & FABBRO, F. 1994. "Verbal memory during simultaneous interpretation: effects
of phonological interference". Applied Linguistics 15 (4).

DILLINGER, M. 1994. "Comprehension during interpreting: What do interpreters know that

bilinguals don't?". Bridging the Gap: Empirical Research in Simultaneous Interpretation ed.
by S. LAMBERT & B. MOSER-MERCER, 155-189. Amsterdam & Philadelphia: John
Benjamins.

DE VEGA, M. 1987-1990. CRONOLAB. University of La Laguna. CICYT PB86-0480 and

DGICYT PB88-0425.

GERNSBACHER, M.A. 1990. Language Comprehension as Structure Building. Hillsdale,

New Jersey: Erlbaum.

GILE, D. 1995. Basic Concepts and Models for Interpreter and Translator Training.
Amsterdam & Philadelphia: John Benjamins.

GILE, D.1997. "Conference interpreting as a cognitive management problem". DANKS, J.H.;

SHREVE, GM.; FONTAN, S.B. & MCBEATH, M.K. eds, 196-214.

HUNT, E.B., FROST, N.& LUNNEBORG, C. 1973. "Individual differences in cognition: a

new approach to intelligence". Psychology of Learning and Motivation ed. by G. BOWER, 87-
122. New York: Academic.

KINTSCH, W. 1988. "The role of knowledge in discourse comprehension: A construction-

integration model". Psychological Review 95 (2), 163-182.

KINTSCH, W. & VON DIJK, T.A. 1978. "Toward a Model of Text Comprehension and
Production". Psychological Review, 85 (5) 363-394.

PADILLA, P. 1995. Procesos de atención y memoria en interpretación de lenguas.

Unpublished Doctoral Dissertation. Universidad de Granada.

PADILLA P., BAJO M.T., CANAS J.J. & PADILLA F. 1995. "Cognitive processes of
memory in simultaneous interpretation". Topics in interpreting research ed. by J. Tommola.
61-71. Turku: Centre for Translation and Interpreting.

ROSCH, E. 1975: "Cognitive representations of semantic categories". Journal of Experimental

Psychology: General 104, 192-233.

ROSCH, E. & MERVIS, C.B. 1975. "Family resemblances. studies in the internal structure of
categories". Cognitive Psychology 7, 575-605.
AT THE CROSSROADS OF TRANSLATION
STUDIES AND IMAGOLOGY

Nedret Kuran-Burçoglu
Bogaziçi University, Istanbul, Turkey

Zusammenfassung
Der Artikel befaßt sich mit dem Einfluß von "image of the other" (Bild vom Anderen)
auf den Übersetzungsprozeß. Der Einfluß kann an drei Stellen wirksam werden: (1) vor
dem Übersetzen, (2) während des Übersetzens und (3) während des
Rezeptionsprozesses. Diese drei Stellen repräsentieren Bereiche, in denen die zwei
Disziplinen Übersetzungswissenschaft (Translation Studies) und Imagologie
(Imagology) einander fruchtbar beeinflussen und dadurch ihre Perspektive erweitern
und der Forschung neue Wege weisen können.

Résumé
Quel est l'impact de "l'image de l'autre" sur le processus de traduction? Cet impact
peut avoir lieu à trois moments différents: 1) avant la traduction, 2) pendant, puis 3)
lors de la réception. La traductologie et l'imagologie pourraient s'influencer avec profit
en abordant ces impacts, de façon à élargir leurs perspectives et à ouvrir aussi de
nouveaux champs de recherche.

Resumen
Este articulo estudia el impacto de las imágenes sobre el fenómeno traductor. Antes de
centrarse en ejemplos, distingue tres posibles ámbitos de influencia entre las dos
disciplinas. Éstos son 1) antes de la traducción, 2) durante el proceso traductor y 3) en
el proceso de recepción. Finalmente, ilustrando cada ámbito con ejemplos, pretende
demostrar lo importante que resulta tanto para los Estudios de Traducción como para el
Estudio de la Imagen considerar las influencias reciprocas para asi ampliar sus
perspectivas y abrir nuevas areas de investigación.
144 Nedret Kuran-Burçoglu

1. Introduction

Roland Barthes states in his book Mythologies that images are not God-given,
but created by men; he goes on to say that

"meanings are produced through the codes at work in

representations and that while meanings might appear to be
natural, obvious and immanent, they are in fact produced: they are
constructed through identifiable processes of signification at work
in all representations." (Barthes, in Kuhn 1987:5)

Indeed there is much evidence demonstrating that images are created by man. If
the processes used in creating the images are properly assessed, several
constraints will be observed that are significant in shaping the final product.
These are temporal and spatial, as well as socio-cultural and political
constraints, i.e. value judgements, prejudices, conventions, habits, traditions,
political and power relationships. Within the signification process, these are
concretized as representations, the corpus of which will eventually create the
image. I use "concretized" in the sense of Iser (1972, 1976). "This term is
sometimes used synonymously with realization, but it carries the strong
implication of making actual what was actually virtual" (Mario Valdes in
Makaryk 1994:527).
A critical approach to this creation process challenges the reliability of
the images. Roland Barthes' view that the meanings attached to the images may
well appear "natural" and "immanent" deepens this scepticism and concern even
more, as it suggests that images may be misleading and thus insidious because
they seem "innocent" and "naive" at first sight.
This conclusion justifies Edward Said's criticism (1978, 1993) of the so-
called "orientalist gaze," as well as Frantz Fanon's (1986) and Homi Bhabha's
(1994) cautious approach to images that fall within "postcolonial discourse".
Another example can be cited from the Image Studies carried out by
imagologists in the analysis of "Nazi discourse". Feminist, gay and lesbian
studies are also greatly concerned about the attitude of the public towards their
subject matter. Such examples indicate how powerful images can be in shaping
human relationships by various means and on various levels. Imagology, which
investigates the genesis of "the image of the other," as well as the socio-cultural
constraints influential upon its creation and transformation processes, tries to
highlight these complex phenomena.
Within the scope of Translation Studies, "the image of the other
(culture)" and its imagological analysis should, in my mind, not be
underestimated, as this image very often plays a formative role in the translation
phenomenon, and the translations in turn may have an initiating, formative or
 Translation studies and imagology 145

transforming effect on the emerging or already existing image of the other.

Hence, this dual influence must, in my opinion, be considered by both of the
disciplines, Translation Studies and Imagology, as it would bring new insight
and awareness to these fields, adding a significant dimension which seems to
have been neglected so far, particularly by Translation Studies.
This paper investigates the impact of images on the translation
phenomenon and will try to explicate the intersecting points of Translation
Studies and Imagology, two closely related significant branches of Comparative
Cultural Studies.
This is done by focusing on those stages of the image creation process
and the translation process where the two disciplines may intersect. In this
paper, the issue will be discussed from the perspective of Translation Studies,
focusing on the three stages of the translation process where the image of the
other has a potentially formative role. I will give some examples demonstrating
the influence of the image of the other in translation.

2. Intersecting points of the two disciplines

The stages of the translation phenomenon where the image of the other may
have an impact are the following. (See also Kuran-Burçoglu 1995, 1998,
forthcoming (a).)
(1) Prior to the translation process. The translator's choice of the text s/he is
going to translate may be directly or indirectly guided or influenced by the
image of the other. This decision (among others) is designated by Gideon Toury
as a "preliminary norm" in Translation Studies (Toury 1980, 1985).
(2) During the translation process. During the translation process the translator
has to make a number of decisions which are reflected in the product, i.e. the
target text, in terms of omissions, additions or lexical choices. During this stage,
designated the "encoding process" in Semiotics and Communication Studies,
the image of the other, in the mind of the translator, can influence the target
text.
(3) During the reception process of the target text. The reader's choice as to
which translated text to read, as well as the so-called reception process of the
target text, designated as such by Communication and Reception Studies, may
also be influenced by the image of the other.

3. The impact of "the image of the other" on the translation phenomenon

After having pinpointed the potential intersecting areas of the two disciplines, I
would like to cite examples for each case.
146 Nedret Kuran-Burçoglu

3.1 Before translation

The first of two examples illustrating the first case is taken from the history of
Turkish translation, which has two significant eras: the Tanzimat and the
Cumhuriyet. The former is the so-called "Renovation era," which started with
the Declaration of Tanzimat in 1839 and lasted into the 1850s, while the latter
began in the years following the Declaration of the Turkish Republic in 1923,
and lasted through the 1940s. These eras mark the two periods of radical
transition in Turkish culture from Eastern to Western cultural norms and values.
It can be justifiably claimed that the Tanzimat era had paved the way for the
Cumhuriyet era. During these two periods of modernization in Turkish culture,
which were naturally accompanied by a series of reforms, translation gained
impetus. Special institutions such as the Encümen-i Danis, and the Tercüme
Bürosu, in the 1840s and 1940s respectively, were set up to select works that
should be translated within the scope of a program encompassing the translation
of world masterpieces into the Turkish language. Among them, works from
Western literature occupied a significant place. In the Cumhuriyet era they were
thought to serve as models for a new nation that was undergoing a cultural
restructuring. Translators were encouraged by competitions, as well as by
commissions from such state institutions as the Ministry of Education and the
Ministry of Culture. During the Tanzimat era, as well as during the Cumhuriyet
era, "Western culture" was associated with concepts such as developments in
technology, rationalism, positivism, secularism and modernism by the Turkish
elite, comprised principally of medical doctors and military men, who naturally
linked them to the Enlightenment. These concepts emerged as a positive image
of the other as opposed to a slightly negative image of Eastern culture, which
was then associated with ideas of backwardness, superstition and
fundamentalism.
The second example illustrating influence prior to the translation itself is
taken from German cultural history during the German Enlightenment period
starting in the 1720s, when the society was undergoing significant change.
German intellectuals emphasized the effectiveness of translations in reshaping
the culture of the German community according to models taken from particular
European countries whose cultures were highly esteemed. These countries were
France and England. While Johann Christopher Gottsched - German
philosopher, theologian and literary scholar - was in favor of French neo-
classical models, German poets like Bodmer, Breitinger and Gotthold Ephraim
Lessing were in favor of English models, as they thought that the English values
would be more appropriate for them to adopt than any others (Kuran-Burçoglu
1995). These judgements, diverse though they were, were actually both based
on favorable images of those particular cultures, and the priorities of these
intellectuals, quite influential men in their country, played a significant role in
 Translation studies and imagology 147

the choice of the literary works to be translated into German in that period of
German history.
These two examples show how the image of the other was instrumental
in shaping period norms in particular countries at turning points of their cultural
histories; it also affected the choice of translators, either directly through
commission by the state and its institutions, or indirectly through public opinion
shaped by policy makers and/or influential people, such as the intellectuals of
the country.

3.2 During translation

For the second intersecting stage of the two disciplines, the translation process
itself, I will cite two examples based on translations of the Flemish poets
Timmermanns and Karel van de Vosteijne into the German language. The
examples are taken from Johan Soenen's article "Imagology and Translation"
(1997). In this article Soenen also emphasizes the significant impact of the
image of the other on translation. He draws attention to particular stylistic
changes, mitigations and intensifications in content, omissions and additions
made by the translator during the translation process, which transform the
original text wittingly or unwittingly according to "the stereotyped image."
Soenen considers those as constraints that stem from "image-bound
interpretations" on the part of the translator (Soenen 1997:128-129)
He first elaborates on the "stereotyped Flemish image" of the period in
Germany, a mixture of sensuality and mysticism, and then points to the
popularity of German translations of the Flemish author Timmermanns' works
into German, which confirm and justify "the existing Flemish Image" in the
minds of the German public.
Soenen then discusses the case of Karel van de Vostaijne, whose works
have also been translated into German in such a way that they reflect "the
existing Flemish image" in the minds of the target public, which actually
represents a distortion of the source text. In order to satisfy the receptor
audiences and justify "their existing stereotyped image" the translator has
transformed the innocent, almost platonic love scenes of the original poems into
erotic scenes through using stylistic elements and emotive words that reflect
sensuality and mysticism (ibid.).
These first two intersecting areas are called "the linking points of
Comparative Literature and the Translation phenomenon" by the imagologist
Hugo Dyserinck (1981). The impact of the image of the other in these two areas
inevitably leads to the third potential influence area, the reception stage.
148 Nedret Kuran-Burçoglu

3.3 During reception

At this third stage, under certain conditions, the following kinds of results can
be observed.
If the target reader is unfamiliar with the source culture, i.e. if translation
is for her/him the first means of getting to know the foreign culture, and hence
the image of the other is not developed yet, translation may strongly influence
the newly emerging image. In this case the image of the other which is in the
minds of the commissioner and the translator will, via the skopos (Vermeer
1990) and its implementation, have a particularly strong impact on the creation
of the image of the other in the mind of the reader. Translations done from the
languages of remote cultures for the first time represent channels through which
these foreign cultures are introduced into the polysystem, in Itamar Even-
Zohar's (1979) terms, of the target culture. Examples of this situation from
former centuries are abundant, but there are also recent cases.
Here I will cite an example that is taken from another recent study done
by Johan Soenen on the image of the Turkish literature in Flanders. The list of
translated works from Turkish literature into Flemish shows that up to recent
times Dutch and Belgian readers have not been familiar with this literature, but
that during the 1980s and 90s translations from this literature have flooded the
book markets. The reception research Johan Soenen did on Dutch and Belgian
readers has revealed that the Turkish literature is received as "the reflection of a
world of harsh realities, brutal current affairs, as well as the existing opposition
to it" by the reading audience (Soenen, forthcoming). The author of this study
explains this "one-sided," "stereotypical," "caricatural," "incomplete," and
"subjective" image (ibid.) as a natural result of the works that have been
selected for translation from the inventory of current Turkish literature by the
publishing houses, i.e. the commissioners, as well as by the translators. These
were books written mainly by authors who for socio-political and economic
reasons have migrated to Belgium or to other European countries. It is the
ideological content of the works - the violation of human rights, problems of
democracy, high inflation rate, the Kurdish problem, as well as religious
fundamentalism and the disadvantaged position of women - rather than their
literary quality that seems to have attracted the commissioners and the
translators to select them for translation, as this would justify the "stereotyped
image" of the Turk in the minds of European readers in the 1990s (ibid.) and
thus guarantee a success. The sudden increase in the number of such books that
were translated from Turkish to various European languages after the military
coup of 1980 in Turkey indicates that translation has certainly played a
significant formative or transformative role in the image of the Turk in Europe
today.
 Translation studies and imagology 149

Images of the other can be maintained, reinforced, strengthened,

challenged, or questioned, modified, transformed or completely changed during
the reception process of the translated texts. These diverse situations can be due
to time and space constraints, as well as the socio-cultural background and
knowledge of the receptor public; but they may also be due to the above-
mentioned constraints initiated by the translator and the commissioner, whose
shaping or changing function should by no means be underestimated.

4. Conclusion

I hope that by pinpointing the potential intersecting spheres of Translation

Studies and Imagology, and by exemplifying certain situations in translation, I
have shown how significant an analysis of the image of the other is for
Translation Studies. It is an aspect that should be integrated into this
interdiscipline.

References

BARTHES, Roland. 1957. Mythologies. Paris: Le Seuil. Translated by Annette Lavers, 1972.
Mythologies. New York: Hill and Wang.

BARTHES, Roland. 1964. Eléments de sémiologie. Paris: Le Seuil. Translated by Annette

Lavers & Colin Smith. 1967. London: Cape.

BHABHA, Homi K. 1994. The Location of Culture. London & New York: Routledge.

BRINKER-GABLER, Gisela (ed.)1995. Encountering the Other(s). State University of New

York Press.

DYSERINCK, Hugo. 1981. Komparatistik. Bonn: Bouvier Verlag.

EVEN-ZOHAR, Itamar. 1979. "Polysystem Theory". Poetics Today 1.

FANON, F. 1952 Peau noire, masques blancs. Paris: Le Seuil. Translated in 1986. Black
Skin, White Masks. London: Pluto.

ISER, Wolfgang. 1972. The Implied Reader. Baltimore: John Hopkins.

ISER, Wolfgang. 1976. The Act of Reading. Baltimore: John Hopkins.

KUHN, Anette. (1985). 1987. The Power of the Image. London, New York: Routledge and
Kegan Paul.
150 Nedret Kuran-Burçoglu

KURAN, Nedret. 1992."Ties between Contemporary Approaches to Translation and Various

Theories of Literature". KOINE, Annali della Scuoli Superiore Per Interpreti e Traduttori San
Pellegrino 2(1-2). 219-229.

KURAN-BURÇOGLU, Nedret. 1995. "Çagdaç Alman Çeviribilimcilerin Yaklasimlari"

Çeviri ve Çeviribilim Üstüne Söylemler ed. by RIFAT, Mehmet, 33-54. Istanbul: Düzlem
Yayinlari.

KURAN-BURÇOGLU, Nedret (ed.). 1997. Multiculturalism: Identity and Otherness.

Istanbul: Bogaziçi University Press.

KURAN-BURÇOGLU, Nedret. 1998. "Çeviri ve Yazin Kuramlari Arasindaki Baglar".

Yabanci Diller Egitim Bölümü Dergisi ed. by TAPAN, Nilufer et. al., 137-148. Istanbul:
Istanbul University Press.

KURAN-BURÇOGLU, Nedret. (Forthcoming, a). "Öteki imgesinin Olusmasinda Çevirinin

Belirleyici Rolü" (Le role déterminant de la traduction dans la constitution de l'image de
l'autre) 1er Colloque International de Traduction: Aspects Culturels de la Traduction. 1997.
Istanbul:Yildiz Technical University Press.

KURAN-BURÇOGLU, Nedret (ed.) (Forthcoming, b). The Image of the Turk in Europe from
1923 to 1990s.

MAKARYK, Irena R. (ed.) 1994. Encyclopedia of Contemporary Literary Theory. Toronto:

University of Toronto Press.

SAID, Edward. 1978. Orientalism. New York: Vintage Books.

SAID, Edward. 1993. Culture and Imperialism. New York: Vintage Books.

SOENEN, Johan. 1997. "Imagology and Translation". KURAN-BURÇOGLU, Nedret (ed),

125-138.

SOENEN, Johan.) "The Image of Turkish Literature in Flanders". KURAN-BURÇOGLU,

Nedret (ed.), forthcoming (b).

TOURY, Gideon. 1980. In Search of a Theory of Translation. Tel Aviv: The Porter Institute
for Poetics and Semiotics.

TOURY, Gideon. 1985."A Rationale for Descriptive Translation Studies". The Manipulation
of Literature ed. by Theo HERMANS, 16-41. London & Sydney: Croom Helm.

VERMEER, Hans. 1990. Skopos und Translationsauftrag - Aufsätze. Heidelberg: Universität.

 PART IV

Translation effects
FOREIGNIZING STRATEGIES IN DRAMA
TRANSLATION
THE CASE OF THE FINNISH OLEANNA

Ritva Leppihalme
University of Helsinki, Finland

Zusammenfassung
Eine exzessive Treue zu ausgangstextuellen linguistischen Anweisungen kann bei
Dramenübersetzung Probleme durch Erschwerung der Relativierung der adressierten
Inhalte beim Zielrezipienten hervorrufen. Das Stück Oleanna (1992) von David
Mamet war Anlaß für beträchtliche Debatten in der englischsprachigen Welt, während
die finnische Produktion in Helsinki (1994) nur geringen Einfluß hatte. Dieser Artikel
beinhaltet die Behauptung, daß die finnische Rezeption wenigstens teilweise durch die
bewußte Übersetzer-Entscheidung zur sprachennahen Imitation des Dialogs von Mamet
hervorgerufen war.

Résumé
Au théâtre, une traduction à la fidélité excessive peut mettre le public-cible en difficulté
face aux questions soulevées par l'oeuvre originale. Si la pièce de David Mamet,
Oleanna (1992), suscita de multiples controverses dans les pays anglophones, en
Finlande en revanche, sa traduction (1994) ne souleva guère de polémique. Cette tiède
réception peut être attribuée en partie à la décision du traducteur d'imiter, verbalement,
le dialogue de Mamet.

Resumen
El publico receptor de una traducción teatral puede experimentar dificultades para
identificarse con el tema si la traducción ha mantenido una fidelidad excesiva a los
recursos lingüisticos del texto original. La obra de David Mamet, Oleanna (1992),
suscitó una gran polémica en el mundo angloparlante mientras que una producción
fmlandesa puesta en escena en Helsinki (1994) tuvo poca repercusión. El presente
articulo sugiere que la recepción que tuvo en Finlandia se debió, al menos en parte, a la
decision consciente del traductor de seguir literalmente a nivel verbal los diálogos de
Mamet.
154 Ritva Leppihalme

1. Introduction

Recently, writings on drama translation have focused largely on the rewriting

needed when there is a new directorial interpretation or directorial subversion,
for example the feminist production of Shakespeare's Taming of the Shrew
discussed in Knowles (1997); or when there is a considerable cultural and/or
time gap between source and target texts, as with many non-Western
productions of Western classics, or the "multicultural" theatre briefly
commented on by Bassnett (1998:106); or when the translation is intermedial,
as when a play written for the stage is translated for the radio (ibid:96-98). But
what happens when none of these applies, when a translated play meets with an
audience familiar with its source culture through frequent exposure in all the
media, there is a topical theme, a realistic mise en scène, and practically no time
gap? Then the linguistic and cultural transfer might be expected to be fairly
smooth. However, the case of the first and so far only Finnish production of
Oleanna suggests that the translation of even this deceptively simple text can be
fraught with problems.
David Mamet's (1947- ) play Oleanna (1992) may be widely known by
reputation simply as a tale of harassment involving a male professor and a
female student at an American university, but critics and theatregoers have
recognized that it also deals with issues of language and power in a thought-
provoking way. At the "David Mamet at 50" conference in Las Vegas in 1997,
there were twelve presentations on various aspects of the play. In one critic's
words, it "ultimately deals far more with power trips than harassment" (Kerns
1997:1); for another, it was "one of the most stimulating experiences I've had in
a theater. In two acts, [Mamet] succeeded in enraging all of the audience"
(Ebert 1995:1). Both the critical success of the play and the heated debates
among audiences in the English-speaking world no doubt contributed to
decisions in theatres in other countries to produce the play in translation.
In Finland, Oleanna was produced in the Finnish National Theatre in
Helsinki in early 1994, fairly soon after its tumultuous reception in the English-
speaking world. The reception in Finland was a great deal more subdued,
however: there was certainly no public discussion of the issues the play raises.
Lack of interest in a play may, theoretically speaking, be due to many factors
(such as the theme of the play, the actors' performance, the director's
interpretation, or the amount of advertising and of media interest generally, to
name only some of the more obvious ones), but I would argue that in this case,
the translation strategies chosen by the translator, Michael Baran,1 may well be
at least partly the explanation. While studies of theatre translation and theatre
semiotics often make the point that it is not possible to separate text from
performance as the text is only one element in the theatrical polysystem
(Aaltonen 1996:92), the analysis below will bypass the other elements and focus
 Drama translation 155

on the translation, to show that the foreignizing strategies consciously chosen

by the translator in fact conflict with the dramatic aims of Mamet's dialogue.

2. Comparison of source and target texts

2.1 The pragmatic level

The immediate impression given by the Finnish translation of Oleanna as soon

as the characters try to communicate is that of strange and unnatural language.
It might be argued that this is precisely the desired effect, and that the audience
is expected to feel that the language is unnatural, reflecting as it does the
inability of the two characters to find a common language. But the awkwardness
of the dialogue in the Finnish version hardly reflects that; rather, it sounds like a
form of interlanguage, with neither character a native speaker of Finnish:

Example (1)
Source text:

JOHN: Is that what you want to talk about?

CAROL: . . . to talk about. . . ?
JOHN: [...] Don't you think . . . ? (Pause)
CAROL: . . .Don't I think. . . ?
JOHN: Mmm?
CAROL: . . . did I. . . ?
JOHN: . . . what?
(Mamet 1993 a: 3)

Target text:

JOHN: Siitäkö sinä haluat puhua?

Ts that what you want to talk about?'
CAROL: . . . puhua . . . ?
'. . . to talk about. . . ?'
JOHN: [...] Eikö sinusta . . . ? (Pause)
'Don't you think . . . ?'
CAROL: . . . eikö minusta . . . ?
'. . . don't I think. . . ?'
JOHN: Mmm ?
'Mmm?'
CAROL: . . . sanoinko . . . ?
'. . . did I say . . . ?'
156 Ritva Leppihalme

JOHN: . . . mitä?
\ . . what?'
(Mamet 1993b:2)

Word for word, the target-text lines certainly replicate the fits and starts of
Mamet's dialogue, but without its basis of familiar domestic speech patterns.
Thus the translation deprives the audience of a disturbing feeling of recognition:
that this is how many of their own conversations go, too.
Such pragmatic differences between American and Finnish conversation
styles are particularly noticeable on the stage, where the presence of actors
speaking the language of the audience creates the illusion that the characters
belong to the same language culture as the audience. While dramatic dialogue is
not intended to represent authentic conversation with the same degree of
accuracy as do transcripts made for conversation analysis, and Mamet's
dialogue is clearly stylized, there is still a need for interaction that makes sense
in pragmatic terms. Mechanical repetition of the interlocutor's words does not
always work very well in the translation as Finnish speakers would sometimes
choose to repeat other words in the lines: in the above extract, for example, the
repetition of puhua 'to talk' suggests wonder at the idea that with two people in
a room, one perhaps wishes to talk about something to the other! What is more,
many conversational strategies are different in Finnish. The pace of Finnish
speech is usually slower than that of (American) English and the toleration of
silences between speeches much higher (Kukkonen 1993:100-111). In Finnish
conversation there would likely be less of a scramble for the turn to speak and
less of a desire to interrupt one's interlocutor; in fact, Finnish speakers
commonly use silences to indicate non-comprehension and disagreement. In an
alternative Finnish translation of Oleanna, the techniques used by Mamet to
indicate lack of communication could perhaps at times be replaced by other
strategies.
Any representation of dialogue in Finnish needs to be in one of two
alternative registers. The dialogue of the play can be written in formal literary
language — for a long time the only possible style on the stage except the
country dialects spoken mostly by comic characters. Or the characters could be
made to speak colloquial urban spoken Finnish, the language of real-life Finnish
students and professors. The latter is a choice often made when contemporary
plays, either domestic or translated, are produced, and has been used
successfully in Finnish productions of Mamet's plays both before and after the
production of Oleanna. The use of the formal register almost inevitably results
in a distancing effect, as this is the language variety that tends to be spoken in
public, in front of audiences, but not in face-to-face encounters. The translator's
choice of the formal register in the dialogue of John and Carol may well be
symptomatic of the distance between the two speakers; but it also, at the same
 Drama translation 157

time, distances the speakers from the audience. If this distance is increased
further by lines that sound un-Finnish, this is also likely to have an effect on the
responses of the audience. I will return to this point below (section 3).
A further difference between the translated dialogue and real-life Finnish
speech is the use of question tags and pragmatic particles. If English usage is
reproduced, as in this translation, so that nearly every well, all right, and yes is
retained and rendered by their dictionary equivalents, their communicative
significance is overlooked. As Östman (1982:144) points out, "the manner of
expressing pragmatic information varies from one language and culture to
another, and [...] the need for expressing this information via explicit
manifestations or implicit suggestions will also vary across languages". The
literal approach adopted by the translator also leads to a striking absence of the
Finnish pragmatic particles that would be common in precisely the sort of
situationsMamet'ss text represents, where speakers find neither understanding
nor agreement.

2.2 The semantic-lexical level

One relevant difference between English and Finnish is the opacity of learned
English vocabulary compared to the transparency of corresponding words in
Finnish. Towards the end of Act I, Carol is uncertain about John's word
predilection (Mamet 1993a: 31), checking whether it means "liking". In Act III
she requests that John explain to her the word indictment (63). A little later
(66), she erupts in fury at the word transpire and demands that he say happen
instead. While Finnish also has its share of Greek and Latin loan words, these
are generally either fully domesticated everyday loans (such as hotelli 'hotel',
televisio 'television'), alternatives to more common transparent words (for
example traditio 'tradition' = perinne) or strictly technical terms in various
special fields (diskurssi 'discourse', metafora 'metaphor', elektroni-
mikroskooppi 'electron microscope'). Finnish semantic equivalents of vast
numbers of learned loanwords in English are by contrast of vernacular origin.
With such difference in lexis, is it possible to convey the implications of
John's choice of opaque words — partly perhaps the reason why he is labelled
elitist in Carol's report to the tenure committee? Can a distinction be made
between Carol's preferred everyday vocabulary and the academic vocabulary
which is part of the professorial mask behind which John tries to hide? If we
look at the three instances singled out above, we find indictment rendered with
a technical legal term, kanne 'action, suit' (Mamet 1993b:37); and predilection
with lukkarinrakkaus (18), a jocular and somewhat dated expression of roughly
the required meaning. Both renditions work reasonably well even though they
are derived from mainly domestic roots: it is conceivable that Carol might be
uncertain about the precise meaning of kanne and unfamiliar with the other
158 Ritva Leppihalme

expression, which is more appropriate to someone of her parents' or

grandparents' generation. Still, the latter term particularly fails to convey John's
pretensions to intellectual superiority in his role as professor. With the third
word, transpire, at which Carol's fury erupts:

Example (2)
CAROL: "Happen"?
JOHN: Yes.
CAROL: Then say it. For God's sake. Who the hell do you think that you are?
(Mamet 1993a: 66)

the rendition into Finnish is weak. John is made to say Ja mitä seuraisi 'And
what would transpire?' (Mamet 1993b:39). Carol violently demands that he use
the word tapahtua 'happen' instead. Unhappily, both words are equally
standard items of Finnish vocabulary with no difference in distribution or
register. This deprives Carol's outburst of all logic in the framework of the play.
From the functional point of view and as an indication of the different
vocabularies it might be a useful translation strategy to pepper John's lines with
a number of scholarly words in Finnish by way of compensation, even where
such words do not occur in precisely the same lines in English. The occurrences
in the source text of words like paradigm, which Carol also queries (Mamet
1993a:45), are helpful in that such words are part of Finnish academic
vocabulary as well and hence can be retained (nearly) unchanged. An item of
restricted register, even with a different meaning, could be substituted for And
what would transpire?, making Carol's sudden rage more intelligible. Such a
strategy would be justified on the grounds that functional equivalence takes
precedence here over semantic equivalence.

3. Discussion

This paper has established that the global translation strategy chosen by the
Finnish translator of Oleanna ignores several pragmatic differences between
American English and Finnish, and does not always result in appropriate
semantic renderings for crucial words in the dialogue. The dialogue thus sounds
unnatural and distant, with the occasional rendering even puzzling for the
audience. Thus the literal translation weakens the dramatic effect of the play.
Mamet's dialogue is known to be carefully crafted, and in Oleanna it
emphasizes the miscommunication caused by the different speech styles of the
two characters. Their interruptions, repetitions and evasions try to give an effect
of authentic conversation and hence to bring the spectators to a realization that
despite the torrent of words there is little true contact between the two
 Drama translation 159

characters — and that this might be true of their own lives as well. The
translator's decision to imitate Mamet's dialogue as closely as possible on the
verbal level is doubtless based on respect for the author's use of language, but it
fails to take into account the differences in source and target language cultures:
it does not enable the target audience to recognize familiar speech patterns and
so to become aware of the lack of communication. Though the words of the
characters are in Finnish, the conversation is not Finnish but foreign.
Judging by a short paper by the translator (Baran 1998:36-39), it is a
deliberate decision on his part to translate plays as literally as possible. "I would
rather be too literal than provide creative approximations", he says when
discussing his principles, though without reference to any particular play
(translation RL). Baran also addresses the choice of the formal "book language"
register when he asserts that the informal variety he rejects aims at "making the
translated discourse sound tougher, more fluent, more modern" (38), and that
such "fluency strives for a feeling of familiarity rather than for the truth" (38).
He thus rejects familiarity as a goal, preferring to foreignize.
Foreignizing strategies may of course be legitimate and even
commendable under certain circumstances. Lawrence Venuti (1995) argues that
fluency in translation is the result of domesticating strategies and that a fluent,
"invisible" translation often expresses repression of cultural differences,
ethnocentrism, and even xenophobia. In his view foreignizing strategies "open
up a space for something that's less ethnocentric" (47). But Venuti admits that
his argument mainly applies to a situation where the target culture knows very
little about the source culture (48), in other words where the source text
represents a little-known and peripheral culture in the eyes of the target
readership; and that is clearly not the case where a contemporary American play
by an acclaimed and well-known author is translated for theatregoers in what
has (rightly or wrongly) been called the most Americanized country in Europe.
The reception of the play in Finland can be characterized as indifferent.
While the play's issues of sexual harassment2 and political correctness may well
be less explosive in Finland, with a tradition of considerable equality between
the sexes and a relatively homogeneous population, surely miscommunication,
lack of contact and abuse of power are equally thorny problems in both
language cultures. And yet, the Finnish production failed to provoke much
discussion or debate. When I saw the translated version of the play in Helsinki it
was evident that during the intermission and after the performance there was
neither excitement nor heated discussion among members of the audience,
contrary to what had been reported to be the case in American theatres where
"loud arguments broke out" (Ebert 1995), and the play provoked "vigorous
after-show discussion and debate" (Gist 1996). What commentary there was in
Helsinki seemed to focus on Carol as a student presumed to be neurotic. The
larger issues were thus overlooked. Of course, there are cultural differences in
160 Ritva Leppihalme

theatre-going norms, and "loud arguments" in public could hardly be expected

in Finnish theatres; still, couples and groups in the audience could well have
discussed the themes in the interval and expressed different opinions on them
without breaking any norms.
Finnish theatre critics, aware of the play's significance in the English-
speaking world, gave the production reasonable reviews, but the play folded
after 43 performances on the smallest stage of the theatre (seating 154). The
Finnish reviews generally praised Mamet's play and noted that its theme of
sexual harassment was topical and controversial. They also recognized that the
play deals with power, struggle for equality, freedom of speech and political
correctness as well as problems in communicating. These issues were seen to be
of universal importance: "We are all involved in the problems it takes up"
(Kajava 1994). The actors were mostly praised for giving intense and disciplined
performances, though a few critics thought them too stolid. Practically every
review noted that Oleanna had been an international success. While some
thought the Finnish production "simple and clean" (Yliluoma 1994), others
regarded it as "clinical" (Majapuro-Joutsamo 1994) or "static, strongly based on
the word" (Ruuskanen 1994). There were occasional complaints that the
dialogue "tasted of paper" (for instance Yliluoma 1994) - whether this was the
fault of the author, the translator or the actors was not made clear. The
translation itself was rarely commented on (in a discussion of Finnish
professional translators on their electronic mailing list, Translat, in 1998 it was
agreed that reviewers in general may neither be, nor think themselves to be,
competent to evaluate translations). In sixteen reviews3 the translation was
mentioned only four or five times, always briefly, with opinions ranging from
"artfully ambivalent" (Aulavuo 1994) (the only positive comment) to
"unnaturally stiff' (Tirkkonen 1994).

4. Conclusion

The lack of interest in the play in Finland may well, as argued in this paper, be
related to its translation. In Christiane Nord's words,4 cultural distance is
incompatible with the appelative function. While faithful on the verbal level, the
Oleanna translation disregards pragmatic considerations and differences in
source and target semantic-lexical resources, and thus deprives the play of much
of its dramatic tension as it fails to create human contact between the characters
and the audience. Some familiarity with conversational analysis and with
intercultural differences in conversational styles might have contributed to a less
literal but more effective version, closer to what Mamet was aiming at. A
misguided attempt to respect the language of a famous dramatist thus led to a
translation that did less than justice to his text.
 Drama translation 161

Notes

1. Baran is the theatre's dramaturge. A dramaturge in a Finnish theatre adapts texts for use
in the theatre, and can also write and translate such texts (Aaltonen 1996:28, 213, 219).
While it would be fruitless to speculate on circumstances which may have had an effect
on this particular translation (such as the amount of time at the translator's disposal or
the expectations of the director, Kurt Nuotio), it can be noted that a dramaturge is
generally thought to hold more power in a theatre than do actors or those translators who
are not directly connected with production (Aaltonen 1996:54).

2. In 1995, an inquiry at Helsinki University showed that 2.7% of students and 6.8% of staff
had encountered some sexual harassment in the past two years. C. 75% of the
respondents were female. I would like to thank Teija Mankkinen for providing these data.

3. Kindly provided by the Finnish National Theatre. All translations from Finnish are mine.

4. In a lecture on conventions and norms in translation, given at the University of Helsinki

in March 1997.

References

AALTONEN, S. 1996. Acculturation of the Other. Irish Milieux in Finnish Drama

Translations. Joensuu: Joensuu University Press.

AULAVUO, A. 1994. "Syytettynä: mies". Pohjolan Sanomat, 14 Feb. 1994.

BARAN, M. 1998. "Ruumiinavauksia ja muita harrastuksia". Käännetyt illuusiot:

Näytelmäkääntäminen suomalaisessa te atterissa ed. by S. AALTONEN, 36-39 Tampere:
Tampere University Press.

BASSNETT, S. 1998. "Still Trapped in the Labyrinth: Further Reflections on Translation and
Theatre". Constructing Cultures. Essays on Literary Translation ed. by S. BASSNETT & A.
LEFEVERE, 90-108. Clevedon: Multilingual Matters.

EBERT, R. 1995. "'Oleanna' film doesn't match play's intensity". Available on Internet:
http://www.kentuckyconnect.com/heraldleader/movies/mq/oleanna.html.

GIST, R. 1996. "Review of Oleanna by David Mamet". Available on Internet:

http://midget.towson. edu/~gist. oleanna.html.

KAJAVA, J. 1994. "Kun hiiri saa kissan hampaisiinsa". Helsingin Sanomat, 14 Feb. 1994.

KERNS, W. 1997. "Mamet's Oleanna challenges cast, prompts audience debate". Available
on Internet: http://www.lubbockonline.com/ news/ 061397/mamets.html
162 Ritva Leppihalme

KNOWLES, R.P. 1997. "Focus, Faithfulness, Shakespeare, and The Shrew: Directing as
Translation as Resistance". Essays in Theatre 16(1), 33-52.

KUKKONEN, P. 1993. Kielen silkki. Hiljaisuus ja rakkaus kielen ja kirjallisuuden

kuvastimessa. Helsinki: Helsinki University Press.

MAJAPURO, M. 1994. "Kamppailu kahdelle näyttelijälle". Hämeen Sanomat, 17 Feb. 1994.

MAMET, D. 1993a. Oleanna. Royal Court Writers Series. London: Methuen.

MAMET, D. 1993b. Oleanna. Trans. Michael Baran. Unpublished typescript. Finnish

National Theatre, Helsinki.

ÖSTMAN, J-O. 1982. "Pragmatic particles in an applied perspective". Neuphilologische

Mitteilungen 83, 135-153.

RUUSKANEN, A. 1994. "Missä kulkee raiskauksen raja?" Kaleva, 15 Feb. 1994.

SCHAFFNER, C. & KELLY-HOLMES, H. (eds). 1995. "Debate". Cultural Functions of

Translation, 32-54. Clevedon, England: Multilingual Matters.

TIRKKONEN, M-L. 1994. "Ilmiö nimeltä poliittinenkorrektius". Me Naiset, 25 March 1994.

VENUTI, L. 1995. "Translation and the Formation of Cultural Identities". SCHAFFNER, C.

& KELLY-HOLMES, H., 9-25.

YLILUOMA, R. 1994. "Oleanna mittailee valtapeliä ja seksuaalisen häirinnän rajoja".

Demari, 15 Feb. 1994.
TRANSLATION AND THE SHAPING OF MODERN
DEMOCRACY. CONCEPTIONS OF LIBERTY AND
AUTHORITY IN NINETEENTH-CENTURY ENGLISH
TRANSLATIONS OF PLATO'S PROTAGORAS
Alexandra Lianeri
University of Warwick, UK

Zusammenfassung
Der Artikel untersucht die Rolle der Übersetzung bei Formung der Konzeptionen von
Demokratie im 19. Jh. Es wird davon ausgegangen, daß die Ideen der Demokratie
weitgehend als Produkt von Übersetzung konstruiert waren. Sie zielte auf die
Einbürgerung der Transzendenz von historisch spezifischen Interpretationen durch
Argumente für deren "Treue" zu einem a priori definierten Original. Sowohl Rhetorik
als auch Praxis der Übersetzung dienten als Mittel zur semantischen Transformation
von Demokratie als eine inhärent ambige Konzeption, die gleichzeitig Glorifikation
von autonomem politischem Gedanken und Gehorsam gegen autoritativen Prinzipien
ausdrückte.

Résumé
Quel est le rôle de la traduction dans la mise en place des conceptions de la démocratie
en Angleterre au 19ème siècle? Selon notre recherche, la notion de démocratie est
largement le fruit de traductions qui ont tenté de dépasser des interprétations
historiquement marquées, en prétendant rester "fidèles" à un original défini à priori.
La pratique traductionnelle, y compris dans sa rhétorique, a donc servi de moyen pour
transformer le sens de démocratie comme concept intrinséquement ambigu - servant à
la fois à la glorification de la pensée politique per se et à la défense de principes
autoritaires.

Resumen
Este ensayo examina el papel de la traducción en la conceptualización de democracia
en Inglaterra durante el siglo XIX. Propone que las ideas democráticas constituyen en
gran medida un producto de la traducción que intentaba establecer la transcendencia
de interpretaciones históricamente especificas al alegar su "fidelidad" a un original
defmido a priori. Tanto la retórica como la práctica traductora sirvieron como medio
para la transformación semántica de la democracia como un concepto intrinsecamente
164 Alexandra Lianeri

ambiguo que expresaba simultáneamente una exaltación del pensamiento politico

autónomo y la adherencia a principios autoritarios

1. The semantic ambivalence of modern democracy

Democracy is an exemplary case of a concept which has been multiply

understood and transformed by conflicting interpretations. The construction of
the Athenian city as a self-instituting body whose political thought and practice
manifested a belief in the ability of the people to think, evaluate and determine
their social and political life, has not merely inspired ardent critics or
passionate advocates; it has also enabled the creation of such a notorious
semantic ambiguity that democracy has either come to be described as virtually
meaningless or as signifying all things to all people.
In what follows I intend to argue that the rhetoric of democracy, in
conjunction with the actual translations and rewritings of classical texts
(Lefevere 1992), served as a means for introducing the notion of authority into
the liberal political discourses of nineteenth-century England and contributed
to the legitimation of democracy as an internally divided concept, which
praised the potential of individuals for autonomous judgement while presenting
historically specific political values as existing beyond contingent criteria of
evaluation. I shall examine the implications of reflections on translation
articulated by thinkers such as Matthew Arnold and John Stuart Mill, as well as
the discursive ways in which liberal democratic thought was inscribed in
translations of the Platonic dialogue Protagoras. Two main hypotheses will be
suggested. The first is that any reference to conceptions of democracy as
transcendent is less innocent and more historically constructed than it might at
first appear to be, and tends to conceal the specific social and cultural processes
through which democracy was constituted as a product of "translation",
manipulation and change rather than a static and unalterable "original"
concept. This hypothesis is predominantly a historical one. The second is that
such recognition of the historicity of both "translation" and "democracy",
which implies the relativisation of our own definitions and interpretations of
these concepts, does not preclude either the potential of critical appraisal of
their cultural and social function or the employment of justifiable conceptual
standards of evaluation, which would enable us to recognise the moral and
political implications of specific translations of democracy against the
presupposed assumption of certain transhistorical human needs, abilities and
rights. This would be a philosophical and political hypothesis.
 The shaping of modern democracy 165

2. Translation faithfulness and the constitution of liberal democracy

2.1 The political meaning and uses of translation faithfulness

After the first decades of the 19th century, translation was explicitly described
by Victorian thinkers as a discourse which pointed outside its immediate
historical context, to a realm of knowledge that was perceived as detached
from the ideological conflicts and cultural contingencies of English society.
Conceived as a means of conveying intact the remoteness of the original
(Bassnett 1991:68), translation had "no pretension to any other merit" as John
Stuart Mill put it, "than that of fidelity" (Mill 1978: 42); a "merit" which not
only implied the assertion of textual transparency, but also located translation
beyond the boundaries of a transient cultural present, in a conceptual space
whose commensurability with this present was also the proof of its
participation in a universal, essentially ahistorical - rather than national or
culturally specific - Weltanschauung. In a typical articulation of this creed
Matthew Arnold, in his famous essay on the translation of Homer, drew a
parallel between the "union of the translator with his original" and the "union
of the human soul with the divine essence", arguing that such a union alone is
the presupposition of a good translation, which can take place "when the mist
which stands between them [the translator and the original] - the mist of alien
modes of thinking, speaking, and feeling on the translator's part - 'defecates to
a pure transparency', and disappears" (Arnold 1960:103). The religious
allusions of this metaphor do not merely set the presuppositions of "good"
translations. By describing faithfulness as the expression of unity between the
human and the divine, Arnold seeks further to set the criteria for the validation
of certain translations as direct realisations of the universal meaning inscribed
in canonical source texts, such as the Homeric epics to which he refers. The
claim to "transparency", which implies, as Lawrence Venuti has argued, the
self-concealment and illusory naturalisation of translation (Venuti 1995:1-42),
is simultaneously the key to the sublimation of the translated text, which
becomes capable of approaching the "truth" at the moment it is deprived of
authenticity and autonomy, and becomes subsumed to the authority of the
original.
Paradoxically, this conception of translation was employed in the 19th
century in order to dispute religious justifications of political truths and values,
challenge the God-given authority of kings and aristocracy, and substitute the
unquestionable supremacy of divine knowledge for the fallible sensibility of
political subjects. A rhetoric which strictly determined the boundary between
"faithfulness" and "betrayal", and defined transparency as the proof of the
translation's validity, was evoked in order to shape and justify the disparate
rewritings and significations of democracy, which later brought about the
notorious semantic ambiguity and contestability of the concept.
166 Alexandra Lianeri

Hence for liberal thinkers of the time, the category of "unfaithful"

translations included a long tradition of works up to the beginning of the
nineteenth century, which identified the Athenian democracy with a state of
anarchy and social instability, and interpreted the notion of political self-
institution as a total absence of moral, legal and political authority. Within this
tradition, a society which institutionalises its collective autonomy and
recognises itself as the only source of ethical and political norms was described
as "lacking" a vital distinction between a realm of political excellence, which
would be the source of political decisions, and the realm of the "common
people", whose action should be subsumed - freely or forcefully - to the
recognised authorities of a community.

2.2 Mill's aim

The refutation of this model of political thought, which had long been
established as authoritative and accurate, was actualised through the
"realisation" of the mistakes, limitations and often political prejudices of
previous translators. This is precisely the point J. S. Mill emphasised when he
decided to publish a series of translations of Platonic Dialogues, which began
with one of the most significant ancient political debates on democracy: that
between Socrates and the Sophist Protagoras, included in Plato's homonymous
Dialogue. For Mill, the truth of Plato's political thought was obscured by the
faults of conservative translations and the political biases of interpreters and
commentators who exaggerated and transformed aspects of the source text "for
the base purpose of discrediting free institutions and freedom of inquiry" (Mill
1978:42-43). In his attempt to oppose these interpretive models, Mill was all
too eager to stress the fidelity and transparency of his translation: "it is the
object of these papers" he asserts, "not to explain or criticise Plato but to allow
him to speak for himself' (Mill 1978:60). The rhetoric which highlights Mill's
"invisibility" was a premise of no passing significance. Fully acknowledged as
a dominant norm in the 19th century, the claim to the self-annihilation of the
translator in the face of canonical source texts was the necessary
presupposition for the recognition of the value of Mill's work and the
shortcomings of its predecessors.
Mill's conception of faithfulness evoked a double standard of
evaluation, on the basis of which he could reject all "negative" appraisals of
classical democracy as "mistranslations", while legitimising as accurate the
identification of Greek political thought with his contemporary liberal ideals.
His claim, which would result in a radical semantic transformation of the
concept of democracy, did not merely consist in the use of translation as a
means of political propaganda, which employed the same ideological weapon -
the notion of transparency - as its predecessors. Liberal thought did constitute
a historically "faithful" interpretation of classical democracy during the 19th
 The shaping of modern democracy 167

century, and it actually established - eventually - the values of liberty, equality

and human emancipation as "real" political ideals.
In this context, the notion of translation faithfulness functioned
metonymically at a level which was much deeper than the direct legitimation of
one translation practice over another. Adopted by political discourses which
aimed to exalt the value of individual judgement as well as the right to self-
determination, the allusion to the transcendent origin of translation introduced
the notion of a priori evaluative standards to which this judgement was
subjected. Mill's reference to faithfulness was not merely grounded on a selfish
intention to justify his translation and by extension his personal role in shaping
the political principles of his time. The assumption inscribed in his translation
rhetoric was the confirmation of the very existence of authoritative political
truths, to which we can appeal irrespective of diversified historical
interpretations of political concepts and values. While for Mill it is up to the
ability of enlightened individuals to discover and rewrite these truths, the
validity of potential interpretations is determined on the basis of their ability to
participate in immanent political judgements, precisely as in Arnold's belief
that a transparent translation participates in the divinity of a canonical original.
This idea, which establishes the value of Mill's translations as political texts on
the basis of their fidelity to "real" democracy, implies a submission of
individual autonomy to the impersonal power of a priori political truths, an
annulment of contingent historical judgement before the universal.
The reconciliation of liberty with authority implied in Mill's approach
to translation outlines an intrinsic tension in every attempt by Western
democratic thought to constitute a historically justifiable system of political
principles, and in the need to evoke some higher presuppositions and beliefs in
order to legitimise and validate this system. An analogous tension, expressed in
the Platonic dialogue Protagoras, is inherited in 19th-century translations of
the classical text. Yet the point at which these translations clearly depart from
Greek conceptions of democracy is when they seek to sustain a final resolution
of this tension, which identifies "real" authority with the excellence of
individual thinkers, a political élite clearly distinguished from the main body of
a social community. After that moment Western democracy acquires a
fundamental propensity towards authoritarian institutions and principles.

3. Conceptions of liberty and authority in translations of Plato's

Protagoras

3.1 A glorification of liberty

When John Stuart Mill translated Protagoras in 1834 he decisively - and to an

extent consciously - fought for the universal, natural rights of the human
168 Alexandra Lianeri

subject: the political principles which were for him actualised in the freedom
and equality of the Athenian democracy. This aim determined the translator's
position towards the two main figures in the dialogue, i.e. Socrates and
Protagoras, whose opinions on the nature of political juxtapose two political
philosophies. The first, expressed by the ancient Sophist, is a moderate support
of relativist thought, a determination of political virtue as primarily existential
rather than essential, in the sense that this virtue can only be defined by
reference to historical human knowledge and evaluation rather than some
absolute and immanent standards. For Protagoras truth can only be human, and
justice, politics and ethics have only man as their measure. The opposing
argument, developed by Socrates, echoes Plato's hostility to the insecure nature
of human, temporal truths and seeks to identify the indelible features of
political knowledge which stand beyond their specific conceptions and
transformations by different men and cultures.
Within a context of strife over the political rights and abilities of
individual citizens, the Protagorean idea that human knowledge should be
considered as the only legitimate measure of political action could justify, for
Mill, not only the acquisition of freedom and the potential for political
participation, but also the renunciation of unquestioned obedience to political
authority lying outside human standards for critique and debate. Perceived as
an early support of liberal democratic ideals, Protagoras's translation found a
discursive location in a broader ideological attempt to restitute a widespread
Victorian image of the Sophistic movement as mischievous and immoral while
(re)constructing its unity with nineteenth-century social struggles and political
debates on democratisation. This move took, in the translation, the form of a
search for the "true" meaning of the ancient notion of 'Sophist', which had
"misleadingly", as Mill argues, become "significative of quibbling and deceit"
(Mill 1978:43). Mill suggested the retranslation of the notion as "a teacher of
wisdom" and explicitly endorsed Protagoras's political insight and
contemporary relevance. For, as he argued in the introduction to his translation,

although Protagoras is confuted, and made to contradict himself again and

again [in the dialogue] [...] what he utters is by no means either absurd or
immoral, but, on the contrary, sound and useful good sense, forcibly
expressed, or, at the lowest, an able pleading in favour of the side he espouses
(Mill 1978:44).

Mill's liberalism found a definitive expression in Protagoras's

description of the Athenian political processes. His political hostility to a life
submitted to blind obedience and intellectual apathy in the face of authority,
and his commitment to the ideal of autonomous political thought, are expressed
in the meticulous translation of Protagoras's speech on the nature of "political
virtue" (320d-328d). The Sophist's argument initially defends the idea that the
 The shaping of modern democracy 169

main components of political virtue, justice and prudence can be expected from
all men reared in democratic communities and constitute an indispensable
presupposition of social coherence and reproduction. Based on the Athenian
example he states precisely this point:

Mill follows the passage relatively closely in his translation, changing only the
notion of political into social virtue:

The Athenians and others ... are ready, when the subject is social virtue, which
depends wholly upon justice and prudence, to listen to all advisers; because of
this virtue all should be partakers, or states cannot exist (Mill 1978:49, my
italics).

We will return to the significance of this change later. The idea expressed in
Protagoras's argument that all citizens should be partakers of "justice" and
"prudence", and therefore all are worthy advisors when it comes to public
issues, was the core principle of the Athenian democracy and provided strong
intellectual support to Mill's political creeds and democratic principles. This
idea, which appears recurrently in Protagoras's speech, is meticulously
rendered by the translator, in opposition to a significant part of Socrates's
thought which is either condensed or entirely omitted. In particular Mill
summarises a speech of several pages (332a-334c) into a mere description of it,
in which he informs his readers that "Socrates forces Protagoras ... to admit
that [prudence] is the same thing with wisdom, that is the same
thing with justice, or at least inseparable from it" (Mill 1978:54). This speech
constitutes a significant part of Socrates's argument that all components of
political virtue are ultimately reducible to "knowledge" and wisdom, in the
sense that knowledge of the "good" and the "evil" is the only presupposition
and legitimisation of human virtue. Mill was not initially negative to the
privileged position knowledge occupied in Socrates's thought. What he
rejected in his translation was the totalising conception of knowledge,
expressed in the source text, as a system of absolute moral and political truths,
and the implied devaluation of critical political judgement and endorsement of
political heteronomy and submission. Against this idea, which dissociated
political virtue from shared social experience and knowledge, Mill juxtaposed
Protagoras's glorification of human political capacity, the belief expressed in
his translation that "it is thus with virtue: all men can teach it" (Mill 1978: 51)
170 Alexandra Lianeri

3.2 An authoritative turn

Despite his faith in human political potential Mill was not willing to endorse a
total relativisation of political principles, which could easily follow the
recognition of equal validity in diversified conceptions of morality and politics.
His attempt to demonstrate the possibility of a self-instituted, rational
democracy, and to sustain a secularised model of political thought, was
versatile and complex enough to accommodate the notion of authority,
however differentiated this notion was from Socratic idealism. Following
Protagoras' thought he sough to suggest that no system of morality, and indeed
no democratic society, could be maintained when devoid of some shared
principles and creeds capable of justifying the very necessity of political
justice, equality, and the self-conscious, responsible conduct of human life.
Hence while Protagoras's speech ostensibly highlights the significance
of personal political opinion, it also points out the existence of universal
principles and historical laws as an indispensable feature of the Athenian
democracy. Justice, prudence, and respect, which are for him the predominant
presuppositions of political virtue, are neither disputable nor historically
relative. As he claims in an important deviation from rational argumentation
which employs the Greek mythical tradition, these values were given by Zeus
to all people when societies were initially formed, and it was Zeus who ordered
men that whoever fails to partake of these values should be killed, as a public
pest (322c-d). Articulated in a time in which myths had certainly lost their
literal value, Protagoras's narrative had more a symbolic than a truly religious
meaning and function. It served to express a system of historically stable
standards of Athenian democratic thought, on the grounds of which individuals
could be educated to become politically capable and responsible citizens. As
Protagoras quickly observes after narrating the myth, the Athenians do not
consider Zeus's gifts as natural and spontaneous, but as the result of careful
education. It is precisely this education, which takes place in the context of
democratic institutions and is guarded collectively by the community, which
creates citizens able to articulate and implement political judgement; as well as
to subsume their selfish interests, when necessary, to the common good. In the
Protagorean conception of the Athenian democracy, political autonomy and
liberty are the indispensable rights of citizens rather than of individuals, and
authority lies with those principles which have been capable of nourishing the
moral and political responsibility of those citizens among themselves as well as
towards the entire city (cf. 323d-328d).

3.3 Mill's translation

On the treatment of this issue Mill's translation departs from the source text at
two main points: it replaces the notion of the citizen by the notion of the
 The shaping of modern democracy 171

individual, and it consequently locates authority not in the collective

responsibility of the city but in the enhanced ability of meritorious individuals
to understand and articulate real political truths and values. Let us examine the
translation of the following passage, in which Protagoras asserts that men who
have been brought up within laws and societies able to cultivate political virtue
are partakers and potentially "creators" of justice:

Mill translated the passage as follows:

All civilised men, even the most unjust, if compared with men among whom
there is no training, no tribunals, no laws, with the wild men [of whom poets
tell us,] would appear a perfect master in virtue (Mill 1978:51).

The translation introduces two significant transformations into the source text.
Mill uses the notion of "civilised men" in order to render the Protagorean
description of men reared within laws and society, and he further substitutes
the idea that these men are educated to be themselves just, as well as "creators"
of justice, in the phrase "perfect master in virtue". The ideal of political
creation expressed in the source text constituted a notion of man not as
individual but qua citizen, whose education derives precisely from his
participation in the political processes of the community and who is able to
express - as part of a social body - moral and political judgements. The
necessity of active political participation is absent from Mill's thought. For
him, as for the majority of 19th-century liberal thinkers, democratic
government requires the agreement rather the participation of the governed in
politics. Democracy itself should be based, Mill argued, not on the often
mistaken and vulgar judgement of the mass of the people, but on the
consensual establishment of an educated and progressive élite which is capable
of political creation: "The idea of a rational democracy is not that the people
themselves govern, but that they have security for good government [...] the
best government (need it be said?) must be the government of the wisest, and
these must always be a few" (quoted in Burns 1969:294). Not surprisingly, the
translation excludes the people from this potentiality of political judgement by
emphasising their "mastery in virtue" rather than their capacity for political
creation acquired in a democratic society. Precisely as in the first extract from
the translation examined in this essay, the virtue of the majority is presented to
be "social" rather than "political" in nature. This substitution accords with
Mill's life-long dedication to a fully extended participatory democracy, which
172 Alexandra Lianeri

does not nevertheless recognise the responsibility of all citizens to express and
implement a collective ethics and politics, that is, the responsibility of a self-
instituting society expressed in the source text.
What is worth some further attention here is the notion of the "civilised
man", introduced in the translation as the subject of social virtue. While it is
beyond the scope of this essay to examine the full range of meanings which
constituted the concept of "civilisation" in modern civil societies, it is
important to note at least one aspect of these meanings, which becomes evident
in Mill's approach to the Platonic dialogue: the identification of civil virtue
with the pursuit of personal interests and the wilful legitimisation of egotistic
utilitarianism and hedonism. Mill's conception of the "principle of utility" -
initially articulated by his father, James Mill, and Jeremy Bentham - as the
quest for individual pleasure and fulfilment of personal desires, is expressed in
the translation of the following Socratic statement:

... pleasure is the same thing with good, and pain with evil: and if a pleasure is
bad, then it is because it prevents a greater pleasure, or causes a pain which
exceeds the pleasure: if a pain is good, it is because it prevents a greater pain,
or leads to a greater pleasure (1978:58).

The translation of the passage stays fairly close to the source text. Yet the idea
expressed by Socrates at this point is part of a broader argument developed in
the dialogue, which ultimately seeks to refute such a relativistic conception of
virtue and maintain that political qualities are intrinsically "good", as
expressions of an a priori political truth, rather than judged as good by
temporal standards of pleasure. It is only by omitting a significant part of this
argument, as mentioned above, that Mill is able to declare, when commenting
on Socrates's thought at the end of his translation, that "the principle of utility,
- the doctrine that all things are good or evil, by virtue solely of the pleasure or
the pain which they produce, - is as broadly stated, and as emphatically
maintained against Protagoras by Socrates, as it ever was by Epicurus or
Bentham" (Mill 1978:61). The doctrine of utilitarianism, which Mill reads in
the Platonic text, illustrates the value of individuality and selfhood, and
advocates the indisputable democratic right of freedom to fulfil one's personal
goals and desires. Yet the conception of people as individuals who are devoted
to the maximisation of their own satisfaction involves an implicit deification of
this satisfaction, a fatalistic submission to self-centred impulses and
inclinations which are presented as the standard for the judgement of one's life
and conduct. One cannot help noting that this standard of civil societies, which
 The shaping of modern democracy 173

essentially disqualified the distinction between "morality" and "immorality"

and subsumed political responsibility to private interests, is not essentially
distanced from the moral indifference of totalitarian regimes to the fate of their
human victims, or from the contemporary devotion to mass consumption and
profit, which disclaims all responsibility for human suffering and destitution.

3.4 Jowett 's translation

The liberal glorification of individuality was unavoidably inclined to adopt an

authoritarian political viewpoint when it came to the resolution of conflicting
interests within a community. Having sustained a political system which
created economic and social inequalities, liberal democracy had to reconcile
the oppositional claims of diversified and hierarchically stratified groups, each
of which sought either the transformation or the perpetuation of existing social
divisions. Without recourse to an ethics of political responsibility, such
reconciliation was sought in the reconstitution of a moral vocabulary which
aimed to naturalise hierarchical social structures and legitimise the authority of
dominant groups as indispensable to the maintenance of social coherence and
reproduction. A belief in the "naturalness" of human inequality - already
prevalent in Mill's thought - was furthered after the middle of the nineteenth
century and was explicitly endorsed in Benjamin Jowett's significant
publication of translations of Plato in 1871.
Jowett sought "to represent Plato as the father of Idealism, who is not to
be measured by the standard of utilitarianism or any other modern
philosophical system", but could provide an alternative to individualistic ethics
and lack of authority, which was for him a menace to social order and
coherence (Jowett 1892: xi). His interpretation of Protagoras established an
irreconcilable division between the two interlocutors; it represented the one as
the advocate of a common and contingent knowledge, and the other as the
seeker of eternal ideals, and Jowett expressed his clear preference for the latter.
"The truth of Protagoras", Jowett argued, "is based on common sense and
common maxims of morality, while that of Socrates is paradoxical or
transcendental, and though full of meaning and insight, hardly intelligible to
the rest of mankind" (Jowett 1892: 122).
The translator advocated through his work his firm belief in the
existence of absolute standards of political morality which were not to be found
among the vulgar opinion of the common people, but were established and
justified by the intellectual and political élites of a society. This belief is
articulated in the translation of a description of the Athenian democracy by
Protagoras, in which the Sophist points out that the Athenian city teaches its
young members its laws and constrains them both to govern and be governed
according to them:
174 Alexandra Lianeri

Jowett translated the passage as follows:

the state compels them [the children] to learn the laws ... [and] the city outlines
the laws ... and compels us to exercise and obey authority in accordance with
those ... and he who transgresses [the laws] is to be corrected, or, in other
words called into account, which is a term used ... seeing that justice calls men
to account (Jowett, 1875:151, my italics).

The translation introduces some substantial changes to the source text. By

replacing the notion of the city by the notion of the state, it constitutes a
distinction between the state, as a locus of compelling power, and the main
body of a social community, a distinction which was absent from the Athenian
democracy. Consequently Jowett transformed the description of the Athenian
citizens as capable of governing and being governed according to laws, by
presenting the political institutions of the city as a process of exercising and
obeying authority. Finally, in the translation it is not the city, i.e. the social
community, which is both the source and the guardian of its laws, as is
articulated in the source text. What it is implied through the use of the passive
voice and the replacement of the polis - which is the subject of punishment in
the original - by the notion of "justice", is the location of legal authority in an
a priori defined conception of justice, whose binding power exists irrespective
of the political thought and action of the city itself. The people cannot have an
active role in this process precisely as they cannot have a creative one in moral
and political life. At best the common man, described, as we saw, by
Protagoras as a creator of justice, is for Jowett an "artificer" of it (1875:152-3),
necessarily subsumed to moral and political principles which transcend his or
her intellectual potentiality.

4. Conclusion

The reference to transcendent political ideals coexisted in English conceptions

of democracy with the exaltation of the human potential to judge and freely
contest the validity of political systems and processes. In conclusion, I will
briefly evaluate the relation of this reading to 19th-century translations, and
acknowledge both its indebtedness to and its distance from conceptions of
democracy articulated by our political predecessors.
 The shaping of modern democracy 175

The contemporary potential to recognise liberty and equality as the

fundamental rights of a people is as much the product of interpretations of
democracy as it was for 19th-century translators. Each of these historical
interpretations - or, indeed, translations of the concept - should not be seen, as
José Lambert and Clem Robyns have argued, as the final component of a static
dichotomy between the source and target text; rather, each constitutes a sign in
itself, subjected to other interpretations, whose formation as part of discursive
systems is related to the interaction of different codes and normative models in
the target community (Lambert & Robyns, forthcoming). Articulated in a
conceptual framework which has been predominantly constructed by liberal
ideas, our interpretation of the Athenian democratic values was fundamentally
enabled by those past translations which have already established the validity
of democratic principles, in opposition to the negative appraisals of democracy
expressed before the 19th century. So was our potential to describe and
criticise the adoption of an authoritarian perspective by English translators,
because this potential presupposed and actualised the empowerment of human
reason and judgement which took place during the last few centuries. In other
words, our reading of Greek democracy as well as its interpretive
transformations stemmed from a conceptual context within which democracy
had already been shaped as the product of a series of translational processes.
This suggestion does not imply the negation of a historically concrete
and relatively stable viewpoint from which we can attempt to understand and
evaluate previous translations of democracy. What was recognised as a chain
of mutually related interpretations of a historically specific sign, i.e.
democracy, does not constitute an endless semiotic process, but is, for
pragmatic reasons, halted at a certain stage with the establishment of a "final
logical interpretant", which either corresponds to a conventional cultural unit
or establishes a partially new one (Lambert & Robyns, forthcoming). The need
for the self-reflective constitution of a "final interpretant" on the basis of which
we can describe translations of democracy and appraise their political
implications is both of an epistemological and moral nature. From an
epistemological perspective it would enable us to indicate the disparate
meanings democracy acquired throughout history, and to highlight cultural
transformations and manipulations of the concept in relation to what we
interpret as its initial meaning and value. From a moral perspective, it would
sustain a critical vocabulary, which would consciously maintain the tension -
inherited from the whole history of democratic thought - between the
conviction that political truths and practices have to be shaped as human, and
therefore contingent, creeds and morals and the need to evoke some more
stable standards of appraisal, capable of verifying this conviction itself as well
as the specific collective choices and laws of every "democratic" community; a
tension which seems to be at once the origin and the predicament of
democracy.
176 Alexandra Lianeri

References
ARNOLD, Matthew. 1960. "On Translating Homer". On The Classical Tradition ed. by R.H.
SUPER. Ann Arbor: The University of Michigan Press.

