Evaluating The Evaluator

Evaluating the Evaluator
This book offers a theoretical framework for assessing translation q uality

grounded in supportive argumentation. The volume outlines a system-
atic framework for translators and translation critics to substantiate their
decisions and judgements on a translation’s quality and, in the case of
negative criticism, put forward a more effective translation solution. The
book traces the decision-making process underpinning translation prac-
tice, considering the different factors surrounding a particular translation
to inform the most appropriate translation strategy, such as the temporal
and geographical relationship between source and target texts, special
provisions required by clients, time frame, qualifications, and sociocul-
tural and political issues. The framework posits that such factors should
underpin any arguments used by the translator in adopting a given strat-
egy and, in turn, that any criticism of a translation’s quality must be in
line with the same argumentative structure. Applied to a corpus of trans-
lation examiners’ reports of translation, the book demonstrates how this
framework can act as a tool to be scaled to fit the needs of the different
actors of a translation – translators, critics, and scholars. This book will
be of interest to scholars in translation studies and practising translators.
Hansjörg Bittner served as a lecturer at Eastern Mediterranean University

(English literature) and at the University of Hildesheim (translation
studies). His publications cover poetics, translation theory, and audiovi-
sual translation. A practising translator, he is currently a lecturer at Hof
University of Applied Sciences, teaching business English and technical
English.
Routledge Advances in Translation and Interpreting Studies
38 Complexity Thinking in Translation Studies

Methodological Considerations
Edited by Kobus Marais and Reine Meylaerts
39 Translating and Interpreting in Korean Contexts

Engaging with Asian and Western Others
Edited Ji-Hae Kang and Judy Wakabayashi
40 Hybrid Englishes and the Challenges of/for Translation

Identity, Mobility and Language Change
Edited by Karen Bennett and Rita Queiroz de Barros
41 Translating the Visual

A Multimodal Perspective
Rachel Weissbrod and Ayelet Kohn
42 Using Computers in the Translation of Literary Style

Challenges and Opportunities
Roy Youdale
43 Concrete Poetry
Translation and Transmission
Edited by John Corbett and Ting Huang
44 Humour in Audiovisual Translation

Theories and Applications
Margherita Dore
45 Evaluating the Evaluator

A Novel Perspective on Translation Quality Assessment
Hansjörg Bittner
For a full list of titles in this series, visit https://www.routledge.com/

Routledge-Advances-in-Translation-and-Interpreting-Studies/book-series/
RTS
Evaluating the Evaluator
A Novel Perspective on Translation
Quality Assessment
Hansjörg Bittner
First published 2020
by Routledge
52 Vanderbilt Avenue, New York, NY 10017
and by Routledge
2 Park Square, Milton Park, Abingdon, Oxon, OX14 4RN
Routledge is an imprint of the Taylor & Francis Group, an
informa business
© 2020 Taylor & Francis
The right of Hansjörg Bittner to be identified as author of this
work has been asserted by him in accordance with sections
77 and 78 of the Copyright, Designs and Patents Act 1988.
All rights reserved. No part of this book may be reprinted
or reproduced or utilised in any form or by any electronic,
mechanical, or other means, now known or hereafter invented,
including photocopying and recording, or in any information
storage or retrieval system, without permission in writing from
the publishers.
Trademark notice: Product or corporate names may be
trademarks or registered trademarks, and are used only for
identification and explanation without intent to infringe.
Library of Congress Cataloging-in-Publication Data
Names: Bittner, Hansjörg, 1963– author.
Title: Evaluating the evaluator : a novel perspective on
translation quality assessment / Hansjörg Bittner.
Description: 1. | New York : Routledge, 2020. |
Series: Routledge advances on translation and interpreting
studies ; 44 | Includes bibliographical references and index.
Identifiers: LCCN 2019043177 | ISBN 9780367417130
(hardback) | ISBN 9780367815882 (ebook)
Subjects: LCSH: Translating and interpreting—Evaluation.
Classification: LCC P306.2 .B58 2020 | DDC 418/.02—dc23
LC record available at https://lccn.loc.gov/2019043177
ISBN: 978-0-367-41713-0 (hbk)
ISBN: 978-0-367-81588-2 (ebk)
Typeset in Sabon
by codeMantra
Contents
List of Figures ix
List of Tables xi
Acknowledgements xiii
1 Introduction 1
2 The Quality of Translation: Different Approaches 6

Juliane House 7
Malcolm Williams 13
Heidrun Gerzymisch-Arbogast and Klaus Mudersbach 17
Ernst-August Gutt 21
Other Approaches to Translation Quality 26
Approaches to Translation Quality in the Twentieth
Century 27
Approaches to Translation Quality in the
Twenty-First Century 32
3 Preliminary Assumptions 50
Defining Translation Quality and Translation
Quality Assessment 50
Some Thoughts on Measuring Quality 51
Some Thoughts on Achieving Good Quality 54
Revisiting Overt and Covert Translation 56
Refining the Concept of Overt-Covert Translation:
Overview 59
Detailed Discussion 61
The Problem of Subjective Evaluation 67
The Problem of Subjectivity from a Philosophical
Point of View 67
vi Contents
The Process of Translation Quality Assessment 70
The Problem of Subjectivity from a Translation
Studies Point of View 72
How to Curb the Subjective in Translation
Towards Evaluating the Evaluator 76
Some Views on How to Evaluate Translations 76
The Evaluation of Commented Translations 82
4 Quality Factors of Translation 96

The “Translator’s Daffodil” 96
Underlying Assumptions 99
Factor Categories 100
Source Text 103
Language Pairs in Translation and Source
Text Analysis (Nord) 103
Sender and Sender’s Intention 105
Audience and Medium 106
Place and Time of Communication 107
Motive for Communication and Text Function 109
Source Text Defects 110
Summary of Source Text Factors 110
Text Form 112
Intratextual Factors 113
Subject Matter and Content 114
Text Composition and Non-Verbal Elements 116
Lexis, Syntax, and Suprasegmental Features 116
Selected Text Types 118
Poetry, Drama, and Comics 118
Audiovisual Translation 121
Summary of Text Form Factors 126
Client 127
Client Roles 128
Deadlines 129
Glossaries 130
Stipulations and Specifications 131
Motivation 132
Summary of Client Factors 133
Translator 133
Overview and Translation Tools 134
Translation Competence 135
Qualification and Motivation 138
Summary of Translator Factors 138
Contents vii
Culture 139
The Relativity of Culture 140
Cultural Norms 141
Translation in Different Cultures at Different
Times in History 143
Borderline Cases: Unlikely Successes and
EU Translations 146
Summary of Culture Factors 148
Politics 149
Power Relations 150
Censorship 152
More Power Issues 154
Politics and the Evaluation of Translation Quality 155
Summary of Politics Factors 157
5 The Principle of Argumentation 170

The Need for an Argumentative Translation
The Need for Translation Theory 173
The Need for Argumentation in Translation 175
Summary – The Need for an Argumentative
Translation Quality Assessment 177
Translation Decisions 177
Decision-Making and Translation 179
Defining the Translation Strategy 181
Summary of the Decision-Making Process in Translation 186
The Argumentation Process 187
A Theory of Dialectical Structures 188
Dialectical Structures in Translation Quality Assessment 195
Examples of Argumentation 204
Criticising the Translation of a Young Adult Novel 205
Criticising the Translation of a Survey 211
6 Evaluating the Evaluator 224

The Corpus: Examiners’ Reports of
Commented Translations 224
Methodology 225
Looking at Formal Characteristics 226
Looking at Aspects of Translation Quality Assessment 228
Source and Target 228
Errors and Achievements 230
Evidence and Arguments 232
Summary of Methodological Issues 240
viii Contents
Results 240
Formal Characteristics 244
Aspects of Translation Quality Assessment 246
Source and Target 246
Errors and Achievements 247
Evidence and Arguments 250
Summary of Results 267
7 Conclusion 273
Index 277
Figures
3.1 Overt–covert overview (general aspects) 60

3.2 Overt–covert overview (specific aspects) 66
3.3 Objectivity and subjectivity in TQA 71
3.4 Evaluation focus for a commented translation 82
3.5 Commented translation triangle 83
3.6 Evaluating the quality of a commented translation
written as a BA thesis 87
4.1 The “translator’s daffodil” 98
4.2 Factor matrix 102
4.3 Overview of translator-related factors 134
4.4 Quality factors → arguments → translation decisions 158
5.1 Argument structure – critique of an excerpt from Ein
Schatten wie ein Leopard 211
5.2 Arguments – first solution: “Sind Sie geschieden?” 214
5.3 Arguments – second solution: “Haben Sie sich schon
einmal scheiden lassen?” 215
5.4 Arguments – third solution: “Sind Sie jemals geschieden
worden?” 215
5.5 Arguments – fourth solution: “Sind Sie schon einmal
geschieden worden?” 215
6.1 Correlation between grades of theses and
lengths of reports 245
Tables
4.1 Summary of source text factors 111

4.2 Technical issues relevant to subtitling 123
4.3 Summary of client factors 133
4.4 Summary of culture factors 148
5.1 Fundamental concepts and ideas of translation quality
assessment 178
5.2 Decision process terminology 186
5.3 The decision-making process in translation 187
5.4 Matrix of discursive goals 193
5.5 List of arguments – critique of an excerpt from Ein
Schatten wie ein Leopard 210
5.6 Translation strategy arguments – survey question
“Have you ever been divorced?” 213
6.1 Comparison of source and target – criteria 229
6.2 Summary of questions and response categories 241
6.3 Comparison of grades and error
counts for translations 249
6.4 Frequency of parameters for questions 10a and 10b 253
6.5 Criticism of satisfactory solution supported
by irrelevant argument 259
6.6 Criticism of satisfactory solution with
acceptable alternative 260
6.7 Criticism of satisfactory solution with a
wrong alternative 261
6.8 Arguments derived from error descriptions 265
6.9 Summary of results with regard to
guiding questions 268
7.1 Uses of the argument-based TQA theory 275
Acknowledgements
I thank those who have supported me during this project. First of

all, there is Professor Klaus Schubert, who commented on individual
chapters and gave useful advice. Then there are former colleagues
and, particularly, students from the translation studies department of
the University of Hildesheim, who provided a thought-provoking foil
against which I would test my assumptions about translation quality
assessment. A German version of the section on “The problem of subjec-
tive evaluation” was published in Studia Translatorica 5. I am grateful to
the editor of this journal, Anna Małgorzewicz, for giving me permission
to include the English version of that article in the present monograph. I
should also like to say “Thank you” to those examiners (of commented
translations written to fulfil the requirements for a BA degree at the
University of Hildesheim) who consented to the anonymous use of their
reports in the chapter discussing the evaluation of the evaluator. Finally,
I thank my family, who had to put up with me writing rather than going
with them on holiday.
1 Introduction
In The Translator’s Turn, Douglas Robinson writes:
It seems undeniable that translation is largely an intuitive process.

Good translators choose words and phrases by reference not to some
abstract system of intellectualized rules, which most of us have never
internalized in the first place, but rather to “messages” or impulses
sent by the body: a given word or phrase feels right. Intuitively, not
just for the translator but for all language users, sense is not cogni-
tion but sensation.
(Robinson 1991:xii)
We beg to differ. When translators choose words and phrases on the basis
of what feels right, they follow some translation strategy – a strategy
they may have acquired not from any intellectualised rules but, unwit-
tingly, from experience and assumptions. For, otherwise, the translator
would not be able to distinguish a good translation solution from a bad
one. Using a translation strategy without being aware of it, the translator
can state whether a rendering is satisfactory or not, but he or she cannot
say why. The theory presented here is designed to help translators make
translation decisions consciously, enabling them to provide arguments
for and against different potential solutions.
The idea is to achieve an unassailable translation – a translation that
harmoniously envelops the source text, the target text, the translation
brief, and the target audience, and that is sustained by a web of interre-
lated arguments which will carry it (the translation) through the fiercest
storms of criticism. Critics, evaluators, or revisers of translations need to
fall back on some prescriptive concept of translation quality and trans-
lation quality assessment (TQA), if they wish to discuss the quality of a
given target text as objectively as possible. That concept will stand pitted
against the translator’s concept of how to translate. And which of the
two concepts is more convincing depends on their respective argumenta-
tive foundations. Again, our theory is intended to furnish the framework
for such an argumentative foundation.
2 Introduction
Given a theory that can handle translation quality in all its facets and
that takes into account the different conditions in which the translation
has been produced, there is no aspect to do with translation that might
not be analysed systematically:
• translators can analyse potential translation solutions and decide

which to take,
• evaluators can analyse the translators’ achievements and judge them,
• translation scholars can analyse the evaluators’ findings to see if
they are appropriate.
Thus, we propose a theoretical tool which benefits those actors in a

translational environment that are involved in, or have an interest in,
translation quality.
So, why is translation quality important? It is important, because it is
key to the success or failure of cross-language communication. In a busi-
ness environment – where translation is “not an end but a means to an
end” (Al-Abbas 2009:229) – we can say with Peña Pollastri that “[trans-
lation quality] is relevant in the market of translation since it is one of
the features that defines the acceptability of certain translation products
and the rejection of others” (Pollastri 2009:239). The problem is how to
distinguish the acceptable translation products from the unacceptable
ones and to determine what makes for a successful target text. In this
context, Fritz Paepke remarks that the success of translation depends
on the oscillating mix ratio of concrete information and the linguistic
perception of such concrete information (cf. Paepke 1994:113). In other
words, what counts with regard to translation quality are both content
and the way in which content is conveyed. How these two variables can
properly be assessed has been investigated in many a TQA theory. We
propose another such theory in this book. Here is a brief outline of what
will be discussed in the following chapters.
Chapter 2 deals with different approaches to translation quality and
TQA. While four views on the subject are analysed in detail – those of
Juliane House, Malcolm Williams, Heidrun Gerzymisch-Arbogast and
Klaus Mudersbach, as well as Ernst-August Gutt – some 60 p apers that
discuss translation quality from various angles are looked at in passing,
as it were. They include theoretical accounts with and without exam-
ples as well as corpus analyses and other empirical studies. The discus-
sion of these approaches to translation quality and translation quality
assessment prepares the ground for our own TQA theory in that it
helps to garner interesting ideas and build the foundation on which an
argument-based model can thrive.
Chapter 3, then, presents a few preliminary assumptions. It is di-
vided into four sections. The first section defines translation quality
and translation quality assessment. It covers different notions of quality
and how quality can be measured, before considering the question of
Introduction 3
how good quality can be achieved in translation. The second section
relates to Juliane House’s concept of overt and covert translation. Here,
we argue that the principle underlying the overt–covert distinction is
theoretically more intriguing than is evident from House’s explanation
of the concept. In the third section,1 we tackle the problem of subjective
evaluation. A fter looking at subjectivity and objectivity from a philo-
sophical angle, we analyse the options afforded by translation quality
assessment to establish the extent to which the individual steps of the
TQA process can be objective. This is followed by a discussion of other
scholars’ attempts to render translation quality assessment as objective
as possible. Finally, we offer our own suggestions of how the subjective
in TQA can be curbed. The last section paves the way for the evaluation
analysis in Chapter 6. Here, we first present different views on how to
evaluate translations, before coming up with a framework that can be
used to analyse the evaluations of translated texts.
In Chapter 4, we develop what might be regarded as the holistic back-
ground to our TQA theory. Assuming a wide-angle perspective, we pick
up and pin down the different factors which, in some way or other, may
affect the quality of the translated text. The factors are grouped around
a centre that represents the actual translation. We refer to the resulting
flowery shape as the “translator’s daffodil”. The individual factor groups
are, then, discussed in detail. Christiane Nord’s extratextual factors as
well as defects in the source text feature prominently in the source text
group. The group referred to as “text form” deals with Nord’s intratex-
tual factors and selected text types, that is, poetry, drama, and comics
as well as audiovisual translation. Any factors relating to the client are
analysed under “client roles”, “deadlines”, “glossaries”, “specifications
and stipulations”, and “motivation”. In the translator group of factors,
we consider those quality aspects that are directly related to the trans-
lator as a human being or to the translator’s task of producing the ac-
tual translation. They include the translator’s qualification, competence,
and motivation as well as any translation tools. The next section on
culture-related factors is about cultural norms and translation in differ-
ent cultures at different times in history. It also tackles borderline cases
of translation. Under the heading of “politics”, we are concerned with
translation in political environments dominated by censorship and/or
specific power structures and how these aspects have an effect on the
quality of the translation product. All factors discussed in this chapter
are distinguished as to whether or not they are relevant to translation
quality and translation quality assessment, and whether or not they are
known to the evaluator of the target text.
Following this account of the overall translation setting with its
various factors relating to translation quality and translation quality
assessment, we are finally in a position to develop our own argumenta-
tive TQA theory. In Chapter 5, we demonstrate how translation quality
can be established on the basis of arguments that draw upon any relevant
4 Introduction
f actors referred to in Chapter 4. The key assumption is that the quality of
a translation solution is only as good as the arguments supporting it. The
first section shows that other translation scholars, too, have emphasised
the need for argumentation in translation. In the second section, then, we
develop the framework within which argumentation can be used to de-
termine the quality of a translated text. This framework takes as a point
of departure Klaus Schubert’s concept of the decision-making process
and applies it to translating. As a result, we get a translation strategy on
the basis of which the translator can then render a given source text into
a target text. This is when, in the third section, the argumentation pro-
cess comes into play. It draws on the most important elements of G regor
Betz’s theory of dialectical structures, providing a tool with which the
relative quality of individual translation solutions can be established.
How this works is demonstrated with two examples: the first deals with
a critic’s criticism of a young adult novel translated from English into
German, and the second reveals how the translator should make a con-
scious decision between different potential translation solutions.
In Chapter 6, our TQA theory is applied to a corpus, namely, the
examiners’ reports of commented translations written as part of the bach-
elor’s degree course “International Communication and Translation” at
the University of Hildesheim. While looking at all aspects of the reports,
we focus particularly on argumentation as a key requirement of TQA: to
what extent do examiners support their positive and negative criticisms
with appropriate arguments, and how do they do it? The second sec-
tion specifies the parameters used to analyse the reports, and the third
section presents the results of the analysis. The overall approach is both
descriptive and prescriptive: on the one hand, we state what we find in the
reports; on the other hand, we relate these findings to certain desiderata
extracted from various requirements put forward by different translation
scholars. The results of the analysis reflect the actual situation of TQA
in an academic environment and could provide a framework of criteria
according to which the quality of a commented translation should be
assessed. Finally, Chapter 7 wraps up the ideas put forward in this book.
By way of concluding our introduction, we will briefly discuss the termi-
nology associated with translation quality assessment. In general, the con-
text defines the meanings of words. Thus, “translation” – apart from being
determined by what is accepted as translation – may sometimes refer to the
process of translating and sometimes denote the result of that process, that
is, the actual translation product. Where the context is not sufficient to
indicate the intended meaning, an unequivocal expression is preferred, for
example, “translation process” or “translated text”. As regards the verbs
“assess” and “evaluate”, and the corresponding nouns “assessment” and
“evaluation”, they are used interchangeably without suggesting variation
in meaning. However, unlike “evaluator”, the term “assessor” tends to
be typically employed in an accounting or legal context – which is why
Introduction 5
we refrain from using the latter expression.2 Depending on the context,
we occasionally refer to “revisers” when talking about a professional
translation environment and to “examiners” when dealing with academic
translation quality assessment. Definitions are provided when they are
deemed necessary – such as the definition of what constitutes a relevant
factor (cf. Chapter 4). As for acronyms, these are generally avoided, with
the exception of TQA for “translation quality assessment”. After these pre-
liminaries, we can now begin to develop a novel perspective on translation
quality assessment in order to be able to evaluate the evaluator.
Notes
1 The content of this section is also available in German (cf. Bittner 2014). Yet,
despite its earlier publication, the German version has been prepared on the
basis of the English original.
2 This is not to say that “assessor” might not also be used in a TQA context,
as has been done, for instance, in Schäffner (1998:4) and Kim (2009:133).
References
Al-Abbas, Suleiman (2009). “The Standard of Translation Quality in the
Jordanian Universities Programmes and the Prospective Role of Atlas
Global Center for Studies and Research”. In: Forstner, Martin et al. (eds.)
(2009:229–238).
Angelelli, Claudia V.; Holly E. Jacobson (eds.) (2009). Testing and Assessment
in Translation and Interpreting Studies: A call for dialogue between research
and practice. Amsterdam/Philadelphia: John Benjamins (American Transla-
tors Association Scholarly Monograph Series v. XIV).
Bittner, Hansjörg (2014). “Das Problem der Subjektivität bei der Beurteilung
von Übersetzungen”. In: Studia Translatorica, vol. 5, 21–35.
Forstner, Martin; Hannelore Lee-Jahnke; Peter A. Schmitt (eds.) (2009).
CIUTI-Forum 2008: Enhancing Translation Quality: Ways, Means, Meth-
ods. Bern: Peter Lang.
Kim, Mira (2009). “Meaning-Oriented Assessment of Translations: SFL and
Its Application to Formative Assessment”. In: Angelelli, Claudia V.; Holly E.
Jacobson (eds.) (2009:123–155).
Paepke, Fritz (1994). “Textverstehen – Textübersetzen – Übersetzungskritik”.
In: Snell-Hornby, Mary (ed.) (1994:106–132).
Peña Pollastri, Ana Paulina (2009). “Evaluation Criteria for the Improvement
of Translation Quality”. In: Forstner, Martin et al. (eds.) (2009:239–260).
Robinson, Douglas (1991). The Translator’s Turn. Baltimore and London: John
Hopkins University Press.
Schäffner, Christina (1998). “From ‘Good’ to ‘Functionally Appropriate’: As-
sessing Translation Quality”. In: Schäffner, Christina (ed.) (1998:1–5).
Schäffner, Christina (ed.) (1998). Translation and Quality. Clevedon: Multilin-
gual Matters.
Snell-Hornby, Mary (ed.) (1994). Übersetzungswissenschaft – Eine Neuori-
entierung. 2nd ed. Tübingen and Basel: Franke.
2 The Quality of Translation
Different Approaches
The quality of translation has been tackled from many different perspec-
tives. These perspectives are as varied as the concepts of “ translation”
and “quality” are manifold: scholars have focussed on translation as a
process or as the result of that process; and they have discussed transla-
tion quality in terms of acceptability, adequacy, or optimisation. Accept-
ability relates to the question of who has to use the translation and who
is in a position to judge the quality of the translation, whereas adequacy
has to do with the degree to which the target text and context can be said
to be equivalent to the source text and context. Optimisation u sually con-
cerns the translation process. Furthermore, the q uality of translation can
be approached from a purely theoretical point of view or on the basis of
an empirical study. It is patent that the many ways in which the above el-
ements can be combined yield a large number of different perspectives on
the quality of translation. This chapter will provide an overview of such
perspectives: four approaches – the works of Juliane House, Malcolm
Williams, Heidrun Gerzymisch-Arbogast and Klaus Mudersbach, as well
as of Ernst-August Gutt – will be analysed in detail, many others will
be discussed in a more cursory fashion. The purpose is to identify and
describe the key ideas and to give the o ccasional hint as to how the re-
spective approaches could benefit from suitable a rguments – the central
aspect of the theory presented in this book.
In trying to account for the many different ways in which translation
quality becomes the object of academic investigation, we propose a
system that looks at two distinguishing features: research focus and
methodology. This system covers not only the assessment side of transla-
tion quality but also its more general aspects. Most approaches deal with
translation quality either on the basis of the translated text or on the
basis of the translation process; occasionally, though, the focus may be
on quality issues that are only remotely related to translation as a product
or process. In some cases, both product and process may be affected by
the examination of a very specific issue (as, for example, in Martín de
León 2007 or in Gerisch/Bastian 2007); in other cases, the role of trans-
lation as product or process is rather marginalised (as in Muñoz Martín /
Conde Ruano 2007, Forstner 2007, or Fox 2009). As for methodology,
Approaches to Translation Quality 7
we juxtapose theoretical and empirical approaches. They can be defined
as follows: an approach is considered empirical if it involves either test
subjects (as in surveys and translation experiments) or a text corpus; an
approach is considered theoretical if it involves neither test subjects nor
a text corpus. Note that this definition does not cater for “empirical” in
its etymological meaning, that is, relating to knowledge derived from
experience (cf. Williams 2004:130, 2009:13).
Our review of selected contributions to the discussion of transla-
tion quality will begin with House (1997), Williams (2004 and 2009),
Gerzymisch-A rbogast (1994) and Gerzymisch-Arbogast/Mudersbach
(1998), as well as Gutt (1991). All of these approaches focus on trans-
lation as a product and would be considered theoretical in their over-
all methodology (even if some of Juliane House’s findings are based on
empirical studies). They have been chosen for their comprehensive dis-
cussion of the subject at hand. In spite of their very different perspec-
tives, all of them are relevant to translation quality assessment – either
directly or indirectly. Our critical analysis of the above-mentioned books
will shed some light on the advantages and disadvantages of the views
put forward by the respective authors. The four approaches can serve as
foils that occasionally reveal our own position with regard to transla-
tion quality and translation quality assessment, thereby anticipating in a
rather sketchy way the concept presented in Chapter 5.
Juliane House
In 1997, when her book Translation Quality Assessment – A Model
Revisited appeared, Juliane House had already been involved with
translation quality for more than 20 years: her PhD thesis had been sub-
mitted to the University of Toronto, Canada, in 1976 and was published
in 1977, with a second edition following in 1981 (cf. House 1997:VII).
Meta Translator’s Journal printed a concise version of her theory in
1977; other publications that deal with the same subject are House
(2001, 2009, 2014, 2015). The following analysis of House’s approach
to translation quality assessment takes as a starting point her book of
1997, a book that “[retains] the essential features of the original model”
(House 1997:VII) while at the same time taking account of “review com-
ments […] and new views, theories and developments inside translation
theory” (ibid.). Our discussion of House’s theory of translation quality
assessment will be complemented by improvements suggested in House’s
most recent publications on the subject (cf. House 2014, 2015).
Equivalence is at the bottom of House’s theory. The need for an
equivalence relation between the source text and the target text arises from
the “double-binding nature” (House 1997:24) of translation, bound as it
is to its source and target cultures. It is through the concept of e quivalence
that a translation fulfils the “demands of invariance” (House 1997:25),
8 Approaches to Translation Quality
capturing the shifting ground of the tertium comparationis. Which element
is to remain invariant in translation and thus establish the equivalence
relation between the source and target texts “must be decided in each
and every individual case by the goal, the purpose of the translation”
(ibid.). House’s defence of equivalence against criticism – be it justified or
unjustified – underlines the crucial role this concept plays in her TQA
approach (cf. House 1997:26).
Closely related to equivalence is another concept favoured by House:
the distinction between overt and covert translation. While
in overt translation the function of the translation is to enable its

readers access to the function of the original in its linguacultural
setting through another language, […] the function of a covert
translation is to imitate the original’s function in a different dis-
course frame.
(House 1997:29)
The implications of this distinction will be discussed in more detail be-

low. For the time being, it is important to see that for both overt and
covert translation to be able to accommodate the concept of equivalence
presupposes not only a flexible notion of equivalence but also a wide
definition of translation. Juliane House is prepared to embrace equiva-
lence in all its varied manifestations; she does not, though, regard the
concept of translation as equally variable.
Indeed, for House, a translation will only be accepted as a translation
if the translator does not apply a “cultural filter” where this is unwar-
ranted. The random application of such a filter would run counter to the
translation principle of source text loyalty. Juliane House explains:
The concept of a “cultural filter” is a means of capturing socio-

cultural differences in shared conventions of behavior and commu-
nication, preferred rhetorical styles and expectation norms in the
two speech communities. These differences should not be left to in-
dividual intuition but should be based on empirical cross-cultural
research. Given the goal of achieving functional equivalence in a
covert translation, assumptions of cultural difference should be
carefully examined before interventions in the original’s meaning
structure is [sic] undertaken. The unmarked assumption is one of
cultural compatibility, unless there is evidence to the contrary.
(House 2001:251)
The cultural filter is used in covert translation to elicit from the target text
reader a response similar to that of the source text reader. A target text
that deviates from the source text in any of the pragmatically relevant
dimensions without this being motivated by target culture preferences
is regarded as a covert version (cf. House 1997:73). Similarly, an overt
version is distinguished from an overt translation by the fact that, here,
“a special function is overtly added to a [translation text]” (House
1997:73) as, for example, in adaptations of literary works for children.
The distinction between a translation and a version allows House to
posit relatively narrow equivalence boundaries for a translation, with
any violation of these boundaries being regarded as d etrimental to the
quality of the translation.
The basis of House’s judgements on the quality of a translation is fur-
nished by a set of eight dimensions or categories slightly adapted from
Crystal and Davy, which, in her revised model, are subsumed “under the
simplifying Hallidayan ‘trinity’ Field, Tenor, Mode” (House 1997:107).
It is with respect to these categories that equivalence has to take e ffect.
Any failure to establish equivalence is counted as an error and reduces
(or, at least, changes) the quality of the translation. House, in some-
what awkward terminology, distinguishes between “overtly erroneous
errors” and “covertly erroneous errors” (House 1997:45), later also
called “overt errors” (for example, House 1997:130) and “covert errors”
(House 1997:74), respectively. While the latter refer to errors occur-
ring as the result of a functional or dimensional mismatch, the former
concern obvious breaches of the target language system and mistakes
resulting from “a mismatch of the denotative meanings of source and
translation text elements” (House 1997:45). To evaluate the quality of
a translation, both overtly and covertly erroneous errors (or, for that
matter, overt and covert errors) are listed, providing the basis of the final
statement of quality.
In detail, Juliane House’s translation quality assessment comprises
(1) an analysis of the original, (2) a statement of function of the original
text, (3) a comparison of the original and the translation, and (4) a state-
ment of quality. First, the original is analysed along the register catego-
ries of FIELD, TENOR, and MODE: FIELD “refers to the nature of the
social action that is taking place, it captures ‘what is going on’, i.e., the
field of activity, the topic, the content of the text or its subject matter”
(House 1997:108); TENOR “refers to who is taking part, to the nature
of the participants, the addresser and the addressees, and the relation-
ship between them in terms of social power and social distance” (House
1997:108–109); MODE “refers to both the channel – spoken or writ-
ten […] and the degree to which potential or real participation is allowed
for between the interlocutors” (House 1997:109). This is complemented
by a close look at GENRE, “a socially established category character-
ized in terms of occurrence of use, source and a communicative purpose
or any combination of these” (House 1997:107). Second, in stating the
function of the original text, the findings of the analysis of the above four
categories are summarised and presented in a coherent argument. Third,
the original and the translation are compared along the lines of the four
categories of FIELD, TENOR, MODE, and GENRE, harking back
to the results of the analysis and trying to make out any mismatches.
Fourth, just as the statement of function summarises the analysis of the
original, so does the statement of quality summarise the mismatches
found in the comparison. The statement of quality brings in the notion
of overt and covert translation and, on this basis, provides a critique of
the translation. Here, House draws her (generally convincing) conclu-
sions from the mismatches discovered, furnishing the occasional broad
explanation as to why the translator has produced them.
Consider, for example, the discussion of Jill Murphy’s Five Minutes’
Peace, a children’s book with elephant characters that has been trans-
lated from English into German (cf. House 1997:122–131). In her
analysis of the original, House is guided by the distinction between
FIELD, TENOR, MODE, and GENRE, listing under these headings
any relevant textual manifestations and pertinent descriptions of linguis-
tic effects. However, this procedure has the disadvantage of tending to be
a little confusing as some of the aspects discussed under one dimension
may also be analysed – and, indeed, are analysed – under another. The
element of humour, for example, is mentioned three times: under FIELD,
under TENOR, and under GENRE. Similarly, the allocation of a partic-
ular feature to one of the linguistic categories (such as lexical, syntactic,
or textual) is sometimes arbitrary – an issue raised by Brotherton with
regard to the original model of 1977 (cf. Brotherton 1981, referred to
in House 1997:102). Juliane House admits that Brotherton is right; yet,
she defends her approach, saying that “this is due to the nature of lan-
guage and does not point to a basic shortcoming of the model” (House
1997:102). True enough: it is not the model itself which is at fault, here,
but its application. Unwanted repetitions and variable allocations can be
avoided if the evaluator takes as a point of departure the text itself and its
textual units, analysing them with reference to the language used and the
corresponding dimensions of FIELD, TENOR, MODE, and GENRE.
In other words, rather than forcing the text into the straitjacket of the
TQA model, the evaluator should redevelop the model from every text
and translation to be analysed (for a different solution to the problem, cf.
House 2015:126). This holds for both the analysis of the original and the
comparison of the original with the translation.
The question of subjectivity has always been a bone of contention in
translation quality assessment. While it is clear that subjectivity should
be avoided as much as this can be done, it should be equally clear that
the complete elimination of the subjective element in TQA is practically
impossible. (See Chapter 3 for a detailed discussion of the problem of
subjectivity.) Juliane House is fully aware of this problem. She writes that
“the ultimate judgment of quality resulting from the analyses contain
necessarily a hermeneutic, subjective component” (House 1997:103),
and, in more detail, she distinguishes between two steps:
the first relates to analysis, description and explanation based on
knowledge (of linguistic conventions), and empirical research, while
the second relates to judgments of values, to social and moral ques-
tions of relevance and appropriateness and, of course, to personal
preference or taste.
(House 1997:166)
The first step should render the second step intersubjectively verifiable in
that the analysis, description, and explanation provide other evaluators
(or the translator) with an argument which they can either embrace or
reject. House recognises that “[i]t is difficult to pass any ‘final judge-
ment’ of the quality of a translation that fulfils the demands of scientific
objectivity” (House 2014:261) and points out:
To judge is easy: to understand is less so. If we can make explicit

the grounds of our judgements, on the basis of an argued set of
procedures […], then, in the case of disagreement, we can talk and
discuss: if we do not, we can merely disagree.
(House 1997:166–167)
Yet, notwithstanding these insightful comments, there is still room for

improving House’s TQA model.
The mismatches established through a close comparison of the orig-
inal and the translation serve as a basis for the final assessment of a
translation’s quality. While generally producing interesting overall
results – revealing, for example, in the above-mentioned translation
of Five Minutes’ Peace the covert strategy of the translator and/or
publisher – House’s mismatch search seldom comes up with an individ-
ual translation alternative.1 In those cases where she does offer what she
thinks might be a better translation, House usually does not give any
reasons why the alternative is better than the original translation. A case
in point can be found in her analysis of excerpts from an English trans-
lation of Walter Benjamin’s philosophical essay “Die Aufgabe des Über-
setzers”. The opening sentence, “Nirgends erweist sich einem Kunstwerk
oder einer Kunstform gegenüber die Rücksicht auf den Aufnehmenden
für deren Erkenntnis fruchtbar” (House 1997:182) has been translated
by Harry Zohn, “In the appreciation of a work of art or an art form,
consideration of the receiver never proves fruitful” (House 1997:185).
Juliane House comments on this translation: “Less precise terms are
chosen: I: first sentence: receiver vs. der Aufnehmende: the processual
aspect is not captured in the translation (‘Perceiver’ as suggested by
Jacobs might be more adequate)” (House 1997:143). The question why
“perceiver” might be more adequate than “receiver” is left unanswered,
since the reason given (namely, that the translation fails to capture the
processual aspect) would suggest a rather clumsy solution along the lines
of “receiving person” or “perceiving person”. In the above context, the
term used in the German original, “Aufnehmenden”, suggests passivity
rather than activity, the effect-producing agent being the work of art; at
the same time, “Aufnehmenden” is not reminiscent of the receiving end
of the communication process – an important point in an essay which
argues against the communicative function of translation. With regard
to the two options “receiver” and “perceiver”, the following arguments
might now be put forward: while “receiver” in the sense of “someone
who receives” agrees with the passivity suggested by “Aufnehmenden”,
yet, unlike the German term, conjures up the relevance of the process
of communication; the other term lacks the communicative connota-
tion but calls for a more active involvement of the person denoted by
“perceiver”. Depending on the relative significance of the two criteria,
we could make a case in favour of either expression.
In Five Minute’s Peace, the translation of the following sentence gives
rise to criticism. “She emptied half a bottle of bath-foam into the water,
plonked on her bath-hat and got in” (House 1997:177). Britta Groiß
translated this, “Sie leerte eine halbe Flasche Badeschaum in das Wasser,
setzte die Badehaube auf und stieg in die Wanne” (House 1997:178).
Here, Juliane House spots a mismatch between the predicates: “setzte
die Badehaube auf vs. plonked on her bath-hat” (House 1997:129).
She lists this together with two other mismatches under the heading of
“Social Attitude” (ibid.) and explains, “Loss of informal style in some
instances, which reduces the humourous [sic] effect” (ibid.). While this
observation is certainly true, House does not provide any alternative
that would turn the mismatch into a matching equivalent. Unless an
alternative translation can be found which clearly makes for a better
match with the original, there is little use in criticising the first trans-
lation as a mismatch. In this case, the meaning of “plonked on”, that
is, put on hurriedly or clumsily, could be captured using the German
verb “aufstülpen” – however, not without drawbacks. A rendering such
as “stülpte die Badehaube auf” would sound a little old-fashioned –
too old-fashioned, perhaps, for children in the late twentieth century.
Less outmoded is the version with a prepositional phrase, “stülpte die
Badehaube auf den Kopf”, which has the disadvantage of adding an
element not mentioned in the source text and thereby slightly increasing
the prominence of the whole predicate. It is, then, not always easy to
undo one mismatch without producing another. The example shows that
what at first glance might appear to be an instance of normalisation –
that is, when “the author’s creativity has been neutralized” (Bowker
2001:355) – may prove sheer necessity and, in the end, remain the pre-
ferred translation alternative.
The central question is this: if a mismatch detracts from translation
quality, would any possible translation avoiding the mismatch auto-
matically improve the quality of the target text? As the above analysis
reveals, there is no clear answer in the affirmative because the alter-
native translation might give rise to a mismatch at a different level of
equivalence. The translation of a particular text segment is sometimes
good with regard to one quality aspect but bad with regard to another
(cf. Gerzymisch-A rbogast 1994:149–150). This is why an evaluator
should always compare one possible translation solution with at least
one other possible translation solution.
In summary, Juliane House’s approach is a valuable contribution to trans-
lation quality assessment, a contribution that helps to reduce the impact of
subjectivity in evaluating translation by furnishing a p ragmatic-linguistic
basis on which the decisions of the evaluator can be traced and verified.
This pragmatic-linguistic basis is supported by empirical studies –
conducted by House “over the past twenty years” (House 1997:79) – which
investigate typical English and German language use in different contexts.
The findings of this contrastive research help to substantiate the use of a
cultural filter in covert translation and justify deliberate deviations from a
straightforward (literal) translation prompted by preferences of the target
culture.2 Still, our analysis of House’s approach has highlighted a few
weak points. These weak points are not beyond remedy: the sometimes
confusing overlap in the discussion of the four categories can be avoided
if the text units rather than the categories are taken as the organising
principle of the translation analysis; the discussion of mismatches can be
expanded to include alternative translation solutions.
Malcolm Williams
In his book Translation Quality Assessment: An Argumentation-C entred
Approach, Malcolm Williams sets out to “explore the a pplication of one
particular aspect of discourse analysis – argumentation theory – to TQA
and develop an assessment framework to complement existing micro-
textual schemes, with specific reference to instrumental translation in
a production context” (Williams 2004:XVII–XVIII). He laments what
he calls an “assessment chaos” (Williams 2004:XIV) and identifies “the
problems and issues that stand in the way of consensus and coherence in
TQA” (ibid.). These include questions such as: What are the relevant as-
sessment criteria and how can the ratings for different criteria be merged
into one overall rating? When it comes to assessing language errors
and grading them, how accurate should be the translation of a word or
phrase and how serious is a particular error? Williams (2004) presents
an approach to translation quality assessment that finds answers to most
of these questions – albeit at the expense of a rather complex framework,
which is why he presents an updated model in 2009. The following
discussion will draw upon Williams (2004) as well as Williams (2009).
Dissatisfied with microtextual approaches to TQA and their often
inconclusive results regarding the quality of a translation, Williams
f ocusses on the argument structure of a text rather than individual
words and phrases. His method “draws primarily on philosopher
Stephen Toulmin’s analysis of argument structure” (Williams 2009:12)3
and starts from the assumption that texts are made up of arguments.
Williams explains:
[E]very instrumental source text contains an argument macrostruc-

ture and […] it is this structure that the translator must preserve in
the target text. This does not mean that other textual features and
functions are not present and are not important. What it does mean
is that preserving the argument macrostructure is the overriding
consideration for TQA.
(Williams 2009:11)
The idea behind this move away from the microtext to the macrotext is
that evaluators will be less likely to disagree about the argument struc-
ture of a text than about individual language issues. Williams claims
that “the evaluator proceeds according to very exacting criteria that
leave little margin for variation and inconsistency between assessments
(assuming consistency in evaluator competence)” (Williams 2004:150).
Brian Mossop, in a review of Williams (2004), agrees: the argument of
a text, he writes, “will usually not be a matter for debate, whereas a sys-
tem in which the score relies heavily on counts of minor error [sic] will
be more open to subjectivity” (Mossop 2004:189–190).
So how does the argument structure work? Based on six elements,
it indicates how a text unfolds its argument. These elements are claim,
grounds, warrant, backing, qualifier, and rebuttal. For three of these
elements (claim, qualifier, and rebuttal), there are alternative e xpressions
(discovery for claim, modalizer for qualifier, and exception and
restriction for rebuttal) that are used depending on the content of the
respective element of the argument structure. The arguments themselves
are characterised by organisational relations such as problem–solution
or question–answer, by conjunctions and their various (and sometimes
ambiguous) functions, by types of argument (for example, definition
or comparison), figures of speech, and narrative strategy (for example,
depersonalisation). Since this “argument schema” presented in Williams
(2004) is somewhat confusing in its complexity, Williams (2009) pro-
poses a simplified model that refers only to the two major argument
macrostructure features: claims and grounds. No matter which model
is used, the analysis and comparison of the argument structure in the
source and target texts constitute the pivotal element in Malcolm Wil-
liams’s TQA approach.
Once the argument structures of source and target texts have been
analysed and compared, the results need to be quantified. This is done
with a special rating scale based on an error definition that distinguishes
between critical errors, major errors, and minor errors. Williams, refer-
ring here to errors as defects, writes that
in an argumentation-centred TQA (ARTRAQ) model,
• defects impairing translation of the argument macrostructure

will be characterized as critical;
• other transfer defects conventionally considered major (con
tresens, charabia, etc.) will be characterized as major but will
be deemed not to render the translation unusable;
• other transfer defects will be characterized as minor.
(Williams 2009:13)
Thus, the priorities are clear: if a translation fails to preserve the argu-
ment structure of the source text, that translation will be considered a
failure. Any other errors, whether they concern the denotative meaning
of a word or a stylistic device, are not regarded as serious – as long as
they do not affect the text’s argument structure.
The performance of the translator is assessed with a “Weighted
A RTRAQ Grid” (Williams 2004:155, 2009:18). This is a table where
the parameters or groups of parameters – those relating to the argument
structure as well as those to do with microtextual issues – are individually
rated. The parameters are weighted depending on the type of text and
the purpose of the translation, with weightings amounting to a total of
ten. Then their quality in the actual target text is established as a mark
out of ten. The final score for each parameter is the product of the qual-
ity mark multiplied by the weight factor. This is eventually compared
to the minimum requirements. Note that, due to the central role played
by the argument structure, the minimum score required for the argu-
mentation parameters necessarily equals the maximum. The final result
is interpreted in terms of a set of translation quality standards. Wil-
liams (2004) distinguishes between maximum or publication standard,
information standard, minimum standard, and substandard. The mini-
mum standard would typically be achieved if the only element rendered
successfully is the argument structure. Williams (2009) adds to these a
student standard for educational purposes.
Any critique of Williams’s approach must take into consideration the
comprehensive practice-oriented goal of the overall model. This goal
implies an analysis of the entire source and target texts and the quan-
tification of the analysis results. While other, notably academic, TQA
models stop short of assessing the whole target text and shrink from
quantifying the final result, ARTRAQ – designed as it is for application
in a practical TQA environment – finds a way to include both compre-
hensiveness and quantification. The focus on argument structure helps
to reduce subjectivity, yet, it does not altogether do away with it. Thus,
there remains a subjective element with regard to the selection and
weighting of the assessment parameters and the numerical evaluation
process. Nevertheless, Williams does succeed in reducing s ubjectivity
to a minimum: first of all, his focus on argument structures provides a
common denominator where different evaluators are likely to agree4;
second, the rating of microtextual errors in terms of a positive quality
mark (rather than just counting and assessing individual mistakes) tends
to reduce the impact of the subjective element within the already low-
profile analysis of the microtext. In short, while it would be too much to
say that the argumentation-centred TQA model developed by Malcolm
Williams neutralises the role of the evaluator, it certainly achieves what
one might call controlled subjectivity.
As it is, Williams’s practice-oriented theory is supposed to reduce
the time needed to assess a translation without having to be content
with analysing samples. However, a full-blown translation quality as-
sessment based on a complete text analysis may not always be feasible
as the following comment suggests: “With respect to TQA procedures
and the issue of full text versus sample analysis, the evaluator will save
time by reading the complete translation to identify problem areas and
restricting detailed TQA to passages containing argument macrostruc-
ture components” (Williams 2009:14). Another bone of contention is
Williams’s assertion that “microtextual models are not designed to as-
sess each passage as an integral part of a whole” (Williams 2004:XVII),
because in our opinion this need not be an inherent defect: why should
not a TQA model based on microtext analysis incorporate macrotex-
tual and contextual elements? Any translation solution that strikes the
evaluator as inappropriate or out of place ought to be assessed against
its immediate and wider context (which includes textual meaning and
argumentative structure), considering cultural exigencies as well as
translation strategies. Equivalence between the source and the target
text would be n egotiated in an analysis that starts from the micro-
textual findings and relates them to the relevant macrotextual and
transtextual issues.
A more serious problem is Williams’s marginal reference to transla-
tion strategy and to cultural aspects of translation. In a short critique of
the ARTRAQ model, Juliane House writes:
The drawback of the standardized, norm-based procedure which

[Williams] suggests is that, in assuming the universality of argu-
mentative structure, he totally disregards the context- and culture-
boundness of texts. Even if such universality did exist, there might
still be culture-conditioned differences in the degree of explicitness
of argumentative structures in texts.
(House 2009:223)
This is an important point: while it may well be argued that there is
no instrumental text without an argument structure, there will almost
certainly be culture-related differences with regard to the explicitness of
the argument structure or the argument structure itself, depending on
the language pair involved in the translation process. In cases in which
an explicit message should be rendered more implicit, thereby adapting
it to the target culture so as not to offend the receiver, Williams’s model
fails to account for the cultural difference. Although it is said to “re-
spond to the functional requirement of an assessment/evaluation system
adaptable to different purposes and client needs” (Williams 2004:149),
ARTRAQ does not discuss the importance of a cultural filter or a trans-
lation strategy. Yet, as it constitutes an assessment framework that can
“complement existing microstructural schemes” (Williams 2004:XVII),
there might be some potential for combining Williams’s approach with
aspects from other TQA models.
Heidrun Gerzymisch-Arbogast and Klaus Mudersbach

Unlike House (1977, 1997) and Williams (2004, 2009), Heidrun
Gerzymisch-Arbogast and Klaus Mudersbach do not focus on translation
quality: in Methoden des wissenschaftlichen Übersetzens (Gerzymisch-
Arbogast/Mudersbach 1998), they develop a text-based translation
process, instructing the translator how to translate. F ollowing their
instructions would, of course, yield a high-quality translation result. In
this way, they furnish a tool that can be used to assess the quality of a
translation. This is also evident from the fact that Methoden des wis-
senschaftlichen Übersetzens emerged from working with Gerzymisch-
Arbogast’s Übersetzungswissenschaftliches Propädeutikum (1994),
an introduction to translation which also tackles the issue of transla-
tion criticism (cf. Gerzymisch-Arbogast/Mudersbach 1998:9). We will
discuss the methods proposed by Gerzymisch-Arbogast/Mudersbach
(1998) together with the approach to translation criticism derived from
Gerzymisch-Arbogast (1994).
Central to the translation procedure presented by Gerzymisch-
Arbogast/Mudersbach (1998) is the combination of three translation
methods: Aspektra, Relatra, and Holontra. 5 These three methods
complement each other in their concerted creation of the target text
(cf. Gerzymisch-Arbogast/Mudersbach 1998:41). They cover different
dimensions and, thus, help to organise the complex translation process:
while Holontra deals with knowledge systems that are manifest in the
source text and which need to be developed and rendered in the target
language on the basis of what the translator knows, Aspektra focusses
on specific aspects within the source text, that is, conspicuous points
at the text segment level that are weighted and then brought out in the
target text accordingly. Relatra assumes an intermediate position be-
tween the other methods in that it describes in a network the relations
between the various text segments. By comparing the networks of the
source text and any target text variants, the translator can check the
global proportions of these networks against each other to see if they
match (cf. Gerzymisch-Arbogast/Mudersbach 1998:58).
The three methods are applied to a free verse poem, “Für höchstes
Gut.”, by Klaus Mudersbach – a poem that reflects in a complex pattern
the consequences of the Chernobyl nuclear catastrophe. After a close
analysis along the lines of Holontra, Aspektra, and Relatra, the German
source text is translated into five languages: English, French, Italian,
Russian, and Spanish. It is assumed that the translations are published
in the feuilleton of a national paper or in an anthology of contempo-
rary poetry (cf. Gerzymisch-Arbogast/Mudersbach 1998:89). Against
this background, any rendering of the poem would be expected to be
based on the principle of overt translation: the original would be repro-
duced in the target language in such a way that the target text reader
is given a glimpse of the source text and source culture. The English
rendering, however, does not fully comply with this principle as is shown
by the deliberately covert translation of concepts relating to the know
ledge system “milk”. Here, what is typical of the German milk system is
translated into something t ypically English: for example, “Axel-frisch-
milch” (Gerzymisch-Arbogast/Mudersbach 1998:88) in the original
becomes “Dale Farm fresh milk” (Gerzymisch-Arbogast/Mudersbach
1998:143) in the translated poem. This contradicts the individual goal
for the English translation, namely, to keep all text components invari-
able in the translation process – as far as this is possible (cf. Gerzymisch-
Arbogast/Mudersbach 1998:112).
What is surprising is not so much the unexpected use of a cultural
filter but the lack of a proper argument in favour of such a filter for
the English translation. The question of how to translate the words and
expressions of the knowledge system “milk” is tentatively answered
only for the French rendering of the poem. Here, Gerzymisch-A rbogast
and Mudersbach argue (or rather Joëlle Philippi, the translator of the
French rendering, argues) that, in principle, there are two possible
approaches: an overt translation with the target text retaining the set-
ting of the source culture; and a covert translation, where the target text
is adapted to a target culture setting. They list the following advantages
of an overt translation: a greater credibility as the target text preserves
the coherence between the author’s stance and the scenes described; a
greater freedom for the translator when rendering the slogans on the
milk container since the translator can invent them (an argument that,
we believe, is supportive of a covert rather than an overt approach); an
easier handling of the cultural particulars, which are simply taken the
way they are. The disadvantages of an overt translation are: a possi-
ble alienation of the reader as he or she may not be able to relish the
specifics of the source culture manifestations; an unintended effect on
the target text reader, namely, that he or she gets the impression of not
being affected – as in the case of the Chernobyl accident. By contrast, in
a covert translation of the knowledge system “milk”, the disadvantages
of the overt approach would turn into advantages. The difficulty to find
target culture elements that retain the intended statement of the source
text is considered the only argument against a covert translation. Thus,
Gerzymisch-A rbogast and Mudersbach conclude that for the translation
of the “milk” system the covert approach is most suitable. They do not
give any specific reasons for their decision, though.6
Emphasising the need for a comparison of the different culture-specific
knowledge systems in the source and target cultures, Gerzymisch-
Arbogast and Mudersbach take the result of such a comparison as the
starting point for the development of any translation strategies and
solutions (cf. Gerzymisch-Arbogast/Mudersbach 1998:114). To us, this
appears to be the wrong order. A translation strategy should be governed
by the goal of the translation and defined prior to any holistic consider-
ations of knowledge systems made out in the source text; such consider-
ations can come into play in the implementation of a strategy, particularly,
when an overt translation solution, though required, seems to be inappro-
priate. Indeed, when the general aim of the translation is to retain the
background of the source culture and keep invariant as many source text
elements as possible, there is no need for a comparison of knowledge sys-
tems as the original system is to be preserved in the target text.
Another important feature of the approach by Gerzymisch-Arbogast
and Mudersbach is the weighting of knowledge systems in the Holontra
method and of aspects in the Aspektra method. This results in a ranking of
knowledge systems and aspects based on a given purpose of the translation.
The individual priorities thus revealed become relevant when the target
text is produced (cf. Gerzymisch-Arbogast/Mudersbach 1998:339–340).
However, the detailed grades assigned to each weighted knowledge system
or aspect seem to us too arbitrary to be fully comprehensible. At the same
time, the strict distinction between knowledge systems and aspects blurs
the fact that a particular aspect – rather than being weighted only against
other aspects – may as well run counter to the weighting of a particular
knowledge system. The complexity of the weighting procedure results in
an unwieldy translation programme. What should count is not the individ-
ual percentage but the relative weight of a particular aspect or knowledge
system when compared with a conflicting aspect or knowledge system. As
long as there is no difficulty in translating the source text in line with the
prevailing strategy, the weighting remains irrelevant. It is only when two
or more translation solutions vie with each other and a decision has to be
made in favour of one of them, that the weight of the aspects or knowledge
systems concerned has to be taken into account.
In spite of the above criticism, we consider Methoden des wissen-
schaftlichen Übersetzens to be an insightful addition to translation
theory. It integrates the varied steps of the translation process into one
coherent and largely verifiable modus operandi, presenting the student
of translation with a useful learning tool. The usefulness of this tool
would still be enhanced if the methods were simplified to the extent
set out in the above discussion. That the process is too time-consuming
to be practicable for professional translators is also recognised by
Gerzymisch-Arbogast/Mudersbach (1998:52, 1998:327). Nevertheless,
the process produces a translation that, in all its facets, is based on well-
founded, informed decisions. This is also true for the English translation
of Klaus Mudersbach’s poem (despite the strange mixing of overt and
covert translation strategies), since every micro-level rendering emerges
as the result of a thorough analysis combining Holontra, Aspektra, and
Relatra. It is obvious that such an approach may also be used in transla-
tion criticism. This is what Heidrun Gerzymisch-Arbogast (1994) does
in her book Übersetzungswissenschaftliches Propädeutikum.
In order to devise a system that accounts for any translation preferences
in the light of macro-level and micro-level decisions, Gerzymisch-
Arbogast (1994) looks at specific translation problems from a m acro-level
and micro-level perspective. Her approach anticipates to some degree the
three methods propounded in Gerzymisch-Arbogast/Mudersbach (1998),
yet, it draws more explicitly on linguistic techniques which can help to
explain the details of the translation process. The book is designed as an
introduction to translation, in particular, to the methodological prob-
lems of translation (cf. Gerzymisch-Arbogast 1994:9) and as such uses
many text examples to demonstrate the possible solutions. A recurrent
text – one which is discussed at macro- and micro-levels – is a one-page
excerpt from the novel Lemprière’s Dictionary by Lawrence Norfolk
and its German translation by Hanswilhelm Haefs. Soon after its pub-
lication, this translation caused a highly controversial but rather poorly
argued debate about its merits and demerits (cf. Gerzymisch-A rbogast
1994:18–21 for a summary). Gerzymisch-Arbogast takes this debate as
an opportunity to develop her translation approach into a tool which can
be used to make statements relating to the quality of a translation.
Gerzymisch-Arbogast’s translation critique hinges on a matrix that
combines individual text passages with what she calls “aspects”. These
aspects are textual properties with variable values or different levels of
manifestation. They are established in a three-step process: first, the
translation is read through and any conspicuous points in the text are
marked with the intention to find out how the text creates the effect it
has on the reader; second, the source text passages corresponding to the
conspicuous features found in the translation are analysed with a view
to determining whether they are also conspicuous within the context of
the original; third, both source and target texts are scanned for addi-
tional aspects that have escaped notice so far, for example, some aspects
relating to the artistic or linguistic design of the overall text. The result
is a list of aspects which is then applied to the source and target texts,
yielding a matrix that reveals the differences between the original and
its translation for each aspect (cf. Gerzymisch-Arbogast 1994:148–149).
The aspect matrix provides the basis for the quality assessment of
a translation. If applied carefully to a complete text or to representa-
tive excerpts of a text, the approach can be used to produce a well-
argued and intersubjectively verifiable translation critique. However,
as Heidrun Gerzymisch-Arbogast points out, such a critique will not
extend to the translation as a whole but only to individual aspects, so
that the translation quality may be good with regard to one aspect and
bad or indifferent with regard to another (cf. Gerzymisch-A rbogast
1994:149–150). The comparison and discussion of the aspect values
is generally c onvincing – especially, in those frequent cases in which
Gerzymisch-A rbogast proposes an alternative translation solution to the
not-so-successful rendering of a given aspect. There are but two minor
drawbacks to this method: first, it is very time-consuming; second, the
question of a translation strategy (for example, in terms of an overt or
covert translation) is not given enough prominence. As a TQA model,
Gerzymisch-Arbogast’s approach is clearly more useful in an academic
environment than in the world of professional translation.
Ernst-August Gutt
While Heidrun Gerzymisch-Arbogast and Klaus Mudersbach (1994 and
1998) provide a hands-on guide to translation and translation evaluation,
Ernst-August Gutt (1991) – in his book Translation and Relevance –
presents a theoretical guide to translation and translation evaluation.
By drawing on Dan Sperber and Deirdre Wilson’s relevance theory, he
gains a fundamental understanding of the principles of translation. Gutt
sets out to “formulate a general theory of translation” (Gutt 1991:vii);
yet, in the end he is surprised that “relevance theory alone is adequate –
there seems to be no need for a distinct general translation theory” (Gutt
1991:viii). We will examine whether this assumption is true and to what
extent Gutt’s approach can be used in a TQA environment.
In his discussion of existing translation theories and more practice-
oriented approaches to translation, Gutt first addresses the question whether
there could ever be a science of translation and comes to the conclusion that
one of the main problems with the scientific investigation of transla-

tion seems to lie in the fact that not only linguistic factors, but many
other factors need to be taken into account. Since these factors be-
long to a variety of different areas of life, there is a question whether
a comprehensive account of translation in the form of a coherent and
homogeneous theory can ever be achieved.
(Gutt 1991:5)
In other words, the quantity and variety of factors involved in the trans-
lation of a source text seem to render a unified scientific theory virtually
impossible. Gutt further shows that the domain of translation is diffi-
cult to determine: its outline often remains unspecified because there
are too many products and processes for which the term “translation”
is used; alternatively, a definition is given, which, however, excludes all
phenomena (whether referred to as “translations” or not) that fail to
comply with the criteria of the definition; finally, there is Descriptive
Translation Studies, where culture-specific criteria are used to outline
the notion of translation (cf. Gutt 1991:5ff). Gutt criticises that the
descriptive approach “either […] leads to the abolition of the intercul-
tural study of translation or […] does in fact rely on non-culture-specific
criteria for determining its domain” (Gutt 1991:7–8). The conclusion
that translation studies does not furnish a satisfactory and comprehen-
sive explanation of its own subject matter is one motive for Gutt to look
for an explanation elsewhere.
Another motive can be found in his analysis of the concepts of evalu-
ation and equivalence. Gutt shows that equivalence – and, in particular,
functional equivalence – cannot easily be accounted for by a theoret-
ical framework since the units which are expected to be equivalent in
translation present themselves as rather elusive textual phenomena that
change depending on the target text and the respective target audience.
Gutt writes:
[I]f it turns out that each individual phenomenon – which here is

not only each text, but potentially each instance of translating it for
a particular audience – may require its own theory of equivalence,
then this means that these phenomena cannot be accounted for in
terms of generalizations at all, and that they actually fall outside the
scope of theory.
(Gutt 1991:12)
Equivalence not only appears to be difficult to harness from a theoreti-

cal point of view but also does not provide a sound basis for evaluation.
Looking at Juliane House’s (1981) approach to translation quality as-
sessment, Gutt argues that the concept of equivalence can at best pave
the ground for a comparison of source text and target text; however,
ultimately “the notion of equivalence itself is inadequate for evaluating
translations” (Gutt 1991:14).
Gutt believes to have found a solution to the above problems. He
abandons the text-orientedness of traditional translation studies, claim-
ing that, with relevance theory, “the host of different factors noted as
important in recent years […] are naturally covered in the only way in
which they can have an influence on translation anyway – and that is as
part of our mental life” (Gutt 1991:20). According to Gutt, relevance
theory is supposed to provide “a natural basis for an empirical account
of evaluation and decision-making” (Gutt 1991:21) in translation. This
is said to be true for both micro-level and macro-level decisions, so that
relevance theory can help to answer, for example, the central question
whether to translate a text covertly or overtly. A theory of communica-
tion, relevance theory – when applied to translation – accounts for the
translation process essentially as an ostensive communication process.
This means that the author of the source text and the translator are
regarded as communicators who deliberately communicate a message
to an audience. While this seems rather straightforward, there is more
to it than a mere communication process. Gutt explains, “Relevance
theory approaches communication from the point of view of compe-
tence rather than behaviour: it tries to give an explicit account of how
the information-processing faculties of our mind enable us to commu-
nicate with one another” (Gutt 1991:20). It should be noted that the
communication process is not regarded as a simple matter of fact but as
dependent on the communicative competence of the communicator(s)
and the audience. As a consequence, any context implied in the com-
munication “does not refer to some part of the external environment of
the communication partners […] it rather refers to their ‘assumptions
about the world’ or cognitive environment” (Gutt 1991:25). Within
this framework of ostensive communication, the principle of relevance
plays a crucial role. The idea is that the receiver of a message expects
“his attempt at interpretation [to] yield adequate contextual effects at
minimal processing cost” (Gutt 1991:30). The communicator, on the
other hand, has the responsibility to communicate in such a way that
the receiver can easily interpret the message in the way intended by the
communicator (cf. Gutt 1991:32).
The principle of relevance as applied to the process of translation has
significant implications. The translator as communicator translates a
source text with the intention to fulfil the expectations of the readers of
his or her translation: for the readers, in trying to comprehend the target
text, should not have to make an unnecessary processing effort (cf. Gutt
1991:101–102). It is against this background that Gutt presents his ex-
amples of translation evaluation. Here is one in which Gutt starts with
a quotation from Adams:
At a climactic moment in Stendhal’s Le rouge et le noir (Book II,

ch. 19), Julien Sorel, after weeks of solitary suffering, has finally
climbed back into Mathilde de la Mole’s good graces, and so under-
takes once more the perilous ascent, via a ladder, to her midnight
bedroom. She receives him with ecstatic, unbounded delight, crying,
‘C’est donc toi!’ And just here C. K. Scott-Moncrieff – for whose
extraordinary gifts as a translator I have, as a general rule, only the
highest respect – slips on the insidious banana peel, and translates,
‘So it is thou!’ What girl of high social rank and free social manners
ever greeted a lover that way?
(Adams, Robert M. 1973:14, Proteus, his lies,
his truth: discussions of literary translation,
New York: Norton; as quoted in Gutt 1991:102)
To analyse this translation critique, Gutt distinguishes two perspectives:

that of the translator, Scott-Moncrieff; and that of Adams, the critic,
as a reader or – to use Gutt’s expression – as the audience. From the
translator’s perspective, the translation can be justified along the
following line of argument (cf. Gutt 1991:103):
1 Since Adams obviously considers Scott-Moncrieff to be a good

translator, it is unlikely that the target text solution, “thou”, has
been selected by chance. Quite the contrary: the form must have
been chosen on purpose because it is rather uncommon in modern
English.
2 As “thou” is so uncommon, it creates a special contextual effect.
It resembles the French pronoun “tu” in that it is also a singular
pronoun, which would not necessarily have been the case with the
common “you”. In this way, the translator wanted to convey – by
way of implicature – the intimate relationship between Julien and
Mathilde.
To shed light on the position of the audience, Gutt provides additional

information. He quotes Adams’s explanation, who indicates that, for an
English readership, “thou” is associated with obsolete or e cclesiastical
language and certainly does not express intimacy. Gutt takes these
remarks as a starting point for a relevance-theoretical analysis (cf. Gutt
1991:104–106):
1 The fact that “thou” is an obsolete pronoun which is at best used in

an ecclesiastical context implies storage in a less accessible part of
the memory. This, in turn, means more processing effort to retrieve
that pronoun.
2 A reader who invests more processing effort would normally expect
to derive a special benefit from the retrieval of an item that is diffi-
cult to get at.
3 However, it is important to consider that readers try to make sense
of what they read. For it is not simply the first interpretation that
springs to mind which counts but the first interpretation which is at
the same time consistent with the principle of relevance. That is to
say, the processing cost invested in retrieving the meaning of “thou”
must yield adequate contextual effects.
4 Given the strong effort required to process “thou”, the reader is
likely to react in one of four different ways: he or she may simply
read on without having fully understood Mathilde’s exclamation;
if there have been other obstacles before, he or she may break off
and reject the whole translation as not worth reading; he or she
may try hard to make sense of the translator’s choice and, perhaps,
misinterpret the use of “thou” as being ironic; having some know
ledge of French, the reader might recognise the underlying pronoun
“tu”, derive the contextual implications regarding the relationship
between Julien and Mathilde, and continue reading.
The above detailed analysis of this fairly straightforward translation prob-

lem shows clearly how the principle of relevance could be employed in
translation assessment. Gutt, however, stops short of giving an evaluative
account of individual translations. It seems that, despite his discussion
of evaluation in the first chapter of his book, he does not consider the
achievement of his theoretical account to lie in an application to TQA.
Evaluation is just one of many aspects in translation for which rele-
vance theory can serve as a useful analytical tool. This becomes clear in
the following quotation:
I tried to show that the principles, rules and guidelines of transla-

tion are applications of the principle of relevance; thus the proposal
is that all the aspects of translation surveyed, including matters of
evaluation, are explicable in terms of the interaction of context,
stimulus and interpretation through the principle of relevance, a uni-
versal principle believed to represent a psychological characteristic
of our human nature. Thus the main contribution of this book is a
reductionist one on the theoretical level – issues of translation are
shown to be at heart issues of communication.
(Gutt 1991:188)
The central message of the book, which Gutt states already in the pre
face, is that “since the phenomena of translation can be accounted for
by this general theory of ostensive-inferential communication, there is
no need to develop a separate theory of translation” (Gutt 1991:189).
While we would agree that relevance theory can be a core element when
it comes to analysing and evaluating a translation, there are several ob-
jections to such a reductionist claim. The first objection concerns the
fact that the process of translation cannot always be explained in terms
of the principle of relevance – at least not in a straightforward manner.
For the translation process is not normally restricted to an author and a
translator, who – as communicators – have to achieve an optimum effect
for a readership that seeks maximum benefit from a minimum process-
ing effort. The client of a translation may not read the target text but still
stipulate that the translator follow certain rules (which might contradict
what the translator considers to be relevant). The second objection has
to do with the perspectives on relevance of the translator and his or her
readership: what the translator regards as relevant to his or her reader-
ship may not be appreciated by all readers. Thus, readers with a different
view of what a translation should look like will probably be unsatisfied
with the translator’s result.7 The third objection relates to the reduction-
ist character of Gutt’s relevance theory. Even if we granted that the prin-
ciple of relevance were at the bottom of all translation, there would still
remain a vast number of factors – linguistic as well as non-linguistic –
by which relevance in a particular translation situation would have to
be negotiated. These factors are altogether missing in Gutt’s theory, al-
though some of them turn up piecemeal in the discussions of individual
translation solutions.
As far as translation evaluation is concerned, Gutt’s major achieve-
ment consists in providing an extremely insightful framework that can
help to justify individual translation decisions at the top evaluation level.
Such justification, however, requires argumentative support at lower
levels – support that should also be accounted for by a theory which, at
least indirectly, purports to provide a basis for the analysis of transla-
tion quality. Given these points of criticism, Gutt’s approach is certainly
worth considering as a tool that can be used in combination with the-
oretical models dedicated to translation quality assessment. Relevance
theory seems not particularly suited to the more practical challenges of
TQA, where complete translation texts have to be assessed. Here, too, it
may well serve as a complementary tool.
This concludes our detailed review of the translation theories put
forward in House (1997), Williams (2004, 2009), Gerzymisch-Arbogast
(1994) and Gerzymisch-Arbogast/Mudersbach (1998), as well as Gutt
(1991). We have found that the approaches presented in these books –
different as they are – make valuable contributions to the ongoing
discussion of translation quality. While House (1997) and Williams
(2004, 2009) arguably provide the most practical methodologies, Gutt’s
(1991) would be considered the most theoretical approach. The ana-
lytical schemes proposed by House and by Gerzymisch-Arbogast and
Mudersbach stand out for their linguistic meticulousness. Williams’s
is less detailed because it foregrounds a text’s argument structure and
marginalises the microtextual features. The relevance-theory-based ap-
proach by Gutt is strikingly insightful; yet, it neglects the factors that de-
termine any relevance-related decisions. Many of the drawbacks pointed
out in the analyses can be offset by adding an argumentative basis and/
or combining the advantages of one approach with those of another.
Other Approaches to Translation Quality

Having discussed in some detail the works of Juliane House, Malcolm
Williams, Heidrun Gerzymisch-Arbogast and Klaus Mudersbach, as
well as Ernst-August Gutt, we will now present further approaches to
translation quality – some of which resemble the above in their method-
ological design, whereas others assume a completely different point of
view. While not pretending to be exhaustive, the outline sketched below
is designed to serve as a rough guide through the maze of perspectives
on translation quality. The idea is that it will illuminate some of the sim-
ilarities and differences between the various approaches with regard to
the two distinctive features: research focus and methodology. Since these
features are not sufficiently refined to suggest a presentational structure,
we will organise this section more or less chronologically.
Despite the chronological order, this overview is not intended as an
outline of the history of translation quality assessment. The eternal
dispute between a translation that clings to the source and a trans-
lation that turns to the target shows in the time-honoured works of
Luther (cf., for example, Luther 1969:15–20), Dryden (cf., for exam-
ple, Dryden 1900:237), Tytler (cf., for example, Tytler 1907:7–8), and
Schleiermacher (cf., for example, Schleiermacher 1813:152) as well as
in the translation concepts promoted by more modern theorists such as
Nida (cf., for example, Nida 1964:165–166), House (cf., for example,
House 2001:250), Newmark (cf., for example, Newmark 1991:10–11),
and Ladmiral (cf., for example, Ladmiral 1993:288). This dispute con-
stitutes the ever fleeting background against which the quality of a
translation had to be assessed in the past and still has to be assessed in
the present.
Approaches to Translation Quality in the Twentieth Century

There have been numerous attempts at capturing translation quality in its
slick appearance. Our outline begins with the third Congress of the In-
ternational Federation of Translators held in 1959 in Bad G odesberg –
the proceedings of which were published in 1963 by Edmond Cary and
Rudolf Walter Jumpelt. The contributions which we present here are
characterised by very insightful comments on translation and transla-
tion quality. Take, for example, Jumpelt, who points out that “[e]ach
decision involves a distinction between alternatives, and in order to dis-
tinguish objectively criteria are needed” (Jumpelt 1963:269). Eventually,
“each criterion will be not a rigid, immutable factor but rather a vari-
able” (ibid.). Such considerations will play a major role in the approach
to translation quality laid down in this monograph. Kandler and Zilahy
furnish philosophical wisdoms such as “[E]ven a good translation is
bound to be wrong in some respect” (Kandler 1963:296) or “A transla-
tion is considered good when it arouses in us the same effect as did the
original” (Zilahy 1963:285). All three articles look at translation quality
from so general a theoretical perspective that the question whether the
focus is on translation as a product or on translation as a process can be
regarded as rather irrelevant.
Machine translation is a type of translation that poses particular
challenges for the quality of the translation result, especially, when it
is carried out fully automatically. In a machine translation study, “An
Experiment in Evaluating the Quality of Translations”, John B. Carroll
(1966) asked test subjects to evaluate machine and non-machine trans-
lations from Russian into English. The evaluation criteria were compre-
hensibility and information content of the translated text, resulting in a
nuanced statistical analysis. Since machine translation affects the central
issues of the present work only marginally, there will not be many more
references to quality discussions with this type of translation. The vast
majority of the approaches to translation quality analysed here have to
do with human translations.
In her highly influential book Möglichkeiten und Grenzen der Über-
setzungskritik, Reiß (1971) paves the way for a holistic TQA approach.
Unlike House (1997) or Williams (2004), she is not so much concerned
with the application of her assessment system to whole texts; rather,
she refers to translation criticism in general, providing a sound theoret-
ical basis for any translation critic. While the oft-quoted text t ypology8
constitutes the framework of Reiß’s theory, the approach would be in-
complete without two sets of linguistic and extra-linguistic factors. Reiß
emphasises that an objective and adequate assessment hinges on the
combined effect produced by the text-typological as well as linguistic
and extra-linguistic features of the source text (cf. Reiß 1971:24). The
interplay of these features defines the opportunities of translation criti-
cism; the limitations of translation criticism become evident when trans-
lations are so specialised as to be hardly recognisable as translations and
when subjective aspects restrict the validity of the critic’s judgement. On
the whole, Reiß’s approach is very insightful; however, it requires some
fine-tuning if it is to be used in practical TQA.
Van den Broeck (1985) yields general insightful information on trans-
lation quality assessment without going too much into detail about spe-
cific assessment criteria. Focussing on translation as a product rather
than a process, he writes:
In my view, translation criticism, despite the subjective element in-

herent in value judgements, can be an objective account if it is based,
at least implicitly, on systematic description. The starting point for
this description will be a comparative analysis of the source and
target texts. Furthermore, a thorough description demands that
not only text structures but also systems of texts be involved in the
comparison. It is only at this point that the critic’s value judgement
can come into operation. However, in the confrontation of his own
critical standards with the norms adopted by the translator, the
critic should clearly distinguish one from the other. His evaluation
should take account not only of the translator’s poetics but also of
the translational method adopted by the translator in view of the
specific target audience envisaged, and of the options and policies
followed in order to attain his purpose. The final outcome of this
confrontation will be the reviewer’s critical account.
(Broeck 1985:56)
Furthermore, there is the distinction between “obligatory shifts” and

optional shifts” in translation (Broeck 1985:57) – the former being “rule-
“
governed, i.e. imposed by the rules of the target linguistic and cultural sys-
tem”; the latter being “determined by the translator’s norms” (ibid.). The
notion of a “shift” implies an invariant from which the translator devi-
ates. This invariant serves “as a tertium comparationis” (ibid.) and is also
referred to as “the Adequate Translation” (ibid.). Van den Broeck points
out that “the Adequate Translation is not an actual text, but a hypothet-
ical reconstruction of the textual relations and functions of the [source
text]” (ibid.). In order to reconstruct the textual relations and functions
of the source text, he advocates an analytical process by which can be de-
scribed “such features, on various levels of description, as are functionally
relevant for the structural relationships within the source text and for the
structure of the text as a whole” (ibid.). While the thus established tertium
comparationis between the source text and the target text is about as close
as one can get to an intersubjectively verifiable b asis of translation, it is
nonetheless dependent on the analytical skills of the translator. In other
words, different translators may well arrive at different tertia compara-
tionis or, for that matter, at different Adequate Translations. Still, with
his emphasis on acknowledging the role of the translator, van den Broeck
sketches an overall picture of translation criticism which may well serve as
the foundation of a full-blown TQA theory.
From a more practical perspective, Williams (1989) looks at trans-
lation quality assessment in a professional setting. In describing the
third version of the assessment standards and procedures developed by
the Canadian government’s Translation Bureau, Williams tries “to ex-
plain the thinking behind those standards and procedures, specifically
in response to the criticism levelled against any systematizing of TQA”
(Williams 1989:15). The issues discussed include: the reliability of the
client’s judgement, the question of what constitutes an acceptable trans-
lation, the problem of subjective evaluation, the concept of consequence
of error, the distinction between major and minor errors, the establish-
ment of acceptable and unacceptable quality levels or ratings, the practi-
cal implementation of TQA, and a model assessment. The point of view
adopted in Williams (1989) is very similar to that of his later publica-
tions (Williams 2001, 2004, 2009): it reveals an interest in the q uality
of translation as a product and presents a coherent theoretical TQA
framework; however, it does not yet feature the argumentation-centred
approach developed particularly in Williams (2004).
“Up until now, translation quality assessment has remained qualita-
tive rather than quantitative”, writes Fan (1990:43). By this, he does
not simply allude to the fact that many TQA approaches do not count
mistakes; rather, he points out that the terminology used to evaluate a
translation – such as excellent, good, fair, and poor – is too vague. In
an attempt to specify the quality of a translation more precisely, Fan re-
lates the evaluative expressions to bands of numerical values and devises
a mathematical model by which pre-defined translation criteria can be
measured (including information content, imagery shift, stylistic levels,
affective elements, metalinguistic aspects, rhetorical devices, and, in
poetic translation, metrical qualities):
The proposed mathematical model […] is based on the general prin-

ciples of the fuzzy subset theory and follows an approach which
might be called analysis-synthesis fuzzy evaluation. The results so
obtained are sorted out, using the statistical method, that is, by
counting the number of Excellents, Goods, Fairs, and Poors given
by a judge for a given number of randomly selected sample units of
translation (normally sentences). And the final result will be arrived
at by some appropriate mathematical operations.
(Fan 1990:48)
In this way, translation quality can, indeed, be represented in terms of

a numerical value calculated for the translation as a whole on the basis
of randomly sampled translation units. Fan demonstrates his a pproach
analysing a Chinese translation of Iacocca: An Autobiography by Lee
Iacocca. Yet, while the mathematical model ensures objective handling
of weights and values assigned to individual criteria and translation
solutions, the criteria themselves and any weights and values are subject
to individual preferences. Fan acknowledges that there is some scope
regarding the use of the criteria. “The criteria”, he writes, “may be
added or deleted, amended or redefined, and the weights can be altered
and redistributed accordingly, to suit different evaluative purposes” (Fan
1990:54). Ideally, there should be “several judges […] to assess the qual-
ity of the same work of translation, so that the result might be more
reliable and therefore more authoritative” (ibid.). Fan’s method would
benefit from utilising a specific rubric that provided explicit guidance as
to what weights and values should be given to criteria and translation
solutions.
In general, text-based approaches to translation quality assessment
may either rely on an existing model such as that of Juliane House or
work with a model specifically designed to suit the researcher’s needs
and preferences. This is what Barghout (1990) does when he develops
his “rhetorical model” (Barghout 1990:100). Distinguishing between
obligatory meanings, extended meanings, and accessory meanings in
literary, non-literary, and hybrid texts, he emphasises the importance
of text analysis and looks at “semantic shifts” (Barghout 1990:132) in
the translation of various texts from Arabic into English and vice versa.
Though clearly theoretical in its general approach, Barghout’s contribu-
tion to translation quality assessment also includes “an experiment con-
ducted to test proficiency in text analysis and text translating” (Barghout
1990:154). In another investigation into the quality of English-Arabic
translations, Mohamed Benhaddou acknowledges that “[t]he present sit-
uational/textual dimensions model is an adaptation of House’s (1981)
situational model” (Benhaddou 1991:118). Benhaddou’s empirical study
would be referred to as product-oriented. A third work which relates to
an English-Arabic context, albeit in quite a different way, is the doctoral
thesis by Al-Bustan (1993). After an extensive presentation of the funda-
mentals of translation quality, he looks at the process of translation from
an empirical perspective, analysing in a survey the quality and efficiency
of more than 200 translation agencies in Kuwait and the UK.
While Benhaddou (1991) adapts Juliane House’s model for the pur-
poses of his TQA analysis, Rodrigues (1996) presents a straightforward
application of the Housian approach, analysing a translation into Bra-
zilian Portuguese of an essay by Tom Wolfe, “The Me Decade And The
Third Great Awakening”. Unlike Rodrigues (1996), Hönig (1998) does
not evaluate a particular translated text but supports his functionalist
view of translation quality with a few telling examples. From a different
perspective, Larose (1998) – whose astute deliberations on translation
quality assessment have been acknowledged by Williams (2004:9–11) –
is concerned with the appropriateness of the different frameworks used
to assess the quality of a translation. By contrast, Horton (1998) demon-
strates his idea of translation quality assessment by applying a general
integrative approach to the German translation of an English Rolex ad-
vertisement. This turns out to be a rather difficult undertaking since
the target text is “a hybrid text […] produced by a combination of two
essentially distinct processes of text manipulation: translation and adap-
tation” (Horton 1998:107).
Similarly, Al Qinai (2000) analyses an Austin Rover brochure ren-
dered from English into Arabic. He “develop[s] an eclectic practical
model that can be empirically tested for analyzing the linguistic and
situational peculiarities of ST [source text] and TT [target text] in the
pre-translational phase and the post translational [sic] assessment of TT
quality” (Al Qinai 2000:499). The model employed is similar to that of
Juliane House; that it “can be empirically tested” must be regarded as
a suggestion for further studies as Al Qinai (2000) does not analyse the
responses of test subjects, nor does he examine a text corpus. In her very
interesting essay “Translation Quality Assessment: Where Can Theory
and Practice Meet?”, Susanne Lauscher (2000) assumes a decidedly
holistic point of view and rejects reductionist approaches as insufficient:
The ultimate goal seems to be the establishment of a conclusive list
applicable to all translations, and the criteria mentioned typically
refer to correct language use in terms of the target language sys-
tem […]. Once that definitive list has been created, it is assumed, it
should be possible to make prescriptive judgements about individ-
ual translations being ‘good’ or ‘bad’. However, practical definitions
of translation quality […] suggest that such lists are not sufficient
for determining translation quality in a professional setting, where
quality seems to depend on a variety of very diverse factors.
(Lauscher 2000:150)
Which factors should be relevant to the quality of a particular transla-

tion has to be decided by the parties involved in the translation project,
so that, ultimately, “[t]ranslation quality assessment and the judgement
of translation are a matter of communication, co-operation and con-
sent” (Lauscher 2000:164).
Approaches to Translation Quality in the Twenty-First

Century
The theorising text-focus of Lauscher’s approach can also be found in
Halliday (2001). Here, too, the general idea is that the question whether
a translation is good “must depend on a complex variety of different
factors that are constantly shifting in their relationship one to another”
(Halliday 2001:14). Halliday looks at how the concept of equivalence
can be harnessed to render it useful in the discussion of translation
quality and propounds the following definition: “A ‘good’ translation
is a text which is a translation (i.e. is equivalent) in respect of those
linguistic features which are most valued in the given translation con-
text” (Halliday 2001:17). Equivalence is only one aspect among many in
Beverly Adab’s analysis of the translation of advertisements from French
into English and vice versa. She establishes a framework of criteria in
which macro- and micro-textual features are complemented by invari-
ability of the overall message, overall potential impact, and adequacy for
purpose (cf. Adab 2001:152).
More concerned with translation as a process is Pinto’s (2001) out-
line of the process-related quality factors in documentary translation.
She clearly favours a functionalist approach and maintains that “[o]ne
important measurement of quality should be the clarity and readability
of the final product” (Pinto 2001:298). Yet, as her article is supposed
to provide “an overview of relevant developments” (Pinto 2001:288)
concerning “the difficulties involved in integrating translating models
and quality systems” (ibid.), Pinto does not investigate in depth but pre
sents glimpses of quality issues in the translation process, such as quality
management, client satisfaction, staff satisfaction, staff management,
and the use of documentation by the translator. Occasional quality as-
pects of the translated text (readability criteria, consistency) are con-
tained within the process-oriented framework. Another non-empirical
and process-oriented approach is presented by Gummerus/Paro (2001).
They look at how organisational design affects the quality of screen
translation in the context of a broadcasting company. The factors rele-
vant to good translation quality include criteria for the recruitment and
training of translators as well as for revising and editing translations.
Particular attention is given to questions such as “How is the trans-
lation production organized? Is translation seen as an integral part of
programme production? What is the position of the translators in the
organization?” (Gummerus/Paro 2001:133).
While some TQA approaches are general in the sense that they do
not restrict their outlook on translation quality, others clearly focus on
either a professional or an educational environment. The articles re-
ferred to in this paragraph all look at translation quality in the context
of translator training at a university. In her discussion of pedagogical
aspects of translation evaluation, Hannelore Lee-Jahnke (2001) empha-
sises the importance of the function of the original text and its purpose,
and recommends that errors be assessed within the broader context of
the overall message, the structure of the text and its effect on the reader
(cf. Lee-Jahnke 2001:267). Lynne Bowker (2001) presents a practical
and objective approach to translation evaluation, particularly to the
evaluation of specialised translations. She argues:
Given that translation evaluation entails making judgements about

appropriate language use, it should not rely on intuition, anecdotal
evidence or small samples; rather, such studies require empirical
analyses of larger bodies of authentic text, as found in the corpus-
based approach.
(Bowker 2001:346)
Bowker describes in detail the design of a special evaluation corpus.

