See discussions, stats, and author profiles for this publication at: https://www.researchgate.net/publication/329102414
Computer-Assisted Translation (CAT) Tools in the Translator Training Process
Book · November 2018
DOI: 10.3726/b14783
CITATIONS
READS
26
10,942
1 author:
Michał Kornacki
University of Lodz
22 PUBLICATIONS 66 CITATIONS
SEE PROFILE
All content following this page was uploaded by Michał Kornacki on 06 August 2021.
The user has requested enhancement of the downloaded file.
1
Table of contents
MICHAŁ KORNACKI
DEPARTMENT OF TRANSLATION STUDIES, INSTITUTE OF ENGLISH STUDIES
FACULTY OF PHILOLOGY, UNIVERSITY OF ŁÓDŹ
COMPUTER-ASSISTED TRANSLATION (CAT) TOOLS IN
THE TRANSLATOR TRAINING PROCESS
1
Table of contents
Table of contents
Table of contents ............................................................................................................... 1
Acknowledgments ............................................................................................................ 3
Chapter 1. Introduction ..................................................................................................... 4
Chapter 2. The role of a translator in the modern world................................................. 11
Chapter 3. Translation competence ................................................................................ 18
3.1. Linguistic and general competence ...................................................................... 18
3.2. Translation competence ....................................................................................... 19
3.2.1. The PACTE model of translation competence .............................................. 23
3.2.2. The EMT model of translation competence .................................................. 29
3.2.3. The TransComp model of translation competence ........................................ 34
3.2.4. Comparison of the presented models ............................................................ 36
3.3. Translation competence vs. translator’s computer competence ........................... 37
3.3.1. Computer competence skills.......................................................................... 38
Chapter 4. Translator training ......................................................................................... 52
4.1. Developing translation competence ..................................................................... 52
4.1.1. Translation competence acquisition model ................................................... 52
4.1.2. Pedagogical perspective ................................................................................ 54
4.1.3. Creating technology-oriented translation course ........................................... 62
4.2. University training programmes .......................................................................... 75
4.3. In-class activities .................................................................................................. 78
4.4. Challenges for translation trainers ....................................................................... 80
4.5. Translation strategies in the translation classroom .............................................. 81
4.5.1. Direct translation procedures ......................................................................... 87
4.5.2. Oblique translation procedures ...................................................................... 90
4.5.3. Translation procedures in the context of CAT tools ..................................... 92
Chapter 5. Computers in translation ............................................................................... 94
5.1. Machine Translation (MT) ................................................................................... 94
5.1.1. Rule-based MT (RBMT) ............................................................................... 97
5.1.2. Statistical Machine Translation (SMT) ......................................................... 98
5.1.3. Hybrid MT ..................................................................................................... 99
5.2. Computer-assisted translation ............................................................................ 100
5.3. Terminology disambiguation ............................................................................. 104
5.4. CAT-based translation process .......................................................................... 116
5.5. Translator vs. computer in contemporary professional translation .................... 122
Chapter 6. Research ...................................................................................................... 129
6.1. Aims of the research study ................................................................................. 129
6.2. Courses available on the market ........................................................................ 129
6.3. Determining justification for the research .......................................................... 134
6.4. Methodology behind the main research ............................................................. 138
6.4.1. CAT environment ........................................................................................ 138
2
Introduction
6.4.2. Timeframe .................................................................................................... 139
6.4.3. Test groups ................................................................................................... 139
6.4.4. Test documents ............................................................................................ 139
6.4.5. Translation memories................................................................................... 140
6.4.6. Errors in translation memories ..................................................................... 142
6.4.7. Stages, phases, and preparation of students ................................................. 150
6.4.7.1. Preparation of students .......................................................................... 151
6.4.8. Data collection methods ............................................................................... 155
6.5. Course of the research ........................................................................................ 162
6.5.1. Stage I Phase I .............................................................................................. 163
6.5.2. Stage I Phase II ............................................................................................ 164
6.5.3. Stage II Phase I ............................................................................................ 165
6.5.4. Stage II Phase II ........................................................................................... 165
6.6. Data and analysis ................................................................................................ 166
6.6.1. Time ............................................................................................................. 166
6.6.2. Errors............................................................................................................ 168
6.6.3.1. Stage I Phase I results............................................................................ 169
6.6.3.2. Stage I Phase II results .......................................................................... 170
6.6.3.3. Stage II Phase I results .......................................................................... 171
6.6.3.4. Stage II Phase II results ......................................................................... 171
6.6.4. Analysis........................................................................................................ 172
7. Results and conclusions ............................................................................................ 173
References ..................................................................................................................... 180
Appendices .................................................................................................................... 195
Appendix 1: List of procedures for CAT-based translation ...................................... 195
Appendix 2: Main research source document – Phase I ............................................ 195
Appendix 3: Main research source document – Phase II .......................................... 200
Appendix 4: Student results per stage ....................................................................... 204
Introduction
3
Acknowledgments
This book would not be possible had it not been for a number of people whom I would
like to thank here heartily. First of all, I would like to thank Professor Łukasz Bogucki
for his support and supervision over my research, as well as for accommodating this
contribution in the Łódź Studies in Language series. I am also thankful to dr Paulina
Pietrzak for her constant inspiration and guidance throughout the project.
I wish to thank Professor Piotr Stalmaszczyk and Professor Gary Massey for their
invaluable comments and advice regarding the main research, as well as for their
continuing support in my endeavours. Thanks are also due to Professor Konrad
Klimkowski and Professor Łukasz Grabowski for their deep insight and critical
comments on an earlier version of this work, and to Tamás Ritter and Peter Reynolds
(Kilgray) for providing the CAT tool (memoQ) for the research. I am also thankful for
help and advice from my colleagues at the University of Łódź.
Finally, I want to thank the one person without whom all of this would not be
possible – my wife, Ania. Thank you.
Michał Kornacki
4
Chapter 1. Introduction
Translation has been employed in various forms since the dawn of civilisation. The very
first organised human cultures needed translation, initially oral and then written following
the invention of writing systems. Eugene Nida (1988: 23) associates the beginning of
translation with Septuagint, which was probably the first translation of Hebrew Old
Testament into Greek (ca. 3rd century BCE). Douglas Robinson (1997, 2002: 7), on the
other hand, finds beginnings of translation in techniques employed by Marcus Tullius
Cicero (106-43 BCE) in De optimo genere oratorum. Cicero translated two Greek letters
into Latin, thus explaining the process:
I did not translate them as an interpreter, but as an orator, keeping the same ideas and the
forms, or as one might say, the “figures” of thought, but in language which conforms to our
usage. And in so doing, I did not hold it necessary to render word for word, but I preserved
the general style and force of the language. For I did not think I ought to count them out to
the reader like coins, but to pay them by weight, as it were (translated by H. M. Hubbell,
available at https://www.loebclassics.com/view/LCL386/1949/volume.xml).
The demand for translators resulted in a more conscious approach to translator education.
Anthony Pym (2009: 1) suggests that origins of more extensive training programmes
could be found in the “elaborate Chinese institutions for the translation of Buddhist texts
(4th to 9th centuries), in the “House of Wisdom” in the 9th century Baghdad, or in cathedral
chapters as in the 12th century Toledo.” Conflicts fostered translation since it was
necessary to understand one’s enemy. Even European colonists in America employed
basic translator training in order to train captured natives to serve them as interpreters and
help them in contacts with the local population. One common thing was that translator
training was carried out locally, at the place it was needed, which meant that state borders,
places where civilisations met, witnessed more rapid development in the field. Again,
Pym (2009) names such examples as schools training French interpreters in
Constantinople (1669), the Diplomatic Academy of Vienna (1754), Al-Alsun school of
translation in Egypt (1835) or an increasingly fast development of translation schools in
China (19th century), with Yan Fu in charge of several translation schools in China,
starting from 1896.
It is worth noting that initially translator training was state-controlled to ensure both
translation quality and allegiance of translators, especially in Europe (see Caminade and
Introduction
5
Pym 1998; Pym 2009); however, such attitude was not uniform around the world. On the
other hand, in Spanish America “translator training was more closely related to sworn
translation which was a tool used to maintain juridical regime, employed in turn to control
colonies” (e.g. Translation curriculum at the Law Faculty, University of Uruguay, 1885)
(Pym 2009). It shows that translator training programmes were very diversified,
depending on local political or institutional requirements.
The 20th century and events like the World War II (the Nuremberg trials in
particular, which hinted at the future role of interpreters in international institutions)
emphasised the importance of translation in national and cross-country communication.
The first programmes aimed at “training professional translators and interpreters were
introduced at the University of Geneva, Switzerland, in 1941, Vienna, Austria, in 1943,
Mainz-Germersheim, Germany, in 1946 and Georgetown, USA, in 1949, for example”
(Schäffner and Adab 2000: vii). The start of the Cold War was a real turning point for
Translation Studies. Suddenly everyone realised the import of knowledge that could be
gained by translating foreign data, like military strategies, plans, and other classified
information. The need to be able to translate vast quantities of data in short periods of
time led, at least partially, to Warren Weaver Memorandum in July 1949. The
memorandum was crucial for the development of machine translation (MT) as it outlined
aims and methods of linguistic research in the field of MT way before computer
capabilities became known to researchers, first in the US and then around the world (see
MT News International, no. 22, July 1999).
The rise in general awareness and need both for regular translators and machine
translation led to the creation of specialised academic translation programmes. Those
were meant to produce highly skilled individuals who could provide translation services
in the civilian world on the one hand, and research and develop suitable MT solutions on
the other. In 1954 Georgetown MT research team presented their working MT system
(see Hutchins, 1994). In 1964 ALPAC (Automatic Language Processing Advisory
Committee) was formed to study MT and evaluate overall progress in computational
linguistics. The committee had a very negative impact on MT in general. Over the next
20 years the research in the field was severely hampered but ultimately led to the
development of, among others, translation memory technology, first commercially
employed by SDL in the 1990s.
The European Union and its need for legislation and other documents to be rendered
in the official languages of all member states significantly increased the demand for both
Michał Kornacki
6
highly skilled human translators/interpreters and automated translation systems.
However, even more important drive factors behind this growth of importance of
translation included a total overhaul of economies of developed countries in the 1980s
(privatisation and deregulation) (see Castells, 1980), the “advent of computer technology
(silicone semiconductors and computer on a chip), and emergence of computer software”
(Cronin, 2003), which made it possible to develop computer-assisted translation tools.
What is more, translation as such started to be seen in terms of a vocation, which
facilitated integration of translator training programmes into university structures.
According to Pym (2009), unemployment amongst students resulted in increased demand
for translator training programmes since young people saw this as an opportunity to find
a job, even though there was no rise in demand for full-time translators and interpreters.
There are countries, though, where demand for quality translators exceeds the capacity of
an education system, e.g. China (see Schäffner and Adab 2000; Pym 2009). In the end,
and thanks to the access to all kinds of computer tools facilitating translation, a significant
portion of all translation traffic is translated by underqualified people, prone to errors and
major setbacks. Data from a number of European Commission surveys show that
freelance (self-employed) translators constitute about 74% of the total number of
translators in Europe. Apart from Slovakia, no European country regulates the profession
of translator. Therefore, even an unqualified person can work as a freelance translator.
Some states do impose specific requirements on sworn or authorised translators, but those
requirements vary from state to state, if they are in force at all (see Pym, Grin, Sfreddo
and Chan, 24 July 2012).
The book will discuss student-computer interaction in the context of computerassisted translation (CAT) course. An attempt will be made to show that the currently
employed CAT teaching methods may result in students relying too much on the CAT
software, and that the incorporation of problematic aspects of external data verification
leads to increased awareness and better quality of student translations.
The primary goal of the book is to diagnose how CAT tools work and to identify
students’ needs in regard to CAT skills acquisition process. What is more, the book will
attempt to define teaching goals for the translator training process in the context of the
use of CAT tools.
CAT tools are state-of-the-art computer programs that allow human translators to
translate faster and with better quality (Bundgaard, Christensen and Schjoldager, 2016).
However, their advantages are to be clearly seen in the case of certain types of (mostly
Introduction
7
scientific and technical – see Fernández-Parra, 2010) texts due to the fact that such texts
follow similar style and structure patterns and, as a result, are repetitive to a certain
degree. In order to stay competitive on the market (the need to work with client-provided
translation memories; optimise work time to translated pages ratio; and increase output
volume and rates per page), a translator is left with no choice but to utilise the latest
translation technology (Pym, 2009; Christensen and Schjoldager, 2016). A variety of
different tools is available on the market at the moment of writing this book (2017). They
include advanced formatting-capable word processors, TM-based solutions, automated
glossaries, machine translation, corpora, and more. A successful translator is required to
know and accommodate these resources in order to achieve best possible results.
However, one has to be aware of the dangers the technology brings about and know
how to negate them. The quality of contemporary computer resources may lead to
misplaced trust in the feedback obtained from them, occasionally resulting in lower
quality translation. Such over-reliance brings about questions whether computers could
replace humans in translation. In fact, some researchers agree that while combination of
various computer resources “can be expected to replace fully human translation in many
spheres of activity” (Pym, 2012, p. 487). At the same time, translators are expected to
become posteditors, refining computer translation (idem.). Even today some types of
specialised texts allow CAT users to revise and approve automated translation output,
providing their own translation only when necessary.
The book assumes that over-reliance on computer resources, especially in the case of
CAT tools, may lead to [serious] errors in translation (see Doherty, 2016), especially in
the case of students, graduates and less experienced translators in general. In order to
prove this theory, a research has been conducted, aiming to show that the users of CAT
tools are prone to accept proposed translation with no proper revision, even though these
suggestions may contain errors1. Such errors may result from a number of reasons. These
may include, for example, incorrect context, or previous translation errors in external
resources (Barbu, 2015), i.e. translation memories (e.g., MyMemory2, which can be
1
2
Both Bowker (2005) and Doherty (2016) suggest that many translators have adopted the so-called
“blind faith” in the previously used human translation, stored in translation memories. As a result, they
assume (often erroneously) that the data is of high quality and does not require as close evaluation as
in the case of translation from scratch. Doherty (2016: 954) claims that this tendency is “compounded
by the reduced remuneration for using TMs” due to the fact that (in theory) less translator interaction
is needed when using partially translated sentence matches.
MyMemory (https://mymemory.translated.net/) is a translation memory containing over 1 billion
bilingual segments in over 6000 language pairs (Barbu, 2015). See Trombetti (2009) for core
Michał Kornacki
8
accessed directly from many CAT tools via proper plugins, e.g. in memoQ3; or resources
obtained from translation service ordering parties [TSOP]). The research focuses on the
errors found in the TSOP-delivered translation memories. It analyses how they were
processed by test subjects. Finally, the book summarises conclusions to be drawn from
the study in relation to the translator training process.
Firstly, an attempt is made to confirm that such errors appear in authentic professional
conditions and that they are of significance statistically. This was done on the basis of a
questionnaire directed to professionals specialising in CAT-based translations. The
purpose of the survey was to determine whether the abovementioned scenarios take place
as well as the scope of errors encountered in external translation memories, and the degree
to which the linguistic quality of these resources affects the translation process.
Secondly, the main research is conducted in order to determine how trainee
translators deal with the problem of faulty external resources in a live project. The
research was conducted in two stages. The first stage took place in 2015/2016 academic
year and involved a group of 22 test subjects. The second stage took place in 2016/2017
academic year with 18 participants. Each stage was divided into two phases in order to
check initial and final CAT skills after a course on CAT tools. Both test groups included
2nd year MA students undergoing CAT programme during their final semester at the
University of Łódź.4 The first group learned how to operate the software. In addition to
the content covered by the first group, the second group discussed dangers posed by CAT
tools and the methods of avoiding them in the process of computer-assisted translation.
While translators have no influence on the quality of received translation memories,
they do have full control over what will be done with the problem during the translation
itself. The study assumes that a significant portion of the previously committed errors is
retained in translations performed by CAT mechanics-oriented (see Chapter 7.1) trainee
translators (I stage) to a much greater extent than in the case of CAT specifics-oriented
(see Chapter 7.1) trainee translators (II stage). Such result would imply that it is not
enough to understand how to operate a CAT tool since it will not yield the desired
outcome. Therefore, it is critical to learn how to use advantages of CAT tools and
consciously avoid any disadvantages they may have. Developing a set of best practices
3
4
principles behind the project and Barbu (2015, p. 9) for outline of methods used to build MyMemory,
and how they affect number of possible errors.
www.memoq.com
Characteristics of both groups and research material will be discussed later in the course of the book.
Introduction
9
regarding the practical use of CAT tools may significantly improve the overall translation
quality through the reduction of the risk of committing an error.
The results allow to draw conclusions regarding initial weaknesses of emerging
translators and propose solutions to be incorporated in the translator training process. The
proposed assumption of over-reliance on computer-assisted translation tools may signal
the emergence of a serious problem to be faced by both students and teachers of
translation, especially in the context of the translator→post-editor evolution of the
profession, as suggested by Pym (2013).
The book starts with a brief outline of the history of computers and evolution of
computer tools for translators. Chapter 2 hints at the role of translators in the modern
world. This is followed by a discussion on the issue of translator competence seen as a
number of theoretical and practical assets that build translator’s identity and form their
translation workshop. It covers a number of translation competence models in order to
provide an overview of what it is to be a translator. Apart from that, a few words are
devoted to the translator training process in Chapter 4, with focus on the application of
technology in translation, as well as the role of the translation trainer. Chapter 5
introduces the concept of computer-assisted translation in detail, discussing its place in
Translation Studies, elements of machine translation in CAT tools, and attempts to
redefine it in the 21st century. The need for redefinition is necessary due to fact that the
development of new technological assets and their integration result in the notion of
computer-assisted translation losing its clarity, especially in the sense of where computer
assistance ends and computer/machine translation starts. Chapter 6 recapitulates the
research goals and outlines research methods, both considered and selected for the study.
Furthermore, it presents data analysis and results. Finally, Chapter 7 summarises
conclusions drawn from the research and discusses whether the main assumptions behind
the book have been proven correct. It proposes suggestions that may be beneficial for
future translator training and suggests a new direction for future research on the issue.
Chapter 2. The role of a translator in the modern world
11
Chapter 2. The role of a translator in the modern world
Before we can venture into the translator training process, CAT tools or translation
competence, a few words need to be devoted to the profession of a translator and their
identity in the modern world.
The history of translation can be traced back to the dawn of human civilisation, as
was mentioned in the Introduction. Its evolution was inextricably linked to the evolution
of politics, trade, industry, and social needs of human beings. What is more, it was shaped
by conflicts, especially the global ones like World War II which was followed by the
Nuremberg trials that have shown the importance of interpreters to the broader audience.
The evolution of translation as profession resulted in creation of numerous norms
(see, for example, Schäffner, 1998; Gile, 1999; Toury, 1999), some of them regarded of
paradigms of modern translation, e.g. translation sense for sense, not word for word,
proposed by Cicero in the 1st century BC. The norms, as well as their origins, are relative,
however. For example, the norm proposed by Cicero is regarded as one of the most
profound discoveries in the world of translation. However, there is no evidence that
translation sense for sense was not a common norm before him (Osers, 1995). What is
more, it can be said that “translation ‘verbum de verbo’ never, at any time in the history
of translation, was the translation norm, but rather the practice of incompetent translators”
(idem.: 54). The justification of this statement is logical – it is hard to believe that a
translation consisting of a string of words in the SL word order would be comprehensible
and acceptable enough to validate the use of translators (especially in the world of ancient
politics which is reported in the annals of history as much more violent than today).
Therefore, it is logical to assume that translators had to be competent and produce good
(useful) translations to be kept around.
One reason to explain the widespread idea that early translation was word-for-word
is that initially the demand for translation was rather low due to localised trade and
political life. Therefore, translation (as it was then) cannot be regarded as an occupation
but rather as a set of additional skills on behalf of happenstance translators – travellers,
traders, soldiers and other people who had the chance to learn a foreign language. Of
course, such a statement does not mean that there were no actual translators. There are
examples of works of translation in ancient history which was not related to the basic
desire for communication but rather originated from spiritual, scientific or artistic needs
(for example, partial translations of the Sumerian Epic of Gilgamesh into Southwest
Asian languages in the 2nd millennium BC, Cohen, 1986). Much younger feats of
translation include Xuanzang’s translations of Indian Buddhist texts into Chinese (some
of which continue influence daily Buddhist forms of reverence even today, Lusthaus,
2000); Toledo School of translating many of the philosophical and scientific works from
Classical Arabic (Burnett, 2001) which allowed to make them known across Europe; or
numerous translation of Bible into English over the ages.
Ultimately, all of these (and more) achievements and the ever increasing pace of
development of the human civilisation have led to the establishment of the framework for
the profession of the translator as we know it today.
At the moment of writing this book (2017), there are no formal requirements to
become a translator in Poland. Obviously enough, if someone wants to become a sworn
translator, certain requirements have to be met. Individuals willing to become an officially
recognised translator have to be graduates of philological or non-philological studies and
pass “a State specialist examination in written legal translation and oral court interpreting:
(TEPIS, 2017: online resource). As sworn translators, the individuals have the “duty not
to refuse a commission of a court or the police, to keep a register of commissions received,
to comply with the rules of good practice, professional ethics, and confidentiality, and to
upgrade professional qualifications” (idem.). Other European countries implemented
similar legislation to control domestic sworn translation. However, the Polish Act is one
of the most “demanding among similar legislative instruments passed by other European
states” (idem.).
None of the above requirements apply if a person wants to become a regular
translator. Virtually anyone can become a translator – no education is needed. Such
statement may seem a bit radical, especially when both the number and scope of
translation programmes at Polish universities are taken into account. It is true, however.
We could argue that the market verifies who a translator is. A translator sells his/her
services (not only linguistic – translators can also be advisors, DTP specialists, or
whatever the job requires them to be). Alas, there is more to that. There are people who
enter various projects and translate content for free, as voluntaries. There are people who
subtitle TV shows and put them on the Internet, receiving only recognition in return. The
list of examples would be long. Therefore, let us agree that for the benefit of this study a
translator will be understood as a human being translating for a living, whose language
Chapter 2. The role of a translator in the modern world
13
skills have been verified by the actual translation buyers. Such a person can be referred
to as a professional translator.
The professional vs non-professional distinction is rather important in terms of
discussing translator training process. Students attend translation courses at universities
around the world for a number of reasons. The author of the book runs a translation course
at the University of Łódź, called Computer Application in Translation. Interestingly, with
the beginning of each new academic year students are asked to answer a few questions,
the replies to which are to give the tutor an overview of their expectations. Each year
about a third of students replies that they want to become professional translators. The
rest wants either to improve their language skills or they chose translation track because
they do not want to teach English at schools. The first answer is quite obvious and rather
expected. Of the remaining two, the latter will not be discussed because it raises questions
about the reasons why they have chosen English philology. The former, however, is quite
interesting. Students claim that translation helps them master L2 language, a statement
that has found support in numerous research papers time and time again (see, for example,
Duff, 1994; Malmkjær, 1998; Leonardi, 2009; Dagilienė, 2012; Bahri and Mahadi, 2016;
Cho and Kim, 2017). In fact, translation as a method of language learning may contribute
not only to the development of such language skills as speaking, listening, writing and
reading, but also to development of translation competence in its broadest sense (see
Campbell, 2002; Pym, Malmkjær and del Mar Gutiérrez-Colón Plana, 2012; Pym, 2017).
English has become lingua franca of the modern world, and it is not important whether
one is Polish, German, Chinese or Arabic – international business relies on English.
Millions of people use it daily as their L2 language, frequently requiring translation.
These are not professionals in the light of the definition provided above. However, they
use the same skills in order to achieve a similar effect, becoming a sort of translators
themselves.
In contrast, professional translators are linguists, fully proficient in one or more
language pairs, profiting from extensive experience in translation and/or interpreting, and
verified by the market. Their main role is to render a message (textual or spoken) from
one language into another language. A professional translator is a language-service
provider (Pym, 2003), a person who can assist in the preparation of a document, advise
and assist a client in matters concerning both the source and translated content. S/he can
provide linguistic assistance (depending on the type of a document and its expected
target), counsel (e.g. during community interpreting sessions), technical services
(document preparation and editing), or any other service required by the translation job.
Contrary to non-professionals who frequently translate “by instinct,” professionals
are fully aware of the linguistic aspect of the trade, combining linguistic research (e.g.
mining for terminology) and practice (e.g. the use of specialised tools that augment the
translation process). They are cognizant of devices and strategies that may be employed
in order to convey a message between languages, frequently with cultural aspect in the
background. The cultural aspect of the trade is visible especially in interpreting, e.g.
business talks. Venuti (2008) discusses the notion of translator’s invisibility which is a
much easier code to live by in written translation than in interpreting. Clients frequently
ask translators for opinions or ask them “to negotiate” in their stead. Such situations are
not something decided upon by the translator, they are somewhat forced on them. Either
course of action requires a translator to be charismatic and strong enough to perform
according to expectations or refuse the job. If they choose to accept the job, they
frequently exert a real influence on the event, losing their ‘invisibility’. As a result, the
translator becomes a person, not a tool as some people would like to see him or her.
Interestingly, globalisation, the Internet and development of technology support
anonymity, especially in professions that do not require direct (in person) contact with a
client. Translation is one of them. Unlike interpreting which requires translators to meet
their clients, written translation can be largely performed from within four walls of one’s
apartment. A client needs not to see or hear the translator – their medium of
communication is via e-mail. It is fast, convenient, and stands up for an official record of
a conversation, so any arrangements made over e-mail can be binding. As a result, most
regular translators can be recognised nowadays mainly by their e-mail and name on an
invoice. Sworn translators are notable exceptions here as they have to meet their clients
since they are required to see the original copy of the translated document (TEPIS, 2017)
and the easiest way to do it is by presenting it in person.
To sum up, it can be said that professional translators can be divided into two
groups: visible (i.e. interpreters and sworn translators who, in a way, represent the
community in person) and invisible (i.e. regular translators who frequently have no direct
contact with clients. However, how do they affect the world? Why does the world need
them? The importance of translation around the world can be summarised in a few general
points:
Chapter 2. The role of a translator in the modern world
15
1. lack of proficiency in English
The first argument is the most obvious one. It is the primary goal of translation –
to enable one person to understand a message provided in a foreign language, or
two people to enter into a thought exchange with the help of a translator.
Interestingly, translation is all about communication, albeit it does not have to be
a two-way process. For example, audio-visual translation allows a Mr Smith, who
is British, comprehend a TV programme rendered originally in Portuguese. It is
an example of passive communication, i.e. working only one way.
2. preference for the native language
According to a survey by Common Sense Advisory (CSA), conducted on “more
than 3,000 global consumers in 10 non-Anglophone countries in Europe, Asia,
and South America, 75% [of respondents] prefer to buy products in their native
language” (CSA, 2014: online resource). Unless someone is bilingual or
multilingual, they will always find information provided in their mother language
easier to comprehend and handle. Hence the rapid growth of the localisation
industry, which helps to prepare global products and services enter local markets.
3. foundation for the global economy
Translation services allow not only local parties to benefit from foreign goods and
content but also local goods and content to be distributed abroad. For example, a
research conducted by Everline (a business e-lender) and the Centre for Economic
and Business Research reported over 880,000 UK-based companies planning to
expand overseas by 2025 (Everline, 2015). As a result, a fifth of all British
companies in England will require translation services to market and sell their
products and services abroad. Of course, the Brexit issue may affect their
decisions in this regard. Nonetheless, there is a huge demand on translation
market, not only in Poland, but worldwide.
4. lingua franca of the international trade
At the moment English is the dominant language of the international trade.
However, if we look at the chart of most numerous speakers per language, we can
see a bit different picture.
Speakers (in MLN)
1600
1400
1200
1000
800
600
400
200
0
Figure 1. The number of native speakers (including bilinguals) per language (Ammon, 2015; as
cited in the Washington Post, 2015)
The data presented in Figure 1 does not deny the role of English in the world, but
at the same time, it shows that it is not dominant if we take the number of native
speakers into consideration. If we consider point 2 above again, we can see that
anyone willing to do business or take leisure in Asia, for example, would have to
consider using translation services.
There are many countries in the world developing rapidly and entering the
global market. They need translation services into English, Chinese, German or
whatever market they want to make business with, but, on the same hand, the
target markets need translation services to make business with them. The need for
translation services in the EU alone is staggering at the moment, and all portents
show that it is going to grow (Pym, Grin, Sfreddo, and Chan, 2012).
5. a medium spreading ideas and information
The importance of translators in spreading ideas and information cannot be
denied. For example, while many scientific papers are prepared and published in
English, many require to be translated from scratch or revised due to poor
language proficiency of the authors. By merely translating, the translators help
shape the history of humanity.
What is more, translation helps preserve knowledge. During the Middle
Ages (or Dark Ages, as some call this period in the history of western civilisation)
much of the ancient knowledge was lost. It was Arabic translators who kept the
ideas of ancient Greek philosophers alive throughout the period. Had it not been
for them, all of this knowledge might have been lost.
Chapter 2. The role of a translator in the modern world
17
The fact is that the world as we know it would look much different if the translation
industry was smaller. The role of translation in the contemporary world is unprecedented.
On a micro scale, it enables people to use goods and services of foreign origin. On a
macro scale, it enables states to function, communicate with each other, make treaties and
enter economic partnerships. Heads of states, their representatives and other staff remain
in constant contact, mainly thanks to translators and interpreters. The EU institutions
alone employ around 4300 translators and 800 interpreters on its permanent staff
(European Union, 2017). At the same time, the US Bureau of Labor Statistics (20162017; online) reports that the
[e]mployment of interpreters and translators is projected to grow 29 percent from 2014 to
2024, much faster than the average for all occupations. Employment growth reflects
increasing globalization and a more diverse U.S. population, which is expected to require
more interpreters and translators.
Demand will likely remain strong for translators of frequently translated languages, such
as French, German, Portuguese, Russian, and Spanish. Demand also should be strong for
translators of Arabic and other Middle Eastern languages and the principal Asian
languages: Chinese, Japanese, Hindi, and Korean.
Globalisation and relatively easy human migration require the market to grow and to
support literary, technical, medical, legal, and regular translators, as well as interpreters.
Most of the people who know one or more foreign languages assume that they can
translate. Undeniably, to some extent such a statement is true. However, there is a
difference between doing something and doing it right. In order to do something right,
one either has to be gifted or train hard. Any successful training requires correct approach,
grounded in the most up-to-date theories and frameworks. Therefore, the discussion on
translation competence – a set of skills that enable an individual to produce successful
translations – has to follow in the next chapter.
Chapter 3. Translation competence
The previous chapter discussed the role of the translator in the modern world. It was
established that the translator is an individual who is either a linguist or proficient enough
in terms of language use that s/he can produce successful translations. The following
chapter introduces the concept of general and linguistic competences, and uses them to
build a context for translation competence, i.e. a set of skills that turn linguistically apt
individuals into translators.
3.1. Linguistic and general competence
Before the discussion on the specialised area of translation competence can begin, it is
vital to define the concept of competence, as set in the context of language studies. In a
general sense, translation competence can be understood as linguistic competence, i.e.
native speaker’s knowledge of their language (Malmkjær, 2009). In this regard, linguistic
competence is something we acquire unconsciously as we grow up and mature. Such
process consists in “interaction between (i) linguistic input data received by the languageacquiring individual and (ii) the default or initial state of the language faculty, which
contains innate knowledge called Universal Grammar” (idem: 122). Universal Grammar
features (i) “a set of innately endowed grammatical principles which determine how
grammatical operations apply in natural language grammars” and (ii) “a set of
grammatical parameters which impose severe restrictions on the range of grammatical
variation permitted in natural languages (perhaps limiting variation to binary choices)”
(Radford, 2004: 17). Carr (2006: 333) supplements this definition, saying that it is “a set
of semantic primitives, out of which specific word meanings are constructed.” Therefore,
it can be stated that linguistic competence is the ability to process and produce language,
basing on certain linguistic standards which surround us through our immediate society
and which we grow up with. It should be noted that linguistic competence, as opposed to
the broader concept of general competence, concerns inherent ability to use language, not
the act itself.
The difference between the linguistic competence and its broader concept has been
visualised by Malmkjær (2009: 123) in Table 1 below:
Chapter 3. Translation competence
19
Table 1. Competence in linguistics and other contexts.
General Competence
Linguistic Competence
is variable between individuals
is identical in each individual
can be improved by teaching
is acquired, not learned
includes aspects of performance
is opposed to performance
includes skills
is knowledge
The linguistic competence denotes native speaker’s knowledge of their language
(Radford, 1988; Malmkjær, 2009) while general competence is closely related to
performance or “the actual use of language in concrete situations” (Chomsky, 1965: 4).
The following section attempts to define the notion of competence in Translation Studies.
3.2. Translation competence
Schäffner and Adab (2000) agree that the 20th century brought a consensus amongst
experts in the field of Translation Studies that translating is a complex activity, involving
expertise in a number of areas and skills. Not only do translators need to have language
and technical skills to perform the task at hand, but also they need to possess the
knowledge, often very specialised, know-how and aptitude which allow them to produce
reliable translations (Kelly 2005, EMT Group, 2009).
Yet, what is translation competence? Can it be defined with any degree of certainty?
Lörscher (2012) reminds that the concept of natural translation goes back to Harris (1977)
and Harris and Sherwood (1978). It is defined as “the translation done by bilinguals in
everyday circumstances and without special training for it” (Harris, 1977: 99). Therefore,
Harris and Sherwood (1978: 160) consider translation to be an innate skill of any
bilingual. In this respect, translation competence is seen as an aspect of bilingualism.
Harris (1977, cited in Lörscher, 2012) elaborates that
[…] all translators have to be bilingual and […] all bilinguals can translate. [...] In addition
to some competence in two languages Li and Lj, they all possess a third competence, that of
translating from Li to Lj and vice versa. Bilingualism is therefore a triple, not a double,
competence: and the third competence is bi-directional.
Although such an approach may seem viable, the notion that all bilinguals can translate
has not been proven so far (see Grosjean, 2001). Malmkjær (2009) mentions research by
Toury (1984) and in Think Aloud Protocol Studies, which clearly suggest that translation
is a learned, not innate, skill. Such a claim can be proven by the fact that not all translators
are able to produce socially acceptable translations, especially considering the fact that
social norms change over time.
Toury (1986) proposes that the innate predisposition to translate is co-extensive
with bilingualism. In his opinion, translation competence does not develop “quasiautomatically” and in parallel with the development of bilingualism. In turn, bilingualism
is regarded to be a necessary factor in developing translation competence, but not
sufficient by any means. Toury (1986) argues that transfer competence, discussed further
on, must be built up in addition to an individual’s bilingual competence. It comprises the
individual’s ability to transfer texts equivalently on various levels according to a given
purpose/aim and with regard to sense, communicative function(s), style, text type, and/or
other factors; or to deliberately violate postulates of equivalence for a certain purpose
(Lörscher, 2012).
Hönig (1991), on the other hand, brings forth the concept of an ideal translation
process which is built on the model of translation competence composed of two main subcompetencies:
•
associative competence and
•
the competence to develop a “macro-strategy” (Hönig, 1991).
It is an interesting fact that the core principles of Hönig’s “macro-strategy” were later
incorporated in the strategic competence of the 2003 PACTE model of translation
competence (see section 3.2.1. in this book).
According to the model presented above, translators start by reading the source text
(ST), but their reception of the text is completely different from that of an ordinary reader.
Hönig (1991) advocates the idea that the “text reception is influenced by the translation
task they have in mind” (Göpferich, 2009: 15). The source text is then processed by the
translator, both consciously and unconsciously. The analysis of the structure, style,
Chapter 3. Translation competence
21
content and expectations with regard to the text lead to the development of a macrostrategy. Such macro-strategy may develop “more or less automatically on the basis of
the translator’s professional experience” (idem.). It is desirable that the macro-strategy
develops before the actual translation takes place, but it is not always the case. The
process of translation follows, which is supported by rules and strategies based on actual
experience of the translator. At this stage the second sub-competence is employed – the
associative (or transfer) competence (Hönig, 1991: 80). When both sub-competencies are
combined, it is possible to arrive at a target text eventually.
The example of Hönig’s model of an ideal translation process (Figure 2, next page)
is a good illustration of the complexity of the issue. Pym (2003) discusses the notion of
translation competence as a puzzle that linguists have been deliberating on for the last 30
years. The issue is that it cannot be clearly defined since it can mean completely different
things depending on the background and experience of a given linguist. Pym (idem.)
brings up a number of different approaches, ranging from a claim that there is no
translation competence at all (Lörscher, 1991: 2), the notion of one supercompetence
(Wilss, 1982), to many sub-competencies involved in the actual process of translation
(PACTE, 2003). The last approach seems reasonable due to its flexibility. It is not fixed
and can adjust in parallel with the needs of the market. This way its components remain
current and the obsolete ones are removed.
The minimalist approach proposed by Pym (2003) focuses on the translation itself
as the main constituent of the supercompetence. He proposes the following two-fold
functional definition of competence (Pym 1991: 489):
The ability to generate a series of more than one viable target text (TT1, TT2 … TTn) for a
pertinent source text (ST) [This corresponds to what Hönig calls associative competence.];
and
The ability to select only one viable TT from this series quickly and with justified confidence.
[This corresponds to Hönig’s macro-strategy and the ability to employ it consistently.]
Figure 2. Hönig’s model of an ideal translation process (1991: 79)
What is important in such a division is the fact that it advocates translation as the act
itself. Pym (2003: 489) claims that “the […] translational part of their [translators’]
practice is strictly neither linguistic nor solely commercial”. At the same time, he
acknowledges that translators require to “know a fair amount of grammar, rhetoric,
Chapter 3. Translation competence
23
terminology, computer skills, Internet savvy, world knowledge, teamwork cooperation,
strategies for getting paid correctly” (idem.) and more. These skills (or rather sets of
competencies) often decide on the success of a translation project, even though they are
not strictly concerned with translation themselves.
Therefore, it could be argued that translation competence is a notion that covers all
kinds of skills and knowledge required (to a lesser or greater extent) to produce correct
and successful translations, the total sum of translator’s knowledge required to render a
text in a target language. In this light, the set of sub-competencies, as proposed by PACTE
(2003), should be seen as a set of major categories, by no means constant ones, that could
(and should) be divided further.
The multifarious aspect of translation competence, advocated in this book, has been
shared with variations by a number of academics around the world (among others,
Hejwowski [2004], Neubert [2000], Kelly [2005], and Shreve [2006])5. The most widely
recognised models of translation competence have been those of the PACTE and EMT
research groups, as well as for the TransComp project.
The PACTE model, proposed by the PACTE Group (2003) lists bilingual, extralinguistic, instrumental/professional, psycho-psychological, transfer and strategic subcompetencies. The EMT model, proposed by the EMT Group (2009) lists translation
service provision, language, intercultural, info mining, technological, and thematic
competencies. The TransComp model, proposed by Göpferich (2009: 21-23), lists
communicative, domain, tools and research, translation routine activation, psychomotor
competence, and strategic competencies.
3.2.1. The PACTE model of translation competence
The PACTE acronym stands for Process in the Acquisition of Translation Competence
and Evaluation, and it serves as a name for a group of researchers concerned with the
acquisition of translation competence in written translation. The group was formed in
1997 and is based in Barcelona, Spain. It was a member of the Institut de Neurociències of
the Universitat Autònoma de Barcelona (2001-2009) and is now a member of GReCO
(Grup de Recerca en Competències) of the Universitat Politècnica de Catalunya.
PACTE initiated the Thematic Network TREC6 (Translation, Research,
Empiricism, Cognition), which brings together experts in empirical and experimental
5
6
See Piecychna (2013: 143)
Visit http://pagines.uab.cat/trec/ for more details
research in translation from around the world, and coordinated it for two years. Their
research concerns translation both in and out of the foreign language, and is primarily
based on their own language combinations, i.e. Spanish and Catalan ↔ English, French
and German. Being a heterogeneous group linked by “the need for more information
about how trainee translators learn to translate in order to create better teaching
programmes, improve evaluation methods and unify pedagogical criteria” (PACTE,
2003), they proposed a translation competence model based on their own experience and
conclusions. The PACTE model has been used by researchers around the world as a
reference model ever since.
In order to attempt to propose a unified model of translation competence, the
researchers analysed the issue in the context of existing models and empirical research.
Not only did they use research on translation, but also they tried to approach the problem
from different angles (e.g. considering translation as an act of communication, they tested
their data against communicative competence). As a result, they decided that all previous
translation competence models were not based on “validated empirical research”
(PACTE, 2003), i.e. there was no data which could be used to define translation
competence components, if any, and connections between those components. The group
decided to provide such research. They started with drawing a line between “competence”
(the knowledge in the background) and “performance” (the actual process of translating).
What is more, they assumed that translation competence is not the same as bilingual
competence (which is, in fact, one of the several sub-components of translation
competence), which the concept of natural translation could suggest.
What is more, they assumed that each successful translator has to have ‘expert
knowledge’ and be able to devise a ‘strategy’ to deal with a given translation job (PACTE,
2003).
When summed up, their basic assumptions were:
•
Translation competence is qualitatively different from bilingual competence;
•
Translation competence is the underlying system of knowledge needed to translate;
•
Translation competence is an expert knowledge and, like all expert knowledge,
comprises declarative and procedural knowledge; the latter is predominant;
•
Translation competence is made up of a system of sub-competencies that are interrelated, hierarchical and that these relationships are subject to variations.
Chapter 3. Translation competence
•
25
The sub-competencies of translation competence are considered to be: a language
sub-competence in two languages; an extra-linguistic subcompetence; an
instrumental/professional sub-competence; a psychophysiological sub-competence;
a transfer sub-competence; and a strategic sub-competence.
(PACTE, 2003)
The idea behind such division was that translation is a process involving a number of
various skills that we either develop or acquire. Some of them are practical, others refer
to our general knowledge. When they work in unison, people are able to translate
effectively.
Nevertheless, it does not mean that they are of equal importance. For example,
strategic competence is needed since it compensates for all shortcomings of other
competencies, and instrumental competence is not always required for some types of
documents. What is more, in their model PACTE allowed for variations in the scope and
role of competencies, depending on context (time, place, technical conditions, etc.),
translator’s individual experience (years worked, familiar domains, etc.), language
combination, specialisation (e.g. technical, legal, etc.), and directionality (to/from a given
language). The reasoning behind such freedom was based on the following examples:
•
in inverse translation the instrumental/professional sub-competence gains
importance;
•
the strategies used by the translator vary according to the distance between the
language pairs used in the translation;
•
in each translation speciality greater importance will be given to different
psychological abilities (logical reasoning in technical translation, creativity in
literary translation);
•
a greater degree of automation may be expected when the translator is very
experienced;
•
the translation context (translation brief, time, etc.) may require a certain
subcompetence to be activated (instrumental/professional, psycho-physiological,
etc.).
(PACTE, 2003)
Flexibility of this model reinforces its significance. The notion of translation competence
is so complex that it cannot be defined in simple, limited terms due to the fact that every
human translator is an individual and their individual translation competence will vary
from person to person.
The outline of basic assumptions behind translation competence was not enough.
The model would be incomplete without a study on how to build, or acquire, translation
competence. PACTE described the problem as a process (conscious or unconscious) to
develop, restructure or refine existing skills, or competencies. Thus, the group’s definition
of translation competence acquisition is:
1.
A dynamic, spiral process that, like all learning processes, evolves from novice
knowledge (pre-translation competence) to expert knowledge (translation
competence); it requires learning competence (learning strategies) and during the
process both declarative and procedural types of knowledge are integrated,
developed and restructured.
2.
A process in which the development of procedural knowledge and, consequently, of
the strategic sub-competence are essential.
3.
A process in which the translation competence sub-competencies are developed and
restructured.
(PACTE, 2003)
Again, the acquisition model also does allow for variations, as each learning process is
slightly different. PACTE refined their definition by adding:
1.
[sub-competencies] are inter-related and compensate for each other;
2.
[sub-competencies] do not always develop in parallel;
3.
[sub-competencies] are organised hierarchically;
4.
variations [between sub-competencies] occur in relation to translation direction,
language combinations, specialisation and the learning context.
(PACTE, 2003)
It means that the acquisition of translation competence is subject to external factors like
teaching methods, or the learning context (learning techniques, form of the course, etc.).
It also means that the process varies not only between individuals, but also between
languages (e.g. learning pace may be different when translating from foreign to native
language than vice versa). What is more, it may be the case that some sub-competencies
take precedence over others, depending on the text specialisation (technical, law, etc.).
And again, due to its flexibility, the model is successful.
Chapter 3. Translation competence
27
The research based on these principles, and conducted by PACTE (for details see
PACTE, 2003) led to the formulation of basic sub-competencies of the PACTE model of
translation competence. The sub-competencies are defined as follows:
• Language sub-competence in two languages is occasionally referred to as
bilingual competence (PACTE, 2005). That does not necessarily mean that a
successful translator has to be bilingual. It merely hints that his socio-linguistic,
lexical-grammatical, textual and pragmatic knowledge in each language should
be equal to that of a true bilingual. As it is, it can be summarised as “the underlying
system of knowledge and abilities necessary for linguistic communication in both
languages” (PACTE, 2003). For communication to take place at all, regardless
whether in oral or written form, the message has to be understood and then a new
one produced. It is what the language sub-competence allows for. Apart from that,
it is the most basic element of translation competence, critical to the ability to
translate between languages.
• Extra-linguistic sub-competence is understood as implicit or explicit knowledge
about the world in various areas of knowledge, i.e. general and specific areas of
knowledge, knowledge about translation, bicultural knowledge, encyclopaedic
knowledge, and subject knowledge (see PACTE, 2003; Ressurrecció, Piorno and
Izquierdo, 2008).
• Instrumental/professional sub-competence can be defined as the knowledge
and abilities used for practising translation, i.e. using documentation, various
sources, new technologies, knowledge about the market and its condition
(PACTE, 2003: 93; 2005: 619). This competence encompasses secondary
translator skills, the ones that are optional for the translation itself, but vital for
the overall success of any given translation today. Competition between
freelancers on the global market and pursuit for translation quality in the era of
the Internet do force translation professionals to search and employ new tools
(ranging from vocabulary to translation management systems) and solutions in
their everyday work. This sub-competence is especially important as it covers
technical skills which are increasingly often required by outsourcers7.
7
Based on personal experience of working as project manager in translation agency in Łódź between
years 2008-2015
• Psycho-physiological sub-competence includes cognitive and behavioural
(memory, attention span, perseverance, critical mind, and so on), and
psychomotor mechanisms.
They include: (1) cognitive components such as memory, perception, and attention
and emotion; (2) attitudinal aspects such as intellectual curiosity, perseverance,
rigor, critical spirit, knowledge of and confidence in one’s own abilities, the ability
to measure one’s own abilities, motivation, etc.; (3) abilities such as creativity,
logical reasoning, analysis and synthesis, etc.
(PACTE, 2003)
It encompasses all traits that would allow an individual to produce a high-quality
translation.
•
In 2003 the PACTE Group refined their 1998 model of translation competence,
which considered transfer, or translation, sub-competence as the actual ability
to render a source text in the target language or to understand the source and
produce the target, “taking into consideration the purpose of translation and the
characteristics of the receptor” PACTE 2003: 48). The new approach defined it as
a “combination of all the sub-competencies, i.e. translation competence: the
ability to carry out the transfer process from the source text to the production of
the target text in function of the receptor’s needs and the purpose of the
translation” (p. 57). As a result, the transfer sub-competence gained recognition
as the total sum of other sub-competencies that allows to produce a successful
translation.