BASSNETT-MC GUIRE, Susan. [1980] 1991. Translation Studies, revised edition, London &
New York: Routledge.

BURNS, J. H., 1969. "J. S. Mill and Democracy 1829-1861". Mill A Collection of critical
Essays, 280-328, ed. by J. B. SCHNEEWIND. London: Macmillan. First published in Political
Studies 5,1957.

ELYOT, Thomas. [1531] 1883. The Boke Named The Governour, edited by Henry HERBERT.
& Stephen CROFT, 2 vols. London: Kegan Paul.

JOWETT, Benjamin. 1875. The Dialogues of Plato. Translated into English with Analyses and
Introductions. Vol. 1. Humphrey Milford: Oxford University Press.

JOWETT, Benjamin. 1892. The Dialogues of Plato. Translated into English with Analyses and
Introductions, 5 vols. Humphrey Milford. Oxford University Press.

LAMBERT, José and ROBYNS Clem. Forthcoming. "Translation" Semiotics. A Handbook on

the Sign-Theoretic Foundations of Nation and Culture ed. by Roland POSNER, Klaus
ROBERING, and Thomas A. SEBEOK (Handbooks for Linguistics and Communication
Science). Berlin and New York: de Gruyter.

LEFEVERE, André, 1992. Translation, Rewriting and the Manipulation of Literary Fame,
Routledge: London and New York.

MILL, John Stuart. [1834] 1978. Notes on Some of the More Popular Dialogues of Plato: The
Protagoras. Essays on the Philosophy and the Classics (Collected Works of John Stuart Mill.
Vol. xi) ed. by J. M. ROBSON. University of Toronto Press

PLATO. 1927. Protagoras, edited by G P. GOOLD, with an English translation by W. R. M.

Lamb. Cambridge Massachusetts & London: Harvard University Press & William Heinemann
Ltd.

QUENNELL, Peter. (ed.) 1950. Byron. A Self Portrait. Letters and Diaries 1789 to 1824.
London: John Murray.

VENUTI, Lawrence. 1995. The Translator's Invisibility. A History of Translation. London &
New York: Routledge.

WILLIAMS, Raymond. [1976] 1983. Keywords: a Vocabulary of Culture and Society. Revised
edition. London: Fontana Press.
TRANSLATING LINGUISTIC MARKERS OF
IDEOLOGY

Tiina Puurtinen
University ofJoensuu, Savonlinna School of Translation Studies,
Finland

Zusammenfassung
Dieser Artikel ist ein Versuch, Ideen und Ergebnisse der kritischen Diskurs analyse auf das
Übersetzen anzuwenden. Das Ziel ist, einige Sprachstrukturen - und zwar Passivisierung,
Nominalisierung und premodifizierte Attributkonstruktionen - zu besprechen, die Ideologie
in englischen und/oder in finnischen Texten reflektieren können, sowie auch die
ideologischen Folgen von bestimmten sprachlichen Entscheidungen des Übersetzers zu
illustrieren..

Résumé
Nous chercherons ici à appliquer à la traduction certaines idées et certains acquis de l'analyse
de discours (AD) critique. Notre objectif est double: d'une part, discuter de quelques
structures linguistiques, à savoir la passivisation, la nominalisation et les constructions
participiales attributs prémodifiées, qui peuvent refléter une idéologie dans les textes anglais
et/ou finnois; d'autre part, illustrer les conséquences idéologiques de certains choix
linguistiques opérés par le traducteur.

Resumen
Este trabajo aplica a la traducción los conceptos y resultados presentados por analistas del
discurso criticos. El objetivo es el estudio de algunas estructuras lingüisticas, en particular
la construcción de la pasiva, la nominalización y las construcciones atributivas de participio
que pueden reflejar una ideologia en los textos en inglés y en finlandés. También se propone
ilustrar las consecuencias ideológicas de determinadas decisiones lingüisticas del traductor.
178 Tiina Puurtinen

1. Introduction

Use of language always reflects the user's attitudes, beliefs and viewpoints more
or less inconspicuously. Language use also embodies values and power relations.
Ways in which such ideological meanings are encoded in linguistic expression have
for the past two decades been the subject of critical discourse analysis (CDA; see
e.g. Fairclough 1990, 1995). CDA aims at revealing how ideology affects linguistic
choices made by a text producer and how language can be used to maintain,
reinforce or challenge ideologies. The concept of ideology itself is ambiguous and
has been variously defined (for criticism of the definitions, see Solin 1995,
Thompson 1984); in this paper, following Simpson (1993:5), ideology refers to
taken-for-granted assumptions, beliefs and value systems shared collectively by
social groups. Every text has an ideology, because it must be based on some values
and ways of thinking. The embedding of ideology in a text is not necessarily
intentional; neither the writer nor the reader need to be conscious of it, especially
if the values and beliefs are those that we have internalised as members of a
particular society. Reflection of such unconscious ideologies in a text can,
however, be particularly effective, because it reinforces generally accepted notions
and attitudes (or so called "common sense", see Fairclough 1990:77-108, Fowler
1987:490) instead of challenging them.
Embedded ideological meanings are interesting for Translation Studies,
firstly because ideologies of societies and cultures are different, and secondly
because the lexico-grammatical realisation of ideology is likely to vary in different
languages. However, very few CDA-oriented studies of translation have been
carried out so far (Hatim & Mason 1997, Knowles & Malmkjær 1989). This paper
discusses some linguistic structures (passivisation, nominalisation, premodified
participial constructions) which can reflect ideology in English and/or Finnish texts,
and illustrates the ideological consequences (often unintentional) of certain
linguistic choices made in translation.

2. Operation of ideology in texts

Ideological content in a text can be explicit or implicit. Explicitly expressed

ideological meaning is naturally intentional. Good examples of such texts are
argumentative texts, such as newspaper editorials, which include clear statements
of the author's opinions, and explicitly didactic children's books and fairytales,
whose underlying moral and ethical principles, for instance, are verbalised within
the narrative itself or in extra-narrative statements (such as prefaces). Implicit
ideology, on the other hand, is often unnoticed and even unintended; it consists of
the writer's possibly subconscious assumptions and generally accepted values
which underlie the writer's linguistic choices. Linguistic expression can thus reflect,
 Markers of ideology 179

and thereby reinforce, values which are unquestioningly accepted by a society, and
by the writer as a member of that society. Of course a text producer can also
deliberately challenge such generally held beliefs by making the text implicitly (or
explicitly) reflect an opposing ideology. For example an author of a children's
detective story might portray female characters as more active than male ones, and
thus more visible linguistically. However, the reverse is likely to be more common
in children's literature. (See Knowles & Malmkjær 1996 for related examples; and
also Hollindale 1988, summarised in Knowles & Malmkjær 1996:65-68.)
Thompson (1990:59-67, summarised in Knowles & Malmkjær 1996: 46-
59), describes five modes of operation of ideology: legitimation refers to the
representation of power relations as legitimate; in dissimulation power relations
are hidden, denied or obscured; through unification diverse social groups are
brought together e.g. by using such symbols of unity as national flags and anthems,
whereas fragmentation is a way of differentiating between groups; through
reification power relations are represented as if they were natural and permanent.
Operation of ideology is served by a number of strategies, which can be realised
in text by micro-linguistic structures (involving single words or phrases) or macro-
linguistic structures (involving a clause or more). The strategies can either operate
implicitly in a text, or they can rely on explicit verbalisations. In the following, I
will concentrate on reification, and three associated strategies, passivisation,
nominalisation, and premodified (participial) constructions (the last one is my
addition to Thompson's list). The reason for focusing on these particular strategies
is that, first, they are realised by specific linguistic forms, which can be easily
identified in a text, and second, both their internal structure and their usage in
various text types is different in English and Finnish. It must be emphasised,
however, that none of the linguistic forms discussed in the following is always an
indicator of ideology in a text and should not automatically be interpreted as such.
CDA-based research of translated and non-translated texts must be supported by
a good knowledge of the values held by the surrounding culture.

3. Reification and related strategies

According to Thompson (1990:65-66), reification is a mode of operation of

ideology through which power relations which are transitory states are represented
as if they were timeless, natural and permanent. This definition could be extended
to include also values, beliefs and customs considered self-evident and legitimate
in a culture or society. In Fowler (1991:80) the term reification seems to have a
more limited scope: it refers to the conceptualisation of a process or quality as a
thing, as a fixed, inanimate entity. Related terms used by other scholars are
objectification (Fowler & Kress 1979) and mystification (Fowler 1991:80). Here,
objectification and mystification are regarded as subordinate processes subsumed
180 Tiina Puurtinen

under the more extensive concept of reification. Reification is supported in texts

by, among other things, passivisation, nominahsation, and at least certain kinds of
premodified constructions.

3.1 Passivisation

Passivisation is a syntactic transformation which can have a strong effect on the

way we perceive events and actions. It is part of the transitivity system of language
(Halliday 1986), which represents processes, participants and circumstances.
Passive sentences can be an effective neutralising (or mystifying) means of
representing actions, processes and participants in such a way that causal relations
and responsibility for actions become obscure, as participants can be deleted from
the sentence. The object of the action, and not the agent or actor as in active
sentences, usually occupies the first position in the clause. Thus passivisation
makes it possible to thematise the object and leave the agent responsible for the
action completely unmentioned. Transitivity analysis has been very fruitful in CDA
studies (see e.g. Tony Trew 1979).
Example 1 shows how the tone of an argumentative text may change if an
English passive (underlined) is translated with a different form and modality into
Finnish. It is from a student translation of an article about the rainforest fires in
Indonesia, published in The Ecologist (Manser-Fonds 1998).

(la) Mitsubishi and MarubenifromJapan, HIAG from Switzerland, Pirelli, Nestlé,

among others, are called upon to take responsibility for what has happened in the
name of 'development', as are consumers called upon to buy local resources instead
of the above-mentioned products. (Manser-Fonds 1998)

Japanilaisten Mitsubishin ja Marubenin, sveitsiläisen HIAG:in, sekä Pirellin,

Nestlén ym. on kannettava vastuu siitä, mitä on tapahtunut 'kehityksen' nimissä, ja
samoin kuluttajien on ostettava paikallisia tuotteita edellä mainittujen sijaan.
(word-for-word transi: ...Nestlé among others must bear the responsibility for what
has happened in the name of 'development', and similarly consumers must buy local
products.)

The source text's passive forms seem to imply that in addition to the writer, there
are other people and groups (The Ecologist magazine, environmental groups and
organisations) behind the demand, although none of them is explicitly mentioned.
The Finnish translation, on the other hand, does not give the same impression but
is likely to be interpreted simply as the writer's personal opinion that consumers
and big companies have a moral obligation to act in the recommended manner. (Of
course, this may be just my way of reading between the lines, and other
interpretations are possible.) An alternative translation, which retains the
 Markers of ideology 181

implication of an unnamed group or groups appealing to companies and

consumers, might read as follows:

(lb)...sekä Pirelliä, Nestléä ym. on kehotettu/ on vaadittu kehotetaan / vaaditaan

kantamaan vastuu...ja samoin kuluttajia on kehotettu / on vaadittu / kehotetaan
/vaaditaan ostamaan paikallisia tuotteita...
(...Pirelli, Nestlé among others have been urged to/are urged to bear the
responsibility... and similarly consumers have been urged to/are urged to buy local
products...)

Example 2, which is from a student translation of an article about new technologies

published in The Futurist (Mayur & Daviss 1998), illustrates the use of the Finnish
passive (impersonal) voice and an equivalent agentless infinitive structure instead
of active verb forms; the sentence also includes replacement of the transitive verbs
draining and concentrating by the Finnish intransitive verbs eivät autioidu and
kerääntyessä, which implies that (without new technologies) rural areas tend to
become desolate spontaneously, and workers move to towns, instead of the
country (government, industry etc.) draining people from rural regions and
concentrating them in towns and ghettos, as in the source text.

(2) Thanks to these new technologies, it's now possible for countries to develop
sound, broad-based economies without industrializing, without draining people from
rural areas and concentrating low-paid workers in company towns or urban ghettos,
and without degrading and exhausting their land, air, and water. (Mayur & Daviss
1998; 46-47)

Uusien teknologioiden ansiosta nykyisin on mahdollista kehittää eheitä

laajapohjaisia talousmalleja siten, että maaseutualueet eivät autioidu
matalapalkkaisten työläisten kerääntyessä tehdaskaupunkeihin tai
kaupunkigettoihin, ja siten, että kehitysmaiden maaperää, ilmakehää ja vesivaroja
ei kuluteta loppuun.
(Thanks to new technologies now it is possible to develop sound broad-based
economic models so that rural areas do not become desolate with low-paid workers
gathering in industrial towns and urban ghettos, and so that the land, atmosphere
and water resources of developing countries are not exhausted.)

The use of the passive vs. active voice and the related issue of role
assignment can also be a method of supporting characterisation in fiction. Knowles
& Malmkjær (1996:79-80) show that in The Secret Garden by F. Hodgson
Burnett, role assignment correlates with the increased activity of one of the main
characters. At the beginning of the story Mary Lennox is less active and less often
in the role of actor, i.e. things happen or are done to her rather than by her, while
towards the end she becomes more active and tends to occupy the actor role in
182 Tiina Puurtinen

sentences. In Hatim & Mason's (1997:153-158) example of what they call

"maximal translator mediation", a Spanish source text on the history of the
Mexicans and its English translation (which appeared in the Spanish- and English-
language editions of the UNESCO Courier) are shown to differ significantly in
terms of transitivity: the source text sentences tend to have inanimate actors as
theme, whereas the target text has human actors in theme position and involved in
action processes. Although one might think that in this way the indigenous peoples
of Mexico are placed in an active role, actively moulding their own destiny like in
the source text, in the target text they are in fact engaged in so-called
"supervention processes" {to exhibit, to meet with, to be engaged in), which
involve no active participation but the processes just seem to happen by
themselves. Although passive verb forms are not used, the impression created by
the translation is that of a passive people observing historical events.

3.2 Nominalisation

Like passivisation, nominalisation can also be used to obscure agency,

responsibility and causality (see Fairclough 1990:120-125, Fowler 1991:79-80),
as participants can be deleted or given a peripheral syntactic status. An activity
which is carried out by one person or group to another is transformed into an event
or state, which has no named participants. Nominalisation is a grammatical
metaphor (Halliday 1986, Ravelli 1988; see also Puurtinen 1993 or 1995:96-103
for grammatical metaphors in translated children's fiction), a marked, incongruent
form of encoding, whereby an action or a process is denoted by a nominal
construction instead of a verb, which can be regarded as the unmarked form.
(Similarly qualities, whose congruent realisations are adjectives, can be
metaphorically expressed with nouns, and clausal relations, normally realised by
conjunctions, can be expressed by nonfinite verb forms, for instance.)
Nominalisation transforms a process into a state or an entity and presents an
activity as something static and unalterable, because it is not under the control of
a named agent. Nominalisation makes the depicted process seem abstract and
remote, whereas the congruent realisation by a verb appears to be more directly
related to the concrete event. The use of grammatical metaphors is likely to
correlate with both interpersonal distance (distance between text producer and
receiver) and experiential distance (distance between the text and social reality; e.g.
language used in casual conversation vs. language used in a scientific paper on
casual conversation): in written texts both interpersonal and experiential distance
are greater than in spoken texts, and grammatical metaphors such as
nominalisations are likely to be more frequent (Eggins et al. 1993). Grammatical
metaphors are typical of, and well suited for, certain text types and target groups.
In example 3, from a student translation of a Newsweek article about forest
management (Gupte 1997), the source text infinitive structure {calls...to control)
 Markers of ideology 183

has been turned into the nominal form pyynnöt kontrolloimisesta ('requests for
controlling'), which obscures agency; who should control the clear-cutting and
logging practices: industrialized or developing countries?

(3) Developing countries, which control 60 percent of the world's forests, have
resisted calls from the industrialized nations to control their clear-cutting and
logging practices. (Gupte 1997)

Kehitysmaat, joiden omistuksessa on 60 prosenttia koko maailman metsistä, ovat

torjuneet teollisuusvaltioiden pyynnöt avohakkuiden ja puunhakkuun harjoittamisen
kontrolloimisesta.

An infinitive structure equivalent to the source text formulation, perhaps combined

with a stronger noun (kehotukset or even vaatimukset kontrolloida, 'calls to
control') would make the meaning clear.

3.3 Premodified participial constructions

Premodified nominal phrases purport to tell the readers what they already know;
the information in modifiers is taken as known, and thus there is no need to
foreground it in a way which invites the reader to object to it (cf. e.g. President
Clinton's outrageous behaviour angered the Americans vs. President Clinton
behaved outrageously and that angered the Americans - the first, premodified
construction implies that Clinton's behaviour is generally considered outrageous,
whereas the second version leaves this view open to discussion and disagreement).
Premodified constructions can be seen as a method of classification (see Fowler &
Kress 1979): premodified nominal phrases seem to be lexical items, unitary
elements, which cannot be split up. Finnish premodified participial attribute
constructions are similar to nominalisations in that they can turn a process into a
mere modifier by embedding the verb as a participial attribute. In the following
Finnish translation of a sentence in Rudyard Kipling's The Jungle Book (1987) the
verb nukkua ('to sleep') appears as a premodifier to those who sleep, and seems
to classify gipsies and woodcutters as a group of people who always tend to or
have to sleep outside. The premodified construction is like a lexical unit, a natural,
fixed entity. The source text's postmodification creates no such impression but
seems to be just an explanatory addition referring to incidental behaviour. (See also
Viertiö 1998 on various functions and interpretations of participial attribute
constructions in sports news in the Finnish press.)

(4) It was the noise that bewilders woodcutters and gipsies sleeping in the open, and
makes them run sometimes into the very mouth of the tiger. (Kipling 1987:11)
184 Tiina Puurtinen

Se ääni juuri eksyttää ulkosalla nukkuvat puunhakkaajat ja mustalaiset ja saa heidät

joskus juoksemaan suoraan tiikerin kitaan. (Pennanen& Jaskari 1965:5)

Finnish participial constructions are a neat way of conveying a lot of information

in a compact package, and therefore they can be considered a very useful form of
expression well suited for several types of texts. In English-Finnish translations
they are often used to replace English relative clauses, probably because an
equivalent Finnish relative clause appears too long and clumsy or disrupts the
smooth rhythm of the text. Occasionally, however, this translation strategy may
result in unintended ideological discrepancies between the source text and the
target text, as in the above example.

4. Conclusion

The above discussion is intended as a brief illustration of the potential implications

and hidden ideological meanings of just a few linguistic structures, which are not
necessarily always used to manipulate the reader on purpose. Nevertheless,
intentional manipulation by linguistic means is possible and has been shown to
happen by critical discourse analysts. Passivisation and nominalisation, among
other things, have been termed "masking devices" by Ng and Branac (1993),
because they present true information in an incomplete way, so that the reader may
get a distorted picture of the depicted event. Moreover, when propositions are not
made explicit but must be inferred from the text, it is difficult to identify and to
reject them (see also Fairclough 1990:140-168 on presuppositions). This may lead
to uncritical, unconscious acceptance of the content of a text.
It is important to bear in mind that a particular linguistic form should not
automatically be interpreted as an expression of a particular ideological meaning.
The use of passivisation and nominalisation, for instance, may be due to a text type
which happens to be characterised by the frequent occurrence of these structures.
Passivisation and nominalisation are also useful tools contributing to the
information structure of a text: they make it possible to thematise something which
is not the actor in the sentence. In other words, it is not the case that all choices,
either by a writer or a translator of a text, are ideologically motivated, nor do
isolated occurrences of certain linguistic constructions form a solid enough basis
for ideological interpretation. Only when a range of linguistic features seem to form
some kind of pattern (as in Trew 1979), can we attempt to uncover an underlying
ideological motivation on the part of a writer or a translator.
In translation such a pattern may also come about unintentionally (which
is probably the case with the UNESCO Courier translation analysed by Hatim and
Mason, mentioned in section 3.1. above). The frequent rendering of English
relative clauses by Finnish participial attribute constructions because of their
 Markers of ideology 185

brevity, or the translation of English passive sentences by Finnish active verb

forms, may result in unintended ideological colouring, and consequently the source
and target text readers may get a completely different perspective on the events
and processes portrayed. It is crucial for translators and trainees to be aware of the
potential ideological significance of linguistic forms, not only in literary texts in
which the role of linguistic features is emphasised, but also in other kinds of texts,
where priority is given to content over form.

Sources of examples

BURNETT, F. Hodgson. 1992/1911. The Secret Garden. London: Sainsbury Walker.

GUPTE, Pranay. 1997. "Anfashionable Forests". Newsweek 6 October, 19.

KIPLING, Rudyard. [1894] 1987. The Jungle Book. New York: Penguin. Transl. into Finnish by Eila
Pennanen & Juhani Jaskari 1965. Viidakkokirjat. Porvoo: WSOY.

KLEINER, Kurt. 1995. "Language Deaths 'Bad for Us A l l " . New Scientist 4 March, 15.

MANSER-FONDS, Bruno. 1998. "Why Are the Forests Burning?". The Ecologist 28(1), 8.

MAYUR, Rashmi & Bennett DAVISS. 1998. "The Technology of Hope". The Futurist, October,
46-51.

References

EGGINS, Suzanne, Peter WIGNELL & J.R. MARTIN. 1993. "The Discourse of History: Distancing
the Recoverable Past". Register Analysis. Theory and Practice ed. by Mohsen GHADESSY, 75-
109. London & New York: Pinter.

FAIRCLOUGH, Norman. [1989] 1990. Language and Power. London & New York: Routledge.

FAIRCLOUGH, Norman. 1995. Critical Discourse Analysis: The Critical Study of Language.
London & New York: Longman.

FOWLER, Roger. 1987. "Notes on Critical Linguistics". Essays in Honour of Michael Halliday ed.
by T. THREADGOLD & R. STEELE, 481-492. Amsterdam & Philadelphia: John Benjamins.

FOWLER, Roger. 1991. Language in the News. Discourse and Ldeology in the Press. London &
New York: Routledge.

FOWLER, Roger & Gunther KRESS. 1979. "Critical Linguistics". Language and Control ed. by
Roger FOWLER, Bob HODGE, Gunther KRESS & Tony TREW, 185-213. London: Routledge &
Kegan Paul.
186 Tiina Puurtinen

HALLIDAY, M. A. K. [1985] 1986. An Introduction to Functional Grammar. London: Edward

Arnold.

HATIM, Basil & Ian MASON. 1997. The Translator as Communicator. London & New York:
Routledge.

HOLLINDALE, P. 1988. "Ideology and the Children's Book". Signal 55.3-22. Reprinted in 1992,
Literature for Children: Contemporary Criticism ed. by P. HUNT, 19-40. London & New York:
Routledge.

KNOWLES, Murray & Kirsten MALMKJÆR. 1989. "Translating Ideology: Language, Power and
the World of the Tin Soldier". ELR Journal 3: Language and Ideology. 205-241. Birmingham:
University of Birmingham.

KNOWLES, Murray & Kirsten MALMKJÆR. 1996. Language and Control in Children's
Literature. London & New York: Routledge.

NG, Sik Hung & James J. BRANAC. 1993. Power in Language. Verbal Communication and
Social Influence. Newbury Park, London & New Delhi: Sage Publications.

PUURTINEN, Tiina. 1993. "Kieliopilliset metaforat lastenkirjojen suomennoksissa". Virittäjä 4,

546-563.

PUURTINEN, Tiina. 1995. Linguistic Acceptability in Translated Children's Literature. Joensuu:

University of Joensuu.

RAVELLI, L.J. 1988. "Grammatical Metaphor: An Initial Analysis". Pragmatics, Discourse and
Text. Some Systematically-Inspired Approaches ed. by Erich H. STEINER & Robert VELTMAN.
London: Pinter, 133-147.

SIMPSON, Paul. 1993. Language, Ideology and Point of View. London & New York: Longman.

SOLIN, Anna. 1995. "Ideology, Power, Interpretation - An Assessment of the Social Theory of
Critical Linguistics". The New Courant 4 ed. by Eija VENTOLA & Anna SOLIN, 162-180. Helsinki:
University of Helsinki.

THOMPSON, J.B. 1984. Studies in the Theory of Ideology. Cambridge: Polity Press.

THOMPSON, J.B. 1990. Ideology and Modern Culture: Critical and Social Theory in the Era
of Mass Communication. Cambridge: Polity Press.

TREW, Tony. 1979. "Theory and Ideology at Work". Language and Control ed. by R. FOWLER
et al. 94-116. London: Routledge & Kegan Paul

VIERTIÖ, Annastiina. 1998. "Mitä lymyääkäänurheilu-uutistenmääritteissä". Kielikello 1, 16.

LES SIMPLIFICATIONS NARRATIVES DANS UNE
TRADUCTION FRANÇAISE DE JOSEPH ANDREWS

Kristiina Taivalkoski
Université de Helsinki, Finland

Abstract
The article deals with shifts in reported discourse, on the basis of an analysis of the
French translation (1743) of Fielding's Joseph Andrews. Types of reported discourse
are classified according to the model proposed by McHale (1978). The analysis shows
that the translator transformed the rarer types into more common types, a result that
corroborates Toury's law of growing standardisation (1995). The author suggests that
the reasons for these shifts are to be found in the norms of the dominant literary
system in the target culture. The shifts have repercussions that are not only stylistic
but also thematic.

Zusammenfassung
Der Artikel befaßt sich mit Verschiebungen (shifts) in erlebter Rede, anhand einer
Analyse der französischen Übersetzung (1743) von Fieldings Werk Joseph Andrews.
Die Klassifizierung von Typen erlebter Rede erfolgt nach dem Modell von McHale
(1978). Die Analyse zeigt, daß der Übersetzer die selteneren Typen in generellere
Typen transferiert hat. Dieses Ergebnis bestätigt Tourys Gesetz der wachsenden
Standardisierung (1995). Der Verfasserin zufolge liegt der Grund für diese
Verschiebungen in den Normen des dominierenden literarischen Systems in der
Zielkultur. Die Verschiebungen weisen Nachwirkungen auf, die nicht nur stilistisch
sondern auch thematisch sind.

Resumen
El articulo aborda las desviaciones en el estilo indirecto basándose en el análisis de
una traducción al francés (1743) de la obra de Fielding, Joseph Andrews. Se clasifican
los tipos de estilo indirecto según el modelo propuesto por McHale (1978). El anâlisis
muestra que el traductor transformé los tipos menos frecuentes a tipos mâs corrientes,
un resultado que confirma la ley de estandarización creciente de Toury (1995). El
autor propone que estas desviaciones se deben a las normas impuestas por el sistema
literario dominante en la cultura término. Las desviaciones presentan repercusiones
tanto estilisticas como temáticas.
188 Kristiina Taivalkoski

1. Introduction

En 1743, l'abbé Desfontaines (1685-1745) - journaliste, critique littéraire,

traducteur et écrivain - traduisit en français le roman Joseph Andrews, paru un
an auparavant en Angleterre. C'était l'époque des belles infidèles en France.
L'attitude des traducteurs français envers la littérature anglaise était
ambivalente. D'une part, une vraie anglomanie régnait en France. Plusieurs
ouvrages philosophiques et romans anglais furent ainsi traduits en français.
D'autre part, les traducteurs français de l'époque étaient tellement influencés
par le classicisme rigoriste du XVIIe siècle (le bon goût, les bienséances, le
vraisemblance, la clarté) qu'ils modifiaient les ouvrages traduits pour qu'ils
soient conformes aux normes du classicisme. Voici comment par exemple le
traducteur Du Resnel justifiait cette domestication littéraire:

Ils [les Anglois] aiment à donner à penser jusques dans leurs

moindres Ecrits, & croient faire plaisir au Lecteur de lui laisser
toujours quelque chose à deviner. Nous voulons qu'on nous
épargne la peine de la recherche, & trouver tout sans qu'il en
coute rien à notre empressement. (Du Resnel, 1758:249)

La modification des ouvrages anglais ne touchait pas seulement au «contenu»

mais aussi à leur «forme» c'est-à-dire aux moyens stylistiques et narratifs
utilisés dans les originaux. L'un des moyens narratifs est le discours rapporté,
manière dont les paroles des personnages sont présentées par le narrateur.
Dans ce qui suit, je vais discuter des simplifications du discours
rapporté relevées dans mon corpus. Celui-ci1 consiste en deux chapitres et
plusieurs extraits plus courts de la troisième édition de la traduction française
(1750) et de la troisième édition de l'original (1743) sur la base de laquelle la
traduction fut faite en 1743. J'ai divisé ce corpus en 265 unités de traduction
dont 144 contiennent du discours rapporté. L'analyse a été faite avec le modèle
d'analyse traductionnelle que j'ai développé à partir de celui de Kitty van
Leuven-Zwart. J'ai appliqué dans l'élaboration de mon modèle les concepts du
narratologue Brian McHale qui distingue entre sept types différents du discours
rapporté. Mon objectif principal a été de tester la loi de la standardisation
croissante de Gideon Toury (1995:267-274), loi corroborée par mon étude.
Selon cette loi, dans la traduction d'un texte, les textèmes du texte de départ ont
tendance à être transformés en répertorèmes de la langue (ou bien de la
culture) d'arrivée:

in translation, source-text textemes tend to be converted into

target-language (or target-culture) repertoremes. (ibid.:268)

Par textème, Toury entend le rôle particulier que joue une unité sémiotique
 Simplifications narratives 189

dans un texte, et par répertorème, il désigne les unités du «répertoire» ou

ensemble d'éléments codifiés d'une communauté culturelle. Ces concepts sont
assez vagues et doivent être rendus opérationnels pour que l'on puisse les
appliquer dans une analyse empirique. Je définirai plus loin ce que je considère
comme textème et comme répertorème dans le discours rapporté de mon
corpus.

2. De trois à sept formes du discours rapporté

Dans son ouvrage Narrative Fiction (1983), Shlomith Rimmon-Kenan

énumère sept types différents du discours rapporté, initialement distingués par
Brian McHale en 1978. Cette répartition est plus polyvalente que la tripartition
traditionnelle entre discours direct, discours indirect et discours indirect libre.
Les quelques études qui ont analysé les modifications du discours rapporté
dans la traduction (par exemple celle de Rachel May, 1994) se basent sur cette
tripartition. En outre, d'autres chercheurs ont observé des cas de disparition du
plurilinguisme2 dans la traduction (par exemple les études faites par le groupe
GRETI, ou Groupe de recherche en traductologie, au Canada) - ce qui amène
souvent également des glissements dans le discours rapporté - mais leurs
études ont porté sur le discours direct. Par conséquent, le phénomène n'a pas
été sondé en profondeur.
La typologie de McHale (1978:258-260) part de l'opposition entre la
diegesis et la mimesis fondée par Socrate dans la République de Platon. Par
diegesis, Socrate désigne les cas où le poète est clairement le locuteur et
n'essaie pas de faire croire que quelqu'un d'autre parle. Par conséquent, la
diegesis signifie le discours indirect. Dans la mimesis, par contre, le poète crée
l'illusion que ce n'est pas lui qui parle mais les personnages. Il s'agit donc du
discours direct que Platon considère comme trompeur (Rimmon-Kenan
1983:106-110). Il est à noter que ces termes ont été utilisés dans plusieurs sens
dans l'histoire de la narratologie. Shlomith Rimmon-Kenan, qui a une attitude
purement descriptive vis-à-vis de l'opposition entre la diegesis et la mimesis,
utilise ces termes dans le sens platonicien. Elle note que même dans le discours
direct, les paroles des personnages sont citées par le narrateur. Par conséquent,
il n'y a que différents degrés de diégétique dans le discours rapporté.
Le classement de McHale des sept types du discours rapporté varie de
la représentation «purement» diégétique à la représentation «purement»
mimétique. Tous ces types ont des caractéristiques linguistiques et stylistiques
dont les modifications peuvent produire aussi un glissement dans le discours
rapporté. Je vais maintenant énumérer ces types. Les exemples sont pris de
mon corpus3, à l'exception de l'exemple concernant le discours direct libre,
septième type du discours rapporté:
1. le résumé diégétique (diegetic summary): The Consequence of this
190 Kristiina Taivalkoski

Interview was a Whisper through a hundred Visits, which were separately

performed by the two Ladies (--). (JA, I, iv, 13-14, c'est moi qui souligne)
2. le compte rendu diégétique (summary, less 'purely' diegetic): And now the
witty Spark, seeing his Jokes take, began to rise in Spirits; and turning to
Adams, challenged him to cap Verses (--). (JA, II, xi, 174, c'est moi qui
souligne)
3. le discours indirect (indirect content-paraphrase): Upon which Adams,
with a Look full of ineffable Contempt, told him, he deserved scourging for his
Pronunciation. (JA, II, xi, 174, c'est moi qui souligne)
4. le discours indirect mimétique (indirect discourse, mimetic to some
degree): he began to revile them, saying, 'that Robberies on the Highway were
now grown so frequent, that People could not sleep safely in their Beds, (—).'
(JA,II,xi,173)
5. le discours indirect libre (free indirect discourse): Which he bid him do,
and he would light his Pipe in the mean time. (JA, II, xi, 173, c'est moi qui
souligne)
6. le discours direct (direct discourse): Slipslop seeing Lady Booby enter the
Room, cried, Help! or I am ravished (--). (JA, IV, xiv, 208)
7. le discours direct libre (free direct discourse) par exemple «Fainy's head
suddenly got very light. Bright boy, that's me, ambition and literary taste....
Gee, I must finish Looking Backward ... and jez, I like reading fine, an' I could
run a linotype or set up print if anybody'd let me. Fifteen bucks a week ...
pretty soft, ten dollars' raise.» (McHale 1978:260 et Rimmon-Kenan 1983:110)

Je fais une distinction entre les types non mixtes et les types mixtes du discours
rapporté. Les types non mixtes ne contiennent que du discours du narrateur
(résumé diégétique, compte rendu diégétique, discours indirect) ou du
personnage (discours direct, discours direct libre). Les types mixtes, par
contre, contiennent aussi bien du discours du narrateur que du personnage. Tels
sont le discours indirect mimétique et le discours indirect libre. Le tableau (1)
récapitule mon classement:

Tableau 1. Répartition des types du discours rapporté en types mixtes et en

types non mixtes

type non mixte type mixte

résumé diégétique discours indirect mimétique
compte rendu diégétique discours indirect libre
discours indirect
discours direct
(discours direct libre n'apparaît
pas dans mon corpus)
 Simplifications narratives 191

3. Diégétisation, mimétisation, standardisation

J'ai classé les types mixtes du discours rapporté comme textèmes et les types
non mixtes comme répertorèmes, parce que les formes mixtes ont été rares
dans la littérature occidentale du XVIIIe siècle. Selon Ann Banfield (1982:225-
235), elles ont certes fait des apparitions sporadiques dans l'histoire de la
littérature. Cependant, les premiers auteurs pour lesquels l'usage de ces formes
est généralement reconnu sont La Fontaine en France, Henry Fielding et Jane
Austen en Angleterre, et Goethe en Allemagne. Le premier auteur qui en
développa un art systématique ne fut pourtant que Flaubert au XIXe siècle.
Comme j'ai testé la loi de la standardisation croissante dans mon étude,
il m'a fallu définir ce que je considère comme standardisation: dans ce
contexte, elle est le glissement d'un type mixte du discours rapporté vers un
type non mixte. Le tableau (2) illustre tous les glissements possibles - en
théorie - entre les types mixtes et non mixtes du discours rapporté et leur
valeur vis-à-vis de la loi de la standardisation croissante. Les glissements entre
les types non mixtes sont neutres, les glissements de types mixtes vers les types
non mixtes sont des standardisations (en caractères gras) et les glissements de
types non mixtes vers les types mixtes et les glissements entre les types mixtes
sont des anti-standardisations (en italiques):

Tableau 2. Glissements possibles entre types mixtes et types non mixtes

1 i non mixte non mixte (pas de glissement)

2 mixte → mixte (pas de glissement)
3 non mixte→mixte
4 mixte → non mixte
5 non mixte 1→non mixte 2 4 j
6 mixte 1 → mixte 25
7 non mixte → Ø6
8 mixte → Ø
9 Ø → non mixte
10 Ø → mixte

Le tableau (3) montre la proportion des types de glissement, dans le corpus

particulier que j'ai étudié. Le numéro en-dessous de chaque colonne désigne le
type de glissement tel qu'indiqué dans le tableau (2).
192 Kristiina Taivalkoski

Tableau 3. Proportion des types de glissement dans le corpus analysé.

J'ai rapporté le nombre d'apparitions (au-dessus de chaque colonne) de chaque

type de relation à la totalité des unités de traduction de mon corpus contenant
du discours rapporté, soit 144. Cela m'a donné les pourcentages suivants:
a) cas neutres (colonnes 1, 5, 7 et 9) = 90%
b) standardisations (colonnes 4 et 8) = 10%
c) anti-standardisations (colonnes 2, 3, 6 et 10) = 0%
Par conséquent, mon corpus ne présente aucun cas de discours rapporté qui
infirmerait l'hypothèse de la standardisation croissante.

4. Mécanismes des glissements narratifs

Après la présentation de mes résultats, je procède à la description des sept

types du discours rapporté pour montrer les mécanismes à l'œuvre dans les
glissements du discours rapporté d'un texte narratif traduit. Les trois premiers
types sont non mixtes (discours du narrateur), le quatrième et le cinquième type
sont mixtes (discours du narrateur et du personnage) et les deux derniers types
sont non mixtes (discours du personnage).
Le type les plus diégétique du discours rapporté est le résumé
diégétique qui annonce que l'énoncé a eu lieu mais sans préciser le sujet ou la
nature de la discussion. Dans l'exemple suivant, l'énoncé dans le texte
d'arrivée (TA, 1ère ligne) est représenté par le résumé diégétique tandis que
dans le texte de départ (TD, 2ème ligne), il s'agit du compte rendu diégétique
étant donné que le narrateur indique la nature de la discussion des personnages.
 Simplifications narratives 193

TA: & elle s'y entretenoit quelquefois avec lui;

TD: and [she would] converse with him in great Familiarity.
(AV, I, iv, 20; JA, I, iv, 13, c'est moi qui souligne)

La raison de cette modification est euphémique. Le personnage de Lady Booby

est moins dévergondé dans la traduction que dans l'original et, par conséquent
le fait qu'elle badine avec son valet a été omis. Cela produit ici également un
glissement narratif: on ne peut plus deviner le sujet de la discussion dans le
texte d'arrivée.
Le type suivant du discours rapporté, le compte rendu diégétique diffère
du résumé diégétique par le fait qu'il annonce également le sujet de la
discussion. Il est plus informatif. Voici un exemple où le compte rendu
diégétique est transformé en discours indirect dans le texte d'arrivée:

TA: Aussi-tôt il ordonna qu'on les allât chercher.

TD: he ordered them into his Presence.
(AV, II, xi, 257; .JA, II, xi, 176)

Le troisième type diégétique est le discours indirect traditionnel dans lequel le

narrateur fait une paraphrase de l'énoncé du personnage sans pour autant en
imiter le style ou la forme. Par conséquent, le lecteur ne peut reconstruire
l'énoncé qui a «eu lieu» au niveau de Yhistoire, ce qui est le facteur commun
des trois formes les plus diégétiques du discours rapporté. Dans mon corpus, il
arrive que le traducteur remplace le discours indirect par le discours direct. En
voici un exemple:

TA: Adams entendant parler d'écrou, lui dit modestement: "Vous

ne voudriez pas, Monsieur, me condamner sans m'entendre?"7
TD: Adams then said, 'he hoped he should not be condemned
unheard.' (AV, II, xi, 258; AV, II, xi, 176)

La différence entre les glissements du discours indirect en discours direct et les

glissements entre les types plus diégétiques du discours rapporté mentionnés
ci-dessus est que ceux-là ne pourraient pas passer inaperçus par le traducteur.
Je suis d'avis qu'ils sont causés dans mon corpus par les normes littéraires de
l'époque: le discours direct était plus apprécié parce que c'était le type
d'énonciation utilisé dans la tragédie, genre canonisé dans la France du XVIIIe
siècle.8 Selon mes calculs, la proportion du discours direct est de 82 % dans le
texte d'arrivée alors que dans le texte de départ elle est de 73 %. Cela est
probablement dû au fait que le traducteur a voulu recourir dans sa traduction à
un phénomène qui plaisait aux lecteurs français de l'époque.
Les deux types suivants dans le classement de McHale sont le discours
indirect mimétique et le discours indirect libre. Dans ces formes, plus
194 Kristiina Taivalkoski

mimétiques, il est question du mélange du discours du narrateur et du discours

des personnages. Par conséquent, la voix est à la fois celle du personnage et
celle du narrateur. Dans ces formes mixtes, le narrateur préserve plusieurs
aspects de la forme et du style de l'énoncé du personnage et il serait possible
de reconstruire «l'énoncé originel»9. La différence importante entre le discours
indirect mimétique et le discours indirect libre est que ce dernier omet toujours
une proposition incise et la conjonction que. Plusieurs indices linguistiques et
extralinguistiques peuvent révéler qu'il est question de formes mixtes du
discours rapporté. Les indices grammaticaux sont les décalages temporels ou
déictiques. Les autres indices des formes mixtes du discours rapporté peuvent
être stylistiques et idéologiques, manifestant l'attitude subjective d'un
personnage particulier. Par exemple, les apparitions d'expressions idiolectales
appartenant à un personnage dans le discours indirect témoignent qu'il s'agit
du discours indirect mimétique. Par conséquent, lorsqu'il y a des mélanges de
style ou de niveaux de langue dans le discours rapporté d'un texte, il y a
souvent aussi des formes mixtes. Il s'ensuit que si le traducteur modifie ces
éléments, il modifie nolens volens également le type du discours rapporté.
Voici deux exemples de glissements des formes mixtes du discours rapporté en
formes non mixtes. Dans le premier cas, il y a un glissement du discours
indirect mimétique en discours direct. Dans le deuxième, la neutralisation de
l'idiolecte du personnage (jargon juridique) amène une transformation du
discours indirect mimétique en discours indirect.

TA: qu'il leur parla ainsi: "Les Voleurs de grand chemin se sont
tant multipliés depuis quelque tems, que les gens ne peuvent
dormir en sureté dans leur lit.
TD: then he began to revile them saying, 'that Robberies on the
Highway were now grown so frequent, that People could not
sleep safely in their Beds, (--).' (AV, II, xi, 257; JA, II, xi, 173,
c'est moi qui souligne)

TA: [le juge de paix] donna ordre d'écrouer les Prisonniers.

TD: [the Justice] ordered his Clerc to make the Mittimus.
(AV, II, xi, 258; JA, II, xi, 176, c'est moi qui souligne)

Le texte de départ de mon corpus contient aussi bien du discours indirect

mimétique (12 apparitions dans mon corpus) que du discours indirect libre (2
apparitions) alors que le texte d'arrivée n'en contient pas du tout. Cela est
probablement causé par les normes littéraires du système dominant (le
classicisme) dans la culture d'arrivée qui n'acceptait ni les décalages temporels
ni le plurilinguisme (ou la polyphonie) qui sont justement les indices des
formes mixtes du discours rapporté. Voici ce que dit notre traducteur, d'un ton
ironique, sur les décalages temporels et sur le plurilinguisme:
 Simplifications narratives 195

MAINTENANT. Cet adverbe joint à un préterit fait un très-bon

effet. Exemp. "Je sçai qu'autrefois à Tyr, & que maintenant à
Carthage vous avez méprisé les recherches de bien des Rois.
(Eneid. t. 2. p. 9.) (Desfontaines 1728:106)

PRETERIT indéfini: Le Trad. de Virgile (Georg. p. 389) dit:

"Saturne ne fut pas le seul à qui l'antiquité ait donné la faucille
pour symbole." Ait donné est ici pour donna: ou bien fut est mis
pour est. Tout cela est bon. (Desfontaines 1728:137)

Les détails de ce livre [traduction de Pamela] sont petits, sans

être bas. On n'y entend point le jargon dégoutant d'un bas
domestique, ou d'un homme de la lie du peuple. Si un
domestique y parle, c'est simplement & raisonnablement; car le
bon sens est de tous les états, & il plait dans la bouche de quelque
personnage que ce soit, pourvu qu'il ne soit point avili par des
pensées & des expressions qui sentent trop le bas peuple. &
qu'un honnête homme n'entend pas volontiers.
(Desfontaines 1757:366, c'est moi qui souligne)

Cependant, les glissements des formes mixtes du discours rapporté en formes

non mixtes sont fréquents dans littérature, encore aujourd'hui. Rachel May
(1994:90, 110-114 et passim) qui a étudié les traductions anglaises de la
littérature russe est d'avis que les traducteurs ont tendance à faire des
modifications qui font disparaître la coexistence de voix dans le texte. Les
études faites par le GRETI témoignent de tendances similaires dans les
traductions françaises de Faulkner (cf. Vidal 1991).
Les types les plus mimétiques du discours rapporté sont le discours
direct et le discours direct libre. Le discours direct est une citation du dialogue
ou du monologue des personnages, qui crée l'illusion du «mimesis pur» même
s'il est toujours stylisé d'après McHale (1978:259). Le discours direct libre est
le type le plus mimétique du discours rapporté. Comme je l'ai dit plus haut, ce
type, qui est caractéristique du monologue intérieur et peut dévier de
l'orthographe et de la typographie normales en imitant l'aspect phonétique des
énoncés, n'apparaît pas dans mon corpus. Cela n'empêche que ce type se prête
aussi facilement aux glissements narratifs.
J'ai déjà mentionné que le texte d'arrivée contient davantage de
discours direct que le texte de départ. Le traducteur a en effet plusieurs fois
transformé le discours indirect ou bien le discours indirect mimétique en
discours direct. Cela rend le ton du texte d'arrivée plus dramatique, comme
dans une tragédie. Parfois, il a même ajouté une réplique en discours direct
dans sa traduction alors que le texte de départ ne contenait pas du tout de
discours rapporté. Voici un exemple d'un tel glissement:
196 Kristiina Taivalkoski

TA: Il faut , dit-elle, que je révele ce fatal sécret, je ne puis plus

le garder.10
TD: nor could she conceal these boiling Passions longer, without
bursting. (AV, IV, xiii, 289; JA, IV, xiii, 198)

Cet exemple montre que le texte d'arrivée diffère stylistiquement du texte de

départ. Le texte d'arrivée a été mimétisé dans le processus de traduction de
façon qu'il ressemble davantage à une pièce de théâtre que le texte de départ.
Voici encore deux exemples de glissements du discours indirect (premier
exemple) et du discours indirect mimétique (deuxième exemple) en discours
direct qui montrent que cette tendance est assez systématique dans mon corpus:

TA: elle dit tout bas à Pamela. "Je me trouve incommodée, ma

chere Niéce voudroit-elle bien se charger d'entretenir Mylord &
mon Neveu, pendant que je me reposerai?"
TD: she whispered Pamela, that she was taken a little ill, and
desired her to entertain her Husband and Beau Didapper.
{AV, IV, xiii, 288; JA, IV, xiii, 198)

TA: Toutes les femmes me sont indifférentes, repliqua Joseph.

TD: Joseph returned, that all Women he had ever seen, were
equally indifferent to him. {AV, I, v, 24; JA, I, v, 15)

5. Conclusion: la disparition du plurilinguisme peut mener à la

standardisation narrative d'un texte

Je peux conclure que la narration, exprimée par le discours rapporté, est moins
polyvalente dans le texte d'arrivée que dans le texte de départ de mon corpus,
ce qui corrobore l'hypothèse de la standardisation croissante de Toury. Les
raisons de cette simplification narrative ne résident pas dans l'incompétence du
traducteur mais dans son opinion sur ce qui est acceptable dans la culture
d'arrivée, ce que le traducteur dit de façon explicite dans son Dictionaire
Néologique. Il est à noter que les glissements narratifs dans le discours rapporté
sont toujours possibles dans les traductions littéraires actuelles. Ils peuvent
passer inaperçus par le traducteur dans des cas de glissement entre des types du
discours rapporté apparentés comme le discours indirect et le discours indirect
mimétique. La disparition du plurilinguisme - de sociolectes, d'idiolectes, de
jargons professionnels - a des répercussions également sur la narration d'un
texte.
Les répercussions de la standardisation du discours rapporté dans la
traduction ne sont pas insignifiantes. C'est que les différents types du discours
rapporté jouent souvent un rôle particulier dans l'ensemble d'un ouvrage
littéraire. Ce rôle n'est pas uniquement stylistique. Les types du discours
 Simplifications narratives 197

rapporté peuvent, par exemple, contribuer à la description des personnages

(idiolectes) et donner au lecteur des indices de l'attitude du narrateur ou bien
de l'auteur implicite envers les personnages. Fielding utilise souvent les formes
mixtes du discours rapporté pour railler quelques personnages de ses ouvrages;
le narrateur mélange les expressions idiolectales des personnages à son propre
discours - rhétoriquement impeccable - pour souligner le ridicule de leurs
expressions. Par conséquent, les glissements dans le discours rapporté peuvent
faire disparaître cette ironie fine du narrateur et altérer sa personnalité. En
outre, il va de soi que le lecteur du texte d'arrivée aura une impression
différente de l'ouvrage que celui du texte de départ.

Notes

1. Le texte d'arrivée de mon corpus comprend 2513 mots, comptés par l'ordinateur, tandis
que le texte de départ en contient 3342. Cela s'explique par le fait qu'il y a plusieurs
omissions dont la plus grande ne comprend pas moins de 532 mots. En outre, le texte
d'arrivée contient une grande addition de 201 mots.

2. Par plurilinguisme, j'entends ici la coexistence de plusieurs registres - dialectes,

sociolectes et idiolectes - dans un texte littéraire.

3. J'utiliserai l'abréviation JA pour désigner le texte de départ de mon corpus et l'abréviation

AV pour le texte d'arrivée

4. Par l'indication «non mixte 2», je veux dire autre type non mixte du discours rapporté.

5. Par l'indication «mixte 2», je veux dire autre type mixte du discours rapporté.

6. Par 0, je veux signaler addition (Ø →) ou omission (→ Ø).

7. Cette réplique est suivie par l'addition suivante qui rend le style du texte d'arrivée encore
plus pathétique: «Ce seroit une injustice criante, contraire à l'administration de la Police
en Angleterre.» {AV, II, xi, 258)

8. Voir aussi Lefevere (1992: 87-92). Le même motif - les exigences de la poétique française
du XVIIIe siècle - expliquerait les changements que fit Houdart de La Motte dans sa
traduction de l'Iliade d'Homère en 1714. Selon Lefevere, l'épopée n'avait pas bonne
presse dans la France du classicisme. Par conséquent, La Motte jugea l'Iliade selon les
critères de la tragédie, genre prédominant à l'époque. Comme on le sait, sa traduction
déclencha la seconde phase de la Querelle des Anciens et des Modernes en France.

9. Selon Brian McHale (1978: 256-257) reconstruire «l'énoncé originel» est une illusion dans
la littérature parce qu'il n'y a pas d'énoncé réel derrière l'énoncé du texte. Le seul énoncé
qui existe est dans le texte.

10. Le pathos de cette réplique est augmenté par l'addition suivante dans le texte d'arrivée:
«Son poids m'accable; en le révélant, je trouverai peut-être quelque secours.» {AV, IV, xiv,
289)
198 Kristiina Taivalkoski

Références

BANFIELD, A. 1982. Unspeakable sentences: Narration and representation in the language

of fiction. Boston & London: Routledge & Kegan Paul.

DESFONTAINES, P.-F. G. 1728. Dictionaire Néologique. A l'usage des beaux Esprits du

Siécle. Avec l'Eloge historique de Pantalon-Phœbus. Par un Avocat de Province. Amsterdam:
Michel Charles le Cene.

DESFONTAINES, P.-F. G. 1757. "Romans." Esprit de l'abbé Desfontaines, ou Réflexions sur

differens genres de science et de littérature. Tome 4: Avec des Jugemens sur quelques Auteurs
& sur quelques Ouvrages tant Anciens que Modernes éd. par LA PORTE, J. De, 305-379.
Londres: Clement, Libraire de la Cour.

DU RESNEL, J. F. du Bellay. 1758. "Discours préliminaire du traducteur." Œuvres diverses de

Pope. Traduites de l'anglois, 231-273. Tome 2. Amsterdam & Leipzig: Arkstee & Merkus.

FIELDING, H. 1743. The History of the Adventures of Joseph Andrews, and of his Friend Mr.
Abraham Adams. Written in Imitation of the Manner of Cervantes, Author of Don Quixote.
London: A. Millar.

FIELDING, H. 1750. Les Avantures de Joseph Andrews, et du ministre Abraham Adams,

Publiées en Anglois, en 1742. Par M... Feilding [sic], Auteur de l'Enfant Trouvé; traduites en
François, Par l'Abbé des Fontaines, sur la troisiéme Edition. Londres: Meyer.

LEFEVERE, A. 1992. Translation, Rewriting and the Manipulation of'Literary Fame. London
& New York: Routledge.

LEUVEN-ZWART, K. VAN. 1989/1990. "Translation and Original. Similarities and

Dissimilarities." Part 1, Target 1(2), 151-181. & Part 2, Target 2(1), 69-95.

MAY, R. 1994. The Translator in the Text. On Reading Russian Literature in English.
Evanston: Northwestern University Press.

McHALE, B. 1978. "Free Indirect Discourse: a survey of recent accounts." Poetics and Theory
of Literature 3, 249-287.

RIMMON-KENAN, S. 1983. Narrative Fiction: Contemporary Poetics. London: Methuen.

TOURY, G. 1995. Descriptive Translation Studies- and beyond. Amsterdam & Philadelphia:
John Benjamins.

VIDAL, B. 1991. "Plurilinguisme et traduction - Le vernaculaire noir américain: enjeux,

réalité, réception à propos de The Sound and the Fury." TTR (Traduction Terminologie
Rédaction) 4(2), 151-188.
 PART V

Computer aids in the translation process

SEEKING TRANSLATION EQUIVALENTS:
A CORPUS-BASED APPROACH

Jeremy Munday
University of Bradford, U.K.

Zusammenfassung
In diesem Artikel wird ein praktischer Ansatz beschrieben hinsichtlich der
Verwendung umfangreicher Computer-Korpora sowohl in zweisprachiger
Lexikographie als auch im Bereich des Übersetzens. Folgendes wird dargelegt:
(1) Methoden der Ermittlung von Übersetzungsäquivalenten für Einträge in
zweisprachigen Wörterbüchern, Spanisch-Englisch, wo der Lexikograph sich
zwischen Quasi-Synonymen zu entscheiden hatte;
(2) mögliche Anwendungen für professionelle Übersetzer. Aus bereits veröffentlichten
Literatur-Übersetzungen werden Beispiele herangezogen und die jeweiligen
Entscheidungen der Übersetzer anhand großer "KontrolP'-Korpora überprüft.
Der Artikel diskutiert auch mögliche negative Implikationen der
Korpuslinguistik, legt aber auch nahe, daß die immer schneller zunehmende
Verfügbarkeit von Computer-Korpora für Forschungszwecke spannende
Möglichkeiten sowohl für Übersetzer als auch für die Wissenschaftskritik wird
bereitstellen können.

Résumé
Cet article décrit une approche pratique pour l'usage des grands corpus informatisés
en lexicographie bilingue et en traduction. Il présente:
(1) des méthodes afin de déterminer les équivalences pour les entrées de dictionnaires
bilingues espagnol-anglais, où le lexicographe a dû choisir entre plusieurs quasi-
synonymes;
(2) des applications possibles pour le traducteur professionnel. Des exemples sont tirés
de traductions littéraires publiées, pour comparer les choix des traducteurs par rapport
à de larges corpus de contrôle.
L'article traite des répercussions négatives possibles de la linguistique de
corpus, mais il suggère aussi que la disponibilité de plus en plus grande de corpus
informatisés pour la recherche fournira d'énormes possibilités non seulement au
traducteur professionnel mais aussi au chercheur critique.
202 Jeremy Munday

Resumen
Este articulo describe un enfoque prâctico para el uso de los grandes corpus
informatizados tanto en la lexicografia bilingüe como en la traducción. En él se
plantean:
(1) métodos para determinar equivalencias para palabras cabeza de articulo en
diccionarios bilingües español-inglés, cuando el lexicógrafo ha tenido que elegir entre
cuasi-sinónimos;
(2) posibles aplicaciones para el traductor profesional. Se sacan ejemplos de
traducciones literarias ya publicadas, comprobando las selecciones hechas por los
traductores y comparândolos con ejemplos extraidos de grandes corpus "control".
Aunque el articulo baraja las posibles repercusiones negativas de la lingüistica
del corpus, acaba concluyendo que la disponibilidad cada vez más grande de los
corpus informatizados para la investigacion proporcionarâ posibilidades de enorme
interés no solo para el traductor profesional sino también para el critico académico.

1. Introduction

The electronic corpus of naturally occurring texts in machine-readable format

has become a central concept in the computer-assisted analysis of language.
However, research has so far tended to focus either on computational stylistic
analyses of distinct but relatively small corpora, such as Jane Austen's
narrative style (Burrows 1987), or on large, representative databases of
language such as the British National Corpus and COBUILD dictionary
projects. This paper will suggest ways of bringing the two kinds of corpora
together in translation, using the large corpora as "controls" against which to
evaluate specific examples from texts on which the bilingual lexicographer,
translation critic or translator is working, and serving as an aid to the selection
of a translation equivalent. The examples below, focusing for reasons of space
on the translation of individual lexical items, will be based on my own
experience as a lexicographer, as a researcher on translation and as a practising
translator.1

2. What is a computerized corpus?

The first large computerized corpus of texts designed for specific large-scale
lexicographic use was put together for the COBUILD English Dictionary
project at Birmingham, U.K. This has now developed into the Bank of English
(BoE). Other British publishers have now followed this line, notably with the
British National Corpus (BNC). The BNC comprises a variety of texts
(newspapers, serious and light fiction, non-fiction, and some spoken
transcriptions) totalling, in January 2000, around 110 million words.
In-house, the large dictionary publishing companies have bigger SUN
 A corpus-base approach 203

systems for the analysis of their corpora. These corpora are now increasingly
available for researchers: the British National Corpus, requiring 4-6 Gb of hard
disk space, is currently available in CD-Rom form to research institutions for
around $400 or via a single-user Internet subscription for around $110 per year.
Smaller-scale commercially-produced research tools for the linguistic analysis
of researchers' own corpora are also now more widely available. Most of these
programs can be easily used by non-IT specialists. A very good example is
Oxford University Press's user-friendly and very effective WordSmith (Scott
1996), which contains a range of useful tools for the analysis of texts and can
be used on most PCs and with little training. The researchers' corpora can be
their own academic texts or books, newspaper collections purchased on CD-
Rom, or literary texts for which the researcher has sought copyright to hold on
disk, as was the case with my own analysis of Garcia Márquez stories in
translation (see 5.2 below).

3. Why use a computerized corpus?

Lexicographers used to have to rely on their intuition or on a small number of

examples laboriously trawled from printed matter when putting together a
dictionary entry. In contrast, as Sinclair (1991: 42) stresses, the quality of
linguistic evidence provided by a large computerized corpus, particularly
regarding typical collocations and grammatical patterns, is vastly superior to
the analyst's. For example, the comparison of frequency of words in a corpus
can help decide which headwords to include, and concordances (see 4.1 below)
can indicate what the most common uses of a headword are and assist in the
selection of "natural" examples. The concept of typicality is crucial both to
lexicography and translation studies; the bilingual dictionary, for instance,
prefers the most widely applicable translation equivalent that is "insertable in a
wide range of contexts" (Manley 1985: 252), while Baker (1993, 1995), who
urges the use of computer corpora in translation studies research, focuses on
the need to identify typicalities of translated language compared to non-
translated language. The differences noted may reveal elements of the process
of translating and the norms at work, which links with Toury's (1995)
influential work on descriptive translation studies.
Specific examples can be evaluated against the larger corpora, which
thus act as a kind of "control". A control corpus is necessary to check the
markedness of the individual examples in both languages and, in the case of
translation criticism, to try and gauge whether any shifts that do take place are
the result of the typical idiolect of the translator or adherence to typical TL
collocation patterns. Very little work seems to have been done in this area. The
difficulty often is that there is no immediately comparable database against
which to judge the ST and TT, since the corpora are either too small, or vary in
204 Jeremy Munday

geographical provenance or genre from the texts under investigation. Many

factors and variables can come into play in the analysis of just a single
translation.

4. Basic tools of corpus linguistics

The great strength of computers is their accuracy and speed in counting and
ordering stored information; and the easiest linguistic phenomena for the
computer to count and order are individual words. The most straightforward
definition of words, which is the one that computers find easiest to assimilate
(see Baker 1993) and which will be used here, is that of an orthographic word-
form, that is, a group of typographical characters with a space on either side
(e.g. an, orthographic and word-form are three examples of words). WordSmith
is able to count the word-forms, discounting hyphens as required, to provide
basic statistics for a text (text length in words, sentences, paragraphs, etc.), to
give a frequency list of the word-forms either in alphabetical order or in
descending order of frequency, and to provide concordances, basic collocation
statistics and type-token ratio (the number of distinct word-forms as a
percentage of the total word-forms in a text). Some of these measurements are
potentially extremely useful in computational stylistics to describe a writer's
"fingerprint", and in descriptive translation studies, where the texture and
structure of STs and TTs can be compared (see, for example, Holmes 1994 on
authorship attribution and Munday 1997, 1998 on translation). The present
paper, however, focusing on seeking small-phrase translation equivalents, will
limit itself to two specific and related areas, those of concordances and
collocations.

4.1 Concordances

The KWIC (KeyWord In Context) concordance displays examples of the

analyst's particular search term (the keyword) on screen together with the co-
text of some 40 characters to either side of that keyword. Simple concordances
can be produced within seconds by the concordancer tool in WordSmith. The
keyword, which the analyst enters, appears in the centre of the screen, and all
the concordance lines can be sorted alphabetically according to the words that
occur to either the right or the left of the keyword. For example, part of the
concordance for the word-form fond in the COBUILD Bank of English corpus
can be seen in Figure 1 below. It is ordered according to the first word to the
right of fond. This is especially useful for seeing grammatical patterns and
most frequent collocates and will be further discussed below.
 A corpus-base approach 205

Figure 1. Part of a concordance of "fond"

was an energetic literary hostess but also a fond aunt to her nephews and nieces.
e family fortunes. They had all bade him a fond farewell, then settled down
had seen the look on Jimmy's face. These fond farewells were too damn facile,
hing for the other person. I've told you how fond I am of Lena. Have you so little
"the food, however, remains a glowingly fond memory for both. Back at
ves of French rosarians or the daughters of fond nurserymen. Sometimes,
Atlantic and married a British barrister fond of amateur theatrics, Cecil
are the right person to know. You are very fond of Anna, rightly so. She is also
rica's out-of-touch marketing man. He was fond of bragging that his relations
Then perhaps he should, Angela. I'm very fond of Gerald." "But then everyone
ou, if I may. I do have a pet dog. I am very fond of her, and it would grieve me
e another to put matters right." "I got very fond of him. And since he was uniq
a had called him Uncle Mark and had been fond of him, unlike most of her other
ham may be our Judas, but he's too fly, too fond of himself, too ambitious to
und a honey-pot. Don't get me wrong. I'm fond of my wife and my boys are
kept trying for a French twist, Michael was fond of recalling, but it kept Connie
but that was a long time ago as he was fond of reminding her. And what can
a good dose of asphalt, starting, as he was fond of saying, with his own back
All his, and no one else's. Pearle wasn't fond of the rabbits, which was just as
our German cousins and was particularly fond of them, though the most loyal
right behind the house. Mrs Roosevelt was fond of this room, with its lovely

Almost any search term can be called up. Thus, a lexicographer may search for
the particular term which forms the entry word being compiled, while
translation analysts may use word frequency lists or their own initial reading of
the text to decide which search terms to examine. It is important to realize,
however, that the intellectual work and interpretation of the concordances is
provided by the user, the computer simply being a very useful tool in
facilitating the analysis.

4.2 Collocation and frequent structures

Collocation is the study of words (the collocates) which frequently occur

206 Jeremy Munday

together (in the case of fond, as will be discussed in the analysis below, it
would be of and words such as farewell and memory). It shows the
paradigmatic axis of language. The most obvious adjacent collocates may be
visible from a concordance, but WordSmith also has a basic collocation
program, which calculates and presents in tabular form the most frequent
collocates of any search term. Much more sophisticated analytical programs
are employed by in-house lexicographers. Yet, if our goal is to use the corpus
as a kind of interactive thesaurus to seek out typical or appropriate translation
equivalents, then either of the packages is useful, since their main function in
our case would be to suggest ideas to the translator which might otherwise
have been overlooked. Thus, statistical results for collocates of rain (verb and
noun) point out the most common uses to be acid rain, heavy rain, wind...rain,
pouring rain, rain forest, rain... down, torrential rain, driving rain, rain...
snow, rain... falling, through... rain, began... rain, cold... rain. Such a list can
be useful for a translator seeking the translation of a collocate within the co-
text of rain (see the "oblique, wind-driven rain" example in 5.2 below).