Without recourse to corpus analysis, Rosenmund (2001) examines the
possibilities of objective evaluation. His recipe is called “Lastenheft” –
specifications which are drawn up prior to the actual translation pro-
cess and which stipulate certain points that must be fulfilled for the
translated text to be acceptable. Truly empirical is Waddington’s (2001)
comparison of four different methods of evaluation (Method A: error
analysis with a tripartite grouping of mistakes; Method B: error analysis
that takes into account “the negative effect of errors on the overall qual-
ity of the translations” (Waddington 2001:314); Method C: a holistic
approach; Method D: a combination of methods B and C). Here, five
correctors apply the four methods to the translations of 64 students.
The result is that “all four methods have proved to be equally valid in
spite of the considerable differences that exist between them” (Wadding-
ton 2001:322). While Waddington (2001) focusses on the validity of
the four methods, Waddington (2004) uses the same study to look at
the methods’ reliability. The results show that “methods based on error
analysis are more reliable than holistic ones” (Waddington 2004:34) and
that the combination of error analysis with a holistic method “greatly
improves the latter’s reliability” (ibid.).
Another empirical study is presented in Sharkas (2005). The overall
aim of this rather substantial work is “to identify factors that affect the
quality of English-Arabic scientific translations for the general public”
(Sharkas 2005:31). Using readers’ evaluation tests, he is in a position
to identify translation problems and establish the quality of a transla-
tion. In an approach that looks at the quality of translated texts from a
linguistic and purely theoretical perspective, Melby et al. (2005) com-
pare two concepts of meaning in translation: they claim that “given two
contrasting approaches to what is a good translation and two corre-
sponding perspectives on meaning, the preferred approach to translation
will lead [them] back to the preferred perspective on meaning” (Melby
et al. 2005:404). In their opinion, a functionalist approach based on a
“triad-semiogenic model” (Melby et al. 2005:440) can best capture the
many factors involved in the discussion of translation quality. The ques-
tion whether the object of investigation is the translated text or the pro-
cess of translation can be considered rather irrelevant in Secară (2005),
who gives an overview particularly of metrical approaches to translation
evaluation. The last article to be discussed in this paragraph presents
again an empirical study: Waddington (2006) investigates the impact
of errors on translation quality. He quotes several authorities on how to
deal with translation errors and concludes:
In spite of these caveats about the difficulty of gauging the effect of

translation errors, and while recognising that this difficulty will vary
considerably depending on the type of text that has been translated, I
still insist that, if we agree that translation is essentially a communi-
cative activity, one of the most logical ways to judge the importance
of translation errors is by endeavouring to calculate their effect on
translation quality. The evaluator is asked not just to assess the quality
of the dead letter of the translation, but to judge how successfully the
translation works as a dynamic vehicle of communication which pur-
ports to transmit the content (in the widest possible sense of this word)
of the source text to the new target text reader.
(Waddington 2006:67)
This approach is applied to the evaluation of student translations, so its

focus on translation as a product is obvious. Similar to what he did in
his papers of 2001 and 2004, Waddington tries to “verify the quality of
this method as compared to a more traditional one which fixes the pen-
alty for different types of mistakes a priori” (Waddington 2006:67). He
finds, though, that the former method fails to “[provide] more reliable or
valid results” (Waddington 2006:71) than the latter.
In 2006, the Leipzig International Conference on Translation &
I nterpretation Studies (LICTRA) had translation quality as its central
theme. The proceedings were published in the following year. Among
the many contributions, we will present a selection that reflects the wide
range of approaches to translation quality. Both Behr (2007) and Didaoui
(2007) investigate translation-related processes from a non-empirical
point of view: the former deals with the special requirements for the
translation of questionnaires and discusses team translation, back trans-
lation, and pretests as methods that can help to ascertain the quality of a
translated survey; the latter is mainly concerned with revision and “the
‘useful quality’ concept” (Didaoui 2007:79). This concept is about the
quality needed for a particular translation: within the framework of an
effective translation management, important texts should be translated
by experienced translators, whereas less important texts may be trans-
lated by less experienced translators.
Standardisation and terminology are two key aspects relevant to
the translation process. Budin (2007), Schmitz (2007), and Galinski
(2007) tackle them from different perspectives. Budin (2007) looks at
the development of international standards for translation quality and
locates translation quality within a four-dimensional context model in
which the four dimensions are interdependent: research and education;
professional practice and services; businesses, industries, and clients;
and standardisation and certification. Schmitz (2007) provides an
excellent overview of the various ways in which terminology manage-
ment contributes to translation quality. Galinski (2007), then, combines
the approaches of Budin and Schmitz in that he focusses on terminol-
ogy standardisation as an aspect of translation quality. From a more
practical point of view, Kurz (2007) describes translation quality man-
agement at SDL International. One very specific element of the q uality
management process is analysed by Gerstner (2007): she examines
proof-reading problems and suggests solutions for them. Since she pro-
vides many e xamples of translations, her approach is as much product-
oriented as it is process-oriented.
Kupsch-Losereit (2007) conceives of translation as a product. Theo-
rising about assessment parameters, she rejects the notion of equivalence
as insufficient when it comes to evaluating a target text and, instead,
favours cultural and discursive contextualisation and decontextualisa-
tion. On a more practical note, Kingscott (2007) recommends the use of
client specifications as a benchmark for the evaluation of translations.
He argues “that the basis for translation quality evaluation must be
Purpose-oriented Explicit or Implicit Specification (PEXIS)” (Kingscott
2007:322) and advocates evaluation on the basis of a metric scale rang-
ing from 0 to 9, since “using metrics forces revisers into being objective”
(Kingscott 2007:323). Equally non-empirical and with a similar focus
on the evaluation of the target text, the three articles by Hagemann
(2007), García Álvarez (2007), and Jüngst (2007) assume a decidedly
educational perspective: the first claims that, when assessing the quality
of a translation (particularly, in a translation studies context), the eval-
uator should try to also account for creative solutions rather than focus-
sing on errors and mistakes; the second demonstrates how a translation
commentary can play an important role in TQA; and the third proposes
ways to increase quality awareness among students of translation.
While most of the contributions to LICTRA 2006 can be categorised
as theoretical rather than empirical, there are some approaches that do
rely on test subjects or corpora. Martin (2007), for example, uses the
thinking-aloud method in a study with four student groups to elabo-
rate her error classification model. The thinking-aloud method is also
used by Göpferich (2007) in her reverbalisation experiment. Assisted by
TRANSLOG, a software that “records (logs) all keystrokes and mouse
clicks during writing processes as well as the time intervals between
them” (Göpferich 2007:211), she tries to determine the comprehensibil-
ity of a text. Interestingly, the experiment does not involve translation
or the evaluation of a translated text; rather, it requires that the test sub-
jects rewrite a technical German text on diabetes in such a way that it is
comprehensible for readers who do not have prior knowledge of the sub-
ject. The result shows “where the text is not optimally skopos-adequate”
(Göpferich 2007:219), skopos adequacy being the key to measuring the
translation quality of pragmatic texts. No test subjects feature in the
empirical research by Ivanova et al. (2007): they analyse two corpora
of Spanish and German employment contracts, using the text analysis
program Atlas.ti. As a result, they find out what collocations are most
appropriate in the given context. Tackling a rather specific text genre,
Gerisch/Bastian (2007) examine translation quality in European adver-
tising texts. While this sounds as if the focus were on translation as a
product, the authors stress the importance of process-governed quality
assurance. The results – based on pilot studies carried out between 2004
and 2006 – are used to provide prescriptive recommendations.
Not quite as specific as the analysis of advertisements is the descrip-
tive approach by Rodríguez Rodríguez (2007): her object of research
are literary texts, in particular, Rowland’s translation of Lazarillo de
Tormes. Drawing on criteria such as the type of text, coherence and co-
hesion, situation, purpose as stated by the translator, and acceptability
within the target system, Rodríguez Rodríguez aims at “a flexible
model of analysis and evaluation which can be limited and redefined a
posteriori […], accounting for the specific features of the analysed text”
(Rodríguez Rodríguez 2007:19). In his essay “Quality Management for
Translation”, Thelen (2009) looks at the overall translation process and
investigates the usefulness of “quality forms” (or “tools”) such as SAE
J2450 and LISA QA model 3.1. By way of example, here is one of several
insightful remarks on the subject:
[T]here is no guarantee whatsoever that […] the human being

perating the tool does this in an objective way: he may overlook
o
errors or interpret errors as non-errors or find errors not grave
enough to be classified as errors, or he may assign a perceived error
in the wrong error category, etc.
(Thelen 2009:204)
Unlike Thelen (2009), Jan Pedersen – in his essay of 2008 on quality

assessment in subtitling – focusses on translation as a product. Using
speech act and skopos theory as well as his concept of extralinguistic
cultural references (ECRs), Pedersen constructs a practical tool which
helps the evaluator of interlingual subtitles to analyse and assess a par-
ticular translation solution. ECRs often constitute a problem to the
translator or subtitler, who “has to make an active decision on how to
bridge a gap between two cultures” (Pedersen 2008:102). Such a gap
can be bridged by two different types of transfer strategies: minimum
change strategies and intervention strategies. Speech act theory comes
into play when analysing a film’s communication structure. However,
since “the traditional speech act scenario is too simplified” (Pedersen
2008:107), Pedersen uses a participant model that integrates all kinds of
audience of a particular utterance: the addressee of that utterance, side
participants and overhearers in the film, an intermediate studio audience
(often marked by canned laughter), and the final film or TV audiences
of the source and target cultures (cf. Pedersen 2008:110). This frame-
work helps the translator to establish the primary illocutionary point,
that is, the most important objective to be achieved by the sender of a
particular utterance. The framework also provides the basis for a hier-
archy of translation priorities, with the skopos of the utterance (that is,
the original sender’s primary illocutionary point) being paramount, fol-
lowed by the speaker’s primary illocutionary point, his or her secondary
illocutionary point, and other considerations (cf. Pedersen 2008:112).
Pedersen convincingly illustrates his approach with several examples
of Swedish and Danish subtitles taken from different films. He rightly
starts from the assumption that subtitles have to make sure that, as far
as this is possible, “the [target text] audience gets as much of the illocu-
tionary effect as the original audience” (Pedersen 2008:111). A transla-
tion is regarded as erroneous, if it “renders neither the primary nor the
secondary illocutionary point” (Pedersen 2008:112).
In 2009, Claudia V. Angelelli and Holly E. Jacobson published a
volume dedicated to the question of how translation quality can be
measured. In their introductory essay, they provide an outline of TQA
approaches, claiming that “none of the models of translation quality
presented thus far addresses the ‘how to’ of effectively and accurately
measuring quality” (Angelelli/Jacobson 2009:2–3). We will present
two approaches from their book. Conceiving of translation as “both a
process and a product” (Angelelli 2009:13), Claudia Angelelli devises
a rubric in which she specifies five translational features: source text
meaning, style and cohesion, situational appropriateness, grammar and
mechanics, and translation skill. Each of these rubric elements reflects
the translator’s ability at five different levels, from “inability” (or a
similar term) to “masterful ability” (or a similar expression). Whereas
A ngelelli’s working draft for a rubric is based on theoretical r easoning,
Eyckmans et al. (2009) carry out an empirical study in which they
compare three evaluation methods: a holistic, an analytical, and a cal-
ibration method. Sceptical of criterion-referenced approaches and the
use of assessment grids, Eyckmans et al. (2009) argue that these “fall
short in adequately reducing the subjectivity of the evaluation, since the
identification of dimensions of translation competence in itself is pre-
eminently subjective” (Eyckmans et al. 2009:74). What they prefer is a
norm-referenced approach. This is how it works: in a translation pretest,
so-called “items” are established, that is, those aspects that are partic-
ularly difficult to translate. For these “items” the evaluators agree on
translations that are acceptable or not acceptable (that is, correct or not).
The method is called the “Calibration of Dichotomous Items-method
(CDI-method)” (Eyckmans et al. 2009:76).
Less specific than the concepts presented in Angelelli’s and Jacobson’s
collection of essays is the account of quality in translation given by
Daniel Gouadec. Taking a theoretical perspective of translation as both
process and product, he provides an interesting framework for the more
specific approaches to translation quality. “The basic idea”, he writes
with quality standard EN 15038 in mind, “is that the quality of the
transaction is ‘good’ if and when both the provider and providee are
satisfied with the translation provision process and, of course, its result”
(Gouadec 2010:270). In this definition of good quality, the relevance of
the translation result looks like an afterthought. And, indeed, in trying
to convey the whole picture, Gouadec focusses more on the translation
process than on the translation product. He acknowledges the impact of
secondary aspects on translation quality when he claims that
good practices undoubtedly reduce the risk of poor quality but

do not suffice by themselves to guarantee quality for two reasons.
The first reason is that accidents happen even in the most highly
“quality-assured” environments. The second reason is that fulfilling
all of the quality requirements would make the cost of translation
unbearable to many work providers. No wonder quality assurance
nearly always comes second to economic considerations.
(Gouadec 2010:271–272)
Gouadec’s approach somewhat neglects the assessment of translation
quality
because the answers are quite simple. Once the grades have been
set and characterized, they may be used as a system for assessing
quality. In a professional setting, no one goes into intricacies: the
translator’s performance is “rotten/lousy”, “poor”, “satisfactory”,
“good” or “excellent” (for instance), and people and businesses have
any number of criteria to judge and justify their judgement […].
(Gouadec 2010:274)
We would argue, however, that, while setting the grades may still be
simple, defining the criteria and assessing a translation in the light of
these criteria is not at all straightforward. Gouadec stops short of ana-
lysing the quality of actual translated texts: he does not come to terms
with the very essence of translation quality assessment.
The year 2010 also saw the publication of … making the mirror
visible … by Miriam Acartürk-Höß, who analyses and compares six
different translations of W. H. Auden’s poem “If I Could Tell You”. She
starts from the assumption that translation criticism should first furnish
a detailed scientific description of the similarity relationship between
the source text and its translation – which may include an attempt to
trace and justify individual translation solutions. Only then can the
translation solutions be assessed (cf. Acartürk-Höß 2010:73). On the
basis of the principle of alterity, Acartürk-Höß develops a comprehen-
sive parameter grid covering both cotextual and contextual parameters.
The critic’s task is to define for a particular poem those parameters or
parameter combinations that he or she deems relevant to the most plau-
sible interpretation. If the location of the parameters in the parameter
grid are different for the original poem and for a translation, the result-
ing parameter shift can be analysed and assessed. Given the intricacies
of poetic translation and evaluation, it is no wonder that Acartürk-Höß
refrains from ranking the six renderings of Auden’s poem according to
their relative quality. Her rather specific approach is product-oriented
and non-empirical.
Near the opposite end of our spectrum of TQA approaches, we
find Ilse Depraetere’s collection of essays, Perspectives on Translation
Quality, published in 2011. In this volume, most contributions are based
on an empirical study. Part I is about translation quality in the transla-
tion training context. Depraetere/Vackier (2011), for example, investi-
gate the hypothesis that formal quality is indicative of overall quality.
They come to the conclusion that this hypothesis “should not be dis-
missed altogether” (Depraetere/Vackier 2011:49), since “there is a rather
strict correspondence between the number of formal and non-formal
errors” (ibid.). In their Spanish-French study of number and gender
agreement errors in student translations, Núñes-Lagos/Moulard (2011)
advocate the integration of error analysis into translation seminars.
Delizée (2011) also deals with student translations. What is interesting
is that she makes an effort to account for professional skills (in addition
to the usual translation-related skills), listing the following: the ability
to work rigorously, the ability to work in a timely fashion, the ability to
revise, the ability to communicate and present arguments, the ability to
self-evaluate, and the ability to cooperate. Quite a different approach is
presented by Gledhill (2011), who advocates a lexicogrammar approach
to checking quality. He asks a rather challenging question: “[B]etween
two potentially equivalent translations, is it possible to identify which
one is best?” (Gledhill 2011:71). To answer this question, he compares
two target text versions – both of which are acceptable – and e valuates
them on lexicogrammatical grounds verified by a detailed corpus analy-
sis. Part II, then, focusses on the evaluation of machine translation, with
studies by Depraetere (2011b) and De Sutter (2011). Part III is about
quality assurance in the translation workflow. Here, Debove et al. (2011)
perform a contrastive analysis of five automated QA tools – to men-
tion only one article. Part IV is concerned with domain-specific qual-
ity in legal translation (Bulcke/Héroguel 2011) and literary translation
(Vanwersch-Cot 2011). The latter provides a literary translator’s account
of how he works, emphasising that “[s]elf-assessment is […] one of the
essential tools of production” (Vanwersch-Cot 2011:262).
Next, our cursory discussion of TQA approaches features three very
different contributions to translation quality. From a rather specific
perspective, Bittner (2011) analyses the German subtitles in a chapter
of the 1974 mystery film Murder on the Orient Express, pinpointing
unsatisfactory translation solutions and making suggestions for im-
provement within the constraints imposed by subtitling. Vahid Dastjerdi
et al. (2011) propose a semiotic model of poetry translation assessment,
claiming that “it is essential not to consider meaning as something sta-
ble” (Vahid Dastjerdi et al. 2011:344). They apply their model to a poem
by Forough Farrokhzad, translated into English as “Another Birth”.
While the non-empirical approaches by Bittner (2011) and Vahid Dast-
jerdi et al. (2011) deal with translation as a product, Martin (2012) looks
at translation as a process, focussing on revision. In his well-argued es-
say, he makes a case for self-revision and contends that “a translator
begins the revision process with a better understanding of the document
than does the reviser” (Martin 2012, no pagination).
In 2014, Sylvia Reinart published a rather comprehensive account
of translation criticism: Lost in Translation (Criticism)? – Auf dem
Weg zu einer konstruktiven Übersetzungskritik. The book combines
theory and practice not in terms of a theoretical TQA framework and
its application but as a thorough discussion of translation quality and
translation criticism in which both theoretical and practical aspects are
duly emphasised. After an introductory chapter, Reinart gives an over-
view of the situations and occasions when translation criticism is needed.
This is followed by some general considerations relating to translation
culture and the consequences for translation criticism, before Chapters
4–6 furnish an interesting analysis of the various ways in which trans-
lations can be criticised, with a particular focus on text-typological
approaches. Reinart then asks the crucial question whether a new model
of translation criticism is needed. She finds that, on the whole, the exist-
ing models are too rigid because of their self-contained analysis systems.
Instead, there should be less restrictive catalogues of criteria (cf. Reinart
2014:85). In Chapter 8, Reinart proceeds to establish the criteria of
translation criticism by analysing and discussing different text types and
forms of translation including the various kinds of interpreting. Trans-
lation, here, is clearly seen as a product: what counts is the target text
in its textual dimension (cf. Reinart 2014:118). Throughout her book,
Reinart contrasts specialist and literary translations, stressing that the
latter cannot be satisfactorily assessed unless the vagaries of individual
interpretation are taken into account. Around 100 pages are devoted
to the discussion of special forms of translation, most of which depend
in some way on electronic media. They include audio description, the
translation of comics, subtitling, dubbing, voice-over, and interpreting
in the media. Quality criteria even cover technical issues, such as when,
in the process of translating an XML document, the XML tags are de-
leted, or when a simultaneous interpreter has to cope with insufficient or
defective technical equipment. Another special criterion are ethical con-
cerns faced by the translator, which may result in the ultimate question
whether or not to translate a particular text at all (cf. Reinart 2014:345–
346). This comprehensive account of the factors reflecting the potential
of translation criticism is followed by a chapter that probes the limits of
translation criticism. Towards the end of the book, Reinart returns to
the practical considerations involved in translation criticism by looking
at how the theoretical insights can be implemented in a highly varied
professional practice. The central message of this rewarding book is that
translation studies does not need just one additional model of transla-
tion criticism but a whole range of such models (cf. Reinart 2014:375).
One model that fits neatly into the range of TQA models outlined
in this chapter is presented in Anna Pavlova’s essay “Strategie der
Ü bersetzung und Beurteilung der Übersetzungsqualität” (Pavlova 2014).
This essay discusses the intricate relationship between strategic consid-
erations in translation and the quality of the target text. It emphasises
the complexity of the evaluation process (cf. Pavlova 2014:269) with a
particular focus on the options a translator has in dealing with diffi-
cult translation problems. Pavlova points out that there are situations
in which the translator can choose from several translation options and
situations in which the solution of the target text is enforced by objective
linguistic or cultural constraints (cf. Pavlova 2014:257). What is import-
ant, here, is that the critic should take the respective translation situa-
tion into account: rather than just comparing source and target texts,
the critic must consider whether there are any options available to the
translator or whether a seemingly unsatisfactory solution is, perhaps,
the only possible solution. Pavlova recommends that a translation be as-
sessed from two different angles: one mark should be given on the basis
of an objective comparison of the target text with the translation brief
or with any other explicit or implicit factors evident from the translator’s
strategy; another mark should be given on the basis of the translator’s
freedom of choice, that is, on the basis of the options the translator has
at his or her disposal (cf. Pavlova 2014:259). Despite this focus on the
translator’s perspective, the approach is product-oriented besides being
non-empirical.
With this we finish our overview of approaches to the quality of transla-

tion. In spite of Erich Steiner’s claim that “a translation evaluation is one
of the central goals of a theory of translation, and a central component
of a model of translation” (Steiner 2004:86), the work of translation
scholars such as Hans J. Vermeer or Christiane Nord has not been in-
cluded. This is because their main focus is not on the quality but on
other aspects of translation. Still, some of their works will find mention
in other chapters of this book. Many of the principles and insights fur-
nished by the scholars presented in this chapter will help us to establish
our own theory of translation quality and translation quality assessment
and, in this way, furnish the basis for a comparative study of evaluations
written to assess bachelor’s theses designed as commented translations.
Notes
1 This has also been noted by Lauscher (2000:155).
2 What exactly constitutes a straightforward or literal translation is, however,
difficult to determine, because it is practically impossible to give a general
definition of the degree to which a target text should adhere to the various
characteristics – stylistic and otherwise – of the source text. If grammatical
correctness were the only criterion, one would have to put up with a ren-
dering such as “Er spielte Klavier, Zeitung lesend” for “He was playing the
piano, reading the newspaper”, whereas a more common translation would
be syntactically less close to the original: “Er spielte Klavier und las dabei
Zeitung”. Also the tertium comparationis between a source text item and its
potential translations may be difficult or even impossible to establish, as the
scope of what a source text term signifies is often different from the scope
of what the corresponding target text term signifies. Compare, for instance,
the following remark made by Wilhelm von Humboldt in the introduction
to his Agamemnon translation, published in 1816:
Man hat schon öfter bemerkt, und die Untersuchung sowohl, als die Er-
fahrung bestätigen es, dass, so wie man von den Ausdrücken absieht,
die bloss körperliche Gegenstände bezeichnen, kein Wort Einer Sprache
vollkommen einem in einer andren Sprache gleich ist. Verschiedene
Sprachen sind in dieser Hinsicht nur ebensoviel Synonymieen; jede drückt
den Begriff etwas anders, mit dieser oder jener Nebenbestimmung, eine
Stufe höher oder tiefer auf der Leiter der Empfindungen aus.
(Humboldt 1969:80)
The ways in which words produce meaning through denotation and conno-
tation are too varied to permit of exactly the same meaning in any source
language expression and its target language equivalent.
3 Williams refers to the 1964 paperback edition of Toulmin’s The Uses of
Argument (Toulmin 1958).
4 While the actual argument structure may be clear, problems might arise
with different levels of the argument structure. Thus, if a target text fails to
capture the argument structure of the source text at the level of the para-
graph (thereby distorting the meaning of one particular paragraph), should
such a blunder be assessed as spoiling the whole translation? There has to
be a distinction between argument structures that do not affect the text as a
whole and argument structures that do. Such a distinction, however, would
not be as clear-cut as one might wish for and, as a result, detract from the
goal of objective assessment.
5 In Gerzymisch-Arbogast (2001:229), these methods are referred to as the
“itemized perspective”, the “relational pattern perspective”, and the “holis-
tic pattern perspective”, respectively.
6 The original discussion is slightly more detailed. It can be found in
Gerzymisch-A rbogast/Mudersbach (1998:162–164).
7 Hans Hönig strikes a similar note when he comments:
What the relevance-theory model, as applied by Gutt, fails to see, how-

ever, is the relevance of the ideas recipients have about “good” and “bad”
translation. Clearly, if a recipient has the idea that the quality of a trans-
lation can be assessed by back translation it will be futile to point out to
him or her that the translated text is “relevant” to him or her as it is since
s/he will not accept that it is a translation.
(Hönig 1998:24–25)
8 The distinction is between informative texts (focusing on content), expres-

sive texts (focusing on form), operative texts (focusing on the reader), and –
somewhat separately – audio-media texts, which depend on technical media
such as radio or television.
References
Acartürk-Höß, Miriam (2010). …making the mirror visible … Deutsche Über-
setzungen englischer Lyrik (W. H. Auden). Versuch einer Verwissenschaftli-
chung der Übersetzungskritik. Frankfurt/Main etc.: Peter Lang. (Beiträge zur
anglo-amerikanischen Literatur, vol. 7.)
Adab, Beverly (2001). “The Translation of Advertising: A Framework for Eval-
uation”. In: Babel, vol. 47, no. 2, 133–157.
Al Qinai, Jamal (2000). “Translation Quality Assessment. Strategies, Para-
metres [sic] and Procedures”. In: MetaTranslators’ Journal, vol. 45, no. 3,
497–519.
Al-Bustan, Suad Ahmed (1993). Quality And Efficiency Factors in Translation.
University of Glasgow: Thesis.
Angelelli, Claudia V. (2009). “Using a Rubric to Assess Translation Ability:
Defining the Construct”. In: Angelelli, Claudia V.; Holly E. Jacobson (eds.)
(2009:13–47).
Angelelli, Claudia V.; Holly E. Jacobson (2009). “Introduction: Testing and As-
sessment in Translation and Interpreting Studies: A call for dialogue between
research and practice.” In: Angelelli, Claudia V.; Holly E. Jacobson (eds.)
(2009:1–10).
and practice. Amsterdam/Philadelphia: John Benjamins. (American Transla-
tors Association Scholarly Monograph Series v. XIV.)
Baker, Mona; Gabriela Saldanha (eds.) (2009). Routledge Encyclopedia of
Translation Studies. 2nd ed. London and New York: Routledge.
Barghout, Mohamed Abdel-Maguid (1990). Translation Quality Assessment:
An Application of a Rhetorical Model. University of Salford: Thesis.
Behr, Dorothée (2007). “Übersetzung und Qualitätsüberprüfung in
vergleichender Umfrageforschung”. In: Schmitt, Peter A.; Heike E. Jüngst
(eds.) (2007:29–39).
Benhaddou, Mohamed (1991). Translation Quality Assessment: A Situational/
Textual Model for the Evaluation of Arabic/English Translations. University
of Salford: Thesis.
Bittner, Hansjörg (2011). “The Quality of Translation in Subtitling”. In: trans-
kom, vol. 4, no. 1, 76–87. Viewed: 12 August 2019. www.trans-kom.eu/
bd04nr01/trans-kom_04_01_04_Bittner_Quality.20110614.pdf
Bowker, Lynne (2001). “Towards a Methodology for a Corpus-Based Approach to
Translation Evaluation”. In: Meta Translators’ Journal, vol. 46, no. 2, 345–364.
Broeck, Raymond van den (1985). “Second Thoughts on Translation Criticism.
A Model of its Analytic Function”. In: Hermans, Theo (ed.) (1985:54–62).
Budin, Gerhard (2007). “Entwicklung internationaler Normen im Bereich der
Translationsqualität bei ISO/TC 37”. In: Schmitt, Peter A.; Heike E. Jüngst
(eds.) (2007:54–65).
Bulcke, Patricia Vanden; Armand Héroguel (2011). “Quality issues in the field
of legal translation”. In: Depraetere, Ilse (ed.) (2011:211–248).
Carroll, John B. (1966). “An Experiment in Evaluating the Quality of Trans-
lations”. In: Mechanical Translation and Computational Linguistics, vol. 9,
nos. 3 and 4, September and December 1966, 55–66.
Cary, Edmond; Rudolf Walter Jumpelt (eds.) (1963). Quality in Translation,
Proceedings of the 3rd Congress of the International Federation of T ranslators
(Bad Godesberg, 1959), New York: Macmillan/Pergamon Press.
Chiaro, Delia; Christine Heiss; Chiara Bucaria (eds.) (2008). Between Text and
Image: Updating Research in Screen Translation. Amsterdam/Philadelphia:
John Benjamins.
De Sutter, Nathalie (2011). “MT evaluation based on post-editing: a proposal”.
In: Depraetere, Ilse (ed.) (2011:125–143).
Debove, Antonia; Sabrina Furlan and Ilse Depraetere (2011). “A contrastive
analysis of five automated QA tools (QA Distiller 6.5.8, Xbench 2.8, ErrorSpy
5.0, SDL Trados 2007 QA Checker 2.0 and SDLX 2007 SP2 QA Check)”. In:
Depraetere, Ilse (ed.) (2011:161–192).
Delizée, Anne (2011). “A global rating scale for the summative assessment of
pragmatic translation at Master’s level: an attempt to combine academic and
professional criteria”. In: Depraetere, Ilse (ed.) (2011:9–23).
Depraetere, Ilse (2011b). “A contrastive analysis of MT evaluation techniques”.
Depraetere, Ilse (ed.) (2011). Perspectives on Translation Quality. Berlin and
Boston: de Gruyter Mouton.
Depraetere, Ilse; Thomas Vackier (2011). “Comparing formal translation evalu-
ation and meaning-oriented translation evaluation: or how QA tools can(not)
help”. In: Depraetere, Ilse (ed.) (2011:25–50).
Didaoui, Mohammed (2007). “Translation Quality: A Pragmatic and Mul-
tidirectional Approach.” In: Schmitt, Peter A.; Heike E. Jüngst (eds.)
(2007:79–90).
Dryden, John (1900). Essays of John Dryden, ed. W. P. Ker, vol. 1. Oxford:
Clarendon.
Eyckmans, June; Philippe Anckaert; Winibert Segers (2009). “The perks of
norm-referenced translation evaluation”. In: Angelelli, Claudia V.; Holly E.
Jacobson (eds.) (2009:73–93).
Fan, Shouyi (1990). “A Statistical Method for Translation Quality Assessment”.
In: Target, vol. 2, no. 1, 43–67.
Forstner, Martin (2007). “Paralipomena zur Diskussion über die Qualität von
Translationsstudiengängen unter dem Bologna-Regime”. In: Schmitt, Peter
A.; Heike E. Jüngst (eds.) (2007:124–157).
Fox, Brian (2009). “The Quest for Quality: Perceptions and Realities”. In:
Forstner, Martin et al. (eds.) (2009:23–27).
Galinski, Christian (2007). “Normung, insbesondere Terminologienormung,
als Einflussfaktor der Qualität von Übersetzungen”. In: Schmitt, Peter A.;
Heike E. Jüngst (eds.) (2007:158–174).
Gambier, Yves; Henrik Gottlieb (eds.) (2001). (Multi)media Translation: Con-
cepts, Practices, and Research. Amsterdam and Philadelphia: John Benjamins.
Gambier, Yves; Luc van Doorslaer (eds.) (2010). Handbook of Translation
Studies, vol. 1. Amsterdam/Philadelphia: John Benjamins.
García Álvarez, Ana María (2007). “Der translatorische Kommentar als Eva
luationsmodell studentischer Übersetzungsprozesse”. In: Schmitt, Peter A.;
Gerisch, Gordon; Bastian, Sabine (2007). “Übersetzungsqualität in europäischen
Werbetexten”. In: Schmitt, Peter A.; Heike E. Jüngst (eds.) (2007:185–200).
Gerstner, Johanna (2007). “Probleme beim Korrekturlesen: Praxisperspek-
tiven”. In: Schmitt, Peter A.; Heike E. Jüngst (eds.) (2007:201–209).
Gerzymisch-Arbogast, Heidrun (1994). Übersetzungswissensschaftliches
Propädeutikum. Tübingen. Basel: Francke.
Gerzymisch-Arbogast, Heidrun (2001). “Equivalence Parameters and Evalua-
tion”. In: Meta Translators’ Journal, vol. 46, no. 2, 227–242.
Gerzymisch-Arbogast, Heidrun; Klaus Mudersbach (1998). Methoden des wis-
senschaftlichen Übersetzens. Tübingen and Basel: Francke.
Gledhill, Christopher (2011). “A Lexicogrammar approach to checking quality:
looking at one or two cases of comparative translation”. In: Depraetere, Ilse
(ed.) (2011:71–97).
Göpferich, Susanne (2007). “Optimizing Reverbalization as a Target-group-
centered Empirical Method for Translation Quality Assessment”. In: Schmitt,
Peter A.; Heike E. Jüngst (eds.) (2007:210–221).
Gouadec, Daniel (2010). “Quality in translation”. In: Gambier, Yves; Luc van
Doorslaer (eds.) (2010:270–275).
Gummerus, Eivor; Paro, Catrine (2001). “Translation Quality. An Organiza-
tional Viewpoint”. In: Gambier, Yves; Henrik Gottlieb (eds.) (2001:133–142).
Gutt, Ernst-August (1991). Translation and Relevance: Cognition and Context.
Oxford and Cambridge: Basil Blackwell.
Hagemann, Susanne (2007). “Zur Evaluierung kreativer Übersetzungsleistun-
gen”. In: Schmitt, Peter A.; Heike E. Jüngst (eds.) (2007:237–255).
Halliday, M. A. K. (2001). “Towards a theory of good translation”. In: Steiner,
Erich; Colin Yallop (eds.) (2001:13–18).
Hermans, Theo (ed.) (1985). The Manipulation of Literature. London: Croom
Helm.
Holz-Mänttäri, Justa; Christiane Nord (eds.) (1993). Traducere Navem.
Festschrift für Katharina Reiß zum 70. Geburtstag. Tampere: Tampereen
Yliopisto.
Hönig, Hans G. (1998). “Positions, Power and Practice: Functionalist
Approaches and Translation Quality Assessment” In: Schäffner, Christina
(ed.) (1998:6–34).
Horton, David (1998). “Translation Assessment: Notes on the Interlingual
Transfer of an Advertising Text”. In: IRAL – International Review of Ap-
plied Linguistics in Language Teaching, vol. 36, no. 2, 95–119.
House, Juliane (1977). “A Model for Assessing Translation Quality”. In: Meta
Translators’ Journal, vol. 22, no. 2, 103–109.
House, Juliane (1981). A Model for Translation Quality Assessment. 2nd ed.
Tübingen: Narr (first published in 1977).
House, Juliane (1997). Translation Quality Assessment: A Model Revisited.
Tübingen: Narr.
House, Juliane (2001). “Translation Quality Assessment: Linguistic Descrip-
tion versus Social Evaluation”. In: Meta Translators’ Journal, vol. 46, no. 2,
243–257.
House, Juliane (2009). “Quality”. In: Baker, Mona; Gabriela Saldanha (eds.)
(2009:222–225).
House, Juliane (2014). “Translation Quality Assessment: Past and Present”. In:
House, Juliane (ed.) (2014:241–264).
House, Juliane (2015). Translation Quality Assessment: Past and Present. Lon-
don and New York: Routledge.
House, Juliane (ed.) (2014). Translation: A Multidisciplinary Approach. Hounds
mills: Palgrave Macmillan.
Humboldt, Wilhelm von (1969/1816). “Einleitung zu ‚Agamemnon‘” In: Störig,
H. J. (1969:71–96).
Ivanova, Vessela; Elke Krüger; Encarnación Tabares; Mirjam Reischert; Karin
Vilar Sánchez (2007). “Kontrastive Mikrofunktionsanalyse als Mittel zur
Verbesserung der Translationsqualität von Fachtexten”. In: Schmitt, Peter A.;
Jumpelt, R. W. (1963). “Methodological Approaches to Science Translation”.
In: Cary, Edmond; Rudolf Walter Jumpelt (eds.) (1963:267–281).
Jüngst, Heike Elisabeth (2007). “Übersetzungsunterricht als Erziehung zur Qua
litätskontrolle”. In: Schmitt, Peter A.; Heike E. Jüngst (eds.) (2007:289–294).
Kandler, Günther (1963). “On the Problem of Quality in Translation: B asic Con-
siderations” In: Cary, Edmond; Rudolf Walter Jumpelt (eds.) (1963:291–298).
Kingscott, Geoffrey (2007). “Translation quality assessment”. In: Schmitt, Pe-
ter A.; Heike E. Jüngst (eds.) (2007:317–325).
Kupsch-Losereit, Sigrid (2007). “Zur Evaluierung von Übersetzungen: P arameter
der Bewertung”. In: Schmitt, Peter A.; Heike E. Jüngst (eds.) (2007:333–343).
Kurz, Christopher (2007). “Translation Quality Management at SDL Interna-
tional”. In: Schmitt, Peter A.; Heike E. Jüngst (eds.) (2007:344–350).
Ladmiral, Jean-René (1993). “Sourciers et ciblistes”. In: Holz-Mänttäri, Justa;
Christiane Nord (eds.) (1993:287–300).
Larose, Robert (1998). “Méthodologie de l’évaluation des traductions”. In:
Meta: Translators’ Journal, vol. 4, no. 2, 163–186.
Lauscher, Susanne (2000). “Translation Quality Assessment: Where Can The-
ory and Practice Meet?”. In: The Translator, vol. 6, no. 2, Special Issue:
Evaluation and Translation, 149–168.
Lee-Jahnke, Hannelore (2001). “Aspects pédagogiques de l’évaluation des tra-
ductions”. In: Meta Translators’ Journal, vol. 46, no. 2, 258–271.
Luther, Martin (1969/1530). “Sendbrief vom Dolmetschen”. In: Störig, H. J.
(ed.) (1969:14–32).
Martín de León, Celia (2007). “Translationsqualität und Kosten-Nutzen-
Analyse”. In: Schmitt, Peter A.; Heike E. Jüngst (eds.) (2007:395–404).
Martin, Charles (2012). “The Dark Side of Translation Revision”. In: Transla-
tion Journal, vol. 16, no. 1, January 2012 [online]. Viewed: 12 August 2019.
http://translationjournal.net/journal/59editing.htm
Martin, Silke Anne (2007). “Modell zur Fehlerklassifikation in der Über-
setzung”. In: Schmitt, Peter A.; Heike E. Jüngst (eds.) (2007:405–416).
Melby, Alan K.; Alan D. Manning; Leticia Klemetz (2005). “Quality in trans-
lation: a lesson for the study of meaning”. In: Linguistics and the H uman
Sciences, vol. 1, no. 3, 403–446.
Mossop, Brian (2004). Book Review: Malcolm Williams. Translation Quality
Assessment: An Argumentation-Centred Approach. Ottawa: University of
Ottawa Press, 2004, 188 pages. In: TTR: Traduction, Terminologie, Rédac-
tion, vol. 17, no. 2, 185–190.
Muñoz Martín, Ricardo; Conde Ruano, Tomás (2007). “Effects of Se-
rial Translation Evaluation”. In: Schmitt, Peter A.; Heike E. Jüngst (eds.)
(2007:428–444).
Newmark, Peter (1991). About Translation. Clevedon, Philadelphia and Ade-
laide: Multilingual Matters.
Nida, Eugene A. (1964). Toward a Science of Translating. Leiden: Brill.
Núñes-Lagos, Carmen; Nathalie Moulard (2011). “Number and Gender
Agreement Errors in Student Translations from Spanish into French”. In:
Depraetere, Ilse (ed.) (2011:51–69).
Pavlova, Anna (2014). “Strategie der Übersetzung und Beurteilung der Über-
setzungsqualität”. In: trans-kom, vol. 7, no. 2, 256–271. Viewed: 12 Au-
gust 2019. www.trans-kom.eu/bd07nr02/trans-kom_07_02_05_Pavlova_
Uebersetzungsqualitaet.20141219.pdf
Pedersen, Jan (2008). “High felicity: A speech act approach to quality assess-
ment in subtitling”. In: Chiaro, Delia et al. (eds.) (2008:101–115).
Pinto, María (2001). “Quality Factors in Documentary Translation”. In: Meta
Reinart, Sylvia (2014). Lost in Translation (Criticism)? – Auf dem Weg
zu einer konstruktiven Übersetzungskritik. Berlin: Frank & Timme.
(Transkulturalität – Translation – Transfer, volume 5)
Reiß, Katharina (1971). Möglichkeiten und Grenzen der Übersetzungskritik.
München: Max Hueber (Hueber Hochschulreihe 12).
Rodrigues, Sara Viola (1996). “Translation Quality: a Housian Analysis”. In:
Meta Translators’ Journal, vol. 41, no. 2, 223–227.
Rodríguez Rodríguez, Beatriz Ma (2007). Literary Translation Quality Assess-
ment. München: Lincom Europa.
Rosenmund, Alain (2001). “Konstruktive Evaluation: Versuch eines Evaluation-
skonzepts für den Unterricht”. In: Meta Translators’ Journal, vol. 46, no. 2,
301–310.
gual Matters.
Schleiermacher, Friedrich (1813). “Ueber die verschiedenen Methoden des Ueb-
ersetzens”. In: Abhandlungen der Königlichen Akademie der Wissenschaften
in Berlin. Aus den Jahren 1812–1813. (1816). [Separately paginated sec-
tion:] Abhandlungen der philosophischen Klasse der Königlich-Preußischen
Akademie der Wissenschaften aus den Jahren 1812–1813. B erlin: Realschul-
Buchhandlung, 143–172.
Schmitt, Peter A.; Heike E. Jüngst (eds.) (2007). Translationsqualität. Frank-
furt/Main etc.: Peter Lang.
Schmitz, Klaus-Dirk (2007). “Translationsqualität durch Terminologiequalität –
wie und wo sollte Terminologiearbeit den Übersetzungsprozess unterstützen”.
In: Schmitt, Peter A.; Heike E. Jüngst (eds.) (2007:537–552).
Secară, Alina (2005). Translation Evaluation – a State of the Art S urvey.
Viewed: 12 August 2019. http://citeseerx.ist.psu.edu/viewdoc/summary?doi=
10.1.1.126.3654
Sharkas, Hala (2005). Genre and Translation Quality: Perspectives in Q uality
Assessment of English-Arabic Translations of Popular Science Genres.
University of Portsmouth: Thesis.
Steiner, Erich (2004). Translated Texts: Properties, Variants, Evaluations.
Frankfurt/Main etc.: Peter Lang.
Steiner, Erich; Colin Yallop (eds.) (2001). Exploring Translation and
M ultilingual Text Production: Beyond Content. Berlin and New York:
Mouton de Gruyter.
Störig, Hans Joachim (ed.) (1969). Das Problem des Übersetzens. Darmstadt:
Wissenschaftliche Buchgesellschaft.
Thelen, Marcel (2009). “Quality Management for Translation”. In: Forstner,
Martin et al. (eds.) (2009:195–212).
Toulmin, Stephen E. (1958). The Uses of Argument. Cambridge: Cambridge
University Press.
Tytler, Alexander Fraser (1907/1791). Essay on the Principles of Translation.
Everyman’s Library. Essays. London/New York: J. M. Dent & co./E. P.
Dutton & co. First Published in 1791, with a Second, Enlarged Edition and a
Third, Revised Edition Published in 1797 and 1813, Respectively.
Vahid Dastjerdi, Hossein; Yasamin Khosravani; Masoud Shokrollahi; Nasim
Mohiman (2011). “Translation Quality Assessment (TQA): A S emiotic Model
for Poetry Translation”. In: Lebende Sprachen, vol. 56, no. 2, 338–361.
Vanwersch-Cot, Olivier (2011). “The Problem of Self-Assessment in Literary
Translation”. In: Depraetere, Ilse (ed.) (2011:249–263).
Waddington, Christopher (2001). “Different Methods of Evaluating Student
Translations: The Question of Validity”. In: Meta Translators’ Journal,
vol. 46, no. 2, 311–325.
Waddington, Christopher (2004). “Should Student Translations Be Assessed
Holistically or through Error Analysis?”. In: Lebende Sprachen, vol. 49,
no. 1, 28–35.
Waddington, Christopher (2006). “Measuring the Effect of Errors on Transla-
tion Quality”. In: Lebende Sprachen, vol. 51, no. 2, 67–71.
Williams, Malcolm (1989). “The Assessment of Professional Translation Qual-
ity: Creating Credibility Out of Chaos”. In: Meta Translators’ Journal, vol. 2,
no. 2, 13–33.
Williams, Malcolm (2001). “The Application of Argumentation Theory to
Translation Quality Assessment”. In: Meta Translators’ Journal, vol. 46,
no. 2, 326–344.
Williams, Malcolm (2004). Translation Quality Assessment: An Argumentation-
Centred Approach. Ottawa: University of Ottawa Press.
Williams, Malcolm (2009). “Translation Quality Assessment”. In: Mutatis Mu-
tandis, vol. 2, no. 1, 3–23.
Zilahy, Simon Pietro (1963). “Quality in Translation”. In: Cary, Edmond;
Rudolf Walter Jumpelt (eds.) (1963:285–289).
3 Preliminary Assumptions
The discussion in Chapter 2 of a large variety of approaches to trans-

lation quality and translation quality assessment has revealed the
occasional problem issue. We have suggested possible solutions for some
of these issues. The solutions, though of a rather tentative nature, have
reflected our own view of TQA, namely, that translation quality assess-
ment is a highly intricate undertaking. In order to pave the way for a
theory that can successfully handle the multifarious elements involved
in the evaluation of a translated text, we will have to clarify four points.
First, the concepts of translation quality and translation quality assess-
ment have to be defined for our purpose as they can be regarded from
various perspectives. Second, a range of possible translation strategies
needs to be described in the context of the overall translation situation.
This description will be based on Juliane House’s distinction between
overt and covert translation. Third, the problem of subjectivity in TQA
requires special attention since translation quality assessment always in-
volves a human component. Finally, since the theory proposed here is
used to compare evaluations of commented translations in an academic
context – we will provide an overview of the framework within which
these evaluations can be analysed and assessed.
Defining Translation Quality and Translation

Quality Assessment
Quality is an elusive concept. Take, for instance, the following defini-
tion: “By quality, I mean that, according to which, certain things, are
said to be, what they are” (Aristotle 1853:26). Although the context of
Aristotelian logic would warrant a more sophisticated reading of that
quotation, we content ourselves with noting that quality is not predom-
inantly a matter of something being good or bad; rather, it is a matter
of something having other attributes than something else and thus be-
ing different in some respects. More than 2,300 years after Aristotle,
Tomoko Hamada writes that “the construction and interpretation of a
term such as ‘quality’ has a social existence that is guided by a person’s
habitus and is embedded in time and space” (Hamada 2000:295). This
Preliminary Assumptions 51
does not necessarily mean that the concept of quality is difficult to pin-
point; it does mean, however, that quality is subject to a large number
of contextual factors which have to be taken into account. Quality is a
relative concept. From a pragmatic point of view, quality may, for exam-
ple, be determined by the extent to which a product or service meets the
requirements for which it was produced or provided. What counts, in the
end, is client satisfaction (cf. Wiesmann 2007:637). In principle, this is
also true of translation quality.1
Yet, who are the clients of a translation? While, strictly speaking, the
client of a translation is the person or institution that has ordered the
translation, one may also regard the reader of the target text as a client.
As long as someone orders a translation just for himself or herself, there
is a good chance for the translator to be able to create a target text that
meets the expectations of the client. Considering the many different rea-
sons for which someone might need a translation, the minimum quality
requirement arguably consists in the fact that a given source text is trans-
lated at all. In such cases, even a machine translation – though fraught
with errors and mistakes – might do the job; whereas a perfect rendering
that is submitted after a stipulated deadline may be altogether useless
(cf. Ahrend 2006:33–34). If, however, the client is represented by a more
or less specific readership, then client satisfaction can only be predicated
on general assumptions about that readership. And if we accept the view
held by Reiß and Vermeer that a text is fully constituted only in reading
(cf. Reiß/Vermeer 1991:90), then to identify quality in terms of client
satisfaction becomes very difficult, indeed. Even if translation quality is
tested empirically (as has been done, for instance, by Carroll 1966), the
results can only show a detail of the whole picture. To shed some light
on the whole picture, this section closes in on translation quality from
several different angles before, finally, suggesting definitions for good
quality in translation as well as for translation quality assessment.
Some Thoughts on Measuring Quality

Our first glimpse of quality in translation reveals a fundamental dis-
tinction: from a purely descriptive point of view, quality reflects the way
in which a target text works; from a prescriptive point of view, quality
is that by which a good translation can be distinguished from a bad
one. These two quality concepts are interrelated, though, and, therefore,
easily mixed up. It is when quality is given a value that the prescriptive
mode takes over from the descriptive mode.2 The transition is usually
smooth, because the term “quality” often seems to imply some kind of
value judgement. As Emsel points out, gauging the quality of a trans-
lation appears in a sense contradictory since a good translation is un-
marked, that is, characterised by a zero value that indicates the absence
of errors and mistakes (cf. Emsel 2007:100). Quite true: it is only when
someone finds fault with a translation that the standard assumption of a
one-hundred-percent quality is reduced to a lower percentage. In other
words: the good quality of a translation is taken for granted (despite the
frequent examples of highly unsatisfactory renderings).
If translation quality is measured in terms of errors and mistakes,
the question is what constitutes an error or mistake. In theory, it is
clear that judging the quality of a translation “is ultimately a matter
of agreement and consensus” (Lauscher 2000:149). In practice, how-
ever, a set of criteria is needed to distinguish between “acceptable”
and “unacceptable” both for individual translation solutions and for
the target text as a whole. This set of criteria should account for trans-
lation in context, since translating “is much more than a mere linguis-
tic code-switching operation” (Nord 2005:180). The evaluator has to
take a close look at the overall translation situation and, in particu-
lar, at the specifications of the translation job (cf. Bühler 2000:368),
since quality requirements are variable, and one and the same target
text can be acceptable in one translation situation and unacceptable in
another. When Ilse Depraetere writes in the introduction to her book
Perspectives on Translation Quality, “There are three issues that are
important when it comes to translation: quality, quality, quality”
(Depraetere 2011a:1), she seems to be exaggerating; however, this ex-
aggeration is put into perspective when we consider that “quality” is a
relative concept and that Part II of her book deals with the evaluation
of machine translation.
The point is that for translation quality to be convincing, it does
not have to be maximum quality but should be tailor-made to suit the
needs of the client or reader (cf. Schubert 2007:14). Hans Hönig writes
that the translation of a text is acceptable if it works – that is to say
for those who can demonstrate a justifiable interest in the use of the
translated text (cf. Hönig 1995:74). Quality, here, becomes a matter
of acceptability. While this is certainly a practicable view, it does pose
a few problems. First of all, the question, what is acceptable or useful
and what is not, remains a moot question: a text that is comprehensible
despite a large number of obvious mistakes – would it be acceptable
or not? Second, a lot depends on text type: what counts as a major
defect in the translation of a novel (for example, stylistic incongruities)
might be seen as rather irrelevant in the translation of a user manual.
Whatever the solution in such cases, the key point in Hönig’s statement
is that the quality issue is decided by the reader or client (or whoever
has a justifiable interest in the use of the translated text). If an external
evaluator is to assess a translation, he or she will have to take the read-
er’s or client’s interest into account.
If usefulness is the overarching goal of translation, how can this goal
be achieved in the translation process? Although there is a lot to be said
about optimising the translation process, we will just broach one or two
issues. The following quotation from Jiri Stejskal may serve as a starting
point:
In judging the quality of Edith Grossman’s translation of Don Quix-

ote, no sensible reader will demand that Grossman be a certified
translator, that she follow a standard defining a quality translation
process, and that the novel satisfy the Society of Automotive Engi-
neers’ Translation Quality Metric.
(Stejskal 2009:291)
In other words, one cannot standardise for all kinds of translation the
conditions under which a translator should produce a high-quality tar-
get text. What might be a desirable requirement in technical transla-
tion (for instance, that the translation be produced using specialised
software)3 may be absolutely irrelevant in literary translation. The con-
ditions of the translation process as well as the competence of the indi-
vidual translator and his or her attitude to work – all these aspects have
a share in the quality of the target text. Foregrounding one and neglect-
ing the others may jeopardise the successful completion of the transla-
tion process. There is no criterion – such as a qualified translator or a
translation agency certified to EN 15038 – that alone would suffice to
guarantee good translation quality (cf. Sinner/Morales Tejada 2015:112;
Kurz 2009:149). By way of example, we quote Allison Beeby-Lonsdale,
who notes that “the assumption that native speaker equals quality still
prevails” (Beeby-Lonsdale 2009:86). Any translator trainer would agree
that, while native speakers of the target language are more likely to pro-
duce a good translation than non-native speakers, there are also many
instances of bad translations by native speakers. Translation quality is
too complex to be reduced to a native/non-native dichotomy.
Yet, despite its irreducible complexity, translation quality is not un-
amenable to rules that govern the way in which the quality of a target
text can be established. Since not all the factors or parameters determin-
ing translation quality are known to the evaluator, their number can be
related to the outcome of the evaluation. The more quality parameters
are known to the evaluator, the more convincing can be the evaluation of
the target text. At the same time, though, the process of establishing the
quality of the translation becomes more complex and, thus, more unpre-
dictable, because some parameter values and their interplay may be sub-
ject to variation. If, however, only few quality parameters are known to
the evaluator, translation quality assessment will be easier but less con-
vincing. To illustrate the effect that lack of knowledge can have on the
assessment of a translation, let us look at two fictitious situations. In the
first, a proofreader complains of substandard terminology in the trans-
lated text, not knowing that the terms in question have been provided
by the client. Changing the terminology might render the document
incompatible with other related documents from the same client. The
second situation relates to the translation of literature. Here, judging a
literary translation without having read the original seems to be com-
mon practice (cf. Granzin 2010). This practice may lead to unwarranted
criticism of stylistic stumbling blocks in the target text, as the critic is
unaware that the source text contains similar idiosyncrasies.
Some Thoughts on Achieving Good Quality

At this point, it is appropriate to remind ourselves that quality can be
viewed from two different perspectives: quality can be that which merely
distinguishes one thing from another, or quality can imply a value
judgement. In literary translation – and in the translation of poetry, in
particular – some quality parameters are so much subject to the prefer-
ences of the individual evaluator that any quality judgement claiming
one translation to be better than another is valid only on the basis of
specific assumptions. This is why Acartürk-Höß is rather reluctant to
criticise individual translations of W. H. Auden’s poem “If I Could Tell
You” – her analysis of Radegundis Stolze’s first (not very convincing)
translation remains remarkably non-judgemental (cf. Acartürk-Höß
2010:272–273). Whenever poetic translation calls for an artistic render-
ing, a quality statement which goes beyond a mere description of what
the translator has done can only rely on the same fleeting criteria that are
drawn upon for the evaluation of any work of art.
By contrast, a poem may also be translated for a purpose other than
that of achieving some high-flown verbal effluence in the target language.
Thomas Hardy’s poem “Neutral Tones”, for instance, has been set to
music by British composer Nicholas Maw. A performance of the work
with the tenor Philip Langridge and guitarist Stephen Marchionda was
recorded and published on audio CD by Chandos together with other
compositions. The CD cover contains the original poem and translations
into French and German (cf. Britten et al. 2005). It is obvious that the
purpose of these translations is not for the reader to indulge in poetic
finesse, but for the listener to get an idea what the singer is singing. The
quality of the German version of Hardy’s poem, translated by Andreas
Klatt, would probably be regarded as not particularly good if viewed
from Parnassian heights4; from the pragmatic perspective of the CD lis-
tener, however, the quality of the translation is sufficient.
Whether in literary or technical translation: good quality can often be
achieved in several ways. Peter Newmark, for example, claims that “ten
different versions of the same text may be equally acceptable” (Newmark
1981:129). This is certainly true considering that even a textual frag-
ment such as “Let’s dance, said Vivien, looking at my husband […]”
translated by 33 students, yielded 27 different German versions (cf.
Wilss 1982:224–225), all of which might be regarded as acceptable.
How many of them would be taken to be equally acceptable depends on
the degree of subtlety with which the passage is analysed. Wilss him-
self seems to be in favour of the first three versions, which he suggests
“would probably come out top” (Wilss 1982:225) in terms of equiva-
lence. Yet, he gives no particular reasons and refrains from discussing
the quality of the 27 renderings in more detail. Such a discussion would
require a very sophisticated TQA model, indeed.
As we look at a target text from a qualitative point of view, are we ac-
tually evaluating the translation or the translator? Surely, in many cases
we are evaluating both since it seems obvious that the competence of the
translator shows in the quality of his or her translation. In translator
training, the mark given for a particular achievement reflects the trans-
lator’s ability to translate rather than the suitability of the translation for
the purpose for which it was done. This is because translator training
has education as its foremost objective. If a student passes a translation
test, it does not necessarily mean that his or her translation would be ac-
cepted in a real-world scenario. However, judging the translator on the
basis of a translation may also be problematic, especially, if more than
one translator is involved in a translation project or if the client’s specifi-
cations are detrimental to the quality of the target text or if proofreaders
and/or editors tamper with the target text. The competence of the trans-
lator (or translators) plays an important role whenever the focus is on
the translation process. In larger projects, good translation skills have to
be complemented by a variety of other quality factors. As Eivor Gum-
merus and Catrine Paro write with a screen translation context in mind:
“Translation quality is […] not only a matter of the competence of the
individual translator, but a result of good co-operation between all the
people involved in the production process” (Gummerus/Paro 2001:139).
The quality of a translation crafted as the result of a complex production
process reflects not only the competence of the translator(s) but also the
more or less smooth interplay of all other factors involved.
For our purposes, we will look for translation quality only in the tar-
get text because it is here that the quality of a translation manifests itself.
Translator competence is but one factor out of many that has an influ-
ence on translation quality. Certification as a qualified translator may be
accepted as proof of the ability to translate well; yet, such certification
is of no use unless it shows in the translated text. Given that quality
is an elusive concept and that quality is relative (as we have pointed
out at the beginning of this section), any definition of good quality in
translation has to account for different opinions. Marcel Thelen hardly
exaggerates when he claims that “there are perhaps as many ideas about
what is good quality translation as there are persons involved in trans-
lation training” (Thelen 2009:204). This immanent subjectivity is also
reflected in the following definition: good quality in translation is the
perception of a translation as most appropriate within the context in
which it functions.5 The vagaries of such perception have to be captured
by a suitable TQA system.
Good quality presupposes an effort to achieve good quality, or in the
words of Brian Fox: “The essential ingredient of quality is the desire to
improve” (Fox 2009:24). In a translation context, that means: “A good
translator will […] strive at the highest possible degree of perfection in
conformity with the situation” (Kandler 1963:296). In other words, a
good translator will do his or her best. Notwithstanding the translator’s
keen commitment, “the effort to create a ‘perfect translation’ is wasted
effort” (Al-Bustan 1993:92); simply, because there is no such thing as a
perfect translation, no one best target text for any given source text –
from a pragmatic as well as from a theoretical point of view. While
there may be sample translations in translator training that purport to
represent something like the best translation, a translation whose qual-
ity the students should aim to achieve, there can be no doubt that such a
“best translation” merely reflects the ability of the translator trainer and
that another translator trainer would have come up with another sample
translation.
In spite of this seemingly frustrating perspective, we intend to pro-
pose a TQA scheme that does make it worth one’s while to aim for
the unattainable. If the goal of producing the best translation cannot
be reached, we should be content with striving for a better translation.
And if we get ever better translations, then the result will ultimately
be an approximation to the best translation. All we need is a tool for
the evaluator to decide which of two translations is better. Such a tool
would have to account for the fact that the number of factors relating to
the assessment of a translation is indefinite and that different evaluators
have different preferences when judging the quality of a translation. In
applying the tool, an evaluator would try to gauge the impact of as many
quality factors as possible in order to reach a quality judgement that has
the potential to convince other evaluators. The assessment system is, in
fact, open to other ideas and opinions so that several evaluators may
work on the evaluation of one translated text. With this in mind, we
can now propose a tentative definition of translation quality assessment.
For our purpose, translation quality assessment is an iterative process in
which one or more evaluators draw upon as many factors as they deem
relevant to the quality of a particular translation in order to weigh in-
dividual translation decisions at micro- and macrotextual levels against
potentially better solutions and, thus, to be able to judge and improve
the quality of the target text.
Revisiting Overt and Covert Translation

One central theoretical element in Juliane House’s theory of trans-
lation quality assessment is the distinction between overt and covert
translation. While overt translation enables the target text reader to ex-
perience the source text in its linguistic and cultural dimension from
a target language and target culture point of view (secondary level
function), covert translation hides the source text in that it presents the
target text as if it were an original in the target culture (primary level
function). Overt translation – being “similar to a citation or quotation”
(House 1997:112) – can, therefore, be regarded as “a case of ‘language
mention’” (ibid.), where the source text is left “as intact as possible”
(House 1997:68). Covert translation, by contrast, is “a case of ‘language
use’” (House 2001:250), where “the translator must attempt to re-create
an equivalent speech event” (ibid.) and try to achieve “true functional
equivalence” (ibid.). With regard to the practice of overt and covert
translation, this means that
in an overt case, the translator has to make, as it were, as few changes

as possible, and will be held openly accountable for the degree of
success with which this is achieved; in a covert translation, the
translator is implicitly licensed to make as many substantive changes
as necessary, and in fact is only likely to be “caught out” and held
accountable in the case that not enough change was wrought, such
that the resultant text is in fact perceived to be a translation.
(House 1997:164)
In other words: overt translation is generally source-oriented, whereas

covert translation tends to be target-oriented. It will be demonstrated,
though, that this conclusion does not necessarily follow from the defini-
tion of overt–covert translation.6
House provides a useful overview of how the two modes of transla-
tion can be distinguished (cf. House 1997:115). She relates the principles
of overt and covert translation to the functional and linguistic levels
of the target text, answering the question whether strict equivalence is
the translational goal. While the genre is preserved in both translation
modes, the levels of register and language/text may be changed in a covert
but not in an overt translation. A change of genre – for example, render-
ing an English novel into a German play – would no longer be regarded
as a translation but as a version. However, given that the overt–covert
distinction “is a cline, not an ‘either-or’ dichotomy” (House 1997:30),
the overview described above fails to do justice to the many intermediate
forms of translation. In this section, we will develop a more refined dis-
tinction between overt and covert modes of translation.
Before setting out to refine the overt–covert translation model, we need
to briefly go back to Juliane House’s notion of a version as opposed to a
translation. As has been shown in Chapter 2, an overt version violates
an otherwise overt translation at the language/text and/or register levels
to accommodate a specific audience such as children or adolescents; a
covert version is inadequate as a covert translation because “the appli-
cation of the cultural filter is unjustified” (House 1997:73). The distinc-
tion between a version and a translation, while perfectly clear from a
theoretical point of view, may pose a few problems in practice, because
the threshold from what would still be regarded as a translation to what
would be considered a version is difficult to define. This is particularly
obvious in covert mode. Juliane House herself strongly cautions against
“making quick decisions about what is a covert version and what is
a translation” (House 1997:117), as “it may be difficult in practice to
make an unambiguous judgment” (ibid.). For this reason, we will not
adopt this distinction but accept versions as translations if they are pre-
sented as such.7
To what extent a source text is to be translated overtly or covertly has
to be defined in the translation strategy. That is to say, the translation
strategy specifies the way in which certain elements of the source text
are rendered in the target text. It is at this point that a clear distinction
should be made between the overt–covert principle and the implications
of that principle as presented by Juliane House: while the overt–covert
principle merely refers to the essential dichotomy between a target text’s
primary level function or secondary level function (that is, between a
target text conceived of as an original text and a target text regarded
as the translation of an original, respectively), the implications of that
overt–covert principle relate to the way in which genre, register, and lan-
guage/text are handled in translation. Unfortunately, the above distinc-
tion between the overt–covert principle and that principle’s implications
is often blurred, as the following example shows.
Having thoroughly analysed Britta Groiß’s German translation of Jill
Murphy’s children’s book Five Minutes’ Peace, House concludes:
The translation can […] be described as a covert one with a cultural

filter having been applied. One wonders, however, if the translator’s
choice might not have been different, i.e., why did she not opt for
an overt translation? 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? (In my corpus of 52 children’s books I
found that all the translations examined are covert translations.) Is
it possible that children in their intelligent and imaginative capac-
ities to learn and be exposed to the strange world of the original
are largely underrated? Why is there not a greater respect for the
original children’s book – especially if the original is a little literary
masterpiece?
(House 1997:131)
What is interesting, here, is that Juliane House regards the translation

as covert, although the overall frame of the translated text should be
overt – at least, if the book mentions the translator’s name on the title
page. In that case, the reader is aware that he or she is reading a trans-
lated text, and awareness of that fact is a key feature of overt translation.
However, the mere visibility of the translator in the target text and, thus,
of the translation qua translation is but a superficial indicator of overt-
ness that ought to be supported by further elements regarded as typical
of overt translation in a particular culture.
Refining the Concept of Overt-Covert Translation: Overview

In using the overt–covert framework to define the strategy for an individ-
ual translation, we need to be constantly aware of the difference between
the immediate significance of overtness and covertness in translation, on
the one hand, and any conclusions drawn from an overt or covert mode
of translating, on the other. The immediate significance of overtness in
translation consists in the fact that, here, the target text constitutes an
explicit reference to the source text; the immediate significance of co-
vertness in translation consists in the fact that, here, the target text tries
to conceal the presence of a source text and seems to exist independently
of it. That an overt translation should, therefore, stick as closely as pos-
sible to the source text and a covert translation depart from the source
text, if necessary, might, however, be too hasty a conclusion. For differ-
ent cultures will invariably look at translation from different perspec-
tives: while the overt–covert principle can be applied to any translation
in any culture (even though there may well be cases in which it is not
clear whether or not a translation is, or should be, recognised as such),
the implications – source orientation in overt and target orientation in
covert translation – are not universally applicable. In some cultures, the
distinction between overt and covert translation may not have any effect
on the way in which a source text is rendered into a target text; in other
cultures, the effect may be less pronounced or altogether different. For
the purpose of refining the above concept of overt and covert transla-
tion, we will assume the same English-German perspective that Juliane
House takes in developing her idea. Thus, we consider it appropriate in
overt translation to imitate the style of the original within the limits set
by the various rules of the target language, and in covert translation to
deviate from the original where that is necessary in order to come up
with a target text that best fits into the target culture. In short, overtness
in translation is, here, regarded as synonymous with target orientation,
covertness as synonymous with source orientation.
In addition to the above distinction between the overt–covert p rinciple
and the implications of that principle, there is another crucial distinction
to be made when it comes to refining Juliane House’s theory of overt vs.
covert translation. This distinction concerns the question as to when
the various factors contributing to the decision in favour of an overt or
c overt translation become relevant. If we conceive of the overt–covert
concept as a means to enable the translator to define his or her transla-
tion strategy for a particular translation, we can distinguish two phases
in the translation process: first, the relevance of overt–covert criteria
before the definition of the translation strategy; second, the relevance
of overt–covert criteria after the definition of that strategy. While some
criteria will be crucial to the definition of the translation strategy, others
may unfold their significance only after that strategy has been defined.
Usually, those criteria that help to define the translation strategy will af-
terwards be relevant only insofar as their use needs to be consistent with
the chosen mode of translation. In establishing a translation strategy, we
draw upon both the overt–covert principle and the implications of that
principle: while the principle itself is most relevant for defining a strategy
at the macrotextual level, the implications of the principle play a role
mostly in determining the translational approach at the microtextual
level following the definition of the overall strategy.8
The above distinctions – namely, between overt or covert translation
and what this implies as well as between elements that are relevant be-
fore and elements that are relevant after the translation strategy has
been defined – is depicted in Figure 3.1. While the overt–covert principle
and its implications may easily be distinguished from each other, the
distinction between what is relevant before and what is relevant after
defining the translation strategy is less clear-cut; for to define a transla-
tion strategy, the findings derived from the overt–covert principle will
often be complemented by findings derived from the implications of that
principle – as will be shown in the discussion below.
before defining
before defining
the translation
the translation
the fundamental
translation visible, translation hidden,
strategy
strategy
principle of
reference to the no reference to the
overt and covert
source text source text
translation
overt covert
translation translation
after defining the
after defining the
stylistic the implications stylistic

translation
translation
strategy
consequences of of the principle consequences of

strategy
overt translation: of overt and covert translation:

usually source- covert usually target-
oriented style translation oriented style
Figure 3.1 Overt–covert overview (general aspects).

