Vermes Albert

BASICS OF TRANSLATION THEORY

AND PRACTICE

EGER, 2021
 Vermes Albert

Basics of Translation Theory and Practice

The Conceptual Foundations of Translation

as Communication
 Vermes Albert

Basics of Translation Theory and Practice

The Conceptual Foundations of Translation as Communication

Az első kiadás (2009)

javított, elektronikus változata

Eger, 2021
A tankönyv megírását az Oktatási és Kulturális Minisztérium által nyújtott
0109/2006. nyilvántartási számú
Kiegészítő Posztdoktori Ösztöndíj támogatta.

Lektorálta:

Dr. Heltai Pál

egyetemi docens

ISBN 978-963-496-186-4

A kiadásért felelős
az Eszterházy Károly Egyetem rektora
Megjelent az EKE Líceum Kiadó gondozásában
Kiadóvezető: Nagy Andor
Felelős szerkesztő: Domonkosi Ágnes
Műszaki szerkesztő: Nagy Sándorné

Megjelent: 2021-ben, elektronikusan

 TABLE OF CONTENTS

Introduction ........................................................................................................9
1. Pages from the History of Translation .......................................................10
1.1. Why we need translation ........................................................................10
1.1.1. The story of Babel .......................................................................10
1.1.2. Which language was the proto-language? ...................................13
1.1.3. What was the Proto-World language like? ..................................14
1.2. The beginnings of translation theory ......................................................15
1.2.1. Cicero and Jerome .......................................................................16
1.2.2. Étienne Dolet ...............................................................................17
1.2.3. Dryden and Tytler ........................................................................18
1.3. Translation in Hungary...........................................................................18
1.3.1. Early translations in Hungarian ...................................................18
1.3.2. Thinking about translation in Hungary ........................................20
1.3.2.1. Cardinal Pázmány ...............................................................20
1.3.2.2. The 19th century...................................................................21
1.4. Study questions ......................................................................................22
2. The Present: Translation in the European Union .....................................24
2.1. The European Union ..............................................................................24
2.1.1. European institutions and bodies .................................................24
2.1.2. Languages in the EU....................................................................26
2.1.3. Language policy in the EU ..........................................................26
2.2. Interpretation in the EU ..........................................................................28
2.2.1. Conference interpreting ...............................................................28
2.2.1.1. Types of interpreting ...........................................................28
2.2.1.2. Language regime .................................................................29
2.2.1.3. The conference interpreter’s language combination ...........29
2.2.2. The Interpreting Directorate of the European Parliament ...........30
2.2.3. The Directorate for Interpretation of the European Court of
Justice ..........................................................................................31
2.2.4. The Directorate-General for Interpretation of the European
Commission.................................................................................31
2.3. Translation in the European Union ........................................................32
2.3.1. The Translation Directorate of the European Parliament ............33
2.3.2. The Translation Department of the Council ................................33
2.3.3. The Directorate-General for Translation of the European
Commission.................................................................................34
2.3.4. The Translation Directorate of the European Court of Justice ....35

5
 2.3.5. The Translation Division of the European Central Bank ............ 36
2.3.6. The Translation Centre for the Bodies of the European Union ... 36
2.3.7. The Interinstitutional Committee on Translation and
Interpretation ............................................................................... 37
2.4. Language technology ............................................................................. 38
2.4.1. Language technology at the DGT of the European Commission 39
2.5. Conclusions ............................................................................................ 39
2.6 Study questions ....................................................................................... 40
3. Basic Concepts and Problems of Translation ............................................ 41
3.1. General questions ................................................................................... 41
3.1.1. The problem of translatability: translation, language, culture ..... 41
3.1.2. Translation quality assessment .................................................... 44
3.1.3. Translation theory........................................................................ 45
3.1.3.1. What is translation? ............................................................. 46
3.1.3.2. The process of translation ................................................... 48
3.2. Equivalence-based approaches to translation......................................... 50
3.2.1. The concept of equivalence ......................................................... 50
3.2.2. Normative approaches ................................................................. 53
3.2.3. Descriptive approaches................................................................ 55
3.2.4. Problems with the notion of equivalence .................................... 56
3.3. Communicative-functional approaches to translation ............................ 57
3.3.1. Polysystem theory ....................................................................... 58
3.3.2. Skopos theory .............................................................................. 59
3.3.3. The theory of translatorial action ................................................ 60
3.3.4. Problems with the communicative-functional approaches .......... 61
3.4. Conclusions ............................................................................................ 61
3.5. Study questions ...................................................................................... 62
4. The Basics of Communication..................................................................... 64
4.1. The translator’s mind ............................................................................. 64
4.1.1. The case of a linguistic savant ..................................................... 64
4.1.2. Smith and Tsimpli’s theory of the mind ...................................... 65
4.1.3. Christopher’s translation problems and the mind ........................ 68
4.2. Communication: basic definitions.......................................................... 68
4.3. Coded and ostensive-inferential communication ................................... 74
4.3.1. Coded communication ................................................................. 74
4.3.2. Ostensive-inferential communication.......................................... 75
4.3.2.1. Ostension............................................................................. 76
4.3.2.2. Inference ............................................................................. 77
4.3.2.3. Relevance ............................................................................ 79
4.3.2.4. Context ................................................................................ 81
4.4. Linguistic communication...................................................................... 82