5. The corpus in action

In this section, a number of example case studies will be presented to suggest

how, in practice, an electronic corpus and the related computerized tools may
help in the selection and evaluation of translation equivalents. The first
examples (section 5.1) are based on bilingual dictionary work, where the
lexicographer, due to space constraints, is often forced to choose between
several different near-synonyms. Section 5.2 will then apply similar techniques
to ST-TT comparisons.

5.1 Uses in bilingual lexicography

One very straightforward way of using a large control corpus is to determine

the frequency of a word-form. This is useful in choosing between two
candidates for the translation of a single word. For instance, the choice between
the near-synonyms involve and entail may be aided by the figures from the
Bank of English corpus which suggest that involve is around ten times more
frequent than entail. Analysis of many complex candidate groups is facilitated
by collocation statistics, but there is often a degree of subjectivity on the part of
the lexicographer in evaluating the figures and instances. A case in point is a
lexicographer faced with translating the Spanish adjective apegado, which
could be translated by attached, fond or devoted. Example sentences are often
given to the lexicographer to translate as an aid to selecting the best translation
equivalent. The most appropriate example(s) are retained in the final version of
the dictionary as illustrative uses for the user, though space constraints often
 A corpus-base approach 207

mean that the examples are cut or reduced:

(1) Es un chico muy apegado a su familia.

He is a boy who is very his family.

(2) Estân muy apegados a lo rural.

They're very country life.

(3) No creo que dimita porque esta muy apegado al poder.

I don't think she'll resign because she's very power.

Frequency analysis of the translation candidates in the COBUILD corpus

indicates that attached (the adjective, not verb forms such as "I attached the
wire to the frame") occurs 3000 times, devoted occurs 2789 times and fond has
1282 occurrences. Yet fond is nearly twice as frequent as devoted in the spoken
corpus. At this point, closer examination of sample instances can help
determine syntactic patterns, typical collocates and the connotations of the
words. Figure 1 above had shown an extract from the concordance of fond.
The most striking pattern, which can be seen visually on the concordance
printout, is 'fond of somebody/something/doing something". The other
candidates similarly have patterns attached to and devoted to.
Further analysis of the concordances and collocations points to typical
collocates or lexical fields in which these words occur. So, for fond of
collocates are family relatives (aunt, wife, him/her, etc.) and positive emotive
words such as farewell, memory and recollections. Negative uses are few
("fond of bragging", "a fond pretence"). Devoted collocates especially with
words in the lexical field of family ("devoted to you/him/her the family", "a
devoted young man/family man/mother," etc.). On the other hand, attached is
clearly the least emotive. Although there are a few examples in the corpus of
"attached to the kids/your home", the majority are neutral or negative
collocations, such as "stigma attached" or links that have negative outcomes ("I
vowed never to let myself become attached to objects again!").
Such an albeit brief analysis of typical collocates can assist the
lexicographer in selecting the most appropriate translation. Moreover, the
analysis has to be brief in the lexicographical workplace, since there are
usually demanding deadlines to be met. In the case of example 1 above, the
most appropriate translation may be devoted to (his family), or fond of if the
sense was "liked" rather than "committed to". In examples 2 and 3, where less
emotion is involved ("country life" and "power"), a less emotive word might
be used, perhaps 'fond of country life" and "attached to power". How many of
these translations are included in the final dictionary entry will depend on the
space constraints. If all three can be included, they will probably be
"discriminated" (differentiated) by their typical collocates or lexical fields. If
208 Jeremy Munday

only one is possible, then the lexicographer will mostly choose the one that
covers the widest lexical field, or, as Manley suggested, "is insertable in a wide
range of contexts". In our case, this may end up being attached to since it does
fit all three example sentences and is the most common of the three words,
even if it may not be the first-choice translation in all cases.

5.2 ST - TT pair comparison

The following collocation examples are taken from a larger ST-TT

comparative study (Munday 1997) of Edith Grossman's American English
translation Strange Pilgrims (SP) of a collection of Garcia Marquez short
stories, Doce cuentos peregrinos (GM). One idea was to see what shifts had
occurred in the process of even a close and accurate translation, and what this
might tell us about the "norms" adopted by the translator. The BNC was used
as a control corpus, against which to measure some of the individual lexical
items from the TT.
In general, there seems to be a trend for the TTs to calque the
collocation of the original Spanish, even when this produces a less normal TT
collocation. This is evident in "splendid fabric" (SP: 180), the translation of
"telas espléndidas" (DC: 236), with no examples in the English-language
corpora. The most typical and "insertable" collocations given by the statistical
packages would be beautiful or rich fabric.
Sometimes, however, the corpus supports the translator's intuition. The
rain, in Maria dos Prazeres, is described as "llovizna de vientos sesgados"
(DC: 137), translated as "oblique, wind-driven rain" (SP: 97). The use of the
latinate oblique in the TT seems intuitively a little formal, while the translator
has also explained the role of the wind by the addition of the past participle
"wind-driven". Collocation statistics show no examples of "oblique rain" but
four examples of "wind-driven rain", much to the surprise of the analyst. They
also throw up a variety of alternative and more statistically frequent
collocations of rain: pouring, torrential, driving, lashing, pelting and squally.
Perhaps "driving rain" or "lashing rain" would work well here, combining in
one adjective the two ideas conveyed separately by oblique and wind-driven.
This is not to say that wind-driven is a poor translation, merely that access to
collocation statistics could help the translator to assess possible translations a
little like an interactive thesaurus. A translator who had access to a corpus
might enter the search term rain or fabric and read down the list of frequent
collocates until he/she found one that fitted into the context of the translation.

6. Conclusion: The horizons of corpus linguistics

The description of previous work in this field has not only isolated different
 A corpus-base approach 209

techniques for the analysis of translation; it has also attempted to touch on

questions of how the role of the critic is being affected by working in a
computerized field. Much of the analysis is still based on counting lexical
items, since counting "is the least contestable mode of analysis" (Holmes
1994:87). Yet there are dangers in this. Certainly, the computer's supreme
ability to calculate individual word-forms runs the risk of focusing on a one-to-
one equivalence at word level, whereas professional translators normally work
with larger chunks of language. In a comparative analysis of a ST and its
translation, the mere act of checking also runs the risk of homing in exclusively
on intuitively "odd-sounding" collocations which are perceived to be translator
"errors" or "translationese": what sounds "odd" to the analyst is a subjective
judgement, hence the role of the control corpora in supporting or refuting that
intuition.
However, the control corpora, which need to be large and representative
in order to check collocations, can never be fully comprehensive. They deal in
typicalities and probabilities; if the collocation in question occurs infrequently
in the corpus, this may mean that it is improbable, but not impossible.
Moreover, literature often intentionally distorts norms and plays with words.
An unusual collocation in the TT may very well be thus because the ST
collocation was an infrequent form or even a novelty. For this reason, the ideal
would be to check the ST collocation first against a Spanish corpus and then to
compare the frequency of use of the original term against the frequency of the
translated term, checked against an English corpus. Although at the moment
the Spanish corpora are mainly too small to give great detail about the
collocations of less frequent words, a large and broad-based corpus is currently
being put together under the direction of the Real Academia. This should offer
much better comparative information in the future. Such developments also
mean that the methods of corpus linguistics will increasingly need to be
included in translation training courses if future translators are to make the
most of computerized language resources.

Note

1. My thanks to Oxford University Press ELT Dictionaries and COBUILD for allowing me
to access their electronic corpora and to use examples from them. Thanks also to Jonathan
Cape (Random House) for allowing me to hold the Garcia Marquez stories on disk for the
period of my research.

References

BAKER, M. 1993. "Corpus Linguistics and Translation Studies: Implications and

Applications". In M. BAKER, G. FRANCIS & E. TOGNINI-BONELLI 1993, 233-50.
210 Jeremy Munday

BAKER, M. 1995. "Corpora in Translation Studies: An Overview and Suggestions for Future
Research". Target 7(2), 223-43.

BAKER, M , FRANCIS, G. & TOGNINI-BONELLI, E. (eds). 1993. Text and Technology: In

honour ofJohn Sinclair. Amsterdam & Philadelphia: John Benjamins.

BURROWS, J. F. 1987. Computation into Criticism: A study of Jane Austen 's novels and an
experiment in method. Oxford: Oxford University Press.

HOLMES, D. I. 1994. "Authorship Attribution". Computers and the Humanities 28(2): 87-106.

MANLEY, J. 1985. "Processing of Excerpts for the Bilingual Dictionary". Symposium on

Lexicography II. Edited by K. HYLDGAARD-JENSEN & ZETTERSTEN, A." 245-54. Tübingen:
Niemeyer.

MUNDAY, J. 1997. Systems in Translation (PhD thesis). Bradford: University of Bradford.

MUNDAY, J. 1998. "A computer-assisted approach to the analysis of translation shifts". Meta
43(4), 542-56.

SCOTT, M. 1996. WordSmith (Computer program). Oxford: Oxford University Press.

SINCLAIR, J.M. 1991. Corpus, Concordance, Collocation. Oxford: Oxford University Press.

TOURY, G. 1995. Descriptive Translation Studies and beyond. Amsterdam & Philadelphia:
John Benjamins.

Main control corpora

(BNC) British National Corpus. http://info.ox.ac.uk/bnc/index.htrnl

(BoE) Bank of English corpus. http://titama.cobuild.collins.co.uk/boe_info.html

Illustrative texts

(DC) GARCIA MÁRQUEZ, Gabriel. 1992. Doce cuentos peregrinos. Madrid: Mondadori.

(SP) GARCIA MÁRQUEZ, Gabriel. 1994. Strange Pilgrims. (Translated by Edith Grossman).
London: Penguin UK.
MACHINE TRANSLATION FOR TRANSLATORS?

Anja Schwarzl
University of Salzburg, Austria

Zusammenfassung
In den letzten Jahren kamen vermehrt Übersetzungsprogramme auf den Markt, die auf
handelsüblichen Personalcomputern laufen. In diesem Aufsatz geht es um die
Übersetzungsqualität dieser sogenannten semi-professionellen Systeme. Kriterien
werden aufgelistet, die bei der Entscheidung, ob sich ein Text für maschinelle
Übersetzung (MÜ) eignet oder nicht, hilfreich sein können. Beispiele für MÜ werden im
Rahmen eines allgemeinsprachlichen Tests und eines Vorbearbeitungstests angeführt.
Ein weiterer Schwerpunkt ist die Rolle der MÜ in der modernen Kommunikation wie
WWW und e-mail.

Résumé
Ces dernières années, on a mis sur le marché des systèmes de traduction automatique
(TA) qui peuvent fonctionner sur des ordinateurs personnels standard. Notre
contribution s'interrogera sur les performances de ces systèmes semi-professionnels.
Nous chercherons à définir des critères pour savoir si on peut recommander la TA pour
certains textes. Nos exemples de réalisation en TA sont de trois sortes: des phrases de
langue générale, un texte technique pré-édité et des documents disponibles en ligne.

Resumen
Desde hace unos afios, existen sistemas automâticos de traducción compatibles con un
ordenador estándar. Este trabajo se centra en el funcionamiento de estos sistemas
denominados semiprofesionales. Este trabajo procurará establecer unos criterios para
determinar si se puede recomendar el uso de la traducción automâtica para ciertos
textos. Se muestran ejemplos del funcionamiento de traducción automâtica en una
prueba de lenguaje general y en otra de preedición. Un tercer tema de este trabajo se
centra en la importancia de la traducción automâtica en los nuevos medios de
comunicación tales como Internet y el correo electrónico.
212 Anja Schwarzl

1. What is MT?

This paper presents an overview of the current performance of semi-

professional machine translation (MT) programs and their usefulness for a
(professional) translator. First of all it has to be made clear what is understood
by MT. In the literal sense of the word, machine translation might refer to any
translation by a machine, so a CD-ROM dictionary lookup might be regarded as
machine translation as well. But the term machine translation is used only for
translation on something higher than word level, i.e. the program must have a
syntactic and semantic component as well. Then a further distinction is made
according to the level of automation:

1 MAHT: MACHINE ASSISTED HUMAN TRANSLATION, the (human) translator is

assisted by the machine, probably in the form of CD-ROM dictionaries, spelling
and style checks, etc. This type of translation is sometimes also referred to as
CAT: COMPUTER ASSISTED TRANSLATION.
2 HAMT: HUMAN ASSISTED MACHINE TRANSLATION; the translation is done by
the machine, but the process is halted occasionally and the computer prompts
for human input in order to solve ambiguities.
3 FA(HQ)MT: FULLY AUTOMATIC (HIGH QUALITY) TRANSLATION; the whole
translation process is covered by the machine. Apart from highly recommended
pre- or post-editing, no human assistance is required.

Most of the semiprofessional systems use a three-step-translation model

consisting of analysis, transfer, and synthesis. So-called first generation
programs use a direct method, whereas products of the second generation are
indirect systems. As the name suggests, the direct method means that the source
language is directly transferred into the target language. The necessary
components (modules) for such a system are (1) a bilingual dictionary, (2)
transfer rules, usually word-for-word substitution, and (3) generation rules for
the target language. Indirect systems use an intermediate representation of the
sentence structure as a result of the analysis process and as the basis for
generation. This intermediate representation can either be language dependent
(a transfer system) or independent (an interlingua system: see Hutchins &
Somers 1992). With an interlingua system, a maximum amount of source-
language analysis has to be carried out.

2. Semiprofessional systems

Other than the established, powerful, and hardware-intensive MT programs

(e.g. LMT, Logos, Systran and several Japanese systems espcially designed for
 Machine translation 213

English/Japanese), which are economical only for huge quantities of texts, in

recent years MT systems have been designed that can be run on DOS-PCs.
Their translation performance is way below the bigger (and at least 10 or 100
times more expensive) programs, but it may be sufficient for smaller translation
tasks. (Schmitt 1996:193; translation by A.S.)

This paper concentrates specifically on the so-called semiprofessional MT

programs, i.e. programs that can be run on standard home computers. This is
done on purpose, as the 'big' software - which is of course of better quality -
requires a hardware standard and/or has costs that are so high that the individual
user can hardly afford them. Only programs that are probably available in bigger
bookstores and are thus widely distributed are included here.
The four programs that have been used for this paper are:
1. Langenscheidt T1. This is bilingual (English/German/English - other
language pairs are also available). It comes with integrated LSP dictionaries,
which are hierarchically structured. If a text is defined by the user as belonging
to the domain of Computer Software, Tl will look for translation equivalents
first in the dictionary for software, then in the respective dictionaries for
Electronic Data Processing, General Technical Terms and finally in the
General Vocabulary. Translation problems are highlighted in colors: unknown
terms are red, compounds blue and ambiguities green. With a mouse-click on a
green word, the user opens a window with all the possible alternatives and can
thus select the proper one.
2. Globalink. The basic language of Globalink is English. On installing the
program, the user has the option of installing various other languages (German,
Spanish, Italian, French, Portuguese). All of these languages can be translated
to and from English only, a translation e.g. from Spanish to German is not
possible. It does not offer a structured dictionary, but uses semantic features
(such as: animated, country, day, family, human, literal, nationality, etc.) for
its lexical entries in order to establish a certain level of concord. The package
also includes a Conversation Utility for online chats, a Translation Utility for e-
mail translation and a WebPage translator for translating web pages. The latter
facility not only translates the text of a html-document, but also preserves the
links of the original document.
3. Pons Personal Translator Plus98. This is a bilingual program
(English/German/English - other language pairs are also available). It offers
several useful features. Before starting the translation process the user can scan
the text for lexical items that are not in the dictionary. They can then be added
before the translation, and errors, like assigning a wrong sentence structure
because of unknown words, are less likely. The dictionary can be edited and up
to 16 additional dictionaries, e.g. for a special type of source text, can be
created individually. Furthermore, the user can define various options for the
214 Anja Schwarzl

translation, such as assigning a text to a certain domain, or whether the program

is supposed to look for reference words for deictic expressions beyond sentence
borders, or whether the text is British or American English. The system also
offers a translation memory, i.e. a sentence archive. Another positive aspect of
this program is that it preserves the character and paragraph format of the
original text.
4. Transcend Euro Pack. This is a multilingual program, offering the same
languages as Globalink. The dictionary, however, cannot be edited or even
looked at. The text that has to be translated has to be put into the clipboard. For
the translation the user has to select the source and target language respectively
and press the start button. The original text in the clipboard is then replaced by
the translated text.
This study does not provide exhaustive coverage of the subject, but
merely seeks to show when and how such programs can be useful for the
professional tranlsator.

3. MT for me?

Whether MT is a useful application for the professional translator depends on

various factors. The basic criteria for efficient MT use could be listed as
follows: (1) required quality of the translation; (2) quantity of source text(s);
(3) quality of source text(s); (4) level of source text(s): average and maximum
sentence and phrase length; (5) required speed. (See Melby 1997.)
Re (1): the quality of MT output at its best reaches indicative translation
quality. It functions best with shorter and simpler sentences and less specialised
terminology.
Re (2): several MT programs can be 'trained', i.e. the dictionaries can be
modified/extended and previously translated phrases are stored as patterns for
upcoming translations (translation memory). This can be extremely useful when
you have a number of documents of the same type, e.g. in business
communication or for manuals.
Re (3): the source text has to be well written; any error in spelling,
grammar (and style) is bound to lower the level of the output.
Re (4): there are various formulae for calculating the difficulty level of a
text, e.g. Flesch, Gunning. As became obvious during this research these
formulas cannot be used to predict the output quality of a machine-translated
text, as 'difficulty level' is not the same for humans and computers. Reading
Ease Scores take into account the average sentence and word length, as longer
words are considered harder to understand, or more elaborate. For a computer,
word length does not matter at all, but sentence and phrase length have a
considerable influence on output quality.
 Machine translation 215

Re (5): the one positive aspect of MT at the moment is that it is fast.

100 English sentences from the British National Corpus (BNC) were translated
(~ 2,400 words / 184 lines / 5 pages) into German and Spanish and each process
took less than a minute with Globalink; Langenscheidt took about 4 minutes.
Interestingly, the other direction, i.e. back into English, took longer on both
systems.

4. Examples of MT output

This section reports on some small-scale tests carried out on the three MT
programs. Two kinds of test were done: (1) a test on general language using the
BNC, and (2) a kind of pre-editing experiment with a technical text.

4.1 Test 1

Sample sentences for general language were acquired from queries in the British
National Corpus (BNC), which is a 100-million word collection of samples of
written and spoken language from a wide range of sources, designed to
represent a wide cross-section of current British English, both spoken and
written (URL: http://info.ox.ac.uk/bnc/).
The query was done for English word light, because it is an everyday
word which nonetheless exhibits a lot of ambiguities. First it shows category
ambiguity: it can be a noun, a verb or an adjective. The adjective itself is
lexically ambiguous, as the opposite of both heavy and dark. The first 100
random sentences of the BNC query's result were used for translation. As three
of them showed too many characteristics of spoken English, i.e. ellipsis, fillers,
incomplete sentences, they were excluded, which left a total of 97 sentences.
It is remarkable that with the same program some categories that were
correct in one language were wrong in another, which may be regarded as
evidence of the use of a transfer and not an interlingua system. In all examples
here an asterisk (*) in front of a translation marks an incorrect translation.

(1) [...] may be equipped with light armour [...]

This was translated by the same program as
[...] werden vielleicht mit einem leichten Panzer [...] ausgestattet (correct)
*[...] puede equiparse con armadura de luz [...] (wrong)

However, in some cases the computer did remarkably well:

(2) to light a cigarette anzünden / encender

to make light of something leichthin abtun
216 Anja Schwarzl

In other cases, again, some versions were correct and others were not:

(3) to come to light (two occurrences)

zum Vorschein kommen I *kommen um zu beleuchten
* venir a encender (two occurrences)

Table 1 shows the outcome of this test, regarding just the category of the word
light, regardless of the quality of the rest of the output. The first column lists
the programs and the respective translation directions; the second and third
column give the absolute number of sentences, where light was translated
correctly or wrongly respectively; the last column gives the percentage of
translation mistakes.

Table 1. Category rating of light in German and Spanish translations

correct wrong % wrong

Globalink: Eng-Ger 86 11 10.67
Transcend: Eng-Ger 81 16 15.52
Langenscheidt: 87 10 9.7
Eng-Ger
Pons: Eng-Ger 83 14 13.58

Globalink: Eng-Sp. 86 11 10.67

Transcend: Eng-Sp. 86 11 10.67

4.2 Test 2

This test was basically a pre-editing experiment. Its aim was to find out how
sentence length effects the quality of the translation outcome. I took a
paragraph from a technical text in German (Eppinger & Herter 1993:84) and
had it rated by the Word97 ™ spelling and grammar check. Then I rewrote the
paragraph with the main aim of shortening the sentences while keeping the same
level of vocabulary and information. Below are the original and the pre-edited
text, both of them with the readability statistics from Word97 ™:

Text 2a (ORIGINAL):
1 Es existieren viele Verfahren, die diese Datenreduktion (von analog nach digital)
auf verschiedene Weise zu erreichen versuchen: Verfahren der Signalformcodierung
quantisieren und komprimieren ausschließlich das Sprachsignal, ohne auf die
charakteristischen Merkmale der Spracherzeugung, wie die unterschiedliche
 Machine translation 217

Generierung von Konsonanten und vokalen oder den Einfluß des Vokaltrakts,
einzugehen.
2 Verfahren, die diese Merkmale der Spracherzeugung mit einbeziehen, werden
Quellencodierungsverfahren genannt, und es lassen sich mit ihnen wesentlich niedrigere
Bitraten erzeugen, da das Codierverfahren dem Wesen des Sprachsignals angepaßt ist
und dieses somit besser erfaßt wird.
3 Eine Mischung beider Verfahren, die Verfahren der hybriden Codierung, versuchen
die Vorteile beider Verfahren miteinander zu verknüpfen.

Vocabulary (100 = very complex): 69.0

Syntax (100 = very complex): 60.0
Flesch (0 = very complex): 0.0
Words per sentence, average 33.3

Text 2b (PRE-EDITED):
1 Eine Datenreduktion ist erforderlich, wenn ein analoges Signal in ein digitales
Signal umgewandelt wird. Viele Verfahren versuchen, diese Datenreduktion zu
erreichen. Sie tun dies auf unterschiedliche Weise. Die Signalformcodierung codiert nur
das Sprachsignal. Sie geht nicht auf die typischen Merkmale des Signals ein. Diese
Merkmale sind z.B. die Generierung von Konsonanten und Vokalen. Ein weiteres
Merkmal ist z.B. der Einfluß des Vokaltrakts.
2 Es existieren auch Verfahren, die diese Merkmale mit einbeziehen. Diese werden
Quellencodierungsverfahren genannt. Damit können niedrigere Bitraten erzeugt werden.
Dieses Codierverfahren ist dem Wesen des Sprachsignals angepaßt. Das Sprachsignal
wird daher besser erfaßt.
3 Es gibt auch eine Mischung der beiden Verfahren. Dies nennt man hybride
Codierung. Sie verknüpft die Vorteile beider Verfahren miteinander.

Vocabulary (100 = very complex): 73.0

Syntax (100 = very complex) 8.0
Flesch (0 = very complex) 21.0
Words per sentence, average 7.66

The Flesch Reading Ease Score has a scale ranging from 0 (very complex) to
100 (very easy). It can be seen that the Flesch Reading Score and the Syntax
complexity changed considerably in the second text, whereas the complexity of
vocabulary remains almost the same. The first aim was to shorten the sentence
length while keeping the vocabulary standard. The second aim was to reduce
the frequency of deictic expressions (e.g. pronouns). The three underlined
words in each text were not found in the Word97™ dictionary. To illustrate the
result it will be sufficient to present just the respective translations of two
programs (Langenscheidt, Pons):
218 Anja Schwarzl

Langenscheidt: Text 2a
1 Many procedures exist, that this data reduction (of analogous according to digital)
in different way, attempt to achieve: Procedures of the waveform codification quantize
and pack exclusively the voice signal, without generation of consonants different to the
characteristic features of the voice generation as those and vokalen or the influence of
the vowel tract to come in.
2 Dealt, the these features of the voice generation with may include, are source
coding techniques mentioned, and it can be made itself bit rates lower with them
considerably since the enciphering procedure is adapted to the being of the voice signal
and consequently this is registered better.
3 A compound of both procedures, the procedures of hybrid codification, do the
advantages of both procedures attempt with each other to associate.

Langenscheidt Tl: Text 2b (pre-edited)

1 A data reduction is required if an analogous signal is converted into a digital signal.
Many procedures attempt to achieve this data reduction. They do this in different way.
The waveform codification only enciphers the voice signal. She does not deal with the
typical features of the signal. These features e.g. are from consonants and vowels
generation. A further feature e.g. is the influence of the vowel tract.
2 Procedures also exist, that these features with may include. These are mentioned
source coding techniques. In this way, lower bit rates can be made. This enciphering
procedure is adapted to the being of the voice signal. Therefore, the voice signal is
registered better.
3 There is also a compound of the two procedures. Is called hybrid codification this.
She associates the advantages of both procedures with each other.

Pons: Text 2a
1 It many methods, the this data reduction ( exist of analogous after digital ) try to
reach in different way: Or the influence of the vowel section spends the signal form
coding of Quantisieren in travelling and exclusively compresses the speech signal,
without on the characteristic features of the speech generation, as the various generation
of consonants and vocal, come in.
2 methods which include these features of the speech generation are called methods of
source encoding and have fundamentally lower bitrates it produced with them for
themselves since the method of encoding is adapted to the nature of the speech signal
and this is included therefore better.
3 Methods, the methods of the hybrid coding, more twoly try to connect the
advantages of both methods together for a mixture.

Pons: Text 2b (pre-edited)

1 A data reduction is required if an analogous signal is changed into a digital signal.
Many methods try to reach this data reduction. You do this in various way. The signal
form coding codes only the speech signal. She doesn't come in on the typical features of
 Machine translation 219

the signal. E.g. these features are the generation of consonants and vowels. E.g. a
further feature is the influence of the vowel section.
2 Methods which include these features also exist. These are called methods of source
encoding. Lower bitrates can be produced with that. This method of encoding is adapted
to the nature of the speech signal. The speech signal is included therefore better.
3 This one also gives both methods a mixture. One calls this hybrid coding. She
connects the advantages of both methods together.

It is very obvious that in both cases the translation of the second text (2b) is
much better, though still far from perfect.

5. MT and CMC (Computer Mediated Communication)

Communication horizons are expanding continually. There are more and

more participants in electronic communication, and this communication is
nevertheless getting faster and faster. But here we face a great paradox:
telecommunication has not solved but highlighted problems that arise from
cultural and language barriers. The delays and high costs of human translation
that used to be accepted are no longer bearable. Time-consuming human
translation is felt to undermine the progress of technology. The former 'barriers
to world-wide communication', which were cost, poor quality, long distances
and languages, have changed considerably. With modern technology, the only
barrier left is language (see O'Hagan 1996: 5).
"The European market for online services will be the fastest growing in
the world over the next two years" (PC Online, May 1996). This of course
implies that universities and big companies - whose staff usually have a good
knowledge of English - will no longer be the main users of the Internet. The
main user will increasingly be 'Everyman' - and this is one of the main
differences from 'standard' media, which are locally restricted and can thus
easily be designed for their intended audience.
More and more Web developers will thus (have to) turn their English-
only sites into multilingual resources. In fact, a term meaning 'creating
multilingual websites' already exists: localization. This can be done in two ways.
The first possibility implies multilingual databases, meaning that various copies
of each document are produced, in different languages. This what is done
nowadays, "Click here to read this document in English" - this solution
consumes disk space (as all the documents have to be stored somewhere), but
delivery and downloading are very fast.
The second possibility is localization via online translation, which means
that documents are written and stored only in one language and are translated
online when downloaded. This occupies very little disk space, as only one copy
220 Anja Schwarzl

of each document is stored, but delivery takes considerably more time because
of MT. This would be the only option for e-mail communication.
In both cases, however, texts and documents have to be translated in a
way which is 'as effective as possible'. The most obvious difficulties (apart from
MT problems) are deciding which languages should be made available;
differences of character set (even among European languages, let alone others);
representation conventions (e.g. date, time, certain symbols and images); and
the necessity of multilingual aligned hypertext. At least today, the solution via
MT is certainly out of the question, or only a temporary alternative. As 'written
proof you might like to check the much-discussed Starr Report on the
AltaVista Homepage (http://starr-report.altavista.com). A warning notifies the
user that the translations of the document have been done without human
intervention and that therefore the quality of these translations may not fully
capture the precision of the original legal text. Additionally, MT would only be
worth it - and thus only be provided - for large language groups, in Europe e.g.
for English, French, Spanish and perhaps German. But CMC will probably be an
(economic) pusher for MT research and development.

References

EPPINGER, B. & HERTER, E. 1993. Sprachverarbeitung, p. 84. Wien: Carl Hanser Verlag.

HUTCHINS, W.J. & SOMERS H.L. 1992. An Introduction to Machine Translation. London:
Academic Press.
LAUER, A. et al. (eds) 1996. Übersetzungswissenschaft im Umbruch: Festschrift für
Wolfgang Wilss zum 70. Geburtstag. Tübingen: Gunter Narr.
MELBY, A. 1992. "The Translator Workstation." Computers in Translation. A practical
appraisal ed. by John NEWTON, 147-165. London: Routledge.
MELBY, A. & WARNER, T. 1995. The Possibility of Language. A Discussion of the Nature
of Language, with implications for Human and Machine Translation. Amsterdam &
Philadelphia: John Benjamins.
MELBY, A. 1997. "Should I use Machine Translation?" Web-Document; URL:
http: //www.ttt.org/theory/mt4me. html)
O'HAGAN, M. 1996. The Coming Industry of Telecommunication. Topics in Translation 4.
Clevedon: Multilingual Matters.
SCHMITT, P.A. 1996. "Computereinsatz in der Translation". LAUER, A. et al., 187-196.
ASPECTS OF UNIVERSAL GRAMMAR IN HUMAN
VERSUS MACHINE TRANSLATION

Martin Kaltenbacher
University of Salzburg, Austria

Zusammenfassung
Viele Unterschiede zwischen Sprachen lassen sich auf eine konträre Auswahl von
binären universalgrammatikalischen Optionen zurückfuhren. Die strukturellen
Unterschiede zwischen Deutsch und Englisch, die sich durch drei unterschiedliche
Parameter ergeben, werden beschrieben, und vier Gruppen von Englisch als
Zweitsprache-Lernern sowie zwei Übersetzungprogramme werden auf korrekte
Übersetzung hin getestet. Universalgrammatikalisches Wissen ist nicht in
Softwarealgorithmen integriert, was in maschineller Übersetzung zu einer Reihe von
Fehlern führt.

Résumé
Nombre de différences entre les langues sont dues aux sélections différenciées dans les
options binaires disponibles en grammaire universelle. On décrit ici des différences
structurales entre l'allemand et l'anglais, selon trois paramètres distincts. Les
traductions de quatre groupes d'apprenants en anglais seront comparées à celles
produites par deux logiciels de TA. La non-intégration du savoir de grammaire
universelle dans les calculs informatisés mènent à divers types d'erreur en TA.

Resumen
Muchas de las diferencias entre las distintas lenguas se basan en contrastar selecciones
entre las opciones binarias distintas que se encuentran en la gramática universal. En
este articulo se describen las diferencias estructurales entre el alemán y el inglés que
tienen su origen en tres parâmetros distintos. También se han realizado y evaluado
pruebas de traducción correcta con cuatro grupos de estudiantes de inglés como
segunda lengua y con dos programas de traducción automática. Se producen errores de
distinta indole en la traducción automâtica ya que los conocimientos de gramâtica
universal no se integran en los algoritmos del software.
222 Martin Kaltenbacher

1. Universal grammar and its role in machine translation

According to Chomskyan theory all human languages share a common set of

abstract rules and principles called universal grammar. This set of rules is part of
a general language acquisition device that human beings are genetically supplied
with. It provides us with an innate language learning programme enabling us to
learn our native language as quickly and proficiently as we do (Chomsky 1965:
47ff.). Within universal principles languages vary along the lines of so-called
parameters. These account for a wide range of structural differences among
individual languages (cf.Cook & Newson 1996, Towell & Hawkins 1994). The
strength of this theory lies in the proposition that many differences between
languages can be explained by different selections of binary options available
within the common set of universal rules. This means that many syntactic
differences in languages can be tracked down to a limited number of individual
selections of binary options: e.g., from the parameters P1, P 2 ... Pn language L1
chooses the options P1+, P 2 -, ... Pn+ whereas L2 selects the settings P1-, P2+, ...
Pn-.
In other words the grammar of a language equals the total sum of a
limited number of either/or options. What this means for computerization, e.g.
in translation software, is easy to explain: as soon as all principles and
parameters can be isolated and adequately explained, many syntactic differences
between a given pair of languages would become computable, thereby also
making translation from any given source language into any given target
language (TL) an easy task for translation software. Since language is the
output of a process of binary option selection, all the software designer would
have to do is to include this process of option selection into the software
algorithm. However, over the years universal grammar has turned out to be far
more complex and abundant than originally expected. Neither the total number
nor the actual configuration of many of its principles and parameters have yet
been fully determined and many of them remain controversial. Moreover,
software engineers have not yet acknowledged the range of advantages that a
grammatical theory based on binary operations, such as the principles and
parameters theory, has to offer them.
This paper looks at three different parameters and some of the structural
differences between German and English as generated by the individual
parameter values. According to current theory, diverging parameter values are
one of the most common reasons for errors both in spontaneous foreign
language production as well as in written text production, such as translation
(see White 1989, Kaltenbacher 1997). Of course, such errors do not normally
occur in professional translations carried out by proficient speakers of the target
as well as the source language. They are, however, quite common in translations
produced by second language learners and in machine translation (MT).
 Universal grammar 223

Four groups of learners of English as a second language were tested and

their results were compared with the translations generated by two so-called
semi-professional machine translation programs. The computer programs
(Globalink 6.0, hereafter GL and Langenscheidt Tl 3.0, hereafter LS) run on
standard PCs and can be purchased at a price of around 300-400 German
marks. The four groups consisted of a group of 27 11-12-year-old pupils at an
Austrian grammar school with two years of formal English education; a group
of 41 pupils, aged 13-14, with four years of formal English training; a group of
28 pupils aged 15-16 with six years of formal English education; and a group of
seven adults who had taken part in a two-year evening course of English at an
adult training centre. The test-sentences were all kept very simple as far as
syntactic structure, morphology and vocabulary are concerned. This should
ensure that the universal phenomena tested constitute the only problem area for
translation in the individual sentences.

2. Adjacency condition on case assignment

The first parameter tested is the so-called adjacency condition on case

assignment. For a long time this had been considered to be a parameter of its
own, but Pollock (1989) has pointed out that it appears to be embodied within
the parameter of verb movement. The English parameter value [+adjacency]
requires the complement within a VP to follow its head (the governing verb)
immediately, as in (la) and (2a). No adverbial must occur between them, as can
be seen from the ungrammaticality of (lb) and (2b):

(1) (a) He quickly entered the room.

(b) *He entered quickly the room.
(2) (a) I went to school by bike.
(b) *ƒ went by bike to school.

The German parameter value [-adjacency] allows verbal complements to

follow their heads, but they can also be separated from them by adverbials of
any kind. Consequently, the German equivalents (3) - (4) to the four sentences
above are all perfectly acceptable.

(3) (a) Schnell betrat er das Zimmer.

(b) Er betrat schnell das Zimmer.
(4) (a) Ich fuhr in die Schule mit dem Rad.
(b) Ich fuhr mit dem Rad in die Schule.
224 Martin Kaltenbacher

In such a case the parameter setting of English is said to generate a 'subset' of

the possible German structures. Therefore, German-speaking learners of L2-
English can be expected to produce a significant number of ungrammatical
translations in English, assuming that they transfer their L1-parameter value into
the target language. The following five German test-sentences (5-9)
incorporating a VAO (verb-adverbial-object) word order were presented to the
test-persons for translation. The same sentences were translated by the two
computer programs GL and LS.

(5) Ich werde in Wien ein neues Haus kaufen.

I will buy a new house in Vienna.
GL: I will buy a new house in Vienna.
LS: I will buy a new house in Vienna.
(6) Mike spielt mit seinem Bruder Tennis.
Mike plays tennis with his brother.
GL: Mike plays tennis with his/its brother.
LS: Mike plays tennis with its brother.
(7) John spielt manchmal Piano.
John sometimes plays the piano.
GL: John sometimes plays piano.
LS: John sometimes plays piano.
(8) Er betrat schnell das Zimmer.
He quickly entered the room.
GL: He/it crossed the room quickly
LS: He entered the room fast.
(9) Sie putzte sorgfältig ihre Schuhe.
She carefully cleaned her shoes.
GL: She/it cleaned her/its/their shoes carefully.
LS: It cleaned its shoes carefully.

The 2nd-year learners produced most TL-violations of all learner groups

here. In test-sentences (6) and (8) they produced more violations than correct
translations, the preferred structures being the word-by-word translations *Mike
plays with his brother tennis and *He entered quickly the room. The adult
group also produced quite a large number of TL-violations in test-sentence (8),
although they were very accurate in the other test-sentences. Naturally, groups
with more years of formal English training produce more target-like
translations. The results rendered by MT suggest that the relevant software has
been programmed to master this aspect of universal grammar. Both programs
appear to be capable of determining clause elements like verb, object and
adverbial, and of applying the word-order changes necessary to arrive at the
proper English structures. But appearances may be deceptive. Further, more
subtle tests with the two programs convey an altogether different picture.
 Universal grammar 225

(10) Er sah letzte Nacht einen Film.

He saw a film last night.
GL: *He/it saw last night a film.
LS: He saw a film last night.
(11) Mike sah diese Nacht einen Film.
Mike saw a film this night.
GL: *Mike saw this night a film.
LS: *Mike saw this night a film.
(12) Mike kaufte diesen Morgen ein Auto.
Mike bought a car this morning.
GL: *Mike bought these mornings a car.
LS: !Mike bought a car for these tomorrow.

Apparently the parsers integrated in the software are capable of

identifying noun phrases (NPs) {letzte Nacht (10), diese Nacht (11), but they
cannot always determine their syntactic functions within the clauses (see Arnold
1994). They appear to follow the general analytic strategy of assigning the
function object to the first NP after the verb, no matter whether it is an object
or an adverbial. As soon as the adverbial is represented by a prepositional
phrase (PP), as in (13) and (14) or by an adverb phrase (15) and (16), it is
correctly analysed as such and consequently the necessary word-order changes
are applied:

(13) Er sah in der letzten Nacht einen Film.

He saw a film last night.
GL: He/it saw a film in the last night.
LS: He saw a film in the last night.
(14) Mike kaufte an diesem Morgen ein Auto.
Mike bought a car this morning.
GL: Mike bought a car in this morning.
LS: Mike bought a car on this morning.
(15) Er sah heute nacht einen Film.
He saw a film tonight.
GL: He/it saw a film tonight.
LS: He saw a film tonight.
(16) Mike kaufte heute morgen ein Auto.
Mike bought a car this morning.
GL: Mike bought a car this morning.
LS: Mike bought a car this morning.
226 Martin Kaltenbacher

3. Head position parameter

The second parameter to be examined is the head position parameter. Here

German deviates from the more common parameter settings. Generally, the
head position within a phrase is either to the left [head initial] or to the right of
its complement [head final] across all phrasal categories. In English the
parameter value is set to [head initial], so a complement will always follow its
governing head to the right, as in a play by Shakespeare (NP), in the garden
(PP), quick as the wind (AP (adjective phase)), bought a car (VP).
German has also selected a [head initial] parameter value, but only for the
phrasal categories of NP, PP and AP . Within VPs the head position is assumed
to be final, which is best reflected in the position of finite verbs in subordinate
clauses: ein Drama von Shakespeare, im Garten, schnell wie der Wind, ...als er
ein Auto kaufte (when he a car bought > 'when he bought a car').
As regards the head position in VPs, the relevant parameter values for
English and German exclude each other (cf. Towell and Hawkins 1994: 79ff.;
Cook and Newson 1996: 218ff.). In such a case translation errors are usually
not to be expected at an intermediate or advanced learner level. The structural
differences between the two languages are so obvious that the target structure is
normally acquired at a very early stage of L2-acquisition. In MT errors should
not occur either, for it should be easy to program the software to account for
the different verb positions in English and German. Both assumptions are
confirmed by the translations of test-sentences:

(17) Ich kann nicht zu deiner Party kommen, weil ich mir den Fuß gebrochen
habe.
I cannot come to your party because I have broken my foot.
GL: I cannot come to your party because I have broken myself the foot.
LS: I can not come to your party because I burst the foot for myself.
(18) Weißt du, wann er morgen kommen wird?
Do you know when he will come tomorrow?
GL: Do you whiten, when he/it will come tomorrow?
LS: Do you know when he will come tomorrow?
(19) Er sagt der Polizei, was er über den Mord weiß.
He tells the police what he knows about the murder.
GL: He/it tells the police, what he/it knows about the murder.
LS: He tells to the police what he knows via the murder.
(20) Es tut mir leid, daß du einen Unfall hattest.
I am sorry that you had an accident.
GL: I am sorry that you had an accident.
LS: It does me suffer that you had an accident.
 Universal grammar 227

The learner groups as well as the computer programs make the necessary
adjustments to the verb position. Among the total number of learner translations
there is only one case of TL-violation, in which the learner produced *I am
sorry that you an accident had for (20). Surprisingly, the number of missing
translations is quite high even in the more advanced test-groups. A possible
explanation for this avoidance strategy could be the greater length and higher
structural complexity of hypotactic sentences. At first sight translation software
again seems to be able to reset the parameter value appropriately. But the two
programs get into trouble when translating sentences that involve NP-movement
such as wh-extraction, as in (21):

(21) Wer meinst du, hat das Geld gestohlen?

Who do you think has stolen the money?
GL: *Who do you think, has the money stolen?
LS: *Who mean you, the money stole? (LS)

If the test-sentence is restructured as a declarative sentence consisting of a

matrix main clause and an embedded main clause functioning as the object of
the matrix clause {Du meinst, er hat das Geld gestohlen), both programs
translate correctly. The corresponding interrogative form {Meinst du, er hat das
Geld gestohlen?) is also translated correctly.
Apparently GL and LS are both programmed to treat English as an SVO
language, but when they cannot find the subject of the embedded clause after it
has been extracted from its original position, they do not know where to put the
verb, as no trace is left. The complexity of test-sentence (21) made it rather
difficult for the learner groups to translate, too. Several subjects offered no
translation, but of course a computer cannot do this. The strategy it employs if
it cannot analyse a source text sufficiently is to take recourse to a so-called
'robust strategy' (Schmidt 1998: 133). It then tries to translate the sentence on
the basis of incomplete information. Usually this means translating the source
text word by word and leaving the individual clause elements in the same
positions as in the source text.

4. Pro-drop parameter

The pro-drop parameter setting for German and English is identical: [- pro-
drop]. Among other phenomena (Chomsky 1981: 140), this implies that
sentences in both languages require an overt subject either in the form of a
referential subject or in the form of an expletive pronoun {Es regnet 'It is
raining'). (Haider 1993; Quirk et al. 1985.)
228 Martin Kaltenbacher

Test results showed that although there is no difference between the L1

and the L2 parameter setting here, both learners and the two programs have
difficulties with the expletive pronouns.

(22) Letztes Jahr waren sehr viele Unfälle.

Last year there were very many accidents.
GL: *Very many accidents were last year.
LS: *Very many accidents were last year.
(22) 1975 gab es kein BSE.
In 1975 there was no BSE.
GL: 1975, there was no BSE.
LS: *1975 gave it no BSE.

In test-sentence (22) the human translators, even those from the more advanced
test-groups, quite often fail to insert the necessary expletive there. Instead, they
produce Last year were many accidents, which simply reflects the German
sentence structure. This violates the English subject before verb requirement,
which in this case demands the expletive there. The computer programs apply
an SV word order in (22), but their output is not grammatical either. Sentence
(23) exhibits a similar problem area for L2-learners, although the human
translators performed better here than in (22), probably because school curricula
provide for particular training on the German phrase es gibt and its English
counterpart there is. For MT the phrase es gibt seems to be particularly difficult
to translate. The two programs produce contradictory results. Neither of them
manages to produce 100%-correct translations of the phrase in varying
contexts. Test-sentence (23) is translated correctly by GL, whereas LS produces
a word-by-word translation of the German sentence structure. In other contexts,
however, the two programs give different results. More subtle tests have shown
that MT solutions can be affected by the presence of morphologically
ambiguous forms, as in test-sentence (24), where GL parses der Straße as a
dative:

(24) Am Ende der Straße gibt es ein nettes Restaurant.

GL: *It gives in the end the street a nice small restaurant.
LS: At the end of the street, there is a nice small restaurant.

Tests with other sentences of this kind, using acronyms instead of nouns, also
revealed inconsistencies in the programs.
 Universal grammar 229

5. Conclusion

The mastery of universal rules and the options they impose on different
languages is one of the keys to correct translation. One of the reasons for MT's
limited efficiency in coping with universal grammar, or with syntactic structure
in general, can be identified in the limited significance that was assigned to the
importance of abstract universal principles in the early days of MT. Schwarzl
(1998: 116) emphasizes that "for a long time linguistics has been left out in the
general MT research ... MT researchers consider[ed] themselves engineers and
the often abstract linguistic theories were only of limited practical use". Human
translators are innately endowed with knowledge about universal principles, and
gradually reset their L1-parameter setting to the appropriate L2-setting along
the route of foreign language acquisition. In MT-software such knowledge is
only partially integrated in syntactic analysis and generation modules. One of the
main deficiencies of MT is the inability of the software to assign the proper
syntactic functions to different grammatical forms. Human translators can easily
determine the syntactic functions of phrases within a sentence owing to their
understanding of context and their knowledge of the world. Hubmayer (1998:
133) points out that human translators are subconscioulsy aware of many facts
that for them 'go without saying', whereas MT must have recourse to actual
words implanted in a lexical database. MT seems to master some of the
processes of syntactic shift required by diverging parameter settings between
two individual languages. When an error is produced its origin is not always of a
purely syntactic nature. Very often it is the ambiguity of morphological or
grammatical forms that prevents MT from correctly analysing the syntax of the
source sentence. These shortcomings will have to be reconsidered and rectified
by both linguists and software engineers in co-operation. How this can be
achieved, however has to be left on the agenda.

References

ARNOLD, D., BALKAN, L., LEE HUMPHREYS, R., MEJER, S. & SADLER, L. 1994.
Machine Translation. An Introductory Guide. Oxford: Blackwell Ltd.

CHOSMKY, N. 1965. Aspects of the Theory of Syntax. Cambridge, Massachusetts: MIT

Press.

CHOMSKY, N. 1981. Lectures on Government and Binding. Dordrecht: Foris Publications.

COELSCH-FOISNER, S. & W. GÖRTSCHACHER (eds), Text and Context: Essays in

English and American Studies in Honour of Holger M. Klein. Rheinfelden: Schäuble Verlag
230 Martin Kaltenbacher

COOK, V. & NEWSON, M. 1996. Chomsky's Universal Grammar. An Introduction. Oxford:

Blackwell Publishers Ltd.

HAIDER, H. 1993. Deutsche Syntax - generativ. Vorstudien zur Theorie einer projektiven
Grammatik. Tübingen: Gunter Narr Verlag.

HUBMAYER, K. 1998. "Möglichkeiten und Grenzen der maschinellen Übersetzung". S.

COELSCH-FOISNER& W. GÖRT SCHACHER (eds), 129-143.

KALTENBACHER, M. 1998. "Parameter Resetting in Second Language Acquisition: a Myth

or Reality?". S. COELSCH-FOISNER & W. GÖRTSCHACHER (eds), 171-182.

POLLOCK, J.Y. 1989. "Verb Movement, Universal Grammar, and the Structure of IP".
Linguistic Inquiry 20. 365-424.

QUIRK, R., GREENBAUM, S., LEECH, G. & SVARTVIK, J. 1985. A Comprehensive

Grammar of the English Language. London: Longman Group Ltd.

SCHMIDT, P. 1998. "Automatisches Dolmetschen". Handbuch Translation ed. by M.

SNELL-HORNBY, H. HONIG, P. KUßMAUL, & P. SCHMITT. Tübingen: Stauffenburg
Verlag.

SCHWARZL, A. 1998. "Machine Translation and the Difficulty of Evaluation". S.

COELSCH-FOISNER & W. GÖRTSCHACHER, (eds), 115-127.

TOWELL, R. & HAWKINS, R. 1994. Approaches to Second Language Acquisition.

Clevedon: Multilingual Matters Ltd.

WHITE, L. 1989. Universal Grammar and Second Language Acquisition. Amsterdam &
Philadelphia: John Benjamins.
 PART VI

Studies of a text type

DOCUMENTARY FILM TRANSLATION:
A SPECIFIC PRACTICE?

Eliana P.C. Franco

KUL, Leuven, Belgium

Zusammenfassung
In der Forschung audiovisueller Übersetzung kommt generell der Übersetzung von
Dokumentarfilmen eine untergeordnete Rolle zu. Dieser Artikel stellt fest, daß der
Kategorie von Dokumentarfilmübersetzung eine eigene Übersetzungsspezifik
zugrundeliegt, der mehr Beachtung zu schenken ist. Den Ausgangspunkt für die
Betrachtung der Dokumentarfilmübersetzung bildet das Genre, das wichtige
übersetzungsaktbezogene Fragen aufwirft, welche sich von Fragen zur Übersetzung von
fiktiven Texten unterscheiden. Die Untersuchung der Praxis vom
Dokumentarfilmübersetzen zeigt, daß diese Übersetzungsaktivität nicht in dem Ausmaß
von objektiver Handlung geprägt ist, wie oft angenommen. Als Erklärung dafür dient,
daß das übersetzerische Verhalten primär kulturgebunden definiert wird.

Résumé
Les documentaires n'ont jamais eu une grande place dans les travaux touchant la
traduction audiovisuelle. Pourtant, selon l'auteur, ils mériteraient une attention
particulière. D'abord comme genre - ce qui soulève nombre de questions quand on en
vient à leur traduction. Ensuite comme genre appelant un type particulier de transfert,
loin d'être aussi objectif qu'on le pense mais toute traduction du fait même qu'elle
prend place dans une culture, n'exige-t-elle pas interprétation?

Resumen
La traducción de documentales suele ocupar un lugar poco privilegiado dentro de la
investigación en traducción audiovisual. En esta presentación se pretende demostrar que
esta actividad constituye una práctica especifica que merece una atención especial. En
primer lugar, se analiza el tema desde una perspectiva de género, lo cual implica varios
aspectos importantes a tener en cuenta durante el proceso traductor que difieren de los
aspectos encontrados en los productos de ficción. Después, se analiza el tema desde una
perspectiva practica, revelando asi que la traducción de documentales no es tan objetiva
234 Eliana Franco

como habitualmente se supone puesto que el comportamiento traductor se encuentra

defmido, en primera instancia, por una cultura.

1. Introduction

Before answering the question addressed in the title of the paper, the corpus
upon which my answer is based should be defined and justified. The corpus
consists of documentaries about Brazil, produced by West Europeans, and
broadcast by cable television channels in Belgium. The decision to investigate
documentaries focusing on my home country/culture (Brazil) came about as a
result of my observation that what Belgians - and other western and Central
Europeans - seem to be interested in, and what they know, about Brazil is no
longer "expected" topics such as football and carnival, but has to do with a very
specific issue: Brazilian street children.
Two years of videotaping (from November 1995 until December 1997)
led to some interesting findings about programmes in general and documentaries
in particular, such as: (a) programmes about Brazil were frequently broadcast (a
total of 80 programmes were recorded over two years); (b) in terms of
production, the vast majority of these programmes were produced by West
Europeans (85% of the total); (c) in terms of topics discussed, the most
frequent one (27% of the programmes) was the issue of "the street children", a
trend that justifies Europeans' current interest in Brazil and also confirms the
social relevance of such programmes (this issue featured in about 27% of the
documentaries recorded on Brazil); and last but not least, (d) most of the
documentaries (66%) made use of a hardly explored audiovisual mode of
transfer, namely "voice-over".
A preliminary investigation of the 22 documentaries which constitute the
corpus indicated that, in general terms, translating documentaries can be viewed
as a specific practice mainly because of aspects considered typical of the genre,
which differ from those involved in the translation of feature films (either
subtitling or dubbing). Many of the translation decisions, however, seem to be
determined not by genre-specific factors but by the host culture. The following
brief account of some of the aspects involved in documentary film translation
suggests that the activity remains to be explored by researchers.
 Documentary film translation 235

2. The nature of the material

The material to be translated in documentaries usually consists of local

interviews and sometimes of commentaries too, if the audiovisual product is
broadcast by a second channel and in a language different from that of its origin,
or when the product is broadcast by the bilingual channel ARTE.
Documentaries belong to the category of "serious television
programmes" because these are about "reality ". As a consequence, viewers of a
documentary about another country/culture immediately assume that they are
receiving factual, truthful information through local images, reliable local
testimonies and accurate commentaries, and of course, that the translation is as
objective as possible.
It is well known, however, that current thought within
Film/Communication/Cultural Studies (see e.g. Winston 1995) and Critical
Discourse Analysis (Fairclough 1992, 1995) challenges the "authenticity" of
documentary information, attributing to it notions of reality construction or
representation. One of the direct consequences of such an approach is to
suggest the active participation of translation in this process of
construction/representation, therefore challenging its objectivity.
But despite the massive attack on the documentary's implicit claim of
impartiality, the role of documentary translation remains almost unquestioned by
scholars, suffering from a kind of marginalization within audiovisual translation
research, as can be confirmed by simply comparing the number of studies
devoted to the factual genre as opposed to those devoted to fiction. Out of the
1241 entries that constitute the latest edition of Yves Gambier's Language
Transfer and Audiovisual Communication Bibliography (1997), only 21 refer
explicitly to the non-fictional genre; among these, six refer to "documentaries"
and 11 refer to the "voice-over" mode of transfer commonly used in these
programmes. Such low figures seem to result from two main facts: first, the
strong tradition of literature, which has certainly influenced audiovisual
translation researchers' preference for the fictional genre and consequently for
the modes of transfer mostly found in this genre (i.e. dubbing and subtitling);
second, what I call a "false belief' among researchers, that "translating facts is a
straightforward, non-problematic activity". Traditionally claimed to be
objective, deprived of the artifices of literary language or cinematic invention,
documentaries could be nothing but a boring research topic.
236 Eliana Franco

3. The translation models

Unlike scripts for the translation of fiction, documentary scripts are very often
post-production texts, a fact which is a consequence of the regular procedure
adopted by documentary filmmakers in which shooting precedes scriptwriting.
Scriptwriting of factual material, therefore, may mean either the transcription of
the original from the videotape before actual translating or straightforward
translation from the tape.
Thus, the important point to be recognized is that, contrary to fiction,
the videotaped material acquires a pivotal position in the translation of
documentaries, being the source text from which original transcriptions and/or
first-translated versions derive. So, in the case of a German documentary where
interviews with Brazilians are translated directly from the videotape, when this
documentary is imported by France the available models for the French
translator are the original tape and the German written version. If a transcription
of the original is also available, the number of candidates - i.e. written source
models at the disposal of the translator of the French version - consequently
increases.

4. Modes of transfer

As already suggested, the study of translated documentary films also leads us to

the "discovery" of a mode of transfer which is specific to the genre: the "voice-
over". (This is not the case for some Central and East European countries, such
as Poland, where the voice-over mode is commonly used in television output in
general.) Because documentaries have raised so little interest in translation
research, the few accounts of its typical mode prove quite confusing, owing to
inconsistent terminology as well as incongruent definitions. The origin of the
problem seems nevertheless to lie in the hardly acknowledged fact within
Translation Studies that the term "voice-over" was actually borrowed from its
predecessor field, namely Film Studies, and that the term's meaning does not
entail the translation activity.
Despite terminological divergences, there is at least some consensus
about the implicit function of this mode which, like subtitling, provides a kind of
"authenticity illusion" through the simultaneous presence of the original
counterpart. Moreover, the conventional procedure of making seconds at the
beginning and at the end of original utterances audible to the viewers (cf.
Luyken 1981:80) assures them that "what is being said is exactly what is being
told". Nevertheless, unlike subtitling, "voice-over" is an oral mode of transfer,
and unlike dubbing, it is not lip-synchronized because its primary function is not
mimetic but objective reproduction, as required by the genre.
 Documentary film translation 237

5. Documentary translation behaviour as a cultural practice

As a consequence of the general language policy adopted by different West

European communities/countries, the voice-over translation of documentaries is
mostly evident in those communities/countries where dubbing of fiction is the
prevailing norm. These are mainly the French and German channels available on
the Belgian cable television. Out of the 22 documentaries of the corpus, a small
sample of three programmes was selected for the observation of translation
practices. Besides discussing similar topics (street/disadvantaged children), each
programme had two language versions (French and German) that were
broadcast by the same channel (ARTE) (see references).
The idea was to observe whether, despite the language, the translated
versions followed similar practices which could point to the active role of the
broadcasting institution as regulating such practices. The analysis, however,
revealed recurring differences between the French and German versions, which
led to the conclusion that the target culture seems to be more important in
defining practices than the institution itself.
As for any sample, observations derived from it should not be viewed as
conclusive, but just as first impressions or hypotheses which will be compared
with the analysis of other documentaries in a later stage of my research. What
follows is a summary of the main differences concerning the orality and
objectivity of French and German voice-over translations of three
documentaries broadcast by ARTE.

5.7 Orality

Regarding orality as expressed by voice, gender, stress and intonation, French

versions proved much more mimetic than their German counterparts, except for
the reproduction of stress, which was poorly exploited in both languages. In all
French versions Brazilian speakers were given voices which matched not only
their sex but also their tone of voice as defined by age. Apart from one
documentary (Carlinhos Brown), where the number of speakers was quite
limited, translation readers in the two other German versions always performed
the roles of more than one Brazilian speaker. In Hundert Morgen Paradies,
for example, there are only two translation voices for the many participants in
the documentary.
One could argue that this procedure is purely economically motivated in
view of the large number of participants, which would otherwise require a large
number of translation voices. Nevertheless, considerations about the concept of
a factual genre as a "non-mimetic genre" could also have meant that translated
content would be seen as more relevant than translation performance. The rare
or absent translation of music observed in German versions may also strongly
238 Eliana Franco

support "relevance according to genre" as a guiding principle of these versions.

Although songs are a constant in Carlinhos Brown and Surfavela, these are
never translated in the German versions but frequently translated in their French
counterparts. It is worth mentioning, however, that songs are generally
subtitled.
An interesting related procedure observed in the German versions of
documentaries in which local interviews are punctuated by original
commentaries is the frequent conversion of the interviewees' answers into
indirect speech and the automatic insertion thereof into the commentary. Along
with the economy and relevance motivations suggested above, this translational
procedure could be viewed from the perspective of Film Studies as being
rhetorically motivated, that is, as a way of enhancing the authority of the
commentator.
With respect to orality as expressed by the interaction between visual
and verbal information there were also divergences between the French and
German versions of the same documentary. Whereas the former tended to keep
the interactive relationship between word and image established by the original,
the latter tended to convey visual output verbally, providing images with a
purely illustrative function. As a consequence, the verbal output of the German
versions became denser, giving the impression that they "translated more" than
their French counterparts. A good example is found in one of the first scenes of
the documentary Carlinhos Brown, which portrays the composer and his
musical project with street children. In this scene the speaker Pintado do Bongó
- Carlinhos Brown's master in the art of playing instruments - conveys meaning
basically through images. Here we find Pintado, Carlinhos, a man and a woman,
all holding different instruments. The French and the German texts read as
follows (information between parentheses refers to images; pauses are indicated
by dots):

(la) La musique, elle EST ! Je suis venu au Candeal...la musique crée à

travers moi, avec les timbales, les bongos, et tout ce qui est...battements de
main...ça a commencé par là... (image of Pintado beating the sticks of his
drums against each other) ensuite avec ça... (Pintado now plays the drum)/ et
puis... (he points to CB who starts playing another instrument; then he points to
the man who starts playing his instrument, and then to the woman who starts
playing her instrument) c 'est parti... (now all play and sing together).
'Music, it IS ! I came to Candeal...music comes out of me, with the timbals, the
bongos, and everything that is...clapping...it [music] began here...(image of
Pintado beating the sticks of his drums against each other) afterwards with
this...(Pintado now plays the drum) and then...(he points to CB who starts
playing another instrument; then he points to the man who starts playing his
 Documentary film translation 239

instrument, and then to the woman who starts playing her instrument) it all
starts...(now all play and sing together)'

(lb) Die Musik kommt aus meinem Innern. Ich kam nach Candeal und machte
Musik. Die Musik lebte schon immer in mir. Für mich ist alles Musik -
Trommeln, Bongos, das Klatschen und das Schlagen der Stöcke. Das
Klatschen ist der Rhythmus damit fängt alles an. Und dann setzen die
Trommeln ein.
'The music comes from my inner self. I came to Candeal and made music. The
music has always lived in me. For me everything is music - drums, bongos,
clapping and the beating of the drum sticks. Clapping is the rhythm from which
everything takes off. And then come the drums'.

Comparing both translations we notice that the German version is a continuous

text which not only verbalizes images but also avoids the rhythm of the
speaker's discourse through the restructuring of sentences, providing the viewer
with a version which resembles written much more than spoken language. The
relevance of genre as expressed by objective, clear information seems again to
be the prevailing principle.

5.2 Objectivity

Surprisingly enough, the German versions did not prove very objective as far as
the meaning of foreign/exotic information was concerned. The German texts
contained some semantic shifts which seem to have been motivated by two
distinct factors. One is the cultural/linguistic gaps which exist between German
and Brazilian Portuguese. For example, in Surfavela, a Portuguese production
about a Brazilian social project which aims at integrating street children into
society through surfing, the surfers' jargon becomes the language of the
documentary. It is characterized by specific vocabulary, slang terms (many of
them derived from English) and idiomatic expressions. Owing to the absence of
a "surfing culture" in Germany (where the word Surfing generally refers to
windsurfing) and the consequent lack of vocabulary and expressions to support
the translation, the transfer of meaning became partly impaired and the German
version of Surfavela resulted in a text that was difficult to understand in the
recipient culture.
Secondly, some semantic shifts in the German translations seem to have
resulted from deliberate attempts to deviate from the original intended message.
One example can be found in Carlinhos Brown, and the comparison of the two
language versions (with French reproducing the main idea of the original) makes
my point clear:
240 Eliana Franco

(2a) Quand je crée un groupe comme Lactomia avec des enfants, je pense à
quoi ? A cette arnaque autour des enfants de rue du Brésil.
'When I create a group like Lactomia with the children, I think of what? Of this
farce about the street children of Brazil.'

(2b) Ich habe die Gruppe Lactomia für die Kinder Brasiliens gegründet. Ich
möchte ihnen helfen, denn die Stras senkinder haben hier ein sehr hartes
Leben.
'I created the Lactomia group for the children of Brazil. I wanted to help them
because the street children have a very hard life here.'

By deviating from the original message, the German translation provides it with
a negative perspective about the source culture which reveals the opposite of
what the speaker is trying to convey to his foreign audience. It is fair to suggest
that such a behaviour was ideologically motivated.
Interestingly enough, the few instances of semantic shifts observed in the
French versions seem to have resulted from a clear attempt to eliminate bias.
These are mainly found in the German production Hundert Morgen Paradies,
where commentary and questions put to interviewees often take on an
evaluative tone throughout the documentary, which aims to emphasize the poor
living conditions of landless families, especially the children. Let us have a look
at an example:

(3 a) "Ihr schlaft doch nicht alle auf diesem schmalen Bett? "
'You don't all sleep in this small bed?'

(3b) "Vous dormez tous dans ce petit lit?"

'Do you all sleep in this small bed?'

Attitude is conveyed here through the emphatic negative (doch nicht) in the
German interviewer's original question (3a). The utterance then becomes more
an expression of the interviewer's indignation about the interviewee's living
situation than an actual question that needs to be answered. By omitting the
negative, the French version completely eliminates the evaluative tone of the
utterance.

6. Conclusion

In this brief account of documentary translation I have outlined some of the

aspects which suggest that the activity should be viewed as a specific practice.
Firstly, there are considerations about the documentary genre - such as factual
material, post-scriptwriting and mode of transfer - which give another
 Documentary film translation 241

dimension to audiovisual translation in terms of objective, model and

expression, at the same time moving away from the stigma that "audio-visual
translation appears to be essentially synonymous with subtitling or dubbing"
(Mailhac 1998:208).
Secondly, examples of documentary translation in practice show that
voice-over versions may be influenced by the receptor culture's own
interpretation of the concept of objective rendering. The fact that simultaneous
different language versions of the same documentary may mean "different texts"
in terms of information content and expression reinforces the idea that
documentary translation plays an active role in the representation of the
documentary reality within a given culture.
Thus, the translation of "real life" as seen on television is not as obvious
as we tend to assume, and its objectivity should no longer be taken for granted
in audiovisual translation research.

References

FAIRCLOUGH, N. 1992. Discourse and Social change. Cambridge: Polity Press.

FAIRCLOUGH, N. 1995. Media Discourse. London/New York: Edward Arnold.

GAMBIER, Y. (ed.) 1997. Language Transfer and Audiovisual Communication. A

Bibliography. Turku: Centre for Translation and Interpreting. First edition: 1994.

LUYKEN, G. M. & al. 1991. Overcoming Language Barriers in Television. Dubbing and
subtitling for the European audience (Media Monographs no. 13). Manchester: The European
Institute for the Media.