Detailed Discussion
The factors that bear upon translation strategy within the overt–covert
framework can be roughly divided into four categories: form, content,
audience, and the author of the source text. Form is concerned with text
type or genre,9 overall text structure and layout, style (that is, the way
in which language is employed in the source text, which includes syntax,
lexis, grammar, register, rhyme, metre, tone, voice etc.); content relates
to semantic considerations, the text topic and text function; the audience
reflects, to some extent, the purpose of the translation and defines the
cultural distance between source and target; the source text author is
relevant in that his or her status in the source and target cultures influ-
ences the significance of form and content elements. These factors can
affect the translation strategy in two ways: in the source text, they help
the translator to get his or her bearings in the overt–covert spectrum; in
the target text, they reflect the relative overt- or covertness of the transla-
tor’s response as well as the consistency of that response. Any other fac-
tors such as censorship or overriding instructions given by the client may
have an additional impact on translation strategy: they will mostly affect
the form and/or content elements within the overt–covert framework.
If we look at the four factor categories in more detail, we notice that
form is the most intricate and specific, whereas content, audience, and
author operate at a more global level. A translator – in developing his
or her strategy for the translation of a particular text – would first con-
sider the genre of that text. In most cases, genre will decide the overall
strategic framework of a translation whereas style and, perhaps, au-
thor will either strengthen or weaken the degree of overtness or covert-
ness within that framework. If genre fails to provide the decisive clue,
then style is usually the most important criterion. Content may play a
subsidiary role in defining the translation strategy. Audience is crucial
only in special cases, when the translation brief specifies a target text
audience completely different from the source text audience: here, the
style of the source text will have to be changed in the target text – the
translation is then stylistically covert. Mostly, however, audience will
come into play after the overall translation strategy has been defined.
The same holds true for the overall text structure and layout, which
may have to be adapted in a covert setting. An example would be the
translation of a manual whose overall text structure or layout does not
meet the requirements of the target culture. What has to be kept in
mind throughout the present discussion is that any rules and recom-
mendations put forward for one specific factor or factor category are
necessarily complemented (and possibly limited or overridden) by the
rules and recommendations put forward for the other factors or factor
categories.
Whether to translate a text overtly or covertly is usually decided on
the basis of form. Thus, the text type or, for that matter, the genre of
the source text signals to the translator which strategy he or she should
pursue as the overall translation strategy. Literary texts, legal and his-
torical documents, for example, should be translated overtly; a covert
translation is typical of user manuals, cookery books, travel guides, and
the like. Somewhere in the middle on the cline between overt and covert
would be journalistic texts. Characterised as they often are by a style
that combines the functional and the elegant, the operative and the so-
phisticated, reader-orientation and author-showiness, journalistic texts
have to be closely analysed with regard to their individual merits and
demerits – linguistic and otherwise – before a translation strategy can be
defined. It is here that content may become relevant: a text about some
real-life things and facts typical of the source culture would be trans-
lated within an overt framework, because the target text is expected to
provide a glimpse of the source culture; a text about some aspects of the
human condition in general (in which cultural idiosyncrasies are more
or less irrelevant) would lend itself to a covert translation. Thus, in the
most overt approach, a source culture element little known in the target
culture would be translated using the original term or an appropriate
translation without providing additional explanatory information. If
such additional information is given, the retention of the source culture
element is regarded as overt, while the explanation of that element is
considered covert. The most covert approach would be to replace the
source culture element in the translated text by a corresponding target
culture element. Although a general strategy indicates the direction into
which a translator should push his or her translation, the decisions at the
microtextual level are largely determined by the form factors operating
at that level.
If we look at how the various aspects of form are implemented across
the overt–covert spectrum, the most conspicuous linguistic renderings
will be found at the overt end. An interlinear translation, for instance,
mirrors the grammatical and syntactic structures of the source text,
paying little or no attention to the grammatical and syntactic rules of
the target language. As the original and translation are presented line by
line one above the other – with the reader usually studying the source
text in the upper line with the help of the word-by-word rendering in the
lower – the function of the translated text could not be more obvious.
Indeed, the term “target text” might be rather inappropriate here, since
it suggests a relative independence that is not achieved in interlinear
translation. However, linguistic idiosyncrasies may also occur in ordi-
nary overt translation, depending on the preferences of the translator.
Thus, a source-text-oriented translation scholar such as Peter Newmark
claims, “A language such as English would gain by the literal translation
of many foreign key-words, idioms and possibly even proverbs, most of
them not particularly culture-bound” (Newmark 1991:35). In a trans-
lation from German into English, he suggests that “Geburtstagskind”
be rendered as “birthday child”, “er ist nicht mehr der Jüngste” as “he’s
no longer the youngest”, and “guten Appetit” as “good appetite” (ibid.).
Usually, these expressions would – even in an overt translation context –
be translated as “birthday boy/girl”, “he’s getting on a bit”, and “enjoy
your meal”, respectively.
An overt translation may well contain a few covert features. Some-
times, as in the above examples, these might be preferable to the corre-
sponding overt alternatives. In other cases, a covert rendering may be
deplored, as we have seen in Juliane House’s analysis of Britta Groiß’s
German translation of Jill Murphy’s children’s book Five Minutes’
Peace. A typical overt translation would imitate the style of the source
text in the target text to the extent that idiosyncratic source text features
become idiosyncratic target text features. Stylistic normalisation in the
target text is kept to a minimum – however, what constitutes that min-
imum is for the translator to decide. Similarly, the syntax of the source
text is emulated in the target text in terms of theme-rheme structures
and syntactic patterns inasmuch as these are compatible with the syntax
of the target language. Again, in the last analysis, what is compatible
and what is not has to be decided by the translator. Any such decisions
should be based on sound reasoning rather than mere gut feeling.
Moving on from overt to covert translation, we can say that, here, the
target text is streamlined in the sense that any features reminiscent of the
source language are brought in line with target language rules and re-
quirements. While this streamlining procedure is fairly straightforward
at the level of grammar, it may be less obvious for syntactic, semantic,
or stylistic issues. According to Juliane House, even a covert render-
ing has to respect the “double-binding relationship” (House 1997:12)
of all translation10; the translator, in trying to cover up the origin of
the target text, is not free to do as he or she pleases. Normalisation,
for instance, should be based on the application of a cultural filter that
has been empirically tested. An interesting example of normalisation is
given by Lynne Bowker, who evaluates student translations of the French
sentence “Il [DVD] va envoyer au tapis CD-Rom, CD-Audio et VHS”
(Bowker 2001:355) with the help of corpus analysis:
In the student translations, a number of students translated envoyer

au tapis by ‘replace’ and some others selected ‘overtake’. These trans-
lations are examples of normalization – the author’s creativity has
been neutralized. The basic semantic meaning has been captured,
but the translation lacks the flair and zeal of the original. An exam-
ination of the Quantity Corpus […] reveals that four texts do indeed
use relatively straightforward terms such as ‘replace’ or ‘outsell’ to
describe the relationship between DVD and CDs, but other texts
contain much more interesting and forceful descriptions, including
‘attack’, ‘challenge’, ‘invade’, ‘make obsolete’, ‘push aside’, ‘put the
squeeze on’ and ‘sound a death knell for’. Placed alongside these
texts, a translation such as ‘replace’ would cause this passage, which
in the source text stood out as being forceful and vibrant, to pale
into insignificance. These concrete examples can be used to demon-
strate to students that un-normalized translations are possible, and
in this case preferable, given the skopos of the target text.
(Bowker 2001:355)
In covert mode, a normalised translation is acceptable; yet, it is not nec-

essarily the best solution. A rendering that goes beyond normalisation
or explicitation – for example, by changing the source text’s genre or
any substantial part of its form or content – results in the most radically
covert form of translation, one that may not be generally recognised as
a “translation”.
Having discussed the most essential criteria for the definition of a
translation strategy, we will now turn to three less obvious factors: au-
thor, audience, and content. The author’s status is important, because it
rubs off on the source text: a Charles Dickens or Henry James – known
as they are for their characteristic writing styles – would generally com-
mand more respect from the translator than some little-known writer.11
What a Dickens or James wrote would be regarded as “sacrosanct”
(House 1997:76) and any tampering with the most typical source text
features condemned as sacrilege. The more famous or historically sig-
nificant an author is, the more overt should normally be the translation
strategy for his or her works. This does not mean that the debut novel
of some (as yet) unknown writer would be translated covertly; it means
that the work of a lesser-known writer is more likely to be subjected to
a few covert solutions within an otherwise overt strategy than the work
of a well-known author. Even a stylistically inconspicuous newspaper
comment written by a famous contemporary writer would probably be
handled overtly in translation on account of the author’s standing. The
role of the source text author may, then, become particularly relevant
to defining the translation strategy, if his or her exceptional personality
(calling for an overt rendering) is joined by a genre and/or style that
would otherwise be handled covertly.
In translation, the audience helps to define the sociocultural differ-
ences between the source text and the target text, and it is the reader’s
reception of the text in his or her specific cultural context which deter-
mines the need for any changes of information content from the source
text to the target text. Any information that the source text implies for
the source text reader may have to be made explicit for the target text
reader; and any information that is supplied in the source text but is
obvious to the target text reader could be omitted in the target text.
The handling of such information in translation is closely related to the
overt–covert distinction in that the strategic course delimits the scope of
the translator’s options. A source text such as a historically significant
work of prose fiction would have to be translated overtly, that is, the text
in all its dimensions “must remain as intact as possible given the neces-
sary transfer and recoding in another language” (House 1997:68). As a
consequence, the target text proper should not add or omit any informa-
tion provided in the source text. If any additions need to be made, they
should be made as footnotes or end notes; omissions are to be avoided,
even if a particular piece of information is not required for an appro-
priate understanding of the target text. In a covert translation setting,
any such additions or omissions are justified by the fact that the target
text – seeking to conceal the existence of the source text – is to be read
like an original: whatever smacks of translation is to be smoothed out
using a cultural filter.12 Thus, while the audience is usually less import-
ant when it comes to defining a specific translation strategy within the
overt–covert framework (unless the translation brief specifies a target
audience completely different from that of the source text), its socio-
cultural background is absolutely crucial to the implementation of that
strategy in a covert setting.
Similarly, content is a factor that is particularly relevant after the
overall translation strategy has been defined. It is closely linked to the
needs of the target audience and subject to the following rules: in an
overt translation, the target text deals with the same cultural context
as the source text (which is often a source culture context); in a covert
translation, the target text typically deals with a target culture context,
which means that the content of the source text may have to be adapted.
Whether the source text content has to be changed in the target text
and if so, to what extent, depends on the quality of the content. The
translator will have to find out in which way the content of the source
text is embedded in the source culture, and how it can be rendered in
the cultural context of the target text. Is what the source text refers
to specific to the source culture? Is there an equivalent in the target
culture? If both questions are answered in the affirmative, then the
covert procedure of adapting the source text content in the target text
should be straightforward. A typical example in an English-G erman
translation would be the conversion of non-metric measurements into
metric measurements. Answering “no” to the second question calls
for the usual tricks and techniques such as borrowing, omission, ex-
planation, compensation, etc. – depending on the requirements of the
individual translation.
The above discussion can be summarised in a diagram that shows the
central criteria relevant within the overt–covert framework.
literary form: genre non-fiction
translation strategy
before defining the
special form: style etc. inconspicuous
before defining the
source text
well-known little known
author
reference to a no reference to a
content
specific culture specific culture
not applicable … audience … unless specified
form: overall
not applicable … … unless specified
structure, layout
overt covert
translation translation
change of genre
identical genres form: genre
possible
after defining the
after defining the

consistently overt form: style etc. consistently covert
source text
degree of overtness degree of covertness
author
source may be
source retained content
adapted
defines handling … audience … of content
form: overall major changes
few changes if any
structure, layout possible
Figure 3.2 O
vert–covert overview (specific aspects).
The six criteria – genre, style, source text author, content, audience,
and overall text structure and layout – are analysed as to their influence
on translation strategy and their consequences for the implementation of
a particular strategy. It is clear that before defining the translation strat-
egy, the focus is on the source text, whereas afterwards it is on the target
text and its translational opportunities. The question whether a specific
overall translation strategy should be termed “overt” or “covert” cannot
always be answered definitively, as some aspects may be overt while oth-
ers are covert. There is no one criterion that overrides all others: overt-
ness or covertness is an issue that has to be decided individually for each
of the above criteria. This, however, has to be seen against the backdrop
of our initial distinction between the overt–covert principle and the im-
plications of that principle: while the overall translation framework may
be clearly definable as overt (because of a target text that is marked as
a translation) or covert (because of a target text that is not recognised
as a translation), the actual translation mode is often less obvious and
may sometimes run counter to the overtness or covertness of the overall
framework. Whether the overall framework of a translation is overt or
covert depends on the various criteria only inasmuch as these criteria
help to specify the function of the text in terms of the primary and sec-
ondary level functions described by Juliane House. Any answer to this
question will also have to take into account the translation brief and the
supposed translational expectations in the target culture.
The Problem of Subjective Evaluation
Translation quality assessment is essentially a human affair. And since
human beings are intrinsically fallible, there is no possibility for TQA to
eliminate error or, for that matter, judgemental variance. Several evalu-
ators evaluating the same translation will hardly ever produce identical
results – neither when identifying any mistakes and weighting them, nor
when marking the translator’s achievement.13 While (1) identifying a
mistake (or, in a reward system, an exceptionally well-translated pas-
sage) is to a large degree determined by the context of the translation,
(2) weighting any deviation from an expected standard and (3) marking
the target text as a whole seem rather arbitrary. The reason for this is
obvious: in the above three-step evaluation process, identifying a devi-
ation requires less input by the evaluator than weighting it or marking
the translation as a whole, because any judgement in the first step is
more immediately prompted by the translation situation at hand than a
judgement in the second or third step. For weighting and marking de-
pend not only on why something has been identified as a mistake or as
an exceptionally good rendering, but also on how far such a deviation
affects the usefulness (or any given purpose) of the target text. In order
to assess the extent to which a particular mistake detracts from the over-
all quality of a translation, the evaluator needs an assessment grid, for
example, a series of weighted metrics as proposed by Geoffrey Kingscott
(cf. Kingscott 2007:322) or the rubric drawn up by Claudia Angelelli
(cf. Angelelli 2009:40–41). By way of discussing the involvement of the
evaluator in more detail, we will first look at the problem of subjectivity
from a philosophical point of view.
The Problem of Subjectivity from a Philosophical

Point of View
In his essay “Das Problem der Objektivität in der antiken Philosophie”,
Olof Gigon writes about subjectivity:
Subjektiv werden wir eine Äußerung nennen, in der ausschließlich
der einzelne Mensch sich selber ausspricht und die den Charak-
ter der Verbindlichkeit weder anstrebt noch zu erreichen vermag;
Verbindlichkeit bedeutet hier die Bindung an eine Wirklichkeit, über
die in der Weise gesprochen wird, daß das Einverständnis des Adres-
saten mit dem, was gesprochen wurde, erwartet werden darf.
(Gigon 1976:11)
We will call an utterance subjective, in which only the individual

human being expresses him- or herself and which neither strives, nor
is able, to achieve the character of commitment; commitment, here,
refers to the process of connecting with a reality which is talked
about in such a way that the addressee can be expected to agree with
what has been said.
[my translation, H. B.]
Here, subjectivity is characterised by the absence of a harmonising link

that integrates an utterance and its relation to the world. By implication,
objectivity involves general agreement about the meaning of a given ut-
terance or perception (see also Radnitzky 1976:189). It is conceived as
the ability to perceive the world as it is in itself, that is, as it is outside the
perceiver’s point of view (cf. Gigon 1976:11).
Analysing the images prepared for scientific atlases, Lorraine Daston
and Peter Galison trace the history of objectivity through the nineteenth
and twentieth centuries. From their vantage point,
[o]bjectivity preserves the artifact or variation that would have been

erased in the name of truth; it scruples to filter out the noise that
undermines certainty. To be objective is to aspire to knowledge that
bears no trace of the knower – knowledge unmarked by prejudice or
skill, fantasy or judgment, wishing or striving. Objectivity is blind
sight, seeing without interference, interpretation, or intelligence.
(Daston/Galison 2010:17)
Confusingly, in this quotation, objectivity appears to be an obstacle on

the way to achieving truth and certainty. However, truth and certainty
cannot be found if those seeking them tamper with the results of ex-
periments that, in themselves, conform to the requirements of objectiv-
ity. For instance, a photographic image that resembles a regular shape
must not be reinterpreted into something equalling that shape; rather,
it must be taken for what it is: an irregular shape. Objectivity is at
risk whenever truly human faculties such as judgement or interpretation
come into play.
While it is generally agreed that, in a scientific context, the subjective
should be eliminated as far as possible to make room for a more objec-
tive perspective, the question remains whether this goal can be achieved.
Thomas Nagel, in his seminal essay “Subjective and Objective”, d escribes
the process that would negotiate subjectivity into objectivity:
At one end is the point of view of a particular individual, having a

specific constitution, situation, and relation to the rest of the world.
From here the direction of movement toward greater objectivity
involves, first, abstraction from the individual’s specific spatial, tem-
poral, and personal position in the world, then from the features
that distinguish him from other humans, then gradually from the
forms of perception and action characteristic of humans, and away
from the narrow range of a human scale in space, time, and quan-
tity, toward a conception of the world which as far as possible is
not the view from anywhere within it. […] The distinction between
subjective and objective is relative. A general human point of view is
more objective than the view from where you happen to be, but less
objective than the viewpoint of physical science.
(Nagel 1992:206)
Although “abstraction from the individual’s specific spatial, temporal,

and personal position in the world” suggests that subjectivity can be
overcome in a limited way if a consensus is reached among different indi-
viduals, Nagel is wary of such a solution. He believes that the subjective
is, in general, “intersubjectively available” (Nagel 1992:207) and that,
therefore, “the transition to a more objective viewpoint is not accom-
plished merely through intersubjective agreement” (Nagel 1992:208).
Rather than trying to reach a consensus, one should strive for “ externality
or detachment” (ibid.) to achieve objectivity. This is why “scientific
measurement interposes between us and the world instruments whose
interactions with the world are of a kind that could be detected by a
creature not sharing the human senses” (Nagel 1992:209). The question
remains: what can be done in those cases where no instrument is avail-
able for objective measurement?
Nagel is aware of this predicament. He acknowledges that “the consis-
tent pursuit of greater objectivity runs into trouble […] when it is turned
back on the self, as it must be to pursue its comprehensive ambitions”
(Nagel 1992:210). There seem to be three solutions, all of which are
more or less inadequate (cf. Nagel 1992:210–211): reduction (that is, ex-
plaining things in terms of specific criteria), elimination (that is, denying
the existence of the subjective), and annexation (inventing a new element
of objective reality such as the will, the ego, the soul, or the command of
God). Is there no better solution? Nagel explains:
The only alternative to these unsatisfactory moves is to resist the

voracity of the objective appetite, and stop assuming that under-
standing of the world and our position in it can always be advanced
by detaching from that position and subsuming whatever appears
from there under a single more comprehensive conception.
(Nagel 1992:211)
In other words, the subjective should be recognised as an irreducible ele-

ment in the pursuit of knowledge. While “accepting the polarity without
allowing either of its terms to swallow the other” is certainly not an
easy task to accomplish, it should be at least “a creative one” (Nagel
1992:213).
To sum up the results of our philosophical digression, there is no way
to dispense with the subjective altogether. While striving to achieve
objectivity is epistemologically desirable, it is certainly no panacea for
problems arising from methodological overestimation. As the bias of
subjectivity will never be completely eliminated, it is important that its
presence be recognised. Subjectivity and objectivity, as we see them, are
not two mutually exclusive concepts; they rather constitute the extreme
positions on a cline with an infinite number of positions in between.
The question is: how much of the objective can be achieved in a given
research method? A complete detachment of the matter at hand from
the human individual investigating it would be an ideal approximation
to objectivity. This can be achieved, for example, in studies based on
scientific measurement, where an instrument supplies results that are
as objective as they can be under the circumstances defined by the re-
searcher. Such results should, then, also be reproducible by other scien-
tists. However, the more dependent scientific investigation is on input
from human beings, the more susceptible it is to subjective influence. In
such cases, the objective is best emulated by intersubjective agreement
and consensus building. With these insights, we are now in a position to
analyse different approaches to TQA and their efforts to overcome the
subjective.
The Process of Translation Quality Assessment

As has been briefly mentioned in the introductory paragraph to this
section, the process of translation quality assessment can be subdivided
into three steps. While the first step is mandatory, the second and third
steps may be optional, depending on the purpose of a specific TQA
task. Let us look at these steps in more detail. The first step consists
in a close reading of the target text and its comparison with the source
text (if available) as well as with the translation standard expected by
the evaluator. As a result of these comparisons, the evaluator notes
deviations from the expected standard, which may be negative (as in
the case of errors, mistakes, and substandard solutions) or positive
(if a translated passage exceeds expectations). In setting the expected
translation standard, the evaluator is guided by several factors,14 some
of which are fairly objective whereas others are less so. Most objec-
tive are usually the graphic representations of the target text and the
source text. Relative objectivity might also be claimed by factors such
as spelling and grammatical rules of source and target languages, any
unequivocal specifications of the translation job including a deadline,
any obvious political constraints such as censorship, and any clearly
established lingual-cultural conventions (ideally supported by empirical
studies). Provided that a given translation situation is fully known in all
its aspects, the most subjective factors to be taken into account are the
interpretation of textual meaning, and, of course, the translation strat-
egy and process. Yet, these factors are still less subjective than those
required in the second and third steps.
Once the quality of a translation has been established in line with
the first step, the TQA process may already be finished as the evaluator
should now be able to provide useful feedback, thus, helping the trans-
lator to improve his or her translation skills. Some TQA procedures,
however, call for quantification, for example, the assessment of written
exams in translator training, or the screening of translations prepared for
government agencies or high-quality language services. Such quantifica-
tion is furnished in the second and third steps – counting and weighting
the findings of the first step, and marking the overall achievement. How-
ever, the validity of the results of the second and third steps hinges on the
thoroughness with which the first step has been carried out. The more
convincing the results of the first step, the more reliable can be the quan-
tifications in the second and third steps. While marking the translation
as a whole is possible without counting any deviations and weighting
them, completing the second step before tackling the third helps to avoid
what some translation theorists frown upon as impressionistic (cf., for
instance, Al Qinai 2000:497; Eyckmans et al. 2009:75; Lee-Jahnke
2001:206). The third step is likely to be more subjective than the sec-
ond, because assigning a specific mark and determining the threshold
between pass and failure is more or less arbitrary, whereas weighting a
translation error can often be justified by its degree of non-compliance
with the intended function of the target text. Figure 3.3 gives an over-
view of the relative objectivity or subjectivity of the three steps in trans-
lation quality assessment.
The question now is: What can be done in each of the three steps to
render translation quality assessment as objective as possible? To find
an answer to this question, we will first present an overview of how
other scholars see the problem of subjectivity and then look at differ-
ent approaches to TQA, examining their strategies to approximate
objectivity.
objective subjective
step 1 step 2 step 3
Figure 3.3 O
bjectivity and subjectivity in TQA.
The Problem of Subjectivity from a Translation
Studies Point of View
In their article on ideology in the Routledge Encyclopedia of Translation
Studies, Peter Fawcett and Jeremy Munday write:
In the scientific and technological atmosphere of the early and

mid-twentieth century, there was for a time a feeling that linguistic
theory had provided a ‘scientific’ basis for grounding translation in
a way that should eliminate subjective evaluations of ‘accuracy’ and
transfer of meaning.
(Fawcett/Munday 2009:139)
This feeling has all but disappeared as the inevitability of the subjective
in translation quality assessment is generally acknowledged. In 1966,
John B. Carroll – whose machine translation study still reflects a more
sanguine outlook on translating as an objective science – admits that
“[t]he evaluation of the adequacy of a translation must rest ultimately
upon subjective judgments, that is, judgments resulting from human
cognitions and intuitions” (Carroll 1966:55). Some two decades later,
Malcolm Williams finds that “in the eyes of many academics and prac-
titioners, translation quality assessment (TQA) is too subjective or too
rigid to yield valid, reliable results” (Williams 1989:13). According to
Reiß/Vermeer, some subjectivity is always attached to any value judge-
ment (cf. Reiß/Vermeer 1991:144). Not even the concept of equivalence
in translation can claim to be objective, because what is considered
equivalent in the source and target texts is subject to the individual
translator’s or evaluator’s discernment. As Jeremy Munday puts it, “the
whole question of equivalence inevitably entails subjective judgement
from the translator or analyst” (Munday 2001:43).
Most straightforwardly does Vermeer give in to the inevitability
of the subjective when he states that no evaluation is objective (cf. Ver-
meer 2006:403).15 Yet, while some theorists go for intersubjective agree-
ment as the best solution to the evaluator’s dilemma – see, for example,
Gerzymisch-Arbogast (1994) and House (1997) – Christina Schäffner asks:
Is intersubjective agreement possible at all? With each assessor h

aving
a specific aim, depending on the factors of the assessment context, and
applying different assessment criteria, the answer will have to be ‘No’.
(Schäffner 1998:4)
Still, if a translation has to be evaluated, this evaluation should be clear

to the translator and to other evaluators. In order to achieve clarity of
judgement, Ammann recommends that the necessary subjectivity be met
by a procedure based on pre-defined theoretical premises and a specified
method – with the premises to be stated for each criticism (cf. Ammann
1990:213). Similarly, Rodríguez Rodríguez (2007) favours a methodical
procedure to approximate objectivity:
The approach must be systematic in order to be objective and avoid

the problems which arise in the attempt to balance the theory with
the practice of translation. As this suggests, there is no doubt that
Translation Criticism requires the implementation of some assess-
ment criteria and of a systematic scheme of analysis of the two texts
to reach the evaluation of the target texts; these criteria must be
established a posteriori in each analysis, for they are closely related
to the specific characteristics which the text displays.
(Rodríguez Rodríguez 2007:6)
What is interesting, here, is the notion that, in order to be objective, the

criteria should be defined on the basis of the features revealed in the
source and target texts. This recalls van den Broeck, who insists that “it
is the critic’s first duty to acknowledge the translator’s norm as objec-
tively as possible before (or while) confronting the reader with his own
set of norms” (Broeck, van den 1985:60).
While Ammann focusses on theoretical premises that depend on a
methodological argument and Rodríguez Rodríguez stresses the impor-
tance of a systematic procedure, Mossop (1989) suggests empiricism as
a means to partly offset the subjective in TQA. “‘Objective’ translation
evaluation”, he writes, “usually refers to an evaluation system that will
let different evaluators arrive at similar conclusions” (Mossop 1989:55).
Mossop’s idea is that “the criteria for assessment be drawn from the
translational norm of the target culture” (Mossop 1989:56); and “[t]o
identify the norm, what is really needed is an empirical study” (Mossop
1989:59). Other explicit or implicit advocates of empiricism in transla-
tion quality assessment include Hönig (1998:32) (“the speculative ele-
ment will remain – at least as long as there are no hard and fast empirical
data”) and Bowker (2001:346) (“A corpus-based approach to translation
evaluation […] is empirical and therefore objective”).
In addition to a systematic approach and an empirical basis, there is
a third, rather unlikely alternative of achieving a more objective TQA
result: quantification. Although – as we have argued above – the use of
numbers for the assessment of a target text involves human arbitrariness,
theorists have pointed out that “[t]he purpose of quantification is to cre-
ate a more objective, transparent and defensible assessment” (Williams
2001:335) and that “using metrics forces revisers into being objective”
(Kingscott 2007:323). The reason is obvious: evaluation techniques such
as counting and weighting individual mistakes or exceptional trans-
lation solutions demand a balanced comparison of the points at issue
both intratextually and intertextually. And this balanced comparison, in
turn, helps to assess a translated text more evenhandedly.
How to Curb the Subjective in Translation
Quality Assessment
We have now identified three methods that can be said to contribute in
some way to a more objective translation quality assessment: a system-
atic selection and use of TQA criteria, an empirical approach, and quan-
tification of the evaluation results. These methods increase the degree
of objectivity in that they interpose a pre-defined scheme between the
evaluator and the target text. While any such scheme is intrinsically sub-
jective (having been designed by a human being), its potential to be used
repeatedly by different human beings detaches the scheme from the indi-
vidual, rendering it interindividual and, thus, more objective. For exam-
ple, the linguistic dimensions employed by Juliane House can easily be
adapted to the requirements of different evaluators and evaluations. The
concise definition of the scheme ensures that certain important analysis
parameters are covered by those who wish to use the scheme; never-
theless, there still remains ample scope for variation as the parameters
need to be filled with content by the human evaluator and not all human
evaluators would come up with the same solutions as Juliane House in
her “model analyses” (House 1997:121). A given evaluation – even if it
follows a clear system – just constitutes an offer to the translator, and
to any other evaluators, to trace the line of argument proposed in the
evaluation and then to either agree or disagree. Any issues that cannot be
agreed on are indicative of the residual subjective. A final judgement is
suspended ad infinitum, because any evaluation is essentially provisional
by dint of its inherently subjective nature.
Empirical analyses used in TQA may complement a criterion-
referenced evaluation, providing an additional instrument to minimise
the subjective. One approach that can be said to be empirical is the CDI
method by Eyckmans et al. (2009), a norm-referenced approach which
involves the calibration of dichotomous items (see Chapter 2 for a brief
outline). Here, empiricism shows in the pretest to establish the dichoto-
mous items and in the agreement required between the evaluators: “for
each translated segment it is agreed between graders which alternatives
are acceptable and which are not” (Eyckmans et al. 2009:76). While
it does considerably reduce the subjective element, the norm-referenced
approach falls short of eliminating it completely: for the validity of the
common denominator that has to be found among the evaluators does
not necessarily extend beyond their number. The important question
here is: how convincing are the arguments on which the consensus has
been built? Empiricism in TQA usually refers to corpus analysis, which
can help the evaluator to justify a translation solution as acceptable or
to reject it as unacceptable. What is more, the results of a corpus anal-
ysis can serve as crucial evidence in deciding which of two acceptable
translation solutions is best.16 How to set up an evaluation corpus has
been described in detail by Bowker (2001). Yet, however insightful the
findings of a corpus-based approach to TQA, it is important to keep
in mind that “[a] corpus should not be seen as a replacement for com-
petence and critical judgement on the part of evaluators, but rather as
an aid to help them make sound and objective judgements” (Bowker
2001:361). How to identify noteworthy passages in the target text and
how to judge them using a corpus analysis tool depends on the evalua-
tor: his or her competence in handling a given TQA task constitutes the
irreducible subjective in the evaluation of a translated text.
If a corpus-based analysis comes relatively close to the objective ideal,
then a metric approach to translation quality assessment approximates
this ideal to a lesser degree as it is more dependent on input from the
evaluator. Take, for instance, the weighted grid proposed by Williams
(cf. Williams 2009:18): while the definition and application of the
parameters for the argument structure and for the various microtextual
issues (which are subsumed under the headings of “Other elements of
transfer”, “Usage/Grammar”, and “Typography”) requires as much sub-
jective input as any criterion-referenced approach, the weighting process
appears to be even more arbitrary. Here, determining the relative weight
of each parameter for a specific translation seems to be fairly straight-
forward compared with the task of assigning a quality value out of ten
for the implementation of the parameters in the target text: the general
range is obvious, the exact figure less so. Once it is established, the sys-
tem benefits from the evaluator’s intensive engagement with the quality
of a particular translation, so that, in the end, quantification leads to a
more balanced view of the individual merits and demerits of the target
text. A similar effect is produced by Angelelli’s draft rubric (cf. Angelelli
2009:40–41), when it comes to matching the quality of a translation
with one out of five levels for each rubric element. Here, the target text
has to be checked against “Source Text Meaning”, “Style and Cohesion”,
“Situational Appropriateness”, “Grammar and Mechanics”, and “Trans-
lation Skill”. The degree to which the translated text meets these general
parameters is specified in terms of pre-defined levels, from inability (or
a similar expression) to masterful ability (or a similar expression). The
rubric provides a framework within which a quality judgement is sup-
ported by its relative unfitness for the next higher or lower level. Thus,
the subjective effort needed to understand and apply such a framework
(or to create another one) helps to achieve more objectivity.
By way of conclusion, here are again the most important findings
about the subjective and objective in translation quality assessment. As
the human element is necessarily involved in translation and evaluation,
and as the individual cannot detach him- or herself from his or her self,
the subjective has to be acknowledged as an inevitable ingredient in any
TQA recipe. However, the impact of the subjective in the evaluation of a
translated text can be reduced by interposing between the evaluator and
the translation to be evaluated a clearly defined system of parameters,
by making use of empirical studies to substantiate the claim that one
translation solution is better than another, and by using an assessment
grid that categorises and quantifies any conspicuous features noticed in
the target text. The objective ideal remains unattainable; yet, we may get
within sight of it, if we follow a clear approach and are, at the same time,
aware of the intrinsic limitations of that approach.
Towards Evaluating the Evaluator

Once we have established our TQA model, we can put it to good use: it
can help assess the quality of translations, and it can serve as a tool to
analyse the evaluations of commented translations. This book focusses
on the latter application. Therefore, the TQA model needs to be comple-
mented by a framework that caters specifically for the different ways in
which the evaluation of a commented translation may fulfil its purpose.
Such a framework will be outlined in this section.
Some Views on How to Evaluate Translations

In his paper “Goals of a revision course”, Mossop (1992) distinguishes
between translation and revision skills. Trying to teach students how
to revise a translated text, he finds that they need “a concept of them-
selves as editors” (Mossop 1992:82). If the students start revising with
a translator’s frame of mind, they “seem to be asking themselves: ‘How
can I improve the text?’ or ‘How would I have translated this text?’”
(ibid.). However, “[t]hese are the wrong questions to ask. Any text can
be improved; the question is: ‘Does it need to be improved?’” (Mossop
1992:83). Students “must overcome the habit of assuming that, of two
possible wordings, one must be right and the other wrong, or at any
rate one must be better” (ibid.). In this context, the “justification of a
change” (Mossop 1992:84) is crucial.
Mossop’s remarks are very much to the point: the reviser – or, for that
matter, the evaluator – of a translation must be able to abstract from his
or her own preferences and assume a perspective that is as objective as
possible. In other words, the changes that the reviser deems necessary
should be accepted as necessary by other translation experts and, ideally,
also by the translator him- or herself. Such acceptance is greatly helped by
clear and succinct arguments that underpin the need to change the target
text. While Mossop’s message is patent, at least one of his statements is
not quite correct. Thus, his claim that any text can be improved is illogical
as it implies the infinite improvement of an ultimately finite entity (the
target text). To improve a text, however, is to make it better, which calls
for a qualitative comparison of the versions before and after the improve-
ment. Improving a text makes sense only as long as the improved version
is generally accepted as – or unequivocally shown to be – better. Mossop’s
statement would be unassailable if it ran: any text can be changed.
The notion that, of two alternative translation solutions, one is better
than the other reflects the translator’s frame of mind since he or she has
to choose one alternative and would want to choose the better one. Such
a choice, whether it is made in a translation or in a revision context, re-
quires argumentative support: if this support is generally convincing, it
can be said to justify the choice.
The above question whether a translated text needs to be improved
not just reveals a dichotomy between the perspectives of a translator
and a reviser but is also indicative of a juxtaposition between theory and
practice or, to be more precise, between an academic and a professional
translation environment. Thus, Joanna Drugan writes:
[Juliane House] sees the fundamental question driving academic

work on quality as, ‘How do we know when a translation is good?’
The equivalent fundamental question for the profession would rather
seem to be, ‘How do we know when a translation is good enough?’.
(Drugan 2013:41–42)
These questions correspond to Brian Mossop’s “How can I improve

the text?” and “Does it need to be improved?”, respectively. While the
first questions (“How do we know when a translation is good?” and
“How can I improve the text?”) suggest that a translated text needs to be
changed to achieve the quality that it can achieve, the second questions
(“How do we know when a translation is good enough?” and “Does
it need to be improved?”) point to a strategy whereby a translated text
should be accepted as satisfactory if it can be accepted as satisfactory.
This distinction is summarised in the following quotation by Jiří Levý:
Translation theory tends to be normative, to instruct translators on the

OPTIMAL solution; actual translation work, however, is pragmatic;
the translator resolves for that one of the possible solutions which
promises a maximum of effect with a minimum of effort. That is to
say, he intuitively resolves for the so-called MINIMAX STRATEGY.
(Levý 2000/1967:156)
Whether in a professional or in an academic context: both the revisers

and the evaluators of translations will have to get their bearings between
the conflicting priorities of what is possible and what is necessary. They
both need to ensure, however, that their positioning is not random.
For the reviser or evaluator of a translation to position him- or her-
self with regard to acceptable and unacceptable translation solutions
presupposes a clear concept of what a translation should look like in
a given translation situation. Yet, what constitutes an error is not al-
ways obvious, as Paul Kußmaul demonstrates with a pertinent exam-
ple from a text that contains a description of the experiences made by
students travelling in Asia. In the sentence “Graduates of many British
and American Universities, and from many other parts of the world, are
postponing entry into their national work-forces to venture overseas and
experience the world at first hand”, the final phrase was rendered by a
student as “um zunächst einmal die Welt kennenzulernen”. This solution
resulted in a controversy:
The two teachers assessing the translation agreed that a more faith-
ful version would have been um die Welt aus eigener Anschauung
kennenzulernen, but they did not agree on how to grade the error.
Teacher A’s verdict was that this was a very serious error. She ar-
gued that at first hand never had the meaning of the adverb zunächst
(“first of all”). The notion of physical experience expressed in at first
hand, she said, was completely lost in the translation. The student in
her opinion displayed a serious deficiency in foreign language com-
petence. A knowledge of basic idioms such as “at first hand”, she
argued, was an indispensable requirement for students before they
embarked on a course of professional translation.
Teacher B agreed that at first hand does not have the same meaning
as “first of all”, but he argued that the translation, although incor-
rect as a literal translation, makes sense within the context and does
not distort the meaning of the text. The temporal notion, he said,
was indeed supported by the immediately preceding context. As far
as the physical experience expressed in the English idiom was con-
cerned, this was rendered […] by the detailed scenic description of
students travelling in Asia. One might even say that the phrase “at
first hand” was redundant in the text. In the translation there was
thus no loss of information. As a consequence, one should not talk
of an error here and the candidate should not be penalized.
(Kußmaul 1995:127–128)
Kußmaul contrasts the foreign language teacher’s view, which “is centred
on the word or phrase as an isolated unit” (Kußmaul 1995:128), with
the professional translator’s view, which emphasises “the communicative
function of the word, phrase or sentence in question” (ibid.), and then con-
cludes: “It may very well be that what is a mistake from a language teaching
point of view is no mistake from a communicative point of view” (ibid.).
Yet, where do we draw the dividing line between the two approaches?
What are the criteria that define the communicative function in a given
translation situation? In the above example, the argument that the phrase
“at first hand” is redundant (because the text is about personal experi-
ences rather than experiences made through some medium) leads to the
conclusion that no information is lost – which is, here, regarded as the
ultimate criterion of translational adequacy. Of course, the requirements
of the communicative function would also be met, if “at first hand” was
properly rendered as “aus eigener Anschauung”, but this is not the point.
The point is that, given the above source text and a corresponding trans-
lation brief, there is no need for the translator to translate individual
words and phrases as literally as possible – provided that the evaluator
assumes the point of view of the professional translator. If the evaluator
assumes the point of view of the foreign language teacher (a view that
may make sense in a translation exercise as part of a language course
or of literary studies), he or she should communicate the corresponding
expectations to the student translator in advance.
In the discussion about errors in translation, Anthony Pym introduces
an interesting distinction, juxtaposing binary and non-binary errors:
A binary error opposes a wrong answer to the right answer; non-

binarism requires that the target text actually selected be opposed to
at least one further target text, which could also have been s elected,
and then to possible wrong answers. For binarism, there is only
right and wrong; for non-binarism there are at least two right an-
swers and then the wrong ones.
(Pym 1992:282)
Spotting a binary error, the evaluator can clearly say, “It’s wrong!”,
whereas in the case of a non-binary error, the evaluator would show a
more sophisticated response of the type, “It’s correct, but …” (cf. Pym
1992:282). The binary/non-binary distinction harks back to Pym’s defi-
nition of translation competence as “a process of choosing between
viable alternatives” (Pym 2003:491, see also Chapter 4). All truly trans-
lational errors are, then, “non-binary by definition” (Pym 1992:283) in
that the translator has either not generated enough viable alternatives
or not selected the best viable alternative. Typical binary errors, on the
other hand, would be spelling mistakes or obvious false friends.
However plausible, the binary/non-binary distinction can only be-
come relevant if the underlying errors are identified on the basis of a
sound argument. For the question whether an error is obvious (that is,
binary) or less obvious (that is, non-binary) cannot always be answered
immediately: what has, on first appearances, been considered a binary
error might eventually turn out to be a non-binary error or even no
error at all. It all depends on the arguments put forward to defend a
translation solution as reasonable or to expose it as an error. What is
more, different evaluators will often come up with different arguments
(see Paul Kußmaul’s example, above), and different arguments will re-
sult in different assessments of translation solutions. The distinction be-
tween binary and non-binary errors is, therefore, directly related to the
argumentative basis on which an evaluator regards a target text unit
as right, wrong, or something in between. What one evaluator would
call a non-binary error may, thus, be binary to another evaluator. The
dichotomy between non-binary and binary errors is straightforward
only from a theoretical perspective; in practice, it should be replaced by
a cline that acknowledges the difficulty of categorising some errors as
either binary or non-binary. For example, how do we categorise o bvious
errors that can be corrected in several ways? The error’s obviousness
would make it binary; the different possibilities of correction would
point to a non-binary error.
Different error categories might suggest different repercussions on the
usefulness of the translated text. As a binary error is easily recognised as
such, its impact would tend to be less strong than that of a non-binary
error which, if it goes unnoticed, is more likely to cause a detrimental
effect. When grading errors, the guiding question is: “How far-reaching
is the error?” (Kußmaul 1995:130). It is complemented by questions such
as “Does [the error] distort the sense of a sentence, of a passage or even
of the whole text [?]” (ibid.) or “Does it inhibit or even destroy commu-
nication?” (ibid.). Binary errors, being non-translational errors, appear
to be more easily assessed in terms of their quantitative impact than
non-binary errors, because “if we agree that evaluation is a quantita-
tive, non-binary concept […], there is no simple and convenient way of
grading” (ibid.). That a binary error, too, can have a very serious effect
on the quality of the translation is shown in the following example:
In a translation test, students had to translate the title and first

passage of a book on handwriting. The title ran
The Hidden Language of your Handwriting.
It was translated in one case as
Die verborgene Sprache ihrer Handschrift.
I have no objection to the phrasing of the sentence, but not writing
ihrer with a capital letter changes its meaning. A back-translation
would give us The hidden language of their handwriting or The
hidden language of her handwriting. I asked my students what they
would expect of a book with this title, and one of them said it re-
minded her of titles of novels and another was reminded of titles
of detective stories. Obviously, this is not merely an orthographic
mistake but an error in sentence meaning if not text meaning and
cannot be treated lightly.
(Kußmaul 1995:145)
Kußmaul implies that what counts is the perspective of the reader: in a

professional context, the translation should, first and foremost, ensure
that the target audience get what they should get – in the above example,
they should get a book about handwriting. Raising expectations that
the book is a novel or detective story clearly indicates a serious error,
although it may be argued that the seriousness of the error depends on
the actual situation in which it is encountered. Potential readers of the
book in question browsing through the shelf with non-fiction books
on handwriting in a bookshop would immediately recognise the mis-
take or even overlook it. Without this context, however, the title might
easily be misunderstood. Kußmaul’s insistence that “this is not merely
an orthographic mistake” is true from the reader’s perspective; from the
perspective of the translator, the error is certainly just a spelling mistake.
That, in translator training, errors should generally be identified on
the basis of the professional translator’s rather than the foreign language
teacher’s view does not mean that their grading should necessarily be
restricted to the professional perspective. Kußmaul’s harsh judgement in
the above example that failure to capitalise “ihrer” constitutes “an error
in sentence meaning if not text meaning and cannot be treated lightly”
(Kußmaul 1995:145) could be challenged on the grounds that students are
not yet translators. Hannelore Lee-Jahnke, for one, emphasises that the
evaluation of a translation should not only cover the quality of the target
text but also help improve the students’ translation competence (cf. Lee-
Jahnke 2005:128). Similarly, Susanne Hagemann finds that it makes little
sense to pretend that translating as a student of translation studies is the
same as translating as a professional translator (cf. Hagemann 2007:238).
This attitude could be reflected in a more lenient grading – a grading that
tries to capture the seriousness of the error by tracing the reasons that
might have led to it rather than assessing the negative effect it may have
on the reader. After all, the evaluation of a student’s translation should
provide feedback on how well he or she did and whether he or she has im-
proved. An industry-standard evaluation that distinguishes only between
acceptable and unacceptable translations would provide reassurance to
the best students and leave the others at a loss.
Another aspect that might help students of translation improve their
performance is positive evaluation. The idea is to acknowledge what is
good in the target text in order to boost the student’s confidence with
regard to his or her ability to translate. Hagemann (2007:245) suggests
that students should get a reward for creative solutions in that the points
awarded for these solutions may partly offset the error count. Creativity
in translation is discussed at great length in Kußmaul (2000). It refers
to the ability of the translator to deal with translation problems – from
phrases or clauses that cannot be rendered literally to source text units
that require a highly sophisticated workaround in the target language.
Whether creativity in translation can always be unequivocally established
and whether a creative solution is always better than an uncreative one:
these questions need not concern us, here. The point to be made is that
since being creative is an essential characteristic of a good translator such
an ability should be recognised in the evaluation of a translated text.
Errors and/or creative solutions can be counted (with penalties either
fixed in advance for different types of mistakes or based on the mistakes’
effect on the quality of the translation) or they may serve as a vague point
of reference for an impressionistic overall assessment of the translation.
Waddington (2001, 2004, 2006) has studied the validity and reliability
of these evaluation methods (see Chapter 2 for a brief discussion) and
finds that “methods based on error analysis are more reliable than ho-
listic ones” (Waddington 2004:34). There is, however, another aspect
that has to be taken into consideration: time. For the task of reading
through students’ translations and assessing their quality can be very
time-consuming, depending on how carefully the evaluator evaluates the
translated text. Thus, Kinga Klaudy suggests that time could play a key
role in the evaluation of translated texts. He writes: “In my view, the
only correct criterion for quality assessment of students’ translations is
the amount of time required to transform them into print-ready texts. If
the revision takes more time than the translation itself, the translation is
bad” (Klaudy 1996:202). Of course, such an evaluation method would
need to be based on a standardised revision process.
This section has provided an outline of some views on translation eval-
uation, covering aspects such as binary and non-binary errors and the
importance of the evaluator’s point of view. In the next section, we will
try to bring together the various strands of these aspects in a framework
that can be used to analyse the evaluations of commented translations.
The Evaluation of Commented Translations

Before establishing a framework that can be used to analyse the evalua-
tions of commented translations in an academic context, we need to look
at the distinctive setting in which these evaluations take place. Students
of translation studies in Germany sometimes can write – and choose
to write – a so-called commented translation as a thesis towards their
translation studies degree. This thesis typically includes an analysis of
the source text, the actual translation, and a commentary on selected
translation decisions. The student usually gets a detailed translation
brief, which may involve additional tasks such as providing a relevant
glossary or presenting the target text in a pre-defined layout. Occasion-
ally, the student is supposed to conceive of the specifications him- or
herself with the proviso that they be reasonably realistic.
The evaluator of the thesis evaluates not only the actual translation but
also the analysis of the source text, the commentary, the bibliography, and
formal aspects such as the title page and the table of contents. Since the
purpose of this kind of thesis is for the student to demonstrate his or her
translation and academic writing skills together with the related linguistic
knowledge, the focus of the evaluation can be depicted in a diagram.
table of source text translated

title page commentary bibliography
contents analysis text
Figure 3.4 Evaluation focus for a commented translation.

Assessing the quality of the translated text is at the centre of the eval-
uation. The source text analysis and the commentary are interesting,
because they ought to be relevant to the translated text in that the for-
mer furnishes the basis on which the source text is rendered into the
target language, whereas the latter helps to confirm the choices made in
the translation. They also reveal the student’s ability to write academic
papers. The remaining aspects are less important since they are not re-
lated to the success of the translation. While the bibliography reflects the
student’s use of academic resources, the title page and table of contents
provide a first glimpse of the student’s diligence and carefulness in pre-
paring the thesis.
Ideally, the source text analysis, the translated text, and the commen-
tary should be interrelated: the analysis helps to define the translation
strategy, which, in turn, forms the basis of the translation; individual
solutions are, then, justified in the commentary with reference to the
source text analysis and the resulting translation strategy. This triangu-
lar correlation is shown in Figure 3.5.
translation
or
yf
can
teg
be
stra
jus
the
tifi
es
ed
ish
by
abl
est
source text analysis ← is argued on the basis of ← commentary
Figure 3.5 C
ommented translation triangle.
The significance of the commented translation triangle is such that

strict adherence to its operations is likely to result in a good translation.
A source text analysis that is performed not as an end in itself but to find
out how the source text can best be turned into a target text lays the
foundation of a successful translation; and a target text solution that is
supported by a comment on the grounds of what has been established
in the source text analysis cannot easily be rejected as inadequate. Of
course, there may also be comments that explain some aspect of the
translation without having recourse to the results of the source text
analysis – such comments may be interesting, yet, they will not help to
justify the translation. Similarly, any results of the source text analysis
that are not relevant to the translation can be regarded as useless, even if
they convey an otherwise thought-provoking piece of information.
Inasmuch as the analysis of the source text and the commentary con-
tribute to the outcome of the translation, they are also important to the
evaluation of the translated text: they help the evaluator understand the
translator’s intention – which is absolutely necessary, if one wishes to do
justice to the translator’s efforts in producing the translation. The evalu-
ation, as Raymond van den Broeck points out, “should take account […]
of the translational method adopted by the translator in view of the spe-
cific target audience envisaged, and of the options and policies followed
in order to attain his purpose” (Broeck, van den 1985:56). The corre-
sponding information is provided in the analysis of the source text. The
evaluator can, then, adopt the translator’s standard for the translation
at hand, provided that it agrees with the overall translation situation,
that is, the translation brief and any translational requirements gleaned
from the source text analysis. It may well be that, faced with a perfectly
convincing approach by the translator, the evaluator has to readjust his
or her own translation standard (cf. also Reiß 1971:12).
When it comes to assessing the quality of the actual translation, the
evaluator has quite a few options. For example, he or she may set the
source and target texts side by side and compare them sentence by sen-
tence; or the evaluator may content him- or herself with a thorough
reading of the target text. In an ideal evaluation, both procedures are
applied: while the comparison ensures that the target text is an accurate
translation of the source text and complies with the specifications for
that particular translation, the perusal of the target text as a text in
its own right helps capture its coherence and consistency. Whether to
go for the complete evaluation programme or to opt for just one of the
two procedures or to combine them in any proportion – the evaluator
has to decide this issue in line with the requirements of the evaluation
task at hand, any time limits or deadlines to be met, and his or her own
preferences. In general, the comparison of source and target would seem
to be more important in an overt translation setting, where the reader
of the target text expects to catch a glimpse of the source text, and less
important in a covert translation setting, where the target text is read as
if it were an original text composed independently of any source text.
The evaluation process may be cut short, if the quality of the translation
becomes obvious already after having read part of the text. It is not to
be expected that the evaluator indicates explicitly which of the above
options he or she chooses. A comparison of the source and target texts –
whether partial or in full – would show implicitly in any examples that
juxtapose a source text unit and the corresponding target text unit.
Unless the evaluation report states that the evaluation process has been
cut short, there is no way of establishing the facts in this respect.
At the level of the translation solution, the evaluator has to decide on
the evaluation method: whether to focus on translation errors or on suc-
cessful solutions to translation problems or on a combination of the two.
Most common seems to be the evaluation on the basis of errors, possibly
complemented by the occasional praise for exceptionally good solutions
(cf. Hagemann 2007:237). Concerning the evaluation of a commented
translation written as a BA thesis, we need to ask to what extent the
advantages of an approach that brings out the positive aspects of a trans-
lation (namely, boosting the student’s confidence) are applicable. The
answer to this question hinges on whether the student actually reads the
report written by the evaluator. While constructive feedback is probably
most effective in the evaluation of translation exercises performed in
class during the study programme, any feedback given in the evaluation
of the final thesis will reach only those few students who ask for such
feedback.
Regarding the treatment of errors, the evaluator again has several
options. One is the choice between counting errors and judging their
frequency in a wholesale fashion. The first alternative is mandatory in
the evaluation of a written exam in which several students translate the
same text. Here, the sameness of the task for each student makes the
resulting translations comparable. The situation is different for a com-
mented translation submitted as a thesis at the end of a bachelor’s degree
course: here, the individual source text has specific characteristics that
determine how difficult it is to translate. Since the level of difficulty var-
ies with each source text, it makes little sense to compare the number
of errors for one translation with that of another (even if the number is
calculated in relation to the total number of words or characters).
An evaluator who counts the errors in the translated text should also
weight them, because different errors have different effects on the qual-
ity of the translation, depending on the type of error and the context in
which it occurs. Hannelore Lee-Jahnke suggests a tripartite classifica-
tion, distinguishing between quality, creativity, and skopos as general
evaluation categories. In the first category (quality), the evaluator looks
at spelling, grammar, content etc.; the second category (creativity) is
about acknowledging positive aspects such as the fortuitous translation
of culture-specific references, and the successful rendering of register,
metaphors, puns, and similar translation problems that require some
creative thinking on the part of the translator; in the third category (sko-
pos), the evaluator checks whether the target text achieves its purpose
and complies with the norms and conventions of the target language
(cf. Lee-Jahnke 2005:128). It can be assumed that third-category errors
are most likely to affect the proper understanding of the target text as
a whole, whereas the impact of first-category errors is usually confined
to smaller text units and second-category errors merely detract from a
smooth reading of the target text.
In addition to such a general categorisation, errors have to be weighted
individually. As the above discussion of Paul Kußmaul’s evaluation
concept with the “guiding question, how far-reaching is the error?”
(Kußmaul 1995:153), has shown, the impact of an error on the under-
standing or readability of the translated text is not always obvious. One
aspect that comes into play, here, is the perspective assumed by the eval-
uator: does he or she look at the translation from the point of view of
a professional reviser or through the eyes of the student translator? The
first perspective would focus on the question whether the target text
meets the requirements of the reader, whereas the second perspective
would put more emphasis on figuring out why the translator translated
as he or she did. Which of the two perspectives an evaluator adopts will
have an impact on the weighting of the errors and, thus, on the grading
of the translation. Ideally, the evaluator should provide a description of
his or her error weighting system.
An interesting decision to be made by the evaluator concerns the ques-
tion whether he or she should actually improve on those solutions in the
student’s translation that are regarded as errors. Here, we are talking
about non-binary errors. The possible argument that the student should
do the correction him- or herself in order to benefit from the increased
effort is, however, not particularly convincing. First of all, non-binary
errors are often difficult to pinpoint so that, in some cases, the student
would be at a loss to find a better solution, even if the evaluator indicated
why he or she finds fault with the translation; the student might doubt
whether there really is a better solution. Second, in the context of a BA
thesis, the evaluation report of the commented translation constitutes
a final assessment of the student’s ability to translate. This assessment
needs to be supported by sufficient evidence for the student to accept
the final mark. Leaving the student to find out for him- or herself how a
particular error could be corrected might be useful for translations that
have to be prepared in class; it is rather impracticable for a commented
translation written as a BA thesis.
Figure 3.6 gives a rough outline of the various aspects an evaluator
might or should consider when writing his or her report on the quality
of a commented translation prepared as a BA thesis. In the diagram, a
triple arrow is used for those evaluation aspects that bear directly on
the quality of the translated text. Double and single arrows show that
the overall result is also determined by the student’s academic writing
skills and linguistic knowledge as well as by his or her diligence and
carefulness in preparing the thesis. The number of arrows is supposed
to reflect the importance of the respective evaluation aspects on the fi-
nal result: here, translation aspects come first, followed by aspects that
indicate academic skills and formal issues. While this seems to us the
most plausible order, other rankings (for instance, putting academic
formal source text analysis target text

aspects and commentary
↓ ↓↓ ↓↓↓ ↓↓↓
diligence academic translator’s comparison of close reading of
and writing intention source and target target text (partial
carefulness skills and texts (partial or in and/or or in full)
↓↓↓
linguistic full)
knowledge
↓ ↓↓ ↓↓↓
↓ ↓↓ focus on successful
focus on errors and/or
↓ ↓↓ ↓↓↓ solutions
↓ ↓↓ ↓↓↓
↓ ↓↓
↓↓↓
yes classifying no
↓↓↓
↓ ↓↓
↓ ↓↓ ↓↓↓
↓ ↓↓ and/or
↓ ↓↓ ↓↓↓
↓ ↓↓ ↓↓↓
yes counting no
↓↓↓
↓ ↓↓
↓ ↓↓ ↓↓↓ ↓↓↓ ↓↓↓
↓ ↓↓ weighting and/or holistic assessment
↓ ↓↓ ↓↓↓
↓ ↓↓
↓↓↓
yes weighting no
↓↓↓
↓ ↓↓
↓ ↓↓ ↓↓↓ ↓↓↓
↓ ↓↓ predefined and/or useful only if errors
individual of a comparable
↓↓↓
↓ ↓↓
↓ ↓↓ weighting system impact are counted
↓ ↓↓ ↓↓↓
↓ ↓↓ professional student
translation
↓↓↓
↓ ↓↓ reader’s translator’s
↓↓↓
assessment
↓ ↓↓ view view
↓ ↓↓ ↓↓↓ ↓↓↓
↓ ↓↓ more strict more lenient
↓ ↓↓ ↓↓↓
↓ ↓↓ error justification
↓↓↓
↓ ↓↓ correction? requirement
↓ ↓↓ ↓↓↓
→→→→ RESULT
↓↓
↓↓
↓↓
↓↓
↓↓
↓↓
↓↓
↓↓
↓↓
↓↓
↓↓
↓↓
↓↓
Figure 3.6 E
valuating the quality of a commented translation written as a BA
thesis.
skills on a par with translation quality) are not inconceivable and may
be favoured by the evaluator or imposed by the relevant section in the
exam regulations. Regarding the question whether errors should be
corrected the answer is “yes”, because any translation criticism ought
to be justified. And what justification could be better than a more
convincing translation solution? Harking back to the demands of var-
ious translation scholars, the problem of justifying one’s translation
criticism – and, indeed, any translation solution – will be tackled in
Chapter 5.
This chapter has prepared the ground for our theory of translation
quality assessment and its application to the evaluations of commented
translations. We have come to terms with the following four aspects:
1 The key concepts – translation quality and translation quality

assessment – have been discussed and defined. These definitions are
specifically designed to suit the purpose of the theory to be devel-
oped in Chapters 4 and 5.
2 We have set up a framework (based on Juliane House’s distinction
between overt and covert translation) that provides guidance to the
translator and the evaluator with regard to building a translation
strategy. Such a strategy serves as one normative cornerstone in the
assessment of a translated text.
3 The common objection that the evaluation of translation must nec-
essarily contain a subjective element has been put into perspective by
philosophical and translational considerations. The discussion has
shown that there is need for a theory of translation quality assess-
ment that can cope with the irreducible subjective.
4 Since our theory of translation quality assessment will be used to
analyse evaluation reports written for commented translations in
the context of a BA thesis, we have – in the fourth section – provided
a framework that outlines the various options of the evaluator of
a commented translation. This framework paves the way for our
analysis of evaluation reports in Chapter 6.
With this, we can now move on to analyse the factors that can play a role
when it comes to evaluating translation quality.
Notes
1 As Jamal Al Qinai puts it, “the reception of [the target text] is the ultimate
assessment of quality” (Al Qinai 2000:517). María Pinto even sees a devel-
opment: “The early notions of the concept of quality were centered on the
fulfilment of certain specifications; this conception gave way to the philoso-
phy of suiting individual needs, to end up with user satisfaction as the basic
principle” (Pinto 2001:290).
2 Lauscher (2006:60) gives an interesting account of what it means to assign a
value to the quality of a translation.
3 This is what Nancy Matis refers to as “[t]echnical QA” (Matis 2011:147),
that is, technical quality assurance.
4 For a rendering that does aspire to Parnassian heights, cf. Bittner (2010).
5 Cf. also Bittner (2011:78).
6 That Juliane House does not distinguish between an overt–covert principle,
on the one hand, and any conclusions to be drawn for the translation
strategy, on the other, is patent also from the following quotation:
The choice of an overt or a covert translation depends not only on the
translator or on the text to be translated, or on her subjective interpreta-
tion of the text, but also on the reasons for the translation, the implied
readers and on publishing and marketing policies, i.e. factors which have
nothing to do with translation as a linguistic procedure.
(House 2015:142, similarly also House 2014:260)
Here, overt translation is equated with source orientation, whereas covert

translation is equated with target orientation.
7 This recalls Gideon Toury’s idea of translation as being determined by the
culture in which it takes place. “Thus, when a text is offered as a translation,
it is quite readily accepted bona fide as one” (Toury 1995:26). This issue is
discussed in more detail in Chapter 5.
8 What we refer to, here, as “overall strategy” is not unlike Gideon Toury’s
“initial norm” (see Chapter 4): only that the initial norm is defined by a given
translation culture, whereas the overall strategy is defined by the translator
for a particular translation. The overall translation strategy will differ from
the initial norm, if the translator – for whatever reasons – fails to stick to the
initial norm prevalent in the relevant translation culture.
9 Juliane House defines “genre” as follows: “genre is a socially established
category characterized in terms of occurrence of use, source and a commu-
nicative purpose or any combination of these” (House 1997:107). This defi-
nition is also suitable for our use of the term. We take it to be synonymous
with “text type” in its narrower sense (see also Chapter 4).
10 Although she admits that the overt–covert distinction “goes some way to-
wards getting out of the double-bind in that a relative leaning towards orig-
inal or translation is implicit in the distinction” (House 1997:30), Juliane
House’s critical discussion of skopos theory and related approaches (cf.
House 1997:11–16) clearly shows that, to her, “translation is a double-bind
operation” (House 1997:77) in which both source and target play an import-
ant role. Note that the concept of the double bind as presented by J uliane
House differs from the double bind referred to by Jaques Derrida, who con-
ceives of it as the simultaneous need for, and impossibility of, translation (cf.
Derrida 1985:184).
11 Juliane House also acknowledges the potential impact of the author, when
she writes: “to achieve in translation a second original, hiding its source
[…] can be done relatively easily with authorless texts or texts that have dis-
pensable authors […]” (House 1997:163). In other words, significant authors
should be translated overtly.
12 For a concise description of the “cultural filter” concept, cf. House
(2001:251–252). Note that our use of “cultural filter” is less restrictive, since
we do not distinguish between a translation and a version.
13 In this context, criticism of literary translations can be very illuminating,
as has been shown by Hönig in his analysis of Marcel Reich-Ranicki’s
Der Dolchstoß des Übersetzers – Saul Bellows Roman „Der Dezember des
Dekans“ und seine deutsche Fassung, Frankfurter Allgemeine Zeitung of
4 December 1982 (cf. Hönig 1995:75–76) and by Gerzymisch-Arbogast
(1994:18–21), who writes about the controversy caused by Hanswilhelm
Haefs’ German translation of Lawrence Norfolk’s novel Lemprière’s Dic-
tionary. On a more scientific note, Huang has analysed the marks given by
60 markers for a sentence with an artificial mistake and finds that “[t]he dif-
ference in inaccuracy tolerance rates among markers who may have the au-
thority to accept or reject a translation is unacceptable” (Huang 2007:290).
Yet, for a study that examines the effects of serial translation evaluation
such as the one by Muñoz Martín/Conde Ruano (2007), differences in
evaluation are absolutely essential (or else there would be no need for such
a study).
14 See Chapter 4 for an elaborate analysis of the elements affecting the assess-
ment of translation quality.
15 Similarly unequivocal is the statement by Ilse Depraetere, who writes:
“It is an obvious point that the evaluation of a translation is subjective”
(Depraetere 2011b:107). And Radegundis Stolze, who emphasises the cen-
tral role played by the translator, insists: “An inherent characteristic of all
individual action is subjectivity. Any external, quasi ‘objective’ factors can-
not fully a ccount for the result of an individual act which also includes social
motivation and personal experience” (Stolze 2011:138).
16 Cf. Gledhill (2011) for a convincing demonstration of how corpus analysis
helps to answer the question, “between two potentially equivalent transla-
tions, […] which one is best?” (Gledhill 2011:71).
References
Acartürk-Höß, Miriam (2010). …making the mirror visible …Deutsche Über-
setzungen englischer Lyrik (W. H. Auden). Versuch einer Verwissenschaftli-
chung der Übersetzungskritik. Frankfurt/Main etc.: Peter Lang. (Beiträge zur
anglo-amerikanischen Literatur, vol. 7).
Ahrend, Klaus (2006). “Kriterien für die Bewertung von Fachübersetzungen”.
In: Schippel, Larisa (ed.) (2006:31–42).
Al-Bustan, Suad Ahmed (1993). Quality and Efficiency Factors in Translation.
University of Glasgow: Thesis.
Al Qinai, Jamal (2000). “Translation Quality Assessment. Strategies, Parametres
[sic] and Procedures”. In: MetaTranslators’ Journal, vol. 45, no. 3, 497–519.
Ammann, Margret (1990). “Anmerkungen zu einer Theorie der Übersetzungs
kritik und ihrer proaktischen Anwendung”. In: TextconText, vol. 5, 209–250.
Angelelli, Claudia (2009). “Using a rubric to assess translation ability: De-
fining the construct”. In: Angelelli, Claudia V.; Holly E. Jacobson (eds.)
(2009:13–47).
and practice. Amsterdam/Philadelphia: John Benjamins. (American Transla-
tors Association Scholarly Monograph Series vol. XIV).
Aristotle (1853). The Organon, or Logical Treatises. Translated by Octavius
Freize Owen. Vol. I. London: Henry G. Bohn.
Becker, Werner; Kurt Hübner (eds.) (1976). Objektivität in den Natur- und
Geisteswissenschaften. Hamburg: Hoffmann und Campe.
Beeby Lonsdale, Allison (2009). “Directionality”. In: Baker, Mona; Gabriela
Saldanha (eds.) (2009:84–88).
Bittner, Hansjörg (2010). “Translating rhymed poetry – Thomas Hardy’s N eutral
Tones”. In: Sahel, Said; Ralf Vogel (eds.) (2010: 20–31). Viewed: 12 August
2019. http://biecoll.ub.uni-bielefeld.de/index.php/nlk/article/view/341/434
Bittner, Hansjörg (2011). “The quality of translation in subtitling”. In:
trans-kom, vol. 4, no. 1, 76–87. Viewed: 12 August 2019. www.trans-kom.
eu/bd04nr01/trans-kom_04_01_04_Bittner_Quality.20110614.pdf
Bowker, Lynne (2001). “Towards a Methodology for a Corpus-Based Approach to
Translation Evaluation”. In: Meta Translators’ Journal, vol. 46, no. 2, 345–364.
Britten, Benjamin; Nicholas Maw; John Dowland (2005): Songs for Tenor and
Guitar (audio CD). Colchester: Chandos.
Bühler, Hildegund (2000). “Translationswissenschaft und Praxis: ein Brücken-
schlag?!”. In: Kadric, Mira et al. (eds.) (2000:363–373).
Carroll, John B. (1966). “An Experiment in Evaluating the Quality of Trans-
lations”. In: Mechanical Translation and Computational Linguistics, vol. 9,
no. 3–4, September and December 1966, 55–66.
Proceedings of the 3rd Congress of the International Federation of Transla-
tors (Bad Godesberg, 1959), New York: Macmillan and Pergamon Press.
Cole, Robert E.; W. Richard Scott (eds.) (2000). The Quality Movement &
Organization Theory. Thousand Oaks, London and New Delhi: Sage
Publications.
Daston, Lorraine; Peter Galison (2010). Objectivity. New York: Zone Books.
(first published in 2007).
Depraetere, Ilse (2011a). “Introduction”. In: Depraetere, Ilse (ed.) (2011:1–5).
Depraetere, Ilse (2011b). “A contrastive analysis of MT evaluation techniques”.
Depraetere, Ilse (ed.) (2011). Perspectives on Translation Quality. Berlin and
Boston: de Gruyter Mouton.
Derrida, Jaques (1985). “Des Tours de Babel”. Translated by Joseph Graham.
In: Graham, Joseph (ed.) (1985:165–207). (French original in the Appendix,
209–248).
Dollerup, Cay; Anne Loddegaard (eds.) (1992). Teaching Translation and
Interpreting – Training, Talent and Experience. Papers from the First Lan-
guage International Conference, Elsinore, Denmark, 31 May–2 June 1991
(Copenhagen Studies in Translation). Amsterdam and Philadelphia: John
Benjamins.
Dollerup, Cay; Vibeke Appel (eds.) (1996). Teaching Translation and Interpret-
ing 3 – New Horizons. Papers from the Third Language International Con-
ference, Elsinore, Denmark, 9–11 June 1995. Amsterdam and Philadelphia:
John Bejamins.
Drugan, Joanna (2013). Quality in Professional Translation. London and New
York: Bloomsbury Academic.
Emsel, Martina (2007). “Mikrokosmos der Verbbedeutung in Sprache und
Ü bersetzung – ist diese Qualität messbar?”. In: Schmitt, Peter A.; Heike E.
Jüngst (eds.) (2007:100–111).
Eyckmans, June; Philippe Anckaert; Winibert Segers (2009). “The perks of
norm-referenced translation evaluation”. In: Angelelli, Claudia V.; Holly E.
Jacobson (eds.) (2009:73–93).
Fawcett, Peter; Jeremy Munday (2009). “Ideology”. In: Baker, Mona; Gabriela
Saldanha (eds.) (2009:137–141).
Fox, Brian (2009). “The Quest for Quality: Perceptions and Realities”. In:
Forstner, Martin et al. (eds.) (2009:23–27).
Gambier, Yves; Henrik Gottlieb (eds.) (2001). (Multi)media Translation: Con-
cepts, Practices, and Research. Amsterdam and Philadelphia: John Benjamins.
Gerzymisch-Arbogast, Heidrun (1994). Übersetzungswissensschaftliches
Propädeutikum. Tübingen and Basel: Francke.
Gigon, Olof (1976). “Das Problem der Objektivität in der antiken Philosophie”.
In: Becker, Werner; Kurt Hübner (1976:11–40).
Gledhill, Christopher (2011). “A Lexicogrammar approach to checking quality:
looking at one or two cases of comparative translation”. In: Depraetere, Ilse
(ed.) (2011:71–97).
Graham, Joseph (ed.): (1985) Difference in Translation. Ithaca and London:
Cornell University Press.
Granzin, Katharina (2010). “Arme Stieftochter Übersetzungskritik”. In: taz.de
(31.7.2010) [online]. Viewed: 14 August 2019. www.taz.de/!401238/
Gummerus, Eivor; Paro, Catrine (2001). “Translation Quality. An Organizational
Viewpoint”. In: Gambier, Yves; Henrik Gottlieb (eds.) (2001:133–142).
Hagemann, Susanne (2007). “Zur Evaluierung kreativer Übersetzungsleistun-
gen”. In: Schmitt, Peter A.; Heike E. Jüngst (eds.) (2007:237–255).
Hamada, Tomoko (2000). “Quality as a Cultural Concept – Messages and
Metamessages”. In: Cole, Robert E.; W. Richard Scott (eds.) (2000:295–312).
Helm.
Hönig, Hans G. (1995). Konstruktives Übersetzen. Tübingen: Stauffenburg.
Hönig, Hans G. (1998). “Positions, Power and Practice: Functionalist Approaches
and Translation Quality Assessment” In: Schäffner, Christina (ed.) (1998:6–34).
Tübingen: Narr.
243–257.
House, Juliane (ed.) (2014:241–264).
House, Juliane (2015). Translation Quality Assessment: Past and Present.
London and New York: Routledge.
House, Juliane (ed.) (2014). Translation: A Multidisciplinary Approach. Hound-
smills: Palgrave Macmillan.
Huang, Harry J. (2007). “Scandals in Translation Quality Assessment”. In:
LACUS Forum, vol. 33, 285–296.
Kadric, Mira; Klaus Kaindl; Franz Pöchhacker (eds.) (2000): Translationswis-
senschaft. Festschrift für Mary Snell-Hornby zum 60. Geburtstag. Tübingen:
Stauffenburg.
Kandler, Günther (1963). “On the Problem of Quality in Translation: B asic Con-
siderations” In: Cary, Edmond; Rudolf Walter Jumpelt (eds.) (1963:291–298).
Kingscott, Geoffrey (2007). “Translation quality assessment”. In: Schmitt, Pe-
ter A.; Heike E. Jüngst (eds.) (2007:317–325).
Klaudy, Kinga (1996). “Quality assessment in school vs professional transla-
tion”. In: Dollerup, Cay; Vibeke Appel (eds.) (1996:197–204).
Kurz, Christopher (2009). “Translatorisches Qualitätsmanagement als verant-
wortungsvolles Handeln”. In: Lebende Sprachen, vol. 54, no. 4, 146–155.
Kußmaul, Paul (1995). Training the Translator. Amsterdam and Philadelphia:
John Benjamins.
Kußmaul, Paul (2000). Kreatives Übersetzen. Tübingen: Stauffenburg.
Lauscher, Susanne (2000). “Translation Quality Assessment: Where Can
T heory and Practice Meet?”. In: The Translator, vol. 6, no. 2, Special Issue:
Evaluation and Translation, 149–168.
Lauscher, Susanne (2006). “Translatqualität – ein Konsens”. In: Schippel, Lar-
isa (ed.) (2006:55–73).
Lee-Jahnke, Hannelore (2001). “Présentation”. In: Meta Translators’ Journal,
vol. 46, no. 2, 2001, 206–208.
Lee-Jahnke, Hannelore (2005). “Unterrichts- und Evaluierungsmethoden zur
Förderung des kreativen Übersetzens”. In: Lebende Sprachen, vol. 50, no. 3,
125–132.
Levý, Jiří (2000/1967). “Translation as a Decision Process”. In: Venuti, Law-
rence (ed.) (2000:148–159). Originally published in 1967 in To Honor Ro-
man Jakobson on the Occasion of his Seventieth Birthday, vol. 2. The Hague:
Mouton, 1171–1182.
Matis, Nancy (2011). “Quality Assurance in the translation workflow – A pro-
fessional’s testimony”. In: Depraetere, Ilse (ed.) (2011:147–159).
Mossop, Brian (1989). “Objective Translational Error and the Cultural Norm of
Translation”. In: Meta Translators’ Journal, vol. 2, no. 2, 55–70.
Mossop, Brian (1992). “Goals of a revision course”. In: Dollerup, Cay; Anne
Loddegaard (eds.) (1992:81–90).
Munday, Jeremy (2001). Introducing Translation Studies: Theories and Appli-
cations. London and New York: Routledge.
Muñoz Martín, Ricardo; Conde Ruano, Tomás (2007). “Effects of S erial
Translation Evaluation”. In: Schmitt, Peter A.; Heike E. Jüngst (eds.)
(2007:428–444).
Nagel, Thomas (1992). Mortal Questions. Cambridge: Cambridge University
Press. (first published in 1979).
Newmark, Peter (1981). Approaches to Translation. Oxford: Pergamon Press.
Newmark, Peter (1991). About Translation. Clevedon, Philadelphia and Ade-
laide: Multilingual Matters.
Nord, Christiane (2005, 1991). Text Analysis in Translation: Theory, Meth-
odology, and Didactic Application of a Model for Translation-Oriented
Text Analysis. 2nd ed. (1st ed. 1991). Translated by C. Nord and P. Spar-
row. Amsterdam and New York: Rodopi. (Amsterdamer Publikationen zur
Sprache und Literatur 94.) Original: Christiane Nord (1988). Textanal-
yse und Übersetzen: Theoretische Grundlagen, Methode und didaktische
Anwendung einer übersetzungsrelevanten Textanalyse. 4th ed. 2009.
Tübingen: Groos.
Pinto, María (2001). “Quality Factors in Documentary Translation”. In: Meta
Pym, Anthony (1992). “Translation error analysis and the interface with lan-
guage teaching”. In: Dollerup, Cay; Anne Loddegaard (eds.) (1992:279–288).
Pym, Anthony (2003). “Redefining Translation Competence in an Electronic
Age. In Defence of a Minimalist Approach”. In: Meta Translators’ Journal,
vol. 48, no. 4, 481–497.
Radnitzky, Gerard (1976). “Bedeutung des Objektivitätsbegriffs in Wissen-
schaftstheorie und Forschungspolitik”. In: Becker, Werner; Kurt Hübner
(eds.) (1976:189–223).
München: Max Hueber. (Hueber Hochschulreihe 12).
Reiß, Katharina; Hans J. Vermeer (1991). Grundlegung einer allgemeinen
Translationstheorie. 2nd ed. Tübingen: Niemeyer.
Sahel, Said; Ralf Vogel (eds.) (2010). 11. Norddeutsches Linguistisches Kollo-
quium. Viewed: 27 August 2019. http://biecoll.ub.uni-bielefeld.de/index.php/
nlk/issue/view/36
Schäffner, Christina (1998). “From ‘Good’ to ‘Functionally Appropriate’: As-
sessing Translation Quality”. In: Schäffner, Christina (ed.) (1998:1–5).
gual Matters.
Schippel, Larisa (ed.) (2006). Übersetzungsqualität: Kritik – Kriterien –
Bewertungshandeln. Berlin: Frank & Timme. (Arbeiten zur Theorie und
Praxis des Übersetzens und Dolmetschens, vol. 8).
Schubert, Charlotte (2007). “Grußwort zur Tagungseröffnung LICTRA 2006”.
In: Schmitt, Peter A.; Heike E. Jüngst (eds.) (2007:14–15).
Sinner, Carsten; Beatriz Morales Tejada (2015). “Translatologische Perzep-
tionsstudien als Grundlage der Bestimmung gelungener Übersetzungen”. In:
Lebende Sprachen, vol. 60, no. 1, 111–123.
Stejskal, Jiri (2009). “Quality Assessment in Translation”. In: Forstner, Martin
et al. (eds.) (2009:291–300).
Stolze, Radegundis (2011). The Translator’s Approach – Introduction to Trans-
lational Hermeneutics. Berlin: Frank & Timme (Arbeiten zur Theorie und
Thelen, Marcel (2009). “Quality Management for Translation”. In: Forstner,
Martin et al. (eds.) (2009:195–212).
Toury, Gideon (1995). Descriptive Translation Studies and Beyond. Amster-
dam and Philadelphia: John Benjamins.
Venuti, Lawrence (ed.) (2000). The Translation Studies Reader. London:
Routledge.
Vermeer, Hans J. (2006). Versuch einer Intertheorie der Translation. Berlin:
Frank & Timme.
Waddington, Christopher (2001). “Different Methods of Evaluating Student
Translations: The Question of Validity”. In: Meta Translators’ Journal,
vol. 46, no. 2, 311–325.
Waddington, Christopher (2004). “Should student translations be assessed ho-
listically or through error analysis?”. In: Lebende Sprachen, vol. 49, no. 1,
28–35.
Waddington, Christopher (2006). “Measuring the effect of errors on translation
quality”. In: Lebende Sprachen, vol. 51, no. 2, 67–71.
Wiesmann, Eva (2007). “Qualität der Rechtsübersetzung – der Beitrag von
J USLEX”. In: Schmitt, Peter A.; Heike E. Jüngst (eds.) (2007:637–647).
Williams, Malcolm (1989). “The Assessment of Professional Translation Qual-
ity: Creating Credibility out of Chaos”. In: Meta Translators’ Journal, vol. 2,
no. 2, 13–33.
Williams, Malcolm (2001). “The Application of Argumentation Theory to
Translation Quality Assessment”. In: Meta Translators’ Journal, vol. 46,
no. 2, 326–344.
Williams, Malcolm (2009). “Translation Quality Assessment”. In: Mutatis Mu-
tandis, vol. 2, no. 1, 3–23.
Wilss, Wolfram (1982). The Science of Translation: Problems and Methods.
Tübingen: Narr.
4 Quality Factors of
Translation
The translator’s task is not as straightforward as it is commonly taken

to be. Translating a source text into a target text involves countless de-
cisions at macro- and microtextual levels – decisions which, in turn, are
influenced by an indefinite number of factors.1 That is why two transla-
tions done by different human translators will hardly ever be the same:
even if the translators are given the same information about the source
text and how to translate it, they will handle that information differ-
ently in line with their respective ontogenetic backgrounds. What factors
become relevant to the translation of a particular source text depends,
then, not only on the input the translator gets but also, and perhaps to a
greater degree, on his or her ability to put such information to good use.
This chapter will reveal how the various factors may affect the task of
the translator and it will pave the way for an argumentative framework
of translation quality assessment. On this basis, we can then analyse the
evaluation reports of commented translations.
The “Translator’s Daffodil”

In her theory of translational action, Justa Holz-Mänttäri (1984) analy-
ses the elements of the translation process from an action point of view.
Here, the translator assumes a key role in the overall translation process
as the producer of the translation. He or she is an agent who co-operates
with other agents to achieve a common goal by means of communicat-
ing a message. The message is conceived of as a mental concept geared
towards the common goal (cf. Holz-Mänttäri 1984:58); its vehicle is
the message carrier, that is, the translation. Holz-Mänttäri points out
that the producer of the message carrier, when conceiving the message
carrier, has to capture and consider all relevant influencing factors and
weigh them against each other so as to find suitable compromises (cf.
Holz-Mänttäri 1984:72). An equally complex translation framework
is provided by Hanna Risku: she looks at the cognitive foundations of
translation and comes up with a model by which the translation pro-
cess can be represented as a problem-solving process. This model is used
to establish a competence profile of the translator (cf. Risku 1998:131).
Quality Factors of Translation 97
Both Holz-Mänttäri and Risku try to account not only for the large
number of elements constituting, and operating on, the translation situ-
ation in its most inclusive sense, but also for the fact that these elements,
rather than being static, are intrinsically dynamic (cf. Holz-Mänttäri
1984:38, Risku 1998:131, 2004:33). The translator, though situated at
the centre of a multitude of changing environmental factors, is subject to
the dynamism of these factors and has to develop his or her translation
competences accordingly. However, the translator’s environment must
not be reduced to a source of information and a problem space: rather, it
is part of the problem’s solution (cf. Risku 2004:72).
The central position of the “situated translator” (House 2014:117) is
also a major characteristic of Ralph Krüger’s cognitive perspective on
specialised translation. Krüger first provides an overview of the devel-
opment of translation theory, outlining “the three cognitive scientific
paradigms of symbol manipulation, connectionism and situated cogni-
tion” (Krüger 2015:273). These paradigms are aptly depicted in terms
of a computer metaphor, a brain metaphor, and an ecosystem metaphor,
respectively. Krüger then presents his “Cologne Model of the Situated
LSP Translator” (ibid.), in which he distinguishes, and specifically ac-
counts for
• the translator as the central agent of the system,