6
 4.4.1. Language and inferential communication ...................................82
4.4.2. Descriptive and interpretive language use ...................................84
4.5. Study questions ......................................................................................86
5. Translation as Interpretation ......................................................................87
5.1. Translation as interpretation: a first approximation ...............................87
5.2. Interpretation ..........................................................................................88
5.3. Translation as interlingual interpretive use ............................................91
5.3.1. Direct and indirect translation .....................................................93
5.3.2. Conditions and corollaries of direct translation .........................100
5.4. Conclusions ..........................................................................................100
5.5. Study questions ....................................................................................101
6. Culture and Translation ............................................................................102
6.1. Introduction ..........................................................................................102
6.2. Culture and culture-specific expressions ..............................................102
6.3. Translation operations and translation strategies .................................104
6.3.1. Translation operations ...............................................................104
6.3.2. Translation strategies .................................................................107
6.4. The use of the four operations in implementing strategies...................108
6.4.1. Total transfer .............................................................................108
6.4.2. Logical transfer ..........................................................................109
6.4.3. Encyclopaedic transfer ..............................................................110
6.4.4. Zero transfer ..............................................................................112
6.5. Conclusions ..........................................................................................113
6.6. Study questions ....................................................................................113
7. Translation Competence: What Makes a Good Translator ...................114
7.1. The minimalist concept of translation competence ..............................114
7.2. Complex translation competence .........................................................115
7.2.1. Linguistic competence ...............................................................115
7.2.2. Subject-related competence .......................................................116
7.2.3. Intercultural competence ...........................................................120
7.2.4. Transfer competence..................................................................120
7.2.5. Communicative competence......................................................121
7.2.6. Knowledge of translation theory ...............................................122
7.3. The final question .................................................................................127
7.3.1. Instinct .......................................................................................127
7.3.2. Experience .................................................................................128
7.3.3. Knowledge .................................................................................129
7.4. Study questions ....................................................................................130
References .......................................................................................................131

7
 INTRODUCTION

This textbook is meant primarily for students of English at the undergraduate

level who study translation as part of their degree programme. It is designed to
complement their proficiency as language professionals specialised in the
English language with an indispensable practical skill that they will inevitably
need in their professional life.
The book contains material for a one-semester lecture course, the aim of
which is to acquaint students with the basic concepts, problems and principles of
translation, which is seen as a special interlingual and intercultural form of
communication. After a brief overview in the first two chapters of the past and
present of translation as a practical activity, the rest of the book presents the
fundamental characteristics of linguistic communication and of translation as an
interpretive form of communication across languages and cultures, and
introduces the basic categories and terminology of translation theory with an eye
on how these can be utilised in the practice of translation work.
Readers of this book are assumed to have a fairly advanced command of the
English language as well as some familiarity with the elementary notions of
linguistics. The theoretical parts of this textbook will no doubt prove challenging
for anyone just getting acquainted with the study of translation, but hopefully the
conceptual foundations laid down here will serve to aid readers in developing a
deeper understanding of the very practical issues of translation work.

9
 1. PAGES FROM THE HISTORY OF TRANSLATION

1.1. Why we need translation

Of course, we need translation because people speak different languages and

they do not understand each other. So they employ linguists called translators
who help them to overcome the problems caused by the differences between
languages. In another sense, however, the question posed in the section title can
be read in this way: Why do people speak different languages? One fascinating
answer is found in the Bible, in the story of Babel.

1.1.1. The story of Babel

Figure 1. The Tower of Babel by Pieter Brueghel the Elder, 1563

Figure 1 depicts the famous tower from the biblical story of Babel, which has
become a myth of the origin of translation (Robinson 1998: 21). The story is
found in Genesis 11:1–9 of the Old Testament and reads as follows (New
International Version, published by The Committee on Bible Translation, 1978).

10
1 Now the whole world had one language and a common speech.
2 As men moved eastward, they found a plain in Shinar and settled
there.
3 They said to each other, “Come, let’s make bricks and bake them
thoroughly.” They used brick instead of stone, and tar for mortar.
4 Then they said, “Come, let us build ourselves a city, with a tower that
reaches to the heavens, so that we may make a name for ourselves
and not be scattered over the face of the whole earth.”
5 But the Lord came down to see the city and the tower that the men
were building.
6 The Lord said, “If as one people speaking the same language they have
begun to do this, then nothing they plan to do will be impossible for
them.
7 Come, let us go down and confuse their language so they will not
understand each other.”
8 So the Lord scattered them from there over all the earth, and they
stopped building the city.
9 That is why it was called Babel – because there the Lord confused the
language of the whole world. From there the Lord scattered them
over the face of the whole earth.

Figure 2. Ziggurat remains at Ur

11
Such towers were in fact erected in historical Mesopotamia and are known as
ziggurats. Mesopotamian ziggurats were staged temple towers that were built to
lift worshippers closer to their gods. The biblical ziggurat is believed to have
been situated in the ancient city of Babylon.
The story of the Tower of Babel is a symbol of human sin. The building of
the tower is seen as humankind’s rebellion against God by attempting to ascend
to heaven. According to the Bible, God punished humanity for this act by
confusing their language. This is where the city’s name comes from: Babel
means confusion.

Figure 3. Confusion of the Tongues by Gustave Doré (1866)

12
The confusion of tongues (confusio linguarum) is the fragmentation of human
language after the collapse of the Tower of Babel. Thus the story implies that
originally all people spoke one common language, which was then confused and
broken up by God as a punishment. The proto-language is the assumed
common source from which present-day languages are derived. This ancient
common language, or proto-language, was lost as a result of human sin and
people fell into linguistic diversity so that they may not understand each other
any longer.
What the story relates, then, is how the original linguistic unity of mankind
was turned into linguistic diversity, a situation that prevents people from
understanding each other, because unity threatened God’s hegemony (Robinson
1998: 22). But then this also implies that God did not intend to allow people to
communicate by means of translation, either. Then the translator, who works to
restore unity and understanding, must be seen in this context as an agent of
renewed human rebellion against the divine plan. We may wonder whether this
will earn people another punishment in the future...

1.1.2. Which language was the proto-language?

The human proto-language spoken prior to the events described in the story was
assumed to have split into seventy or seventy-two dialects, depending on
tradition. During the Middle Ages, Hebrew was widely considered to be this
language, used by God to address Adam in Paradise, and used by Adam as
nomothete (name-giver, Genesis 2:19) to give things names, especially by
Christian scholastics. Therefore this hypothetical proto-language is also
sometimes referred to as the Adamic language.
History has preserved a number of anecdotes about people who attempted to
discover the origin of language by experiment. The first such tale was told by
Herodotus, who relates that Pharaoh Psamtik (probably Psammetichus I) had
two children raised by deaf-mutes. He wanted to see what language the children
ended up speaking in the company of their speechless guardians. When the
children were brought before him, one of them said something that sounded to
the pharaoh like bekos, the Phrygian word for bread. From this, Psamtik
concluded that Phrygian was the first language.
According to another story, King James V of Scotland also tried a similar
experiment. His children were supposed to have ended up speaking not Phrygian
but Hebrew.
Frederick II of Prussia and Akbar, a 16th century Mughal emperor of India are
also said to have tried a similar experiment. However, the children that they tried
these experiments with did not speak. (Of course, this is what we would expect
today in light of the psycholinguistic data on child language acquisition that
have been gathered in the past decades.)