MAILHAC, J.P. 1998. "Optimising the linguistic transfer in the case of commercial videos".
Translating for the Media (Papers from the International Conference, Berlin, November 22-
23, 1996) ed. by Yves Gambier. Turku: Centre for Translation and Interpreting, 207-223.

WINSTON, B. 1995. Claiming the Real The documentary film revisited. London: British
Film Institute.

Documentaries

Hundert Morgen paradies. Brasiliens landlose nehmen sich ihr Recht (1996), a German
production (1996) by Lourdes Picareta and Philip Siegel, broadcast by WDR on 14/05/96 and
by ARTE on 10/05/96.

100 arpents de terre de paradis, French version broadcast by ARTE on 10/05/96.

242 Eliana Franco

Carlinhos Brown (1996), a French production by Claude Santiago, broadcast in French and
German by ARTE on 26/10/96.

Surfavela (1996), a Portuguese production by Joaquim Pinto and Nuno Leonel, broadcast in
French and German by ARTE on 05/11/96.
TRANSLATION AND MASS CULTURE

Kirsten Malmkjær, Middlesex University, UK

John Milton, Universidade de Sao Paulo, Brazil
Veronica Smith, Universität Klagenfurt, Austria

Zusammenfassung
Übersetzungen fur den Massenmarkt stellen ein Problem für die
Übersetzungswissenschaft dar. In diesem Artikel wird eine Analyse der
Gegenwartsszene der Massenmarkt vorgenommen und betrachtet, wie Übersetzung in
der Realität "gemacht" wird: Angefangen wird mit einem allgemeinen Überblick über
die Übersetzung für Massenmarkt anhand von Buchklub-Übersetzungen in Brasilien
(JM). Nach dem Überblick folgt eine Untersuchung darüber, wie ein dänischer
Erfolgsautor eine Massenleserschaft auf dem englischsprachigen Markt erreichte (KM).
Anschlieend wird eine Untersuchung über die Naturalisierung der Kinder- und
Jugendliteratur für den deutschsprachigen Markt vorgestellt (VS).

Résumé
Les traductions pour la culture de masse pose un certain problème à la traductologie.
Nous nous proposons de voir comment est réalisée la traduction pour le marché de
masse, en étudiant trois exemples distincts: les traductions pour un club de livres au
Brésil (JM), la trajectoire de l'auteur danois P. Høeg devenu célèbre dans le monde
anglophone (KM), et la façon dont la littérature pour enfants est adaptée, naturalisée
pour les marchés germanophones (VS).

Resumen
Las traducciones para los mercados de masa constituyen un cierto problema para los
Estudios de Traducción. Este articulo hace un balance de la situación actual del
mercado de masas para ver cómo se realiza realmente la traducción. Se utilizará para
ello un estudio general de la traducción para el mercado de masas, utilizando como
ejemplo traducciones realizadas para un club de libros en Brasil (J.M.). Después se
hizo un seguimiento con estudios de cómo un conocido escritor danés de grandes éxitos
consiguió llegar a un amplio publico en el mercado angloparlante (K.M.), y cómo se
naturaliza la literatura infantil para el mercado alemân (V.S.).
244 Malmkjær, Milton & Smith

1. Characteristics of the translation of mass fiction (J.M.)

The illustrations in this section are taken from studies into the translations of the
Clube do Livro book club and the the translation of classic fiction in Brazil
(Milton 1995).

1.1 Team translation

Rather than being the work of an individual, the condensed or adapted

translation, or for that matter the dubbed or subtitled film, or the translation
made within industry, will be the work of a team. It is a mere part of the
assembly line. The name of the "translator" may not appear on the work. If it
does, it may be a pseudonym: a highbrow translator may not wish to have his or
her name associated with the work, or it may even be an invented name for a
team.
Historical parallels can be found in medieval translation, where
adaptation, omissions, retellings, alterations, etc., were a normal part of what
we now call translation. Edgar Morin (1977) believes that the concept of the
creator essentially belongs to the nineteenth century. In mass culture the
producer in some ways revives the old collectivism of epic poems, or of the
workshops of famous painters such as Rafael and Rembrandt. Translation teams
or "factories" are not at all new: they were already producing translated novels
in the 18th and 19th centuries, translating popular French novels in many
countries.

1.2 Standardization

Standardization, or Fordism, is an important factor in the production of factory

novels and translations. Different forms of standardization can be found: a)
theme: the work is tailored to suit the tastes of the reader; b) language: sub­
standard language and dialect is cut out; c) style: the work should not deviate
from a strict narrative style; d) size: the Brazilian Clube do Livro publications
were standardized, after 1960, to 160 pages; e) weight: a low weight to cut
postal costs will be an important economic factor in publishers that depend on
postal services.

1.3 The role of the author

Commercial production ignores the so-called sacredness of the author. Walter

Benjamin's well-known essay "The Work of Art in the Age of Mechanical
Reproduction" (1992) emphasizes the fact that the contemporary mechanical
possibilities of reproducing the object will change our relationship to the work
 Translation and mass culture 245

of art and destroy the ritualistic and magical elements surrounding the original.
With film and photography there is never a single original. The very nature of
both arts is essentially reproductive: films and photographs cannot be collected
as paintings can. Moreover, the cost of a film is so high that it must be made
available to the highest number of people possible. Indeed, the nature of film is
highly democratic as it will make all myths and great authors and artists
available to cinema-goers.
The fragmentation of a film is very different to the unity of a painting or
a poem. The director, responsible for the final product, will depend on the
expertise of many other different areas, sound, photography, etc., of which he
will have only a limited knowledge. He will be more of a manager of the final
product. Compare this with the painter or novelist, who will have all elements
of the work of art under their control. The final result of the film or photograph
will come about as the result of the combination of a number of very different
elements. Edgar Morin points out that in spite of this standardization, the
culture industry does not sell soap powder, and that within a certain framework,
there is a demand for originality. New products must always be invented. This,
for Morin, is the "dynamic contradiction" of the culture industry (Morin
1977:28).
Though Benjamin's essay concentrates on the reproducible arts of film
and photography, his essay also helps us to understand commercial translation,
where the editor will coordinate the work of original author, translator,
copydesk, illustrator and finance department.

1.4 Commercial considerations

Definite commercial strategies will be used. The translation will be directed to a

definite market. There will be markets for different segments of society. Non-
condensed, carefully produced translations will be directed towards markets
from a higher social class. (Adapted) translations of Pride and Prejudice and
Wuthering Heights are often directed towards a market of female readers.
Similarly, Moby Dick, Kidnapped, Huckleberry Finn and Gulliver's Travels are
usually directed towards the juvenile market.
Deadlines are all-important. The product must be released on time even
though it may have a few flaws, particularly if is not a product which is aimed at
an erudite market. Meeting deadlines is much more important that perfectly
accurate copy.
A study by Anikó Sohár (forthcoming) on the translation of mass fiction
in Hungary has shown considerable remnants of the translators' own notes and
queries remaining in the final version, showing that absolutely no revision took
place. For example, empty lines and the translator's notes, e.g., "(error in
original)" remain in the published translations! Similarly, in the Brazilian Clube
246 Malmkjær, Milton & Smith

do Livro study, very simple spelling mistakes were found, especially in the
names of authors: Virginia Wolff, Charlotte Bronté, George Elliot, and Kunt
(sic!) Hamsun.

7.5 Choice of texts

The great novel tradition, particularly the 19th-century novel, is particularly

reusable in different shapes and guises: condensations, films, cartoons, etc.
Many of the most popular authors — Swift, Dickens, Jane Austen, Scott,
George Eliot, the Brontès, Balzac, Melville, Stevenson — were already read by
mass audiences. The 19th-century realist tradition was much closer to popular
taste than modernism. By contrast, the 20th-century modernist tradition of the
novels of Joyce, Faulkner, Lawrence and Virginia Woolf hardly lend themselves
to adaptation. The 19th-century texts join other standard mythical texts which
have been used and reused in a variety of forms: The Iliad and The Odyssey,
Faust, Shakespeare's tragedies, Oedipus, Don Quijote, etc.

1.6 The recycling strategy

It may often be cheaper to recycle an already existing translation than to

commission a new one. The same novel, in the same translation, may even be
directed towards different markets. Thus we see the same (or slightly modified)
translation appearing in different guises. Ediouro published two almost identical
condensations of Pride and Prejudice in 1970. The small format edition,
"translated" by Nair Lacerda, is slightly longer and its language is slightly more
formal than the large-format illustrated edition "translated" by Paulo Mendes
Campos and aimed more at the juvenile market.
As Viagens de Gulliver a terras descorthecidas was published in a
heavily cut version by Ediçoes Cultura in 1940, "Portuguese translation by
Henrique Marques Junior, scrupulously revised and modernized". "Scrupulously
revised" is a euphemism for "enormous cuts were made". An identical
translation, except for a few uses characteristic of the Portuguese of Portugal
which were brazilianised, was published as Viagens de Gulliver by Jackson in
1957. This time the translator was Cruz Teixeira!
The Clube do Livro often relies on previously published translations.
Both the Clube do Livro (1953), here a double edition with smaller print, and
the Ediçoes Cultura (1943) translation of Ivanhoe seem to be heavily calqued
on the Gamier translation of 1905. Similarly, the Clube do Livro translation of
Silas Marner (1973) is a slightly updated version of the Martins version (1942).
Even the misspelling of the author's name (George Elliot) is copied!
 Translation and mass culture 247

1.7 Marketing strategies

Commercialism is not beyond a few tricks — the reader may not be treated with
all that much respect. Translations were generally introduced as special
translations, a euphemism for "contains many cuts".

1.8 The pseudotranslation

A common marketing trick is that of introducing an original work as a

pseudotranslation. The most famous pseudotranslation is Don Quijote, which
Cervantes pretended to be a translation from the Arabic. Cervantes had much
more freedom to make the satire of romances de cavalaria if he pretended it
was a translation. Macpherson' s Ossian was one of the greatest of literary
hoaxes. The Scottish poet, James Macpherson, pretended to have discovered
and have translated Celtic poems from the 4th Century AD. For a while
Macpherson was one of the most celebrated figures in the literary world.
Most frequently, the reason for pseudotranslations is commercial. Anikó
Sohâr (forthcoming) documents "The Mystery of Wayne Mark Chapman - a
case of Fictitious Translation in Hungary". Wayne Mark Chapman is the name
given to the fictitious author of a series of best-selling science fiction novels in
Hungary written by a team of authors. In post-communist Eastern Europe,
popular fiction from the West has considerable prestige. Science fiction written
by Hungarian authors would not be attractive to potential purchasers.
Obviously, therefore, it pays to invent the foreign author, the original work, and
even give biographical information on him. Wayne Chapman lives in Concord,
New Hampshire (maybe a literary joke), and his first books, Blood Season, and
Banners, were published by Pengdragon Publishing Co. Inc., London. Sohâr,
after a considerable amount of literary detective work, managed to discover that
these novels were in fact pseudotranslations and that neither the author nor the
publishing house existed. In addition, after the series began to be successful, the
editors took fewer pains to insist that these books were originally written in
English.

1.9 The value of book club translations

Such translations are ephemeral, throwaway, not catalogued by libraries and

information networks. Jerusa Pires Ferreira had little luck in her search for O
Livro de Sao Cipriano in national libraries (Ferreira, 1993: xxi). Indeed a
librarian in the Bibliothèque Nationale in Paris was annoyed and said that it was
unthinkable that this kind of work should be found in the national library of
France!
248 Malmkjær, Milton & Smith

2. Høeg at home and abroad (K.M.)

There has been a steady stream of Danish literature into English since the early
19th century, but relatively few authors have reached the British (or American)
best seller lists. The most recent writer to do so is Peter Høeg (b. 1957).
When Høeg published Forestilling om det tyvende Àrhundrede
('Vision/performance of the 20th century', 1988), it was likened by reviewers to
the magical realist work of Allende and Garcia Marques, and its ironical tone to
that of the older, socially critical Danish novelist, Hans Scherfig (1905-1979).
It was first published in the UK in 1996 as The History of Danish Dreams,
translated by Barbara Haveland.
Høeg's second publication was a collection of short stories, Fortællinger
om natten ('Tales about the/at night', 1990; translated as Tales of the Night).
Critics note the similarity to Karen Blixen and the shock effect in the short
stories which is certainly retained in De måske egnede ('The perhaps suitable',
translated as Borderliners) published in Denmark in 1993 and in the US in
November 1994. This book accuses an array of well-known, contemporary
educationalists, many of whom are still living, of having ruined the lives of very
many children in their care, but it is also a more general attack on the Danish
establishment's way of dealing with difference and failure. It is not, however, the
book that initially made Høeg's name in English.
That book was Frøken Smillas fornæmmelse for sne (1992), published
in the United States, translated by Tiina Nunnally as Smilla's Sense of Snow, in
1993. Høeg wanted alterations made in the translation before publication in
Britain, but Nunnally refused to let the altered translation go out under her
name. Consequently, the name F. David was invented for the version published
in Britain as Miss Smilla 's Feeling for Snow (Guido Waldman, personal
communication). Smilla has been described as a thriller, as the Danish answer to
Jules Verne, as a crime/science fiction novel, and as a socially critical fable of
suspense. It has now been made into a film.
Kvinden og aben ('The Woman and the Ape') was published on 1 April
1996. This book has been seen as reflecting the animalism of our times, a
critique of modern civilisation. Critics have noted echoes of The Jungle Book,
King Kong, and Planet of the Apes.

2.1 Høeg and the concept of (in) visibility

The Danish editor at Farrar Straus and Giroux, Høeg's American publisher, is
reported (Korse 1994) to have said that what is special about Høeg is that his
style is at once Danish and international; and it is possible that the combination
of this mixture of Danishness and international-ness with Høeg's subject matters
is a major factor in making him a best seller in English. This raises some
 Translation and mass culture 249

questions about the troubled notion of the translator's invisibility (Venuti 1995).
This notion is not simple, but part of what Venuti seems to mean by it is a kind
of fluency. Venuti also wants to link the question of visibility versus invisibility
to the question of foreignising versus domesticating (1995:21): "Anglo-
American culture ... has long been dominated by domesticating theories that
recommend fluent translating." He claims that in the postwar period, when
modernism was marginalised in English-language translation, "Most translators
chose a fluent, domesticating method that reduced the foreign text to dominant
cultural values in English" (1995:225), and he seems to hold that this trend is
still in force.
However, Høeg's texts in English are not inscribed with dominant
English values, at least not obviously so. They are made fluent in English to
some degree, but there are plenty of markers of Danishness in the language, in
particular in the names of persons and places. Other markers of linguistic
foreignness are transferred from the originals, with respect to which they are
also foreign (though perhaps less so in Danish than in English in the case of
names from Greenland). But the fact that cultural difference is marked up in the
translations, which are therefore (and for other reasons) visible as translations,
may actually be more a domesticating feature, than the opposite (see below). So
there is some difficulty in placing Høeg with respect to the oppositions Venuti
sets up; and it may also be the case that Venuti's concepts are rather too broad
and imprecise to be used for analytic purposes.
Better, perhaps, to approach the work with something like Lefevere's
notion of a typology of changes (1992: 61) in mind, and try to determine which
adjustments to the receptor culture have been made in the translations, and
which foreign elements have been retained, and then try to decide whether these
can be said to serve domestication or foreignisation, and if they can, how and in
what sense. Such a typology might include the translation-oriented categories of
Additions, Omissions and Changes, because it is by adding, omitting and
making changes over and above those demanded by the basic translation
process that translators domesticate or foreignise. There are, however, no
significant omissions, so far as I am aware, in the Høeg texts.
The typology should probably also include Setting as a category because
of its obvious connection with culture, and Innovation because it often amounts
to the type of generic change which, as Lefevere (1992: 23) points out, is
desirable when there is a perceived need for it "in the environment of a literary
system for that system to be or remain functional" in the sense of having "an
impact on the environment by means of the works it produces, or the rewritings
thereof'. But it is obvious from what was said above about the reception of
Høeg's work that he has been considered innovative, so no further use will be
made of this category of analysis.
250 Malmkjær, Milton & Smith

2.2 Additions

2.2.1 Extratextual
The Høeg translations, but not the originals, offer some information about the
author. This highlights the foreignness of the writer and may therefore be called
foreignising; but it constitutes domestication insofar as it conforms to a British
norm for book presentation. The translation of Smilla includes, on the page
immediately before the narration begins, a map of Copenhagen which is
presumably intended to help the reader follow Smilla around the town and
which obviously highlights the foreign setting for the story. The History of
Danish Dreams has a historical setting and includes historical characters. Many
of these are listed at the end of the translation in a "Biographical List of Real-
Life Characters". Obviously, the list highlights the foreignness of the setting,
and so does the paragraph which introduces the list, which mentions Danes and
Denmark repeatedly.

2.2.2 Intratextual
F. David, in Smilla, inserts explanatory phrases like "the children's book" in
front of "Rasmus Klump on the Ice Cap"(p. 12), the title of a book which every
Dane would recognise as a children's book. It could be argued that the mere
fact of having to categorise a book is foreignising, but in general this strategy as
employed by David promotes transparency and fluency, for example when she
inserts "his summer home" in front of "on Klitvej in Skagen" (Target Text p.
16; Source Text p. 24, my translation):

TT Here they haven't put in a tennis court. But not for lack of space. It's
probably because Loyn has a a couple of them in his back garden in Hellerup,
and two more at his summer home on Klitvej in Skagen.

ST In here they haven't made a tennis court. But not for lack of space. It is
probably because Loyn has a couple in his back garden in Hellerup, and two
more on Klitvej in Skagen.

F. David's explanation of a feature of Esaja's speech seems to be an attempt at

domesticating something which cannot help but be foreignising. The only way
to avoid a sign of foreignness here would be complete re-writing, or leaving out
the passage altogether (TT p. 46; ST p. 54):

TT "Do you," he said, addressing me formally in Danish, "think that I could

sleep here?"

ST - Do you [De] think, he said, - that I could sleep here?

 Translation and mass culture 251

This clearly highlights the book's Danish origin.

2.3 Changes and retentions

2.3.1 Punctuation
Where Høeg uses punctuation sparingly, particularly in Borderliners and in The
History of Danish Dreams, to create a run-on, stream-of-consciousness effect,
Haveland's translations normalise the punctuation to a degree - tidy it up, to
make for a somewhat calmer, more controlled, conventional reading, and also
to make it conform more closely to English norms. This promotes transparency,
and is domesticating.

2.3.2 Names
Haveland retains almost all Danish names, though she translates some names of
institutions. These translations clearly provide some information about the
nature of the organisations, but it may seem foreignising, and so, of course, is
the retention of the Danish names, which is the primary strategy used in all the
translations. F. David adjusts the name of the child Esajas, who becomes Isaiah,
in accordance with English Biblical tradition, but not with Greenlandic Biblical
tradition, which is Danish based. Obviously, this is an instance of acculturation
towards English.

2.4 Setting

According to The Observer, Borderliners "strikes at the heart of Danish pride

at the benevolence and good sense of the country's institutions ..." (quote from
the back cover; translation by K.M.). Notice that the British newspaper's
comment refers specifically to Denmark. This book, though shocking, is
acceptable in translation partly because its powerful social criticism is explicitly
of an aspect of Danish life and only implicitly about schooling, control, lust for
power, and liberation through rebellion in general. The Danish setting distances
the readership from the disturbing treatment of some widely relevant, topical
issues. This seems interesting, because it amounts to a kind of reverse
domestication.
The same phenomenon can be observed in the case of The History of
Danish Dreams. The book begins in Danish, "This is the Danish dreams'
history", and the English title is obviously drawn from this beginning. The
change of title serves two purposes. Least importantly, perhaps, it frees the
translator from worrying about the troublesome, ambiguous noun forestilling.
More interestingly, it localises the novel: it announces explicitly that it is about
Denmark.
252 Malmkjær, Milton & Smith

Høeg ties his fictions to history - primarily Danish history - both

through the use of historical events and characters, and through extensive
literary allusion. To have made them transparent would have required total
geographical and historical reinvention, and, having eschewed that option, the
translators have, in a sense, been absolved from having to make any overall
decision about whether to foreignise or whether to domesticate. The books are
and remain foreign, and this, of course, offers Høeg innovation on a plate with
respect to English culture, where Danish history and literature are not in the
forefront of most readers' minds. However, given Høeg's topics, this is an
advantage: the overtly foreign settings ensure that the innovation will not
offend. And in this lies a good measure of covert domestication.

3. Naturalising a national icon: Enid Blyton's boarding school stories

(V.S.)

Apart from its entertainment value, children's literature is nowadays considered

to fulfil an important role in the acquisition of reading maturity. In addition to
the underlying aim of broadening young readers' vocabulary and training their
ability to deal with more complex syntactic structures than they meet in
everyday conversation, it often represents a source of moral and didactic
principles. Studies on the translation of children's literature tend to focus on the
fact that the young reader may not yet be able to cope with different cultures
and environments. An extreme form of domestication or rather naturalisation
seems to be the order of the day, so any potentially puzzling detail is filtered out
and the norms of the target culture rather than of the source culture are
adopted. (For a critical discussion, see Oittinen 1995.) As Puurtinen (1994:84)
found: "Children with their imperfect reading abilites and limited experience of
life are not expected to tolerate as much strangeness and foreignness as adult
readers." While in a more polemical vein, House (1997:131) asks,

Why do translators of children's books feel licensed to change as they see fit
instead of providing the children with access to the original? [...] Is it possible
that children in their intelligent and imaginative capacities to learn and be
exposed to the strange world of the original are largely underrated?

Why should this be the case? And do the Enid Blyton stories to be discussed
here confirm these findings? A superficial reading of many of the children's
books available in translation might lead to the conclusion that the main stock-
in-trade of children's literature is the exciting story rather than the creative use
of language. This of course also applies to many mass-media translations where
translators are explicitly told to translate freely, as Smith and Klein-Braley
 Translation and mass culture 253

(1985:133-4) discovered when a set of publisher's translation guidelines for the

translation of romantic fiction came into their hands. 1 As even the story-line can
be changed, one wonders why publishers actually bother with translations at all.

3.1 Enid Blyton and the reception of her writing

These criticisms of translation practice have to be placed fairly in the context of

the writer chosen for analysis. Enid Blyton was one of the most prolific writers
of children's literature and is still one of the best-selling children's authors in
English, although she has been criticized on the grounds of sexism, racism and
linguistic impoverishment. Blyton has been translated into 40 languages. Her
popularity among children is said to stem from her ability to put herself in the
situation of the children she describes, reflecting their interests and anxieties, as
well as to write a gripping story. Good and bad, right and wrong, sensible and
silly are all clearly differentiated in her writing but without drowning the sense
of fun. When approaching the task of translating Blyton, these are the qualities
which should be reproduced. In addition, strategic decisions on how to deal
with cultural aspects have to be made, especially when the setting of the stories
and its concomitant associations play a key role in creating the overall
atmosphere: in the case in point the institution of the British single-sex boarding
school.
The two six-volume series of school stories considered here, St. Clare's
and Malory Towers, were mainly written during the 40s. The schools described
represent an elite microcosm sheltered from events going on in the world
outside, where mainly upper-class girls develop their personalities and discover
their vocations, leaving school with career aspirations and plans for the future,
which do not include merely slotting into the wife-and-mother role. By tackling
problems independently without too much help from adults, the girls' characters
are moulded, so that by the end of the series even the most incorrigible snobs
see the error of their ways. The tediousness of school routine is alleviated by
midnight feasts and pillow fights, and jokes played on the teachers, who are on
the whole good-natured and understanding.
The books were not translated into German until the 60s and 70s.
Various translators worked on the books and were actually named in the early
editions (Schneider Verlag). Later editions make no acknowledgement of the
translator, attributing the "adaptation" (Bearbeitung) to the publisher. This
appears to be less a strategy for emphasising the translator's invisibility than to
disguise the fact that the publisher now produces German originals based on the
Blyton characters: Blyton's original six-volume series (in German: Hanni und
Nanni and Dolly) now comprise 18 or 19 books each.
254 Malmkjær, Milton & Smith

3.2 The problems of cultural adaptation

Two main problems arise here in translation: firstly, how to transfer the
culturally-saturated setting of the English boarding school, and secondly, how
to update the text for contemporary readers. It is conceivable that the timing of
the translations into German, still relatively close to the end of the Second
World War, actually influenced the translation strategy. Otherwise it is hard to
imagine why the brown school uniform of Malory Towers should be changed to
dark blue! On the other hand there is evidence that the decisions taken were not
really based on any principled consideration of the setting; a consistent strategy
of domestication has not been adopted: the setting is naturalised. Blyton
located Malory Towers in a deserted part of Cornwall and has her chief
protagonist travel from London to get there. The translators omit references to
these real places but go on to translate details of the setting so literally that it
turns entirely into a fantasy landscape, since no German-speaking country offers
a similar topography.
An aspect which causes considerable confusion is the way sports are
dealt with. In Malory Towers one of the main sporting activities is the game of
lacrosse, played outdoors on a field with a stick. Despite the fact that this game
is not actually played by very many English native-speaker readers and so may
be as exotic for them as for German-speaking readers, the translators decide to
opt for a hybrid strategy of translating references to the game as Tennis or
Handball or Turnen (Gymnastics). (In the examples below, MT = Malory
Towers, SC = St. Clare's, HN = Hanni und Nanni, D = Dolly; numbers refer to
the volume in question.)

"I can only find one of my lacrosse boots." SC2.

"Und ichfindenur einen Turnschuh." HN2.

The problems this confused strategy leads to come to a crux in the fifth volume,
where a whole chapter is devoted to the description of a lacrosse game. From
the very start, the German version fails to mention which sport we are dealing
with (not even calling it handball), so we can only deduce from passing
references what is going on. But this is easier said than done. Tests done with
various readers show that during the reading process, football, handball and
basketball are all called to mind, only to be discarded when the sentence "wurde
ihr den Ball aus dem Netz geschlagen" ('the ball was forced out of her net') is
reached, leaving the readers completely in the dark. This passage shows what
can result when the ramifications of the options available are not followed
through consistently.
 Translation and mass culture 255

Throughout the books, the time-honoured strategy of omission is

adopted. The result is a bland storyline without the details of description and
dialogue which explain and justify the characters' actions. One of the important
features of boarding-school life is the ethos which drives the community of
schoolgirls and teachers. Passages containing statements of the shared values
tend to be left out as, for example in MT5:

"Shows what Malory Towers does to you!" said Darrell. "Still, I suppose any
good boarding-school does the same things - makes you stand on your own feet,
rubs off your corners, teaches you common sense, makes you accept
responsibility. (p. 110)

"Das zeigt nur, wie wohl uns Möwenfels tut!" meinte Dolly. (D5, p. 103).

Or in MT2:

"I'm going to the post with this for Daphne. But don't go and split on me, Gwen.
I know it's against the rules." Gwendoline stared at Mary-Lou in surprise. "You
breaking the rules!" she said. "I don't believe you ever did that before. You're
mad to think you can get to the post and back in time." (p. 126).

"Ich gehe für Diana zur Post, um das Päckchen aufzugeben", erwiderte Marlies.
Evelyn starrte sie ,berrascht an. "Jetzt noch?" fragte sie. "Du kannst niemals
pünktlich zum Abendessen zurück sein, wenn du jetzt noch zur Post gehst!'" (D2,
p. 119).

Again, in MT2, three whole pages in which the heroine Darrell confesses to the
headmistress that she had wrongly accused another pupil Ellen of stealing and
cheating at exams are missed out, while later the confession of the real thief is
dramatically curtailed. In SCI a chapter where Isabel (Nanni) accidentally finds a
set of exam questions which will give her an unfair advantage is left
untranslated. The moral lessons to be learned in these passages, which might
also be relevant for the new readership, i.e. that cheating, boasting, lying,
stealing and lack of consideration for others' feelings are not acceptable forms
of behaviour, do not figure in the German text.

3.3 Treatment of humour

Though these examples may give the impression that these stories are moral
treatises for the edification of young girls, it must be said that humour is an
important element in them. In both series, the French mistress is one of the main
comic characters. Mam'zelle is represented as a good-hearted but
temperamental figure, fond of theatrical gestures, whose command of English is
256 Malmkjær, Milton & Smith

somewhat shaky and who is the favoured target for tricks and practical jokes, as
she has herself a sense of humour. This character is partially reproduced in the
German text in a similarly stereotypical fashion. Some of the language errors are
also taken over and adapted to fit the German language, such as,

"We shall gobble together about it." "Gabble, Mam'zelle, you mean," said Sally,
with a giggle. (MT5, p. 20).

"Wir werden zusammen davon Schwätzern." "Schwatzen, Mademoiselle, meinen

Sie", sagte Susanne kichernd. (D5, p. 27).

Nevertheless the German translator cannot sustain the effort of linguistic jokes
and leaves many of them simply untranslated. Elsewhere the language "errors"
ascribed to Mam'zelle are simply misunderstood by the translator, e.g. the word
"chest" appears to have been understood as "trunk" (= large suitcase), rather
than "heart":

"You will soon be their friends, n'est-ce pas? Tomorrow they will - what do you
call it - they will take you to their chests and you will be one of them." MTS, p.
20).

"Du wirst bald mit ihnen Freundschaft schliessen, n'est-ce pas? Morgen werden
sie dir ihre Sachen zeigen, und du wirst ihnen gehören." (D5, p. 27).

One of the highlights of MT5 is when, after years of having practical jokes
played on her, Mam'zelle takes revenge and plays a joke on the girls. Having
confiscated a catalogue of items for practical jokes, she sends for some fake
monster teeth and wears them to a sports event. She revels in the astounded
reactions of all she meets, including colleagues. The urge to laugh, however,
becomes too great and she decides to return indoors to remove the teeth. On
the way she meets various other people, including the headmistress with some
parents, and then in the company of the girls in her class she can hold out no
longer and explodes with laughter, dropping the teeth. Only then do the girls
realise that their teacher has played a "treek" on them, and her popularity is
reconfirmed. This whole chapter is omitted in German. It is not quite clear why:
perhaps the idea of a teacher resorting to such undignified behaviour is not
compatible with German school stories, perhaps the translation was just
growing too long. In any case, the readers are cheated of one of the most
hilarious incidents in the whole of Malory Towers.
This brief comparison of these English and German texts confirms some
of the tentative conclusions put forward by House:
 Translation and mass culture 257

there may well be differences as regards children's books between the German
and English linguacultures: there is a German tendency towards depicting a
different role relationship between children and adults, there is [...] less humour,
a far greater explicitness and a greater need to impose edifying ideas and
ideology on the stories told in children's books. (1997:130)

Certainly the degree of humour found in the English originals of Blyton's stories
is not found in the German translations, and interestingly enough the degree of
explicitness in the German translations often does exceed that of the originals.
In view of the fact that much of the detail of the originals is omitted, this
conclusion may surprise, but in fact often the detail is replaced by a single
sentence which summarises and at the same time makes actions more explicit
than they would have been. The edification which House finds in her German
data is less present in these texts. This may be the result of an attempt to update
the books and situate them in a social context in which the boarding school and
the class considerations attached to it do not play an equivalent role.
In direct comparison, the German translations come across as less exciting from
a child's point of view. It is all the more surprising, therefore, that they enjoy
such enormous success with German-speaking readers. A critic from India
(Hindu Index, 26.04.97), documenting Blyton's success in India, claims that
Blyton has no real competitor among local authors, apart from those who have
chosen to clone her. Could it also be the case in the German-speaking world,
that, regardless of the dubious quality of the translations, Enid Blyton simply
fills a niche that no-one else has succeeded in filling?

Note
1
The guidelines advocate the following procedures: Sie müssen frei übersetzen. Übernehmen
Sie nicht die oft schwülstige, überladene Ausdrucksweise des Originals. [...] Wann immer Sie
im Original auf inhaltliche Fehler stoen, bitte ändern. [...] Gelegentlich reagieren die
Hauptpersonen unverständlich, oder ihre Gefühle sind nicht nachvollziehbar. Auch das
dürfen Sie ändern.[...] Sie dürfen nicht nur, sondern Sie müssen sogar sehr frei übersetzen. (A
free translation is essential. Do not reproduce the sometimes excessive and over-exaggerated
language of the original... Whenever you encounter content errors, correct them...
Occasionally the main characters act oddly or their emotional reactions do not seem natural...
This should also be changed. A free translation is not just recommended: it is a "must".)

References

BENJAMIN, Walter, 1992. "The Work of Art in an Age of Mechanical Reproduction".

Illuminations ed. by ARENDT, H., 217-252. (Transi. H. Zohn). London: Fontana,
258 Malmkjær, Milton & Smith

FERREIRA, Jerusa Pires, 1993. O Livro de Sao Cipriano: uma Legenda de Massas. Säo
Paulo: Perspectiva.

HOUSE, Juliane, 1997. Translation Quality Assessment. A Model Revisited. Tübingen:

Gunter Narr.

KORSE, Jens. 1994. Review of De måske egrlede. Det Fri Aktuelt, 16 November.

LEFEVERE, André, 1992. Translation, Rewriting and the Manipulation of Literary Frame
London & New York: Routledge.

MILTON, John, 1995. "A Traduçao de Romances 'Classicos' do Ingles para o Português no
Brasil" ('The Translation of 'Classic' Novels from English to Portuguese in Brazil'),
Trabalhos em Lingüistica Aplicada, 24, Instituto de Estudos da Linguagem, UNICAMP,
Campinas, Brazil.

MILTON, John, 1996. "The Translations of O Clube do Livro". TradTerm 3, 47-65. Säo
Paulo: Universidade de Sao Paulo.

MORIN, Edgar, 1977. Cultura de Massas no Século XX: O Espfri to do Tempo 1- Neurose.
Rio de Janeiro: Editora Forense Universitaria.

OITTINEN, Riita, 1995. I Am Me, I Am Other Tampere: Tampere University Press.

PUURTINEN, Tiina, 1994. "Dynamic style as a parameter of acceptability in translated

children's books." Translation Studies. An Interdiscipline ed. by SNELL-HORNBY, M.,
PÖCHHACKER, F., KAINDL, K., 83-90. Amsterdam & Philadelphia: John Benjamins.

SMITH, V. & KLEIN-BRALEY, C. M., 1985. In other words... Munich: Hueber.

SOHÀR, Anikó (forthcoming). "The mystery of Wayne Chapman - a Case of fictitious

translation in Hungary". Proceedings of the "Practices of Literary Translation" Conference,
Norwich, 1996. Ed. by HALE, T.

VENUTI, Lawrence. 1995. The Translator's Invisibility: A History of Translation. London &
New York: Routledge.

Publications by Enid Blyton

St. Clare's, Vols. 1-6. Methuen, 1942-1945. German translations: Hanni und Nanni, 1-6.
Schneider Verlag, 1965-1970.

Malory Towers, Vols. 1-6. Methuen: 1946-1951. German translation: Dolly, 1-6. Schneider
Verlag, 1966-1967.
 Translation and mass culture 259

Publications by Peter Høeg

Forestilling om det tyvende Arhundrede. 1988. Copenhagen: Munksgaard/Rosinante. English

translation by Barbara Haveland, The History of Danish Dreams. 1995 and 1996. London:
The Harvill Press.

Fortællinger om natten. 1990. Copenhagen: Munksgaard/Rosinante. English translation by

Barbara Haveland, Tales of the Night. 1997. London: The Harvill Press.

Frøken Smillas fornemmelse for sne. 1992. Copenhagen: Munksgaardt/Rosinante. English

translations by Tiina Nuunally, Smilla's Sense of Snow. 1993. New York: Farrar, Strauss &
Giroux Inc., and by F. David, Miss Smilla's Feeling for Snow. 1996. London: The Harvill
Press

De måske egnede. 1993. Copenhagen: Munksgaard/Rosinante. English translation by Barbara

Haveland, Borderliners. 1994. London: The Harvill Press.

Kvinden og aben. 1996. Copenhagen: Munksgaard/Rosinante. English translation by Barbara

Haveland, The Woman and the Ape. 1996. London: The Harvill Press.
TEXT, IMAGE, AND TRANSLATION
THE EXAMPLE OF ADVERTISING IN GERMAN AND
IN BRAZILIAN PORTUGUESE IN A GLOBALIZED
CONTEXT

Masa Nomura
University of Sao Paulo, Brazil

Zusammenfassung
Die vorliegende Arbeit behandelt die Art der Stereotypen, auf die die Werbung in
einem globalisierten Kontext zurückgreift. Eine Auswahl von Werbetexten im
Deutschen und im brasilianischen Portugiesisch soll dazu dienen, Fragen der Identität
und der Fremdheit anhand des Stereotypenbegriffs zu erläutern. Die in Werbetexten
gefundenen Stereotypen stimmen nur teilweise mit der allgemeinen Vorstellung
überein, was unter "Deutschsein" und "Brasilianersein" zu verstehen ist.

Résumé
Ce travail traite de quelques similitudes et différences entre les stéréotypes culturels
utilisés dans la publicité alors que la mondialisation est en cours. On a analysé une
sélection de textes publicitaires en allemand et en portugais brésilien dans la
perspective des questions d'identité et de différence, la notion de stéréotype éclairant
ces questions. Les stéréotypes rencontrés dans les textes étudiés confirment, en partie
seulement, ce qui est entendu par les expressions "être allemand" ou "être brésilien".

Resumen
Este trabajo habla de algunos puntos comunes y diferentes entre los estereotipos de
que se vale la publicidad en un contexto globalizado. Una selección de textos
publicitarios en alemân y en portugués brasileño fue analizado con la fmalidad de
discutir cuestiones relativas a identidad y diferencia y la noción de estereotipo que
ilumine estas cuestiones. Los estereotipos encontrados en los textos confirman
parcialmente la noción consensual de lo que sea "ser alemân" y "ser brasileño".
262 Masa Nomura

1. Introduction

When the Portuguese discovered Brazil in 1500, the first powerful image they
had of the new land was the colourful presence of parrots in a green landscape,
so they named it 'the land of parrots'. Reports by early European travellers
mention 'naked Indians eating human flesh'. These examples are mentioned
merely to illustrate the tendency that predominates in the field of advertising
for the tourist industry nowadays, especially the one directed at the European
public that has retained the emphasis on the exoticism of the 'Southern lands'.
Even presumably more informed depictions of what could be called the Third
World embody ideological barriers. As Lennox (1994:185) critically points
out, the lack of interest in the historical and cultural background of less
developed countries has led to a trend towards overemphasising negative
aspects, such as underdevelopment, poverty, social disarray, etc. Exoticism and
poverty are aspects that still mark the European view of Third World peoples.
In spite of the differences, the progressive intertwining of international
and economic relations imposes increasingly closer and more frequent contacts
with foreign cultures. This is confirmed by the expansion of markets beyond
geopolitical borders, a trend that requires companies to commercialise the same
products and services in different cultural environments simultaneously. One
should also consider issues normally studied in marketing and advertising
schools (Garcia 1990, Marcondes and Ramos 1995), such as the way the media
influence consumers, producing changes in deeply engraved habits or instilling
the compulsion to buy often superfluous products and services. Understanding
the discourse of international advertising thus involves questions that concern
culture-specific data and norms. Multinational companies seek to operate a
coherent production and marketing strategy on a global basis. Nevertheless, in
Europe and North America it is increasingly realised that translations of
avertising texts must be adapted, not translated (Schmitt 1985:2).
Advertising has now become a rich field of research also for translation
studies (cf. Reiss 1976, Snell-Hornby 1986, Azenha 1994, Azenha, Dornbusch
& Nomura 1997). The awareness of the confrontation between identities
sensitises the translator to the continuum that exists between reality and text:
"Advertising [...] is nothing but a part of a wider reality, namely, the set of
social relations gathered together at a specific historical moment" (Garcia
1990:9; translation M.N.). It is clear nowadays that whenever cultural barriers
are crossed advertising texts need to be rewritten or 'adapted'.

2. Stereotypes

The cognitive concept of shared world knowledge comprises the speaker's

notions about typical facts, situations and values generally prevailing in the
 Text, image and translation 263

cultural community to which he/she belongs. German and Brazilian cultural

communities share, to a large extent, the same core of world knowledge, yet
there are relevant differences in behaviour that can be detected and understood
in their own terms only after some experience of the everyday life within a
cultural community is gained by individuals of another culture.
Part of this shared world knowledge consists of stereotypes, as
instances of human cognitive processing of perception and decoding. They
assist in the formation of the basic beliefs and values of a cultural community
at a given historical moment and place (Arras 1998:259). Some stereotypes are,
to some extent, contiguous to the more social notion of prejudice. The negative
connotations are understandable since prejudices are but hostile attitudes
towards members of a group about which one has reservations (ibid.: 264).
A certain feature may be chosen as specially 'culture-bound' so as to
catch the attention of an advertising professional and be transformed into an
icon for marketing. For instance, the statement that we Brazilians are rather
superficial. According to this belief, Brazilians tend to look for whatever is
considered fashionable, provided it relates to important values for their social
group. Men, for instance, are said to be very much concerned about things that
reinforce their masculinity; women, on the other hand, are said to be interested
in eternal youth and in whatever could prevent them from losing their male
partners. Such views should by no means be taken seriously as defining
features of 'the Brazilians'. These examples are meant only to illustrate the
notion of stereotype itself. They do not take into account, for example, the
recent accomplishments of women in their struggle for equal rights with men,
or the recent changes in the concept of manhood.
The fact is, however, that the advertising industry profits from these
stereotypes by launching products that underline this desire to 'patch' beauty or
power. For example, in cosmetics, the aesthetic results that could be produced
by a body cream are thought to be much more effective than informative data
about its low toxicity to the skin or the environment. Or, when presenting a
brand new car, the emphasis can lie on its colour or form rather than on
technical. In so doing, advertising seeks to stress the enviable social status one
might acquire by owning such a product.
As for the Germans, the self-image they most care about is concerned
with their trade-mark image. "Made in Germany" indicates the technical
perfection of German technology and the reliability of their security norms,
especially those concerning automobiles and machines. "Feiern. Made in
Germany. Astra Caravan Edition 100." is the title of a commercial to sell the
newest car model produced by Opel {Focus 16, April 99). Qualities normally
attached to the German people, such as self-discipline, order, efficiency and
punctuality, may be stressed in advertising, as can be seen in this text for the
German WestLB bank: "Umstellung auf den Euro - sicher, effizient,
pünktlich:] {Focus 47, Nov. 98).
264 Masa Nomura

3. Advertising texts

3.1 Selection of material

Samples of advertising texts were selected from Brazilian (100 texts) and
German magazines (another 100 texts). Both corpora have a similar journalistic
format (2 weekly magazines in each country: Veja and IstoE, Der Spiegel and
Focus), the same large circulation (approximately 1 250 000), the same type of
contents (news briefs, politics, sports, culture, new trends), and the same target
public (average working men/women, usually seeking to be well-informed
about all kinds of issues). Each one dominates the market over similar printed
media; they are opinion formers and known to be serious and reliable in their
treatment of the news. The advertising texts they carry are presumably
successful, thanks to the great number of potential consumers they reach.

3.2 Analysis: the manipulation of wishes

The enormous influence exerted by advertising on the market and on people's

wishes is conveyed through the creative interplay between verbal and
nonverbal signs. This semiotic association was taken into account in the
analysis of our corpus. Our sampling shows two kinds of methods employed
both in Brazil and in Germany: product appreciation by means of expressive
resources taken from technical languages; and the act of suggesting what to do,
directly addressed to the consumer. The former seems to be more extensively
used in Germany (90% of the German examples), while the latter one is
preferred in Brazil (70% of the Brazilian examples).

3.2.1 Product appreciation

Product appreciation often resorts to the expressive devices employed in
technical language. A Brazilian advertising text for the Chevrolet company, for
instance, makes use of nominalisation:

Estrutura com sistema de absorçao de impactos programado. VECTRA.

('Structure with programmed impact absorption system. VECTRA.') (IstoE
1508, Aug. 98)

Advertising in German may appeal to the creative procedures of word

formation, a very typical feature of the German language (Bödeker 1971:11f):

Alfa Romeo 156. Kraft, die souverän kontrollierbar ist: Vierlenker -

Vorderradaufhängung, direkte Servolenkung, leistungsstarke Bremsen und
ABS. Und für langanhaltenden Fahrspaß sorgt die durch Verzinkung und
Hohlraumversiegelung korrosionsgeschützte Karosserie. (Focus 47, Nov. 98)
 Text, image and translation 265

Loan words and the use of codeswitching (especially English) are

frequently used in commercials for conferring a tone of modernity to the
advertised product and making the reader feel in tune with the world. English is
a kind of password used everywhere to stress the universality of a product:

Der Ventura Effekt. Der neue Eagle Ventura. Der High-Performance-Reifen

der Zukunft: eine Top-Kombination aus souveränem Fahrverfahren durch
optimales Handling sowie herausragender Strassenhaftung [...]. (for
Goodyear tyres, Focus 16, April 99)

Perspektiven entstehen aus Energie. Bayernwerk. Wir powern Ideen.; "Take

your world with you, es war noch nie so leicht." (Motorola mobile phone,
both in: Spiegel 17, April 99)

A Brazilian advertisement for the Jeep Cherokee: "Motor 4.0L High Output,
suspensao Up Country, traçao Selec-Trac © 4X4 Full-Time com cambio
automático, ar-condicionado e duplo air-bag." (IstoE 1508, Aug. 98);
"MontBlanc. The Art of Writing." (Veja 36, Aug. 97)

Comparatives, superlatives and pompous words, instead of the usual

names of products and services, enhance the appreciation of the product by
increasing its prestige. One of the consequences of product appreciation is that
the object of discourse appears more important than it really is.

Cerveja Brahma. A numero um. ('Brahma beer. The number one.', Veja 17,
April 99)
Das einzig wahre Warsteiner Spitzen-Pilsener der Premium-Klasse. (Der
Spiegel 17, April 99)

3.2.2. Suggesting what to do

Suggesting and cajoling are speech acts embodied in affirmative, imperative
and interrogative sentences directly addressed to the reader, in order to provoke
his/her vanity and his/her desire for a better life. These extremely efficient
strategies of seduction are especially noticeable in Brazilian advertising. Most
texts begin by directly addressing the reader in structures like "voce quer? ...
voce vai..." ('do you want? ... [then] do so'), as shown in the following
example:

Se voce quer mais da vida - nada mais justo que voce tenha um cartao à
altura. - Faça algo de novo na sua vida. ('If you want more from life -
nothing is more important than a highly recommended credit card. - Do
something new in your life.') (IstoÉ 1499, June 98)
266 Masa Nomura

Examples of this kind were difficult to find in the German sample.

The intention of stirring up new interests in potential consumers and, in
some way, of influencing their behaviour is often visible in questions asked in
a tone of friendly advice, as shown in this text for a new brand of coffee:

Por que nao mudar o seu ritmo? Por que nao ficar bonita e gostosa? Por que
nao mudar o seu dia-a-dia? ('Why don't you change your rhythm of life?
Why not be pretty and sexy? Why don't you change your routine?') (Nescafé,
Veja 36, Sept. 97)

In German commercials we often find sentences that denote inner convictions

conveyed in a very positive and convincing tone, so as to reinforce one's sense
of confidence in the statement made by the text producer:

Sie werden es mit allen sechs Sinnen erleben. Alle anderen nur mit einem.
Sehen Sie die Welt mit anderen Augen. Fühlen Sie die Sonne warm auf der
Haut. Atmen Sie den frischen Duft der Natur. Und gemessen Sie es, unterwegs
zu sein. Das neue Saab 9-3 Cabrio gibt Ihnen die souveräne Sicherheit, alles
perfekt unter Kontrolle zu haben. [...]. (Der Spiegel 8, Feb. 98)

Fostering confidence in the product is another objective. This can be

conveyed by resources like direct speech, euphemisms, and the positive
assessment of the advertised item given by celebrities, i.e. trustworthy people
who guarantee the ethos of the advertisement and the reliability of the product
(Hantsch 1972: 95). These so-called 'qualified judges' can be famous
personalities, fictitious characters with honorific titles, anonymous
connoisseurs. Old traditions or cliché-ideas can also unleash certain
associations with reliability and confidence. Popular sayings can assume the
function of authorities, for it is a widely-held opinion that folk wisdom is
distilled in them, being, therefore, accepted by all as a measure of quality. Such
expedients can be easily found in the advertising texts of both cultures.
In Brazilian advertising, we often find ambiguous statements. This
expedient allows for an interplay with an image that, in fact, has nothing to do
with the product itself For example, a commercial bank asks: "Problemas com
cheques voadores?" ('Are you having problems with flying cheques?'). This
question is illustrated with a bird cage with a 'flying' cheque inside, in a clear
reference (for Brazilians!) to bouncing cheques (not a rare occurrence in an
economy like the Brazilian, where such speculation is quite common). We
could not find any ambiguous statements in our German sample. We assume
that this is perhaps due to the trust a German citizen has in the solid German
economy and the absence of any doubts about the quality of products and
services offered in Germany.
 Text, image and translation 267

Another way of drawing the attention of potential consumers is to

suggest something completely different by means of the skilful use of
wordplays combined with images. This intersemiotic interplay must make use
of the shared knowledge of both text producer and text recipient in order to be
successful. In a Brazilian commercial for a dishwasher, a woman watches her
husband washing dishes in the kitchen. The text says: "Nenhuma mulher quer
um hornern bom de pia" ('No woman wants a man [who is] good at [the] sink').
"Ele é bom de pia" means 'he washes the dishes very well'. For Brazilians this
expression arouses an immediate association with the expression "Ele é bom de
camd" ('he is good in bed, he is a wonderful lover'); this meaning is reinforced
by the wordplay between pia ('sink') and pica, which is phonetically close to
the popular name for the male organ. A very successful pun in German
involves the well-known name Bill Gates and the German expression "Wie
geht es weiter?" in this advertising for a computer magazine: Wie GATES
weiter? Computer-News? COMPUTER BILD. (Focus, 97).

3.3 Image

3.3.1 Visual resources

The appealing effect of many advertising texts is obtained via visual signs.
From a cultural point of view, the most interesting visual elements are those
that can draw people's attention. All elements - illustrations, colour, the
available space - are intended to construct, at the iconic level, the symbols of
desire. Appealing directly to the senses, these symbols lead to other symbols
that connote power, supremacy, riches, and beauty.

3.3.2 A selected image typology in advertising

Food (20% of the Brazilian. total, 3% of the German total). It is quite
remarkable how many advertising texts in Brazil use food to promote sales of
an immense variety of non-edible products: a sandwich for advertising a radio
programme, pizza for a new call service from the telephone company, lemon
slices for a cleaning product, a cup of hot coffee for an automobile, a popcorn-
carpeted room for a television set, a glass of wine for a credit card. German
advertising texts do not make much use of this device, but if they do, it is not in
the same way as Brazilians do. A German commercial for a car, for instance,
shows very objectively the capacity of the car's boot, stating the exact amount
of luggage it can hold, whereas a similar text in Brazil shows the boot of a car
full of gifts and bottles of champagne. One example in German of the use of
food in association with a non-edible product has a rather matter-of-fact effect
on the text recipient: the photo of a blue-coloured banana by a printer machine
(Lexmark) points out that this printer can reproduce anything in any colour.
Family and children (Br. 15%, Ger. 12%). This theme is often exploited
in both cultures: images of a happy middle-class family, husband, wife, a little
268 Masa Nomura

boy and a little girl; mother and children, a young father looking after his son;
many infants, preferably cute and blond with blue eyes, selling a large number
of different products. A Brazilian advertising agency was very successful in
launching a campaign for a multinational manufacturer of dairy products
(Parmalat) with children of several ethnic backgrounds dressed up as different
mammals. The appeal to many people's feelings of tenderness towards children
in their infancy paves the way for the sale of the advertised goods.
Sexuality (Br. 15%, Ger. 13%). The explicit or suggested resort to
sexuality is highly appreciated among advertisers in general. In Brazil, a
tropical country with a long coastline full of sunny beaches, images of young
men and women with perfect tanned bodies in the skimpiest swimsuits are
widely used in advertising. To appreciate naked bodies is therefore rather
natural in Brazil and has no special erotic effect on Brazilian onlookers.
However, it immediately catches one's attention if the advertising plays on
social taboos or prejudices. A black-and-white couple, both young, beautiful
and sexy, was chosen to sell a new brand of coffee; he is black and she is
blonde, or vice-versa, in six image versions. The verbal text by a cup filled
with hot coffee draws one's attention to the perfect match between the two:

Chegou o café Parmalat. O café à altura do nosso leite. ('Parmalat coffee has
arrived. The coffee to match our milk.')

Although miscegenation is (in theory) socially accepted in Brazil, the very fact
of presenting it so clearly in the black-and-white interplay caused a great
impact on the text recipients.
The automobile as a symbol of status (Br. 30%, Ger. 50%). Shining cars
in metallic colours, an image exploited both in Brazilian and in German
advertising, are used for the purpose of arousing the covetousness of the
potential buyer. In both cultures, cars are often depicted against a vast
landscape, as in many advertisements for car brands such as Cherokee, Toyota,
Mitsubishi and others. Such an arrangement emphasises the sense of conquest
of new spaces caused by the possession of that particular object of desire.
Characters belonging to folk fantasy and the world of arts (Br. 12%,
Ger. 15%)). A knowledge of history and mythology is to some extent
incorporated in German folk fantasy and often appears in German advertising:
characters from fairy tales, witches, goblins, giants, kings, queens, as well as
portrayals of classical paintings, famous sculptures, boroughs, castles and
churches. In Brazil, popular heroes and heroines come from the sports world
(Ayrton Senna, car racing champion; Ronaldinho, football star; Gustavo
Kuerten, tennis champion) or from the television entertainment world; the
mixture of fictional and real characters plays an important role in people's
fantasy. One of the most popular icons of Brazilian television show-business,
 Text, image and translation 269

Xuxa, blonde, beautiful, young and rich, gains millions of dollars by serving as
a model for many advertising agencies.
Religious symbols (Br. 8%, Ger. 7%). In Brazil, as well as in Germany,
religious symbols are very common in advertising. The image used for a
German car manufacturer shows the interior of a Gothic cathedral to symbolise
the vast interior space of its vehicles. In Brazil, a country marked by
Catholicism and religious syncretism, images of Jesus Christ, the Pope and
popular Catholic saints, as well as gods and goddesses of the Afro-Christian
religion, may serve to attest the quality of some goods.
These examples show that the messages produced in the advertising
industry are structured, above all, around the interplay between image and
verbal text: the higher the number of visual and/or intertextual points of contact
around which such associations are built, the stronger the persuasive power of
the message. Owing to the richness of specific cultural references, the task of
the translator of such advertising texts on a global scale will be very complex.
The difficulty for translation is proportional to the extensiveness of these
multiple connections.

4. Some consequences for translation

There are two main consequences that should be mentioned. One is the need
for adaptation (in the broadest sense of the term) with due appreciation of
culture-specific conditioning factors, stereotypes, and the peculiarities of the
advertising industry. The second is the way we are forced to reflect on the
notion of 'translating' itself, since, in the case of advertising, it is the appealing
element that will have to be translated, even if this means nearly recomposing
the entire text.
Secondly, the translation of advertisements highlights question of the
creativity involved in the process of rewriting texts, so that they can operate
with images related to intrinsic values of the target-culture, provided the
original message is preserved. It also underlines the importance of mastering
the expressive means geared to a specific translation job. The translation of
advertising reveals conditioning factors that are present, to a varied extent, in
the translation of any kind of text (political speeches, religious texts,
manifestos, legal documents, user's instructions). It therefore stresses the
hybrid character of texts. It also calls for reflection on the translator's role as a
re-creator, not just a simple mediator of the information exchanged between the
several parties involved in the communication process (source-text recipient,
target-text producer, mediator of the negotiation between advertiser,
advertising agency, publishers etc.). Finally, it invites the reconsideration of the
question of 'authorship' in translation (individual work, team work, the role of
the advertising agencies) and of ethics (the sanction imposed by the company
270 Masa Nomura

on the 'quality of the translation', the responsibility for the published content,
the method of diffusion etc.).

References

ARRAS, U. 1998. "Stereotype und Vorurteile - mit besonderer Berücksichtigung des

Fremdsprachenunterrichts". Pandaemonium Germanicum. Revista de Estudos Germânicos 2,
257-288. Säo Paulo: Humanitas.

AZENHA, J. 1996. "Kulturelle Aspekte bei der Übersetzung technischer Texte - Deutsch-
(brasilianisches) Portugiesisch. Theorie und Praxis". Die Unterrichtspraxis/Teaching German,
73-76. New Jersey: AATG.

AZENHA, J., DORNBUSCH, C, NOMURA, M. 1997. "Imagem, texto, sensibilizacâo,

criatividade". Pandaemonium Germanicum 1, 99-119. São Paulo: Humanitas.

BÖDEKER, J. D. 1971. Sprache der Anzeigenwerbung. Ein Arbeitskurs zum Verständnis

appellativen Gebrauchs der Sprache. 2nd ed. Karlsruhe: Lehrerheft der Arbeitsunterlagen für
den Deutschunterricht.

GARCIA, N. J. 1990. Sadismo, sedução e silêncio. Propaganda e controle ideológico no

Brasil: 1964-1980. São Paulo: Ed. Loyola.

HANTSCH, I. 1972. "Zur semantischen Strategie der Werbung". Sprache im technischen

Zeitalter 42. 93-112.

LENNOX, S. 1994. "Enzensberger, Kursbuch, and 'Third Worldism'. The Sixties'

Construction in Latin America". Neue Welt/Dritte Welt. Interkulturelle Beziehungen
Deutschlands zu Lateinamerika und der Karibik ed. by BAUSCHINGER, S. & COCALIS, S,
185-200. Tübingen-Basel: Francke Verlag,

MARCONDES, P. & RAMOS, R. 1995. 200 anos de propaganda no Brasil: do reclame ao

cyber-anúncio. Säo Paulo: Ed. Meio & Mensagem.

REISS, K. 1976. Texttyp und Übersetzungsmethode. Der operative Text. 3rd ed.: 1992.
Heidelberg: Niemeyer.

SCHMITT, P. 1985. "Interkulturelle Kommunikationsprobleme in multinationalen Konzernen.

Ein Bericht aus der Sicht des Übersetzers". Lebende Sprachen 1, 1-9.

SNELL-HORNBY, M. 1988. Translation Studies: An Integrated Approach. Amsterdam &

Philadelphia: John Benjamins.
RECEPTION FACTORS IN TRANSLATED
ADVERTISEMENTS

Cristina Valdés
University of Oviedo, Spain

Zusammenfassung
Der folgende Artikel beschäftigt sich mit der Rezeption von Werbeanzeigen, die für
den spanischen Markt übersetzt wurden, und schlägt eine grundlegende
Herangehensweise an die Beobachtung vor, dass einem Text bei der Rezeption durch
verschiedene Kulturen unterschiedliche Bedeutungen verliehen werden. Eine
deskriptive Analyse der ins Spanische übersetzten Anzeigen zeigt, dass die
Übersetzung fast immer eine Verändung der einzelnen Textelemente erfordet, da es
sich um multimediale Texte handelt, welche verbale und nonverbale Komponenten
enthalten. In allen untersuchten Beispielen seigt sich, dass verschiedene Faktoren und
die jeweilige Rezeption der Anzeigen zu unterschiedlichen Übersetzungsstrategien
fuhren.

Résumé
Cet article présente quelques-uns des facteurs touchant la réception de publicités
traduites pour le marché espagnol et propose une approche générale du fait que les
éventuelles cultures de réception donnent une signification différente à un même texte.
Les textes multimédia combinent des composants verbal et non-verbal: leur traduction
exige presque toujours une transformation de ces composants, comme le prouve
l'analyse de publicités traduites en espagnol. Dans tous ces cas, différents facteurs
déterminent les stratégies de traduction.

Resumen
El siguiente articulo presenta algunos de los factores que afectana la recepción de
anuncios traducidos para el mercado español y ofrece una aproximación general al
hecho de que distintas culturas receptoras asignan un significado diferente al mismo
texto. Ya que se trata de textos de naturaleza multimedia que combinan un
componente verbal y uno no-verbal, su traducción supone casi siempre una
transformación de dichos componentes, como demuestra el análisis descriptivo de un
numero de anuncios traducidos al espafiol. En todos estos casos, factores diferentes y
su recepción dan lugar a varias estrategias traductoras.
272 Cristina Valdés

1. Introduction

The starting point for this study is a consideration of advertisements as

multimedia texts, given their semiotic complexity (Reiss 1981:125). They
combine pictures, sounds and words in both oral and written realizations and in
a variety of forms: they can be made up of songs, music, dubbed voices,
subtitles, typography, photographs, cartoons, etc. Colour, size, position, music,
movement or light intensity can all influence the way an advertisement is
perceived. All these elements are culture-specific, since cultures assign
different interpretations to them out of convention.
The main factor influencing the production and the translation of an ad
is the receiver. All translation decisions depend on the presupposition the
translator has about the target consumer's interpretation of the message.
However, it is not always easy to communicate the intended message of
multimedia texts, as they are characterized as having a heterogeneous,
anonymous and geographically dispersed mass audience. Moreover, the
viewers' reaction to an advertisement is never passive but active, since they
construct meaning from the interaction between their sets of values, beliefs and
expectations and the complex internal structure of the text. "The meanings are
affected by what texts and audiences bring to them" (Brierley 1995:204).
John Corner has identified three levels of meaning which audiences
may bring to texts, and which texts convey: denotation, connotation and
preferred reading (in Brierley 1995:204). The simplest meaning is the
denotative one, as it is shared by everyone (a tree is recognized as a tree); the
second meaning, the connotative meaning, entails certain difficulties for
translation since it exists when a combination of signs stands for something
else that a particular culture or group share. When these signs are transferred to
another context, the target may assign different meanings and associations to
them. Receivers also bring their personal history and values to the text in a
preferred way; an advertisement may remind people of memorable things in
their own lives. When the reader does not understand the code, Stuart Hall calls
this an "aberrant decoding" (Brierley 1995:205), which can occur especially in
the case of some global advertisements which fail to hit the right cultural mark.
Marketing texts quote hundreds of examples of failed translations of
advertising texts.
It is therefore difficult to define only one meaning of an advertisement,
as there are as many interpretations as audiences, who respond to texts in
different and often unpredictable ways. The same applies to translation. The
translator can only propose one possibility among many others, which must be
meaningful to the target readers in a particular context. The text is given a
meaning only at the moment of reception. The multiple possibilities of
encountering different target texts disappear in advertising, owing to the high
cost of the production of this type of text and to its temporary nature. There can
 Translated advertisements 273

be, however, different versions of the same target text depending on the
medium employed (e.g. television or printed magazines).
In advertising, the feedback process is indirect and delayed; marketing
experts employ different methods to analyse consumers' responses in many
countries. This does not imply an individual reaction but a group response,
from particular segments of society which might react differently to the same
ad. Reactions can be found in letters sent to the editor in magazines, surveys
among television viewers, etc. Although the feedback effect is not immediate,
changes in the campaign may be required after its first reception in the target
market. A well-known example of this was a printed advertisement promoting
a new soap powder which showed a picture of dirty clothes on the left, a box of
soap in the middle and clean clothes on the right. The soap did not sell well in
the Middle East. Why? Obviously because the advertisers forgot that in that
part of the world, people usually read from right to left. The main obstacle here
was not the linguistic component but the placing of the pictures.

2. Reception factors in translated advertisements

The first goal is to produce a text that will be considered 'acceptable' by the
target text receivers. The translator knows the target culture conventions for
that particular text type, as regards both verbal and non-verbal elements, and
works according to these conventions. Potential consumers have preconceived
ideas which affect the reception of ads. Once they identify the target text as an
ad, they relate it to their expectations about this type of text. They also bring to
it the knowledge they possess about the product and brand, and about the sort
of ads for that brand which they are used to seeing. If there is a dramatic
change in any one of these points, their reactions may be unexpected.
The translation of advertisements may entail the transformation of any
textual element, and these changes usually depend on receivers' expectations
as regards prototypes, that is, conventionalized advertisements for particular
products or brands, or on their lifestyle. Texts must be understood and must be
relevant to the target audience if they are to be convincing. Other decisions
made by the translator which involve a transformation in the target text are
related to changes in the non-verbal component (images and sounds) or to
restrictions imposed by the media. The examples discussed below show how
receivers influence the strategies adopted by translators.

2.1 A different target audience: Iberia

Two advertisements for the airline Iberia which appeared in Spanish and in
English in 1997 share similarities as regards the colours, the picture, and the
content of the text; however, there is a clear difference between them, which
274 Cristina Valdés

has resulted from a change of focus in the translation: they were produced with
two different target audiences in mind. The Spanish source text was published
in the weekly magazine supplement El Semanal, which addresses middle-class
readers of both sexes. The slogan Lideres entre Europa y Latinoamérica
introduces the main theme of the text and the advantage the Spanish airline
company intends to promote: it acts as the link between Spain and the rest of
Europe and also with Latin America, as is explained in the first paragraph. The
translation into English, however, entails a change in the target of the ad since
the audience of the text is now those readers of the international publication
Time who might wish to fly to Spain and Latin America, as the slogan
indicates: Leading the way to Spain and Latin America. A different target has
meant different translation strategies: while in the Spanish text the verb used is
unimos (El Semanal, 26th October 1997), given the links between Spain and
Latin America, in the English version the translator simply "offers" (Time, 17th
November 1997) destinations in those countries.
It is the last part of the text that most clearly shows the difference in
focus between both advertisements: the Spanish text mentions the wish of the
company to become the best company in the world for the Spanish audience:
Alli donde tú vayas. Donde tu quieras llevarnos (El Semanal, 26th October
1997), so the target addressed is the Spanish audience for which Iberia is Tu
primera compania, emphasizing the familiarity of the audience with the
company. However, the English text has been written for an international
audience, so the translator has adopted the perspective of a foreigner flying to
Spain and Latin America: We 'd like to be the airline of your choice when flying
to Spain and Latin America. In this case the company is "No. 1 to Spain and
Latin America" and not from Spain to Latin America as it appears in the
Spanish text. A different recipient has thus given rise to a change of strategies.