• the co-operation partners,
• the working world and professional status as social factors,
• explicit governing instruments,
• computer-aided translation (CAT) tools,
• electronic data processing in general,
• general work-related tools and aids,
• physical and psychological factors.
While clearly embracing the ecosystem approach and acknowledging the

ground-breaking work of scholars like Justa Holz-Mänttäri and Hanna
Risku among others, Krüger criticises that theories of situated transla-
tion (1) emphasise the necessity to permanently reconstruct knowledge,
thereby ignoring the benefits of the connectionist paradigm, notably, the
explanatory power of the frames-concept; and that they (2) neglect the
role of the actual text work to be done by the translator (cf. Krüger
2015:288–289). Such neglect would be wholly inappropriate in a con-
text that deals with the evaluation of translation as a product. Thus,
our own approach in this chapter combines a clear focus on translation
quality assessment with a comprehensive view of the overall translation
situation. It is depicted in a diagram dubbed the “translator’s daffodil”,
whose open-ended categories are designed to cover all those aspects of
translation that are also covered by any theories of translational action
or situated translation.
98 Quality Factors of Translation
text form
client culture
target text
politics translator
source
text
Figure 4.1 T
he “translator’s daffodil”.
The concept of the “translator’s daffodil” was first introduced in an es-

say discussing the quality of translation in subtitling (cf. Bittner 2011).
Here, the target text is at the centre of the translation process, with six
groups of factors influencing it. The resulting diagram looks like a flower
with six petals, hence the name.
The “translator’s daffodil” signifies
what the translator should be aware of when translating a text. Un-

like those diagrams that, in trying to describe the translation pro-
cess, reveal a dichotomy between the source and target texts and
their respective cultures with translation coming in between (for
two examples, see Nord 2005:38–39), the “translator’s daffodil” re-
flects the translator’s focus on the target text. [….] The reason why
any assessment of quality in translation should focus on the target
text is obvious: it is here that the quality of a translation becomes
manifest.
(Bittner 2011:77)
The grouping of factors surrounding the target text is more or less arbi-
trary. There is not necessarily a clear assignment of individual factors to
specific groups: for example, the role of the target text reader as a factor
influencing the translation process may be explicitly defined by the client
in the translation brief, or it may have to be inferred from the source
text (as will often be the case in literary translation). What is more, the
various factors also influence one another, thereby weakening or rein-
forcing their combined impact on the target text. The translator and the
evaluator in their respective assessment of the quality of a translation
will try to take in as many factors as possible.2 The more factors they
account for, the more solid can be the qualitative foundation on which
the target text rests.
Underlying Assumptions
The evaluator’s task is to emulate the translator: to look at the translation
from a perspective that follows the translator’s point of view (cf. Albrecht
2009:32). This emulation, however, implies difference rather than same-
ness in that the evaluator’s perspective constitutes a position above the
translator’s frame of reference. While the translator has to actually trans-
late a source text (from scratch, as it were), the evaluator takes the trans-
lator’s solutions as a starting point. This is the kind of vantage point from
which the evaluator judges the decisions made by the translator. Ideally,
the evaluator should be highly aware of the various factors involved in a
particular translation, thus, having at his or her disposal at least as much
relevant information as the translator.3 Moreover, he or she ought to be
as skilled a translator as the translator him- or herself. The distinction
between the roles of evaluator and translator is not as clear-cut as it might
seem at first, if we consider that the translator also usually evaluates his
or her own translation. Thus, our system of factors affecting translation
is as much an evaluator’s daffodil as it is a translator’s daffodil.
When trying to trace the six-petalled pattern of the “translator’s
daffodil”, looking at each group of factors in detail, we will be con-
stantly aware that any point made or fact stated must necessarily be of
a highly tentative nature. This is because, in their effect on the qual-
ity of a translation, the vast majority of factors cannot be reduced to
any hard-and-fast rules or definitions. As Hans Vermeer writes, there
are no invariants in translation. The principle to which he subscribes is
that of a relative relativism that extends also to theories and models of
translation (and beyond) and leaves no room for fixtures. Any convic-
tions and actions are based on assumptions, any interaction constitutes
a non-binding offer – one that can be accepted or rejected depending
on the assumptions of the recipient (cf. Vermeer 2007:174). While, in
principle, we agree with Vermeer’s position, at the same time, we also
go along with Juliane House’s more pragmatic view of translation. She
finds fault particularly with Vermeer’s concept of the text as existing
only in the various readings of individual readers: “Such a relativistic
view […] is anathema to anybody – including myself – hypothesizing
that a text embodies some autonomous meaning, and that this meaning
cannot exclusively be seen as emanating from the mind of the individual
reader” (House 1997:14–15). From our point of view, the idea of a text’s
“autonomous meaning” is not incompatible with the notion of a relativ-
ism in which there are no invariants.
If we regard both positions as the extreme ends of a continuum that
reflects the variable interplay between a text and its recipient, we can
easily see how Vermeer’s subjective relativism as well as House’s objec-
tive textual meaning can be justified. Poetic texts, for example, where
language is often used as an end in itself, would tend to be subject to
interpretation by the individual reader, thereby eliciting different re-
sponses, whereas texts that fulfil a referential function are likely to be
written in such a way that their meaning is clear to whoever reads them.
More towards the middle of the continuum, there are those texts that
can be misunderstood to a greater or lesser degree – whether due to an
incompetent writer or an incompetent reader or because of deliberate
ambiguity. When it comes to evaluating the quality of a translation, the
target text would, of course, find its place on the continuum just as any
other text. Thus, the reception of the text by the evaluator as reader
would reflect as much idiosyncratic subjectivity or general objectivity as
the text would seem to suggest. The actual process of evaluation, then,
is rendered more difficult in that the target text should be assessed in the
context of its own formation. The sheer impossibility of equally and ad-
equately accounting for all factors involved will tilt the balance towards
a more relativistic rather than objective undertaking. Therefore, our dis-
cussion, in this chapter, of individual factors within the various groups
of factors cannot be exhaustive; its aim is merely to provide a map that
will help the evaluator to get his or her bearings.
Factor Categories
The set of factors relevant to translation quality contains a subset of
factors relevant to the assessment or evaluation of translation quality,
because relevance to the quality of a translation does not necessarily im-
ply relevance to the evaluation of that translation’s quality. This reason-
ing may not be immediately obvious: for, after all, any factor that bears
upon translation quality will change that quality and, thus, have an ef-
fect on our judgement of a particular translation’s quality. However, in
evaluating a target text, we consider any factors that directly affect the
usability of that target text in the target culture; we do not consider any
factors that are irrelevant to the usability of the translation. The follow-
ing examples will make this clear:
1 A poorly paid translator produces a translation which is in some re-

spects unacceptable. We may conclude that the poor payment of the
translator affects the quality of the translation, since the translator
has to take on more translation jobs to earn his or her living and,
therefore, cannot spend as much time on one translation (and on the
quality assurance of that translation) as he or she would like to. Nev-
ertheless, when we assess the quality of that translation, poor payment
does not affect our assessment, as it is not a quality criterion. If it
were, we would have to view the translator’s poor payment in terms of
“mitigating circumstances” and judge the translation more leniently.4
2 A translator who has to work against a tight deadline may have to
go without some of the usual quality assurance measures in order to
deliver his or her translation on time. The result will be a translation
that is not as good as it would have been, if the translator had had
enough time to check the quality of the target text. As in the first
example, one would argue that, here, a tight deadline affects the
quality of a translation but not the evaluation of that quality.
3 Regarding the same situation as in the second example: if timely de-
livery of the translation is crucial to the use of that translation, then
meeting the deadline with an imperfect target text is better than
coming up with a perfect one after the deadline has expired (cf. also
Ahrend 2006:34). In this case, delivering the translation on time is
not only a quality but also a quality assessment criterion – one that
overrides all other criteria provided that the target text can be used
despite the errors and mistakes it contains.
On the basis of the above considerations, we define: A factor is relevant

to the assessment of translation quality if it is directly related to the
usability of the translation in its textual dimension. Any factors that
constitute obstacles which the translator should overcome when trans-
lating a particular source text (and which are not directly related to the
usability of the translation in its textual dimension) merely affect the
quality of a translation but not its assessment. The above definition is
tentative in the sense that there may well be cases in which it is not clear
whether or not a particular factor is relevant to the evaluation of a given
translation. However, any such cases do not detract from the general
validity of the definition.
Another distinction with regard to the quality factors concerns the
evaluator of a translation: there are those factors that are known to
the evaluator, and there are other factors that remain unknown to the
evaluator. Most obvious are the factors relating to the target text, factors
that are either directly visible in the target text (for instance, spelling and
grammar) or that can be inferred from the target text (such as the cul-
tural context). As regards the unknown factors, these are not easily pin-
pointed in a general way. They will usually extend to those factors that
guide the translator’s actions without clearly manifesting themselves in
the target text or that are also unknown to the translator (for example,
specific background knowledge about the source text – invaluable in po-
etry translation but not always available). It is clear that the distinction
between factors known to the evaluator and factors unknown to the
evaluator also applies to the translator as evaluator; only that, in general,
the number of factors known to the translator is higher than the number
of factors known to an external evaluator, because the translator is more
immediately involved in the translation process than the evaluator. For
instance, an external evaluator is usually not aware of all those factors
that concern the translator and his or her personal situation during the
production of the translation to be evaluated. Yet, while any such factors
may significantly affect the quality of the translation, they are not rele-
vant when it comes to evaluating that translation.
The juxtaposition of (1) translation quality factors that are relevant
to quality assessment and translation quality factors that are not rele-
vant to quality assessment and (2) translation quality factors that are
known to the evaluator and translation quality factors that are not
known to the evaluator is illustrated in Figure 4.2. 5 Here, the only
obvious value is the target text in its graphic and linguistic dimension
as a factor that is both known to the evaluator and relevant to the
evaluation of the translation quality. The other three values given in
square brackets are possible examples of translation quality factors
that are relevant to evaluation and unknown to the evaluator, irrele-
vant to evaluation and unknown to the evaluator, as well as irrelevant
to evaluation and known to the evaluator. These examples are based
on the assumption of a particular translation situation. For instance,
that the source text is not available to the evaluator may sometimes
be a problem of literary translation assessment (cf. Granzin 2010);
graphic and linguistic

known [qualifications of the
elements of the target
factors translator]
text
[circumstances under [source text not

unknown
which the target text available to the
factors
was produced] evaluator]
irrelevant relevant
factors factors
Figure 4.2 Factor matrix.

the circumstances under which a target text was produced are usually
neither relevant to the evaluation nor known to the evaluator (although
they do affect the quality of the target text); the translator’s qualifica-
tions may well be known to the evaluator, yet they should not interfere
with the evaluator’s judgement of the quality of the translation. On the
basis of the above factor matrix we can now analyse the various groups
of factors in more detail.
Source Text
A translated text, by definition, is always based on a source text (or, in
special cases, on several source texts).6 This fact entails a very specific
relationship between the two texts, a relationship that is often described
in terms of equivalence.7 Yet, equivalence is a rather vague and often
shifting criterion for the relative success of a translation, because both
the equivalence unit and the equivalence level can vary considerably.
Still, since there is no translation without a source text, the quality of a
target text depends – to a large degree – on the way in which the transla-
tion has been created from the source text. Both source and target texts
are at the very centre of the process of translation. Most importantly, the
source text furnishes the language from which the translator is to trans-
late. Other factors that affect the relationship between the source text
and the target text may include the intended – or implied – readers of the
source text as well as the circumstances that gave rise to the composition
of the source text. In this section, we will look at the various source
text-related factors and specify for each of these factors, whether or not
they are relevant to the evaluation of translation quality and whether or
not they are known to the evaluator.
Language Pairs in Translation and Source

Text Analysis (Nord)
The quality of a translation has to be assessed against the background of
the languages involved. Usually, there is a source language, and a target
language that is different from the source language.8 The difference be-
tween the source language and the target language is one of many factors
that affect the degree of difficulty for the translator in performing a par-
ticular translation. Translating a text into an unrelated language would
seem to be more difficult than translating the same text into a closely
related language, because in the first case the translator has to bridge a
larger gap between the respective grammatical systems of the source and
target languages than in the second case (cf. Carbonell 1996:83; Mäl-
zer 2013:263). While this argument may be perfectly logical as far as it
goes, there may be numerous exceptions in which the translation into a
related language turns out to be more difficult than the translation into
an unrelated language. This is due to the fact that translation does not
consist in a code shift from one language to another (as some advocates
of a linguistic approach to translation seem to suggest, for example, Kade
1968 or Freigang 1978) but generally involves a change from one cultural
system to another. Thus, while the language pair defined by the source
and target texts does certainly have an impact on the translation process
and, thereby, on the outcome of that process – or else, there would be no
use for language pair-related translation guides such as Wolf Friederich’s
Technik des Übersetzens (Friederich 1969) – this impact manifests itself
mostly in false friends and other encroachments of the source language
on the target text, errors which are more likely to occur when the source
and target languages are closely related or when the target language con-
tains many borrowings from the source language. In this way, the lan-
guage pair has a potential effect on the quality of a translation, though
without being relevant to the assessment of that quality.
The source text exists prior to the target text.9 Anything related to the
source text may affect that source text and thereby change the ground
on which the target text is produced. This recalls Christiane Nord’s Text
Analysis in Translation (Nord 2005) with its detailed discussion of ex-
tratextual and intratextual factors. The extratextual factors, which are
relevant to our discussion in this section, include the sender and/or text
producer, the sender’s intention, the audience, the medium or channel,
the place of communication, the time of communication, the motive for
communication, and the text function. The intratextual factors, which
will be dealt with in the section on text form, include: the subject matter,
the content, any presuppositions, the composition of the text, non-verbal
elements, lexis, sentence structure, and suprasegmental features. Though
not all of these factors will be relevant to all translations, Nord’s theory
caters for almost any kind of text and any kind of translation, making
sure that basically all texts to be translated – whether print texts, manu-
scripts, comic strips, audiovisual texts, etc. – can be thoroughly analysed
in order to achieve the best translation results.
While Nord’s text analysis is certainly very detailed, it is also partly
redundant in the sense that relevance of some extratextual factors to
translation is often established in combination with intratextual factors
or in comparison with the target text. For example, if a sender is im-
portant because he or she is known for a peculiar style of writing, then
this peculiarity will probably also show in the intratextual features of
lexis and/or syntax. Or the choice of a medium for the source text is rel-
evant to translation because that medium needs to be compared with the
medium specified for the target text: unless there is a complete change
of medium (say, from an audiovisual to a print medium), the medium
will be relevant only because it is associated with a different audience
or a different style of expression. In many such cases, then, relevance to
translation is only indirect.
Sender and Sender’s Intention
The following examples taken from Nord (2005) will demonstrate the
extent to which the extratextual factors can affect translation quality.
The first example relates to the role of the sender:
If ex-Prime Minister Edward Heath writes an editorial in a B ritish

newspaper, British readers will immediately know what political
party the author belongs to. If the text is translated and published
in the German weekly paper DIE ZEIT, many German readers may
not be able to “classify” the author as easily. If, however, the clas-
sification is relevant for the comprehension and/or interpretation of
the article, the information has to be supplied in a few introductory
lines or even in the text itself, if possible.
(Nord 2005:51)
Omission of the information that Sir Edward Heath was a conserva-

tive Prime Minister of the United Kingdom might, then, detract from
a proper understanding of the target text. Therefore, such an omission
would not only have a negative impact on the quality of the transla-
tion; it would also be relevant to the assessment of that quality because
it reduces the usability of the translation. However, it should be noted
that supplying additional information relates to what Nord calls pre-
suppositions: they “comprise all the information that the sender expects
(= presupposes) to be part of the receiver’s horizon” (Nord 2005:106).
The information about the sender is, thus, pertinent to translation qual-
ity assessment only as a situational presupposition. It is usually known
to the evaluator.
By contrast, the intention of the sender is usually not evident but has
to be gleaned from information supplied by the source text or the trans-
lation situation. Furthermore, the sender’s intention is not a factor that
impacts the quality of the target text on a stand-alone basis, as the fol-
lowing example shows:
If a text is published in a newspaper on the pages specially devoted

to political commentaries (which in quality papers is often separate
from news and reports), this medium of publication can be taken as
a clear hint that the sender’s intention was that of “commenting” on
recent political events or tendencies.
(Nord 2005:56)
Given a corresponding translation brief, the target text would also be a

commentary. What matters, here, in terms of translation quality is the
way in which the sender’s intention is captured in the target text, pro-
vided that the source text function reflects the sender’s intention. If the
source text function does not reflect the sender’s intention, the translator
will have to decide whether to stick to the function of the source text
and neglect the apparently different intention of the sender, or whether
to modify the original presentation of the message in the target text so
as to render the translated text compatible with the sender’s intention.
Such change will be particularly appropriate whenever the quality of the
source text is unacceptable. A company homepage, for instance, which is
obviously intended as an invitation to potential customers to do business
with that company, clearly needs to be translated in a way that reflects
this intention – whether or not that intention comes across in the source
text. It is in such specific cases that the sender’s intention plays a role for
the evaluation of a translation, albeit in conjunction with the actual text
function and other intratextual factors.
Audience and Medium

“Audience” is the next item on Nord’s list of extratextual factors. Usu-
ally, it is more or less known to the evaluator, concluded from informa-
tion provided by the source text or the translation brief. As “readers”,
the audience features already in the above Edward Heath example of the
role played by the sender in the translation process; and it even comes up
twice: as British source text readers and as German target text readers.
Ultimately, the difference between the source and target text audiences is
indicative of the difference between the source and target text cultures –
an aspect that will be discussed in the section on culture, below. Never-
theless, we will quote one example from Nord in which the relevance of
the audience or reader becomes particularly obvious:
In his article “Translation as a Decision Process” […] Jiři Levý refers

to his book Umění překladu. In the German version of this article,
the book is cited under the same title, although there is a German
translation available […], which would probably be of greater inter-
est to the average German reader than the Czech original.
(Nord 2005:57)
Here, it is the different language (rather than the different culture) of the
target text audience compared to the source text audience which is crucial
to an analysis of the example. While Nord’s argument is certainly valid
in its own right, it does presuppose a translation strategy in which ease of
comprehension is paramount, whereas loyalty to the source text plays a
subordinate role. After all, the English source text – like the German target
text – provides its readers only with the title of the Czech original! A trans-
lation that tries to preserve the source as a historically important text might
give the title of the German book together with further details in a footnote.
It is clear that, in this case, consideration of the audience is relevant not
only to translation quality but also to translation quality assessment. Yet,
whether the solution described in the above example would be regarded as
good or bad depends on the strategy chosen by the translator.
The medium or channel of the source text has only an indirect impact
on translation quality, that is to say, as an element that reinforces or
weakens the effect of other factors relevant to translation. Christiane
Nord writes:
The dimension or medium is relevant because it provides some clues

as to the size and identity of the addressed audience.
(Nord 2005:64)
In addition, the specification of the medium may give some clue as to

the sender’s intention (e.g. in the case of a poster or a picture post-
card) and to the motive for the communication (e.g. in the case of a
death announcement in a newspaper).
(ibid.)
For the translator it is important […] to take into account the fact
that the “same” media may have quite different functions in another
culture.
(ibid.)
As a general rule […] the medium determines the receiver’s expecta-

tions as to text function.
(Nord 2005:65)
These quotations implicitly confirm what we have mentioned above

about the relevance of the medium to translation quality assessment:
namely, that a source text medium will be relevant to the evaluation of
a specific target text, if it is different from the target text medium, either
in kind or in function. Yet, this relevance will be only indirect, with the
medium having repercussions on other, more immediate factors. The
evaluator generally knows the medium of the source text as well as that
of the target text.
Place and Time of Communication

As for the place of communication, its effect on translation quality like-
wise hinges on the difference between source and target as the following
examples show:
The distance between London and Liverpool is much “shorter” as

perceived by a Texan than by an Englishman.
(Nord 2005:68)
In the case of newspaper articles, the place where the paper is pub-
lished is normally taken to be the place of text production as well.
Therefore, readers of the Sunday Times can assume that the infor-
mation “Mortgage cut in sight” refers to Great Britain […]. If corre-
spondents send their reports from somewhere else, the place of text
production is usually specified […].
(Nord 2005:69)
The first quotation is an example of what Nord calls “relative geography”

(Nord 2005:68), which refers to the fact that, in the target culture, the
sense of place and place-related quantities may be different from that of the
source culture. This would have to be taken into account in a translation
geared towards the target culture. Note that, here, the effect of the place of
communication on translation quality is inextricably intertwined with the
expectations of the reader in his or her cultural setting. Regarding Nord’s
second example, certain implicit assumptions about the geographical scope
of a report may have to be referred to explicitly in the target text in order
to avoid misunderstanding. Failure to do so would be detrimental to the
quality of the translated text. Thus, the place of communication for the
source and target texts can certainly be a factor relevant to the assessment
of a translation. However, it may not always be obvious to the evaluator.
Like the place of communication, the time of communication cannot
always be clearly established. Where it can be established, it sometimes
turns out to be crucial to certain translation decisions. Most import-
ant in this connection is the relation between the time of publication
of the source text and any temporal references in that text. In an in-
strumental translation (as opposed to a documentary translation) (cf.
Nord 2005:80), temporal specifications that are relative to the time of
publication of the source text may have to be changed so as to fit the time
of publication of the target text. In the following example, translation
quality and the evaluation of that quality are immediately affected as the
usability of the target text depends on a readjusted rendering of time: an
article in a German local newspaper published towards the end of April
announces a specific event as happening “at the last weekend of next
month”. This article is translated into Turkish for a monthly magazine
issued at the beginning of the month in which the event referred to in
the newspaper article is to take place. It is obvious that the specification
“at the last weekend of next month” would have to be changed to “at
the last weekend of this month” (or another unequivocal specification).
Other cases in which the time of communication plays a role seem to af-
fect translation quality less severely. Christiane Nord refers to “the prob-
lems involved in translating or re-translating old texts” (Nord 2005:72).
She points out that different approaches are possible, depending on “the
prevailing translation tradition or concept” (ibid.). Here, an evaluator
would be expected to check the consistency of the translator’s approach.
Motive for Communication and Text Function
Closely related to the time of communication and to the overall trans-
lation situation is the motive for communication: why – or for what
occasion – has the source text been written? Nord acknowledges that
finding out which event has motivated a certain text is not always easy
and, therefore, “not always relevant to translation” (Nord 2005:76).
Still, she comes up with an interesting example:
On March 12th, 1984, the Spanish daily paper El País published

a commentary under the title “El Día de la Mujer” (International
Women’s Day). It is the motive for text production this title alludes
to and not the subject matter, because the text deals with the situa-
tion of working women in Spain in 1984. The newspaper reader was
expected to be familiar with the occasion, International Women’s
Day, since it had been commented on quite frequently at the time.
If the text is to be translated, it is the motive for translation (as well
as the dimensions of time and place) that has to be taken into ac-
count. Only a few days later the date will have been pushed into the
background by other events, and a title like “International Wom-
en’s Day” will arouse specific expectations about the subject matter,
which the text cannot meet.
(Nord 2005:76)
Here, the motive of communication is relevant to the quality of trans-

lation and to its assessment, yet, only in combination with other source
and target text factors such as the time of communication and the
content.
Nord’s final extratextual factor, text function, is derived from infor-
mation supplied by the “pragmatic relationships between sender, re-
ceiver, medium, and motive” (Nord 2005:82). The evaluator is usually
aware of the main text function and any subordinate functions, as they
are primarily deduced from the source text itself. Yet, they are also sup-
ported by any information about the sender, the sender’s intention, the
medium, the place and time of communication, and the motive for com-
munication. Because of the interrelatedness of text function and other
extratextual factors, it is difficult to imagine a translation whose quality
depends on the proper rendering of text function alone. A good example
has already been given in the discussion of the sender’s intention: a com-
pany homepage whose function should be to attract potential customers
but which fails to properly fulfil that function would have to be adapted
in translation to meet that requirement. Here, the intratextual factors
relating to textual style do not produce the desired effect: they need to
be optimised in the target text in order to achieve the intended conative
text function.
Source Text Defects
A rather common cause of a defective target text is a defective source
text. The defects of a source text will most likely affect translators who
conceive of translation as a code-switching exercise and, thus, produce
a target text that is as literal as possible and as free as necessary.10 By
source text defects, we refer not so much to grammatical errors or spell-
ing mistakes (as long as they do not detract from meaning), but to fac-
tual, terminological, and stylistic errors. In most translation situations,
the translator would be required to come up with a target text that works
for the target text reader. That is to say, if the source text fails to meet
the reader’s expectations, this does not mean that the target text should
also fail to meet the expectations of the reader. Particularly, in a covert
setting, the quality of the target text should be at least as good as that of
the source text – in the case of a defective source text, it should be better.
Examples of source text defects are: a wrong date given for an event
that took place at some time in the past, a wrong term used to refer to
a specific machine part, a style that fails to address the intended audi-
ence. The translator would be required to provide the right date, use the
correct term, and adjust the style of the target text so as to address the
intended audience. These examples show that a defective source text
can have a serious impact on the quality of the translation and, thus, is
very relevant to translation quality assessment. An evaluator who can
compare source and target should be aware of any source text defects:
he or she would be expected to know enough about the source text to
tell whether it contains any errors – which might require doing some
research.
Summary of Source Text Factors

To conclude our discussion of source text factors, we can say that, gen-
erally, they become relevant to TQA only in connection with the corre-
sponding factors of the target text. Under certain circumstances, they
may have an immediate influence on the usability of the translation and
therefore affect the judgement of an evaluator more or less directly. Of-
ten, though, they play only a minor, contributory role in translation as-
sessment, with factors such as lexis or syntax being to the fore. In very
special translation contexts, the source text may even be almost com-
pletely irrelevant to an assessment of the target text. Thus, Ira Torresi
writes about advertising:
Another factor which makes it conceptually difficult for translation

scholars to engage in a systematic analysis of advertising material is
the current practice, adopted by several multinational companies, of
developing local campaigns simultaneously from a brief that avoids
Table 4.1 Summary of source text factors
Factor Is the factor relevant to the evaluation of translation quality? Is the factor known to the evaluator?
Languages involved No. Different languages, it is true, will have different quality Yes.
criteria, but whether we translate an English text into
German or Chinese does not affect the usability of the
translation in its textual dimension.
Sender Only under certain circumstances and usually in combination Ideally, yes. Of course, there will also be many cases
with other factors: for example, if additional information in which the sender or producer of the source text
about the sender is crucial to a proper understanding of the is not known to the translator, let alone to the
target text. evaluator.
Sender’s intention Only in very specific cases and in combination with other Usually, yes. In many cases, the sender’s intention
factors: for example, if the sender’s intention as deduced can be inferred from other textual and contextual
from the context has not been implemented in the actual factors.
text function.
Audience Yes, but only in combination with other factors such as Usually, yes. The audience can be identified on
presuppositions. the basis of other factors. The definition of a
text’s readership can range from a very specific
readership to a very general one.
Medium Only in combination with other factors and if the target text Yes.
medium is different from the source text medium.
Place of Only under certain circumstances and usually in combination It depends. Often, the place of communication is not
communication with other factors: for example, if an implicit place of known to the evaluator.
communication needs to be made explicit in the target text.
Time of Yes, depending on time differences between the publication It depends. Often, the time of communication is not
communication of the source and target texts, and on temporal references known to the evaluator.
made in the source text.
Motive for Only in very specific cases and in combination with other Only in specific cases.
communication factors: for example, if the motive for communicating the
message of the source text is different from the motive for
communicating the target text.
Text function Yes, but in combination with other factors that constitute the Usually, yes. In many cases, the text function can be
function of the source text. inferred from other textual and contextual factors.
Source text defects Yes. Yes, provided the evaluator has access to the source
text.
culture-specificity as much as possible. In this process […] there is
no single advertisement or campaign that can be easily recognised
as a ‘source’ text.
(Torresi 2009:7)
And with respect to news gathering and dissemination, Jerry Palmer

points out that “an act of news translation undertaken at the editing
stage is frequently – if not usually – based upon more than one ‘original’
text, with these texts commonly summarized and amalgamated in the
same process as translation” (Palmer 2009:188). However, these cases of
translation in advertising and in news gathering and dissemination are
special and should not detract from the importance of the above source
text factors, since the latter mostly furnish the basis on which microtext
decisions are taken.
Keeping in mind that any categorisation of individual source text fac-
tors is subject to certain assumptions about the overall translation situ-
ation (assumptions that may turn out to be wrong), we will summarise
the findings of this section in a table.
Text Form
Under “text form”, we subsume all those criteria that concern the textual
features affected by the process of rendering a source text into a target
text. These criteria range from the text type or genre via Nord’s intratex-
tual factors (subject matter, content, presuppositions, composition of the
text, non-verbal elements, lexis, sentence structure, suprasegmental fea-
tures) to the smallest elements of language such as syllables, phonemes,
morphemes, and letters. In short, text form covers what constitutes both
the source text and the target text as texts. And while the elements of
text form are evident to the evaluator (as they should be, at least in the
translated text), most of them also have a direct impact on translation
quality and play a crucial role in translation quality assessment.
Before discussing individual text form factors, we should shed some
light on the connection between a text typology and the dichotomy of
fictional and non-fictional texts, as both distinctions – that between dif-
ferent text types and that between fiction and non-fiction – may have
a decisive effect on the way texts are translated. A text typology such
as that by Katharina Reiß (1971) – juxtaposing texts characterised by
content, texts characterised by form, and texts characterised by an
appellative function (plus audio media texts) – would count literature
under texts characterised by form. While this is generally appropriate,
the typological system fails to do justice to the essential distinction be-
tween fictional and non-fictional texts. This has been pointed out by
Greiner (2004:12, 21–22), who emphasises that the language of fiction
is a non-referential language: the question whether a statement is true or
false does not arise in a fictional text. Thus, when we translate fiction,
certain aspects of translation such as the translation brief, the target
audience, or the function of the translated text are of little or no con-
sequence to the process of translation (if we discount special cases such
as novels adapted for children): the target text, like the source text, will
be a piece of fiction, of literature aspiring to aesthetic perfection. What
is important is the overall form of the source and/or target text, its ex-
pressive force generated by the interplay of all those factors that hold the
text together and give it the potential for aesthetic effect. While these
factors are, of course, also relevant to non-fiction, here, they unfold their
intrinsic power only as part of the fictional world created in the text. The
real-world expressive or aesthetic function of a fictional text is, thus,
realised in the text’s non-referential representations as devised by the
writer or translator.
The term “text type” is slightly ambivalent: it may refer to the individ-
ual categories of texts in text typologies (such as the one by Katharina
Reiß, mentioned above) or to the types of text found in an empirical
context. In German, the former are usually called “Texttyp”, the latter
“Textsorte”. Christiane Nord aptly remarks:
[T]ext can be classified on various levels of generalization. It is there-

fore not surprising that some authors specify text types as “newspa-
per reports”, “sermons”, or “resolutions”, while others prefer a more
general categorisation into “informative”, “expressive”, or “opera-
tive” texts (cf. Reiß 1971).
(Nord 2005:78)
Which kind of “text type” we refer to will have to be concluded from

the context in which the term is used. For the purpose of discussing
text types as factors of text form, however, we will take the term with
its more specific meaning. Our discussion will first focus on Chris-
tiane Nord’s intratextual factors, before turning to some text type
examples. For each of the text form factors, we will specify if they are
relevant or irrelevant to the evaluation of translation quality; we will
not specify if they are known to the evaluator, because that depends
on whether the evaluator has access not only to the target but also to
the source text.
Intratextual Factors
Christiane Nord’s intratextual factors of source text analysis are inter-
dependent in the sense that no factor can be discussed without at least
implying the influence of other factors. With such a proviso in mind, we
will now conceive of examples in which one specific intratextual factor
plays a dominant role in a given translation situation.
Subject Matter and Content
The first factor is that of subject matter, which is usually inferred from
the text content. Here, the relationship between the title and content of
a text is of particular interest. While, in many cases, a straightforward
rendering of the title is most appropriate, some titles are too idiosyn-
cratically suited to the source language and/or culture to be translated
into a near-literal title. In English, journalistic reports, for instance, may
sometimes feature titles that are grammatically complete sentences. In
German, a more succinct title is generally preferred in such a case.
When it comes to translating the titles of literary works or films, we
often find that the translation has nothing, or very little, to do with the
original. For example, how should the title of Heinrich Böll’s short story
“Anekdote zur Senkung der Arbeitsmoral” be translated into English?
The relevant criteria can be outlined as follows:
1 As the short story can be regarded as a literary classic, the overall

translation strategy should be overt, that is, the target text reader
would expect the translation to afford him or her a glimpse of the
original.
2 The level of equivalence between the target and the source text
would be that of aesthetic and stylistic achievement.
3 The translator may, therefore, have to forego a literal translation, if
such a translation fails to do justice to the aesthetic or stylistic merits
of the original.
4 What clearly needs to be preserved in the target text is the semantic
link between the title and the subject matter or content of the short
story.
5 Stylistically, the title of Böll’s short story is typically German, char-
acterised as it is by the combination of a noun and a prepositional
phrase, whose regular sequence of unstressed and stressed syllables
makes for pleasant reading.
The central question, here, is whether the translator should present the
target text reader with a translation that merely imitates the grammat-
ical patterns of the German original or with a translation that tries to
capture the German reading experience.
Leila Vennewitz opts for an almost literal rendering, “Anecdote Con-
cerning the Lowering of Productivity” (Böll 1986:628). This is a rather
clumsy title as opposed to the smooth nominal style and regular lilt of
the German original. My own suggestion would be a translation that
captures the semantic essence of the story, preserves the rhythmicity of
the original title, and adds a typically English flavour: “To Work or Not
to Work” (Bittner 1997a:91). Thus, for the translated title to comply
with the requirements of the target language, the translator has to ensure
that the title relates to the subject matter of the text; if it fails to do so,
then this will have a negative impact on the quality of the target text and
directly affect the assessment of the translation.
Content is an important analysis factor, because it is here that the
reader (that is, the translator) comprehends the text at the global level
not only with its denotative but also with its connotative meanings.
Failure to fully understand these meanings may result in a translation
that is not as good as it might be. At this point, content is inextricably
linked with another factor, presuppositions, since comprehension is de-
pendent on the reader’s horizon. Christiane Nord illustrates this with
an example:
Meisl comes to Vienna on business for the first time in his life, and
in the evening he wants to go to see a play at the famous Burg The-
atre. So he asks the lady in the booking office: “What is on tonight?”
And she answers: Twelfth Night or What You Will. “Oh well,” says
Meisl, “I would prefer The Blue Danube.”
(Nord 2005:97)
The effect of this joke hinges on a triple misunderstanding: (1) Meisl

obviously does not know “that Twelfth Night or What You Will is the
title of a Shakespeare play” (Nord 2005:97); (2) he is also not aware
“that it is not usually the theatre audience that decides on which play is
going to be performed” (ibid.); and (3) he fails to realise “that the Burg
Theatre is not a theatre where operettas are performed” (ibid.). Crucial
to a proper understanding of the joke is the knowledge that “What You
Will” is not to be taken literally. Thus, in the text, Meisl must remain
ignorant of the fact that What You Will is a play by Shakespeare; in
other words, a translator who fears that the target audience may not
understand the punchline cannot simply add the required information.
Even if such information were marked as extraneous, it would spoil the
joke. The only possible solution is an explanation given in a footnote.
This shows that the translator and the evaluator have to consider not
only the assumed understanding of the target text reader but also the
wider translation situation for them to be able to decide whether addi-
tional information should be provided at all, and if so, how this should
be done. While adding explanatory information is usually not a problem
in journalistic texts, for example, such a procedure would generally be
frowned upon in literary translation. The potential mistakes a translator
can make in relation to presuppositions are, then, of two kinds: either
failing to supply information that is necessary for the target audience
to properly understand the meaning of the text, or giving an additional
explanation in the text where such a procedure is not warranted. In this
way, content and presuppositions directly affect translation quality and
translation quality assessment.
Text Composition and Non-Verbal Elements
The next factor among Christiane Nord’s intratextual factors is text
composition. How a text is composed is relevant to translation quality
particularly in those cases in which the macrostructure of the source text
would have to be modified in the target text because of culture-specific
conventions. For example, the position of any warnings to be included in
a German user manual must comply with the requirements specified in
the relevant standard – whatever their position in the source text. Text
composition is also important when several translators translate differ-
ent parts of one and the same source text.
The German version of the textbook on linguistics edited by André

Martinet […] was produced by two translators: Chapters 1 to 25
were translated by I. Rehbein, and Chapters 26 to 51 by S. Stelzer.
Each of the chapters is an independent text and, at the same time,
part of a larger unit, whose characteristics have to be taken into
account by both translators.
(Nord 2005:111)
In translation projects in which two or more translators are involved in

the translation of a larger text, translation quality may be detrimentally
affected if the translators fail to co-ordinate their terminology choices.
Here, text composition is complemented by lexis as the most immediate
criterion for translation quality assessment.
Non-verbal elements are defined as “[s]igns taken from other, non-
linguistic, codes, which are used to supplement, illustrate, disambiguate,
or intensify the message of the text” (Nord 2005:118). They include “pho-
tos, illustrations, logos, special types of print, etc.” (ibid.). Whenever such
elements are subject to different representations in the source and target
cultures, the translator may have to make the corresponding changes – at
least in a covert translation setting. Types of print, for example, which
are medium-specific rather than culture-specific, would be adapted ac-
cording to the requirements of the target medium. In this case, one might
argue, though, that selecting the right font and font characteristics for
the target text is not necessarily the task of the translator but should be
done by the editor. Non-verbal elements, thus, have only a limited impact
on translation quality and may often be neglected in the evaluation of a
translated text. However, they are absolutely essential to the translation
of comic strips or audiovisual material, which will be discussed below.
Lexis, Syntax, and Suprasegmental Features

The most obvious factor relevant to translation quality and translation
quality assessment is lexis. A word from the source text may easily be
rendered in a way that is not quite adequate. This could be a matter of
meaning, of register, or of carelessness: the translator might misunder-
stand a particular expression in the source text, might fail to observe the
intended register or simply overlook an important word or get the spell-
ing in the target text wrong. One particular aspect of lexis, terminology,
assumes utmost significance in specialised translations. This involves
not just finding the right signifier for the right signified but also several
other issues. What criteria determine the selection of a particular term,
if there is more than one option to choose from? The following possible
factors instantly spring to one’s mind: register – one option is colloquial,
another option is formal; corporate language – even if one’s own solu-
tion appears to be terminologically better than the alternative expression
used in the company, the latter may have to be selected for the sake of
uniformity; consistency – the question when a varied use of words is
preferable to a consistent use of words may not only depend on text type
and word class but on the individual translation context. What is more,
typical collocations should also be observed, particularly, in specialised
translations. This concerns set phrases as well as those collocations in
which one component can be replaced by another without rendering the
expression as a whole unintelligible.11
Compared with lexis, the sentence structure of the source text is a
factor whose relevance to translation quality is less pronounced, if we
leave out obvious mistakes. This is due to the fact that the syntax of the
source text can sometimes be translated into different sentence struc-
tures without conceding an immediately recognisable loss of quality.
Still, the translation of sentence structures requires sophisticated con-
siderations. Does a particular sentence structure in the target language
have the same effect as the corresponding sentence structure in the
source language? Should an ambiguous syntax be rendered in a way
that reflects this ambiguity, or should it be disambiguated? What about
theme-rheme structures and figures of speech – how important is their
retention in translation? It is clear that in order to find answers to these
and other syntax-related questions, the translator and the evaluator will
have to check the overall translation context for relevant quality criteria.
Like syntax, which can be used as a means to foreground almost any
information in what is known as cleft sentences, the suprasegmental fea-
tures help to highlight certain aspects of a text, “framing [its] phonolog-
ical ‘gestalt’ or specific ‘tone’” (Nord 2005:132). Prosody and intonation
are not just applicable to spoken language but may play a role also in
written text; they are not limited to poetry. Among the devices by which
texts indicate prosodic or intonational emphasis (that is, deviations from
the norm), Christiane Nord includes syntax, punctuation, inverted com-
mas, italics, phonetic spelling, and even sound patterns such as ono-
matopoeia, alliteration, and assonance (cf. Nord 2005:133–138). While
intonation is certainly present in all texts, it is relevant to translation
only in those cases in which a particular pattern of intonation stands out
as unusual. Here, the translator has to carefully assess the necessity and
possibility of producing an equivalent pattern in the target language. In
a source text featuring a predominantly informative function, the occa-
sional occurrence of alliteration or assonance may be altogether fortu-
itous, with no particular effect intended. To try to reproduce the same
or a similar sound pattern in the target text would then be not only an
unnecessary effort, but might even result in a less satisfactory solution.
If the suprasegmental features of a source text are conspicuous in any
meaningful way, they are relevant to translation, to translation quality,
and to translation quality assessment.
Selected Text Types

As has been pointed out in Chapter 3, the genre or text type is often key
to a translator’s decision about his or her translation strategy. This is
because there are cultural norms and conventions that govern how we
translate a specific text type. In Germany, for example, a novel – at least
one of acknowledged literary excellence – would be translated with par-
ticular attention paid to the style of the source text. Thus, the fact that
the text to be translated is a classic novel predefines the translator’s op-
tions of a translation strategy. On the basis of the source text genre and
against the background of any relevant translation conventions, an eval-
uator may, then, dismiss as inappropriate a translation that obviously
ignores such conventions for the genre to be translated. In another exam-
ple, an English document (of whatever kind) which needs to be produced
in court in Germany and, therefore, counts as a legal document has to be
translated into German in such a way that the judge gets a precise idea of
what the original is like, both in form and in content. Here, the text type
(legal document) combines with the translation purpose (production in
court) and the pertinent conventions to implicitly outline the overall ap-
proach for the translator. If the assessment of the translation reveals that
the translator has not followed the required path of translation, the tar-
get text is possibly useless for the purpose for which it was made. In both
examples – novel and legal document – text type as such is not a factor
that bears on translation quality directly. However, if we include a wider
context in which the text type is regarded in conjunction with both the
source text and the target text, and in which any translation conventions
function as a yardstick against which to gauge the actual translation in
its textual form, then, the genre or text type can be counted among the
factors relevant to translation quality assessment.
Poetry, Drama, and Comics

One text type in which prosody and intonation assume a key role is
poetry. Here, even the smallest constitutive elements of language are
important: syllables, morphemes, phonemes, and sometimes even letters.
Their individual attributes help to produce rhyme, assonance, rhythm,
and metre.12 Poetry whose visual image of clauses, phrases, words, and
letters is laid out on a page to create meaning is usually referred to as
concrete poetry. What makes any kind of poetry difficult to translate –
and even more difficult to assess in translation – is the dominance of the
poetic function.13 The implicit aspiration of the poet to some remote aes-
thetic ideal makes the analysis of poetry and its translation an extremely
complicated undertaking.14 The translator’s task is to try and capture
that undefinable something which constitutes the poetic quality of the
poem – a task that can be compared to the futility of trying to catch the
colours of a rainbow: clearly determining the elements that make up
the overall aesthetic impression is impossible, as the poem will always
be more than the sum of its parts.15 This insurmountable obstacle not-
withstanding, the translator still has to analyse what he or she can find
in the poem in terms of linguistic and poetic features and also in terms
of meaning.
A translator who tries to account for as many of these features as is
practically possible hopes to create in the target language an approximate
reflection of the poetic ideal expressed in the original. If successful, this
kind of translation will also have some aesthetic value in its own right.16
However, being aware of the above dilemma that the source text ideal
cannot be achieved by a target text that tries to reproduce the features
of the original, the translator can also opt for a different strategy: he or
she may regard the translation of the poem as the recreation in the target
language of the original’s overall aesthetic impression. How this overall
aesthetic impression is to be recreated is up to the translator as a poet:
any rhyme and metre may be changed or discarded altogether, other
poetic features may be introduced, and the meaning may be modified.
It is clear that such a translation (and, here, our definition of translation
has to be stretched to infinity) can no longer be judged on the basis of a
comparison with the source text, but has to be accepted and assessed as
an independent poem. The fact that another poem in another language
gave rise to the creation of the poem in the target language is no longer
relevant to the assessment of that poem’s quality. The evaluation of any
poetry that has been translated in such a way is as subjective as the eval-
uation of any original poetry. Only a target poem that acknowledges its
origin by retaining at least some formal features of the original can be
assessed on the basis of those criteria by which the translator intended to
translate the source poem. In general, these criteria will always include
the formal aspects of poetic design and the rendering of meaning. What
counts in the end is the extent to which the translator succeeds in merg-
ing form and meaning into a satisfactory whole.
The language elements to be accounted for in poetry translation may
have to be complemented by theatrical components, when a play written
in verse is to be translated. The situation is similar for a play composed
in prose: here, the aspects typical of the theatre may add to the difficulty
of translating a piece of literature. In principle, there are two possibili-
ties, as Anderman points out:
Unlike the translation of a novel, or a poem, the duality inherent in

the art of the theatre requires language to combine with spectacle,
manifested through visual as well as acoustic images. The translator
is therefore faced with the choice of either viewing drama as litera-
ture or as an integral part of a theatrical production.
(Anderman 2009:92)
If viewed as literature, the play in the target language will be translated

as any other piece of literary prose or verse. However, if the play should
be translated for the stage, possibly for a particular performance, the
factors relevant to the translation process and, thus, to the quality of
the translation will have to cover all the opportunities and limitations
arising from the speeches of the characters (the primary text consisting
of dialogues and monologues) as well as from stage directions and the
stage setting (specified in the secondary text or by the stage director).
This additional set of stage-related factors complicates the task of the
translator, but it does not necessarily make it more difficult: what the
audience can see on stage may not have to be referred to in what the au-
dience can hear. What counts in the end is the performability and com-
prehensibility of the translated play. Its ultimate success may, in some
cases, be determined by the response of the audience, which can, thus,
become an indicator of translation quality.17 The evaluator of a play
translated for the stage would definitely have to watch the performance
in order to be able to make a sound judgement and come up with a
well-argued review.
Like the translation of a play for the stage, the translation of com-
ics involves additional visual elements.18 Far from being a mere extra,
these elements – the panels – constitute the essence of the comic strip
(cf. Reinart 2014:248). They are an integral part of the story told – and,
in general, they must not be changed. If the source text is displayed in
speech bubbles of a given size, the target text usually has to fit in the
same speech bubbles. Whether this space is sufficient depends on a va-
riety of factors: the grammatical requirements of the target language
compared with the source language, the font and font characteristics
stipulated for the target text in relation to those given in the source text,
and last but not least the semantic opportunities afforded the translator
by the interplay of panel picture and text. These semantic opportunities
are rivalled by semantic restrictions: on the one hand, what is depicted in
the panel need not be mentioned in the text; on the other hand, whenever
the source text makes an essential reference to the scene in the picture,
there is little scope, if any, for the translator to replace a source-culture
specific by the corresponding, yet, different specific of the target culture.
For example, a panel showing a bull in a china shop and a source text us-
ing the idiom “like a bull in a china shop” could not easily be translated
into German, as the German equivalent features an elephant in the china
shop. In the translation of comics, the translator always has to compare
the information given in the pictures with the information given in the
texts and, on the basis of that comparison, establish the story of the
comic strip, before trying to recreate the same story in the target lan-
guage for the target culture by refilling the speech bubbles or any other
dedicated text spaces. An evaluator will have to consider the space limits
of comic translation when judging the quality of the target text.
Audiovisual Translation
Audiovisual translation comprises a variety of translation forms includ-
ing subtitling, audio description, dubbing, voice-over, and their varied
applications.19 They all have in common a medium that transports a
combination of audio and visual elements.20 It is this combination that
the translator has to analyse before he or she can begin to translate: be-
cause the elements that need to be translated are always embedded in a
context of other elements that are not translated. Therefore, the transla-
tion has to be done in such a way that the relation between the translated
elements and those that remain untranslated is left intact. While audio-
visual translation is relevant to several different media such as film (in-
cluding feature films, documentaries, and live broadcasts), the theatre,
the opera, dance performances, and museum exhibitions, we will focus
on interlingual and intralingual subtitling as well as audiodescription in
a film context.
The analysis and evaluation of any of these types of audiovisual
translation has to take as a point of departure the polysemiotic envi-
ronment of the film medium: it comprises audio elements that are verbal
or non-verbal as well as visual elements that are verbal or non-verbal.21
This can be depicted in an overview:
• audio + verbal (language A) → for example, dialogue, voices, and

songs
• audio + non-verbal → for example, sound effects and music
• visual + verbal (language A) → for example, captions, banners, and
other text displays
• visual + non-verbal → that is, the moving image of the film
In film translation, we distinguish three different target groups: (1) peo-

ple who do not understand the language of the film, (2) people who can-
not hear the audio elements of the film, and (3) people who cannot see
the visual elements of the film. Interlingual subtitling, intralingual sub-
titling, and audio description, respectively, help to satisfy the different
translation needs of these target groups. Thus, in interlingual subtitling,
all verbal elements of the film – whether audio or visual – are translated
from the language of the film (language A) to a language understood by
the target audience (language B):
• audio + verbal (language A) → visual + verbal (language B)

• visual + verbal (language A) → visual + verbal (language B)
Since most of what needs to be squeezed into the limited space of the
subtitles is the content of the spoken dialogue of the film and since the
reception of spoken language generally takes less time than the reception
of written language, the translator and subtitler usually has to condense
the source material in the translation process. This is also true of in-
tralingual subtitling. Here, the translation renders audio code as visual
code:
• audio + verbal (language A) → visual + verbal (language A)

• audio + non-verbal → visual + verbal (language A)
The translation process of intralingual subtitling is, in a sense, reversed

in audio description, where the visual is rendered audible:
• visual + non-verbal → audio + verbal (language A)

• visual + verbal (language A) → audio + verbal (language A)
While audio description consists mainly in describing the moving image

of the film between dialogues, any text that appears in the film will also
have to be described. After this brief overview, we will now look more
closely at the various factors that have an effect on quality and/or the
evaluation of quality in interlingual subtitling, intralingual subtitling,
and audio description.
As has been demonstrated in Bittner (2011), the quality of interlingual
subtitles is as much subject to the whole gamut of quality factors as any
other translation. Only that the relevance of these factors is qualified
by the presence of several technical issues that must not be ignored by
the translator. These issues include all those aspects of subtitling to do
with the timing, placement, size, format, and sequence of the subtitles.
A subtitler who fails to pay attention to the technical issues is likely to
produce a negative response from those who are supposed to read the
subtitles. Table 4.2 provides a summary of the technical issues relevant
to subtitling and of what may happen if they are ignored.
Many technical parameters will be pre-set in the subtitling soft-
ware, so that the subtitler can focus on the linguistic side of subtitling.
Table 4.2 T
echnical issues relevant to subtitling
Technical issue relevant to subtitling Possible effects if the issue is disregarded
The subtitle should appear with the onset of the subtitled dialogue. If the subtitle appears significantly before the onset of the
subtitled dialogue, the audience may be irritated as the subtitle
spoils a dialogue-free sequence of the film.
If the subtitle appears significantly after the onset of the subtitled
dialogue, the audience may be wondering if the first part of the
dialogue is actually covered by the subtitle.
The subtitle should be displayed long enough for the audience The audience might miss out on important information, if they
to be able to read it completely (minimum: about 1.5 seconds, cannot read the whole subtitle. Subtitles that stay on the
maximum for a two-line subtitle: about 6 seconds). screen for more than six or seven seconds may be perceived as
disturbing the enjoyment of the film (cf. Reinart 2014:269).
Ideally, subtitles should not go over shot changes, or, if they have Any overlapping of subtitles and shot changes (particularly, if
to, they should remain visible for a certain time after the shot it lasts only for a fraction of a second) may be confusing to
change (for example, one second or half a second). the audience, making it more difficult for them to follow the
information provided in the subtitles.
Subtitles are usually restricted to a maximum of two lines with Subtitles taking up much space diminish the pleasure derived
37–40 keystrokes each, depending on the size of the screen and from the visual image of the film.
the font size.
Subtitles should be placed on the screen in such a way that they Subtitles violating this placement rule diminish the pleasure
impinge as little as possible on the picture of the film. In most derived from the visual image of the film.
cases, the subtitle is, therefore, placed at the bottom of the
screen with the shorter of two lines above the longer one.
The layout of subtitles should be such that a dialogue within Failure to mark a dialogue that occurs within a subtitle will
one subtitle is immediately recognised. This can be done by render comprehension of that subtitle more difficult.
centre alignment in combination with horizontal strokes at the
beginning of each line.
The subtitles have to be presented in such a way that they can be Letters displayed without an outline or a box as background
read without difficulty, a central question being whether the may be very difficult to read whenever the screen picture
letters stand out from the background of the picture on the underneath the subtitles assumes a colour similar to that of the
screen. subtitle letters.
There should be a certain minimum distance (for example, of four If one subtitle follows another without an interval in between,
frames) between two subsequent subtitles. the viewer’s eye may fail to “register the appearance of a new
subtitle” (Ivarsson/Carroll 1998).
Yet, the restrictive environment calls for a number of additional criteria
that need to be observed. They are summarised below:
• Ideally, the sequence of subtitles should account for the viewer’s

reading rhythm and the rhythm of the film.
• Subtitles should be grammatically correct.
• The syntax should be easy to follow; higher syntactic boundaries
should coincide with the end of a subtitle or subtitle line.
• There should be a close correlation between the film dialogue and
the subtitle, as the audience may also have some knowledge of the
source language.
• Basically all spoken language is subtitled, even if it contains redun-
dant information.
• Written language in a film (posters, banners, signs, etc.) should also
be subtitled.
(For most of the above criteria, cf. Ivarsson/Carroll 1998.)

It is clear that non-compliance with these rules either increases the
difficulty of reading the subtitle or leaves the audience at a disadvantage
as to the content of the film. The subtitler, in addition to condensing the
spoken dialogue into written language, will have to make some com-
promises when attempting to reconcile all of the above rules plus any
general translation criteria. The evaluator’s task is to understand the
subtitler’s implicit preferences and to judge the subtitles on a broad basis,
weighing any substandard translation against at least one other well-
argued alternative.
The quality factors relevant to interlingual subtitles are also relevant
to intralingual subtitles, that is, subtitles for the deaf and hard of hear-
ing. 22 Nevertheless, subtitling for the deaf and hard of hearing is subject
to additional restrictions. First of all, one major problem arises from the
fact that the target group of intralingual subtitles is inhomogeneous.
Those who are deaf from birth often learn sign language as their mother
tongue; their reading skills may be limited. Those who become deaf or
hard of hearing later in life still remember the language they have grown
up with; as a result, their reading and lip-reading skills are often bet-
ter than those of the first group. In order to accommodate both target
groups, subtitlers have to make compromises that balance the different
needs. 23 The common ground, here, comprises the subtitling of sounds
and noises that cannot be seen and sometimes marking long stretches of
film that are without any sound or dialogue. In some cases, the subtitles
or subtitle lines indicate the respective speaker through different colours
or shifting, that is, with the subtitle placed under the character to be
identified. 24 Yet, other measures to improve the comprehension of films
for the hearing-impaired are less unequivocal. For example, while re-
placing difficult expressions with simpler synonyms might seem a good
idea for deaf people with lower reading skills, this technique may be
slightly annoying to people who depend on lip-reading or who can still
hear the occasional fragment of what is being said by the protagonists
on the screen. The translator and the evaluator have to be aware that
subtitling for the deaf and hard of hearing may sometimes mean cop-
ing with the impossibility of serving two masters. Ideally, the evaluator
should watch the subtitled film as a deaf or hearing-impaired viewer
would watch the film: unless the evaluator is hearing-impaired, he or she
might switch off the film’s audio component.
Audio description is about as difficult to evaluate as subtitles for the
deaf and hard of hearing. In this intersemiotic form of translation, the
audio describer is expected to follow a variety of different rules telling
him or her what to describe and how to describe it. However, these rules
may vary to some extent depending on who issues the corresponding
guidelines.25 The following criteria are, therefore, to be applied with
a good deal of common sense, based on the actual film sequence to be
described:
• The description of the visual elements in a film should cover what

is necessary for a proper understanding of the story; however, the
description should not encroach upon the film’s atmosphere.
• In principle, the film should be described between dialogues.
• The audio describer is supposed to describe from a neutral, third-
person point of view. There should be no explicit narrative perspec-
tive, for example, through expressions such as “we see”. 26
• The text of the audio description should be spoken in an unemo-
tional voice to keep the audio narrator in the background: the film’s
narration should unfold unimpaired by the audio description.
• The description itself has to be factual, with as little interpretation
as possible.
• The individual words chosen for the description should be specific
and to the point. Vague, general expressions are to be avoided.
• A simple sentence structure is preferable to a complex sentence
structure and the tense to be used is usually the present tense.
• Colours should feature in the description since, by association, blind
or visually impaired people do have a sense of what the colours
mean.
• In general, information should be given only when it appears on
the screen, not beforehand. This is particularly important in cases,
when such information would give away the plot.
• Filmic expressions (such as those relating to camera perspective)
should be used with care or avoided altogether.
Obviously, some of the above criteria cannot easily be assessed objec-

tively. Especially, the first criterion refers – rather vaguely – to the actual
art of audio description, which consists in capturing the elements of a
film scene in such a way that the gist of that scene is conveyed without
omitting any essential elements or confusing the audience by provid-
ing too detailed information. The main task of the evaluator is to judge
whether any deviation from the rules is justified by the situation in the
film, and whether the effect thus produced is better than that of the
rule-abiding alternative. To assess the audio description of a film (or,
indeed, any other audio description), the evaluator should be able to
assume the perspective of a visually impaired listener, or at least be very
familiar with that perspective.
Summary of Text Form Factors

Before summarising the results of this section on text form, we will take
a cursory glance at one vast area of translation-related activity that has
hitherto been altogether excluded from our discussion. The area we are
referring to is that of localisation, defined by Anthony Pym very broadly
as “the adaptation and translation of a text (like a software program)
to suit a particular reception situation” (Pym 2004:1). The question is
whether, as Pym argues, translation is part of the concept of localisa-
tion, or whether the two exist side by side with the occasional overlap:
For us, the concept of localization includes translation. This could

be a contentious issue, since it can equally be argued that translation
is a set of practices that only occasionally intersects with localization
processes (software is often localized but not entirely translated),
and straight one-on-one translation processes can often be found
without any visible context of localization (as in the translation of
a novel). 27
(Pym 2004:3)
We would agree with Pym’s second proposition rather than going along
with the somewhat revolutionary idea (resulting from the unconditional
subordination of all translation to the concept of localisation) that
translation theory may as well be “re-baptized as localization theory”
(Pym 2004:57). From our point of view, localisation usually involves
translation, and that translation can be, and should be, evaluated as
one requirement of many localisation requirements (cf. also Schubert
2007:113). The language side of the translation has to be analysed in
the context of the ever moving distribution process in localisation. Here,
“[a] localized text is not called on to represent any previous text; it is
instead part of one and the same process of constant material distri-
bution, which starts in one culture and may continue in many others”
(Pym 2004:5). Economic factors, technical issues, and project manage-
ment needs may often dominate the overall quality assurance process,
assigning to translation the subservient role of an accessory whose claims
to qualitative excellence are dwarfed by comparison. Yet, marginalisa-
tion within the localisation process need not imply disregard for quality
issues from a theoretical perspective. The translation part of localisation
would have to be assessed within the framework of all criteria – whether
linguistic, economic, or otherwise – that affect the actual target text.
The intricacies of translation as part of localisation lie, however, beyond
the scope of this chapter.
Our summary of text form factors does not include a table in which
each factor is listed with an explanation as to whether or not it is rel-
evant to the evaluation of translation quality and whether or not it is
known by the evaluator. This is, because all text form factors are, or can
be, relevant to the evaluation of translation quality – albeit often in com-
bination with other factors. Since text form factors, by definition, are
manifest in – and, to some extent, constitutive of – both the source and
target texts, they play a decisive role in the assessment of a target text’s
quality. The text form factors of the target text are known to the evalua-
tor; those of the source text should ideally be known to the evaluator. An
evaluator who does not have access to the source text or who disregards
the source text may not be able to adequately judge the quality of the
target text (particularly, in literary translation).
Client
The role of the client in translation quality assessment covers more than
just ordering from a translator the translation of a particular text. 28 The
client qua client may stipulate that the translation, as a product and/or
process, meet certain specifications. This is often particularly important
in technical translation. 29 With reference to this context, Klaus Schubert
locates the client as one agent among others in the overall process of
translation:
Especially in the technical field, the translation work is additionally

controlled. The controlling influences can originate from any other
agent directly or indirectly involved. These include the initiator
(the agent who orders the translation to be made), the informants
(agents from whom researched information is obtained or who au-
thored it), members of the translation team, agents carrying out
secondary tasks such as terminology work, but also agents outside
the translation process proper, such as the speciality community
or industry whose best practice rules exert an influence, authors
of textbooks and handbooks in technical translation, teaching
staff in translator training, standardizing bodies, authorities and
legislators.
(Schubert 2010:353)
Here, the “initiator” corresponds to our concept of “client” in the physi-
cal sense of a person or organisation that triggers the translation process
through a translation order. We are now interested in how such a person
or organisation may contribute to translation quality.
A closer look at the above quotation shows that apart from the agents
most directly involved in the translation – the “initiator” and the “mem-
bers of the translation team” – there are other agents who exert influ-
ence on the translation mostly either through the translation team (the
translator) or through the initiator (the client) or through both. In fact,
many of the ancillary agents – from informants via terminology experts,
best practice proponents, and standardising bodies to authorities and
legislators – may play a role as a force influencing both the client and the
translator. (Only the authors of textbooks and handbooks on technical
translation as well as the teachers in translator training would seem to
influence primarily the translator rather than the client.) The informa-
tion provided to the client through any of these agents, then, may have
a direct impact on the translation and, possibly, also on the quality of
the translation.
Client Roles
The role of the client is more complicated, if we consider that the trans-
lation can also be ordered from a translation agency, which, in turn,
orders the translation from one of its translators. Here, the translator
deals with the translation agency as his or her client, whereas the agency
deals with the initiator of the translation process as its client. Any client-
related translation quality factors will add up on their way from the
initiator-client via the agency-client to the translator: thus, a tight dead-
line stipulated by the initiator will end up with the translator as an even
tighter deadline, because the agency also needs time to check the target
text before delivering it to the original client. Supporting information
may be provided by either client; what is supplied by the initiator will
reach the translator via the agency. The general impact of deadlines and
other factors on translation quality and, perhaps, on the assessment of
that quality is likely to be more pronounced in a translation situation
with a translation agency acting as the translator’s direct client.
Despite the many different ways in which the client may have an effect
on translation quality, the client’s role is not necessarily an essential role.
While a client is central to translation in a business context (since no-
body would translate, say, a financial report, unless someone asked them
to), the client all but disappears when someone translates a poem just for
pleasure. In that case, the role of the client might be said to be located
in the person who translates, so that the usually distinct roles of client
and translator coincide: here, the translator is his or her own client, just
as the translator always is – or should be – the first evaluator of his or
her translation. While, in an ordinary translation situation, the client
specifies the translation task at hand and the translator tries to fulfil that
task to the best of his or her ability; in a translation situation without
a client, the specifications are defined by the translator. The success of
the translation depends, then, on the degree to which the translator can
meet the self-imposed translation requirements. Where the translator is
his or her own client, the quality factors discussed in the following sec-
tions are applicable only in a limited way, if at all.
Deadlines
When ordering the translation of a source text, the client often specifies
the day, and sometimes the time, by which the target text has to be deliv-
ered. Setting a deadline may have a negative effect on the quality of the
translation, if it allows the translator only a minimum time span within
which to complete the translation. A deadline that gives the translator
plenty of time will not affect translation quality.30 There is no doubt that
a translator who has to rush in order to finish the translation on time is
more likely to produce a defective target text than a translator who has
enough time also to go through the quality assurance process for the
translated text. Of course, this reasoning is valid only if all other para
meters remain equal; that is to say, we must not compare, for instance,
the performance of two translators with different translating abilities.
What may happen, if the client stipulates a deadline that can hardly be
met, even by a professional translator using state-of-the-art CAT tools,
will be analysed with an example.
A company requires the translation of a product catalogue from
German into English. On Thursday at around ten o’clock in the morn-
ing, they place a corresponding order with a professional freelance
translator. As the English catalogue is to be presented at an industrial
fair in Scotland on Saturday and before that needs to be proofread and
printed, the deadline for the translator is Friday, 11 am. The amount to
be translated leaves the translator with three options: (1) working al-
most without interruption for 24 hours, (2) working very hard but eating
and sleeping as necessary, and (3) outsourcing part of the task to other
translators (if agreed by the client) and putting everything together on
Friday morning. Using the first option, the translator will probably meet
the deadline; however, his or her performance will suffer from mistakes
made due to lack of sleep, so that the target text might be in need of revi-
sion by the client. With too little time in the second option, the translator
may have to dispense with some research and other quality assurance
measures, so that, again, the quality of the translation is not as good as
it could be. The third option might be the best, provided that the other
translators are reliable and that all translators share their terminology
solutions as they translate. Yet, there will still be plenty of work to do
for the managing translator in terms of harmonising the various parts
of the catalogue, turning them into one consistent target text. If, in any
of these options, the deadline cannot be met (say, the company gets the
translated catalogue only at 1 pm), then the catalogue may have to go di-
rectly to the printer’s without having been properly proofread. Whatever
the procedure, the company ordering the translation of the product cata-
logue at such short notice is clearly putting the quality of the translation
at stake. Insufficient time for translation is, thus, likely to be detrimental
to the quality of the target text.31 However, as this factor does not di-
rectly show in the translation (what shows in the translation are errors
and mistakes, which could also have been caused by something else),
insufficient time does not count as a criterion when it comes to assessing
the quality of the target text.32 What is more, an evaluator is possibly
not aware of the time problem (unless it is the client or an agent directly
involved in the translation process who does the evaluation) – a consid-
eration that holds for most quality-related factors to do with the client.
Glossaries
Apart from setting the deadline, some clients also give the translator ad-
ditional information. This may include a glossary (mono- or bilingual),
related texts in the target language, or specifications including technical
standards or legal requirements. Such information is designed to help
the translator produce a better translation, and, in general, it achieves
its objective. A bilingual glossary of technical terms, for instance, facili-
tates the terminology side of the translation process and helps to ensure
terminological consistency both within the translated text and between
the target text and other documents of the client. The resulting good
quality of the translation (in this respect) is, thus, due to the provision
of a bilingual glossary (at least to some extent). Since the consistent use
of key terms becomes obvious directly in the target text, it is relevant
not only to the quality of the translation but also to the assessment of
that quality – as long as the evaluator has at his or her disposal also a
copy of the glossary (or knows the client’s terminology). To an external
evaluator unfamiliar with the terminological preferences of the client,
the effect of the glossary remains invisible as any consistency within the
target text is expected as the translational norm.
Two points, though, render the above considerations less straightfor-
ward than they seem to be. First, what happens if the glossary provided
by the client contains the occasional substandard translation? And sec-
ond, as regards terminological consistency, does that refer only to con-
sistency within the target text or also to consistency between source
and target? In the first case, the translator will discuss any contentious
translations with the client and deal with them as reflected by the out-
come of that discussion. Weighing the pros and cons of a substandard
translation that occurs also in other texts by the client against those
of a perfect terminological solution that is not used elsewhere in texts
by the client, the evaluator will have to examine the overall situation
carefully, before he or she can draw any conclusions. In the second case,
the problem is solved more easily: in a covert translation setting, any
inconsistency in the source text should be corrected in the target text; in
other words, consistency within the target text takes precedence over the
consistency between the source text and the target text. Inasmuch as the
client is accountable for any consistency issues, the client directly affects
the quality of the translation. This influence extends also to translation
assessment, particularly, if the client evaluates the target text.
Stipulations and Specifications

Besides deadlines and terminological requirements, the translation brief
may also specify other criteria. For example, the translation brief may
stipulate that the translator be qualified or certified, or even that he or she
be sworn in by a particular court; and it may stipulate that the translator
work in accordance with specified translation standards and use a par-
ticular translation software. Very often, the language of the target text
has to be the translator’s mother tongue. These and other stipulations are
intended to ensure a high-quality translation in that they help to control
the translation process to a certain degree. That such translation brief
specifications have a positive effect on the quality of a target text cannot
be denied. Still they cannot guarantee a good translation result. These
factors, then, are relevant to translation quality, but they are not relevant
to the evaluation. In a limited way, this is also true of specifications that
have no direct bearing on the translation job at hand, for instance, the
requirement that the translator provide the client with the translation
memory and/or dictionary created in the course of the translation. Here,
the influence on translation quality is indirect: if a translation agency
needs to give translation work for a client to a translator other than the
one who is usually doing the job, such a translation memory and/or dic-
tionary can achieve a certain degree of consistency in the translation of
individual terms as well as frequently used phrases and sentences.
Most potent are those client specifications that override the usual
translation standards. Where the translation brief specifies, for example,
that a source text written for adults should be translated and adapted
for children, such a specification implicitly outlines the quality needed
for the target text in terms of certain textual and linguistic properties.
It is, then, on the basis of these properties that an evaluator will have to
judge to what extent the adaptation requirement has been satisfactorily
met in the target text. The specification in the translation brief has, as
it were, an umbrella function for all those criteria that are considered
relevant to the translation and adaptation of a text for children. Since
the implementation of these criteria is directly visible in the target text,
such a specification can be said to be directly relevant not only to the
quality of the translation but also to its evaluation. Also, it is obvious
that such a specification constitutes a quality factor that is known to the
evaluator.
Motivation
Not known to an external evaluator but certainly to the client as evalua-
tor is the final quality factor discussed in this section. We are concerned
with the client’s influence on the motivation of the translator and on
the way in which this influence adds another piece to the jigsaw puzzle
of a perfect translation. Even though the link with the actual quality of
the target text is only indirect, the client’s influence on the translator’s
motivation and, thus, on translation quality can be considerable: for a
translator who is demotivated can hardly be expected to do his or her
best. One key motivational force in this context is payment. A well-
paid translator is more willing to work as diligently as he or she can,
and also will more readily provide additional quality services such as
pointing out to the client any defects in the source text or commenting
individual translation decisions for the benefit of the client. A poorly
paid translator – apart from probably being not particularly enthusiastic
about his or her work – will have to take on more translation jobs to
make ends meet so that he or she can spend less time on one job, which,
in turn, may have a negative effect on translation quality.
However, apart from payment, there are other, emotive means by
which the client can motivate the translator, thereby paving the way
for a good translation. These include, for example, positive feedback,
recommendations to other potential clients, and also constructive
criticism. Any such response on the part of the client will show the
translator that his or her work is appreciated and that he or she can
be proud of it. A translator whose professional skills are recognised
in this way will not want to jeopardise his or her good reputation by
submitting mediocre translations. Of course, the above instruments of
motivation are used for follow-up orders, as feedback, recommenda-
tions, and constructive criticism become relevant only after the first
translation order has been processed. While implicit recognition of the
translator’s good work is already obvious from the fact that the client
places a follow-up order with the translator, motivating the translator
by explicitly showing appreciation for his or her linguistic achievement
is probably more powerful still in laying the ground for a high-quality
translation. Ideally, both the emotive and the pecuniary means of mo-
tivating the translator should go hand in hand to ensure a positive
setting for the production of the target text. Despite their impact on
translation quality, the motivational factors are not relevant to the
evaluation of translation quality, because their relation to the usability
of the target text is only indirect.
Table 4.3 Summary of client factors
Factor Is the factor relevant to the Is the factor known to the

evaluation of translation quality? evaluator?
Deadlines Only in those cases, in which the No, unless it is the client or an
target text cannot be used, if it agent directly involved in
is submitted after the deadline the translation process who
has expired. evaluates the translation.
Glossaries Only in those cases in which the Only in those cases in which
evaluator has a copy of the the evaluator has a copy of
glossary. the glossary.
Stipulations Stipulations with regard to the No, unless it is the client or an
translator’s qualifications are agent directly involved in
not relevant to the evaluation the translation process who
of translation quality. evaluates the translation.
Specifications Specifications that override the The evaluator must know the
usual translation standards specifications in order to be
are clearly relevant to the able to evaluate the quality
evaluation of translation of the translation.
quality.
Motivation No. No, unless it is the client who
evaluates the translation.
Summary of Client Factors

The various client-related quality factors of translation are summarised
in Table 4.3.
Translator
Usually, the translator is held accountable for any mistakes found in
the translation, because he or she is seen as the immediate producer of
the target text. In this function, the translator is also responsible for the
quality of the translation: he or she is, thus, the target text’s first eval-
uator. While this would suggest that the translator is in full control of
the translation process, the fact that the translator always operates in a
wider context – whether economic, cultural, or otherwise – curbs the
translator’s power over the quality of his or her translation.33 Still, the
translator makes most of the decisions required to produce the target
text. This has certain consequences, as Ben van Wyke explains:
If translators accept the fact that the original will always be trans-
formed by the intervention of their work, they will also have to ac-
cept the fact that, contrary to the prevalent requirement that they
do otherwise, they will always be visible as they leave marks of the
decisions they have made.
(Wyke 2010:113)
The visibility of these marks as being the translator’s has implications
for the evaluation of the target text: the translator will get the blame for
any translation solutions that fail to live up to the evaluator’s expecta-
tions; yet, although the visibility of the translator’s marks may enable
the evaluator to draw some conclusions about the translator and/or the
translation situation, there is usually no direct relation between a target
text solution and any of the features attributed to the role of the trans-
lator. The factors discussed in this section, then, are relevant to trans-
lation quality, but they are not normally relevant to translation quality
assessment.
Overview and Translation Tools

Within the context of the “translator’s daffodil”, we subsume under
“translator” all those quality factors that are directly related to the
translator as a human being or to the translator’s task of producing a
target text. Since the task of producing the target text involves not only
the actual translator but, possibly, also other agents such as a reviser
or proofreader, 34 many of the factors referred to in this section will re-
late to both the translator and the reviser. This is certainly true of any
personal assets and skills such as translation competence, qualification,
and motivation. Other factors include tools that may be used by the
translator and by the proofreader to facilitate their respective tasks.
The translator may either use a CAT tool (together with a translation
memory and/or a terminology manager) or a machine translation (MT)
tool, whereas the proofreader could make use of spellchecking and/or
grammar-checking programs. Figure 4.3 provides an overview of the
translator-related quality factors.
• CAT-tool with • MT-tool • spell-checker

translation • grammar checker
Essential tools memory and/or
terminology
manager
Translator Proof-reader or reviser
• Translation competence
Personal assets
• Qualification
and skills
• Motivation
Figure 4.3 Overview of translator-related factors.
Other tools might be added, such as word processing or project man-

agement software. The tools mentioned in the overview all relate to trans-
lation quality: CAT tools, spellchecking tools, and grammar-checking
tools are designed to improve the overall quality of the target text,
whereas MT tools primarily speed up the actual translation process. As
the MT output is usually not satisfactory, it requires post-editing – how
much, depends also on whether the source text has been edited before
being fed into the machine.35 Pre- and post-editing as well as properly
training the MT engine can be regarded as the actual quality assurance
measures in MT.
Computer-aided (or computer-assisted) translation tools help the
translator to translate more consistently: identical translation units will
be translated in exactly the same way, and specific words or phrases
can be stored in, and retrieved from, a terminology database. A CAT
tool also ensures that no source text units are accidentally skipped in
translation. Despite these obvious advantages, there are possible pitfalls
when using a CAT tool. If the translation of already translated units
is done automatically, a unit that requires a different translation (be-
cause it occurs in a different context) may turn out to be substandard
(or even wrong) in translation. Uwe Reinke points out another possible
drawback, namely, that the demands on target text coherence may not
be fully met, because CAT tools invite a limited focus on the clause
or sentence as the unit to be translated and lead translators to accept
100 percent matches suggested by the tool (cf. Reinke 1997:104). Simi-
larly, spellcheckers are helpful, when it comes to spotting mistakes that a
proofreader might overlook36; yet, if they suggest the occasional alterna-
tive that is not correct in a given context, the not-so-attentive user of the
tool will easily cause a wrong word to creep into the text. Being aware of
such pitfalls is the best way to avoid them. Such awareness is, therefore,
part of the translator’s competence.
Translation Competence
What is a competent translator? In trying to find an answer to that
question, we will first look at two definitions of translation compe-
tence. The PACTE Group37 defines translation competence as “the un-
derlying system of knowledge required to translate” (PACTE 2011:318).
This system of knowledge includes a bilingual sub-competence com-
prising pragmatic, socio-linguistic, textual, grammatical and lex-
ical knowledge; an extra-linguistic sub-competence of general world
knowledge, domain-specific knowledge, as well as bicultural and en-
cyclopaedic knowledge; translation-specific knowledge and skills to do
with how translation works in theory and in practice; an instrumental
sub-competence relating to the use of documentation resources and any
information and communication technologies relevant to translation; a
strategic sub-competence that serves to control the translation process
and ensures its efficiency; and, finally, psycho-physiological compo-
nents, which cover cognitive aspects such as memory, perception, atten-
tion and emotion, and attitudinal aspects such as intellectual curiosity,
perseverance, rigour, the ability to think critically, and also general
skills such as creativity and logical reasoning (cf. PACTE 2011:319).
Obviously, this approach to translation competence is an analytical ap-
proach, in which the overall knowledge and skills of a translator are
scrutinised minutely.
Pym (2003), by contrast, offers a minimalist definition. He claims that
a translator needs basically two abilities:
• The ability to generate a series of more than one viable target

text (TT1, TT2 ,…, TTn) for a pertinent source text (ST);
• The ability to select only one viable TT from this series, quickly
and with justified confidence.
(Pym 2003:489)
In short, translation competence is “a process of choosing between vi-

able alternatives” (Pym 2003:491). The above approach is a synthetic
approach in the sense that any other aspects referred to in a multicom-
ponent model (such as the one by PACTE) will invariably add up to form
the basis of Pym’s two core abilities:
We propose that, together, these two skills form a specifically trans-

lational competence; their union concerns translation and nothing
but translation. There can be no doubt that translators need to know
a fair amount of grammar, rhetoric, terminology, computer skills,
Internet savvy, world knowledge, teamwork cooperation, strategies
for getting paid correctly, and the rest, but the specifically transla-
tional part of their practice is strictly neither linguistic nor solely
commercial. It is a process of generation and selection, a problem-
solving process that often occurs with apparent automatism.
(Pym 2003:489)
A practically minded critic may point out that an individual trans-

lation solution need not be based on the selection of one target text
unit from a whole range of possible target text units; it may as well
be the only obvious solution. However, in theory, there is no such
thing as an obvious solution. Even the easiest of translations involves
a selection process in which the translator, whether consciously or un-
consciously, discards several viable and non-viable alternatives. Pym’s
concept is intriguing, because it focusses on the essence of translation
competence.
The distinction between two essential skills, on the one hand, and
several contributory factors, on the other, bespeaks a translation compe-
tence model which consists of elements that bear directly on the quality
of the target text and elements that affect translation quality only indi-
rectly. Whether a translator has succeeded in picking an option from a
set of previously generated translation solutions will become evident in
the target text’s macro- and/or microstructural manifestations. Here,
the translator’s selection shows directly in textual aspects such as lexis,
syntax, or overall structure. As a result, Pym’s two essential elements of
translation competence not only have a direct impact on translation qual-
ity but are also relevant to the assessment of that quality. And because of
its visibility in the target text, the translator’s choice is also known to the
evaluator. What the evaluator usually remains ignorant of, though, are
the other viable target text units that the translator rejected.38
Those aspects of translation competence that determine the ability
of the translator to make the right choices can be found among the
sub-competences listed by the PACTE Group. Perhaps, the most prom-
inent of these are the bilingual and extra-linguistic sub-competences,
because it is here that the translator first needs to assess his or her own
abilities: is he or she sufficiently well-versed in the source and target
languages and cultures (both geographically and domain-specifically) to
take on the translation at all (cf. also Witte 2000:55)? In fact, to know
when not to accept a translation job because of a lack of competence in
the specialised field covered by the source text can certainly be regarded
as a key skill of the translator, since overestimation of his or her own
abilities inevitably leads to translations of inferior quality. The transla-
tion of a particular text may, however, also be refused on other grounds.
Thus, a refusal should be taken into consideration or is even imperative,
if the translator does not have enough time for the translation, or if the
content of the source text violates the ethical standards of the trans-
lator or of the professional association of which he or she is a mem-
ber.39 However, the question of when not to accept a translation job
need not be discussed further in the context of this book, as there can
be no translation quality and no translation quality assessment without
a translation.
Once the translator has accepted a translation job, he or she needs to
be able to fall back on the whole range of skills and knowledge related
to translation competence as described by PACTE (2011:319). Whether
skills or knowledge – any of the competences and sub-competences may
adversely affect the quality of the translation in its textual and situational
dimension. While insufficient bilingual, extra-linguistic, and instrumen-
tal sub-competences are more likely to have a negative effect on the target
text as text, inadequate translation-specific knowledge and/or an under-
developed strategic sub-competence will bear upon the translation process
and situation, which, in turn, may affect the target text indirectly. The
psycho-physiological components mentioned by PACTE are important to
both the target text and the translation situation: a translator, however
skilled, whose attitude to work, language, and business is characterised
by carelessness, negligence, and inattention to detail, will probably pro-
duce a translation of inferior quality and, perhaps, even jeopardise the
timely delivery of the target text.40 With the psycho-physiological compo-
nents extending to concepts as vague as emotion, perception, and rigour,
the number of aspects which may come into play when analysing the
translation process and the result of that process, is almost infinite.
Qualification and Motivation
In view of the key importance of the translator’s competence, the ques-
tion of the translator’s qualification might be considered redundant. Af-
ter all, what counts is the ability of the translator to translate a particular
source text into a target text, no matter whether or not the translator
is formally qualified as a translator. And, indeed, one may find compe-
tent translators that have no formal qualification and qualified transla-
tors whose competence leaves much to be desired.41 In some translation
contexts, the typical translator may not have the qualification of a
translator. Thus, Jerry Palmer writes about the translators in the news
gathering and dissemination business, where translations often consist
of summaries made from several source texts: “Translation ‘in the field’
is usually done by personnel who are not specialized in translation and
interpretation […], for whom translation is only part of the job descrip-
tion, and may not even be its most important part” (Palmer 2009:187).
In the end, it is the result – that is, the quality of the target text – that
matters. Such considerations notwithstanding, any study on the quality
of professional vs. non-professional translations would start from the
assumption that there is a general correlation between the qualification
of a translator and his or her translation competence.
Like the translator’s psycho-physiological aptitude and professional
stance as part of his or her translation competence, the translator’s mo-
tivation affects the quality of both the target text and the translation
process: a high motivation will have a positive effect on the psycho-
physiological factors, which will then help the translator to produce a
translation of good quality. We distinguish between two kinds of moti-
vation: external motivation and self-motivation. While self-motivation is
closely related to the psycho-physiological factors and, to some extent,
reflects the translator’s attitude towards his or her work, external mo-
tivation requires input from a third party. If he or she is a member of a
translation team, the translator may get encouragement from other team
members. Motivation may also come from the translator’s family, who
show respect for his or her work. Usually, most important is the kind
of motivation that clients provide when they give positive feedback and
pay an adequate price for the good service they get. Whatever the effect
of a particular kind of motivation on the quality of a translation, it will
be difficult to estimate its share in the overall result. Motivation of the
translator, while being relevant to translation quality, remains irrelevant
to translation quality assessment.
Summary of Translator Factors