Such definition is supported by Shäffner (2000: 148) who defines transfer
sub-competence as one “specific to translation and [integrating] all the other subcompetences.”
•
Strategic sub-competence is responsible for solving problems and the efficiency
of the process of translation (PACTE, 2005). It is employed throughout the entire
process and is primarily concerned with:
➢ planning and overseeing the project at various stages; activating other subcompetencies when needed;
➢ predicting, finding and problem-solving;
➢ compensating for deficiencies in other sub-competencies.
Chapter 3. Translation competence
29
The PACTE model of translation competence has gained wide recognition in the world
of Translation Studies. This book focuses on one of its sub-competencies – the
instrumental/professional competence – which is what allows translators to make use of
sophisticated computer software.
3.2.2. The EMT model of translation competence
Since its inception, the European Union had a huge demand for quality translation due to
its multi-national form. Various solutions were implemented to meet this demand,
including machine translation (e.g. EUROTRA, see Commission of the European
Communities, 1994, for more details). Even the ALPAC report did not discourage MT
development in Europe to the extent it did in the USA. However, MT was not enough to
deal with the scale of the problem. In the EU human translation has always been of critical
importance due to communication and legal requirements. The rapid growth of the EU
between 2004 and 2007 (12 new countries joined in) presented the need for unification
of translator training programmes and setting a set of minimum requirements for
translators in order to set a certain standard. In order to address the issue the European
Master’s in Translation (EMT) programme was founded. It is a partnership project
between the European Commission and higher-education institutions offering master’s
level translation programmes. The EMT not only regulates the rapid evolution of the
profession but also sets standards to be included in translation programmes within the EU
(due to their diversity and potential incompatibility [see EMT Expert Group, 2009a, for
more information]). The abovementioned standards were proposed by the EMT expert
group working on the EMT programme, and formed on the initiative of the DirectorateGeneral for Translation (the European Commission’s in-house translation service) in
April 2007 (Chodkiewicz, 2012). The main goal of the group was to produce:
•
a generic description of the tasks and competences of translators to match the needs of the
translation industry and public bodies, such as the EU institutions;
•
a draft of a European model curriculum that addresses these requirements and could thereby
enhance the status and quality of the translation profession
(EMT, 2009b: 1)
As a result, two documents were drawn. The first one involved the list of competencies
that should be acquired. The second one listed criteria for admission of the translation
programmes by the EU leading universities to the EMT network (Chodkiewicz, 2012).
From now on, all universities, who wanted to promote their own translation programmes,
had to meet certain conditions which would qualify them to join the EMT network and
use the EMT logo.
The EMT model of translation competence “revolves around the notion of
Translation Service Provision Competence” (Klimkowski, 2015a). Such approach is
consistent with the notion of translator being not exclusively an individual who translates,
but rather a language-service provider (Pym, 2003). Therefore, the EMT model highlights
the prominence of the business aspect of the profession of translator, i.e. providing full
language service which includes translation.
The EMT model, or framework as the EMT group calls it, defines translation
competence as a “combination of aptitudes, knowledge, behaviour and know-how
necessary to carry out a given task under given conditions” (EMT Expert Group, 2009a).
In short, it combines all the skills that make translator a language-service provider. Those
sets of skills, or sub-competencies, are quite similar to the PACTE model, with the
exception of the translation service provision component, of course. The importance of
the Translation Service Provision Service can be clearly seen in the EMT group’s
visualisation of their framework:
Figure 3. Graphic representation of the EMT competence framework (EMT Expert Group, 2009a: 4).
Unlike the PACTE model, which, according to Klimkowski (2015a: 70) can be
questionable due to the fact that it represents “mainly declarative kind of knowledge,”
and raises doubts from both professional and didactic perspective, the EMT framework
presents a “very detailed list of skills” (ibid., p. 70), based both on declarative and
Chapter 3. Translation competence
31
procedural knowledge. However, Klimkowski (idem.) argues that the framework should
not be treated as a fixed point of reference in the process of translator training. He suggests
that
it is rather an indication that the proposals like the ones discussed above should be interpreted
as a point of departure for developing educational solutions tailored to the needs of a given
learning community (students, teachers, academic institutions, representatives of the local
markets, etc.)
(Klimkowski 2015a: 71).
This statement is in accord with the fact that the profession of translator undergoes
evolution (Pym, 2013), and any attempt at setting a fixed framework is doomed to failure
in the long run. The profession has to evolve in order to accommodate market
expectations. It is true for the vast majority of specialised professions. Globalisation has
warped space-time of translation, producing a hybrid profession combining competencies
of the translator, manager, researcher, IT specialist, etc. As a result, competencies evolve,
rising to prominence depending on market and professional conditions.
Nevertheless, the framework has gained generally positive evaluation from
language researchers, mostly due to its accuracy, flexibility and focus on service
provision. The core competencies, named by the EMT Expert Group, include:
•
Translation service provision sub-competence is divided into two dimensions
which, combined, encompass many competencies covered by previous models of
translation competence (Chodkiewicz, 2012). The two dimensions are:
interpersonal and production dimension. The interpersonal dimension concerns
translator’s social role (see Kiraly, 2003), everyday relations between translators
and their clients (as set in the context of market demand and marketing/selfpromotion strategies). The dimension also includes teamwork and adherence to
professional standards (Chodkiewicz, 2012; PACTE, 2003). The production
dimension, on the other hand, concerns the actual process of translation in
accordance with the current conditions, i.e. client’s preferences. Sometimes
translators may be required to make some decisions8 for the client on the one hand,
and justify them using appropriate language on the other (see PACTE, 2003).
8
Klimkowski (2015a) suggests that it is fairly common for a translator to support a customer in many
ways, including decisions regarding the translation itself and the fact that the “clients […] expect us
•
Language sub-competence, similarly to the 2003 PACTE model, is the
background knowledge and abilities which enable to start and carry out a
communication process. It is critical to the ability to translate.
•
Two-dimensional
intercultural
sub-competence
includes
intercultural
dimension, or extra-linguistic competence similar to the one found in the PACTE
model (PACTE, 2003); and textual dimension which is responsible for “the ability
to analyse the macrostructure and coherence of a text and reproduce it according
to the conventions of a particular genre and rhetorical standards” (Chodkiewicz,
2012: 40). It is one of the challenges of translation to read a text, process it and
translate accordingly.
•
The next competence listed in the EMT model concerns information mining. It
includes all skills required to find information through plain search (e.g. Google
search engine) or research (online and offline corpora, terminology, professional
support [ProZ, help of experts in a given field], etc.), and the knowledge how to
use this knowledge in the context of current translation project (see also Nord,
1991; PACTE, 2003).
•
Technological sub-competence is similar to the instrumental/professional subcompetence in the 2003 PACTE model. It covers the ability to acquire, master and
use both software and hardware solutions which assist in translation (increasing
quality and productivity), research (methods to access and process data), deliver
completed projects (cloud, FTP servers), and more (see PACTE, 2003).
•
Thematic sub-competence, on the other hand, concerns understanding the
document and its theme better (Chodkiewicz, 2012). Any given translation is
focused specifically on a given document and it is critical for any professional
translator to develop their knowledge, skills, terminology, etc. in relation to the
theme of the document (see also Nord, 1991; PACTE, 2003).
Chodkiewicz (2012) points out that some competence components of the EMT model
were not included in the previous models. The skills in question are practical in nature,
mostly very specific, and falling into the category of translation as a language-service
provision (document pre- and post-processing, meeting deadlines and working underpressure, or adhering to client’s specific requirements, to name a few), as well as
[translators – MK] to offer them all kind of support in text construction, reviewing business
presentations, hearing the client’s employees deliver their presentations as training before their
business conferences, etc.” (Klimkowski, 2015a: 71).
Chapter 3. Translation competence
33
information mining (finding and evaluating the quality of one’s sources). Apart from that,
the model seems to pay more attention to technological aspect of translation competence
(knowing pros and cons of machine translation, and the knowledge how to use it in
translation). The same applies to the intercultural dimension, and to its textual dimension
in particular – the model stresses the importance of the actual editing of the content of a
document (EMT Expert Group, 2009: 4-7).
The key difference between EMT and previous models of translation competence
is that the competencies presented here are completely reorganised and refocused on
specific practical, and market-oriented, skills. Contrary to the 2003 PACTE model, there
is no transfer sub-competence. What is more, it is not focused on language or strategic
competence, but rather on the overall ability to provide a client with high-quality
language-service that translation is a part of (Chodkiewicz, 2012).
In order to understand the importance of such re-prioritization of translation subcompetencies, it is necessary to have actual experience with translator-client relations.
There is a saying on the Polish translation market that you can have your translation fast,
cheap, and good – pick any two. It seems that market demand and competition between
translators did revise it. Today quality is a must-have, short deadlines are expected, and
price can be accommodated. While translators use the Internet to mine for data, clients
use it to mine for translators, better [non-translation] service and lower translation rates.
Hence the importance of the translation-as-a-service approach. It is especially true of
sworn translators who meet clients face to face more often than freelancers working from
home. On such occasions, sworn translators have to adhere to the rules governing the
profession (e.g. 2005 Sworn Translator’s Code, see TEPIS, 2017) and social etiquette
more closely since they advertise their services with their own person. What is more, due
to their professional experience, frequently they are expected to help and advise clients
on documents, their translation and future use. The translator is seen as a professional
who “knows” and whom people trust with their documents so that they will not only
receive a quality translation service, but quality translation services tailored to their
particular needs9. Clients expect translators to know how local offices work, what they
pay attention to (layout-wise) and what sort of translation they will accept (languagewise). They are not interested in the process, but in the effect. Thus, a successful translator
is required to provide a full service, and translation is only one of its elements.
9
Based on personal experience of working as project manager in translation agency in Łódź between
years 2008-2015
Inclusion of this sub-competence in the EMT framework assures that all the partner
universities include it (at least partially) in their translation programmes, thus drawing
students’ attention to the problem of service provision and client-translator relationship
(e.g. quoting, composing coherent and unified quote messages to clients, addressing
complaints).
3.2.3. The TransComp10 model of translation competence
The last model of translation competence to be outlined in this paper was prepared for the
TransComp research on the development of translation competence. The research was
conducted by Susanne Göpferich and was conducted on 12 students of translation and 10
professional translators with at least 10 years of professional experience in translation
and/or interpreting.
The TransComp model is predominantly based on the 2003 PACTE model of
translation competence, as well as Hönig’s model of an ideal translation process (see
Hönig, 1991) and Kiraly’s psycholinguistic model of the translation process (see
Kiraly, 1995).
Figure 4. Göpferich’s translation competence model (Göpferich, 2009: 21)
The model includes the following sub-competencies (Göpferich, 2009: 21-23):
•
Communicative sub-competence in at least two languages is akin to the
PACTE group’s “bilingual subcompetence.” Apart from lexical and grammatical
knowledge, it includes pragmatic knowledge as well (“the knowledge about genre
10
TransComp was a research project focused on the development of Translation Competence. It was
based on English and German, and supervised by Susanne Göpferich. It was funded by the Austrian
Science Fund (project No. P20908-G03 (September 2008–August 2011).
Chapter 3. Translation competence
and
situation-specific
35
conventions
in
the
respective
cultures”).
The
communicative aspect of this sub-competence is best seen in source-text reception
(comprehension) and target-text quality. In unison, they allow a successful
translator to find and evaluate semantic equivalence between source and target
units, a most useful skill in the process of segment aligning in CAT tools.
•
Domain sub-competence relates to the extra-linguistic (general and domainspecific knowledge) competence, as suggested by the PACTE Group. In this
model, however, it is necessary to understand the source and produce the target
text, and assess what sort of external knowledge is required to complete a given
project (and where to find it).
•
Tools and research sub-competence is the same as the instrumental/professional
sub-competence to be found in the 2003 PACTE model. It “comprises the ability
to use translation-specific conventional and electronic tools, from reference works
such as dictionaries and encyclopaedias (either printed or electronic), term banks
and other databases, parallel texts, the use of search engines and corpora to the
use of word processors, terminology and translation management systems as well
as machine translation systems” (Göpferich, 2009: 22).
•
The translation routine activation sub-competence has not been outlined in the
previous models. It comprises the knowledge and the abilities to use previously
acquired, language-pair specific, skills in order to carry out specific transfer (or
translation) operations, ultimately leading to acceptable target units.
•
Psychomotor abilities required for reading and writing (with electronic tools)
constitute the psychomotor sub-competence. Note that these skills concern the
actual ability to read and write, not to use the software itself (which falls into tools
and research category). The general idea is that the more advanced these skills
are, the less attention is required to the process of reading and writing. As a result,
more cognitive capacity is left for other, more demanding tasks. This somewhat
pictures a translator as a kind of machine with limited resources. Refinement of
one process reduces its resource consumption and allows other “processes” to use
the freed resources in order to achieve better overall quality.
•
Finally, strategic sub-competence is the one that controls the others. It prioritises
tasks and employs other sub-competencies when necessary. Thus, it leads to
development of higher-level strategies for future use. This competence
corresponds to the strategic competence by the PACTE group (2003).
3.2.4. Comparison of the presented models
All presented models draw up from the previous research on translation process and
translation competence. PACTE’s ongoing research concerning translation competence
has been widely recognised and accepted; at the same time, the EMT Expert Group went
further by focusing on translation as a language service in order to accommodate actual
market expectations. TransComp, on the other hand, can be seen as a combination of the
two in that it is closely based on the 2003 PACTE model and the more practical aspects
of the EMT model, like translator’s social responsibility, or working under pressure
(Göpferich, 2009). The table below presents all three discussed models. While the central
sub-competencies of each model have been highlighted with small caps, the one
competence that is virtually the same in all three models has been highlighted with bold
face.
Table 2. PACTE, EMT, and TransComp translation competence models compared
PACTE
language
sub-competence in two
languages
extra-linguistic sub-competence
instrumental/professional
sub-competence
psycho-physiological subcompetence
transfer sub-competence
STRATEGIC SUB-COMPETENCE
EMT
TransComp
TRANSLATION
communicative subcompetence in at least two
languages
domain sub-competence
tools and research subcompetence
translation routine activation
sub-competence
psychomotor sub-competence
STRATEGIC SUB-COMPETENCE
SERVICE PROVISION SUBCOMPETENCE
language sub-competence
technological
sub-competence
intercultural sub-competence
information mining
thematic sub-competence
As can be seen, all models agree on the importance of technology and list it as one of the
key competencies to be had by a successful translator. This set of computer-bound skills
is critical for any language professional working in the contemporary market. The key
competence of the EMT model, translation service provision sub-competence, would be
unattainable had it not been for technological means (both tools and skills) which serve
as its founding blocks.
The next section attempts to outline the instrumental/professional sub-competence
if we use PACTE [2003] nomenclature.
Chapter 3. Translation competence
37
3.3. Translation competence vs. translator’s computer competence
The previous discussion outlined sub-competencies of three prominent models of
translation competence. It has been established that one of them, related to tools and
technology, is virtually the same throughout the three models. Since all the skills included
in this sub-competence are computer-related, it is only natural to use Kiraly’s (2000)
suggestion and refer to them collectively as “translator-relevant computer competence,”
hereinafter referred to as TCC. Kiraly (2000: 125) says that “the goal [of the translator
training process] would be to make sure that students acquire the objectively identifiable
set of sub-skills that together comprise translator-relevant computer competence, which
they should be able to apply to real translation tasks once they leave the institution.” All
the models outlined above agree on the importance of the computer skills. Kiraly (2000)
points out that it is possible for a teacher, in extremis, to transfer only those skills that
s/he believes they will need, and the majority of those are purely practical skills involving
computer hardware and software. Of course, such extremely technological approach may
not be valid for a general translation course curriculum. Nonetheless, it shows that
technology should not be disregarded since it contributes heavily to the overall success
of a translation trainee. Brooks (in Bowker, 2002: 4) supports this claim by saying that
(…) experience has shown that graduates who are conversant with CAT technology are a a
real advantage when it comes to working in highly technologized translation environments
such as the software industry and the organisation of the European Union. This trend is also
increasing in other sectors, such as the telecommunications and automotive sectors.
Brooks (idem: 57) adds that
(…) the development of CAT tools has not reduced the need for language professionals.
Rather, improvements in efficiency and reductions in translation costs brought about by the
use of these tools have stimulated a demand for localised products and created jobs for
translators who are skilled at using technology.
Technology is no longer an asset outside translation. Some of its aspects may be of
secondary importance, but other have become essential for translation to exist in the
modern sense at all.
With that in mind, the discussion on TCC has to allow for the fact that it is a very
specialised (though quite extensive) area of expertise, which deals primarily with
proficiency in utilising computer hardware and software for the purposes of translating.
The goal of these skills is to support translators in their work, not translate for them.
3.3.1. Computer competence skills
Before venturing into particular subsets of TCC, it is worth asking if this kind of
competence is relevant for a translation course.
In his paper, Redefining Translation Competence in an Electronic Age, Anthony
Pym (2003: 481-482) presents the problem his students voiced in relation to his Advanced
Translation class.
[W]e [students] are not really translating […]. But I quickly reply, we have learned how to
use Revision tools and comments in Word; we have discovered a few good tricks for Internet
searches; we have found out about HTML; we can create and localise fairly sophisticated
websites; we can do wonderful things with translation memories… and these are the things
that the labour market is actively looking for. All that, I insist, is part and parcel of translating
these days. No, some still reply, what we want is lists of false friends, modulation strategies,
all the linguistic ricks, plus some practice on a few really specialised texts… and that, my
more critical students believe, is the invariable hard core of what they should be learning in
the translation class.
The problem he faced is quite common in practical translation classes. Does it mean that
teachers of translation should take primarily students’ expectations into consideration?
The answer is yes, and no. Yes, because translation classes, regardless of the type of skills
they try to teach, should be primarily focused on the translation itself and any class devoid
of translation component should not be treated as a translation class. On the other hand,
the role of translation trainers11 is to prepare their students to start a professional career
after graduation, which means entering highly competitive market, which demands
certain extra-translational skills. Pym (2003: 493-494) says that “technology will always
be one or two steps ahead of any multicomponent list. […] [E]lectronic tools are simply
techniques that speed up and broaden the production of alternative target texts” or are
used to eliminate them. It is not possible to introduce a universal set of skills to be taught
11
The word “trainer” is understood here as a person teaching translation as a profession, and not as an
academic discipline.
Chapter 3. Translation competence
39
since translation market and its demands are in constant shift. One could argue that it
would be more viable to analyse current trends and try and foresee how they will change
in the near future. Above all, it is important to consider students as individuals, their
expectations, and then adapt their translation courses accordingly.
Having said that, let us go back to the issue of the TCC. When asked about the place
of computer skills in translation, we can say that they are not vital to the process of
translation in the sense that translation can be done without them, using pen and paper.
However, such form of translation is neither fast nor quality-oriented. Kiraly (2000: 123124) argues against such form of translation classes since the handwritten translation is
inferior to computer-based translation in almost every way. What is more, all three
outlined models of translation competence name technology as one of the competencies
of a translator. In order to explain the problem, a line should be drawn between translating
for educational purposes, and commercial, or professional, translation. In the first case,
all means and methods of translation can be justified, provided they yield an acceptable
outcome in the end. In the second, only those methods of translation are justified which
enable translators to deliver high-quality translations in short periods of time and stay in
line with the competition on the market. Here computer skills show their value. In order
to survive on the market, “both freelance and staff translators use computer-based
translation memories, desktop publishing software, access to the Internet and a variety of
on- and off-line electronic resource” (Kiraly, 2000: 123). At the moment of writing this
book, the above words by Don Kiraly are 16 years old. Moreover, if anything, they are
even more true today than they were when Kiraly wrote them. Globalisation, the Internet,
new CAT tools, and cloud services push demands put on the translator even further.
Translation quality, speed, and flexibility (e.g. ability to work in groups or to work with
online translation packages) have become traits not so much desired, as expected of the
translator by translation outsourcers12, and constitute an indisputable asset for any
translation graduate.
Bogucki (2009: 52) names three key categories that base on computer application
in translation: word-processors, electronic dictionaries and glossaries and translation
memory-based software. While this list is still current, the author believes it should be
further extended to encompass a broader range of skills/applications that are in everyday
12
Based on personal experience of working as project manager (PM) in translation agency in Łódź
between years 2008-2015 and feedback from other PMs (Agata Sadza, Translateria; Jakub Pięciński,
Lionbridge).
use. Especially so when Pym’s (2003) idea of translator as a language-service provider is
considered. The need to update the list is also clearly visible through the prism of client
expectations, as described by Risku, Pein-Weber and Milošević (2016). Clients expect
the work to be done and are not necessarily concerned with the process itself. Therefore
it is up to translators to decide whether they want to work with their clients directly or
indirectly through an agent (a translation agency), who will consume part of expected
profit. It is a question of business specialisation: one can focus solely on translation as a
freelancer or go one step further and become translator-project manager, undertaking
larger projects that require translation, advanced client management, work distribution
(management of other freelancers), document handling, and quality assurance (Sikes,
2011; Straub, 2013; Risku, Pein-Weber and Milošević, 2016). The proposal of updated
list of computer application in translation (as language-service provision) can be found in
Table 3 below.
Bowker (2002: xxii) argues that “in its broadest definition CAT technology can be
understood to include any type of computerised tool that translators use to help them do
their job. This could encompass tools such as word processors, grammar checkers, e-mail,
and the World Wide Web (WWW).” Building upon this thought, we can say that any
technology-based tools that help to translate can be called CAT. Alas, such approach may
be too broad. Therefore, Bowker proposed to differentiate skills between human
translation (HT), computer-assisted translation (CAT), and machine translation (MT).
Table 3. Bowker’s overview of technology used in translation
HT
• Word-processors
• Spelling and grammar
checkers
• Electronic resources (e.g. CDROMs)
• Internet (e.g. WWW, e-mail)
CAT
• Data-capture tools
• Corpus-analysis tools
• Terminology-management tools
• Translation memories
• Localisation and Web-page
translation tools
• Diagnostic tools
MT
• Machine
translation
systems
Some of the technology mentioned in the table above reaches beyond the area of
translation, as anyone can use OCR tools to digitalise documents, for example. What is
more, the list is not, and will never be, complete as it is dictated by demands of the market
and the ever-evolving image of the translator as a language-service provider (Pym, 2003).
Chapter 3. Translation competence
41
While such approach seems valid, the goal of the list below is to combine skills of
the translator (i.e. ones which facilitate, or enable, in extreme cases translation) and of a
language-service provider. It has been compiled from personal experience as a translator,
and the list of computer skills found most desirable in apprentices who wanted to undergo
an apprenticeship in the translation agency the author worked in13:
Table 4. List of translator-relevant computer competence skills
AREA OF EXPERTISE
SKILLS
Computer hardware
rudimentary computer handling (storage devices; Internet
connection, e.g. in client’s office; etc.)
advanced computer tools handling (simultaneous translation
equipment handling)
text-formatting (word-processors; preparing proper, high-quality
source for translation)
handling various file formats (identifying file extensions and
opening software; document conversion)
handling optical character recognition (OCR) systems
online sources (terminology, machine translation, online file
converters, etc.)
CAT systems (use of, translation memories, term bases, server
projects, CAT revision tools)
localisation software
subtitling software
DTP and graphic software
quoting of documents (processing character-, line-, or wordcount;
CAT-based document analysis and pricing)
e-mail based communication (presenting quote to a client; processing
orders; processing complaints; establishing a strong business
relationship with the client)
marketing
company management (e.g., invoicing)
Computer software
Service provision
The list presented above denotes certain areas of expertise since the expanse of the
Internet, the amount of different software available on the market, and new demands put
on translators every day would make it impossible to name and discuss each and every
aspect of TCC. Nevertheless, some examples should be given so as to provide insights on
what these categories include and how they correspond to the notion of translator’s
computer competence.
13
Project manager in Translateria Translation Agency, 2008-2015
Hardware: Rudimentary computer handling
Rudimentary computer handling, as the name suggests, is all about handling physical
tools a translator may require. It is a set of skills allowing to operate storage devices (USB
sticks, external drives, CD-ROMs, and more). Apart from that, especially in case of
interpreters providing services on location, it may be required to know how to handle
various types of computers and keyboards (e.g. PC vs. Apple), how to establish wired or
wireless connection to the Internet at client’s location, and so on. A translator has to be
versatile and always prepared for the unexpected, especially when an interpreting job
turns out to be an interpreting-translation-and-consultancy job. The way translator
handles him/herself on such occasions is a testimony of his experience, and contributes
to his/her professional image.
Hardware: Advanced computer handling
Interpreting jobs require translators to use external equipment that allows for the process
of translation to take place. It does not really matter whether it is consecutive or
simultaneous interpreting, the interpreter involved has to be aware of what equipment is
necessary, how to operate it, and what optimum operating conditions are. In theory, events
requiring the assistance of a simultaneous interpreter are always covered by a specialised
company providing both equipment and support. In practice, at least in Poland, it is not
always the case and the interpreter has to use his/her own judgement and experience to
provide the service in the first place.
Software: Word-processors and importance of advanced formatting techniques
Text editing, or more particularly text formatting, denotes a set of skills that are primarily
concerned with visual aspect of translation. In the 21st century all of translation is done
using computer software, the most basic and widespread of which are word-processors
(MS Word, LibreOffice’ Write or Apple Pages, to name a few). They are the most basic
tools needed to perform “written” translation. Whereas some, especially older generation,
of translators are nostalgic about using a classic typewriter for everyday work, the
demands of the market (short deadlines, the universal requirement to deliver translation
almost instantaneously after it is completed and the fact that the target translation should
be editable [for review and proofreading purposes]) make it virtually impossible to use
anything else other than a computer for the job.
Chapter 3. Translation competence
43
In fact, a translator is no longer a person who only translates a given text. The
definition of the profession shifted from a pure linguist to someone deeply involved in
cultural and technical communities. Pym (2003: 491) suggests that today the “languageservice provider is currently the most apt” name for the profession. What does it mean?
Well, in the case of text formatting and using word-processors, it means that translators
are frequently required, depending on their assignment, not only to translate but also to
edit and improve on the original visual aspect of the text so that it looks neat and tidy
(professional) and can be used by its owner (the client) with no further work involved.
Clients often expect translators to improve on the visual aspect of the document. As a
result, language professionals frequently have to deal with sloppily written and formatted
documents which require not only visual but also content improvement. It could be argued
that it is not the role of a translator to alter the document in order to improve it contentor visual-wise. Yet, we should not forget the above-mentioned claims. Today the
profession requires more flexibility, especially that clients tend to appreciate all-in-one
package services (translation, proofreading, editing), most commonly provided by
translation agencies (see Sikes, 2011). Such translation can then be used “as is”, i.e. as an
end product. What is more, it seems only natural to say that if a translated document is
formatted in the same way as the source text, the client is more likely to ask for another
translation in the future.
Therefore, despite the fact that computer-assisted translation tools and machine
translation solutions do receive more and more recognition, word-processors still
constitute an unshakable foundation of a translator’s workshop. Moreover, if we take into
account the fact that a significant number of projects involve translation from noneditable content (JPGs, PDFs, CDRs, and other graphical files), the knowledge how to
process the source in order to produce a successful end product becomes vital.
Software: Utilising optical character recognition software and graphical solutions for
regular and computer-aided translation
Another set of computer skills that can be categorised as part of TCC are skills related to
the processing of non-editable data, like images, for example. Although regular written
translation concerns primarily editable content, a significant portion of the total volume
of translation involves files that have to be either translated “on-the-fly” (i.e. source text
is used for reference only and is not involved directly in the process of translation) or the
translator may decide to pre-process them in order to use them in specialised software,
like computer-assisted translation (CAT) tools. The latter approach has many advantages
since it allows to benefit from the effort in the future. If a non-editable document is
translated “on-the-fly”, all gained experience, terminology and knowledge related to a
given topic will be stored only in the translator’s memory. When converted to the editable
format and translated with the application of specialised software, like CAT tools, all that
knowledge is safely stored on a computer and is ready for future use.
A. Optical character recognition (OCR). OCR systems are frequently saviours in
the case of large documents (manuals, contracts, etc.) which are delivered for
translation in the form of a scan (e.g. PDF, JPEG or TIFF image). Basically, OCR
software converts scanned image containing textual data into text and saves it as
an editable file (Bowker, 2002). For someone who works with CAT tools,
converting such a text to an editable format may be of crucial importance as it
may turn out that they have translated similar document before and can use
previous translations and solutions stored in a translation memory (TM).
Competences involved in this case are twofold. First, the translator needs to have
linguistic competence in order to analyse the document and decide whether they
will benefit from its pre-processing on account of owned translation memory.
Second, they need to know how to handle the conversion process itself. The
conversion process will not be analysed here, though. Suffice to say that it may
include cleaning, straightening, brightening, graphical editing of the original,
followed by proper application of the OCR Software (Abbyy Finereader,
OmniPage, Readiris, and more) and final post-processing (addressing conversion
issues, further cleaning, spellcheck, fixing layout, and more). Knowing which
document is suitable for conversion, what problems the converter may find and
how to address them (before, during and after the process) are very handy assets
indeed.
B. Graphical solutions in translation. Another skill to be mentioned in this section
is related to the ability to edit graphical content in a translation package.
Following the idea that a translator is a language service provider, we may
ask what happens when one is to translate a project which contains elements which
cannot be edited in a conventional way (i.e. in a word-processor). Examples here
might include posters, flyers, brochures, static Internet content and so on. Both
Chapter 3. Translation competence
45
personal experience and feedback from project managers working in translation
agencies show that consciousness of the translation workflow among clients is
low and translators often have to deal with PDF, JPG, PNG, TIFF formats (the
four given here being just a few examples), the content of which is fixed. It is rare
for clients to send already extracted textual content for translation (instead of plain
images). As was mentioned above, clients tend to expect full service (Pym, 2003;
Risku, Pein-Weber and Milošević, 2016). Therefore, in such a case translators
face basically two options. They can either translate the content “on-the-fly” and
outsource the graphic editing part of the job or they can edit the source file using
specialised software (Adobe Photoshop, CorelDRAW, Inkscape, Adobe
Illustrator, and so on) in order to provide their client with the ready product. Both
options have their advantages and disadvantages.
Table 5. Advantages and disadvantages of outsourcing graphic editing
•
•
•
•
•
•
TRANSLATORS OUTSOURCE
TRANSLATORS EDIT GRAPHIC
GRAPHIC EDITING
SOURCE THEMSELVES
Advantages
No need for sophisticated skills
No need for specialised software
Time saved on editing can be used for
other pure translation projects
Disadvantages
Risk of errors in the final document as
most
DTP
(desktop
publishing)
specialists are not language professionals
and may not be proficient in source/target
language
Part of the total remuneration is spent on
DTP services
Translators have to depend on time and
skills of third parties, which may have an
influence on the final deadline.
•
•
•
•
•
Advantages
Full control over content modification
The translator keeps entire remuneration
Full control on overall progress of the
project development and deadlines
Disadvantages
Need to acquire specialised skills and
software, some of which can be very
expensive
DTP is time-consuming
The distinction presented above clearly shows that graphic (or DTP) computer
skills are entirely optional for a successful translator. True, they give more control
over the process, but time and financial requirements to employ those may turn
out to be too high to justify their worth. Nonetheless, other than that it is worth to
consider them as they may turn out to be extremely beneficial in the long run.
Software: Computer-assisted translation (CAT) systems
In the previous section, it has been established that some skills (like DTP, for instance)
are entirely optional for a translator. Such statement is true, but only if we put it in a right
context – the market demand. In this particular case, it may turn out that the number of
jobs involving DTP is relatively low (many clients being DTP agencies who need
translation and not the DTP service) and not worth investment. On the other side of the
spectrum are tools which not only facilitate the process of translation but, frequently are
mandatory to be given a translation assignment (e.g. due to client requirements)
(Christensen and Schjoldager, 2016). The latter group includes CAT tools (memoQ,
Memsource, Trados, Déjà Vu, Wordfast, to name a few). The idea of CAT tools is that
they are computer programs that allow to create databases (translation memories) which
store all our translations (aligned into segments on sentence-level) and allow to speed up
translation process and increase translation quality through translation memory-based
propagation of existing translation memory entries. CAT tools will be discussed in detail
later in the course of this book.
Software: Importance of high-quality source documents for translation
This section combines skills mentioned in the previous three sections, but it is
nevertheless important to touch upon the issue of the quality of the source documents,
especially in the context of utilising CAT tools.
It was mentioned before that the source documents submitted for translation are
frequently far from ideal. They can be non-editable, partially non-editable or their
formatting may be messy. All of these issues are irrelevant when translation is done “onthe-fly,” i.e. without the use of specialised software. Yet, in the case of a CAT tool the
form of the source document has tremendous importance. As was mentioned before, CAT
tools rely on translation memories, which in turn rely on proper segmentation. If a
translator processes a source which cannot be cleanly divided into segments, then the
translator risks being not able to use the TM (or use it in a limited scope) for future
translations due to mismatches in the database, e.g.
Chapter 3. Translation competence
47
Figure 5. Example of bad segmentation in memoQ
It the example above we can see that the sentence “Adjust the pressure regulating valve
through the full range up to 7.5 tonnes and reduce the load to zero again before
toolsetting” was divided into two segments. In this case, the source – a scanned manual –
was converted to editable form using Abbyy Finereader, an OCR tool. Due to poor visual
quality of the original, some errors occurred and “artefacts” were introduced in the
editable copy. When imported to memoQ, a CAT tool, the sentence was divided into two
segments between ”-“ and ”^” characters. Such division forces the translator to translate
two separate segments instead of one. This may result in the programme not finding a
correct match in the database when translating a similar document in the future, thus
making the entire process pointless. Of course, it is possible to “fix” segmentation from
within the CAT tool, but it is a complicated and time-consuming process (when the entire
document is taken into consideration) which should be limited to a minimum, if possible.
Therefore, a successful translator should know relations between word-processors, OCR
systems and CAT tools so that they know how to work with the source document at each
stage of the process of translation. Successful pre-editing of a document may significantly
contribute to the overall speed of the actual translation, at the same time assuring that
segmentation will be very precise. CAT tools became standard elements of translator’s
workshop. Therefore, they should be utilised to their fullest extent and it is up to the
translator to expand their computer competence to such a degree that they can utilise all
features of the CAT tools, producing reliable resources (TMs) for the future use.
Software: Software localisation
The process called software localisation denotes the adaptation of a software product or
a webpage to the linguistic, cultural and technical requirements of the market the software
is addressed to (e.g. Esselink, 2000). The process includes a broad range of activities
involving a significant number of both translation and computer skills. Language
Scientific, a leading US-based technical translation company, lists the following major
steps in the software localisation process:
•
Preparation of Files prior to Localization
•
User Interface Localisation: Translation and Adaptation of User Interface
•
User Interface Localisation: Testing and Delivery
•
Localising Online Help and Translating Documentation
•
Client Review and Implementation of review changes
•
Update Translation Memory, prepare for the next software release
(Language Scientific, 2016)
Each of these steps requires a translator to employ a range of skills, the majority of which
are not related to the linguistic process of translation. These involve handling the
localisation software, project management, software engineering, testing and desktop
publishing, to name a few.
Due to its complexity and amount of labour required, it is rarely the case that a
translator works alone on a project, unless it is a relatively easy task of translating a small
website. Such an “easy” task can be extremely demanding for an individual translator
who is required to do everything himself, be it editing of XHTML code or extracting
content strings from a computer database, for example. In the case of larger projects, the
process is always run by a development team consisting of a wide range of specialists.
Nevertheless, in order to produce accurate translations that can be used within the
project (a number of restrictions apply which have to be taken into consideration), a
translator has to be aware of other processes running in parallel to the translation itself.
Some issues cannot be resolved by other specialists, who are not language professionals
and do not have language expertise. Therefore, it is up to the translator to employ their
computer skills in order to achieve the satisfying effect. It is especially so due to the fact,
that unlike regular translation, which is performed on the basis of a finalised source
document, software localisation often concerns the actual development of the given
product (so as to enable simultaneous release in a number of languages).
It is important to note that software localisation is a process which aims to produce
a product that has the feel of one originally written and designed for a given market. SDL
Translation Zone (2016) lists
a number of points that have to be considered, as well as the language, in order to effectively
localize a software product or website: measuring units, number formats, address formats,
time and date formats (long and short), paper sizes, fonts, default font selection, case
Chapter 3. Translation competence
49
differences, character sets, sorting, word separation and hyphenation, local regulations,
copyright issues, data protection, payment methods, currency conversion, taxes.
All of these issues have to be taken into account and then used within a highly structured
localisation project, involving the cooperation of many specialists and employing a
number of computer skills that are required for the process of localisation to be a success.
Software: Online resources
Widespread globalisation and easy access to information virtually from anywhere force
modern translators to rely heavily on the Internet, which is filled with information and
tools that language service providers require every day. To use the Internet or not is no
longer an option, it has become a necessity – starting with exchange of communication
between translators and outsourcers, through mining for terminology in online
dictionaries and professional forums, to working on online translation projects (CAT
tools), the translator has no other choice but to accept the fact that we are surrounded by
information technology and have to use it. What is more, as of 2015, 1st year BA students
coming to study translation were born between 1995 and 1996 and, as a result, they grew
up in the era of the Internet. For these people, who plan their professional career as
translators, the Internet and its vast knowledge repositories are quite obvious resources.
In many cases, they do not have to acquire new skills in this respect but rather to refine
those they already have.
Paradowska (2015) agrees with Gambier (2009) and Gouadec (2007) that one of
the most important resources for translation is the Web itself and, as a result, web
searching skills are of utmost importance to any professional translator. It fits perfectly
into the multicomponent PACTE and EMT models of translation competence. Web
searching, or web behaviour, can be defined as “performing goal-driven actions aimed at
meeting information needs for translation problem solving” (Enríquez Raido, 2014: 62,
cited in Paradowska, 2015: 227). Translators search for terminology, user manuals,
working descriptions, etc. using information retrieval systems, like widely accessible
search engines. It is a repetitive process as the same query may be placed in different
contexts. Yet, “most search engines present similar interfaces allowing people to submit
a query, receive a set of results, follow a link, explore the information, and [if not satisfied
– MK] modify the query.” (Choi, 2010: 1). It is worth noting, however, that although user
interfaces of modern search engines are similar and the same queries may be entered in
most of them, the results they provide may be different due to different systems and
algorithms working in the background. Knowing how to ask questions, and refine our
queries, is a skill few professional translators would neglect.
Software: Utilising machine translation in regular (non-CAT) translation
Machine translation, which has been discussed in detail in chapter 2 of this book, is the
kind of asset everyone has heard about, but few people have used it. While MT plugins
provide MT results in real time in CAT tools, regular application of MT in the process of
translation is limited to one of the tools available one Internet (Google Translate, Yahoo
Babel Fish, SDL FreeTranslation or WorldLingo). While these tools were designed for
people with no working knowledge of a given language in order to allow them to
comprehend online resources (e.g. translate foreign web pages), they are used also by
students of translation (Paradowska, 2015) and, as a result, by graduate translators.
Disregarding the application of such MT in professional translation, the knowledge how
to use those tools may prove to be useful in comparative research and data mining from
languages other than the language pair of the translation project.
Service provision: Quoting
One of the critical computer-related skills any translator should have is the ability to
evaluate and quote a document. Depending on the tools a translator uses, the quote may
be character-, line-, (word processor) or word-based (CAT tools). The price depends on
many factors, all of which have to be considered before a quote is sent. Such issues as
“urgency, emergency, language, the directionality of translation, complexity, special
requirements in terms of hardware, skills, software,” (Gouadec, 2007: 214) or additional
services may require final price alteration.
What is more, the quoting itself may require extra payment. There are cases when
a translator receives a JPG/PNG/TIFF/etc. file which cannot be quoted automatically (due
to the fact that none of these formats may contain editable text), and the translation order
is price-dependent. Another example is a situation where a translator receives a contract
with MS Word-added comments, and only the comments are meant for translation14. Such
a situation does require a translator to know what software to use, and in what fashion, in
order to determine the final price of translation.
14
A problem to be solved with the use of appropriate filters on import to CAT tools environment.
Chapter 3. Translation competence
51
Service provision: e-mail based communication
Although composing e-mails may, and well should, be considered a skill of
communicative writing, it deserves to be mentioned aside other TCC skills. It is the final
step before actual translation can commence. Therefore, it is important to list it as one of
the computer-based service provision skills. It is how most translators contact their
clients, a sort of a name card that a client associates with the translator. Hence, the import
of e-mail communication, especially considering the fact that it may be also used to
process complaints, if any.
Service provision: Company management
The last item to be discussed concerns a set of skills in computer handling, which covers
the management of one’s own business, provided a translator has established his/her own
economic activity. Pros and cons of such solutions are a topic for another discussion.
Suffice to say that those who do run their own company have to handle invoicing
software, project management software, and probably the basics of HTML or CMS15 in
order to market their own activity. It is yet another area of computer expertise that adds
to the notion of translator as language-service provider.
Considering the above-listed examples of computer sub-competences, one may wonder
what the future will bring. Mackenzie (2004: 32) reminds that the skills we, translators,
consider standard, or natural, to the process of translation, were not “imaginable in the
1970’s or early 80’s.” She also mentions marketing as one of the key competencies of a
successful translator, as there would be no professional cooperation between the translator
and a client, had it not been for marketing of his/her services. Therefore, it seems prudent
to list these skills as crucial, if not critical, to a successful professional career, and use it
either as a reference for self-study or as part of the curriculum in translation training
programmes. This way teachers can “prepare students for working life by giving them at
least an introduction to these realities” (Mackenzie, 2004: 33). The issue will be discussed
further in the next chapter which focuses on the development of translation competence,
translator training, and translation strategies.
15
Content Management System, e.g. Wordpress, Drupal, Joomla.
Michał Kornacki
52
Chapter 4. Translator training
The previous chapter discussed the notion of translation competence and how it shapes
the translator. The discussion would not be complete, however, without an overview of
how translation competence can be developed, how translators are trained, and what
translation strategies they apply, both in training and in the professional career.
4.1. Developing translation competence
The development of translation competence is a process that raises many doubts among
experts in Translation Studies. In the following section, an attempt will be made to outline
how translation competence can be acquired and developed.
4.1.1. Translation competence acquisition model
The PACTE (2000: 104) model presented on the next page is based on the PACTE
group’s revised model of translation competence. It assumes that everyone has innate pretranslational competence that serves as a starting point for future development. Next,
through experience and learning strategies, sub-competencies are developed, integrated,
and prioritized depending on a given specialization or other conditions, specific for
individual translator (specific client needs, etc.). Once the sub-competencies are
developed, a trainee translator enters novice stage. The PACTE Group comments:
Thus, the novice stage in the development of translation competence could be defined as the
stage when the sub-competencies have been acquired, at least partially, but they do not
interact with each other. Therefore, the development from novice to expert is not only a
question of acquiring the missing subcompetencies, but also of re-structuring the existing
sub-competencies to put them at the service of the transfer [competence].
(PACTE, 2000)
The development of sub-competencies results in accumulation of declarative knowledge.
However, what is more important, it opens the way to the restructuring of existing
knowledge and mastering it (PACTE, 2000).
Chapter 4. Translator training
53
This model of acquisition of translation competence is in accordance with the
Schneckenberg & Johannes (2006) Steps to the Acquisition of Competence. It starts with
information. The information, through networking, turns into knowledge. If the
knowledge can be applied in live context, then it is an ability. An ability mixed with
attitude (motivation and right values) leads to action. If an action yields a desirable result,
it is competence. Competence combined with responsibility leads to professionalisation
(Schneckenberg & Johannes 2006: 30).
Figure 6. The PACTE group’s translation competence acquisition model
Although this model was proposed as an outline to research on eCompetence in higher
education, it does seem valid in the context of translation competence, since it confirms
the PACTE model of translation competence acquisition. By developing subsequent subcompetencies and practising translation, a trainee expands his/her knowledge, adding new
skills and adjusting old ones. Once and the sub-skills are in place, the final refinement
may commence.
Michał Kornacki
54
Figure 7. Steps to the Acquisition of Competence (Schneckenberg & Johannes 2006: 31)
4.1.2. Pedagogical perspective
According to Marczak (2016), translation competence models can be categorised into
pedagogical models (see Schäffner & Adab, 2000), revolving around the problems found
in the translator education process, and empirical models (see PACTE, 2003), which have
been verified empirically. Since the latter group has been discussed before, let us take a
closer look at the pedagogical approach and see how translation competence develops.
As was mentioned before, translation competence is acquired through a
combination of learning strategies used to develop and refine sub-competencies (PACTE,
2003). The process should be goal-oriented in the sense that the goal of each translation
course is “to train students to reach a level which will allow them to join the professional
market as novices and once there, increase their specialisation depending on the field in
which they find employment” (Way, 2000: 132). Although some students of translation
may prefer to translate lyrics or comics in the classroom, such fields of expertise will not
yield acceptable financial return when translation becomes the main occupation of a given
individual. Therefore, any translator training process should be focused on real-life
translation assignments that constitute main bulk of translation jobs offered on the market.
Kiraly (2000: 122) remarks that authenticity in the translation classroom is critical to any
result-oriented translation programme. Only by working on authentic assignments do
students develop proper habits in translation. They know what to do and how to deal with
any potential problems. What is more, they know what problems to expect. Thus, they
develop their translation sub-competencies, and are able to refine them through the
Chapter 4. Translator training
55
exercise with texts that they may be asked to translate in the future. Such exercises may
include a broad range of documents:
•
official texts, like arrest warrants or financial statements, for example. Even
though these documents are frequently bound to be translated by sworn
translators in order to be officially recognised, it is a widespread practice to
subcontract such assignments to regular translators and authenticate them after
translation16;
•
business texts, e.g. leaflets, reports, statements, official correspondence, and
more;
•
technical texts, e.g. manuals, product description and specifications, and more;
•
specialised texts, e.g. medical papers, reports, summaries; art, and more;
•
regular documents, e.g. correspondence, marketing, and more;
•
other texts, depending on trainer’s experience;
A sample real-life scenario may require students to respond to a client requiring short
translation in an express mode (order and delivery on the same day, frequently under two
hours), and fulfil the order. Such projects are often initiated by phone, which can be
conducted in a group. The trainer becomes the client and raises the issue. Students are
required to come up with all possible questions that may help them with processing the
order, providing the client with all vital information, e.g. price, or special conditions, at
the same time. Once the trainer is satisfied that students covered all areas, students may
commence with the translation. When finished (texts are selected based on their difficulty
and length so that students can translate them under one hour), translation is delivered to
the client by e-mail for checkup. A short discussion follows. The entire scenario should
fit within one class session. Depending on the course, it is welcome to introduce followup exercises, e.g. processing client’s complaint to the provided service (translation
quality, communication errors – no translation received, etc.) in order to allow students
to deal with the problem individually or in groups. Unfortunately, the scope of authentic
scenarios to be conducted in class is limited by time constraints, and many of them have
to be home based, with no immediate feedback from the trainer.