2.2 A new product in the target market: Advanced Night Repair, by Estée
Lauder

My next two advertisements, one published in the Spanish weekly supplement

El Semanal (21st December 1997) and the other in the United Kingdom
monthly catalogue of the firm Debenhams (Debenhams, Spring/Summer 98),
exemplify how translators have different presuppositions about how easily
receivers will understand certain parts of the source text, which leads to the use
of a compensating strategy to convey that information. The example also
shows how the multimedia nature of the advertisement provides the key to the
viewer's interpretation. The advertised product is a skin recovery complex
launched internationally by the cosmetic firm Estée Lauder. The Spanish
translator adds to the target text the phrase Complejo Protector y Restaurador,
which explains what the product is used for and at the same time compensates
the audience for their lack of understanding of the information given in English
 Translated advertisements 275

in the accompanying picture of the bottle. This strategy is very often adopted
by translators when the product is new in the target market and its use is not
well known. The visual image of the product can help to indicate what the
promoted product is, but the advertisement can make more explicit its
advantage in the cosmetic market.
Although the Spanish translator keeps the core of the message of the
source text, s/he adds some extra information because of the skopos of the
target text, which is not only to sell but to introduce a product which is
unknown to the target audience. A long fragment has been added to emphasize
the benefits of the introduction of the product in the Spanish market:

Si hubiera comenzado a utilizar Night Repair cuando se creó en 1982, su piel

no pareceria hoy 15 años mayor. Con el lanzamiento de Advanced Night
Repair, una fórmula aún mas avanzada, Estée Lauder consigue detener el paso
del tiempo sobre su piel.

In this fragment the translator urges potential Spanish customers to buy the
product as soon as possible in order to make up for lost time.
Translators also adopt this kind of explicitation strategy (see e.g.
Fawcett 1997:100) when the promoted product is closely related to a culture
which is distant from that of the receivers and therefore there may be more
information which needs to be made explicit. The translator fills these cultural
gaps in order to bring readers closer to the advertisement, but preserves the
foreign flavour of the product. An example of this strategy is provided by an
advertisement of Scotch Glenfiddich whisky translated into Spanish. The
central concept of the campaign highlights the origin of this whisky in the
Highlands and the importance that the water of a certain valley has in the
making of the liquor. The Spanish translator, presupposing the lack of
understanding among non-Gaelic speakers, added extra information about the
name of the product: Los nombres dificiles sugieren el misterio. Glenfiddich,
en gaélico, es el Valle del Ciervo'. If the target were a Gaelic-speaking
community, this addition would obviously be redundant. Compare also the
explanation of the English name in: The Highlands. Las Tierras Altas. Región
escocesa de extraordinaria belleza donde se elabora el mejor whisky del
mundo.
A similar strategy can be seen in the following advertisement for a
Finnish vodka:

Toma el sol de medianoche. Hay algo mógico en Finlandia. Un sol que no se

oculta nunca en verano. Un trigo que madura en esos dias largos y luminosos.
Un vodka que hace brillar tus noches. Las hace intensas, alegres, vibrantes.
Pon Finlandia en tu vaso y pondras algo de su magia en tus noches.
276 Cristina Valdés

The explanation given to Spanish readers by the translator is linked to the

picture of a midnight sun, a natural phenomenon not very familiar to the
Spanish audience since it is a culture-bound concept.

2.3 Different target culture habits: Tia Maria

The next example is a pair of advertisements for the liqueur Tia Maria. Both
the English source text and the Spanish target text present the same visuals but
with a different slogan. While the English source text invites readers to choose
between drinking Tia Maria "Neat, mixed or over ice?", the Spanish translator
has adapted his/her translation according to the drinking habits in the Spanish
culture, where such drinks are either consumed neat or with ice: ¿Solo o con
hielo? The strategy employed by the Spanish translator has thus been
influenced by the target culture norms: while the image follows the
globalization principle, the linguistic message has been adapted according to
the target.
Sometimes a different use of a product in the target culture entails a
change of the image as well, and a new message in translations. An example of
this can be found in the Kellogg's TV advertisements in England and the
United States on the one hand and in Spain on the other hand. While in the first
case, cereals are commonly consumed at breakfast by both adults and children,
this product was launched in Spain mostly as breakfast food for children, so
translators need to be aware of this usage when addressing their texts to the
different targets.

2.4 Phonetic factors: Spearmint

The paralinguistic features of advertisements may sometimes pose difficulties

in the reception of the intended message. Receivers have to make an extra
effort to respond to an image or a sound which appears incoherent. The well-
known brand of chewing-gum Spearmint by Wrigley's Jr. Company had its
name translated in Germany to Speermint for phonetic reasons. The first
reaction to the English brand name was negative because the German audience
did not recognize the phonetic equivalent in German for that English spelling
of the brand name; and it was thus difficult to pronounce. This can be called a
case of "phonetic shock". The meaning of the English brand name Spearmint
resulted from the anchoring relationship between the logo of the brand, a black
spear, and the reference to its flavour. The German translator thought it better
to change the name to Speermint in order to retain this relationship and the
holistic meaning. There are many examples of the need to translate brand
names for phonetic reasons which are often quoted in marketing books.
A different strategy related to prosodic features was employed after the
first reception of the English translation of a Spanish-created television
 Translated advertisements 277

commercial. The text was produced, translated into English and dubbed by an
English native speaker in an agency in Madrid. However, when sent to the
London office it had to be translated and dubbed again. The reason was that the
actor responsible for the dubbing had to play the role of a lower-class London
shopkeeper but neither the refined style nor the intonation used was appropriate
for a shopkeeper. Once more, the multimedia nature of the commercial, the
combination of sound, image and text, imposed restrictions on the translation.
With globalization in mind, advertisers seem to rely nowadays less on
linguistic written forms in television commercials and more on music,
particularly on English pop songs. But still, unless the song is very popular,
hearers in contexts other than where the song originated will fail to recognize
the meaningful relationship established between the lyrics and the whole
message of the spot.

2.5 Typography: L'Oréal Perfection

Typography is one of the means most often manipulated by advertisers in order

to support the intended message: for its connotative potential, its appeal, and
because typography can help maximize the effectiveness of reception. As Cook
points out (1992:84), "many successful brand names are inextricably connected
with the typeface in which they are written". The particular circumflex accent
of the French brand Lancôme, for example, is so well known that viewers
immediately relate the perfume Poème to that brand, thanks to the circumflex
alone.
Sometimes translations entail a transformation of the typefaces
employed, in order to compensate the target reader for the connotations the
words evoke in the source text. Two advertisements of the cosmetic range
L'Oréal Perfection in English and Italian can illustrate how the interaction
between typography, some pictorial elements and some words in the Italian
translation account for the evocations expressed only by words in the English
source text. The skopos of both advertisements is the same, to promote the
1996 spring collection of new colours by the cosmetic brand L'Oréal,
characterized by a futuristic reference to galactic shades. Colour, an essential
feature in these advertisements, is preserved in both texts. The English source-
text producer emphasizes the symbolic value of colour, by playing with the
rich semantic field related to light in English. Nouns and adjectives such as
shine, millennium, mega-shimmer, iridescent, spectrum, Blue Moon, Lunar
Lilac, Astro Bronze or Galactic Green convey the same connotations as the
colour of the text. The typefaces also reinforce this futuristic atmosphere with
their simple long greyish and blue characters.
The Italian translator, in order to preserve the same clues to trigger
similar interpretations in the Italian audience, emphasized the role of
typography to compensate for the loss of the connotative value of the English
278 Cristina Valdés

words related to shine. With that purpose in mind, the headline Cybershine is
translated into Italian as Cyberflash, plus a subheading in futuristic graphics,
La nuova moda conquista lo spazio, which has been added in the Italian
version. The cold colours employed and the main paragraph also retain the
connotations of words such as magnetica, futuro scintillante di tecnologia
spaziale, etc., and therefore they produce a similar effect on Italian viewers.
Another addition made in the translation shows a photograph in which some
lipsticks and nail-polish bottles appear as if they were astronauts floating in the
galaxy. This picture is closely related to the clothes worn by the model.

2.6 Stereotyped visuals: Werther 's Original and Orbit

In order to show how advertising translation, as a type of multimedia

translation in general, depends on the interplay between the different semiotic
codes which give shape to the text and the reception they meet, I will mention
two interesting examples from television. Both commercials present some
changes in the visual component, transformations made on the grounds of the
reception of the image. The first one is a commercial for Werther's Original
sweets on one of the Spanish television channels and its English equivalent on
an English television channel. The camera movements and the text are
basically the same, highlighting the sentimental bonds these sweets establish
between grandfather and child. However, the commercial is different in terms
of the actors. The role of the grandfather is played by two actors of about the
same age, bald, with moustache, a deep voice and traditional clothes. They
share most physical features on a denotative and a connotative level. However,
the two actors playing the role of the child show different visual features. The
Spanish commercial has been adapted to the visual stereotype of a Latin
audience: the boy has darker hair and eyes than the boy in the English text.
This strategy of adapting the visual aspect of the advertisement results from the
need to make the target audience identify with the actor enjoying the sweet.
This is known as the mirror technique and it often relies on cultural
stereotypes, "creating the illusion of looking at a perfect version of yourself as
you may become by using the product" (Vestergaard & Schrøder 1985:84).
A more interesting example is that of the television spots of the
chewing-gum brand Orbit in English and Spanish versions. At first sight both
commercials look the same: there are two young couples enjoying themselves
at a fair, eating some sweets. However there is a subtle difference which one
can perceive on closer examination: in the first scene of the Spanish spot, the
dark-haired couple appears on the left side of the screen; in the second scene
the camera frames them more closely on the right side of the screen. In the
following scenes the camera follows them until it returns to the first shot,
where both couples leave the fair and the dark-haired one appears on the left
side of the screen once again. In contrast, the English commercial focuses on
 Translated advertisements 279

the blonde couple from the beginning; the camera makes the same movements
but this time framing the blonde girl. At the beginning and end of the
commercial the blonde couple is shown on the left side, occupying the position
the dark-haired couple had in the Spanish text. The message and the actors are
the same in both texts but the change of the target audience has meant some
transformations in the final presentation of the advertisement. The medium has
played a significant role in this example, since the use of different camera
movements is one of the advantages of television, for it allows the framing of
certain segments of the picture rather than others, to place more emphasis on
them.

2.7 A complex example: American Airlines

The next example once more reveals the complexity involved in multimedia
translation and how translation strategies depend on the intended reception of
the text. The four advertisements I will describe are part of American Airlines'
international campaign "Smiles" in Europe and Latin America. The translation
of the original English text has not only meant a linguistic transfer into
Swedish, French and Spanish but also a transformation of the semiotic
interplay of non-verbal elements.
The text focuses on the picture of a smiling blonde American Airlines
air-hostess, Susan Kent, who is directly facing the viewer. The photograph is
the same for the English, Swedish and French advertisements, while in the text
for the Latin American market the air-hostess is called Isabel Philippi and
represents the stereotypical image of Latin women: dark skin and hair. The
smile is kept, as it serves as the persuasive central concept of the campaign.
The visual manipulation is required to achieve the viewer's identification with
the stereotyped image of the model, but this is done retaining the smile.
There are also some other changes in the smaller pictures related to the
skopos and the target audience of the advertisement. While in the Spanish
translation the promoted American Airlines service between Latin America and
United States is addressed to a general audience, the English source text, as
well as the Swedish and French versions, specifically target businessmen, as
we can infer from the references to work in these texts. This change of target
alters the structure of the Spanish translation, and the offer of five-course
menus and facilities for businessmen in two of the smaller pictures disappears
in the other versions. Other subtle transformations have to do with the different
conventions as regards clothes: for example, in the Latin culture men do not
usually wear braces but they do wear a necktie, as is depicted in the picture.
The marketing experts who mention this campaign state that it "ran in
Europe, basically changing only the language" (Wells et al. 1995:738), a
statement that a closer look at the four texts has proved untrue. The purpose
here has not been to make a detailed analysis of the four advertisements but to
280 Cristina Valdés

show these general factors related to the reception of a multimedia translated

text.

3. Conclusions

The study of these advertisements has shown how certain modifications in the
visual and oral components are required in order to convey the intended effect
of advertisements on the target audience. Multimedia texts offer a good
example of how cultural conventions and intended text reception affect
translation strategies in what marketing experts regard as "a globalized
context."
The idea of a "one-audience" world has prevailed in advertising so far.
However, evidence from translations shows that transformations often take
place. In spite of all the globalization attempts at making a one-culture world
by means of employing the same visual, oral and written elements, translators
still feel the need to adapt their messages to their target receivers and they thus
challenge the tendency of globalization, as the different strategies I have
mentioned have revealed.

References

BRIERLEY, S. 1995. The Advertising Handbook. London & New York: Routledge.

COOK, G. 1992. The Discourse of Advertising. London & New York: Routledge.

CRYSTAL, D. 1994 [1987] The Cambridge Encyclopedia of Language. Cambridge:

Cambridge University Press.

FAWCETT, P. 1997. Translation and Language. Linguistic Theories Explained. Manchester:

St. Jerome Publishing.

MYERS, G. 1994. Words in Ads. London: Edward Arnold.

REISS, K. 1981. "Type, Kind and Individuality of Text. Decision Making in Translation".
Poetics Today 2(4), 121-131.

SNELL-HORNBY, M. 1996. Translation und Text. Vienna: WUV-Universitätsverlag.

VESTERGAARD, T. & K. SCHRØDER. 1985. The Language of Advertising. Oxford &

Cambridge: Basil Blackwell.

WELLS, W., J. BURNETT & S. MORIARTY. 1995. (3rd edition) Advertising. Principles and
Practice. New Jersey: Prentice Hall.
CREATIVITY IN LEGAL TRANSLATION: HOW
MUCH IS TOO MUCH?

Susan Sarcevic
University of Rijeka, Croatia

Zusammenfassung
Während Kreativität im juristischen Übersetzen im allgemeinen bestritten ist, vertritt die
Autorin den Standpunkt, gewisse Abweichungen vom Ausgangstext seien nicht nur
akzeptabel, sondern sogar nützlich, vor allem wenn dadurch Klarheit und richtige
Betonung gefördert und die durch die Paralleltexte angestrebten Folgen herbeigeführt
werden. Beispiele aus den zweisprachigen kanadischen Bundesgesetzen zeigen, wie die
französisch sprechenden Übersetzer ehemalige als sakrosankt angesehene Regeln
verletzen und neue Texte schaffen, die weder beim Lesen noch im Aussehen den
englischen Texten entsprechen.

Résumé
La créativité en traduction juridique prête plutôt à controverse. Notre point de vue est
que les modifications apportées à l'original sont acceptables, sinon même souhaitables,
pour arriver à plus de clarté et souligner ce qui est important, les textes parallèles
atteignant ainsi les objectifs recherchés. On citera des exemples de la législation
fédérale bilingue du Canada: les traducteurs (en français) transgressent désormais des
règles naguère considérées comme sacro-saintes, produisant de la sorte des textes
nouveaux mis à distance de leurs homologues en anglais.

Resumen
Aunque la creatividad en la traducción juridica suscita polémica, este trabajo propone
que cierta "desviación del original" es permisible y hasta deseable para lograr mayor
claridad y énfasis con el fin de conseguir que, en la practica, los textos paralelos
obtengan los resultados deseados. Los ejemplos citados de la legislación federal
canadiense bilingüe muestran cómo los traductores francófonos en la actualidad violan
unas reglas previamente consideradas sagradas, produciendo textos nuevos que no se
parecen a sus homólogos en inglés ni en su aspecto ni en su lectura.
282 Susan Sarcevic

1. Introduction

Although translation creativity has aroused increasing attention in recent

translation studies, it is still considered a "smoke-screen concept" (Wilss
1996:49). Not surprisingly, most of the existing studies on translation creativity
are devoted to literary translation (Kussmaul, in this volume and 1998:39; also
Kussmaul 1995:39-53). This, however, does not exclude the possibility of
creativity in non-literary translation. In regard to technical texts, Durieux
maintains that translators can and should be creative. When it comes to legal
texts, she is more cautious, acknowledging that such texts are subject to special
restrictions (1991:169). Long considered the most restrictive of all translations,
legal texts remained a bastion of literal translation for centuries. This is
particularly true in regard to legislative texts, the primary source of the law.
Without a doubt, the mere idea that translators of legislative texts can be
creative is controversial (Koutsivitis 1991:141). Taking the view that creativity
manifests itself differently in different areas of translation (cf. Wilss 1988:11),
this study attempts to show that translators of legislative texts can be creative
and still respect the restraints of the profession. The legislative texts analyzed
are authentic legal instruments having the force of law in two or more
languages. Hence, we are dealing with authoritative translations having equal
status with the original(s) for the purpose of interpretation by courts of law. For
this reason, lawyers no longer regard such texts as mere translations and even
avoid using the term translation, frequently referring to them as "parallel" texts
instead. Accordingly, the term parallel texts is used here to denote the texts of a
legal instrument authenticated in two or more languages (see Sarcevic
1994:301; cf. Sager 1998:77).

2. New developments in legal translation

Fearing that any deviations from the original could pose a threat to uniform
interpretation and application, lawyers traditionally obliged translators to
preserve the letter of the law by reproducing the wording and syntax of the ST
as closely as possible. This view was finally challenged early in the twentieth
century by Virgile Rossel, a law professor who dared to translate the German
text of the Swiss Civil Code into "natural" French. Accused of heresy for
having altered the letter of the law, Rossel defended his "revolutionary"
translation by invoking the principle of language equality. In his opinion, the
French-speaking population of Switzerland had the right to have their laws
written in the genius of the French language (see Sarcevic 1997:36-40; also
Dullion 1997:379-381).
 Legal translation 283

New developments in legal translation are particularly evident in Canada,

one of the few countries that is not only bilingual but also bilegal, i.e., it has
more than one legal system. Above all, it is the clash between the common law
in the English-speaking provinces and the civil law in the French-speaking
province of Quebec that makes Canada a mecca for legal translators. Despite
Canada's long history in legal translation, serious reform did not take place until
the so-called silent revolution that shook Quebec in the I960's (see Gémar
1995:65-70). In response to the Quebecers' demands for language equality,
Francophone translators of federal legislation were finally given a green light to
produce a linguistically pure text in the genius of the French language. This,
however, was only an initial step. New demands for equal treatment of the
French texts of federal legislation gradually led to the introduction of methods
of bilingual drafting that have radically changed the role of the legal translator.
In the late 1970's Alexandre Covacs, jurilinguist in the French division
of the Legislation Section of the Ministry of Justice, proposed five methods of
bilingual drafting that combine translation and drafting in various ways and
degrees: alternate drafting, shared drafting, double-entry drafting, parallel
drafting, and joint drafting. The initial phase of bilingual drafting is usually
alternate drafting. In my opinion, the term alternate does not refer to the
alternate use of drafting and translation but rather to the fact that the source
text alternates. Namely, some parts of the original are drafted in English, others
in French and so on. These parts are then exchanged and translated. Shared
drafting differs from alternate drafting in that each co-drafter draws up half of
the text, then translates or has the other half translated. As a result, neither text
can be designated as the principal source or target text. In double-entry drafting
(named after the accounting method), the same person produces both texts,
preferably part by part in tandem. In contrast, parallel drafting requires lengthy
consultation and collaboration by the co-drafters who produce both texts
simultaneously yet independently. After preparing a detailed outline of the draft
text, they retire and draft large parts or even the entire text, then meet again to
compare, modify, and coordinate the two texts.
According to Covacs, joint drafting is the ultimate goal. Here the co-
drafters prepare the outline and also do the drafting jointly. Proceeding section
by section, they write, compare, coordinate, and revise both texts. Though time
consuming, this method is considered ideal because the co-drafters collaborate
throughout the entire process, thus guaranteeing unity of thought, yet assuring
that each text is formulated in the genius of that language (see Covacs 1982:93).
In brief it can be said that the new methods of bilingual drafting
coordinate the production of parallel texts by incorporating the translator into the
drafting process, thus defying traditional translation which presumes that the
source and target texts are always produced at a different time and place (as
defined by Vermeer 1986:33). As the process continues, the traditional
284 Susan Sarcevic

designations of "source" and "target" text no longer apply and the translator is
gradually converted into a co-drafter. The question to what extent co-drafting
can be regarded as translation is dealt with elsewhere (Sarcevic 1997:105-108).
Here it suffices to say that co-drafting is not traditional translation but involves
new forms of text production (cf. Vermeer 1996:34, where he now argues for a
broad definition of translation).

3. New freedom in legal translation

Co-drafting has brought translators of federal legislation in Canada new freedom

deemed impossible just a short time ago. Co-drafting also enables and even
encourages translators to be creative. As a note of caution, it should be
emphasized that the new freedom has not been won at the expense of reliability. As
equally authentic texts, both the English and French texts of federal legislation are
approved and authenticated by the Canadian Parliament, and as such are presumed to
contain the same substance, express the same legislative intent and lead to the same
results in practice (cf. Didier 1990:221). Above all, it is the results that count in
legal translation. Accordingly, it follows that the main emphasis in a receiver-oriented
approach to legal translation is on the interpretation and application of parallel texts
by courts of law. Multilingualism in the law can be effective only if translators strive
to produce a new text that will be interpreted and applied by the courts in the same
manner as the other parallel text(s) of that instrument. It is thus legitimate to ask how
and to what extent legal translators can be creative and still promote uniform
interpretation and application by the courts.
In Canada the equally authentic texts of federal legislation continue to be
printed side by side; however, the French version is no longer a mere image of the
English text. The two texts neither look nor read the same. In their new role as co-
drafters, legal translators can be said to enjoy the "freedom of an artist," i.e., the
"freedom to use to the fullest extent everything that language permits" (Driedger
1982:4). Although creativity in legal translation focuses primarily on the creative
use of language, translators/co-drafters of federal legislation in Canada do not stop
here. Testing the limits of their new freedom, Francophone translators not only reject
the accepted linguistic patterns of English common law legislation but also dare to
defy the composition and style of legal rules formulated in the common law tradition.
In this sense, creativity in the Francophone texts of Canadian federal legislation
entails the breaking of common law drafting norms. Generally speaking, the ultimate
goal of Francophone translators is to incorporate an increasing number of civil law
drafting practices into the French texts of Canadian federal legislation. Accordingly,
their "creative products" are not totally original as their creativity is derived from
civil law legislation and, in particular, from French drafting practices (not to be
confused with Neubert's notion of derived creativity, 1997:17-20). Nonetheless, the
 Legal translation 285

new French texts of Canadian federal legislation are "unusual" and contain an
"element of surprise" for all those accustomed to reading common law legislation.
From this point of view, Kussmaul's definition of creativity as a "shift from the
original" that is "functionally adequate" (1998:39) can also apply to these texts. In
my opinion, creativity in legal texts can be deemed "functionally adequate" only
when the "shifts" do not interfere with the uniform interpretation and application of
the parallel texts in question. On the contrary, they should promote uniform
interpretation and application by achieving clarity and proper emphasis, ensuring that
the texts will lead to the same results in practice. While creativity in legal translation
entails innovation, it should not be used for the sake of innovation itself Creativity is
at its best when legal translators use innovative means to compensate for differences
between legal systems and languages, with a view to promoting uniform
interpretation and application. Hence, creativity in legal translation is not simply free
translation.

4. Examples of creativity in legal translation

Regulatory in nature, legislative texts consist mainly of legal rules formulated in

accordance with accepted drafting practices that tend to vary from system to
system. For the most part, legal rules prescribe legal actions in the form of
obligations, permissions, and authorizations whose performance is subject to
certain conditions. As a result, the basic logical structure of legal rules is
expressed by the formula: if P, then Q, which means that Q shall be performed
only in cases where the conditions constituting P are fulfilled. Thus it is only
natural that legal rules are most commonly formulated in conditional sentences.
Greatly simplified, it can be said that the majority of legal rules consist of two
main parts: the fact-situation (P) specifying the conditions to be fulfilled and the
statement of law (Q) prescribing the legal action to be performed. Since P is a
precondition to Q, it was earlier held that the fact-situation must precede the
statement of law. This is the case for example in section 67(1) of the Canada
Labour Code published in the Revised Statutes of Canada, 1970, hence prior to
the reform:

Fact-Situation: (P) Statement of Law: (Q)

Where the Minister is authorized to require the Minister may require the information to
a person to furnish information under this be furnished by a notice to that effect served
Part or the regulations/ personally or sent by registered mail
addressed to the latest known address of the
person for whom the notice is intended, and
such person shall furnish the information
within such reasonable time as is specified
in the notice.
286 Susan Sarcevic

Looking at the French translation of the above provision, also in the Revised
Statutes of Canada, 1970, we see that the translator faithfully reproduces the
wording and syntax of the ST, even preserving the capitalization of English
words. A notable exception, at least from the legal point of view, is the last part
of the statement of law that is formulated as a separate sentence in the French
text, thus breaking the common law drafting rule that each section must be
formulated as a single sentence:

Fact-situation: (P) Statement of law: (Q)

Lorsque le Ministre est autorisé à exiger il peut les exiger au moyen d'un avis à
qu'une personne fournisse des cet effet signifié personnellement ou
renseignements en vertu de la présente adressé par courrier recommandé à la
Partie ou des règlements, / dernière adresse connue du destinataire.
Cette personne doit fournir les
renseignements dans le délai raisonnable
spécifié dans l'avis.

At about the time co-drafting was introduced, the Labour Code was undergoing
a total revision, as a result of which the entire French translation was revised.
While the English text of the above provision remained practically unchanged
(now section 253(1)), major interventions occur in the French translation of the
provision published in the Revised Statutes of Canada, 1985. Daring to break
rules previously considered sacrosanct, the translator/co-drafter completely
reformulates and restructures the translation, making it read more like a French
provision:

Le ministre peut, dans le cadre de la présente partie ou de ses règlements, exiger

certains renseignements au moyen d'un avis soit signifié à personne, soit envoyé
par courrier recommandé à la dernière adresse connue du destinataire. Celui-ci
est tenu de s'y conformer dans le délai raisonnable qui y est fixé.

The boldest innovation in the French text is undoubtedly the departure from the
basic conditional sentence, thus proving that a variety of surface structures can
be used to express a legal rule without disturbing the thought process. Instead of
expessing the conditions of the fact-situation in an introductory lorsque-clause as
earlier, the translator/co-drafter places them in an adverbial phrase (dans le cadre de
la présente partie ou de ses règlements) embedded in the statement of law
between the main verb and its modal (peut exiger), thereby qualifying the legal
action. Of course, there are restrictions to the translator's creativity. In particular, he
or she must ensure that the new surface structure expresses the deep structure of the
intended logical relations (here: if/then), and that the elements of the fact-situation
and statement of law are clearly formulated and easily recognizable.
 Legal translation 287

In the past, legal translators in Canada had practically no decision-making

authority. Among other things, they were not permitted to delete words or phrases of
the ST, including ones deemed unnecessary or out of place in the TT. This is no
longer the case in co-drafting. In the revised French provision above, the
translator/co-drafter deletes words and phrases which appear in the fact-situation
and are repeated in the statement of law in the English text (the minister, require,
furnish information). To achieve greater concision without sacrificing precision,
words such as to that effect and addressed are also deleted. As earlier, the person
for whom the notice is intended is rendered as destinataire. The new formulation
of the second part of the statement of law is particularly effective. Again it is
formulated in a separate sentence; however, this time celui-ci is used instead of
repeating destinataire; the legal imperative shall is expressed less directly with est
tenu; instead of literally translating the phrase to furnish the information, the new
French text simply says est tenu de s'y conformer, which leads to the same result.
Finally, the English phrase as is specified in the notice is rendered freely as qui y
est fixé. The result is a loosely textured text that expresses the same substance
in a clear and concise manner (on French drafting, see Bocquet 1994:15-16).

5. How much creativity is too much?

Such interventions make it clear that the creative intentions of the Francophone
translators/co-drafters are not limited to language but also include the right to
create a new text with a distinctive style based on civil law drafting practices.
This is particularly evident when one compares the English and French texts of
section 253(1) of the Canada Labour Code, as revised and amended in the
Statutes of Canada, 1993:

Where the Minister is authorized to Le ministre peut, dans le cadre de la présente

require a person to furnish information
under this Part of the regulations, the
Minister may require the information
to be furnished by a notice to that
effect served personally or sent by
registered or certified mail addressed
to the latest known address of the
person for whom the notice is intended,
and that person
(a) where the notice is sent by
registered or certified mail, shall be
deemed to have received the notice on
the seventh day after the day on which
it was mailed; and
partie ou de ses règlements, exiger certains
renseignements au moyen d'un avis signifié
à personne ou par courrier recommandé ou
certifié à la dernière adresse connue du
destinataire;
en cas de signification par courrier, l'avis est
réputé avoir été reçu par le destinataire le
septième jour qui suit celui de sa mise à la
poste; le destinataire est tenu de s'y
conformer dans le délai raisonnable qui y est
288 Susan Šarcevic

(b) shall furnish the information within fixé.

such reasonable time as is specified in
the notice.

This time the substance of the provision has been amended by inserting a time
specification into both texts. In English, the time specification and the existing
requirement are set off from the main part of the text by paragraphing, a
technique frequently used in common law legislation to enumerate a list of
contingencies, alternatives, requirements or conditions to be applied
cumulatively or alternatively. Although the device serves as a visual aid to
comprehension by indenting blocks of text, Francophone translators/co-drafters
contend that it is not always used to its best advantage. As a result, they reject
paragraphing whenever possible and incorporate the clauses into the main body
of the provision, as has been done in the French text of the amended provision.
As seen above, the rejection of paragraphing by Francophone translators/co-
drafters results in two very different looking texts, each with its own distinctive
style. Although such changes affect the composition of the French text, they are
usually tolerated as long as the text is clear and the substance unaltered.
As far as the substance is concerned, the question arises as to whether it
is unduly altered by acts of creativity resulting in semantic shifts. Whereas
common law provisions tend to be particular, civil law provisions are formulated
in general terms. Convinced that they can express the same idea in more general
terms, Francophone translators/co-drafters began to use a generic term to
express a series of subordinate terms in the English text. For example, in section
9 of The Official Languages Act (Statutes of Canada, 1988), the English
expression "all rules, orders and regulations governing the practice or
procedure..." is rendered in French as "les textes régissant la procédure et la
pratique. .."In this case, the generic term textes is acceptable because it includes
rules, orders and regulations. Such changes, however, involve a high degree of
risk as the translator must be absolutely certain that the generic term covers all
the subordinate terms, yet is not too broad so as to unintentionally widen the
scope of application, allowing the text to be applied in unintended situations. To
avoid possible misinterpretation, Francophone translators/co-drafters sometimes
identify the items covered by the generic term in an appositive phrase and use
the generic term in a subsequent subsection. For example, the subject of
subsection 1 of section 11 of The Official Languages Act, A notice,
advertisement or other matter is rendered as Les textes - notamment les avis et
annonces in the French version. Instead of repeating the subject and part of the
qualifier in a lengthy cross-reference as in subsection 2 of the English text, the
translator/co-drafter simply uses the generic term:
 Legal translation 289

Where a notice, advertisement or other\ Il est donné dans ces textes égale
matter is printed in one or more\ importance aux deux langues officielles.
publications pursuant to section (1), it
shall be given equal prominence in each
official language

Reformulating a provision in general terms allows the translator/co-drafter to

avoid repetition, thus achieving greater concision in the French version of federal
legislation. The following provision from section 4(3) of the Canada Evidence
Act (Revised Statutes of Canada, 1985) shows to what extent semantic shifts are
tolerated in the French text for the sake of achieving concision. Whereas the
English text consists of a compound sentence containing two co-ordinate clauses,
the same effect is achieved in a simple sentence in the French text by using the
generic term conjoint:

No husband is compellable to disclose Nul ne peut être contraint de divulguer une

any communication made to him by his communication que son conjoint lui a faite
wife during their marriage, and no wife is durant leur mariage.
compellable to disclose any
communication made to her by her
husband during their marriage.

6. Determining the acceptability of creative translation

Creative translation requires considerable language and legal competence on the

part of the translator/co-drafter. Generally speaking, creative translation is not
acceptable if it poses a threat to the uniform interpretation and application of
the parallel texts in question. In particular, an act of creativity endangers
uniform interpretation and application if it alters the substance, makes the text
unclear or ambiguous, muddles the legislative intent, changes the scope of
application or raises any doubts that the legal effects will not be the same. When
making such decisions, the translator/co-drafter must always take account of the
communicative situation of reception in the particular communicative act, i.e.,
the situational factors relating to the interpretation and application of the
parallel texts in question.
In a nutshell, the translator/co-drafter should take account of the
following factors when determining the acceptability of an act of creativity: the
jurisdictions where the text is applicable, the legal systems in these jurisdictions,
the courts with jurisdiction to decide disputes in matters governed by the piece
of legislation, the law in which the judges are trained and the rules of
interpretation they apply. For example, Canadian federal legislation is applied in
290 Susan Šarcevic

the entire territory of Canada; hence, two legal systems come into play:
common law and civil law. A system of federal courts has been established to
decide disputes in both English and French in all matters governed by federal
statute. Canadian federal law is common law; thus the sitting judges are trained
in the common law and apply common law rules of interpretation. The fears of
Canadian common law lawyers therefore appear to be justified: the increasingly
bold creativity by Francophone translators/co-drafters has reached its upper
limits of acceptability.
Multilingual communication in the law can be successful only if there is
interaction between text producers and receivers, i.e., those who interpret and
apply the parallel texts of a given instrument. Among other things, this means
that the drafting techniques used in each of the parallel texts should be
compatible with the rules of interpretation applied by the relevant courts. Since
common law rules of interpretation are applied by federal judges, including the
justices of the Supreme Court of Canada, it follows that drafting elements that
interfere with such rules pose a threat to uniform interpretation and application
and are thus unacceptable in federal legislation. In keeping with this reasoning,
common law lawyers have warned translators/co-drafters to refrain from using
elements of civil law drafting that are incompatible with the methods of common
law interpretation. In the interest of guaranteeing the harmonious co-existence
of the two texts, Francophone translators/co-drafters have apparently agreed to
strike a balance with the English text and the common law in federal statutes.
This, however, is not their last word.

7. Can the Canadian model serve as a precedent?

Without a doubt the Canadian model serves as a "source of inspiration" for

legal translators worldwide (Levert 1995:266). This, however, does not mean
that such acts of creativity are acceptable in other multilingual jurisdictions. The
question to what extent creativity is acceptable in parallel texts depends on the
communicative situation in the particular jurisdiction concerned. In Switzerland,
for example, several co-drafting methods are also used; however, the results are
very different in the sense that Swiss translators are much more conservative
when it comes to creativity (see Sarcevic 1997:102, 272).
Canada is clearly an exception. Here the struggle for equal treatment of
the French texts of federal legislation has pushed the tolerance of creativity to
the upper limits. To compensate for the new freedom enjoyed by translators/co-
drafters, Canadian lawyers have resorted to legal means by building checks and
balances into the judiciary system, thus enabling greater control over the
interpretation and application of parallel texts. In particular, the Supreme Court
has developed a sophisticated system of bilingual interpretation that enables
 Legal translation 291

judges to resolve interpretation problems with a high degree of pragmatism (see

Beaupré 1986:44-64; cf. the methods of multilingual interpretation at the
European Court of Justice in Volman et al. 1988; on freedom and constraints in
the translation of EU legislation, see Koutsivitis 1991:145-147). Fearing that
federal judges had too much discretion, lawmakers later enacted special
legislation intended to ensure that the two equally authentic texts of federal
legislation have the same effect in all parts of Canada (see Sarcevic 1997:274;
also Sarcevic 1989:287). Similar legislative measures were recently enacted in
Hong Kong, which has been officially bilingual since 1986 and bilegal since its
return to Chinese rule in 1997 (see Pasternak 1996:44). Such measures,
however, are by no means a carte blanche for unlimited creativity by
translators. Skilled translators/co-drafters always take account of the situational
factors of the particular act of communication when determining where, how,
and to what extent they can be creative and still respect the restraints of the
profession.

References

BEAUPRÉ, M. 1986. Interpreting Bilingual Legislation. Toronto: Carswell.

BOCQUET, C. 1994. Pour une méthode de traduction juridique. Prilly: CB Service.

COVACS, A. 1982. "La réalisation de la version française des lois fédérales du Canada".
Langage du droit et traduction ed. by J.Cl. GÉMAR, 83-100. Montréal: Linguatech & Conseil
de la langue française.

DIDIER, E. 1990. Langues et langages du droit. Montreal: Wilson & Lafleur.

DRIEDGER, E. 1982. A Manuel of Instruction for Legislative and Legal Writing. Ottawa:
Ministry of Justice.

DULLION, V. 1997. "Lorsque traduire, c'est écrire une page d'histoire: La version française du
Code civil suisse dans l'unification juridique de la Confédération". L'histoire et les théories de
la traduction, 371-388. Berne: ASTTI & Genève: ETI.

DURIEUX, C 1991. "Liberté et créativité en traduction technique". La Liberté et traduction

ed. by M. LEDERER & F. ISRAEL, 169-179. Paris: Didier.

GÉMAR, J.-CI. 1995. Traduire ou l'art d'interpréter, tome 2: Application. Sainte-Foy: Presses
de l'Université du Québec.

KOUTSIVITIS, V. 1991. "La Traduction juridique: liberté et contraintes". La Liberté en

traductioned.by M. LEDERER & F. ISRAEL , 139-150. Paris: Didier.

KUSSMAUL, P. 1995. Training the Translator. Amsterdam & Philadelphia: John Benjamins.
292 Susan Šarcevic

KUSSMAUL, P. 1998. "Types of Creative Translating". Volume of Abstracts, EST Congress -

Granada September 1998. [Full paper in the present volume, p. 117-126]

LEVERT, L. 1995. "Bijuridisme et bilinguisme législatifs: Un enjeu et un pari". Français

juridique et Science du droit ed. by G. SNOW & J. VANDERLINDEN, 255- 266. Bruxelles:
Bruyant.

NEUBERT, A. 1997. "Postulates for a Theory of Translatio". Cognitive Processes in

Translation and Interpreting ed. by J. DANKS, G. SHREVE, S. FOUNTAIN & M.
MCBEATH, 1-24. London: Sage.

PASTERNAK, V. 1996. Chinesisch als Rechtssprache in Hongkong. Frankfurt am Main: Peter

Lang.

SAGER, J. 1998. "What Distinguishes Major Types of Translation?" The Translator 4(1),
69-89.

SARCEVIC, S. 1989. "Conceptual Dictionaries for Translation in the Field of Law".

Internationaljournal of Lexicography 2(4), 277-292.

SARCEVIC, S. 1994. "Translation and the Law: An Interdisciplinary Approach". Translation

Studies, An Interdiscipline ed. by M. SNELL-HORNBY, F. PÖCHHACKER & K. KAINDL,
301-307. Amsterdam & Philadelphia: John Benjamins.

SARCEVIC, S. 1997. New Approach to Legal Translation, The Hague, London, Boston:
Kluwer Law International.

VERMEER, H. 1986. "Übersetzen als kultureller Transfer". Übersetzungswissenschaft - eine

Neuorientierung ed. by M. SNELL-HORNBY, 30-53. Tübingen: Francke.

VERMEER, H. 1996. A Skopos Theory of Translation. Heidelberg: TEXTconTEXT.

VOLMAN, Y., EIJSBOUTS, W.T. & VAN MONTFRANS, M. 1988. La cour et les langues.
Amsterdam: Universiteit van Amsterdam.

WILSS, W. 1988. Kognition und Ubersetzen. Tübingen: Niemeyer.

WILSS, W. 1996. Knowledge and Skills in Translator Behavior. Amsterdam & Philadelphia:
John Benjamins.
FRANK WEDEKIND'S SEX TRAGEDY LULU
IN ENGLISH AND FRENCH VERSIONS
THE CASE FOR AN INDIVIDUALISTIC VIEW OF
LITERARY TRANSLATION

Rainer Kohlmayer
University of Mainz, Germersheim, Germany

Zusammenfassung
Die Schlußszene von Wedekinds Lulu-Tragödie stellte die englischen, amerikanischen
und französischen Übersetzer vor eine schwere Aufgabe. Zum einen konnte die
schockierende Reihe von 'Freiern' religiöse, ethnische und kulturelle Normerwartungen
verletzen. Zum andern erwies sich die sprachliche Hybridität als ästhetisches
Rezeptionshindernis. Der Aufsatz gibt einen ersten Überblick über das breite Spektrum
der individuellen übersetzerischen Konzepte und plädiert für eine stärkere
Berücksichtigung der ästhetischen Autonomie der Übersetzer.

Résumé
La tragédie Lulu de Frank Wedekind (1864-1918) fut à l'origine très à l'avant-garde.
Le texte, qui mélangeait plusieurs langues, allait à l'encontre de bien des normes
morales et esthétiques de l'époque. Les traductions américaines, anglaises et françaises
présentent une multitude d'approches individuelles très différentes. Il est assez étonnant
que même les versions anglaises les plus récentes occultent complètement la polyphonie
linguistique de la pièce originale de Wedekind.

Resumen
El drama sexual de "Lulu" de Frank Wedekind (1894) fue el primer intento de los
escritores alemanes de crear una obra teatral realmente cosmopolita tanto por su tema
(el poder global del sexo) como por su polifonia lingüistica (al mezclar varios idiomas y
formas de expresión). Este articulo trata de las actitudes personales de los traductores
ingleses y franceses y de sus decisiones frente al texto rebelde de Wedekind.
294 Rainer Kohlmayer

1. The individualistic approach

The aim of my paper is mainly to describe how Wedekind's English and French
translators intervened linguistically, aesthetically, and ideologically in order to
make the Lulu-tragedy cross the linguistic and cultural border from Germany
into France, England, and the United States in the course of the 20th century.
My approach is, in certain respects, rather similar to that of the
Manipulation Group, especially to Lefevere's and Toury's widely accepted
thesis that translators tend to adapt their work to the norms of target cultures or
target cultural polysystems, i.e. mainly to the norms of the ruling aesthetics and
to the power structure of the ruling institutions. What I reject, however, at least
within the framework of literature, is the determinism that very often seems to
crop up in explanations of translators' and readers' behaviour. Changeable
attitudes, habits, and conventions are defined as norms, and it is frequently
taken for granted that the existence of a so-called norm is reason enough for a
translator to stick to it, or that the norm expectancy of theatre audiences or
reading publics is the most important basis of literary success or failure. Norms
are thought to govern literary translators' behaviour. Therefore the individual
translator's behaviour is thought to be of interest only insofar as he or she
follows or contradicts those pre-existing norms.
In my opinion, literature is definitely not a closed system with built-in
causalities. Literature is not made up of a universe of norms to be broken or
stuck to by writers and translators. Literary translators are, as a rule, not merely
functionaries of target norms or target groups. They have a considerable
amount of autonomy. Not even during the Nazi epoch in Germany, when the
German literary system was regulated as never before, did translators'
behaviour become entirely predictable. There were translations in favour of and
against the ruling powers of that time (see e.g. Kohlmayer 1994;1996a).
Taking into consideration the fact that the system approach cannot fully
explain, let alone predict a literary translator's behaviour, we ought to devote
more time and energy to finding out about literary translators as individual
persons and decision-makers. I do not deny the pressure of social norms, yet I
do not accept that literary translators are deconstructed into ego-less adherents
to, or mere fulfillers of, pre-existing norms. Literature and art are, as long as
they are original products, first and foremost something individual, and thus, by
definition, something un-systematic and even ab-normal.
I thus adhere to an "individualistic" theory of literary translation. For me,
all literary translations are, first and foremost, individual compromises. The
translator's basic maxim - a de facto Gricean maxim - is: "Be prepared to
compromise". The quality of the individual compromise depends, to put it in a
nutshell, on four characteristic features of the literary translator: on his or her
 Lulu in English and French versions 295

individual knowledge and skill, conscience and courage. (See also Kohlmayer
1996b, 1988; and Gombrich 1993.)

2. A brief survey of the German and English versions of Lulu

Today, Wedekind's Lulu exists in two standard German versions. According to

Wedekind's express wish in his preface to the 1913 edition of his collected
works, only the last edition, of 1913, was to be regarded as the final
("endgültig") and standard version of the Lulu-tragedy (Wedekind 1989:180),
and so it remained until ten years ago. That version was the result of almost
twenty years of revisions and adaptations of the original manuscript to the
claims of German censors, critics, and theatres. It consists of two separate parts,
Erdgeist and Die Büchse der Pandora, with four and three acts, respectively. It
was this version that was translated into English and French and several other
languages from 1914 onward. There are two American versions, the first one by
Samuel A. Eliot (1914), a second one by Carl Richard Mueller (1967), as well
as two British versions, one by Stephen Spender (1952), another by Charlotte
Beck and Peter Barnes (1971).
However, in 1988, exactly 70 years after Wedekind's death, another
version turned up: Wedekind's original and uncensored five-act manuscript of
Lulu, as he had written it in Paris and London from December 1892 to June
1894. The original version was premiered at the Deutsches Schauspielhaus
Hamburg in January 1988, with considerable success; a critical edition was
published in 1990 (Wedekind 1990), the English translation, by Edward Bond,
came out in 1993, and the American version, by Eric Bentley, in 1994.
I shall deal both with Wedekind's original text of 1894 and the two
recent English versions thereof, namely Bentley's and Bond's, as well as with
Wedekind's 1913 version and four of its translations into English (by Eliot,
Spender, Mueller and Beck & Barnes). I shall, however, limit myself to just a
few critical aspects.

3. Multilingualism and heterogeneity in Lulu

First I discuss which particular features of Wedekind's earlier and later versions
presented the most obvious difficulties for translators.

3.1 Linguistic polyphony

Perhaps the most striking and surprising feature of Wedekind's original text of
1894 is its linguistic hybridity. The first three acts, set in Germany, are written in
296 Rainer Kohlmayer

German. The fourth act, however, set in Paris, makes abundant use of the
French language. Most of the nouveaux riches, pimps, journalists, prostitutes,
etc. assembled in Act Four speak French. I quote a few characteristic lines from
the beginning of Act Four:

Madeleine de Marelle. Mais je ne puis pas le croire. C'est insensé. - On y était

souvent si gai. On y valsait, on y jouait, on y montait à cheval - on faisait les
cent coups. - Dire que tout est fini maintenant!
Bianetta Gazil. Moi, je n'y ai été qu'une seule fois - avec mon ancien amant, le
duc de Brétigny. Il aimait à y aller...
Rodrigo (das volle Glas in der Hand). Mesdames et Messieurs - excusez -
Mesdames et Messieurs - vous me permettez - soyez tranquilles - c'est le - (zu
Ludmilla Steinherz) Was heißt Geburtstag?
Ludmilla Steinherz. L'anniversaire...
Rodrigo. Heißen Dank. [...]
Bianetta Gazil De quel pays êtes vous?
Rodrigo. Je suis - Autrichien...
Madeleine de Marelle. Mais vous n'avez pas l'accent Autrichien...
Rodrigo. Madame - vous êtes un ange! (Wedekind 1990:74).

The fifth act, set in London, and ending with Lulu's murder at the hands of Jack
the Ripper, contains passages in German, French, Swiss German, as well as
more than 250 dialogue passages in English.
In one of his notebooks of the time of writing, Wedekind commented on
his use of different languages by referring to Shakespeare:

Shakespeare hat seinem Heinrich V. bereits einige Scenen in französischer

Sprache eingefugt. Wir leben drei Jahrhunderte nach Shakespeare und wir sind
Deutsche. Wir sind die gebildetste Nation der Welt (Wedekind 1990:210).

Apparently he thought at the time of writing that there might be a sufficiently

large circle of literary polyglots in Germany who would be able to read and
understand a multilingual text.
The deeper reason motivating Wedekind's mixing of languages was
certainly that he wanted to produce world theatre in a very real sense. The
choice of cities like Berlin, Paris, and London for the settings, and the use of
German, French, and English were in tune with Wedekind's universal message:
the alienation, repression, and destruction of female sexuality in the male-
dominated, capitalist world. In the later version, however, Wedekind translated
the foreign language passages back into German, thereby reducing his ambition
considerably. Only the Swiss dialect passages were left untouched.
 Lulu in English and French versions 297

3.2 Pidgin English

A second obvious difficulty for translators was Wedekind's pidgin English.

Whereas Wedekind's French in the fourth act is very good, the English in the
fifth act is quite poor. Wedekind had specially gone to England for half a year in
January 1894 in order to gain some first-hand knowledge of the milieu, and he
did actually write the fifth act during his stay in London. In the fifth act Lulu,
living with Schigolch and Alwa in a miserable attic somewhere in London, is
forced to prostitute herself in order to survive. Now, Wedekind's English in
Lulu is perfectly understandable despite numerous mistakes and recurrent
interlingual errors. Yet there is no doubt that Wedekind was aware of his
weakness in English. In a way, he turned it into an advantage by portraying only
characters whose broken English might have resembled his own. Nobody would
expect Lulu to converse in perfect English with her customers, nor could her
exotic customers, with the possible exception of Jack the Ripper, be expected to
have a native speaker's command of English.

Lulu. Why will you not stay all night with me, Sir?
Jack. I am married man.
Lulu. You say, you missed the last bus and that you have spend the night with
one of your friends...
Jack. Time is money. - How much do you want?
Lulu. Pound...
Jack. Good evening. - (geht zur Thüre)
Lulu (hält ihn zurück). Stay! - Stay! (Wedekind 1990:127).

Most of the time Wedekind's level of English is suitable for the characters he
portrays. In terms of aesthetics, the very lack of normal verbal communication
in the last act can be interpreted as an additional dimension of modernity. I
quote from Edward Bond's introduction:

Lulu barters with her customers in English. The play shows an African
speaking it. It is becoming the world language - English is the language of
capitalism. [...] The impression is strangely modern, like a multi-language film
with subtitles (Bond 1993:63).

3.3 Lulu's customers

A third feature creating problems for the linguistic transfer of the play into other
cultural systems was the heterogeneity of Lulu's four customers in Act Five and
the considerable critical potential inherent in their portrayal. The sequence with
Lulu's four customers in the last act can be seen as a kind of quick-motion
298 Rainer Kohlmayer

repetition of her four husbands. Lulu's customers represent four typical

variations of male sexuality, or, rather, four types of sexual repression and
exploitation. Certainly, Jack the Ripper is the most violent embodiment of
sexual repression of the four. Yet by putting him in a line with Lulu's three
previous customers, Wedekind points out a kind of 'family resemblance' of male
behaviour.

4. A wide spectrum of individual translational attitudes and decisions

I shall discuss Lulu's first three customers in some detail, together with the
corresponding translations, before commenting on Bond's and Bentley's recent
English versions.

4.1 Mr. Hopkins

The first "gentleman" (Wedekind 1990:128) Lulu brings up from the street is
Mr. Hopkins, a pious Christian, judging from the book Schigolch and Alwa find
in the pocket of his coat: "Lessons for those - who are - and those who want to
be - Christian Workers" (Wedekind 1990:113). According to Lulu, he is "totaly
[sic] mad" (Wedekind 1990:128). Mr. Hopkins does not only refrain from
uttering a single word himself, he also forbids Lulu to speak by laying a finger
to his lips, putting his hand over her mouth, and so on. Mr. Hopkins'
pantomime is not only funny, it is also a Nietzschean attack against Christianity:
Mr. Hopkins' strange behaviour demonstrates the Platonic splitting up of man
into body and soul. Mr. Hopkins' mind does not want to know or hear about
what his body is doing.
In the 1913 version Wedekind slightly emphasised the ideological
content of the scene by changing the title of the book to "Ermahnungen für
fromme Pilger und solche, die es werden wollen" (Wedekind 1989:165;
"Exhortations to pious pilgrims and those anxious to become such" in Spender
1952:162), thereby evoking the well-known Christian tradition of the soul's
pilgrimage on earth.
Interestingly enough, the French version of 1969 cut out the passage
with the prayer-book so that, for the French public, Lulu's customer was no
longer identifiable as the archetypal Christian (Jouve 1969:172). And,
surprisingly, the English version by Beck and Barnes in 1971 also neutralised
the anti-Christian potential of the scene, by transforming the character into a
"deaf mute", who "puts his forefinger to his lips and ears and shakes his head
making strange rasping noises" (Beck & Barnes 1971:75). This character was
obviously physically handicapped, whereas Wedekind's Mr. Hopkins was, as it
were, mentally or spiritually handicapped, i.e. handicapped by Christianity.
 Lulu in English and French versions 299

The other translators, however, did not interfere with or neutralise

Wedekind's original ideological message. Nor did Bond and Bentley.

4.2 Kungu Poti

Lulu's second customer, Kungu Poti, is another, far more embarrassing

caricature of male sexuality, this time of the brutal, vulgar, barbaric variety. He
is a black African, the son of the Sultan of Ouaoubée, he has six wives in
London, "three English and three French" (in the later version Wedekind adds
"two Spanish"), all of whom he declares "too stylish for me" (Wedekind
1990:119). In her conversation with Jack the Ripper, Lulu refers to Kungu Poti
as "a nigger" (128). It is quite clear that Wedekind's Kungu Poti is an extremely
negative stereotype of primitive sexuality. He actually wants to rape Lulu, and
kills Lulu's fourth husband before escaping.
The figure certainly poses enormous problems for translators with a
political, or, rather, intercultural conscience. It is remarkable that Wedekind, in
the subsequent version, made the figure of Kungu Poti even more repellent. In
the 1913 version Kungu Poti speaks a pidgin German that is more primitive than
in Wedekind's English version and, what is even more revolting, according to
Wedekind's stage direction, he is supposed to "belch" all the time (Wedekind
1989:171).
One might have expected that the clearly racist, Eurocentric portrait of
the barbaric African might have become ever more embarrassing for the English
translators in the course of the 20th century. If we have a look at the
translators' solutions we discover, however, that their behaviour is far from
uniform, let alone predictable. In 1914, the first American translator, Samuel A.
Eliot, Jr., tactfully transformed the belching into a "hiccough" (Eliot 1914:71),
whereas in 1952 Stephen Spender (Spender 1952:167) and, following him
rather closely in 1967, the American Carl Richard Mueller, translated quite
literally (Mueller 1967:160). Jean-Pierre Jouve, in 1969, introduced Kongo (as
he calls him) as "un nègre élégant" and duly omitted the stage direction with the
belching which would hardly have been in character (Jouve 1969:178).
Charlotte Beck and Peter Barnes, adapting the play for a performance at the
Nottingham Playhouse in 1970, made obvious efforts to render the character of
Kungu Poti less racist. He does not belch, he is less brutal, and he is both
funnier and wittier, leaving the stage with the line "d'coming King goes!". The
adapters even have Lulu run after him, calling "Come back" (Beck and Barnes
1971:78). Bond corrects Kungu Poti's broken English only minimally. Bentley,
on the other hand, lets him speak normal conversational English.
It is evident that we have here a wide spectrum of individual
translational attitudes and decisions, ranging from political recklessness
300 Rainer Kohlmayer

(Spender, Mueller) to a certain concern for political correctness avant la lettre

(Eliot, Jouve, Beck and Barnes).

4.3 Dr. Hilti

Lulu's third customer, Dr. Hilti, is the most complex character of the four
(Wedekind 1990:122-125). After the archetypal Christian and the extremely
barbaric variety of male sexuality, Wedekind presents us here with the capitalist
and calvinist variety of male sexuality. It is important to realize that Dr. Hilti
belongs to the uppermost layer of European society and culture. On the one
hand, he is a young professor of philosophy from Zurich, a "Darwinian", as he
says, coming from a wealthy aristocratic family, and about to get married to a
Swiss millionairess from Basel. On the other hand, he is absolutely illiterate in
emotional and erotic matters. For Dr. Hilti, sexuality is nothing but man's
instrument for the production of offspring. He comes upstairs with Lulu because
now, after getting engaged, he is in a hurry to find out how the thing works. He
is the quintessential image of what Max Weber, in his sociology of Calvinism's
Protestant ethic, termed "innerwordly asceticism" ("innerweltliche Askese"). All
these geographical, social, financial, and ideological features are symbolically
united in the strong Swiss German dialect he speaks.
A further difficulty for translators, on top of the cultural and linguistic
complexity of Dr. Hilti's character, is added by the multilingual conversation
between him and Lulu. I quote a short passage from Wedekind's original
manuscript in order to demonstrate Wedekind's stunningly avantgarde
technique of mixing pidgin English, French, Swiss German, and Standard
German:

Lulu. [...] But you are not English?

Dr. Hilti. No. -I am only here the last two weeks. - Are you borne in London?
Lulu. No Sir. -I am French...
Dr. Hilti. Ah, vous êtes Française?
Lulu. Oui monsieur, je suis Parisienne. [...]
Dr. Hilti. I am coming from Paris, where I was staling for eight days. [...] I am
from Zurich in Switzerland. [...]
Lulu. Alors vous parlez l'Allemand?
Dr. Hilti. Sprachän Sie töütsch?
Lulu. Un petit peu seulement, parce que mon ancien amant était Allemand. - Il
était de Berlin, je crois.
Dr. Hilti. Tonnärwättär - wia miach thas fröüt, thas Sie töütsch spreachän!
Lulu [...]. Komm, Süßer, komm. - Du bleibst bei mir die Nacht. (Wedekind
1990:123).
 Lulu in English and French versions 301

In the 1913 version the scene was cut by more than half, the multilingual
passages disappeared, and the mixture of languages was reduced to German and
Swiss German (Wedekind 1989:174f). But even so Dr. Hilti contained
sufficient snags for translators. Here is a brief review of the translators'
individual solutions.
The first American translator, Eliot, reduced Dr. Hilti's social status
substantially, perhaps because of the mistaken belief that Dr. Hilti's strong
Swiss dialect might somehow be a sign of vulgarity, whereas the very opposite
is true: it signifies wealth. So for Eliot, Dr. Hilti is no longer the puritan,
provincial product of Switzerland's financial aristocracy, he simply comes "of a
very old country family", and speaks "in the broadest north-country vowels"
(Eliot 1914:74). Besides, he is no longer a Privatdozent, but merely a "tutor".
His fiancée undergoes the same fate of social relegation: she is no longer a
wealthy Basel aristocrat's daughter with two million to her credit but owns
merely "a hundred thousand" (75). Whereas Wedekind's Dr. Hilti is a
provocative caricature of the puritan ideal which combines sexual abstinence
with social and financial success, Eliot's Dr. Hilti is merely a boorish provincial.
Stephen Spender, Carl Richard Mueller, Beck and Barnes, and Eric Bentley
translate Hilti's Swiss dialect into perfectly normal English, with only Bentley
adding a footnote to inform the reader about Hilti's Swiss dialect, "a form of
speech Wedekind found amusing" (Bentley 1994:194). He does not inform the
reader about the social and financial connotations of the Basel dialect. Beck and
Barnes cut the Hilti scene by about half. The French translator leaves it out
completely, and so does, amazingly, Edward Bond, although in his preface he
had praised Wedekind's multilingual modernity.
It is surprising that none of the English versions managed to do justice
to Wedekind's Dr. Hilti, linguistically and ideologically the most interesting
character in the last act.

4.4 Jack the Ripper

Lulu's last customer, Jack the Ripper, seems to have been less of a problem for
the English and French translators. The most obvious aesthetic and ideological
changes occurred in Beck's and Barnes' adaptation, where Jack the Ripper
becomes mythologized into something like a high priest of the theatre of cruelty.
The adapters add the following stage direction: "As in a dream he slowly raises
the bloodstained knife high like a chalice and licks it" (Beck & Barnes 1971:82).
302 Rainer Kohlmayer

5. Bond's and Bentley's different versions of Lulu

I conclude with a brief discussion of Bond's and Bentley's recent translations of

Wedekind's original version of the play.

5.1. Different interpretations

Both translations are, in my view, important works of art; both are accompanied
by original poems by the translators themselves, thereby demonstrating the
extent to which they had become involved creatively in their work, and both
translators offer detailed interpretations of the play, with Bentley's introductory
essay (and translation) reacting critically to Bond's previously published
interpretation. Bond's interpretation is Marxist, Bentley's is, broadly speaking,
Feminist. He sees Wedekind as the "playwright of a sexual revolution" (Bentley
1994:26).

5.2. A linguistic monoculture

Both translators do away with Wedekind's mixture of languages and dialects.

English is the only language spoken throughout, with no more than a handful of
French passages cropping up in the Paris Act. Thus both translations are, in
their internal linguistic structure, far more uniform and homogeneous than
Wedekind's original text of 1894. The linguistic hybridity of the original has
disappeared, despite the fact that for Wedekind, at the time of writing, it was a
characteristic feature of modern world literature. In a footnote Bentley explains:
"In Wedekind's original, the French characters speak in French. But one cannot
ask a British or American audience to accept this [...]" (Bentley 1994:127). In
operas, however, they often do accept it. Both Bond and Bentley seem to
favour the idea of the linguistic monoculture, at least as far as theatre texts are
concerned.

5.3 A correct English

Both translators correct Wedekind's weak English. Most of the time they
rewrite Wedekind's sentences even when they are perfectly correct, and most of
the time they do so irrespective of the person speaking. Bond is somewhat more
restrained in this respect, anxious to preserve some of the "strange,
hallucinatory effect" (Bond 1993:XXVII) of Wedekind's English in the last act,
whereas Bentley turns Lulu into quite a proficient speaker of English in her
dealings with her customers. Both translators seem to reject the idea of
presenting the readers with speakers of pidgin English.
 Lulu in English and French versions 303

5.4. A clear manipulation

Where Bentley keeps all of Wedekind's characters, Bond cuts out several
characters in Act Four (as well as Dr. Hilti in Act Five), in addition to changing
the end of Act Four. By cutting the complete story of a twelve-year-old girl
consenting to become a prostitute in Act Four, Bond reduces the sexual theme,
and, by adding a new ending to Act Four, he reinforces the anti-capitalist
content of the play, in keeping with his Marxist interpretation.

5.5 Conservative translations

Each translation and each interpretation is a coherent piece of work, the result
of an individual aesthetic and ideological compromise and concept. Yet Bond's
Marxist and Bentley's Feminist interpretations of Wedekind's Lulu coincide in
one important point: both eliminate or reduce the linguistic hybridity as well as
the heterogeneity of the characters in Wedekind's play. If cultural hybridity is an
important feature of our time, Wedekind's original play from 1894 must be
considered more avantgarde than the English versions published a hundred years
later.

References

BECK, C. & BARNES, P. 1971. Frank Wedekind. Lulu. A sex tragedy. Adapted by Peter
Barnes from Frank Wedekind's Earth Spirit and Pandora's Box. Translation by Charlotte
Beck, with an Introduction by Martin Esslin. London: Heinemann

BENTLEY, E. 1994. Frank Wedekind's The First Lulu. English version by Eric Bentley. New
York/London: Applause Theatre Books.

BOND, E. 1993. Frank Wedekind. Plays: One. Lulu: A Monster Tragedy, translated and
introduced by Edward Bond and Elisabeth Bond-Pablé. London: Methuen Drama.

ELIOT, S.A. 1914. Frank Wedekind. Erdgeist (Earth-Spirit). /Pandora's Box. Translated by
Samuel A. Eliot, Jr. GLEBE 2 (3/4). New York: Albert and Charles Boni.

GOMBRICH, E. H. 1993. A Lifelong Interest. Conversations on Art and Science with Didier
Eribon. London: Thames & Hudson.

JOUVE, P.J. 1969. Frank Wedekind. Lulu. Version française et adaptation par Pierre Jean
Jouve. Héricourt: Editions L'Age d'Homme.

KOHLMAYER, R. 1988. "Der Literaturübersetzer zwischen Original und Markt. Eine Kritik
funktionalistischer Übersetzungstheorien". Lebende Sprachen 33(4), 145-156.
304 Rainer Kohlmayer

KOHLMAYER, R. 1994. "Oscar Wilde's Society Comedies and the National Socialist
Message". New Comparison 17(Spring), 11-22.

KOHLMAYER, R. 1996a. Oscar Wilde in Deutschland und Österreich. Untersuchungen zur

Rezeption der Komödien und zur Theorie der Bühnenübersetzung (Theatron 20). Tübingen:
Niemeyer.

KOHLMAYER, R. 1996b. "Wissen und Können des Literaturübersetzers. Bausteine einer

individualistischen Kompetenztheorie". Übersetzerische Kompetenz. Beiträge zur
universitären Übersetzerausbildung in Deutschland und Skandinavien ed. by A. F.
KELLETAT, 187-205. Frankfurt: Peter Lang.

MUELLER, CR. 1967. Frank Wedekind. The Lulu Plays. Earth Spirit. Pandora's Box. Death
and the Devil. Newly Translated and with an Introduction by Carl Richard Mueller. New
York: Fawcett Publications.

SPENDER, S. 1952. Frank Wedekind. Earth-Spirit / Pandora's Box. Translated from the
German by Stephen Spender. London: Vision Press.

WEDEKIND, Frank. 1989 [=1913]. Lulu. Erdgeist. Die Büchse der Pandora. Herausgegeben
von Erhard Weidl. Stuttgart: Reclam.

WEDEKIND, Frank. 1990 [=1894]. Die Büchse der Pandora. Eine Monstretragoedie.
Historisch-kritische Ausgabe der Urfassung von 1894. Herausgegeben, kommentiert und mit
einem Essay von Hartmut Vinçon (Pharus III). Darmstadt: Jürgen Hausser.
 PART VII

Culture-bound concepts
SOBRE LA TRADUCIBILIDAD DE LOS
MARCADORES CULTURALES

Leticia Herrero
Universidad de Alicante, Espana

Abstract
Culture markers (culture-bound terms) are elements that are so tied to a particular culture
that exact equivalents cannot usually be found in another language, without the concept
losing its cultural identity. Various translation strategies are nevertheless available, which
respect the cultural origin of such terms to different degrees. The translator's decisions
about what strategies to use naturally affect readers' understanding of the translation, but
they also affect the image of the source culture which is created in the target culture.