We will conclude this section with a short discussion of a phenomenon
that affects translation quality and translation quality assessment not at
the root, but at the top. Self-translation, as the phenomenon is called,
refers to “the act of translating one’s own writings into another language
and the result of such an undertaking” (Grutman 2009:257). It is clear
that the author who translates his or her own work has, as a transla-
tor, the advantage of knowing his or her own intention as the author
of the original. Especially, in literary translation, this appears to be a
most forceful argument that affords an external evaluator little room
for manoeuvre. The evaluation of such a translation would most likely
focus on a descriptive account of the relation between source and target.
Here, the power of the translator as his or her own evaluator operates at
the level of translation quality assessment and thereby commands good
quality for the target text – provided that the self-translator is as well-
versed in the target language as he or she is in the source language. In
principle, though, also self-translating authors are subject to the range of
factors discussed in this section.
As in the discussion of text form factors, there is no need to sum-
marise the findings of this section in tabular form. With the exception
of translation competence – notably, the key competence as defined by
Anthony Pym – the translator-related factors are not relevant to the eval-
uation of translation quality. While translation competence shows quasi-
tautologically in the translation decisions made in the target text, any
qualitative benefits reaped from the use of CAT tools or spellcheckers
could, and should, also have been achieved without using these tools.
The question whether the translator and proofreader are qualified and
motivated to complete their respective tasks is irrelevant to the evalu-
ation of the target text. As regards the evaluator’s knowledge of these
factors, he or she sees the translator’s competence only in the translation;
the evaluator remains ignorant of all other factors unless he or she is
explicitly informed about them.
Culture
What is culture? It certainly is an elusive phenomenon that is difficult
to define, because it manifests itself at different levels, from the clearly
visible (such as clothing) via the semi-visible (such as underlying cus-
toms) to the invisible (such as preferred patterns of thought and ac-
tion).42 Starting from the way the term “culture” is used, David Katan
points out:
Until the birth of anthropology, culture referred exclusively to the

humanist ideal of what was considered ‘civilized’ in a developed so-
ciety. Since then, a second meaning of culture as the way of life
of a people has become influential. With the development of disci-
plines such as cultural studies, a third meaning has emerged which
attempts to identify political or ideological reasons for specific
cultural behaviour […]. Hence, depending on the definition adopted,
culture may be formally learnt, unconsciously shared, or be a site of
conflict.
(Katan 2009:70)
Inasmuch as translation deals with the possibility of transferring mean-

ing from one language and culture to another, culture is conceived of as
the unconsciously shared way of life of a people. It consists in the modes
of behaviour of an individual, a society, or parts of a society at any given
period of time (cf. also Vermeer 2006:158). Where translation enters the
realm of politics and power, Katan’s third concept of culture becomes
relevant. This will be discussed in the corresponding section, below.
However sophisticated or elaborate a definition of culture may be,
[u]ltimately, culture has to be understood not only as a set of levels

or frames but as an integrated system, in a constant state of flux,
through which textual signals are negotiated and reinterpreted ac-
cording to context and individual stance.
(Katan 2009:73)
Thus, any analysis of how culture can affect translation quality will
have to cope with the fact that the ground of cultural norms and rules
is constantly shifting. This section will (1) look at the shifting ground
of culture in more detail, (2) outline the various norms of translation as
defined by Gideon Toury, (3) give examples of translational norms in
different cultures at different times in history, and (4) discuss selected
borderline cases of translation at the crossroads of culture and transla-
tion quality assessment.
The Relativity of Culture

A translator who translates from a source language into a target lan-
guage will have to take into account any differences between the respec-
tive source and target cultures. Such linguistic and cultural differences
constitute the skeleton of the translation situation, in which the trans-
lator typically belongs to – or feels at home in – the target culture. The
foreignness or otherness of one culture (the source culture) can only be
experienced by comparison with another culture (the target culture) (cf.
Witte 2000:109). It is through such a comparison that the translator
establishes, assesses, and negotiates any differences between the source
and target cultures with regard to an adequate translation. These differ-
ences can be textual or non-textual. They affect translation in that they
force the translator to find answers to practical questions relating to the
translatability of the source text, the most pertinent translation strategy,
and the handling of presuppositions.
Culture, however, is more than what a translator can take into ac-
count in his or her attempt to produce a high-quality translation. Con-
sider the following apt quotation by André Lefevere:
The way in which translations are produced matters because trans-

lations represent their originals for readers who cannot read those
originals. In other words: translations create the ‘image’ of the orig-
inal for readers who have no access to the ‘reality’ of that original.
Needless to say, that image may be rather different from the reality
in question, not necessarily, or even primarily because translators
maliciously set out to distort that reality, but because they produce
their translations under certain constraints peculiar to the culture
they are members of.
(Lefevere 1996:139)
A translator can only translate within the boundaries of his or her cul-
tural knowledge and experience. Although he or she should try to be
consciously aware of as many cultural issues as possible, there will al-
ways be a number of self-evident rules or norms of culture with which
the translator will comply intuitively. Of course, with culture being “in
a constant state of flux” (Katan 2009:73), it is impossible to pinpoint
all the elements of a given culture. Since culture operates at different
levels, from the individual to a whole nation (and beyond), any descrip-
tion of a particular culture is subject to the idiosyncratic experience of
that culture by the person writing the description. Thus, within certain
limits, there will be variation in how the individual members of a culture
view that culture. Young people, for instance, regard some modes of be-
haviour or some expressions as perfectly acceptable, while older people
deem the same totally unacceptable.
Cultural Norms
Still, culture smoothes over any idiosyncratic preferences, for the indi-
vidual to be able to live in his or her community (cf. Vermeer 2006:356).
This smoothing over occurs as a result of cultural norms.43 Such norms
are also relevant to translation, as Christina Schäffner explains:
Translation being defined as socially contexted behaviour requires an

explanation of the socio-cultural constraints which determine transla-
tors’ behaviour. These constraints can be absolute rules or pure idio-
syncracies [sic] as the two extremes, with norms as a graded continuum
in between. Some norms may be more forceful and closer to rules,
whereas others only exert a rather weak influence. Moreover, norms
are not fixed once and for all but can change in the course of time.
(Schäffner 2010:236–237)
This quotation contains two crucial ideas about norms in translation:
first, norms are neither arbitrary nor definite; second, norms are pre-
scriptive to some extent, or else they would not exert any influence on
the translation process. The second idea runs counter to the concept of
descriptive translation studies, which stresses “that norms are a category
of descriptive analysis and not, as the term might imply, a prescriptive set
of options which are thought by the analyst or scholar to be desirable”
(Baker 2009:190). While the descriptive translation scholar wishes to
substantiate or refute any intuitive claims that something believed to be
a translational norm is, indeed, such a norm, the translator regards any
perceived norm as prescriptive and relies on a more or less subjective
interpretation of that norm in order to produce the best possible target
text. It is evident that both perspectives on translational norms in a given
culture are important: the descriptive approach helps to understand the
factual basis of translation; this basis should then be used to tell the
translator how to translate.
Rule-like norms are more likely to be recognised and complied with
than norms which exert but a weak influence. Toury distinguishes three
types of translational norms: an initial norm, preliminary norms, and
operational norms (cf. Toury 1995:56–58). The initial norm serves to
define whether the translator should stick to the norms of the source text
(and source language and culture) or whether the norms of the target
language and culture are more appropriate. The preliminary norms do
not concern the individual translator and his or her particular transla-
tion project, but “the existence and actual nature of a definite t ranslation
policy […] and the directness of translation” (Toury 1995:58).
Translation policy, here, refers to choices to do with what is translated
(source text types, individual source texts, authors, source languages,
etc.), whereas the directness of translation involves “the threshold of
tolerance for translating from languages other than the ultimate source
language” (ibid.).44 For any decisions made during the actual act of
translation, the translator reverts to operational norms. These relate to
the selection of linguistic material for the target text as well as to any
changes made in the course of the translation, for example, in terms of
large-scale omissions.
So, how do these types of norms affect translation quality assessment?
The initial and operational norms are most important, here, because
they are directly involved with the source and target texts as texts: a
foreignising mode of translation applied to a source text type that is
usually handled in a domesticating way (or vice versa) and the appropri-
ateness or inappropriateness of any choices made in the target text have
an immediate impact on the quality of the translation. By contrast, the
preliminary norms are less likely to be relevant to the quality of a target
text, because relevance would imply a significant deviation in terms of
translation policy. Here, the translation of a particular source text or
source text type would have to be extremely uncommon, or even forbid-
den, in the target culture for such a fact to matter to an evaluator who is
bound by the relevant preliminary norm.
To conclude our discussion of cultural norms in translation, we can
say that the evaluator of a target text should clearly be aware of any
relevant norms and use them as the yardstick against which to gauge
the quality of a translation. In doing so, the evaluator may, however,
have to adapt some of the norms to fit the purpose of the evaluation. For
example, a norm that allows a certain number of minor mistakes for the
translation of a source text to be regarded as good or acceptable in a par-
ticular business context would not be applicable with a zero-defect qual-
ity approach.45 Thus, inasmuch as the compliance or non-compliance
with a translational norm is directly visible in the target text, that norm
is regarded as relevant to translation quality assessment.
Translation in Different Cultures at Different

Times in History
Translation culture, like any culture, changes with history and geogra-
phy: in different countries at different times, there have prevailed differ-
ent concepts of translation. Often, the central question is whether the
translation of a particular text or text type in the culture of a particular
place at a particular time is more geared towards the source text and
source culture or towards the target text and target culture. In other
words, a given translation context needs to be probed for the initial
norm prevailing in it. For example, in Chinese translation history (cf.
Hung/Pollard 2009), there seem to have been not infrequent changes
from one initial norm to the other, whether during the time of the sutra
translations or in the 1920s and 1930s: sometimes a word-for-word
rendering was preferred, sometimes a free translation. While the most
important argument in favour of the second method is readability, the
first approach is – in the case of the sutra translations – supported by “a
belief that the sacred words of the enlightened should not be tampered
with” (Hung/Pollard 2009:372), and – in the case of the debate in the
1930s – by an argument that the Chinese language should benefit from
the “objective of appropriating from European languages through trans-
lation wording and grammatical devices” (Hung/Pollard 2009:376).
Similarly, Baker/Hanna describe a word-for-word and a sense-for-sense
translation for the Abbasid period in the Arab world, with the latter
approach turning out to be more successful than the former (Baker/
Hanna 2009:333).
Whether readability of the target text is more important in transla-
tion than faithfulness to the source text also with regard to idiomatic
elements depends on a culture’s tolerance of foreignness in that cul-
ture’s language. An interesting case in point is the situation in Japan. As
Masaomo Kondo and Judy Wakabayashi write:
Today the literal approach seems to be the more popular form in

Japan, and free translation is generally considered in a rather nega-
tive light. […] the approach adopted by many contemporary transla-
tors who are willing to sacrifice natural Japanese for ‘fidelity’ to the
original is based simply on the belief that literal translation equates
with faithful translation. There is also considerable tolerance of lit-
eral translation on the part of readers, who have long been accus-
tomed to a form of Japanese which is heavily influenced by Chinese
and who expect translations to be unidiomatic.
(Kondo/Wakabayashi 2009:475)
Note that the period covered by this statement is the latter part of the
twentieth century, in spite of the date given in the reference (the text also
appeared in the first, 1998 edition of the Routledge Encyclopedia of
Translation Studies). Yuri Furuno agrees that “acceptance of foreignness
on the part of Japanese readers […] seems to have existed throughout Ja-
pan’s history” (Furuno 2005:147); yet, from her vantage point of writing
in the twenty-first century, “Japanese translation norms today are mov-
ing away from a source text orientation” (Furuno 2005:151). Her empir-
ical studies lead her to conclude that “[i]n response to the contemporary
readers’ increasingly target-oriented expectation towards translated
texts, the traditional source-oriented translation approach and writing
style may continue to change” (Furuno 2005:159). That means, a trans-
lator into Japanese is free to depart from the source text only inasmuch
as this is necessary for a readable target text. An approach based on
Vermeer’s skopos theory would probably be considered inappropriate.46
The preference for one initial norm is subject to two forces: the ex-
pectation of the target text readers pulls the initial norm in one direc-
tion, whereas the authority and influence of successful translators and
theoreticians through their translations and treatises push the norm in
the same, or in the opposite, direction. Of course, these pushing and
pulling forces are not altogether independent of each other, because the
target text readers are influenced by the translations they read and the
translators are influenced by the readers’ responses to their translations.
Whenever an initial norm is supported by a theory of the prescriptive
sort, the pushing force can gain additional momentum, with ideology
paving the way. And when one such theory stands pitted against an-
other, the initial norm and its translational consequences become tossed
about by conflicting interests. This seems to have been the case in the
German-speaking world of the eighteenth century, which saw a remark-
able literary theory dispute mainly between Gottsched in Germany, on
the one hand, and Bodmer and Breitinger in Switzerland, on the other.
The dispute also extended to translation issues, as Harald Kittel and
Andreas Poltermann explain:
Gottsched maintained that a good translation had to be in agreement

with the principles of enlightened, normative poetics. If the original
or source text did not conform with these rules, the translator was
duty-bound to improve, expand or abridge. The translation had to be
a German text, through and through. Breitinger, in contrast, main-
tained that there were no superfluous words in literary works of art.
In his Kritische Dichtkunst (1740) he rejected many of the more pre-
sumptuous claims of the Enlightenment, thus preparing the way for
English in preference to French literature and its ideals. Anticipating
Herder and Humboldt, he argued that the mentalities of different
nations are reflected in the peculiarities of their respective languages.
Therefore, a translation must not violate the ‘thoughts’ (Gedancken)
of the original or deviate from its source in any other way.
(Kittel/Poltermann 2009:414–415)
With his ideas being “developed by Friedrich Gottlieb Klopstock (1724–

1803) and Johann Gottfried Herder (1744–1803)” (Kittel/Poltermann
2009:415), Breitinger seems to have prevailed in the debate.
Gottsched’s notion of translation as being confined by the rules of
the target language recall the French phenomenon of the “belles in-
fidèles” in the seventeenth and eighteenth centuries. These “free dy-
namic translations […] aimed to provide target texts which are pleasant
to read” (Salama-Carr 2009:406), that is, target texts which were sup-
posed to improve on the quality of the source text. While this seems to
have been the dominant approach to literary translation at that time, it
was nevertheless “not universally accepted” (Salama-Carr 2009:407):
the examples of a more source-oriented approach given by Salama-Carr
mostly refer to the translation of religious texts. Such historical transla-
tion theories and practices are not so much relevant to the quality of an
individual translation than to the assessment of that quality. Since trans-
lators usually translate, or try to translate, as they are required to do by
the translational zeitgeist of the time and culture they live in (a zeitgeist
which, in turn, is defined by the actual practice of leading translators),
any translation that goes against the grain may be judged as of inferior
quality. This would certainly be the case, if the evaluator accepted the
translational norm prevalent at the time as his or her own evaluation
standard. Yet, if the target text is of outstanding quality in other re-
spects, notably in those that lend the translation an undeniable aesthetic
forcefulness, the evaluator may as well acknowledge the aesthetic qual-
ity and ignore the initial norm dominant at the time as being irrelevant.
The standard by which an evaluator judges the quality of a translation
may, thus, be (1) the same as that of the translator, (2) the same as that
of the translational zeitgeist, or (3) a standard different from both that
of the translator and that of the translational zeitgeist. The third, rather
theoretical, option seems to be most unlikely, because an evaluator’s
standard or norm that is supported neither by the translator’s norm nor
by the prevailing norm of a culture at a given time will hardly provide
a sound basis for a convincing assessment of the target text. Usually,
the translation norms of the target culture and of the translator should
coincide – if that is not the case, the evaluator will have to find out why
before establishing his or her own norm or set of norms.
Borderline Cases: Unlikely Successes and EU Translations

When judging a translation, should the success of the target text be a
quality criterion? Should, for instance, the wide circulation of a trans-
lated text be accepted as proof that this text meets the readers’ expec-
tations and, therefore, must be of good quality? These questions are
answered in the negative: for instance, the sales of literary translations
depend not only on textual quality but also on other factors such as
advertising and reviews. In some cases, the economic success of a trans-
lation remains a mystery. Thus, it is amazing that a book which had
received scathing criticism on a large scale, such as the German trans-
lation of Lemprière’s Dictionary, sold more than 100,000 copies (cf.
Hönig 1995:124). Another interesting example is furnished by some of
those books by Nigerian writer Amos Tutuola which are largely based
on Yoruba mythology, as Paul Bandia explains:
Tutuola […] literally translated some Yoruba mythology into En-

glish. In an attempt to capture Yoruba syntax in English (and given
that he was a public service clerk with just an elementary school ed-
ucation), he produced curious syntactic patterns that endeared him
to European readers.
(Bandia 2009:318)
Apart from the fact that Tutuola’s achievement would not be regarded
as proper translation, the strongly foreignising element of syntactic id-
iosyncrasy operates at an aesthetic level which readers are free to like
or dislike and whose significance should be recognised by an evaluator.
The positive reception of errors has also been reported for non-literary
translation as shown in the following quotation:
In some communication situations, errors are expected and regarded

as acceptable and even ‘fun’. This is, for example, the case in Danish
tourist brochures translated into poor German. In spite of the errors
the brochures retain a high degree of usability.
(Hansen 2010:386)
Here, despite the claim that in such a context “errors are expected and
regarded as acceptable”, these errors would not count as part of the
aesthetic design of the brochures and, therefore, not be accepted by a
reviewer.
An evaluator would do well to be aware of the actual cultural and
economic context of a translation, whether or not this context should
be accounted for in the assessment of that translation’s quality. Thus, in
an assessment of Richard Francis Burton’s nineteenth-century transla-
tion of The Arabian Nights, it is useful to know that, evidently, “Burton
furnishes the image of the Arabs, and of Arab culture, literature, and
even language, he wished to render, consciously or not: the image that
was expected of the Orient and, of course, of his work as a translator”
(Carbonell 1996:80–81). On the other hand, in the case of the Poems
of Ossian, eighteenth-century pseudo translations by James Macpher-
son (who insisted that he translated what he wrote, but who failed to
provide evidence in terms of the original), the translator’s motive for
presenting his prose poetry as translation, albeit interesting, is of no
consequence to an assessment that can only tackle the literary merit of
the supposed target text.47 In both cases, the translators wrote their
texts in line with the (assumed) cultural expectations of their target
audience, who wanted to experience the Arabic, oriental atmosphere
of the Arabian Nights and the ancient Celtic flavour of the Poems of
Ossian, respectively. In both cases, the authenticity of the content of
these texts – derived from the fact that they are translations (or per-
ceived to be translations) – is more important than the adherence to an
initial norm of translation.
In the above examples of the influence of culture on translation, the
identification of source and target cultures has been taken for granted:
in a translation context in which a Polish text needs to be rendered into
Italian, for example, the translator has to grapple with the problems
resulting from the difference between the Polish and Italian cultures.
There are, however, some translation situations relating to certain text
types or forms of translation, in which cultural difference gives way to
cultural similarity. We are talking about translations in international
organisations or institutions such as the European Union. Here, it seems
that hybridity, which refers to target text features that are unusual for
the receiving culture, is a common phenomenon. Concerning the assess-
ment of such foreignising translations, Ji-Hae Kang writes:
Hybridity in EU translations, in particular, has been associated less

with ‘translationese’, or lack of translational competence, and more
with a convergence between cultures or institutional patterns of be-
haviour. […]. EU translations are ‘intracultural’ in that they are re-
flective of a distinct EU culture that cannot be accounted for by the
dichotomous conceptions of source and target cultures […].
(Kang 2009:144)
It should be noted, however, that the creation of a distinct EU culture
is clearly dependent on the institutionalisation of translation. Although
the cultural exchange between countries within the European Union or
within the European community of shared values causes the cultures of
these countries to become more similar in some respects, the translation
of texts outside the institutional boundaries remains very much exposed
to cultural differences.
Summary of Culture Factors

This section has shown that culture – however varied in its differ-
ent manifestations – plays a crucial role in translation. Cultural gaps
between source and target may have to be bridged depending on the
translational approach (which, in turn, is often dependent on text
type), and the initial norm of the target culture has to be taken into ac-
count. An evaluator can find any cultural gaps by scanning the source
text for culture-specific items. These items are defined by Javier Franco
Aixelá as
[t]hose textually actualized items whose function and connotations

in a source text involve a translation problem in their transference
to a target text, whenever this problem is a product of the nonexis-
tence of the referred item or of its different intertextual status in the
cultural system of the readers of the target text.
(Franco Aixelá 1996:58)
Table 4.4 Summary of culture factors
Factor Is the factor relevant to the Is the factor known to the

evaluation of translation evaluator?
quality?
Initial norm Yes. It should show in the Yes. In order to evaluate the
overall strategy of a target text, the evaluator
successful translation. needs to be aware of the
prevailing initial norm in
the relevant culture at a
specific time in history.
Preliminary No. Possibly.
norms
Operational Yes. Operational Yes. In order to evaluate the
norms norms should govern target text, the evaluator
microstrategic decisions needs to be aware of the
and be reflected in the lexis, prevailing operational
syntax, etc. of the target norms in the relevant
text. culture at a specific time in
history.
The culture-specific items are usually identified in the process of
translation. That this identification process can also follow the rigour
of a methodological regime has been demonstrated by Floros (2003).
The transfer of the culture-specific items from the source text to the
target text clearly affects the quality of the translation as well as the
assessment of that quality. The underlying initial norm of transla-
tion and any general cultural aspects that may have an influence on
the translation process and on individual decisions by the translator
will affect the perspective from which the evaluator looks at the tar-
get text. Inasmuch as they find their way into the evaluator’s norm
(against which the quality of the translation is measured), they are
more immediately relevant to the evaluation of the target text than
to its quality. The quality factors relating to culture are summarised
in Table 4.4.
Politics
For our purpose, politics can be regarded as “[o]ne of the main aspects of
culture” (Al-Taher 2008:1), where “culture is mixed with ideologies and
interests” (ibid.). Since translation takes place in a cultural environment,
it is also influenced by the ideologies and interests of that environment.
Translating can, thus, be depicted as a political act. The following quo-
tations, taken from the book Translation, Power, Subversion by Román
Álvarez and M. Carmen-África Vidal, will illustrate this.
Translation always implies an unstable balance between the power

one culture can exert over another. Translation is not the production
of one text equivalent to another text, but rather a complex process
of rewriting that runs parallel both to the overall view of language
and of the ‘Other’ people have throughout history; and to the influ-
ences and the balance of power that exist between one culture and
another.
(Álvarez/Vidal 1996:4)
This statement reflects a perspective characterised by postmodernist

thought. Translating does not take place in a vacuum; it takes place
against the backdrop of ideology.
If we are aware that translating is not merely passing from one text
to another, transferring words from one container to another, but
rather transporting one entire culture to another with all that this
entails, we realize just how important it is to be conscious of the
ideology that underlies a translation.
Rather than being free to decide how to render a given source text into a
target text, translators are constrained in many ways:
by their own ideology; by their feelings of superiority or inferior-

ity towards the language in which they are writing the text being
translated; by the prevailing poetical rules at that time; by the very
language in which the texts [sic] they are translating is written; by
what the dominant institutions and ideology expect of them; by the
public for whom the translation is intended.
Of course, it is not only the translator who is affected by the vari-

ous aspects of a translation-related politics. All agents involved in the
translation – whether on the producing side (such as the translator(s),
proofreader(s), any editor, or even the author of the source text) or on
the receiving side (such as the client and the readers) – are subject to the
multifarious influences of the politics by which they are surrounded.
Power Relations
The question, here, is: who or what influences the agents involved in the
translation? And who or what has the power to effect changes to a trans-
lated text and thereby determine its form and content? By way of answer
to these questions, we will give a brief outline of the politics and power
relations we consider relevant to translation. In principle, power can be
exerted by humans (as individual power) and by systems (as structural
power). Both kinds of power play a role in the politics of translation pro-
duction and translation assessment. The translator as the first and most
immediate agent involved in the composition of the target text usually
operates within an economic setting that pecuniarily restricts his or her
options of how to handle translation jobs and, in the end, of how to
translate. After all, the translator often has to make a living from trans-
lating. Therefore, he or she is not likely to refuse a translation job, even
if it is for reasons that might turn out to be detrimental to translation
quality – such as a tight deadline or low payment or a technical source
text about a subject matter with which the translator is not familiar.
The power relationship between the client and the translator is generally
determined by the economics of supply and demand48: the former is in a
more powerful position than the latter, if there are many other translators
available for a particular translation job; the latter is in a more powerful
position than the former, if a particular translation job cannot be done by
someone else. This situation has been described by Theo Hermans:
In conditions where individual translators cannot easily be dis-

pensed with, because alternatives are unavailable or too expensive,
for example, the translator’s clients have no option but to trust not
only the translator’s technical expertise but also his or her personal
and ideological loyalty. The translator’s power is [sic] such cases is
symbolic as well as material. Loyalty may have its price, and may
depend on whose side the translator is ultimately on.
(Hermans 1996:38–39)
Here, the power relationship between the client and the translator goes
beyond the effects of economic principles: it does not just determine the
price of the translation service, but affects the ideological issues involved
in disseminating the source text content in the target culture. The ques-
tion is, whether the translator has any scope to translate as he or she
pleases, or whether the translator is altogether trammelled by prescrip-
tive norms.
Of course, for the translator to be able to exert power in the process
of translation, the source text would first have to provide such an oppor-
tunity by furnishing a discourse that in some way represents a bone of
contention between the source and target cultures. Basil Hatim quotes,
here, the subtly chosen phrase “immigrants and their offspring” (Hatim
2009:91), which British politician Enoch Powell “was fond of using in
preference to, say, ‘immigrants and their children’” (Hatim 2009:92).
After listing a number of alternative expressions that might have been
used instead of the insinuatingly offensive noun “offspring”, Hatim
analyses:
Linguistic forms such as those from Powell’s speech are intertex-

tually seen by translators in terms of (a) a pre-discoursal linguistic
norm in which synonymy could be said to exist (e.g. offspring = chil-
dren); (b) an unmarked, register-based discourse (offspring = +legal);
and (c) a marked, imported discourse which involves the hijacking
of the normal discourse of (b), because Powell is not a lawyer but a
politician. The competition of the various discourses can ultimately
be reconciled by arriving at a reading which, while institutionally
sound (the text producer could not be taken to court for libel), is in-
tertextually pernicious: in the particular context under study, Pow-
ell’s remarks are dehumanizing and reminiscent of statements often
heard within racist discourse such as “they breed like rabbits”.
(Hatim 2009:92)
A translator translating the phrase “immigrants and their offspring”

into an immigrant language has the following options: depending on the
opportunities afforded by the target language, he or she can render the
word “offspring” by a term with approximately the same characteris-
tics and connotations as those shown in Hatim’s analysis; he or she can
use a near-synonym that reflects the offensive innuendo of the original
and is actually offensive; he or she can opt for a neutral synonym that
completely dispenses with intimations of offensiveness. The decision in
favour of one solution has to be made by the translator, who, in such a
case, is guided first and foremost by his or her attitude to translation.
Censorship
The restrictions encountered by the translator may be immanent in his
or her ethical stance or in the target culture, or they may be enforced
by the client. The latter situation seems to be not uncommon, as André
Lefevere writes: “Ideology is often enforced by the patrons, the peo-
ple or institutions who commission or publish translations” (Lefevere
1992:14). Here is an example: at a conference on comic translation,
which took place in Hildesheim, Germany, from 31 October to 2 No-
vember 2014, the first translator of the German Asterix comics, Gudrun
Penndorf, mentioned that she was not allowed to use vulgar or coarse
language but had to prepare her translation in line with a special list of
words provided by the publisher.49 Yet, most likely, the publisher would
not have bothered about providing such a list, were it not for the cultural
norms predominant at the time (the 1960s) in the target community. The
force that bears on translation, here, is that of censorship.
As Denise Merkle points out, “[c]ensorship has been justified on
aesthetic, moral, political, military and religious grounds” (Merkle
2010:18). In translation, the subtlest form of censorship is arguably
the non-accidental alteration of source text content in the target text,
whereas the most radical form is a total ban:
Famous instances of banned individual translations belonging to di-

verse literary domains (canonical texts, high and lowbrow culture,
children’s literature) and demonstrating the range and practices of
censorship include the following: Macchiavelli’s The Prince (banned
in France in 1576); The Thousand and One Nights (banned in the
USA in 1927); H. B. Stowe’s Uncle Tom’s Cabin (banned in Rus-
sia in 1852); and Lewis Carroll’s Alice’s Adventures in Wonderland
(banned in China in 1931), among many others.
(Billiani 2009:29)
Rather than prohibiting the publication of a translated text in order to

prevent people from reading what they should not read, censors may
also demand that the content of the source text be modified or selectively
deleted in translation. Nike Pokorn – who thoroughly investigated the
socialist impact on translation, comparing the translations of children’s
books before, during, and after the socialist era – notes that
[t]ranslators in all Socialist states, as in other totalitarian regimes,

were subject to different forms of censorship, ranging from punitive,
repressive or postcensorship to different forms of preventive or prior
censorship, as well as to the self-censorship of the translator. De-
spite the variety of different forms of censorship, it is argued here
that Socialist translation in different cultural and linguistic environ-
ments nevertheless purged the translated text of the same or similar
elements.
(Pokorn 2012:1)
Many of these elements were of a religious kind. The mention of God in

the source text, for instance, is avoided in translation through the use
of non-religious nominal expressions or passive constructions. 50 Other
elements that are frequently censored include obscene or offensive lan-
guage. For instance, “1950s male translators working in the US simply
excised [Simone de Beauvoir’s] daunting descriptions of awkward con-
traceptive contraptions and erotic love scenes” (Flotow 2009:124).
As has already been intimated in the Pokorn quotation above, censor-
ship need not be based on official regulations. Francesca Billiani argues
that
censorship has to be seen not as an institutional set of rules, or even

as an overtly repressive means of controlling public opinion and dis-
courses, but rather as a set of unwritten rules, shaped both by cur-
rent habitus and by the symbolic capital a text enjoys in a certain
field.
(Billiani 2009:28)
A translator whose text is to survive in an environment governed by

strict norms will have to stick to these norms by avoiding whatever dis-
pleases and anticipating whatever pleases the authorities in charge and/
or the readership. Censorship operates in the translator’s head – whether
consciously or unconsciously – and the strictness of its enforcement
seems to be directly proportionate to the size of the target audience,
as Denise Merkle points out: “Generally speaking, the broader the in-
tended audience is, the more rigorous the censorship” (Merkle 2010:18).
Or, metaphorically speaking: the more people a textual virus can infect,
the more effort is needed to eradicate or contain that virus.
While censorship is often regarded as something negative – being as-
sociated with dictatorial regimes and other forms of repression – the
change of source text content in the target text may also be seen in a
rather positive light. This is probably true for “the reception of rewritten
children’s literature – often positively connoted through the choice of
the noun ‘adaptation’ – and of sanitised subtitled movies for the gen-
eral public (GP)” (Merkle 2010:19). Here, rather than being repressive,
“censorship is perceived […] as in the best interests of the ideological
positioning of a larger socio-political entity (for example, children’s lit-
erature, GP films)” (ibid.). Yet, who defines what is good or bad for a
given readership or audience? To what extent, if at all, is there a con-
sensus among translators, publishers or any other agents involved in the
production of a particular text as to the inherent quality of controversial
passages? These questions have no obvious answer; they show that cen-
sorship is not always easily found out.
More Power Issues

Translation is also subject to influences that stem from power relations
on the fringes of censorship and beyond. These influences often ema-
nate from the target texts themselves. In certain intercultural power
constellations, for example, translations may be used as “an instru-
ment of colonial domination, producing hegemonic representations of
the colonized” (Brownlie 2009:79). 51 What is at work, here, is not so
much the translator’s intention to affirm the colonial facts, but the as-
sumed cultural superiority of the target culture over the source culture.
Furthermore, the relative significance of a culture is often reflected in
its literary achievements and their renderings in other languages. André
Lefevere writes:
Translators not infrequently use their translations to influence the

evolution of the poetics of their time. […] The compromises transla-
tors find between the poetics of the original and the poetics of their
culture provide fascinating insights into the process of acculturation
and incontrovertible evidence of the extent of the power of a given
poetics.
(Lefevere 1992:26)
Yet, what does “the power of a given poetics” consist in? And to what
extent does the power of poetics depend on the power of the language
and culture underlying such poetics? There is no straightforward answer
to these questions: while the power of a given poetics would seem to
consist in the ability of a culture’s poetics to spread to other cultures by
way of translation, the successful adaptation of the poetics of a source
culture to the requirements imposed by the poetics of the target culture
might equally be considered indicative of a powerful poetics. Which one
is stronger: a language and culture that is able to adapt or a language
and culture that remains unchangeable like a bedrock? Johann Wolf-
gang von Goethe – philosophising about literature and life – seems to
have been in favour of the former: the power of a language, he writes, is
not that it rejects the foreign but that it devours it (cf. Goethe 1950:789).
It is often in (literary) translations that the analysis of power relations
can yield the most interesting results.
In order to detect power in translated texts, it is necessary to analyse
the circumstances that gave rise to the translation in the first place and
to examine any signs and traces of power in the wording of the final
rendering. Luise von Flotow describes how to scan translated texts for
gender-related power issues:
When gender serves as a lens for the micro-analysis of individual

translations, the focus is on the minute details of language that (may)
reflect the gendered aspects of a text, or seek to conceal them (often
in the case of homosexual writings). Translations can be shown to
be sensitive to such manifestations of gender, exaggerate them or
ignore and obscure them. Often, the translation effects discerned
through such analyses provide clues about the cultural and political
literary climate of the translating culture, or can be understood as a
facet of this climate.
(Flotow 2009:124)
While, in principle, the analysis of any translation may yield interesting

insights into the power relations depicted in the translated text and
into the power relations operating between the cultures connected by
the translation, there are certain texts and translation situations that
are of particular interest in this context. In general, it seems that the
texts most affected by translation politics are literary texts. According
to Lefevere, translations of central texts (such as the Bible) are under
particular scrutiny, because any tampering with what is regarded as the
“original” might subvert the culture in which that text is seminal; and
translations from peripheral cultures into what might be called central
cultures may tend to absorb the cultural idiosyncrasies of the source
into the target in such a way that the source is no longer recognisable
in the target:
It is in the treatment of texts that play a central role within a culture

and in the way a central culture translates texts produced by cul-
tures it considers peripheral, that the importance of such factors as
ideology, poetics, and the Universe of Discourse is most obviously
revealed.
(Lefevere 1992:70)
And such revelation of politics in translation brings with it a qualita-

tive difference to an assumed translational ideal without that ideological
influence.
Politics and the Evaluation of Translation Quality

Measuring the qualitative impact of politics on a given translation may
be fairly straightforward, as in the case of open censorship, or it may
require a close analysis of the translated text, if the rules that prohibit
certain textual manifestations operate at subliminal levels. Even after a
close comparative analysis of the source and target texts, the share that
politics may have had in controlling the translation process and the out-
come of that process will hardly ever be established for certain. This is
due to the fact that a translator’s decisions usually remain unexplained.
Moreover, they may not even be “decisions” in the narrower sense of
the term: what appears to have been paradigmatically selected from a
range of alternatives may actually have been inserted in the target text
unthinkingly as the result of some translational automatism. Finding
the source of, or reason for, a particular translation solution is interest-
ing, especially, if the translation process and its optimisation constitute
the object of research; its relevance to translation quality assessment,
though, is mostly reflected in the possible knowledge of why a solution
has been selected, which, in turn, may help to establish the underlying
translation strategy.
When it comes to evaluating the quality of a translation that has been
shaped to some degree by political influences, the question is whether
the evaluator is aware of the influences under which the target text has
been produced. With censorship often working covertly, this question
will, in many cases, have to be answered in the negative. Where such
censorship can be taken for granted with considerable certainty, the
evaluator should make allowances for the trouble caused to the transla-
tor by such political influence. This is certainly the point of view taken
by Anna Pavlova, who gives the following example:
Wenn im Translat von Heinrich Bölls Gruppenbild mit Dame die Üb-
ersetzerin Ludmila Černaja die Erwähnung der Liebe zum gleichen
Geschlecht im Translat weglässt, so ist es keine freiwillige Entschei-
dung ihrerseits und auch kein Nachweis ihrer mangelnden Profes-
sionalität, sondern ein Zugeständnis gegenüber der Sowjetzensur.
(Pavlova 2014:259)
When, in the translation of Heinrich Böll’s Gruppenbild mit Dame,

the translator Ludmila Černaja leaves out a reference to homosexual
love, then this is not a voluntary decision on her part, nor proof of a
lack of professionalism, but a concession made to Soviet censorship.
As an evaluator, Pavlova finds that, from an objective point of view,

many translations of Böll, Dürrenmatt or Frisch from the Soviet era are
blatantly bad (cf. Pavlova 2014:259). However, she does not blame the
translators, because they could not help it. Pavlova regards as crucial in
this context the degree of freedom or the boundedness of the transla-
tional decision (ibid.). In other words: what options are left to the trans-
lator (or whether there are any options left to the translator) should play
a role in the assessment of a translation.
Summary of Politics Factors
While it would be presumptuous to suggest better alternatives in a trans-
lation situation governed by Soviet censorship (or a similarly obvious
control system), such a procedure may cautiously be followed in cases
where censorship or other restrictive powers are less obvious and subject
to speculation. The evaluator has to find a compromise between two op-
posing views: on the one hand, the translator should be excused for any
unsatisfactory translation solution that is due to political powers acting
on the translator and the translation process; on the other hand, any
unsatisfactory solution which the evaluator can improve on casts a harsh
light on the translator and his or her translation skills. The assessment
of the target text, here, stands and falls with the evaluator’s ability to
determine the impact of any power relationships at work in the trans-
lation and its environment. Unless it is clear that not the translator but
some political machinations can be held responsible for a poor transla-
tion solution, any such solution should not be counted in the assessment
of the overall target text. That is to say, politics should be relevant to
the assessment of translation quality, if it leaves the translator without
any chance to stave off any negative effects on him- or herself: a bad
translation in such a case is only as bad as it might have been without
the political influence. However, politics is irrelevant to TQA, if it can-
not be made out unequivocally and remains subject to speculation: any
substandard translation will then get the criticism it deserves.
This brings to an end our discussion of the various groups of factors that
have an impact on translation quality and the assessment of that quality.
What remains to be looked at is the role played by the evaluator(s) in the
evaluation process: including the translator and any proofreaders as work-
in-progress evaluators, as well as a client, reader, or reviewer as evalua-
tors of the finalised translation. Whether as translator, proofreader, client,
reader, or reviewer – the role of the evaluator is crucial, because it is he or
she who defines the yardstick against which to gauge the quality of a trans-
lation. The evaluator analyses a given target text within the framework of
the “translator’s daffodil”, culling whatever information he or she can get
that is relevant to the quality and evaluation of the text. This information
is used together with the evaluator’s knowledge of the world to set up a
standard for the nonce. That is to say, the evaluator’s standard for the
assessment of a particular translation is always unique. The standard will
change immediately, if there is another evaluator or another target text.
The quality factors discussed in this chapter will be used to form ar-
guments and argument structures which, in turn, will either substanti-
ate the translator’s decisions or reveal their inadequacy (see Figure 4.4).
In Chapter 5, we will demonstrate how the evaluator can assess the
quality of individual translation solutions by drawing upon any of the
above-mentioned factors relevant to a given translation problem.
Strategy 1 Solution A1 Solution B1

Translation decisions
… … …
↑ ↑ ↑
Arguments and argument structures: show whether translation decisions are appropriate
↑ ↑ ↑
Quality factors: make up the content of arguments
uality factors → arguments → translation decisions.

Figure 4.4 Q
Notes
1 Basil Hatim stresses the importance of the context when he writes: “[I]n
translation, there is hardly a decision taken regarding any element of lan-
guage in use at whatever level of linguistic organization, without constant
reference being made to the text in which that element is embedded” (Hatim
1997:12). Olga Horn reminds us that the conditions of the translation pro-
cess should be taken into consideration when evaluating a translated text,
which includes factors such as the target audience as well as the ideas of the
client and of the source text author:
[E]ine Bewertung [darf sich] nicht auf den Vergleich zwischen Ausgangs-
und Zieltext beschränken, sondern muss außerdem die Bedingungen
des Übersetzungsprozesses berücksichtigen. Denn auf die Gestaltung
des Translats hat nicht nur der Übersetzer Einfluss, sondern auch sol-
che Faktoren wie der angestrebte Rezipientenkreis, die Vorstellungen des
Auftraggebers und des Originalautors.
(Horn 2006:111–112)

Hans Vermeer (2006:295–302), in his holistic attempt at analysing trans-
lation in context, provides a seemingly endless list of factors that affect
the process of translation. And Fritz Senn (1994:54) finds that the num-
ber of aspects relevant to translation cannot be limited (“Die Zahl der für
das Ü bersetzen wesentlichen Gesichtspunkte läßt sich nicht beschränken”).
Jeremy Munday is more specific:
A translator/interpreter as [sic] an active participant in the communica-
tion process, one who ‘intervenes’ not as a transparent conduit of mean-
ing but as an interested representer of the source words of others and in
a communicative situation constrained and directed by extratextual fac-
tors including commissioner, brief, purpose, audience expectation and
target text function. In addition, the translator or interpreter brings his/
her own sociocultural and educational background, ideological, phra-
seological and idiosyncratic stylistic preferences to the task of rendering
a source text in the target language.
(Munday 2012:2)
2 This is reminiscent of Melby’s “translation parameter tree” (Melby 2012, no
pagination) – 21 parameters that should be specified prior to the production
phase of a translation.
3 Any translator-related information, while usually unknown to the evalua-
tor, is, of course, known to the translator. In most such cases, though, this
information is not relevant to the evaluation of a translation.
4 While poor payment of the translator is certainly not a criterion that would
be drawn upon in the evaluation of a translation’s quality, there may be
other “mitigating circumstances” which the evaluator should take into ac-
count, such as any restrictions due to censorship.
5 A different categorisation of quality criteria has been proposed by Sylvia
Reinart, who distinguishes between criteria that are derived from the source
text and criteria that are imposed on the translation from outside (cf. Rein-
art 2014:38).
6 This is true even of pseudo translations. Here, the (lost or non-existent)
source text is the topic of a (fictitious) story of how the translator came to
write the translation. Those who believe that the story is true will accept
the target text as translation; those who do not will deny the target text
the status of translation. In other cases, the source text cannot be identi-
fied, as “there is no proper source text that was exclusively produced in
one language and culture, and conforming to source-culture-specific text-
typological conventions” (Schäffner 1997:208). Here, Schäffner refers to the
Manifesto for the Elections to the European Parliament of June 1994 – a
document that was first drafted in German, then revised and amended us-
ing English, French, German, and Italian as working languages, and finally
translated into the other languages of the EU member states (cf. Schäffner
1997:194). Two more examples of translations that do not rely on a clearly
identifiable source text are given at the end of this section.
7 For a systematic approach to the concept of equivalence, cf. Koller (2011).
Given our concept of translation as that which is regarded as a translation,
we are not taking sides in the debate on whether the term “equivalence” ad-
equately captures the nature of translation. Cf., for example, Snell-Hornby
(1994:13ff), Neubert (1994:87), Kußmaul (1994:224f), and Chesterman
(1994:155–156). However, we would certainly agree with Anthony Pym
that “[e]quivalence is involved in the psychology of translators when they
work as translators, and indeed in the work of quality controllers whenever
they evaluate the work of translators as translators” (Pym 2004:67). The
mere notion of translating something into something else implies a relation-
ship between the “somethings” that retains a minimum degree of sameness,
usually referred to as “equivalence”. Holger Siever (2010:81) suggests that
“Implikation” (that is, implication) might be a more appropriate term, signi-
fying as it does a unilateral rather than bilateral relationship. Yet, while the
concept of implication may not succeed in replacing the concept of equiva-
lence, it may at least complement it (cf. Schreiber 2014:18).
8 In addition to interlingual translation, there is also intralingual translation
and intersemiotic translation (cf. Jakobson 1959:261). These will be dis-
cussed in the section on text form.
9 While this may seem too obvious to be worth mentioning, some postmodern
approaches to translation would seem to dissolve – or reduce to insignificance –
the temporal and spatial difference between source and target. Consider, by
way of example, Clive Scott’s notion of translation as “a mode of reading”
(Scott 2012:10), where “every reader should be a translator and […] no other
translator can translate our reading for us, although other translators may
change the way we read” (ibid.).
10 On this mode of translating, cf. Vermeer (2003:20–21).
11 Many interesting examples of terminology issues in translation can be found
in Reinart (2014:105–118).
12 Cf. Bittner (1997b:21ff) for a detailed discussion of the relation between
rhythm and metre.
13 The poetic function is most acutely defined by Roman Jakobson, who writes:
“The poetic function projects the principle of equivalence from the axis of
selection into the axis of combination” (Jakobson 1960:358). In non-poetic
language, words are selected on the basis of equivalence on the paradigmatic
axis of selection; in poetic language, words are selected on the basis of equiv-
alence on the syntagmatic axis of combination. Thus, our choice of words
in non-poetic language is governed by semantic suitability alone, whereas
in poetry the linguistic structures assume an importance that outshines the
impact of lexical meaning.
14 So complicated, indeed, that the following remark by Jean Boase-Beier is
not surprising: “The central question that all studies of the translation of
poetry have asked, implicitly or explicitly, is whether poetry can be trans-
lated” (Boase-Beier 2009:194). Poets seem to be particularly sceptical in this
respect. Percy Bysshe Shelley, for one, writes about the vanity of translating
poetry: “[I]t were as wise to cast a violet into a crucible that you might dis-
cover the formal principle of its colour and odour, as seek to transfuse from
one language into another the creations of a poet” (Shelley 1821/1840:782).
15 Nevertheless, most poetry can be translated. How this has been achieved
from English (and French) into German is well documented in Wittbrodt
(1995). After all, what is possible in practice cannot be impossible in the-
ory: “Es kann nicht theoretisch unmöglich sein, was praktisch möglich ist”
(Schneiders 2007:72).
16 As Burton Raffel remarks rather apodictically: “Poetry is not footnotes, is
not definitions. It either is poetry or it is false coin. And poetry in translation
is either poetry born anew or it is nothing at all” (Raffel 1971:115).
17 Anderman (2009:94–95) mentions the reviewers’ responses to two different
translated versions of Anouilh’s Antigone.
18 A useful overview of comics and their translation is given by Reinart
(2014:247–261). That the translation of comics is, so to speak, a form of
translation in its own right as it cannot be categorised unambiguously with
any other form of translation is shown by Mälzer (2014:636).
19 See Mälzer (2013) for a detailed discussion of the different texts and codes
involved in audiovisual translation, particularly, in subtitling and dubbing.
Jüngst (2010) and Reinart (2014:236–337) provide a good general overview
of the various forms of audiovisual translation. For a comprehensive discus-
sion of subtitling, cf. Díaz Cintas/Remael (2007). A code of good subtitling
practice has been drawn up by Ivarsson/Carroll (1998). Kurz (2006) pre
sents the historical and theoretical aspects of dubbing, before analysing the
German dubbing version of the film “Lock, Stock and Two Smoking Barrels”.
20 The voice-over technique is not restricted to audiovisual media but can also
be used, for example, in a radio broadcast.
21 The polysemiotic environment of a film is also mentioned, for example, in
Heinze (2005:13).
22 A more detailed account of subtitling for the deaf and hard of hearing can be
found in Jüngst (2010:123–137).
23 The requirement that subtitles need to be carefully timed for the audience
to be able to read them in full seems to become comparatively insignificant
in live subtitling, a form of subtitling for the deaf and hard of hearing in
which live broadcasts on TV (such as sports events, news, interviews, and
the weather forecast) are subtitled during the broadcast, sometimes with a
delay of a few seconds after the actual live event. As Pablo Romero-Fresco
points out, “broadcasters, companies and deaf associations seem to choose
verbatim (and therefore faster) subtitles, whereas academics and researchers
usually prefer edited (and slower) subtitles” (Romero-Fresco 2011:16). The
preference of companies and broadcasters is clear, since verbatim subtitling
is more cost-efficient than edited subtitling. That even deaf associations are
in favour of verbatim subtitles, although studies have shown that the amount
to be read in such a case may often be incompatible with the reading speed
of the target audience (cf. Romero-Fresco 2011:112ff), can be explained by
an indefinite fear that the deaf or hard of hearing may miss out on important
information, if what is said is summarised in the subtitle. Romero-Fresco
writes: “[M]any deaf viewers equate editing to censorship and therefore
support verbatim subtitles, regardless of their speed, as the only method
to provide them with full access to the original content” (Romero-Fresco
2011:118). In the United Kingdom, the Office of Communications (Ofcom)
monitors the quality of live subtitles. Measuring subtitling speed, latency
(that is, the delay between speech and subtitle), and the number and type
of errors, Ofcom responded to complaints from viewers in 2013 by consult-
ing on proposals to improve the quality of live subtitling with stakeholders,
subtitle users, subtitling providers, broadcasters, and academics. See Ofcom
(2013, 2014a, 2014b) for a detailed presentation of the quality measurement
procedure and the results. For a study analysing the quality of German live
subtitles, cf. Kurch et al. (2015).
2
4 Pablo Romero-Fresco adds to this the possibility of providing “information
about the tone or volume of the speakers’ utterances (angry, sad, low, loud)”
(Romero-Fresco 2011:17).
2
5 A thorough comparison of audio description guidelines can be found in Bitt
ner (2012).
2
6 Such narrative neutrality notwithstanding, Harald Kautz-Vella (1998:21)
suggests that the audio describer experiment with the narrative devices used
in radio plays, and Dosch/Benecke (2004:34) give an example (“Bibi Blocks-
berg”) in which the audio narrator speaks in the role of a central character
of the film.
27 For a very clear description of the various distinctions between localisation
and translation, see Nauert (2007).
28 Note that, in translator training, the trainer takes on the role of the client
when providing the brief for the translation to be produced and any instruc-
tions that seem appropriate to the task at hand.
29 Klaus Schubert points out that the adjective in “technical translation” is
potentially ambiguous as it “can relate to content either from technology
and engineering or from any specialized domain” (Schubert 2010:350). In
his article, he uses the word in the first sense.
30 One might argue that giving a translator plenty of time to translate a speci-
fied source text has a positive effect on the quality of the translation. How-
ever, since a good translation is unmarked (or marked by a zero value) in
the sense that it is characterised by the absence of defects, errors, and mis-
takes, a translation is expected to be a good translation by default (cf. Emsel
2007:100) – sufficient time being but a standard prerequisite for translation.
31 Time is also an issue in literary translation. Javier Franco Aixelá gives the
following example:
In many countries, literary translators complain (we certainly do so a

lot in Spain) about working conditions that force them to translate very
fast and with nearly no time for revision. These conditions, together with
the lack of specialized training in a country that, like Spain, has not
had a university degree in translation until recently, explain a number
of incongruities and obvious misunderstandings. For example, one that
appears in a translation of Dashiell Hammett’s Red Harvest, where ‘a
parson named Hill’ becomes ‘un tal Hill’ (someone named Hill), because
of a probable confusion between ‘parson” and ‘person’, and not as the
result of any sort of conscious decision on the part of the translator.
(Franco Aixelá 1996:67)
32 There is one exception to this rule, namely, if the translation cannot be used
after the deadline has expired. This case has been discussed in the “Factor
Categories” section.
33 That the translator is not as independent an agent in the translation process
as it may seem at first has also been pointed out by André Lefevere: “Trans-
lators function in a given culture at a given time. The way they understand
themselves and their culture is one of the factors that may influence the way
in which they translate” (Lefevere 1992:14).
34 We will use these terms interchangeably. The reviser or proof-reader is, as it
were, the second internal evaluator within the translation process (after the
actual translator).
35 The extent to which quality is affected by machine translation seems to be
rather uncertain, as Joanna Drugan points out: “Of unclear impact for qual-
ity is the recent increase in MT use. Virtually without exception, translators
claimed that they would always prefer to translate texts from scratch, often
referring to quality as their justification” (Drugan 2013:33).
36 “A computer never mistakes a comma for a full stop; a human’s tired eyes
can easily do so” (Drugan 2013:31).
37 “In Spanish, the acronym PACTE corresponds to Process in the Acquisition
of Translation Competence and Evaluation” (PACTE 2011:317).
38 Sometimes, especially in literary translation, the translator comments on
his or her strategy or on individual translation decisions, thereby enabling
the evaluator or critic to assess the target text on the basis of the translator’s
assumptions and preferences. Such a basis of evaluation is recommended, for
example, by van den Broeck (1985:60–61) and by Reinart (2014:349–350).
39 For a more detailed discussion of the question when not to translate, cf.
Reinart (2014:344–346).
40 The translator’s ethics, too, play a role, here: “Being ethical”, writes Ben
van Wyke, “does not involve simply declaring fidelity, but, instead, sorting
through difficult decisions and taking responsibility for those taken” (Wyke
2010:113). From a slightly different angle, Cecilia Alvstad, Adelina Hild,
and Elisabet Tiselius look at how the translator’s behaviour is “guided by
both cognitive constraints […] and culturally acquired norms […]” (Alvstad
et al. 2011:2).
41 “Translation degrees or professional qualifications do not guarantee excel-
lent translations. Many such qualifications are assessed by essays on transla-
tion theory rather than hands-on practice” (Drugan 2013:70).
42 This distinction harks back to the so-called iceberg model, which is de-
scribed in Katan (2009:70–72).
43 The cultural norms also include translation standards (or other standards
relevant to translation) that have been explicitly set up by institutions, or-
ganisations, professional bodies, etc. in order to improve the quality of the
translation process and product. These translation standards are usually
more absolute and unequivocal than most other norms.
4
4 This phenomenon is also known under the name of relay translation. James
St. André defines it as “the translation of a translated text (either spoken or
written) into a third language (for example, from Chinese to English, then
from English to French)” (St. André 2009:230).
45 In professional translation, the fundamental question of translation quality
is not “How do we know when a translation is good?” (House 2001:243)
but rather “How do we know when a translation is good enough?” (Drugan
2013:42).
46 The situation in Japan in the twentieth century recalls Gideon Toury’s ob-
servation that “translations which deviate from sanctioned patterns […] are
often tolerated by a culture to a much higher extent than equally deviant
original compositions” (Toury 2005:4).
47 Cf. Bittner (1997b:59–60) and Prill (2003:11–12) for further information.
48 In a rather specific case, Andy Stauder discusses the economic characteristics
of screen translation on the German market and shows that the “factors of
a socio-economic dimension […] have a significant bearing on translation
quality” (Stauder 2014:93).
49 Oral communication made by Gudrun Penndorf in her conference presentation.
50 Cf. Pokorn (2012:127) for a few examples.
51 One example is given by Jeremy Munday, who states that “in the British
Victorian era […] where target culture values were deemed to be superior,
Edward Fitzgerald’s enormously influential translation of The Rubaiyat of
Omar Khayyām (1859) completely reworked and rearranged the Persian
[source text]” (Munday 2012:40).
References
Abrams, M. H. (ed.) (1986). The Norton Anthology of English Literature.
5th ed., vol. 2. New York/London: W. W. Norton & Company.
Ahrend, Klaus (2006). “Kriterien für die Bewertung von Fachübersetzungen”.
In: Schippel, Larisa (ed.) (2006:31–42).
Albrecht, Jörn (2009). “Literarisches Übersetzen – Linguistische Grundlagen”
In: Zybatow, Lew N. (ed.) (2009:3–44).
Al-Taher, Mohammad Anwar (2008). The Translation of Intertextual Expres-
sions in Political Articles. PhD thesis, University of Salford.
Álvarez, Román; M. Carmen-África Vidal (1996). “Translating: A Political
Act”. In: Álvarez, Román; M. Carmen-África Vidal (eds.) (1996:1–9).
Álvarez, Román; M. Carmen-África Vidal (eds.) (1996). Translation, Power,
Subversion. Clevedon, Philadelphia and Adelaide: Multilingual Matters.
Alvstad, Cecilia; Adelina Hild; Elisabet Tiselius (2011). “Methods and Strate-
gies of Process Research: Integrative Approaches in Translation Studies”. In:
Alvstad, Cecilia et al. (2011:1–9).
Alvstad, Cecilia; Adelina Hild; Elisabet Tiselius (eds.) (2011). Methods and
Strategies of Process Research: Integrative Approaches in Translation Stud-
ies. Amsterdam and Philadelphia: John Benjamins Publishing Company.
Anderman, Gunilla (2009). “Drama Translation”. In: Baker, Mona; Gabriela
Saldanha (eds.) (2009:92–95).
Baker, Mona (2009). “Norms”. In: Baker, Mona; Gabriela Saldanha (eds.)
(2009:189–193).
Baker, Mona; Sameh Fekry Hanna (2009). “Arabic Tradition”. In: Baker, Mona;
Gabriela Saldanha (eds.) (2009:328–337).
Bandia, Paul (2009). “African Tradition”. In: Baker, Mona; Gabriela Saldanha
(eds.) (2009:313–320).
Billiani, Francesca (2009). “Censorship”. In: Baker, Mona; Gabriela Saldanha
(eds.) (2009:28–31).
Bittner, Hansjörg (1997a). “Translating Heinrich Böll’s ‘Anekdote zur Senkung der
Arbeitsmoral’”. In: Norwich Papers V, 91–95.
Bittner, Hansjörg (1997b). The Metrical Structure of Free Verse. University of
East Anglia: Thesis.
Bittner, Hansjörg (2011). “The Quality of Translation in Subtitling”. In: trans-
bd04nr01/trans-kom_04_01_04_Bittner_Quality.20110614.pdf
Bittner, Hansjörg (2012). “Audio Description Guidelines – a Comparison”. In:
Norwich Papers, vol. 20, 41–61.
Boase-Beier, Jean (2009). “Poetry”. In: Baker, Mona; Gabriela Saldanha (eds.)
(2009:194–196).
Böll, Heinrich (1986). The Stories of Heinrich Böll, translated by Leila Venne-
witz. London: Secker & Warburg.
Brownlie, Siobhan (2009). “Descriptive vs. committed approaches”. In: Baker,
Mona; Gabriela Saldanha (eds.) (2009:77–81).
Carbonell, Ovidio (1996). “The Exotic Space of Cultural Translation”. In: Álva-
rez, Román; M. Carmen-África Vidal (eds.) (1996:79–98).
Chesterman, Andrew (1994a). “Quantitative Aspects of Translation Quality”.
In: Lebende Sprachen, vol. 39, no. 4, 153–156.
Díaz Cintas, Jorge; Aline Remael (2007). Audiovisual Translation: Subtitling.
Manchester: St. Jerome.
Dosch, Elmar; Bernd Benecke (2004). Wenn aus Bildern Worte werden – Durch
Audio-Description zum Hörfilm. 3rd ed. Munich: Bayerischer Rundfunk.
Drugan, Joanna (2013). Quality in Professional Translation. London/New
York: Bloomsbury Academic.
Jüngst (eds.) (2007:100–111).
Floros, Georgios (2003). Kulturelle Konstellationen in Texten: Zur Beschreibung
und Übersetzung von Kultur in Texten. Tübingen: Narr. Jahrbuch Übersetzen
und Dolmetschen, Band 3 (2002).
Flotow, Luise von (2009). “Gender and sexuality”. In: Baker, Mona; Gabriela
Saldanha (eds.) (2009:122–126).
Franco Aixelá, Javier (1996). “Culture-specific Items in Translation”. In: Álvarez,
Román; M. Carmen-África Vidal (eds.) (1996: 52–78).
Freigang, Karl-Heinz (1978). “Überlegungen zu einer theoretisch-linguistisch
fundierten Methodologie der Übersetzungswissenschaft”. In: Wilss, Wolfram
(ed.) (1981:150–170).
Friederich, Wolf (1969). Technik des Übersetzens. München: Hueber.
Furuno, Yuri (2005). “Translationese in Japan”. In: Hung, Eva (ed.)
(2005:147–160).
Gambier, Yves; Luc van Doorslaer (eds.) (2010). Handbook of Translation
Studies, vol. 1. Amsterdam and Philadelphia: John Benjamins.
Goethe, Johann Wolfgang von (1950). Goethes poetische Werke, vollständige
Ausgabe, zweiter Band. West-östlicher Divan, Epen, Maximen und Reflex-
ionen. Stuttgart: Cotta.
Granzin, Katharina (2010). “Arme Stieftochter Übersetzungskritik”. In:
taz.de (31.7.2010) [online]. Viewed: 14 August 2019. http://www.taz.
de/!401238/
Greiner, Norbert (2004). Übersetzung und Literaturwissenschaft. Tübingen:
Gunter Narr. (vol. 2 of the series Grundlagen der Übersetzungsforschung).
Grutman, Rainier (2009). “Self-Translation”. In: Baker, Mona; Gabriela Sal-
danha (eds.) (2009:257–260).
Hansen, Gyde (2010). “Translation ‘Errors’”. In: Gambier, Yves; Luc van
Doorslaer (2010:385–388).
Hatim, Basil (1997). Communication Across Cultures: Translation Theory and
Contrastive Text Linguistics. Exeter: University of Exeter Press (Exeter Lin-
guistic Studies).
Hatim, Basil (2009). “Discourse Analysis”. In: Baker, Mona; Gabriela Saldanha
(eds.) (2009:88–92).
Heinze, Hendrik (2005). “Good Subtitling Practice”: Konventionen der Unter-
titelung in Film und Fernsehen. München: Grin Verlag.
Hermans, Theo (1996). “Norms and the Determination of Translation: A The-
oretical Framework”. In: Álvarez, Román; M. Carmen-África Vidal (eds.)
(1996:25–51).
Helm.
Holz-Mänttäri, Justa (1984). Translatorisches Handeln: Theorie und Methode.
Helsinki: Suomalainen Tiedeakat.
Horn, Olga (2006). “Elemente des Substandards in Viktor Pelevins Roman Ča-
paev i Pustota und ihre Wiedergabe in der Übersetzung von Andreas Tret-
ner”. In: Schippel, Larisa (ed.) (2006:109–129).
Tübingen: Narr.
243–257.
House, Juliane (ed.) (2014:241–264).
House, Juliane (ed.) (2014). Translation: A Multidisciplinary Approach. Hound-
smills: Palgrave Macmillan.
Hung, Eva (ed.) (2005). Translation and Cultural Change: Studies in History,
Norms and Image-Projection. Amsterdam and Philadelphia: John Benjamins
Publishing Company.
Hung, Eva; David Pollard (2009). “Chinese Tradition”. In: Baker, Mona; Gabri-
ela Saldanha (eds.) (2009:369–378).
Ivarsson, Jan; Mary Carroll (1998). Code of Good Subtitling Practice. Available
online from ESIST (European Association for Studies in Screen Translation).
Viewed: 16 August 2019. www.esist.org/wp-content/uploads/2016/06/Code-
of-Good-Subtitling-Practice.PDF.pdf
Jakobson, Roman (1959, 1971). “On Linguistic Aspects of Translation”. Se-
lected Writings II: Word and Language, pp. 260–266. The Hague and Paris:
Mouton.
Jakobson, Roman (1960). “Closing Statement: Linguistics and Poetics”. In:
Sebeok, Thomas A. (ed.) (1960:350–377).
Jüngst, Heike Elisabeth (2010). Audiovisuelles Übersetzen – Ein Lehr- und Ar-
beitsbuch. Tübingen: Narr.
Kade, Otto (1968). Zufall und Gesetzmäßigkeit in der Übersetzung. Beihefte
zur Zeitschrift Fremdsprachen: Zeitschrift für Dolmetscher, Übersetzer und
Sprachkundige 1. Leipzig: Verlag Enzyklopädie.
Kang, Ji-Hae (2009). “Institutional Translation”. In: Baker, Mona; Gabriela
Saldanha (eds.) (2009:141–145).
Katan, David (2009). “Culture”. In: Baker, Mona; Gabriela Saldanha (eds.)
(2009:70–73).
Kautz-Vella, Harald (1998). Audiodeskription – Filme beschreiben für Blinde.
Berlin: Rubikon.
Kittel, Harald; Andreas Poltermann (2009). “German Tradition”. In: Baker,
Koller, Werner (2011). Einführung in die Übersetzungswissenschaft. 8th ed.
Tübingen: Francke.
Kondo, Masaomo; Judy Wakabayashi (2009). “Japanese Tradition”. In: Baker,
Krüger, Ralph (2015). “Fachübersetzen aus kognitionstranslatologischer Per-
spektive: Das Kölner Modell des situierten Fachübersetzers”. In: trans-
kom, vol. 8, no. 2: 273–313. Viewed: 16 August 2019. www.trans-kom.eu/
bd08nr02/trans-kom_08_02_01_Krueger_Modell.20151211.pdf
Kurch, Alexander; Nathalie Mälzer; Katja Münch (2015). “Qualitätsstudie zu
Live-Untertitelungen – am Beispiel des ‚TV-Duells‘”. In: trans-kom, vol. 8, no.
1, 144–163. Viewed: 16 August 2019. www.trans-kom.eu/bd08nr01/trans-
kom_08_01_07_Kurch_Maelzer_Muench_Duell.20150717.pdf
Kurz, Christopher (2006). Filmsynchronisation aus übersetzungswissenschaft-
licher Sicht – Eine kontrastive Synchronisationsanalyse des Kinofilms „Lock,
Stock and Two Smoking Barrels“. Hamburg: Verlag Dr. Kovač.
Kußmaul, Paul (1994). “Übersetzen als Entscheidungsprozeß. Die Rolle der
Fehleranalyse in der Übersetzungsdidaktik”. In: Snell-Hornby, Mary (ed.)
(1994:206–229).
Lefevere, André (1996). “Translation and Canon Formation: Nine Decades of
Drama in the United States”. In: Álvarez, Román; M. Carmen-África Vidal
(eds.) (1996:138–155).
Lefevere, André (ed.) (1992). Translation/History/Culture: A Sourcebook.
L ondon and New York: Routledge.
Mälzer, Nathalie (2013). “Der Einfluss der Übersetzungsmodalitäten auf den
filmischen Dialog”. In: trans-kom 6 [2] (2013): 260–284. Viewed: 16 Au-
gust 2019. www.trans-kom.eu/bd06nr02/trans-kom_06_02_01_Maelzer_
Modalitaeten.20131212.pdf
Mälzer, Nathalie (2014). “Was ist eigentlich Transmediale Übersetzung?”. In:
Forschung und Lehre, vol. 8, no. 14, 636–637.
Melby, Alan K. (2012). Structured Specifications and Translation Parameters
(version 6.0). Viewed: 16 August 2019. www.ttt.org/specs/
Merkle, Denise (2010). “Censorship”. In: Gambier, Yves; Luc van Doorslaer
(2010:18–21).
Munday, Jeremy (2012). Evaluation in Translation. Critical Points of Transla-
tor Decision-Making. London and New York: Routledge.
Nauert, Sandra (2007). Translating Websites. MuTra 2007 – LSP Translation
Scenarios: Conference Proceedings (EU-High-Level Scientific Conference Se-
ries). Viewed 16 August 2019. www.euroconferences.info/proceedings/2007_
Proceedings/2007_Nauert_Sandra.pdf
Neubert, Albrecht (1994). “Translatorische Relativität”. In: Snell-Hornby,
Mary (ed.) (1994:85–105).
Nord, Christiane (2005, 1991). Text Analysis in Translation: Theory, Method-
ology, and Didactic Application of a Model for Translation-oriented Text
Analysis. 2nd ed. (1st ed. 1991), Translated by C. Nord and P. Sparrow. Am-
sterdam and New York: Rodopi. (Amsterdamer Publikationen zur Sprache
und Literatur 94.) Original: Christiane Nord (1988). Textanalyse und Üb-
ersetzen: Theoretische Grundlagen, Methode und didaktische Anwendung
einer übersetzungsrelevanten Textanalyse. 4th ed. 2009. Tübingen: Groos.
Ofcom (2013). Measuring the Quality of Live Subtitling – Statement. Viewed:
16 August 2019. http://stakeholders.ofcom.org.uk/binaries/consultations/
subtitling/statement/qos-statement.pdf
Ofcom (2014a). Measuring Live Subtitling Quality – Results from the First
Sampling Exercise. Viewed: 16 August 2019. www.ofcom.org.uk/__data/
assets/pdf_file/0019/45136/sampling-report.pdf
Ofcom (2014b). Measuring Live Subtitling Quality – Results from the Second
Sampling Exercise. Viewed: 16 August 2019. www.ofcom.org.uk/__data/
assets/pdf_file/0024/47625/qos_second_report.pdf
PACTE (2011). “Results of the Validation of the PACTE Translation Compe-
tence Model: Translation Problems and Translation Competence”. In: Alv
stad, Cecilia et al. (eds.) (2011:317–343).
Palmer, Jerry (2009). “News Gathering and Dissemination”. In: Baker, Mona;
Gabriela Saldanha (eds.) (2009:186–189).
Pavlova, Anna (2014). “Strategie der Übersetzung und Beurteilung der Über-
setzungsqualität”. In: trans-kom, vol. 7, no. 2, 256–271. Viewed: 12 Au-
gust 2019. www.trans-kom.eu/bd07nr02/trans-kom_07_02_05_Pavlova_
Uebersetzungsqualitaet.20141219.pdf
Pöckl, Wolfgang; Ulrich Prill (eds.) (2003). Übersetzung und Kulturwandel.
Edition Praesens (no place name).
Pokorn, Nike K. (2012). Post-Socialist Translation Practices: Ideological Strug-
gle in Children’s Literature. Amsterdam and Philadelphia: J. Benjamins (Ben-
jamins Translation Library, vol. 103).
Prill, Ulrich (2003). “Übersetzen: Beschreiben, Erfinden und Umschreiben”. In:
Pöckl, Wolfgang; Ulrich Pröll (eds.) (2003:9–15).
vol. 48, no. 4, 481–497.
Pym, Anthony (2004). The Moving Text: Localization, Translation, and Distri-
bution. Amsterdam and Philadelphia: John Benjamins.
Raffel, Burton (1971). The Forked Tongue – A Study of the Translation Process.
The Hague/Paris: Mouton.
Reinart, Sylvia (2014). Lost in Translation (Criticism)? – Auf dem Weg zu einer
konstruktiven Übersetzungskritik. Berlin: Frank & Timme (Transkulturalität –
Translation – Transfer, vol. 5).
Reinke, Uwe (1997). “Integrierte Ubersetzungssysteme – Betrachtungen zu
Übersetzungsprozeß, Übersetzungsproduktivität, Übersetzungsqualität und
Arbeitssituation”. In: Lebende Sprachen, vol. 42, no. 3, 97–106.
München: Max Hueber (Hueber Hochschulreihe 12).
Risku, Hanna (1998). Translatorische Kompetenz: kognitive Grundlagen des
Übersetzens als Expertentätigkeit. Tübingen: Stauffenburg.
Risku, Hanna (2004). Translationsmanagement. Interkulturelle Fachkommu-
nikation im Kommunikationszeitalter. Tübingen: Narr.
Romero-Fresco, Pablo (2011). Subtitling through Speech Recognition: Respeak-
ing. Manchester: St. Jerome.
Salama-Carr, Myriam (2009). “French Tradition”. In: Baker, Mona; Gabriela
Saldanha (eds.) (2009:404–410).
Schäffner, Christina (1997). “Where is the source text?”. In: Schmidt, Heide;
Gerd Wotjak (1997:193–211).
Schäffner, Christina (2010). “Norms of Translation”. In: Gambier, Yves; Luc
van Doorslaer (2010:235–244).
Schippel, Larisa (ed.) (2006). Übersetzungsqualität: Kritik – Kriterien –
Bewertungshandeln. Berlin: Frank & Timme (Arbeiten zur Theorie und
Schmidt, Heide; Gerd Wotjak (eds.) (1997). Modelle der Translation: Festschrift
für Albrecht Neubert. Frankfurt/Main: Vervuert.
Schmitt, Peter A.; Heike E. Jüngst (eds.) (2007). Translationsqualität. Frankfurt/
Main etc.: Peter Lang.
Schneiders, Hans-Wolfgang (2007). Allgemeine Übersetzungstheorie: Verste-
hen und Wiedergeben. Bonn: Romanistischer Verlag (Abhandlungen zur
Sprache und Literatur 169).
Schreiber, Michael (2014). “Implikation statt Äquivalenz? Anmerkungen zu ei-
nem Vorschlag”. In: Studia Translatorica 5, 11–19.
Schubert, Klaus (2007). Wissen, Sprache, Medium, Arbeit: Ein integratives
Modell der ein- und mehrsprachigen Fachkommunikation. Tübingen: Gunter
Narr.
Schubert, Klaus (2010). “Technical Translation”. In: Gambier, Yves; Luc van
Doorslaer (2010:350–355).
Scott, Clive (2012). Literary Translation and the Rediscovery of Reading.
Cambridge [etc.]: Cambridge University Press.
Sebeok, Thomas A. (ed.) (1960). Style in Language. Cambridge: MIT.
Senn, Fritz (1994). “Literarische Übertragungen – empirisches Bedenken”. In:
Snell-Hornby, Mary (ed.) (1994:54–84).
Shelley, Percy-Bysshe (1821/1840). “A Defence of Poetry”. In: Abrams, M. H.
(ed.) (1986:777–792).
Siever, Holger (2010). Übersetzen und Interpretation. Die Herausbildung der
Übersetzungswissenschaft als eigenständige wissenschaftliche Disziplin
im deutschen Sprachraum von 1960 bis 2000. Frankfurt/Main etc.: Peter
Lang.
Snell-Hornby, Mary (1994). “Einleitung – Übersetzen, Sprache, Kultur”. In:
Snell-Hornby, Mary (ed.) (1994). Übersetzungswissenschaft –Eine Neuorien-
tierung. 2nd ed. Tübingen and Basel: Franke.
St. André, James (2009). “Relay”. In: Baker, Mona; Gabriela Saldanha (eds.)
(2009:230–232).
Stauder, Andy (2014). “Screen Translation: Quality through Transparency, De-
mocratisation and Openness to Science?”. In: Studia Translatorica, vol. 5,
91–99.
Torresi, Ira (2009). “Advertising”. In: Baker, Mona; Gabriela Saldanha (eds.)
(2009:6–10).
Toury, Gideon (1995). Descriptive Translation Studies and Beyond. Amster-
dam and Philadelphia: John Benjamins.
Toury, Gideon (2005). “Enhancing Cultural Changes by Means of Fictitious
Translations”. In: Hung, Eva (ed.) (2005:3–17).
Vermeer, Hans J. (2003). “Die sieben Grade einer Translationstheorie”. In:
Studia Germanica Posnaniensia XXIX, 19–38.
Vermeer, Hans J. (2006). Versuch einer Intertheorie der Translation. Berlin:
Frank & Timme.
Vermeer, Hans J. (2007). Ausgewählte Vorträge zur Translation und anderen
Themen. Berlin: Frank & Timme.
Wilss, Wolfram (ed.) (1981). Übersetzungswissenschaft. Darmstadt: Wissen-
schaftliche Buchgesellschaft. Wege der Forschung, vol. 535.
Wittbrodt, Andreas (1995). Verfahren der Gedichtübersetzung: Definition,
Klassifikation, Charakterisierung. Frankfurt/Main: Lang.
Witte, Heidrun (2000). Die Kulturkompetenz des Translators: begriffliche
Grundlegung und Didaktisierung. Tübingen: Stauffenburg.
Wyke, Ben van (2010). “Ethics and Translation”. In: Gambier, Yves; Luc van
Doorslaer (2010:111–115).
Zybatow, Lew N. (ed.) (2009). Translation: Neue Entwicklungen in Theorie und
Praxis: SummerTrans-Lektionen zur Translationswissenschaft – IATI-Beiträge
I. Frankfurt/Main etc.: Peter Lang (Forum Translationswissenschaft, vol. 11).
5 The Principle of
Argumentation
Theories of translation quality assessment usually establish a framework

of translation-related criteria, which is then applied to a particular trans-
lation. Yet, since the concepts of translation and translation quality are
variable and sometimes elusive, these theories are potentially susceptible
to contradictory views. They have, it is true, been conceived to cover a
wide range of types of, and opinions on, translation; still, they leave loop-
holes for those critics who do not agree with the proposed theory or who
choose to discuss the quality of a translated text from an altogether dif-
ferent angle. Thus, while these theories of translation quality assessment
explain the evaluation procedure and thereby achieve a certain degree of
intersubjective verifiability, their preconceived and often implicit assump-
tions about translation render them vulnerable to attacks from differently
minded critics. The question is, can a TQA theory be devised that accom-
modates even the most diverging opinions? Our answer is: yes, it can.
If any preconceived and often implicit assumptions about translation
and translation quality go against the ideal of objective evaluation, then
it makes sense to do without such assumptions. To begin with, we refuse
to define “translation” in any way that restricts the term’s scope of ref-
erence. What is meant by “translation” (or by the equivalent expression)
in a given culture determines what is accepted as a translation. Gideon
Toury writes:
Strange as it may sound to the uninitiated, there is nothing too per-

verse in claiming that a text’s position (and function), including
the position and function which go with a text being regarded as
a translation, are determined first and foremost by considerations
originating in the culture which hosts them. In fact, this is the most
normal practice of the ‘persons-in-the-culture’ themselves. Thus,
when a text is offered as a translation, it is quite readily accepted
bona fide as one, no further questions asked.
(Toury 1995:26)
A definition of translation as that which is presented as one includes any

target text based on a source text or on the assumption that a source
The Principle of Argumentation 171
text “must have existed” (Toury 1995:34). The texts covered by this
definition include concepts as different as adaptation, audiodescription,
and pseudo translation – to name only three.
With translation being defined on such shifting ground, we have no
foundation on which to base any translational assumptions. The only as-
sumption that can be made is that there are no invariants in translation
(cf. Vermeer 2007:174, discussed in Chapter 4). As a consequence, we
need to accept that the quality of a translated text cannot be adequately
assessed using an a priori assessment system. No evaluation criteria can
be taken for granted unless they are directly derived from the trans-
lated text to be evaluated and from the circumstances surrounding the
production of that text. In other words, what is needed in order to ade-
quately assess translation quality is a system that uses criteria drawn a
posteriori from the givens of the overall translation situation. The range
of these givens and their potential significance has been outlined in our
description of the translator’s daffodil (cf. Chapter 4). Any facts relative
to the translation and the translation situation constitute the pool of
quality factors that may be used to evaluate the quality of a particular
target text. The evaluator selects those factors he or she considers rel-
evant to the quality of the translation and arranges them in line with
an argumentation system that indicates why one translation solution is
better than another.
What we need is a theory that does not presuppose any translation-
related attitudes or perspectives, a theory that is, indeed, totally unre-
lated to the topic of translation (and might, therefore, be applied to other
subjects as well). Such a theory can be found in the realm of logic. Its
challenge would be to cater for all of the different opinions on transla-
tion and integrate all the different evaluations of a specific target text. It
would have to be able to clarify contentious translation decisions, such
as when the translator prefers one solution and the evaluator or critic or
reviser another. That the quality of translation revisions is not always
better than the quality of the original translation has been found out, for
example, by Peter J. Arthern, who states that “revisers may fail to correct
substantive mistakes, or may even introduce substantive mistakes and
they may also leave or introduce ‘formal’ errors of translation or layout”
(Arthern 1983:54). On the one hand, this quotation refers to obvious
mistakes (that is, mistakes that would be acknowledged as mistakes by
the translator or reviser who made them); on the other, it may also refer
to differences of opinion such that the translator or reviser who made the
“mistake” regards the corresponding translation solution as correct. In
the first case, a strictly logical system of translation quality assessment
should clearly state why a particular translation solution is to be regarded
as a mistake or an error; in the second case, the system’s line of argument
may either produce an unequivocal result as in the first case or show that
both the translator’s and the reviser’s solutions are equally acceptable.1
172 The Principle of Argumentation
Our logical system of translation quality assessment features argu-
ments and lines of argument whose number, structure, and content in-
dicate whether, and in what respect, one translation alternative is better
than another. In this context, the principle of relativity is crucial: the
quality of one translation solution cannot be measured in absolute terms
but has to be compared to the quality of another solution. Good transla-
tion quality can only be better translation quality, just as bad translation
quality can only be worse translation quality. There is no use dismissing
a translation solution as unacceptable unless a better alternative can be
produced. The ability to come up with at least one alternative solution
has been shown to be the translator’s key competence (cf. Pym 2003:489,
discussed in Chapter 4). This is, then, necessarily also a key competence
of the evaluator, since the evaluator of a translation should be as skilled
a translator as the translator him- or herself.
The comparison of at least two translation solutions forms the centre
of our logical approach – an approach in which logic does not aspire
to a theory of thought but serves as a humble instrument by which the
results of thinking can be verified (cf. Bayer 2007:99). Furthermore, the
approach opens up the opportunity of a dialogical exchange: the trans-
lator and the reviser may vie with each other for the most convincing
arguments in favour of the best translation. After all, the basic principle
of argumentation is, “The one who has the last word laughs best” (Dung
1995:322). It is likely that he or she will come closest to the best transla-
tion who is able to consider the most arguments – not just pros but also
cons. Klaus Bayer’s caveat that the involvement with possible counter-
arguments to one’s own point of view has a paralysing and unsettling
effect (cf. Bayer 2007:202) is probably less applicable to the reasoning
of translators and translation evaluators inasmuch as they have not yet
formed an opinion of what the best translation solution is, but use argu-
mentation to arrive at such an opinion.
The logical approach to translation quality assessment will be dis-
cussed in more detail below. First, we will trace the need for an ar-
gumentative TQA theory through a number of recommendations and
desiderata expressed in pertinent remarks and observations by various
translation scholars. This is followed by an account of the theoretical
implications of making translation decisions, before the last section fea-
tures the actual argumentation process together with a few examples.
The Need for an Argumentative Translation