13
 During the Renaissance, Hebrew had lost its status as the original language of
Paradise, and was considered just one of the seventy languages derived from the
confusion of tongues. Priority was now claimed for the alleged Japhetite
languages (see Genesis 10:5), which were never corrupted because their
speakers had not participated in the construction of the Tower of Babel. Among
the candidates for a living descendant of the Adamic language were Gaelic,
Tuscan, Flemish, Swedish and German.
Later linguists realised that they had been asking the wrong question about
the original language. The right question is not which language was the common
ancestor but what it was like. This realisation led to a systematic comparison of
modern and ancient languages based on phonological, lexical and syntactic
correspondences, which in the 19th century resulted in the proposal for a
supposed common ancestor of most European and some other languages: Proto-
Indo-European. By analogy, the term Proto-World language refers to the
hypothetical latest common ancestor of all the world’s languages, an ancient
language from which all modern languages and language families – and usually
including all known dead languages – derive.

1.1.3. What was the Proto-World language like?

The modern concept corresponding to the Adamic language is that of the Proto-
World language, but rather than positing divine inspiration, historical linguists
assume that it arose from proto-linguistic forms of communication. The origin of
language (glottogony, glossogeny) is a topic that has been written about for
centuries, but because of the ephemeral nature of speech there is almost no data
on which to base conclusions on the subject. We know that, at least once during
human evolution, a system of verbal communication emerged from proto-
linguistic or non-linguistic means of communication, but beyond that little can
be said. No present-day human group, anywhere, speaks a “primitive” or
rudimentary language. While existing languages may differ in the size and
content of their lexicons, all human languages possess the grammar needed for
effective communication, and can translate, invent, or borrow from other
languages the vocabulary needed to express the full range of what any speaker
may want to communicate.
The Proto-World language would have been spoken roughly 200,000 years
ago, the time suggested by archaeogenetics for the phylogenetic separation of
the ancestors of all humans alive today, mainly by analysis of mitochondrial
DNA. Proto-World would not necessarily be the first language spoken, but only
the latest common ancestor of all languages known today, and it already may
have gone through a long evolution, and even may have existed alongside other
languages of which no trace survived into historical times. For example, it is
disputed whether or not Homo neanderthalensis had the faculty of speech. If

14
they did, their language in all probability would not have been derived from
Proto-World as defined above.
The above arguments make assumptions about a Proto-World language based
on anthropology, human migration patterns and the assumed faculty of speech in
prehistoric humans. Direct linguistic statements about the Proto-World language,
however, due to the time depth involved, are reckoned impossible by most
historical linguists by any available method. For this reason, proposals for
attributes of Proto-World are considered to be on the fringe of linguistic studies,
so much so that in 1886 the Linguistic Society of Paris even banned discussion
of the origin of language, considering it to be a question to which no scientific
answer could be given. Many people also question the theory underlying the
whole research enterprise, referred to as the theory of monogenesis. The theory
of monogenesis is the assumption that all known languages derive from a
common ancestor.
They suggest that language may have developed independently in different
groups of early humans from proto-linguistic means of communication, thereby
disputing the existence of a Proto-World language. This debate is essentially
about the definition of the term language, and about whether the system of
communication employed by human beings at the time of Mitochondrial Eve,
the assumed common female ancestor of humankind, qualified as a language in
the narrow sense.

1.2. The beginnings of translation theory

There existed no systematic theories of translation before the end of the 19th
century. Thoughts about matters of translation were mainly published in
translators’ forewords or appendices to their translations, in which they
explained to the reader the problems and solutions that arose during the
translation of the given work of art (Tarnóczi 1966: 15). However, these early
ideas lacked the kind of systematicity and consistency that is expected from
what we can call a theory in the modern sense.
Until the middle of the 16th century, there were not even separate words for
the written and oral modes of translation. The German term Dolmetschen,
French interprète, Italian traduttore were all used to refer to practitioners of both
kinds of activity. In French, for example, the use of the expression traducteur for
translators of the written word was suggested only as late as 1539 (Tarnóczi
1966: 15).
All this is not to say, however, that people made no attempt whatsoever at
drawing such distinctions or summing up their ideas about translation in earlier
times. As a matter of fact, thinking about translation has a very long tradition.
The earliest such attempts that we know of date from antiquity.

15
 1.2.1. Cicero and Jerome

Cicero, in the 1st century BC, was the first European to summarise the principles
of translation in his De optimo genere oratorum (46 BC). In this work, he
expresses his objection to translating a text in a word-for-word (verbum pro
verbo) manner because he thinks it is more important to be faithful to the sense
than to the words and thus he advocates a form of free translation. Additionally,
he likens his job of translating the speeches of Greek orators to that of the orator
rather than to that of the interpreter. He writes: “I did not translate them as an
interpreter but as an orator… 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” (Cicero 1949: 365, cited in Venuti 2004: 14).
This opposition seems striking in the sense that it contrasts two activities
which apparently belong to completely different categories: the interpreter plays
the role of mediator by conveying ideas that already have been formulated in
verbal form, whereas the orator creates something new by putting into words
ideas that have not been previously manifested in words (Hell 2004: 28). What is
it, then, that makes translation in Cicero’s eyes a rhetorical activity different
from oral mediation?
One difference is the medium of the activity: translation takes place in
writing, interpreting orally. The second is that interpreting has a very limited
time-frame since it normally happens simultaneously with, or right after, the
production of the original text and thus the interpreter has no previous
knowledge of the text interpreted. On the other hand, translation is less linked to
the original in time and the translator normally prepares the translation after the
original has been read and understood fully. As a matter of fact, as Hell notes
(2004: 29), someone who translates a text sentence by sentence without
previously having studied the whole text is more of an interpreter than a
translator in the true sense of the word. Thirdly, Cicero believed that while the
interpreter’s job is to reproduce the original, the translator, like the orator,
produces something new, though on the basis of another text, and is, in fact, the
author of the new text, not a mere mediator. This way, his theory of translation
provides justification for his translational practice by placing translation in
between the two extremes of interpreting and oration.
Cicero’s ideas found a disciple in St. Jerome, who translated the Bible into
Latin (known as the Vulgata) in the 4th century AD, using the original Hebrew
text of the Old Testament. In his Letter LVII entitled De optimo genere
interpretandi (notice the similarity of the title with that of Cicero’s work referred
to above) he writes: “Non verbum e verbo, sed sensum exprimere de sensu,” that
is: “I render not word for word, but sense for sense” (Jerome 2004: 23). In this
letter he argued at length that while the Evangelists’ interpretation and the
Septuagint (the Greek version of the Old Testament) were amazingly different