Zusammenfassung
Kulturmarker (kulturgebundene Termini) sind Elemente, die eine so enge Verbundenheit
mit einer speziellen Kultur aufweisen, daß ohne Verlust der kulturellen Identität dieser
Konzeptionen keine exakten Äquivalente für sie in einer anderen Sprache zu finden sind.
Zur Verfugung stehen jedoch verschiedene Übersetzungsstrategien, die die kulturelle
Herkunft dieser Termini unterschiedlich berücksichtigen. Es ist klar, daß die vom
Übersetzer getroffeneWahl der Strategie das Verständnis der Übersetzung beim Leser
beeinflußt. Darüber hinaus wirkt sich die gewählte Strategie auf das Image der
Ausgangskultur in der Zielkultur aus.

Résumé
Les marqueurs culturels (ou termes culturellement marqués) sont des éléments tellement
liés à une culture qu'il est presque impossible de leur trouver dans une autre langue des
équivalents exacts sans qu'ils perdent leur identité. Néanmoins, existent différentes
stratégies traductionnelles qui respectent, à des degrés divers, leur origine culturelle. Les
décisions du traducteur, quant à la stratégie à adopter, affectent la compréhension du
lecteur mais aussi l'image de la culture d'origine dans la culture réceptrice.

A poco que el lector de este articulo se haya detenido frente a los

escaparates de las librerias o haya echado un vistazo a los suplementos literarios
de los periódicos, habrâ reparado, sobre todo de unos meses a esta parte, en la
308 Leticia Herrero

repentina atracción que parecen haber sentido las editoriales espanolas por las
literaturas de origen asiático, desde la india o hindu hasta la china y japonesa
(aunque la pasión por lo exótico demuestra llegar mås lejos y dar la vuelta al
mundo recuperando lo que, hasta el momento, habia sido apartado por diferente.
La literatura britanica actual ha descubierto entre sus escritores una
comunidad de artistas de origen indio que estân demostrando ser tan herederos de
Thackeray, Austen o Dickens como el más British de los escritores britânicos. El
enorme éxito obtenido en la propia Gran Bretana con libros firmados por Meera
Syal, Anita Desai, Vikram Seth o Rohinton Mistry, sin olvidar el reconocimiento
con el Premio Booker 1997 a la primera novela de Arundhati Roy, The God of
Small Things, son un aval más que suficiente para que las editoriales espanolas se
lancen a la búsqueda y captura de titulos escritos por autores angloindios.
Antes de que los nombres recién mencionados llegaran a Espafia, otros del
mismo origen ya lo habian hecho, pero los motivos que llevaron a traducir los
primeros titulos no coinciden con la politica actual que incita a su traduction, que
claramente responde a la demanda social por estos libros. Es ahora cuando
muchos traductores contratados por las editoriales comparten la ardua tarea de
transferir a nuestra lengua estos textos tan particulares cargados de referencias
culturales; y es esta coincidencia la que invita a retomar la eterna cuestión de los
limites de la traducción y reflexionar sobre los problemas que ineludiblemente
plantea la transferencia de elementos culturales entre lenguas.
Ciertamente, por la parte que les afecta, los Estudios de Traducción no
pueden permanecer ajenos a la creciente publicación de textos hïbridos que se
está produciendo en el mercado bibliográfico internacional. El término hïbrido se
aplica a la literatura de los escritores oriundos de las antiguas colonias para
enfatizar su vinculo con dos culturas, esto es, con una que les es propia por
nacimiento y con otra que lo es por adopción (en el caso que nos ocupa, la india y
la britanica respectivamente). Los rasgos definitorios de la novela angloindia son
radicalmente diferentes a los de un texto firmado por algunos ingleses
contemporâneos como Martin Amis o David Lodge, por lo que es lógico pensar
que la actividad traductora no planteará los mismos problemas en un caso y en
otro. Basândome en este razonamiento, las obras de Hanif Kureishi, Gita Mehta o
Amitav Ghosh constituyen mi punto de partida para reflexionar sobre la
traducibilidad de los elementos culturales.
Todos sabemos que el inglés, el idioma mâs difundido en todo el mundo,
es medio de comunicación incluso en lugares que originariamente hablaban otras
lenguas; el caso de India, donde la lengua de los britânicos permanece como
recuerdo de su pasado colonial, es paradigmâtico. Asi quedaba patente en una de
las multiples entrevistas que Arundhati Roy se viera obligada a atender por el
inesperado éxito de su novela, al describir su uso del inglés como una imposición:
"Después de 300 años de colonialismo" admite la escritora, "fue la lengua la que
me escogió a mi" ("Arundhati Roy. Reflejos de la India". El Pais semanal, 12-7-
98, pâg. 26).
 Marcadores culturales 309

India ha visto cómo el uso del inglés le abre una puerta al exterior que, de
lo contrario, estaria herméticamente cerrada para un mundo que no conoce las
otras dos lenguas mâs importantes del pais. El inglés, en el pasado un elemento de
opresión, posibilita hoy al estado hindu un presente cosmopolita y una creciente
proyección al exterior si consideramos el ritmo de expansion de esta lengua.
Como dijera Salman Rushdie, la lengua inglesa ha dejado de ser exclusiva de los
ingleses (Rushdie 1991: 70) y ahora está en boca de chinos, indios, africanos y
caribeños. Este hecho que los hindúes mâs ortodoxos consideran una contrariedad
y una falta de identidad es el arma de trabajo de los escritores postcoloniales; la
lengua inglesa que una vez los colonizó es ahora dominada por los autores y
puesta al servicio de sus propósitos. El periodo postcolonial es el tiempo de la
recolonización de la lengua y la cultura.
Si vamos a hablar de los problemas que surgen al traducir uno de estos
textos, creo conveniente ofrecer unas nociones, aun muy sucintas, sobre la
literatura postcolonial. La novela postcolonial tiene un poco del género
autobiogrâfico, otro poco del histórico, algunas pinceladas fantásticas, y todo ello
dispuesto de un modo que la critica ha creido semejante al realismo mágico. De
acuerdo o no con tal comparación, si es cierto que la narración se convierte para
el autor postcolonial en un proceso de creación de su propia historicia, historia
entendida en su doble acepción de sucesión de acontecimientos reales a lo largo
de un periodo, y relation de los hechos que conforman un relato. (Esta
ambivalencia lingüística esta ingeniosamente representada en el juego de palabras
inglés [his]story).
El escritor postcolonial recrea la historia de un modo subjetivo, desde
contar su infancia hasta recapitular episodios nacionales. Hace con la historia lo
que el borgiano personaje Pierre Menard con El Quijote: escribirla de nuevo
tiempo después, dedicar sus escrùpulos y vigilias a repetir en un idioma ajeno —
recordemos que el inglés le es propio por imposición— algo ya contado (Borges
1939: 280). La historia final asemeja a una especie de palimpsesto en el que
viejos trazos de otras plumas son todavia perceptibles. La verdad histórica, decia
Borges que dijo Menard segün habia dicho Cervantes, no es lo que sucedió, es lo
que juzgamos que sucedió. Y eso es lo que hace el escritor postcolonial, juzgar y
escribir los acontecimientos segün su propia visión de los hechos, escribir la
historia para demostrar el poder mediático del lenguaje que durante siglos ha
silenciado a las culturas minoritarias y marginales.
Inevitablemente, como consecuencia de esta representación personal, el
lenguaje que la transmite tiene que ser asimismo recreado. Si el lenguaje concede
expresión a la realidad cultural, hay una necesidad imperiosa de inventar un
lenguaje diferente para contar una realidad diferente. El Ingles con I mayúscula,
el Inglés estândar, no esta cualificado para ello y, en su defecto, deben recurrir al
inglés con /' minúscula. Para contar su historia, el escritor juega con la lingüistica,
rompe el lenguaje, arranca las palabras de su contexto natural y las contextualiza
en un nuevo entorno cultural.
310 Leticia Herrero

En tanto en cuanto el traductor no es el autor del texto origen, éste se le

presenta opaco, en ocasiones incluso inaccesible, y su traducción es más una
actividad hermenéutica que una de transferencia de lenguas. Traducir, ya lo
sabemos, no es sólo una actividad lingüistica; la lengua es la representación de
una visión del mundo y, en este sentido, la traducción se complica porque no es
fácil conciliar una lengua con una estructura cultural que no le es propia, siquiera
familiar. El espacio del otro no es siempre abordable desde la lengua meta; la
naturaleza del otro es exótica y ajena, y no siempre es posible llevârsela al lector
meta sin naturalizarla. Recordemos que el mismo Rabindranath Tagore, Premio
Nobel de Literatura en 1912, se tradujo a si mismo, subvirtiendo y manipulando
su propio discurso para ajustarlo al modelo occidental (dicen que escribia de
modo totalmente distinto en bengali y en inglés); de no haberlo hecho, quizâs
nunca se hubiera consagrado como poeta en este lado del mapa.
La actuación de Tagore ejemplifica el eterno conflicto del traductor,
reproducir de manera aceptable en la cultura receptora el contenido del texto
origen permaneciendo dentro de los limites del original. Los textos hibridos, a
caballo entre dos culturas, son un reto para el traductor; en ellos más que en otros,
el profesional descubre que la dificultad de todo proceso de traducción no es
atribuible únicamente al uso de signos lingüisticos diferentes en cada idioma.
La traducción supone siempre un enfrentamiento entre los valores, los
conocimientos y las sensibilidades de los dos sistemas culturales participes,
enfrentamiento que se manifiesta a lo largo del proceso de transferencia como
problemas de traducción. El concepto de problema de traducción deriva,
entonces, de la resistencia que presenta la cultura meta a la recepción de un texto
origen de expresión y contenido propios de otro sistema lingüistico-cultural, y su
origen puede ser lingüistico, pragmâtico, interpretativo y cultural.
Pero algo mâs: el problema de traducción deriva tanto de los
anisomorfismos entre los sistemas lingüisticos involucrados como de la politica
de traducción que dirige el trabajo de los profesionales y que, en la actualidad y
en nuestro pais, prescribe que una traducción ha de ser aceptable en la cultura de
recepción como si de un original se tratara. Traducir, entonces, no es sólo
cuestión de superar las asimetrias existentes entre las culturas, sino de conseguirlo
de un modo aceptable para el lector meta. Hay unas palabras de Valentin Garcia
Yebra que ilustran muy bien lo que pretendo decir, y sus conclusiones son lo
bastante contundentes para reproducirlas aqui. La cita dice asi:

"Supongamos que alguien quiere traducir una novela japonesa costumbrista. Al

lector nativo le parecerán del todo normales muchas de las situaciones y
conductas reflejadas en la novela; probablemente, le serân familiares los nombres
propios que aparezcan en ella. Al lector de esta novela traducida al español tales
situaciones, tales conductas, le parecerân sorprendentes, quizá incluso chocantes,
y los nombres propios le producirán una impresión extrana.
 Marcadores culturales 311

¿Qué debe hacer entonces el traductor? ¿Debe conservar en la traducción las

situaciones y los comportamientos chocantes, y la extraneza de los nombres
propios? En tal caso, no habrá equivalencia funcional. La lectura de la novela
producirá efectos muy dispares en el lector nativo y en el lector de la traducción.
Pero, si se sustituyen las situaciones, los comportamientos y los nombres propios
japoneses por situaciones, comportamientos y nombres propios familiares para
los lectores de la lengua terminal, se puede llegar a cambiar tanto la novela que
resulte 'otra', no 'la misma' en lengua diferente. Sera entonces una imitación, no
podrá ya llamarse traducción" (Garcia Yebra 1994: 388).

Si han tenido ocasión de hojear un original de Salman Rushdie, por

ejemplo, o el anteriormente mencionado The God of Small Things, tan vendido en
los últimos meses, habrân comprobado que dispersas por el texto asoman
palabras escritas en una lengua india que incomodan al lector interrumpiendo la
fluidez de la lectura. Es el rasgo más visible e inmediato de la novela angloindia,
sus páginas en inglés con continuas intromisiones en hindi, urdu, punjabi o tamil.
Estas palabras son la representación de la tradición y de la historia india, el
testimonio de su memoria colectiva, que no sálo producen significado, también lo
contextualizan; el gesto es reivindicativo de escritores apologistas de sus
origenes: escriben porque el mundo les ha ignorado a ellos y a sus culturas, y
ahora quieren presentar su historia sin filtros, sin occidentalizar, sin manipular. El
texto es asi un conjunto de elementos culturales compartidos por una comunidad
que los vincula a su experiencia; para el lector versado, son elementos
profundamente connotativos en tanto coherentes con sus conocimientos, para el
profano, carecen incluso de denotation y son pura forma lingüistica.
El nombre con el que se conoce a estas palabras de contenido
especificamente cultural es diverso según el autor: realia (Bödeker y Fresse 1987;
Koller 1992), cultural terms (Newmark 1982), cultural features (Nida 1964),
kulturerne (Oksaar 1988), culture-marker (Nord 1994), y en nuestro ámbito
nacional, elementos culturales especificos (Franco 1996), referencias culturales
(Mayoral 1994), segmentos textuales marcadamente culturales (Mayoral y
Munoz 1997). Yo me referiré a ellos como marcadores culturales especificos. A
pesar de no ser los unicos responsables de las areas de inequivalencia
interlingüïstica, los marcadores culturales si parecen ser los mâs propensos a
provocar problemas de traducción y, por tanto, a limitar la consecución de la
equivalencia.
Es importante subrayar el carâcter virtual del concepto marcador cultural
ya que, en ningùn momento, podemos asegurar que todo marcador cultural
supone siempre un problema de traducción. Es cierto que son elementos cuyo
contenido muestra potencialmente cierta resistencia a la transferencia, pero esa
potencialidad puede desaparecer en el proceso de traducción concreto.
No es prâctico ni real asociar los marcadores culturales con la idea de
inequivalencia porque "del mismo modo que la equivalencia absoluta no existe"
312 Leticia Herrero

nos tranquiliza Rosa Rabadân, "tampoco existe la inequivalencia total" (Rabadán

1991: 110). Ésta, la inequivalencia, no es un fenómeno que se pueda definir de
manera aislada; sólo es delimitable y defendible sobre elementos
contextualizados y durante el proceso de transferencia a otra lengua. Resulta más
acertado y pragmâtico hablar de los limites de la equivalencia, entendiéndola
como una relación entre los textos dinâmica y dependiente de las circunstancias y
los elementos participes en el proceso de traduction. En resumen, es muy difïcil
delimitar el concepto de marcador cultural. La relatividad de toda situación
comunicativa es una razón sólida para definirlo sólo en su contexto, es decir,
obtener una definición descriptiva del mismo.
Por fortuna, las propuestas para solucionar las dificultades que plantean
estos elementos son más concretas y expresas que su definición. En general, las
estrategias de traducción se resumen en tres posibilidades: (1) incorporar el
elemento original sin modificar en el texto meta; (2) traducirlo lingüisticamente,
es decir, intercambiar los significantes originales por significantes de la lengua
meta que estén relacionados semânticamente; (3) manipular el elemento para
preservar su valor funcional en el texto. Las tres admiten diversos grados de
ejecución que oscilan desde la técnica mâs conservadora de la no traducción o
repetición del marcador cultural hasta otras mâs sustitutorias como la
neutralization o, incluso, tan domesticadoras como la naturalización. En la escala
que limita la actuación del traductor, las técnicas del polo de la conservatión
velan por la adecuación al texto origen; las del polo de la sustitución conceden
gradualmente mayor autonomia al texto traducido frente a su original y priman la
aceptabilidad de aquél entre sus receptores.
Es conveniente ilustrar la teoria con algunos ejemplos, que he tornado de
traducciones espanolas. El caso de la no traducción o repetición del marcador
cultural es el mâs dramâtico, pues no se percibe ningún esfuerzo por transmitir el
contenido. Asi ocurre en el ejemplo n° 1:

(1) —He encontrado algunos sellos interesantes para Kety —dijo Soli
—Soli me ha traido un Scinde Dawke —balbució Kety con un
repentino estallido de animation.
(R. Lucas, Lafàbrica de hielo, pág. 73).

A menos que al lector de la novela le guste el mundo de la filatelia, es

difïcil que reconozca el elemento «Scinde Dawke»; no obstante, a pesar de su
opacidad, en este caso, el contexto facilita informatión que permite desentranar
parte del significado, como pasa en el ejemplo n° 2, en que podemos deducir que
Sashi Kapoor es un personaje atractivo, aunque no conozcamos quien es Sashi
Kapoor.

(2) —Te diré que por dentro no eres feo.

—Si, claro, por dentro soy igualito que Shashi Kapoor.
 Marcadores culturales 313

(H. Kureishi, El buda de los suburbios, pâg. 358)

A lo largo de las novelas, son muchos los casos en que el lector puede
encontrarse perdido ante una referencia cultural, sin contar aquéllos que, debido
al perfil del lector modelo de estas traducciones —un lector de alto nivel
cultural—, son supuestamente reconocibles (por ejemplo: Old Vic, Whitehall,
Indira o Barbara Cartland).
Tampoco la traducción lingüistica del marcador es siempre conveniente
pues, en ocasiones, el resultado confonde al lector, quien reconoce
lingüisticamente el elemento meta, pero semânticamente se le escapa el referente,
como ocurre con los ejemplos n° 3, 4 y 5:

(3) [TO] elephant obedience trials —» [TM] pruebas de obediencia de

elefantes
(R. Lucas, La fábrica de hielo, pág. 163/199)

(4) [TO] black toothpowder → [TM] polvos dentifricos negros

(H. Kureishi, El buda de los suburbios, pâg. 31/45)

(5) [TO] the free radio scheme was a dead duck —» [TM] el plan de la
radio gratis habia fracasado
(S. Rushdie, Oriente, Occidente, págs. 26/29)

La última posibilidad de las tres con las que cuenta el profesional en la

traducción de los marcadores culturales, la de la manipulación del marcador, es la
más agresiva para el original, tal como observamos en la neutralización absoluta
del origen cultural del ejemplo n° 6, en la naturalización del n° 7 y en la omisión
del n° 8:

(6) [TO] the Everyman of the twentieth century —» [TM] el personaje

corriente del siglo veinte.
(H. Kureishi, El buda de los suburbios, pág. 141/182)

(7) [TO] penniless → [TM] sin un céntimo

(S. Rushdie, Hijos de la medianoche, pâg. 118/172)

(8) [TO] (...) instructing Jockey to drive the sahib to the police-station →
[TM] le dio instrucciones a Jockey para que le llevase a la comisaria.
(R. Lucas, La fábrica de hielo, 102/128)

Es fácil observar que todas tienen sus desventajas. El resto de estrategias

de traducción repartidas gradualmente entre las tres posibilidades mencionadas
314 Leticia Herrero

también esconden inconvenientes: por ejemplo, he leido textos en que la lista de

asteriscos remitiendo a un glosario final es tan larga que un comentario del
original hubiera resultado más cómodo que su traduccion; y he encontrado
neutralizaciones parciales que en lugar de hacer más transparente el marcador,
como cabria esperar, lo complican; asi ocurre en el siguiente ejemplo:

(9) [TO] the P.M.'s letter → [TM] el titulo de médico director

(S. Rushdie, Hijos de la medianoche, 15/25)

No es correcto decir que unas estrategias sean mejores que otras, pues, en
primer lugar, habria que especificar para qué o para quién son mejores. La
estrategia de traducción mas conveniente, mas adecuada o más funcional será
aquélla que el traductor asi lo estime en la situatión interlingüistica global. El
análisis de los marcadores culturales en la novela angloindia me ha permitido
observar la cantidad de condicionantes que determinan la labor del traductor. La
naturaleza del nombre, si es propio o comün, —generalmente, el grado de
conservatión es mayor con los nombres propios que con los comunes—, el
historial interlingüistico entre las culturas involucradas —las versiones prefijadas
son respetadas prácticamente siempre—, o el universo cultural de donde proviene
el marcador —los traductores tienden a ser más conservadores con los
marcadores hindúes y más sustitutorios con los británicos— son algunos de los
condicionantes de traduccián de los marcadores culturales, pero es posible
encontrar muchos otros factores que influyen en la decisión del traductor como el
grado de opacidad comunicativa o el de explicatión.
La traducibilidad dependerá, por tanto, del marcador, su contexto y la
situation de transferencia. Hay un ejemplo que me parece tremendamente
ilustrativo de la comodidad con la que se pueden resolver algunos problemas de
traducción; lo cita John T. Brewer en su articulo "The Role of 'Culture' in
Successful Translation" (1988: 24) y esta tornado de una novela alemana. La
escena narra la huida de un asesino del lugar del crimen tras borrar sus huellas y
eliminar todas las pruebas. La puerta de la habitation esta cerrada; el original
describe cómo el criminal la abre con el codo para evitar dejar marcas. Este
detalle, que en el entorno cultural origen de la novela resulta verosimil, para un
lector americano supondria una hazaña por parte del asesino, puesto que
prâcticamente todas las puertas en Estados Unidos tienen pornos redondos que
difïcilmente se prestan a ser abiertos con el codo.
En el caso anterior, la intervention del traductor en el texto supone un
cambio insignificante; el propio Brewer sugiere colocar un panuelo en el bolsillo
de la chaqueta del asesino con el que poder tocar el porno sin dejar impregnadas
las huellas - gesto, por otro lado, muy recurrente en el cine negro. Pero ¿qué
posibilidades encuentra el traductor cuando el problema de traducción tiene un
alcance mayor, cuando supera los limites de la novela y afecta a la forma de
pensar y de vivir de una comunidad cultural?
 Marcadores culturales 315

Mi respuesta, me temo, es tan clara y contundente como decepcionante:

no lo sé. La traducción, que siempre ha pretendido servir de vinculo entre las
culturas, parece rendirse ahora ante la dificultad de unirlas y enfatiza con mas
fuerza sus diferencias. Con elementos claramente contextualizados en su entorno
cultural, nos vemos obligados a desistir ante la imposibilidad de transferir en
términos familiares para el lector meta una realidad que pertenece exclusivamente
a la comunidad cultural origen. La manipulación del texto se presenta, en casos
asi, como la solución más inmediata; incluso el mas estricto de los profesionales
difîcilmente producirâ un texto meta coherente para el receptor si en algün
momento no recuire a ella. El problema es que asi, con la manipulación del texto,
la traducción controla la imagen del original, y cuando esta actuación afecta
reiteradamente a una misma cultura, la traducción puede llegar a convertirse en
un proceso de manipulación cultural.
Por tanto, si reconocemos abiertamente los limites de la traducción y
aceptamos que no todo lo que se dice en una lengua puede decirse en otra sin que
pierda la identidad cultural en el nuevo texto, habremos avanzado en el propósito
de dar voz a todas las culturas, de representarlas en su propio lenguaje.
Precisamente, para luchar contra la mediatización de otras culturas a través del
lenguaje, a finales de 1996, vio la luz un proyecto de la editorial Macmillan en
India que lleva por titulo "Modern Indian Novels in [English] Translation". El
objetivo primordial de esta colección de libros es hacer llegar al lector ajeno a
India representaciones literarias fidedignas sobre el pais y su cultura. Los libros
seleccionados han sido escritos en diversas lenguas indias - bengali, hindi,
punjabi, urdu, tamil, etc. - y su traducción al inglés la lleva a cabo un traductor
indio, lingüistica y culturalmente nativo. El motivo que induce al equipo editorial
a elegir un traductor nativo es mantener el contenido original de la obra a pesar de
las dificultades que ello conlleve. Es casi lógico suponer que el traductor inglés
podria desanimarse ante una realidad india que no encuentra fácil de expresar en
su lengua y optaria por manipular el contenido, por britanizar el texto origen. El
traductor indio, sin embargo, recuire a notas a pie de pagina, introducciones y
todas las herramientas que tenga a su disposition para no deshacer el sabor local
del texto, para explicar, por ejemplo, qué es "Holi" sin conformarse con decir que
es un festival, o que "the Indian Mutiny" no fue simplemente una rebelión india.
Iniciativas como la de Macmillan hacen reflexionar sobre la subyugación
a la que se ha sometido a ciertas literaturas - entiéndase culturas - marginales. No
debiéramos permitir que esto siguiera sucediendo. En una recensión del nuevo
libro publicado por V. S. Naipaul, (India. A Wounded Civilization. Traducción
española: India. Una civilización herida, Debate, 1998. Trad. de Flora Casas)
Carlos Pujol escribia:

"Kipling es siempre muy inglés al hablar de la India [...] y sin duda por eso lo
entendemos muy bien; no sólo habla un idioma muy accesible [...], sino que lo
ve todo a través de unas gafas inglesas [...]. Naipaul quiere presentarnos sus
316 Leticia Herrero

personajes tal cual son, no tal como los podemos comprender en Europa, se
identifica con ellos, entra maravillado en su comprensión intima, por lo común
enigmática para nosotros; señala, registra, deja hablar, reúne apasionadamente
materiales que no interpreta [...]. El escenario es tan vasto e intrincado que su
actitud desalienta un poco al lector, que echa de menos esa mania tan occidental
de simplificar las cosas para hacernos la ilusión de que las dominamos.'' (ABC
literario, 30-1-1998, pág. 11)

Repito: "esa mania tan occidental de simplificar las cosas para hacernos la
ilusión de que las dominamos". Creo que ya esta todo dicho. Como escribiera el
Nobel Saul Bellow en una de sus novelas: "For God's Sake, open the universe a
little more". La traducción, o en tal caso, la no traducción de los términos
culturales es un paso hacia ello.

Referendas bibliográficas
BORGES, J. L. 1939. "Pierre Menard, autor de El Quijote". Obras completas, 274-281.
Barcelona: Circulo de Lectores.

BREWER, J. T. 1988. "The role of 'Culture' in Successful Translation". Literature in Translation.

From Cultural Transference to Metonymic Displacement, ed. by P. TALGERI & S. B. VERMA,
21-26. Bombay: Popular Prakashan.

FRANCO AIXELA, J. 1996. Condicionantes de traducción y su aplicación a los nombres propios

(Inglés-Espanol). Universidad de Alicante. Tesis inédita.

GARCIA YEBRA, V. 1994. Traducción: historia y teoria. Madrid: Gredos.

KUREISHI, H. 1990. The Buddha ofSuburbia. London: Faber and Faber. Traducción espanola: El
buda de los suburbios. Barcelona: Anagrama, 1994 (1996). Trad. de Mónica Martin Berdagué.

LUCAS, R 1992. The Ice Factory. London: Heinemann. Traducción espanola: La fäbrica de
hielo. Barcelona: Anagrama, 1995. Trad. de Maribel de Juan.

NORD, Ch. 1994. "It's Tea-Time in Wonderland: Culture-markers in Fictional Texts".

Intercultural Communication. Proceedings of the 17th International L.A.U.D. Symposium, ed. by
H. PÜRSCHEL, 523-538 Frankfurt: Peter Lang.

RABADAN, R. 1991. Equivalencia y traducción. Problemática de la equivalencia translémica

inglés-espanol. León: Servicio de Publicaciones de la Universidad de León.

RUSHDIE, S. 1981 (1995). Midnight's Children. London: Vintage. Traducción espanola: Los
hijos de la medianoche. Madrid: Alfaguara, 1984 (1989). Trad, de Miguel Saenz.

RUSHDIE, S. 1991. Imaginary Homelands. Essay and Criticism 1981-1991. London: Granta.

RUSHDIE, S. 1994 (1995). East, West. London: Vintage. Traducción espanola: Oriente,
Occidente. Barcelona: Plaza y Janés, 1997. Trad. de Miguel Saenz.
EL MUNDO DE [LA FILO]SOFIA
LA TRADUCCIÓN DE CONCEPTOS CULTURALES Y
LA FUNCIÓN DEL TEXTO - EL CASO DE LA
TRADUCCIÓN DE EL MUNDO DE SOFIA DE JOSTEIN
GAARDER.

Åse Johnsen
Universidad de Bergen, Noruega

Abstract
The paper analyses the English and Spanish translations of Jostein Gaarder's novel
Sophie's World (Norwegian original Sofies verden, 1991), focusing on the work's
double function as indicated by its subtitle: A Novel about the History of Philosophy.
The two translations differ in their strategies for translating cultural aspects. The
English version seems more to emphasize the novel's didactic function as a history of
philosophy, while the Spanish version presents it more as a work of fiction set in
Norway.

Zusammenfassung
Der Artikel besteht aus einer Analyse der englischen und spanischen Übersetzung des
Romans Sofies Welt (norwegisches Original Sofies verden, 1991) von Jostein Gaarder.
Im Fokus der Betrachtung steht die in dem Untertitel indizierte Doppelfunktion des
Romans: Roman über die Geschichte der Philosophie. Die zwei Übersetzungen
unterscheiden sich in bezug auf die gewählten Strategien zur Übersetzung kultureller
Aspekte. Die englische Version scheint in größerem Maß die didaktische Funktion des
Romans als eine Geschichte der Philosophie hervorzuheben, während die spanische
Version den Roman eher als ein in Norwegen situiertes fiktives Werk präsentiert.

Résumé
Notre contribution analyse les traductions anglaise et espagnole du roman de J.
Gaarder: Le monde de Sophie (original norvégien Sofies Verden, 1991), en mettant
l'accent sur la double fonction de l'ouvrage, telle qu'elle est révélée par son sous-titre:
Roman sur l'histoire de la philosophie. Les deux traductions diffèrent dans leurs
stratégies pour rendre les aspects culturels. La version anglaise semble faire ressortir
318 Åse Johnsen

davantage la fonction didactique du livre comme histoire de la philosophie tandis que la

version espagnole le présente plus comme oeuvre de fiction située en Norvège.

Considerando el proceso de traducción como un proceso de toma de

decisiones entre diferentes alternativas de expresión en la lengua y cultura meta,
la presente comunicación discutirâ el papel que juega la función o las funciones
del texto en ese proceso y cómo el traductor, mediante sus elecciones, puede
manipular el énfasis que se les da a las diferentes funciones del texto en la
versión traducida. También se verá que el acercamiento funcionalista de la
traducción puede ser el más adecuado también para los textos literarios.
La primera decision que ha de tomar el traductor es elegir el método de
traducción: una traducción exotizante o una adaptación. La discusión sobre
estos dos métodos se remonta a los tiempos de Cicerón (ano 46 aC), de San
Jerónimo en el siglo IV y de Schleiermacher en el siglo IXX. Esta dicotomía
todavia sigue vigente en los estudios de traducción con distintos nombres:
House distingue entre overt y covert translation (1977:106-107), Reiss entre la
traducción filológica y la comunicativa (1981:125) y Nord entre documentary
translation (traducción documento) e instrumental translation (traducción
instrumento) (1991:72-73). Común a esos pares de métodos es que los
primeros (fig. 1) se caracterizan por ser una especie de documento en donde el
lector meta (LM), perteneciente a una situación de comunicación meta (SM), es
testigo de una comunicación anterior entre el emisor (E) y el lector origen (LO),
dentro de una situación sociocultural origen (SO), y los segundos (fig. 2), al
contrario, producen un texto meta que sirve como instrumento para una nueva
comunicación directa entre el emisor origen y el lector meta.

Figure 1. Figure 2.

Segün Schleiermacher, el método exotizante es mås apto para la traducción de

textos literarios y cientificos, y el método de adaptación para los textos de la
vida cotidiana. Para los funcionalistas, sin embargo, no es el tipo de texto el que
debe decidir el método a seguir, sino la función que la traducción ha de tener en
 El mundo de la [filo]sofia 319

la cultura meta. Los funcionalistas toman como punto de partida que una
traducción es un acto comunicativo intercultural, y que como todo acto
comunicativo tiene su propósito. El propósito o la función del texto origen
(TO) puede diferir del propósito o la función del texto meta (TM), y lo que
debe guiar al traductor en su elección de método no es la función del TO, sino
la función o el propósito que ha de tener el TM, o sea, el skopos de la
traducción. Segûn Nord, la convención actual de la traducción literaria sigue
siendo la traducción documento y exotizante, con la excepción de muchos libros
infantiles (Nord, 1997:103). Prueba de aquello son también las fuertes criticas
que hubo en Noruega cuando se supo que en la traducción al inglés de El
Mundo de Sofia muchas referencias a la cultura noruega habian sido adaptadas
a la cultura angloparlante/norteamericana. Nord (1997:80) propone un
acercamiento funcionalista a la traducción literaria para mostrar que, teniendo
en cuenta los aspectos funcionales, una traducción instrumento o una
adaptación es muchas veces la que da el resultado más adecuado para la
situación comunicativa meta.
Tomando como punto de partida lo arriba expuesto, decidi acercarme a
la traducción inglesa de El Mundo de Sofia para ver si se trataba de una
tendencia etnocéntrica por parte de los angloparlantes o si las elecciones de la
traductora podian basarse en otro fundamento, por ejemplo un acercamiento
funcionalista a la traducción. Como punto de comparación elegi la versión
española de la novela.
La novela fue publicada por el escritor noruego, Josten Gaarder, en
1991. Gaarder ha sido profesor de filosofia, pero en Noruega es conocido sobre
todo por sus novelas para adolescentes. En El mundo de Sofia, que tiene como
subtitulo: Novela sobre la historia de la filosofia, Gaarder combina bien sus
dos profesiones, ya que el libro es una novela de ficción para adolescentes, y a
la vez una introducción a la historia de la filosofia para el mismo público. Asi
por lo menos se ha de entender la intención original del texto, aunque el libro
llegó a ser un best-seller y ha llegado a fascinar a un publico que va mâs allá de
los adolescentes noruegos. En poco tiempo llegó a alcanzar gran éxito también
entre un publico adulto y en pocos años la editorial vendió los derechos de
traducción a nada menos que 43 lenguas. No es nada raro que fuera un escritor
noruego, y ademâs profesor de filosofia, quien haya tenido la idea de escribir
una introducción a la historia de la filosofia para adolescentes, dado que en
Noruega la filosofia no entra en el curriculum ni de la escuela bâsica ni de los
institutos, impartiéndose solamente como asignatura optativa en algunos
institutos o escuelas populares. Sólo es en la universidad, donde la filosofia es
una asignatura obligatoria.
En El Mundo de Sofia conocemos a Sofia, una chica de 15 años, que un
dia empieza a recibir cartas del filósofo Alberto Knox, quien, a lo largo de la
novela, va introduciendo a Sofia a la historia de la filosofia mediante cartas,
videos, encuentros nocturnos, etc. A mitad del libro se introduce a otra chica,
320 Åse Johnsen

Hilde, y a su padre Albert Knag, que parecen ser personajes de fictión dentro
de la realidad ficticia del libro. A medida que avanza la historia, resulta, sin
embargo, que la "realidad" es el mundo de Hilde y su padre, y la fiction el
mundo de Sofia y Alberto Knox. Resulta, ademâs, ser el padre de Hilde quien,
mediante el mundo de Sofia, quiere iniciar a Hilde en los grandes pensamientos
filosóficos, tal como se puede pensar que ha sido la intention de Gaarder:
introducir a su hijo de 15 años entonces y a los demâs adolescentes noruegos a
la historia de la filosofia.
Como novela, el libro tiene, por tanto, una estructura compleja, ya que
se juega con la "realidad "/fictión dentro de la fictión, pero la novela también
tiene una parte o función didâctica dada por el subtitulo: Novela sobre la
historia de la filosofia. Y es basândome en esa doble función del libro: la
novela como fictión y como texto didâctico, como yo me aproximo al análisis
de las traducciones de la novela.
Segün los functionalistas, la función del texto meta no necesariamente es
la misma que la función del texto original. ¿Cuál es el caso de El Mundo de
Sofial Si bien la función del texto original, es la de ensenar la historia de la
filosofia mediante un código artistico, las traducciones o los textos metas, ¿han
de tener, necesariamente, la misma función? Los jóvenes en el mundo
hispanohablante quizâ no lean el libro como una introduction a la filosofia, dado
que estudian la filosofia en el bachillerato o COU, sino mâs bien como una
novela de fictión sobre una chica noruega que vive en una situatión
sociocultural noruega. Los lectores norteamericanos, que al igual que los
noruegos, generalmente no entran en contacto con la filosofia como disciplina
escolar hasta llegar a la universidad, quizâ no lean el libro como un libro
noruego sino como una introductión a la filosofia, etc. Tal como veremos
seguidamente, el traductor, mediante su manera de traducir puede, hasta cierto
punto, decidir como el lector meta ha de leer el texto traducido, enfocando o
bien la fictión o bien la historia de la filosofia con su función didâctica.
En 1997, la traductora Gülay Kutal, que ha traducido la novela al turco,
publicó un articulo en el que critica a la traductora Paulette Möller por haber
cometido graves faltas en la traductión al inglés (Samtiden num. 2/3, 1997).
Señala, entre otras cosas, que se ha traducido una mention a la novela Victoria
de Knut Hamsun por Of Mice and Men de John Steinbeck, Henrik Wergeland
por Byron y hasta en una ocasión se ha reemplazado Noruega por Francia. El
articulo causó muchas reacciones en Noruega y cambiar a Hamsun por
Steinbeck fue caracterizado como una estupidez (Hansen, Aftenposten,
29.05.97) y casi tan criminal como asesinarlo (Kutal, Samtiden, num. 2/3,
1997). En un principio, el autor mismo se adhirió a la critica, pero en un articulo
posterior en la mencionada revista Samtiden (num. 5/6, 1997), Gaarder subraya
que los cambios que se han hecho en la versión inglesa pueden defenderse, ya
que se ha intentado mantener el mensaje filosófico existente detrâs de los
ejemplos.
 El mundo de la [filo]sofia 321

Las reacciones que hubo en Noruega también se pueden ver en relación

a que Noruega es una nación pequena, orgullosa de su literatura y su herencia
cultural, las que incluyen nombres tan conocidos como Henrik Ibsen, Knut
Hamsun y Edvard Grieg. Con sólo 4,5 millones de noruegoparlantes la
traducción es imprescindible para poder difundir nuestra cultura.
Los aspectos interesantes de la critica son, por lo tanto, los culturales, la
distancia que puede haber entre el mundo del texto y el mundo real de los
respectivos lectores metas y qué soluciones se han elegido para reducir la
distancia entre estos dos mundos en las traducciones.
Según Nord (1997:87) hay tres posibilidades de distancia cultural entre
el mundo del texto y la cultura meta:
-El mundo del texto corresponde a la realidad de la cultura origen.
-El mundo del texto no corresponde a la realidad de la cultura origen,
siendo un caso especial de este grupo cuando el mundo del texto corresponde a
la cultura meta.
-El mundo del texto corresponde a la cultura origen, pero ha sido
desculturado mediante referencias explicitas a otra realidad: "Erase una vez en
un pais lejano..."
¿A qué realidad y a qué mundo corresponde entonces El Mundo de
Sofial A mi modo de ver, hay dos mundos en esta novela: El mundo de Sofia y
el mundo de la filosofia. El marco de la historia contada en la novela es
Noruega y la realidad de dos chicas jóvenes de nuestros dias: Hilde y Sofia. La
otra parte de la novela, la historia de la filosofia, sin embargo, es una historia
reconocible para todos los lectores, y es la historia de la filosofia de la Europa
occidental.
Dado que Sofia es una joven noruega, muchos de los ejemplos que
emplea Alberto Knox para explicarle las ideas filosóficas a Sofia estân tornados
de la cultura noruega. De manera que en algunas ocasiones los dos mundos
coinciden parcialmente o se entrecruzan:

Figure 3.

En esta intersección se refleja si el traductor traduce el texto solamente como

novela de ficción o si también tiene en mente la función didâctica del texto. La
322 Åse Johnsen

ficción, es decir la historia sobre Sofia e Hilde, causa problemas de traducción

de los aspectos culturales como en cualquier libro de ficción que se traduzca a
otra lengua/cultura. La parte de la filosofia, la que presenta una realidad menos
ficticia y más "verdadera", representa un mundo que los lectores meta
comparten con los lectores origen y su traducción no parece haber causado
grandes dificultades en lo que se refiere a las referencias culturales del mundo
de la filosofia occidental. En ninguna de las traducciones a las que me he podido
acercar (la inglesa, la espanola, la alemana, la sueca, la francesa y la árabe), ha
dejado de ser Sofia una chica noruega ni la filosofia ha dejado de ser la filosofia
de la Europa occidental. Las referencias culturales que aparecen en la
intersection, donde se cruzan los dos mundos, tienen, sin embargo, un trato
distinto en la traducción inglesa y la espanola.
Para ilustrar lo arriba expuesto, paso a mencionar algunos ejemplos
sacados del capitulo sobre la Edad Media de la novela.
Hablando sobre cómo Europa llegó a cristianizarse y cómo se
desarrollaron las diferentes naciones durante la Edad Media, Alberto Knox le
pregunta a Sofia: "¿Qué habria sido Europa sin la Edad Media? ¿Una provincia
romana?" Y sigue: (En los ejemplos el subrayado y la negrita son mios.)

(l)Men den klangbunn som ligger i navn som Norge, England eller
Tyskland, er nettopp dette bunnløse dyp som kalles middelalderen. Det
svømmer mange feite fisker i dette dypet selv om vi ikke alitid klarer å fa
øye på dem. Men Snorre var et middelader-menneske. Det var Olav den
hellige også. Og Karl den store. For ikke å nevne Romeo og Julie,
Benedikt og Årolilja, Olav Åsteson eller trollene i Heddal-skogen.
(p.173)

(la)La resonancia que tienen nombres como Inglaterra, Alemania o

Noruega se encuentra precisamente en esta inmensa profundidad que se
llama Edad Media. En esta profundidad nadan muchos peces gordos,
aunque no siempre los veamos. Snorri fue un hombre de la Edad Media,
también lo fueron Olaf el Santo* y Carlomagno. Por no decir Romeo y
Julieta.

Olaf el Santo, rey de Noruega, se convirtió al cristianismo e impulsó la

religión en el pais. Fue asesinado en la batalla de Stiklestad en 1030.
(N.de las T.)(p. 209)

(lb)Yet the resonance in such names as England, France or Germany

is the very same boundless deep we call the Middle Ages. There are
many shining fish swimming around in those dephts, although we do not
always catch sight of them. Snorri lived in the Middle Ages. So did
 El mundo de la [filo]sofïa 323

Saint Olaf and Charlemagne, to say nothing of Romeo and Juliet, Joan
of Arc, Ivanhoe, the Pied Piper of Hamelin, (...) (p. 143)

En primer lugar se ve que la versión espanola altera el orden de los paises

mencionados: 'Noruega, Inglaterra o Alemania', poniendo a Inglaterra en primer
lugar y a Noruega en el último; la versión inglesa, por su parte reemplaza a
Noruega por Francia y altera también el orden. En las versiones sueca, alemana
y francesa no hay ningún cambio, es decir, hay una traducción literal. Los
personajes que se mencionan como elementos importantes de la cultura noruega
de la Edad Media, es decir 'Olav den Hellige' (el rey Olaf el Santo), 'Benedikt y
Årolilja, y Olav Åsteson' (personajes de la poesia medieval) y 'trollene i Heddal-
skogen' (los trolls en los bosques de Heddal), tienen un trato muy diverso según
las traducciones. La versión espanola tiene una nota a pie de página que explica
quién fue Olaf el Santo, y veremos que esa nota también tiene relevancia para el
ejemplo siguiente. Los tres últimos elementos, pertenecientes a la literatura
noruega, son omitidos sin mâs en la version espanola, mientras que la versión
inglesa propone elementos conocidos para un lector angloparlante. Las demás
traducciones proponen varias soluciones, entre ellas Tristan e Isoida, los
Nibelungos, Abelardo y Eloisa, Blancanieves o los trolls en los bosques
noruegos entre otros, siendo la mayoria adaptaciones a elementos conocidos
para los lectores meta.

(2)Til Norge kom ikke kristendommen før på 1000-tallet, men det ville
være en overdrivelse å si at Norge ble et kristent land etter slaget på
Stiklestad (p 173)

(2a)El cristianismo no llegó a Noruega hasta el ano 1000, pero seria una
exageración decir que toda Noruega se convirtió en pais cristiano
después de la batalla de Stiklestad. (p. 209)

(2b)Christianity didn't come to Norway, by the way, until the eleventh

century. It would be an exaggeration to say that the Nordic countries
converted to Christianity at one fell swoop. (p. 144)

En este ejemplo, la versión espanola es una traducción fiel al original, y en la

nota del ejemplo anterior se ha explicado qué sucedió en la batalla de Stiklestad
y cuândo se produjo ésta, mientras que la traducción inglesa reemplaza Noruega
por los paises Nórdicos, tratando asi Noruega y los paises Nórdicos como una
entidad, y suprime la mención de la batalla de Stiklestad, poniendo una
expresión idiomâtica para decir "de un solo golpe". Las demâs traducciones
mantienen tanto Noruega como la batalla de Stiklestad.
El ultimo ejemplo es el mâs curioso por ser el que causó mayor numero
de criticas y también por ser el que mâs claramente muestra el enfoque de las
324 Åse Johnsen

traducciones, como función novela fiction o función didáctica. El filósofo

Alberto Knox le explica a Sofia la relación entre la obra de creación de Dios y la
Biblia según Santo Tomás de Aquino. Cito de la novela espanola, que es una
traducción fiel al texto noruego (p. 184), salvo la nota a pie de pagina, anadida
por las traductoras.

(3) - Déjame añadir otra imagen. Si lees una novela, por ejemplo
Victoria de Knut Hamsun ...
- De hecho la he leido...
- ¿Conoces algo sobre el autor leyendo simplemente la novela que ha
escrito?
- Al menos puedo saber que existe un autor que la ha escrito.
- ¿Puedes saber algo más de él?
- Tiene una visión bastante romántica del amor.
- Cuando lees esta novela, que es creación de Hamsun, obtienes una
impresión de la naturaleza de Hamsun. Pero no puedes contar con
encontrar datos personales sobre el autor. Por ejemplo, ¿puedes saber
mediante la lectura de Victoria la edad que tenia el autor al escribir la
novela, dónde vivia o cuântos hijos tenia?
- Claro que no.
- Ese tipo de datos los podrás encontrar en una biografia sobre Knut
Hamsun. Solamente en una biografia, o autobiografia, sabrás mâs acerca
del autor como "persona".
- Si, asi es.
- Más o menos asi es la relación entre la obra de creación de Dios y
la Biblia. Sólo mediante la observatión de la naturaleza podemos
reconocer que hay un Dios. No resulta dificil ver que ama las flores y los
animales, si no, no los hubiera creado. Pero sólo en la Biblia
encontramos informatión sobre la persona de Dios, es decir, en su
"autobiografia".
- iQué ejemplo más bueno!

*Knut Hamsun (1859-1952). El novelista noruego mâs importante de todos los

tiempos. Premio Nobel de Literatura en 1920. (N. de las T.) (pp. 223 - 224).

En la traducción inglesa se ha cambiado el ejemplo de Victoria de Knut

Hamsun por Of Mice and Men de John Steinbeck, cambiando también la frase: -
Tiene una visión bastante romántica del amor. por He seems to care about
outsiders. (p. 153)

De esa manera, la traducción inglesa cumple la función didâctica: se

compara la relación entre Dios y la Biblia con una novela y su autor, conocidos
para un publico angloparlante, y asi el ejemplo sirve como comunicación directa
 El mundo de la [filo]sofîa 325

sin que haya desviaciones a la cultura noruega. La traducción espanola, sin

embargo, con la nota a pie de página, desvia la atención del lector de las ideas
de Tomás de Aquino hacia información bibliogrâfica sobre un escritor noruego
que sólo sirve como un ejemplo en la novela, y que en toda la novela solamente
aparece en esta ocasión.
Hay mâs ejemplos en la novela, pero el espacio no me permite
comentarlos aqui. Para resumir se puede decir que la versión inglesa parece
haberse traducido con una clara función didâctica, adaptando las referencias
culturales relevantes para la comprensión de la filosofia. La versión espanola,
sin embargo, parece haber dado mâs énfasis a la función de la difusión de la
cultura noruega, ya que en las notas y explicaciones se añade información sobre
los personajes noruegos, acercando asi al lector a la cultura noruega. La función
de difiusión de la cultura noruega no fue, sin embargo, una intención inicial por
parte del autor, ya que no sabia que se iba a traducir a tantas lenguas. El autor
mismo subraya que no debemos olvidar que El mundo de Sofia ante todo se
basa en una idea pedagógica y que las referencias a la cultura noruega se
mencionan como ejemplos de exponentes noruegos de ideas culturales
europeas. Dice además que: 'Yo no sabia en aquel momento que El mundo de
Sofia se iba a traducir a tantas lenguas' (Samtiden, 5/6, 1997:103), queriendo
con eso quizâ decir que, de haberlo sabido, hubiera elegido otros ejemplos mâs
universales. Gaarder dice, sin embargo, que está orgulloso de haber llevado a
personajes noruegos a muchos rincones del mundo mediante la novela.
Tal como yo lo veo, si se quiere mantener la función didâctica en este
caso, el método mâs adecuado una traducción-instrumento, produciendo asi un
texto que se dirija directamente a los lectores meta. Si asi se hace, la novela no
dejarâ de ser una historia de ficcion que tiene lugar en Noruega, ni Sofia dejarâ
de ser una joven noruega, ni la historia de la filosofia dejarâ de ser la filosofia de
todos los tiempos, pero la novela seguirâ manteniendo la función didâctica del
texto original. Si, por el contrario, se elige un método exotizante o una
traducción-documento, puede que el texto pierda su función didâctica,
desviando la atención del lector desde la comprensión de la filosofia a elementos
de la cultura noruega poco o nada relevantes para la filosofia en si.
Las dos versiones de El Mundo de Sofia aqui analizadas son muestras
de que el traductor, mediante su election de método de traducción, puede
manipular la finition del texto traducido, y también de que, como bien dice
Nord, la traducción-instrumento o adaptación puede ser la mâs adecuada
también para un texto literario.
326 Åse Johnsen

Referencias bibliográficas
GAARDER, J. 1991. Sofies verden. Roman om filosofiens historie. Oslo: Aschehoug.

GAARDER, J. 1991. El mundo de Sofia, Novela sobre la historia de lafilosofia. Traduccion

al espanol por Kirsti Baggethun y Asunción Lorenzo. Madrid: Ediciones Siruela.

GAARDER, J. 1993. Sofies Welt. Roman über die Geschichte der Philosophie. Traducción al
alemån por Gabriele Haefs. München: Carl Hanser Verlag.

GAARDER, J. 1993. Sofies Värld. Roman om filosofins historia. Traducción al sueco por
Mona Erikson. Stockholm: Rabén & Sjögren.

GAARDER, J. 1995. Sophie's world. A Novel about the History of Philosophy. Traducción al
inglés por Paulette Møller. (Published in arrangement with Farrar, Strauss & Giroux Inc.)
London: Phoenix.

GAARDER, J. 1995. Le monde de Sophie. Roman sur l'histoire de la philosophie. Traducción

y adaptación al francés por Hélène Hervieu y Martine Laffon. Paris: Éditions du Seuil.

GAARDER, J. 1995. 'alam sofi (1996) Traducción al arabe por Haiat Al Houaiek Aatia.
Stockholm: Dar Al-Muna.

GAARDER, J. 1997. "Kommentar". Samtiden 5/6 1997.

HANSEN, E. 1997. "Verden minus Ivar Aasen". Aftenposten, Oslo 29.5.97.

HOUSE, J. 1977. "A Model for Assessing Translation Quality". Meta 22(2), 103-109.

KORSVOLD, K. 1997. "Laget for å nå amerikansk ungdom", Aftenposten, Olso 29.5.97.

KUTAL, G. 1997. "Hva er galt med Sophie's World?" Samtiden 2/3

MØLLER, P. 1997. "Sophie's World - Et gensvar". Samtiden 5/6

NORD, C. 1991. Text Analysis in Translation. Amsterdam & Atlanta: Rodopi.

NORD, C. 1997. Functionalist Approaches Explained. Manchester: St. Jerome.

REISS, K. 1981. "Understanding a text from the translator's point of view". The Bible
translator 32 (1), 124-134.

REISS, K & H. J. VERMEER: (1984/1991, 2 ed.). Grundlegung einer allgemeinen

Translationstheorie. Tübingen: Niemeyer. Traducción al espanol por Sandra Garcia Reina y
Celia Martin de León: 1996. Fundamentos para una teoria funcional de la traducción.
Madrid: Akal Ediciones.
INSTITUTIONAL REFERENCES FROM THE
ENGLISH-SPEAKING WORLD IN THE SPANISH
PRESS

Anne Martin
University of Granada, Spain

Zusammenfassung
In dieser Studie werden die Mechanismen der spanischen Presse, mit denen sie sich auf
politische, wirtschaftliche und soziale Institutionen englischsprachiger Länder bezieht,
analysiert. Die Untersuchung basiert auf einem Corpus verschiedener Textabschnitte, die
unter vier Gesichtspunkten analysiert wurden: von der Textangabe ausgehend, nach Themen
organisiert, unter dem Aspekt der Übersetzunsmethoden und schließlich nach Zeitung und
Autor. Das Ergebnis der Analyse zeigt, daß eine große Tendenz zur wortwörtlichen
Übersetzung besteht, sowie zu Anleihen, obwohl innerhalb dieser allgemeinen Tendenz doch
interessante Unterschiede zwischen den verschiedenen Zeitungen auftreten.

Résumé
Cette étude analyse les mécanismes déployés par la presse écrite espagnole pour se référer
aux institutions politiques, économiques et sociales des pays anglophones. Elle s'appuie sur
un corpus d'extraits de textes qui ont éte analysés en quatre temps: référence originale,
thèmes majeurs, procédés de traduction et, finalement, quotidien et auteur. Les résultats
montrent une forte tendance à la traduction littérale ou à l'emploi d'emprunts, même si, au
sein de cette tendance générale, on constate des différences intéressantes entre les journaux
analysés.

Resumen
Este trabajo analiza los mecanismos desplegados por la prensa escrita española a la hora de
referirse a las instituciones politicas, económicas y sociales de los paises de habla inglesa. Se
basa en un corpus de segmentes de texto que se analizaron en cuatro fases: por referencia
original, por temas afines, por procedimientos de traducción y finalmente, por periódico y
autor. Los resultados demuestran una tendencia mayoritaria hacia la traducción literal asi
como los préstamos, aunque dentro de esta tendencia general, se constatan interesantes
diferencias entre los periódicos analizados.
328 Anne Martin

1. Introduction

This paper reports on a research project which analyses the translation strategies
used by the main Spanish daily newspapers when referring to the institutions of the
English-speaking world. The aim of the project was to describe the strategies used
and establish whether any pattern in that use emerged, and if so, to examine any
possible underlying causes.
Our initial interest in this subject was a result of the impression that the
Spanish reader with no knowledge of the English language and customs of at least
the U.S and U.K. would be at a loss to understand many of the issues covered by
the daily press regarding the English-speaking world. We were surprised by the high
number of translations that did not seem to reflect the source culture adequately or
in which the original English term was most obvious, to the detriment of clarity in
Spanish. This trend seemed especially obvious when dealing with those aspects of
British life which have or are perceived to have a specific cultural or semiotic
weight, and presumed to have no target-culture equivalent. Our hypothesis was
therefore that these references were not usually translated adequately. In theory, we
felt that combinatory solutions such as couplets, triplets and functional formulations
were the stategies most suited to informative journalism (Martinez Albertos
1989:69), but our impression was that they were not widely used in the press and
that this was detrimental to clarity.
We thus sought to verify this hypothesis through a corpus-based study
compiled in April 1992 from the three main Spanish dailies at that time: El Pais,
ABC and Diario 16. The text segments registered in the corpus are mainly from
articles by the foreign correspondents of these three newspapers. The final corpus
has a total of 756 entries, each containing what we have defined as an institutional
reference, that is a reference to an institution related to the social, political,
economic or sporting life of a country (Dictionnario de la Real Academia Española,
DRAE 1992:175).
The analysis was carried out in the framework of the functionalist theory of
translation (Reiss & Vermeer 1984, 1996; Nord 1991), viewing translation as an
offer of information in one language about information that has been offered in
another, and the sociological approach to journalism (Martinez Albertos 1989,
1991) of the Spanish Escuela Complutense.

2. Analysis of the corpus: Method and phases

Once the corpus was complete, it was analysed from four different angles. The first
phase of analysis involved completing all the information concerning each entry in
the corpus, and consequently we felt it was necessary to find the original reference
designated by each of our entries. This was a prerequisite if we were to analyse the
 Institutional references 329

treatment each had received in transmission into Spanish. In many cases, the original
reference was of course obvious. However, in others, it was difficult and sometimes
even impossible to determine, as with the case of Comité de Revisión Civil {Diario
16). (Corpus examples are given in bold; underlining indicates italics in the original.)
This stage enabled us to establish which original references had received more than
one translation in Spanish, and this in turn enabled us to compare the different
versions.
The second stage of analysis examined the corpus from the standpoint of
translation strategies or options. This phase responded directly to the objective of
this project, which was to systematise the description of the different methods used
to transmit these references in Spanish. We must point out that in this phase, and
indeed in general, we adopted a product-based, descriptive approach, similar to that
of Franco Aixelá (1995), Mason (1994a, 1994b, 1995) and Mayoral & Muñoz
(1996). Thirdly, the corpus was analysed on the basis of subject area, that is to say
grouping together the terms or expressions whose originals came from the same
subject area, as is common in literature on the subject of translation procedures, for
example all the political parties, all the names of government departments
(Newmark 1981:73, 75-77; 1988:99-100; Santoyo 1987:48,49; Álvarez Calleja
1991:226). Finally, the corpus was analysed according to the origin of each
reference, that is to say by newspaper and by author.
Whilst the first two stages of analysis - by original reference and translation
procedure - aimed at establishing how the references had been dealt with and
whether there was any set pattern in the treatment they had received, the last two
stages of analysis - by subject area and by newspaper and author - were aimed at
revealing possible motivations behind this treatment.
We found evidence of nine general procedures in the corpus. These are not
to be considered as water-tight compartments, and there are many cases of overlap
between the categories. But they do represent general trends.

3. Results

Our nine categories are as follows:

3.1 Loans (represented in 25.39% of the corpus entries)

In our corpus we found many loans of all kinds. It was inevitable that we should find
proper nouns and in fact they were most numerous, ranging from the names of
newspapers, TV and radio networks, museums, galleries to football teams:

el "Sunday Times" (Diario 16), el The New York Times (El Pais), la Music
Televisión (El Pais), el San Francisco Museum of Modern Art (ABC), los Chicago
330 Anne Martin

Bulls (El Pais), el Liverpool (El Pais).

We also found many words and expressions simply left in the original:

el "New York Stock Exchange" (ABC), Oxbridge (ABC), constituency (Diario 16).

3.2 Literal Procedures (21.56% of the corpus)

Under literal procedures we included both calques and morphological translations or

cognates:

"entrismo" de "militantes" (en el Partido Laborista) (ABC), estación central de

policia (ABC), la Oficina Galesa (Diario 16).

3.3 Explanations /Couplets (13.22% of the corpus)

In this category, we included all the solutions with two or more component parts,
whether one of the components was in brackets or not. Many different combinations
of the component parts were found, for example, original expression and initials in
brackets, loan plus translation, translation plus functional explanation. The common
denominator in this category was that in all of the combinations, the transparency of
at least one of their components permitted the identification of the original English
reference:

Centra para el Control de Enfermedades (CDC) (El Pais), National Heritage

(Patrimonio Nacional) del Reino Unido (El Pais), "Carta de los Ciudadano", uno de
los inventos del "majorismo" (ABC).

3.4 Established Translations (7.9% of the corpus)

This category includes all those translations sanctioned by use, even though they
may be loans, calques etc. In fact, the majority were literal translations:

Departamento de Estado (El Pais, Diario 16), el Partido Conservador (El Pais,
Diario 16, ABC), la Cámara de los Comunes (El Pais, Diario 16, ABC).

3.5 Functional Versions (7.8% of the corpus)

This category refers to neutral versions, that is to say, solutions that explain the
function of the English original, without necessarily mentioning the original name of
the reference through a loan or literal translation, for example:

sindicato de actores del Reino Unido (El País), sindicato de futbolistas (El Pais), el
 Institutional references 331

departamento del Ministerio britanico del interior que controla el consumo de

estupefacientes y su clasificación (ABC), el banco central estadounidense (El Pais).

3.6 Classifiers (7.4% of the corpus)

The only combinatory solutions not included in category 3 above. The following are
typical examples of this category:

"el periódico The New York Times" (El Pais), "la cadena CNN" (El Pais).

3.7 Habitual Translations (6.08% of the corpus)

We included this category because we felt that there were various non culture-
specific institutional references that already had their set translation, such as alcalde
for Lord Mayor, reina for queen, senado for senate, and which needed to be
accounted for in some way. There is a difference between established translations
and this category since the former tends to be used for references which exist
exclusively in the culture of origin but which are widely known in the target culture,
for example Partido Demócrata, Partido Laborista, whereas in this category we
included universal references that are not exclusive to the source culture:

concejal (El Pais), bolsa (El Pais, Diario 16, ABC), senador (El Pais, Diario 16).

3.8 Partial Translations (4.16% of the corpus)

We included this category because of the large number of entries, which are
probably very common in the press, made up of two parts: one generic part which is
translated, together with a proper noun, for example:

centra Médico de Hackensack (El Pais), Universidad de Kansas (ABC), Museo de

Arte Moderno de San Francisco (El Pais).

We wished to distinguish this procedure from those entries in which the entire
expression was left in the original, which occurred in many cases.

3.9 Cultural Equivalence (3.43% of the corpus)

This category includes approximate translations which refer to a person or entity

carrying out a similar or identical function in the target culture:
332 Anne Martin

la federación inglesa de fútbol (El Pais), parlamento autónomo (ABC, Diario 16),
socialistas británicos (ABC).

4. Discussion of results

We feel that two clear tendencies emerge from the corpus: solutions that are mainly
explanatory, and those that favour identification of the original over explanation of
its meaning or function. Having said that, we do not wish to imply that those
solutions which tend towards identification do not communicate functional
information, or vice versa. In any case, the corpus bears witness to the fact that, on
numerous occasions, combinations of these two purposes appear to be present. As
we have stated above, we felt that those solutions that gave maximum information
in a discreet way would be the most suited to the aims of informative journalism
(Martinez Albertos 1989:69; 1991:1002). Therefore, these would be the
combinatory solutions. We also felt that functional formulations would in principle
be more suitable than literal procedures.
As we had anticipated, this theoretical ideal was not entirely borne out by
the results of our analysis. The two most common options found in our corpus are
those that, at least in principle, give precedence to identification: literal translation
(21.56%) and loans (25.39%). This trend is also reflected in the combination
solutions, where literal translations and loans make up the majority of the
components.
We even noted some cases in which certain literal translations which actually
have a different meaning in Spanish have become so firmly entrenched with use that
they can no longer be expressed in any other way and the two meanings coexist. An
obvious example of this is Secretario de Estado norteamericano, the firmly
established way of referring in Spanish to the U.S. Secretary of State, even though
Secretario de Estado means something quite different in the Spanish government
hierarchy. This example also shows us the importance of the press in disseminating
certain usages. On the other hand, the more communicative solutions were not so
frequent, especially those that did not include a specific reference to the original
term or expression.
As far as the distribution of these different translation solutions is concerned,
there was a tendency to use the least explanatory solutions in the economics sections
of the three dailies, and this tendency was even more marked in the sports pages. It
almost seemed as though this was not a question of non-specialist information but
rather the transmission of specialized information to an expert reader. Curiously
enough, this trend was also apparent in the cultural sections of ABC and Diario 16.
Obviously, the classification as we have described it so far does not take into
account the communicative value in context of the solutions adopted. Solutions that
are included in the same category may be suited to the communicative needs of one
 Institutional references 333

contextual situation and not to another. We therefore felt it was necessary to carry
out complementary analyses in order to take the context into account. Such analyses
included grouping together all those references that had received various different
translations within one article. By grouping together these sequences, it became
apparent that often different translations had been used for questions of clarity, or as
a style-varying device. Some translations which, seen in isolation, appeared to be
rather obscure, became clear when seen as part of a sequence. Often a full
explanation had been offered on the first appearance in the same text and the
variation was simply due to stylistic reasons. This phenomenon was most frequent in
El Pais,

5. Differences between newspapers

We shall now go on to discuss the following phase of analysis, that is, analysis by
newspaper and by author. The results of this stage show that, like the general
corpus, the most common strategies used in each individual newspaper are literal
translation and loans. However, despite this similarity of the results in general terms,
there are certain disparities between the different newspapers, which perhaps
constitute the most interesting aspects in the search for some kind of motivation
behind the use of one option or another. Functional and explanatory solutions in El
Pais total 22.12%) of its total corpus, whereas they only account for 14.56%) of
ABCs total corpus. At the same time, the figures for literal procedures are slightly
lower in El Pais than in the other two publications. Diario 16 seems to follow a
middle course between the other two dailies, except in the case of partial
translations, which is a solution that it hardly used at all, and the non-classifiable
entries, the vast majority of which are to be found in Diario 16.
On closer examination, the values of the different solutions within each
category used by each paper become apparent. Many of the loans in El Pais are
names of sporting teams or initials, the meaning of which had already been explained
in a couplet or triplet earlier in the text. By contrast, in ABC, the loans registered are
of an entirely different nature: names of buildings, clubs or cultural centres, the
generic part of which was left in the original, for example The National Gallery of
Art, San Francisco Museum of Modern Art, el Metropolitan Correctional Center
or institutions like Eton and Oxbridge, transcribed without any form of explanation,
as are institutions whose names could very easily have been translated into Spanish,
such as "Hong Kong Trade Development Council" or "New York Stock
Exchange". In Diario 16 the patterns are not as clear cut. Generally speaking the
majority of the loans are metonymic, although one correspondent consistently
transcribes original words and terms which he perhaps perceives as being culturally
unique; this is not in fact the case, as an adequate translation could have been found
for many of these terms, which include constituency, devolution and hung
334 Anne Martin

parliament.
The use of unusual terms in the target language can create the totally false
impression of a difference between cultures that in fact does not exist. The use of
such source-culture lexis hides the fact that the concept itself does in fact exist in the
target culture, attributing a semiotic status to the source term which it actually does
not have. This seems to be particularly the case with British political institutions with
archaic names but similar functioning to their European counterparts.
This difference of value also exists in the category of literal procedures.
Whilst in El Pais, the literal translations tend to be fairly clear, transparent and
connotation-free, 'in Diario 16 and ABC we observe a mixture of clear and adequate
literal translations with those that were simply incorrect or nonsensical. In ABC for
example, we found the following:

entrismo de militantes en el Partido Laborista; estación central de policia;

comisionado de policia; master en artes; doctorado enfilosofia;cabeza de estado;
manifiesto for election manifesto.