Quality Assessment
Do translators need translation theory? And why do we need an ar-
gumentative approach to translation quality assessment? This section
undertakes to find appropriate answers to these questions.
The Need for Translation Theory
If translators were asked about translation theory – whether they know
any and whether they make use of it – the result to be expected would
probably depend on the translators’ training: those with a university de-
gree in translation would have some knowledge of translation theory
and would use that knowledge when translating; those who have not
had translation training at university level would tend to know little,
if anything, about translation theory and translate more or less on the
basis of code-switching. In trying to gauge the divide between transla-
tion theory and professional practice, Katan (2017) conducted a survey
in which translators had to answer questions about their training, their
work as translators, and their opinion on the importance of various
translation-related issues. Almost 900 respondents completed the whole
questionnaire, most of them educated at university in translation and/or
interpreting and/or languages. One of the results is that translators “fail to
see the relevance of translation theory” (Katan 2017:149). For instance, in
answering the question “What makes ‘translation’ a profession?” (Katan
2017:123), only three percent mentioned that translation “has its own
theory” (Katan 2017:124). Practitioners, it appears, set much less store by
theoretical aspects of translation than translation scholars.
Why do we need, then, translation theory, if translators get on with-
out it? Or do they? Perhaps, they use translation theory unknowingly –
taking for granted that translation means rendering a text from one
language into another and that, therefore, a good translation should be
as literal as possible and as free as necessary. 2 This principle suggests
itself by dint of the common-sense meaning of “translation”. It implies
that, since what translation means is clear to everybody, everybody can
translate, provided he or she understands the source language and is able
to write in the target language. In a sense, the generally acknowledged
(or assumed) dictionary meaning of “translation” constitutes a theory of
translation in its own right and is possibly the most widely used transla-
tion theory. Two points should be made, here: first, the above principle
is subject to variation, as an emphasis on “possible” and “free” will
give the translator some leeway to deviate from the original; second, the
principle lacks any indication as to when a translation should be literal
and when free – and, if free, in which way. By contrast, when Lauri
Honko, writing about translating oral epics, finds that “the translation
should be ‘as verbatim as possible and as readable as necessary’” (Honko
2000:33), he makes a conscious decision to “exclude […] the strategy
of ‘functional equivalence’” (ibid.). He wants to get “as close as possi-
ble to the original text” (ibid.), because his “target group […] consists
of informed scholars” (ibid.). Translation theory may be ignored, but
it cannot be avoided. It is inherent in the translation decisions and the
translator would do well to put it to good use.
Andrew Chesterman agrees, insisting that “there can be no practice
without theory” (Chesterman 1994:93). He has in mind the inevitability
in translation of having to select one solution from a number of alterna-
tive solutions – which echoes Pym’s minimalist definition of translation
competence. Pym also regards translation theory as unavoidable:
The minimalist approach basically sees translating as a process of

producing and selecting between hypotheses, and this is in itself a
mode of constant theorization. If thought through, the model is ac-
tually claiming that translators are theorizing whenever they trans-
late; theorization is an important part of translation practice. The
model also implies that whole translation approaches may be related
to translating in two ways: they may help translators produce more
alternatives than they would otherwise have thought of (pointing
out the existence of a problem is often the most important task of
theorization), and/or they may help them eliminate possible alterna-
tives. Theories would thus be productive and/or reductive, and both
kinds are obviously necessary.
(Pym 2003:492)
A translator has to make translation decisions, just as any writer has to

decide how to express what he or she wants to say. It may be argued that
some of these decisions, if not all, can be made unconsciously or sub-
consciously, in other words, the translator makes little effort to find the
best solution. However, even an intuitive decision needs some, perhaps
subliminal, theoretical foundation. This is obvious, when we look at
what might be a typical translation approach: a translator generates in
his or her mind possible target text units until he or she hits on a solution
that fits his or her notion of what the target text unit should be like; the
translator then discards the first, unneeded alternatives and chooses the
last one. The more potential solutions the translator generates, the more
conscious will probably be the selection process. The translator’s notion
of what a target text unit should be like constitutes the translation the-
ory that no translator can do without.
Thus, if translation theory is inherent in translating, it makes sense to
use translation theory as a vehicle to further the translation process and
to support its result. Furtherance of the translation process is achieved,
if the translator makes his or her translation decisions consciously, that
is, if he or she knows why a particular solution is viable and why it is
preferable to another viable solution. The translation result is supported
in that the translator who makes use of a translation theory can justify
his or her solutions and defend them against criticism. In the next sec-
tion, we will show that, in principle, this also applies to evaluators of
translations.
The Need for Argumentation in Translation
Whenever opinions on translation solutions differ, there has to be a the-
oretical basis on which this difference can manifest itself. The following
statement by Juliane House is very much to the point: “If we can make
explicit the grounds of our judgements, on the basis of an argued set
of procedures […], then, in the case of disagreement, we can talk and
discuss: if we do not, we can merely disagree” (House 1997:166–167).
The discussion of a translation problem can only be fruitful, if the par-
ticipants (for example, an evaluator and a translator) find some common
ground for their arguments. Yet, if the participants differ already on
the most fundamental assumptions about how a particular text should
have been translated, then they can only agree to disagree. The common
ground might consist in “an argued set of procedures”, such as the trans-
lation quality assessment tool proposed by Juliane House. Any disagree-
ment, then, arises not about the tool itself, but about issues relating to
the application and/or interpretation of the tool. The challenge is to find
a tool that (almost) everybody can accept. Such a tool could be furnished
by argumentation.
The need for argumentation in translation quality assessment has been
recognised – directly or indirectly – by a number of translation scholars.
Looking at the possibilities of potential and actual translations, Erich
Steiner, for one, points out that the overriding question is, “where […]
are the ‘good’ translations, and, equally importantly, why are they the
good ones?” (Steiner 2004:97). Radegundis Stolze writes that even
though the translator may translate intuitively, he or she should after-
wards be able to give reasons for individual translation decisions (cf.
Stolze 1992:215 and 2001:240). Likewise, the EMT expert group lists as
one competence for professional translators the ability to “[know] how
to justify one’s translation choices and decisions” (EMT expert group
2009, no pagination). Brigitte Horn-Helf is more specific when she states
that the translator must be able to justify his or her translation decisions
to non-experts in translation (such as judges and lawyers), using plau-
sible arguments that are theoretically sound (cf. Horn-Helf 1999:299).
The ability to give reasons why a translation solution fails to be satisfac-
tory is even more important for evaluators of translations, as Alexander
Künzli remarks, because the evaluator usually has to justify his or her
solutions to an expert, that is, a translator (cf. Künzli 2014:20). Brian
Mossop’s “Learn to justify changes” (Mossop 1992:84) takes the same
point of view: the evaluator must be able to give reasons for any im-
provement he or she wishes to make to a translation solution. Similarly,
Katharina Reiß demands that translation criticism – whether positive
or negative – should always be supported by a detailed argument and
pertinent evidence (cf. Reiß 1971:12).
However, justifying one’s changes to individual translation solutions
is not enough, as Reiß points out: the evaluator or critic should also
come up with a solution that is evidently better than the translator’s
(cf. Reiß 1971:12). Chesterman agrees. Taking as a point of departure
Karl Popper’s view of the scientific method (with an initial problem that
is provisionally solved by a tentative theory through error elimination,
the result of which presents another problem that needs to be solved), he
explains:
Error Elimination includes the necessity of replacing an inadequate

item with one that you think is better. Criticism thus includes the
suggestion of improved versions, which themselves are then subject
to the same critical process. And alternative versions themselves
must be justified, defended, corroborated.
(Chesterman 1994:92)
Whether the evaluator’s solution is better than the translator’s depends

on the arguments supporting them. In this way, the evaluator’s solu-
tion may also turn out to be less convincing than that of the translator.
Therefore, justifying one’s criticism and providing what one would re-
gard as a better solution should always go hand in hand. Justifying one’s
criticism should never go without a suggestion for improvement just as
a suggestion for improvement should never go without justifying one’s
criticism. Providing what one thinks is a better solution does not neces-
sarily imply the reason why it is a better solution.
Beyond the demand to justify translation changes and to provide an
alternative that can be shown to be better, translation scholars have in-
dicated some additional aspects that we consider to be crucial to an
argumentative evaluation framework. There is, for example, Jiří Levý’s
insightful remark about translation being a decision-making process:
From the point of view of the working situation of the translator at

any moment of his work (that is from the pragmatic point of view),
translating is a DECISION PROCESS: a series of a certain number
of consecutive situations – moves, as in a game – situations imposing
on the translator the necessity of choosing among a certain (and very
often exactly definable) number of alternatives.
(Levý 2000/1967:148)
To understand why the translator translated as he or she did, the eval-

uator would need to trace the translator’s decision-making process to
the very first translation decision.3 Arguments cannot become effective
in a vacuum but need a framework in which one argument can relate to
another. In reconstructing the decision-making process of the translator,
the evaluator creates such an argumentative framework.
Within the framework, the individual arguments have to be filled with
content based on criteria that are relevant to the evaluation of a transla-
tion. These criteria or factors (discussed in Chapter 4) are not easily de-
fined, as M. A. K. Halliday notes: “It is notoriously difficult to say why, or
even whether, something is a good translation, since this must depend on
a complex variety of different factors that are constantly shifting in their
relationship one to [sic] another” (Halliday 2001:14). The complexity of
this variety of different factors is such that they can neither be limited
in number nor determined in advance: in translation quality assessment,
according to Sylvia Reinart, it will be necessary to move on from closed
systems of analysis to more open lists of criteria (cf. Reinart 2014:85);
and “these criteria must be established a posteriori in each analysis, for
they are closely related to the specific characteristics which the text dis-
plays” (Rodríguez Rodríguez 2007:6). What is more, “each criterion will
be not a rigid, immutable factor but rather a v ariable – something that
changes as a function of something else” (Jumpelt 1963:269). A logical
consequence is, then, Vermeer’s d ictum – mentioned at the beginning of
this chapter and discussed in Chapter 4 – that there are no invariants
in translation (cf. Vermeer 2007:174). Relativity will, thus, be the un-
derlying principle of our argumentative approach to translation quality
assessment.
Summary – The Need for an Argumentative Translation

Quality Assessment
Several translation scholars have put forward concepts and ideas
that underpin the need for translation theory and a translation
quality assessment based on arguments that reflect the translator’s
decision-making process. Table 5.1 provides an overview of these con-
cepts and ideas and the corresponding references. They point to a
translation quality assessment in which argumentation plays a key
role: argumentation determines the range of factors relevant to the
assessment of a particular target text, and it underlies the criticism of
translation decisions as well as their justification. Thus, in the rest of
this chapter, we will develop an argument-based theory of translation
quality assessment.
Translation Decisions
A translated text is based on a large number of translation decisions at
various macro- and microtextual levels. The significance of such deci-
sions has been recognised by Rudolf Walter Jumpelt, who is one of the
first scholars to make reference to the decision-making process in trans-
lation (cf. Schubert 2011:752). He points out that translation is a process
of selection between complex variables and emphasises that a theory
Table 5.1 F
undamental concepts and ideas of translation quality assessment
Concepts and ideas References
Translation practice always involves Chesterman (1994:93), Pym

translation theory. (2003:492)
A translator must be able to justify his or Steiner (2004:97), Stolze
her translation decisions. (1992:215, 2001:240),
EMT expert group
(2009:5), Horn-Helf
(1999:299)
An evaluator should try to understand why Reiß (1971:12), Broeck
the translator translated as he or she did. (1985:56)
An evaluator must be able to justify any Steiner (2004:97), House
changes he or she would want to make (1997:166–167), Künzli
to a translation. (2014:20), Mossop
(1992:84), Reiß (1971:12)
Saying why a particular translation unit Reiß (1971:12), Chesterman
should be changed is not enough: the (1994:92)
evaluator must also provide a better
solution.
Just providing a better solution is not [No reference: a conclusion
enough either, because usually the better drawn from the two
solution does not indicate why it is previous concepts]
better.
Translating is essentially a decision- Pym (2003:489), Levý
making process. (2000/1967:148)
Translation decisions are made and Halliday (2001:14), Jumpelt
criticised on the basis of many different (1963:269)
factors, which are variable, depending
on their relationship to one another.
These factors cannot be defined in Rodríguez Rodríguez
advance but must be established (2007:6)
for each translation and translation
quality assessment on the basis of the
source text and the overall translation
situation.
As a consequence, the number of factors Reinart (2014:85)
relevant to the evaluation of a particular
translation varies.
of translation should describe the aspects that govern the translator’s

choice between possible translation solutions rather than the solutions
themselves (cf. Jumpelt 1961:186). The overall decision-making process
has been described from a more general perspective by Klaus Schubert:
Die grundlegende Vorstellung lautet, dass eine Entscheidung darin

besteht, aus einer Menge vorgegebener Lösungen eine (oder, je nach
Art der Aufgabe, mehrere) auszuwählen. Ein Entscheidungsprozess
wird somit als eine Kette von Auswahlentscheidungen betrachtet.
Ich nenne die Menge aller Lösungen, unter denen bei einer solchen
Auswahlentscheidung gewählt werden kann, den Entscheidungs-
raum. Jede potenzielle Lösung besitzt Entscheidungsmerkmale.
Dies sind inhärente oder zugewiesene Eigenschaften oder Attribute
der Lösungen, die es erlauben, sie zu unterscheiden. Die Entschei-
dung erfolgt mit Hilfe von Entscheidungsregeln, die mit Entschei-
dungskriterien versehen sind. Der eigentliche Entscheidungsprozess
erfolgt durch einen Abgleich zwischen den Kriterien der Regeln und
den Merkmalen der Lösungen. Die Gesamtheit der Regeln ist der
Entscheidungsmechanismus.
(Schubert 2007:245)
The basic idea is that a decision consists in selecting one or more

solutions from a given set of solutions (depending on the task to be
accomplished). The decision-making process is, thus, regarded as a
chain of selection decisions. I call the set of all solutions from which
to choose when making such a selection decision the decision space.
Each potential solution has decision attributes. These are inherent
or assigned characteristics or features that make it possible to dis-
tinguish between individual solutions. The decision is made with
the help of decision rules, which are furnished with decision crite-
ria. The actual decision-making process compares the criteria of the
rules with the attributes of the solutions. All the rules taken together
are referred to as the decision mechanism.
[my translation, H.B.]
In this section, we will take a closer look at the above description of

the decision-making process and explain its significance in a translation
context.
Decision-Making and Translation

The overall process is referred to as a decision-making process (or a de-
cision process). Within this decision-making process, a decision could be
either a selection process or the result of that process. Here, the term is
used with the former meaning. The outcome of the selection process is a
solution. If the task is to select one solution, then the number of decisions
in a decision-making process equals the number of solutions generated
by that process. Likewise, the number of decisions equals the number of
decision spaces. The decision attributes of the potential solutions in the
decision spaces correspond to the decision criteria of the decision rules.
Attributes and criteria could be regarded as two sides of the same coin
since the former are invariably linked up with the latter: the criteria of the
decision rules are used as a template on the basis of which potential solu-
tions can be generated whose attributes match the criteria of the rules to
a greater or lesser degree. The application of a decision rule to a selection
decision in the decision-making process has an effect on subsequent de-
cisions in that it restricts their respective decision spaces. In the sense
that earlier decisions influence later decisions, the decision-making pro-
cess is a hierarchical process. While the overall direction in the decision-
making process is from the top downwards, there are some processes in
which subsequent decisions may also have repercussions on preceding
decisions – suggesting the possibility of a limited bottom-up corrective to
the prevailing top-down process. This corrective comes into play when-
ever the impact of a decision is such that it modifies the decision rule or
the decision criteria established by previous decisions. A decision-making
process is, therefore, an inherently dynamic process.
As Jiří Levý points out, translating is a decision process if considered
from the translator’s point of view (cf. Levý 2000/1967:148); and, sim-
ilarly, Fritz Paepke conceives of translating as an action-oriented deci-
sion process (cf. Paepke 1971:115). It requires decisions at all levels of
language and culture.4 Once the translation process has started, any
decisions are made in consequence of previous decisions and, at the same
time, contribute to the basis on which subsequent decisions are made.
Jiří Levý aptly likens the translation process to a game – “a game in
which every succeeding move is influenced by the knowledge of pre-
vious decisions and by the situation which resulted from them” (Levý
2000/1967:149). That means, consistency is an important factor in the
decision-making process, a factor that concerns, in particular, the trans-
lation strategy. A translation strategy outlines the overall orientation of
a translation, generally distinguishing between source and target ori-
entation.5 Once the translator has decided to translate a source text in
line with a particular translation strategy, he or she should stick to that
strategy throughout the translation.6
While theory would suggest that the number of translation solutions in
the target text equals the number of decisions made in the decision-making
process, this is not the case in practice because, in the course of translating
a given source text, many an intermediate solution produced as the result
of a decision does not appear in the target text. Whether intermediate or
final, every solution implies at least one non-solution – “those words and
utterances that are not selected” (Munday 2012:13), but which “[lurk] be-
hind the actual selection and, if we look, [tell] us much about the values
underlying those words” (ibid.). Jeremy Munday explains that
an individual choice of word or expression does not exist in isolation

but in relation to the other possible choices that the writer or speaker
has discarded or otherwise did not use. […] These become percepti-
ble, for example, in the comparison of one translator’s choice against
another’s or in the study of the forms which are revised at different
stages in the translation process.
(Munday 2012:13)
In the assessment of an individual translation, any of the translator’s
selections is measured against the alternative suggested by the evaluator
and, perhaps, against any unselected solutions (provided that the trans-
lator explains his or her decision in favour of the chosen alternative).
So what does it mean for a translator to make a translation decision?
What should a translator bear in mind when selecting a translation solu-
tion? The most general answer is: as many aspects as possible that are rel-
evant to translation. Thomas Kempa suggests a more specific response.
Looking at textual meaning from a translation-didactics perspective, he
explains sense in terms of three sense-producing elements (semantics,
pragmatic context, and skopos) and rightly claims that an awareness of
all three (rather than a focus on just one) as part of a holistic approach
to translation can help translators (and students of translation) make
more informed translation decisions (cf. Kempa 2013:366). To this end,
a translator – in trying to negotiate the meanings of source and target –
has to take into account both explicit and implicit information. If the
amount of information, both explicit and implicit, is to be the same in
the source text and in the target text, there are four options: explicit
information in the source text may be rendered (1) explicit or (2) implicit
in the target text; and, likewise, implicit information in the source text
may be rendered (3) explicit or (4) implicit in the target text. Rendering
implicit source text information explicit in the target text is referred to
as explicitation; rendering explicit source text information implicit in
the target text is referred to as implicitation. Any change in the amount
of information from source to target would be an addition or omission.7
A translator has to ponder the need for explicitation or implicitation
against the background of the translator’s daffodil (cf. Chapter 4).
Defining the Translation Strategy

The decisions made by the translator in the process of translating should
serve one goal, namely, to achieve the best possible translation qual-
ity under the conditions imposed by the translation situation. These
conditions – the factors relevant to translation quality – have been de-
scribed in our account of the translator’s daffodil in Chapter 4. They
are important throughout the actual translation process, but also play
a crucial role when it comes to defining the translation strategy. The
translator defines the translation strategy before starting the translation
proper. In this way, he or she knows how to translate and, if necessary,
can justify his or her translation decisions on the basis of the translation
strategy.8 Since several decisions are needed to define a translation strat-
egy, the decision-making process in translation can be divided into two
sub-processes: the first sub-process helps the translator to establish the
translation strategy, and the second sub-process assists him or her in the
actual translation. We will analyse the first sub-process in more detail
to demonstrate how the translator can pave the way for a successful
translation by specifying the strategic basis on which he or she intends
to translate a given source text.
In Chapter 3, we discussed Juliane House’s concept of overt and co-
vert translation and suggested a few modifications. This modified ver-
sion of the concept serves as a point of departure for the development of
a translation strategy and for the subsequent translation process with its
individual translation decisions. Once the translator has decided to take
on a particular translation job, he or she considers the overall translation
situation to look for clues that help to define the translation strategy.
What should a translation strategy look like, if quality is the translator’s
foremost concern? The question is what constitutes quality in transla-
tion and who has an interest in the quality of the translated text. Al-
though we conceive of quality strictly in terms of relativity – with the
qualitative impact of one translation solution being always relative to the
qualitative impact of another – there has to be a concrete point of refer-
ence from which to start the decision-making process. Most promising
for a fruitful outcome of the decision-making process in translation is
the assumption that the readers of the target text (and any other stake-
holders that have an interest in the translation) are key to the assessment
of the target text’s quality.9
In the concept of overt and covert translation, the potential readers of
the target text play a central role: in an overt translation, the potential
readers of the target text are aware of the fact that they are reading a
translation and they are interested in the target text as being a transla-
tion; in a covert translation, the potential readers are not interested in
the target text as being a translation (and may not even be aware of the
fact that they are reading a translation) but read it as though they were
reading an original text. Once the translator has decided to translate
a given source text, he or she should try to gauge the expectations of
the target text readers in terms of the overt–covert distinction. The first
decision, then, determines whether the overall framework of the trans-
lation strategy should be overt or covert.10 To this end, the translator
should ask himself or herself a two-part question: (1) will the reader of
the translated text know that he or she is reading a translation, and if
so, (2) will he or she be interested in the fact that the text is based on an
original?
If the first part of the question is answered in the negative, the over-
all framework of the translation strategy is unequivocally covert. The
answer largely depends on the type and function of the source text.
Most business texts such as letters, memos, brochures, websites, or user
manuals – texts in which the referential or conative function is to the
fore – should be translated covertly so that the reader of the target text
does not notice that he or she is reading a translation. By contrast, any
legal deeds or documents to be presented in court are subject to overt
translation, as the target text reader uses the translation as a means to
get at the original. The same holds true for literary texts, especially,
those which stand out for a peculiar style: the more prominent the po-
etic function of the source text, the more overt should usually be the
translation. Another aspect that comes into play, here, is the author of
the source text. If the author is well known, for example, for his or her
literary achievements, the translation of a text written by him or her
would be expected to reflect the characteristics of the original. On the
other hand, the translation of a run-of-the-mill bestseller novel is less
in need of exhibiting typical source text features because that novel’s
primary function is to entertain the reader both in the original and in
the translated version – thus, permitting the rather covert rendering of a
genre that is often translated overtly.
Similarly in the middle of the overt–covert scale are journalistic texts,
for example, reports from popular-science magazines. If the target text
is marked as a translation, the reader will know that it is a translation;
yet, he or she is probably not interested in the target text as being a
translation. Whether or not the reader of the target text is interested in
the target text as being a translation depends on the source text’s status.
Again, the author may become relevant, for instance, if he or she can be
said to be an authority on the subject dealt with in the text. In that case,
the importance of the author’s knowledge is such that the target text
reader would expect to be offered a glimpse of the original. Also content
may be taken into account: a report about a topic concerning a particu-
lar culture and written from an insider’s perspective commands a signif-
icance that should be acknowledged by an overt translation framework;
by contrast, a report dealing with a universal subject might be translated
covertly, as the reader of the translation is probably not interested in the
original (nor even in the fact that such an original exists at all).
Throughout this discussion of the translator’s first decision – the de-
cision that determines the overall framework of his or her translation
strategy – we should keep in mind that the choice of an overt or covert
strategy has no universal consequences for the way in which a source
text is rendered into a target text. Whichever translation strategy has
been chosen, the appropriate translation procedure depends on what the
respective expectations of the readers actually imply in the context of a
particular target culture at a particular time in history. For the trans-
lator (as for the evaluator of a translation), this will often be his or her
own culture in a contemporary setting, so that jumping to conclusions as
regards the implications of the overt or covert framework would seem to
be only too natural. However, from a theoretical perspective, it is neces-
sary to reckon with different translation preferences in different cultures
at different periods in history. The phenomenon of the “belles infidèles”
in seventeenth- and eighteenth-century France is a case in point: here,
the translation of a novel into French was expected to improve on the
source text in such a way that the target text would be “pleasant to read”
(Salama-Carr 2009:406). While many readers at the time were obviously
keen on the content of the originals rather than on their artistic merits,
such an attitude would still be classified as overt, because the readers
were aware that they were reading translated texts and they wanted to
be charmed by the texts’ foreignness. With this cultural and historical
proviso in mind, we can now proceed to the decisions the translator
needs to make, after he or she has decided on the overall strategic frame-
work. These decisions are grounded on a German target culture context
in the early twenty-first century.
Once the translator has come to terms with the overt–covert transla-
tion strategy, he or she can then decide on the general degree of source
or target orientation. If the translator finds that the readers of the target
text are likely to be genuinely interested in the original, he or she will
try to stick as closely as possible to the source text. If the translator has
come to the conclusion that the readers of the target text are not at all
interested in the original, he or she will try to ensure that the translation
reads like an original in the target culture. If the translator’s delibera-
tions have led him or her to conceive of a strategy that combines overt
and covert elements – such as might be the case whenever the potential
readers of the target text know that they are reading a translation but do
not seem to be particularly interested in the fact – the result would prob-
ably be a source-oriented approach with occasional allowances made to
accommodate any specific target culture preferences.
After this second decision, in which the translator gets his or her bear-
ings in terms of source or target orientation, the translation strategy
can still be defined in more detail, if the text type or genre and the
translation brief are taken into account. To give an idea of the decisions
involved, we will look at two examples: the translation of a classical
novel and the translation of a marriage certificate to be presented in
court. The translation of the classical novel is stylistically demanding.
The translator needs to decide at what level and to what extent he or
she should try to capture the stylistic features of the source text in the
target text. Here, the decision space includes a number of potential solu-
tions between the following extremes: sticking as closely to the style
and wording of the source text as is compatible with the grammatical
rules of the target language; and gauging the aesthetic effect of the sty-
listic features of the source text and replacing them in the target text by
stylistic features that produce an equivalent or similar aesthetic effect
for the target audience. A decision in favour of one of these solutions
is likely to be tentative, because only the actual translation process will
show for each instance of a stylistically challenging passage whether the
chosen solution is feasible or needs to be modified. As a result, the imple-
mentation of a solution – that is, the adherence to a specified degree of
source-orientation throughout the translation process – can have three
effects: (1) it may be exactly what has been intended and thereby confirm
the decision in favour of the solution; (2) it may be different from what
has been intended and thereby modify the solution for the nonce; and (3)
it may be different from what has been intended and thereby modify not
only the solution but also the pertinent decision rule. In the third case,
the original decision rule (following from the first three decisions of the
decision-making process) would have called for a decision in favour of
an unsatisfactory solution, the modification of which would then cause
the rule to be changed in a dynamic decision-making process.
The decisions to be taken when translating the marriage certificate are
more straightforward. If the translation brief specifies, or if the transla-
tion situation implies, for example, that the translated marriage certificate
is to be presented in court in Hamburg, the translator should be guided
in his or her translation decisions by the instructions provided in a leaflet
issued by the Department for Internal Affairs of the Free and Hanseatic
City of Hamburg on how to prepare certified translations. The trans-
lator’s third decision would, then, consist in the adoption (rather than
the rejection) of these instructions, which can be regarded as decision
rules. These rules would have to be obeyed during the actual translation
process. They help the translator to implement the right overt translation
strategy. By way of example, here are three of the instructions:
• The layout of the translation should match that of the source text
(cf. Freie und Hansestadt Hamburg 2010:1).
• Crossed-out passages that can still be read should be translated and
marked as being crossed out in the original document (cf. Freie und
Hansestadt Hamburg 2010:2).
• If the translator notices content errors in the document to be trans-
lated, he or she should point them out in a suitable way to avoid any
suspicion of a translation error (cf. Freie und Hansestadt Hamburg
2010:3).
These instructions or decision rules are intended to ensure an optimum

understanding of the source text by reading the target text. In this way,
they help to improve the quality of the translation.
Once the translator has defined the translation strategy for a particular
translation job, he or she can start to translate. The decision-making pro-
cess that has served to set up the translation strategy is continued as the
source text is being translated. Every decision made during the translation
process should follow from the decisions made in defining the translation
strategy. Moreover, they should build on one another in the sense that later
decisions are influenced by earlier decisions. Inasmuch as written texts
unfold their meaning sequentially, the decision-making process should
naturally follow the textual sequence. In other words, the translator is well
advised to start translating at the beginning of the source text rather than,
say, somewhere in the middle. This is why large projects involving several
translators for the translation of one source text require an additional,
co-ordinating decision-making process. It is clear that the discussion of the
decision-making process following the definition of the translation strat-
egy cannot be more detailed without reference to an actual translation.
The concrete decisions made as part of the actual translation process will,
therefore, be discussed together with the evaluation of the target text.
Summary of the Decision-Making Process in Translation

The elements of the decision-making process in translation are listed and
explained in Table 5.2. Table 5.3 depicts the overall process.
Table 5.2 Decision process terminology
Decision process Explanation with reference to a translation

terminology context
Decision-making process Involves all decisions made in translation: the

decision to accept the task of translating a
given source text, the decisions needed to
define the translation strategy, the decisions
that govern the actual translation process and
lead to the final version of the target text
Decision The activity of selecting one solution from at
least two potential solutions in the overall
translation process
Solution At the level of the microtext: a target text unit
selected in the translation process; at the level
of the macrotext: the result of a strategic
translation decision
Decision space The set of potential solutions to a particular
translation decision
Decision attributes Characteristics of a potential translation
solution that determine whether it is selected
as the preferred translation solution
Decision rules Rules that are defined by translation-related
criteria including translation decisions and
that govern the decision-making process in
translation
Decision criteria Decision rule characteristics drawn from the
translation environment outlined in the
translator’s daffodil and from any previous
translation decisions
Decision mechanism The decision rules operating in a translation
context and which include, and are near-
identical to, the translation strategy
Table 5.3 T he decision-making process in translation
Decision Issues to be Grounds on which the decision is made

decided
Start decision To translate or not The translator’s situation, including

to translate a competence, time, willingness,
given source text pecuniary need, etc.
First decision To adopt an Readers’ expectations regarding the
overt or covert target text function in a given culture
framework for at a given period in history: target
the translation text like a translation vs. target text
like an original
Second To adopt a source- Readers’ expectations regarding target
decision or target- text orientation in a given culture at
oriented strategy a given period in history: target text
close to source text vs. target text
adapted to target culture
Third decision To further specify Previous decisions, which lead to
what the conclusions concerning the handling
adopted strategy in translation of textual subcategories
implies such as style, content or layout
Further Concrete Previous decisions, with the translation
decisions translation strategy being defined by the first,
problems second, and third decisions, and the
translator’s daffodil
We should emphasise once again that, far from being independent, a

decision relates to preceding decisions in that it builds on them, and
to later decisions in that it determines their individual scope. The first,
second, and third decisions specify the translation strategy, making up
the core of the overall decision mechanism. Some decision rules may still
emerge during the actual translation process – decision rules that may
even cause the translator to readjust the translation strategy; yet, gen-
erally, the rules of the translation strategy established in decisions one,
two, and three should provide a clear and unchanging guideline for the
translator to solve any problems in the actual translation process. Thus,
the translation strategy serves as a basis for the justification of individ-
ual translation solutions. It also plays a crucial role in the evaluation of
translation.
The Argumentation Process

An ideal translation would consist of translation solutions that are better
than any alternatives found in the respective decision spaces. Thus, it
is necessary to show that whatever can be said in favour of a preferred
translation solution for a given translation unit is more convincing than
what can be said in favour of any other solutions for the same unit. This
implies that whatever can be said against a preferred translation solution
for a given translation unit should carry less weight than what can be
said against any other solutions for the same unit. The translator and the
evaluator have to consider the pros and cons for each potential solution
of a specific translation problem. To come to a conclusion in this reason-
ing process, they need arguments that strengthen the preferred solution
and arguments that weaken the case for each of the possible alternatives.
For the quality of a translation solution is only as good as the arguments
supporting it. What is needed, then, is an argumentation framework
that reveals the relative quality of one translation solution in comparison
with alternative translation solutions. In looking for a system that can be
used to set up such an argumentation framework, we have come across
the theory of dialectical structures by Gregor Betz (2010).11 In this sec-
tion, we will first give a brief outline of Betz’s theory, before presenting
a simplified version for our own purposes.
A Theory of Dialectical Structures

In a discussion, according to Betz, any insight into how the exchange of
arguments works seems to be based on intuition rather than clear rea-
soning, as there is no regulatory framework that could be consulted (cf.
Betz 2010:4). This is why Betz undertakes to contribute with his theory
of dialectical structures to a grammar of reasonable argumentation (cf.
Betz 2010:4). The theory should provide a basis for the analysis and
explanation of all kinds of argumentative exchanges. In this context,
it is important to see that complex and controversial argumentation is
both structure and process at the same time: argumentation is viewed
as a structure inasmuch as it consists of many elements (theses, justifica-
tions, objections, etc.) that are interrelated in a variety of different ways
(characterised by expressions such as proving, contradicting, defending,
or countering); and argumentation is viewed as a process, because it
emerges and develops as the result of a discussion in which the propo-
nents add one idea or opinion to another (cf. Betz 2010:5). However,
while other approaches may tend to be more process-oriented, Betzʼs the-
ory of dialectical structures (as the name indicates) is at bottom a struc-
tural argumentation theory (cf. Betz 2010:8). This has two advantages:
on the one hand, it avoids the problems encountered by the procedural
approaches; on the other, it ties in with the requirements of formal logic
– formal logic being one of three points of departure for Betz’s theory.
The other two are informal logic and theories of defeasible reasoning.12
Combining the advantages of these three logical methods while avoiding
their respective disadvantages: that is what Betz tries to achieve with his
theory of dialectical structures. Like informal logic theories, Betz’s the-
ory should be able to capture and depict the macrostructure of complex
argumentation; like formal logic, it should make possible a precise and
detailed analysis and evaluation of individual arguments by means of
deductive reconstruction; like the theories of defeasible reasoning with
their argumentation systems, it should be useful for the normative eval-
uation of complex real argumentation.13
In developing the theory of dialectical structures, Betz deals with the
following aspects: the structure of complex argumentation, dialectically
coherent positions, justifications, arguments, and dialectical patterns.
The structure of complex argumentation is characterised by the way in
which the “atoms” of this structure, that is, theses and deductive argu-
ments (cf. Betz 2010:52), relate to one another. What determines these
dialectical relations is the internal structure of arguments (consisting of
premisses, intermediate conclusions, and conclusions), on the one hand,
and the logical-semantic relations between the sentences of different ar-
guments, on the other (cf. Betz 2010:53). Betz shows that the many dia-
lectical relations possible between the sentences of different arguments
can be reduced to two relations, namely, the standard support and the
standard attack: while a supporting argument strengthens another ar-
gument to which it relates, an attacking argument weakens it. The two
relations are defined as follows (cf. Betz 2010:56):
• An argument A supports an argument B, if and only if the conclu-

sion of A is equivalent to one of the premisses of B.
• An argument A attacks an argument B, if and only if the conclusion
of A is the negation of one of the premisses of B.
That one argument supports another argument is shown by a continu-

ous arrow (→); that one argument attacks another argument is shown by
a dashed arrow (⇢). It should be noted, however, that “support” and “at-
tack” are not success verbs, that is, they do not necessarily achieve what
they intend to achieve. A supported argument may still collapse, just as
an attacked argument may thrive. The point to be made, here, is that a
dialectical structure is characterised by relations of support and attack,
and that these relations affect the evaluation of the overall structure and
its arguments (and theses). Indeed, Betz defines a dialectical structure
as a set of arguments and theses plus the attack and support relations
prevailing between them (cf. Betz 2010:61).
The dialectical structure of complex argumentation usually involves
several proponents with different positions. These positions have to be
dialectically coherent – a dialectically coherent position being one that
can reasonably be taken in a debate (cf. Betz 2010:70). Betz distinguishes
between two kinds of positions: those with an assignment of sentences
and those with an assignment of arguments. If the proponents’ posi-
tions are defined on the basis of sentence assignment, the following three
conditions must be met for any of the positions to be dialectically coher-
ent (cf. Betz 2010:70):
• condition of equivalence (that is, equivalent sentences are given the

same truth value);
• condition of deductivity (if all premisses of an argument are true,
then also the conclusion of that argument is true)14;
• consistency (contradictory sentences are given complementary truth
values).
Three different conditions apply, if the proponents’ positions are defined

on the basis of argument assignment (cf. Betz 2010:77)15:
• poise (the accepted premisses and conclusions of the arguments

taken by a proponent do not contradict one another);
• closure (if an argument contains as premisses only accepted sen-
tences from arguments taken by a proponent, then the proponent
also takes that argument);
• completeness (if a complex argumentation contains both a sentence
and its negation, then a proponent accepts one of these two sen-
tences as a premiss or conclusion of one of the arguments taken by
him or her).
Thus, there are two different ways by which a proponent’s position in a

dialectical structure can be defined: the position may be defined either
by sentences that are assigned to it or by arguments that are assigned
to it. Betz notes that the coherence conditions at sentence level and at
argument level are characterised by what he calls quasi-equivalence
(“Quasi-Äquivalenz” (Betz 2010:92)) and he points out that, for the en-
suing discussion, he would mainly use sentence assignment as the pre-
ferred mode of defining a proponent’s position, because it is easier to
handle from a theoretical point of view and more precise than defining
a proponent’s position through the assignment of arguments (cf. Betz
2010:92–93). Notwithstanding its disadvantages, the argument-based
definition of a proponent’s position is not redundant; it has the benefit
of capturing the basic intuitions governing a debate (where this debate
operates largely at argument level) and, thus, provides a better overview
of the overall argumentation than an analysis at sentence level (cf. Betz
2010:93).
The sentence-assigned and argument-assigned positions feature differ-
ent kinds of entailment relations. While a sentence-assigned position is
characterised by dialectical-semantic entailment relations, an argument-
assigned position is characterised by dialectical-syntactic entailment re-
lations. These entailment relations are distinguished as follows (cf. Betz
2010:107):
• In a dialectical-semantic entailment, a sentence p follows from sen-
tences S, if and only if in all dialectically coherent positions in which
sentences S are true sentence p is also true – or if, vice versa, there is
no dialectically coherent position according to which sentences S are
true and sentence p is false.
• In a dialectical-syntactic entailment, a sentence p follows from sen-
tences S, if there is an argument or line of argument that has p as its
central conclusion and whose free premisses16 are included in S. In
other words, p follows dialectically-syntactically from S, if there is a
“derivation tree” ending in p. This “derivation tree” is a dialectical
substructure featuring only support arrows which support an argu-
ment with sentence p as its conclusion.
Both dialectical-semantic and dialectical-syntactic entailment relations

can be used to determine the degree to which a thesis is justified by the
premisses and conclusions of the arguments that relate to it. However,
since the concept of dialectical-semantic entailment appears to be more
decisive in argumentation theory than the concept of dialectical-syntactic
entailment,17 it is the former that constitutes the point of departure for a
further analysis of how the degree of justification of a thesis (or sentence)
can be established.
A sentence is justified by other sentences in that its truth is attributed
to the truths of the others (cf. Betz 2010:106). Or, more generally, a sen-
tence is justified by other sentences, if it follows from them (ibid.). The
question is, how these findings can help to explain the common impres-
sion that, often, the justification of one thesis is more convincing than
that of another. By way of answering this question, Gregor Betz draws
on the concept of partial implication. The degree of justification of a
thesis can be regarded as the degree of partial implication of a sentence
p in a dialectical structure consisting of sentences S. It is calculated by
dividing the number of dialectically coherent positions in which p and
sentences S are true by the number of dialectically coherent positions in
which sentences S are true (cf. Betz 2010:116). The crucial figure, here,
is the number of dialectically coherent positions – a number that varies
depending on whether the dialectically coherent positions are based on
sentence assignment or on argument assignment. If they are based on
sentence assignment, there will be more dialectically coherent positions
than if they are based on argument assignment. Furthermore, due to the
context-relatedness of dialectical structures, the degree of justification
of a sentence p changes, whenever the dialectical structure is modified
in certain ways. The changes include the following events in particular
(cf. Betz 2010:117):
• If we insert independent arguments (that is, arguments that do not

contain any premisses or negations of premisses which are already
part of the debate) with conclusion p into the dialectical structure,
then this will increase the degree of justification of p ceteris paribus
(that is, provided that the added arguments do not support or attack
any arguments other than the thesis expressed in sentence p).
• If we insert independent arguments with conclusion non-p into the
dialectical structure, then this will – ceteris paribus – decrease the
degree of justification of p.
• If we attack arguments that falsify p (because they have non-p as
their conclusion), then this will – ceteris paribus – increase the de-
gree of justification of p.
• A new argument that attacks another argument only changes –
ceteris paribus – the degree of justification of p, if the insertion of
the argument that is being attacked has already changed the degree
of justification.
These characteristics of the degree of justification in a dialectical struc-

ture largely correspond to our everyday notion of the concept: support-
ing arguments increase the degree of justification, attacking arguments
decrease it, and counterarguments increase it (cf. Betz 2010:117). In spite
of this apparent correspondence between theory and practice, Betz cau-
tions against too sanguine an expectation that any decision-making pro-
cesses could reasonably be based on the degree of justification, or that
the degree of justification could be used in any straightforward way to
calculate the expected benefits of a particular action (cf. Betz 2010:121).
Still, he does apply the method to show that one explanation of a hy-
pothesis is better than another (cf., for example, Betz 2010:208).
To what extent it makes sense to rely on degrees of justification is
discussed in connection with arguments or, to be more precise, with
acts of argumentation. When it comes to assessing such acts of argu-
mentation, the principle of relativity is crucial, because the results of the
assessment hinge upon the concrete dialectical structure at the time of
the assessment. Any new argument inserted in the debate changes the
dialectical structure and, thus, the basis on which it should be assessed
(cf. Betz 2010:128). On such uncertain ground, what discursive goals
should the proponents of the debate try to achieve? The first discursive
goal should be dialectical coherence (cf. Betz 2010:129). That is to say,
a proponent has to ensure that his or her position in the debate is not
inconsistent, in other words, that he or she does not contradict him-
or herself. Another discursive goal consists in justifying one’s own sen-
tences in the dialectical structure or the sentences of another proponent.
Gregor Betz refers to a burden of proof that needs to be discharged (cf.
Betz 2010:130–131). A burden of proof of the first order is one in which
a proponent finds out for him- or herself why a sentence p (that does not
follow dialectically-semantically from other sentences accepted by him
or her) is true; he or she discharges the burden of proof by giving reasons
to justify the assumption that sentence p is, indeed, true. A burden of
proof of the second order is one in which a proponent justifies a sentence
p that has been put forward by another proponent, yet does not follow
dialectically-semantically from the other sentences accepted by that pro-
ponent. The strategy of discharging first- or second-order burdens of
proof is regarded as a reasonable discursive goal. A third discursive goal
of sensible argumentation consists in trying to achieve a high degree of
justification for one’s own core position (cf. Betz 2010:132). The more
robust one’s core position is, the less likely will it be subject to revision
as a result of changes to the dialectical structure.
Given that the three discursive goals can be achieved using a defensive
or an offensive strategy, there are, in total, six goals that the propo-
nents can reasonably pursue in argumentation (cf. Betz 2010:135) as is
shown in Table 5.4. The question which of these discursive goals should
be pursued and whether a defensive or an offensive argumentative strategy
is preferable cannot be answered unequivocally (cf. Betz 2010:137–140).
What is interesting, though, are those strategies by which a proponent
can change the degree of justification of a given sentence. Betz suggests
that the effect of a supporting or attacking argument is related to that
argument’s distance from the central thesis (cf. Betz 2010:141–143). Hav-
ing tested various argumentation dynamics with a computer program, he
concludes that, in relation to a given thesis, an argument is most effective
if it supports or attacks the thesis directly rather than indirectly through
other arguments. However, this is not to say that attacking or supporting
arguments further away from the central thesis makes no sense; depend-
ing on how the debate develops, other sentences at different places in the
dialectical structure may also be of interest and deserve special attention.
Table 5.4 M
atrix of discursive goals
Discursive Defensive strategy Offensive strategy

goal
Dialectical My position is dialectically The positions of the

coherence coherent. other proponents are
dialectically incoherent.
Burdens of The theses that I wish to Certain theses follow
proof justify follow dialectically- dialectically-semantically
semantically from the from the sentences
sentences accepted by me accepted by the other
(burden of proof of the first proponents (burden of
order). proof of the second order).
Robustness The partial position that The partial positions that
constitutes the core of my represent the core claims
point of view features a high of the other proponents
degree of justification. feature a low degree of
justification.
Given the six discursive goals and the various strategies used to
achieve them, the question arises whether there are certain dialectical
patterns that should be embraced or avoided. One question, in partic-
ular, is whether a circular argument (also known as petitio principii or
begging the question)18 automatically results in a fallacy. Betz distin-
guishes between three different kinds of circular dialectical patterns –
redundant, maximally redundant, and strictly circular argumentation
(cf. Betz 2010:154–156) – and analyses their potential to contribute to
any discursive goals. The general claim that producing a circular argu-
ment is a mistake is challenged by the thesis that, due to the evaluative
underdetermination of dialectical structures, the expediency or inexpe-
diency of such a structure can only be established on a case-by-case
basis. While being inexpedient in some cases, circular argumentation
patterns can indeed serve a rational purpose: Betz gives examples in
which a circular structure is used to safeguard dialectical coherence, to
discharge burdens of proof, and to increase the robustness of a dialecti-
cal position (cf. Betz 2010:158–161).
In this way, the analysis of dialectical structures can yield valuable in-
sights into the quality and effectiveness of arguments and argument pat-
terns in a debate or controversy. Gregor Betz’s theory helps to illuminate
any activities and texts (also beyond the realm of philosophy) inasmuch
as they serve an argumentative purpose (cf. Betz 2010:178). Yet, how
should a text- or activity-based argumentation be reconstructed and
evaluated? Betz points out that any real-life argumentation is almost
inadvertently underdetermined; that is to say, whenever we engage in a
controversial discussion either orally or in writing, we hardly ever spell
out individual arguments, let alone the dialectical relations between
them. Often individual premisses are left out, and sometimes even the
conclusions are not explicitly mentioned. In other words: the sketchy
(and patchy) presentation of real-life argumentation needs to be comple-
mented by quite a few dialectical components that are merely implied
in an actual debate, before the argumentation and its arguments can be
analysed using the theory of dialectical structures.
This reconstruction of the original discussion in terms of a dialectical
structure that can, then, be analysed and evaluated calls for a hermeneu-
tic process (depicted in the form of a trefoil) in which the three elements
of validity, plausibility, and functionality are constantly reconsidered.
Validity implies the reconstruction of arguments as deductive argu-
ments, plausibility refers to the principle of charity, and functionality
has to do with the way in which an argument fits into the dialectical
context. Thus, only a deductive argument ensures that the premisses
justify the conclusion; the principle of charity demands that any dia-
lectical components that have to be added should be selected in line
with the most rational and sensible interpretation of the proponents’
intentions, meaning that arguments should be reconstructed with as few
premisses as are necessary to render the argument as strong as possible;
similarly, the suitability of an argument for a given dialectical context
relates to questions such as whether its reconstruction fulfils an obvi-
ously intended purpose (for example, of attacking another argument) or
whether its reconstruction ensures that the proponent achieves his or her
discursive goal. Betz recommends that in order to meet the requirements
shown in the hermeneutic trefoil, the reconstructor should first draw a
rough sketch of the overall dialectical structure before working out the
details.19
To conclude the discussion of Betz’s theory of dialectical structures,
we should ask whether, or to what extent, the proposed method of ar-
gumentative analysis and evaluation can actually claim to be objective.
Considering that the objective may be construed as a point of view from
nowhere, as “an assumption that everything must be something not to
any point of view, but in itself” (Nagel 1992:208), the theory of dialec-
tical structures might be too dependent on perspectives to achieve at
least a minimum of objectivity: it is not the dialectical structure that is
checked for coherence, but the various positions taken within that struc-
ture; not the dialectical structure determines whether arguments make
sense or are rather inexpedient, but the discursive goals proponents try to
achieve in the debate; the reliance on specific points of view culminates
in the evaluative underdetermination of dialectical structures, according
to which nothing can be said about the rationality of an argumentation
unless it relates to concrete points of view and discursive goals. At the
same time, however, the evaluation of points of view within the dialecti-
cal structure is in itself independent of any point of view and, therefore,
objective: whether a proponent’s position in a given dialectical structure
is dialectically coherent does not depend on the perspective from which
that question is answered; the evaluator’s point of view is similarly irrel-
evant, when it comes to defining whether one thesis follows dialectically-
semantically from another, or when the degree of justification of a thesis
is being established. Thus, while the process of reconstructing the in-
dividual arguments of a debate and the relations between them does
invariably contain a subjective element, the analysis and evaluation of
the dialectical structure (once it has been reconstructed) are objective
inasmuch as they rely only on the tools provided by Betz’s theory. 20
Dialectical Structures in Translation Quality Assessment

Having discussed Betz’s theory of dialectical structures in detail, we will
now provide an outline of how this approach can be used to evaluate
translation decisions. The key question is: how can translation quality
be measured using a dialectical argument structure? To answer this
question, we will generally assume the perspective of the translator as
an evaluator of his or her own translation.21 In trying to find target
language equivalents for the corresponding source language units, the
translator is faced with the problem of selecting the best solution from
a range of possible solutions. This range may be characterised by abun-
dance, by relatively straightforward correspondence, or by dearth de-
pending on (1) whether the translator has a large number of potential
solutions available to him or her, (2) whether he or she is presented with
just one obvious solution, or (3) whether there is no immediate solution
so that the translator has to resort to special translation techniques (such
as omission, source text retention, explanation, compensation, etc.). In
the second case, the selection process is easy because any alternatives
are clearly less appropriate than the one obvious solution. In the first
and third cases, however, the selection process is difficult, either because
there are many good alternative solutions or because there are several
special translation techniques that need to be considered. Whether easy
or difficult, the selection process is at the core of translation. To maxi-
mally control this process, the translator should be aware of the argu-
ments that make him or her select one solution over another.
In order to decide which of several potential translation solutions the
translator should select, he or she needs to compare the solutions’ argu-
mentative bases. To this end, each of the potential solutions is assumed
to be the best solution, and these assumptions are then cast into the form
of central theses. If there are, for example, two potential solutions –
solution “x” and solution “y” – that vie with each other in one and the
same translation context, the resulting central theses would run:
• Thesis 1: Given the specific translation context, “x” is the best trans-
lation solution.
• Thesis 2: Given the specific translation context, “y” is the best trans-
lation solution.
These two theses are, of course, mutually exclusive: only one of the solu-
tions can be the best in the given translation context. The translator’s
task is to find as many arguments as possible that support or attack solu-
tion “x” and as many arguments as possible that support or attack solu-
tion “y”. This task challenges the translator’s ability to take a neutral
perspective and not to play favourites. He or she needs to have a genuine
interest in finding the best solution – an interest that leaves no room for
subliminal preferences. While considering two irreconcilable positions
at the same time may be difficult, the general boost in translation quality
that comes with the translator’s ability to make a well-argued decision
when selecting the better of two alternative solutions is certainly worth
the effort.
Here, we should add a few remarks about the practical application
of Betz’s theory of dialectical structures to translation quality assess-
ment. While all elements of the theory are, in principle, relevant to TQA
(except for the degree of justification, as will be shown below), the fact
that their actual implementation in the evaluation process has to be fea-
sible in a real-life translation and evaluation context renders the simpli-
fication of several aspects inevitable. Neither a translator as his or her
own evaluator nor an external evaluator would be likely to go through
any of the formal arguments and argument patterns to arrive at a de-
cision, which of two alternative translation solutions is better than the
other. An external evaluator who has to record the results of his or her
evaluation will have to give his or her opinion in writing; yet, he or she
would stop short of providing a schematic argument structure to illus-
trate individual results. A translator – who as an evaluator of his or her
own translation is only responsible to him- or herself – will go through
the arguments and argument structures (which are supposed to deter-
mine the choice of a particular translation solution) in his or her mind
only. The arguments will not take the form of premisses followed by a
conclusion; rather, they will be informal arguments that one would use,
for instance, in a discussion.
To render the present theory as practicable as possible, we will con-
tent ourselves with informal arguments when it comes to supporting
or attacking a potential translation solution. Only if there is any doubt
whether an informal argument actual fits in the argument structure,
do we demand that the arguments of this structure be reconstructed in
line with the formal requirements set out in Betz (2010). It follows that,
indirectly, the informal arguments must also comply with Betz’s formal
requirements: a translator or evaluator who supports or attacks a trans-
lation solution should do so with an argument that is reasonable from
a common-sense perspective. Chances are that this informal argument
will, then, stand the test of being turned into a formal argument. Still,
as they are inherently underdetermined, informal arguments are more
likely to be subject to debate than formal arguments.
Before discussing in more detail the process of finding the best transla-
tion solution for a given source text unit, we should briefly explain why
it makes sense to talk about arguments and argument structures, even
though much of what determines the quality of a translation solution
is down to intuition. However, intuition usually does not operate in a
vacuum (with the possible exception of eureka moments and strokes of
genius): it depends on a cognitive framework within which the transla-
tor’s (and, to some extent, the evaluator’s) mind can produce the intu-
itive ideas that are then used to find a translation solution. While the
actual process of translating is not as straightforward as some attempts
at formalising it would seem to suggest, 22 the many loops and switches
between conscious cognition and unconscious intuition need not be
altogether erratic. The more clearly the translator defines the strategic
framework within which the translation process should take place, the
more easily will he or she be able to come up with a satisfactory solution.
If intuition can hold on to the right framework, it is probably more pro-
ductive and successful than without such a framework (cf. also Hönig
1995:47). Translation practice, therefore, can benefit from a theory that
helps the translator to define the co-ordinates of the translation task at
hand and to gauge the quality of any potential solutions.
The same is true of evaluation practice. The translator is always also
his or her own evaluator, because he or she constantly checks whether a
potential solution generated by intuition is actually fit to be used as the
final solution. Such checking is part of the selection process which Pym
(2003) regards as the second essential competence of a translator – the
first being the generation of alternative solutions. While it is probably
too simplistic to equate the generating of several target text units with
intuition and the selecting of one of these units with cognition, the dis-
tinction between these processes can be said to be indicative of the fact
that, in translation, conscious and unconscious processes are intricately
intertwined. In the case of an external evaluator, the ability to generate
viable solutions and to pick the most convincing of these solutions is as
important as it is in the case of the translator. Since the evaluator should
be able to justify any criticism and suggest an alternative that is proven
to be better than the translator’s solution, he or she needs to understand
the translator’s decisions. The present theory of translation quality as-
sessment is designed as a tool for the evaluator to thoroughly understand
why a translator translated as he or she did.
So how would a translator or evaluator go about determining whether
one translation solution or another is actually better? How would he or
she collect arguments and arrange them to build the argument struc-
tures? And how do these argument structures indicate which of the al-
ternatives is actually preferable? Before answering these questions, we
should point out that the overall evaluation process can only work, if it is
embedded in a comprehensive framework such as the one outlined in the
previous chapter on translation decisions. Without this framework, the
arguments and argument patterns might not attain the stringency that is
necessary for them to be convincing. The problem is – as Hönig points
out (cf. Hönig 1995:50) – that any arguments based on rules which are
regarded as absolute may turn out to be unfounded and, therefore, inef-
fective: the rules make sense, but they are not applicable in all contexts.
Since there are no absolute rules in translation, no invariables that
a translator can cling to in any translation situation, each rule used to
formulate an argument must be derived from the actual translation sit-
uation. The translation situation – with the source text and the transla-
tion brief, in particular – serves as the starting point for the translator
to seek out the right arguments for the right translation decisions. Both
the translator and the evaluator must come to terms with the fundamen-
tal issues that concern the scope within which the source text should
actually be translated. They are those issues which need to be decided
in order to define the translation strategy: notably the first, second, and
third decisions of the decision-making process described above. Just as
the first, second, and third decisions of the decision-making process de-
limit the range of any concrete microtext decisions in the course of trans-
lating, so should the arguments put forward to justify the first, second,
and third decisions support the arguments put forward to justify any
microtext decisions. An argument that supports a microtext decision
without being supported by a strategic argument is pointless.23
The requirement that any arguments which support or attack a partic-
ular translation solution have to be directly based on the arguments of
the translation strategy has several implications:
• First, since an argument not supported by any strategic arguments

is useless, there is no need to blindly generate as many arguments as
possible.
• Second, the strategic development of the target from the source via
a plausible line of argument ensures a minimum quality that usually
would go hand in hand with an acceptable translation solution –
provided that the strategic framework is not objectionable. This
means that an evaluator who approves of the translator’s strategy
and finds that a decision at the level of the microtext is in keeping
with that strategy should, generally, also approve of the microtex-
tual translation solution.
• Third, among the arguments that make up the translation strategy
we may distinguish between mandatory arguments and optional ar-
guments or between more important and less important arguments.
While this distinction is not covered by Betz’s theory of dialectical
structures, it may turn out to be relevant when it comes to decid-
ing difficult translation issues. By saying that translation arguments
need to be based on the arguments of the translation strategy, we
mean that they have to be based on all mandatory or important
arguments of the translation strategy. They need not be supported
by optional or less important arguments. The distinction between
mandatory and optional or between important and less important
arguments for the translation strategy requires that these arguments
are marked accordingly – it is particularly relevant to arguments
supporting the third decision. If necessary, the arguments may also
be ranked by importance.
• Fourth, a comparison of the arguments and/or argument structures
underlying two potential translation solutions for the same source
text unit will yield either of three results: both argument structures
are grounded in a proper translation strategy; neither argument
structure is grounded in a proper translation strategy; one argu-
ment structure is and the other is not grounded in a proper trans-
lation strategy. In the first case, both translation solutions would
be acceptable; in the second case, a third solution would have to be
found; and in the third case, the decision is made in favour of the
solution supported by the translation strategy.
• Fifth, the decision in favour of one solution is based on the num-
ber of arguments or on individual preference, if either alternative
is backed by the translation strategy. In this case, the solution with
more supporting (and fewer attacking) arguments is likely to be se-
lected. If the number of supporting and attacking arguments is the
same for both potential solutions, then the translator’s preference
will come into play. He or she will weigh the arguments as to which
are more important and which less. Note that this is a decision only
a translator will have to make; an evaluator who is faced with an
acceptable target text unit need not think about a better alternative.
• Sixth, the argument structures used to make microtext decisions
in translation are so much trammelled by the rules of the decision-
making process that any two structures for the same translation unit
can easily be compared on the basis of the number of arguments that
directly support or attack the potential solutions in question (always
provided that these arguments spring from the translation strategy).
This is one reason why there is no need to establish the degree of jus-
tification for any of the argument structures in the decision-making
process of translation. 24 Another reason is that the results obtained
by calculating the degree of justification reflect an exactness that the
hermeneutic process of reconstructing the formal from the informal
arguments could never achieve.
The above-sketched translation quality assessment theory follows a

ecision-making system within which alternative translation solutions
d
are strengthened or weakened by well-founded arguments. Below, we
will provide examples of arguments at each decision level. Note that,
in terms of the theory of dialectical structures, the decisions can be re-
garded as central arguments. In the three cases discussed below, the start
decision is neglected, because it merely signifies that the translator un-
dertakes to accomplish the translation task.
Case 1: Translating a Website
A British online translation agency would like its English website to
appear also in other European languages. They ask us to translate it into
German. We are aware that the website, and particularly the homepage,
has been designed to attract custom. The text has a conative function: its
message should be relevant to visitors of the website and encourage them
to do business with the agency (first argument). The visitors are only
interested in the text insofar as it relates to their needs (that is, finding a
suitable translation service provider); they are not interested in the fact
that they are reading a translation; they don’t care about the E nglish orig-
inal (second argument). Thus, the overall translation strategy is covert
(first decision). From the first and second arguments, we also conclude
that the translation should be oriented towards the target text reader and
his or her cultural background (third argument and second decision). In
line with such target orientation, we make assumptions about what a
typical reader might expect when reading the target text: the target text
should be written in such a way that it is easily understood (fourth argu-
ment) and that it is relevant to the reader’s situation (fifth argument). The
fourth argument implies the use of a varied but familiar vocabulary (sixth
argument), an idiomatic style (seventh argument), and a straightforward
syntax with a balanced use of co-ordinated and subordinated clauses
(eighth argument). The content of the source text should be retained in
the target text – so that any extra-linguistic cultural references made in
the original also appear in the translation. Target culture adaptations are
possible or necessary only for such cultural references that are used as
examples or with a figurative meaning, or that are explicitly required to
be changed in the target text – for instance, if a British contact address
should be replaced by a corresponding address in Germany (ninth ar-
gument). Another conclusion to be drawn from the first two translation
decisions is that the quality of the target text should be at least as good
as that of the source text; in other words, the target text should live up to
the expectations of its readers, even if the source text fails to do so (tenth
argument). Arguments four to ten make up the third decision. The argu-
ment structure so far constitutes our translation strategy.
Now we are ready to start the actual translation. Below the heading, the
homepage begins: “We know that communication is crucial to the success
and future growth of our business”. This sentence is grammatically im-
peccable, syntactically a little awkward, stylistically slightly verbose, and
teleologically unacceptable. Let us look at the issues in detail. The syntax
features an insignificant main clause, which could be omitted with no loss
of information: what they know is obvious by dint of their mentioning
it. The adjective “future” is redundant, because growth in this context
can only refer to the future. The most important issue is the last one: the
website, and especially the homepage, of a company – if it is to attract
custom – must focus on the needs of the potential customer or client. The
above introductory sentence, however, does just the opposite: it focusses on
the need of the translation agency. Here, a missing letter reorients the in-
tended focus as the possessive pronoun should be “your” instead of “our”.
A literal rendering based on the sentence as it stands might be acceptable,
if that sentence appeared in a management report. The actual context,
however, clearly points to a mistake in the source text. While the argument
in favour of a literal translation would seem to be intrinsically strong (after
all, a translator who translates the source as it is cannot really be blamed
by the client), it lacks the support of a convincing translation strategy. If
we take the above translation strategy for granted and prove that the argu-
ments which point to the source sentence as being completely inadequate
for the purpose for which it was written are more convincing than any
arguments that might be proposed in favour of a largely literal translation
of the source sentence as it is, then we get the following optimised source:
Communication is crucial to the success and growth of your business.
This can be rendered into German as Kommunikation ist für den Erfolg
und das Wachstum Ihres Unternehmens von entscheidender Bedeutung.
Case 2: Translating User Instructions
The translation into German of the user instructions for a vacuum
cleaner will be based on a strategy similar to that for the previously dis-
cussed website. The first decision points to a covert rendering, because
those who read the user instructions do not care about the existence of
an English original; they are just interested in the instructions as a means
to help them use the vacuum cleaner appropriately. Therefore, the trans-
lation should be geared towards the target culture (second decision), so
as to fulfil its instructional purpose most effectively. The arguments put
forward as part of the third decision include, among others, a uniform
terminology within the target text and the requirement that any errors
or inconsistencies in the original must be smoothed out in the transla-
tion. On the basis of this strategy, the translator should improve on the
source text. If the source text uses different terms to refer to one and the
same part of the vacuum cleaner (for instance, “dust cassette” and “dust
container”), then the target text should have only one expression to refer
to that part, because terminological variation might be confusing to the
user and should, therefore, be avoided.
Case 3: Translating a Journalistic Essay
In October 2014, NUVO, Canada’s premier lifestyle magazine, pub-
lished “Selfie Overload”, a short essay by Daniel Bettridge about the com-
mon habit of taking pictures of oneself. The text is stylistically interesting,
sprinkling witty everyday language with slightly extravagant expressions.
Its rendering into German for a German lifestyle magazine would be co-
vert, since the readers of the target text would not be interested in the
source text but would read the translation as if it were an original written
in the target language (first decision). Thus, the task of the translator is to
produce a text that embraces all the positive aspects of the original and
discards whatever is not so good (second and third decisions). Among the
positive aspects is certainly the brilliant style, which shows the text to
be equally entertaining and informative. In this context, however, much
depends on the translator’s interpretation. He or she needs to find answers
to the following questions when translating individual source text units:
1 Of the two functions conveyed in the source text, which one should
be dominant in the source unit: the entertaining or the informative
function?
2 Depending on the answer to the first question, what should be the
dominant function in the target text for the given unit?
While the answer to the first question entails a judgement on the quality
of the source text, the answer to the second tackles the consequences for
the target text. On the basis of a covert translation framework, a source
text unit that is stylistically not quite in line with the source text func-
tion should be improved in the target text. On the other hand: even if the
source text unit is deemed perfect, a stylistically equivalent translation
might be regarded as inappropriate in the target text. These consider-
ations pave the way for the arguments that are needed to decide individ-
ual translation issues.
One source text expression, “society’s penchant for self-portraiture”
(Bettridge 2014, no pagination), gives rise to an interesting argumen-
tative seesaw regarding the translation of the somewhat conspic-
uous compound “self-portraiture”. A literal rendering is possible,
yielding “Selbstporträtierung”. This solution can be attacked as be-
ing an abstract noun, which is stylistically not desirable if there is a
concrete alternative. Such an alternative would be “Selbstporträts”
(“self-portraits”). The corresponding argument garners support from
the translation strategy as the latter emphasises the importance of the
text’s informative function. At the same time, however, the source
(and target) text’s entertaining function provides the basis of a line
of argument that supports “Selbstporträtierung” as the preferred
solution: much of what makes for entertainment in the source text
is caused by a style that occasionally hampers the reader’s progress.
Here, the text draws attention to itself and it does so on purpose.
Thus, there are two possible lines of argument: the first one empha-
sises the importance of using concrete nouns to support the informa-
tive function of the text (“Selbstporträts”); the second one points up
the advantage of using a rather unusual noun to cater for the enter-
taining function of the text (“Selbstporträtierung”). In the end, the
translator has to decide which line of argument is more convincing.
Note: a third argument – namely, that the source text uses a less com-
mon noun so the target text should do the same – is irrelevant because
of the covert translation framework.
The three cases show that finding the right arguments to make the
right decisions is crucial already at the pre-translation stage, when the
translator needs to define his or her translation strategy. While the trans-
lation of user instructions or a company website leaves little doubt as to
which decisions are right and which would be wrong, there are other
translation situations in which the decision-making process at the pre-
translation stage is less straightforward. Clear decisions are supported
by clear arguments and need not be challenged by any arguments in
favour of alternative decisions. It is only when the source text and the
translation situation indicate less unequivocally whether an overt or
covert framework and a resulting source or target orientation is more
appropriate, that the strategic issues need to be decided by weighing
the pros and cons for each of the alternatives. For example, a report
originally published in The Guardian is translated into German and
published in the German weekly Der Freitag: that the translated text is
marked as a translation is intrinsically pointing to an overt rendering,
because the reader of the translation will know that he or she is reading
a translation. The question is, however, whether the reader is also inter-
ested in the English source text. The answer to this question depends
on the translator’s assessment of the situation: would the reader of the
translation prefer to read the original, and what aspects of the source
text would render it more attractive to the reader than the target text?
An external evaluator should carefully analyse the translator’s position
with regard to the translation strategy and accept it unless there is a
good reason not to.
This section has shown how arguments and argument structures can
be used to consciously develop a translation strategy, which then helps
the translator to achieve a good translation result. In the next section,
we will look at this issue solely from the point of view of the evaluator,
analysing translation criticism against the background of the decision-
making process in translation and the concomitant argument structures
implied.
Examples of Argumentation
Since evaluators are provided with the translation they are supposed to
evaluate, they start the evaluation process from a perspective that has
advantages and disadvantages compared to the translator’s perspective.
On the one hand, the evaluator approaches the translation with a fresh
mind, a mind that has not been numbed by translation routine and that
will notice errors which the translator did not notice; on the other hand,
the evaluator usually does not deal with and discuss the source text in
as much detail as the translator, who would do plenty of background
research which the evaluator would not. Charles Martin agrees that,
in this respect, the translator has an advantage over the evaluator (or
reviser):
[I]t should be noted that even when revisers are fully competent
they do not begin revision with the same knowledge of the source
text that the translator has, unless they first read carefully through
the document, which this author suspects is rarely the case, at least
in the private sector, given the time and cost constraints. Although
self-revision and other-revision apparently involve doing the same
thing, a translator begins self-revision from a much broader and in-
formed perspective since he or she has already read each sentence of
the text, given thought to it and has acquired an overall perception
of the document.
(Martin 2012, no pagination)
The evaluator should, therefore, strive to come to grips with the source
text in as much detail as possible to be able to make an informed judge-
ment about the quality of the translation.
Thus prepared, the evaluator is then in a position to praise and/or
criticise the target text. Ideally, he or she will observe the following key
points:
1 Try to understand why the translator translated as he or she did.

2 Give reasons why a particular translation solution is – or is
not – satisfactory.
3 Provide a better solution, if necessary, and state why it is better.
The first point includes, in particular, an assessment of the translation

strategy inasmuch as such a strategy can be derived from the target text
or from any explicit statements made by the translator. In this context,
knowledge of the translation norms prevalent in the target culture and
of the criteria that determine how the source text should be translated is
essential. Such knowledge, however, will only partly be made explicit in
the actual translation critique; the most obvious aspects are likely to be
taken for granted. By way of example, we will look at two instances in
which a translation solution is subjected to interesting criticism.
Criticising the Translation of a Young Adult Novel

To begin with, here is a remarkable piece of translation criticism – one
that complies with the above requirements to an extent seldom achieved
in this genre. In her discussion of the quality of young adult novels trans-
lated from English into German, Sabban (2009) focusses on the quality
of orality in representations of speech and how that quality has been
retained in translation. The discussion takes into account – and, indeed,
starts from – the fact that translations make up a sizeable percentage of
German publications of children’s and young adult literature and that
young adult novels, in particular, are often used at German secondary
schools as reading matter in German classes. While Sabban’s study cov-
ers two American novels, Myron Levoy’s A Shadow Like a Leopard
(German: Ein Schatten wie ein Leopard) and Sonia Levithin’s Incident
at Loring Groves (German: Die Tote im Wald), her paper discusses in
detail six excerpts from the first novel only. There are three parts: first, a
general analysis of the problem in the context in which it occurs; second,
a close look at instances of feigned orality in the source text; third, the
evaluation of the translation excerpts in the light of the criteria defined
in the first and second parts. To analyse the arguments put forward in
the paper, we will follow a two-step procedure: first, we will align them
with the decision-making process; then, we will show in a diagram how
these arguments combine to form a coherent argument structure.
The first decision deals with the question whether the readers are in-
terested in the fact that they are reading a translation. To answer this
question, the readership needs to be defined. For a young adult novel,
the target audience are adolescents from age 12 (cf. Sabban 2009:63).
They would look for a text that is convincing as a novel and a good read
(cf. Sabban 2009:70). The translation framework is, thus, covert. With
regard to the second decision, the translator should produce a target text
that is functionally equivalent to the source text and geared towards
the target language (cf. Sabban 2009:70). The second decision, like the
first, is based on an intuitive assumption of what the young adult target
audience would likely expect from the reading of Ein Schatten wie ein
Leopard. This assumption could be said to be indirectly supported by
the fact that students who have read the novel have shown a positive
response (cf. Sabban 2009:64).
Before analysing the consequences of the first two decisions for the stra-
tegic issues tackled in the third decision, we should mention a few special
arguments that are relevant in Annette Sabban’s paper. As the paper is
published in a volume that deals with language as a key qualification, the
quality of the translation is not only assessed in relation to how it affects
young adults as readers who enjoy reading a gripping novel. It is particu-
larly assessed against the background of young adults having to read the
novel as a requirement for their German class. This setting does not change
any aspects of the strategy that is supposed to determine the translation,
but it increases their relevance: a stylistic translation requirement which is
important to all young adults reading the target text becomes even more
important, if the audience consists of students who are supposed to peruse
the text in order to acquire literary and linguistic competences. This is
why the quality of linguistic expression is absolutely essential to the qual-
ity of the translation: the nuances of language can have a major influence
on how the target text is interpreted and they also affect – consciously or
subconsciously – the language skills of the students, who are inevitably
exposed to the language of the translation (cf. Sabban 2009:68). The as-
sumption that these considerations are not necessarily those of the transla-
tor (whose target audience would probably be teenagers who read for fun)
does not make any difference to the translation strategy.
The arguments which, as part of the third decision, define the trans-
lation strategy in more detail refer to the translation of direct speech,
because that is the object of the study. The relevant criteria include the
sociolect, the register, and the feigned orality revealed in such speech (cf.
Sabban 2009:68). Given these criteria and a close analysis of pertinent
passages in the source text, the task of the translator can be defined as
follows (cf. Sabban 2009:70):
1 He or she should try to achieve an orality that suggests authentic

spoken language.
2 The representation of speech should be such that it fits into the situ-
ational context in which it occurs – particularly with regard to the
adolescent characters and their specific social backgrounds.
3 The speeches must be constructed in such a way that they fit into the
co-text and make sense.
These three arguments form the basis on which individual target text
passages are analysed. They perfectly capture the requirements for the
translation of direct speech. We will see, however, that aspects such
as the plausible and faithful portrayal of the characters also apply to
non-direct speech text passages. The analysis then, since it focusses on
those translated passages that are not satisfactory, will reveal how a bad
translation can seriously affect the reception of the target text.
To demonstrate the operation of the arguments put forward in assess-
ing the quality of the translation, we will look at two sentences from the
third excerpt (cf. Sabban 2009:72):
• Some of the boys called him Book-Eyes because of his glasses and
squint. And his endless books.
The translation goes:
• Manche Jungen nannten ihn wegen seiner Brille und seines ständi-
gen Blinzelns auch Brillenschlange. Und dann sein ewiges Gerede
über Bücher.
This is a very interesting example, even without the representation of

speech. The German text is not acceptable and Annette Sabban clearly
shows why. She attacks the translator’s solution with arguments that
follow from the ideal translation strategy. 25 The first two arguments
prove that the meaning of the target text does not quite correspond to
that of the source text. While this proof is in itself insufficient because
of the covert strategy framework (in which semantic deviations are often
justifiable), Sabban combines the semantic mismatches with references
to a more serious shortcoming as a result of the mismatches. The boy
referred to here, Felipe, does not endlessly talk about books, and the cor-
responding German expression (“sein ewiges Gerede”) has negative con-
notations that are not implied in the original (cf. Sabban 2009:72). The
fact that a squint is a sight defect (whereas constant blinking – “
ständiges
Blinzeln” – is not) is mentioned only in parentheses; more important is
the (negative) impression conveyed by the target text that Felipe might
have a tic (cf. Sabban 2009:72). To this we can add a syntactic drawback:
while, in the original, the sentence “And his endless books” hinges on
“because of” in the previous sentence (which ensures text cohesion), the
German “Und dann sein ewiges Gerede über Bücher” is syntactically
unrelated to the previous sentence. Such unrelatedness might be an op-
tion if the sentences hung together semantically – h owever, that is not the
case: “Und dann sein ewiges Gerede über Bücher” (And then his endless
talk about books) seems to be a mere afterthought, a random addition
to the previous criticism that Felipe was a “Brillenschlange” (four-eyes).
These findings have serious repercussions on how the boy is perceived
by the reader. According to Sabban, one student wrote in her reading di-
ary that Felipe was a swot. This characterisation is due to the translation
(other passages in the target text have a similar effect); however, it does
not fit into the context. Why would the hero of the novel, Ramon, who
does not only do everything to be accepted in a youth gang but who also
has an interest in reading and writing – why would he admire a swot? In
the English original, Ramon experiences through Felipe that it is possi-
ble to succeed, even if one has to put up with teasing and contempt.26 In
other words, the target text paints a rather incongruous picture of the
relationship between Ramon and Felipe.
Another aspect to be criticised is the use of the term “Brillenschlange”
(four-eyes). In the German translation, Felipe got his nickname not only
because of his glasses but also because of his constant blinking (“we-
gen […] seines ständigen Blinzelns”). Yet, the term “Brillenschlange”, as
Sabban (2009:73) proves with a corpus search, only refers – somewhat
disdainfully or humorously – to people who wear glasses; it has nothing
whatsoever to do with blinking one’s eyes, nor does it reflect the keen
learner associated with the rather uncommon source text expression
“Book-Eyes”. There is another, slightly weaker, argument that could be
raised against the use of “Brillenschlange” in this context: although the
term can refer to both men and women, it is far more commonly used
with reference to women, as a corresponding image search would show.
“Brillenschlange”, therefore, does not make sense, here. As Annette Sab-
ban rightly emphasises, those readers who know the expression will get
confused unless they merely skim through the passage; whereas those
who do not know it will get the wrong idea of what it means (cf. Sabban
2009:73).
The analysis of the above excerpt shows that the translation is in-
adequate and should be changed in several respects. However, unlike
with many other issues discussed in her paper, Annette Sabban does not
really provide a better translation solution in this example. There may
be two reasons for this: first, the general focus on the purpose of the
translation as reading matter for students in secondary school German
classes renders pointing out the problems caused by the reception of the
target text more important than pointing out how the target text could
be improved; second, the changes necessary in the translation are, in
this case, either fairly obvious or idiosyncratic. Two improvements are
merely hinted at (cf. Sabban 2009:72): the addition of Felipe’s endless
talk (“sein ewiges Gerede”) should be dropped and “squint” should be
translated as “Sehfehler” or “Schielen”. Considering the above criticism,
we can translate, for example: Manche Jungen nannten ihn Buchauge –
wegen seiner Brille und seines Sehfehlers. Und wegen seiner unzähligen
Bücher. The expression “Buchauge” has connotations similar to that of
“Book-Eyes”; its grammatical number (singular rather than the plural,
“Buchaugen”) seems to be more typical of nicknames for individuals.
Another aspect that has been left out in the discussion of the above
excerpt is an attempt to explain why the translator, Elisabeth Epple,
translated the passage as she did.27 What might have caused her to se-
lect the term “Brillenschlange” is its idiomaticity – an idiomaticity that
seems to outshine the consequences of the term’s negative connotations.
Idiomaticity is also part of the explanation for the added noun phrase
“sein ewiges Gerede”; yet, disambiguation might have played a role, too.
While “his endless books” could refer to shelves full of books owned
by Felipe or just to the vast number of books he has read, the German
phrase denotes that books are on Felipe’s mind and that he bothers oth-
ers by talking about them. For both translation solutions, their negative
impact on the characterisation of the boy and, thus, on the interpreta-
tion of the novel is more important than their idiomaticity. This aspect is
not explicitly mentioned in Annette Sabban’s analysis of the appropriate
translation strategy but becomes obvious in the arguments put forward
against the shortcomings of the actual target text.
By way of summarising the findings in this discussion of Annette Sab-
ban’s translation evaluation, we will first draw up a list of the various
arguments before arranging them in a way that reveals their interrelat-
edness. The arguments are numbered consecutively (A1, A2, …, An).
In addition to the arguments actually stated in Annette Sabban’s paper,
we distinguish between arguments (prefixed with an “i”) that can be
regarded as implied by any of the explicit arguments and arguments
(prefixed with an “a”) that we have added. Moreover, we should point
out whether an argument is part of the translation strategy (marked
“ts”) and, if it contributes to the third decision in the decision-making
process, whether it is a mandatory (“m”) or an optional (“o”) argument.
In Table 5.5, arguments A9–A15 directly relate to the thesis that the
target text unit quoted above is the best translation of the correspond-
ing source text unit. The last argument, A15, reflects what might have
been the translator’s perspective. It supports the translator’s solution;
yet, its impact amounts to nothing, since it is cancelled out by arguments
A8–A14. Being a mandatory argument of the translation strategy, A8
ensures that A15 does not count, if the attack of arguments A9–A14 is
well founded – which it is. This is shown in Figure 5.1.
The ideal translation strategy is defined by the arguments of the first
three decisions (A1–A8) and forms the basis on which arguments A9–
A14 can attack the thesis. Argument A15 (which is not an acceptable
argument, because it lacks the support of the translation strategy) has
Table 5.5 List of arguments – critique of an excerpt from Ein Schatten wie ein Leopard
Argument ID Argument text
A1ts Since the target audience of the translated novel are young adults from age 12, it is likely that they expect a text that is
convincing as a novel and a good read.
iA2ts The target audience are probably not interested in the fact that they are reading a translation; the strategic framework is,
therefore, covert.
A3ts The translator should produce a target text that is functionally equivalent to the source text and geared towards the target
language, because the readers, who have shown a positive response, would want to read the novel as an independent text.
A4tsm The representations of speech should reflect an authentic spoken language.
A5tsm The representations of speech should appropriately reflect the character of the speaker.
A6tsm The speeches must be constructed in such a way that they fit into the co-text and make sense.
A7tsm With slight modifications, arguments A4–A6 also apply to non-direct-speech text: expressions must be appropriate in the
context in which they are used (in terms of language, content, and the characterisation of the protagonists in the novel).
iA8tsm If idiomaticity or authenticity vies with content and/or an appropriate characterisation of the protagonists, then the latter
are more important than the former.
A9 In the target text unit, the expression “sein ewiges Gerede” has negative connotations and, thus, distorts the
characterisation of Felipe as intended in the source text.
A10 In the target text unit, the expression “ständiges Blinzeln” would suggest that Felipe has a tic and, thus, distorts the
characterisation of Felipe as intended in the source text.
aA11 In the target text unit, the two sentences are syntactically unrelated and semantically only loosely connected, whereas they
are closely linked syntactically and semantically in the corresponding source text unit: this reinforces the effect described
in A9.
iA12 In the target text unit, the term “Brillenschlange” has negative connotations and, thus, distorts the characterisation of
Felipe as intended in the source text.
A13 In the target text unit, the term “Brillenschlange” is at least partly explained by the expression “ständiges Blinzeln” – an
explanation that is wrong (as can be shown with a corpus search), because the expression merely refers to an outer
characteristic (the wearing of glasses); the term “Brillenschlange”, therefore, does not make sense to the attentive reader.
aA14 In the target text unit, the term “Brillenschlange” refers to a boy, whereas the term is more commonly used with reference
to girls or women.
aA15 In the target text unit, all expressions used are very idiomatic.
The target text unit “Manche Jungen ... über Bücher” is the best translation of
Thesis:
the corresponding source text unit “Some of the boys ... endless books”.
⇡ ⇡ ⇡ ⇡ ⇡ ⇡ ↑
4th decision: A9 A10 A11 A12 A13 A14 A15
↑
3rd decision: A4 A5 A6 A7 A8
↑
2nd decision: A3
↑
1st decision: A1 → A2
Figure 5.1 A rgument structure – critique of an excerpt from Ein Schatten wie
ein Leopard.
been crossed out. If we replace the actual translation solution by the

above-mentioned alternative, then arguments A9–A14 would be irrele-
vant and could be replaced by two supporting arguments: the alternative
solution would ensure the syntactic integration of the second sentence
and a fitting characterisation of Felipe.
Criticising the Translation of a Survey

In his book Verstehen und Übersetzen, Paul Kußmaul presents his read-
ers with interesting translation problems. One such problem concerns
the translation into German of the question “Have you ever been di-
vorced?” (Kußmaul 2010:39). This is an example taken from the Eu-
ropean Social Survey 2004. According to Kußmaul, the translator first
came up with “Sind Sie jemals geschieden worden?” (ibid.) – which is
a literal translation. This solution, however, is not the best solution, as
Kußmaul explains:
Das Risiko, auf die Frage „Sind Sie jemals geschieden worden“ eine
falsche Antwort zu bekommen, ist wahrscheinlich gering, man kön-
nte sich aber vorstellen, dass die Befragten den Eindruck bekommen,
geschieden zu werden, sei eher etwas Ungewöhnliches oder vielle-
icht sogar moralisch Verwerfliches. Für einen guten Katholiken ist
es das ja auch. Eine spontane Antwort wäre dann: „Nein, natürlich
nicht.“ Und möglicherweise würde eine solche Antwort dann durch
die Frage suggeriert. Dies lässt sich als Szene beschreiben. Durch die
Frage im Ausgangstext wird eine neutrale und wertfreie Szene sug-
geriert. Wertfreiheit ist entscheidend bei Umfragen, um verfälschte
Antworten zu vermeiden.
(Kußmaul 2010:182)
The risk of receiving the wrong answer is probably low, when the
question is “Sind Sie jemals geschieden worden?” – however, respon-
dents may get the impression that being divorced is something rather
unusual or even morally reprehensible. For a good Catholic, that’s
what it is. A spontaneous answer would then be: “No, of course
not”; and such an answer would possibly have been suggested by
the question. This can be described as a scene [in terms of Fillmore’s
scenes and frames]. The question in the source text suggests a neu-
tral and non-judgemental scene. Neutralness is crucial in surveys to
avoid getting a distorted response.
This explanation clearly states why the first solution can still be im-
proved on. Paul Kußmaul suggests the following translation: “Sind Sie
schon einmal geschieden worden?” (Kußmaul 2010:182). Below, we will
demonstrate how a detailed analysis of the source text following the
pattern of the decision-making process can pave the way for what may
be regarded as a best translation.
The first decision concerns the question whether the overall translation
strategy should be overt or covert. In the above example, the respondents
are not in the least interested in the source text; they probably want to
be able to answer the questions without any difficulty and finish the sur-
vey as quickly as possible. The overall translation strategy is, therefore,
covert. The outcome of the first decision together with the assumptions
about the respondents’ expectations renders the outcome of the second
decision obvious: the translation should be target-oriented, taking into
account the respondents’ cultural background and their knowledge of the
world. In the third decision, these strategic requirements are further spec-
ified with regard to the above source text. The question “Have you ever
been divorced?” has to be translated in such a way that the target text
• preserves the source text’s noncommittal view of the respondent’s

marital status,
• is perfectly idiomatic (so as to ensure easy understanding),
• retains the neutral perspective of the source text as to who divorced
whom,
• does not imply any divorce-related value judgements.
Unlike the arguments inherent in the first two decisions, the requirements
derived from the third decision represent arguments that directly bear on
the selection of the best possible translation solution. These four arguments
are mandatory, because of the high demands made by the overall transla-
tion situation: a survey with unequivocal questions that the respondents
should be able to answer as truthfully as possible. A translation solution
can, therefore, only be satisfactory, if it is supported by all four arguments.
The arguments developed as a result of the translator’s decision-making
process define the translation strategy. They are summarised in Table 5.6.
Table 5.6 Translation strategy arguments – survey question “Have you ever
been divorced?”
Argument ID Argument text
A1ts The respondents of the survey are not aware of the fact that
they are reading a translated text; the strategic framework
is, therefore, covert.
A2ts The respondents probably prefer questions that are easy
to understand; therefore, the translation should be
target-oriented.
A3tsm The source text is noncommittal with regard to the
respondent’s marital status: the target text must also be
noncommittal in this respect.
A4tsm The source text is perfectly idiomatic and easy to
understand: the target text must also be perfectly
idiomatic and easy to understand.
A5tsm The source text does not imply that either the respondent or
his or her partner filed for divorce; the target text must
maintain the same neutral perspective.
A6tsm The source text does not suggest that having been divorced
is unusual or morally reprehensible; the target text must
maintain the same neutral perspective.
While the arguments are reasonably precise, they do allow for some
interpretational leeway. For example, the question whether or not a
potential translation solution is idiomatic cannot always be answered
with certainty: what is idiomatic and what is not often depends on
the intuition of a competent native speaker, and discussions with stu-
dents have shown that such intuition is variable. In some cases, even a
corpus search may not be conclusive. This is when the argumentative
approach to translation and translation quality assessment reaches
its limits and decisions have to be made intuitively – as we will see
below.
Having thoroughly analysed the source text and come up with a num-
ber of arguments for the translation strategy, the translator must now
try to find a target text that is supported by the strategic arguments. To
this end, the translator will intuitively generate what strikes him or her
as potential translation solutions. With idiomaticity as the guiding prin-
ciple, the solutions may include the following:
1 Sind Sie geschieden? [Are you divorced?]