Schäffner (2000) argues, that although post-graduate translation courses are quite
common, they are usually time-limited (majority are one-year courses) and focus on
specialised translation skills, taking language competence for granted. Thus, not all sub16
Based on own experience as a translator and project manager in a translation agency
Michał Kornacki
56
competencies have sufficient time to develop, and some skills are rather hinted at than
used in practice. Schäffner (2000: 146) reminds that
(…) translation competence is a complex notion which involves an awareness of, and
conscious reflection on all the relevant factors for the production of a target text that
appropriately fulfils its specified function for its target addressees. Such competence requires
more than a sound knowledge of the linguistic systems of L1 and L2. In addition, it involves,
at the very least, knowledge of communicative and text-typological convention in the source
and target language cultures, subject and culture-specific knowledge, and (re)search skills.”
Thus, it seems valid to claim that the translator training process should start at an
undergraduate level. It does not mean that a person cannot reach expert level in translation
if the training is taken after graduation (or any point of life for that matter). The author
believes that sufficient determination allows any individual to reach expertise in any field.
However, when student training and market conditions are considered it seems prudent
to say that the earlier the start, the better the effects are.
In accord to what was said previously, in order to achieve full translation
competence, a number of specific sub-competencies must be developed. Below the author
presents a number of competencies following Schäffner’s (2000) division. As will be
seen, they can be found in translation competence models proposed later by PACTE
(2003), EMT Expert Group (2009) and TransComp (2009), although not always as
separate sets of sub-competencies.
Development of linguistic competence in the languages concerned develops
mainly through language skills modules and focusing on linguistic structures, as well as
communicative use of language. Schäffner (2000) reminds that translators should be
equally competent in L1 and L2. Therefore, stress needs to be put on basic linguistic
concepts (morpheme, word, syntax, etc.) and semiotics. Attention to mother tongue is
critical to the development of the linguistic competence since the form of translation
courses, which are foreign language-oriented by definition, poses a risk of neglecting
one’s own native language, thus developing gaps in knowledge and lowering its quality.
Fenyő (2005: 71) defines translators as “bilingual mediators who mediate not only
between languages but also between cultures.” Therefore, cultural competence, or
general knowledge about historical, political, economic, and cultural notions concerning
a given country, can be regarded as an essential part of the translator’s competence since
Chapter 4. Translator training
57
it allows for the cross-cultural transfer to take place. The cultural competence develops
through course modules devoted solely to this type of knowledge.
The knowledge about text types, genres and general conventions and regularities
governing written text in a given language, can be referred to as textual competence.
This type of competence develops twofold. On the one hand, students learn how to
structure text depending on its type and target audience (Schäffner, 2000). The
communicative purpose of the textual message is paramount at this stage of development.
On the other hand, it is vital to practice written translation on the basis of authentic texts
(Kiraly, 2000; Way, 2000) which allow students to learn types, forms, and formats of
documents that require professional translation. Although global market demands are
paramount to develop skills required to translate most common types of texts on the
market, Way (2000) suggests that each translation course should include a variety of other
texts in order to provide students with a general understanding of the market.
The development of domain/subject specific competence, or knowledge of the
relevant subject, or area of expertise, is time-constricted. It is developed through exposure
to various types of specialised texts which require translation. Since most translation
programmes offer 3-year BA or 2-year MA studies, the amount of time that can be
devoted to specialised texts is limited. Therefore, students primarily learn how to manage
terminology (extraction, storage and retrieval, etc.) and use it in specific contexts. The
development of this competence takes place mainly outside a classroom, and through
personal experience. Neubert (2000: 9) argues that a successful translator does not have
to be an expert in a given specialised field. It is important, however, to know where to
look for information, how to find it and “succeed in the process of information transfer
and communication” (Martínez and Fabler, 2009: 92).
(Re)search competence covers the problem-solving capability of a translator. It
allows to predict, evaluate and solve translation problems using previously developed
strategies. It can be developed by exposure to translation assignments which require
logical and creative thinking. Such scenarios force students to search for, or come up
with, strategies that will allow them to deal with the problem. Again, this competence
develops mostly outside the course, during real translation projects. Classroom can give
students some basic guidelines, which have to be refined later, when applied to real
commercial assignments.
Similarly to the 2003 PACTE model, the list provided by Schäffner (2000) and
presented above includes transfer competence as critical to the translation process. It is
Michał Kornacki
58
unique in that it alone is “specific to translation and integrates all other sub-competencies.
It refers to the ability to produce target texts that satisfy the demands of the translation
task” (Schäffner, 2000: 148). Combined with the (re)search competence, the two
contribute to the overall translation competence the most since they cover procedural
knowledge. While this competence can develop on its own through translating, it is worth
noting that course-delivered guidelines may speed up the process and provide a novice
translators with insights for the future. Such approach may ultimately turn out more valid
than learning from one’s mistakes.
The outline presented above attempts to identify lead competences and hint at how
they can be developed. Academic scholars raise questions regarding the purpose of
division of translation competence into several sub-competencies. Malmkjær (2009: 122)
asks whether it is “useful to include so many features as part of translation competence.”
and why “so many scholars of translation competence proceed in this fashion.” Pym
(2003) suggests that either it is related to the need “to establish the discipline as something
separate from linguistics and language training” while Malmkjær (2009: 125) theorises
“that people include so many features in the definition because the task of a translator is
indeed very complex.”
The main problem with translation competence is that it is not something constant.
It changes, evolves, constantly adapts in order to accommodate market demands and new
resources available to translators. What is more, the existing models of translation
competence (PACTE, 2003; EMT, 2009) are inadequate in terms of translation teaching
(Piotrowska, 2015). Piotrowska argues that translation pedagogy “remains peripheral and
enjoys less of a prestige in comparison to more theoretically-oriented sub-branches of the
discipline” (idem: 20). She argues that most translation educators are not trained to teach
translation as a profession. While the theoretical approach has always been regarded with
great esteem by language scholars and course designers, it is the practical approach that
attracts students to translation courses17.
Both Kiraly (2014) and Piotrowska (2015) agree that static model of a translation
course, with teacher as an instructor and student a passive learner, has become obsolete.
The pace at which local and global translation markets change requires ingenuity,
creativity and adaptation from a translator. A translation course that utilises static, or
linear, progress in developing translation competence may be informative, but, ultimately,
17
Based on interviews with the 2nd year BA students of English at the University of Łódź, at the inception
of their translation programme.
Chapter 4. Translator training
59
it is too slow for students to develop traits that will allow them to survive professional
competition after graduation. A model where the teacher is a supporter, nudging students
to learn how to discover solutions to problems, and refine them after successful
application, is needed. Kiraly and Piotrowska (2014) call for such innovation in translator
education, especially considering the fact that with no established teaching programs for
educators, translation teachers have to rely on their own experience and common sense.
While such situation has its advantages (for students learn actual real-life scenarios) it
relies too heavily on the skills of an individual educator. Lack of standardisation may
result in lower quality of certain courses. Especially, if we follow Kiraly and Piotrowska’s
(2014) “emergent curriculum” idea, which revolves around purposeful, dynamic, and
ever-changing nature of the learning process. Such approach has the potential to activate
students to deal with unpredictable learning
An emergentist view not only allows but requires teacher to climb down from their pedestals
of authority, and it implies an obligatory change in their roles from distillers and transmitters
of knowledge to guides and companions on the student’s road to experience. Syllabus design
is no longer a task to be accomplished by a teacher alone prior to the start of a course; it
becomes a tentative plan that emerges with new challenges and unexpected turns, and one
leading to unpredictable goals as a course progresses.
(Kiraly, 2015: online resource)
As can be seen, such problems as translation competence, its acquisition, and
development (both from student’s and educator’s point of view), are difficult to evaluate
due to the constant change of demands placed on translators, as well as point of view. The
emergentist view, proposed by Kiraly (2015), shows that although such translation
competence models as that of PACTE (2003), EMT (2009) or TransComp (2009), can be
regarded as milestones of translation competence discussion and are highly successful, a
new learner-oriented approach is required.
Kiraly and Piotrowska (2014) observe that there are several issues with the
competence-based
approach
to
translation
competence,
especially
different
understanding (or distribution) of sub-competencies.
Rather than seeing competence as discernable products, that is fragmented skills and
knowledge, the underlying conceptualisation is that both learning and competence are
Michał Kornacki
60
holistic
and
autopoietic
(dynamic,
unpredictable,
self-generating
and
self-
maintaining) processes.
(Kiraly and Piotrowska, 2014: online resource)
In order to address the issue, a new model of translation competence has been proposed
at the School of Translation, Linguistics and Cultural Studies (Fachbereich Translations-,
Sprach- und Kulturwissenschaft18 – FTSK), Johannes Gutenberg University Mainz.
Figure 8. A Model of Emergent Translator Competence (Kiraly and Piotrowska, 2014)
18
http://www.fask.uni-mainz.de/
Chapter 4. Translator training
61
According to Kiraly and Piotrowska, the model of emergent translator competence was
partially inspired by the view of competence acquisition proposed by Stuart and Hubert
Dreyfus (1980). But also by complexity theory, which has come to inform a post-modern
understanding of learning as a largely autopoietic (i.e. self-initiating, self-sustaining and
perpetually dynamic) process of becoming – rather than a static set of learnable facts and
piecemeal skills: (Kiraly and Piotrowska, 2014: unpaginated). The model presented in the
Figure 7 is built on the concept of a number of sub-competencies gradually merging over
time into a single super-competence (suggested earlier by Wills, 1982; Pym, 2003). The
approach can be considered evolutionary. It is based on sub-competencies, which can be
developed and refined over time, and takes into consideration existing structures within
institutions. However, it rejects typical cause-and-effect relationship between teaching
and learning, proposing instead that “learning is a holistic, emergent, self-perpetuating
and embodied lifelong process that proceeds both within the individual and within
communities of practice at different levels” (Kiraly and Piotrowska, 2014: unpaginated).
The shift from instruction to discovery in translator’s development should eventually
allow for the development of super-competence (i.e. “a unified capacity or capability for
professional, expert and flexible workplace performance as the learner progresses (…)
towards their more advanced levels of proficiency, expertise, and mastery” (idem).
Of course, the model has yet to be thoroughly tested, but it appears convincing from
the perspective of translator education. The development of translation competence is a
process, and, therefore, has to be treated as such. Nonetheless, earlier reflections on the
issue should not be forgotten. They bring insights from the history of translation discourse
that should be considered even today, when technology opens new ways of development.
That being said, the paramount role of pedagogy is emergence and development of
competences should be recognised. While many skills can flourish without any external
influence, market-oriented tuition is vital to achieving long-term goals of any translation
trainee (qualifications to become a successful professional) (Piotrowska, 2015). Multiple
competence-based approach to translation competence, suggested and discussed by no
small number of academics and researchers so far, has generally been accepted and forms
the basis for future discussions. It is doubtful, considering what has been mentioned
above, that one fixed model of translation competence will be formed due to ever
changing expectations of the market, development of computer technologies (e.g. MT)
and the role of the translator in modern society.
Michał Kornacki
62
4.1.3. Creating technology-oriented translation course
Understanding how translation competence is organised, and how individual subcompetencies interact with each other is paramount in order to plan and run a successful
course. A technology-oriented translation course poses a greater challenge due to a
number of reasons. To start with, basic requirements for such course are quite high.
Students have to be able to use a computer laboratory with enough workstations for each
student to work individually. “The very fact that each student has immediate and direct
access to a complete set of electronic tools obviously promote the integration of extensive
hands-on experience in the classroom” Kiraly, 2000: 126). Secondly, the organisation of
the course is much more difficult. By default, the technology course will require students
to work with data provided by the teacher. While the regular course can rely on printed
handouts, the same cannot be applied to a digital class. Therefore, the teacher has to
incorporate a learning management system (LMS), like Moodle for example, in order to
share class resources amongst students. Other options include sharing via cloud drive, or
running own website with all the required resources stored locally on the website’s server.
Such approach grants the teacher maximum flexibility in creating the layout and
distributing the content of the course. However, it requires considerable IT (website
creation, file uploading via FTP) and content management experience.
Of course, an ideal solution would be to have all the computers in the lab networked,
and open to the teacher’s workstation. However, as experience shows, it cannot always
be achieved due to technical problems, or simply, a local IT policy.
Kiraly (2000: 126) suggests “a relatively inexpensive alternative to the multiworkstation arrangement” in the form of an overhead projector linked to the teacher’s
computer. Such solution allows students to see visual instruction and learn through a
presentation. However, such approach is hardly collaborative and does not allow students
to practice presented skills immediately after they were presented.
While this solution may be appropriate for a regular translation course, a course
devoted to technology in translation will not be successful without proper tools to work
with.
Method of instruction
For a course to be successful, appropriate methods of instruction have to be selected.
Kiraly (2000) chooses four models of interaction with students, proposed earlier by
Chapter 4. Translator training
63
Paulson and Paulson (1992) and Paulsen (1995), that can be used in the computer
classroom:
1.
One-to-many, in which communication takes place between one teacher and a group
of learners (e.g. in online lectures);
2.
Many-to-many, where communication takes place between learners (with or without a
teacher as moderator), in the form of discussion, role-play activities, brainstorming and
project groups.
3.
One-to-one, in which communication takes place between one teacher and one learner,
or between two learners (e.g. online peer tutoring, or e-mail tandems);
4.
One-alone, involving autonomous learning with the help, for example, of online
databases and journals, online applications and software libraries, and news groups.
(Kiraly, 2000: 127)
The order of the models is not accidental, as it reflects stages of competence acquisition
which allows learners to ascend the ladder of competence acquisition in their pursuit of
mastery. The ladder, or “scaffolding,” as Kiraly names it, can be visualised thus:
>>>LEARNING>>>
GRADUATE
Students mastered the content provided in-class and
are proficient enough to identify, find, and process
additional information (e.g. from the Internet) when
necessary
Students are proficient enough to undertake
individual assignments. The teacher provides one-toone assistance if needed.
The theory is put into practice while students work in
groups in order to achieve the desired goal. The
teacher acts as a moderator.
Novice learners are introduced to the concept of
computer-based translation through instructions and
lectures provided by the teacher.
NOVICE
Figure 9. The ladder of competence acquisition
One-to-many method, the first model on the list above, represents purely transmissionist
method of teaching. It includes pre-recorded knowledge (either a lecture or video
presentation, e.g. on YouTube) to be delivered to students, whose role is to listen and
Michał Kornacki
64
process the knowledge presented. During in-class activities, the teacher becomes a
manager, or coordinator, encouraging groups of students, and helping them achieve
desired goals. At the same time, the teacher receives first-hand experience of what
students know and how they react to the previously provided knowledge, and problems
that may arise. In addition, Kiraly (2000: 127) points out that this method may be
especially successful in a diverse group of students with different starting skills, where it
is necessary to “restructure [the course content] for the class as a whole” in order to reach
pre-determined teaching goals.
Due to the fact that the range of knowledge transmitted this way is limited, and new
content can be added only to a small extent (e.g. expanding pre-recorded video tutorial
may require too much time to be worth the effort), this method of instruction should be
used cautiously (e.g. to fill in the gaps in students’ knowledge, and between students),
“and followed immediately by, or interspersed with active student involvement” (Kiraly,
2000: 128). Such practice will not only allow students to learn by practice “what the
teacher has just presented, but also go beyond it” (idem.).
Many-to-many model is aimed to allow students to work in groups in order to
develop strategies for dealing with complex, authentic tasks that they may be exposed to
in the future. Students are required to show their individual expertise on group forum so
that their insights may benefit the group as a whole. However, in order to simulate
teamwork over a distance, students are required to communicate and collaborate online,
using e-mail or communicators like Skype or WhatsApp. Currently, establishing a group
chat poses no challenge at all and such class setting allows to practice translator-translator
“communicative interaction at a distance, in the absence of many of the normal discourse
and information-carrying cues like gestures and body language” (idem.).
One-to-one model of interaction includes teacher-student and student-student
exchange. In the first scenario, the teacher acts as a mentor supporting a student in
tackling specific tasks, or providing individual feedback when necessary. This relation
bears the risk of the teacher becoming an instructor which, at this step of the ladder, may
be unproductive. At this stage, the role of the teacher is to support the student and suggest
ways to overcome a problem, rather than providing the solution.
The second scenario involves two students working on the same project, but
pursuing goals of their own. Students may support themselves in order to achieve final
success. Again, the model fails if one of the students becomes the instructor. Therefore,
it would seem prudent to make an effort to pair students of equal, or at least similar,
Chapter 4. Translator training
65
proficiency so that neither has visible advantage over the other. Unfortunately, it may be
extremely difficult (and frequently impossible) to achieve due to physical constraints of
the classroom and capabilities of students themselves. The teacher has to take into
consideration their language proficiency, class-related skills, and character traits, which
makes the task labourious and not 100% effective at best.
One-alone is the final step in the learning process outlined above. When a student
reaches the level of proficiency that allows him/her to carry out tasks covered in the
classroom before, it is required to provide him/her with authentic tasks that will require
them to use those skills in an entirely new context. It may be also referred to as adaptation.
The profession of translator is exceptionally demanding in terms of ability to find
solutions to new problems and adapt to new conditions. Therefore, it is at this stage that
students learn how to work alone, without the support of the teacher and the group, in
order to achieve acceptable results (produce an acceptable translation).
Kiraly (2000) points out that the sequence of instruction methods is not, and cannot
be, fixed. It will change depending on the course content, the initial proficiency of a given
group of students, and the overall progress of the course (some stages may not be present
at all).
Course content – static or dynamic?
It might seem that most of the students coming to the university are computer-proficient.
Real-life experience shows, however, that this is not entirely true. Students do have some
working knowledge of how to handle a computer and an office suite. However, they use
computers, and mobile devices for that matter, mostly for communicative purposes. As a
result, using them for translation purposes may not be as obvious, as they think.
In the previous sections, I proposed and described computer-based resources that
could be included in translation curriculum. While some researchers believe that “basic”
(e.g. document formatting) aspects of computer use may be discovered by students on
their own (Bowker, 2002), the author’s experience as project manager at translation
agency in Łódź, Poland, shows that the skills in question remain largely undiscovered
even in freelance translator, not to mention students of translation, if not covered at some
point of their education. Moreover, the fact is that it is the document formatting-related
Michał Kornacki
66
set of skills that enables a translator to deliver successful translations on micro-level19.
The importance of computer-based skills, even the most basic ones, has been outlined by
Kiraly (2000: 130):
Customers of translation services today can and often do expect to receive a professional
quality text formatted according to their specifications. In many cases it has become too
expensive to pass on a poorly designed and formatted text to a copy editor or secretary for
revision – text layout is often part of the translator’s work. The translator-in-training must
acquire a sophisticated set of computer-based skills in order to meet the demands of his or
her future employers and clients.
Therefore, it is critical to initiate students into the computer-based classroom and allow
them to develop all the hardware and software skills that will enable them to enter the
market after graduating successfully. As with any other course, it is important not to
overload them with complex content from the beginning and structure the course content
accordingly to their needs. It may be prudent to analyse initial skills of students in case
the content of the course has to be refined (either to exclude some skills or to include new
ones). The figure below presents a sample analysis of proficiency in text formatting that
I use with 2nd-year BA students, attending Computer Application in Translation at the
Institute of English Studies, University of Łódź, Poland.
Test your knowledge of the most basic editing features of word processors
This exercise is meant to test your initial skills in the use of a word processor (MS Word or
LibreOffice Writer)
Please follow the instructions below and carry them out. Once you are done, save the
document and send it over to michal.kornacki@uni.lodz.pl
Instructions:
Change the margins of this page to 1,7 centimetres, add 1 centimetre as the gutter margin.
Enter the month you were born here:
*Right* justify this line. Use the align text right command, not the space bar.
Center this line. Use the center command, not the space bar.
Copy this line and paste it at the very bottom of your document (use keyboard shortcuts
Ctrl+C Ctrl+V).
Convert this line to bold face
Convert this line to italics.
Underline the word appears where it appears in this sentence.
19
Micro-level should be understood as delivering high quality translations, formatted according to
customer’s wishes, in small numbers (e.g., by part-time translators). Macro-level would include bulk
translations using CAT tools, done by professional full-time translators. In this sense, macro-level
translation means staying competitive on the translation market.
Chapter 4. Translator training
67
Underline this entire sentence.
Use the numbering command to list 4 foods you like. Place the list under this line:
On the numbered list you just created, change the numbers to lower case Roman Numerals.
Use the bullets command to list 3 places you like. Place the list under this line:
On the bullet list you just created, change the shape of the bullet to a black square.
On the bullet list you just created, use the sort button to put the words in alphabetical order.
Use the thesaurus command to change the word hot to a different word.
Use superscript to change 6th, 7th, 8th & 9th so they are correctly formatted.
Change all the text on this line to uppercase.
Change the font face of this line to Cooper Black and change the size to 24 points.
Green highlight the one word in this sentence that starts with the letter g.
Insert an outside border around this sentence.
Insert an outside border around the fourth word in this sentence.
At the end of this line type the total number of words in this document.
At the end of this line type the total number of paragraphs in this document.
Add one level of indent to this sentence.
Change font colour of the word colour to blue.
Convert this line to small caps.
Add a left indent of 5 cm to this sentence.
Change the line spacing between the previous 10 lines to 1.5 line.
Set a 5 cm, dotted tab to this line.
Figure 10. Basic editing skills – evaluation20
Evaluated students are required to carry out the instructions in the document within a
limited timeframe, save it and send it over e-mail to the teacher’s address. Their works
are then compared with a pre-formatted template using “Compare” function in MS Word.
Results show that most students have limited knowledge of word processor features that
would enable them to format a document according to client’s requirements.
This, in turn, leads to a question whether course content should be static or dynamic.
Static content is pre-defined and does not change throughout the course. Everything is
planned in advance and knowledge is passed on to students on the basis of a predetermined schedule. The main disadvantage of this approach is that no extra content is
allowed and the teacher has to follow a narrow timeframe in order to discuss all the
content.
A dynamic course allows for greater freedom of the curriculum. While there is a
list of goals to be achieved, the teacher can add extra content depending on group’s
interests and proficiency. There is no fixed plan, just general outline of skills to be
20
Based on Formatting Frenzy exercise, published by Bristol Public Schools at:
https://www.bristol.k12.ct.us/uploaded/memorial_ms/UA/Computer_Science/documents/Formatting
_Frenzy/grade_8/Formatting_Frenzy_1.docx.
Michał Kornacki
68
developed. The main advantage of this approach is that students may learn much more
from their teacher if they only want to. The main disadvantage is that some goals may not
be reached due to time lost on the extra content. What is more, depending on the
university, in-house regulations may effectively make it impossible to run a dynamic
course. If an academic institution forces a given course content on the teacher, there may
be no room for extra (dynamic) content. However, if there is a chance to develop a course
by oneself, it is worth to consider leaving some room in case there are requests from
students for extra content. Such extra content may involve, for example, discussion on
annotation in sworn translation as a bonus to a class on editing a document in order to
retain the original layout.
It seems that the best option would be to have a detailed static course prepared with
enough free time slots to accommodate any extra wishes from the students. This way the
teacher makes sure that students follow certain requirements and develop skills required
to receive a positive final grade. At the same time, there is room to add extra content
basing on individual needs of the group.
Learning computer skills – examples
Discussion on how to learn particular examples of computer-based applications is a
subject too broad to cover here. Nevertheless, it is necessary to at least outline general
areas of study.
The first such area, according to Kiraly (2000), is research for translation
purposes. The job of the translator is as much about looking for information, as it is about
translating. Kiraly lists such resources as “mono- and bilingual dictionaries, (…)
encyclopaedias, specialised reference works, internal company glossaries, parallel texts,”
(2000: 133) as examples of tools that can be used to help the translator achieve goodquality translation in less time. Students have to know when, how and where to look for
information and find relations as well as dependencies. A good example of research
proficiency can be an authentic situation in which a client submits a hardcopy (or scanned
image) of a technical manual, and wants to have it translated with the use of CAT tools
(for the sake of obtaining a translation memory). For that to happen, the translator has to
convert the file to DOCX file format and fix all possible conversion errors. The process
is long since it requires significant post-processing of the converted document. A possible
way to go around the problem is to look for that particular manual on the Internet. There
is high chance that it is possible to find an editable version of the document online and
Chapter 4. Translator training
69
use it (after content verification) in the translation job. The Internet contains vast
resources and knowing how to identify and find them is an invaluable asset.
Terminology and its management constitute the next important area to discuss.
Saying that a translator has to know terminology may seem cliché unless we put it in a
context. A translator translating for the EU, for example, has to know and use exact
terminology, according to given standards. This stems from the fact that business has
been largely internationalised over the last several decades, and some standards in
information exchange are needed. These concern mostly terminology, the importance of
which can be seen in the fact that European Union created a separate system for
terminology exchange, called Inter-Agency Terminology Exchange (IATE).
IATE is the EU inter-institutional terminology database system. IATE has been used in the
EU institutions and agencies since summer 2004 for the collection, dissemination and shared
management of EU-specific terminology. The project partners are:
•
European Commission
•
Council
•
Parliament
•
Court of Auditors
•
Economic & Social Committee
•
Committee of the Regions
•
Court of Justice
•
Translation Centre for the Bodies of the EU
•
European Investment Bank
•
European Central Bank
(IDABC, 2006)
Its objectives include:
•
IATE aims to provide a web-based infrastructure for all EU terminology resources,
enhancing the availability and standardisation of the information. These resources
include any existing data compiled by the participating organisations and the three
major existing databases (EURODICAUTOM: the European Commission's
multilingual term bank born from two existing lexicographic tools, the Dicautom
and the Euroterm; TIS: Terminological Information System of The General
Secretariat of the Council of the European Union and EUTERPE: Exploitation
Unifiée de la Terminologie au Parlement Européen or 'European Parliament onestop terminology management system').
Michał Kornacki
70
•
IATE was also developed to serve as a vehicle for the application of advanced
language processing technology to multi-lingual terminology management.
•
It aims to support interactivity through the provision of:
➢
The possibility for a user to carry out modifications and add entries directly
to the central database;
➢
In-built validation procedures to ensure quality;
➢
The development of management and reporting tools;
➢
A messaging system to be used as a communication mechanism between the
actors in the terminology workflow.
(IDABC, 2006)
The above shows the importance of terminology management, whether it is a small
corporation or a union of states. A translator can only benefit from using terminology
management software or accessing an external database. Of course, there are many
solutions available on the market, both commercial and open-source. At the University
of Łódź, students practice skills included in this category through Termbases feature of
memoQ CAT tool. Students learn the importance of terminology, how to construct their
own specialised termbases during the process of translation, how to create termbases
depending on the resources provided by a client (e.g. word lists), as well as to look for
terminology on the Internet, download and process it, and add it to the existing termbases.
It is vital, I believe, to teach students how to be resourceful in terms of terminology
management and the importance of contextualised termbase entries, especially in
specialised translation.
One of the must-have skills of any successful translator is editing. When combined
with proofreading, it provides an individual with technical ability to produce translations
according to modern standards regarding document form and format. As was mentioned
before, thorough testing would be welcome before actual course begins in order to avoid
discussing known skills. Alas, if it turns out that students find it challenging to follow
client requirements regarding the layout of a document, or they cannot recreate the layout
of a printed source document, an activity-based instruction is suggested.
It can be expected that students will require feedback if they are to learn from their
mistakes. This feedback can be provided by the teacher or by peers. The latter is especially
effective in practising translation and proofreading skills. Students perform translation
and then swap places in order to proofread their transactions (a visually-proficient way of
doing so is to use Track Changes function in a word processor). On the one hand, they
Chapter 4. Translator training
71
learn how to provide the proofreading service; on the other, the activity shows them errors
visible at trainee level. As a result, the student arrives at a roughly final version (translated
and proofread), which then can be sent over to the teacher for final proofreading.
Another useful exercise, to be completed entirely by the student, involves
proofreading of a text previously altered by the teacher, e.g.
ORIGINAL
ALTERED TEXT
The Ugly Duckling by Hans Christian Andersen
(a fragment)
Once there was a duck who had just hatched a
brood of ducklings; one of them had been longer
coming out of the shell than the others, and when
it came it was very ugly. But its mother did not
love it less on that account; mothers never think
their little ones ugly. It could swim very well, so
she knew it was not a young turkey, as an old duck
had said it might be, and she took it with all the
rest of the brood to the farm-yard to introduce it
into good society. An old turkey, who was very
grand, came up to the duck, and said, “Your
children are all pretty except one. There is one
ugly duckling. I wish you could improve him a
little.” “That is impossible, your grace,” replied
the mother, “he is not pretty; but he has a good
disposition, and swims even better than the
others.” “Well, the other ducklings are graceful
enough,” said the turkey, “pray make yourselves
at home, here.”
But how could the ugly duckling do so? The
whole farm-yard laughed at him. The ducks
pecked him, the fowls beat him, the girl who fed
the poultry drove him away with a stick.
The Ugly Duckling by Hans Hristian Andersen
(the fragment)
Once there was a duck who has just hatched a
pack of ducklings; one of them had been longer
coming out of the shell than the others, and when
it come it was very ugly. But its mother did not
love it lesser on that account; mothers never think
their little ones ugly. It does swim very well, so
she knows it was not a young turkey, as an old
duck had said it might be, and she took it with all
the rest of the brood to the farm-yard to introduce
it into good society. The old turkey, who was very
grand, came up to the duck, and told, “children are
all pretty except one. There is one ugly duckiling.
I wish you could improve him a little.” “That is
impossible, your grace,” replied the mother, “he
is not pretty; but he have a good disposition, and
swims even better than the others.” “Well, the
other ducklings are graceful enough,” the turkey,
“pray make yourselves at home, here.”
But how could the ugly duckiling do so? The
whole farm-yard laughed at him. The ducks
pecked him, the fowls beat him, the girl who fed
the poultry drove him away with a stuck.
Figure 11. Proofreading exercise – flawed source and unaltered original
This exercise requires simple grammatical proofreading. Students are required to read the
source and correct it if necessary. When finished, both documents are “compared” using
the Compare function in MS Word. If all the errors were found and corrected, no changes
are marked. If some errors were skipped, or a new one added, they will be visible in the
Michał Kornacki
72
comparison output. It is not possible to present such output here due to text-formatting
issues. Therefore, another option will be used, that of Diff tools, like text-compare.com:
Figure 12. Compare output (Diff tools – www.text-compare.com)
This allows the student to see where changes are and discover what their nature is. At this
stage, it is quite easy to draw conclusions and learn from clearly marked examples.
The courses on translation technology for the 2nd-year MA students, conducted by
the author, include aspects of website localisation, which require students to learn the
structure of Wordpress CMS-based web pages and to translate the content from the web
page. Wordpress is an example of open-source web software that can be used to create
blogs and web pages. The importance of Wordpress can be seen in the following numbers:
•
over 409 million people view more than 23.0 billion pages21 each month
(Wordpress.com)
•
as of 29/06/2016, almost 26% of all the websites and blogs on the Internet use
Wordpress (Smith, 2016)
The conclusion is that a quarter of all the websites on the Internet use Wordpress. The
introduction of this CMS into the translation classroom not only allows students to get
familiar with one of the most popular website creation suites and to use other systems
(e.g. Joomla, Drupal) due to their similarity and the transfer of skills. What is more, the
software has modal architecture, which comes handy when completing a project for a
client (if the website has not multilingual module added, a conscious translator can add
one and use it to translate the content).
21
Hosted on Wordpress.com. The figure does not take into account self-hosted websites
Chapter 4. Translator training
73
In order to learn how to localise a Wordpress-based website, students are given
access to a website and are asked to produce a web page containing the content of their
own choice. It may be a hobby, politics, history, etc. They are tutored in basic HTML
language if necessary and the use of graphical user interface (GUI) to be used to add the
content. Once done, students are again tutored on the modal architecture of Wordpress
with attention put to localisation plugins (e.g. WPML or Polylang). The next step is the
peer-performed localisation of the new content – students learn in practice how to
translate CMS-based content.
As a bonus to this activity, students learn basic HTML and develop CMS handling
skills, which may allow them to publish their website one day in order to market their
services.
Computer-assisted translation classroom allows the teacher to teach students
more advanced applications of computers in translation, focused on the commercial
aspect of translation. There are some prerequisites to comply with, however. First and
foremost, a student has to have a working knowledge of file format conversion, editing,
and optical character recognition. Due to the fact that CAT tools require editable source
text to work on, it is sometimes required to prepare such a source out of a hardcopy (or
images). The layout of the source document has to be brought back, and the text itself
prepared with CAT tools in mind. Any other knowledge, like HTML or XML languages,
comes as a bonus.
Once the pre-requisites are met, students learn how to use CAT systems to translate.
Initially, they deal with short texts. As their competence progresses, they move on to
longer texts, requiring to build a termbase and to work in groups. They receive
assignments requiring translators to work in a team and share ideas and terminology in
order to achieve a common goal. Once this stage is completed, students learn more
esoteric features of CAT tools, like content filtering, for example22. Finally, students are
asked to carry out individual projects, which require them to provide translation,
translation memory and termbase as separate files, a test of their collective skills and
creative thinking.
The last area of computer-based skills to be mentioned here, in the context of the
learning process, is subtitling. The Institute of English Studies, University of Łódź,
Poland, provides specialised courses on audiovisual translation (AVT). The subtitle22
Content filtering can be used, for example, to import only comments to a text for translation,
disregarding the actual content of the document, if so the customer wishes.
Michał Kornacki
74
related part of the AVT course discusses the theory behind subtitles and requires students
to test this theory in practice. For that, they need to learn all the technical background of
making subtitles using such free tools as Subtitle Workshop or Subtitle Edit. Initially,
students are instructed how to operate the software, and first exercises are based on
synchronisation of subtitles with no timecodes. The next step requires students to translate
regular (time-coded) subtitles from one language to another, which is followed by a live
in-programme translation (i.e. student translates on-the-fly what is spoken in a video,
places the subtitle, and adjusts the timecodes). Video samples are largely obtained from
YouTube or Vimeo, with care taken to use those that are not subtitled (either by human
subtitler or by Google automatic captioning23, based on Google’s speech recognition
system). The final assignment requires a group of student to subtitle a longer piece of
video (about 1 hour). The process requires them to manage the entire project between
themselves, including distributing such roles as project manager, transcriber, translator,
terminology manager, proofreader, and subtitler. The task requires them to combine
previously learned skills in order to achieve a common goal that they are collectively
responsible for.
As can be seen, the content of a technology-based course largely depends on the
professional experience of the teacher who designs activities for students. The tools we
teach, and use, today influence solutions yet to come. Computer-assisted translation has
already revolutionised translation market in terms of quality, output volume and time
spent on translation. Speech recognition systems (like the one Google uses for automatic
captioning of YouTube videos), a novelty some 15 years ago (Kiraly, 2000), is used quite
successfully to subtitle online videos in a number of languages. Today, professional
success depends on many things, and conscious use of computer technology in translation
is one of them. What is more, as Kiraly (2000: 139) observes, “by acquiring a firm fell
for the interplay of the computer-based tools that are used by the community of translators
today, students will be well prepared to continue learning how to use the tools of
tomorrow”. Students need to learn how to use computer tools in order to exist on the
modern market. Additionally, they have to understand how to use them in order to find,
learn and use new tools. The inclusion/incorporation of problem-solving, set in the
context of authentic assignments, will allow students to acquire “the lifelong learning
skills that will ensure their ability to adapt dynamically to the tools of the profession as
23
As of 2016, Google automatic captioning is available in English, Dutch, French, German, Italian,
Japanese, Korean, Portuguese, Russian, and Spanish (Google, 2016).
Chapter 4. Translator training
75
they evolve in the future” (Kiraly, 2000: 139). The following chapter moves on to the
issue of translator training.
4.2. University training programmes
Most of foreign languages departments at universities in Poland and abroad offer
translation courses as part of their degree (BA, MA, or both) programmes. While
translation is often used in practical language classes as a method of checking language
acquisition, “translation tasks have increasingly been seen as training activities in
themselves, imparting skills that are specific to translation as a mode of communication”
(Pym, 2009: online resource). This shift becomes more and more visible and can be seen
in many edited volumes and monographs alike, e.g. Malmkjær (1998), Baer and Koby
(2003), Malmkjær (2004), Tennent (2005), Kearns (2008), House (2016). As a result, and
due to increasing market demand, translation has been on the incline for some time now.
In order to address that demand, universities around the world had to adapt a new
approach to the discipline and provide adequate training programmes. At the moment,
European students have several options in this regard: 3-year BA and (optional) 2-year
MA programmes (most European universities; BA programme in Turkey lasts 4 years24),
or 5-year BA/MA programmes in some cases (e.g. Fine Arts). According to Klimkowska
(2013: 45), in Poland, where the research takes place, we can name “the following forms
of translator education:
•
first-cycle (undergraduate) studies
•
second-cycle (supplementary) studies
•
specialised post-graduate studies
•
specialised courses and trainings25”
As far as the rest of the world is concerned, “the need to adapt to existing local structures,
coupled with required language learning at the university level, has led to a clear
predominance of programmes at Masters level” (Pym, 2009: online resource).
It is worth mentioning that some of the higher education institutions offering MA
courses in translation in Europe are associated in the European Master’s in Translation
(EMT) network, which at the same time is a quality label for MA university programmes
24
25
Education System in Turkey. http://www.studyinturkey.com/content/sub/education_system.aspx
Translation mine, MK.
Michał Kornacki
76
in translation26. The role of the network is to “improve the quality of translator training
in order to enhance the labour market integration of young language professionals”
(European Commission, 2017). The core of the project is the EMT translator competence
profile, drawn up by European experts in translation. The profile lists a critical set of
competences required by translators to achieve success in today’s market. For that reason,
it is used by higher education institutions as a model for their translation programmes. As
a result, the EMT network helps to train highly skilled translators and interpreters,
improve the quality of translation and, as a result, the status of translators in the European
Union.
As of 2017, the EMT lists 2 Polish institutions: Wydziałowe studia magisterskie w
zakresie tłumaczenia specjalistycznego i zawodowego (MA in Translation: PolishEnglish-French-German), Adam Mickiewicz University in Poznań; and Magisterium na
kierunku Lingwistyka stosowana, specjalność nauczycielsko-tłumaczeniowa (MA in
applied linguistics, foreign language teaching and translation), University of Warsaw.
Pym (2009: online resource) reports that “since the 1990s there have been strong
arguments in favour of moving translator training away from general modern-language
programmes, in many cases resulting in independent programmes exclusively for the
training of translators and/or interpreters.” It is clearly visible in the fact that while many
institutions incorporate translation as additional content to their existing programmes,
many others provide separate translation tracks for students (e.g. the Institute of English
Studies at the University of Łódź offers its 2nd-year BA students a choice of vocational
track: teaching or translation). What is more, more and more institutions offer specialised
post-graduate studies in translation (the University of Silesia, the University of Warsaw,
the University of Wrocław, the University of Łódź, Adam Mickiewicz University,
Jagiellonian University in Kraków, Nicolaus Copernicus University, Warsaw School of
Applied Linguistics, Vistula University in Warsaw – as listed by TEPIS [2017]). Usually,
they are 1- or 2-year courses aimed to train professional translators and interpreters.
Depending on the programme, both 2-year MA and 1- or 2-year post-graduate
studies in translation focus on those skills that can be used by translators and interpreters
in the authentic work environment. Therefore, their goal may be to support students to
access specific market niches (e.g. audiovisual translation or software localisation). The
26
European Master’s in Translation (EMT) explained. https://ec.europa.eu/info/education/europeanmasters-translation-emt/european-masters-translation-emt-explained_en
Chapter 4. Translator training
77
attention is paid more to the translation process and research skills than to language
training.
Independent Masters-level programmes, on the other hand, can be more focused on
the skills used by translators and/or interpreters. They might thus be expected to cater to
specific market niches or skill sets such as audiovisual translation, literary translation or
localisation. In practice, however, these programmes still tend to offer general approaches
to translation, albeit without the language training that is offered in the first years of the
full programmes. This ‘general Masters’ approach has been proposed as a model for a
European Masters in Translation. In fact, such approach is behind the EMT model of
translator’s competence. Moreover, while the EMT has been designed to train translators
and facilitate contact with the professional market, the fact remains that the influence of
academic training on the profession is limited. Rather, it is the other way round – the
market shapes the training programmes by its demands.
While the idea of the market affecting the training may sound questionable,
ultimately it may be positive for the students due to the fact that, as mentioned by Bowker
(2004), Chesterman and Wagner (2004), Gouadec (2007) or Pym (2009), the popular
impression is that the university training does not serve the needs of the market. The
reasons are twofold: translation teachers are not always translators themselves (and thus
their knowledge is only theoretical, not practical), and the translation courses have to fit
into a wider curriculum restrictions set by national education institutions, which is not
always possible. In fact, professional translators and translation agencies are frequently
critical about university training programmes, claiming that they are inefficient, too
theoretical and not up to date with market demands. They argue that institutions fill the
market with graduates with no practical skills that would allow them to cope with
authentic problems. While those voices should be heard and considered, one has to admit
that there are numerous experienced translators teaching practical and theoretical
translation. Trainers understand the need for contact with authentic assignments and
authentic people working in the profession. Therefore, they share their own experience
and invite professionals to the classroom, discuss translation project workflow, authentic
scenarios, helping to develop a set of skills that can match demands of the market. Pym
(2009) mentions that it is desirable for professional organisations to enter the classroom
and share their expectations. Recruitment on the basis of portfolio alone too often ends
with all required skills being “developed in-house rather than at the university” (idem.:
online resource). What is more, the translation market is not coherent in its expectations
Michał Kornacki
78
towards new translators, depending on whether the main area of interest concerns
translation, interpreting, or specialisation. As a result, there is no consensus on the issue,
and that is the reason why such initiatives as the EMT network are so desperately needed
and most welcome.
4.3. In-class activities
There are many competing ideas of what should happen in a classroom devoted to
translation. One of the most notable approaches made by Kiraly (2000) brings forward
the distinction between translation competence and translator’s competence. According
to Bernardini (2004), it can be categorised as a difference between translator training and
translator education. According to Pym (2009: online resource), the ‘training’ part (or
translation competence) is to be associated with “the (mostly linguistic) skills needed to
produce an acceptable translation […], the acquisition of which will always be a
combination of instruction and practice.” In short, these are the core skills that allow
translators to work and are most valued. On the other hand, ‘translator education’ (or
translator’s competence) comprises a “wide range of interpersonal skills and attitudes
[…] in addition to purely technical skills” (idem.). This is all the extra-linguistic
knowledge that allows a translator to function on the market. Such extra skills include
teamwork, interaction with other translators, terminologists, reviewers, project managers,
and end clients. This type of training requires more than to absorb data – it requires to
learn to find and assess the information by themselves, and to use it correctly. “[T]hey
should not just absorb professional norms from seeing their translations corrected; they
should be able to discover the norms and ethical principles, mostly through work on
‘authentic’ professional tasks or while on work placements, contributing to debates on
these issues as they go along” (idem.). Therefore, it is critical that young translators can
develop on many planes simultaneously; development of only technical skills may prove
to be a serious handicap in their career. Higher education institutions need to consider this
in the context of their long-term training programmes. Young translators have to learn
how to translate AND how to function in professional communities of translators in order
to succeed on the market.
The approaches like the one by Kiraly (2000) or Bernardini (2004) allowed to
challenge classical approach to translator training. Especially so when we consider that
Chapter 4. Translator training
79
the creation of situations where everyone thinks they can translate, basically through free
online machine translation, creates an urgent social need for training in how to use those
technologies (the basics of post-editing), when to trust them, and more especially when not
to trust them. That training can be called translator training, but a lot of it could also be called
language teaching, since the technologies can also be used as language-teaching tools (Pym,
2017: online resource).
In fact, the translator training may benefit from language-acquisition classes in a number
of ways. Many teaching techniques can be transferred to the translation classroom (e.g.
terminology searches). González Davies (2004) suggests, for example, the acting-out of
communicative situations and the use of discussion forums. In fact, the author describes
numerous teaching techniques that can be used for the training of translators. According
to Pym (2009: online resource), one of the most important issues in the translation
classroom concerns “the use of oral translation situations.” While it is commonly believed
that conference interpreting is way more difficult than written translation and, therefore,
it should be taught after written translation skills, the doctrine is being challenged. It is
mainly due to the fact that “the greater context-dependence of spoken language makes
translation purposes all the more obvious” (idem.).
In case of written translation, the diversification of courses includes diversification
of content through various combinations of the translation instructions, groupwork,
partial translation, gist translation, use of parallel texts, translation in social and/or
discursive contexts, to name a few (see, for example, Nord, 1997, 2005; House, 2000).
An interesting approach is presented by Kiraly (2000), Gouadec (2007) and Klimkowski
(2015b) who recommend all of the translation teaching techniques to be incorporated into
large, semi-authentic projects which would require students to work in small groups and
with pre-assigned roles (translator, reviewer, terminologist, project manager).
Michał Kornacki
80
4.4. Challenges for translation trainers
The translation industry is heavily affected by new technological developments,
especially in the IT. As a result, the higher education institutions are bound to include
those in their translator training programmes. They have to consider the impact of
translation memories, machine translation and content management systems as they have
a profound influence on the way translators do their job. Since they cannot be seen as
‘tools’ anymore (Pym, 2009), they have to take their rightful place in the training
programmes. It cannot be denied that they affect the nature of the translation process.
Contrary to the classical approach in which translators work with continuous texts, the
new technology forces an approach that requires professionals to work with “pretranslated discontinuous chunks and databases, and thus increasing the importance of
review processes” (Pym, 2009: online resource). What does it mean? Mossop (2003: 20)
said that “[i]f you can’t translate with pencil and paper, then you can’t translate with the
latest information technology.” If we refer this statement to the process of translator
training, we could say that you cannot teach state-of-the-art translation if you do not have
a deep understanding of how the technology can be used to improve the translation
process. Not only do we need specialised programmes at MA level, or as advanced postgraduate short-term courses, but also highly specialised individuals who can transfer their
knowledge in such areas as audiovisual translation, computer-assisted translation,
interpreting or localisation (Pym 2009, 2017; Kornacki, in print).
The changes in the market call for a new approach to translator training. Not all
students choose translation track to become translators. Therefore, a pro-student oriented
teacher cannot set the same learning objectives for everyone. Students vary in skills and
competences. Therefore, an effort should be made to develop self-regulatory skills in
students. Zimmerman (2000: 14) defines self-regulation as “self-generated thoughts,
feelings, and movements that are arranged and cyclically adapted to the fulfilment of
personal goals.” Instead of pre-set goals, students need to be given a set of tools allowing
them to define their own goals. Only then can they achieve success in professional life.
Thus, the role of the teacher is focused on providing students with insights on the current
situation on the market in terms of employment options, competition, business relations
and, in case of translators, pros and cons of running own business enterprise.
Chapter 4. Translator training
81
Another important issue to be dealt with the use of technology in translator’s
workshop. The role of the teacher is not so much as to teach students how to handle
needed software (word-processors, optical-character recognition tools, basic graphics
editing) but to show students that they need those skills in order to succeed as a translator.
Teaching has become secondary (albeit still very important) to presenting and justifying
practical requirements for the job. It includes CAT tools, mentioned before. A necessity
in many cases, CAT tools are tremendous assets for many translators and translation
agencies which, at the same time, set traps for the unwary. Hence the need for teachers
who are (or have been) also practitioners of the trade since it is them who know where,
when and what problems to expect when utilising CAT tools.