In Diario 16, we found Ministro de Deportes, creating the idea in Spanish that there
is a Secretary of State for Sports in the U.K., and la Oficina Galesa, creating the
idea that there is no ministry in charge of Welsh affairs.
This trend is reinforced when we examine the type of loan and literal
translation that formed part of the combinatory solutions. In El Pais, the majority of
these are composed of original initials together with a functional or literal translation
of the meaning of those initials, e.g. Servicio Nacional de Salud (NHS). By contrast,
in ABC most of the combinations consist of literal translations plus explanation.
Some of the explanations offered are rather curious, for example: Carta de los
Ciudadanos, carta de naturaleza del "majorismo " o farol sin consecuencias -
según vaya todo, offering very little real information. Having said this we must point
out that such solutions cannot be considered as a majority trend. In the couplets and
explanations found in Diario 16, the same kind of loans mentioned above were
apparent {hung parliament, constituency, devolution), although on this occasion
they were accompanied by an explanation.
It is interesting to consider the reasons why these differences should arise. In
the case of El Pais, we feel that the results are due in part to the existence of a style
book (1990) that stipulates clear guidelines with respect to any explanations to be
offered, for example regarding the use of initials and foreign words or expressions.
There is also a clear distinction made between the different genres ofjournalism. It is
curious that failure to comply with the style book regulations is notorious in El Pais,
and is even alluded to and denounced by the paper's own journalists themselves
(Arias in El Pais, 25/6/95). However, our results show a clear difference in so far as
consistency and harmonization of style is concerned if we compare El Pais to the
other two newspapers studied. Furthermore, the number of journalists signing
 Institutional references 335

articles from which the entries in our corpus were extracted in El Pais is higher than
in the other two dailies, a situation which could have led to greater fragmentation of
criteria in El Pais. Such fragmentation is present in Diario 16. Diario 16 supposedly
has a twenty-page internal style book (Fernández de Beaumont 1988:391) which we
were unable to obtain and the existence of which was even denied on several
occasions by Diario 16 journalists. We were therefore not at all surprised that this
newspaper proved to have the widest variety of solutions, including many of those
that we could not classify, in addition to a total lack of consistency in the use of
italics, inverted commas and capitals. In short, Diario 16 gives the impression of
having fewer internal style regulations together with the absence of any kind of
common philosophy about the kind of journalism the paper represents. As far as
ABC is concerned, there was no style book at the time our corpus was compiled,
although the widespread custom was to follow the guidelines established in the
Manual de Espanol Urgente published by the Agencia Efe press agency (1992,
1994).
We have mentioned the existence of a common philosophy of journalism
and we consider that part of the motivation behind the use of certain translation
solutions could be related to this aspect. This is very clearly stated in El Pais' style
book (El Pais 1990:15-21) and we feel that the results we have mentioned stem
from this clear statement of intent concerning the type of journalism to be pursued
by El Pais. We are of course referring to the higher number of functional (that is to
say, communicative) solutions, the objective nature of explanations, the greater
consistency of solutions regardless of the author of the article, consistency in the use
of markers, greater number of sources, together with other aspects such as the clear
difference between the different genres (information, opinion and analysis) which is
fundamental if a newspaper is to project any kind of united common philosophy.
In ABC this does not occur and it would seem that this publication pursues a
different kind ofjournalism which is more personal and individualistic in nature, with
more opinion and less information. The kind of journalism we see in ABC seems
directly targeted at a previously identified reader with whom there is a high degree
of complicity, rather than the hypothetical general reader specified by text books on
journalism (Martinez Hernando 1990:47).
Moreover, the ideological function of the journalism prevalent in ABC is
more noticeable than in the other two publications, a fact which has been pointed
out by other authors (Martinez Hernando 1990:170-1). In this sense, there would
seem to be a certain hostility towards the U.K. in ABC in general, which is shown by
drawing attention in a sarcastic tone to minority and possibly anachronistic aspects
of British life and traditions. Some of the translation solutions found in ABC would
seem to aim at ridiculing the source culture rather than offering information about
elements of that culture. This is the case with the biased nature and total lack of
functionality of some of the explanations to be found in ABC, for example:
336 Anne Martin

Canciller del Ducado de Lancaster, un puesto de rango simbólico que en su dia

desempenara Tomas Moro, or Carta de los Ciudadanos, un invento del
"majorismo".

This particular explanation contrasts noticeably with that offered in El Pais:

Carta de los Ciudadanos, la inconcreta declaración presentada en julio pasado con

el objetivo de proteger los derechos individualesfrentea los abusos de las grandes
compañias y del estado.

In these cases it almost seems as though the solutions given in ABC seek to
perpetuate an image of British society as being mildly eccentric and largely
anachronistic, instead of explaining the function of the references in objective terms.
Infrequent expressions are used to designate phenomena that are perfectly familiar
to the Spanish speaking reader, such as entrismo, Canciller del Ducado de
Lancaster. In this way, cultures are falsely depicted as being widely different, when
this is simply not the case. In Spain too, political parties are infiltrated by the left and
there are ministers without a specific portfolio.

6. Conclusions

To summarize our conclusions we could say that the most frequent trends noted in
our corpus were loans and literal procedures, whilst the least frequently used were
functional solutions and cultural equivalents, although the proportions differed
according to the newspaper in question, which was in line with the hypothesis that
we set out to test. Contrary to our hypothesis however, this did not generally affect
the clarity of the Spanish version. In most cases the solutions given clearly
transmitted at least one part of the original meaning, although the few cases of lack
of clarity detected were almost always due to the use of literal procedures. In this
sense, our original impression proved to be false, as the incorrect or nonsensical
versions were in a definite minority. There would seem to be a clear trend towards
the inclusion of the form of the original term through the use of literal options, and
only when these have been ruled out for whatever reason are the other options
considered. The impression is that literal translation is the norm, and when it is not
used this is due to a process of elimination.
This priority given to literal translation no doubt responds to the concept
that society in general has of the nature of translation: an exclusively linguistic
operation. This misconception may be due to the fact that most people's conscious
contact with translation is limited to foreign language classes in which translation is
used as a way of underlining the syntax of the language studied and not as a vehicle
for communication (Nord 1991:156; Newmark 1988:8; Bassnett-McGuire 1980:2).
Be that as it may, it is also clear that the high number of literal procedures and loans
 Institutional references 337

reflects the pressure brought to bear on Spanish by the Anglo-American culture and
its language. Moreover, it is indicative of the knowledge of this language and culture
that the average Spanish reader is thought to possess, which would be unthinkable
for other languages and cultures.

References

AGENCIA EFE. 1992. Vademécum de Español Urgente I. Madrid: Agencia Efe.

AGENCIA EFE. 1994. Manual de Espanol Urgente (10th ed.). Madrid: Ediciones Cátedra.

LVAREZ CALLEJA, M.A. 1991. Estudios de Traducción (inglés-espanol): Teoria, practica,

aplicaciones. Madrid: UNED.

ARIAS, J. 1995. "^Por qué traducimos tan mal los vocablos ingleses?" El Pais, 21.06.95.

BASSNETT-MCGUIRE, S. 1980. Translation Studies. London: Methuen.

EL PAIS. 1990. El Pais: Libro de Estilo. Madrid: Ediciones El Pais.

FERNANDEZ DE BEAUMONT, J. 1988. Los libros de estilo en la prensa de prestigio.

(Funciones de las normas de redacción de El Pais) (2 vols.). Doctoral dissertation. Facultad de
Ciencias de la Información, Universidad Complutense de Madrid.

FRANCO AIXELÂ, J. 1995. "Specific Cultural Items and their Translation". Translation and the
Manipulation ofDiscourse. Selected papers of the CERA Research Seminars in Translation Studies
1992-93. ed. by P. JANSEN, 109-123. Leuven: CETRA, The Leuven Research Center for
Translation, Communication and Cultures.

MARTINEZ ALBERTOS, J.L. 1989. El Lenguaje Periodistico. Madrid: Paraninfo.

MARTINEZ ALBERTOS, J.L. 1991. "Periodismo". Diccionario de Ciencias y Técnicas de la

Comunicació ed. by A. BENITO, 1003-1023. Madrid: Paulinas.

MARTINEZ HERNANDO, B. 1990. Lenguaje de la prensa. Madrid: Eudema.

MASON, I. 1994a. "Techniques of Translation Revisited: a Text-Linguistic Review of 'Borrowing'

and 'Modulation'". Estudios sobre la traducció ed. by A. HURTADO ALBIR, 61-72. Castellón:
Publicaciones de la Universidad Jaume I, Castellón.

MASON, I. 1994b. "Modulation and Other 'Techniques' Revisited". Turjuman, Journal of

Translation Studies. Vol.3(2), Oct. 1994, 51-60.

MASON, I. 1995. "Translating 'Techniques': A Procedural Approach". Cross-Words, Issues and

Debates in Literary and Non-Literary Translating ed. by I. MASON & C. PAGNOUILLE, 59-67.
Liège: L3 (Liège, Language and Literature).
338 Anne Martin

MAYORAL, R. 1992. "La ampliación de la información en la traducción". Lingüistica

Antverpiensia, 26, 45-71.

MAYORAL, R. & R. MUNOZ. 1996. "La traducción como transmisión de información. Problemas
y estrategias comunicativos en la traducción de un texto asmimstrativo". Lecture given at the V
Curso superior de Traducción Inglés/Espanol University of Valladolid, Spain, 18th April 1996
(photocopy).

NEWMARK, P. 1981. Approaches to Translation. London: Prentice Hall.

NEWMARK, P. 1988. A Textbook of Translation. Oxford: Pergamon Press.

NORD, C. 1991. Text Analysis in Translation. Amsterdam/Atlanta: Rodopi.

Real Academica Espanola. 1992. Dictionnario de la lengua espanola (DRAE) 21th ed. Madrid:
Espasa Calpe.

REISS, K, & H. VERMEER. 1984. Grundlegung einer allgemeinen Translationstheorie.

Tübingen: Max Niemeyer Verlag.

REISS, K, & H, VERMEER. 1996. Fundamentos para una teoria funcional de la traducción.
Madrid: Akal.

SANTOYO, J.C. 1987. "La 'traducción' de los nombres propios". Problemas de la traducción.
Mesa Redonda - noviembre 1983, 91-97. Madrid: Fundación Alfonso X El Sabio.
BRITISH ACCULTURATION OF ITALIAN
THEATRE

Stefania Taviano
University of Warwick, UK

Zusammenfassung
Im Fokus dieses Artikels stehen die wesentlichen Aspekte der britischen Übersetzung
und Rezeption vom italienischen Drama. Der Transfer vom italienischen Theater in die
anglophone Kultur wird unter Berücksichtigung der britischen Haltung zur
mediterranen Kultur, insbesondere zur italienischen Kultur, analysiert. Es wird gezeigt,
wie und bis zu welchem Ausmass der Respons zum italienischen Drama durch die
Gegensätzlichkeit britischer und italienischer Kultur und durch den Diskurs des Südens
beeinflusst wird. Dieses Herangehen an italienische Kultur kommt in zwei offensichtlich
divergierenden Herangehensweisen mit einer Prädominanz in der Zielkultur zum
Ausdruck. Diese werden hier analysiert.

Résumé
Nous aborderons certains aspects de la traduction et de la réception en Grande-
Bretagne du théatre italien. L'analyse porte sur les attitudes britanniques envers les
cultures méditerranéennes, en particulier la culture italienne. Nous démontrerons
comment et jusqu'où la réponse au théatre italien est marquée par l'opposition entre les
deux cultures en présence et par un certain discours méridional. Deux approches, en
apparence divergentes, prédominent dans la culture d'arrivée.

Resumen
Este articulo se centra en los principales aspectos de la traducción británica y la
recepción de las obras teatrales italianas. Se analiza la transferencia del teatro italiano a
la cultura anglosajona teniendo en cuenta la actitud britânica hacia las culturas
mediterráneas, con especial énfasis en la cultura italiana. Se demuestra cómo y en qué
medida se ve afectado el teatro italiano por los contrastes entre la cultura britânica e
italiana y por el discurso del sur. Este enfoque de la cultura italiana se hace patente en
dos enfoques aparentemente divergentes de la cultura término que se analizarân en este
articulo.
340 Stefania Taviano

1. Introduction

This article analyses some of the difficulties inherent in British translations and
productions of Italian plays. I argue that these obstacles are not so much related
to the exotic character of Italian theatre itself, but that they are, on the contrary,
generated by the British attitude to Mediterranean cultures, in particular Italian
culture. Starting from the notion of faithfulness to the original, I show how the
reception of Italian theatre is influenced by an opposition between British and
Italian culture and by what has been called a discourse of the south. This
attitude to Italian culture reveals itself in the two diverging strategies adopted in
British productions of Italian drama: one adheres to a stereotypical image of
Italian culture, which involves reducing Italian characters to caricatures; and the
other consists of transferring Italian plays to a British milieu to ensure the
domestication of those aspects of Italian drama which are deemed too exotic to
be accepted by target spectators.

2. Faithfulness

The "cultural turn" in Translation Studies (Bassnett 1998:123-140) has brought

a change from the evaluative method of comparing translations and originals to
the study of textual/extratextual constraints and selection processes which
influence and shape translation. Factors such as the criteria which determine the
selection of a certain text, the role of editors and publishers (Venuti 1995), the
reception of foreign texts by the receiving culture, are now the "primary focus"
of attention.
But despite this shift, critical discourse surrounding British productions
of Italian theatre continues to be characterised by a normative approach built
around concepts such as faithfulness, authenticity and truthfulness to the
original. This attitude is clearly expressed by Michael Coveney in his review of
the 1992 English production of Dario Fo's The Pope and the Witch, in which,
paraphrasing Arnold Wesker, he writes that

All productions (..) are by definition distortions, or transformations, or

realisations of what the author wrote. The most obvious distortion is that
rendered by translation into another language. Dario Fo's The Pope and the
Witch at the Comedy must charitably be categorised a misfired instance
perpetrated by translator Ed Emery and adaptor Andy de la Tour. (Coveney
1992)

Patrice Pavis points out that the notion of faithfulness related to the mise en
scène, predominant in critical discourse, "is pointless". He observes that the
 British acculturation 341

concept of faithfulness to the author's ideas and intent is meaningless since it

implies that "the text has an ideal and fixed meaning, free from any historical
variations" (Pavis 1992:26). When applied to the mise en scène of foreign texts
the argument proves to be equally meaningless as it is based on the assumption
that, for example, Dario Fo's original texts can only be interpreted and staged in
one right way, and that foreign productions should follow the steps of Italian
productions (as if they were all the same), regardless of cultural differences and
historical changes.
Basing their arguments on the concept of faithfulness to the original,
several reviewers have criticised British productions for a lack of Italian
character, while some critics have praised other productions for being
"genuinely" Italian. For example, the 1991 English production of Dario Fo's
Accidental Death of an Anarchist, adapted and interpreted "through a British
filter" by Alan Cumming and Tim Supple (Kaye 1991), was criticised by Hilary
Hutcheon because of "a distinct absence of fiery machismo and passion" and for
not being "Italian enough" (Hutcheon 1991). Conversely Martin Esslin praised
Zeffirelli's 1978 production of De Filippo's Filumena for making it "all seem
perfectly natural and genuinely Italian" (Esslin 1978:29).
Both directors and adapters alike have often defended their work by
highlighting the Italianness or Englishness of their interpretation. Mike Ockrent,
who directed the 1983 British production of Inner Voices by Eduardo De
Filippo, claimed to have preserved the original Neapolitan setting of the play
because he believed that the subject could not be adapted for British audiences
without losing its significance (quoted in What's On and Where to Go 30 June
1983). On the other hand, Peter Tinniswood, who adapted De Filippo's Napoli
Milionaria in 1991 for the Lyttelton Theatre, set the play in Liverpool and
replaced Neapolitan dialect by Scouse. He justified his approach in the
following way: "There was no point in my adapting it unless I brought some of
my character in — otherwise you may as well get a eunuch in to do it. I think I
brought in a certain frenzy of dialogue" (O'Keeffe 1991).

3. British and Italian culture

The discursive framework based on the above-mentioned concepts of

faithfulness and authenticity stems from awareness of the differences between
Italian and British culture. One of the main concerns often emerging in critics'
discourse is the difficulty, in some cases expressed as impossibility, of
translating and transplanting foreign characters and themes, especially Italian
ones into English. When Saturday, Sunday, Monday by Eduardo De Filippo
was staged at the Old Vic in 1973, directed by Franco Zeffirelli, Martin Esslin
wrote that De Filippo was so rooted in the tradition of Neapolitan theatre and
342 Stefania Taviano

culture, his themes and characters so local "that it seems almost impossible to
imagine how such a localised folk theatre could ever be transplanted into
another language, ever be performed by actors of another nationality and
background" (Esslin 1973: 41). Esslin's comments can be specifically applied to
the UK since his concern seems to reflect the British belief in a profound
unlikeness between the two cultures. This view about the nature of the relation
between Italy and the UK belongs to what has been defined as the discourse of
the south, to which I will come back later in this paper.
An anonymous critic in the Times Literary Supplement (1925), writing
on the English translation of Pirandello's plays, identifies the opposition
between British and Italian culture as a more universal condition, whereby the
exoticism of Italian theatre will often need to be erased, no matter what the
target cultural context is. Furthermore, the fact that we read the same kind of
lost-in-translation comments separated by a period of fifty years — the TLS
critic writes in 1925, while Martin Esslin writes in 1973 — and that they have
been applied to two very different playwrights, seems to confirm the existence
of a general British cultural approach affecting the reception of Italian drama in
the UK. While Esslin's concern about the Neapolitan nature of Eduardo De
Filippo's work, about its strong links to Naples society and culture and the
subsequent difficulties in translating it into a foreign language can be up to a
point understood, on the other hand, Luigi Pirandello's theatre and his
reflections about the human condition have a more universal value, even if his
work reflects to a certain extent, as for every writer, his native cultural milieu
— Sicily. In other words, despite the uniqueness of each playwright's work and
the differences in their approach to drama, the overall British attitude towards
Italian dramatists is rendered uniform by the above-mentioned cultural frame
through which Italian cultural products are perceived.
But despite the emphasis given to the foreign identity of Italian theatre
as opposed to British theatre, despite the difficulties of presenting Italian plays
to British audiences, as seen in the reviews mentioned, productions staged in the
UK by Italian companies in the original language have been highly successful. In
1925 Pirandello's company, the Teatro d'Arte, came to London to present four
of his plays: Sei personaggi in cerca d'autore, Enrico IV, Vestire gli ignudi,
Così è se vi pare. The productions were very successful and several reviewers
emphasised the response and involvement of the audience despite the fact that
the performances were in Italian. The enthusiasm initiated by the Teatro
dell'Arte company continued for several years, creating what was defined as
"the cult of Pirandello" {Times 1926). Equally successful were the Italian
production of Carlo Goldoni's The Servant of Two Masters directed by Giorgio
Strehler during the 1967 Peter Daubeny's World Theatre Season at the
Aldwych, and the production of Eduardo De Filippo's Napoli Milionaria by his
company during the 1972 World Theatre Season. Writing on this production,
 British acculturation 343

Irving Wardle (1972) pointed to the "universal" value of De Filippo's theatre.

The approach here is the exact opposite of the "contrastive" one dominating so
much of the reception of Italian culture in the UK, illustrated in the reviews
quoted above. What is even more interesting is that a production done in Italian,
by definition less accessible to English-speaking audiences, provoked a different
response than many other English productions affected by a more patronising
representation of Italian cultural values. Similarly, Dario Fo and Franca Rame's
workshops at the Riverside Studios in 1983 confirmed their success with British
audiences, and the Fo-Rame Theatre Group proved to be a stimulating
opportunity for collaboration between Italian and British theatre.

4. Stereotypes about Italians

I have emphasised elsewhere (Taviano forthcoming) the influence of English

stereotypes about Italians on the reception of Italian theatre and showed how
this is made explicit in critics' discourse and above all by translation strategies
adopted in British productions. These stereotypes and clichés belong to a
particular discursive framework, the discourse of the south, which emerged in
late 18th-century British culture. This discourse played a major role in the
construction of a coherent cultural, national and imperial British identity, since
this identity was based on "its opposition to the Orient, but also to Europe and
the Mediterranean." Britain defined itself by what it was not, as argued by
Stefania Arcara (1998:xv). She writes that "the South is for nineteenth century
Britain simultaneously an exotic, distant, and different entity, opposed (and
often inferior) to itself." If 19th-century Britain claimed a kinship with the
Mediterranean, in particular with Italy and Greece, seen as the "cradle of
civilisation" from which "Britain and the north proceed in a genealogical line of
progress", on the other hand superiority over the "decayed" Mediterranean
countries was equally emphasised.
This opposition to a primitive Mediterranean South as an integral part of
a British identity is still present in the 20th century, as testified by various texts
about Italy and in particular by travel guides. William Ward's book Getting it
Right in Italy, A Manual f or the 1990s is such a book. The author claims his
book is "the first to offer genuine insight and practical assistance in
understanding the Italians"; it is so important that "no visitor will be safe on the
streets without Getting it Right in Italy" (Ward 1990:v). Every aspect of Italian
culture analysed in the book is presented through the opposition between
British and Italian culture. Most themes are in fact introduced through formulas
such as: "unlike the Anglo-Saxon..." and "in Italy there is no equivalent of...".
Another example of a patronising discourse about Italians is The New Italians
by Charles Richards. Commenting on "two of the stock images of the Italian:
344 Stefania Taviano

child adoring and highly emotional", Richards adds the following joke: "They
talk as we know with their hands. When a social psychologist went round
orthopaedic wards he found those patients with arm injuries were tongue-tied:
they could not talk if they could not move their hands" (Richards 1994:xiv).
For the 1982 British production of De Filippo's Natale in Casa
Cupiello, under the English title Ducking Out, at the Greenwich Theatre, the
adaptor Mike Scott transposed the action from Naples to Lancashire, and Eric
Shorter in the Daily Telegraph makes an interesting comment: "The evening is
enjoyable in an instructive way, because it shows how different in temperament
and humour, character and culture, religion and social manners the British are
from the Italians" (Shorter 1982). Here this particular play and performance are
interpreted as just another example which proves how different these two
cultures are. The domesticating translation strategy of replacing the original
setting by one familiar to British audiences, although considered by some as a
means of avoiding the use of national stereotypes about Italians, becomes a way
of confirming the opposition between source and target culture and of
diminishing the possibilities of interaction between the two societies. The Punch
critic emphasises the consequences of such an approach:

Just as Chekhov's Three Sisters would look a bit odd in Cheltenham trying to get
to London, so does De Filippo's extended family of Neapolitans; they seem to be
auditioning for some minor Ayckbourn comedy about yet another disastrous
family Christmas and it does definitely diminish them as characters. (Ö) Their
behaviour, instead of being rooted in Italian tradition and plausibility, becomes
oddly arbitrary in Lancashire. (Punch 1982)

The opposition between Catholics and Protestants, which was one of the main
elements of the above-mentioned discourse of the south, and the subsequent
attribution of inferiority to Catholic cultures by the British, is another element
which continues to characterise the reception of Italian theatre. Andy de la
Tour's adaptation of Dario Fo's The Pope and the Witch in 1992 was not very
successful mainly because, according to numerous critics, the author's attack on
the Catholic Church was unappealing to English audiences. Reviewers felt that
British people are too "enlightened" and emancipated to be interested in a play
like The Pope and the Witch, which can only be a success in a primitive country
like Italy. It becomes evident that the discourse based on the notion of British
superiority, as opposed to the backwardness of Italy, is still predominant in
20th-century multicultural Britain. The fact that critics choose to focus primarily
on the attack against the Catholic Church, which is only one aspect of Fo's
critique of Italian society expressed in the play, leaving out more central ones
such as drug addiction, proves once more that the interpretation of Italian plays
continues to be manipulated to fit this patronising cultural frame. On the same
 British acculturation 345

line, the fact that Andy de la Tour explains in a programme note that he had to
cut out many of the Italian references from the script because they would not be
understood by a British audience, can be interpreted as a "veiled apology" for
an unsuccessful production (Stanfield 1992). The insurmountable differences
between Italian and British cultures are in fact often exploited by directors and
adapters to justify their domesticating approach.

5. A process of appropriation

The relationship between British and Italian culture, or rather the British, and
also American, approach to Italian cultural values, leads on to the issue of the
relation between minority and majority cultures and the latter's tendendy to
appropriate and transform the first (Pavis 1992). This issue has been emphasised
by several critics and scholars analysing the reception of Italian theatre and of
Italian culture in general. Clive Barker condemns "the insular determination of
the British theatre to solve its own problems in its own way" (Barker 1981:28),
and Malcolm Rutherford puts it in more explicit terms by saying that "some of
the English seem to think the Italians are funny just because they are Italians"
(Rutherford 1992).
It cannot be denied that an act of interpretation is involved in any mise
en scène and equally in any translation from one language into another. As Pavis
argues, "a text does not speak on its own, it has to be made to speak. [...] It is
not possible to neutralise the stage so that the text can speak on its own, or be
heard without mediation or without distortion" (Pavis 1992:45-6). Similarly
Sirkku Aaltonen (1996:74) argues that acculturation is an inherent part of the
process of drama translation.
In the case of the British reception of Italian theatre it is more a process
of "appropriation" which seems to emerge, connected to the uneven relation
between majority and minority cultures. What Pavis observes about the relation
between European and non-European theatre emphasising the risk of
Eurocentrism can be applied to the relation between British and Italian theatre,
where there is equally a risk of Anglocentrism.
The insular tendency of British culture and the re-emergence of a
discourse of the south reveals itself in several fields and in several forms, and
the producing of Italian plays is only one of them. In some cases it is expressed
through the emphasis put on Italian accent and gesticulation. In 1925 a British
production of And That's The Truth by Luigi Pirandello was staged, and the
extensive use of clichés is emphasised in the following review: "the English
actors were evidently determined to be more Italian than Italians. [...] Every
gesture, every movement, was so exaggerated as to convey the idea that Italians
are all born film actors, perpetually posing for "close-ups" {Times 1925). The
346 Stefania Taviano

tendency to accentuate the intensity of speeches and dialogues in Pirandello's

plays seems to have been recurrent from the first appearance of his theatre in
Britain to the present day. Both in terms of the written text and its theatrical
rendering, the Latin temperament has often been exaggerated to the extent of
appearing melodramatic, revealing a process of acculturation in which the
foreign text is rendered through dominant domestic values. That the
presentation of Italian plays to British audiences adheres to commonly shared
views about Italians and their culture is testified by a number of productions,
not only of Pirandello's plays, but also of Eduardo De Filippo and Dario Fo's
work.
Benedict Nightingale (1973), in his review of the English production of
De Filippo's Saturday, Sunday, Monday, questions the effects of the actors'
arms waving and the "generalised Italian tinge" of their accents. Another similar
case is represented by the 1982 British production of Pirandello's Liolà at the
New International Theatre in London, directed by Fabio Perselli. Liolà was
originally written in Sicilian dialect and therefore transferring it into English is a
difficult task for the translator. Jennifer Lorch in a review of the production
comments as follows:

I found it regrettable that in an attempt to present 'sicilianità' the director had his
company speak in foreign accents, thus maintaining an offensive tradition
developing in the English stage presentation of Italians. This gelato/spaghetti
English, a reflection of the British view of the Italian urban proletariat uneasily
absorbed into its culture, is simply not the best way to present the peasants of
Agrigento. (Lorch 1983:101)

There are numerous productions of Italian plays which have generated this kind
of critique, testifying to a widely adopted stereotypical presentation of Italian
culture to British audiences, and to a common tendency to reduce Italian
characters to caricatures. But throughout the history of English language
productions of Italian plays, the opposite kind of approach has also emerged,
that is to say the tendency to diminish those foreign aspects of Italian culture
which are not easily acceptable by British audiences by transposing Italian plays
to a British context. As we have seen, this is the case of Peter Tinniswood's
production of Napoli Milionaria, but there are several more examples of the
adoption of this strategy.
In 1985 Luigi Pirandello's Six Characters was staged in Bristol, directed
by George Brandt, who explained: "I played down the specifically Italian aspect
of Six Characters without making it aggressively English" (Brandt 1987:5), and
he added that he did so: "to denationalise the play a little". In 1987 another
production of the same play was presented at the National Theatre, this time
based on the adaptation by Nicholas Wright, who does not know Italian. Giulia
 British acculturation 347

Ajmone Arsan, commenting on the production, writes that "the play has been
anglicised", and concludes that this interpretation is "patronising towards the
public" (Arsan 1987:35). Katharine Worth (1987:36), in a review of the same
performance at the National Theatre, commented on the "homey context" of the
play, which opened it up for "popular consumption in the English theatre".
The anglicising strategy of reducing a "disturbing phenomenon" to a safe
and familiar event is a different but equally patronising result of the discourse of
the south. Given the cultural superiority and authority of the target culture,
British interpreters and adaptors of Italian texts can either choose to offer a
stereotypical representation of the source culture or minimise, and even cut,
those parts which are simply too unusual and exotic from a British perspective
to be culturally accepted.
In Nicholas Wright's adaptation, the actors rehearse Hamlet instead of
The Rules of the Game, since the latter is not very well-known in Britain. As
was explained in the programme, the choice was made 'from the familiar classic
repertoire', therefore the disorientating effect caused by the use of Pirandello's
own play was lost. The audience was presented instead with a familiar play and
was not made uncomfortable as happens in the source text. The play undergoes
further changes in this adaptation; above all the conflict between reality and
fiction, on which Six Characters is based, is essentially ignored and Nicholas
Wright comments: "We decided not to break down the barrier between the
public and the stage" (Arsan 1987), without justifying his choice. The
conclusion is also different: instead of the daughter who runs away laughing,
which was considered too dramatic an ending, this adaptation concludes with
the Director sitting at his desk and simply asking for more light. All these
alterations make the play more accessible and acceptable to British audiences as
they eliminate typically Italian aspects, considered too disturbing for the British
public, as emerges from Giulia Ajmone Arsan's following comment:

Having failed to grasp the meaning of such a conclusion, Nicholas Wright just
thought it was excessively dramatic, echoing those who opposed the first London
production of the play because it was judged to be too upsetting. Indeed
throughout the play the intention to tone down the exuberance of Latin passion is
evident. Isn't it much more English, discreet, restrained, to conclude the play
with the Director asking for a light while quietly sitting at his desk? (Worth
1987)

This process of acculturation, which inscribes Pirandello's theatre within

dominant English values, appears to have characterised other recent
productions of his plays, and although criticised to a certain extent, it is
considered, as in the above-mentioned review, a solution to the difficulties of
presenting Pirandello's work to British audiences. Whatever the opinion of
348 Stefania Taviano

"ardent Pirandellians" (Worth 1987), the resulting success indicates that, in

more than one case, the anglicised Pirandello has been better received than the
Italian Pirandello with all his disturbing peculiarities. (Taviano & Lorch,
forthcoming).

6. Conclusion

The two diverging approaches to Italian theatre analysed in this article cannot
be considered simply as the result of specific translation strategies, of choices
made by one particular translator or director, but they need to be put into
context and analysed in relation to a more general British attitude to the
Otherness of Italian culture, based, as we have seen, on the discourse of the
south originating in the late eighteenth century. What is important to understand
is that this approach is not confined to the theatre - though it cannot be denied
that it can become more explicit through mise en scène and the critical discourse
of reviewers - but emerges in other fields such as travel writing and also in the
translation and reception of Italian literature in general. Zigmunt Barânski, in an
article about the diffusion of Italian literature in the UK, reveals the
predominance in British culture of the image of Italy as a "pre-industrial country
and idyllic place" (Barânski 1993:260 translation by S.T.) and identifies it as the
main reason for the limited distribution and success of Italian literature in the
Anglophone world. This aspect of the British representation of Italian culture
cannot be separated from the idealisation of the same culture from a British
point of view, as confirmed also by John Julius Norwich's book The Italian
World (1983).
It becomes clear, then, that the British reception of Italian theatre
represents a fundamental field of investigation in the analysis of the relation
between these two cultures, but it also may be instrumental in the identification
of contemporary aspects of a British discourse of Otherness affecting the
reception of a number of Mediterranean cultures.

References

AALTONEN, Sirkku. 1996. Acculturation of the Other. Irish Milieux in Finnish Drama
Translation. Joensuu: Joensuu University Press.

ARCARA, Stefania. 1998. Constructing the South: Sicily, Southern Italy and the
Mediterranean in British Culture, 1773-1926. Doctoral thesis in British Cultural Studies,
University of Warwick.
 British acculturation 349

ARSAN, Giulia Ajmone. 1987. "Six Characters at the National Theatre". The Yearbook of
the British Pirandello Society 7, 32-35.

BARÂNSKI, Zigmunt. 1993. "La diffusione della letteratura italiana contemporanea in Gran
Bretagna". The Italianist 13, 255-265.

BARKER, Clive. 1981. "Right You Are (If You Could Only Think So)". The Yearbook of the
British Pirandello Society 1, 26-33.

BASSNETT, Susan. 1998. "The Translation Turn in Cultural Studies'. Constructing Cultures
ed. by in S. BASSNETT & A. LEFEVERE, 123-139. Clevedon: Multilingual Matters.

BRANDT, George W. 1987. "Six Characters in Bristol". The Yearbook of the British
Pirandello Society 7, 1-10.

COVENEY, Michael. 1992. "The Pope and the Witch". Observer 19 April.

ESSLIN, Martin. December 1973. "Saturday, Sunday, Monday". Plays and Players, 41-42.

ESSLIN, Martin. January 1978 "Filumena". Plays and Players, 28-29.

HUTCHEON, Hilary. 1991. "Accidental Death of an Anarchist". Tribune 11 January.

KAYE, Paul. 1991. "Rebirth of an Anarchist". London Student 10 January.

LORCH, Jennifer. 1983. "Liolà". The Yearbook of the British Pirandello Society 3, 100-101.

NIGHTINGALE, Benedict. 1973. "Famiglia Faces". New Stateman November.

NORWICH, John Julius. (ed). 1983. The Italian World. London: Thames & Hudson.

O'KEEFFE, Brendan 1991. "Bray of Naples". What's On 26 June.

PAVIS, Patrice, 1992. Theatre at the Crossroads of Culture. London & New York:
Routledge.

PUNCH. 17 November 1982. "Ducking Out".

RICHARDS, Charles. 1994. The New Italians. London: Michael Joseph.

RUTHERFORD, Malcolm. 1992. "The Pope and the Witch". Financial Times 15 April

SHORTER, Eric. 1982. "Ducking Out". Daily Telegraph 16 December.

STANFIELD, Keith. 1992. "The Pope and the Witch". City Limits 30 April.

TAVIANO, Stefania. Forthcoming. "The Reception of Dario Fo and Franca Rame in the
UK". In Justice and Morality - Visions of change in European Theatre, ed. by Edward M.
BATLEY & David BRADBY, University of London.
350 Stefania Taviano

TAVIANO, Stefania & LORCH, Jennifer. Forthcoming " Producing Pirandello in England",

The Yearbook of the British Pirandello Society, 1999.

The Times. 18 September 1925. "And That's The Truth".

The Times. 18 June 1926. "The Pleasure of Honesty".

Times Literary Supplement. 26 March 1925. "And That's The Truth".

VENUTI, Lawrence. 1995. The Translator's Invisibility. London & New York: Routledge.
WARD, William. 1990. Getting it Right in Italy. A Manual for the 1990s. London:
Bloomsbury.

WARDLE, Irving. 1972. "Napoli Milionaria". The Times 9 May.

What's On and Where to Go. 30 June 1983. "Voice Choice".

WORTH, Katherine. 1987. "Six Characters at the National Theatre (2)". The Yearbook of the
British Pirandello Society 7, 36-43.
 PART VIII

Translation history
TOWARDS A HISTORY OF TRANSLATION IN A
(POST)-COLONIAL CONTEXT: AN AFRICAN
PERSPECTIVE

Paul Bandia
Concordia University, Montreal, Canada

Zusammenfassung
Der Verfasser analysiert das Übersetzen zur Kolonialzeit in Afrika und besonderes die
Tradition der mündlichen Literatur. Im Anschluss daran erörtert der Verfasser die
Ambiguität der Diskurse über die Literatur nach der Kolonialisierung: zum einen
erzeugten diese Diskurse die Hierarchie zwischen dem Zentrum (Kolonisator) und der
Peripherie, zum anderen Hessen sie das aus dem Kontakt der Sprachen und Kulturen
stammende Kreolische zu. Bei dem letzteren Standpunkt, mit einer Sprache, die
zwischen zwei Sprachen oder als dritter Kode fungiert, wird das Übersetzen selbst
gleichzeitig Quelle und Ergebnis der Mischsprache - die Frage über die
Ausgangstexte wird auf eine andere Weise gestellt.

Résumé
L'auteur s'attache à considérer le rôle de la traduction en Afrique, au temps de la
colonisation - notamment envers la tradition littéraire orale. Il s'interroge sur les
ambiguités des discours sur la littérature, après la colonisation: en effet, ces discours
tantôt reproduisaient la hiérarchie entre centre (colonisateur) et périphérie, tantôt ils
assumaient la créolisation issue du contact des langues et des cultures. Dans cette
dernière perspective, avec une langue constamment «entre deux» ou «troisième code»,
la traduction elle-même devient à la fois source et résultat du métissage - posant
autrement, entre autres, la question des textes de départ.

Resumen
Este articulo analiza la historia de la traducción desde una perspectiva postcolonialista,
basada en una evaluación critica de algunas prácticas de traducción. Evalua el impacto
del hibridismo lingüistico sobre la teoria de la traduccion tradicional, seãlando las
limitaciones de ésta y haciendo hincapié en la necesidad de explorar una dimensión
etnogrâfica en la búsqueda de una teoria de la traducción mâs global.
354 Paul Bandia

1. Introduction

The history of translation is increasingly being studied from a more

comprehensive standpoint, with the distinct advantage of bringing pluralism to
bear on the subject across the world and hence of achieving some measure of
breadth and objectivity (Delisle and Woodsworth 1995). In the quest for a
comprehensive history of translation, new perspectives have developed, such
as the parameters of gender and postcolonialism. This paper deals with the
history of translation from a (post)-colonial perspective, based on a critical
assessment of some translation practices as well as the ideological
underpinnings that guide or determine choice and orientation in the translation
of the (post)-colonial experience (Rafael 1988; Cheyfitz 1991; Niranjana
1992). Traditional translation theory has often emphasized a clear demarcation
between source language and target language. However, in the postcolonial
contexts the issue is not that clear-cut, given the linguistic complexities
whereby for historical reasons a plethora of languages coexist within the same
geographical space, interacting with one another, modifying and subverting
one another, often resulting in hybrid languages which defy the traditional
definition of source or target language (Mehrez 1992; Robinson 1997).

2. Colonial translations

Translation in colonial Africa was carried out by various types of colonialist

inscribers such as administrators, missionaries, linguists and anthropologists.
Translation at this time is viewed as the process by which colonialism inscribed
African oral literature, whether for historical or philological reasons.
According to Austen (1990), "African oral texts produced during the colonial
era present a major paradox for translation" (29). On the one hand, they are
considered "'authentic' African documents, composed in an indigenous idiom"
(29). On the other hand, they were produced through the "intermediacy of
colonial scribes acting for their own purposes via not only the language but
also the medium (writing) of European domination" (29). It is however
important to study the very large body of African oral literature "captured" by
colonialist efforts at transcription and translation, and the resulting colonialist
view of African literature.
In colonial Africa, administrators of various kinds, acting as translators,
were responsible for many of the early collections of African oral texts. These
administrators concerned themselves professionally with the control of
Africans and thus viewed the study of indigenous literature as a valuable key to
"native psychology" (Austen 1990:31). In the words of one of the most
respected French folklorist/administrators, Equilbecq (1972), "It is necessary to
know those whom one wishes to dominate" (22). However, "these literary
 Towards a history of translation 355

efforts were also motivated by a more amateur humanist concern for an African
cultural essence endangered by colonialism itself' (Austen 1990:31). Hence, as
noted by Equilbecq, "These traditions are the supreme vestiges of the primitive
beliefs of the black race and, on this basis, deserve to be saved from oblivion"
(1972:22). There is therefore an unbalanced power relationship which in such
an imperialist enterprise as colonialism is bound to dictate the criteria and
parameters guiding the translation process. As a consequence, the authenticity
of 'translated' or 'recorded' oral texts needs to be clarified.
The authenticity of African oral texts produced by pre-tape-recorder
scholars has come under serious attack as these texts had to be 'recorded'
during performances of African oral works. However, some colonial
administrators hardly made the effort to attend such performances, and even
when they did it is doubtful that they understood. Those informants who were
recognized as performers within their own culture seem to have been the
Africans most conveniently available to administrators, particularly African
subordinates of the colonial regime itself. As pointed out by Austen: "Even
when the immediate narrators spoke in their own languages, the translators
were inevitably government interpreters, whose own imperfect European idiom
was then further edited by the administrator/author of the ultimately published
text" (1990:31). One can only imagine the influence of colonialist inscription
on these collected narratives and their obvious philological shortcomings. The
context of performance of the narratives is hardly known and there is no
indigenous text to control the translation.
The attitude of missionaries towards African oral narratives was
significantly different from that of administrators in many ways. For one thing,
the missionaries had no interest in preserving indigenous culture which, by and
large, embodied the very values they sought to eradicate in their bid to convert
the natives to Christianity. However, the missionaries eventually realized that
the inculcation of Christianity and other related secular western values into
African society required mastery of local languages and a good grasp of those
elements of indigenous belief and practice which could help establish the
foundation for spreading Christianity in Africa.
Therefore, missionaries, unlike administrators, recorded oral literature
texts under more reliable conditions, partly due to their commitment to learn
local languages for purposes of evangelization (particularly Bible translation).
The missionaries produced bilingual texts, translated by themselves and
derived from skilled performers. They even produced texts exclusively in local
languages. Vernacular translations (in local languages) were thus available, and
moreover, mission catechists and school teachers were often Africans who
participated in the translation of these religious texts.
However, missionary colonialism also sought to shape the literature
received by the Africans by constructing texts compatible with Christian
conceptions and omitting facts which conflicted with Christian values (Austen
356 Paul Bandia

1990:33). Mpiku (1972:128-9), for example, describes the "détournement" of

traditional tales when a Christian interpretation is tacked on to them. This
deliberate text 'construction' is most prevalent in the missionaries' recounting
of African creation myths which emphasize the monotheistic dimension of
African belief Furthermore, for religious reasons, the missionaries did not
record the narratives, poetry and trickster tales which they considered less
"moral", thus producing an "innocent" version of African literature that
reinforces the "primitive" vision of the administrators (Austen 1990:33).
Linguists and anthropologists are two groups of European scholars who
were also particularly interested in indigenous Africa during the colonial era.
Linguists were mainly interested in cataloguing the very numerous languages
of the continent and analyzing their vocabulary, grammar, syntax and
phonology. Their enterprise was consistent with the aim of the missionaries to
translate their own religious texts into African vernaculars. The work of the
linguist required extensive recording of texts as the basic material of language
study. Colonial-era linguists recorded African oral narratives with the greatest
scrupulousness, in phonetic script for the indigenous version, and literal
interlinear or adjacent prose versions for the translations. However, the amount
of oral literature recorded was limited and the context of performance was
essentially unnatural, that is in a laboratory. For instance, German linguists
often transported African informants into metropolitan seminars.
The colonial-era students of African literature also included the first
generations of professional anthropologists who travelled regularly between the
western metropolises and the "exotic" societies which were their object of
study. The main contributions to African oral literature by colonial-era
anthropologists came from scholars who were employed at the time by either
museums or colonial governments. These itinerant anthropologists often lacked
the required theoretical sensitivities and intensive involvement in a local
culture and language to be able to record and translate the oral literature.
However, some anthropologists based for long periods in a single territory
could develop great proficiency in local languages and thus undertake valuable
research on the oral tradition. Yet even the works of these anthropologists can
be criticized on the basis of their overt subjugation to colonial purposes (for
example, see von Laue 1976). Nonetheless, some itinerant anthropologists did
produce significant amounts of texts. A case in point is Leo Frobenius, a
prolific recorder of African narratives in the colonial era who was driven by a
"German romantic vision of Africa's anti-rational civilization, and
commitment to a radical theory of cultural diffusionism" (Austen 1990:35)
(and, of course, its political implications) (see Jahn 1972).
Based on this compressed survey of colonial inscription, it can be said
that colonial projection of African literature was essentially couched in a
hegemonic discourse which failed to account fully for the African subject it
 Towards a history of translation 357

was constructing. This agenda of hegemony can inform research in translation

studies which seeks to explore the power differential of imperialism.
In our research, we seek to determine what kind of African texts emerge
from the colonial processing of oral literature by examining the texts
themselves rather than emphasizing the history of their production. We analyze
in what ways the texts are shaped to conform to a colonialist version of African
history and culture, and how elements of African expression considered
subversive of the colonial project are suppressed. Therefore, it is less the
historical reconstruction than the transcribed, or translated, literary texts which
are of interest, together with the relationship between them and various issues
of political and cultural hegemony viewed from a translation perspective.

3. Post-colonial translations

Translation activity in postcolonial Africa can be divided into three main

categories: religious translation, public service translation and literary
translation. Religious translation began in the colonial era and continued well
into the postcolonial period. European and American missionaries, from
various denominations, have set up stations in different parts of Africa and
have continued to learn local languages for purposes of evangelization and
especially for the translation of the Bible and other religious texts. Public
service translation has continued to flourish as a consequence of the legacy of
colonialism, namely the adoption of European languages as official languages.
For the purposes of this discussion, we will dwell on literary translation, which
allows us to explore more readily the rapport between translation and the
power of hegemony.
Postcolonial discourse regarding African literature has promoted a
binarized model of the world and has maintained a center-periphery polarity
which both exaggerates and simplifies the effects of the colonial imposition of
European languages (Barber 1996:3). The experience of the colonized is often
determined by his or her relation to the metropolitan center. The postcolonial
writer conveniently articulates African literature in a defiant yet accessible
idiom, 'writing back' as it were in a language the ex-colonizers can understand
even if the European language of writing is often expressed in a modified
register with an essentially African flavour. Postcolonial criticism represents
European language writing as an off-shoot of imperial domination; however,
portraying the African writer as a happy contributor to the Great Western
Tradition via an idiom of compulsion can mask the stark power relations
between the center and the periphery. As a consequence of the imposition of
European languages on the colonies and the displacement or devaluing of
indigenous languages and literatures, the colonial subject was culturally and
linguistically dispossessed, leading to a deep loss of self-esteem and cultural
358 Paul Bandia

confidence. The colonial subject was forced to learn the colonizer's language
in which he or she was initially trapped and constrained. Postcolonial writers
found themselves languageless, gagged by the imposition of European
languages on their world (Barber 1996:6).
However, the postcolonial writer made a comeback by radically re-
appropriating the colonial language culture. In the words of Barber (1996), "the
periphery now takes on the culture and language of the center and transforms
it, breaking it, infusing it with local registers and refashioning it so that it
speaks with the voice of the marginalized" (6). Instead of one hegemonic
English, for instance, we get a plurality of local Englishes. Marginality has thus
become an unprecedented source of creative energy. And the literature of the
margins is thus at the frontiers of postmodern aesthetics (Tiffin : x). In
postcolonial criticism of African literature, the colonial subject is thus
portrayed as having finally found a 'voice'; the colonized has mastered and
subverted the colonial codes as a means of asserting his or her identity. If the
African writer has to write in a European language, he or she has to transform
the language to make it bear the burden of African experience.
In The African Palimpsest (1991), Zabus makes an extensive and
sustained attempt to bring postcolonial analysis to bear specifically on African
texts. She carries out a detailed analysis of the expression of indigenous
linguistic repertoires within a range of anglophone and francophone novels.
She assesses the role of linguistic hybridization in African Europhone
literature, not only as a means of providing "a distinctively African cocktail",
but also as an attempt "to resist, recreate and subvert the dominant (European)
language" (Barber 1996:8). The African writer is shown to draw on thematic
elements from oral as well as written repertoires in order to achieve the effect
of orality through specific techniques of writing. African Europhone writing
thus takes on its special Africanity, its ability to express specifically African
experience in an alien tongue. African writers crack the standard forms of
metropolitan European languages and re-form them into unique new patterns.
One has the distinct impression that expressions in indigenous languages are
made to live on through translated borrowings and echoes. According to Barber
(1996), the analysis of oral and indigenous repertoires as a pool of resources
for written Europhone texts precariously straddles the world of orature and that
of literature and bridges the two by translating the one into the other (1996:10).
An example of the impact of indigenous repertoires on African
Europhone texts can be seen in the practice of relexification. According to
Zabus, relexification is an "imaginative world-creating attempt at forging a
new literary aesthetic medium or 'third-code' out of the alien, dominant
(European) lexicon. It is also an attempt at appropriating the Other's language
in order to make it one's own and bend its otherness or fixity to artistic and
ideological exigencies" (1991:103).
 Towards a history of translation 359

Gabriel Okara's The Voice is a conscious experiment in syntactic

relexification. Throughout the novel, the texture of the Ijo language (spoken in
Nigeria) does textual violence to standard metropolitan English. The novel is
replete with morphosyntactic innovations such as the postponement of the verb
or of the negative, which are characteristic of the Ijo language.

- Who are your people be? ... If you are coming-in people be, then come in.
(Okara 1964:26-27).
- Et vous les gens, qui pouvez-vous être? ... Si vous devez être des gens à
entrer, alors entrez. (Okara 1985:15).

- You cannot a thing I have done not put on my head; How can you on my
head put a thing that happened not? (Okara 1964:66).
- Vous ne pouvez pas une chose que point n 'ai accomplie me faire endosser;
Comment pouvez-vous me faire endosser une chose qui point ne s'est passée?
(Okara 1985:60).

- These are my answering words to your questioning words (Okara 1964:56).

- Voilà les mots qui font réponse à tes mots qui faisaient question ( Okara
1985:48).

Zabus emphasizes the fact that, as a diachronic practice, relexification

differs from mother-tongue interference, calquing and loan-translation. Okara
does not translate from his native Ijo language, i.e. he does not seek
equivalence, nor does he aim at recoding the original according to the norms of
the target language. Unlike translation, relexification does not operate from the
language of one text to the other but from one language to another within the
same text. Jean Sevry acknowledges the difficulty of translating The Voice into
French in the foreword to his translation, La voix. He admits to "bending" the
French language in an attempt to reproduce Okara's unusual syntax and, being
somewhat dependent upon the genius of the French language, resorts to the use
of archaic negative structures such as "que point n'ai accomplie," "qui point ne
s'est passée," to compensate for the lack of a similar syntactic play in French.
The African writer thus creates a "third code" in the "overlapping space
between other tongue and mother tongue" - a palimpsest where "behind the
scriptural authority of the European language, the earlier, imperfectly erased
remnants of the African language can still be perceived" (Zabus 1991:104-
105).
The hybrid state of postcolonial society provides us with a great
potential for linguistic creativity and adds another dimension to our
understanding of translation activity. Indeed, the mixing of cultures and
languages in itself presupposes translation. However, it is a form of translation
that defies the sort of preconceived notions about source and target language,
360 Paul Bandia

and the concept of equivalence, espoused by traditional translation theory. As

Mehrez points out:

These postcolonial texts, frequently referred to as 'hybrid' or 'métissés'

because of the culturo-linguistic layering which exists within them, have
succeeded in forcing a new language that defies the very notion of a 'foreign'
text that can be readily translated into another language. With this literature
we can no longer merely concern ourselves with conventional notions of
linguistic equivalence, or ideas of loss and gain which have long been a
consideration in translation theory. For these texts written by postcolonial
bilingual subjects create a language 'in between' and therefore come to
occupy a space 'in between'. (1992:21)

Our explorations into an ethnographic theory of translation deal squarely with

what Mehrez refers to as the space 'in between' characteristic of the
postcolonial context. The debate over writing in a major world language and
writing in a local third-world language (see Ngugi wa Thiong'o 1986) leads
one to examine the creation of a 'third code', that is a language 'in between'
which, as mentioned earlier, is hardly accounted for by traditional translation
theory. Some of the issues raised have to do with the following sets of
questions.
First, the African writer of Europhone literature always translating?
How does one define the writer's source and target languages since he or she is
indeed the locus of both linguistic experiences, namely the language of the
imaginary original versions and the language 'in between' which is the
language of writing. In other words, how does one define source and target
language in this context of linguistic layering and hybridity?
Second, as the translation is supposedly operated upon an oral tradition
narrative, how does one assess the authenticity of the source text in order to
evaluate the quality of the target text? A main characteristic of the kind of
translation practiced by African writers is the frequent absence of "real
originals", that is, the absence of written versions of oral narratives.
Third, how do postcolonial translations fare in the current context of
globalization? In other words, how does a language 'in between' enable a
postcolonial society to assert its identity, and does this count for much in a
global economy which still promotes linguistic hierarchy, as evidenced in the
hegemonic influence of western metropolitan varieties of English or French,
for example?
These are some of the issues explored in our quest for a history of
translation in Africa. We seek to deconstruct and demystify the imperialist
subtexts of ethnographic translation theory. Our research opposes the
assimilative sense-for-sense translation associated with colonial discourse and
a form of radical literalism (as practiced by some African writers) associated
 Towards a history of translation 361

with transformative postcolonial discourse (Niranjana 1992; Robinson 1996).

Niranjana points out that "Frantz Fanon and others have written about the
revolutionary potential of a historical sense in the hands of the colonized. This
sense is crucial also for a practice of translation in the colonial/postcolonial
setting. By reading against the grain of colonial historiography, the
translator/historian discovers areas of contradiction and silent resistance that,
being made legible, can be deployed against hegemonic images of the
colonized" (1992:76). Niranjana thus emphasizes the revolutionary importance
of what she refers to as "retranslation", which is indeed a decolonizing strategy
used by the counterhegemonic translator, working from the vantage position of
a postcolonial subject.

4. Conclusion

As argued in Robinson (1996), language is undoubtedly one of the instruments

of imperial domination, and all colonial regimes had language policies of one
kind or another. Gérard (1981:176) draws a strong distinction between "two
utterly different patterns" in relation to African-language expression:
repression by French, Portuguese, Spanish and Italian colonial powers, and
encouragement by British and German powers. It will be interesting to see how
these different colonial linguistic policies have influenced the history of
translation in Africa.
The more recent period of translation history in Africa has been
characterized by a progression toward 'othering' the foreign (colonial)
language, in an attempt to reduce its role while rescuing African orature from
(post-)colonial domination and glottophagia. The glottopolitical 'othering' of
the European language in African Europhone literature aims to reinstate the
oral, traditional and African-language expression which has virtually been
erased or silenced by colonial glottophagia. A study of the history of
translation in Africa will throw light on the various options that African writers
are envisioning as a means to produce a literary 'third code' and consequently
claim their rightful place on the world literary stage.

References

AUSTEN, R. 1990. "Africans speak, colonialism writes: the transcription and translation of
oral literature before World War II". Cahiers de Littérature Orale, 28: 29-53.

BARBER, K. 1996. "African Language Literature and Postcolonial Criticism". Research in

African Literatures, 26(4): 3-30.
362 Paul Bandia

CHEYFITZ, E. 1991. The Poetics of Imperialism: Translation and Colonization from "The
Tempest" to "Tarzan ". New York: Oxford University Press.

DELISLE, J. & WOODSWORTH, J. (1995). Translators through History. Amsterdam &

Philadelphia: John Benjamins.

EQUILBECQ, F.V. 1972. [1913] Contes populaires d'Afrique Occidentale. Paris:

Maisonneuve et Larose.

GÉRARD, A. 1981. African Language Literatures: An Introduction to the Literary History of

Sub-Saharan Africa. London: Longman.

JAHN, J. 1972. Leo Frobenius: the Demonic Child. Austin: University of Texas Press.

LAUE, H. VON. 1976. "Anthropology and Power; R. S. Rattray among the Ashanti". African
Affairs 75: 33-54.

MEHREZ S. 1992. "Translation and the Postcolonial Experience: The Francophone North
African Text". Rethinking Translation ed. by L. Venuti, 120-138. London & New York:
Routledge

MPIKU, M. Y. 1972. "Introduction à la littérature kikongo". Research in African Literatures 3:

117-161.

NGUGI, W. T. 1986. Decolonising the Mind: the Politics of Language in African Literature.
London: James Curry.

OKARA, G. 1964. The Voice. London: Heinemann.

OKARA, G. 1985. La voix. Trans. by J. Sevry. Paris: Hatier.

NIRANJANA, T. 1992. Siting Translation: History, Post Structuralism, and the Colonial
Context. Berkeley: University of California Press.

RAFAEL V. L. 1988. Contracting Colonialism: Translation and Christian Conversion in

Tagalog Society Under early Spanish Rule. Ithaca: Cornell University Press.

ROBINSON, D. 1996. Translation and Taboo. Dekalk: Northern Illinois University Press.

ROBINSON, D. 1997. Translation and Empire. Manchester: St. Jerome Publishing.

TIFFIN, H. 1991. "Introduction". Past the Last Post: Theorizing Post-Colonialism and Post-
Modernism ed. by I. Adams & H. Tiffin, vii-xvi. Hemel Hampstead: Harvester Wheatsheaf.

ZABUS, C. 1991. The African Palimpsest: Indigenization of Language in the West African
Europhone Novel. Amsterdam & Atlanta: Rodopi.
LA CRITIQUE DE LA TRADUCTION LITTÉRAIRE
EN TURQUIE (1940-1992)

Emine Bogenç Demirei & Hülya Yilmaz

Université Technique de Yildiz, Istanbul, Turquie

Abstract
The paper surveys the history of literary translation criticism in Turkey during 1940-
1992. The policies, objectives and contributions of three representative translation
journals are analysed: Tercüme (Translation, 1940-66), "Yazko Çeviri" (Yazko
Translation Magazine, 1981-84) and Metis Çeviri (Metis Translation Magazine, 1987-
92). Early progress was followed by a stagnation phase, despite the attempts to
establish an objective and theoritically based discipline of translation criticism.

Zusammenfassung
Die Untersuchung befasst sich mit der Kritik von übersetzten literarischen Werken in
der Türkei. In diesem Rahmen wurden drei Übersetzungsperiodika, die in den Jahren
1940 bis 1992 erschienen sind, untersucht. Diese sind "Tercüme"
(Übersetzungszeitschrift "Übersetzung", 1940-66), "Yazko Çeviri"
(Übersetzungszeitschrift Yazko, 1981-84) und "Metis Çeviri" (Übersetzungszeitschrift
Metis, 1987-92). Diesem Zeitabschnitt mit der aufsteigenden Tendenz in der
Übersetzungskritik der literarischen Werke folgt jedoch trotz intensiver Bemühungen
hinsichtlich einer objektiv, wissenschaftlich und theoretisch fundierten
Übersetzungskritik eine Ruheperiode.

Resumen
En este trabajo de investigatión sobre la critica de la traducción literaria en Turquia se
han analizado tres revistas de traducción: "Tercüme" (Traducción, 1940-66), "Yazko
Çeviri" (Traducción Yazko, 1981-84), "Metis Çeviri" (Traducción Metis, 1987-92),
representantes de la modalidad de la traducción durante el periodo que va desde 1940
hasta 1992. A través de sus politicas, objetivos y aporte a la critica de la traducción,
estas revistas proporcionan una vista panorámica de la evolución/regresión de la
critica de la traducción durante el mencionado periodo. Después de un periodo de
progreso viene una época de estancamiento, a pesar de todos los esfuerzos realizados
para lograr una critica a la traducción objetiva y teóricamente cientifica.
364 Demirei & Yilmaz

Revues de traduction en Turquie depuis 1940:

*Tercüme (Traduction) 1940-66: 87 numéros
*Çeviri (Traduction) 1979: un seul numéro
*MEB Düçiin, Bilim, Egitim (Ministère de l'Education nationale,
Philosophie, Science, Pédagogie) 1979: un seul numéro
*Baglam (Lien) 1979 : durant une année
*Yazko Çeviri (Traduction Yazko) 1981-84: 18 numéros
*Diin ve Bugiln Çeviri (La Traduction hier et aujourd'hui) 1985:
2 numéros
*Metis Çeviri (Traduction Métis) 1987-92: 21 numéros
*Çeviribilim ve Uygulamalari (La Traductologie et ses pratiques):
depuis 1992
*Tömer Edebiyat Çeviri (Traduction littéraire Tömer): depuis 1994
*Çeviribilim Tömer (Traductologie Tömer): depuis 1995

Cette liste des revues parues en Turquie, depuis 1940 jusqu'à nos jours, permet
de voir, d'après les titres, l'évolution du domaine de la traduction. Notre
recherche sur la critique de la traduction, se limitera à l'analyse de trois revues
Tercüme, Yazko Çeviri, Metis Çeviri1, représentatives de l'époque et de la
situation de la traduction uniquement littéraire.
La critique de la traduction littéraire en Turquie a-t-elle évolué ou bien
régressé depuis 1940? Afin de répondre à cette question, nous nous sommes
proposé d'analyser en particulier les politiques de ces revues, leurs objectifs et
leur apport à la critique de la traduction littéraire.
Quelle était la situation de la critique de la traduction avant 1940? Les
premiers exemples de critique apparaissent d'abord sous forme de "reproche" et
de "querelle"2. Entre 1923 et 1940, la critique se limite à présenter les œuvres,
avec des jugements subjectifs. A partir de 1940, la critique devient une mise en
valeur de l'œuvre et de l'écrivain. Cette évolution est due en premier lieu au
manque d'œuvres traduites et aussi au fait que le concept de "critique" est à
l'époque étranger à la société turque.

1. Tercüme - Traduction (1940-1966)

1. 1 Création de la revue en 1940

La situation idéologique, politique et culturelle de la Turquie avant 1940 a

entraîné un mouvement visant à l'universalisation, défini comme un besoin
primordial d'accéder à la pensée et à l'art universels par le truchement de la
littérature. Pour une bonne structuration de la vie culturelle en Turquie et pour
 Critique de la traduction littéraire 365

une intégration à la civilisation universelle, on organisa un mouvement de

traduction systématique, en particulier des classiques de la littérature mondiale.
L'objectif principal de ce mouvement "humaniste" était d'enrichir d'une part la
langue turque et d'autre part la bibliothèque nationale. Ainsi sur l'initiative de
Hasan-Ali Yücel, homme d'Etat et de lettres, ministre de l'Education nationale
de l'époque, un Bureau de Traduction fut créé, afin de planifier ce mouvement
de traduction et de publier une revue bimensuelle intitulée Tercüme
(Traduction). Les membres de ce Bureau ont d'abord projeté la traduction des
classiques de la littérature mondiale. Lors de la préparation des listes des
œuvres à traduire, la valeur a primé sur le souci commercial, comme l'affirma
Bedrettin Tuncel, un des fondateurs de ce Bureau de Traduction: '"Nous
sommes convaincus de l'inutilité de traduire en turc des œuvres dont la valeur
commerciale aurait plus d'importance que la valeur littéraire."3

1. 2 Débuts de la critique de la traduction

A la suite de cette activité de traduction systématique, grâce à une

augmentation du nombre des œuvres traduites, la critique de la traduction
prend forme. Afin d'améliorer le niveau de la qualité des critiques de
traduction, la sélection se fait selon des critères précis pour éliminer "les
dilettantes de la langue" et "les pirates de la langue" comme les a surnommés
Nurullah Ataç, grand écrivain et traducteur de l'époque. La critique cesse alors
d'être "une chasse à l'erreur". Par ailleurs, la sélectivité des critiques entraîne la
sélectivité dans le choix des œuvres à traduire.
Au niveau de la forme, le fait que les traductions étaient publiées
accompagnées du texte original, montre combien cette revue s'ouvrait à la
critique. Dans cette perspective, Nurullah Ataç souligne l'importance qu'on
accorde à la critique de la traduction, tout en mettant l'accent sur son côté aussi
difficile que celui de la traduction elle-même:

La critique de la traduction est ausi difficile à faire que la traduction elle-

même. Nous avons rencontré bon nombre de mauvaises traductions pour
lesquelles il nous a été très difficile de déterminer les erreurs. Dès lors, nous
souhaitons que la place accordée à la critique soit aussi large que celle de la
traduction.4

Cette ouverture d'esprit entraîne la multiplication du nombre des critiques de

traduction. Peu à peu, les critiques ressentent le besoin de baser leur jugement
sur des critères et des normes. A une époque où la théorie de la traduction
n'existait pas encore, ils se sentent désormais obligés de proposer chacun leurs
critères. Le danger de la subjectivité menaçant, ils se mettent à la recherche
d'une certaine normalisation. Ce besoin est surtout formulé dans le chapitre
réservé aux "critiques des critiques" qui occupe une place importante dans la
366 Demirei & Yilmaz

revue. Nusret Hyzyr, l'un des critiques influents de la revue alors, exprime
souvent son désir de déterminer les critères en critique de la traduction. Il
appelle les autres critiques et traducteurs à établir ces normes. C'est en quelque
sorte un premier pas vers la création d'une théorie de la critique de la
traduction .