16
“in both words and syntax” (p. 26), they were both justifiable “so long as the
truth lay open to understanding” (p. 28).
While such a theory properly underscores the creative aspect of translation, it
also raises problems, as we will see later in Chapter 3, in devising criteria for
evaluating the quality of translations. Basically, the source of this problem is the
assumption that the same sense, or content, can be expressed in different ways,
in different linguistic forms, even in different languages. If so, however, what is
the explanation for the fact that no two translators in the same language will ever
produce identical translations of the same text? Is it because both are possible
paraphrases of the same content or because one or the other is not a perfect
translation? And if so, how do we find the “perfect translation” and how do we
evaluate the relative merits of the different translations? These questions will be
picked up in more detail in Chapter 5.

1.2.2. Étienne Dolet

In France, Étienne Dolet of Lyon published a brief essay on translation in 1540

under the title La manière de bien traduire d’une langue en aultre. He offered
five basic rules of translation. (1) The translator has to be absolutely clear about
the topic and content of the text to be translated. (2) The translator has to have a
perfect command of the language he or she is translating from, and also of the
language he or she is translating into. (3) The translator must not translate in a
word-for-word manner. Word-for-word translation is proof of the the translator’s
imperfection and deficiency of mental capacities. The translator must not follow
the word order of the original but should take care of its sense to express the
author’s intention. (4) The translator should avoid using less-known forms of
expression in the translation. (5) The translation should be formulated in clear,
elegant language, paying attention to the appropriate linking of words (cited in
Tarnóczi 1966: 14–15).
In Dolet’s thinking, it is apparent from his rules 3 and 5 that the quality of the
translation depends to a great degree on how it manages to convey the sense of
the original and also on how well it is written in terms of the style of its
formulation. But this is not what makes his contribution really significant
compared with Cicero, who by likening the translator to the orator, also
emphasised, if not explicitly, the importance of the style of the translation.
Dolet’s originality is revealed by the ideas that he brings into focus in rules 1,
2 and 4. In the first rule he says that the translator needs to be clear about the
topic of the original, which suggests that he recognised the importance of having
appropriate background knowledge in making sense of a text. In the second rule
he highlights the necessity of having perfect knowledge of both the source and
the target language. Even today people tend to forget that an ability to do good
translation presupposes that (a) one is a good listener/reader in the source

17
language and also that (b) one is a good communicator in the target language,
which is not evidently so even in the case of a native speaker. Finally, the fourth
rule emphasises the importance of using the kind of language in the translation
that the reader can readily understand. This is obviously a reaction to the general
practice of Dolet’s contemporaries of producing very precise verbatim
translations, which often borrowed foreign expressions from the original texts.

1.2.3. Dryden and Tytler

At the end of the 17th century, John Dryden emphasised that in his translations of
Ovid’s Epistles and The Aeneid he intended to enrich his own English language
(Tarnóczi 1966: 16). He thus considered translation not only as a way of
transferring ideas but also as a means of expanding the expressive potentials of a
language.
A century later, in 1791, Alexander Fraser Tytler published his Essay on the
Principles of Translation, in which he listed three main requirements for a good
translation. (1) The translation must be a perfect rendering of the original’s
ideas. (2) The manner and style of the translation must reflect the character of
the original text. And (3) the translation must have all the ease of an original
piece of writing (Tytler 1791, cited in Tarnóczi 1966: 16). We have already seen
that of these requirements the first and the third had been voiced earlier by
Cicero or Dolet, for instance. Tytler’s novelty appears in the second
requirement, where he clearly points out that the translation should also respect
the original’s style. In other words: not only the content of the original but also
the manner in which that content was formulated should be respected.
To summarise, from Cicero on, we can discern a clear tendency for
translators to lay emphasis not only on the status of the translation as a replica of
the original ideas but also on the fact that the translation, at the same time, needs
to be linguistically and stylistically on a par with the original piece of writing.

1.3. Translation in Hungary

1.3.1. Early translations in Hungarian

In the Middle Ages, Latin was the official language of the Kingdom of Hungary.
All official documents, laws, decrees and orders, notices and letters were written
in Latin. (There were even poets, like Janus Pannonius in the 15th century, who
only wrote in Latin.) This of course was very useful in view of the country’s
international relations since at the time Latin was the lingua franca in Europe,
fulfilling the function that English has today. On the other hand, this situation
greatly hindered the development of a literary form of the Hungarian language.