The above is inextricably linked to the fact that the pace of developments in
technology and changes in the market dictate the course of action of translation (if not
all) teachers. In author’s opinion, it is the greatest challenge for translation trainers – to
remain one step ahead of the market and prepare students (both in terms of hard skills and
mental attitude – belief in own skills and ability to adapt to the market) for professional
life after graduation. Therefore, it is important to prepare future translators accordingly,
also in terms of choices they will have to make in everyday work. The issue of translation
strategies appears crucial both for translation theory and practice. Therefore, it is the main
subject of the next section.
4.5. Translation strategies in the translation classroom
The previous sections have presented the importance of developing translation
competence in terms of practical skills to be applied in the translation process. The
significance of those skills cannot be denied; it is vital to remember that it is not enough
to know a language and how to operate the software in order to be a translator. Each
successful translator has to be able to make conscious decisions and follow certain
procedures in order to achieve the ultimate goal. It is paramount to introduce those
‘guidelines’ into the translator training process, and subsequently the translation
classroom, lest the translation course produces individuals who have all the necessary
tools to produce a translation but lack the method to do so. Therefore, the following
section focuses on translation strategies as a theoretical background for both trainee and
professional translators.
Michał Kornacki
82
The discussion starts with an attempt to disambiguate the term. Translation
strategies represent a number of approaches to similar translational problems. Płońska
(2014: 67) observes that “not only is the term strategy used to describe different concepts
but also various terms are used to express the same meaning” (see also Chesterman, 1997;
Hejwowski, 2004). Strategies entail a whole range of mental and practical processes –
“forms of explicitly textual manipulation” (Chesterman, 1997: 89) – that can be grouped
into methods (organised ways of conduct, employing translation techniques to achieve a
certain goal), procedures (dealing with the smallest units of a text, like a word), and
techniques (or how methods and procedures are applied in the translation process).
However, there is no consensus among academics as to the exact nature of translation
strategies. Venuti (1988) uses the concepts of foreignisation and domestication to discuss
the issue, proposing that translation strategies are all about developing a method to
translate a given text. Lörscher (1991: 8) proposes to define the notion as a “potentially
conscious procedure for solving a problem faced in translating a text, or any segment of
it.” He uses the problem-centred approach to discuss the way translators deal with such
issues, defining a new division between ‘strategic’ and ‘non-strategic’ behaviour of the
translator (idem.; Chesterman, 1997). Cohen (1998) supports Lörscher’s (1991) line of
thought by arguing that it is consciousness that helps to draw a line between strategies
and non-strategies. On the other hand, Bell (1998) juxtaposes global (concerning the
entire texts) and local (concerning sentence-level translation units) strategies. It draws
partially from Newmark’s (1988) division who made a clear distinction between methods
and procedures in translation (which could be equated with Bell’s ‘global’ and ‘local’
division). In Newmark’s framework, translation methods should be considered in the
context of whole texts, while translation procedures concern smaller parts of texts, like
phrases and sentences. His list of translation methods includes (idem.):
•
word-for-word translation, retaining the word order of the SL with individual
words translated by their most common meanings, with no regard for the context;
•
literal translation, converting SL grammar structures to their TL equivalents, again
with individual words translated by their most common meanings, with no regard
for the context;
•
faithful translation, striving to reproduce SL meaning within the constraints of the
TL;
Chapter 4. Translator training
•
83
semantic translation, again endeavouring to reproduce SL meaning within the
limits of the TL with attention paid to the aesthetic value of both SL and TL text;
•
adaptation, adapting cultural references in the SL text to the cultural background
of the TL audience;
•
free translation, reproducing the SL text in the TL with no regard for the form,
style and/or content of the SL;
•
idiomatic translation, reproducing the message of the SL using target-specific
proverbs and expressions, carrying the same overall meaning as the SL text;
•
and communicative translation, reproducing contextual meaning of the SL text in
the TL.
On the other hand, Newmark (idem.) lists the following procedures:
•
transference – transfer of an SL word to a TL text;
•
naturalisation – adaptation of an SL word to a standard pronunciation first, and
regular morphology in the TL second;
•
cultural equivalent – replacement of a culture-specific word in SL text with a TL
one;
•
functional equivalent – a process which focuses on the function of a given word
and requires a translator to use a culture-wise neutral word;
•
descriptive equivalent – explains the meaning of “culture-bound terms” in several
words (p. 83);
•
synonymy – a “near TL equivalent” (p. 84).
•
through-translation – can be referred to as calque or loan translation.
•
shifts, or transpositions – changes of the grammar form between SL and TL texts
•
modulation – a process which focuses on the perspective of SL and TL texts. It
concerns reproduction of the original message in the target text with attention to
the norms governing the TL;
•
recognised translation – takes place when a translator “uses the official or the
generally accepted translation of any institutional term” (p. 89).
•
compensation – a process which allows a translator to add extra meaning in one
sentence in order to compensate for the loss of meaning in another one;
•
paraphrase – explains the meaning of the culture-bound terms in greater detail
than in the case of a descriptive equivalent;
Michał Kornacki
84
•
couplets – a combination of two different procedures;
•
notes – providing extra information in the TL;
•
componential analysis – “comparing an SL word with a TL word which has a
similar meaning but is not an obvious one-to-one equivalent, by demonstrating
first their common and then their differing sense components” (p. 114).
Płońska (2014: 68) notes that “Newmark’s classification of translation methods and
procedures partially overlaps that of Vinay and Darbelnet (1958/2000) but is much more
detailed. It is also based on the opposition between literal and free translation.” This
overlap may result from the fact that the discussion on tackling problems in translation
can be traced back to Cicero (Robinson 1997, 2002) and his deliberations on translating
sense-for-sense, instead of word-for-word. Each subsequent work published over the
years added to the general knowledge on the subject, stimulating new thought and
approaches (see Barbe, 1996). However, while the distinction between procedures is more
or less clear in the context of the practical application, their division in the light of the
Translation Studies is not, e.g. Lörscher (1991) uses a cognitive approach to the problem
while Chesterman (1997) advocates textual approach. The fact is that literal (or wordfor-word translation of the source text) and free (i.e. sense-for-sense) translation can take
on many different forms, depending on the context and the background of their defining
researcher, e.g.
(…) word-for-word translation vs. sense-for-sense translation, source-oriented translation vs.
target-oriented translation, direct translation vs. oblique translation (by Vinay and Darbelnet),
adequacy vs. acceptability, formal equivalence vs. dynamic equivalence (by Eugene Nida),
semantic translation vs. communicative translation (by Peter Newmark), overt translation vs.
covert translation (by Juliane House), documentary vs. instrumental translation (by
Christiane Nord), foreignization vs. domestication (by Lawrence Venuti), etc.
(Sun, 2012: online resource)
The division presented above reflects various approaches and corresponds to different
goals of translation. For example, the foreignization vs domestication pair (as suggested
by Venuti (1995, 1998) above addresses the issues of cultural intervention. The author
provides examples of Roman translators ‘domesticating’ Greek poetry into Latin by
exchanging Greek background with the Roman one, removing the Greek author, and
“passing the translation off as a text originally written in Latin” (1998: 241). On the other
Chapter 4. Translator training
85
hand, he proposes foreignization in order to “register the linguistic and cultural
differences of the foreign text,” thus “making the translated text a site where a cultural
other is not erased but manifested” (p. 242). The example of Venuti shows that translation
methods and procedures are two sides of the same coin. They are selected on the basis of
needs and aims of translators, with different approaches having the same origins in the
free and literal translation. It should be noted, however, that the ultimate division between
‘literal’ and ‘free’ has been one of the key issues in Translation Studies and, at the same
time, the subject of criticism over the ages (Sun, 2012; Hatim and Munday, 2004). Hatim
and Munday (idem: 230) follow Steiner (1975/1998) by claiming that the debate is
“ultimately sterile since it does not encourage further examination of the internal and
external contextual constraints which affect the translation strategy and function.” In this
view, both concepts can be regarded as mere starting points for a more detailed discussion
on translation strategies.
Many researchers used this basic division in order to develop their own taxonomies
of translation procedures. One of them, suggested by Newmark (1988), was presented at
the beginning of this section. However, it is very detailed and a simpler classification may
be of use. For example, a classification suggested by Vinay and Darbelnet27 (1958/2000)
proposes two main methods of translation. They distinguish between direct translation,
having much in common with word-for-word (literal) translation, and oblique translation,
allowing the translator to interpret (sense-for-sense).
27
The classification proposed by Vinay and Darbelnet is almost 60 years old when this book is
composed. Nonetheless, it has aged well and remains in high regard (see, e.g. Chesterman 1997; Bell
1998; Hatim and Munday 2004; Sun 2012; Wach 2013; Płońska 2014; Pym 2014; Pym and TorresSimón 2014; Munday 2016) even though the authors themselves saw that, for example, their idea of
equivalence is not always possible to attain. What it means is that the rendering of a source language
phrase with vocabulary counterparts in the target language does not guarantee a successful translation
since external context has to be taken into account (Panou 2013). Pym (2014) suggests that the reason
behind the classification being so successful and valid even today is that the categories it uses are
somewhat general and blurred. (…) “students struggle to distinguish between “transposition”,
“modulation” and “adaptation”, and, as mentioned, the nature of “equivalents” was never really
explained” (idem.: 12). As a result, the classification can be used with any metalanguage, provided
proper attention to the source text is paid (idem.).
Michał Kornacki
86
DIRECT TRANSLATION
OBLIQUE TRANSLATION
(word-for-word, or literal translation)
(sense-for-sense, or free translation)
techniques:
• borrowing
• calque
• literal translation
techniques:
• transposition
• modulation
• equivalence
• adaptation
Figure 13. Taxonomy of translation procedures according to Vinay and Darbelnet (1958/2000)
The approach presented in this book focuses on procedures since CAT tools do not
advocate looking at a text as a whole. While some CAT tools attempt to negate the issue
(e.g., live preview feature in memoQ, see Figure 14), not all of them do so, or the preview
feature is not easy to use (e.g. SDL Trados). The main problem is that majority of display
screen is taken up by the main editor with only tiny part (if at all) reserved for the preview.
As a result, a different type of approach is proposed. CAT tools pre-process the text,
dividing it into smaller translation units, and translator works segment by segment28, thus
focusing on particular problems at hand, instead of looking at a bigger picture.
Figure 14. Context preview in Memsource.
28
It is worth noting that in CAT tools a segment constitutes a unit of translation. While in general a
segment fits into the Newmark’s (1988) idea of “natural unit of translation” since it is frequently made
of a single sentence. See Chapter 5 for more details.
Chapter 4. Translator training
87
Waliński (2015: 56) notes that “[translation] procedures can be employed at three levels
of language: (a) the lexicon; (b) the grammatical structures; and (c) the message, which
stands for higher elements of text, including, besides sentences and paragraphs, certain
situational utterances that convey broader meanings.” Some utterances may carry an extra
meaning, which is not immediately obvious. In her paper on humour, Brzozowska (2010)
brings the example of Asian tigers which, apart from the literal meaning, may have other
meanings in the context of economic and demographic potential of China and its influence
on the global market.
4.5.1. Direct translation procedures
Vinay and Darbelnet observe that, while it is possible to “transpose the source language
message element by element into the target language,” sometimes “translators may also
notice gaps, or lacunae, in the target language which must be filled by corresponding
elements, so that the overall impression is the same for the two messages” (2000: 85). In
such a case, the translator can use either a parallel concept or category to convey the
meaning. It can be done with direct translation procedures:
a) borrowing – generally regarded as the simplest of all procedures, it involves using
foreign words and phrases in the target text. Frequently, the technique is used to
deal with new technologies entering a given locale rapidly. In such situation, the
SL name for a device or phenomenon starts to be used before a localised version
is produced. As a result, it becomes recognisable and popular under the foreign
name, and local name is not needed. Examples of lexical forms borrowed directly
from English and used daily in Polish include, e.g. ‘smartfon’ (English
smartphone) and ‘tablet’ (with derivatives like ‘phablet’) (see Ni, 2009)
Vinay and Darbelnet (1958/2000) note that borrowing may also be used to
achieve a certain stylistic effect, like introducing the feel of the SL in the target
text. This technique may be used to show the nationality of a character through
language, e.g. in the Polish dubbing to the movie “Sing” by the Illumination
Entertainment, Gunter, the pig, uses German Ja! instead of Tak (Yes), which
combined with his accent makes it clear to the audience that he is German. In this
case, a translator opted to leave the foreign elements (within an already foreign
text) intact in order to retain the desired effect. On the other hand, there are many
terms in Polish which have been borrowed from English and function in their SL
Michał Kornacki
88
form despite having a Polish equivalent, e.g. ‘Dział Human Resources’ instead of
‘Dział Kadr’.
b) calque – this technique is a type of borrowing which involves literal translation of
an SL item into TL. Wach (2014) lists two types of calque: semantic and
structural. A semantic calque preserves its foreign meaning in the TL, and is quite
common in computer terminology, e.g. ściągać – download, used to denote the
process of data transfer from an online server to a local computer. Structural
calques involve creating new formations in Polish (including the way they are
used in the target language) by borrowing foreign morphological structure and
meaning. Examples of both types are presented below:
Table 7. Semantic and structural calques according to Wach (2014: 162-163)
Semantic
administrator – administrator
aplikacja – application
bufor – bufor
ikona – icon
katalog – folder
pakiet – package
Structural
poczta elektroniczna – e-mail
baza danych – database
twardy dysk – hard drive
pamięć wirtualna – virtual memory
podział czasu – time stamp
system operacyjny – operating system
A different approach is presented by Smith (2006), who proposes the following
division of calque types:
•
phraseological calque – in which phrases are translated word-for-word
(e.g. flea market – pchli targ);
•
syntactical calque – in which syntactical construction (or function) of the
SL is replicated in the TL (e.g. healthy food – zdrowa żywność);
•
loan-translation – in which a word is translated morpheme-by-morpheme
into the TL (e.g. skyscraper – drapacz chmur);
•
semantic calque – in which extra meanings of a given SL word are
transferred to the TL word with primary meaning retained in the TL (e.g.
[computer] mouse - myszka);
•
morphological calque – in which inflexion of an SL word is transferred to
the TL word (e.g. to click – kliknąć).
Chapter 4. Translator training
89
As was noted previously, borrowing and calque are so strongly related that calque
can be viewed as a type of borrowing (Vinay and Darbelnet, 1958/2000) since it
is not always immediately clear where the border between the two lies. Many
cases constitute an amalgam of the categories mentioned above. It may be
connected to the fact that Polish has a very long history of borrowing expressions
from French, Russian, German, and more recently, English. These include
business, sport, technology, culture, and other areas. One of the reasons behind
the number of borrowings and calques in Polish is that they are relatively
straightforward solutions to problems encountered in the translation process.
However, as such, they should be applied with caution. Even though many
expressions came to be popular words, used in various context, they are not always
correct and should be avoided in translations, e.g. to eat in fast-food [restaurant]
can be translated as zjeść coś w fast-foodzie, which is not right and should rather
be rendered as zjeść coś w restauracji/barze.
c) literal, direct, or word-for-word translation technique is used to render source text
in TL one word at a time with focus on linguistic rules of the target language.
According to Munday (2016), it provides best results when applied to two
languages of the same family, e.g. French and Italian, preferably sharing the same
culture. In fact, the notion should be considered from two very different angles:
good and bad practice. It is a good practice to use the technique in the technical
translation of scientific, technical, technological or legal texts, i.e. texts focused
primarily on the direct meaning and not the style of the text. The problem with
this technique may arise from the fact that its application may render text (or parts
of it) incomprehensible, e.g. mistranslation of idioms, or when applied in
translation between analytic (non-inflecting) and synthetic (inflecting) languages,
like English and Polish, for example. In extreme cases, it may render even the
grammar completely unintelligible and oblique translation must be used (see
Vinay and Darbelnet, 1958/2000; Munday, 2016).
Literal translation strategy has many uses, e.g., in CAT-based translations
because the majority of texts translated with the use of CAT tools are specialised
texts, and the tools themselves enforce bit-by-bit (segment by segment)
translation. Texts that may read awkwardly when literature is considered are
perfectly acceptable in the case of technical documents, e.g. manuals.
Michał Kornacki
90
The technique can also be seen in machine translation results, yet to much
lesser extent than some time ago. Hutchins (1995) mentions that early MT systems
were based on vast databases of texts and their translations. As a result, their
output was limited to providing direct, or literal, translations of queried strings.
Subsequent development of tools allowed to use better grammar structure, but
many words were retained in the original language. Today, hybrid MT systems
combine these technologies with automated post-processing. Yet, even such
results, good as they are, are frequently used by translation agencies as a rough
first translation to be post-edited by a human translator (see Pym, 2012).
The strategies described above have many advantages and are the obvious choice in
multiple situations. However, if a target text is still lacking in quality after the three
procedures were used, oblique translation procedures have to be employed.
4.5.2. Oblique translation procedures
Vinay and Darbelnet (1958/2000) note that, due to metalinguistic and structural
differences, it may not be possible to transpose certain stylistic effects into the target.
Moreover, if it were, the transposition would ruin syntax and/or lexis. As a result, more
complex methods have to be applied that grant translators a “strict control over the
reliability of their work” (p. 84). Such control can be asserted through oblique translation
procedures:
a) the “method called transposition involves replacing one word class with another
without changing the meaning of the message” (p. 88). However, in the case of
transposition as a translation technique, it is a word class, not a sign, that is being
replaced without changing the meaning of the phrase or sentence. It is a very
versatile technique that can be applied extra- and intra-linguistically (Waliński,
2015). For example, a sentence reading ‘He announced he would start in the
upcoming election’ can be transposed as ‘He announced his start in the upcoming
election’. Both utterances can be translated, depending on the context and style,
as ‘Ogłosił, że wystartuje w nadchodzących wyborach’ and ‘Ogłosił swój start w
nadchodzących wyborach,’ which are examples of intralinguistic transposition in
Polish in their own way. The fact that they carry the same meaning allows for
their extralinguistic transposition, as shown in the figure below:
HE ANNOUNCED HE WOULD START
IN THE UPCOMING ELECTION
91
=
HE ANNOUNCED HIS START
IN THE UPCOMING ELECTION
TARGET
TRANSLATION
SOURCE
Chapter 4. Translator training
OGŁOSIŁ, ŻE WYSTARTUJE
W NADCHODZĄCYCH WYBORACH
=
OGŁOSIŁ SWÓJ START
W NADCHODZĄCYCH WYBORACH
Figure 15. Example of extralinguistic transposition
It can be argued that transposition allows to add a different stylistic value to the
original expression. As a result, it is primarily meant to improve the final
readability, or elegance, of the target text.
b) modulation “is a translation technique in which there is a change in semantics and
point of view” (Tardzenyuy, 2016: 52). It involves changing the message
depending on the context, e.g. in cases when literal or transposed translations do
not sound right in the target language, even though they are correct grammar-wise.
The use of this technique may be obligatory or optional (see Munday, 2016),
depending on the wider linguistic context. Obligatory modulation may be seen in
grammatically pre-established contexts, e.g. ‘Ona ma 12 lat’ should be translated
as ‘She is 12 years old,’ and not ‘She has 12 years’ as word-for-word translation
might suggest. Examples of optional modulation may include, e.g. ‘it is not
difficult to show’ → ‘łatwo jest pokazać’ [lit. ‘it is easy to show’] (see Vinay and
Darbelnet, 1958/2000). Munday (2016) notes that the difference between
transposition and modulation is that modulation is “the touchstone of a good
translator,” and transposition “shows a very good command of the target
language” (Vinay and Darbelnet, 1958/1995: 246).
c) equivalence involves reproduction of (frequently extra-textual) meaning of the SL
through an entirely different method. Most obvious examples of this technique
include exclamations: ‘Oi’ – ‘Hej’, or ‘Ouch!’ – ‘Auć!’; expletives: ‘Damn!’ –
Michał Kornacki
92
‘Cholera jasna!’; onomatopoeic sounds: ‘Woof woof’ – ‘Hau Hau’, ‘Oink Oink’
– ‘Chrum Chrum’; proverbs: ‘Good riddance to bad rubbish!’ – ‘Baba z wozu,
koniom lżej!’; and idiomatic expressions: ‘to see pink elephants’ – ‘widzieć białe
myszki.’ Unlike the previous two oblique translation strategies mentioned above,
due to the fact that modulation addresses specific cases of equivalence, including
whole sets of proverbs, idioms, expressions, etc., this technique is regarded as
fixed in most cases.
d) adaptation, the last oblique technique to be mentioned here, involves adaptation
of the message for the benefit of the TL audience who may lack cultural
background to understand the message fully. Vinay and Darbelnet (1958/2000)
call it a kind of ‘situational equivalence,’ since it is triggered in certain contexts,
like real life or a movie scene. In fact, Jarniewicz (2000) suggests that adaptations
are quite common in book and movie translations. Bogucki (2013: 51) lists
examples from Shrek movie and its dubbing by Bartosz Wierzbięta, probably one
of the most significant and explored examples of movie dialogues adaptation in
Poland: “(…) ‘the Muffin Man’ becomes ‘Muchomorek,’ a character from a
Czechoslovakian cartoon that was all the rage in Poland in the 1970s. Similarly,
‘awful cheese’ becomes ‘ser Podlaski,’ a type of cheese available all over Poland.”
4.5.3. Translation procedures in the context of CAT tools
While it could be agreed that general translation rules and strategies apply to CAT-based
translations as well, one should consider the implications of using a sophisticated
computer programme in translation. On the one hand, the translator deals with a body of
text that is to be translated from SL to TL and all proper language and translation
procedures have to be considered. On the other hand, the translator has to be fluent in the
use of computer technology in general, and a given CAT tool in particular. What is more,
CAT tools promote segmentation of the source and target texts. Hence, sentence-based
(or smaller) segmentation29 reduces the applicability of translation methods in favour of
translation procedures since it is difficult to see a block of text over segment boundaries.
What is more, segmentation forces translator to work on separate sentences (some CAT
tools hide translated segments), which results in reduced fluency of the translated text and
further limiting the use of translation procedures, e.g. extra-sentential compensation. This
29
See Chapter 7 of this book for more information on segmentation.
Chapter 4. Translator training
93
may be partially negated by the fact that CAT tools are used predominantly in specialised
and technical translation. However, it cannot be argued that it is a problem.
By their nature, CAT tools also reduce the use of translation procedures in general.
If a CAT tool finds a match in a translation memory, it propagates the match in translation
(frequently translators have the option to change propagation settings) automatically
applying it throughout the target document (or project). What it means is that it also
results in propagation of errors, if there are any in the translation memory. Since the
translation was done automatically, the translator could not use any procedures, resorting
to use a previous translation (by himself or by some third party).
The following section continues the discussion on the issue of the translator’s role
in the process of CAT tools-based translation in the context of assisted and automated
translation. Furthermore, it introduces and discusses in detail such concepts as CAT tools
and translation memory, their advantages and disadvantages regarding the quality of
translation. What is more, it tries to assess the role of a human being in modern translation,
with a focus on what ‘assisted’ translation actually means.
Chapter 5. Computers in translation
The discussion concerning computer-based translation would be incomplete without at
least one brief attempt at outlining how such a method of translating came to be. Since
the research presented in this paper is focused on computer-assisted translation (CAT),
attention needs to be paid to its genesis, development and application.
As was mentioned previously, translation memory technology – a technology
behind all CAT systems – is fairly new, considering the entire history of translation. It
started to be used commercially in the 1990s by the SDL company (Bowker, 2002;
Bloodgood and Strauss, 2014; Doherty, 2016). However, the entire idea of having a
machine to help translate, or replace a translator altogether, is much older.
5.1. Machine Translation (MT)
The concept of machine translation is based on an idea that a mechanical device is able
to produce a translation without any human assistance. It is important to understand the
value of the word “any” in this statement. True machine translation process denies the
possibility of human involvement. Of course, such MT variations as MAHT (machineaided human translation) and HAMT (human-aided machine translation) do exist and are
successfully employed, but their scope is somewhat unclear and “the term computerassisted translation can cover both of them. (…) [T]he central core of MT itself is the
automation of the full translation process” (Hutchins 1995).
Disregarding 17th-century dreams (Hutchins, 2010) about translating machines, it
can be said that modern history of machine translation started in the early 1930s. In 1933
George Artsrouni designed a device which used paper tape in order to find word
equivalents between two languages. However, it was Petr Petrovich Troyanskii, a Russian
scholar and educator, who formulated the basis for modern study on machine translation.
According to Hutchins (idem: 434), he:
Chapter 5. Computers in translation
95
envisioned three stages of mechanical translation: first, an editor knowing only the source
language was to undertake the ‘logical’ analysis of words into their base forms and syntactic
functions; secondly, the machine was to transform sequences of base forms and functions
into equivalent sequences in the target language; finally, another editor knowing only the
target language was to convert this output into the normal forms of his own language.
Troyanskii envisioned both bilingual and multilingual translation.
This view is consistent with the idea of HAMT which requires human pre- and postprocessing of text, while the translation itself is done by a machine.
Regrettably, Troyanskii was known only in Russia and his research had no
influence on the global thought on the issue. His ideas came well before the Weaver’s
memorandum (1949). The Warren Weaver’s memorandum30 had a huge impact on the
future development of machine translation. Weaver suggested the application of
cryptography, statistical methods, Claude Shannon’s information theory and exploiting
the logical features of languages. As it turned out, the reception of the memorandum was
varied. Some people rejected it outright, arguing that the complexity of any given
language is much more than any machine can successfully deal with. However, there were
also those who found the idea interesting and challenging. One of those was Erwin Reifler
who proposed how basic word-for-word translations can be used. He introduced the idea
of pre- and post-editing, and the use of language rules. It was then that statistical approach
to machine translation issues was proposed by Abraham Kaplan at the Rand Corporation.
In 1952 the first MT conference was held at the Massachusetts Institute of
Technology (MIT). The conference resulted in many new ideas put forward, mostly
concerning pre- and post-editing, “micro-glossaries as means of reducing ambiguity
problems” (selecting appropriate target lexical items), “and for some kind of syntactic
structure analysis” (Hutchins, 2006: 376). In addition, one of the conclusions of the
conference was that a public presentation of a working MT system was needed in order
to attract funding. Such presentation was held on 7 January 1954 at Georgetown. The
demonstration – an MT system used to translate a pre-selected sample from Russian to
English, basing on a closed set of 250 words and a few grammar rules – attracted a great
deal of attention. The presented quality of translation assured substantial funding in the
USA, as well as new MT projects blossoming around the world.
30
From Language Today, no.6, March 1998, pp. 22-23
The following years saw a massive increase in general interest in machine
translation. USA, many European countries and the Soviet Union had their own MT
projects which included a broad range of research on “formal linguistics (particularly in
the Soviet Union), semiotics, logical semantics, mathematical linguistics, quantitative
linguistics, and nearly all of what would now be called computational linguistics and
language engineering” (Hutchins, 2006: 377).
Over time more and more researchers became aware of numerous and complex
linguistic problems that could not be overcome. There was no appropriate technology
available yet. What is more, several prominent individuals (e.g. Bar-Hillel, see Green,
Heer, and Manning, 2015) claimed that the general concept of MT was flawed. At the
time it was believed that the overall goal of MT was to create a fully automated highquality translation system (FAHQT), which would, in turn, produce a translation of equal
quality to that produced by a human translator. However, ongoing research suggested that
the idea was not only unrealistic with the available technology – it was impossible in
principle.
As a result, MT funding bodies in the USA saw the need for comprehensive analysis
of the current research and its future application. Thus, the Automatic Language
Processing Advisory Committee (ALPAC) was formed in order to examine the situation.
The 1966 ALPAC report said that “MT was slower, less accurate and twice as expensive
as human translation and that there is no immediate or predictable prospect of useful
machine translation” (Hutchins 1995, online). The most important conclusion, in the
context of this paper, was the recommendation for development of machine aids for
translators (e.g. automatic dictionaries), and the continued support of basic research in
computational linguistics, which are, in fact, the core principles of CAT tools. The shift
of focus from MT, which turned out to be costly, time-consuming and of low quality
(according to the report), to other computer-based aids for translators, put an end to the
research on MT in the USA and most of the world.
The following years saw a decline in MT research in the USA. It was continued in
Canada and Europe. Canada required reliable translation solutions to support its
government’s bicultural policy. Europe had to deal with the rapidly increasing volume of
translation within the European Community.
There were some major breakthroughs like the creation of Météo, for example,
which was first employed to translate Canadian weather forecasts in 1976. It was
successfully used until September 2001, after which it was replaced by a competitor
Chapter 5. Computers in translation
97
programme. Other important systems included French TITUS, Chinese CULT, Japanese
ATLAS, notoriously successful SYSTRAN, in-house Xerox Corporation MT and an MT
system developed by the Logos Corporation (see Hutchins 2010).
Numerous solutions and successes fuelled extensive research on MT. The
development of computers and new research in computational linguistics allowed to
create new forms of data processing, like corpus-based or statistical machine translation.
While all the systems mentioned above were (and some of them still are) successful
in translation for specific purposes, researchers (and service providers) realised that while
these tools can be used to translate vast amounts of data, it is still not enough and the
majority of translation is done by human translators. Moreover, they did not need a system
which either replaces them or reduces them to pre- and post-editor role. What they needed
were tools assisting a translator (dictionaries, termbases, word processors, text converters,
etc.), not replacing him. The role of MT changed from a utopian system meant to replace
human translator to a component of translator’s workshop. Apart from that it was still
successfully employed in domain-restricted systems for specific purposes (Xerox,
Microsoft) and as a service for non-translators, which needed to understand the content
without the need for professional translation (auto-translation of web pages, free online
MT systems like Google Translate) (Hutchins, 1995, 2006).
5.1.1. Rule-based MT (RBMT)31
A rule-based approach to MT is one of the very first strategies ever developed. “More
complex than translating word to word, these systems develop linguistic rules that allow
words to be put in different places, to have different meaning depending on context, etc.”
(Costa-Jussà, Farrús, Mariño, and Fonollosa, 2012: 248). Apart from rules (grammatical,
lexical, and stylistic), a standard RBMT system includes software used to process these
rules and a significant number of bilingual dictionaries for each language pair. It can
grow, and it is easy to maintain. Its main advantage is the fact that it allows for deep
syntactic and semantic analysis of a body of text, but it requires significant knowledge
and a great number of rules. It is a huge drawback. An RBMT system, in order to provide
acceptable results, requires a lot of time and linguistic resources to build. As a result, it is
very expensive. What is more, when a rule needs refining, there is no guarantee that such
31
Examples of RBMT systems include: the 1954 Georgetown experiment; SYSTRAN
refinement will influence the overall accuracy of the system (Costa-Jussà, Farrús, Mariño,
and Fonollosa, 2012).
5.1.2. Statistical Machine Translation (SMT)32
SMT, as the name suggests, utilises a statistic approach to multilingual corpora. An SMT
system uses numerous large corpora in order to find statistical patterns and rules, which
can be later implemented in the profession of translation. The greater number of corpora
available for the system, the better the results (text fluency retained in most cases).
Contemporary SMT systems use phrases, rather than words, like building blocks, and
produce translation “using the overlap in phrases” (Costa-Jussà, Farrús, Mariño, and
Fonollosa, 2012: 249). The biggest drawback of SMT systems is that they require
significant amounts of data to be processed on-the-fly, making them computer-resource
dependent. As a result, significant computational power is required (SYSTRAN, 2016).
Table 8. RBMT vs SMT
RULE-BASED MT
+ Consistent and predictable quality*
+ Out-of-domain translation quality
+ Knows grammatical rules
STATISTICAL MT
– Unpredictable translation quality
– Poor out-of-domain quality
– Does not know grammar
+ High performance and robustness
+ Consistency between versions
– High CPU and disk space requirements
– Inconsistency between versions
– Lack of fluency
– Hard to handle exceptions to rules
+ Good fluency
+ Good for catching exceptions to rules
– High development and customisation costs
+ Rapid and cost-effective development costs
provided the required corpus exists
*) System strengths marked with bold font face
Rule-based MT provides good out-of-domain quality and is by nature predictable.
Dictionary-based customization guarantees improved quality and compliance with corporate
terminology. But translation results may lack the fluency readers expect. In terms of
investment, the customization cycle needed to reach the quality threshold can be long and
costly. The performance is high even on standard hardware.
32
Examples of SMT systems include: Google Translate, Microsoft Translator
Chapter 5. Computers in translation
99
Statistical MT provides good quality when large and qualified corpora are available.
The translation is fluent, meaning it reads well and therefore meets user expectations.
However, the translation is neither predictable nor consistent. Training from good corpora is
automated and cheaper. But training on general language corpora, meaning text other than
the specified domain, is poor. Furthermore, statistical MT requires significant hardware to
build and manage large translation models.
(SYSTRAN, 2016)
Considering the requirements towards a successful MT system, and pros and cons of the
two systems presented, it is clear that another system is required. In order to accommodate
all requirements, a concept of a hybrid MT was born.
5.1.3. Hybrid MT
Hybrid MT system is basically a combination of the best feature of both RBMT and SMT
approaches. Unlike before, here a source text is processed in one of two ways:
•
translation is done using the rule-based system, and then the output is adjusted
with the help of the statistical system, or
•
rule-based system is employed to pre-process a source text, then an SMT system
translates it, and finally the RBMT system takes over again in order to adjust the
output.
As a result, an observable increase in output quality, as well as processing flexibility and
control, can be noticed. What is more, a combination of the two approaches allows to
reduce the amount of data required to train the software. Additionally, with decreased
size of constituent corpora, less computing capacity is required (SYSTRAN 2016).
All the research in the field of MT33 influenced the statistical approach to bilingual
text alignment34, which in turn allowed translators to store and access previous
translation. This led to the creation of “translation memory,” the core element of each and
any CAT tool.
33 See Costa-Jussà, Farrús, Mariño, and Fonollosa (2012), Costa-Jussà (2015), Sin-wai (2015), Shiwen,
Y. and B. Xiaojing (2015), Yang, L. and Z. Min (2015) and SYSTRAN (2016)
34 See Gale W. A. and K. W. Church. 1991. A program for aligning sentences in bilingual corpora.
Technical Report 94. AT&T Bell Laboratories. Statistical Research.
5.2. Computer-assisted translation
The ALPAC report and the resulting shift of focus in MT research allowed researchers to
distance themselves from the utopian assumptions behind MT and work both on linguistic
and computational solutions that could be applied to machine translation. It led, at least
partially, to corpus-based and statistical research, which laid the foundations for
computer-assisted translation.
The corpus-based structure as well as new data storage and retrieval methods
allowed to lay principles for translation memory (TM). TM, a bilingual corpus itself,
allows to access and use previous translations in current projects. A powerful tool in itself,
it was not enough. Translators required adaptable tools that could be used to facilitate the
translation process. This need was recognised by the market, and first CAT tools started
to emerge. The first proto-CAT tool was the Translation Support System (TSS) developed
by Automated Language Processing Systems (ALPS) in the mid-1980s. However, the
market was not yet technologically ready for such tools. Most translators still used
typewriters. However, soon typewriters started to be replaced with home personal
computers (PCs) equipped with word processors, and later connected to the Internet. By
the mid-1990s the flexible and multi-purpose PCs were so popular that buying one was
no longer a heavy financial burden.
With the background ready, it was easier to introduce new CAT tools to the market.
A number of tools appeared – Trados, Translator’s Workbench, MultiTerm (terminology
database), Translation Manager 2, Transit, Déjà Vu or Eurolang Optimiser
(discontinued). Their effectiveness was varied, but “it was Trados – thanks to successful
European Commission tender bids in 1996 and 1997 – that found itself the tool of choice
of the main players, and, thus, the default industry standard” (Garcia 2015). The
competition was fierce, however, and forced those first CAT tools to evolve. By the end
of the 1990s, such features as translation memory, alignment tools, terminology
management and various file processing filters were available in the most robust systems.
Current research and development of CAT tools is focused on refining existing
features and adding new ones, like, for example, AdaptiveMT introduced in SDL Trados
2017. In fact, MT has never been forgotten. On the contrary, it has been incorporated in
various forms in the majority of contemporary CAT tools as one of the modules to be
Chapter 5. Computers in translation
101
used. Kilgray’s memoQ, for example, allows to choose one (or more) commercial MT
plugin to be employed in the translation grid.
In commercial desktop programmes the majority of these plugins require a regular
fee to be paid in order to use them since they are external solutions. SDL’s new MT
solution, AdaptiveMT is a free MT tool that “learns and improves seamlessly,
continuously and in real-time to save time and money by minimising future post-editing”
(SDL, 2017). It is meant to build one’s own MT system, based on individual style and
translation technique.
Figure 16. MT plugins in memoQ
However, the general CAT idea of MT working in support of a translator was retained.
The use of MT solutions is entirely optional in the translation process. In fact, the most
prominent desktop CAT tools available on the market – Trados, memoQ, or Déjà Vu –
are so sophisticated that they leave no room for brand new programmes. This lack of
room for new classical solutions, easy access to a broadband Internet connection and
mobile expectations in general, have led to the development of an entirely new type
of CAT.
The problem with CAT tools is that they are mostly bound to the MS Windows
environment. So far, OS X and Linux users had no access to commercial CAT tools.
While open source solutions bring computer-assisted translation to these platforms, their
functionality is frequently limited due to lack of compatibility with the rest of the market.
In order to address this issue, some companies approached the problem from an entirely
different angle. They have bet on cloud computing. To appreciate the importance of this
new approach, it is critical to understand what “cloud computing” stands for. Mell and
Grance (2011) define it thus
Cloud computing is a model for enabling ubiquitous, convenient, on-demand network access
to a shared pool of configurable computing resources (e.g. networks, servers, storage,
applications, and services) that can be rapidly provisioned and released with minimal
management effort or service provider interaction.
In the context of a discussion on specialised software for translators, cloud computing
means accessing specialised software via a web browser. This way, the cloud software
crosses system requirement boundaries and can operate on any computer (or tablet, for
that matter – something that is unthinkable in the case of standard CAT tools) all over the
world, provided a compatible (meaning up-to-date) web browser is used, and an Internet
connection is established.
Currently, there are several cloud CAT tools available in Software as a Service
(SaaS)35 model. The most prominent ones are Memsource, Wordbee, XTM Cloud,
MateCat or SmartCAT. Some of them can be used for free (an option for Memsource,
SmartCAT), others require a monthly subscription. All of them offer advanced features
like machine translation modules, client service, or training, for example. The fact is that
even though their technical background is different, the core principles of CAT tools are
retained and they allow to benefit from computer-assisted translation anywhere.
Their power, and that of any professional CAT tool for that matter, is their modality
and ability to use various resources, like term bases and machine translation. In fact, with
core principles set, the current development of CAT tools is focused on refining these
‘aids’ so that the translation results they yield are more precise in the context of actual
document in translation. This stems from the fact that “most commercial MT systems are
technically hybrid systems” (Costa-Jussà, 2015), i.e. they utilise rule-based, statistical
35
SaaS means that the provider can use “applications running on a cloud infrastructure. The applications
are accessible from various client devices through either a thin client interface, such as a web browser
(e.g., web-based email), or a program interface. The consumer does not manage or control the
underlying cloud infrastructure including network, servers, operating systems, storage, or even
individual application capabilities, with the possible exception of limited user-specific application
configuration settings” (Mell and Grance, 2011, see their report on cloud computing for more
information).
Chapter 5. Computers in translation
103
methods and computer adaptive learning solutions. It is in opposition to current MT
research which is dominated by statistical methods and still provides new insights and
solutions to the practical application of MT solution. Yet, as CAT tools are ultimately
practical tools, and designed through the cooperation of linguists, engineers, information
experts and computer scientists, their architecture needs to be adapted to provide a
translator with precise, frequently single, piece of information, to be used in a very
particular context. In such a scenario general rules no longer apply, and new ones need to
be introduced. Therefore, it can be concluded that the current development of computerassisted translation is fuelled by the actual needs of translators and focuses on maximising
speed and translation quality through refinement of the existing solutions like MT, for
instance.
Figure 17. Memsource Web Editor utilising MT module
To sum up, computer-assisted translation has come a long way from the 1966 ALPAC
report and its criticism of machine translation. All things considered, the report made
room for the development of actual computer aids for translators. Redefinition of the
approach to MT ultimately led not only to the creation of CAT tools, but also to adding a
new quality to machine translation itself. No longer is it meant to replace human
translators; instead, it is to support their knowledge and skills in producing high-quality
translations. At the same time, corpus-based and statistical MT serves as the basis for
linguistic research worldwide. At the moment, it is hard to foresee what direction further
development of MT may take. It will surely continue developing as a CAT subsystem.
Apart from that, we may expect a successful implementation of Speech-to-Speech MT
(see, for example, Ellis, Creutz, Honkela and Kurimo, 2008), which is a logical next step
in easing human-human communication (as a follow-up to the success of such tools as
Google Translate). Nevertheless, it is likely that eventually all CAT tools will move to
the cloud entirely, becoming subscription-based tools. And that in itself would fit into
current market trends which see even such IT giants like Microsoft use the subscription
model with services like MS Office365 or the upcoming Windows 10 Enterprise (see
Mehdi, 2016). This stems from the fact that flexibility of both cloud tools and subscription
plans, not to mention included software updates, challenges old software distribution
system. Therefore, it is logical to assume that, with time, even the biggest players on the
market will have to move to the cloud in order to survive.
5.3. Terminology disambiguation
The use of specialised software in the process of translation requires to devote a few
words to define several concepts in order to provide a framework for the discussion on
CAT tools. The list below is not, by all means, conclusive. It merely explains basic
notions discussed in this book.
1. Computer-assisted Translation (CAT) – “a form of translation that makes use
of a software program supporting and facilitating the translation process” (TAUS,
2017). It is sometimes called machine-assisted, or machine-aided, translation (see
Chan, 2004: 139-140).
2. CAT tools – a group of computer programmes whose main goal is to improve
translator’s accuracy and productivity thanks to a number of internal and external
resources, e.g. editors, glossaries, translation memories (TM), machine
translation modules, etc., all of which are integrated into the same translation
environment. Examples of the group include Across Personal Edition, Atril Déjà
Vu X Professional, CafeTran Espresso, Fluency Translation Suite, GeoWorkZ –
Translation Workspace, Heartsome Translation Studio, LogiTerm Pro, MadCap
Lingo, MateCat, memoQ , Memsource Cloud, MetaTexis for Word, MultiTrans
Chapter 5. Computers in translation
105
Prism Freelance, OCLanguage, OmegaT, SDL Trados, Swordfish Translation
Editor, Transit NXT Professional+, Wordbee, Wordfast Anywhere, XTM
Cloud36, and more.
Most translators and translation agencies list the following as main
advantages:
a) Consistency – CAT tools help to make sure a translator uses the same
terminology, abbreviations, product names, etc. consistently throughout the
entire project or series of projects;
b) Cost reduction – the more the translator translates using a CAT tool, the larger
his translation memory grows. Over time, the level of repeated content will
also increase thus allowing the translator to translate faster while retaining the
overall quality of his/her translations. This in turn allows to translate more in
the same amount of time, resulting in higher income;
c) Quality Assurance – most CAT Tools have inbuilt quality assurance features
such as auto-correct, auto highlighting of grammar errors, inconsistent
numbers, missing tags, translations, etc. for the translator to fix;
d) TM Management – CAT tools allow to use more than one TM for a given
project, import and export them. What is more, they allow to align source and
target documents in order to create a TM out of previously translated material;
e) TB Management – CAT tools allow to create and use multiple termbases
which store all the terminology the translator comes across during his/her
work. The terminology is available for immediate re-use as it is looked up in
real-time in the termbase;
f) Machine translation capable – most modern CAT tools allow to use machine
translation as one of their resources;
g) Increased productivity – when using a CAT tool, the translator works through
new content, and the software uses existing TMs based on a pre-set match
threshold. This allows to translate faster in order to meet even the shortest
deadlines;
h) Concordances – a feature that allows the translator (or a customer) to perform
a search query on a TM in order to check for specific or similar translations in
context;
36
Based on ProZ (http://www.proz.com/software-comparison-tool/cat/cat_tools/2) Accessed on:
14/03/2017.
i) Project Management – all projects are stored in one place with the translator
being able to access and review them at any time. What is more, many CATs
offer their own tools for quoting with quotes bound to given projects. Apart
from that, CAT tools allow to create and distribute project ‘packages,’ or bits
of translation, to a number to translators for them to work on the project
simultaneously, and reduce overall completion time as a result;
j) Multi-format friendly – CAT tools allow for a broad range of file formats to
be imported and translated. For instance, memoQ supports the following file
formats: .doc, .docx, .dot, .dotx, .docm, .dotm, .rtf, .ppt, .pptx, .pot, .potx,
.pptm, .potm, .xls, .xlsx, .xlt, .xltx, .xlsm, .xltm, Google Docs, Sheets, Slides,
.htm, .html, .idml (.indd), .icml, .mif (version 8 and above only), .svg, .ttx (presegmented), .sdlxliff, .xml, .xhtm, .xhtml, Android .xml, .xliff (1.2),.xliff
(2.0), .xliff for WordPress, mqxliff (MemoQ xliff), tmx, .dita, .ditamap, .pdf,
.catkeys, .csv, (Magento).csv, .dbk, .desktop, (Mozilla).DTD, .epub,
(Joomla).ini, .json, .lang, .Plist, .po, .properties, (Java).properties, .rc, .resx,
.srt, .strings, .sbv, .sub, .ts, .txt, .wiki, .yaml, .tag, .xtg, .zip (memoQ, n/a.).
Apart from that, CAT tools offer filters to be used with various formats,
adding functionality (e.g., it is possible to download and use SRT filter in
memoQ, which filters out timestamps for a .srt subtitle file and translate the
actual content only);
k) Exclusive projects – many customers require translators to use a particular
CAT tool. Therefore, only users of those tools may take part in certain
projects. A side-effect is that competition is less fierce as not every translator
is a CAT user.
The disadvantages of using CAT tools include:
a) Segmentation – CAT tools force segmentation (see disambiguation of
“segment” below) which may, in turn, entail the obvious danger of translating
sentence for sentence;
b) Error propagation – CAT tools base on TMs and re-use of previously
translated material, either by the translator him-/herself or by third persons.
The feature, while being one of the core features of any CAT tool, carries the
danger of repeating errors stored in the TM;
c) Reduction of rates – while CAT tools allow to translate more thanks to the
benefits of TM, customers may take advantage of the fact that at least part of
Chapter 5. Computers in translation
107
the translation is automated and reduce rates for certain matches. As a result,
translators may end up translating more for less, in the end earning the same
as they would without a CAT tool;
d) Cost and training – CAT tools tend to be costly and complicated to use. When
buying a CAT tool, a translator has to consider its price, time needed to master
its functions and the cost of training;
e) Compatibility issues – CAT tools are not 100% compatible with each other.
As a result, translation agencies tend to hire translators who use the same CAT
tool(s) like them, eliminating other professionals (regardless of quality and
experience) on this very basis.
3. Translation memory (TM) – a linguistic database used to store translated texts
paired with their source equivalents for future reference. CAT tools refer to TMs
in order to check if parts of a currently translated text have been translated before.
If so, the CAT tool suggests to re-use these fragments. In fact, the translator may
choose for the programme to auto-propagate current translation with results found
in the TM. Such feature may have enormous importance for individuals who
translate large numbers of similar texts or work with documents which contain
significant amounts of repetitions. Here, the use of a CAT tool will not only speed
up the process but, thanks to its time-saving qualities; it will also allow to increase
overall translation output.