2 Haben Sie sich schon einmal scheiden lassen? [Have you ever filed
for divorce?]
3 Sind Sie jemals geschieden worden? [Have you ever been divorced?]
4 Sind Sie schon einmal geschieden worden? [Have you ever been
divorced?]
These potential solutions have to be tested against the requirements
stipulated in the translation strategy. While all solutions are idiomatic,
each of the first three violates one requirement and, therefore, lacks
the support of the corresponding argument. This can be described as
follows:
1 The question “Sind Sie geschieden?” relates to the marital status

of the respondent rather than to his or her having gone through a
divorce sometime in the past.
2 The question “Haben Sie sich schon einmal scheiden lassen?”
wrongly suggests that the respondent actually filed for divorce –
whereas he or she may have been divorced without wanting to get
divorced.
3 As has been pointed out above, the question “Sind Sie jemals ges-
chieden worden?” may lead to an untruthful answer, since the re-
spondent feels that having been divorced is unusual and, perhaps,
morally reprehensible.
4 The question “Sind Sie schon einmal geschieden worden?” does not
violate any of the requirements.
Having pondered these issues, the translator is now in a position to se-

lect the fourth solution as his or her best solution.
The argument-based decision-making process can also be depicted in
a diagram for each of the potential translation solutions (see Figures
5.2–5.5). Since the source text consists of only one sentence and since ar-
guments A3–A6 in the above translation strategy present rather exhaus-
tively the requirements for the translation of that particular sentence,
there is no need for additional arguments relating to more specific trans-
lation issues. Any of the mandatory arguments A3–A6 either supports
the thesis that a potential translation solution is the best translation
solution (that is, if the solution fulfils the corresponding requirement)
or attacks the thesis that a potential translation solution is the best trans-
lation solution (that is, if the solution does not fulfil the corresponding
requirement).
The target text “Sind Sie geschieden?” is the best translation of the source text
Thesis:
“Have you ever been divorced?”.
⇡ ↑ ↑ ↑
3rd decision: A3 A4 A5 A6
↑
2nd decision: A2
↑
1st decision: A1
Figure 5.2 A rguments – first solution: “Sind Sie geschieden?”

The target text “Haben Sie sich schon einmal scheiden lassen?” is the best
Thesis: translation of the source text “Have you ever been divorced?”.
↑ ↑ ⇡ ↑
↑
2nd decision: A2
↑
1st decision: A1
Figure 5.3 A
rguments – second solution: “Haben Sie sich schon einmal scheiden
lassen?”
The target text “Sind Sie jemals geschieden worden?” is the best translation of
Thesis:
the source text “Have you ever been divorced?”.
↑ ↑ ↑ ⇡
↑
2nd decision: A2
↑
1st decision: A1
Figure 5.4 A
rguments – third solution: “Sind Sie jemals geschieden worden?”
The target text “Sind Sie schon einmal geschieden worden?” is the best
Thesis:
translation of the source text “Have you ever been divorced?”.
↑ ↑ ↑ ↑
↑
2nd decision: A2
↑
1st decision: A1
Figure 5.5 A
rguments – fourth solution: “Sind Sie schon einmal geschieden
worden?”
Figures 5.2–5.5 show that of the four potential solutions only the last
one is supported by all four mandatory arguments (A3, A4, A5, and A6).
These arguments are supported by the argument derived from the sec-
ond decision (A2), which, in turn, is supported by the argument derived
from the first decision (A1).
In the given translation situation, it is obvious that the translator will
select the last of the four potential translation solutions. However, this
process may be further complicated by additional solutions that spring
to the translator’s mind as an afterthought: as soon as “Sind Sie schon
einmal geschieden worden?” has been selected as the best rendering for
“Have you ever been divorced?”, the translator considers “Wurden Sie
schon einmal geschieden?”, “Sind Sie schon mal geschieden worden?”,
and “Wurden Sie schon mal geschieden?” as equally appropriate trans-
lations. The question is whether the changes in these alternative solu-
tions give rise to an additional optional or even mandatory argument
(one that has so far been overlooked) or whether a decision between
the four options must rely on gut instinct. The solutions can be distin-
guished by two features: tense (present perfect vs. past) and formal-
ness (more formal “einmal” vs. less formal “mal”). Thus, we get the
following:
1 “Sind Sie schon einmal geschieden Present perfect and more formal
worden?”
2 “Wurden Sie schon einmal geschieden?” Past tense and more formal
3 “Sind Sie schon mal geschieden Present perfect and less formal
worden?”
4 “Wurden Sie schon mal geschieden?” Past tense and less formal
The decision whether to use the present perfect or past tense cannot eas-
ily be based on a clear argument. While, in German, the present perfect
is often associated with spoken language and the past tense with written
language, this distinction does not hold for the verb “sein” (“to be”) –
compare “Wir waren in Spanien” (“We were in Spain”) and “Wir sind
in Spanien gewesen”. There is no difference in meaning, but the first
alternative is more common than the second. This is probably also true
of the verb “werden”, albeit to a lesser extent. The ultimate decision in
favour of one tense or the other is, therefore, not particularly relevant to
the quality of the target text. Note that the tense used in the source text
must not play a role in the translator’s decision, because a correspond-
ing argument would be totally irrelevant due to the covert translation
strategy.
As regards the more formal/less formal distinction, a decision should
be made on grounds that go beyond the source text question. While it
may be argued that a less formal question sounds slightly more nat-
ural than a more formal one, such an argument may be offset by an-
other argument which emphasises the need for stylistic uniformity of
the whole survey: if the other questions have been translated in a more
formal style, then the question “Have you ever been divorced?” should
be translated in the same way. Such an argument would be supported
by arguments A1 and A2; however, it should not count as a mandatory
argument, because if, in some case, it contradicts the idiomaticity argu-
ment, the latter takes priority. The uniformity argument would, then,
be an optional argument. Whichever solution the translator chooses, it
is acceptable. The translator could make a random choice, trust his or
her gut feeling, or select a solution along the subtle lines of argument
described above. An evaluator should not find fault with the solution.
This is the point where our argumentative approach to translation qual-
ity assessment – while possibly still relevant to the translator – becomes
insignificant to the external evaluator, who should be satisfied with a
solution that is acceptable.
This chapter has shown how argument structures can be used to illus-
trate the decisions made in the process of translating and in translation
quality assessment. Whether a translator needs to rely on the intricate
interplay between intuitively generating potential solutions and con-
sciously selecting one of these alternatives, or whether an evaluator
needs to assess the translator’s decisions and gauge them against his or
her own standards – the above system can be used to do both: translate
a source text into a satisfactory target text and assess the quality of a
translation in a way that is to a high degree intersubjectively verifiable.
To summarise our argument-based system of translation quality assess-
ment, we can put forward the following principles:
• Translation is essentially a decision-making process; the decisions of

this process are based on argumentative reasoning.
• An argument structure consists of one thesis and one or several ar-
guments relating to that thesis.
• The thesis claims that a particular translation decision or transla-
tion solution is the best decision or solution.
• The arguments relating to the thesis are expressed mentally, orally,
or in writing, using natural language.
• The arguments are informal arguments: they do not follow the for-
mal argument structure of premisses and conclusion. If necessary,
they could be reconstructed as formal arguments.
• It is not necessary to calculate the degree of justification of the the-
sis, because the argument structures themselves reveal which trans-
lation solution is better.
• At least two alternative theses and their corresponding argument
structures are needed to make a translation decision at the level of
the microtext.
• A translation decision favours the thesis that commands the stron-
gest argumentative support.
• In the decision-making process, later decisions follow from earlier
decisions.
• The arguments sustaining the first three decisions constitute the
translation strategy.
• At the level of the third decision, the translation strategy is char-
acterised by mandatory and, possibly, optional arguments. The re-
quirements expressed in the mandatory arguments are binding for
any direct translation decisions.
• The arguments directly supporting a thesis in the context of a mi-
crotextual translation decision must be supported by all mandatory
strategic arguments – otherwise, they are irrelevant.
• A solution that is supported by all mandatory arguments of the
translation strategy is an acceptable solution provided that the
translation strategy is sound.
The application of these principles to the evaluation of commented

translations in a university context will be demonstrated in Chapter 6.
Notes
1 The two solutions may also turn out to be equally unacceptable, if both are
supported by rather weak arguments and a third solution can be shown to
be better than either of the first two solutions.
2 On this mode of translating, cf. Vermeer (2003:20–21).
3 That the evaluator or critic should make an effort to understand the transla-
tor’s decisions is shown, for example, in Reiß (1971) and Broeck (1985).
4 This has been pointed out by Hans Vermeer, who writes: “Translation
fordert Entscheidungen auf allen sprachlichen und kulturellen Rängen”
(Vermeer 1994:35).
5 By contrast, a translation method refers to an individual technique by which
a source text unit is rendered into a target text unit in the face of certain
linguistic or cultural obstacles. Such translation methods are described, for
example, in Vinay/Darbelnet (1995:30ff).
6 Obviously, the translation strategy concept proposed here is prescriptive in
the sense that a translator who follows such a strategy is likely to produce
a good translation. At the same time, however, the strategy concept is ex-
planative in that it helps to explain the decisions that must have been made
(explicitly or implicitly), if the translation is acceptable. That translation
strategy can also be regarded as any cognitive process by which a translator
translates is shown by Wolfgang Lörscher, who has carried out a study to
develop “a descriptive concept of strategy” (Lörscher 1991:72).
7 Cf. Krüger (2013) for a detailed discussion of the concepts of explicitation
and implicitation from a cognitive linguistic perspective. Murtisari (2013),
who looks at the matter from a relevance-theory angle, suggests that the dis-
tinction between explicitness and implicitness is a cline rather than a clear-
cut dichotomy.
8 Similarly, Hönig (1995:55–56) emphasises the need for a macrostrategy
(“Makrostrategie”), which enables the translator to monitor and control his
or her microstrategies in the process of translating.
9 Interestingly, this assumption is also made by Carsten Sinner and Beatriz
Morales Tejada (2015) in a study on translation perception as the basis of
defining successful translations.
10 The overt–covert distinction is presented here as if it were applicable to all
cultures. From our limited English-German perspective, though, such a uni-
versal claim would be presumptuous. One could well imagine cultures in
which the different expectations of readers implied in the distinction be-
tween overt and covert translation do not exist or are irrelevant because a
translated text is always marked as such and read as such. Nevertheless, we
take it that, in a large number of cultures, the above-described difference
in what readers expect from translated texts is relevant to the way in which
translation works in these cultures.
11 Other possible approaches to argumentation would have been Toulmin
(1958) and Freeman (1991).
12 This is comparable to Kenneth McLeod’s distinction between deductive, in-
ductive, and defeasible types of argument. The latter are of a rather “fragile
nature” as they “[lack] the verifiable backing of probability and thus the
strength of inductive arguments” (McLeod 2012:23).
13 The three points of departure are discussed in Betz (2010:29–47); for a sum-
mary of what the theory of dialectical structures is supposed to achieve, cf.
Betz (2010:47).
14 This presupposes that the reasoning behind an argument must be valid, so
that, if all premisses of an argument are true, it is rational to assume that the
conclusion of that argument is also true.
15 The conditions are preliminary in the sense that they will be defined more
precisely in the light of further theoretical considerations (cf. Betz 2010:87–
91). For our purposes, though, the more straightforward definitions suffice.
16 In a dialectical structure, the premiss of an argument is called a “free” prem-
iss, if and only if there is no dialectically valid argument the conclusion of
which negates, or is equivalent to, the premiss (cf. Betz 2010:84).
17 Betz shows that, while dialectical-syntactic entailment always implies
dialectical-semantic entailment, this is not true the other way around (cf.
Betz 2010:109).
18 Cf. Betz (2010:150f.) for a discussion of the terminology.
19 The aspects discussed in this paragraph summarise Betz (2010:180–184).
20 This reflects Betz’s train of thought on the objectivity of the evaluation pro-
cedures in dialectical structures (cf. Betz 2010:192–193).
21 In the case of an external evaluator, the situation is different inasmuch as
the external evaluator evaluates a translated text against the background of
his or her own world knowledge and translational expectations. Here, much
depends on the degree to which the evaluator is willing to, and manages to,
reconstruct the translator’s intention and his or her strategy.
22 One such example can be found, for instance, in Krings (1986:404). By
contrast, Hönig (1995:51) proposes a model that purports to be somewhat
closer to translating as it actually happens.
23 A typical example would be the argument that, of two viable target text
units, one is better than the other because it is closer to the wording of the
source text, even though, in the particular translation situation, closeness to
the source text is absolutely irrelevant to the quality of the translation.
24 The degree of justification is relative to the ratio between directly supporting
and directly attacking arguments: two supporting and one attacking argu-
ment result in a higher degree of justification than three supporting and two
attacking arguments.
25 The – in many places – less-than-satisfactory translation suggests that the
strategy followed by the translator must have been different from the trans-
lation strategy outlined above, which constitutes the evaluator’s suggestion
of how the novel should have been translated.
26 This summarises the slightly more elaborate argumentation in Sabban
(2009:73).
27 The discussions of other excerpts do feature the occasional explanation of
what might have led the translator to come up with such an unsatisfactory
translation. An example can be found in the discussion of the second excerpt
(cf. Sabban 2009:72).
References
Arthern, Peter J. (1983). “Judging the Quality of Revision”. In: Lebende
Sprachen, vol. 28, no. 2, 53–57.
Bayer, Klaus (2007). Argument und Argumentation: Logische Grundlagen der
Argumentationsanalyse. 2nd ed. Göttingen: Vandenhoeck & Ruprecht.
Bettridge, Daniel (2014). “Selfie Overload”. In: NUVO [online]. Viewed:
25 August 2019. http://nuvomagazine.com/magazine/autumn-2014/
selfie-overload
Betz, Gregor (2010). Theorie dialektischer Strukturen. Frankfurt/Main: Vitto-
rio Klostermann. (Philosophische Abhandlungen, vol. 101).
Proceedings of the 3rd Congress of the International Federation of Transla-
tors (Bad Godesberg, 1959), New York: Macmillan/Pergamon Press.
Chesterman, Andrew (1994). “Karl Popper in the translation class”. In:
Dollerup, Cay; Annette Lindegaard (eds.) (1994:89–95).
Dollerup, Cay; Anne Loddegaard (eds.) (1992). Teaching Translation and
Interpreting – Training, Talent and Experience. Papers from the First Lan-
guage International Conference, Elsinore, Denmark, 31 May – 2 June 1991
(Copenhagen Studies in Translation). Amsterdam/Philadelphia: John
Benjamins.
Dollerup, Cay; Annette Lindegaard (eds.) (1994). Teaching Translation and
Interpreting 2 – Insights, Aims, Visions. Papers from the Second Language
International Conference, Elsinore, Denmark, 4–6 June 1993. Amsterdam/
Philadelphia: John Benjamins.
Dung, Phan Minh (1995). “On the acceptability of arguments and its funda-
mental role in nonmonotonic reasoning, logic programming and n-person
games”. In: Artificial Intelligence, vol. 77, 321–357.
EMT expert group (2009). “Competences for professional translators, experts
in multilingual and multimedia communication.” Viewed: 25 August 2019.
https://ec.europa.eu/info/sites/info/files/emt_competences_translators_en.pdf
Freeman, James B. (1991). Dialectics and the Macrostructure of Arguments:
A Theory of Argument Structure. Berlin and New York: Foris Publications.
Freie und Hansestadt Hamburg, Behörde für Inneres (2010). “Merkblatt für
die Anfertigung von beglaubigten Übersetzungen”. Viewed: 25 August
2019. www.hamburg.de/contentblob/2111860/data/merkblatt-ueberset
zungen.pdf
Halliday, M. A. K. (2001). “Towards a theory of good translation”. In: Steiner,
Erich; Colin Yallop (eds.) (2001:13–18).
Helm.
Honko, Lauri (2000). “Text as process and practice: the textualization of oral
epics”. In: Honko, Lauri (ed.) (2000:3–56).
Honko, Lauri (ed.) (2000). Textualization of Oral Epics. Trends in Linguistics.
Studies and Monographs 128. Berlin and New York: Mouton de Gruyter.
Horn-Helf, Brigitte (1999). Technisches Übersetzen in Theorie und Praxis.
Tübingen/Basel: Franke.
Tübingen: Narr.
Jumpelt, Rudolf Walter (1961). Die Übersetzung naturwissenschaftlicher und
technischer Literatur. Berlin: Langenscheidt. (Langenscheidts Bibliothek für
Wissenschaft und Praxis 1).
Jumpelt, Rudolf Walter (1963). “Methodological Approaches to Science Trans-
lation”. In: Cary, Edmond; Rudolf Walter Jumpelt (eds.) (1963:267–281).
Katan, David (2017). “Translation Theory and Professional Practice: A
Global Survey of the Great Divide”. In: Hermes – Journal of Language
and Communication Studies, vol. 22, no. 42–2009, 111–154. doi: 10.7146/
hjlcb.v22i42.96849. Viewed: 25 August 2019. www.researchgate.net/
publication/228683372_Translation_Theory_and_Professional_Practice_A_
Global_Survey_of_the_Great_Divide
Kempa, Thomas (2013). “Anregungen für eine Differenzierung eines funk-
tionalen translationsdidaktischen Verstehens- und Sinnbegriffs aus prag-
matistischer Sicht”. In: trans-kom, vol. 6, no. 2, 345–370. Viewed: 25
August 2019. www.trans-kom.eu/bd06nr02/trans-kom_06_02_04_Kempa_
Sinnbegriff.20131212.pdf
Krings, Hans P. (1986). Was in den Köpfen von Übersetzern vorgeht: Eine em-
pirische Untersuchung zur Struktur des Übersetzungsprozesses an fortge
schrittenen Französischlernern. Tübingen: Narr.
Krüger, Ralph (2013). “A Cognitive Linguistic Perspective on Explicitation and
Implicitation in Scientific and Technical Translation”. In: trans-kom, vol. 6,
no. 2, 285–314. Viewed: 25 August 2019. www.trans-kom.eu/bd06nr02/
trans-kom_06_02_02_Krueger_Explicitation.20131212.pdf
Künzli, Alexander (2014). “Die Übersetzungsrevision – B egriffsklärungen,
Forschungsstand, Forschungsdesiderate”. In: trans-kom, vol. 7, no. 1, 1–29.
Viewed: 25 August 2019. www.trans-kom.eu/bd07nr01/trans-kom_07_01_
01_Kuenzli_Revision.20140606.pdf
Kußmaul, Paul (2010). Verstehen und Übersetzen: Ein Lehr- und Arbeits-
buch. 2nd ed. (1st ed. 2007). Tübingen: Narr Francke Attempto.
Lenz, Friedrich (ed.) (2009). Schlüsselqualifikation Sprache: Anforderungen –
Standards – Vermittlung. Forum Angewandte Linguistik vol. 50, Frankfurt/
Main etc.: Peter Lang.
Levý, Jiří (2000/1967). “Translation as a Decision Process”. In: Venuti, Law-
rence (ed.) (2000:148–159). Originally published in 1967 in To Honor Ro-
man Jakobson on the Occasion of his Seventieth Birthday, vol. 2. The Hague:
Mouton, 1171–1182.
Lörscher, Wolfgang (1991). Translation Performance, Translation Process, and
Translation Strategies: A Psycholinguistic Investigation. Language in perfor-
mance, vol. 4. Tübingen: Narr.
Martin, Charles (2012). “The Dark Side of Translation Revision”. In: Transla-
tion Journal, vol. 16, no. 1, January 2012 [online]. Viewed: 12 August 2019.
http://translationjournal.net/journal/59editing.htm
McLeod, Kenneth C. (2012). Argumentation in biology: exploration and analy-
sis through a gene expression use case. Heriot Watt University: Thesis.
Mossop, Brian (1992). “Goals of a revision course”. In: Dollerup, Cay; Anne
Loddegaard (eds.) (1992:81–90).
Munday, Jeremy (2012). Evaluation in Translation. Critical points of translator
decision-making. London and New York: Routledge.
Murtisari, Elisabet Titik (2013). “A relevance-based framework for explici-
tation and implicitation in translation: an alternative typology”. In: trans-
bd06nr02/trans-kom_06_02_03_Murtisari_Explicitation.20131212.pdf
Nagel, Thomas (1992). Mortal Questions. Cambridge: Cambridge University
Press. (first published in 1979).
Paepke, Fritz (1971). “Sprach-, text- und sachgemäßes Übersetzen”. In: Wilss,
Wolfram (ed.) (1981:112–119).
vol. 48, no. 4, 481–497.
Reinart, Sylvia (2014). Lost in Translation (Criticism)? – Auf dem Weg zu einer
konstruktiven Übersetzungskritik. Berlin: Frank & Timme. (Transkulturalität –
Translation – Transfer, vol. 5).
München: Max Hueber. (Hueber Hochschulreihe 12).
Sabban, Annette (2009). “Übersetzte amerikanisch-englische Jugendliteratur
im Deutschunterricht der Sekundarstufe I und die Ausbildung literarischer
und sprachlicher Kompetenzen”. In: Lenz, Friedrich (ed.) (2009:63–79).
Salama-Carr, Myriam (2009). “French tradition”. In: Baker, Mona; Gabriela
Saldanha (eds.) (2009:404–410).
Schmitt, Peter A.; Susann Herold; Annette Weilandt (eds.) (2011). Translations-
forschung. vol. 2. Frankfurt/Main etc.: Lang, 749–761. (Leipziger Studien zur
angewandten Linguistik und Translatologie 10).
Schubert, Klaus (2007). Wissen, Sprache, Medium, Arbeit: Ein integratives
Modell der ein- und mehrsprachigen Fachkommunikation. Tübingen: Gunter
Narr.
Schubert, Klaus (2011). “Übersetzen als Entscheidungsprozess. Eine Rekapitu-
lation”. In: Schmitt, Peter A. et al. (eds.) (2011:749–761).
Sinner, Carsten; Beatriz Morales Tejada (2015). “Translatologische Perzep-
tionsstudien als Grundlage der Bestimmung gelungener Übersetzungen”. In:
Lebende Sprachen, vol. 60, no. 1, 111–123.
Snell-Hornby, Mary (1994). “Einleitung – Übersetzen, Sprache, Kultur”. In:
Snell-Hornby, Mary (ed.) (1994). Übersetzungswissenschaft –Eine Neuorien-
tierung. 2nd ed. Tübingen and Basel: Franke.
Steiner, Erich (2004). Translated Texts: Properties, Variants, Evaluations.
Frankfurt/Main etc.: Peter Lang.
Steiner, Erich; Colin Yallop (eds.) (2001). Exploring Translation and Multilin-
gual Text Production: Beyond Content. Berlin and New York: Mouton de
Gruyter.
Stolze, Radegundis (1992). Hermeneutisches Übersetzen: linguistische Katego-
rien des Verstehens und Formulierens beim Übersetzen. Tübingen: Narr.
Stolze, Radegundis (2001). Übersetzungstheorien: eine Einführung. Tübingen:
Narr.
Toulmin, Stephen E. (1958). The Uses of Argument. Cambridge: Cambridge
University Press.
Toury, Gideon (1995). Descriptive Translation Studies and beyond. Amsterdam
and Philadelphia: John Benjamins.
Venuti, Lawrence (ed.) (2000). The Translation Studies Reader. London:
Routledge.
Vermeer, Hans J. (1994). “Übersetzen als kultureller Transfer”. In: Snell-Hornby,
Mary (ed.) (1994:30–53).
Studia Germanica Posnaniensia, vol. XXIX, 19–38.
Vermeer, Hans J. (2007). Ausgewählte Vorträge zur Translation und anderen
Themen. Berlin: Frank & Timme.
Vinay, Paul; Jean Darbelnet (1995). Comparative Stylistics of French and En-
glish: A Methodology for Translation. Amsterdam and Philadelphia: John
Benjamins.
Wilss, Wolfram (ed.) (1981). Übersetzungswissenschaft. Darmstadt: Wissen-
schaftliche Buchgesellschaft. (Wege der Forschung vol. 535).
Based on the principles of translation quality assessment set out in

Chapters 4 and 5 and the evaluation framework described in Chapter 3,
we are now in a position to analyse actual evaluations, that is, the ex-
aminers’ reports of commented translations written to fulfil the require-
ments for a BA degree. We take as a point of departure the prescriptive
elements of our theory of translation quality assessment – those aspects
of evaluation that have been regarded as essential by other translation
scholars and for which the above theory provides a coherent rationale.
This is complemented by a descriptive dimension in which we will es-
tablish how evaluators actually evaluate commented translations: what
approaches they use and what their priorities are. Finally, by comparing
the descriptive findings with the prescriptive elements, we will get a first
impression of whether and to what extent the examiner’s reports meet
the scholarly expectations of good translation quality assessment.
The Corpus: Examiners’ Reports of Commented

Translations
Theses written as part of the bachelor’s degree course International
Communication and Translation at the University of Hildesheim may
either discuss a subject related to the study programme or feature the
translation of a 3,000-word text. The second option is referred to as
a commented translation, because it includes not only a detailed ana
lysis of the source text but also a commentary on the translation pro-
cess and/or individual translation decisions. As has been pointed out in
Chapter 3, the commentary ideally serves to defend the translation on
the basis of the source text analysis.1 It may constitute a separate chap-
ter of the thesis (like the analysis of the source text) or take the form of
footnotes or endnotes.
While commented translations are fairly common in International
Communication and Translation at the University of Hildesheim, they
appear to be less common at other German institutions that offer de-
gree courses in translation studies. When I contacted the translation
Evaluating the Evaluator 225
departments at several universities, some said they did not accept com-
mented translations as bachelor’s theses, whereas others told me that rel-
atively few of their bachelor’s theses were commented translations. In the
end, I received examiners’ reports only from the University of Hildesheim.
Each examiner agreed (usually in writing) that I was allowed to use his
or her report(s) for my project.2 To render the texts anonymous, any
personal data of examiners and students have been deleted. The total
corpus comprises 82 reports by 11 different examiners. More than half
of the reports (46) were prepared by one examiner, the others contributed
between one and ten reports. I have refrained from adding to the corpus
any of my own evaluations of commented translations.
Methodology
The analysis of the corpus will cover all aspects of the reports: the for-
mal characteristics of the theses referred to in the reports and the formal
characteristics of the reports themselves, the evaluation of the source
text analysis as well as the assessment of the translation together with
the translation commentary. To describe these aspects, we shall build
a set of clearly defined categories and a corresponding set of suitable
questions by which these categories can be elicited. For all those aspects
that relate to the translation – that is, the source text analysis, the ac-
tual rendering, and the commentary – the questions should also cover
the prescriptive dimension so as to enable a comparison between TQA
desiderata and the actual handling of the translation quality assessment.
Whenever a question lends itself to making assumptions about the out-
come of the investigation, we will express this assumption in terms of a
hypothesis.
As we analyse the examiners’ reports and the way in which they re-
flect the different approaches of assessing the commented translation,
we should be aware of the practical aspects involved in preparing the
report. One such aspect is time. Examiners often have to read several
theses so that in order to do justice to all of them, they have to strike the
right balance between a detailed analysis and a cursory reading of the
commented translations. In other words, time constraints might cause
the examiner to examine the student’s translation not as closely as he
or she would have done otherwise. Such time constraints could be at
fault, if the report is relatively short, if it provides examples only from
the beginning of the target text, or if it discusses the translation without
reference to the source text. Another practical aspect has to do with the
underdetermination inherent in informal argumentation: an informal
argument or a statement often implies other arguments that are not ex-
plicitly mentioned in the report. When evaluating the evaluator, we need
to take such implicit arguments into account.
Looking at Formal Characteristics
The formal characteristics of a commented translation include the fol-
lowing aspects:
• the title page;

• the table of contents;
• the chapters featured in the thesis and the extent to which they fulfil
their respective functions;
• the overall structure, in particular, the order of the chapters and
their numbering;
• the layout;
• mistakes in spelling, grammar, and punctuation outside the actual
translation;
• the length of the thesis.
The questions that arise in this context are, for example:
• Does the title page contain all details required?

• Is the table of contents correct? Does it give the page numbers of
each chapter?
• Does the thesis have an introduction and a conclusion? Does the
bibliography cover all references made in the text?
• Have the chapters been arranged in a plausible order? Is the num-
bering logical?
• Does the text show irregularities in layout? Does it comply with any
stipulated layout requirements?
• Are there frequent mistakes in spelling, grammar, and punctuation
in any of the non-translation chapters of the thesis?
• Does the thesis stay below the required maximum length of 40 pages?
Our task is to find out whether the examiner of a commented translation

provides answers to any such questions. To elicit the formal characteris-
tics listed above, we can ask:
Question 1: What formal characteristics of the thesis does the evaluator
cover in his or her evaluation?
The answer to this question will consist of a list of categories of formal
characteristics. While the above seven categories are intended to be ex-
haustive, there may – in the analysis of the examiners’ reports – well arise
characteristics that would best fit into an altogether different category.
This is why our conclusive list will comprise the following eight categories:
1 title page
2 table of contents
3 chapters
4 overall structure
5 layout
6 mistakes
7 length
8 others
Since the present work focusses on the TQA approaches featured in the
examiners’ reports, the formal characteristics of the commented transla-
tions will not be analysed in detail. Reference will only be made to the fact
whether or not a particular category has been mentioned in the report.
The formal characteristics of the reports themselves include overall
structure and length. With regard to overall structure, we can distin-
guish between a strong and a weak structure: a strong structure has a
distinct outline with paragraphs (and, possibly, headings) that clearly
refer to one particular issue; a weak structure, on the other hand, con-
sists of large units that integrate the discussion of several issues. Since it
may not always be obvious if a given examiners’ report features a weak
or a strong structure, it makes sense to also cater for hybrids: those cases
in which the structure is neither strong nor weak (but something in be-
tween). Thus, we get:
Question 2: Does the report have a weak or a strong structure?
The answer will be one of the following categories:
1 weak
2 strong
3 neither weak nor strong
As regards the length of the reports, we will give the approximate num-
ber of pages. In this context, a first general hypothesis comes into play.
As the report should be proof against any objections that might be raised
by the student, it is likely that commented translations of low quality
require more evidence than commented translations of high quality that
the criticism put forward in the report is, indeed, justified. Moreover,
since the quality of a translation is often assessed based on the number
of errors, commented translations with many errors will probably result
in longer reports. The hypothesis, then, runs:
Hypothesis 1: The length of an examiner’s report is in inverse propor-
tion to the quality of the commented translation.
Of course, the length of a report also depends on the preferences and
time constraints of the examiner. Therefore, it makes sense to test the
hypothesis not only for the complete corpus but also for the reports writ-
ten by each individual examiner. Although the background to the first
hypothesis might suggest a prescriptive dimension even to the formal
characteristics of the examiners’ reports (according to which it would
be better to write longer reports rather than shorter ones), such consid-
eration is out of place, here, because it neglects the complexity of the
overall evaluation process. The prescriptive dimension will become rel-
evant in the next section, when we look at aspects of translation quality
assessment.
Looking at Aspects of Translation Quality Assessment

The focus of the present study is on analysing how evaluators of com-
mented translations back up their TQA approach – in particular, what
principles they follow and what arguments they use. In keeping with our
findings from previous chapters, we will ask whether and to what extent
the examiners’ reports abide by the standards that translation scholars have
called for. The corresponding issues will be discussed under three headings:
source and target, errors and achievements, and evidence and arguments.
Source and Target

A central question in TQA concerns the actual comparison of source and
target: does the evaluator of a translation look at both the target text
and the source text, or does he or she assess the target merely in its own
right? The criteria that may help decide this question have largely been
set forth in Chapter 3. They are summarised in Table 6.1.
The problem with these criteria is that it may be impossible to estab-
lish whether they affect the evaluation of the commented translation,
unless the report explicitly refers to them as the cause of a particular
procedure. A case in point is an insufficient quality of the source text
analysis: here, the evaluator could even stop short of reading the transla-
tion if the source text analysis is a fail. Another clear hint might be the
absence of source text examples – which might indicate that source and
target have not been compared. References to line numbers of the source
text might give a rough idea of what source text sections any examples
were taken from. For instance, if these numbers do not exceed 100 or
150 and if the examples reveal quite a few errors, then it is likely that
the second half or so of the translated text has either not been examined
or deemed unnecessary to prove the quality of the translation. That an
examiner’s report explicitly provides details of the evaluation procedure
is not to be expected. The vague conclusions that could, in some cases,
be made with regard to the comparison of source and target or a mere
perusal of the target text can usually not be put down to any hard and
fast criteria. Therefore, we will ask generally:
Question 3: What hints does the examiner’s report provide with re-
gard to the evaluation procedure, that is, in particular, with regard
to the question whether the source text and the target text have been
compared?
Table 6.1 C
omparison of source and target – criteria
Criteria Explanation
Time constraints Ideally, the evaluator does both: compare source and target for accuracy of the translation and read the
target text on its own for coherence and consistency. However, since the observant evaluator may also pay
attention to coherence and consistency of the target while comparing it to the source, the comparison is
more important than the perusal of the translated text on its own. Thus, if faced with time constraints, the
evaluator should go for the comparison rather than the target text reading. Only exceptional circumstances
might justify skipping the comparative element. These circumstances will be broached in connection with
the other criteria.
Source text type and Much depends on the translation strategy. If the source text and translation brief require a strict overt
translation brief handling of the translation, a comparison of source and target is essential. On the other hand, if the
translation brief demands the covert translation of a source text in which the conative function is
dominant, then the source is relevant only inasmuch as it provides the basis of the content of the target
text. Provided that the evaluator has read the source text and knows what it is about, he or she can
dispense with a sentence-by-sentence comparison of source and target. In a commented translation,
however, the overall situation tends not to favour an extremely overt nor an extremely covert approach. In
other words, a comparison of source and target is generally considered advisable.
Quality of the target A translation that is either extremely good or extremely bad could be argued to require less effort on the
text part of the evaluator than a translation whose quality ranges somewhere between these extremes. If a first
reading of the target text shows that the quality is unacceptable and the only possible result is a fail, then
there may be no need to compare source and target. The case might be similar – though, perhaps, less
unequivocal – if the translation strikes the evaluator as exceptionally good. Rather than dispensing with
either the comparison of source and target or the close reading of the target text on its own, the evaluator
could also opt for a partial evaluation of the commented translation: if the excerpts looked at show a fairly
consistent quality, it can be assumed that the rest is of approximately the same quality.
Quality of the source A recent exam regulation stipulates that the analysis of the source text must achieve at least a pass. This
text analysis regulation could have serious consequences for the evaluation of the translation, because a failed source
text analysis would render an assessment of the target text quality superfluous.
Preference of the Irrespective of the above criteria, an evaluator will have his or her own preferences with regard to the
evaluator evaluation of a commented translation. If he or she deems it sufficient to just compare source and target or
to just read the target text on its own, then the other criteria may have little or no effect on the evaluation.
Due to the vagueness of this question, there is no need to set up any re-
sponse categories. Such categories will be more readily established in the
following chapter, when we look at errors and achievements.
Errors and Achievements

To assess the quality of a translation, the evaluator has the following
options:
1 giving a general impression of the quality without reference to indi-

vidual errors and achievements
2 focussing on errors as detracting from an expected quality standard
3 focussing on achievements as indicators of translation quality
4 taking into account both errors and achievements in the translated text
The first option would be particularly suitable for a target text written
to a very high standard – with hardly any mistakes and no translation
errors. In the second option, the evaluator expects an acceptable transla-
tion as the norm and estimates or calculates the impact that errors have
on the quality of the target text. The focus on achievements in the third
option can be regarded as an effort to emphasise the positive elements
in a translation and, thus, to encourage the student: as with the second
option, the effect of the successful translation solutions on the overall
target text quality can be either estimated or calculated. The same holds
true for the fourth option, which combines the second and third options.
Calculating the impact of errors and/or achievements on the quality of
the translation adds to the objectivity that every evaluator should strive for,
because classifying, weighting, and counting errors and/or successful trans-
lation solutions involves a close analysis of the source and target texts.3
However, since different source texts are not necessarily comparable, the
evaluator ought to establish a framework by which he or she can assess the
level of difficulty of translating a given source text. Whether such an assess-
ment is carried out in a wholesale fashion after a close reading of the text to
be translated or whether the level of difficulty is calculated on the basis of
the number of source text sentences or phrases that pose special translation
problems – the result would have to be a factor by which the number of er-
rors and/or successful solutions could be standardised. For a translation of
average difficulty, the number of errors and the number of successful trans-
lation solutions would be multiplied by one. That factor would be slightly
higher for errors, if the translation was comparatively easy, and slightly
lower, if the translation was particularly difficult. To count achievements,
the adjustment of the factor would be reverted. Obviously, such a detailed
approach is extremely time-consuming, so that it is doubtful whether it is
applied to the evaluation of commented translations.
The first question that should be asked in this context is as follows:
Question 4: How does the evaluator assess the quality of the translation?
The answer to this question will be one of four categories corresponding
to the four options given at the beginning of this section:
1 general impression
2 error focus
3 achievement focus
4 errors and achievements
A closer look will then reveal if the examiner’s report distinguishes be-
tween different kinds of errors and/or achievements:
Question 5: Have the errors and/or achievements been classified and,
if so, how?
Providing answers to this question in terms of predefined categories is
pointless as there are many possibilities of distinguishing between different
errors or achievements. If errors and/or achievements have been classified
in an examiner’s report, we will give a brief description of the error and/
or achievement classes. Apart from being classified, errors and/or achieve-
ments may also be counted. Here, the straightforward question is as follows:
Question 6: Does the evaluator count errors and/or achievements?
In the case of an affirmative answer, we should then also ask:
Question 7: Have the errors and/or achievements been weighted and,
if so, how?
For counting makes sense only if major errors and/or achievements
are weighted differently from minor errors and/or achievements. What
counts in the end is not the number of errors and/or achievements but
the number of points calculated by totalling the weighted errors and/or
achievements. Errors (and, to a lesser degree, achievements) can also be
weighted in an impressionistic fashion, the most straightforward distinc-
tion being that between major and minor errors. As with classification,
the weighting of errors and/or achievements, if applicable, can only be
described individually for each examiner’s report.
Closely related to weighting is the last question under the heading of
errors and achievements:
Question 8: Which evaluation perspective prevails: the professional
reader’s or the student translator’s perspective?
Four answers are possible:
1 the professional reader’s perspective

2 the student translator’s perspective
3 neither perspective (if the issue is handled inconclusively)
4 not applicable
The question about the evaluation perspective might be applicable, if the
examiner’s report focusses on errors and if at least one of the errors can
be interpreted in different ways. A wrong spelling, for example, that is
not easily recognised as such since it inadvertently changes the meaning
of the intended word, would be considered a major error from the profes-
sional reader’s perspective but a minor error from the perspective of the
student translator – after all, it is just a spelling mistake (cf. Chapter 3
for an example). Here, a specific weighting of the error would indicate its
relative seriousness or triviality as would a pertinent explanation in the
examiner’s report. Where such errors are not weighted at all or handled
inconclusively, neither perspective prevails. However, errors that can be
interpreted from different points of view are not very frequent. There-
fore, it is likely that, in most cases, this question will be regarded as not
applicable. Where it is applicable and where the evaluator indicates his
or her point of view through a corresponding interpretation of the error,
we are already touching on aspects of evidence and arguments – which
are the subject of the next section.
Evidence and Arguments

In Chapter 5, we have shown that arguments play a crucial role in
translation – when making decisions in favour of a particular solution
and when assessing the quality of these solutions. We have also shown
that the need for argumentation in translation quality assessment has
been recognised from different angles by several translation scholars.
Their demands have led to the following TQA requirements:
1 An evaluator should try to understand why the translator translated

as he or she did.
2 An evaluator must be able to justify any changes he or she would
want to make to a translation.
3 The evaluator must provide a better solution and state why it is
better.
4 The evaluator’s arguments draw on factors that cannot be defined in
advance but must be established for each translation and translation
quality assessment on the basis of the source text and the overall
translation situation.
These requirements will now be discussed in more detail. The idea is to

generate suitable questions and response categories that can be used to
analyse the examiners’ reports and find illuminating answers as to how
evaluators justify their evaluations of commented translations.
The first requirement refers to the translator’s overall translation strat-
egy and to his or her reasons for preferring one particular translation
solution to another. The question is how the strategy and any preferences
can be recognised in the commented translation. There are two possi-
bilities: they may be explicitly mentioned, as is the case, for example,
when the source text analysis leads to an obvious translation strategy or
the commentary clearly states why an alternative solution has not been
selected; or the translator’s intention may be merely implied in the way
he or she has translated certain source text passages. Whether explicit
or implicit, the translator’s intention can be either included in the exam-
iner’s report or ignored. To include the translator’s intention, the evalu-
ator would have to quote relevant remarks from the source text analysis
or the commentary, or reconstruct the strategy or individual decisions
from the commented translation. While the first option is undoubtedly
explicit, the second may range from obvious attempts at reconstructing
the translator’s intention to mere assumptions about what the translator
might have had in mind. Finally, the evaluator may also criticise the
translator’s failure to explain his or her translation strategy. As we ana
lyse the examiners’ reports, we should ask the following question:
Question 9: Does the evaluator discuss the translator’s intention and,
if so, how?
If the answer is “yes”, the following categories become relevant:
1 Quoting the translation strategy: the evaluator quotes the transla-

tion strategy explicitly mentioned by the translator.
2 Quoting the translator’s reasons: the evaluator quotes any reasons
the translator gives to defend individual translation decisions.
3 Reconstructing the strategy: the evaluator reconstructs the transla-
tor’s strategy.
4 Reconstructing the translator’s choice: the evaluator reconstructs
why the translator chose a particular translation solution.
5 Criticising disregard of a strategy: the evaluator criticises that the
translator does not sufficiently explain his or her translation strategy.
Since the evaluator evaluates, first and foremost, the translated text, and
since the translator’s intention is, in a sense, implied in the way he or she
translated the source text, we expect that, for many examiners’ reports, the
above question will be answered in the negative. In those cases where the
answer is positive, the majority of reports will either criticise the translator’s
disregard of a strategy or refer to the translator’s reasons for a particular
translation solution: while disregard of a strategy is easily made out in the
translator’s source text analysis, his or her reasons for a particular trans-
lation solution can be readily found in the commentary to the translation.
If an evaluator finds fault with a particular translation solution, he
or she has to state why.4 The arguments provided to support any such
criticism of the target text must be sound in the sense that they need to
be founded in a translation strategy derived from the source text analysis
and the translation brief. Any other arguments are irrelevant. For exam-
ple, if the translator intentionally omitted a source text element because
it would be redundant in the target culture, then any argument against
such a procedure would be inappropriate in a covert translation setting.
Also not founded in a translation strategy are quasi-tautological argu-
ments (“this is wrong because it is unacceptable”) and superficial argu-
ments (“this doesn’t sound nice”). The following two questions cover
both negative and positive criticism:
Question 10a: To what extent does the evaluator provide arguments as
to why a translation solution is not satisfactory?
Question 10b: To what extent does the evaluator provide arguments as
to why a translation solution is successful?
Before answering these questions, we need to discuss another distinction
with regard to arguments.
Ideally, each unsatisfactory solution is criticised with at least one rel-
evant argument; irrelevant arguments should not occur. However, some
target text errors might be so blatantly obvious that there is no need for
an explicit argument to attack the corresponding translation solution.
This is usually the case with grammar and spelling mistakes: here, the
arguments can be regarded as implied by the mention of the mistakes. In
answering the question whether and to what extent the evaluator provides
arguments against unsatisfactory translation solutions, both explicit and
implicit arguments must be taken into account. Obviously unsatisfactory
translation solutions are those errors that Pym (1992) refers to as “bi-
nary” (cf. Chapter 3 for a brief discussion of this concept). Apart from
grammar and spelling mistakes, they also include false friends and sim-
ilarly evident translation errors. These errors usually concern individual
words or phrases and tend to have little effect on a proper understanding
of the target text (because they are easily recognised). If in doubt whether
an unsatisfactory solution is an obvious mistake, we will demand that the
evaluator’s criticism be underpinned by an explicit argument.5
Non-binary errors usually call for an explicit argument; and this
argument will often be supported by quoting the error together with
the phrase, clause, sentence, or even paragraph in which it occurs. In
some cases, such a context may imply the argument why the translation
solution is unsatisfactory and, thus, render an explicit argument unnec-
essary. There are, then, two different kinds of implicit arguments: the
arguments implied in binary errors and the arguments implied in those
non-binary errors quoted with a sufficient context. Note that the number
of implicit arguments includes only the most obvious argument; any ad-
ditional arguments that could be devised do not count. Thus, the number
of implicit arguments equals the number of binary errors and sufficiently
contextualised non-binary errors without explicit arguments. Multiple
arguments provided for one error or successful translation solution must
be explicit arguments. If a binary error or sufficiently contextualised
non-binary error comes with an explicit argument, then the argument
implied in the error does not count unless the explicit argument provides
another, less obvious reason for criticising the translation solution.
Admittedly, the question whether an error quoted without an explicit
argument and without context is binary or non-binary is subject to inter-
pretation. If we can imagine a likely context in which a solution that is
considered unsatisfactory can actually be regarded as satisfactory, then
the mere mention of this solution as unsatisfactory does not imply the
required argument. For example, criticising the rendering of English “in-
vite” as German “auffordern” calls for an explicit argument, because
there are contexts in which such a translation is adequate. A non-binary
error quoted without an explicit argument and without context will
probably be confirmed as an error by recourse to the source text, even
if a first impression suggests that the corresponding translation solution
is acceptable. Still, we should allow for the possibility that checking the
source text does not furnish such confirmation.
With regard to successful translation solutions, it could be contended
that they are less in need of being substantiated by an argument than
translation errors, because the point made in mentioning them does not
challenge the judgement of the translator, who would probably agree that
his or her translation solution is, indeed, successful. Although successful
translation solutions are likely to be quoted without an argument, any
arguments that are put forward to underpin a positive judgement can
only be explicit arguments. In other words, a successful translation solu-
tion cannot imply the argument why it is successful. It is to be expected
that positive translation criticism will often dispense with argumentative
support and, possibly, discuss the relevant solutions in a general way.
With this, we are now in a position to specify the insights to be gained
from answering the argument questions 10a and 10b. We are consider-
ing the results from four points of view.
Looking at translation solutions in general:
1 total number of translation solutions referred to in the examiner’s

report
2 number of unsatisfactory translation solutions referred to in the ex-
aminer’s report
3 number of successful translation solutions referred to in the examin-
er’s report
Looking at errors mentioned in the examiner’s report:
4 total number of binary errors

5 number of binary errors with an additional explicit argument that is
different from the argument implied
6 number of binary errors where the implicit argument is made explicit
7 total number of non-binary errors
8 number of non-binary errors quoted with a context sufficient to im-
ply the argument proving the point
9 number of non-binary errors with explicit arguments
10 number of non-binary errors without arguments – neither implicit nor
explicit (calculating the difference between 7 and, in brackets, 8 plus 9)
Looking at successful translation solutions mentioned in the examiner’s

report:
11 number of successful translation solutions with arguments

12 number of successful translation solutions without arguments
Looking at arguments that come up in the examiner’s report:
13 total number of arguments including multiple arguments for one

translation solution
14 number of implicit arguments (adding the results of 4 and 8 and
subtracting 6)
15 number of explicit arguments including multiple arguments (calcu-
lating the difference between 13 and 14)
16 number of relevant arguments (explicit and implicit)
17 number of irrelevant arguments (can only be explicit)
In addition, we can formulate two hypotheses:

Hypothesis 2: The number of explicit arguments exceeds the number of
implicit arguments.
This hypothesis is based on the assumption that the evaluator, when discuss-
ing the translation errors, is likely to pick on the less obvious ones – those
that need to be substantiated by an explicit argument – because the obvious
errors can easily be dealt with in a wholesale fashion. With regard to success-
ful translation solutions, we expect that the third hypothesis is true:
Hypothesis 3: The number of successful translation solutions without
arguments exceeds the number of successful translation solutions with
arguments.
While the above results and hypotheses cover the core aspects of ar-
gumentation in evaluating commented translations, they do not quite
conclude the argument issue – as will be shown in the next paragraph.
Translation errors have so far been regarded as those individual ren-
derings that the evaluator finds fault with, because he or she deems
them unsatisfactory for some reason. But what if a translation solution
that has been judged unsatisfactory is, in fact, perfectly adequate? This
question brings to mind our key proposition that the quality of a trans-
lation solution is only as good as the arguments supporting it (cf. Chap-
ter 5). Argumentation must establish whether a solution is satisfactory
or unsatisfactory. Using the wrong argument to criticise a translation
solution may result in misjudgement. Such misjudgement can emerge
from any of the following four constellations:
1 A satisfactory solution is criticised as unsatisfactory without argu-

mentative support.
2 A satisfactory solution is criticised as unsatisfactory by providing an
irrelevant or even wrong argument.
3 A satisfactory solution is “corrected” by providing another solution
that is equally satisfactory.
that is wrong.
While it is clear that such assessment errors should not occur, they can-
not be altogether ruled out.6 The first and third constellations can easily
come about, when an expression strikes the evaluator as not quite ap-
propriate and he or she can immediately think of an alternative. How-
ever, providing an irrelevant or wrong argument (second constellation)
or even a wrong solution (fourth constellation) to substantiate criticism
of a satisfactory solution would be a very serious error on the part of the
evaluator. The question that elicits any of the four constellations or a
negative response runs:
Question 11: Does the evaluator judge as unsatisfactory translation
solutions that are, in fact, satisfactory and, if so, in which way?
It is to be expected that the negative responses are the rule and the cases
described in the first and third constellations are extremely rare. We are
not likely to find instances that correspond to the second and fourth con-
stellations. In judging the evaluator’s judgement, we should, of course,
make allowance for individual preferences: what we might regard as just
about acceptable or possibly acceptable (depending on a context that the
evaluator does not provide) may already be unacceptable to the evaluator.
Therefore, we will accept the evaluator’s judgement unless there is clear
evidence that the criticised solution is, indeed, satisfactory. The next step
in our analysis of the evaluator’s assessment of translation quality takes
up the question of alternative solutions provided in the examiner’s report.
In Chapter 2, we come to grips with Juliane House’s approach to
translation quality and give two examples in which translation criticism
is not convincing, because it stops short of providing a well-argued alter-
native. Generally, there is no point in saying that a translation solution is
unsatisfactory, if one cannot come up with a better alternative and state
why it is better. This is certainly true for those unsatisfactory translation
solutions that require a clear explicit argument as to why they are unsat-
isfactory. Therefore, we ask:
Question 12: To what extent does the evaluator provide an alternative
to a criticised target text solution, and to what extent does he or she
show that the alternative is, indeed, better?
Only those obvious errors whose mention implies the underlying argu-
ments can do without a correcting alternative. In such cases, the correct-
ing alternative and the argument justifying the alternative can be regarded
as implicit. The above question, then, should yield the following results:
1 number of unsatisfactory solutions for which no alternative is given;

2 number of unsatisfactory solutions for which at least one alternative
is given;
3 number of alternative solutions provided (including multiple alterna-
tives for one translation error);
is given that is not supported by an argument;
is given that is supported by an argument (or arguments).
The first two figures show how far an examiner’s report meets the re-
quirement of providing an alternative solution that is more suitable than
the criticised translation. They allow for both explicit and implicit alter-
natives. An alternative is implied, if it is the only possible alternative (or an
obvious alternative) for the error indicated (for example, the correct spell-
ing of a word); an implicit alternative does not extend to translation errors
that are obvious yet can be corrected in different ways. In combination
with the second result, the third figure indicates whether, in some cases,
the evaluator provides more than one alternative for one unsatisfactory
translation solution. In the fourth result, the number covers only explicit
alternatives, because an implicit alternative would imply the reason why
it is better than the translator’s solution. Finally, the fifth result includes
both explicit and implicit alternatives and arguments. Note: while an im-
plicit alternative always points to a corresponding implicit argument, this
is not necessarily true the other way round. An implicit argument involves
an implicit alternative only if the underlying error is a binary error; an im-
plicit argument derived from the sufficient context of a non-binary error
does not suggest an obvious alternative solution nor does it automatically
double as an argument that supports an explicit alternative. These figures,
then, reveal the extent to which the examiners’ reports comply with the
third TQA requirement mentioned at the beginning of this section.
The fourth requirement is about the factors that can be used to ar-
gue for or against individual translation solutions. These factors may be
drawn from a large variety of sources (as has been shown in Chapter 4)
provided that they emanate from the actual translation situation and are
relevant to translation quality assessment. The different sources have
been outlined in the concept of the translator’s daffodil. This concept
can help the evaluator to cover a wide range of TQA factors and it can
help us to analyse the content of any explicit or implicit arguments in
the examiners’ reports. The translator’s daffodil arranges the following
factor groups around the focal target text:
• source text
• text form
• client
• translator
• culture
• politics
What factors from these factor groups actually play a role in the evalu-
ation of the target text depends on their relevance to the usability of the
translation (cf. the definition given in Chapter 4 of factors relevant to the
assessment of translation quality).
In the context of a commented translation, the factors relevant to the
evaluation of the target text are not likely to be culled from the whole
range of factor groups. The evaluator’s arguments will probably draw
on text form, source text, and culture factors rather than on factors to
do with the translator, the client, or politics. Here are some examples
of factors that can be expected to be prominent in the arguments of the
evaluators:
1 The most obvious errors – grammar and spelling mistakes – relate

to the text form factors lexis and syntax. Corresponding criticism
would be supported by implicit or explicit arguments to the effect
that a particular word or sentence does not comply with the rules
of the target language. Since grammar and spelling mistakes are
usually evident, it is to be expected that they are often dealt with in
a wholesale fashion rather than being listed individually.
2 Another text form factor, meaning, comes into play when the eval-
uator finds fault with an unequivocally wrong translation solution.
As in the first example, the mere mention of the mistake implies the
argument that supports the evaluator’s finding.
3 The time of communication as a source text factor is important
whenever a relative time reference in the source text would need to
be changed in the target text. The corresponding argument should
state that the change is necessary due to the source and target text
functions and the resulting covert translation strategy.
4 Source text defects constitute an important source text factor. Such
defects may extend, for example, to factual errors, inappropriate
terminology, or stylistic inadequacy.
5 Culture factors such as cultural norms are always relevant to trans-
lation, albeit often in a more general way. Their relevance is more
pronounced whenever culture-specific items have to be rendered
from the source into the target language and culture. Here, an ex-
plicit argument would have to explain the appropriate handling of
the translation issue against the background of cultural norms.
Given the complexity of the translation process with a large number of

factors influencing the target text, any question which tries to elicit the
content of the evaluator’s arguments cannot be answered simply by a
set of response categories. Thus, the following question will result in a
description of the argumentative content revealed in the reports:
Question 13: What are the arguments put forward by the evaluator?
Here, the concept of the translator’s daffodil with its factors and factor
groups may serve as a descriptive framework. The description of the
evaluator’s arguments concludes our discussion of aspects and charac-
teristics that can be expected in an examiner’s report. Next, we will
summarise the methodology.
Summary of Methodological Issues

The questions and response categories by which we intend to capture
the most important aspects of the examiners’ reports are summarised in
Table 6.2. The questions give rise to three hypotheses:

implicit arguments.
arguments exceeds the number of successful translation solutions
with arguments.
Whether the hypotheses are true or false will be revealed in the next
section together with the results of the study.
Results
Applying the above set of questions to the corpus is straightforward
in theory but less so in practice. Consistency arguably constitutes the
biggest problem facing the researcher: trying to ensure that errors and
Table 6.2 Summary of questions and response categories
Question Response categories
Question 1: 1 title page

What formal characteristics of the thesis 2 table of contents
does the evaluator cover in his or her 3 chapters
evaluation? 4 overall structure
5 layout
6 mistakes
7 length
8 others
Question 2: 1 weak
Does the report have a weak or a strong 2 strong
structure? 3 neither weak nor strong
Question 3: [not applicable]
What hints does the examiner’s report provide
with regard to the evaluation procedure,
that is, in particular, with regard to the
question whether the source text and the
target text have been compared?
Question 4: 1 general impression
How does the evaluator assess the quality of 2 error focus
the translation? 3 achievement focus
4 errors and achievements
Question 5: [yes + description; no]
Have the errors and/or achievements been
classified and, if so, how?
Question 6: [yes; no]
Does the evaluator count errors and/or
achievements?
Question 7: [yes + description; no]
Have the errors and/or achievements been
weighted and, if so, how?
Question 8: 1 the professional reader’s perspective
Which evaluation perspective prevails: 2 the student translator’s perspective
the professional reader’s or the student 3 neither perspective
translator’s perspective? 4 not applicable
(Continued)
Question Response categories
Question 9: If the answer is “yes”:

Does the evaluator discuss the 1 Quoting the translation strategy: the evaluator quotes the translation strategy explicitly mentioned by the
translator’s intention and, if so, how? translator.
2 Quoting the translator’s reasons: the evaluator quotes any reasons the translator gives to defend
individual translation decisions.
3 Reconstructing the strategy: the evaluator reconstructs the translator’s strategy.
4 Reconstructing the translator’s choice: the evaluator reconstructs why the translator chose a particular
translation solution.
5 Criticising disregard of a strategy: the evaluator criticises that the translator does not sufficiently explain
his or her translation strategy.
Question 10a: Looking at translation solutions in general:
To what extent does the evaluator
1 total number of translation solutions referred to in the examiner’s report
provide arguments as to why a
2 number of unsatisfactory translation solutions referred to in the examiner’s report
translation solution is not satisfactory?
3 number of successful translation solutions referred to in the examiner’s report
Question 10b:
To what extent does the evaluator Looking at errors mentioned in the examiner’s report:
provide arguments as to why a
4 total number of binary errors
translation solution is successful?
5 number of binary errors with an additional explicit argument that is different from the argument implied
6 number of binary errors where the implicit argument is made explicit
7 total number of non-binary errors
8 number of non-binary errors quoted with a context sufficient to imply the argument proving the point
9 number of non-binary errors with explicit arguments
10 number of non-binary errors without arguments – neither implicit nor explicit (calculating the difference
between 7. and, in brackets, 8. plus 9.)
Looking at successful translation solutions mentioned in the examiner’s report:
11 number of successful translation solutions with arguments
12 number of successful translation solutions without arguments
Looking at arguments that come up in the examiner’s report:
13 total number of arguments including multiple arguments for one translation solution
14 number of implicit arguments (adding the results of 4. and 8. and subtracting 6.)
15 number of explicit arguments including multiple arguments (calculating the difference between 13.
and 14.)
16 number of relevant arguments (explicit and implicit)
17 number of irrelevant arguments (can only be explicit)
Question 11: If the answer is “yes”:
Does the evaluator judge as
1 A satisfactory solution is criticised as unsatisfactory without argumentative support.
unsatisfactory translation solutions
2 A satisfactory solution is criticised as unsatisfactory by providing an irrelevant or even wrong argument.
that are, in fact, satisfactory and, if so,
3 A satisfactory solution is “corrected” by providing another solution that is equally satisfactory.
in which way?
4 A satisfactory solution is “corrected” by providing another solution that is wrong.
Question 12: 1 number of unsatisfactory solutions for which no alternative is given;
To what extent does the evaluator 2 number of unsatisfactory solutions for which at least one alternative is given;
provide an alternative to a criticised 3 number of alternative solutions provided (including multiple alternatives for one translation error);
target text solution, and to what extent 4 number of unsatisfactory solutions for which at least one alternative is given that is not supported by an
does he or she show that the alternative argument;
is, indeed, better? 5 number of unsatisfactory solutions for which at least one alternative is given that is supported by an
argument (or arguments).
Question 13: [not applicable]
What are the arguments put forward by
the evaluator?
arguments of the same kind are interpreted in the same way – which
presupposes a clear notion, what “of the same kind” actually means.
The countless forms in which language is used to refer to errors and
express arguments leave some scope as to the classification of the latter.
For example, errors are binary if they are obvious and if they imply one
correct rendering. While the first condition can often be verified with-
out difficulty, the second is subject to a more intuitive approach: does
one particular solution lend itself as the obvious translation or would
other solutions spring to mind with equal force? The situation is similar
with regard to arguments: here, we need to find an appropriate answer
to the question, What counts as an argument? Should we accept as an
argument the mere specification of an error as, say, a vocabulary error?
Moreover, much of what constitutes an implicit argument depends on
our handling of the binary/non-binary dichotomy. In the discussion be-
low, we will try to shed some light on these issues. The results will be
presented following the structure of the methodology.
Formal Characteristics
An easy topic requiring little discussion, the formal characteristics of a
bachelor’s thesis are readily dealt with in the introductory paragraph of the
examiner’s report. Of the 82 reports analysed, only four do not make any
reference to aspects of form. Most of the others comment on the appropriate-
ness of the bibliography and on any mistakes made outside the translation.
None of the response categories to question 1 (What formal characteristics
of the thesis does the evaluator cover in his or her evaluation?) goes without
comment. While the title page is referred to only in one report, the eighth
category (“others”) turns out to be essential as examiners complain about
missing page or line numbers and praise appendices featuring informative
charts or interesting e-mail correspondence – to give just a few examples.
All in all, the formal characteristics of a commented-translation bachelor’s
thesis play but a marginal role in the examiners’ reports.
The formal characteristics of the reports themselves yield a mixed
picture. Given our distinction of an overall structure that is weak or
strong or neither weak nor strong, we find that the majority (54) feature
a strong structure, whereas only 12 show a weak structure. These fig-
ures, however, are clearly influenced by the respective preferences of the
examiners as one examiner wrote 43 of the strong-structure reports. A
similarly strong dependence on examiner preference is reflected in the
length of the reports, which ranges from less than one page to almost
four pages. While examiner 2, for example, always writes about one
page, the reports by examiner 6 comprise between little less than two
pages and almost three pages.
This result partly anticipates the answer to the first hypothesis (The
length of an examiner’s report is in inverse proportion to the quality
of the commented translation), when we consider that the BA theses
evaluated by examiner 2 achieve the whole gamut of grades from 1.0 to
5.0.7 In other words, the length of an examiner’s report does not nec-
essarily reflect the quality of the commented translation – what counts
are the examiner’s preferences as to how much he or she wishes to
write. However, this does not completely refute the first hypothesis.
Looking at reports written by the other examiners, we find that there
is at least a tentative correlation between length and grades: longer
reports often go hand in hand with lower grades. In those two cases
in which a failed bachelor’s thesis is assessed on the basis of a fairly
short report, the reason for the fail lies in the unsatisfactory analysis
of the source text rather than in a bad translation. On the contrary: in
the two reports, the examiners refer to the translation as very accurate
and felicitous, respectively. Figure 6.1 shows the correlation between
the overall grades given for the theses and the length of the reports. It
also reveals a general preference for shorter reports of between one and
two pages.
While 21 reports are up to one page long, 46 are between one and
two, 13 between two and three, and 2 between three and four pages
long. There is a general tendency for lower grades to result in longer re-
ports. Having dealt with the formal characteristics of the theses as they
are discussed in the examiners’ reports and with the formal aspects of
the reports themselves, we will now turn to our central issue: translation
quality assessment.
5.0
4.5
4.0
3.5
grade of thesis
3.0
2.5
2.0
1.5
1.0
0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0
length of report (number of pages)
Figure 6.1 Correlation between grades of theses and lengths of reports.

Aspects of Translation Quality Assessment
Bluntly, the examiners’ reports do not quite meet the requirements of
good translation quality assessment outlined in Chapter 5. This con-
cerns the comparison of source and target, the handling of translation
errors and achievements as well as the argumentative support of indi-
vidual criticism. Of course, the analysis of different reports by different
examiners assessing different translations under different circumstances
will inevitably produce a multifaceted picture of different outcomes. In
this section, we will try to do justice to the many aspects involved in
writing an examiner’s report.
Source and Target

The question as to what hints the examiners’ reports provide with regard
to the respective evaluation procedure can, in most cases, be answered
unequivocally: they compare source and target. In the report, this shows
in the obvious juxtaposition of source translation units and the corre-
sponding target translation units, which are then discussed with regard
to their relative appropriateness. The vast majority of these translation
units consist of an individual word or a short phrase, whereas sentences
or even paragraphs are hardly ever quoted (with one notable exception).
The comparison of source and target is evident in 69 reports, the oth-
ers either imply such comparison through general remarks, for example,
about the translator’s understanding of the source text, or show no evi-
dence of source–target comparison.
No evidence of source–target comparison, however, does not neces-
sarily mean that source and target have not been compared. They prob-
ably have – but to what extent and how carefully cannot be established
even by the closest report analysis. In this context, we should note that
proof of what constitutes the basis of the evaluator’s judgement may
take a variety of forms apart from the mere juxtaposition of source and
target: most essential is a reference to the thesis that has been examined.
While such a reference is usually made in addition to the comparison
between the original and the translation, it sometimes comes without a
quotation from the source text (even if the target text unit discussed is
not in itself obviously erroneous), and sometimes it is used on its own
just with an explanation of what is wrong. This highlights an essential
problem of the present study: in those cases where a reference to the BA
thesis is the mainstay of the examiner’s argument (with insufficient or
no quoting from source and target), the report would have to be read
together with the thesis. As the theses to the reports do not figure in
our study, any doubtful issues may not be fully clarified; however, we
have tried to access the original texts on the Internet whenever that was
possible, so that too short an example given in the report can in many
cases still be assessed in a wider context. While the above problem plays
but a marginal role in the question of source and target comparison, it
features more prominently in the next sections about errors and achieve-
ments as well as evidence and arguments.
Errors and Achievements

Five questions need to be answered under the heading of errors and
achievements: How does the evaluator assess the quality of the transla-
tion? (question 4); Have the errors and/or achievements been classified
and, if so, how? (question 5); Does the evaluator count errors and/or
achievements? (question 6); Have the errors and/or achievements been
weighted and, if so, how? (question 7); and Which evaluation perspec-
tive prevails: the professional reader’s or the student translator’s per-
spective? (question 8). We will look at these questions, in turn.
The answer to question 4 is not surprising: the vast majority of ex-
aminers’ reports (64) focus on errors when evaluating the translation.
In these cases, a good translation is unmarked, that is, characterised by
a zero value that indicates the absence of errors and mistakes (cf. Emsel
2007:100). What detracts from an acceptable translation standard for
a given source text lends itself more readily to objective criticism than
what contributes to it (cf. Greiner 2004:115). Still, there is some recogni-
tion of good translation solutions: seven reports feature both errors and
achievements. These are neglected in the remaining 11 reports, which
assess the target text in an impressionistic way. None of the reports
focusses only on achievements – probably, because a corresponding
evaluation procedure would require more effort than an error-oriented
approach.
Classification of errors can be found in 52 reports. However, such
classification does not consist in a predefined set of error categories but
in grouping the errors found according to different types of errors. In
many cases, errors are classified as a result of being defined and this
definition often doubles as an argument why a particular translation
solution is regarded as an error. In other words, stating what is wrong
with a given target text unit provides the error’s definition, implies its
classification, and suggests the (rather superficial) reason why it is re-
garded as an error. For example, one report distinguishes between er-
rors of expression, vocabulary errors, too literal and too free renderings,
quoting several target text units for each type of error. In about two out
of three cases, the translation is furnished together with the correspond-
ing source text unit. That the four categories are primarily intended to
classify the errors in the target text is obvious from the fact that the
examples are grouped in line with these categories. Another report, by
contrast, gives a rather general characterisation of the contexts in which
the translator tended to make a mistake and then lists the individual
errors together with very little explanatory information. A partial error
classification can be found in yet another report: here, some errors are
referred to individually together with a brief explanation while others
are listed as examples of a general type of error such as word formation
and collocations untypical of German language patterns.
The question whether the examiner counts errors and/or achievements
in the target text is answered in the affirmative for 44 reports with regard
to errors. Achievements are not counted. It should also be noted that the
44 reports were prepared by the same examiner. In all 44 reports, the
error count distinguishes between errors made in the translation and
errors made in the remaining text (that is, the source text analysis and
the translation commentary). The reports also mention the number of
characters including spaces (and sometimes also the word count) of the
source text as well as, occasionally, the number of content words that
have been omitted in translation without justification. In some cases, the
analytical part of the thesis and the translation are graded separately.
All errors are counted in terms of error points, which suggests that the
errors have been weighted. Unfortunately, the reports do not specify
any error-weighting criteria. Thus, the first part of question 7 (Have
the errors and/or achievements been weighted and, if so, how?) can be
answered in the affirmative for the 44 reports in which the errors are
counted, whereas the second part remains unanswered.8
Upon taking a closer look at the error counts, we find that their con-
tribution towards a more objective assessment of translation quality (as
argued in Chapter 3) is not quite clear. As a general rule, we would
expect a low error count for the translation to go hand in hand with a
good grade for the thesis. However, since the translation accounts only
for 50 percent of the whole thesis, the error count for the translation
does not necessarily reflect the overall grade. If possible, we will there-
fore refer to the grades for the translation (rather than for the source
text analysis and the commentary) when analysing the relation between
error count and grade. To render the error counts comparable, we have
to relate them to the respective character counts for the source texts,
calculating the number of errors in the translation for 1,000 characters
including spaces. This figure ranges from 1.5 to 12.1 for all 44 reports
featuring an error count. Considering only the six reports that specify
the grade for the translation (either explicitly or implicitly), we get the
results shown in Table 6.3.
Although the results show a general correspondence between the rel-
ative error count and the grade for the translation, there does not seem
to be one clear underlying scheme that assigns a particular grade to a
given number of errors per 1,000 characters including spaces. Nota-
bly, the translation assessed in example 6 might have been graded too
favourably, considering that the deduced grade could have been even
better and the relative error count is significantly higher than that of
Table 6.3 Comparison of grades and error counts for translations
Example Errors Grade for Explicit or implicit reference

no. per 1,000 translation
characters
including
spaces
1 1.9 1.3 Implicit: the examiner states that the

translation gives a very good overall
impression; however, the co-examiner
specifies the grade for the translation
as 2.0 in a separate report.
2 3.9 2.0 Explicit: the overall grade being 1.7 due
to a very good analysis (1.3).
to a less satisfactory analysis (3.3).
to a less satisfactory analysis (3.7).
to an unsatisfactory analysis.
6 12.1 3.0 Implicit: the examiner remarks that the
considerable defects of the analysis are
largely responsible for the overall grade
of 3.3, which implies a grade for the
translation of 3.0 or better.
the other translation with the same grade (example 5). The reports that
do not grade the translation separately likewise suggest that the error
count has not been employed in a consistent fashion. While 4.2 errors
for 1,000 characters including spaces can fetch a 1.3, a thesis featuring
a translation with an error count of only 2.9 is graded 1.7. This may
partly be due to a different assessment of the source text analysis and
the translation commentary. Another aspect that possibly prevents the
examiner from applying his or her grading scheme consistently to all
translations is the influence of the co-examiner. Thus, the counting of
errors helps to control the assessment of the translation only to a lim-
ited degree.
In connection with weighting errors, the examiner may have to de-
cide whether to take the professional reader’s or the student translator’s
perspective (question 8). Such a decision need not be made explicitly: it
can also be implied in the way a particular error is described and counts
towards the overall result of the translation quality assessment. To give a
brief example: “41,000 lbs of high-grade zinc” were rendered as “41,000
lbs (etwa 18 kg) hochwertigen Zinks” where the correct amount would
be 18,000 kg (in German: 18.000 kg). While such an error could be
regarded as potentially very serious from a professional reader’s perspec-
tive (if there is a risk of the error affecting real-life situations), the grade
(1.7) for the thesis shows that, here, probably the student translator’s
perspective prevailed. In another example, taken from a different report,
the examiner criticises as hardly comprehensible (“kaum verständlich”)
a rendering of the noun phrase “orbiting rock” as “kreisende Feldblock”,
emphasising that “Feldblock” is not a German word. While such criti-
cism is justified from the professional reader’s perspective, a student’s
perspective would recognise here a simple spelling mistake where “Feld-
block” should have been “Felsblock”. The last example concerns the
translation of the following sentence: “The bottom line […] is this: par-
ents matter and the more research psychologists and neuroscientists
do, the more we realise that they matter even more than we knew […]”
(Moorhead 2013, no pagination). In the report, the examiner finds fault
with the translation of “the more research psychologists […] do” as “je
mehr Psychologen […] Forschung betreiben”, recommending “je mehr
Forschung Psychologen […] betreiben” as the better alternative. The er-
ror is rightly classified as an error of expression rather than a more seri-
ous translation error, because the translator’s solution is not necessarily
wrong but ambiguous. Depending on whether “mehr” is regarded as
an adverb or quantifier, the rendering of the clause is right or wrong,
respectively. It is because the wrong reading is the more immediate read-
ing that the translator’s solution is criticised as inadequate. However, the
possibility of a correct interpretation (which, in the given context, is not
unlikely) justifies a lenient handling of the error.
To conclude the discussion of errors and achievements, we will
briefly summarise the results. As expected, most reports focus on er-
rors, with achievements playing a marginal role. The errors are often
classified in that they are defined and grouped in line with the defini-
tion. In the reports of one examiner, the errors are usually counted and
implicitly weighted. Such counting and weighting affects the grade for
the translation only in a limited way because the examiner’s judgement
is complemented by the judgement of the co-examiner. Regarding the
two evaluation p erspectives – the professional reader’s and the student
translator’s – n
either perspective is referred to explicitly in the reports.
Which perspective prevails for a given error assessment can occasionally
be established on the basis of that error’s context and the examiner’s
handling of the error. To what extent reference to errors and achieve-
ments is backed by corresponding arguments will be discussed in the
next section.
Evidence and Arguments

The first question to be answered with regard to evidence and arguments
(question 9) relates to a proper understanding of why the translator trans-
lated as he or she did: Does the evaluator discuss the translator’s inten-
tion and, if so, how? To come straight to the point, in 46 reports, the
evaluator does not discuss the translator’s intention. In the remaining 36
reports, the evaluator refers to aspects to do with the translation strat-
egy favoured by the translator. We distinguish five different approaches:
(1) quoting the translation strategy, (2) quoting the translator’s reasons
for a translation decision, (3) reconstructing the translation strategy, (4)
reconstructing the translator’s choice, and (5) criticising disregard of a
strategy. Of the reports that discuss this issue, 30 do so in such a way
that only one category applies, while six reports bear upon two catego-
ries in their discussion of the translator’s intention.
Quoting the preferred strategy of the translator can be direct or
indirect – either reproducing or mentioning what the translator writes
about his or her strategic approach. For example, in one report, the ex-
aminer uses a direct quote from the thesis to show that the translator
does not stick to her own precepts when she states that units have to be
translated into the metric system and that a space is required between
a figure and the corresponding measurement unit. Another report, by
contrast, quotes only indirectly from the thesis: the translator, it says,
intends to write the target text in such a way that the readers do not no-
tice that they are reading a translation. The implied translation strategy,
here, is a covert strategy. Direct and indirect reference to the transla-
tor’s intentions can also be found for the second approach (quoting the
translator’s reasons). We find, for example, a quotation concerning the
translation of direct quotes through paraphrasing. In another report, on
the other hand, there is but a general reference to an explanation given
by the translator that one translation solution is preferred to another.
Regarding the third approach (reconstructing the translator’s strat-
egy), we come across several examples in which the examiner derives
from the translation the underlying strategy. For instance, the translator
is said to work with a clear focus on the target audience and, in an-
other report, the examiner notes that the translator aims to render the
casual tone of the original into the target language inasmuch as this is
acceptable in the context of the given target publication. The least com-
mon method of tracing the translator’s intention is the fourth approach
(reconstructing the translator’s choice). Here, we hit upon one example
only: the examiner observes that the translator obviously looked up the
phrase in question in a bilingual dictionary, accepting a direct render-
ing of “shed one’s skin” – a phrase used with reference to the animal
kingdom, that is, reptiles. Finally, the translator’s intention can also be
referred to by criticising the translator’s disregard of a strategy. In one
report, the examiner states that the results of the analysis are not suffi-
ciently related to the translation and another report finds fault with the
translator’s neglect of legal differences between the source and target
countries in a text about children and divorce. The above examples are
evidence of the various approaches by which the examiner tries to come
to terms with the intention of the translator.
Following the requirement that the evaluator of a translation respect
the translator’s preferences, we will now look at the requirement that the
evaluator be able to justify any changes he or she would want to make
to a translation. Such need for justification calls for well-founded argu-
ments to support any criticism of translation solutions. While primarily
serving to substantiate negative comments, arguments should also be
used to underpin any reference made to successful solutions. Questions
10a and 10b cover a whole range of issues to do with the argumentative
framework designed to justify the examiner’s criticism. Before giving
specific examples, we will discuss in Table 6.4 the figures for each of the
17 parameters analysed in connection with the questions: To what ex-
tent does the evaluator provide arguments as to why a translation solu-
tion is not satisfactory? and To what extent does the evaluator provide
arguments as to why a translation solution is successful?
Counting the total number of translation solutions referred to in each
examiner’s report is not as straightforward as it may seem at first. Gen-
eral remarks about the translator’s skill of rendering the source text into
the target language do not deal with actual translation solutions unless
they are reasonably specific and accompanied by a page or line reference.
The following quotation is a case in point:
[Die Übersetzerin] hat den Ausgangstext in einem adäquaten

Sprachstil übersetzt. Das Lesen des Textes wird sehr erleichtert,
weil sie häufig aus einem langen Satz im AT mehrere kurze deutsche
Sätze macht. Einziges Manko ist, dass sie häufig vor das deutsche
Genitiv-s (besonders S. 20) einen Apostroph setzt.
[The translator] has translated the source text in an adequate style.