18
 In Hungary, the beginnings of the written literary tradition were closely
linked with translation. Practically all the early texts in the Hungarian language
from the 11th to the 16th century were more or less faithful translations of Latin
originals (Radó 1998: 449). Probably the oldest known literary text in
Hungarian, entitled Halotti beszéd (Funeral Oration), written around 1195, is a
free rhythmic prose translation of the Latin source. Translation remained the
dominant form of literature in Hungarian up to the 16th century.
The most important source text in translation, no doubt, just as in other parts
of Europe, was the Bible. As Woodworth (1998: 102) notes, it was in a class of
its own in this function. Its importance in translation is based on a paradox: it is
a pivotal text in European culture, but it was written in languages that very few
people can understand. The Old Testament was written mainly in Hebrew, with
some parts in Aramaic, while the New Testament was written in Greek.
The first Hungarian translations of parts of the Bible appeared in the 15th
century. The first complete translation of St. Jerome’s Latin Vulgata (a
translation of the Bible from the Hebrew original) was produced by Gáspár
Károli, a Protestant preacher, in 1590, followed in 1626 by a Catholic translation
written by György Káldi.
Translations of other religious texts can be found from as early as the 15 th
century, including the Érdy Codex, from 1527, a collection of Hungarian legends
written for nuns.
The 16th century also brought about the first major wave of non-religious
translations into Hungarian. Radó (1998: 449) mentions the following as
outstanding examples of literary translations from this period. Aesop’s Fables
were translated into Hungarian by Gábor Pesti in 1536 and later by Gáspár
Heltai in 1566. Sophocles’ Electra was translated in 1558 by Péter Bornemissza
in the form of an adaptation, just as Castelleti’s Amarilli from Italian in 1588 and
George Buchanan’s Jephte from a Latin version in 1589 by Bálint Balassi.
In this case, as in many other cases including early translations of the Bible,
the Hungarian translation was based not on the original text but on a translation
of it in Latin or in some other language, which served as an intermediary source.
The language of this intermediary source is called a relay language. This kind
of translation occurs when either the original text is unavailable or when the
translator does not speak the language of the original and needs to rely on a
version produced in a relay language. For example, the first Hungarian
translation of Shakespear’s Hamlet in 1790 by Ferenc Kazinczy was also done
from German as a relay language, not English.
Until the 1780s, as we have seen, Latin served as the official language of
administration in Hungary, just as in other Christian countries throughout
Europe, and thus translations were mostly done from texts in Latin. After the
expulsion of the Turks from Hungary, Latin as an official language was replaced
by German from the end of the 18th century until 1867, when Hungary gained

19
internal self-government within the Austro-Hungarian Monarchy and Hungarian
was also established as an official language of the country. Consequently, in this
period German replaced Latin as the primary source language from which
translations into Hungarian were produced. This was also the time when the
translation of official documents on a larger scale began, followed by the
translation of specialised texts in the 19th century (Radó 1998: 450).
The end of the 19th century saw the birth and flourishing of journals in
various specialised fields, including natural sciences, medicine, history and
sociology, which published many articles translated from foreign languages.

1.3.2. Thinking about translation in Hungary

The first written record of the Hungarian word for translation, fordítás, is in the
Érdy Codex, which comes from 1527. (Before this word occurred, Hungarians
used the word magyaráz to describe the activity.) It will be interesting to start
this overview by comparing the Hungarian word to its equivalents in some other
European languages.
The word fordítani literally means to turn something over. For instance, we
can turn over a page to see what is on the other side. Its use for translation, then,
is a metaphorical extension of the original meaning. It seems to suggest that we
need to turn over a page of the original text in order to see what is behind the
words.
The English word translation comes from Latin translatio, which originally
means ‘transfer’. The Latin word itself was based on the Greek word
metapherein, having the same meaning. (In present-day English, the original
Greek word is preserved in the word metaphor, which means an expression with
a transferred meaning.) Thus all these three words metaphorically mean to
transfer, or to carry over, something from one place to another. In the case of the
transfer of texts, the thing to be carried over, of course, is the meaning of the
text.
The difference between the Hungarian metaphor and the Greek-Latin-English
metaphor for translation might suggest that thinking about translation in
Hungary followed a path different from that in the rest of Europe, but we will
see that in fact it was very much in line with ideas formulated elsewhere in
Europe.
Thinking about translation in Hungary, just as in other parts of Europe, began
in the form of translators’ reflections on their job in forewords to their published
works. The first of these writings appeared in the 17th century.

1.3.2.1. Cardinal Pázmány

Well before Tytler published his Essay, in 1604 Péter Pázmány wrote in a
foreword to his translation of De imitatione Christi by Thomas Kempis that his

20
intention was to produce a text which would read as one that was originally
written in Hungarian by a Hungarian person (“ne láttatnék deákból csigázott
homályossággal repedezettnek, hanem oly kedvesen folyna, mintha először
magyar embertül magyarul íratott volna”) (cited in Taróczi 1966: 259). In other
words, Pázmány, just as Dolet and Tytler, was an advocate of translations that
sound natural in the target language. This mode of translation may be called
target language-oriented.
Target language-oriented translation is when the target text is formulated
linguistically in a way so that it conforms to the norms and conventions of use of
the target language, rather than to those of the source language.

1.3.2.2. The 19th century

In a foreword to his translation of Sallustius, a 4th-century writer and friend of
the Emperor Julian the Apostate, Ferenc Kazinczy pointed out that “the way the
translation needs to be done is determined by the nature of the work to be
translated, the purpose for which we translate, and the readers for whom we
translate” (“mint kelljen fordítani, a’ Munka neme határozza meg, mellyet
fordítunk, a’ czél, mellyért fordítunk, az olvasók kiknek fordítunk”) (cited in
Taróczi 1966: 17–18, my translation).
In this statement, Kazinczy voices very important ideas. He recognises that
the manner of translation is not universally predetermined but is relative to the
genre and function of the original, the purpose that the translation is to fulfil, and
the needs of the target reader. This holy trinity of textual function, purpose of the
translation and intended target reader is still a pivotal issue in modern translation
theories.
The importance of taking into account the reader of the translation was also
emphasised later by Károly Szász, who in his 1859 inaugural speech in the
Hungarian Academy said the translation “must evoke exactly the same, no less
and no other, feeling of pleasure in the reader as he would feel by reading the
original, were its language his native tongue” (“teljesleg ugyanazon, azaz nem
kisebb és nem más gyönyörérzetet kell költenie az olvasóban, mint amit az
eredeti olvasásánál érezne, ha az nemzeti nyelve volna”) (cited in Bart 1981:
258, my translation). This idea would later turn up repeatedly in various forms in
translation theory, which shows that Szász managed to touch on a central
question here.
In the middle of the 19th century, Ferenc Toldy, father of Hungarian literary
history, in his study entitled A műfordítás elveiről (On the principles of literary
translation) distinguished between two kinds of fidelity in translation: fidelity to
the content of the original and fidelity to the form of the original. However, in
contrast with Tytler, he denied the possibility of being faithful to content and
form at the same time. This also came to be a much debated issue in the 20 th