What is more, many clients (including corporations and translation
agencies) require translators to utilise their own TMs in order to assure
consistency of vocabulary and style between previous, current and future
translations. Thanks to the use of the same, albeit updated with each subsequently
translated document, translation memory, the client reduces the risk of
heterogeneous style of new translations, even when performed by a number of
translators.
TMs share several unique properties. They are:
•
directional – source and target languages are fixed;
•
bilingual – contains only two languages;
•
automatic – TM is updated and searched automatically.
A CAT tool monitors the TM content for matches (match definition below) in real
time, providing the translator with translation suggestions based on those matches.
It should be noted that it is possible (sometimes necessary) to use more
than one TM at a time. It allows to create specialised TMs based on a given field
of expertise or individual requirements of a client.
Bogucki (2009) lists three methods of creating a TM: create a new one
from scratch, import existing TM database into our own TM or use alignment
feature of CAT tools, i.e. combine source and target files (source text and
translation, translated without CAT tool), divide them into smaller chunks called
segments, and pair them between two texts. The translator then checks the
alignment for errors, confirms correct matches and exports aligned segments to
translation memory. This way, even professionals working in the field for a
number of years can enter the market of CAT-using translators and increase speed
and volume of their translations.
4. Segment (TU) – it is the basic translation unit (TU) used by CAT tools. Most
commonly a segment equals a sentence. However, the segmentation rules can
frequently be adjusted to meet individual preferences or particular expectations
on behalf of the client. In this respect, a segment can be defined as the shortest
fragment of text that can be automatically extracted by a CAT tool. Other
examples of segments include bullet points, table cells, in-tag strings, sentence
fragments separated by an empty line, etc. See the Figure 18 below for
segmentation settings in memoQ.
Figure 18. Segmentation rules in memoQ
Chapter 5. Computers in translation
109
The discussion on segments in CAT tools requires a few words to be devoted to
the problem of the TU itself. It is one of the most fundamental concepts in
Translation Studies with as many different definitions as there are translation
theorists. Vinay and Darbelnet (1958/2000: 88) suggest that “translators start with
words or units of translation, to which they apply particular procedures.”
Newmark (1988: 65), on the other hand, proposes that “a sentence is a natural unit
of translation” He supports this claim by reasoning that it is unusal to divide a
sentence, provided it is not excessively long. Short sentences are short for a reason
(e.g., to stress something or achieve a certain stylistic or psychological effect).
What is more, he argues that if “long sentences are a part of a writer’s style in an
expressive text, they have to be preserved” (idem.). The discrepancy between the
length of a TU runs parallel to the conflict between free and literal translation.
Literal translation TUs are short since it focuses predominantly on single words.
Free translation is concerned with portions of meaning rather than syntactic
structures that carry it (see Newmark, 1988; Hatim and Munday, 2004). While in
standard translation translators have a full freedom over what they will consider a
TU, the CAT-based environment forces segmentation (TUs) on them. Figure 18
above shows that it can be adjusted to some degree to meet individual preferences
of a translator, but this adjustment is not always clear or easy (see Figure
19 below).
Figure 19. Segmentation editor in Memsource
The segmentation in CAT tools is something their users have to learn to accept
and use. While being a technical necessity (segments constitute bilingual
translation pairs in translation memories) in CAT tools, it does carry dangers for
an inexperienced translator. Declercq (2015: 482) names the following advantages
and disadvantages of segmentation in CAT tools:
(A) Advantages
•
A sense of control on the segment level;
•
Similar pace;
•
Close reading, no interference of non-verbal elements;
•
Added value of term recognition;
•
No formatting issues;
•
Increased accuracy and consistency;
•
Being able to monitor progress;
•
Auto-propagation;
•
Possible copying across of the source segment.
(B) Disadvantages
•
The layout of the source text is lost;
•
No feeling of overall view and alienation from the context;
•
Lack of non-verbal elements affects quality and productivity (listed after
Biau Gil, 2007, in Declercq, 2015);
•
Lack of control;
•
Formatting sometimes still requires editing;
•
A tendency to more literal translation.
CAT tools are best suited to translate technical and specialised texts, like
medicine, for example. This truth stems from the fact that heavy segmentation,
i.e. text delivered sentence by sentence for translation, provides best results with
content-packed sentences, more suited for literal than free translation. Literature,
or poetry for that matter, frequently requires wider context than a manual, for
example. CAT tools make it hard for the translator to translate sense-for-sense,
disregarding the SL syntax. They promote sentence for sentence translation due
to the fact that such a method proves to be successful in specialised texts. If an
inexperienced translator uses CAT tools to translate a novel, for example, the final
outcome may retain most of the meaning of the original, but the style of the
translation may suffer, resulting in the text being awkward to the reader.
Chapter 5. Computers in translation
111
The auto-propagation feature carries the danger of re-using incorrect
translations, stored in the TM, in the current project. What is more, the translator
has to pay attention to formatting which is not 100% visible in the CAT editor but
can change during the process of translation. It is partly due to the fact that in the
course of their everyday work translators have to work with all kinds of different
types of documents with DOC/DOCX, XLS, XLSX and PDF being the most
common ones. The problem is that not all of them have been formatted properly,
and PDFs frequently suffer in-sentence page and line breaks, created
(unconsciously) by their authors during the saving process. The possible
formatting errors are numerous, with encoding issues and certain formatting
incompatibilities unique to the tool that was used to create a document. Figure 20
below presents broken segmentation when translating an instruction of a Dyson
Supersonic™ hair dryer37.
Figure 20. An example of broken segmentation when translating a PDF file with memoQ.
As can be seen, in the above example segments are not sentences but parts of
sentences divided by a hard return and marked with ‘¶’ sign in MS Word.
Translators have limited options to fix such issues, with one of them based on
computer competence discussed earlier in the course of this book. A rather safe,
but time-consuming, option is to adjust the source document manually to prepare
37
A random manual in PDF file, downloaded for the sake of presentation only. Available from:
https://www.dyson.co.uk/medialibrary/Files/Dyson/Support/downloads/UK/HairCare/HD01_Supers
onic_Manual.pdf Accessed on 2 April 2017.
it for translation in a CAT tool. While this resembles pre-processing done for the
benefit of machine translation, in the context of CAT tools it concerns mostly preformatting of the source document. The same can be done in CAT editor while
translating but, depending on the severity of broken segmentation, it may be
slower and more painstaking than editing the document before it is imported into
the CAT tool (in the case of PDF file such process involves OCR and full
formatting recreation).
Students often ask the author about the point of such laborious pre-editing.
Why not translate it as it is? The answer is two-fold. First, translation based on
bits of sentences would probably be of a very poor quality. Secondly, one should
not forget about the main asset of CAT tools, i.e. the translation memory built in
the process of translation. If based on low-quality material, it would not benefit
the translator in the future due to low match probability. Moreover, if a translator
is not concerned with the quality of his TMs then there is no reason to use CAT
tools in the first place for their strength lies in their ability to use our old
translations (stored in TMs) for current and future projects.
5. Match – the term denotes a hit in a translation memory. If a CAT tool finds a
matching sequence of words (in relation to the currently translated text) in the
TM, such sequence is called a match. Matches are always provided as text and
percentage numbers next to the suggestion (to denote the extent to which the
match is identical with the translated sequence). The percentage thresholds can be
adjusted in CAT settings. The classification of matches presented below has been
based on Bogucki (2009), SDL (2017) and Kilgray (2017):
•
Repetition – a segment was translated within a given project and is
repeated later in the document;
•
Context Match – 101% compatibility – a “context match is better than a
100% match. To be a context match, the translation memory segment must
be a 100% match for the document segment and the two segments must
have the same document context. For the document segment and the
translation memory segment to have the same context, they must both have
been preceded by the same segment” (SDL, 2017).
Context matching is possible due to the fact that translation memories
store a lot of information, including:
o the source document segment;
Chapter 5. Computers in translation
113
o the segment with translation;
o the segment preceding the segment in the source document. If such
information is unavailable, CAT tools use other information stored
in the TMs, e.g. information about the placement of the segment in
the TM.
•
Exact match – 100% compatibility – a segment in the translated text has
an identical counterpart in the TM;
•
Good match – 95% – 99% compatibility – “the source text of the segment
is exactly the same as the match, but there are slight differences: numbers,
tags, punctuation marks and spaces might be different;
•
Fuzzy match – 94% – 70% compatibility – a TM segment and the one in
translation are compatible in 70-94%. The 70% threshold (see Silva, 2014)
is constant in many CAT tools (can be adjusted manually), with some
exceptions, like memoQ38 for example;
•
No match – no match was found in the TM.
In addition to the above, some producers list additional categories, e.g. the Kilgray
company (2017) uses the notion of Double Context Match to denote context
matches with the additional condition of having the same ID in the TM and
translated document. Figure 21 below presents match distribution as presented
by memoQ.
Figure 21. Match distribution in memoQ Adriatic
38
memoQ differentiates between three classes of fuzzy matches in order to provide more detailed
statistics of the source document, and provide better results during the process of translation. These
include:
“High fuzzy (85%-95): In average-length or longer segments (8-10 words or more), normally there is
a difference of one word.
Medium fuzzy (75%-84): In average-length or longer segments (8-10 words or more), normally there
is a difference of two words.
Low fuzzy (50%-74%): In average-length or longer segments (8-10 words or more), the difference is
more than two words” (Kilgray, n/a).
6. Termbase – a “term base is a database containing pairs of words or expressions
(terms) in multiple languages. In translation, term bases are used like a glossary
where you can access terms (words and expressions) relevant to your translation
work” (Kilgray, n/a). These are smaller units than segments and, therefore, cannot
be extracted on the basis of segmentation rules. Similarly to TM, termbase (TB)
boasts several unique features. It is:
•
multidirectional – there is no source-target division;
•
multilingual – the database may contain any number of languages;
•
semi-automatic – a CAT tool does an automatic checkup of the TB and
provides search results to the translator who, on the other hand, has to add
each term to the TB manually
7. Translation mechanics and translation specifics – in the course of the book two
types of processes will be referred to. These processes describe the method in
which CAT tools are taught/learned and used. Translation mechanics refers to
understanding how a given type of software works and how to operate it and its
resources (i.e. creation, import and export of TMs, TBs, settings, etc.). It also
concerns solving of technical problems. Translation specifics, on the other hand,
focuses on predicting and handling translation problems in CAT tools. It mainly
concerns developing a set of best practices which, when observed, have a
substantial positive impact on overall quality of translation – a feat to be achieved
by limiting the risk of committing or copying errors.
8. Machine Translation (in CAT tools) – machine translation has been outlined
before in the course of this book. When discussed in the context of CAT tools, it
refers to a range of tools (plugins, apps) that can be configured to access external
MT systems. For instance, memoQ allows to configure and use MT resources
provided by Altlang MT, Crosslang Gateway MT, GoogleMT, IP Translator MT,
iTranslate4.eu MT, KantanMT, Let’s MT!, Microsoft MT, Moses MT, Pangea
MT, Salte Desktop, SYSTRAN MT, tauyou and Tilde MT (memoQ, 2017; see
Figure 22).
Chapter 5. Computers in translation
115
Figure 22. MT configuration panel in memoQ 8 Adriatic.
SDL Trados users may benefit from SDL’s own solutions like SDL Language
Cloud (AdaptiveMT) and SDL BeGlobal MT services. Apart form that, they can
use external applications downloaded from SDL AppStore, e.g. MT Enhanced
Plugin for Trados Studio (SDL AppStore, n/a.). This particular app allows Trados
to “to retrieve translations from either Microsoft Translator or Google Translate,
with added features” (idem.). Memsource uses Microsoft Translator by default
(free of charge), but allows also to configure Google Translate, Google Translate
Premium Edition, Custom Machine Translation Engines, Apertium, CrossLang,
Globalese, KantanMT, Microsoft Translator / Microsoft Translator Hub,
MoraviaMT, NICT, Omniscien Technologies, PangeaMT, PROMT, SDL
BeGlobal, SDL Language Cloud, SYSTRAN, Tauyou, Tauyou Real-time, Tilde
MT and more (Memsource, 2017). Most contemporary CAT tools allow to
configure and use MT services. In this case, MT matches are either autopropagated (if the software is set to do so) or displayed as suggestions alongside
other matches (see Figure 23 below).
Figure 23. MT results in Memsource (100 denotes a match from a TM; MT is a match provided by
machine translation – Microsoft Translator).
MT match quality does not depend on the actual work of the translator (like in the
case of TM). It all depends on the quality of MT engine used and, as can be seen
from the number of different MT engines available for CAT tools, translators have
the option to choose the one they feel satisfied with. What is more, in each case
the translator has to decide whether to accept or alter the suggested match or set
the CAT tool to accept those matches by default.
5.4. CAT-based translation process
The previous sections outlined how students gain their competence as translators and
presented the basic notions of computer-assisted translation. It is vital to describe the
process of CAT-based translation, even if in general terms, try and explain actions taken
by test subjects during the research project.
Most CAT tools follow the same principles when it comes to handling a translation.
The process starts with a document (or a set of documents) sent in by a client. The
translator creates a project, adds the document(s) for translation, translation memory and
a termbase. If the customer provided his/her own TM or TB, the translator has to use them
also for quoting purposes since it may contain segments that will not require translation
in this particular case. TBs are provided in order to assure consistency of vocabulary
between the current and previous projects. If the customer does not provide any resources,
the translator either creates a new TM (for the client) and uses his/her TMs as references,
or uses his existing TM as the main memory to be updated.
Chapter 5. Computers in translation
117
Once the quote is done and the client has accepted the deadline and price conditions,
the work may begin. When a document is imported39 into a CAT tool, its content is
divided into units called segments. Segments are usually single sentences, although users
may redefine segmentation rules within the programme in order to suit their needs. When
a translator translates a sentence, both the source and translation are stored in TM as a
segment pair. A simple example below presents source and target segments in memoQ.
Figure 24. Source and target segments in memoQ
A database entry for this sample translation looks like this:
<tu>
<seg>Ala ma kota.</seg>
<seg>Ala has got a cat.</seg>
</tu>
This is a simplified sample as each CAT tool adds various information to segments
(creation and modification dates, author, document name and type, creation method,
preceding segments, and so on). However, it shows that it is a piece of computer code
(XML), where <tu> tag stands for translation unit and <seg> tag for each segment.
Separately they are useless but combined like in the example they become a powerful
linguistic tool.
When a segment is translated, the translator has to confirm it and, by doing so, save
it to the TM. When it is done, the software automatically proceeds to the next segment.
Figure 25. Finding a match in translation memory
39
In the CAT tools nomenclature documents are not opened but rather ‘imported’ into the system.
Depending on whether it has found a match in the TM, the CAT tool enters (propagates)
the match into the editor and leaves it for the translator to review and confirm as proper
translation. In the case of fuzzy and good matches, more attention needs to be paid as
those require more or fewer changes done to the match structure. Exact (100%) and
context matches usually require proofreading. Repetitions are inserted and confirmed
without translator’s attention needed. In the above example (Figure 25), the programme
automatically searched its TM and found a repetition (the sentence translated a moment
before), which was then automatically propagated as the correct translation. Most CAT
tools offer advanced settings for match processing, including auto-propagation method.
A translator may set the programme to fill translation based on matches and confirm
translations for a given range of matches.
Figure 26. Pre-translate dialogue window with match-linked settings
in SDL Trados 2017.
Such setting carries the danger of committing errors in translation since, contrary to “blind
faith” in the quality of external TMs (Bowker, 2005; Doherty, 2016), such resources may
contain numerous errors40 that will be propagated into the translation and, possibly,
skipped altogether41.
40
41
See an introductory research presented later in the book which probes the issue.
“Frequently it is the case that the translator performs spellcheck after the translation is done. The TM
contains the first version of the translation, i.e. the one with spelling mistakes. While some CAT tools
change segment status to ‘Draft’ [in SDL Trados, author’s note] and update TM once they are
confirmed as translated again. Not all CAT tools do that in this way (e.g., Trados 7, still used by some
translators) and correct mistakes on the final document, without updating the TM. As a result, the
Chapter 5. Computers in translation
119
It should be mentioned at this point that unless segments are locked (i.e. set to readonly mode by the client, for example. See Figure 27 below) the translator has the ultimate
control over each segment and its final translation.
Figure 27. Locked segments in SDL Trados 2017. The translator cannot alter the translation.
It is especially important in the context of the present study. Even though a computer
programme supports the translator with various amounts of linguistic data, not all of it is
of high quality (see Doherty, 2016). Therefore, a successful translator has to consider a
number of things. First of all, translation of a segment has to be performed, or a match
has to be considered. The match may help the translator or may hamper him/her instead,
depending on its quality. All TM-generated errors retained in the final translation become
the translator’s errors, and will be considered as such by the client. In the case of doubt
the translator may assume that the TM carries desired forms of segments and, thus, should
remain unchanged (idem.). However, such an approach requires experience that translator
trainees simply may not have. What is more, the client may be unaware of any errors in
their TM. Therefore, contacting the customer may be the best option in such a case.
Once the translation is done, the translator proofreads the document, correcting any
errors found by him/herself or suggested by the programme. Depending on the CAT tool,
the proofread segments may or may not receive a new status (e.g. in memoQ these are
visualised by ‘V+’ [check mark plus] or ‘VV’ [double check mark]). Contemporary
wordprocessor-independent CAT tools (SDL Trados, memoQ) do not require the raw
translation to be cleaned (see Bogucki, 2009: 65). Each translated segment is confirmed
(updated in the TM) and, unless changed, will appear in its current form in the final
translator ends up with a correct end document and a TM containing spelling mistakes. When the same
TM is used to translate a new document, the translator sees a 100% match and considers it a proper
translation, even though the segment still contains errors. Therefore, all segments (even 100% and
context matches) should be read, but 100% matches frequently are not included in the final quote
(translator gets no money for them), and less attention is paid to them.” (Marcin Świnoga, freelancer.
2016. Personal communication, translation MK.)
document. If the translation was ordered by a translation agency, a translation package
may be the final product of the process (see Figures 28 and 29 for SDL-based translation
workflow). Packages are electronic bundles consisting of a source document, translation
memory, termbase, and corpora related to the subject of the document. All the translator
has to do is to open the package, translate the document, export the package and send it
back to the reviewer (external translator or a translation agency). If errors are found, the
package returns to the translator for revision. Only after it has been corrected does it return
to the ordering party. If the ordering party sent in a regular document for translation, the
process is simpler; the final product will be a document in the target language and, if
requested, TM created in the process, exported to a TMX (Translation Memory
eXchange) file. No additional work is required as the entire translation was stored
bilingually in the TM to be used with some future project.
Chapter 5. Computers in translation
Figure 28. Single-File Translation Workflow in SDL Trados Studio (http://producthelp.sdl.com)
121
Figure 29. Project Package Translation Workflow in SDL Trados Studio (http://producthelp.sdl.com)
5.5. Translator vs. computer in contemporary professional translation
It was mentioned before that technology has become part of our everyday lives, especially
so in the professional translation (see Christensen and Schjoldager, 2016). Translators use
the Internet to contact their clients, deliver translations, mine data, access glossaries,
Chapter 5. Computers in translation
123
machine translation; they use editing, graphics, OCR software and, last but not least, CAT
tools. They are more or less fluent in handling the technology, but it is something they
have to do in order to stay on the market. “[T]he need for technological competencies for
professional translators to remain on top, if not ahead, of change has never been more
evident than it is now” (Doherty, 2016: 962). However, little do we know how certain
areas of this computer ‘expertise,’ or rather ‘competence’, affect the translation process
itself (Bundgaard, Christensen and Schjoldager, 2016). In fact, considering the role CAT
tools have in technical and otherwise specialised translation, it is hard not to notice that
the translators seem to have lost their dominant role in the translation process to the
machine. Already in 2013, Anthony Pym suggested that in the future the role of a human
in the process of translation will (d)evolve to that of a post-editor. Risku (2014: 336) uses
the term of “computer-assisted network economy” to refer to the current market. This
‘future’ begins now. Statistically speaking, at some point in the future human translation
will produce enough number of technical texts that the data will be sufficient to ‘teach’
MT systems to translate certain types of texts automatically.
Actually, in some cases, it has already happened. For example, the Polish version
of the MS Office website is translated into Polish using Microsoft Translator. The online
content starts with a disclaimer reading:
Important: this article has been translated using machine translation, see the disclaimer 42.
English version of the article can be found here. (Microsoft, 2017; translation mine43)
If it already happens, it will surely happen again. Albeit on a much greater scale. The term
“computer-assisted network economy” (Risku, 2014: 336) denotes a global change in the
perception of the translation market which is no longer composed of individuals working
on projects alone, but rather of groups of co-working individuals. While such a change
resulted from the need to share translation content in order to be able to carry out much
larger projects than would be possible individually, the side-effect was that
communication and data exchange methods had to be developed. CAT tools filled this
niche, taking at least partial control over the process of translation away from human
42
43
“ATTENTION: Machine Translation Disclaimer: This article has been translated by a computer
system with no human intervention. Microsoft offers these machine translations to help non-English
speakers enjoy content regarding Microsoft products, services and technologies. Due to the fact that
the article was machine translated, it may contain vocabulary, syntax or grammar errors.” (Microsoft,
2017; translation mine).
All messages on the Polish version of Microsoft webpage are displayed in Polish.
translators. Certain projects may require translators to spend as much (or more) time
proofreading automated translation than translating. Christensen and Schjoldager (2016:
107) suggest that “while the CAT tool is generally expected to aid and support the
translation process, it may also offer resistance and restrain the process in several ways.”
The restraints are not that obvious at first glance, but in no way they are to be
ignored. CAT tools have to be bought and then learned (de Saint Robert, 2008; Garcia,
2008). What is more, they reduce context through segmentation (Vázquez and Vázquez,
2014), create the risk of error propagation (de Saint Robert, 2008; Bogucki, 2009).
Additionally, the translator has to understand how the TM works and be able to maintain
it properly, manage numerous TMs, handle technological aspects of using a sophisticated
computer programme, and deal with significantly more complex cost-calculation and
invoicing system (Vázquez and Vázquez, 2014). However, even those restraints can be
(at least partially) negated through the use of a third party application, like SDL
Wordcount that makes quoting and invoicing much easier.
The discussion on who is the actual translator – human or machine – touches upon
quite a serious problem, one potentially devastating for the freelance translation market
globally. While the idea that a computer might completely replace the human translator
may sound science-fiction (or a bad joke in the best of cases) to some, recent
achievements in computer technology show that the possibility is quite serious. It was
mentioned before that computers already take some workload (and part of freedom) off
translators’ shoulders, and a number of translation theorists (see Pym, 2013) propose the
idea that ultimately translators’ job will be to ‘fix’ minor stylistic or contextual
discrepancies in machine translation output. In this light, the meaning of the notion
‘computer-assisted’ takes on a completely different aspect. What does this ‘assistance’
mean? Is it any help provided to the translator during the process of translation, or does
it mean doing everything, from pre-processing, through translation itself, to postprocessing, with translator assessing (and correcting, if necessary) the final output?
Translation Studies differentiate between the following types of translation methods:
human translation, fully-automated machine translation (FAMT), human-aided machine
translation (HAMT), machine-aided human translation (MAHT) (Kastberg, 2012), and
computer-aided/assisted translation (CAT) (Qun and Xiaojun, 2015), although Chan
(2004) and Kastberg (2012) list CAT as an advanced form of MAHT. The question of
‘assistance’ can be discussed in relation to HAMT, MAHT and CAT only, as FAMT does
not require human interaction to produce translations. The process is entirely autonomous
Chapter 5. Computers in translation
125
and carried out by (primarily) hybrid MT systems that combine the best characteristics of
the ‘pure’ ones, like rule-based, example-based, and statistical MT (see Costa-Jussà,
2015; Qun and Xiaojun, 2015).
Kastberg (2012: 42) defines HAMT as “software developed for the machine to
translate what it can, in the way it can. The human role can be compared to that of a
consultant or an editor, i.e. that the translator corrects or modifies what, in the machine’s
translation suggestions, is unacceptable to him or her.” Human assistance, in this case,
takes on the form of adjusting the translated content. The editing can take place before
(A. pre-editing), during (B. editing) and after translation (C. post-editing). Examples of a
human aiding a machine may include adjusting word-order so that the computer program
can process it smoothly and flawlessly (A), reacting in real time to problems the program
faces (e.g. choosing the right synonym for a given context) (B), and (C) correcting final
output of the computer program (idem.). Translators-post-editors, suggested by Pym
(2013) would use HAMT systems in their everyday work.
According to the definition provided by Chan (2004: 139-140), the MAHT “refers
to a type of human translation with limited assistance from the machine. It does not
remove from the translator the burden of actually performing the translation. The machine
is a tool to be used or controlled at the discretion of the translator.” According to the
definition, terms like computer-aided translation, computer-aided human translation,
computer-assisted translation, computer-assisted human translation, machine-aided
translation, machine-assisted translation, and machine-assisted human translation (idem.)
denote the same thing. In fact, any computer resource that aids the translator in the process
of translation can be referred to as a MAHT tool. The classifying factor is whether the
translator has control over the computer programme and its actions or not. If not, it is an
MT system. If yes, it can be regarded as a MAHT resource (e.g. spelling and grammar
checkers, offline and online dictionaries, corpora, and more). CAT tools incorporate the
above and add termbases, translation memory capability, concordancers, and external MT
resources. Therefore, it is only natural to list them in the MAHT category.
The problem is that the line between MAHT/CAT and FAMT/HAMT is no longer
so clear, especially now. It is still true that human translators are responsible for the
translation process, but to a lesser extent than a decade ago. The progress in process
automation is especially visible in translation. TMs combined with MT allow to pretranslate a document on the basis of match threshold.
While remuneration for a 100% match in many translation agencies stands at 25%
of a regular rate, it is not always the case. More and more companies do not pay for those
at all, deciding that since no actual translation will be involved, no payment is necessary
(see Drexler 2016: 58). The policy is unfair, especially when we consider that it does not
mean that they do not require proofreading, however. Unless arranged with the customer,
the translator is accountable for the overall quality of the entire translation, regardless of
the quality of the translation memory used. Unfortunately, translators rarely have the
upper hand in company – freelancer relationship and their capacity for negotiation is low.
All they can do, on most occasions, is to reject the job (and the money they would earn).
While such solution is possible for experienced and respected translators, many more
simply cannot afford it. The application of combined TM and MT resources may further
fuel the problem of remuneration and translation quality. Doherty (2016: 962) notes that
(…) the lines between human and machine are continually blurred and professional
translators become more reliant and embedded into the translation process that they had
hitherto controlled. (…) With informed and effective use of TMs and MT, many of the known
issues and shortcomings of these technologies can be overcome, especially in terms of
translation quality, to somewhat mitigate the downward trend in pricing for translation
services in line with tighter budgets and deadlines.
It means that the effective use of technology could make up for less than optimal job
conditions. However, an alternative option should be considered. TM and MT resources
shared by the customer (e.g., translation agency) may drop the rates simply to reflect that
the translators have less work translating. The fact is that some customers send in pretranslated documents (sometimes with locked segments so that the translator can translate
only the unlocked ones44).While such policies may be considered to rise out of moneysaving policy, it does not always pay off in terms of quality. In theory, if given a
sufficiently large TM and TB, the translator could translate some specialised text with no
expertise in the subject. It could be done even though such a thing would require flawless
resources to work with (so that no errors are repeated and chances for success would
probably be slim anyway). It means that (technically speaking) contemporary CAT tools
are capable of replacing the human translator in the process of translation. The only
reasons for them not doing so are: a) there are no TMs large enough to allow processing
44
Reported by Marcin Świnoga, a freelancer specialising in CAT-based technical translation. 2016.
Personal communication.
Chapter 5. Computers in translation
127
of each text string possible; and b) we, humans, are not perfect. When we translate,
sometimes we make errors – errors that are saved in TMs or other resources, which in
turn are used to in MT and CAT systems. It all boils down to the fact that a CAT system
(consisting of a CAT tool, internal and external TB, TM, and active MT service) cannot
react (yet) to non-standard (or rather non-recorded45) linguistic problems while pretranslating and, basing on pre-recorded human work, repeats the same errors. If it could
react and adjust pre-recorded translations, there would be no MAHT or HAMT, and
FAMT systems would do all the work instead.
As it is, CAT tools do not translate by themselves and ‘merely’ assist the translator
in the process. In fact, their assistance is so efficient that they may reduce the capacity for
learning in some individuals. It seems to be a natural side-effect of using sophisticated
software which takes much work off the translator’s mind. The programme ‘remembers’
so that human does not have to. CAT tools have been designed to do so. This feature
allows translators to do their job faster and in greater volumes. Moreover, it raises a
question of whether this aspect of ‘assistance’ simplifies translation into the
(re)production of texts in another language. Human-performed translation, rendering a
source text into the target language, is a mental activity (see, for example, Sager, 1994;
Baker and Malmkjær, 1988). In this light, CAT-based translation seems more like a
mechanical process involving the (re)use of the entire blocks of the text translated by
other people, own translations, and machine translated output in order to produce new
translations. No longer does the translator require to learn new things constantly since the
software is there to ‘assist.’ It is a natural response to market demand and the fact that a
translation customer is primarily interested in three properties of a translation: quality,
speed and price. The process is not relevant as long as these three conditions are satisfied.
CAT tools allow translators to meet this demand. They need to translate a given text,
phrase, or find a term only once. Once stored in the TM or TB, it will wait there to be
used again. However, would the CAT user be able to translate with the same quality
without the CAT tool46?
45
46
In theory, if enough translated segments were recorded in a translation memory, a CAT could
successfully pre-translate any specialised document based on context matches (101%). Human
assistance would be required only for document preparation purposes, which could be probably
resolved as well considering the pace in which OCR tools are upgraded. The same could be said for
HAMT systems which by default require less human attention than MAHT systems.
A point of reference may be helpful to understand to what extent a TM-based CAT tool can ‘assist’
(or rather completely replace) a translator. “Since November 2007 the European Commission’s
Directorate-General for Translation has made its multilingual Translation Memory for the Acquis
While such theoretical scenario is not likely – who would stop using CAT tools
once they have learned their usefulness – it is an interesting thing to consider from the
psychological perspective. We tend to rely more and more on technology which facilitates
our daily life. We tend to get frustrated when something does not follow general market
standards and requires more of our attention. It is the same with translators. If there are
tools which will allow them to translate more, faster, and earn more at the same time –
the choice is obvious. If the use of the software is advocated by customers like translation
agencies, they will use it even more. In the end, it is a simple loss and profit account. If
the software allows the translator to earn more with less effort, it is worth the money
invested, even if his/her overall role is diminished by the ‘assisting’ computer programme
(e.g. use of previous translations done by other translators).
To sum up, CAT tools are extremely capable computer programmes designed to aid
the translator in the translation process. However, the may pose a challenge to some users
due to their nature and process of operation. In order to assure best results, they require
no small amount of training and conscious use. Through assistance they allow translators
to improve their translation output and overall quality significantly, at the same time
reducing the amount of control over the process. The amount of control removed from
the translator may be enough for an inexperienced, or less careful individual, to commit
grave errors by re-using faulty translations. The above statement is, at the same time, the
core assumption behind the study described in this book. The next chapter will introduce
two initial studies, to be followed by the main research. Primary assumptions, test groups,
research material, and conduct will be discussed. Furthermore, research results will be
presented, analysed and conclusions drawn.
Communautaire, DGT-TM, publicly accessible in order to foster the European Commission’s general
effort to support multilingualism, language diversity and the re-use of Commission information” (EU
SCIENCE HUB. Accessed on 16/04/2017). As of 7 March 2016, the number of segments containing
Polish language was 5,330,072, or 84,396,336 words (and counting). In case of PL-EN language pair
the number of segments far exceeds 2,000,000 (according to LSPSoftware.com), or almost a half of
the total number of Polish segments in the repository. The resource is freely available to all translators
working in the EU languages for use in their translation projects.
Final remarks
129
Chapter 6. Research
The previous chapters introduced the theoretical background for the principal part of the
book, i.e. the research. The following chapter recapitulates the main aims of the research
and set it in the context of the available free and commercial courses on CAT tools. An
additional section will be devoted to error typology. Next, it presents the structure of the
research study, including the platform, time frames, test groups, test documents,
resources, errors, stages and data collection methods. This is followed by an account of
the course of the research study. Finally, the analysis of the results is provided, closed
with an overall summary and conclusions.
6.1. Aims of the research study
The research study aims to analyse whether trainee translators (students of translation),
taught to use CAT tools in accordance with freely accessible online guides or selected
commercial courses, are effectively prepared to deal with translation projects involving
the use of external translation memories and challenges they may pose.
The main assumption of the book is that trainee and less experienced translators are
prone to over-rely on computer resources, relinquishing a portion of their independence
to a machine, which may lead to [serious] errors in translation (see Doherty, 2016).
Effective empowerment of students to deal with authentic issues with translation memory
quality, not present in resources prepared by software developers and commercial
courses47 (see section 6.2. below), will result in the more conscious use of computer
resources like CAT tools and facilitate entering CAT market by reducing the risk of
unwanted translation quality problems.
6.2. Courses available on the market
For a CAT tool to be successful, it needs to be functional, user-friendly and provided with
comprehensive and easy to follow manual. Each CAT software developer publishes
47
Commercial courses are understood are paid courses focused only on teaching CAT software outside
academic institutions.
online manuals, wikis, and training videos with the aim of allowing the software buyers
to learn and master the computer program. In fact, the materials are so detailed that, given
the user possess at least some computer expertise, they allow to learn all required features
at home and start working in no time. For example, SDL, the developer of Trados, has
prepared and published video tutorials on Trados which include presentation on how “to
use new features in SDL Trados Studio 2017, such as: AdaptiveMT, upLIFT Fragment
Recall, upLIFT Fuzzy Repair, File Type Preview and our Custom Display filters; how to
personalize your translation experience and create your own Studio ribbon tabs across all
views; how to use the project management and SDL package technology; and how to
upgrade your translation memories and use the enhanced search” (SDL, 2017, online
source). SDL’s resource page lists the following videos48:
•
Introducing SDL AdaptiveMT,
•
SDL Trados Studio 2017 Tutorial,
•
SDL Trados Studio 2017 5 Minute Video,
•
SDL Trados Studio 2017 in 2 minutes,
•
AdaptiveMT in SDL Trados Studio,
•
How to use Translation Quality Assessment,
•
Translation Quality Video Introduction,
•
and Introduction to the integration of SDL Language Cloud with SDL
AutoSuggest 2.049.
In addition, each SDL Trados buyer is entitled to take an online certification test which,
if passed, will grant them the title of a ‘certified user.’
Other developers, like Kilgray for instance, follow a similar path. They also publish
training videos on memoQ which cover such topics as:
•
Introduction to memoQ Adriatic,
•
Project Management with memoQ Adriatic,
•
Introduction to Translators memoQ Adriatic,
•
Productivity and Client Satisfaction,
•
Find your muses! – Intermediate course into memoQ productivity,
•
Make it comfy! Adjust memoQ to your needs, Hello memoQ! – an introduction
for individual translators,
48
49
Only SDL Trados related content was listed
http://www.sdl.com/resources/video-gallery.html
Final remarks
•
Start! – beginners guide to the memoQ galaxy,
•
Do you know your LiveDocs?50,
•
and more.
131
Even a cursory analysis of the content of the above-mentioned video materials shows that
they are indeed very informative, allowing to learn features of given CAT tools and to
develop technical skills required to use them. Another benefit of those resources is that
they are mostly free to access and use.
However, those users who do not feel at ease learning about computer software on
their own, or are interested in more advanced features, have the opportunity to buy one
of the commercial courses, available on the market. These usually do not come cheap, at
least for an average Polish translator (not to mention students). However, they allow for
mastering the software under the tutelage of a competent trainer.
The list presented below outlines the learning topics in memoQ course by one of
German translation companies:
•
Introduction to computer-aided translation
•
Technology overview
•
Setting up your work environment
•
Overview of the translation editor
•
Basic editing functions
•
Translating MS Office documents
•
Translating PDF documents
•
Reviewing the translation
•
Time and cost estimation
•
Translating XML in memoQ
•
Translation of InDesign documents
•
Quality assurance
•
Translation of software strings
The same list for memoQ, but as provided by one of Poland-based companies
(information in brackets denotes level of the course):
50
•
Introduction (beginner)
•
Configuration + mechnisms behind suggesting (beginner)
https://www.memoq.com/en/recorded-webinars
•
Offline projects - translation and terminology (beginner)
•
Creating translation memories out of existing documents (beginner)
•
memoQ server from the perspective of a translator (beginner)
•
Versioning (track changes, updating source document during translation)
(intermediate)
•
File import function (exemptions and change of settings) (intermediate)
•
Views and tips and tricks in memoQ server-based projects (intermediate)
•
Settings - configuration (seqmentation and QA) (intermediate)
•
Terminology import (basic and advanced glossary imports) (intermediate)
•
Advanced file import (advanced import filters, multilingual filters) (advanced)
•
Regular expressions for translators and their use in memoQ (advanced)
•
Own translation rules (advanced)
•
Automation: auto-translate rules and muses dictionaries (advanced)
•
Project templates (advanced)
Course contents for other CAT tools available on the market follow a similar pattern.
The reason for presenting all those topics to be covered either during free online
courses, or the commercial ones, is twofold. Firstly, it is meant to show that the presented
samples of both types of courses/trainings are comprehensive sets of resources that allow
to prepare a translator to use the software at a fairly advanced level. The commercial
courses seem to be the obvious choice for users who want to obtain more detailed knowhow in respect of a given tool. Secondly, they show that the main issue of this book, i.e.
the risk of reusing errors from translation memories in current and future translations, is
not reflected in the topics covered by the courses and training videos51. At the same time,
it has to be clearly noted that this statement refers to the courses analysed for the benefit
of this study. It is in no way meant to discredit any of them.
One reason for such an approach to the issue may be that CAT tools have been
designed to follow a certain workflow (see Figure 28 for SDL Trados workflow) which,
if observed, greatly reduces chances for receiving translation memories with errors.
51
The author of the book had the opportunity to use online resources published by SDL, Kilgray and
Memsource; take a commercial course on memoQ, and complete Memsource course for trainers. None
of those courses tackled the problem of dealing with errors in translation memories.
Final remarks
133
However, translation industry, communication issues and frequently human factor
make following the workflow impossible. As a result, errors are not only committed but
also preserved in translation memories. The issue is important also due to the fact that it
is virtually unavoidable in the professional career of a successful translator. The only case
when a translator would face minimal risk of finding errors in translation memories would
be to work exclusively with the self-created TMs, and commit no errors throughout their
career. However, that is hardly the case. Translators have to work with multiple TMs
(local, server-based), frequently composed of translations by many different translators.
Given the proper CAT workflow, these should be reviewed, refined, and made uniform.
It is not always the case (see Yamada, 2011). Sometimes TMs come from clients who
either do not know the proper workflow (as assumed by the software developers), do not
have the means (manpower and other resources) to introduce it, or it is not of primary
concern to them. In either case, resulting TMs may contain a number of errors that
endanger the translation quality. Some CAT translators have confirmed that problem in
the initial survey to the research, reported later in this chapter. Of course, such
experienced professionals will probably know such clients and use their TMs with proper
care and reservation. Nevertheless, the present study is concerned with students and
newcomers to the CAT world – people who have either no experience in translation in
general, or at least no experience in the use of CAT tools. Such people are more prone to
‘reuse’ TM-based errors, especially when they are taught (via free or commercial courses)
how suggestions work and how the software assists the translator. When an inexperienced
translator (a student especially) is given a suggestion which is a bad translation, but at the
same time s/he knows that this translation has been used before and was accepted, the
person may feel inclined to accept it. If, on the other hand, a new CAT user was exposed
to various cases of using TMs (based on authentic situations), they might feel more
independent and empowered to challenge suggestions made by the software. The present
study strives to verify those assumptions.
6.3. Determining justification for the research
Any research involving a discussion on students falling into the trap of errors in
translation memories calls for verification whether those errors are present there at all.
Therefore, the discussion on the main research of the book will be preceded by a report
on findings from a smaller study, conducted before the main one.
The study involved a number of professional translators (CAT users) and members
of ProZ52 translator community. It was based on a survey prepared using GoogleDocs and
published on the Translation Theory and Practice part of the ProZ forums. It was directed
exclusively to professionals working with CAT tools on a daily basis.
The goal of the study was to confirm that the main research assumptions have
justification in authentic situations that translators face in their profession.
A total of 26 translators answered the questions, 1853 of whom declared to be
professional CAT users and their answers were considered in this study.
The list of figures below presents questions and answers:
Figure 30: How long have you been an active translator?
52
www.proz.com
It has to be noted that the survey was directed to a very specific group of people. Hence low number of
responses.
53
Final remarks
135
Figure 31: Please specify the volume of your overall translation jobs carried out using CAT tools (Trados,
memoQ, Déjà Vu, etc.).
Figure 32: How frequently do you work with translation memories provided by your customers (incl.
translation agencies) (results from infrequent to very frequent)?
Figure 33: How frequently do you find errors in translation memories (results from infrequent to very
frequent)?
Figure 34: If you happen to find errors in translation memories, how many are there (on average) (results
from a little to a lot)?
18
16
Number of answers
14
12
10
8
6
4
2
0
grammar
terminology
spelling
misinterpretation
Figure 35: What types of errors (if any) do you find in translation memories provided by your customers?
Figure 36: Do you prefer specific texts for translation (e.g. legal) and you reject those jobs which do not
fall into this category?
Final remarks
137
Figure 37: If you happened not to get a new job for some time, would you consider accepting a translation
job outside your professional experience? (e.g. something you have not translated before).
Figure 38: Have you ever found yourself in a situation when you could not use translation memory
provided by your customer due to its bad quality (i.e. errors)?
The results of the study point to several facts. First of all, they show that errors in
translation memories provided by customers are a very common issue. On average, the
translators reported a moderate amount of errors (subjective opinion) per a translation
memory. Interestingly, despite the fact that most contemporary CAT tools have in-built
spellcheckers, the majority of the reported errors are spelling mistakes. Grammar and
terminology errors come second, with misinterpretation (or lost meaning) being the least
experienced problem (which was to be expected considering the fact that CAT tools are
primarily used to translate technical or otherwise specialised texts which leave little room
for free interpretation).
While almost 60% of respondents reported that they specialise (and, therefore, are
more likely to spot and correct translation errors, e.g. in terminology), almost 70% agreed
that they would consider accepting a job outside their professional experience if it was
justified financially. This implies that student-translators, who have no professional
background, would be even more likely to accept such commissions.
Finally, over 70% of the surveyed translators declared that at least once they were
not able to use a translation memory provided by their customer due to its bad quality,
(i.e. the number of errors) and, subsequently, required a number of corrections made it
time and cost inefficient to use the TM.
The study suggests that at least some portion of TM-based translation does not
undergo the revision process (or at least it is very inaccurate). Alternatively, as the
author’s personal experience shows, the revision may be conducted on DOCX files, not
in a CAT tool, with corrections never reaching the TM.
6.4. Methodology behind the main research
6.4.1. CAT environment
The main research was conducted over a period of two academic years using memoQ
CAT tool, developed by Kilgray. memoQ is a computer-assisted translation environment
(a full suite of tools) developed by Kilgray, a Hungary-based company. Thanks to its
quality, ease of use, and marketing (including Kilgray’s Academic Program), it quickly
gained recognition on the professional market of translation and became second most
popular CAT tool after SDL Trados (according to ProZ.com54).
The reason why memoQ was chosen for the project is that the Institute of English
Studies, University of Łódź, benefits from the Kilgray Academic Program, the goal of
which “is to give the Institution the right to use an unlimited number of memoQ licenses
for educational purposes, free of charge” (www.memoq.com, Internet resource).
Therefore, memoQ was the obvious choice since it could be used both in the classroom
and by students at home (thanks to individual educational licences, shared by the local
memoQ server administrator [the author of the book] within the CAT course).
54
http://www.proz.com/software-comparison-tool/cat/cat_tools/2
Final remarks
139
6.4.2. Timeframe
The research started at the beginning of April 2016 and continued on to the end of May
2017. It was divided into two Stages, each of which was again divided into two Phases.
Both Stages included 2nd year MA students (translation track) over two subsequent years.
The goal was to capture data from both groups following the same initial and different
final principles, and compare the results:
•
Stage I Phase I took place at the beginning of April 2016
•
Stage I Phase II took place at the end of May 2016
•
Stage II Phase I took place at the beginning of April 2017
•
Stage II Phase II took place at the end of May 2017
6.4.3. Test groups
As was mentioned above, test subjects included 2nd year MA students (MATIS – MA in
Translation Studies programme) in their final semester.
•
Stage I Phase I involved 22 students, 18 of which gave consent to analyse their
test data (two recordings could not be used, see section 6.5).
•
Stage I Phase II collected data from the 16 individuals who consented in Phase I.
•
Stage II Phase I involved 18 students, all of whom consented to have their data
analysed (one recording could not be used, see section 6.5).
•
Stage II Phase II involved 17 students, all of whom consented to have their data
analysed in Phase I.
In both cases, groups were mostly female (with only one male in the 2017 group). As a
whole, they were rather homogenous with average computer skills and no previous
experience in the use of CAT tools.
6.4.4. Test documents
Test material consisted of two documents in Polish containing fragments of an
independent expert auditor’s opinion with an accompanying financial statement. Both
documents were translated into English as it is a more popular direction of translation in
Poland. While traditional Translation Studies disapproved, or even rejected, translation
into a non-mother tongue (see, for example, Newmark 1988), market conditions and the
position of English as the new lingua franca have contributed to a growing acceptance of
directionality amongst translation theorists. Snell-Hornby (2000: 37) warns that “it is
particularly on the global level of supracultural communication that translation into
English as a non-mother tongue has become a fact of modern life.” According to Beeby
Lonsdale (1998), “directionality” refers to the fact that translations can be done both into
and from a mother tongue. In the case of the Polish translation market, translations from
Polish into English are quite common due to a low number of English translators
proficient enough in Polish to translate successfully, and those few who can frequently
demand rates which are not acceptable for translation buyers in Poland. While fitting into
the actual condition of the Polish market, translating into a non-native language was an
additional challenge to the test subjects and their use of CAT tools.
Each test group had been exposed to such material before during previous semester
concerned with non-CAT computer tools used in translation where they were asked to
recreate a series of similar (the same subject) documents using optical character
recognition (OCR) and word processing software, and to translate them.
Phase I document was around 1,200 words long, while Phase II document was
around 850 words long. The main bodies (opinion of an independent auditor) were almost
the same (520 vs 517 words), while the remaining balance sheet part was longer in the
Phase I document. This was a deliberate decision on behalf of the author of the book. The
Phase II document was a bit shorter also in order to give students more time to pay
attention to details, if necessary. Additionally, the number of errors in the TM was
reduced from 18 (Phase I) to 15 (Phase II). Both source documents (Phase I and Phase II)
can be found in Appendix 2 and 3, respectively.
6.4.5. Translation memories
The main goal of the research was to test how trainee translators (students) deal with
errors found in translation memories. Since it was impossible to obtain complete authentic
TMs without violating copyrights and proprietary rights of respective third parties, a
decision was made to provide a TM based on an actual translation of test documents into
English, performed by a Łódź-based sworn translator. Legal and financial terms in
translation were additionally consulted with another Łódź-based sworn translator.