Reading of the text is very much facilitated by her often rendering
one long sentence as several short sentences in German. What is un-
acceptable, though, is her frequent use of an apostrophe before the
German genitive -s (particularly on p. 20).
While the first two sentences refer very generally to the way in which the
translator translated the source text, the third describes a common error
and specifies the page number where to find it. Neither the adequate style
nor the splitting of sentences is regarded as a translation solution for our
purpose; the genitive -s error together with the page number, however,
does count as a translation solution.
Parameters 2 and 3 complement each other. They divide the results
for the first parameter into unsatisfactory and successful translation
solutions. As the answer to question 4 has shown, there is a clear pref-
erence for translation assessment based on errors. Achievements are
occasionally referred to, yet, only in addition to errors. Note that the
numbers of reports under parameters 2 and 3 do not quite match those
Table 6.4 Frequency of parameters for questions 10a and 10b
Parameter Frequency
1
Total number of translation solutions referred to in the The number of translation solutions referred to ranges from 0 to 51. Eight reports dispense with examples,
examiner’s report providing only an impressionistic assessment of the commented translation. On average, a report
discusses 14 translation solutions.
2 Number of unsatisfactory translation solutions referred Here, the figures are largely the same as for the first parameter, except for those eight reports that also
to in the examiner’s report feature successful translation solutions (see parameter 3).
3 Number of successful translation solutions referred to Only eight reports refer to successful translation solutions, with numbers ranging from 1 to 3.
in the examiner’s report
4 Total number of binary errors The number of binary errors ranges from 0 to 17, with 26 reports discussing no binary errors at all.
5 Number of binary errors with an additional explicit There is only one report that features binary errors, namely, two, supported by an explicit argument that
argument that is different from the argument implied is not identical to the argument implied.
6 Number of binary errors where the implicit argument While, in 32 reports, there are no binary errors where the implicit argument is made explicit, the
is made explicit remaining 50 reports have up to 17 of such binary errors.
7 Total number of non-binary errors The number of non-binary errors ranges from 0 to 39, with ten reports discussing no non-binary errors at
all.
8
Number of non-binary errors quoted with a context Here, 70 reports do not feature non-binary errors with a context sufficient to imply a corresponding
sufficient to imply the argument proving the point argument. Those that do, generally have between one and four contextualised non-binary errors, an
exception being one report with ten such errors. Note that where non-binary errors are supported by an
explicit argument, that argument takes precedence over the argument implied in any context.
9 Number of non-binary errors with explicit arguments While 14 reports do not refer to any non-binary errors supported by an explicit argument, the remaining
68 reports use explicit arguments for up to 49 non-binary errors.
10
Number of non-binary errors without arguments – Of those 72 reports that quote non-binary errors, only 18 feature non-binary errors without any
neither implicit nor explicit (calculating the difference argument: there are six of these errors in 1 report, three in another 1, two errors in 5 reports, and one
between 7 and 8 plus 9) in 11 reports.
11
Number of successful translation solutions with Four reports provide arguments for between one and three successful translation solutions. The total
arguments number of successful translation solutions with arguments is seven.
12
Number of successful translation solutions without Four reports feature one or two successful translation solutions without giving any supportive arguments.
arguments The total number of successful translation solutions without arguments is five.
Total number of arguments including multiple
13 Not counting the eight reports without examples, there are 74 reports with up to 50 arguments each – the
arguments for one translation solution average being 15 for those reports with examples.
14
Number of implicit arguments (adding the results of 4 The number of implicit arguments ranges from 0 to 12, with 63 reports featuring no implicit arguments.
and 8 and subtracting 6) While this may seem contradictory in view of the binary error count (because binary errors always
imply the reason why they are errors), the discrepancy is explained by the fact that implicit arguments
made explicit no longer count as implicit arguments.
15 Number of explicit arguments including multiple In 11 reports, there are no explicit arguments. In the remaining 71 reports, the number of explicit
arguments (calculating the difference between 13 and 14) arguments ranges from 1 to 50.
16 Number of relevant arguments (explicit and implicit) Not counting the 8 reports without examples, there are 74 reports with up to 48 relevant arguments each.
17 Number of irrelevant arguments (can only be explicit) There are 29 reports featuring between one and four irrelevant, or wrong, arguments.
for question 4, How does the evaluator assess the quality of the transla-
tion? For instance, 11 reports have been said to discuss the translation
in a general, impressionistic way, whereas eight reports do not feature
any examples. In those three reports that are referred to as impression-
istic although they contain at least one example, the overall approach
to translation assessment is dominated by general criticism. And while
seven reports are listed under “errors and achievements”, eight reports
contain examples of successful translation solutions. What is more, the
“errors and achievements” list includes one report without a successful
translation solution (here, the reference to achievements is only general),
whereas two reports with successful translation solutions are not catego-
rised under “errors and achievements” because the successful translation
solutions are presented not so much as achievements than as standard
renderings.9
Parameters 4–10 concern the numbers of binary and non-binary er-
rors with and without supporting arguments. The results are not as un-
equivocal as one would wish because the distinction between a binary
and a non-binary error involves subjective reasoning. For only if an error
is obvious and implies the correct alternative does the error count as a
binary error. While the first criterion is usually fairly straightforward,
the second often leaves room for manoeuvre: as soon as there are several
alternatives that might be used to correct the error, the error would have
to be classified as non-binary. Strictly speaking, an error could only be
binary if it were due to poor grammar or spelling. However, such an
approach would ignore the fact that, in the reports, a large number of
errors is apparently presented without an explicit argument because the
error itself and its correction are regarded as obvious enough not to re-
quire further substantiation. Against this background, we have decided
to count as binary errors all those obvious errors the correction of which
suggests one particular alternative as the best solution. In this way, the
number of non-binary errors presented without an argument (param-
eter 10) has been significantly reduced. To give an example: while the
expression “depicted offensively” is considered readily translatable as
“beleidigend dargestellt” (as suggested by the examiner), which is why
“offensiv dargestellt” is deemed a binary error; the rendering “Deutsches
Messing” for the noun phrase “German brass” is a non-binary error
because, in the military context in which it occurs, the possible correct
translation given by the examiner (“hohe Tiere in der deutschen Armee”)
is not necessarily the most immediate alternative – “hochrangige deut-
sche Militärs” constituting another, equally good solution.
Since binary errors imply both the reason why they are errors and
the correcting alternative, there is no need for the examiner to provide
any explicit arguments. If an explicit argument is given, then it is often
rather superficial, reflecting the argument that is already implied. We
found only two instances of a binary error where an explicit argument
is provided that is different from the error’s implicit argument. The fol-
lowing examples are taken from the same report. The examiner argues:
Schiefe Bilder wirken manchmal unfreiwillig komisch und entspre-

chen damit nicht genau dem Ausgangstext:
“Abgezwackter Strom schleicht durch das Dorf …” (S. 43)
“Die Wurzeln der Krise … seien viel tiefer verankert …” (S. 47)
Here, we regard the introduction to the examples as one argument: the

oblique images that sometimes have an involuntarily comical effect and,
as a result, do not correspond to the source text. The source text pas-
sages are not provided in the report but can be found online in Cooper
(2003, no pagination): “Hijacked electric power snakes through the vil-
lage”; “The crisis […] runs much deeper”. For the binary error in the sec-
ond example, the explicit argument of the introduction is different from
the one implied; for the non-binary error in the first example, the explicit
argument is the only argument. While the first example (“Abgezwackter
Strom schleicht durch das Dorf …”) does not suggest an obvious solu-
tion, especially, since the report fails to provide the source text unit, the
second example is in itself tautological: the roots of the crisis are rooted
much more deeply [.…]. This tautology makes for a binary error and
constitutes the main ingredient of an implicit argument.
There are no implicit arguments in connection with successful trans-
lation solutions as positive criticism is less in need of justification than
negative criticism. Parameters 11 and 12 show that the number of suc-
cessful solutions mentioned in the reports is significantly lower than
the number of errors and that of the eight reports which make specific
reference to achievements in the translation one half provides support-
ive arguments while the other does not. Thus, the evidence furnished
with regard to hypothesis 3 (The number of successful translation solu-
tions without arguments exceeds the number of successful translation
solutions with arguments) is inconclusive. There are, in fact, even more
successful translation solutions with than without arguments. In the
following, we will give two examples of reports that refer to success-
ful translation solutions. In the first, the examiner appreciates that the
translator has made an effort to maintain certain rhetorical devices in
the target text such as alliteration and parallel structures. We interpret
the description of the translation approach (that is, maintaining rhetor-
ical devices) as an explicit argument. In the second, by contrast, a mere
reference to a successful solution as being well translated or representing
an easy-to-understand rendering of the corresponding source text unit
does not provide the reason why the solution is regarded as successful.
Parameters 13–17 deal with argument numbers. That, for a given
report, the total number of arguments can be different from the total
number of translation solutions discussed is due to the fact that some
translation solutions may not be supported by an argument while others
may be supported by several arguments. Since the overall number of
translation solutions referred to in the reports (1,124) is higher than the
total number of arguments (1,102), there are more translation solutions
left without argumentative support than furnished with multiple argu-
ments. The following examples will demonstrate how one error can be
explained using at least two arguments. In one report, the examiner
criticises the translation of “in the near term” as “in naher Zukunft”, ar-
guing that the expression does not refer to a future period from the time
of writing but, in the context in which it is used, implies a short-term
investment. What is more, the phrase “near term” also appears in the
section heading, which is why this reference should also be retained in
translation. Thus, there are two independent arguments why the trans-
lation solution should be regarded as an error. Not quite as independent
are the two arguments presented in another report: rendering “until a
year earlier” as “bis ein Jahr zuvor” is, on one page, criticised as being
grammatically incorrect and four paragraphs further on, on the next
page, as too literal. Whether this double mentioning of the same error is
due to input from the co-examiner cannot be established with certainty.
Though put forward independently, the two arguments are evidently
connected: too literal a translation can be said to result in grammatical
incorrectness. A line of argument in the sense that one argument is sup-
ported by another is rarely found in the corpus. One such example is the
following quotation:
Insgesamt ist es der Verfasserin nicht gelungen, die journalistiche

[sic] „Farbe“ des Ausgangstextes auf Deutsch wiederzugeben. Viele
Ausdrücke werden neutralisiert (z. B. „knobbly arms of plane trees“
> „astreichen Platanen“; „hunter green“ > „dunkelgrün“; „rejigger-
ing the food scene“ > „die Essenskultur neu definiert“; „sprucing
up“ > „modernisiert“; „storied“ > „bekannt“). Die Übersetzung als
Reisebericht wirkt fad und einfallslos.
On the whole, the author failed to reproduce in German the journalistic

“colour” of the source text. Many expressions have been neutralised
(e.g. “knobbly arms of plane trees” > “astreichen Platanen”; “hunter
green” > “dunkelgrün”; “rejiggering the food scene” > “die Essenskultur
neu definiert”; “sprucing up” > “modernisiert”; “storied” > “bekannt”).
As a travel report, the translation is rather unimaginative and boring.
Here, the first argument that many expressions have been neutralised
could be attacked on the grounds that the covert rendering called for
by the overall translation situation leaves room for a translation which
does not imitate the original in every detail. While this is certainly true,
a closer look at the type of text and its implied purpose makes it clear
that some stylistic elements of the source text should be preserved in
translation – notably those pointed out by the examiner in the above
quotation. The second argument, then, consists in the explicit remark
that the text is a travel report, implying that such texts need to be writ-
ten in a captivating style. It supports the first argument.
Parameters 14 and 15 are relevant to the second hypothesis (The num-
ber of explicit arguments exceeds the number of implicit arguments).
With more than 1,000 explicit arguments and less than 60 implicit argu-
ments, the figures clearly confirm the hypothesis.10 This stark difference
reflects a general preference for non-binary errors (of which we count a
total of 855) as opposed to binary errors (totalling 257) and is also due
to the fact that the implicit argument of a large number of binary errors
is made explicit, so that it no longer counts as implicit. An even stronger
(but expected) contrast shows when we compare the number of relevant
(1064) and the number of irrelevant (38) arguments (parameters 16 and
17). While the relevant arguments constitute the norm, the irrelevant
arguments require some explanation as to why we regard them as irrel-
evant. Basically, there are three kinds of irrelevant arguments: (1) those
reflecting an insufficient involvement of the examiner with the transla-
tion situation (which would call for more detailed research or a closer
analysis of source or target language issues); (2) those in which the error
cannot be recognised as such; and (3) those that are simply wrong. They
all deal with unsatisfactory translation solutions that are, in fact, satis-
factory or might possibly be regarded as such. For instance, when the
examiner in one report claims that “improbable” had been translated
too literally as “Bande zu knüpfen”, then this is an error that is totally
incomprehensible and beyond justification: whether the translation is ac-
tually wrong cannot be established on the basis of the argument in the
report.
Our discussion of irrelevant arguments directly takes us to question
11 (Does the evaluator judge as unsatisfactory translation solutions that
are, in fact, satisfactory and, if so, in which way?). For an argument is
often irrelevant because the examiner spots an error where there is none.
As has been pointed out in the previous paragraph, this is probably
due to an insufficient involvement of the examiner with the translation
situation: given that the translator usually took great pains to analyse
the source text in its various dimensions (cf., for instance, Adamzik
2010:283) and to weigh the target language options accordingly, the
examiner may easily overestimate his or her knowledge of the various
circumstances surrounding a particular translation solution. To shed
some light on the way in which evaluators may occasionally misjudge
the translations they are supposed to evaluate, we will look at some
examples pertaining to this issue. Apart from the 53 reports for which
question 11 is answered in the negative, the remaining 29 reports11 can
be put in at least one of four categories, which will be repeated here for
convenience:
1 A satisfactory solution is criticised as unsatisfactory without argu-

mentative support.
2 A satisfactory solution is criticised as unsatisfactory by providing an
irrelevant or even wrong argument.
that is equally satisfactory.
that is wrong.
There is only one report that criticises a satisfactory solution as unsat-

isfactory without giving any reasons. Stating that some terms caused
difficulties for the translator, the examiner comes up with a list of
seven source expressions and their translations. One of these is “beach
resorts – Badeorte”. Since a “Badeort” is referred to as a tourist resort
on the coast even by the Duden,12 the examiner should have accepted
the translation as correct. Note that in categorising the above example
as being without argumentative support, we make a distinction between
the explicit classification of an error and the mere mentioning of an error
that only vaguely implies what one might consider to be a classification
of the error. While many of the errors examined in this study are sup-
ported by an argument that derives from a classification of the error (for
example, as an error of expression or a translation error), the mere state-
ment in the report referred to above, that some terms caused difficulties
for the translator does not constitute a classification and, thus, fails to
qualify as an argument. The line between argument and no argument is
rather thin, but it needs to be drawn.
Regarding the second category, we count 18 reports in which a satis-
factory solution is criticised as unsatisfactory by providing an irrelevant
or even wrong argument. The examples in Table 6.5 give an idea of the
different ways in which examiners misjudge the translations they are
supposed to evaluate. Next, we will consider those cases (relating to the
third category) in which a satisfactory solution is “corrected” by provid-
ing another solution that is equally satisfactory.
Having in mind a ready-made solution for a particular translation prob-
lem, an examiner may easily be tempted to push his or her solution at the
expense of the translator’s and neglect the qualitative potential of the target
text. Seven reports feature in this third category. Table 6.6 presents five
examples. These examples illustrate the examiner’s occasional neglect of
a fundamental TQA principle, namely, that “the critic should clearly dis-
tinguish” between “his own critical standards [and] the norms adopted
by the translator” (Broeck 1985:56).
Table 6.5 Criticism of satisfactory solution supported by irrelevant argument
Quotation Remarks
Es sind nur wenige Vokabelfehler zu Interestingly, Merriam-Webster online defines “grocery store” as “a store that sells food and household
monieren: […] grocery store als supplies: supermarket”. On the basis of this definition and for lack of a better alternative, the translator’s
„Supermarkt“ […]. choice seems perfectly appropriate.
Einige ihrer Ausdrücke erscheinen The implied criticism, here, is that neologisms (“Neuschöpfungen”) should not be used. However, as any
mir Neuschöpfungen zu sein Internet search will show, the three compounds are anything but neologisms, being used by online
wie […] „Zigarettenfälschung“ publications such as spiegel.de or welt.de.
(S. 25), „Gesetzesvollzug“ (S. 28),
„Gesetzesvollstrecker“ (S. 32), […].
Vermutlich ist im AT (Z. 165) ein Here, the examiner’s assumption that the source text should read “weight” instead of “height” is wrong. The
Tippfehler, den [die Übersetzerin] whole paragraph, it is true, deals with weight; however, in that paragraph, “height” – mentioned in the first
nicht erkannt hat: height anstatt sentence – has a contrasting function: “The ramifications of misperceiving a child’s height are fairly benign,
weight – im ganzen Abschnitt dreht but that’s not the case for the equally common misperception that children who are overweight have no
es sich um das Gewicht. such problem” (Friedman 2014, no pagination).
Allerdings wäre es bei einigen While the point made by the examiner is generally justified (namely, that Anglicisms should be translated
Anglizismen durchaus doch or at least complemented by an explanation), being absolutely correct for the two other examples given in
sinnvoll gewesen, diese zu the report; it is rather inappropriate for the translation of “brownout” as “Brownout” because the context
übersetzen oder – falls es sich um in which the term occurs contains the explanation called for by the examiner: “The less severe and more
ein Kulturspezifikum handelt – common form is a fragmentary blackout, or ‘brownout’, which is like a light flickering on and off in the
diese um einen erklärenden brain. Perhaps you remember ordering your drink, but not walking to the bar. Perhaps you remember
Einschub zu ergänzen (z. B. […] kissing that guy, but not who made the first move” (Hepola 2013, no pagination). We assume that the target
„Brownout“ […]). text retains the explanatory context. It certainly does so for the last two sentences, which were quoted by
the examiner in connection with the rendering of parallel structures.
Über Bewährung entscheiden in The source context is: “parole boards are making exactly these kind of predictive decisions every day about
Deutschland Gerichte und keine which prisoner or young offender we are going to release early” (Adams 2013, no pagination). This
„Ausschüsse“ (S. 38 Z. 269). statement is made by Adrian Raine, a neurocriminologist living in the USA, and it refers to the United
States. Regardless of whether the translator’s rendering is appropriate or inappropriate, the examiner’s
argument is irrelevant because the text is not about Germany and an adaptation to a German context in the
translation very unlikely.
Table 6.6 Criticism of satisfactory solution with acceptable alternative
Quotation Remarks
Es sind nur wenige Since “vivid” can mean “producing distinct mental images” as well as “having the appearance of
Vokabelfehler zu monieren: vigorous life or freshness” (Merriam-Webster online) and since both “lebendig” and “bildhaft” are
[…] vivid narratives als possible collocations for “Erzählungen”, the translator’s solution should not be regarded as an error –
„bildhafte Erzählungen“ (S. most certainly not as a vocabulary error.
338 Z. 68) statt „lebendige
Erzählungen“
Es finden sich einzelne As in the previous example, while the examiner provides a good alternative, the translator’s solution is
Vokabelfehler, etwa […] equally acceptable as “sauber” can also mean “allen Erfordernissen, den Erwartungen entsprechend,
neatly als „sauber“ (S. 22 in hohem Maße zufriedenstellend, einwandfrei” (Duden online). This is also true if we consider
Z. 56) anstatt „ordentlich“ the source context, which can be reconstructed from the line numbers given in the report: “All his
clothing was neatly folded and arranged” (Hannaford 2013, no pagination). Either of the likely
renderings by the translator – “sauber gefaltet” or “sauber zusammengelegt” – would be appropriate.
In der Übersetzung sind […] The examiner’s alternative solution is acceptable, but so is the translator’s rendering. In fact, “Eiweiß”
etliche Vokabelfehler zu is the standard German term for “egg white” (“Eiklar” being referred to as Austrian usage by the
monieren: „egg white“ als Duden); the only disadvantage is that it is potentially ambiguous, meaning “protein” as well. Yet, the
„Eiweiß“ anstatt „Eiklar“ context listing other food does not leave any scope for ambiguity.
[…].
Vokabel- und The context is: “Some of the divided families he interviews agree to appear together on camera, mulling
Übersetzungsfehler: […] over what it was that led to the divorce, and how each of them felt as events unfolded” (Moorhead
divided als „getrennt“ (Z. 2013, no pagination). As with the “Eiweiß” example, above, the term used by the translator is at least
310) anstatt „geteilt“ […]. as acceptable as the alternative suggested by the examiner – “getrennte Familien” being even more
common than “geteilte Familien”.
Sie will eine Wiederholung What kind of repetition the examiner refers to is not clear. If the corresponding translator’s argument
vermeiden (S. 28 Fn. regarding the style of the target text is valid (that is, avoiding repetition), then it is more convincing
15) und übersetzt anger than the examiner’s argument, because the general strategy behind the translation of the source text
and rage nur mit „Wut“, should be covert, given its largely informative function. A covert rendering requires that the target
obwohl sich doch „Wut text be optimised to meet the expectations of the readers. Here, the question whether the combination
und Zorn“ anbieten und of the two words in the source (which might be regarded as a hendiadys) should be retained in the
texttreuer wären. target text is rather unimportant: the examiner’s argument in favour of a more loyal rendering is not
supported by the covert translation strategy.
Table 6.7 Criticism of satisfactory solution with a wrong alternative
Quotation Remarks
Weitere Übersetzungsfehler: The term in question occurs in the following context: “If you don’t have long-term goals, Markman warns,
[…] catching up on TV als you run the risk of doing lots of little things every day – cleaning the house, sending emails, catching up
„sich im Fernsehen über on TV – without ever making a contribution to your future” (Webber 2014, no pagination). It is obvious
die neuesten Ereignisse that, here, the phrase in question should have the meaning “to learn about (recent events)” (Merriam-
informieren“ (Z. 67f) Webster online) – which corresponds to the translator’s German rendering. The examiner’s tentatively
anstatt „im Fernsehen nichts suggested alternative is inappropriate because the context calls for a verb that denotes action. However,
verpassen“ o.ä. […]. “not to miss out on something on TV” (“im Fernsehen nichts verpassen”) does not denote such action.
Mir ist nur ein Vokabelfehler The vocabulary error is no such error, not just because “fishy” can mean both “of fish” and “questionable”
aufgefallen […]: fishy stories (Merriam-Webster online), but because the context of the source text and, particularly, of the comic
als „Fischgeschichten“ (S. novel Three Men in a Boat (to Say Nothing of the Dog) referred to in the source (cf. Hammer 2012)
18) anstatt „verdächtige makes it clear that the first meaning is more likely to be relevant, here. A closer look at the relevant
Geschichten“. quotation in the novel would have settled the issue.
Es sind einige Vokabelfehler Since “corporate lawyer” can refer to “a lawyer who works for a corporation” or “a lawyer who specializes
zu monieren: […] corporate in corporate law” (Collins English Dictionary online), the context is decisive in finding the right
lawyer als „Anwalt für meaning. The two occurrences of “corporate lawyer” in the source text are: “David Wertime worked as a
Unternehmensrecht“ corporate lawyer in Hong Kong” and “After four years, Wertime finally realized he was never going to be
(S. 38 Z. 3, S. 40 Z. 1f) a truly great corporate lawyer, because he just didn’t care deeply enough about it” (McGowan 2013, no
statt „Justiziar“ oder pagination). Additional research shows that Wertime worked for law firms Cravath and Milbank – that
„Firmenanwalt“ […]. is, as a lawyer who specialises in corporate law – which is why the translator’s rendering is the correct
one.
In the fourth category, there are four reports that regard a satisfactory
translation solution as unsatisfactory and come up with a wrong alter-
native. In two of the three examples provided in Table 6.7, the examiner
failed to sufficiently research the backdrop to the translator’s solution.
These findings for question 11 give rise to the formulation of a golden
rule of translation quality assessment: never assume that the translator is
wrong unless you can prove it. Such proof would also include providing
a better alternative.
Alternative translation solutions are the subject of question 12 (To
what extent does the evaluator provide an alternative to a criticised
target text solution, and to what extent does he or she show that the
alternative is, indeed, better?). Here, we will look at the following five
categories:
1 number of unsatisfactory solutions for which no alternative is given;

is given;
3 number of alternative solutions provided (including multiple alterna-
tives for one translation error);
is given that is not supported by an argument;
is given that is supported by an argument (or arguments).
The number of unsatisfactory solutions for which no alternative is given

(category 1) ranges from 0 to 20 for one report, whereas the number of
unsatisfactory solutions for which at least one alternative is given (cate-
gory 2) ranges from 0 to 34. The total figures for the whole corpus are
484 for category 1 and 628 for category 2. This general tendency to
provide at least one alternative for an unsatisfactory translation solu-
tion also shows when we compare the figures for individual reports: 37
reports feature more unsatisfactory translation solutions for which at
least one alternative is given than unsatisfactory translation solutions
for which no alternative is given as opposed to 32 reports in which the
unsatisfactory translation solutions without any alternative prevail. Of
the remaining 13 reports, five have the same number of unsatisfactory
solutions for categories 1 and 2, while eight reports do not give any ex-
amples of unsatisfactory translation solutions.
Not surprisingly, the number of alternative solutions provided (cate-
gory 3), is not much higher than the number of unsatisfactory solutions
for which at least one alternative is given (category 2) – 687 vs. 628 –
because usually one alternative is sufficient to prove that a translation
solution is unsatisfactory. Still, there are some instances where the exam-
iner shows that the translator had several options to come up with a sat-
isfactory rendering of a given source text unit. Here is a good example:
the original “… worked with managers of the Mammoth Pacific power
plant complex (known als [sic] MPLP) in Mammoth Lakes, California
to remove silica …” is translated as “… mit den Betreibern des Mam-
moth Pacific Kraftwerks auch abgekürzt als MPLP, in der Stadt Mam-
moth Lakes, Kalifornien an der Extrahierung …”, which, in turn, is
corrected to “in der Stadt Mammoth Lakes (Kalifornien) oder … in der
Stadt Mammoth Lakes, Kalifornien, an der … oder aber … Mammoth
Lakes (USA) …”. The three alternatives are equally possible, indicat-
ing what a satisfactory rendering might look like. Occasionally, though,
mention of a second alternative suggests that the examiner was not fully
satisfied with the quality of the first solution. Thus, “Romanfigur” com-
plements “Hauptperson” as a possible translation solution for “char-
acter” (which has been translated as “Charakter” in the target text),
because “Romanfigur” seems more appropriate (given a likely context
that is not provided in the report). In another example, the second al-
ternative should even replace the first, because the first alternative is not
really an alternative: “This breeds circular thinking” (Friedman 2014,
no pagination) can be made out as the context of a criticism that finds
fault with the rendering of “breeds” as “Arten” (a totally incomprehen-
sible translation, which would require some target context in the report)
and tentatively provides as alternatives first “brütet” and then “erzeugt”.
The latter solution seems perfect, the former unacceptable.
As for categories 4 and 5, we count a total of 91 unsatisfactory solu-
tions for which at least one alternative is given that is not supported by
an argument and 537 unsatisfactory solutions for which at least one
alternative is given that is supported by an argument. The high number
of alternatives supported by an argument includes implicit alternatives
supported by implicit arguments (in the case of binary errors) and many
alternatives where the error argument doubles as an argument for the al-
ternative solution. Admittedly, the majority of these arguments are rather
weak, and the criterion when to accept an error argument as an argu-
ment supporting the alternative solution is not always unequivocal. As
a general rule, a translation alternative and the error argument can only
imply the argument for the alternative if the source text is given – unless
the error is recognised as an error independently of the source text. Crit-
icism such as “Tim Cooks angebliche Genialität liegt nicht in der „Bedi-
enung“ (Z. 390), sondern eher in der Strategie” – evidently referring to a
rendering of “operations” in “Tim Cook’s genius at operations” (Arthur
et al. 2014, no pagination) – provides an alternative not supported by
any argument. When an argument is provided, it is typically identical
to the error argument: in that it states why the translator’s solution is
regarded as an error, the corresponding argument anticipates why the
alternative represents a better solution. When one examiner criticises as
plain wrong a translation of “Stalin cut his political teeth fighting the
oil magnates of Azerbaijan” (Worth 2014, no pagination) as “Stalin hat
seine politischen Zähne … ausgebissen”, adding in parentheses “… hat
seine politischen Sporen verdient”, then this addition is obviously an
alternative intended to correct the wrong solution of the translator.13 By
implication, the error argument, which attacks the translator’s solution,
supports the alternative provided by the examiner. There are not many
cases in which the argument supporting the alternative solution is stated
explicitly. In one report, the examiner finds fault with “mit Menschen,
die gerade VoIP-Anrufe […] tätigten” as a translation for “with people
inside making voip phone calls” and proposes “Menschen, die gerade
per Internet telefonierten” as an alternative. Here, the error argument,
which reproaches the translator for rendering the text as in a technical
translation, is followed directly by the argument in support of the alter-
native, which indicates how a better solution could be found (namely, by
stepping back from the original and choosing a style typical of the target
publication). Reference to the latter argument is, then, made in the intro-
duction to the alternative solution: “im Stil hier besser” – “stylistically
better” – gives the reason why the examiner’s suggestion is preferable to
the translator’s rendering. This brings to an end our discussion of the
five categories under question 12.
Question 13 does not feature any categories. It deals with the trans-
lation arguments as such: What are the arguments put forward by the
evaluator? As has been mentioned already above, many arguments con-
sist in a brief definition and classification of a given error. For instance,
when an error is described as an error of expression or as an error of
translation, then that description does not only define and classify the er-
ror, it also states why a particular target text unit is regarded as an error:
an error of expression indicates that the translator opted for a solution
requiring some stylistic or terminological fine-tuning; an error of trans-
lation is probably more serious as it points to a problem with the trans-
fer of meaning. The arguments supplied are usually underdetermined
because, in order to make sense, they presuppose premisses, other argu-
ments, and intermediate conclusions. For example, when, in one report,
the examiner argues that the translation of “committed” as “begann” is
a vocabulary mistake, we do not just assume a suitable source and target
context but also accept that – whatever the correct solution might be – it
is supposed to correspond to the source in terms of lexis and/or meaning.
More generally, we take for granted that the respective source text unit
should be translated in line with the prevailing norms of the target cul-
ture. Keeping in mind this principle of underdetermination, we provide
in Table 6.8 an overview of the arguments derived from various error de-
scriptions. The list is not intended to be exhaustive; still, it reflects the
range of arguments based on error descriptions. It also reflects the gen-
eral predominance of text form factors (cf. Chapter 4) in translation
evaluation: except for the culture focus of the argument referred to as
Table 6.8 A rguments derived from error descriptions
Argument Error description English description
inadequate or unidiomatic target text “erhebliche Mängel […] bei der deutschen Formulierung” considerably flawed German wording
inadequate or unidiomatic target text “idiomatisch ungeschickte […] Formulierungen” idiomatically clumsy wording
inadequate or unidiomatic target text “massive Schwächen im Ausdruck” severe shortcomings in expression
inadequate or unidiomatic target text “Ausdrucksfehler” errors in expression
inadequate or unidiomatic target text “Raum für eine präzisere Wortsuche” room for a more precise word search
inadequate or unidiomatic target text (Anglicism) “Übernahme von Anglizismen” accepting Anglicisms
inadequate or unidiomatic target text (collocation) “Kollokation” collocation
inadequate or unidiomatic target text (contradiction) “Widersprüche” contradictions
inadequate or unidiomatic target text (inappropriate register) “Wörter aus einem unpassenden Register” words taken from an inappropriate register
inadequate or unidiomatic target text (style) “weniger elegant als das Original” less elegant than the original
inadequate or unidiomatic target text (style) “gestelzte Ausdrucksweise” stilted expression
inadequate or unidiomatic target text (style) “Stilblüten” bloomers
inadequate or unidiomatic target text (tautology) “Tautologien” tautologies
inadequate or unidiomatic target text (too free translation) “zu frei [übersetzt]” translated too freely
inadequate or unidiomatic target text (too literal translation) “stark zeichenorientierte Herangehensweise” a strongly sign-oriented approach
inadequate or unidiomatic target text (too literal translation) “Ausdrücke […] zu wörtlich übersetzt” expressions translated too literally
inadequate or unidiomatic target text (unwanted repetition) “Redewendung […] taucht […] 25x auf” phrase comes up 25 times
inadequate or unidiomatic target text (use of English-German “Verwendung von gemischt englisch-deutschen using mixed English-German expressions
compounds) Ausdrücken”
inconsistent target text terminology “Eine einheitlichere Übersetzung […] wäre übersichtlicher a more consistent translation would have
gewesen.” been clearer
inconsistent target text terminology “Inkonsistenzen” inconsistencies
incorrect grammar “Deutsch ist […] fehlerhaft” German is flawed
incorrect grammar (tense) “Schwäche […] im Gebrauch von Tempora” shortcomings as regards the use of tenses
insufficient target culture orientation “hätte […] versuchen sollen, japanische Begriffe […] zu should have tried to explain Japanese terms
erklären”
stylistic devices of the source text ignored in the target text “erkennt […] einige Stilmittel nicht” fails to recognise some stylistic devices
target text meaning different from source text meaning “Sinnfehler” semantic error
target text meaning different from source text meaning “erhebliche Mängel […] in der Wiedergabe das [sic] AT” considerably flawed rendering of the source
text meaning
target text meaning different from source text meaning “Vokabelfehler” vocabulary mistake
target text meaning different from source text meaning “gravierende Sinnverschiebungen” severe sense shifts
target text meaning different from source text meaning “Übersetzungsfehler” translation error
unjustified omission “Auslassung” omission
unjustified omission “wird hier ohne Kommentar weggelassen” has been omitted without comment
wrong reference “[…] ergibt sich aus dem Ko- und Kontext eindeutig, dass the co- and context clearly show that “they”
„they“ sich auf die Energiespeichersysteme bezieht” refers to the energy storage systems
wrong syntax “syntaktisch falsche […] Formulierungen” syntactically wrong expressions
“insufficient target culture orientation”, the arguments draw on text
form factors and, usually, they just state what is absolutely necessary.
Only occasionally do arguments give explanations that go beyond the
minimum requirement. Two examples will be discussed in the next
paragraphs.
The examiner’s argument is most convincing if it makes reference to
the translator’s strategy. A case in point is the following error discussion:
An einigen Stellen ist die Übersetzung nicht in sich konsistent. Auf

Seite 21 wird „536 hp“ (Z. 227 vom 1. Text) mit „536 PS“ übersetzt,
obwohl „horsepower“ sonst in Kilowatt umgerechnet wird, wie der
Autor im Kapitel 6 „Wichtige Übersetzungsentscheidungen“ (S. 30)
angibt vorzugehen.
In some places, the translation is inconsistent. On page 21, “536

hp” (l. 227 of the first text) is translated as “536 PS”, although
“horsepower” is usually converted into kilowatts, which is how –
in chapter 6, “Wichtige Übersetzungsentscheidungen” [important
translation decisions] (p. 30) – the author says he would proceed.
[my translation, H.B.]
Apart from the fact that the conversion referred to is rightly pointed out
as necessary in the context of a specialist journal such as “Automobil
Industrie”, the argument is significantly reinforced when the examiner
reveals that, here, the translator violates his own strategy. Such an argu-
ment would also retain its force if the target text were to be published in
a non-specialist magazine.
Besides bringing into play the translator’s strategy, the examiner may also
reinforce his or her argument by analysing why a particular error occurs:
„liquor and tobacco stores“ (AT1 Z. 107) wird mit „Likör- und
Tabakgeschäfte“ (ZT Z. 129) übersetzt; „liquor“ ist im Amerika
nischen ein Oberbegriff für sämtliche alkoholische Getränke; es
liegt ein klassischer false friend vor.
“liquor and tobacco stores” (l. 107 of source text 1) is translated as

“Likör- und Tabakgeschäfte” (target text l. 129); in American En-
glish, “liquor” is a generic term for alcoholic beverages in general;
this is a classic false friend.
Here, the argument that “liquor” refers to alcoholic beverages in gen-

eral draws further support from the explanation that the error is a clas-
sic false friend. In other words, the mere statement that the translation is
wrong because the source term in question has a different meaning is com-
plemented by the assumption that the error is due to the similarity between
English “liquor” and German “Likör”. While the false-friend explanation
cannot be expanded into a supporting argument in the Betzian sense (cf.
Chapter 5), it certainly renders the actual error argument more explicit,
providing the reason why the translator translated as she did. With this we
finish the discussion of question 13 and, indeed, of all questions relating
to our analysis of translation evaluation reports in an academic context.
The results of this analysis will be briefly summarised in the next section.
Summary of Results
To summarise the findings of Chapter 6, we will take as a point of de-
parture the 13 guiding questions discussed above. This summary is de-
picted in Table 6.9. Finally, we will state whether the three hypotheses
made in the course of this chapter are true or false.
The first hypothesis cannot unequivocally be confirmed as true or false.
While, in general, examiners tend to write longer reports if the com-
mented translation is of a lower quality, the length of the reports is also
subject to the preferences of the individual examiner. What is more, the
fact that the reports do not just discuss the quality of the translation
but also assess the translator’s ability to analyse the source text further
complicates the relation referred to in the hypothesis.
implicit arguments.
The second hypothesis is confirmed as the reports feature more than
1,000 explicit arguments but less than 60 implicit arguments. However,
rather than serving, first and foremost, to substantiate the assumption of
a translation solution as erroneous, many arguments seem to be primar-
ily designed to define and classify errors.
arguments exceeds the number of successful translation solutions with
arguments.
The figures do not confirm the third hypothesis: of the 12 successful
translation solutions discussed, 7 come with an argument (or argu-
ments), whereas 5 do not. However, the result is not particularly con-
clusive since only eight reports discuss translation achievements – either
with arguments (four) or without (four).
The above results show to what extent the reports meet the TQA re-
quirements stipulated by translation scholars. These requirements have
Table 6.9 Summary of results with regard to guiding questions
Question Results
Question 1: The formal characteristics of the theses are often dealt with in the introduction of the report. The bibliography
What formal characteristics of the thesis does the and mistakes made outside the translation are referred to most frequently. The other characteristics do occur in
evaluator cover in his or her evaluation? the reports; yet, they are comparatively infrequent.
Question 2: Roughly two-thirds of the reports have a strong structure with clearly distinguished paragraphs or sections,
Does the report have a weak or a strong structure? whereas one-third tends towards larger units that integrate the discussion of several issues. The structure of a
report depends on the preferences of the examiner.
Question 3: Evidence of a comparison of source and target shows in 69 reports. The other reports either imply such
What hints does the examiner’s report provide with comparison through general remarks, for instance, about the translator’s understanding of the source text, or
regard to the evaluation procedure, that is, in provide no evidence of source–target comparison.
particular, with regard to the question whether
the source text and the target text have been
compared?
Question 4: More than three quarters of the reports (64) focus on errors. Of the remaining 18 reports, almost 40 percent
How does the evaluator assess the quality of the (seven) look at both errors and achievements, while the rest (11) evaluates the commented translation based on
translation? general impression.
Question 5: Errors have been classified in nearly two-thirds of the reports (52); achievements have not been classified.
Have the errors and/or achievements been classified Classification typically consists in error definition plus grouping of errors. Whether or not errors are classified
and, if so, how? is entirely up to the examiner.
Question 6: Errors are counted in 44 reports, achievements are not counted. All 44 reports were written by the same
Does the evaluator count errors and/or examiner.
achievements?
Question 7: The errors that have been counted have also been weighted. The weighting criteria, however, cannot be
Have the errors and/or achievements been weighted established.
and, if so, how?
Question 8: As most errors do not reveal a specific evaluation perspective, question 8 is usually not applicable. In those cases
Which evaluation perspective prevails: the where such a perspective can be made out, the examiner does not seem to be aware that he or she assumes a
professional reader’s or the student translator’s particular point of view.
perspective?
Question 9: For more than half of the reports (46), the answer to question 9 is “no”. Of those 36 reports for which the answer
Does the evaluator discuss the translator’s intention is “yes”, 18 quote the translation strategy, 12 criticise disregard of a strategy, 8 reconstruct the strategy, 3 quote
and, if so, how? the translator’s reasons, and 1 reconstructs the translator’s choice. As these figures amount to a total of 42,
there are 6 reports where two categories are applicable combining the third and fifth categories, the first and
fifth categories, and the first and third categories.
Question 10a: Here, we will present the total figures for the whole corpus of 82 reports. These figures reflect the overall evaluation
To what extent does the evaluator provide arguments focus. Thus, the fact that of 1,124 translation solutions referred to in the examiners’ reports only 12 are
as to why a translation solution is not satisfactory? successful translation solutions confirms the general error focus pointed out in the answer to question 4.
Question 10b: Of the 1,112 errors, more than three quarters, namely, 855, are non-binary errors, whereas 257 are binary.
To what extent does the evaluator provide arguments Although binary errors, by definition, come with an implicit argument as to why they are errors, in nine out of
as to why a translation solution is successful? ten cases (232) the implicit argument is made explicit. An additional explicit argument that is different from
the argument implied is provided only in two cases. Among the non-binary errors, less than four percent (32)
provide a context that is sufficient to imply the reason why the translation solution counts as an error. Most
non-binary errors (793) are supported by an explicit argument, whereas the number of non-binary errors
without arguments (30) is almost negligible. We conclude that an argument is usually provided for those errors
that need one.
For more than half (7) of the 12 successful translation solutions mentioned in the reports, the examiner provides
an argument why he or she thinks that the translation solution is successful. In the five remaining cases, the
examiner does not state why a particular translation solution is successful.
The total number of arguments (1,102) does not equal the number of translation solutions with arguments
(1,089), because some translation solutions feature multiple arguments. As we do not count implicit arguments
that have been made explicit, the number of implicit arguments (57) is significantly lower than the number
of binary errors (257). Most arguments are explicit arguments (1,045). More than 96 percent (1,064) of all
arguments provided are relevant in that they clearly show why an error is an error. Still, there are some (38)
arguments that are not convincing in this respect and, thus, irrelevant.
Question 11: While for 53 reports question 11 is answered in the negative, there are 30 instances in 29 reports where the
Does the evaluator judge as unsatisfactory evaluator judges as unsatisfactory a translation solution that is, in fact, satisfactory. In 18 cases, this is done
translation solutions that are, in fact, satisfactory by providing an irrelevant or even wrong argument. Seven unjustly criticised translation solutions come with
and, if so, in which way? a satisfactory alternative, four with an unsatisfactory alternative. There is only one example of a satisfactory
solution being criticised as unsatisfactory without argumentative support. Apparently the most common
reasons for such misjudgement on the part of the evaluator are, first, taking for granted what must not be taken
for granted and, second, underestimating the amount of research required to find the right translation solution.
Question 12: Some 56 percent (628) of all unsatisfactory translation solutions referred to in the reports are corrected by
To what extent does the evaluator provide an at least one alternative; approximately 44 percent (484) are not. The total number of alternative solutions
alternative to a criticised target text solution, provided (687) shows that, in some cases, multiple alternatives are used to illustrate the translator’s options.
and to what extent does he or she show that the Occasionally, though, a second alternative serves to conceal the inadequacy of the first. There are 91
alternative is, indeed, better? unsatisfactory translation solutions where the alternative is not supported by an argument, whereas 537
unsatisfactory translation solutions do get such support. This is because, in most cases, the argument that
attacks the error implicitly supports also the alternative.
Question 13: All arguments employed in the reports are informal arguments. They are underdetermined because they imply
What are the arguments put forward by the premisses, additional arguments, and intermediate conclusions. Many arguments are rather weak in the sense
evaluator? that they do not provide a link to the translation strategy (which is usually taken for granted). The evaluators
most commonly refer to inadequate or unidiomatic target text or find that the target text meaning is different
from the source text meaning. Complaints about incorrect grammar often render the implicit argument of binary
errors explicit. The case is similar with unjustified omissions. Arguments criticising a wrong reference or syntax
or inconsistent target text terminology typically call for a larger context. Other arguments attack inconsistent
target text terminology and failure to incorporate stylistic devices of the source text in the target text.
been discussed above. They are mentioned, here, again for the sake of
convenience:
1 An evaluator should try to understand why the translator translated

as he or she did.
2 An evaluator must be able to justify any changes he or she would
want to make to a translation.
3 The evaluator must provide a better solution and state why it is
better.
4 The evaluator’s arguments draw on factors that cannot be defined in
advance but must be established for each translation and translation
quality assessment on the basis of the source text and the overall
translation situation.
The first requirement is met by less than half of the reports. It shows
in the evaluator’s engagement with the translator’s strategy (or lack
thereof) and in his or her trying to find out – and mentioning – why the
translator translated as he or she did. The second requirement is about
errors, that is, those translation solutions that the evaluator would like
to change. Here, the reports usually provide an informal argument to
briefly explain why a solution is wrong. The argument often consists of
one word only and is derived from the definition or classification of the
error. As for the third requirement, more than half of the unsatisfactory
translation solutions discussed in the reports are corrected by the evalu-
ator, who provides an alternative solution. The vast majority of these al-
ternatives are supported by an argument that attacks the unsatisfactory
translation solution. The fourth requirement amounts to the question
what arguments the evaluators put forth to substantiate their judgement.
Here, we have found that the arguments are mostly derived from the
immediate translation situation. The evaluators do not apply an explicit
evaluation framework to the translation. Still, they sometimes neglect
aspects or factors that they should have taken into account – which, in a
few cases – leads to inappropriate criticism.
Notes
1 The range of issues that such a translation commentary may tackle has been
described in García Álvarez (2007): it covers any aspects relevant to the pro-
cess of rendering the source text from the source culture into a target text in
the target culture.
2 A bachelor’s thesis is always checked by two examiners. As a rule, the first
examiner writes the report – possibly with input from the second examiner.
Occasionally, the second examiner also contributes to the text of the report.
3 Cf. the relevant discussion in Chapter 3.
4 Note that, in principle, this also applies to translation solutions which the
evaluator regards as particularly successful.
5 As we have pointed out in Chapter 3, Pym’s distinction between binary and
non-binary errors is not as clear-cut as it would seem at first: an obvious
error that – in spite of its obviousness – permits of several corrections would
be regarded as binary for its obviousness and as non-binary because of the
correction alternatives. In such cases, we will opt for a non-binary error
unless there is one obvious solution that immediately suggests itself as the
correct translation.
6 Cf. also Peter J. Arthern’s essay “Judging the Quality of Revision” (Arthern
1983) briefly referred to at the beginning of Chapter 5.
7 The following grades are possible: 1.0, 1.3, 1.7, 2.0, 2.3, 2.7, 3.0, 3.3, 3.7,
4.0, and 5.0 – with 1.0 and 1.3 denoting a very good result, 1.7–2.3 a good
result, 2.7–3.3 a satisfactory result, 3.7 and 4.0 a sufficient result, and 5.0
an insufficient result. While a 4.0 is still a pass, a 5.0 is a fail.
8 As anticipated, there is also no evidence that examiners explicitly take into
account different levels of translation difficulty.
9 One report, for instance, criticises a totally inadequate translation of the
adjective “tough” in one example and then quotes a sentence in which the
rendering of the same adjective is satisfactory.
10 Although the figures appear to be unequivocal, the result should be taken
with a grain of salt as many explicit arguments are derived from the clas-
sification and definition of errors. Thus, it is doubtful whether the explicit
arguments serve, first of all, as a means to substantiate the less obvious
errors. Rather, the error definition or classification can also be used for ar-
gumentation purposes.
11 While it might seem no coincidence that there are also 29 reports that fea-
ture at least one irrelevant argument, the two sets of reports are not exactly
identical: on the one hand, the reports for which question 11 requires an
affirmative answer also include one report completely without irrelevant ar-
guments (but with two examples of unjustified criticism that is devoid of
any argument), and, on the other hand, the reports for which question 11 is
answered in the negative feature one report with an irrelevant argument in
connection with a successful translation solution.
12 “Fremdenverkehrsort an der Küste” (Duden online).
13 Interestingly, the examiner – in correcting the translator – fails to use a re-
flexive pronoun (Stalin hat sich seine politischen Sporen verdient), which is
also required in the translator’s rendering (Stalin hat sich seine politischen
Zähne ausgebissen).
References
Adams, Tim (2013). “How to spot a murderer’s brain”. In: The Guardian [on-
line]. Viewed: 27 August 2019. www.theguardian.com/science/2013/may/12/
how-to-spot-a-murderers-brain
Adamzik, Kirsten (2010). Sprache: Wege zum Verstehen. 3rd ed. Tübingen:
Francke.
Arthern, Peter J. (1983). “Judging the Quality of Revision”. In: Lebende
Sprachen, vol. 28, no. 2, 53–57.
Arthur, Charles; Jemima Kiss; Samuel Gibbs; Alex Hern; Siraj Datoo (2014).
“Technology predictions for 2014: smarter cities, bigger games”. In: The
Guardian [online]. Viewed: 27 August 2019. www.theguardian.com/
technology/2014/jan/02/technology-predictions-2014-smarter-cities-bigger-
games-drones-tablets
Cooper, Marc (2003). “Behind Globalization’s Glitz”. In: The Nation [online].
Viewed: 27 August 2019. www.thenation.com/article/behind-globalizations-
glitz/
Dollerup, Cay; Anne Loddegaard (eds.) (1992). Teaching Translation and Inter-
preting – Training, Talent and Experience. Papers from the First Language
International Conference, Elsinore, Denmark, 31 May – 2 June 1991 (Co-
penhagen Studies in Translation). Amsterdam/Philadelphia: John Benjamins.
Jüngst (eds.) (2007:100–111).
Friedman, Nick (2014). “Parents Just Don’t Understand”. In: Psychology Today
[online]. Viewed: 27 August 2019. www.psychologytoday.com/articles/2014
07/parents-just-dont-understand
García Álvarez, Ana María (2007). “Der translatorische Kommentar als Eval-
uationsmodell studentischer Übersetzungsprozesse”. In: Schmitt, Peter A.;
Greiner, Norbert (2004). Übersetzung und Literaturwissenschaft. Tübingen:
Gunter Narr. (vol. 2 of the series Grundlagen der Übersetzungsforschung).
Hammer, Joshua (2012). “The Long and Winding History of the Thames”. In:
Smithsonian Magazine [online]. Viewed: 27 August 2019. www.smithsonian
mag.com/history/the-long-and-winding-history-of-the-thames-139049
496/?all
Hannaford, Alex (2013). “A Man Was Murdered Here”. In: Texas Observer
[online]. Viewed: 27 August 2019. www.texasobserver.org/a-man-was-
murdered-here
Hepola, Sarah (2013). “My drinking years: ‘Everyone has blackouts, don’t they?’”.
In: The Guardian [online]. Viewed: 27 August 2019. www.theguardian.com/
society/2015/jun/13/my-drinking-years-everyone-has-blackouts-dont-they
Helm.
McGowan, Kathleen (2013). “Living a Lie”. In: Psychology Today [online].
Viewed: 27 August 2019. www.psychologytoday.com/articles/201301/
living-lie
Moorhead, Joanna (2013). “Children and divorce: ‘I just want to know why
they broke up’”. In: The Guardian [online]. Viewed: 27 August 2019.
www.theguardian.com/lifeandstyle/2013/aug/31/children-divorce-separated-
documentary-olly-lambert
Pym, Anthony (1992). “Translation error analysis and the interface with lan-
guage teaching”. In: Dollerup, Cay; Anne Loddegaard (eds.) (1992:279–288).
Webber, Rebecca (2014). “Reinvent Yourself”. In: Psychology Today [on-
line]. Viewed: 27 August 2019. www.psychologytoday.com/articles/201404/
reinvent-yourself
Worth, Jess (2014). “Ending the oil age”. In: New Internationalist [online].
Viewed: 27 August 2019. https://newint.org/features/2014/11/01/extended-
oil-keynote
7 Conclusion
Weaving our way through a maze of different approaches to, and dis-
cussions of, translation quality and translation quality assessment, we
have gleaned from them in passing, as it were, the occasional idea or
suggestion useful for our purpose of building an argument-based TQA
theory. Juliane House’s concept of overt and covert translation deserves
particular mention in this context as it pervades our argumentative
framework from top to bottom, having been streamlined for the pur-
pose and slightly adapted to our theoretical needs. A second pillar of
our theory is provided by Gregor Betz’s theory of dialectical structures:
in combination with Klaus Schubert’s elements of the decision-making
process, it furnishes the extra-translational basis of our approach to
translation quality. Last but not least, there are a great many ideas pro-
posed by various translation scholars that have made their way into the
TQA theory presented in this book. To name only some of these ideas,
we are indebted to Hans Vermeer’s principle of relative relativity, Chris-
tiane Nord’s extratextual and intratextual factors of translation, and
Anthony Pym’s minimalist translation competence model as well as his
distinction between binary and non-binary errors. We have also bene-
fited from several of the practically minded discussions of examples by
Hans Hönig and Paul Kußmaul, and sought backing for our argument-
based TQA approach with translation scholars such as Katharina Reiß,
Andrew Chesterman, and Jiří Levý. Catering for any kind of translation,
the resulting theory provides an argumentative basis that other theories
of translation quality and translation quality assessment can draw on to
substantiate their claims.
Consider, by way of example, Hans Vermeer’s skopos theory, which
has been criticised for reducing the source text to a mere offer of in-
formation, thus, leaving the translator with too much scope in his or
her decisions as to how to translate. This is why Juliane House, for
one, finds that “skopos theory is not very useful for translation quality
assessment” (House 2015:11). In his essay “Die sieben Grade einer
Translationstheorie”, Vermeer counters such criticism:
Tatsächlich lässt die Skopostheorie Raum für eine gebührende

Beachtung des Ausgangstextes, seiner Form, seiner Inhalte, seiner
274 Conclusion
Bedeutungen und/oder seines Sinns, wenn der Skopos dies erfordert
oder erlaubt. Eigentlich soll das Ausgangstextem in jedem Fall und
jeder Weise so weit wie möglich beachtet werden.
(Vermeer 2003:25)
In fact, skopos theory leaves room for due consideration of the source
text, its form, its contents, its meanings and/or sense when required
or allowed by the skopos. The source texteme should actually be
taken into account as far as possible – in any case and in any way.
In our argumentative theory of translation quality assessment, the skopos

of the translation situation is part and parcel of the translation strategy
that emerges as a result of the first three decisions – regarding (1) the
adoption of an overt or covert framework for translation, (2) source or
target orientation, and (3) implications as to what specific consequences
this has for the actual translation. The skopos can be defined as soon as
the three decisions have been made; it appears to be near-identical to our
concept of a translation strategy. We contend that skopos theory would
benefit from our argument-based approach as the latter can help to sub-
stantiate any claims made by the former in a given translation situation.
After all, the two theories are closely related in that they both rely on the
expectations of the target audience to provide translational guidance.
Our argumentative TQA approach boasts several different applica-
tions, which can be distinguished according to user type. The first user is
the translator as the person who renders a source text into a target text.
Here, application of the theory enables the translator to set up a transla-
tion strategy that matches the specific translation situation – which con-
tributes to a better translation quality as the translator becomes aware
of why he or she translates the way he or she does. At the same time, the
translator acts – even in the process of translating – as the first evaluator
of his or her own translation. To be able to evaluate different potential
translation solutions, the translator will consider the pros and cons for
each of these potential solutions and select the best one. Our theory pro-
vides the underlying rationale for this selection process. The second user
is the external evaluator of a translated text. He or she can draw upon
our argumentative approach to substantiate any criticism of individual
translation solutions. At the same time, however, the external evalua-
tor needs to have in place a clear concept of how a given source text
should be rendered into the target language. To establish such a concept,
the evaluator can rely on the TQA theory presented in this book. The
third user is the translation scholar who wishes to analyse evaluations of
translated texts. How this works has been demonstrated in Chapter 6,
in which we have analysed the examiners’ reports of commented trans-
lations written to fulfil the requirements for a bachelor’s degree.
Conclusion 275
Table 7.1 Uses of the argument-based TQA theory
User User role Use
Translator Producer of the target text Establishing a translation

strategy
Translator Evaluator of the target text Checking and optimising
the quality of individual
translation solutions in the
light of the translation strategy
External Evaluator of the target text Assessing the quality of
evaluator individual translation
solutions and the quality of the
translation as a whole
External Re-translator of the source Establishing his or her own
evaluator text translation strategy to be
able to assess the translator’s
translation and translation
strategy
Translation Analyst of translation Developing a set of parameters
scholar evaluations or translation based on the macro- and
criticism microstrategic principles
of our theory to analyse
translation evaluations
The different uses of our argument-based TQA theory are depicted in

Table 7.1. The user roles of the translator and the evaluator run in par-
allel, so to speak, since both the translator and the evaluator feature one
translating and one evaluating role each. However, there are two import-
ant differences, here. First, for the translator, the translating role takes
precedence over the evaluating role, which has a mere supporting func-
tion. This relation is reversed in the case of the evaluator, who focusses
on the evaluating role but needs to compare the results of the evaluation
with the strategic results developed in his or her translating role. Second,
while the translator – despite Gutt’s application of relevance theory to
translation (cf. Gutt 1991) and Levý’s claim that a MINIMAX strategy
is the preferred translational approach (cf. Levý 2000/1967:156) – will
often go for what he or she regards as the best translation solution, the
evaluator should content himself or herself with an acceptable translation
solution.
Using the present theory in the role of the analyst, we have come to
terms with how academics assess the quality of commented translations.
They tend to do so in line with a MINIMAX strategy by providing as
little argumentative support as is necessary to prove the point they wish
to make. This is why the arguments are often reflected in the description
or classification of an error. Moreover, the examiners’ reports occasion-
ally do not quote enough context for the reader to fully comprehend
the issue in question – such reports are obviously intended to be read in
276 Conclusion
conjunction with the bachelor’s thesis. Our findings could be used to set
up a form for the standardised assessment of commented translations,
an assessment that pays due attention to argumentation and translation
strategy. Future studies of how the quality of translated texts is evalu-
ated might include criticism of literary translations.
Probably the most obvious application of our argumentative TQA
theory consists in evaluating actual translations. These translations could
be renderings of functional text such as brochures or manuals, of news
reports, audio-visual text, or works of literature: there is no limit to the
kind of translation that can be evaluated using the theory presented in
this book. This is due to the fact that its argumentative foundation is to-
tally independent of translation and can also be used for other purposes.
What our TQA theory cannot do is quantify the results of the evaluation
and determine what overall mark the translation should get. However,
the theory’s fundamental design ensures that it can easily be used in
combination with other, more specific TQA approaches – notably those
that rely on quantification to assess the quality of a translation. Finally,
some machine translation (MT) approaches might possibly benefit from
the argument-based insights gained in our theory.
References
Gutt, Ernst-August (1991). Translation and Relevance: Cognition and C ontext.
Oxford and Cambridge: Basil Blackwell.
House, Juliane (2015). Translation Quality Assessment: Past and Present.
L ondon and New York: Routledge.
Levý, Jiří (2000/1967). “Translation as a Decision Process”. In: Venuti,
Lawrence (ed.) (2000:148–159). Originally published in 1967 in To Honor
Roman Jakobson on the Occasion of his Seventieth Birthday, vol. 2. The
Hague: Mouton, 1171–1182.
Studia Germanica Posnaniensia, vol. XXIX, 19–38.
Index
Note: Bold page numbers refer to tables and page numbers followed by “n”
denote endnotes.
Acartürk-Höß, Miriam 39, 54 Benhaddou, Mohamed 31

Adab, Beverly 32 Betz, Gregor 188–95
Adamzik, Kirsten 257 Billiani, Francesca 152, 153
Ahrend, Klaus 51, 101 Bittner, Hansjörg 40, 98, 114, 160n12,
Al Qinai, Jamal 31, 71, 88n1 163n47
Albrecht, Jörn 99 Boase-Beier, Jean 160n14
Al-Bustan, Suad Ahmed 31, 56 Böll, Heinrich 114
Al-Taher, Mohammad Anwar 149 Bowker, Lynne 12, 33, 63–4, 73, 74–5
alternative translation solution Broeck, Raymond van den 28–9, 73,
257–64 84, 162n38
Álvarez, Román 149–50 Brownlie, Siobhan 154
Alvstad, Cecilia 162n40 Budin, Gerhard 35
Ammann, Margret 72–3 Bühler, Hildegund 52
Anderman, Gunilla 120, 160n17 Bulcke, Patricia Vanden 40
Angelelli, Claudia V. 37–8, 75
argument 199–200, 233–5, 252–7, calibration of dichotomous items
264–7; relations 189 method 38
argument macrostructure 14 Carbonell, Ovidio 103, 147
Aristotle 50 Carroll, John B. 28, 72
Arthern, Peter J. 171 Carroll, Mary 123, 124
Auden, Wystan Hugh 39, 54 CDI method see calibration of
audience 61, 64–5, 106–7; see also dichotomous items method
reader censorship 152–4
audiodescription 122, 125–6 Chesterman, Andrew 159n7, 174, 176
audiovisual translation 121–6 client 51, 52, 127–33
author 61, 64 comics 120–1
commented translation 82–4, 224–5
Baker, Mona 142, 143 competence see translation
Bandia, Paul 146 competence
Barghout, Mohamed Abdel-Maguid content 61, 65, 114–15
30–1 corpus 224–5
Bastian, Sabine 36 creativity 81
Bayer, Klaus 172 cultural filter 8
Beeby-Lonsdale, Allison 53 culture 139–43; Arabic 143; Chinese
Behr, Dorothée 35 143; eighteenth century German
belles infidèles 145, 183–4 144–5; Japanese 144; seventeenth
Benecke, Bernd 161n26 century French 145
278 Index
Darbelnet, Jean 218n5 formal characteristics: of commented
Daston, Lorraine 68 translations 226–7, 244; of
De Sutter, Nathalie 40 examiner's reports 227–8, 244–5
deadline 51, 129–30 Fox, Brian 56
Debove, Antonia 40 Franco Aixelá, Javier 148, 161–2n31
decision-making process 179–81, 187 Furuno, Yuri 144
decisions 177–87
defect see source text: defect Galinski, Christian 35
definition: good quality in translation Galison, Peter 68
55–6; quality factors 101; García Álvarez, Ana María 36, 270n1
translation 170–1; translation genre 89n9; see also text type
quality assessment 56 Gerisch, Gordon 36
Delizée, Anne 40 Gerstner, Johanna 35
Depraetere, Ilse 39, 40, 52, 90n15 Gerzymisch-Arbogast, Heidrun 17–21
Didaoui, Mohammed 35 Gigon, Olof 67–8
discursive goals 192–3 Gledhill, Christopher 40
Dosch, Elmar 161n26 glossary 130–1
drama 119–20 Goethe, Johann Wolfgang von 154
Drugan, Joanna 77, 162n35, 162n36, good quality in translation:
162n41, 163n45 definition 55–6
Dung, Phan Minh 172 Göpferich, Susanne 36
Gouadec, Daniel 38–9
ECR see extralinguistic cultural Granzin, Katharina 54, 102–3
reference Greiner, Norbert 112–13, 247
Emsel, Martina 51, 161n30, 247 Grutman, Rainier 139
EMT expert group 175 Gummerus, Eivor 33, 55
entailment relations 190–1 Gutt, Ernst-August 21–6
equivalence 7–8, 103
error: and achievement 85, Hagemann, Susanne 36, 81, 85
230–2, 247–50; binary vs. Halliday, M. A. K. 32, 177
non-binary 79–80, 270n5; Hamada, Tomoko 50
correcting 86; counting 81–2, Hanna, Sameh Fekry 143
85; different assessment 77–9; Hansen, Gyde 146
weighting 85–6 Hardy, Thomas 54
evaluation see translation quality Hatim, Basil 151, 158n1
assessment Heinze, Hendrik 160n21
evaluation perspective 80–1, 231–2, Hermans, Theo 150–1
249–50 Héroguel, Armand 40
evaluator vs. translator 99, 204–5 Holz-Mänttäri, Justa 96–7
explicit information 181 Hönig, Hans 31, 146, 219n22;
extralinguistic cultural reference 37 acceptable translation 52;
extratextual factors 104 criticism of relevance theory 43n7;
Eyckmans, June 38, 71, 74 empiricism 73; macrostrategy
218n8
factor matrix 102 Honko, Lauri 173
factors see translation strategy: factors Horn, Olga 158n1
or quality factors Horn-Helf, Brigitte 175
Fan, Shouyi 30 Horton, David 31
Fawcett, Peter 72 House, Juliane 7–13, 99–100, 175,
film: polysemiotic environment 273; author 64, 89n11; genre 89n9;
121; translation see audiovisual good translation 163n45; overt–
translation covert translation 56–8, 88–9n6
Flotow, Luise von 153, 155 Huang, Harry J. 89n13
Index 279
Humboldth, Wilhelm von 42–3n2 Martin, Charles 40, 204
Hung, Eva 143 Martin, Silke Anne 36
hybridity 147 mathematical model 30
Matis, Nancy 88n3
ideology 149–50 medium 107
implicit information 181 Melby, Alan K. 34, 158n2
interlinear translation 62 Merkle, Denise 152, 153
intratextual factors 104 minimum quality requirement 51
Ivanova, Vessela 36 Morales Tejada, Beatriz 53
Ivarsson, Jan 123, 124 Mossop, Brian 73, 76, 175
motivation 132, 138
Jacobson, Holly E. 37–8 motive for communication 109
Jakobson, Roman 159n8, 160n13 Moulard, Nathalie 39–40
Jumpelt, Rudolf Walter 27, 177–8 Mudersbach, Klaus 17–20
Jüngst, Heike Elisabeth 36, 160n22 Munday, Jeremy 72, 158n1,
justification 175–6, 191–2 163n51, 180
Kandler, Günther 27, 56 Nagel, Thomas 68–9, 195

Kang, Ji-Hae 147 native speaker 53
Katan, David 139–40, 141, Neubert, Albrecht 159n7
162n42, 173 Newmark, Peter 54
Kautz-Vella, Harald 161n26 non-verbal elements 116
Kempa, Thomas 181 Nord, Christiane 52, 104–9, 112, 113,
Kingscott, Geoffrey 35–6, 73 115–17
Kittel, Harald 145 normalisation 12, 63–4
Klaudy, Kinga 82 norms 141–3, 148
Koller, Werner 159n7 Núñes-Lagos, Carmen 39–40
Kondo, Masaomo 144
Krings, Hans P. 219n22 objectivity 68–70, 195
Krüger, Ralph 97 Ofcom 160–1n23
Künzli, Alexander 175 overt–covert translation 8, 56–7, 182;
Kupsch-Losereit, Sigrid 35 principle vs. implications 59
Kurch, Alexander 160–1n23
Kurz, Christopher 35, 53 PACTE 135, 162n37
Kußmaul, Paul 77–8, 80–1, 159n7, Paepke, Fritz 180
211–12 Palmer, Jerry 112, 138
Paro, Catrine 33, 55
language pairs 103–4 Pavlova, Anna 41–2, 156
Larose, Robert 31 Pedersen, Jan 37
Lauscher, Susanne 31–2, 52 Penndorf, Gudrun 152
Lee-Jahnke, Hannelore 33, 71, 81, 85 Pinto, María 32–3, 88n1
Lefevere, André 141, 152, 154, 155, place of communication 107–8
162n33 poetic function 160n13
Levý, Jiří 77, 176, 180 poetry 54, 118–19, 160n14
lexis 116–17 Pokorn, Nike 152–3
limitations of proposed theory 213, 276 politics 149–57
localisation 126–7 Pollard, David 143
Lörscher, Wolfgang 218n6 Poltermann, Andreas 145
power 150–2, 154–5
machine translation 28, 51, 72, 134–5, Prill, Ulrich 163n47
162n35 principles of an argument-based
McLeod, Kenneth 219n12 TQA 172
Mälzer, Nathalie 103, 160n18 Pym, Anthony 79, 126, 136, 172, 174
280 Index
qualification 138 source–target comparison 229,
quality 50–1 246–7
quality factors 100–3, 177, 238–40; source text 103–12; analysis 104;
definition 101; known vs. unknown defect 110
101–2; purpose of 96; relevance to source vs. target orientation 59, 184
assessment 100–1 specifications 131–2
questions: and response categories standardisation 35
241–3; and results 268–9 Stauder, Andy 163n48
Steiner, Erich 42, 175
Radnitzky, Gerard 68 Stejskal, Jiri 53
Raffel, Burton 160n16 stipulations 131
reader 182; see also audience Stolze, Radegundis 90n15, 175
Reinart, Sylvia 123, 162n38, strategy see translation strategy
162n39; categorisation of criteria subject matter 115
159n5, 177; comics 120, 160n18; subjectivity 10–11, 15–16, 67–76;
terminology 159n11; translation philosophical point of view 67–70
criticism 40–1 subtitling 122; interlingual 122–4;
Reinke, Uwe 135 intralingual 122, 124–5; live
Reiß, Katharina 28, 51, 72, 84, 112, 160–1n23; verbatim 160–1n23
175–6 suprasegmental features 117–18
rhetorical model 30 syntax 117
Risku, Hanna 96–7
Robinson, Douglas 1 terminology: management 35;
Rodrigues, Sara Viola 31 decision process 186; use in the
Rodríguez Rodríguez, Beatriz Ma 36, book 4–5
73, 177 text composition 116
Romero-Fresco, Pablo 160–1n23, text form 61–4, 112–27
161n24 text function 109
Rosenmund, Alain 33 text type 113, 118; vs. genre 89n9
Thelen, Marcel 36–7, 55
Sabban, Annette 205–8 time of communication 108
St. André, James 162n44 Torresi, Ira 110, 112
Salama-Carr, Myriam 145, 184 Toury, Gideon 142, 163n46, 170–1
Schäffner, Christina 72, 141, 159n6 TQA see translation quality
Schmitz, Klaus-Dirk 35 assessment
Schneiders, Hans-Wolfgang 160n15 translating: a classical novel 184–5; a
Schreiber, Michael 159n7 journalistic essay 202–3; a marriage
Schubert, Charlotte 52 certificate 185; a website 200–2;
Schubert, Klaus 126, 127, 161n29, fiction vs. non-fiction 112–13; user
177–9 instructions 202
Scott, Clive 159n9 translation: agency 128; brief 131;
Secară, Alina 34 competence 55, 135–7; definition
selection process 196 170–1; method 218n5
self-translation 139 translation quality 1–2, 51–3;
sender 105 perspectives 6
sender's intention 105–6 translation quality assessment 76–88;
Senn, Fritz 158n1 definition 56; options 84–5; process
Sharkas, Hala 34 70–1; requirements 178, 232, 270;
Shelley, Percy Bysshe 160n14 survey 211–17; young adult novel
Siever, Holger 159n7 205–11
Sinner, Carsten 53 translation strategy 1, 58, 66, 181–7,
skopos 273–4 218n6; factors 61; recognition of
Snell-Hornby, Mary 159n7 232–3, 250–1
Index 281
translation theory 173–4 version (as opposed to translation)
translation tools 134–5 9, 57–8
translator 55, 133–9 Vidal, M. Carmen-África 149–50
translator's daffodil 97–9 Vinay, Paul 218n5
usefulness of the translation 52–3 Waddington, Christopher 33–4,

users of proposed theory 2, 34–5, 81–2
274–6, 275 Wakabayashi, Judy 144
Wiesmann, Eva 51
Vackier, Thomas 39 Williams, Malcolm 13–17, 29, 72,
Vahid Dastjerdi, Hossein 40 73, 75
Vanwersch-Cot, Olivier 40 Wilss, Wolfram 54–5
Vermeer, Hans J. 51, 72, 159n10, Wittbrodt, Andreas 160n15
218n2, 218n4; culture 140, 141; Witte, Heidrun 137, 140
no invariants in translation 99, 171, Wyke, Ben van 133, 162n40
177; skopos 273–4; translation
factors 158n1 Zilahy, Simon Pietro 27

Evaluating The Evaluator

Uploaded by

Copyright:

Available Formats

Evaluating The Evaluator

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Evaluating The Evaluator

Uploaded by

Copyright:

Available Formats

What is the book about?

What is the book about?

What theoretical framework does the book outline?

What theoretical framework does the book outline?

Evaluating the Evaluator

This book offers a theoretical framework for assessing translation q ­ uality

Hansjörg Bittner served as a lecturer at Eastern Mediterranean ­University

38 Complexity Thinking in Translation Studies

39 Translating and Interpreting in Korean Contexts

40 Hybrid Englishes and the Challenges of/for Translation

41 Translating the Visual

42 Using Computers in the Translation of Literary Style

44 Humour in Audiovisual Translation

45 Evaluating the Evaluator

For a full list of titles in this series, visit https://www.routledge.com/

2 The Quality of Translation: Different Approaches 6

4 Quality Factors of Translation 96

5 The Principle of Argumentation 170

6 Evaluating the Evaluator 224

3.1 Overt–covert overview (general aspects) 60

4.1 Summary of source text factors 111

I thank those who have supported me during this project. First of

In The Translator’s Turn, Douglas Robinson writes:

It seems undeniable that translation is largely an intuitive process.

• translators can analyse potential translation solutions and decide

Thus, we propose a theoretical tool which benefits those actors in a

in overt translation the function of the translation is to enable its

The implications of this distinction will be discussed in more detail be-

The concept of a “cultural filter” is a means of capturing socio-­

To judge is easy: to understand is less so. If we can make explicit

Yet, notwithstanding these insightful comments, there is still room for

[E]very instrumental source text contains an argument macrostruc-

in an argumentation-centred TQA (ARTRAQ) model,

• defects impairing translation of the argument macrostructure

The drawback of the standardized, norm-based procedure which

Heidrun Gerzymisch-Arbogast and Klaus Mudersbach

one of the main problems with the scientific investigation of transla-

[I]f it turns out that each individual phenomenon – which here is

Equivalence not only appears to be difficult to harness from a theoreti-

At a climactic moment in Stendhal’s Le rouge et le noir (Book II,

To analyse this translation critique, Gutt distinguishes two ­perspectives:

1 Since Adams obviously considers Scott-Moncrieff to be a good

To shed light on the position of the audience, Gutt provides additional

1 The fact that “thou” is an obsolete pronoun which is at best used in

The above detailed analysis of this fairly straightforward translation prob-

I tried to show that the principles, rules and guidelines of transla-

Other Approaches to Translation Quality

Approaches to Translation Quality in the Twentieth Century

In my view, translation criticism, despite the subjective element in-

Furthermore, there is the distinction between “obligatory shifts” and

The proposed mathematical model […] is based on the general prin-

In this way, translation quality can, indeed, be represented in terms of

Which factors should be relevant to the quality of a particular transla-

Approaches to Translation Quality in the Twenty-First

Given that translation evaluation entails making judgements about

Bowker describes in detail the design of a special evaluation corpus.

In spite of these caveats about the difficulty of gauging the effect of

This approach is applied to the evaluation of student translations, so its

[T]here is no guarantee whatsoever that […] the human being

This book offers a theoretical framework for assessing translation q uality

Hansjörg Bittner served as a lecturer at Eastern Mediterranean University

The concept of a “cultural filter” is a means of capturing socio-

To analyse this translation critique, Gutt distinguishes two perspectives:

Figure 3.1 Overt–covert overview (general aspects).

A binary error opposes a wrong answer to the right answer; non-

Figure 3.4 Evaluation focus for a commented translation.

Figure 4.2 Factor matrix.