21
century, having serious repercussions in the work of the famous poet-translators
Mihály Babits, Dezső Kosztolányi and Árpád Tóth.
In 1860–1861, Sámuel Brassai published his views on translation in a series
of eleven articles under the title Mégis valami a fordításról (Yet something
about translation) on the pages of the journal Szépirodalmi Figyelő (Literary
Observer), edited by János Arany. He called attention to the futility of
translating word-for-word by warning: “Instead of vainly struggling, grappling
and wrestling with words, which mostly ends in failure, turn your attention
towards the sentence” (“A szókkal való hasztalan vesződség, küzdés, birkózás
helyett, minek legtöbbnyire kudarcz a vége, fordítsák figyelmöket a mondatra”)
(cited in Tarnóczi 1966: 19, my translation). Here appears another important
question of modern translation theory: the question of what is the size of the
appropriate translation unit. The unit of translation is the stretch of source
language text that the translator focuses attention on at a given moment.
The same idea was underlined, in 1874, in a study called A helyes fordításról
(On appropriate translation) in the journal Magyar Nyelvőr (Hungarian
Language Guard), by Emil Ponori Thewrewk, who wrote: “Strict faithfulness
consists not in translating the word, but in translating the sentence, and not the
external form of the sentence but the sense contained therein” (“A szigorú hűség
nem abban áll, hogy a szót fordítsuk, hanem a mondatot, annak is nem külső
formáját, hanem értelembeli foglalatját”) (cited in Taróczi 1966: 19, my
translation).
Up until this point, there was no dividing line in the thinking of translators
about translation between literary and non-literary forms of translation. Literary
translation, as Tarnóczi (1966: 19) notes, had no special theoretical status and
was simply considered to be one genre of translation. From the end of the 19 th
century on, however, the focus started to shift towards questions of literary
translation, pushing issues of non-literary translation into the background. This
state of affairs remained unchanged right up until the middle of the 20th century,
when social and political changes brought about an increased recognition and
awareness of non-literary forms of translation.

1.4. Study questions

After reading the chapter, answer the following questions. If you cannot
remember, go back and revise the related part.

What is the best-known myth about the origin of translation? What is implied
by this story and what is the irony of the story in connection with translation?
What alternative theories exist concerning the development of human
language?

22
 In what sense is the translator like an orator, rather than an interpreter,
according to Cicero? What ideas did Jerome have about translation? What
problem is raised by Cicero’s and Jerome’s approach to translation?
What novelty appears in Dolet’s rules as compared to Cicero?
What was Dryden’s and Tytler’s contribution to translation theory?
What was the significance of translation in mediaeval Hungary? What
functions did Latin have in mediaeval Hungary?
What were some of the most important milestones of religious translation in
Hungary?
What were some of the most important milestones of non-religious
translation in Hungary?
What new ideas did Cardinal Pázmány bring into translation and what new
ideas about translation occurred in 19th-century Hungary?
In what ways did Hungarian scholars revise issues that are still important in
modern translation theory?

23
2. THE PRESENT: TRANSLATION IN THE EUROPEAN UNION

After a brief survey in the previous chapter of the history of translation in the
western world, it is in order to take a look at translation as it is practiced today.
The primary aim of this chapter is to describe what happens in the European
Union in terms of communication across languages and to show how the
circumstances, the role and the professional status of the translator have changed
as a result of social, political and technological development. All the relevant
information about these issues, summarised in this chapter, can be found on the
various web pages of the EU relating to language matters. (You can start
browsing at https://ec.europa.eu/info/index_en.)

2.1. The European Union

At the time of writing this book in 2009, the European Union (EU) has 27
member states. It was established in 1993 by the Maastricht Treaty and is the
successor to the six-member European Economic Community (EEC), which was
founded in 1957 by the Treaty of Rome. The EU can be described as both a
supranational and an intergovernmental body.
The combined population of the 27 member countries was approximately
492.9 million in 2006. With this number of citizens, which makes up 7.3% of the
world’s population, living on just 3% of the land surface of the Earth, the EU is
one of the most densely populated regions of the world.

2.1.1. European institutions and bodies

The EU is governed by a number of institutions, the most important of which are

the European Commission, the Council of the European Union, and the
European Parliament. The European Commission (whose official name is the
Commission of the European Communities) acts as the executive branch. It is
currently composed of 27 commissioners, one from each member state. The
Commission is responsible for drafting all proposed rules and regulations in
order to co-ordinate European Law. It also controls some agencies and is
responsible for the day-to-day running of the Union. The president of the
Commission is nominated by the Council, and confirmed by the Parliament.
The Council of the European Union, also known as the Council of Ministers,
forms one half of the EU’s legislative branch, the other being the Parliament. It
should not be confused with the European Council, which is composed of the
heads of government and the President of the European Commission. The
Council of the EU is composed of the national ministers responsible for the area

24
of EU law being addressed. For example, European legislation regarding
agriculture would be treated by a Council composed of the national ministers for
agriculture. The presidency of the Council rotates between the member states
every six months.
The other half of the legislative branch is the European Parliament, which is
the only Union body composed of representatives directly elected by the citizens
of the EU member states. Every five years citizens of the member states elect
785 Members of the European Parliament (MEPs). Although they are elected as
national representatives, the members of the Parliament have their seats arranged
according to political ideals rather than nationality. The chair of the Parliament
is elected by its members. The European Parliament holds monthly plenary
sessions of one week in Strasbourg. Additional plenary sessions may be held in
Brussels. Committees and political groups meet in Brussels and elsewhere in
Europe. There are also meetings of parliamentary delegations for international
relations.
The judicial branch of the EU consists primarily of the European Court of
Justice (officialy called the Court of Justice of the European Communities). It is
composed of 27 judges, one judge being nominated by each member state. The
president is elected from among the judges. Below the Court of Justice there is a
lower court called the Court of First Instance, which was created to reduce the
work load of the Court of Justice. It deals with cases brought by individuals or
companies. There is also the European Court of Auditors, which monitors the
EU’s accounts.
Other bodies include the European Central Bank and the European
Investment Bank, which finance projects that contribute to the economic
development of the Union. The European Central Bank (ECB), established in
1998, controls monetary policy within the euro zone, which includes the member
states that have adopted the euro as their currency. The working language of the
ECB is English, although it is located in Frankfurt, Germany. There are also two
advisory committees: the Committee of the Regions (CoR) and the European
Economic and Social Committee (EESC). The EESC is a consultative assembly
concerned with matters such as social policy, social and economic cohesion,
environment, education, health, customer protection, or industry. The CoR is the
EU’s assembly representing local and regional authorities in issues concerning
local or regional policies.
There is no official capital city of the EU. The locations of the institutions are
spread across many cities and several member states. However, Brussels is often
considered to be the actual capital as most of the primary institutions are situated
there, including the Commission and the Council. The Parliament is also partly
based in this city but has its official seat in Strasbourg. Luxembourg city hosts
all the EU’s law courts and a number of other departments and bodies. The ECB
has its headquarters in Frankfurt, Germany.