The translation was then edited in order to introduce most common translation
errors, suggested by professional translators in the initial survey and verified for
authenticity with local freelancers in Łódź.
Final remarks
141
The idea was to use TMs containing errors that are not immediately visible and may
be skipped when not proofread carefully.
The next step was to align the two documents – the source and altered translation –
using memoQ LiveDocs module and the LiveAlign™ technology. It allows to combine
two monolingual documents into one bilingual translation memory. This way, the
translator retains control over links between the actual segments.
Figure 39. Aligning documents in memoQ
As a result, two separate TMs were obtained, one for each test document. Due to time
constraints during the actual research session and the need to provide students with ample
time to proofread the translation and look up unfamiliar words, both TMs were prepared
in order to meet the following requirements:
• each TM was to contain around 50% of 100%+ matches (Exact, Repetitions and
Context matches, according to memoQ terminology). It means that at least 50%
of the test document was already translated and was stored in the TM.
• 70-99% matches (fuzzy matches) made up around 20% of the test document.
• and around 30% of the document required translation from scratch.
The errors were placed in the 100%+ matches (those which had already been translated),
to be suggested and inserted by the CAT tool into the translation. The remaining segments
were either not translated at all or exhibited only partial translation, with no additional
errors added. This course of action resulted from the fact that students tend to pay more
attention to incomplete segments (no or fuzzy matches) since the 100%+ matches are
inserted automatically55 by the program and the remaining ones need translating (see
Kenny and Way, 2001, for example). However, the goal of the study was to check whether
students rely extensively on the quality of the TMs and only 100%+ matches could be
considered to be of high quality since they are full versions of translation performed and
accepted by somebody else. What is more, the text was not long enough to allow for
numerous errors in the TM.
6.4.6. Errors in translation memories
The above section focuses on errors in translation memories. The following paragraphs
explore the issue in a bit greater detail in order to outline what types of errors found its
place in those TMs and why.
The discussion should start with the definition of a translation error. Hansen (2010:
385) notes that the “perception and evaluation of an error as a translation ‘error’ depends
on the theoretical approach to translation and the evaluator’s ethical norms with respect
to translation.” In short, an error may take various forms. Both translation theorists and
associations of translators tried to ‘promote’ their error categories. For example,
Hejwowski (2004) proposed a quite elaborate and detailed classification of errors, which
included:
•
Errors of syntagmatic translation (dictionary equivalents, false friends,
commonly accepted equivalents, calques, unnecessary transfers);
•
Misinterpretation errors (mistaking two SL syntagms or verb frames,
misinterpreting scenes/scripts, misconstructing the text’s modality);
•
Realization errors (errors of the target language, wrong evaluation of the TL
readers’ knowledge, insufficient knowledge of the subject-matter);
•
Meta-translation errors (choice of translation technique, additions, omissions,
two versions, too many or too few footnotes, wrong translation strategy,
corrections, changing the text’s intertextuality).
TAUS (Translation Automation User Society) provides a simpler classification which can
be broken into (www.taus.net, 2017):
55
The feature mentioned in the main text is called auto-propagation. It is a CAT feature that allows the
program to assist translator by inserting 100%+ match into the translation grid. Translator’s job is then
to check the translation and confirm it. The feature can be switched ON and OFF in CAT settings. In
memoQ it is ON by default.
Final remarks
•
143
Language – although it can refer to ambiguous sentences, an error in this category
generally means a grammatical, syntactic or punctuation error;
•
Terminology – a glossary or other standard terminology source has not been
adhered to;
•
Accuracy – incorrect meaning has been transferred or there has been an
unacceptable omission or addition in the translated text;
•
Style – quite subjective, it refers to a contravention of the style guide56.
This classification seems clearer and less complicated than the one by Hejwowski. Yet,
Pym (1992) had already proposed even more concise classification into binary and nonbinary errors. In this division, “there is only right and wrong [answer for binarism]; for
non-binarism, there are at least two right answers and then the wrong ones” (idem.: 282).
Moreover, it needs to be observed that there are degrees of importance when it comes to
errors in translation. Consider a spelling mistake which, according to Pym, is binary. Yet,
its impact on the overall value and quality of translation is limited, unless it affects text
meaning. A mistranslation on the other hand – also a binary error – is much more serious.
Apart from these three examples, it is worth to take a closer look at the classification
by Martinez Melis and Hurtado Albir (2001: 281) who distinguished four categories:
•
errors relating to the source text (ST) and errors relating to the target text (TT)
•
functional errors and absolute errors
•
systematic errors (recurrent) and random errors (isolated)
•
errors in the product and errors in the process
This classification is quite important in the course of this book mainly due to the latter
two categories, which can define the nature (systematic errors) and reasons (errors in the
process) for the errors in translation memories. However, further discussion on the issue
will be continued in the last section of this chapter.
In fact, it all boils down to the fact that an ‘error’ is a kind of deviation from an
established expectation, which have to be formulated before any classification can be
made. Chesterman (1993) notes that expectations can be practical (e.g. conformance to
the requirements of a client, type of text, or behavioural norms defined by the community
of professional translators.
Therefore, the errors found in the TMs used in the research fall into four categories,
as suggested by the ProZ professionals in the initial research. Those categories include:
56
Quoted from https://www.taus.net/knowledgebase/index.php/Error_typology
• grammar.
• terminology,
• misinterpretation,
• and spelling.
This classification aims to reflect authentic types of errors committed in such a special
work environment as CAT tools. The tables below present full list of errors that could be
found in the source TMs used in the research.
Table 9. Errors found in the research TMs: Phase I
Category: Spelling
Number of errors in the text: 2
TM:
… secondary legoslations issued on the basis of it and other binding legal provisions.
Should be:
… secondary legislations issued on the basis of it and other binding legal provisions.
Polish:
… wydanymi na jej podstawie przepisami wykonawczymi oraz innymi
obowiązującymi przepisami prawa.
A common spelling mistake, resulting probably from the fact that “o” and “i” are
neighbours on the keyboard
Desc.:
TM:
Should be:
Polish:
Desc.:
The financial statement has been prepared for the period from 1 January 2013 to 31
December 2012, and the data is compared to the period from 1 January 2012 to 31
December 2021 .
The financial statement has been prepared for the period from 1 January 2013 to 31
December 2013, and the data is compared to the period from 1 January 2012 to 31
December 2012 .
Sprawozdanie finansowe zostało przygotowane za okres od 1 stycznia 2013 r. do 31
grudnia 2013 r., natomiast dane porównawcze obejmują okres od 1 stycznia 2012 r.
do 31 grudnia 2012 r.
Probably an error resulting from fast typing. The translator wanted to type 2013, but
missed the key, resulting in an error.
Category: Grammar
Number of errors in the text: 3
TM:
Auditor[‘s] responsibility
Should be:
Auditor’s responsibility
Polish:
Odpowiedzialność Biegłego Rewidenta
Desc:
No Saxon genitive to denote whose responsibility it is. The errors is frequent when
fast typing. No proofreading.
TM:
An audit consists of performing procedures aiming to obtain evidence on the amounts
and information revealed in the financial statement.
Final remarks
Should be:
Polish:
Desc.:
TM:
Should be:
Polish:
Desc.:
145
An audit consists in performing procedures aiming to obtain evidence on the amounts
and information revealed in the financial statement.
Badanie polega na przeprowadzeniu procedur mających na celu uzyskanie dowodów
badania dotyczących kwot i informacji ujawnionych w sprawozdaniu finansowym.
Wrong use of a word-defining preposition. “Consists of” means that something is
made or formed out of something else, whereas “consist in” means that something is
an essential (or main) part of something else.
In making such risk assessment, we have taken into account internal control
connected with the preparation and fair presentation of the financial statement in order
to design audit procedures that are appropriate in the present circumstances, not in
order to express our opinion on the efficiency of internal control in the unit.
In making such risk assessment, we take into account internal control connected with
the preparation and fair presentation of the financial statement in order to design audit
procedures that are appropriate in the present circumstances, not in order to express
our opinion on the efficiency of internal control in the unit.
Przeprowadzając ocenę tego ryzyka bierzemy pod uwagę kontrolę wewnętrzną
związaną ze sporządzeniem oraz rzetelną prezentacją sprawozdania finansowego w
celu zaplanowania stosownych do okoliczności procedur badania, nie zaś w celu
wyrażenia opinii na temat skuteczności działania kontroli wewnętrznej w jednostce.
This is a clear example of an outline of a certain procedure. The use of Present Perfect
in the TM suggested an account of having performed some action, whereas the Polish
version clearly states that this is how it is done; therefore Present Simple “we take”
should have been used.
Category: Mistranslation
Number of errors in the text: 2
TM:
Should be:
Polish:
Desc.:
Our task is to audit the financial statement and issue our opinion as to its reliability
and the correctness of the accounting records constituting the basis for drafting
therein.
Our task is to audit the financial statement and issue our opinion as to its reliability
and the correctness of the accounting records constituting the basis for drafting
thereof .
Naszym zadaniem jest, w oparciu o przeprowadzone badanie, wyrażenie opinii o tym
sprawozdaniu finansowym oraz prawidłowości ksiąg rachunkowych stanowiących
podstawę jego sporządzenia.“Therein” and “thereof” are two common words found in legal writing. “Therein”
means “in that,” whereas “thereof” means “of that.” The purpose of reversing them
was to check whether students, who were exposed to similar texts before, will be able
to detect this fairly obvious error and correct it. Actually, one does not have to know
to word exactly to see (if reading carefully) that the “-in” ending does not fit the
context.
TM:
12th Economical Division of the National Court Register.
Should be:
12th Economic Division of the National Court Register.
Polish:
XII Wydział Gospodarczy Krajowego Rejestru Sądowego.
Desc.:
This is an obvious example of mistranslation. Polish “gospodarczy” (and
“ekonomiczny”) translates as “economic” into English. “Economical” means
“ekonomiczny” as well, but in a sense of not wasting something. The two English
words are often confused.
Category: Terminology
Number of errors in the text: 7
TM:
The Company has been created for an unlimited period of time.
Should be:
The Company has been established for an unlimited period of time.
Polish:
Spółka została utworzona na czas nieokreślony.
Desc.:
This case is a clear poor choice of words on behalf of a translator. A company is not
created, it is established
TM:
Advances for tangible fixed assets [under construction]
Should be:
Advances for tangible fixed assets under construction
Polish:
Zaliczki na środki trwałe w budowie
Desc.:
The use of omission of “w budowie – under construction.” The error may stem either
from a regular human fault or it may be the result of using earlier suggestion from TM
(during some earlier project) and accepting it without revision.
TM:
money and other pecuniary assets
Should be:
cash and other pecuniary assets
Polish:
środki pieniężne i inne aktywa pieniężne
Desc.:
Here the Polish „pieniężne” should refer to cash, and not to general term “money” –
“pieniądze”
TM:
Revaluation capital
Should be:
Revaluation reserve
Polish:
Kapitał z aktualizacji wyceny
Desc.:
The error results from a literal translation from Polish “kapitał,” which translates
directly as “capital”
TM:
Running expenses / revenues / activities
Should be:
Operating expenses / revenues/ activities
Polish:
Koszty działalności operacyjnej
Desc.:
In this case the word „operacyjne” was translated as “running,” probably on the
account that s/he understood the term as costs of running the company, i.e. all
expenses involved. Afterwards, the word “running” was consequently used for the
remaining five times when it was used in the document.
TM:
Amortisation [and depreciation]
Should be:
Amortisation and depreciation
Polish:
Amortyzacja
Desc.:
“Amortisation” is a literal translation of Polish “Amortyzacja.” However, English
equivalent of this item on the balance list is “Amortisation and depreciation.”
Note!
TM – actual error in the translation memory
Should be – proper translation
Final remarks
147
Polish – Polish version
Desc. – description of the error
[ ] – missing content (omission)
Table 10. Errors found in the research TMs: Phase II
Category: Spelling
Number of errors in the text: 2
TM:
Should be:
Polish:
Desc.:
TM:
Should be:
Polish:
Desc.:
The Management Board of the parent entity is responsible for drawing up a report on
activity and for internal auditing, which is deemed essential to ensure that the
consolidated financial statements are free from mistatement resulting from intentional
fraud or error.
The Management Board of the parent entity is responsible for drawing up a report on
activity and for internal auditing, which is deemed essential to ensure that the
consolidated financial statements are free from misstatement resulting from intentional
fraud or error.
Zarząd jednostki dominującej jest odpowiedzialny również za kontrolę wewnętrzną,
którą uznaje za niezbędną, aby sporządzane skonsolidowane sprawozdania finansowe
były wolne od nieprawidłowości powstałych wskutek celowych działań lub błędów.
A common spelling mistake. The error repeats consistently three times in the
document.
Comparative data was presented based on the Corporate Group’s financial statements
for the financial year ending on 31 December 2012, examined by some other
authorized entity which on 06 March 2012 issued an assessment of those consolidated
financial statements in which it stated no remarks.
Comparative data was presented based on the Corporate Group’s financial statements
for the financial year ending on 31 December 2011, examined by some other
authorized entity which on 06 March 2012 issued an assessment of those consolidated
financial statements in which it stated no remarks.
Dane porównawcze zostały przedstawione na podstawie skonsolidowanego
sprawozdania finansowego Grupy Kapitałowej za rok obrotowy kończący się 31
grudnia 2011 r., zbadanego przez inny podmiot uprawniony do badania, który w dniu
6 marca 2012 r. wydał opinię bez zastrzeżeń o tym skonsolidowanym sprawozdaniu
finansowym.
Another typing error. The translator probably missed number keys when typing.
Category: Grammar
Number of errors in the text: 2
TM:
Should be:
The examination consists of conducting procedures aimed at obtaining
concerning the amounts and information disclosed in the [consolidated]
statements.
The examination consists in conducting procedures aimed at obtaining
concerning the amounts and information disclosed in the consolidated
statements.
evidence
financial
evidence
financial
Polish:
Desc.:
TM:
Should be:
Polish:
Desc.:
Badanie polega na przeprowadzeniu procedur mających na celu uzyskanie dowodów
badania dotyczących kwot i informacji ujawnionych w skonsolidowanym
sprawozdaniu finansowym.
Again, an example of wrong use of a word-defining preposition. “Consists of” means
that something is made or formed out of something else, whereas “consist in” means
that something is an essential (or main) part of something else.
When assessing that risk, we have taken into consideration internal auditing related
to drawing up and reliable presentation of the consolidated financial statements in
order to plan the examination procedures that will be appropriate to the circumstances,
rather than to issue our assessment of the efficiency of internal auditing in the entity.
When assessing that risk, we take into consideration internal auditing related to
drawing up and reliable presentation of the consolidated financial statements in order
to plan the examination procedures that will be appropriate to the circumstances, rather
than to issue our assessment of the efficiency of internal auditing in the entity.
Przeprowadzając ocenę tego ryzyka bierzemy pod uwagę kontrolę wewnętrzną
związaną ze sporządzeniem oraz rzetelną prezentacją skonsolidowanego sprawozdania
finansowego w celu zaplanowania stosownych do okoliczności procedur badania, nie
zaś w celu wyrażenia opinii na temat skuteczności działania kontroli wewnętrznej w
jednostce.
Again, this is a clear example of an outline of a certain procedure. The use of Present
Perfect in the TM suggested an account of having performed some action, whereas the
Polish version clearly states that this is how it is done; therefore Present Simple “we
take” should have been used.
Category: Mistranslation
Number of errors in the text: 3
TM:
OPINION OF [AN INDEPENDENT] EXPERT AUDITOR
Should be:
OPINION OF AN INDEPENDENT EXPERT AUDITOR
Polish:
OPINIA NIEZALEŻNEGO BIEGŁEGO REWIDENTA
Desc.:
The use of omission of “niezależnego - independent.” The error may stem either from
a regular human fault or it may be the result of using earlier suggestion from TM
(during some earlier project) and accepting it without revision.
TM:
Responsibility of [the Management Board and] the Supervisory Board
Should be:
Responsibility of the Management Board and the Supervisory Board
Polish:
Odpowiedzialność Zarządu oraz Rady Nadzorczej
Desc.:
Another use of omission. See above for explanation.
TM:
Under the Accounting Act of 29 September 1994 (Dz.U., issue 152, item 1223) (“The
Accounting Act”), the parent entity Management Board and members of the
Supervisory Board are obliged to ensure that the consolidated financial statements
meet the requirements provided for thereof.
Under the Accounting Act of 29 September 1994 (Dz.U., issue 152, item 1223) (“The
Accounting Act”), the parent entity’s Management Board and members of the
Supervisory Board are obliged to ensure that the consolidated financial statements
meet the requirements provided for therein.
Should be:
Final remarks
Polish:
Desc.:
TM:
Should be:
Polish:
Desc.:
149
Zgodnie z ustawą z dnia 29 września 1994 r. o rachunkowości (Dz. U. z 2009 r. nr 152,
poz. 1223 z późniejszymi zmianami) („ustawa o rachunkowości”), Zarząd jednostki
dominującej oraz członkowie Rady Nadzorczej są zobowiązani do zapewnienia, aby
skonsolidowane sprawozdanie finansowe oraz sprawozdanie z działalności spełniały
wymagania przewidziane w tej ustawie.
As was explained before, “therein” and “thereof” are two common words found in
legal writing. “Therein” means “in that,” whereas “thereof” means “of that.” The
purpose of reversing them was to check whether students, who were exposed to similar
texts before, will be able to detect this fairly obvious error and correct it. Especially so
considering the fact that a very similar error could be found during the Phase I
assignment.
The examination consists of conducting procedures aimed at obtaining evidence
concerning the amounts and information disclosed in the [consolidated] financial
statements.
The examination consists in conducting procedures aimed at obtaining evidence
concerning the amounts and information disclosed in the consolidated financial
statements.
Badanie polega na przeprowadzeniu procedur mających na celu uzyskanie dowodów
badania dotyczących kwot i informacji ujawnionych w skonsolidowanym
sprawozdaniu finansowym.
Another use of omission. See above for explanation.
Category: Terminology
Number of errors in the text: 4
TM:
Value of the company
Should be:
Goodwill
Polish:
Wartość firmy
Desc.:
In this case the translator translated “wartość firmy” literally, which is not correct.
TM:
Money and money equivalent
Should be:
Cash and cash equivalent
Polish:
Środki pieniężne i ich ekwiwalenty
Desc.:
Just like in Phase I, the Polish „pieniężne” should refer to cash, and not to general term
“money” – “pieniądze”
TM:
Revaluation capital
Should be:
Revaluation reserve
Polish:
Kapitał z aktualizacji wyceny
Desc.:
Again, the error results from a literal translation from Polish “kapitał,” which translates
directly as “capital”
TM:
Money flow hedge reserve
Should be:
Cash flow hedge reserve
Polish:
Kapitał z wyceny transakcji zabezpieczających
Desc.:
This error pertains to the use of word “money” where it should be “cash.” It is an
example of translation without checking the nature of the term. What is more, there
are other examples in the text referred to as “cash,” so the error is to test perception of
the student.
TM:
Assessment of hedging instruments
Should be:
Valuation of hedging instruments
Polish:
Wycena instrumentów zabezpieczających
Desc.:
An obvious example of low translator’s competence. The word choice here should be
“valuation” since “assessment” translates as “ocena.” This error was suggested as a
frequently recurring one by one of the Łódź-based translators.
However, in the course of the research the most important factor will not be what types
of errors there were, but whether students were able to detect and correct them despite the
programme suggesting these were correct translations, approved by some other translator
before. Therefore, two scenarios will be considered:
a) a student skips an error which implies either process error (lack of proofreading),
over-reliance on the expected ‘quality’ of the TM, or lack of linguistic
competence, and
b) a student detects and corrects an error, proving that s/he is aware of the analysed
problem and, therefore, negate the dangers of using external TMs through own
translation competence.
6.4.7. Stages, phases, and preparation of students
The actual study was divided into two Stages over two consecutive academic years. In
each case, it was conducted in the summer semester due to the course programme of the
2nd year MATIS - MA in Translation and Interpreting Studies. Each Stage was further
divided into two Phases. Phase I was meant to collect data from both test groups shortly
after they learned basics on how to use CAT tools, but before they had the chance to gain
experience or learn quirks of the trade. Phase I, therefore, took place in the first part of
the semester, whereas Phase II took place towards the end of each respective semester.
The goal of Stage I Phase II was to determine if the actual exposure to CAT tools
and translation projects is enough to raise awareness of the students and assure acceptable
translation quality, as compared to Phase I. What is more, it was to verify the assumption
that the lack of proper training in less obvious aspects (or rather not preparing users for
certain contingencies), as is visible in the official training materials, does pose a
Final remarks
151
significant problem in the context of quality of translation and language-service
provision.
On the other hand, the goal of Stage II Phase II was to prove that proper exposure
to some less obvious aspects of CAT-based translation process yield more satisfying
results in terms of translation, and overall service, quality.
6.4.7.1. Preparation of students
Preparation of students was critical in the course of the research. Grabowski (2009: 67)
notes that
developing ICT [information and communication technologies] skills and practical
knowledge of dedicated translation software is best implemented among students of
technological profiles, whereas students of philology have a humanistic profile. As a result,
teaching advanced skills in operating dedicated translation software, such as translation
memories (TMs), terminology management systems (TMSs), online databases, localization
software, content management systems (CMS), web editing software, DTP, FTP applications
or spreadsheets, to name just a few, usually involves high degree of resistance among
students, whose contact with technologies […] is limited and frequently even none. Given
that professional translators need both humanistic (translator’s skills) and technological
profile (translation skills), students graduating from translation specializations are deficient
in ICT skills and their chances to prosper on the job market decrease considerably.
The observation by Grabowski outlines a very serious problem faced by students of the
humanistic departments. The issue was present also in both test groups. In order to negate
the problem, both groups took two extensive courses: audio-visual translation (AVT) and
computer application in translation. The courses took place over the two semesters
preceding the research. The AVT course focused on subtitling, transcription, and
developing teamwork and project management skills. The computer technology course
concerned primarily the use of word-processors and text formatting, optical character
recognition, desktop publishing, regular expressions, file archiving, and more. As a result,
the discrepancies in computer literacy between students were addressed and negated.
What is more, the preparation provided students with the same starting skills regarding
the use of computers in regular translation (non-CAT).
The preparation for Stage I Phase I and Stage II Phase I during the research semester
was identical in both cases. Content-wise, the programme followed the “Get Started”
official videos which, in theory, allow new users to start using the software. Therefore,
students learned what CAT tools and translation memories are, which was followed by
detailed instruction on the use of memoQ (version 2015), including program settings,
project creation, translation memories, translation editor, termbases, and basics of
quoting. The table below presents examples of exercises covered during this stage.
Table 12. Examples of exercises during the preparatory stage
Expertise
Examples of exercises
program settings
a) change language of the UI
b) download and install Hunspell spellchecking dictionaries
c) manage resource console
d) connect to online servers and use online termbases
e) configure plugins
a) create a project in memoQ
project creation
translation memory
translation editor
termbases
statistics
a) create a TM in order to translate a document
b) translate a document and send your customer both the translation
and the TM (export to TMX)
c) import a TM from several TMX files
d) create a TM using LiveAlign (LiveDocs) and use it to translate a
document
a) navigate the top ribbon to point, name, and describe the
functionality of selected icons (e.g. views, pre-translate, statistics,
QA, and more)
b) practical exercises during which students have to run QA check,
refer to QA errors and fix them
c) review the translation
d) use Views to display only selected segments
e) edit and use auto-translation rules
a) create a termbase, add terms from within the translation editor
b) edit Polish terms for inflection (correct use of the “|” sign)
c) import and export termbases
a) use statistics to quote a job
Following this introductory stage, the Phase I data was collected during a special in-class
assignment. The assignment was designed to take up to two hours (max time limit) on the
account that the relative volume of translation was around 40% of the entire document,
or 1800 characters with spaces per hour. While that much time available may be
questioned, it was determined by taking the following into account:
•
students use computers in a computer lab, not their own devices,
•
the computers in the lab (room 0.15, Faculty of Philology, University of Łódź) are
protected by the DeepFreeze maintenance computer program, which ‘reverts’ the
Final remarks
153
machine to its initial (default) state on Power On, resulting in removal of all
programs and settings changed by students; therefore, each time students have to
set up their memoQ according to their needs and wishes,
•
their progress is recorded, which may be a stress factor,
•
possible technical errors,
•
there is enough time to check unknown terminology on the internet,
•
there is enough time to proofread the document.
The task was to take a translation memory provided by the client, use it to translate the
test document in memoQ, and export both final document and TM.
Based on the code of conduct from the previous classes, repeated before the
assignment started, in case of doubt one should always consult their client (in this case,
the teacher)57. Doubts in this context should be considered not as problems with handling
the software, but issues with the clarity of the task itself. If needed, explanations were to
be asked for in whisper at the teacher’s desk in order not to hint other test subjects and
not to disturb their work.
Examples of unclear situations:
#1
Student:
Do you want to receive your translation in DOCX or PDF?
Client:
DOCX.
#2
Student:
There are errors in the TM. Do you want to have them corrected or left unchanged?
Client:
Please correct all of the found issues.
57
Interestingly, there were no questions during Stage I
The actual differences in the preparation of students started with Phases II. In the case of
Stage I, students continued learning about memoQ. The topics covered included handling
various file types (DOCX – formatting principles, physical extracting of media; PDF –
formatting issues and the need to know OCR tools; XML/HTML documents and working
with tags); advanced file import (using import filters); advanced quoting; creating
termbases on the basis of external terminology banks with the use of spreadsheets (MS
Excel, LibreOffice Calc), converting it to a memoQ-friendly format and importing them
into the program; and, finally, introduction to regular expressions to be used both in
termbase preparation and in the creation of own translation rules. On top of that, students
had to translate various documents both in class and at home. Topic included finance
(financial statements), warehouse product lists, and various short manuals (calculator,
hair drier, washing machine – excerpt), all with variants translated in two or three times
(e.g. battery replacement manual translated in class, its variant at home with the use of
the TM created in class, and finally third variant again in class, but this time using the
TM created over the course of the previous two translations). This way students learned
how matches work.
Towards the end of the semester, students were asked to carry out Stage I Phase II
assignment. All the rules and conditions were the same except for the text document,
which was a variant of the first one (both source documents can be found in the
Appendix).
Preparation for the Stage II Phase II included the core principles of preparation for
Stage I Phase II. However, this time the programme of the course focused not on the
mechanics of the CAT-based translation (understanding how the program works and
utilising its resources, i.e. creation, import and export of TMs, termbases, settings, in
everyday work) but rather on CAT-based translation specifics, i.e. the process of
translation with attention paid to less frequent, and yet authentic, issues and methods of
dealing with them. Above all, it is concerned with the development of a code of conduct
which, when observed, will significantly improve overall translation quality by reducing
the risk of committing or skipping an error.
Therefore, the course was extended to cover the following areas:
•
students were required to proofread matches carefully since the majority of TMs
used in class contained errors (similar to the test documents),
Final remarks
•
155
students were required to present the errors found and explain how they were
corrected upon completion of a given translation job (QA and review process
analysis),
•
students used peer-created TMs, which allowed to include random errors into the
course. All the errors found were discussed.
Example: students were asked to translate a fragment of a washing machine
manual. Afterwards, they exported their TMs and shared them with other students,
who used them for the following project.
•
discussions on the origins of errors in translation memories (e.g. proofreading on
DOC/X without updating the TM).
•
students were asked to use external resources, e.g. procedure lists58 and videos.
Students prepared in this way were asked to carry out the Phase II assignment (conditions
and requirements the same as described above).
Data from both Stages was captured with the use of FlashBack Express screen
recording utility, which is discussed in the next section.
6.4.8. Data collection methods
The method used to capture data during the research was heavily influenced by the work
environment, i.e. computer laboratory. A number of options were considered, i.e.
keystroke logging, eye-tracking, think-aloud protocols, screen recording, and
retrospective interviews (see Ericsson and Simon, 1984; Krings, 1986; Tirkkonen-Condit,
1989; Lörscher, 1991; Jääskeläinen, 2000, 2002; Asadi & Séguinot, 2005; Degenhardt,
2006; Jakobsen and Jensen, 2008; Dragsted, 2010; Grucza, 2012). In the end only one
method was found suitable – recording of a computer screen with the use of specialised
software. Before we can discuss that, however, it seems valid to explain why other
methods were rejected.
From the very inception of the research problem, keystroke logging seemed a
natural choice to capture data while translating. There are numerous programs on the
market, including Translog 2006, “which registers all keystrokes, mouse clicks, and the
time intervals between them” (Göpferich, 2011). It was successfully used in the
TransComp project. However, there were objections against installing keystroke logging
software in a laboratory used by students and teachers for various purposes. Another,
58
Sample procedure list can be found in Appendix 1.
much more problematic issue was that in the case of CAT tools the only way for
determining when somebody is proofreading or deciding whether something is correct or
not would be pauses before confirming a segment by Ctrl+Enter keystroke sequence
(which, again, is not recorded by all programs because both Control and Enter are
function keys). While this could actually make sense – if there was a pause and then a
string of keystrokes correcting an error, we would know it was corrected. If there was no
pause, the there was no proofreading. How about instances when there was a pause but
no correction? Can we assume that somebody performed proofreading and decided
against correcting a given segment? There is no way of knowing that. What is more,
keystroke data would present logs and timestamps between strokes. Such data would be
difficult to follow and refer to the actual content of the document, and there would be no
context, leaving much to the subjective interpretation of the analyst. Therefore, this
method was rejected.
Since keylogging was no longer an option, analysis of TM structure was selected
instead.
It was mentioned previously that TM is a type of database where all translation
segments are stored. The software saves much more information than the actual
translation. It is useful for translators working in teams who have to share their TMs or
simply for backup purposes. Below, there is a full translation unit (TU) sequence as
retrieved from a TMX file (TM eXchange file format).
<tu changedate="20160210T083903Z" creationdate="20160210T083903Z" creationid="Michał"
changeid="Michał">
<prop type="client">Michał Kornacki</prop>
<prop type="project">PhD research</prop>
<prop type="domain">Finances</prop>
<prop type="subject">Statement</prop>
<prop type="corrected">no</prop>
<prop type="aligned">yes</prop>
<prop type="x-document">Bilans_PL_ready.docx-Bilans_tłum_ENG_ready.docx</prop>
<tuv xml:lang="pl">
<prop type="x-context-pre">&lt;seg&gt;Podstawowy przedmiot i czas działalności
Spółki:&lt;/seg&gt;</prop>
<prop type="x-context-post">&lt;seg&gt;handel hurtowy środkami ochrony
roślin,&lt;/seg&gt;</prop>
<seg>Przedmiotem działalności Spółki jest między innymi:</seg>
</tuv>
<tuv xml:lang="en">
<seg>The Company’s principal activity is, in particular:</seg>
</tuv>
</tu>
Final remarks
157
Table 13 explains what all of this data represents:
Table 13. Explanation of TM field names
XML code
changedate
creationdate
creationid
changeid
client, project, domain, subject,
corrected
aligned
x-document
x-context-pre
x-context-post
<seg>Przedmiotem działalności Spółki jest
między innymi:</seg>
<seg>The Company’s principal activity is, in
particular:</seg>
Information
Document modification date
Document creation date
Creating user
Modifying user
Project related data
Shows whether the segment was altered (e.g. by
proofreading)
Segment source - either regular translation or
source-target text alignment
name of source document(s)
Preceeding segment
Following segment
Source segment (comes 1st)
Target segment (comes 2nd)
Most of this information is irrelevant for the research but changedate, creationdate,
corrected and <seg> do promise some findings after closer inspection.
Creationdate shows when a given segment was created, whereas changedate shows
when the segment was modified. If we assume that the research material contained 100%
match errors that would have to be found and corrected by the translator, these two values
alone show if proofreading was done or not. If not, both values will be the same. If
proofreading was involved, the changedate value would be later than the creationdate.
Another important information is the corrected value (“yes/no”) which shows
whether the segment was modified (i.e. proofread and corrected – yes) or not (no).
Finally, the strings following <seg> tags constitute actual source and target texts
and allow to identify segments for analysis.
It is assumed that this method would have yielded results once the amount of
information was reduced to the values mentioned above. In order to achieve that, a
powerful text-editing software would be needed (e.g. open source notepad ++). Using
Find&Replace feature, as well as regular expression59 syntax used to define search
patterns, it is possible to reduce individual TU to:
59 For more detailed information on regular expressions see Mitkov 2003.
changedate=20160210T083903Z
creationdate=20160210T083903Z
corrected>no
<seg>Przedmiotem działalności Spółki jest między innymi:
<seg>The Company’s principal activity is, in particular:
In such a simplified version, it is much easier to find those segments which contain errors;
what is more, the method was tested with positive results. However, it presents only the
final result of translation with no context of how it was achieved. It is assumed that a
glimpse at the actual process of translation would be beneficial. It shows how the
translator behaves when dealing with ‘problematic’ segments, whether they arrive at the
correct version, and if yes – whether they do it immediately or only later during
proofreading stage (if there is proofreading at all). Ultimately, the method was rejected
as well.
Think-aloud protocols (TAP) were first described by Lewis while he conducted
research in IBM (see Lewis, 1982). The method has become popular in psychology,
language studies (e.g. reading and writing), translation research, and is used to record
decision making and process tracing data. It involves the test subject to describe what
he/she is doing, looking at, or thinking while performing a given activity. However, in
this case, students would have no inkling as to what sort of information was required from
them (giving any more or less clear instruction could put the entire research at risk since
students could understand what actually was tested). Apart from that, there were logistical
issues – students would have to perform the assignment individually in order not to hint
or disturb other students. It was impossible due to the limited availability of the computer
lab (regular classes) and time constraints on the part of students.
Another method considered for the research involved retrospective interviews. This
idea was abandoned after a test run in which students who were asked to translate a much
shorter document than the test one (although with similar start resources) and answer
several questions afterwards. The test showed two things: 1) students remembered little
regarding individual segments and it was next to impossible to determine what had
actually happened during the translation process; 2) even though students were asked not
to communicate between each other lest they suggest what is the object of the study and
prepare for it, they actually did. As a result, the data obtained could not be used for
analysis. And, of course, there was the problem of time constraints. Similarly to TAPs,
retrospective interviews would require individual session for each of almost 40 test
subjects – an impossible feat due to both the author’s and student’s daily obligations.
Final remarks
159
The last two options to consider were eye-tracking and screen recording. The eyetracking technique allows to obtain a unique insight into the areas of special interest
during the translation process. It presents translator’s pace, shows the number of
revisions, pauses and their duration, and identifies possible problems during the
translation. The technique is often supplemented by screen recording (by default, since
eye-tracking systems record the screen and overlay it with eye-tracking data – a feature
that can be filtered off) and keystroke logging for triangulating purposes (see
Ehrensberger-Dow & Massey, 2013). The obstacles to using an eye-tracking system were
twofold: a) eye-tracking systems (both the device and software) are very expensive and
require substantial training in their use; b) the use of eye-trackers requires the data to be
processed with a statistical software like Statistica or R (open-source) while the research
did not justify using statistical software to prove the thesis 60; and the logistics of the
experiment – the method would require capturing data person by person, which –
considering the results of the test run mentioned previously – did not seem an appropriate
solution.
As a result, the screen recording technique was selected. As a standalone computer
program, the screen recorder does not capture eye movement like eye-tracker. It captures
only the computer screen in the form of a video file which can be then watched in the
recording software and analysed. A huge advantage of this method over the previous ones
is the fact that it allows to capture data from all test subjects in a session at the same time.
Most researchers report the use of Camtasia Studio for their research purposes (see,
for example, Degenhardt, 2006; Göpferich, 2009; or Sofyan and Tarigan, 2017). While
the software offers very good quality and multiple useful features, it is rather expensive.
Therefore, an alternative was selected – FlashBack Express 5 by BlueBerry Studios. The
BlueBerry studios agreed to provide the Insitute of English Studies with an academic
licence for the Standard Version of the software (Pro version allows for real-time video
compression and multiple export options). Even though it allows to record sound (could
be useful for TAPs) and webcam image (possibly useful to determine what a test subject
is doing at a given moment) (see Figure 40 below), both of these features were disabled
in order to preserve anonymity of the test participants and TAPs would simply not work
60
Many thanks to dr Carlos da Silva Cardoso Teixeira of the School of Applied Language and
Intercultural Studies, Dublin City University, for his help concerning the potential use of eye-trackers
in the research.
in a group environment (disturbance, voice commentaries might affect translation process
of other test subjects).
Figure 40. Welcome screen of the FlashBack Express 5 Recorder
The software was installed on all the computers in the computer lab by the IT staff of the
Faculty of Philology before the research. Operation of the program is quite straight
forward. A user has to select whether a whole screen or a window will be recorded and
choose whether sound and webcam are to be recorded as well. Once done, the user clicks
the “Record” button and the recording starts. Once the assignment is complete, the user
brings up the Flashback Express recording control window, clicks the “Stop” button
(Figure below) and saves the recording in the FBR (FlashBack Recording) format.
Figure 41. Flashback Express recording control window
In addition to screen recording, the software can capture key strokes (available through
Options->Keystroke setting [see Figure below]),
Final remarks
161
Figure 42. Keystroke capture feature in FlashBack Express
and mouse cursor, which is extremely handy in the analysis of video section when a
student does not translate. An eye-tracker would be perfect in such a situation since it
would inform the analyst precisely what a person was looking at and for how long. With
no access to this technology one can at least follow mouse cursor which shows the lines
along which the person was looking at the screen (eyes follow mouse cursor – we do not
move it blindly, but on purpose). The figure below presents a fragment of an actual
recording with the student checking online dictionary. The mouse cursor is highlighted
with a circle; what is not seen in the image below is how the mouse cursor moves along
the student’s eye movement lines while s/he is reading and analysing the search result.
Figure 43. Mouse cursor highlighted in FlashBack Express recording
FlashBack Express proved to be a good choice to capture video data from students. The
standard version used in the study has its drawbacks, though. The captured video can be
watched and analysed only with the use of the program itself as the only available export
option is an uncompressed AVI format, which results in a single 1.5-hour video size
exceeding 30 gigabytes after export. The FBR file size is around 700 megabytes. More
recent versions of FlashBack Express allow to set up compression for the FBR file,
resulting in files (without sound) below 100 megabytes. In the present study, the older
version of FlashBack Express was used, and the captured data analysis process was
slowed down due to large sizes of recorded videos and significant processing power
required to play and navigate the videos.
The next section discusses the progress of each subsequent Phase.
6.5. Course of the research
The research started at the beginning of April 2016. Each Phase of the research started
with an instruction on the use of FlashBack Express. The instruction was published on
the course website and was based on text instructions and screenshots taken from the
program. As was mentioned before, the utility is easy to handle. However, it was
paramount to make sure students remember all steps required to start recording, stop it
and save the video capture file. Additional instruction had to be given in order to make
sure the data capture process went smoothly, e.g. instruction not to use Ctrl+Shift+S
keyboard shortcut while translating in memoQ. By default, the shortcut copies a source
segment to the target one in memoQ, as stops recording in FlashBack Express at the same
time. Since it could have posed some problems, it was decided to refrain from using it.
The next step was to familiarise students with the task and resources. The resources
(source document and translation memory) were to be downloaded from the passwordprotected section of the course website (so that no one can have early access to the files)
and saved on the computer in the computer lab. Once it was done and everybody
confirmed having (and being able to open) both documents, students were asked to start
recording in FlashBack Express and then memoQ. It was mentioned before that by default
memoQ does not use any spelling dictionary, for example. No settings were altered in
memoQ workstations in relation to the default clean install of the program. The idea was
to simulate an authentic scenario when somebody buys the program and starts using it for
Final remarks
163
translation. Therefore, students had to set up spellcheck dictionaries on their own, as well
as terminology and translation memory plugins. The teacher made no hints during the
process and acted only the role of a translation service client. If somebody did not prepare
his/her work environment accordingly (despite the fact that it was covered during the
classes preceding the assignment), no comment was made.
Starting FlashBack Express first and memoQ second marked the start of the
assignment. While working on the project, students could use computer and Internet
resources freely in imitation of authentic conditions they would translate in. If there were
questions regarding the nature of the assignment or some client expectations, students
could contact the client (teacher) by approaching his station and asking the question in
whisper (in order not to disturb others and avoid any unwanted suggestions for others). If
the question was technical in nature (pertaining to the process of translation or memoQ
handling), no answer was given.
In perfect conditions, a translator would be expected to start memoQ, create a new
project, import the document for translation, create a TM and import the contents of the
external TM and create a termbase (optional for the assignment). Then the document
would be translated and most, if not all, TM-based errors detected and corrected either
during the translation or the proofreading/revision stage. Such an outcome on behalf of
the test subjects would disprove the thesis of this book. If, on the other hand, the TMbased errors were skipped and saved again to the current TM (the TM created for the
project was, in fact, a new TM, which would retain errors from the previous one), the
thesis would be confirmed.
Each time a student decided his/her work was finished, s/he would export the
translated document into a DOCX file and present it to the teacher. Such an event marked
the end of the assignment for the person.
6.5.1. Stage I Phase I
Stage I Phase I took place at the beginning of April 2016. A total of 22 2 nd year MA
(MATIS) students took part in it. However, only 18 consented to have their progress data
captured and analysed anonymously. Out of those, two could not be used because the
recording stopped (most probably it was accidentally switched off by a student) early in
the process. As a result, 16 video recordings were suitable for analysis.
The document contained the opinion of an independent auditor. The translation was
from Polish into English. Students had a total of two hours to complete the task (see
section 6.4.7.1. Preparation of students). The students were given instructions and
reported all was clear. The data capture process started with no problems. As a precaution,
students were asked not to communicate with each other. There were some technical
issues with memoQ on a few computers which needed resolving. Kilgray granted the
Institute of English Studies a server license which allows to set up a memoQ server and
use it to create and manage student licences on the basis of Client Access Licence (CAL).
Each computer on computer lab 0.15 has its own licence. However, in case the program
stops responding, it has to be terminated through Windows Task Manager. The server
frees up the licence, but the program cannot connect to the server and activate; as a result,
it cannot be used61. In order to fix the problem, the WiFi/Ethernet card had to be disabled
and licence requested in the offline mode. The program could not connect, obviously, but
the no licence status was reset. Upon re-enabling the WiFi/Ethernet card and the Internet
connection, the program could contact the server normally, receive a licence and activate,
enabling the student to continue work. The two-hour time limit was designed with such
problems in mind.
Time scores of individual students are provided in section 6.6.1. below.
In addition, students had no questions to the client.
6.5.2. Stage I Phase II
Stage I Phase II took place at the end of May 2016. The same group of students as in
Phase I was tested. However, data was captured only from the 16 students whose data
from the Phase I could be used in the research.
Again, the document contained the opinion of an independent auditor, albeit it was
a different document than in the Phase I. The maximum time to complete the task was
again 2 hours. The students were reminded not to communicate with each other. When
all of them reported that they are ready, the assignment began.
Similarly to Phase I, there were some licence problems with memoQ which were
resolved with no greater issues.
Time scores of individual students are provided in section 6.6.1. below.
No questions to the client were asked this time, either.
61
It has to be noted that most probably the fault lies in the local area network, not the program itself.
Final remarks
165
6.5.3. Stage II Phase I
Stage II Phase I took place at the beginning of April 2017. It involved 18 students of the
2nd year MA (MATIS), again their final semester in translation. The data was captured
from all 18 participants. However, it turned out that one FBR file was corrupted and it
was impossible to use it for the research. Therefore, 17 recordings were collected. All
students consented to have their progress anonymised and analysed.
The document was the same one as in the case of Stage I Phase I. Preparation for
this Phase was identical as in Stage I (see Section 6.4.7.1. Preparation of students). The
students received instructions, which had to be repeated in order to make sure students
knew how to handle the recording system. The data capture process started with no
problems. There were some minor issues with students trying to communicate, but those
were quickly resolved. No problems with memoQ took place this time. However, one
workstation stopped responding after about 20 minutes into the assignment and had to be
rebooted. Sadly, due to the DeepFreeze auto-revert utility, all progress was lost (see
section 6.4.7.1. Preparation of students). The student agreed to start again with the time
limit reset to two hours. The person managed to complete the task and record the process
successfully.
There was one case when a student accidentally stopped the recording. However,
the instruction not to use Ctrl+Shift+S was clear enough that she was aware of the
problem and reported the issue to the teacher. The first part of the video was saved, and
then the recording was started again. As a result, the record of the translation process by
this particular students was split into two files.
Time scores of individual students are provided in section 6.6.1. below.
Again, students had no questions to the client.
6.5.4. Stage II Phase II
Stage II Phase II took place at the end of May 2017. The test was conducted on the same
group of students as in Phase I. However, data was captured only from those students
who managed to submit a research-viable material in the previous phase. One additional
recording was stopped accidentally by a student only after approximately 20 minutes and
was discarded. Therefore, 16 video recordings were saved.
The document was the same one as in the case of Stage I Phase II. Preparation for
this Phase took into account translation mechanics, as described in section 6.4.7.1.). Since
it was the second time the students were asked to record their work, no additional
instruction was necessary. The progress of the assignment was rather smooth with only
two students suffering from the licence issue. In both cases the recording was paused, the
problem fixed, and recording resumed. No other problems occurred.
Time scores of individual students are provided in section 6.6.1. below.
Interestingly, this time there were four questions to the client. All questions
concerned errors in the TM and the code of conduct the client expected from the
translator. All of the students were instructed to correct all found errors. The following
section discusses this novelty, as well as other results.
6.6. Data and analysis
Both Stages of the study allowed to capture over 70 hours of data. Due to technical and
human errors, the total number of recordings were 32, or 16 per each Stage.
The following section presents the results and attempts to justify and analyse the
results. The presentation starts with time and then moves on to the actual issue of error
detection and correction.
6.6.1. Time
As was reported before, students had 2 hours to complete the task. All sessions confirmed
that it was enough as no one exceeded the limit. In truth, some students, especially in
Stage II Phase II seemed to be finished sometime before they actually stopped recording,
but kept on checking and double-checking their translations.
The times reported here are noted on the basis of the duration of entire recording
by a student, including preparation, setting up the software, translating, data mining, etc.
The tables below present recording times in Stage I and Stage II.
Table 14. Recording times in Stage I
Phase I
Student
Phase II
00:44:22
S1-01
00:46:08
01:13:26
S1-02
01:12:58
00:59:27
S1-03
01:00:40
00:56:56
S1-04
01:10:03
01:00:48
S1-05
01:16:32
00:54:15
S1-06
01:01:04
Final remarks
167
01:17:35
S1-07
01:07:12
00:59:28
S1-08
01:00:45
00:42:12
S1-09
00:49:02
01:11:05
S1-10
01:13:42
01:09:33
S1-11
01:12:29
00:52:58
S1-12
00:57:48
01:09:28
S1-13
01:05:13
00:48:10
S1-14
00:50:09
00:44:45
S1-15
00:44:22
00:45:06
S1-16
01:09:46
Table 15. Recording times in Stage II
Phase I
Student
Phase II
00:49:26
S2-01
00:57:50
01:28:58
S2-02
01:32:04
01:08:10
S2-03
01:21:40
01:47:51
S2-04
01:30:04
01:17:22
S2-05
01:24:18
01:32:02
S2-06
01:23:04
01:49:05
S2-07
01:35:32
01:05:20
S2-08
01:20:15
01:35:43
S2-09
01:42:08
01:33:40
S2-10
01:42:25
01:22:40
S2-11
01:19:17
01:36:21
S2-12
01:24:27
01:12:54
S2-13
01:35:11
01:38:08
S2-14
01:22:32
00:56:34
S2-15
01:19:34
01:40:50
S2-16
01:44:07
As can be seen, the times in Stage I are shorter than in Stage II. There is no clear
justification for this fact since both groups attended exactly the same courses throughout
their studies and technology-based courses covered the same material. Group mentality
(the need to excel, to be the first, a kind of competition present in the Stage I group, and
not in the Stage II group) may be a factor here.