1. 3 Critères en vigueur dans les critiques de traduction de la revue

Les traducteurs et critiques de la revue s'interrogeaient déjà sur la possibilité ou

non de la traduction, sur l'insuffisance de la langue d'arrivée, sur la fidélité au
texte original, sur le transfert de sens et enfin sur les compétences du
traducteur. Dans leurs critiques, ils s'appuient donc sur les notions de fidélité et
d'infidélité, sur le niveau linguistique, sur la paraphrase, sur les ajouts et les
suppressions; ils réfèrent aussi aux préfaces et aux annotations; ils s'interrogent
sur la validité des traductions de traduction, sur les traductions collectives.

1.4 Changements dans la politique de la revue et évolution de la critique de

traduction

A partir de 1946, à la suite de la traduction des classiques, on note un

renouvellement général au sein de la revue. C'est en même temps l'année où
débutent les critiques de traduction de poèmes. Les traductions des littératures
arabo-persanes, balkaniques et asiatiques (chinoise) prennent alors de
l'importance. Les membres de la revue décident aussi d'élargir leur champ de
traduction en traitant de sujets littéraires contemporains, tout en comblant les
lacunes dans les domaines artistiques et philosophiques. Cela aboutit à une
nouvelle liste d'œuvres à traduire.
L'activité de traduction encouragée au début par le ministère de
l'Education nationale commence à se répandre dans le privé c'est-à-dire les
maisons d'édition, les journaux et autres revues littéraires. Cet élan nouveau
dans le domaine de la traduction comporte des risques dus au souci commercial
qui cette fois prime la qualité de la traduction et du traducteur. Halit Fahri
Ozansoy, écrivain et traducteur de l'époque, annonce clairement, par ces
quelques mots, sa crainte à ce sujet:

Ces dernières années, les maisons d'édition, suite à une demande croissante de
traduction de romans, ont engagé beaucoup de traducteurs. Ainsi de bons et de
mauvais traducteurs ont paru sur le marché. Ma crainte à présent porte sur les
traductions de poèmes à la mode de nos jours, car j'ai bien peur qu'ils engagent
cette fois encore des traducteurs incompétents dans ce domaine. Dans ce cas-
là, tous les domaines littéraires, le lyrique, l'épique et le pastoral auront été
sabotés.6
 Critique de la traduction littéraire 367

Cette crainte préoccupe tous les intellectuels turcs de l'époque, conscients du

danger que peut comporter cette dégradation au niveau de la qualité des
traductions. Sans écarter ce danger, la culture turque ne pourra s'ouvrir
sainement aux cultures mondiales.
Le 29 janvier 1947, avec le changement du ministre de l'Education
nationale, une organisation nouvelle s'annonce au sein de la revue Tercüme7.
Cependant, suite à ces nouvelles décisions, les critiques se plaignent d'avoir
négligé l'enrichissement de la langue turque et proposent la création d'un
dictionnaire à partir des traductions8. Quelques-uns pensent même que ce flux
de traductions risque d'entraver la création d'œuvres littéraires turques.
Le problème des droits de reproduction devient aussi un sujet de
préoccupation; à cet égard, l'ouvrage intitulé Guide de l'édition et de la
librairie édité en France est pris comme modèle9.
Dans les années 50, d'autres projets voient le jour comme la traduction
d'œuvres turques dans les langues occidentales, destinée à faire connaître cette
littérature. L'UNESCO propose d'ailleurs de subventionner les dix premiers
travaux10. Pour commencer, en 1951, on tente de s'attaquer à la poésie (numéro
52 de Tercüme). Dans les autres numéros, on poursuit les efforts avec la prose,
en particulier les récits et les nouvelles.
Malgré ces projets en vue, à partir de cette époque, les critiques
deviennent peu à peu de simples répétitions, s'exprimant avec des clichés.
Elles ne portent plus de jugement, elles se limitent à des analyses soit sur
l'œuvre, soit sur l'écrivain. Seul le dernier paragraphe (ou parfois seule la
dernière phrase) est consacré à la critique de la traduction. Cette évolution est
notable à partir du numéro 45 de Tercüme, en 1948.
Désormais, la revue, en proie à diverses critiques, éprouve la nécessité
de reconsidérer ses objectifs premiers, comme l'indique le professeur Bedrettin
Tuncel, l'un des fondateurs de la revue:

Notre objectif premier était et reste d'enrichir la connaissance de nos lecteurs,

en leur proposant le plus d'exemples possible de traductions

Il met également l'accent sur la responsabilité du traducteur:

Le plus important en traduction est le fait que le traducteur doit choisir le

texte, l'écrivain qui lui convient et qui correspond le mieux à son caractère, à
son état d'âme.11

Malgré les nombreuses tentatives de renouvellement, une dégradation

apparente se fait sentir aussi bien dans la politique de la revue que dans la
qualité et la quantité des critiques de traduction. Cette dégradation entraînée
par les conditions économiques, la hausse du prix du papier et des frais de
publication, le manque de traducteurs, et qui est due aux conflits idéologiques
368 Demirei & Yilmaz

et politiques entre les membres, conduit même à des périodes de suspension, de

un à deux ans. Ainsi la revue sort avec un retard de huit mois en 1948, d'un an
en 1954, de deux ans en 1957, etc. En 1966, avec la fermeture du Bureau de
Traduction par décision politique, l'aventure de la revue Tercüme prend fin.

2. Yazko Çeviri - Traduction Yazko (1981-1984)

2. 1 Situation socio-politique et culturelle du pays dans les années 1980

La période problématique qui s'annonce aura des conséquences socio-

politiques troublantes sur la vie culturelle. Avec le coup d'Etat de 1980, la
nouvelle loi constitutionnelle apporte des "censures primitives" qui selon
Ahmet Cernai, directeur de la revue Yazko Çeviri, "nuisent directement à
l'indépendance culturelle et à la démocratisation de l'art"12. Ce coup porté à la
liberté provoque une diminution considérable du nombre des lecteurs. Dès lors,
un certain fanatisme se fait sentir dans tous les domaines. Pour lutter contre cet
esprit conservateur, les intellectuels sont appelés à s'organiser.

2. 2. Politique de la revue

Yazko, fondée en 1981 par des écrivains et des traducteurs, a été une revue
bimensuelle de traduction, indépendante de l'Etat; elle a eu pour objectif
essentiel "de préserver l'actualité de la problématique de la traduction en tenant
compte de ses divers aspects"13 . A la période de parution de Yazko, l'activité
de traduction est loin d'être alors systématique. Cette désorganisation dans ce
domaine est nettement exprimée par Ahmet Cernai:

Un grand nombre des traductions faites à cette époque peut être qualifié de
"littérature passe-temps". Cette situation est due à un souci de profit
commercial mais surtout au fait que l'on ne sait pas exactement quelle œuvre
traduire.14

Pour mettre fin à cette désorganisation, la revue s'engage à donner une place
importante aux critiques de traduction car les éditeurs de l'époque pensaient
que les traductions n'étaient pas soumises à une critique objective. De fait, ces
critiques se réduisaient alors à des textes de présentation à caractère
publicitaire. C'est pourquoi la revue décide de consacrer une de ses parties aux
"écrits sur la traduction", exclusivement réservée aux critiques. On décide de la
création d'une commission afin de contrôler d'une part la sélection des textes à
traduire et de l'autre les écrits qui seront publiés dans la revue.
En 1982, dans le numéro 7, le comité de rédaction donne des détails sur
les principes et critères qui vont opérer dans la sélection des critiques de
 Critique de la traduction littéraire 369

traduction: perfectionnement de la langue turque, conformité à la politique de

la revue. On refuse en particulier les critiques qui portent sur le traducteur (et
non sur la traduction), c'est-à-dire les textes qui sont plutôt une attaque contre
la personnalité du traducteur ou de simples "bavardages calomnieux"15.
D'ailleurs, Ferit Edgü, écrivain et critique littéraire, affirme sans hésitation que
"la critique de la traduction ne peut être considérée comme un domaine à part"
en Turquie et ajoute:

Les critiques de traduction de nos jours, qui sont seulement une énumération
d'erreurs, ont pour but d'exhiber les connaissances et la culture générale du
critique.16

Pour lui, le critique n'énumère pas seulement les erreurs; il doit être celui qui
sait aussi donner un sens profond à l'œuvre, éclairer le lecteur.

2. 3. Etape importante pour la critique de la traduction

Pour donner un élan nouveau à la critique de la traduction et pour essayer de

définir les principes d'une critique objective, les membres de Yazko décident de
réserver alors une partie de la revue aux théories de la traduction, à la
traductologie. Ainsi, en se fondant sur des bases scientifiques et objectives, la
critique pourrait cesser d'être "une chasse à l'erreur" et occuper une place
importante en tant que sous-domaine de la traduction. Dès lors, les théories de
la traduction, avancées par des auteurs occidentaux, sont prises en
considération et font l'objet d'écrits d'universitaires turcs17. Les premières
réactions basées sur les théories entraînent des polémiques entre traducteurs et
critiques. Pour donner un exemple de ces polémiques, on peut citer la réponse
du traducteur Fatih Özgüven à la critique théorique de Nilüfer Kuruyazici:

A quoi sert de se baser sur des théories scientifiques si vous continuez à faire
'la chasse à l'erreur', si en pratique vous contredisez votre propre hypothèse
théorique.18

Cette critique de la critique, tout en montrant les limites de la théorie mal ou

non encore adoptée par les traducteurs, annonce une période de transition vers
une critique objective.
Malheureusement, au début de 1984, pour des raisons politiques, la
revue Yazko ne peut plus bénéficier de l'aide des chercheurs. Elle cesse
subitement de paraître après le numéro 18 (mars-avril 1984).
370 Demirei & Yilmaz

3. Metis Çeviri - Traduction Métis (1987-1992)

3. 1. Objectifs de la revue

Cette revue trimestrielle est parue pour la première fois en 1987, fondée en
grande partie par des académiciens; son objectif est surtout d'encourager et
d'orienter les jeunes traducteurs, sans négliger les théories de la traduction.
Les conséquences de la traduction sur la langue et indirectement la
culture turques, ont amené cependant les intellectuels et les hommes de lettres
à se pencher aussi sur les problèmes de la structure de la langue turque. En
effet, l'utilisation abusive d'expressions et l'assimilation de règles
grammaticales, issue de langues étrangères, entraînaient, peu à peu, une
déformation inquiétante de la langue turque, en train de devenir "un nouveau
jargon indomptable"19, selon la préface même du premier numéro de la revue.
Un des moyens le plus efficace pour lutter contre cette déformation
linguistique est la critique de la traduction.

3. 2. Influence de diverses disciplines sur la critique de la traduction

La "fonctionnalité", concept nouveau20, commence à être discutée autant en

traduction qu'en critique de la traduction. Dans cette perspective, l'idée qu'un
texte puisse avoir plusieurs traductions met en question la valeur de la critique.
L'interaction entre les disciplines commence aussi à jouer un rôle
important dans l'expansion de la terminologie de la traduction. Les chercheurs
turcs en traductologie empruntent ainsi peu à peu des termes à la linguistique,
et à la sémiotique. Mais cette inflation terminologique en traduction retarde
aussi bien l'assimilation de toutes ces approches nouvelles que leur mise en
application dans les critiques de la traduction.
N'empêche, les nouveaux horizons théoriques redéfinissent les
responsabilités du critique et attribuent des responsabilités au lecteur. Le
critique qui devrait servir de pont entre la traductologie et la pratique de la
traduction ne peut plus négliger la théorie de la traduction. Quant au lecteur, il
devrait être plus sélectif à l'égard des mauvaises traductions et plus actif dans
l'activité de la critique afin d'exercer une certaine pression sur la politique des
maisons d'édition.21
Metis Çeviri, après avoir suscité maintes discussions en traduction,
préfère cesser de paraitre en automne 1992, afin d'éviter le déjà-dit.
 Critique de la traduction littéraire 371

4. Conclusion

En Turquie, l'époque la plus brillante pour la critique de la traduction littéraire

est celle des débuts de la revue Tercüme. Grâce à une politique démocratique
universaliste et surtout à une activité systématique de traductions, la critique de
la traduction a pu alors se développer librement.
Avec Yazko Çeviri, la traduction fait ses premiers pas vers la théorie.
Parallèlement, les membres de la revue tentent de définir une autre approche de
la critique de la traduction. Cependant, la critique qui doit faire face à une
régression de la qualité des traductions et en même temps à des restrictions
imposées par les critères commerciaux, vit une période de stagnation.
La revue Metis Çeviri, quant à elle, profite des apports de la
traductologie mais le fait que les théories ne soient pas appliquées directement
à la pratique empêche leur assimilation et bloque l'évolution de la critique de la
traduction.
Aujourd'hui encore, en Turquie, nous ne pouvons pas affirmer que la
critique de la traduction bénéficie d'une certaine autorité. Le climat socio-
politique et les difficultés de l'interdisciplinarité en sont les causes essentielles.
"La co-existence de nombreux et divers critères d'évaluation et d'approches
offre aux critiques, aux lecteurs et aux traducteurs contemporains un défi",22 ne
favorisant pas le statut de la critique de la traduction.

Notes

1. Comme notre recherche se concentre sur l'analyse de ces trois revues de traduction, nous
nous contenterons de reporter les références dans la partie Notes, sans avoir recours à une
partie Bibliographie. Toutes les traductions sont des auteurs du présent document.

2. Ataç'i Aniş (Commémoration de N. Ataç). Turk Dil Kurumu Yayinlari (Editions de

l'Association de la langue turque). Ankara Üniversitesi Basimevi (Maison d'édition de
l'Université d'Ankara). 1968. p.26.

3. Tuncel, Bedrettin. "Tercüme Meselesi" (Affaire de la traduction). Tercüme 1, 19 mai

1940, p.79.

4. Ataç, Nurullah. "Ikinci Yila Girerken" (Au Seuil de la deuxième année). Tercüme 7, 19
mai 1941, p.3.

5. Hizir, Nusret. "Tercüme Tenkitleri Hakkinda" (A Propos des critiques de traduction)

Tercüme 20-21, 19 septembre 1943, p. 177.

6. Ozansoy, Halit Fahri. "Edebiyat"(Littérature). Tercüme 37, 19 mai 1946, p.96.

7. "Haberler" (Les nouvelles). Tercüme 41-42, 19 mars 1947, pp.435-437.

372 Demirei & Yilmaz

8. Rado, Şevket. "Yeni bir Tercüme Devrine Girerken" (Au seuil d'une nouvelle période en
traduction). Tercüme 45, mai-juin 1948, p.250.

9. Görk, Vecihi. "Telif Hakki" (Droit de Reproduction). Tercüme 46, juillet- août 1948,
pp.338-344.

10. "Önsöz" (Préface). Tercüme 52. mai 1951, p.217.

11. Tuncel, Bedrettin. "Kisa Notlar" (Brèves notes). Tercüme 60, avril-juin 1955, p. 135.

12. Cemal, Ahmet. "Anayasa Taslagi Üzerine birkaç Söz" (Quelques propos sur la nouvelle
loi constitutionnelle). Yazko Çeviri (Traduction Yazko) 8, septembre-octobre 1982, p.6.

13. Cemal, Ahmet. "Başlarken" (En guise d'avant-propos). Yazko Çeviri (Traduction Yazko)
1, juillet-août 1981, p.9.

14.Ibid.

15. Cemal, Ahmet. "Türkiye'de Çevirmenlik" (La profession de traducteur en Turquie). Yazko
Çeviri (Traduction Yazko) 15, novembre-décembre 1983, p.8.

16. /leri, Selim. "Ferit Edgü'yle Çeviri Üzerine" (Reportage avec Ferit Edgü sur la traduction).
Yazko Çeviri (Traduction Yazko) 11, mars- avril 1983, p.85.

17. Voir l'article de Paker, Saliha. "Çeviride 'yanliş/dogru' Sorunu ve Şiir Çevirisinin
Degerlendirilmesi" (Le Problème du juste et du faux en traduction et l'évaluation de la
traduction de poème). Yazko Çeviri (Traduction Yazko) 13, juillet- août 1983, p. 131-139.

18. Özgüven, Fatih. "Thomas Mann Çevirmeninin Cevabi" (Réponse du traducteur de Thomas
Mann). Yazko Çeviri (Traduction Yazko) 15. Op. cit.. p. 153.

19. "Préface". Metis Çeviri (Traduction Metis) 1, automne 1987, p.7.

20. Le terme de fonctionnalité apparait pour la première fois chez les théoriciens tels que
Gideon Toury et Hans J. Vermeer. Par ailleurs, dans les revues que nous avons consultés,
d'autres noms sont cités: par exemple Anton Popovic, Raymond van den Broeck, etc.
Ainsi, apparaissent plusieurs interprétations de la "fonctionnalité".

21. Tapan, Nilüfer; Tanyeri Ergand, Çaglar. "Critique de traduction". Metis Çeviri (Traduction
Metis) 19, printemps 1992, p. 116.

22. Maier, Carol 1998. "Reviewing and criticism". Routledge Encyclopedia of Translation
Studies ed. by Mona Parker, p.209. London & New York: Routledge.
THE DOMESTICATED FOREIGN

Outi Paloposki & Riitta Oittinen

Helsinki University & Tampere University, Finland

Zusammenfassung
Im Fokus dieses Artikels stehen generelle ÜbersetzungsStrategien und ihr zeitlicher
Wandel. Das Hauptanliegen ist zu zeigen, wie die Einbürgerung in unterschiedlichen
Übersetzungen in verschiedenen Zeiten gewirkt hat. Die Konzeptionen der
Einbürgerung und Verfremdung werden hier auch generell diskutiert. Paloposki befaßt
sich mit der Einbürgerung von William Shakespeare's Macbeth in Finnland Anfang
des 19. Jahrhunderts. Oittinen konzentriert sich auf die Einbürgerung mit Blick auf
Kinder, insbesondere bei den drei finnischen Übersetzungen des Werkes Alice's
Adventures in Wonderland von Lewis Carroll.

Résumé
Nous nous attacherons ici aux stratégies globales en traduction, à leur évolution. Nous
montrerons en particulier jusqu'où diverses traductions de différentes époques ont été
"naturalisées", pour discuter plus globalement de l'opposition entre naturalisation et
exoticisation. Paloposki traite de la naturalisation du Macbeth de Shakespeare, en
Finlande, au début du 19ème siècle tandis que Oittinen analyse la naturalisation des
livres pour enfants, surtout dans les trois traductions finnoises de Alice de L. Carroll.

Resumen
En este trabajo, nos centraremos en las estrategias globales de la traducción y su
evolución a través del tiempo. Nuestro principal objetivo sera demostrar que la
"naturalización" se ha utilizado en distintas traducciones en diferentes épocas, pero
también abordaremos los conceptos de "naturalización" y "extranjerización" en
general. Paloposki aborda la naturalización del Macbeth de Shakespeare a la Finlandia
de principios del siglo XIX. Oittinen se centra en la naturalización para niños,
especialmente en las tres traducciones al finlandés de Alice's Adventures in
Wonderland de Lewis Carroll.
374 Paloposki & Oittinen

1. Introduction

Time can play havoc with our ideas of translation. At some point in history, the
most highly regarded translations were those that conformed to the ideal of
'belles infidèles', whereas at some other stage, accuracy was the hallmark of
good translations. In addition to time, place plays an important role in what is
considered a good translation. These two variables form an (albeit simplified)
skeleton for studying two strategies translators have often put to use in their
translations: foreignization and domestication.
Foreignization generally refers to a method (or strategy) of translation
whereby some significant trace of the original "foreign" text is retained.
Domestication, on the other hand, assimilates a text to target cultural and
linguistic values (Robinson 1997b: 116-117; see also Chesterman 1997:28).
The most recent - and probably also most fervent - critic of domestication,
Lawrence Venuti (see eg. 1995:18-22) has attacked domestication as a site of
ethnocentric racism and violence, "...an ethnocentric reduction of the foreign
text to target-language cultural values..." (ibid.:20). His preferred method of
translation for literary texts is foreignization; or "...resistancy, not merely
because it avoids fluency, but because it challenges the target-language culture
even as it enacts its own ethnocentric violence on the foreign text" (ibid. :24).
For Venuti, there are a number of reasons why foreignizing is desirable
(and domestication is to be rejected). For him, domesticated translations
"conform to dominant cultural values" (ibid. :291), whereas foreignization
"challenges the dominant aesthetics" (ibid. :309) (like when he himself
translates the Italian poet Milo di Angelis). Secondly, foreignized translations
"signal the linguistic and cultural differences of the foreign text" (ibid. :311).
And thirdly, foreignization "seeks to restrain the ethnocentric violence of
translation" (ibid. :20). From this, it can be inferred that foreignness as such is
something desirable, and that domestic values, linguistic codes, and aesthetics
are undesirable (as they should be challenged, not conformed to). Looking at
these claims, it is interesting to note that Venuti actually denies that he would
be interested in the foreign as such (ibid. :41-42), making a point of using
"foreignness" only as a strategic tool. As Anthony Pym (1996:166-167) notes,
Venuti's own language use certainly defies easy understanding, and unearthing
aims and causes in his text is rather like the work of a detective.
Venuti's approach has been criticized before (see, e.g., Pym 1996;
Lane-Mercier 1997). In this article, we would like to call attention to a number
of empirical examples of domestication in practice and to point out that they do
not conform to the simple mechanistics of "bad" (domesticated) versus "good"
(foreignized) translations. While Venuti's examples offer interesting insights
into the background of several translators, foreignizers as well as
domesticators, and his analysis can be seen as a refreshing challenge to (some)
Anglo-American literary translation practices, his generalisations are likely to
 The domesticated foreign 375

be less convincing when checked against different kinds of data. It is not only a
question of how texts are translated (whether they are domesticated or
foreignized), but why these strategies have been used. Anthony Pym (1992:
222; 1998: 5-6) addresses roughly the same problem when he contends that the
"what?" and the "how?" questions (what he calls "translation archaeology" and
"translation criticism") are not enough if a study is to be properly historical.
The "how?" questions logically precede the "why?" questions, but it is the
latter that help us understand the phenomena in question.
Lawrence Venuti has recently elaborated on many of his previous ideas
(see, eg., Venuti 1998); among other things, the use of the term
"foreignization" has given way to a new term, "minoritizing translation"
(ibid. : 12), the aim of which is "... never to erect a new standard or to establish a
new canon, but rather to promote cultural innovation as well as the
understanding of cultural difference by proliferating the variables within
English..." (ibid. :11). However, we wish to challenge a view of foreignization
which advocates this method as the only morally acceptable alternative for a
translator conscious of her/his choices and their consequences in a world of
power politics, racism and ethnocentrism. We wish to show that translation is
(yes, quite rightly) a battlefield of many opposing strategies and views, and that
two seemingly opposing strategies can be aiming at similar effects, while one
and the same strategy can be used for diametrically opposed purposes.

2. Shakespeare translation as domestication (O.P.)

Domestication is an elusive term: it can entail a wide variety of different

things, and marking the boundaries between what is domestication and what is
foreignization is nearly impossible. To establish some common ground for a
discussion of domestication, I have tried to identify specific phenomena that
could be studied as instances of domestication. We must bear in mind that if
we call something a domesticated text, there must be at least one alternative
way of rewriting it (following a supposedly "foreignizing" strategy). First,
however, I will provide some background for the location and period I will be
dealing with.

2.1 Finland in the early 19th century

Nearly two hundred years ago, as the result of prolonged wars between Sweden
and Russia, Finland became an autonomous part of the Russian Empire. The
earliest roots of Finnish as a literary language go back to the times of the
Reformation and Bible translation, but it was not until the Russian era that the
first translations of fiction appeared in Finnish, in a situation where cultural
patriotism was slowly changing into national awakening. In the early days of
376 Paloposki & Oittinen

the Russian era, not much changed, though, on the political scene - the
legislation and administrative system remained largely Swedish, according to
the promises made to Finns by their new ruler, Czar Alexander the First. The
Finns could not know for sure, either, whether this was to be their permanent
state, or whether Finland would change hands again and return to Sweden. But
the changed status of the Finnish province, together with a gradual tightening
of the grip of the ruler, gave rise to a more acute awareness about Finland's
identity. The search for identity focused on the Finnish language, the
development, standardization and literary eloquence of which rapidly became
important and even controversial issues. Texts written in Finnish - both
translated and original - started to appear outside the traditional categories of
legal and religious texts (religious texts had hitherto counted for 82% of all
published texts). In 1831, the founding of the Finnish Literature Society
institutionalized the aim of making the Finnish language (which even many
Swedish-speaking Finns had started to call their "mother tongue") one of the
pillars of the future nation. Closely related to this elevated status of the
language was the collection and writing down of songs and poems from the
oral tradition, which started as early as the 17th century.
In addition to this major endeavour of compiling a Finnish epic, all
other kinds of written texts became valuable for the aims of the nationalistic
campaign. Translated fiction was to serve the Finnish cause in creating reading
materials for the public, educating and "improving" people, and for polishing
the language so that original writing could then emerge. Although language is
not the only (and often not even the most important) component of nationalism
(see Hobsbawm 1991:20-22), it was at times a very important element in the
nationalist struggle. Finland is not an isolated example: similar cases of
enhancing the status of national languages have been reported in other parts of
Europe at certain stages of a nationalist awakening, and the Czech and the
Slovak language policies of the time seem rather parallel to that of Finland
(Hansson 1984; Hobsbawm 1991; Schulze: 1993:58-59; Steinberg 1987:203-
204).

2.2 Translated fiction in Finland

The first four books of translated fiction in Finland appeared, quite

independently, in 1834. Among them was Shakespeare's Macbeth, translated
as Ruunulinna. Interestingly, these translations pre-dated the first edition of
Kalevala, the great Finnish epic, which only appeared the following year. I will
here look at Ruunulinna and its Finnish contextualisation.
Shakespeare's plays are fruitful ground for the study of domestication,
as his works have been adapted and domesticated in different ways in different
parts of the world, time and again. True, there are those who claim to look for
the "original" or "essential" Shakespeare, but the fact remains that within the
 The domesticated foreign 377

Shakespeare tradition, there have been highly contextualised versions (for

Shakespeare translation, see, e.g., the volume edited by Delabastita and D'Hulst
in 1993; Heylen 1993; Fridén 1986). Different European traditions adopted
Shakespeare now in neo-classical, now in Romantic versions.
Macbeth was first translated into Finnish by a retired army major, a
veteran of the Swedish-Russian wars (the outcome of which had been the
annexation of Finland to Russia). His keen literary interests included a passion
for the Finnish language and an urge to promote it. Although we can only
speculate on the reasons that led Major Lagervall to translate Macbeth, it does
not seem an altogether arbitrary choice. Among Shakespeare's plays, Macbeth
is the one that was soonest and has been most often translated. If we assume,
like Wolfgang Ranke (1993:168), that there are three levels of significance in
Macbeth - the moral level (individual guilt and penance), the political level
(tyranny and the restoration of order by government) and the metaphysical
level (the natural order threatened by forces hostile to nature) - we can easily
understand the appeal of Macbeth throughout the times to different audiences.
Each generation or group can foreground the elements that seem to apply most
easily to their circumstances; thus, Schiller in his translation stressed the power
of the state and downgraded the metaphysical element, with witches that were
turned into responsible, thinking creatures (Donner 1950:9; Ranke 1993:171,
176), whereas a Swedish translator of Macbeth, Eric Geijer, retained the
supernatural forces in his 1813 version.
In Finland, texts to be translated tended to be chosen from the literary
canon of the day. Shakespeare certainly was well-known in Finland, both
through translations of his works into Swedish and German and through the
performances of travelling theatre companies that put on his plays in Swedish.
His works were discussed in newspapers and in personal correspondence, and
his influence was recorded in memoirs and diaries. Secondly, Macbeth's
themes - war, betrayal, loyalty - in addition to being considered universal,
were clearly very tangible in the Finland of the early 19th century. They were
the personal experiences of a seasoned war veteran who had now become a
Shakespeare translator. Moreover, the Scottish countryside depicted in
Macbeth found an easy counterpart in Karelia, the Eastern part of Finland,
which Lagervall decided to use as a setting for his Macbeth - or Ruunulinna
(Ruunulinna meaning "crown castle" or "royal castle"). For Ruunulinna
definitely was a domesticated text, if ever there was any.
Macbeth's arrival in Finland was greeted by interest and approval by
the literate circles. However, Lagervall's use of the Finnish language and the
runic metre were not altogether approved of, and he was offered the chance to
correct his translation along the lines suggested to him by a "language board"
of the Finnish Literature Society - a chance he refused, publishing the book at
his own expense, instead of having it published by the Society. The Society's
minutes and members' correspondence show that there were no hard feelings
378 Paloposki & Oittinen

afterwards, and even newspaper criticism was largely favorable (even the
British press hailed "Bunulinus" (sic). As for wider audiences, Ruunulinna
obviously did not attract any great readership, and when the first plans for a
Finnish theatre were hatched and Lagervall suggested Ruunulinna as an
obvious choice of a play, he was not taken up on his offer.

2.3 What can be domesticated?

Domestication, in Venuti's terms, refers both to fluency, which, for Venuti,

counts as one of the most striking features of domestication, and to the
inscription of domestic values throughout the rewriting/translation of a text (cf.
Venuti 1995:5-6; 49). This latter takes the form of "adding" or clarifying
things, as in the translations of for example Denham and D'Ablancourt. These
two dimensions of domestication work together: fluency and familiarity
"respected bourgeois moral values" (ibid.:130) - in Matthew Arnold's
conception of translation, in this case. Mapping a textual strategy, like
domestication, onto underlying values (like "bourgeois moral values") is not an
easy task, though. I have therefore tried to identify, as instances of
domestication, lower-level linguistic and stylistic changes in translated texts.
The changes made in Macbeth/Ruunulinna are an example of these.
The most evident change, the one that can be seen instantly, is the
changing of names in Ruunulinna. Both the protagonist and all the other
characters bear Finnish names, and so do all the places. Now, translating names
is a very common practice, both in children's literature and in new translating
cultures, and it is often attributed to a desire to make the names easier to
understand and pronounce (for children; for audiences unaccustomed to foreign
names). Phonological familiarity, easy identification, and comprehension are
thus often thought to be the reasons behind this kind of domestication.
In Ruunulinna, though, facilitating recognition and recall of names may
not be the primary reason for their translation. Finnish names in Ruunulinna
imply more than phonological ease: they stand for a historical claim. It was
important for Lagervall that Finland should have a heroic past, an epic history.
In his epilogue to Ruunulinna, he states the reason for the change of names.
The problem, says Lagervall, that Walter Scott had evoked when asking
whether Macbeth really happened in Scotland can now be settled: Macbeth did
not take place in Scotland, it took place in Karelia, in Finland. Scotland thus
becomes Eastern Karelia, a wild and vast area in Eastern Finland, bordering
Russia. Some of the protagonists are named after legendary Finnish heroes in
the Swedish-Russian wars; the three sisters are Finnish mythological creatures;
and the geographical locations of the different places mentioned in Ruunulinna
are explained in an appendix to the book. Lagervall knew the old Finnish
legends well, he was acquainted with the oral tradition and familiar with
 The domesticated foreign 379

Finland's history. Ruunulinna offers the readers a new way of looking at

Finnish history and mythology.
The play was further localized by the use of Finnish runic verse, the
Kalevala metre, making use of repetition and other poetic devices typical of
Finnish oral tradition. Finnish words are rather long on average and there is a
consistent, marked initial stress. This is why Finnish poetry has made use of
rather original devices - poetic forms such as alexandrines would be highly
improbable in Finnish. The Finnish Kalevala meter is alliterative, unrhymed,
non-strophic; a kind of trochaic tetrameter consistent with the prosodic pattern
of Finnish. Poems are usually formed of pairs of parallel lines where the
second line restates but does not repeat the lexical content of the first (see
Bosley 1997:13; DuBois 1995:15-18).
The Finnish atmosphere created by the use of this verse was further
reinforced by the picture on the title page: a Finnish runic singer sitting by a
lake, accompanying himself with the traditional Finnish instrument, the
kante le.
For an idea of the use of the Finnish meter, consider the following
extract, from Act I, Scene I in Macbeth/Ruunulinna (note that repetition and
the eminent role given to witches in the Finnish version renders this scene in
Finnish longer than Shakespeare's text):

MACBETH I.i. 12-13: All [the three weird sisters]

Fair is foul, and foul is fair:
Hover through the fog and filthy air.

RUUNULINNA I.i.73-78: Muut [the witches]

Liehtokaam, kiehtokaam
Painakaamme paletta,
Vatustakaam valetta,
Nostakaamme painumaan,
Painakaamme nousemaan,
Liehtokaam, kiehtokaam

The overall strategy for translating Macbeth into Finnish in 1834 thus
consisted of changes on two levels: 1) changing the setting (changing and
explaining, if necessary, all names, both people and places) and all references
to historical events, persons etc. to ones from Finnish history or mythology; 2)
replacing the original form with Finnish runic verse.
Domestication is not necessarily dependent on one reason or factor
only. There may be several reasons behind the need or desire to domesticate a
text. I have here looked at one possible explanation for domesticating: the
historical role given to translations in the creation of not only an indigenous
literary tradition but also a history, a heroic past that would justify patriotic
380 Paloposki & Oittinen

feeling. There may have been other reasons behind domestication, less easy to
detect. Translating poetry has often been considered an act of new literary
creation. Perhaps Lagervall was (unconsciously) applying Percy Bysshe
Shelley's famous idea about transplanting the seed of a poem ("The plant must
spring again from its seed or it will bear no flower"; in A Defense of Poetry
from 1821). Ruunulinna is Macbeth born again in a different world, sprung up
from seed. Underlying Lagervall's creative work, there may have been a
conception of art as universal heritage and of other people's work as a seed that
can be transplanted.
While Lagervall uses Macbeth as material for his own literary creation,
he also offers his work for the good of the Finnish language and for the
creation of a literary language, a theatrical tradition and a canon of history. A
tall order, but such was the time that grand visions and high hopes could give
rise to projects like Ruunulinna.
This domesticated Macbeth can thus be seen as an attempt not only at
the improvement of the Finnish language or the enriching of Finnish literature,
but at the creation of a history worthy of admiration on a national scale.
Ruunulinna is far from the imperialist violence attributed to domestication: on
the contrary, it can be seen as one small attempt by one individual to enhance
the status of a minority language with hardly any literary tradition in fiction.

3. Domesticating for children (R.O.)

Within children's literature, domesticating and foreignizing are delicate issues

(see Oittinen forthcoming). Several scholars take a clear stand against
domesticating (adaptations being a case in point here), as they feel it denatures
and pedagogizes children's literature. Another reason for their negative views
about adaptations altogether is the way translation is seen: if translation is
understood as producing "sameness", there definitely is a clear distinction
between translations and adaptations. On the other hand, if all translating is
considered to be rewriting, it is much more difficult to tell one from the other.
In her Poetics of Children's Literature (1986), Zohar Shavit deals with
adaptations in children's literature. She takes into account such issues as time,
place, culture, and even different child images. Even though she does not deal
with translation explicitly, her studies form a sounding board for looking at
translation from the point of adaptation. In this sense, Göte Klingberg's scope
is narrower: in his work on translating and adapting children's books, mainly in
Children's Fiction in the Hands of the Translators (1986), he concentrates on
words and text fragments in isolation, with the goal of formulating strict
translation rules, among them the principle of never domesticating.
Shavit studies versions and adaptations from the standpoint of the status
of children's literature. Regardless of the reasons she finds for adapting
 The domesticated foreign 381

literature for children - appropriateness and comprehension, plus taboos that

need censoring - she contends that adapting (domesticating) is a sign of
disrespect for children. Plots and characters are usually less ambiguous in
children's literature than in stories for adults: characters are either totally good
or totally bad (Shavit 1986:119ff).
Shavit also compares the child images in the different versions of Little
Red Riding Hood by Jean Perrault and the Grimm brothers. The first version by
Perrault appeared in 1697 and it has an unhappy ending, with the wolf eating
up Little Red Riding Hood. In the Grimm version, Riding Hood and her
grandmother manage to get out of the wolf's stomach, and the wolf dies. For
Shavit, the different endings reflect a profound difference in the child concepts
of the Grimm brothers and Perrault. In the times of the Grimm brothers, family,
the child's innocence and the pedagogy of fairy tales were considered very
important. Thus the Grimm version, given the form of a fairy tale, is a moral
tale, where evil is punished. Even the family relations are much closer in their
version: the grandmother loves her grandchild and has sewn a hood for her,
which does not happen in Perrault's tale (Shavit 1986:13, 22). For Shavit, the
different versions of Little Red Riding Hood clearly show that the changes
made to the children's versions are neither minor nor insignificant.
Like Shavit, Klingberg distinguishes between translation and
adaptation. But whereas Shavit's contention arises from an explicit agenda to
heighten the status of children's literature, Klingberg's view is based on an
understanding of translation as producing "sameness". To Klingberg it seems
natural that the function of the translation is always the same as that of its
original. He suggests that as the author of the source text (for children) has
already taken into consideration her/his readers, the only task of the translator
is to keep to the same degree of adaptation as in the original, that is, she/he
should keep to functional equivalence: "The translation should not be easier or
more difficult to read, be more or less interesting, and so on. We could thus try
to find methods to measure the degree of adaptation in the source text and in
the translation and to compare them" (Klingberg 1986:65; see also Nida
1964:159, 167; Nida and Taber 1969:24).
On this basis, any alteration at the translation stage is negative: it is
"manipulating" the word of the original, as Klingberg argues. The attitude is
echoed in Carmen Bravo-Villasante's views: she finds anti-localizing
(foreignizing) the only way to treat foreign material: "The criterion by which
the originals should be adapted to the practices of the country in question so
that they can be understood better, results in distortion of the text" (Bravo-
Villasante in Klingberg et al. 1978:48). Her opinions are a sign of adult worry
about children not learning "enough", not becoming educated "enough" - from
an adult point of view. They show that we adults have little faith in our
children's ability to find knowledge and information by themselves. We
undervalue the role of imagination in learning. Another important issue here is
382 Paloposki & Oittinen

that apart from the names of flowers and capital cities, children learn many
other important things from books, too: children need to be emotionally
involved so that they learn to understand other people's feelings in different
situations. Stepping into someone else's shoes is easier in a book than in real
life.
Seeing adaptation as negative (like Shavit, Klingberg and Bravo-
Villasante all do) is in line with Venuti's ideas on domestication. Yet
adaptations are products of their contexts, and it would be more interesting to
study the justifications made in favour of either adapting (domesticating) or
foreignizing the text. Thus the point is not whether adapting or domesticating is
a negative or positive phenomenon as such. Rather, what is at issue is the
purpose of the whole translation project, the translation situation, and the
translator's child image.
Klingberg - and to some extent Shavit, too - deals with texts and
languages as closed systems with permanent meanings, paying less attention to
the reader's participation and creative understanding. For Mikhail Bakhtin,
texts can be understood in quite a different way: they are unities, the parts of
which are understandable on the basis of the whole (as well as the whole
reading situation) and the relationship between the whole and its parts.
Klingberg and Bravo-Villasante's fears about children "not learning enough"
are akin to "the authoritarian word" described by Bakhtin: given from above, it
would not be understood in an active way (Bakhtin 1990:342). Thus
foreignizing can be, oddly enough, very authoritarian, and against Venuti's
ideas. As Douglas Robinson (1997a: 79-131) points out, foreignizing is akin to
authoritarian rhetoric in schools and politics, designed to stupefy and passivize.
In the following, I will take a closer look at three full translations of
Lewis Carroll's Alice's Adventures in Wonderland. The original story of Alice
was first published in 1865, and has since been interpreted from innumerable
perspectives. As Martin Gardner (1970:8) has pointed out, the Alice books
"lend themselves readily to any type of symbolic interpretation - political,
metaphysical, or Freudian". Yet most scholars seem to agree on one thing: the
story is a parody, intentionally throwing mud on all our "sacred cows" like
school, religion, old age, babyhood, and family life. In Finland, there are three
full "Alice" translations: in 1906 appeared the first translation by Anni Swan;
in 1972, a translation by Kirsi Kunnas and Eeva-Liisa Manner; and in 1995 the
latest translation by Alice Martin (see Oittinen 1997.)
All three translations were created in different situations and served
different purposes, which is easy to understand: Finland has changed a lot in
ninety years. All three translators have domesticated or foreignized their texts
in different ways. Generally, all three have domesticated the British culture and
brought it closer to Finnish culture, language, and society. What differs in these
translations are their different target groups, the future readers of the text. The
 The domesticated foreign 383

first two translations are clearly targeted (domesticated) for child audiences,
while the third, most recent translation is far more adult.

3.1 "Alice" in 1906

When the first Finnish "Alice" appeared in 1906, Finland was part of the
Russian empire. Strong nationalistic pressures in Russia had led to a series of
severe measures against the autonomy of Finland, and more were to come
before independence would be gained 11 years later. In the early 20th century
the number of books originally written in the Finnish language was not yet
very great. There was a need to develop Finland's literary language, and
influences from other languages and other cultures were sought. New themes
and new genres were needed; we could say Finland needed to be "foreignized".
The translator of the first "Alice", Anni Swan, was a children's author
and translator, who took a great interest in the position of the Finnish language.
Swan used foreign literature to nourish the Finnish language and culture.
Swan's Alice is like little girls used to be in early 20 th -century Finland: she is a
polite little country girl with a Finnish name, Liisa. The 1906 translation
clearly mirrors the child image of early 20th-century Finland. Little girls were
supposed to be nice and gentle. She even curtseys in a situation where, in the
1995 version, Alice only solemnly bows. Of the two Alices written for children
(1906 and 1972), this one was written from an adult perspective.
But what is domesticated here really? As I see it, it is the child image:
the story of Alice comes from Great Britain but the child image of the story is
very Finnish. Swan's "Alice" is domesticated for Finnish readers: the story
seems to take place in Finnish surroundings and the main character seems
Finnish, too. On the other hand, there is also an element of foreignization: by
introducing the story of Alice to Finnish readers, Swan also introduces a new
genre, nonsense, to Finland. In this way Swan's translation is, at the same time,
both domesticated and foreignized. I find this "both and" very interesting: the
"two issues" are actually part of the same whole, that of the change taking
place every time a text is translated.

3.2 "Alice" in 1972

When Kirsi Kunnas and Eeva-Liisa Manner's translation appeared, Finland had
gone through a period of industrialization and urbanization, which had changed
the country thoroughly in a decade. This was, of course, reflected in literature,
both in originals and in translations: the themes gradually became more
urbanized and the change showed in lexical choices and in the metaphors used.
The 1970s were also years of radicalism and political movements. People were
more matter-of-fact, more serious; fantasy was not considered good for
children. The 1960s and 1970s were clearly marked by educational ideals.
384 Paloposki & Oittinen

Thus the new Finnish version of the anarchic story of Alice, parodizing
rules and regulations, was certainly something else than a reflection of the
times: it was an attack against the seriousness of the time. As Kirsi Kunnas
says, she resisted many of the phenomena of the time, especially the prevailing
child image. In the 1970s children were supposed to act like small adults - a
way of looking at children and childhood Kunnas did not agree with.
Kunnas and Manner's Alice - still with her Finnish name, Liisa - is a
very capable, impertinent, even impudent girl: she seldom thinks twice, she
hardly ever ponders on things or thinks things over. She speaks abruptly and
responds very quickly. The translation is full of carnivalistic laughter. This
Alice laughs shamelessly at the adult phenomena of the Finnish 1970s. This is
a translation domesticated for child readers with their viewpoint in mind.

3.3 "Alice" in 1995

In 1995, the year when Finland became a member of the European Union, the
latest Finnish "Alice" appeared. In the 1980s and 1990s, the Anglo-Saxon
orientation has been very strong in Finland. Today, watching television, we
know (or we think we know) a lot about countries like the United States and
Great Britain. This certainly adds to our toleration of foreign names, places,
and milieux, even if this knowledge about the foreign may be shallow.
This shows in translations. Translators need not add extra explanations
or domesticate the stories to a great extent. For instance, names need not
necessarily be translated. Unlike the 1906 and 1972 versions, the 1995
translation lets Alice keep her British name. There are also bits and pieces
omitted from the first two translations but included in the newest version: the
references to another culture (such as to Shakespeare) are no longer considered
too strange for Finnish child readers.
In the three Finnish "Alice" translations, the reasons behind the
different solutions and ways of domesticating seem to lie in the translators'
different strategies and different audiences, and different child images. The
first two translations were domesticated to make them more accessible, but the
most recent translation is clearly directed to older readers. Through her
accuracy, Martin, the translator of the 1995 version, has also been able to give
a more thorough picture of the story and even the author's background. But
does this really mean that the translation is foreignized, or that the original has
been foreignized through translation? Is it not always the case that when texts
are translated they always - to a certain extent - become domesticated as well?

3.4 Instances of domestication and foreignization in Finnish "Alices"

Alice Martin has especially underlined the otherness, the foreignness of the
book. While the first two translators have, from a certain point of view,
 The domesticated foreign 385

localized (domesticated) their translations and deleted anything strange for

Finnish child readers, Martin has used a different strategy: she has anti-
localized the story.
One illustrative example of this difference is the scene where Big Alice
has just cried a pool of tears and Small Alice (having changed her size again)
tumbles into the pool that she herself has cried:

As she said these words her foot slipped, and in another moment, splash! she
was up to her chin in salt-water. Her first idea was that she had somehow
fallen into the sea, "and in that case I can go back by railway," she said to
herself. (Alice had been to the seaside once in her life, and had come to
the general conclusion that wherever you go to on the English coast, you
find a number of bathing-machines in the sea, some children digging in
the sand with wooden spades, then a row of lodging-houses, and behind
them a railway station.) However, she soon made out that the sea was in the
pool of tears which she had wept when she was nine feet high. (Carroll
1981:12)

Here we find the description of British seaside life in the 19th century with
quaint things likes bathing machines and wooden spades, from another time
and another culture. The first two translators have omitted the section in bold
altogether, while the third translator has kept all the details and diligently
depicted the sea, the children, the sand, the lodging-houses, and the railway
station.
If we look at "Alice" translations into other languages, we find similar
solutions. In many Italian, German, French, Spanish and Portuguese (Brazilian)
versions the section has been deleted, especially in the versions meant for small
children. This consistency shows that translators have not been sloppy but that
their strategies and audiences differed from those of the original author's.
Throughout her translation, Martin preserves the otherness of the
language, culture, time, place, and gender, while the earlier translators rewrote
the story for Finnish child readers. The 1995 version is much closer to British
culture and history, and also the two sexes are more distinctly present.
Throughout her translation, Martin has wanted to include everything, the whole
story of Alice. It is, paradoxically, this preciseness that makes Martin's text
very funny and very postmodern. Even if the story as such is clearly situated in
19th-century England, it becomes a postmodern combination of old and new,
strange and familiar, even female and male, when it is rewritten in another
language and in another time.
Gender is the issue in my second example. Martin's translation includes
several details that refer to Carroll's love for little girls and hatred of little
boys. The lullaby sung by the Duchess to a baby boy is a good example of the
differences in dealing with gender. The lullaby based on David Bates's original
386 Paloposki & Oittinen

poem is one of the many poem parodies of the book. Bates's original goes like
this:

Speak gently to the little child!

Its love be sure to gain;
Teach it in accents soft and mild;
It may not long remain.

Carroll's parody goes as follows, changing the baby into a baby boy, who later
turns into a pig:

Speak roughly to your little boy,

And beat him when he sneezes:
He only does it to annoy,
Because he knows it teases.
(Carroll 1981:44)

The Finnish translators have each been able to give expression to the nasty
tones of Carroll's parody. Yet both of the earlier translators have left out the
baby's sex and speak only of a child, while Alice Martin's translation speaks of
a baby boy. Martin Gardner, the author of The Annotated Alice, points out that
"it was surely not without malice that Carroll turned a male baby into a pig, for
he had a low opinion of little boys" (Gardner 1970: 84). Only the third
translation reveals this detail not only of the story itself but also of Carroll's
life. "De-sexing"can be seen as an instance of domestication, too. "Otherness"
involves not just language and culture, but issues like gender and being a child
(a child is an other to an adult).

4. Conclusion

Going back to Venuti, and on the basis of the data presented here, it seems
evident that foreignizing and domesticating are contextual phenomena and
need to be studied as such. Even if we agree that translating is always an issue
of power and politics, we feel that translation is more than that. The word
"foreignizing" in itself might be misleading, at least in the context of
translating: every time we translate we necessarily domesticate, one way or the
other. The text becomes part of the target-language culture and literature. The
direction of this cultural transfer also matters: translating into English is
different from translating from English.
Maybe foreignizing is an illusion which does not really exist. Perhaps
we should only speak of different levels and dimensions of domestication.
Anyway, when we speak of domesticating and foreignizing, we cannot avoid
 The domesticated foreign 387

the problem of situation. As Mikhail Bakhtin points out, in every reading act,
we meet with otherness, with other points of view (1990:279-280). This is
meeting with the foreign, which becomes domesticated through translating. If
we see translating from a dialogic point of view, as communication between
human beings, as an attempt to understand, we cannot accept Venuti's views,
as he - intentionally or unintentionally - forgets the future readers of texts. If
we do not translate for our readers, then why translate at all?
Translators normally address different audiences differently (like child
and adult readers): as translators we domesticate for our audiences, taking into
consideration their assumed views and ways of understanding. The future
readers of our translations are our superaddressees: we have a certain image of
them. Of course, these readers never exist in the flesh. Yet they are necessary,
since in this way the translator shapes his/her text into a credible whole. The
translator's child image is one kind of superaddressee.
Throughout this article, we have been referring to different purposes of
texts, different settings, different audiences, and different times. These are all
dimensions of domestication: what is domesticated, how and why. Names can
be domesticated, the setting localized; genres, historical events, cultural or
religious rites or beliefs can be domesticated. Domestication is not an
automatic product of a certain time or place, either: it can be highly
idiosyncratic. We domesticate for Finns, for children, for minority cultures, for
majority cultures, for political ideals, for religious beliefs. Whether it is cultural
imperialism or emergent nationalism, carried out for propriety reasons or for
educational purposes, depends on the situation. Alice (in Wonderland) has
become Liisa in Finland for quite different reasons than Macbeth became
Ruunulinna. Persian poets may have been "shaped" by Edward Fitzgerald
because of his feeling of cultural (imperialist) superiority (see Bassnett
1991:3); classical texts that have been retranslated into Quebecois may have
been treated that way because of a need for self-assertion on the part of a
minority culture (see Brisset 1990). Feminists may get away with "hijacking"
or "womanhandling" texts because these are seen as liberating practices (see
Godard 1990:91, 94; Simon 1996:14-16, 35), whereas their male counterparts
might not so easily find ways of justifying their domestication. Texts may also
be domesticated because of political pressures, censorship, or differing moral
values.
We are not defending any of these practices here, as we feel it is not a
question of justifying one strategy or the other; but they all have their
underlying logic. The point is, domestication does not necessarily conform to
dominant cultural values: it can also bring about the cultural difference
(advocated by Venuti) of a minor language, as in the Finnish Macbeth. Thus,
domestication cannot be explained away with notions such as "wholesale
domestication of foreign values". There is no inherent, tried and tested ethics of
388 Paloposki & Oittinen

foreignizing translation that would make it the only morally justifiable

alternative.
Lawrence Venuti looks at domestication from an Anglo-American
perspective, and his analysis of this "ethnocentric violence" is certainly worth
paying attention to. His generalizations, however, need to be tested on other
situations. What we would like to argue on the basis of our examples is that
first, both domestication and foreignization may spring from a desire to fight
against oppression (or from a desire to oppress, for that matter); and, second,
that foreignness is not the only quality in a text. There are other levels that can
be studied; there may be things that cannot be measured on a bipolar same-
different axis. With regard to foreignness and its place, there might be other
means of bringing over foreign qualities than that of non-fluent translation.

References

BAKHTIN, M. [1981] 1990. The Dialogic Imagination: Four Essays (orig. Voprosy literatury
i estetiki, 1975). Trans. Caryl Emerson and Michael Holquist. Ed. Holquist. Austin: University
of Texas Press.

BASSNETT, S. 1991. Translation Studies. London & New York: Routledge.

BASSNETT, S. & LEFEVERE, A. (eds.). 1990. Translation, History and Culture. London &
New York: Pinter Publishers.

BERMAN, A. 1984. L'épreuve de l'étranger. Paris: Gallimard. (Translation into English by S.

Heyvaert. 1992. The Experience of the Foreign. Albany: State University of New York Press.)

BOSLEY, K. 1997: To the Reader. Skating on the Sea. Poetry from Finland. Edited and
translated by Keith Bosley. Helsinki: Finnish Literature Society.

BRISSET, A. 1990. Pour une sociocritique de la traduction : Théâtre et altérité au Québec

(1968-1988). Longueuil, Québec: Éditions du Préambule.

BURKE, P. & PORTER, R. (eds.). 1987. The Social History ofLanguage. Cambridge:
Cambridge University Press.

CARROLL, L. 1972. Liisan seikkailut ihmemaassa (orig. Alice 's Adventures in Wonderland,
1865). Illust. John Tenniel. Finnish trans. Kirsi Kunnas and Eeva-Liisa Manner. Jyväskylä:
Gummerus.

CARROLL, L. 1981. Alice 's Adventures in Wonderland and Through the Looking-Glass (orig.
1865 and 1871). Illust. John Tenniel. New York: Bantam Books.

CARROLL, L. 1983. Liisan seikkailut ihmemaassa (orig. Alice 's Adventures in Wonderland,
1865). Illust. Tove Jansson (1966). Finnish trans. Anni Swan (1906). Porvoo: WSOY.
 The domesticated foreign 389

CARROLL, L. 1995. Alicen seikkailut ihmemaassa (orig. Alice's Adventures in Wonderland

1865). Ilhust. John Tenniel. Finnish trans. Alice Martin. Porvoo: WSOY.

CHESTERMAN, A. 1997. Memes of Translation. Amsterdam & Philadelphia: John

Benjamins.

DELABASTITA, D. & D'HULST, L. (eds.). 1993. European Shakespeares. Translating

Shakespeare in the Romantic Age. Amsterdam & Philadelphia: John Benjamins.

DONNER, H.W. 1950. Svenska översättningar av Shakespeares Macbeth. Åbo: Akta

Akademiae Aboensis, Ser. A:20.1..

DUBOIS, T. 1995. Finnish Folk Poetry and the Kalevala. New York & London: Garland.

FRIDÉN, A. 1986. Macbeth in the Swedish Theatre in 1836-1986. Malmö: Liber Forlag.

GARDNER, M. [1960] 1970. The Annotated Alice. Alice's Adventures in Wonderland and
Through the Looking Glass by Lewis Carroll. Illust. John Tenniel. Harmondsworth: Penguin
Books.

GODARD, B. 1990. Theorizing Feminist Discourse/Translation. BASSNETT, S. &

LEFEVERE, A. (eds.), 87-96.

HANSSON, S. 1984. Svenskans nytta, Sveriges ära. Litteratur och kulturpolitik under 1600-
talet. Kungälv: Skrifter utg. av Litteraturvetenskapliga institutionen vid Göteborgs Universitet,
11.

HEYLEN, R. 1993. Translation, Poetics and the Stage. Six French Hamlets. London & New
York: Routledge.

HOBSBAWM, E. 1991. Nations and Nationalism since 1780. Cambridge: Cambridge

University Press.

KLINGBERG, G. 1986. Children's Fiction in the Hands of the Translators Studia

psychologica et paedagogica. Series altera LXXXII. Lund: Bloms Boktryckeri Ab.

KLINGBERG, G, 0RVIG, M. & AMOR, S.. (eds.). 1978. Children's Books in Translation.
The Situation and the Problems. Stockholm: Almqvist & Wiksell International.

LANE-MERCIER, G. 1997: "Translating the Untranslatable: The Translator's Aesthetic,

Ideological and Political Responsibility". Target 9(1), 43-68.

NIDA, E. A. 1964. Toward a Science of Translating. Leiden: E. J. Brill.

NIDA, E. A. & TABER, C. 1969. The Theory and Practice of Translation. Leiden: E. J. Brill.

OITTINEN, R. 1997. Liisa, Liisa ja Alice. Tampere: Tampere University Press.

OITTINEN R. 2000. Translating for Children. New York: Garland.

PYM, A. 1992. "Shortcomings in the Historiography of Translation". Babel 38(4), 221-235.

390 Paloposki & Oittinen

PYM, A. 1996. "Venuti's Visibility". (Review Article.) Target 8(1), 165-177.

PYM, A.1998. Method in Translation History. Manchester: St. Jerome Publishing.

RANKE, W. 1993. "Shakespeare Translations for Eighteenth-Century Stage Productions in

Germany: Different Versions of Macbeth". DELABASTITA, D. & D'HULST, L. (eds), 163-
182.

ROBINSON, D. 1997a. What Is Translation? Centrifugal Theories, Critical Interventions.

Kent, Ohio & London. Kent State University Press.

ROBINSON, D. 1997b. Translation and Empire. Postcolonial Theories Explained.

Manchester: St. Jerome Publishing.

SCHULTZE, B. 1993. "Shakespeare's Way into the West Slavic Literatures and Cultures".
DELABASTITA, D. & D'HULST, L. (eds), 55-74.

SHAVIT, Z. 1986. Poetics of Children's Literature. Athens, GA.: University of Georgia Press.

SIMON, S. 1996. Gender in Translation. Cultural Identity and the Politics of Transmission.
London & New York: Routledge.

STEINBERG, J. 1987. "The Historian and the Questione della Lingua." BURKE, P. &
PORTER, R. (eds), 198-209.

VENUTI, L. 1995. The Translator's Invisibility. A History of Translation. London & New
York: Routledge.

VENUTI, L. 1998. The Scandals of Translation. Towards an Ethics of Difference. London &
New York: Routledge.
 Subject Index
A cultural references 308, 311, 321, 328,
action, 50-52
African literature, 354-355, 357
adaptation 239-240, 245, 251, 253-254,
269, 274, 276, 280, 294-295, 318, 325,
345,347,381-382
addition, 250
adequacy, 121-125, 328
advertising, 262-269, 272-273
audience (see reception)
B 380, 382, 384, 386-387
belles infidèles, 188,374
Bible, 21, 67, 68
bilingualism/bilingual, 64, 68, 130-131,
133, 136, 202-203, 206, 212, 235, 283
C
Catalan, 74, 75, 77
categorization, 132-135
Castilian, 75, 77
checking, 40-41, 43, 45, 57
children's literature, 22, 178, 252, 319,
380
Chinese, 22, 24, 67
cognitive apects, 12, 51, 89, 90, 97, 118-
119, 128, 132
collocation, 203, 206-209
commission(er), 53, 146-147, 149
communication, 18, 21, 28, 32, 33, 36, 51
77,158-159,219
competence, 54, 58, 64, 68, 90, 92, 98,
100, 136
comprehension processes, 121, 125, 128-
129, 131, 135, 137, 141
concordance, 204
cooperation, 49, 53, 56
corpus, 202-204, 208-209, 215, 234
court interpreting, 19
creativity, 96, 118, 120, 124, 144, 282,
284-285, 287, 289, 302
(translation) criticism 364-370, 375
332-333, 336, 340, 342, 345, 348, 355
D
Danish, 18, 20, 21, 24, 74, 110, 114, 248
discours rapporté, 188-190, 192-195
discourse analysis, 178
dictionary, 42, 113, 202, 212-213
domestication, 66, 159, 249, 252, 254,
310, 315, 319, 344-347, 358, 375-379,
documentary film, 235, 237, 240
drafting, 40, 44, 47-48, 283, 288-291
drama translation, 153-154, 340, 345-348
dubbing, 236, 241, 247
Dutch, 20, 21, 148
E
English, 18, 21-22, 24, 52, 63, 65, 68, 76-
77, 81, 110, 114, 122-125, 156, 180-181,
183, 193-195, 205, 207-208, 213, 215,
223, 226, 265, 274, 285-289, 302, 308,
322-324
equivalence, 206, 360
errors, 20, 23, 40, 43, 93, 222, 229
Esperanto, 76, 79
expert/expertise, 56, 58, 89, 93, 95, 98,
101
F
faithfullness, 165-166, 340-341
fidelity, 23, 120, 165, 323
Finnish, 52, 57, 155-158, 180-181, 183-
184,376,383
Flemish, 147-148
fluency, 65-66, 249, 253, 374
foreignizing, 66, 155, 159, 318, 361,
374-375, 380, 382, 384-386
French, 20, 68, 75, 121, 193-196, 213,
237-240, 286-289
392 Translation in context

function, 319-320, 325, 328, 370 M

fuzzy number, 28-30, 36
German, 20-22, 24, 66, 122, 125, 146,
213, 215, 217, 223-228, 237-240, 254-
255, 264-266
grammar, 222, 229
Greek, 21-22, 67-68, 81, 169
H names, 251,378, 384
Hebrew, 21, 67-68
human, 256-257, 267
hybrid language/text, 308, 354, 359
I 335
ideology, 178-179
images, 144, 149, 263, 267, 276
imagology, 144-145
indirect translation, 18-19, 23-24
initiator, 53, 55
interlingual translation, 5, 10, 18-19
interpreters, 128-130, 133-134, 138, 140-
141
(conference) interpreting, 18, 20, 129
intersemiotic translation, 5, 9
intralingual translation, 5, 10
intrasemiotic translation, 5, 8
invisibility (see fluency)
Italian, 20, 213,278, 340
machine translation, 212, 222
meaning, 31, 167, 240, 272, 288
methodology, 5-7, 91, 106, 114, 128-129,
204, 223, 263, 328-329
minority language, 74, 82-83
(non)mother tongue, 62-63, 67-68
multilingualism (see register)
N
native speaker, 63-69
narrative, 178, 189,333-335
Neokoiné, 76, 79
newspaper, 109-110, 122, 182, 328, 333-
nominalisation, 182, 184, 264
non-verbal, 273, 278-279
norms, 65, 145, 147, 160, 166, 175, 294,
336
Norwegian, 7, 13-14, 322-324
novel, 246, 309, 311, 320
O
oral literature, 354, 356
orality, 237 (see also paralinguistic
features)
original, 18, 21, 23, 25, 65-66, 165, 245,
340
oversettelse, 4, 13

K
knowledge, 89-90, 93-94, 96-97, 99, 101, paralinguistic features, 276 (see also
229, 263, 273 orality)
passivisation, 180, 184
pidgin, 76, 297
Latin, 21-22, 67, 81 plays, 340, 342, 376
legal translation, 282-284, 289-291 Portuguese/Brazilian Portuguese, 213,
lexicography, 203, 206 239, 246, 264-266
liaison interpreting, 19 (post-)colonialism, 354, 356-357, 360
literal translation, 124, 159-160, 282, 294, premodified constructions, 183
309, 318, 320, 332, 334, 336, 365 problem solving, 43, 57, 110-112
literary translation, 22-23, 208, 246, 248, prototype, 4, 7, 9, 12, 120, 122-123
319,336 pseudotranslation, 36, 247
loans, 329, 332-333, 359 punctuation, 251
localisation, 219
 Index 393

Q target language/text, 64-65, 67, 112, 115,

quality, 41-43, 45, 128, 155, 159, 213-214 128, 155, 374, 376, 387
task(s), 40, 89, 97, 128, 131
R team translation, 68, 244, 269
reading, 43, 131, 137, 140-141, 145, 184, terminology, 57, 75-77, 81, 133, 213
252,381 terminology policy 75-76, 80, 83
reception/receiver, 18-20, 23-24, 28, think aloud protocol, 41, 46, 90, 106, 110
30-31, 33-35, 53-54, 145, 148, 154, 159 third code, 360
197, 241, 251, 253, 272-273, 280, 328, time log, 20, 24
343, 345, 348 time pressure, 108-110, 114, 219, 245
(language) register, 156, 158, 189, 196 training, 42-43, 45-46, 93-94, 98, 100,
296, 302 102, 129, 134, 141, 209
reification, 179 translation (concept of-), 4, 13, 18, 22, 42,
relay, 19-24 164, 165
relexification, 358-359 (category of-) 4, 7, 11-12
reported speech, (see discours rapporté) (typology of-) 5-6
Romance languages, 75, 80 translation pairs, 5-6, 8, 10, 208
Russian, 74 translation policy, 364-365, 367-369, 376
translation procedure (see also strategy)
S 18, 40, 42, 45, 47, 95, 101, 238, 318,
science fiction, 247 329-331
script, 236, 345 translation process, 88-89, 97, 106, 110,
semantic shift (see meaning) 114, 121-122, 125, 129, 139, 145, 196,
sender, 18, 20, 23-24, 28, 30-31, 33-35, 212,249,269,283
54 Translation Studies, 44, 46, 144-145, 178,
short stories, 248 308, 370
simultaneous interpreting, 20, 23, 128 translator, 24, 28, 29, 34-35, 39, 42, 44,
skills, 42, 46, 95, 101, 129, 132, 134, 137, 46, 49, 53-54, 59, 62-63, 65-66, 69, 88,
140 92, 98, 107, 110, 114, 145, 184, 212, 214,
skopos, 55, 124, 148, 275, 279, 319, 328 269,314,367
Slovene, 22 Translog program, 106-109, 111
Spanglish, 76 transparency, 66, 165, 167
Spanish, 76-77, 206, 213, 216, 274-275, Turkish, 146, 148, 364, 370
312-314,322-324,329-331 typography, 58, 277
standardization, 46, 75, 78, 79, 188, 191,
196, 244 V
stereotype, 262, 269, 278, 340, 343, 346 voice over, 235-236, 241
strategy, 24, 51, 98, 100, 110, 113-114,
122, 141, 153, 158-159, 171, 174, 179, W
196, 246, 249-252, 254, 273-275 word, 204, 206, 209, 215-216, 224, 239,
298-301, 312, 343, 360, 374, 387 264, 267
subtitling, 18,234,236,241 work procedures (see translation
support translation 23-24 procedure)
working memory, 128, 137-138, 140
T
tales, 22. 178