25
 2.1.2. Languages in the EU

The EU has 23 official and working languages: Bulgarian, Czech, Danish,

Dutch, English, Estonian, Finnish, French, German, Greek, Hungarian, Italian,
Irish, Latvian, Lithuanian, Maltese, Polish, Portuguese, Romanian, Slovak,
Slovenian, Spanish and Swedish. German is the most widely spoken mother
tongue with about 90 million native speakers, followed by English, French, and
Italian. English is the foreign language spoken by most people, followed by
German and French. 56% of EU citizens are able to hold a conversation in a
language other than their mother tongue. Most official languages of the EU
belong to the Indo-European language family, except Estonian, Finnish, and
Hungarian, which belong to the Uralic language family, and Maltese, which is a
Semitic language. EU official languages are written in the Latin alphabet except
Bulgarian, written in Cyrillic, and Greek, written in the Greek alphabet.
There are also about 150 regional and minority languages, spoken by up to 50
million people. Of these, only the Spanish regional languages, Catalan, Galician,
and Basque, can be used by citizens to communicate with the Council of the
European Union, the Commission, the Economic and Social Committee, the
European Parliament and the European Ombudsman, as well as in the workings
of the Committee of the Regions. Though regional and minority languages can
benefit from EU programmes, protection of linguistic rights is a matter for the
individual member states. A wide variety of languages from other parts of the
world are spoken by immigrant communities in EU countries. Turkish, Maghreb
Arabic, Russian, Urdu, Bengali, Hindi, Tamil, Ukrainian, and Balkan languages
are spoken in many parts of the EU. Many older immigrant communities are
bilingual in the local language and in that of their community. Migrant
languages are not given formal status or recognition in the EU or in the EU
countries and they are not covered by EU language teaching programmes.

2.1.3. Language policy in the EU

The European Union is founded on the principle of “unity in diversity”: the

diversity of cultures, customs and beliefs, and of languages. Article 22 of the
Charter of Fundamental Rights of the European Union, adopted by the leaders
of the EU in 2000, says that the Union shall respect linguistic diversity, and
Article 21 prohibits discrimination based on a number of grounds, including
language. Together with respect for the individual, openness towards other
cultures, tolerance and acceptance of others, respect for linguistic diversity is a
core value of the European Union. This principle applies not only to the 23
official languages of the Union but also to the many regional and minority
languages spoken in the member states. This is what makes the European Union
a unique place. It is not a “melting pot” in which linguistic and cultural

26
differences are toned down, but a place where diversity is thought of as a source
of wealth. This approach is in sharp contrast, for example, with the English Only
Movement in the USA.
To put this principle into practice, the EU established various organisations
that provide linguistic services. The legal basis for the existence and operations
of the EU language services is Council Regulation No 1 of 1958 determining the
languages to be used by the European Economic Community, which lists the
(then 11) official languages and specifies when and for what purposes they are to
be used.
Article 2 of the Regulation enshrines in primary law the principle that the EU
institutions must communicate with their correspondents in the member states in
the official language chosen by the correspondents. Article 3 also says that when
sending documents to a member state or a citizen of a member state, the EU
institutions have to use the official language of that member state. Article 4 of
the Regulation states that “regulations and other documents of general
application shall be drafted” in all the official languages. Here, as in any other
EU legislation, the term drafting is used instead of translation. As Wagner, Bech
and Martínez (2002: 7) note, this is a “logical consequence of the principle
according to which all official languages have equal status”. This means that
officially there are no “original texts” and “translations” but only various
“language versions” of documents. (In reality, of course, there is always an
original document, which is then translated into the other languages.)
The EU provides interpretation, translation and publication services in all of
its official languages, but only legislation and important documents are produced
in all the official languages. Other documents are translated only into the
languages necessary for their use. For internal purposes the EU institutions make
their own language arrangements. The European Commission, for example,
conducts its internal business in English, French and German, and only uses
other languages for public information and communication purposes. The
European Parliament, on the other hand, has members who need working
documents in their own languages, so its document flow is fully multilingual. In
the European Union, language policy is the responsibility of member states, but
EU institutions promote the teaching and dissemination of the languages of the
member states through a number of programmes.
One might think that the costs of communication in such a huge multilingual
organisation as the European Union make up a rather significant part of the
whole budget. However, the total costs of all translation and interpretation in all
of the EU institutions is actually less than 1% of the total EU budget. The overall
cost of all the language services provided by all the EU institutions – including
translation and interpretation – was approximately €800 million in 2006, which
was less than €2 per citizen. Surely no-one would like to say that this amount of
money is ill-spent.

27
 2.2. Interpretation in the EU

The European institutions between them hold tens of thousands of meetings each
year and have almost a thousand conference interpreters on the staff. The
majority of the interpreters are staff members of the European Parliament or the
European Court of Justice, which each have separate interpreting services, or of
the Directorate-General for Interpretation (DG Interpretation) of the European
Commission. The three interpreting services cooperate closely, in particular in
the management of their joint pool of about 2700 freelance interpreters.