On average, the Stage I group needed 58 minutes to complete the assignment in
Phase I and 62 minutes in Phase II. The students needed around 7% more time even
though they had gained some experience between the Stages and the document was
noticeably shorter.
In the Stage II Phase I, the students needed on average 84 minutes to complete the
assignment. In Phase II it was 87 minutes – an increase of around 2%. While results are
similar, the Stage II result is slightly better than the Stage I and students needed less time
compare to Phase I.
6.6.2. Errors
The main object of the study was to register how students deal with problems that could
be found in the TM provided by the customer. The data is reviewed individually stage by
stage and each error is marked as either omitted or corrected. The exact information on
whether an error was skipped (No) or corrected (Yes), is provided in Appendix 4,
numbered according to the lists provided below.
The ordered list of Phase I errors includes:
•
Error #1 - legoslations
•
Error #2 - Auditor[‘s] responsibility
•
Error #3 - the basis for drafting therein
•
Error #4 - An audit consists of performing
•
Error #5 - we have taken into account
•
Error #6 - 12th Economical Division
•
Error #7 - has been created for an unlimited period
•
Error #8 - prepared for the period from 1 January 2013 to 31 December 2012
•
Error #9 - Advances for tangible fixed assets [under construction]
•
Error #10 - money and other pecuniary assets
•
Error #11 - Revaluation capital
•
Error #12 - Running expenses
•
Error #13 - Amortisation [and depreciation]
•
Error #14 - running revenues
•
Error #15 - running revenues
•
Error #16 - running expenses
•
Error #17 - running expenses
•
Error #18 - running activities
The ordered list for the Phase II includes:
Final remarks
•
Error #1 - Opinion of an [independent] expert auditor
•
Error #2 – Responsibility of [the Management Board] the Supervisory Board
•
Error #3 – mistatement
•
Error #4 – provided for thereof
•
Error #5 – mistatement
•
Error #6 – consists of conducting
•
Error #7 – in the [consolidated] financial statements
•
Error #8 – mistatement
•
Error #9 – we have taken
•
Error #10 – 31 December 2012
•
Error #11 – Value of the company
•
Error #12 – Money and money equivalent
•
Error #13 – Revaluation capital
•
Error #14 – Money flow hedge reserve
•
Error #15 – Assessment of hedging instruments
169
The errors are reported in the order of appearance in text/translation process.
Individual results have been collected and an average has been drawn in order to
compare the data and outline a summary.
On the occasion that a student detected an error and corrected it, or rephrased the
segment in a correct way, the segment change was considered as corrected.
6.6.3.1. Stage I Phase I results
The following diagram presents the results of individual students as recorded in Stage I
Phase I. Full results divided into individual error cases can be found in Appendix 4.
Stage I Phase I
120
100
80
60
40
20
0
S1-01 S1-02 S1-03 S1-04 S1-05 S1-06 S1-07 S1-08 S1-09 S1-10 S1-11 S1-12 S1-13 S1-14 S1-15 S1-16
Omitted
Corrected
Errors retained in the TM
Figure 44. Individual results, Stage I Phase I
6.6.3.2. Stage I Phase II results
The following diagram presents the results of individual students as recorded in Stage I
Phase II. Full results divided into individual error cases can be found in Appendix 4.
Stage I Phase II
120
100
80
60
40
20
0
S1-01 S1-02 S1-03 S1-04 S1-05 S1-06 S1-07 S1-08 S1-09 S1-10 S1-11 S1-12 S1-13 S1-14 S1-15 S1-16
Omitted
Corrected
Errors retained in the TM
Figure 45. Individual results, Stage I Phase II
Final remarks
171
6.6.3.3. Stage II Phase I results
The following diagram presents the results of individual students as recorded in Stage II
Phase I. Full results divided into individual error cases can be found in Appendix 4.
Stage II Phase I
120
100
80
60
40
20
0
S2-01 S2-02 S2-03 S2-04 S2-05 S2-06 S2-07 S2-08 S2-09 S2-10 S2-11 S2-12 S2-13 S2-14 S2-15 S2-16
Omitted
Corrected
Errors retained in the TM
Figure 46. Individual results, Stage II Phase I
6.6.3.4. Stage II Phase II results
The following diagram presents the results of individual students as recorded in Stage II
Phase II. Full results divided into individual error cases can be found in Appendix 4.
Stage II Phase II
100
90
80
70
60
50
40
30
20
10
0
S2-01 S2-02 S2-03 S2-04 S2-05 S2-06 S2-07 S2-08 S2-09 S2-10 S2-11 S2-12 S2-13 S2-14 S2-15 S2-16
Omitted
Corrected
Errors retained in the TM
Figure 47. Individual results, Stage II Phase II
6.6.4. Analysis
The research yielded most interesting results. While Phase I scores were virtually the
same in both Stages and neared 90% of the retained TM errors, Phase II clearly shows
that the main assumptions of the book were correct. In the case of Stage I, the extended
exposure to CAT tools has proven to be beneficial for the students, who have improved
in finding and correcting errors by about 10% on the average. However, Stage II results
clearly show that the stress put to translation specifics, i.e. best practices in using CAT
tools, proves to be a more effective approach. Stage II students improved by about 30%
on the average (only around 62% of TM-based errors were retained, see Figure 48 below).
Average number of retained errors
100
90
80
70
60
50
40
30
20
10
0
Stage I
Stage II
Phase I
Phase II
Figure 48. Average number of retained errors between Stages.
What is interesting, while in Phase I in both Stages many students skipped altogether the
review of the 100% match suggestions offered by the program, some of them were found
to actually proofread those matches. However, the proofreading turned out to be very
cursory and frequently too short to allow for a proper revision. As a result, even quite
obvious errors like different numbers in source and target segments were skipped.
In Stage II, the student approach changed visibly. True, there were cases when
students skipped about 90% of errors, but the majority managed to detect and correct
significantly more errors than in Phase I. What is more, they were much more suspicious
about suggestions made by the program and attempted to rephrase some of the correct
suggestions. While in Phase I there was virtually no terminology verification, Phase II
Final remarks
173
saw a significant increase in this respect. An interesting fact is that in Phase I the export
of the translated document to a DOCX file ended the process. In Phase II many students
reviewed the DOCX file again, finding and removing spelling mistakes (own and TMbased). However, some of the students still failed to update the TM after revising the
DOCX file, even though it was a part of the course repeated time and time again as one
of the more frequent reasons behind errors in translation memories.
Ultimately, the results show that the main thesis of the book, i.e. that empowering
students/trainee translators through activities aimed to develop best practices in using
CAT tools, not limited to the translation mechanics, has a significant impact on translation
quality. It is especially true considering around 20% better results than a similar group of
students not prepared in a similar fashion, and only after about two months of intense
exposure to the quirks of the trade.
To recapitulate, the results prove that regular online training materials offered by
software developers, as well as commercial courses based on those, cannot be considered
sufficient to enter the CAT-based translation market. A CAT buyer is not prepared, in
terms of proper procedures and best practices, to use the software to its fullest extent and
not endanger his/her professional image as translators.
7. Results and conclusions
The results of the research seem to confirm the main assumptions of this book – students
who used the information contained in the official training videos, but then were exposed
to authentic quality issues when working with CAT tools, performed much better than
their colleagues who were not prepared in the same way. One can only speculate on the
reasons why the content concerning TM-quality is not introduced in those courses. The
insights taken from the analysis of the actual data lead to two possible options: a) CAT
tools are sophisticated computer programs with highly structured workflow that denies
the possibility of TM-based errors, and b) TM quality comes secondary in the context of
the translation competence of the translator who, given their professionalism, ideally will
be able to detect and correct errors regardless of their exposure to CAT tools.
If we look at the SDL Trados workflow (see Appendix), we can see that the
package-based version is more robust, with the pronounced role of a reviewer who
decides when the translation can be approved, and suggests corrections if that is not the
case. The package-based workflow is a natural choice for companies with established
operation procedures (like quality assurance, for example), because it allows to improve
quality by reducing the possibility of errors on various stages of the translation process.
On the other hand, there is the single document workflow, which concerns primarily
translation of a document with the use of own resources. The use of own resources
reduces the risk of re-using faulty translations (but only if we assume that the translator
makes no errors him/herself).
The problem is that no process is perfect and even package-based workflow cannot
guarantee the quality of translation memory or termbase, as was suggested by
professionals in the initial study. Therefore, not including the analysis of dangers of using
external resources for CAT-based translation into training materials may prove to be of
value when it comes to product marketing, but it certainly does not prepare CAT users
for every contingency. As a result, it does not prepare CAT users for effective work.
The second option considered was the over-reliance on technology. The captured
data revealed that some students did seem to proofread the suggestions made by the
program, but ultimately left them unchanged. On several occasions, the mouse cursor
would move slowly along the text as if the person was reading (which seems perfectly
logical considering that the recordings were single fluid sessions and some segments were
confirmed out of hand). However, with no access to an eye-tracking system, it is
impossible to verify this. If it can be assumed that the revision was there, but there were
no positive results, then there can be no doubt that something else has failed. A number
of factors can be named:
•
lack of conscious use of the software,
•
lack of reflection over own actions and their effects,
•
lack of independence as translators,
•
over-reliance on the machine and accepting suggestions offered by the
program with no second thoughts.
This raises a question of whether CAT tools should be used by “inexperienced” translators
at all. CAT tools are designed to help translators but, as was mentioned previously, they
“assist” the translator, not the other way round. Therefore, if the trainer does not help the
translator develop a kind of readiness and lack of trust in the machine, the ability to
question the content it suggests, and to review this content, the results may be far from
Final remarks
175
perfect and CAT tools may cause more harm than good in certain conditions. In the initial
research, professionals reported cases when the external TM was of such poor quality that
they could not use it for the project. They were “competent” enough to notice the fact and
make a proper decision on that basis. If it was not the case, the translation might continue
with poor result.
The presented problem may be worth considering in future studies because freely
available video courses do not prepare a potential CAT user as they should. True, they
show how to process a document in order to produce a translation. However, it may not
be enough for translators who are yet to develop their own independence and who may
assume that the feedback offered by the software is always of high quality. The training
in the use of CAT tools, or any automated computer resources for that matter, calls for a
new approach to the issue – one which could be tailored to meet individual needs of a
student and focus on the final translation quality, not only on the translation process.
Obviously, the process, or the mechanics, is important since it allows to perform the task
at all. However, if it does not go hand in hand with translation specifics, focused on the
quality, the ultimate success of the translator may be in danger. It has to be emphasized
that even if the use of CAT tools is required by the client, the CAT tools are only a means
to an end, not the end in itself. Ultimately, it is the quality of translation that matters and
brings the client back to the translator.
To sum up, the results of the main study confirm the initial assumptions and allow
to draw conclusions. Free courses may teach an individual how to operate a software, but
they will not teach them how to translate. What is more, such courses do not pay attention
to developing best practices in the use of CAT tools.
It is important to note that the discussion presented in this book addresses an
important issue in the process of translator training in the age of technology. The
contemporary translator is no longer solely a linguist – the trade requires an individual to
be a translator, consultant, IT specialist, graphic designer, marketing manager or
whomever the job requires them to be. There are more and more new tools that can assist
the translator in a number of ways; many of them advertised to be easy and intuitive to
use. However, not all solutions are obvious and easy to handle, especially for those
translators who are less computer literate.
The subject of the book was chosen with the desire to update the process of teaching
CAT tools. This goal is current also because many translation agencies in Poland and
abroad require new translators to use one or more CAT tools. Therefore, a graduate has
to be CAT-literate in order to be able to enter the translation market with any hopes for
success.
The research has shown that free online tutorials may not always be enough of a
training base to prepare the translator sufficiently to produce high-quality translations
and, therefore, achieve commercial success in his/her professional career. CAT tools, as
the name suggests, are devices to be used as an aid, an asset that improves translator’s
chances for success. Therefore, mastering them and the entire CAT-based translation
process is instrumental in achieving this success. By no means is this statement to mean
that it is always the case. Much depends on the person who is to learn how to use the
software. The research indicates that the greater awareness of the translation quality in a
translator, the smaller risk of committing (or copying) an error. However, less competent
and less experienced trainee translators do seem to experience problems in terms of the
effective application of CAT tools in everyday work. It was proven that, when asked to
translate a document using an external translation memory (provided by a customer),
trainees neglect a thorough revision of suggestions provided by the CAT tool, and thus
retain most of the errors (spelling, grammar, terminology, sense) originating from the
TM. What is more, greater exposure to translation with CAT tools did not change their
approach visibly. It was proven that translation mechanics, or mastering the technical
aspect of the tools, is not enough to yield satisfactory results. In order to achieve success,
both in terms of quality, standing on the market and, therefore, ability to attract new
customers and achieving financial satisfaction, the translator has to be trained to be
independent, self-reflective, conscious of the limitations of the software prepared to work
with external TM in terms of revision, form and origin of errors, and general caution
during the process of translation. The change was registered in students who focused on
both translation specifics and mechanics, while free online courses focus only on how to
handle the software, and not use its resources in context.
On the same hand, it should be noted that the final results of the Stage II Phase II
group were not as good as was expected. Most probably this result is related to students’
low experience in the trade. At the moment, it is difficult to speculate whether more
experienced translators would perform any better. However, the findings of this study
will serve as a basis for a future research on translator’s self-regulation, its development
and importance for the profession.
Nevertheless, the study confirmed that the inclusion of external data revision
policies into CAT environment leads to the development of practical aspects of translation
Final remarks
177
competence. As a result, trainee translators are more cautious and conscious of their work.
On average, Stage II Phase II students have shown significant improvement in translation
quality thanks to the application of revision procedures (during and after translation, and
after the translated document has been exported) and data management. Therefore, it can
be assumed that the increased independence and consciousness of own skills and machine
limitations would result in a greater chance for successful entering the highly competitive
translation market with a degree of success.
The study has also shown that the level of complexity of contemporary CAT tools
can be a challenge to many people. The technical aspects of handling such a computer
program are of course important and appropriate time should be devoted to mastering
those. However, the study has also allowed to identify the main problem, i.e. relinquishing
a portion of control over the translation process to the machine. The acronym CAT can
be expanded as Computer-Assisted human Translation; however, many students ended
up translating only untranslated content and assisting the computer in confirming the TM
suggestions (working akin to a semi-automatic MT). CAT tools are not MT systems, even
though they incorporate elements of machine translation. Therefore, relinquishing a part
of control to a computer program may, and frequently does, lead to disaster in terms of
the quality of translator service provision. The main goal of a translator is to provide topquality translations. Over-reliance on CAT systems may result in retaining TM-based
errors throughout a number of documents. It is true that CAT developers have come up
with a certain type workflow (varying from tool to tool), which assures the greatest
efficiency, consistency and relatively low error risk factor. However, the training in CAT
tools requires attention to be put to the risks of not following the workflow and dangers
of trusting the machine too much. Hence the need for insightful courses on translating in
CAT tool, not on using CAT tools.
It is an open secret that not all translation traffic is monitored and revised. What is
more, “revision by another person can only assure quality if this person is truly competent
and the translation/revision process is properly executed” (Martin, 2012). As a result a
significant portion of TMs available on the market, including online open repositories
like MyMemory (http://mymemory.translated.net/), may be full of errors. Most of those
are simple, yet easy to omit, errors that do not require any special expertise (e.g. spelling
mistakes). A translator who completes a regular online CAT course may be prepared in
terms of handling the program, but is prone to place too much trust in the data (translated,
accepted [and maybe even reviewed]) coming from other translators. It could be argued
that such problems could be negated by the translator’s expertise in a given field.
However, the initial study on professional translators indicated that they are willing to
translate documents outside their area of specialisation, under certain circumstances. The
same, albeit to a much greater extent, can be said about graduates and new translators
who do not have an established position on the market and have to translate all jobs given
to them just to stay on the market. The problem was signalled before – the lack of
awareness of the problem demonstrated in the present book may be a detriment to
effective use of CAT tools. A question arises: ‘Is it wise to expose translators to CAT
tools before they have the chance to gain some experience in the trade?’ Expertise in one
or more CAT tools becomes standard, and all new translators entering the market are
expected to know how to use them. Ultimately, if a trainee translator considers
professional career in translation, CAT tools are not an option – they are a necessary
condition to achieve commercial success. The competition on the market forces new
translators to open themselves to new technologies and new ways of translating in order
to compete with experienced translators with an established portfolio of clients.
The quality of a course can be measured in many ways. The average success of the
course participants on the market can be referred to as one of such markers. If we consider
that free online courses are available for everybody, students and professional translators
alike, we can see that students and non-CAT professional translators have different
starting conditions. Therefore, their final results, based on such courses, may be quite
different. It shows clearly that the average success of the course participants does not
reflect the potential of the course. Especially so when we speculate that it is easier for
professional translators to incorporate CAT tools into their translator’s workshop due to
several factors:
a) they have experience in regular translation (i.e. expertise in one or more fields);
b) they have a client base that they translate for and CAT tools are (in theory) a
means to enter new areas of the translation market; and
c) they are already shaped as translators, i.e. they are confident at what they are
doing.
Such people will probably be satisfied with free online courses because they are less timeconsuming than a CAT course at a university, for example. Trainee translators do not
have this background: they are prone to be influenced by other translators’ choices due to
their lack of experience and confidence; they do not have their own client base and are
forced to accept all offered jobs in order to gain the experience and expertise. In their
Final remarks
179
case, as the research suggests, CAT courses focused on how to operate the program do
not work. A different approach is needed – one that will teach the trainee how to use the
program consciously and effectively; when to trust external resources, when to be
cautious and how to make sure that no new errors are committed and transmitted further
with the exported TM. Students have to see that automation is very helpful in translation.
However, it has to be kept in check lest a translator takes responsibility for translation
that is only partially his own, and in fact performed by a machine. Pym (2013) suggested
the evolution of the profession of the translator to that of the post-editor. That may well
be true for the future. However, even in such a scenario translator competence (both in
terms of using tools, linguistic competence, independence, self-awareness, consciousness
and translation skills) is the key because, as Martin (2012) suggested, the fact that a
translation has been revised does not mean it is of high quality.
As a final remark, the author of the book wants to express satisfaction with the data
collection method. Despite minor technical problems, the data collected allowed for
insight into the actual translation process.
As was mentioned before, data collected during the research, as well as practices
developed and the experience gained, will be used to further research the development of
self-efficacy of the translator.
References
Ammon, U. 2014. Die Stellung der Deutschen Sprache in der Welt [The Status of the German
Language in the World]. Berlin: De Gruyter.
Asadi, P. and C. Séguinot. 2005. Shortcuts, Strategies and General Patterns in a Process Study of
Nine
Professionals.
Meta
50
(2).
522–547.
Available
at:
http://www.erudit.org/revue/meta/2005/v50/n4/019861ar.pdf.
Baer, B. J. and G. S. Koby. (eds.). 2003. Beyond the Ivory Tower. Rethinking Translation
Pedagogy. Amsterdam and Philadelphia: Benjamins.
Bahri, H. and T. S. T. Mahadi. 2016. Google Translate as a Supplementary Tool for Learning
Malay: A Case Study at Universiti Sains Malaysia. Advances in Language and Literary
Studies 7(3). [Online] Available from: http://dx.doi.org/10.7575/aiac.alls.v.7n.3p.161
[Accessed on: 09/08/2017]/
Barbe, K. 1996. The Dichtomy of Free and Literal Translation. META XLI 3. 328-337. [Online]
Available
from:
http://www3.uji.es/~aferna/H44/Free-literal.pdf
[Accessed
on:
17/04/2017]
Barbu, E. 2015. Spotting False Translation Segments in Translation Memories. In: C. Orăsan and
R. Gupta (eds.), Proceedings of the Workshop on Natural Language Processing for
Translation
Memories
(NLP4TM),
9–16.
[Online]
Available
from:
http://www.aclweb.org/anthology/W15-5202
Beeby Lonsdale, A. 1998. Direction of Translation (directionality). In: Mona Baker (ed.),
Routledge Encyclopedia of Translation Studies, 63-67. London and New York: Routledge.
Bell, R.T. 1988. Psycholinguistic/Cognitive Approaches. In: M. Baker and K. Malmkjær (eds.),
Routledge Encyclopedia of Translation Studies, 185-190. London: Routledge.
Bernardini, S. 2004. The Theory Behind the Practice. Translator Training or Translator
Education? In: K. Malmkjær (ed.), Translation in Undergraduate Degree Programmes.
Amsterdam and Philadelphia: Benjamins.
Biau Gil, J. R. 2007. What You See Is What You Get? A Pilot Experiment on Access to Visual
Information in Translation Interfaces.
Paper
presented
at
the
12th
Annual
Internationalisation and Localisation Conference, 26–28 September 2007, Dublin, Ireland.
[Online]
Available
from:
www.localisation.ie/resources/conferences/2007/
presentations/Biau_LRC_XII/LRC_XII_slides.pps
Bloodgood, M. and B. Strauss. 2014. Translation Memory Retrieval Methods. In: S. Wintner, S.
Goldwater and S. Riezler (eds.), Proceedings of the 14th Conference of the European
Chapter of the Association for Computational Linguistics, 202–210. Gothenburg:
Association for Computational Linguistics.
References
181
Bogucki, Ł., 2009. Tłumaczenie wspomagane komputerowo. Warszawa: PWN.
Bogucki, Ł., 2013, Areas and Methods of Audiovisual Translation Research. Lodz Studies in
Language vol. 30, Frankfurt am Main: Peter Lang.
Bowker, L. 2002. Computer-aided Translation Technology. [ebook] Ottawa: University of
Ottawa Press. Retrieved from: Google Books
Bowker, L. 2004. What Does it Take to Work in the Translation Profession in Canada in the 21st
Century? Exploring a Database of Job Advertisements. Meta 49(1). 960-972.
Bowker, L. 2005. Productivity vs Quality? A Pilot Study on the Impact of Translation Memory
Systems. Localisation Focus 4(1): 13–20.
Brooks, D. 2000. What Price Globalisation? Managing Costs at Microsoft. In: R. C. Sprung (ed.),
Translating into Success: Cutting-Edge Strategies for Going Multilingual in a Global Age,
43-57. Amsterdam andPhiladelphia: John Benjamins
Brzozowska, D. 2010. Humor in Cultural Discourse: Polish Jokes about China and Japan.
Cultural Perspectives 15. 68–83.
Bundgaard, K., Christensen, T. P. and A. Schjoldager. 2016. Translator-computer Interaction in
Action: an Observational Process Study of Computer-Aided Translation. Journal of
Specialised Translation 25. 106-130.
Bureau of Labor Statistics, U.S. Department of Labor, Occupational Outlook Handbook, 201617
Edition,
Interpreters
and
Translators.
[Online]
Available
from:
https://www.bls.gov/ooh/media-and-communication/interpreters-and-translators.htm
[Accessed on: 13/08/2017]
Burnett, C. 2001. The Coherence of the Arabic-Latin Translation Program in Toledo in the
Twelfth
Century.
Science
in
Context
14(1-2).
249-288.
https://doi.org/10.1017/S0269889701000096Published
Caminade, M. and A. Pym. 1998. Translator-Training Institutions. In: M. Baker (ed.),
Encyclopedia of Translation Studies, 280-285. London and New York: Routledge.
Campbell, S. 2002. Translation in the Context of EFL: the Fifth Macroskill? TEFLIN Journal :
Teaching English as a Foreign Language in Indonesia 13(1). 58-72.
Carr, P. 2006. Encyclopedia of Language and Linguistics: Second Edition, Volume Nine.
Amsterdam: Elsevier.
Castells, M. 1980. The Economic Crisis and American Society. Princeton: Princeton University
Press.
Chan, S. 2004. A Dictionary of Translation Technology. Hong Kong: Chinese University Press
Chesterman, A. 1993. From ‘is’ to ‘ought’: Laws, Norms and Strategies in Translation Studies.
Target vol 5 (1). 1–20
Chesterman, A. 1997. Memes of Translation: the Spread of Ideas in Translation Theory.
Amsterdam and Philadelphia: John Benjamins.
Chesterman, A. and E. Wagner. 2004. Can Theory Help Translators? A Dialogue Between the
Ivory Tower and the Wordface. Manchester: St Jerome.
Cho, S. and S. Kim. 2017. L1 Translation as Scaffolding in Tutor Talk: A Case Study of Two
Korean Tutors. GEMA Online® Journal of Language Studies 17(2), May 2017 [Online]
Available from: http://doi.org/10.17576/gema-2017-1702-02 [Accessed on: 09/08/2017]
Chodkiewicz, M., 2012. The EMT Framework of Reference for Competences Applied to
Translation: Perceptions by Professional and Student Translators. The Journal of
Specialised Translation 17. 37-54.
Choi, Y., 2015. Investigating Variation in Querying Behaviour for Image Searches on the Web.
ASIS&T ‘10 Proceedings of the 73rd ASIS&T Annual Meeting on Navigating Streams in
an Information Ecosystem 17(1). 1-10.
Chomsky, N. 1965. Aspects of Theory of Syntax. Cambridge: Massachusetts. MIT Press.
Christensen, T. P. and A. Schjoldager. 2016. Computer-aided Translation Tools – the Uptake and
Use by Danish Translation Service Providers. Journal of Specialised Translation 25. 89105.
Cohen, A. D. 1984. On Taking Tests: what the Students Report. Language testing 11(1). 70-81.
Cohen, J. M. 1986. “Translation.” Encyclopedia Americana 27. 12.
Commission of the European Communities. 1994. Final Evaluation of the Results of EUROTRA:
a Specific Programme Concerning the Preparation of – the Development of an Operational
EUROTRA System for Machine Translation. COM(94) 69 final. Brussels, 20.09.1994
Costa-Jussà, M. R., Farrús, M., Mariño, J. B. and J. A. R. Fonollosa. 2012. Study and Comparison
of Rule-Based and Statistical Catalan-Spanish MT Systems. Computing and Informatics
31. 245–270
Costa-Jussà, M., 2015. How Much Hybridization Does Machine Translation Need? Journal of
the Association for Information Science and Technology 66(10). 2160-2165.
Cronin, M. 2003. Translation and Globalisation. London: Routledge.
CSA. 1 April 2017. Survey of 3,000 Online Shoppers Across 10 Countries Finds that 60% Rarely
or
Never
Buy
from
English-only
Websites.
[Online]
Available
from
https://www.commonsenseadvisory.com/Default.aspx?Contenttype=ArticleDet&tabID=6
4&moduleId=392&Aid=21500&PR=PR. [Accessed on: 12/08/2017].
Dagilienė, I. 2012. Translation as a Learning Method in English Language Teaching. Studies
about Languages 21. 124-129. [Online]
Available from: http://dx.doi.org/10.5755/j01.sal.0.21.1469. [Accessed on: 08/08/2017]
de Saint Robert, M. J. 2008. CAT Tools in International Organisations. Topics in Language
Resources for Translation and Localisation. Benjamins Translation Library 79. 107-119
Declercq, C. 2015. Translation Technology in the United Kingdom. In: S. Chan (ed.), Routledge
Encyclopedia of Translation Technology, 364-374. London: Routledge
References
183
Degenhardt, M. (2006). CAMTASIA and CATMOVIE: Two Digital Tools for Observing,
Documenting and Analysing Writing Processes of University Students. In: L. Van Waes,
M. Leijten and D. Neuwirth (eds.), Writing and Digital Media. Oxford: Elsevier, 180–186.
Doherty, S. 2016. The Impact of Translation Technologies on the Process and Product of
Translation. International Journal of Communication 10(2016). 947–969.
Dragsted, B. (2010). Coordination of Reading and Writing Processes in Translation: an Eye on
Unchartered Territory. In: G. M. Shreve and E. Angelone (eds.), Translation and
Cognition, 41–61. Amsterdam: John Benjamins. doi: 10.1075/ata.xv.04dra
Drexler, V. 2016. “I don’t translate, I create!” An On-line Survey on Uniformity versus Creativity
in Professional Translations. Anchor Academic Publishing (Verlag).
Dreyfus, S. and H. Dreyfus. 1980. A Five-Stage Model of the Mental Activities Involved in
Directed
Skill
Acquisition.
[Online]
Available
from:
http://handle.dtic.mil/100.2/ADA084551. Copyright: Approved for public release.
Duff, A. 1994. Translation: Resource Books for Teachers. Oxford: Oxford University Press
Ehrensberger-Dow, M. and G. Massey. 2013. Indicators of Translation Competence: Translators’
Self-concepts and the Translation of Titles. Journal of Writing Research 5 (1). 103–131.
Ellis, D., Creutz, M., Honkela, T. and M. Kurimo. 2008. Speech to Speech Machine Translation:
Biblical Chatter from Finnish to English. In: The IJCNLP-08 Workshop on NLP for Less
Privileged Languages, 123–130. Hyderabad: Asian Federation of Natural Language
Processing.
EMT Expert Group, 2009. Competences for Professional Translators, Experts in Multilingual and
Multimedia Communication. [Online] Available from:
http://ec.europa.eu/dgs/translation/programmes/emt/key_documents/emt_competences_tr
anslators_en.pdf [Accessed on 14/08/2016]
EMT Expert Group, 2009. Developing a European network of translation programmes. [Online]
Available
from:
http://ec.europa.eu/dgs/translation/programmes/emt/key_documents/
emt_developing_network_en.pdf [Accessed on 14/08/2016]
Enriquez Raido, V. 2014. Translation and Web Searching. New York: Routledge.
Ericsson, K. A. and H. A. Simon. 1984. Protocol Analysis. Verbal Reports as Data. Cambridge,
MA: MIT Press.
Esselink, B. 2000. A Practical Guide to Localization. Amsterdam and Philadelphia: John
Benjamins
EU
SCIENCE
HUB.
n/a.
DGT-Translation
Memory.
[Online]
Available
from:
https://ec.europa.eu/jrc/en/language-technologies/dgt-translationmemory#More%20details%20/%20Reference%20publication [Accessed on: 16/04/2017]
European Commission. 2017. European Master’s in Translation (EMT) Explained. [Online]
Available
from:
https://ec.europa.eu/info/education/european-masters-translation-
emt/european-masters-translation-emt-explained_en [Accessed on: 24/08/2017]
European Union. 2017. EU Administration - Staff, Languages and Location. [Online] Available
from: https://europa.eu/european-union/about-eu/figures/administration_en [Accessed on:
13/08/2017]
Everline. 2015. Small Business Tracker – Summer 2015. [Online] Available from:
https://www.everline.com/app/uploads/2015/11/Everline-Small-Business-TrackerSummer-2015-presentation.pdf [Accessed on: 12/08/2017]
Fenyő, S. 2005. The Translator’s Cultural Competence. European Integration Studies 4(2).61-72.
Fernández-Parra, M. 2010. The Workflow of Computer-Assisted Translation Tools in Specialised
Translation. Reconceptualizing LSP: Online proceedings of the XVII European LSP
Symposium
2009.
[Online]
Available
http://bcom.au.dk/fileadmin/www.asb.dk/isek/fernandez-parra.pdf
from:
[Accessed
on
07/02/2017]
Gambier, Y. 2009. European Master’s in Translation: Competences for Professional Translators,
Experts in Multilingual and Multimedia Communication. Brussels: European Master’s in
Translation (EMT).
García, I. 2009 Research on Translation Tools. Translation Research Projects 2. 27.
Garcia, I. 2015. Computer-aided Translation Systems. In: S. Chan (ed.), Routledge Encyclopedia
of Translation Technology, 68-87. Amsterdam and Philadelphia: Routledge.
Gile, D. 1999. Norms in Research on Conference Interpreting: A Response to Theo Hermans and
Gideon Toury. In: C. Schäffner (ed.), Translation and Norms, 98-105. Multilingual
Matters.
González Davies, M. 2004. Multiple Voices in the Translation Classroom. Activities, Tasks and
Projects. Amsterdam and Philadelphia: Benjamins.
Google.
2016.
Use
Automatic
Captioning
[Online].
Available
from:
https://support.google.com/youtube/answer/6373554?hl=en [Accessed on: 30/11/2016]
Göpferich, S., 2009. Towards a Model of Translation Competence and its Acquisition: The
Longitudinal Study Transcomp. In: S. Göpferich, A. L. Jakobsen and I. M. Mees (eds.),
Behind the Mind. Methods, Models and Results in Translation Process Research, 11-37.
Copenhagen: Samfundslitteratur Press.
Gouadec, D. 2007. Translation as a Profession. Amsterdam and Philadelphia: John Benjamins.
Grabowski, Ł. 2009. The Role of Computer Skills in the Formation of Modern Translator’s
Professional Competence. V Międzynarodowa Konferencja Edukacyjna ‘Języki obce w
kontekście współczesnych wyzwań i perspektyw.’ Ustroń 2009. [Online] Available from:
http://www.kms.polsl.pl/prv/spnjo/index.php?st=nades [Accessed on 07/08/2017]
References
185
Green, S., Heer, J. and C. D. Manning. 2015. Natural Language Translation at the Intersection of
AI and HCI. acmqueue 13(6). [Online] Available from:
http://queue.acm.org/detail.cfm?id=2798086 [Accessed on: 09/02/2017]
Grosjean, F., 2001. The Bilingual’s Language Modes. In: J. L. Nicol (eds.), One Mind, Two
Languages: Bilingual Language Processing, 1-22. Oxford: Blackwell.
Grucza, S. 2012. The Investigation of Translation Competences and Translation Processes.
Vestnik MGLU, Moskiewski Państwowy Uniwersytet Lingwistyczny, Moskwa.
Hansen, G. 2010. Translation ‘Errors’. In: Y. Gambier and L. van Doorslaer (eds.), Handbook of
Translation Studies, 385–388. Amsterdam: John Benjamins.
Harris, B. and B Sherwood. 1978. Translation as an Innate Skill. In: D. Gerver and H. W. Sinaiko,
(eds.), Language, Interpretation and Communication, 155-170. New York-London::
Plenum Press.
Hatim, B. and J. Munday. 2004. Translation: An Advanced Resource Book. New York: Routledge
Hejwowski, K., 2004. Translation: a Cognitive-Communicative Approach. Olecko: Wszechnica
Mazurska.
Hönig, H. G. 1991. Holmes’ ‘Mapping Theory’ and the Landscape of Mental Translation
Processes. In: K. Leuven-Zwart and T. Naajkens, (eds.), Translation Studies: The State of
the Art. Proceedings from the First James S. Holmes Symposium on Translation Studies,
77-79. Amsterdam: Rodopi.
House, J. 2000. Consciousness and the Strategic Use of Aids in Translation. In: S. TirkkonenCondit & R. Jääskeläinen (eds.), Tapping and Mapping the Processes of Translation and
Interpreting: Outlooks on Empirical Research. Amsterdam and Philadelphia: Benjamins.
House, J. 2016. Translation as Communication across Languages and Cultures. New York:
Routledge.
Hutchins, J. 1995. “The Whisky Was Invisible”, or Persistent Myths of MT. MT News
International 11(1995). 17-18
Hutchins, J. W. 2006. Machine Translation: History. In: E. K. Brown and A. Anderson (eds.),
Encyclopedia of Language & Linguistics vol. 7, 375–383. Boston: Elsevier.
Hutchins, W. J. 1995. Machine Translation: a Brief History. In E. F. K. Koerner and R. E. Asher
(eds.), Concise History of the Language Sciences: from the Sumerians to the Cognitivists,
431-445. Oxford: Pergamon Press.
Hutchins, W. J. 2010. Machine Translation: a Concise History. Journal of Translation Studies,
vol. 13 (1-2). 29-70.
Hutchins, W. J. 1994. The Georgetown-IBM Experiment Demonstrated in January 1954. MT
News International 8 May 1994. 15-18.
Interoperable Delivery of European eGovernment Services to public Administrations, Businesses
and Citizens (IDABC). 2006. IATE: Inter-Agency Terminology Exchange. [Online]
Available from: http://ec.europa.eu/idabc/en/document/2294/5926.html [Accessed on:
30/11/2016]
Jääskeläinen, R. 2000. Focus on Methodology in Think-aloud Studies on Translating. In S.
Tirkkonen-Condit and R. Jääskeläinen (eds.), Tapping and Mapping the Processes of
Translation
and
Interpreting:
Outlooks
on
Empirical
Research,
71–82.
Amsterdam/Philadelphia: Benjamins.
Jääskeläinen, R. 2002. Think-aloud Protocol Studies into Translation: An Annotated
Bibliography. Target 14 (1). 107–136.
Jakobsen, A. L. and K. T. H. Jensen. 2008. Eye Movement Behaviour Across Four Different
Types of Reading Task. Copenhagen Studies in Language 36. 103-124.
Jarniewicz, J. 2000. Przekład tytułu: między egzotyką a adaptacją. In: W. Kubiński, O. Kubińska
and T. Z. Wolański (eds.), Przekładając nieprzekładalne, 477-483. Gdańsk: Wydawnictwo
Uniwersytetu Gdańskiego.
Kastberg, P. 2012. Machine Translation Tools – Tools of The Translator’s Trade. Communication
& Language at Work vol. 1(1). 32-44
Kearns, J. (ed.). 2008. Translator and Interpreter Training. Issues, Methods and Debates.
London: Continuum.
Kelly, D. 2005. A Handbook for Translator Trainers. Manchester: St Jerome.
Kenny, D. and A. Way. 2001. Teaching Machine Translation and Translation Technology: a
Contrastive Study. In The MT Summit VIII Workshop on Teaching Translation 2001, 18
– 22 September 2001, Santiago de Compostela, Spain.
Kilgray. 2017. memoQ 2015 help: Match Rates from Translation Memories or LiveDocs Corpora.
[Online] Available from: http://kilgray.com/memoq/2015-100/help-en/index.html?matchrates-from-translation-m.html [Accessed on: 21/03/2017]
Kiraly, D. 1995. Pathways to Translation: Pedagogy and Process. Kent: Kent State Univ. Press.
Kiraly, D. 2000. A Social Constructivist Approach to Translator Education. Manchester: St.
Jerome Publishing.
Kiraly, D. 2003. From Instruction to Collaborative Construction: a Passing Fact or the Promise
of a Paradigm Shift in Translator Education? In: B. J. Baer and G. S. Koby (eds.), Beyond
the Ivory Tower: Rethinking Translation Pedagogy, 3-27. Amsterdam: John Benjamins.
Kiraly, D. 2006. Beyond Social Constructivism: Complexity Theory and Translator Education.
The Journal of the American Translation and Interpreting Studies Association 1(1).
68-86.
Kiraly, D. 2013. Towards a View of Translator Competence as an Emergent Phenomenon:
Thinking Outside the Box(es) in Translator Education. In: D. Kiraly, S. Hansen-Schirra
and K. Maksymski (eds.), New Prospects and Perspectives for Educating Language
Mediators, Chapter: 10, 197-224. Narr Verlag.
References
187
Kiraly, D. 2014. From Assumptions about Knowing and Learning to Praxis in Translator
Education. In: M. Piotrowska and S. Tyupa (eds.), inTRAlinea Special Issue: Challenges
in Translation Pedagogy.
Kiraly, D. and Piotrowska, M. 2014. Towards an Emergent Curriculum Development Model for
the European Graduate Placement Scheme. The Future of Education 4th edition conference.
[Online] Available at: http://conference.pixel-online.net/FOE/files/foe/ed0004/FP/0366SET281-FP-FOE4.pdf [Access date: 26/11/2016]
Klimkowska, K. 2013. Orientacja na sukces zawodowy studentów kończących studia
translatorskie. Lublin: Wydawnictwo UMCS.
Klimkowski, K. 2015a. Entrepreneurial Training in Translator and Interpreter Education.
Roczniki Humanistyczne 63(11). 67-83.
Klimkowski, K. 2015b. Towards a Shared Curriculum in Translator and Interpreter Education.
Wrocław: Wydawnictwo Wyższej Szkoły Filologicznej we Wrocławiu.
Kornacki, M. in print. Technologia w codziennej pracy tłumacza. In: B. LewandowskaTomaszczyk (ed.), Kompetencje Współczesnego Tłumacza 2017.
Krings, H. P. 1986. Translation Problems and Translation Strategies of Advanced German
Learners of French. In: J. House and S. Blum-Kulka (ed.), Interlingual and Intercultural
Communication: Discourse and Cognition in Translation and Second Language
Acquisition Studies, 263–275. Tübingen: Narr.
Language Scientific. 2015. Software and Website Localization Process Outline. [Online]
Available from: http://www.languagescientific.com/translation-services/softwarelocalization-services/software-localization-process.html [Accessed on: 09 09 2015].
Leonardi, V. 2009. Teaching Business English through Translation. Journal of Language and
Translation 10(1). [Online]
Available
from:
http://www.sejongjul.org/download/download_pdf?pid=jul-10-1-139
[Accessed on: 10/08/2017]
Lewis, C. H. 1982. Using the “Thinking Aloud” Method in Cognitive Interface Design.
(Technical report). IBM RC-9265, IBM Watson Research Center, Yorktown Heights, NY,
USA.
Lörscher, W. 1991. Translation Performance, Translation Process and Translation Strategies. A
Psycholinguistic Investigation. Tübingen: Gunter Narr.
Lörscher, W. 2012. Bilingualism and Translation Competence. SYNAPS – A Journal of
Professional Communication 27/2012. 3-15.
LSP Software. 2017. Pamięci tłumaczeń utworzone z dokumentów Komisji Europejskiej.
[Online]
Available
from:
https://www.lspsoftware.pl/pamieci-tlumaczen-i-
slowniki/pamieci-tlumaczen-utworzone-z-dokumentow-komisji-europejskiej
on: 16/04/2017]
[Accessed
Lusthaus, D. 2000. Buddhist Phenomenology. London: Curzon Press
Mackenzie, R. 2004. The Competencies Required by the Translator’s Roles as a Professional. In:
K. Malmkjær (ed.), Translation in Undergraduate Degree Programmes, 31-38.
Amsterdam and Philadelphia: John Benjamins.
Malmkjær, K. (ed.). 2004. Translation in Undergraduate Degree Programmes. Amsterdam and
Philadelphia: John Benjamins.
Malmkjær, K. 1998. Translation and Language Teaching. Manchester: St. Jerome
Malmkjær, K., 2009. What is Translation Competence? Revue francaise de linguistique
appeliquée XIV(1/2009). 121-134.
Marczak, M., 2016. Developing Selected Aspects of Translation Competence through
Telecollaboration. English for Specific Purposes World 46(15). unpaginated.
Martin, C. 2012. The Dark Side of Translation Revision. Translation Journal 16(1). [Online]
Available
from:
http://translationjournal.net/journal/59editing.htm
[Accessed
on:
15/08/2017]
Mehidi, Y., 2016. Announcing New Subscription Options for Windows 10 and Surface for
Businesses. [Online] Available at: https://goo.gl/bnNTYQ [Access date: 18 November
2016].
Mell, P. and T. Grance. 2011. The NIST Definition of Cloud Computing. Gaithersburg: National
Institute of Standards and Technology: U.S. Department of Commerce.
Memsource.
2017.
Machine
Translation.
[Online]
Available
from:
https://wiki.memsource.com/wiki/Machine_Translation [Accessed on: 16/04/2017]
Memsource. 2017. Memsource Cloud User Manual: Supported File Formats. [Online] Available
from:
https://wiki.memsource.com/wiki/Memsource_Cloud_User_Manual#Supported_
File_Formats [Accessed on: 16/04/2017]
Mitkov, R., 2003. The Oxford Handbook of Computational Linguistics. Oxford: Oxford
University Press.
Montero Martínez, S. and P. Faber. 2009. Terminological Competence in Translation.
Terminological 15(1). 88-104.
Mossop, B. 2003. What Should Be Taught at Translation Schools? In: A. Pym et al. (eds.),
Innovation and E-learning in Translator Training. Tarragona: Intercultural Studies Group.
MT News International, no. 22, July 1999. Warren Weaver Memorandum, July 1949. [Online]
Available
at:
http://www.hutchinsweb.me.uk/MTNI-22-1999.pdf
[Accessed
on:
06/09/2016].
Munday, J. 2016. Introducing Translation Studies: Theories and Applications, 4th edition.
London and New York: Routledge
References
189
Neubert, A. 2000. Competence in Language, in Languages, and in Translation. In: C. Schäffner
and B. Adab (eds.), Developing Translation Competence, 3-18. Amsterdam and
Philadelphia: John Benjamins.
Newmark, P. 1988. A Textbook of Translation. Hertfordshire: Prentice Hall.
Ni, L. 2009. For “Translation and Theories”. English Language Teaching 2(2). 78-83
Nida, E., 1988. Bible Translation. In: M. Baker and K. Malmkjær (eds.), Routledge Encyclopedia
of Translation Studies, 22-29. London: Routledge.
Nord, C. 1997. Translating as a Purposeful Activity. Functionalist Approaches Explained.
Manchester: St. Jerome.
Nord, C., 1991. Text Analysis in Translation: Theory. Methodology and Didactic Application of
a Model for Translation-Oriented Text Analysis. Amsterdam: Rodopi.
Osers. E. 1995. Translation Norms. In: A. Beylard-Ozeroff, J. Kralova and B. Moser-Mercer
(eds.), Translators’ Strategies and Creativity, 53-62. Amsterdam and Philadelphia: John
Benjamins.
PACTE Group. 2000. Acquiring Translation Competence: Hypotheses and Methodological
Problems in a Research Project. In: A. Beeby, D. Emsinger and M. Presas, (eds.),
Investigating Translation. Amsterdam: Benjamins.
PACTE Group. 2003. Building a Translation Competence Model. In: F. Alves (ed.),
Triangulating Translation: Perspectives in Process Oriented Research, 43-65.
Amsterdam: John Benjamins.
PACTE Group. 2005. Investigating Translation Competence: Conceptual and Methodological
Issues. Meta 50 (2). 609-619.
Panou, D. 2013. Equivalence in Translation Theories: A Critical Evaluation. Theory and Practice
in Language Studies 3(1). 1-6. doi:10.4304/tpls.3.1.1-6
Paradowska, U., 2015. Expert Web Searching Skills for Translators – a Multiple-case Study. In:
P. Pietrzak and M. Deckert, (eds.), Constructing Translation Competence, 227-244.
Frankfurt am Main: Peter Lang.
Paulsen, M. F. 1995. An Overview of CMC and the Online Classroom in Distance Education. In:
Z. Berge and M. Collins (eds.), Computer-mediated Communication and the Online
Classroom, vol. III: Distance Learning. Cresskill, New Jersey: Hampton Press
Paulson, F. L. and P. R. Paulson. 1994. Assessing Portfolios Using the Constructivist Paradigm.
Paper presented at the Annual Meeting of the American Educational Research Association
in New Orleans, Los Angeles. 4-8 April 1994. Washington, D.C.
Piecychna, B., 2013. Legal Translation Competence in the Light of Translation Hermeneutics.
Studies in Logic, Grammar and Rhetoric, October, vol. 34. 141-159.
Piotrowska, M., 2015. Revisiting the Translator Competence in the 21st Century. In: P. Pietrzak
and M. Deckert, (eds.), Constructing Translation Competence, 13-25. Frankfurt am Main:
Peter Lang.
Płońska, D. 2014. Strategies of Translation. Psychology of Language and Communication 18(1),
67-74.
Pym, A. 1992. Translation Error Analysis and the Interference with Language. In: C. Dollerup
and A. Loddegaard (eds.), The Teaching of Translation, 279-288. Amsterdam: John
Benjamins.
Pym, A. 2003. Redefining Translation Competence in an Electronic Age. In Defence of a
Minimalist Approach. META XLVIII(4). 481-497.
Pym, A. 2009. Translator Training. Pre-print text written for the Oxford Companion to
Translation
Studies.
[Online]
Available
at:
http://usuaris.tinet.cat/apym/on-
line/training/2009_translator_training.pdf [Access date: 07/10/2015].
Pym, A. 2013. Translation Skill-Sets in a Machine-Translation Age. Translators’ Journal 58(3):
487-503
Pym, A. 2014. Vinay and Darbelnet and the Politics of Translation Solutions. [Online] Available
from:
http://usuaris.tinet.cat/apym/on-line/translation/2014_vinay_darbelnet.pdf.
[Accessed on: 05/06/2017]
Pym, A. 2017. For a Sociology of Translator Training. A talk given at the inaugural conference
of the World Interpreter and Translator Training Association, Guangzhou, China,
November
19,
2016. [Online]
Available
from:
http://usuaris.tinet.cat/apym/on-
line/training/2017_sociology_training.pdf [Accessed on: 27/08/2017]
Pym, A. and E. Torres-Simón. 2014. The Pedagogical Value of Translation Solution Types.
Perspectives 23(1). 89-106
Pym, A. Grin, F., Sfreddo, C. and Chan, A. L. J., 24 July 2012 . The Status of the Translation
Profession in the European Union (DGT/2011/TST), European Commission.
Pym, A. Malmkjær, K. and M. del Mar Gutiérrez-Colón Plana. 2012. Translation and Language
Learning: An Analysis of Translation as a Method of Language Learning in Primary,
Secondary and Higher Education. A Research proposal DGT/2012/TLL. [Online]
Available
from:
http://www.est-translationstudies.org/research/2012_DGT/reports/
2012_technical_proposal.pdf [Accessed on: 10/08/2017]
Radford, A. 1988. Transformational Grammar: A First Course. Cambridge: Cambridge
University Press.
Radford, A. 2004. English Syntax: An Introduction. Cambridge: Cambridge University Press.
Ressurrecció, V. M., P. E. Piorno, and I. G. Izquierdo. 2008. The Acquisition of Translation
Competence through Textual Genre. Translation Journal 12 (4). [Online] Available from:
http://translationjournal.net/journal/46competence.htm
References
191
Risku, H. 2014. Translation Process Research as Interaction Research: from Mental to SocioCognitive Processes. In: R. Muñoz Martín (ed.), Minding Translation. Special Issue 1 of
MonTI, 331–353.
Risku, H., Pein-Weber, C. and J. Milošević. 2016. “The Task of the Translator”: Comparing the
Views of the Client and the Translator. International Journal of Communication 10(2016).
989–1008.
Robinson, D. 1997-2002. Western Translation Theory. From Herodotus to Nietzsche.
Manchester: St. Jerome.
Sager, J. C. 1994. Language Engineering and Translation: Consequences of Automation.
Amsterdam & Philadelphia: John Benjamins
Santander Consumer Bank. 2016. Boosting UK Exports: Helping Businesses Expand Overseas.
Report
2016.
[Online].
Available
from:
https://www.santandercb.co.uk/s3fs-
factsheets/scrn_rdy_santander_report_26pg.pdf [Accessed on: 12/08/2017]
Schäffner, C. 1998. The Concept of Norms in Translation Studies. Current Issues in Language
and Society 5(1-2). 1-9. http://dx.doi.org/10.1080/13520529809615500
Schäffner, C. and B. Adab. 2000. Developing Translation Competence. Amsterdam/Philadelphia:
John Benjamins.
Schneckenberg, D. and W. Johannes. 2006. Understanding the Concept of Ecompetence for
Academic Staff. In: I. Mac Labhrainn, C. McDonald Legg, D. Schneckenberg and J. Wildt,
(eds.), The Challenge of eCompetence in Academic Staff Development, 29-35. Galway:
Centre for Excellence in Learning and Teaching.
SDL
Documentation
Center.
2017.
Product
Help.
[Online]
Available
from:
http://producthelp.sdl.com/sdl%20trados%20studio/client_en/Projects_View/P_About_Pr
oject_Packages.htm#Project_Package [Accessed on: 28/03/2017]
SDL
Trados
Studio
2017
|
SDL.
2016.
[Online]
Available
from:
http://www.sdl.com/solution/language/translation-productivity/trados-studio/what
[Accessed on: 27/11/2016]
SDL Translation Zone. 2015. What is Software Localization?. [Online] Available from:
http://www.translationzone.com/products/software-localization
[Accessed
on:
09/09/2015].
SDL* AppStore. 2017. MT Enhanced Plugin for Trados Studio. [Online] Available from:
http://appstore.sdl.com/app/mt-enhanced-plugin-for-trados-studio/150/
[Accessed
on:
16/04/2017]
Shäffner, C., 2000. Designing a Translation Programme. In: C. Schäffner and B. Adab (eds.),
Developing Translation Competence, 143-156. Amsterdam/Philadelphia: John Benjamins.
Shiwen, Y. and B. Xiaojing. 2015. Rule-based Machine Translation. In: C. Sin-wai (ed.), The
Routledge Encyclopedia of Translation Technology, 186-200. New York: Routledge
Sikes, R. 2011. Rethinking the Role of the Localization Project Manager. In: K. J. Dunne and E.
S. Dunne (eds.), Translation and Localization Project Management: The Art of the
Possible, 235–264. Amsterdam: John Benjamins.
Silva, R. 2014. Integrating Post-Editing MT in a Professional Translation Workflow. In: S.
O’Brien et al. (eds.), Post-editing of Machine Translation: Processes and Applications, 2450. Newcastle upon Tyne: Cambridge Scholars
Sing. 2016. [video] USA: Illumination Entertainment
Sin-wei, C. 2015. The Development of Translation Technology 1967-2013. In: C. Sin-wai (ed.),
The Routledge Encyclopedia of Translation Technology, 3-31. New York: Routledge
Smith, C. 2016. By the Numbers: 29 Amazing WordPress Statistics. DMR [Online]. Available
from:
http://expandedramblings.com/index.php/wordpress-statistics/
[Accessed
on:
30/11/2016]
Smith, M. 2016. The Influence of French on Eighteenth-Century Literary Russian. Oxford, Bern,
Berlin, Bruxelles, Frankfurt am Main, New York, Wien: Peter Lang.
Sofyan, R. and B. Tarigan. 2017. Online Resources Management in Self-Corrections and
Translation Quality. International Journal of Sciences: Basic and Applied Research
(IJSBAR); Vol 35(2)
Statistics on the DGT‐Translation Memory (DGT‐TM). 7 March 2016. [Online] Available from:
http://optima.jrc.it/Resources/DGT-TM_Statistics.pdf [Accessed on: 16/04/2017]
Straub, D. April 2013. What Delights Purchasers of Translation Services... tcworld April 2013.
[Online] Available from: http://www.tcworld.info/rss/article/what-delights-purchasers-oftranslation-services/ [Accessed on: 09/02/2017]
Study
in
Turkey.
2017.
Education
System.
[Online]
http://www.studyinturkey.com/content/sub/education_system.aspx
Available
from:
[Accessed
on:
24/08/2017]
Sun, S. 2012. Strategies of Translation. In: C. A. Chapelle (ed.), The Encyclopedia of Applied
Linguistics. [Online] DOI: 10.1002/9781405198431.wbeal1117. John Wiley and Sons, Inc.
SYSTRAN.
2016.
What
Is
Machine
Translation?
[Online]
Available
from:
http://www.systransoft.com/systran/corporate-profile/translation-technology/what-ismachine-translation/ [Access date: 28/11/2016]
Tardzenyuy, N. C. 2016. Revisiting Translation Strategies and Techniques. International Journal
of Comparative Literature & Translation Studies vol. 4:4(2016). 48-56.
Tennent, M. (ed.). 2005. Training for the New Millennium. Pedagogies for Translation and
Interpreting. Amsterdam and Philadelphia: John Benjamins.
TEPIS.
2017.
The
Sworn
Translator’s
Code.
[Online]
Available
from:
http://www.tepis.org.pl/index.php/tepis-o-sobie/dzialalnosc-pt-tepis/the-polish-society-ofsworn-and-specialized-translators [Access date: 08/02/2017]
References
TEPIS.
193
2017.
Uczelnie
kształcące
tłumaczy.
[Online]
http://www.tepis.org.pl/index.php/linki/uczelnie-ksztalcace-tlumaczy
Available
from:
[Accessed
on:
24/08/2017]
The Washington Post. 2015. The World’s Languages, in 7 Maps and Charts. By R. Noack
and L. Gamio. [Online] Available from:
https://www.washingtonpost.com/news/worldviews/wp/2015/04/23/the-worldslanguages-in-7-maps-and-charts/?utm_term=.3b06f4b6b8ed [Accessed on: 12/08/2017]
Tirkkonen-Condit, S. 1989. Professional vs. Non-Professional Translation: a Think-Aloud
Protocol Study. In C. Séguinot (ed.), The Translation Process, 73-85. Toronto: HG
Publications.
Toury, G. 1999. A Handful of Paragraphs on “Translation” and “Norms.” In: C. Schäffner (ed.),
Translation and Norms, 9-31. Multilingual Matters.
Toury, G. 1984. The Notion of ‘Native Translator’ and Translation Teaching. In: W. Wilss and
G. Thome, (eds.), Die Theorie des Übersetzens und ihr Aufschlusswert für die
Übersetzungs- und Dolmetschdidaktik, 186-195. Tübingen: Narr.
Toury, G. 1986. Natural Translation and the Making of a Native Translator. TEXTconTEXT.
11-29.
Translation Automation User Society (TAUS). 2017. Computer-Assisted Translation. [Online]
Available
from:
https://www.taus.net/knowledgebase/index.php/Computer-
assisted_translation [Accessed on: 20/03/2017]
Trombetti, M. 2009. Creating the World’s Largest Translation Memory. [Online] Available from:
http://www.mt-archive.info/05/Aslib-2009-Trombetti.pdf [Accessed on: 09/02/2017]
Vázquez, L. M. and S. R. Vázquez. 2014. Beyond Traditional Computer-assisted Translation
Tools. A Localisation Perspective. Presentation delivered at Equivalences 2014 «
Traduction et téchnique » conference, 24 October 2014. Kursaal, Bern
Venuti, L. 1995. The Translator’s Invisibility: a History of Translation. London and New York:
Routledge.
Venuti, L. 1998. Strategies of Translation. In: M. Baker (ed.), Routledge Encyclopedia of
Translation Studies, 240-244. London and New York: Routledge
Venuti, L. 2008. The Translator’s Invisibility: A History of Translation. Second Edition. London
& New York: Routledge
Vinay, J.-P. and J. Darbelnet. 1958/2000. A Methodology for Translation [translated by J. C.
Sager and M.-J. Hamel]. In: L. Venuti (ed.), The Translation Studies Reader, 84-93.
London and New York: Routledge.
Wach, S. 2013. Calquing English Terminology into Polish. Academic Journal of Modern
Philology 2(2013). 161-169
Waliński, J. T. 2015. Translation Procedures. In: Ł. Bogucki, S. Goźdź-Roszkowski and P.
Stalmaszczyk (eds.), Ways to Translation, 53-65. Łódź: Wydawnictwo Uniwersytetu
Łódzkiego.
Way, C. 2000. Structuring Specialised Translation Courses: A Hit and Miss Affair. In: C.
Schäffner and B. Adab, (eds.), Developing Translation Competence, 131-141.
Amsterdam/Philadelphia: John Benjamins.
Wills, W. 1982. The Science of Translation. Tubingen: Gunter Narr.
Wordpress.com. 2016. A live look at activity across WordPress.com [Online]. Available at:
https://wordpress.com/activity/ [Accessed: 30/11/2016]
Yamada, M. 2011. The Effect of Translation Memory Databases on Productivity. In: A. Pym
(ed.), Translation Research Projects 3, 63-73. Tarragona: Intercultural Studies Group.
Yang, L. and Z. Min. 2015. Statistical Machine Translation. In: C. Sin-wai (ed.), The Routledge
Encyclopedia of Translation Technology, 201-212. New York: Routledge
Zimmerman, Barry J. 2000. Attaining self-regulation: A social cognitive perspective. In: Monique
Boekaerts, Paul R. Pintrich and Moshe Zeidner (eds.), Handbook of self-regulation, 13–39.
San Diego, CA: Academic Press. https://doi.org/10.1016/B978-012109890-2/50031-7
Appendices
195
Appendices
Appendix 1: List of procedures for CAT-based translation
 prepare a document for translation (formatting, OCR, etc.)
 create a project and enter all required information
 import a document (use filters if necessary)
 add (check) or create a translation memory
 add (check) or create a termbase
 review project settings
 translate the document
 if you do not specialise in the subject and you use external translation memory, review
 match suggestions carefully
 review specialised terminology if you are unsure about the quality of obtained resources
 run QA check and review the document
 update your TM
 export the document
Appendix 2: Main research source document – Phase I
OPINIA NIEZALEŻNEGO BIEGŁEGO REWIDENTA
Dla Zgromadzenia Wspólników firmy.
Opinia o sprawozdaniu finansowym
Przeprowadziliśmy badanie załączonego sprawozdania finansowego firmy. z siedzibą w Warszawie
(„Spółka”), na które składa się wprowadzenie do sprawozdania finansowego, bilans sporządzony na dzień
31 grudnia 2013 r., rachunek zysków i strat, zestawienie zmian w kapitale własnym oraz rachunek
przepływów pieniężnych za rok obrotowy kończący się tego dnia oraz dodatkowe informacje i objaśnienia.
Odpowiedzialność Zarządu
Zarząd Spółki jest odpowiedzialny za prawidłowość ksiąg rachunkowych, sporządzenie i rzetelną
prezentację tego sprawozdania finansowego oraz sporządzenie sprawozdania z działalności zgodnie z
ustawą z dnia 29 września 1994 r. o rachunkowości (Dz. U. z 2013 r. poz. 330 z późniejszymi zmianami)
(„ustawa o rachunkowości”), wydanymi na jej podstawie przepisami wykonawczymi oraz innymi
obowiązującymi przepisami prawa. Zarząd Spółki jest odpowiedzialny również za kontrolę wewnętrzną,
którą uznaje za niezbędną, aby sporządzane sprawozdania finansowe były wolne od nieprawidłowości
powstałych wskutek celowych działań lub błędów.
Zgodnie z ustawą o rachunkowości, Zarząd Spółki jest zobowiązany do zapewnienia, aby sprawozdanie
finansowe oraz sprawozdanie z działalności spełniały wymagania przewidziane w tej ustawie.
Odpowiedzialność Biegłego Rewidenta
Naszym zadaniem jest, w oparciu o przeprowadzone badanie, wyrażenie opinii o tym sprawozdaniu
finansowym oraz prawidłowości ksiąg rachunkowych stanowiących podstawę jego sporządzenia. Badanie
sprawozdania finansowego przeprowadziliśmy stosownie do postanowień rozdziału 7 ustawy o
rachunkowości, krajowych standardów rewizji finansowej wydanych przez Krajową Radę Biegłych
Rewidentów oraz Międzynarodowych Standardów Rewizji Finansowej. Regulacje te nakładają na nas
obowiązek postępowania zgodnego z zasadami etyki oraz zaplanowania i przeprowadzenia badania w taki
sposób, aby uzyskać racjonalną pewność, że sprawozdanie finansowe i księgi rachunkowe stanowiące
podstawę jego sporządzenia są wolne od istotnych nieprawidłowości.
Badanie polega na przeprowadzeniu procedur mających na celu uzyskanie dowodów badania dotyczących
kwot i informacji ujawnionych w sprawozdaniu finansowym. Wybór procedur badania zależy od naszego
osądu, w tym oceny ryzyka wystąpienia istotnej nieprawidłowości sprawozdania finansowego na skutek
celowych działań lub błędów. Przeprowadzając ocenę tego ryzyka bierzemy pod uwagę kontrolę
wewnętrzną związaną ze sporządzeniem oraz rzetelną prezentacją sprawozdania finansowego w celu
zaplanowania stosownych do okoliczności procedur badania, nie zaś w celu wyrażenia opinii na temat
skuteczności działania kontroli wewnętrznej w jednostce. Badanie obejmuje również ocenę odpowiedniości
stosowanej polityki rachunkowości, zasadności szacunków dokonanych przez Zarząd Spółki oraz ocenę
ogólnej prezentacji sprawozdania finansowego.
Wyrażamy przekonanie, że uzyskane przez nas dowody badania stanowią wystarczającą i odpowiednią
podstawę do wyrażenia przez nas opinii z badania.
Wprowadzenie do sprawozdania finansowego
(wszystkie dane liczbowe przedstawiono w tysiącach złotych)
Dane identyfikujące Spółkę
Nazwa Spółki
Nazwa firmy, zwana dalej Spółką
Siedziba Spółki
Aleje Jerozolimskie, Warszawa
Rejestracja w Krajowym Rejestrze Sądowym
Siedziba sądu: Sąd Rejonowy dla m.st. Warszawy w Warszawie, XII Wydział Gospodarczy Krajowego
Rejestru Sądowego
Data:
22 listopada 2001 r.
Numer rejestru: KRS 0000000000
Podstawowy przedmiot i czas działalności Spółki:
Przedmiotem działalności Spółki zgodnie z umową Spółki jest między innymi:
• handel hurtowy środkami ochrony roślin i innymi produktami chemicznymi,
• produkcja dodatków do betonów wszelkich typów,
• handel hurtowy i produkcja materiałów budowlanych,
• usługi pośrednictwa handlowego na rzecz firm grupy.
Appendices
197
Spółka została utworzona na czas nieokreślony.
Okres objęty sprawozdaniem finansowym
Sprawozdanie finansowe zostało przygotowane za okres od 1 stycznia 2013 r. do 31 grudnia 2013 r.,
natomiast dane porównawcze obejmują okres od 1 stycznia 2012 r. do 31 grudnia 2012 r.
Bilans
(wszystkie dane liczbowe przedstawiono w tysiącach złotych)
AKTYWA
Nota
31.12.2013
31.12.2012
281 498,7
76 064,8
Inne wartości niematerialne i prawne
3 494,9
4 327,6
Zaliczki na wartości niematerialne i prawne
680,7
-
4 175,6
4 327,6
Środki trwałe
31 823,3
35 547,8
grunty (w tym prawo wieczystego użytkowania gruntu)
12 437,8
12 437,8
budynki, lokale i obiekty inżynierii lądowej i wodnej
12 621,9
14 150,5
urządzenia techniczne i maszyny
2 305,8
3 347,9
środki transportu
3 930,1
5 070,3
inne środki trwałe
527,7
541,3
Środki trwałe w budowie
234 926,1
8 755,0
Zaliczki na środki trwałe w budowie
-
17 940,4
266 749,4
62 243,2
1 495,6
1 384,2
1 495,6
1 384,2
Aktywa trwałe
Wartości niematerialne i prawne
Rzeczowe aktywa trwałe
Należności długoterminowe
1
2
3
Od pozostałych jednostek
Długoterminowe rozliczenia międzyokresowe
Aktywa z tytułu odroczonego podatku dochodowego
17.3
8 574,0
7 636,5
Inne rozliczenia międzyokresowe
4
504,1
473,3
9 078,1
8 109,8
232 798,2
269 208,7
Materiały
769,8
1 195,5
Produkty gotowe
1 260,9
2 073,8
Towary
101 867,0
92 809,8
Aktywa obrotowe
Zapasy
5
Zaliczki na dostawy
934,1
902,8
104 831,8
96 981,9
8 899,9
8 345,6
8 899,9
8 345,6
102 798,1
89 231,8
84 467,9
78 854,3
społecznych i zdrowotnych oraz innych świadczeń
17 391,8
9 347,2
inne
938,4
1 030,3
111 698,0
97 577,4
15 156,0
73 576,2
15 156,0
73 576,2
15 156,0
73 576,2
1 112,4
1 073,2
514 296,9
345 273,5
Nota
31.12.2013
31.12.2012
9.1
21 313,2
21 313,2
Kapitał zapasowy
6 694,6
6 694,6
Kapitał z aktualizacji wyceny
0,2
0,2
Pozostałe kapitały rezerwowe
115 508,8
112 119,7
12 502,9 1
3
56 019,7
516,8
358 277,2
201 756,7
Należności krótkoterminowe
Należności od jednostek powiązanych
z tytułu dostaw i usług
6.1
Należności od pozostałych jednostek
z tytułu dostaw i usług
6.2
z tytułu podatków, dotacji, ceł, ubezpieczeń
Inwestycje krótkoterminowe
Krótkoterminowe aktywa finansowe
środki pieniężne i inne aktywa pieniężne
Krótkoterminowe rozliczenia międzyokresowe
7.1
8
AKTYWA RAZEM
PASYWA
Kapitał własny
Kapitał zakładowy
Zysk netto
Zobowiązania i rezerwy na zobowiązania
389,1
Rezerwy na zobowiązania
Rezerwa z tytułu odroczonego podatku dochodowego
17.3
661,9
817,9
Rezerwa na świadczenia emerytalne i podobne
10.1
3 197,8
2 459,6
- długoterminowa
1 494,4
1 190,1
- krótkoterminowa
1 703,4
1 269,5
Pozostałe rezerwy
4 567,0
6 024,3
3 744,6
3 697,8
822,4
2 326,5
8 426,7
9 301,8
- długoterminowe
10.2
- krótkoterminowe
10.3
143
Appendices
199
Zobowiązania krótkoterminowe
Wobec jednostek powiązanych
281 173,8
164 082,8
67 995,7
164 082,8
inne
213 178,1
-
Wobec pozostałych jednostek
10 775,0
8 126,4
4 946,1
5 465,5
zaliczki otrzymane na dostawy
2,7
0,2
z tytułu podatków, ceł, ubezpieczeń i innych świadczeń
1 276,0
2 445,2
inne
4 550,2
215,5
137,3
100,4
292 086,1
172 309,6
57 764,4
20 145,3
z tytułu dostaw i usług
11.1
z tytułu dostaw i usług
11.2
Fundusze specjalne
Rozliczenia międzyokresowe
Inne rozliczenia międzyokresowe
- długoterminowe
12.1
4 838,8
813,5
- krótkoterminowe
12.2
52 925,6
19 331,8
57764,4
20145,3
514269,9
345273,5
01.01.2013-
01.01.2012-
31.12.2013
31.12.2012
w tym od jednostek powiązanych
55 024,5
47 299,1
Przychody netto ze sprzedaży produktów
91 456,3
84 312,8
Zmiana stanu produktów
(593,8)
138,6
Przychody netto ze sprzedaży towarów i materiałów
408 275,0
362 879,9
499 137,5
447 331,3
Amortyzacja
(4 836,2)
(5 785,5)
Zużycie materiałów i energii
(26 327,7)
(25 982,8)
Usługi obce
(47 018,6)
(39 702,6)
Podatki i opłaty
(1 812,8)
(1 686,9)
Wynagrodzenia
(37 592,6)
(35 345,3)
Ubezpieczenia społeczne i inne świadczenia
(9 295,6)
(8 383,6)
Pozostałe koszty rodzajowe
(28 427,0)
(27 859,7)
Wartość sprzedanych towarów i materiałów
(325 439,4)
(280 339,9)
PASYWA RAZEM
Rachunek zysków i strat
(wszystkie dane liczbowe przedstawiono w tysiącach złotych)
Nota
Przychody netto ze sprzedaży i zrównane z nimi
13
Koszty działalności operacyjnej
(480 749,9)
(425 086,3)
18 387,6
22 245,0
Zysk ze zbycia niefinansowych aktywów trwałych
654,0
379,8
Inne przychody operacyjne
2 416,2
624,6
3 070,2
1 004,4
Aktualizacja wartości aktywów niefinansowych
(2 386,2)
(11 342,5)
Inne koszty operacyjne
(2 157,9)
(3 408,0)
(4 544,1)
(14 750,5)
16913,7
8 498,9
1 087,7
2410,3
847,0
2 265,5
1 087,7
2 410,3
(501,4)
(172,5)
- w tym dla jednostek powiązanych
(501,4)
(6,6)
Inne
(978,0)
(1 154,6)
(1 479,4)
(1 327,1)
Zysk ze sprzedaży
Pozostałe przychody operacyjne
14
Pozostałe koszty operacyjne
Zysk z działalności operacyjnej
Przychody finansowe
Odsetki
15
- w tym od jednostek powiązanych
Koszty finansowe
Odsetki
16
Rachunek zysków i strat należy analizować łącznic z dodatkowymi informacjami i objaśnieniami, które stanowią
integralną część sprawozdania finansowego.
Rachunek zysków i strat
(wszystkie dane liczbowe przedstawiono w tysiącach złotych)
Nota
Zysk brutto
Podatek dochodowy
17
Zysk netto
01.01.2013-
01.01.2012-
31.12.2013
31.12.2012
16 522,0
9 582,1
(4 019,1)
(6 193,0)
12 502,9
3 389,1
Appendix 3: Main research source document – Phase II
OPINIA NIEZALEŻNEGO BIEGŁEGO REWIDENTA
Dla Walnego Zgromadzenia Wspólników firmy
Appendices
201
Opinia o skonsolidowanym sprawozdaniu finansowym
Przeprowadziliśmy badanie załączonego skonsolidowanego sprawozdania finansowego firmy z siedzibą w
Tarnowie („Grupa Kapitałowa”), na które składa się skonsolidowane sprawozdanie z sytuacji finansowej
sporządzone na dzień 31 grudnia 2012 r., skonsolidowane sprawozdanie z całkowitych dochodów,
skonsolidowane sprawozdanie ze zmian w kapitale własnym oraz skonsolidowane sprawozdanie z
przepływów pieniężnych za rok obrotowy kończący się tego dnia oraz informacje dodatkowe do
skonsolidowanego sprawozdania finansowego, zawierające opis istotnych zasad rachunkowości oraz inne
informacje objaśniające.
Odpowiedzialność Zarządu oraz Rady Nadzorczej
Zarząd jednostki dominującej jest odpowiedzialny za sporządzenie i rzetelną prezentację tego
skonsolidowanego sprawozdania finansowego zgodnie z Międzynarodowymi Standardami
Sprawozdawczości Finansowej, które zostały zatwierdzone przez Unię Europejską i innymi
obowiązującymi przepisami prawa oraz sporządzenie sprawozdania z działalności. Zarząd jednostki
dominującej jest odpowiedzialny również za kontrolę wewnętrzną, którą uznaje za niezbędną, aby
sporządzane skonsolidowane sprawozdania finansowe były wolne od nieprawidłowości powstałych
wskutek celowych działań lub błędów.
Zgodnie z ustawą z dnia 29 września 1994 r. o rachunkowości (Dz. U. z 2009 r. nr 152, poz. 1223 z
późniejszymi zmianami) („ustawa o rachunkowości”), Zarząd jednostki dominującej oraz członkowie Rady
Nadzorczej są zobowiązani do zapewnienia, aby skonsolidowane sprawozdanie finansowe oraz
sprawozdanie z działalności spełniały wymagania przewidziane w tej ustawie.
Odpowiedzialność Biegłego Rewidenta
Naszym zadaniem jest, w oparciu o przeprowadzone badanie, wyrażenie opinii o tym skonsolidowanym
sprawozdaniu finansowym. Badanie skonsolidowanego sprawozdania finansowego przeprowadziliśmy
stosownie do postanowień rozdziału 7 ustawy o rachunkowości, krajowych standardów rewizji finansowej
wydanych przez Krajową Radę Biegłych Rewidentów w Polsce oraz Międzynarodowych Standardów
Rewizji Finansowej. Regulacje te nakładają na nas obowiązek postępowania zgodnego z zasadami etyki
oraz zaplanowania i przeprowadzenia badania w taki sposób, aby uzyskać racjonalną pewność, że
skonsolidowane sprawozdanie finansowe jest wolne od istotnych nieprawidłowości.
Badanie polega na przeprowadzeniu procedur mających na celu uzyskanie dowodów badania dotyczących
kwot i informacji ujawnionych w skonsolidowanym sprawozdaniu finansowym. Wybór procedur badania
zależy od naszego osądu, w tym oceny ryzyka wystąpienia istotnej nieprawidłowości skonsolidowanego
sprawozdania finansowego na skutek celowych działań lub błędów. Przeprowadzając ocenę tego ryzyka
bierzemy pod uwagę kontrolę wewnętrzną związaną ze sporządzeniem oraz rzetelną prezentacją
skonsolidowanego sprawozdania finansowego w celu zaplanowania stosownych do okoliczności procedur
badania, nie zaś w celu wyrażenia opinii na temat skuteczności działania kontroli wewnętrznej w jednostce.
Badanie obejmuje również ocenę odpowiedniości stosowanej polityki rachunkowości, zasadności
szacunków dokonanych przez Zarząd oraz ocenę ogólnej prezentacji skonsolidowanego sprawozdania
finansowego.
Wyrażamy przekonanie, że uzyskane przez nas dowody badania stanowią wystarczającą i odpowiednią
podstawę do wyrażenia przez nas opinii z badania.
Opinia
Naszym zdaniem, załączone skonsolidowane sprawozdanie finansowe Grupy Kapitałowej przedstawia
rzetelnie i jasno sytuację majątkową i finansową Grupy Kapitałowej na dzień 31 grudnia 2012 r., wynik
finansowy oraz przepływy pieniężne za rok obrotowy kończący się tego dnia, zostało sporządzone, we
wszystkich istotnych aspektach, zgodnie z Międzynarodowymi Standardami Sprawozdawczości
Finansowej, które zostały zatwierdzone przez Unię Europejską oraz jest zgodne z wpływającymi na treść
skonsolidowanego sprawozdania finansowego przepisami prawa obowiązującymi Grupę Kapitałową.
Inne kwestie
Dane porównawcze zostały przedstawione na podstawie skonsolidowanego sprawozdania finansowego
Grupy Kapitałowej za rok obrotowy kończący się 31 grudnia 2011 r., zbadanego przez inny podmiot
uprawniony do badania, który w dniu 6 marca 2012 r. wydał opinię bez zastrzeżeń o tym skonsolidowanym
sprawozdaniu finansowym.
AKTYWA
31.12.2012
%
aktywów
31.12.2011
%
aktywów
Aktywa trwale
Rzeczowe aktywa trwale
Nieruchomości inwestycyjne
Wartości niematerialne
Wartość finny
Inwestycje w jednostkach podporządkowanych
Inwestycje dostępne do sprzedaż)'
Pozostałe aktywa finansowe
Należności długoterminowe
Aktywa z tytułu odroczonego podatku dochodowego
Pozostałe aktywa
Aktywa trwałe razem
Aktywa obrotowe
Zapasy
Pozostałe aktywa finansowe
Należności z tytułu podatku dochodowego
Należności z tytułu dostaw i usług oraz pozostałe
Środki pieniężne i ich ekwiwalenty
Pozostałe aktywa
Aktywa trwałe przeznaczone do sprzedaży
Aktywa obrotowe razem
SUMA AKTYWÓW
PASYWA
Kapitał własny
Kapitał zakładowy
Kapitał z emisji akcji powyżej ich wartości nominalnej
Kapitał z aktualizacji wyceny
Kapitał z wyceny transakcji zabezpieczających
Różnice kursowe z przeliczenia sprawozdań finansowych
jednostek podporządkowanych
Zyski zatrzymane
Kapitał własny akcjonariuszy jednostki dominującej
Udziałowcy niesprawujący kontroli
31.12.2012
%
aktywów
31.12.2011
%
aktywów
Appendices
203
Kapitał własny razem
Zobowiązania
Zobowiązania z tytułu kredytów, pożyczek
Rezerwy na świadczenia pracownicze
Pozostałe zobowiązanie długoterminowa
Pozostałe rezerwy
Dotacje rządowe
Przychody przyszłych okresów
Rezerwy z tytułu odroczonego podatku dochodowego
Pozostałe zobowiązania finansowe
Zobowiązania długoterminowe razem
Zobowiązania z tytułu kredytów, pożyczek
Rezerwy na świadczenia pracownicze
Zobowiązania z tytułu podatku dochodowego
Zobowiązania z tytułu dostaw i usług oraz pozostałe
Pozostałe rezerwy
Dotacje rządowe
Przychody przyszłych okresów
Pozostałe zobowiązania finansowe
Zobowiązania krótkoterminowe razem
Zobowiązania razem
SUMA PASYWÓW
1.01.201231.12.2012
Przychody ze sprzedaży
Koszt własny sprzedaży'
Zysk brutto ze sprzedaży
Koszty sprzedaży
Koszty ogólnego zarządu
Pozostałe przychody operacyjne
Pozostałe koszty operacyjne
Zysk na działalności operacyjnej
Przychody finansowe
Koszty finansowe
Przychody (Koszty) finansowe netto
Zysk z udziałów w jednostkach stowarzyszonych
wycenianych metodą praw własności
Zysk przed opodatkowaniem
Podatek dochodowy
Zysk netto
SKŁADNIKI INNYCH CAŁKOWITYCH DOCHODÓW
Wycena instrumentów zabezpieczających
423
%
przychodów
ze sprzedaży
1.01.201131.12.2011
%
przychodów
ze sprzedaży
Rozliczenie instrumentów zabezpieczających
Zmiana wartości godziwej inwestycji dostępnych do
sprzedaży
Podatek odroczony od składników innych całkowitych
dochodów
Różnice kursowe z przeliczenia sprawozdań finansowych
jednostek podporządkowanych
Suma składników innych całkowitych dochodów
Całkowite dochody ogółem
Appendix 4: Student results per stage
S1-01
#1
#2
Stage I Phase I
#3 #4 #5 #6 #7 #8 #9 #10 #11 #12 #13 #14 #15
No
X
X
X
S1-02
#1
#2
#3 #4 #5 #6 #7 #8 #9 #10 #11 #12 #13 #14 #15
No
X
X
X
X
X X X
X X
X
X
X
X
X
Yes
X
X
X
X
Yes
S1-03
#1
No
X
#2
X X
X
X
X
X
X
#3 #4 #5 #6 #7 #8 #9 #10 #11 #12 #13 #14 #15
X
X
X
X
X
Yes
X
X
X X
X
X
X
X
X
X X
S1-04
#1
#2
#3 #4 #5 #6 #7 #8 #9 #10 #11 #12 #13 #14 #15
No
X
X
X
#1
#2
#3 #4 #5 #6 #7 #8 #9 #10 #11 #12 #13 #14 #15
X
X
X X X
X X
X
X
X
X
X
Yes
S1-05
X
No
X
X
X X X
X X
X
Yes
X
X
X
S1-06
#1
#2
#3 #4 #5 #6 #7 #8 #9 #10 #11 #12 #13 #14 #15
No
X
X
X
S1-07
#1
#2
#3 #4 #5 #6 #7 #8 #9 #10 #11 #12 #13 #14 #15
No
X
X
X
S1-08
#1
#2
#3 #4 #5 #6 #7 #8 #9 #10 #11 #12 #13 #14 #15
No
X
X
X
X X X
X X
X
X
X
X
X
X
X
X
Yes
X
X
X
X X
X
X
X
X
X
X X
Yes
X
X
X
X
X
Yes
X
X X
X
X
X
X
X
X
S1-09
#1
#2
#3 #4 #5 #6 #7 #8 #9 #10 #11 #12 #13 #14 #15
No
X
X
X
S1-10
#1
#2
#3 #4 #5 #6 #7 #8 #9 #10 #11 #12 #13 #14 #15
No
X
X
X
X X X
X X
X
X
X
X
X
Yes
X
X
X
X
Yes
X
X
X X
X
X
X
X
X
X
S1-11
#1
#2
#3 #4 #5 #6 #7 #8 #9 #10 #11 #12 #13 #14 #15
No
X
X
X
X
X
X X X
X X
X
X
X
X
X
Appendices
205
Yes
S1-12
#1
#2
#3 #4 #5 #6 #7 #8 #9 #10 #11 #12 #13 #14 #15
No
X
X
X
S1-13
#1
#2
#3 #4 #5 #6 #7 #8 #9 #10 #11 #12 #13 #14 #15
No
X
X
X
S1-14
#1
#2
#3 #4 #5 #6 #7 #8 #9 #10 #11 #12 #13 #14 #15
No
X
X
X
S1-15
#1
#2
#3 #4 #5 #6 #7 #8 #9 #10 #11 #12 #13 #14 #15
No
X
X
X
#1
#2
#3 #4 #5 #6 #7 #8 #9 #10 #11 #12 #13 #14 #15
X
X
X
X
X X X
X X
X
X
X
X
X
Yes
X
X
X
Yes
X
X X
X
X
X
X
X
X X
X
X X X
X X
X
X
X
X
X
Yes
X
X
X X X
X X
X
X
X
X
X
Yes
S1-16
No
Yes
X
X
X
X X X
X X
X
X
X
X
X
X
Stage I Phase II
#3 #4 #5 #6 #7 #8 #9 #10 #11 #12 #13 #14 #15
#1
#2
Yes
X
X
X
S1-02
#1
#2
#3 #4 #5 #6 #7 #8 #9 #10 #11 #12 #13 #14 #15
S1-01
X
No
X
No
Yes
X
S1-03
#1
X
X
X
#2
X X X
X X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
#3 #4 #5 #6 #7 #8 #9 #10 #11 #12 #13 #14 #15
X
No
X
X
X
X X X
X
X
X
X
X
X
Yes
X
X
X
S1-04
#1
#2
#3 #4 #5 #6 #7 #8 #9 #10 #11 #12 #13 #14 #15
No
X
X
X
X
X
X X X
X
X
X
X
X
X
X
Yes
#1
#2
#3 #4 #5 #6 #7 #8 #9 #10 #11 #12 #13 #14 #15
Yes
X
X
X
S1-06
#1
#2
#3 #4 #5 #6 #7 #8 #9 #10 #11 #12 #13 #14 #15
No
X
X
#1
#2
#3 #4 #5 #6 #7 #8 #9 #10 #11 #12 #13 #14 #15
X
X
S1-05
X
No
S1-07
X
X
Yes
No
Yes
X
S1-08
#1
#2
No
X
X
X
X
X
X X
X X X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X X
X
X
X
X
X
X
X
X
#3 #4 #5 #6 #7 #8 #9 #10 #11 #12 #13 #14 #15
X
X
Yes
X
X
X
X
X X
X
X
X
X
X
X X
S1-09
#1
#2
#3 #4 #5 #6 #7 #8 #9 #10 #11 #12 #13 #14 #15
No
X
X
X
X
X
X X X
X
Yes
S1-10
No
X
X
X
X
X
X
#1
#2
#3 #4 #5 #6 #7 #8 #9 #10 #11 #12 #13 #14 #15
X
X
X
X X X
X
X
X
X
X
X
Yes
X
X
X
S1-11
#1
#2
#3 #4 #5 #6 #7 #8 #9 #10 #11 #12 #13 #14 #15
No
X
X
X
S1-12
#1
#2
#3 #4 #5 #6 #7 #8 #9 #10 #11 #12 #13 #14 #15
No
X
X
X
#1
#2
#3 #4 #5 #6 #7 #8 #9 #10 #11 #12 #13 #14 #15
X
X
X X X
X X
X
X
X
X
X
Yes
X
X
X X X
X
X
X
X
X
X
Yes
S1-13
X
X
No
X
X
X X X
X
X
X
X
X
X
Yes
X
X
X
S1-14
#1
#2
#3 #4 #5 #6 #7 #8 #9 #10 #11 #12 #13 #14 #15
No
X
X
X
X
X
X X X
X X
X
X
X
X
Yes
S1-15
X
#1
No
#2
#3 #4 #5 #6 #7 #8 #9 #10 #11 #12 #13 #14 #15
X
X
Yes
X
S1-16
#1
#2
No
X
X
X
X
X X X
X
X
X
X
X
X
X
#3 #4 #5 #6 #7 #8 #9 #10 #11 #12 #13 #14 #15
X
X
Yes
X X
X
X X
X
X
X
X
X
X
S2-01
#1
#2
Stage II Phase I
#3 #4 #5 #6 #7 #8 #9 #10 #11 #12 #13 #14 #15
No
X
X
X
S2-02
#1
#2
#3 #4 #5 #6 #7 #8 #9 #10 #11 #12 #13 #14 #15
No
X
X
X
X X
X X
X
X
X
X
X
X
Yes
X
X
X
X X X
X X
X
X
X
X
X
X
Yes
S2-03
#1
#2
#3 #4 #5 #6 #7 #8 #9 #10 #11 #12 #13 #14 #15
No
X
X
X
X
X
X X
X X
X
X
X
X
X
X
Yes
#1
#2
Yes
X
X
S2-05
#1
#2
#3 #4 #5 #6 #7 #8 #9 #10 #11 #12 #13 #14 #15
No
X
X
X
#1
#2
#3 #4 #5 #6 #7 #8 #9 #10 #11 #12 #13 #14 #15
S2-04
#3 #4 #5 #6 #7 #8 #9 #10 #11 #12 #13 #14 #15
X
No
X
X
X X
X
X
X X
X
X
X
X
X
X
X X
X
X
X
X
X
X
Yes
S2-06
X
X
X
No
Yes
X
X
S2-07
#1
#2
X
X X
X X
X
X
X
X
X
X
#3 #4 #5 #6 #7 #8 #9 #10 #11 #12 #13 #14 #15
X
No
X
X
X
X X X
X X
X
X
X
X
X
Yes
X
X
S2-08
#1
#2
#3 #4 #5 #6 #7 #8 #9 #10 #11 #12 #13 #14 #15
No
X
X
X
S2-09
#1
#2
#3 #4 #5 #6 #7 #8 #9 #10 #11 #12 #13 #14 #15
No
X
X
X
X
X
X X X
X X
X
X
X
X
X
Yes
X
X
X X X
X X
X
X
X
X
X
Appendices
207
Yes
#1
#2
Yes
X
X
S2-11
#1
#2
#3 #4 #5 #6 #7 #8 #9 #10 #11 #12 #13 #14 #15
No
X
X
X
S2-12
#1
#2
#3 #4 #5 #6 #7 #8 #9 #10 #11 #12 #13 #14 #15
No
X
X
X
S2-10
#3 #4 #5 #6 #7 #8 #9 #10 #11 #12 #13 #14 #15
X
No
X X
X
X
X
X
X X
X
X
X
X
X
X
X X
X X
X
X
X
X
X
X
Yes
X
X
X X
X X
X
X
X
X
X
X
Yes
S2-13
#1
No
X
#2
#3 #4 #5 #6 #7 #8 #9 #10 #11 #12 #13 #14 #15
X
X
X
X X
X
Yes
X X
X
X
X
X
X
X
S2-14
#1
#2
#3 #4 #5 #6 #7 #8 #9 #10 #11 #12 #13 #14 #15
No
X
X
X
S2-15
#1
#2
#3 #4 #5 #6 #7 #8 #9 #10 #11 #12 #13 #14 #15
No
X
X
X
X X X
X X
X
X
X
X
X
Yes
X
X
X
X X
X X
X
X
X
X
X
X
Yes
S2-16
#1
#2
#3 #4 #5 #6 #7 #8 #9 #10 #11 #12 #13 #14 #15
No
X
X
X
#1
#2
Stage II Phase II
#3 #4 #5 #6 #7 #8 #9 #10 #11 #12 #13 #14 #15
X
X
X X X
X X
X
X
X
X
X
Yes
S2-01
X
No
Yes
X
S2-02
#1
X
X
X X X
X
X
#2
X
X
X
X
X
#3 #4 #5 #6 #7 #8 #9 #10 #11 #12 #13 #14 #15
X
No
X
X
X X
X
X X
X
X
S2-03
#1
#2
#3 #4 #5 #6 #7 #8 #9 #10 #11 #12 #13 #14 #15
Yes
X
X
X
S2-04
#1
#2
#3 #4 #5 #6 #7 #8 #9 #10 #11 #12 #13 #14 #15
No
X
X
S2-05
#1
#2
#3 #4 #5 #6 #7 #8 #9 #10 #11 #12 #13 #14 #15
No
X
X
X
X
No
X
X
Yes
X
X
X X
X
X
X
X
X
X
X X
X X X
X
X
X X
X
X
X
X
Yes
X
X
X X
X
X
X
X
X
X
X
X
X
X
X
Yes
S2-06
X
X
#1
#2
#3 #4 #5 #6 #7 #8 #9 #10 #11 #12 #13 #14 #15
X
No
X X
X
X
X
S2-07
#1
#2
#3 #4 #5 #6 #7 #8 #9 #10 #11 #12 #13 #14 #15
No
X
X
X
X
Yes
X
X
Yes
X
X
X
X X
X
X X
X
X
X
X
X
X
X
X
X
X
S2-08
#1
#2
#3 #4 #5 #6 #7 #8 #9 #10 #11 #12 #13 #14 #15
No
X
X
X
X
X
X X
X X
X
X
X
X
X
Yes
#1
#2
#3 #4 #5 #6 #7 #8 #9 #10 #11 #12 #13 #14 #15
Yes
X
X
X
S2-10
#1
#2
#3 #4 #5 #6 #7 #8 #9 #10 #11 #12 #13 #14 #15
S2-09
X
No
X
No
X
X
X
X
X X
X X
X X X
X
X
X
X
X
X
X
X
X
X
X
S2-11
#1
#2
#3 #4 #5 #6 #7 #8 #9 #10 #11 #12 #13 #14 #15
No
X
X
X
X
S2-12
X X X
X
X
#1
#2
X
X
Yes
Yes
X
X
X
X
X
X
X
#3 #4 #5 #6 #7 #8 #9 #10 #11 #12 #13 #14 #15
X
No
X
X
X
X
X
X
X
X
X
S2-13
#1
#2
#3 #4 #5 #6 #7 #8 #9 #10 #11 #12 #13 #14 #15
Yes
X
X
S2-14
#1
#2
#3 #4 #5 #6 #7 #8 #9 #10 #11 #12 #13 #14 #15
No
X
X
X
S2-15
#1
#2
#3 #4 #5 #6 #7 #8 #9 #10 #11 #12 #13 #14 #15
No
X
X
X
X
X
X
X X X
X X
X
X
X
X
X
#1
#2
X X
X
X
X
X
X
X
X
X
X
X
X
X
#3 #4 #5 #6 #7 #8 #9 #10 #11 #12 #13 #14 #15
X
X
X
X
X
No
X
X
X
X
X
Yes
Yes
X X
X
X
X
Yes
S2-16
X X
X
Yes
No
View publication stats
X
X
X
X
X
X
X
X
X
X
X
X
X