2.2.1. Conference interpreting

Conference interpreting deals with rendering speeches at meetings from one

language into another. Interpreters are expected to fluently and accurately
deliver what speakers say, preserving the content as well as the manner and tone
of the speech.
Speaking at meetings and at negotiations is a central element in the process
of decision making. Participants at a meeting can express themselves in any one
of a number of passive languages, i.e. those languages from which interpreting is
provided at the meeting, and can listen to what the other participants are saying
in one or more active languages, i.e. those languages into which all speeches at
the meeting are rendered by the interpreters.

2.2.1.1. Types of interpreting

There are three major modes of interpreting. In consecutive interpreting, the
interpreter sits or stands near the speaker, listens to what he or she says and
repeats the whole of what has been said, generally with the aid of notes taken
during the speech. Because of time limitations, this technique is rarely used
when more than two active languages are involved.
In simultaneous interpreting, the interpreter sitting in a special booth listens
to a speech through a headset and translates it into a microphone while the
speaker is speaking.
Chuchotage or whispering interpreting is a variant of simultaneous
interpreting where the interpreter listens to the original speech and at the same
time renders it by whispering into the listener’s ear. This technique is still used
in short meetings for which it is impossible to use the equipment necessary for
simultaneous interpretation.
DG Interpretation and the separate interpreting services of the European
Parliament and of the Court of Justice render speeches delivered in meetings,
using mainly simultaneous interpretation.

28
 2.2.1.2. Language regime
In conference interpreting, there is a difference made between active and passive
languages from the viewpoint of the interpreter. An active language is a
language that the interpreters speak and that delegates at a conference can listen
to. A passive language is a language that the interpreters understand and that is
spoken by the delegates at a conference.
Active and passive languages may be combined in various ways in meetings.
A language regime is a specific combination of passive and active languages at
a meeting.
A meeting with a 23-23 language regime has 23 passive and 23 active
languages. In the European Union institutions this means that all the official
languages are interpreted into all the official languages. Such a regime is called
complete and symmetric. Reduced regime means that interpretation is provided
from less than the full number of official languages. In an asymmetric regime,
you can speak more languages than you can listen to. When we say a meeting
has a 15-3 regime, it means that delegates may speak 15 official languages but
interpretation is only provided into three, for example into English, French and
German.

2.2.1.3. The conference interpreter’s language combination

The AIIC, the international association of conference interpreters, uses the
following definitions. The interpreter’s active languages are the languages that
the interpreter can work into. These include what are called the interpreter’s A
and B languages.
An A language is the interpreter’s native language (or another language
strictly equivalent to a native language), into which the interpreter works from
all her or his other languages in both modes of interpretation, simultaneous and
consecutive.
A B language is a language other than the interpreter’s native language, of
which she or he has a perfect command and into which she or he works from one
or more of her or his other languages.
The interpreter’s passive languages are the languages which the interpreter
can work from into his or her active languages. A passive language is called a C
language. A C language is a language of which the interpreter has a complete
understanding and from which she or he works.
In most cases, conference interpreters work from their passive (C) languages
into their mother tongue (A language). Some interpreters have an excellent
command of a language other than the native language and are able to work into
that language from one or more of their other (C) languages. These interpreters
have a second active (B) language. Some interpreters work into a B language in
only one of the two modes of interpretation.

29
 An interpreter who is able to work from his or her mother tongue into a
second active language is said to do a retour. Some interpreters who have a
retour language work into that language only in the consecutive mode. A small
number of interpreters are able to work from all their languages into a second
active language. An even smaller number of interpreters have more than two
active languages. Interpreters who can work from all their languages into a B
language are said to do a second full booth.

2.2.2. The Interpreting Directorate of the European Parliament

The European Parliament is unique among international organizations in that it

has to provide a full multilingual environment. Since any citizen of the Union
has the right to be democratically elected as a member of the European
Parliament, fluency in a widely used language cannot be expected of MEPs as
would be the case for diplomats or EU officials. The right of Members to follow
debates and express themselves in their own mother language is explicitly
provided for in the Parliament’s Rules of Procedure. The interpreting service of
the European Parliament must provide interpreters for all these meetings and
also supplies interpreters to the Court of Auditors and the Commission services
in Luxembourg, where the secretariat of the European Parliament is based.
The Interpreting Directorate employs approximately 240 permanent staff
interpreters and relies on a reserve of more than 1000 auxiliary conference
interpreters, of whom between 200 and 500 must be recruited each day to cover
its needs. In 2002, the total volume of activity represented 56,000 interpreter
days for the European Parliament organs alone.
Except for face-to-face meetings and for missions away from Brussels or
Strasbourg, practically all EP meetings rely on simultaneous interpreting in
specially equipped meeting rooms, where interpreters work in teams of at least
two interpreters per active language. For a meeting with simultaneous
interpreting in six or more languages there must be three interpreters in each
language booth. Typically, interpreters work from three or more passive
languages into their mother tongue. However, interpreters who are native
speakers of such lesser-known languages as Finnish, Estonian, Hungarian,
Latvian, Lithuanian, etc. also have to work from their mother tongue into
another language.
To work at the European Parliament, training or experience as a conference
interpreter is necessary. Interpreters should have a full university education and
exceptionally good knowledge of languages. Versatility, analytical skills and
complete mastery of the mother tongue are vital. Finally, the almost unlimited
range of subjects covered in parliamentary debate requires extensive general
knowledge and considerable expertise in all areas of European Union activity.

30
 2.2.3. The Directorate for Interpretation of the European Court of
Justice

The main task of the Directorate for Interpretation is to ensure the

comprehension of spoken interventions at hearings before the Court of Justice
and the Court of First Instance by providing simultaneous interpretation by
teams of experienced conference interpreters. It is also responsible for providing
interpreting services for other internal meetings, such as meetings of members of
the judiciary of the Members States or during visits to the Court.
In order to carry out those tasks, the Directorate for Interpretation employs a
core staff of interpreters employed as officials, which is supplemented, where
necessary, by auxiliary conference interpreters.
Each case before the Court of Justice and the Court of First Instance has its
own “language of the case”. That language is chosen by the parties from among
the official languages of the European Union. Moreover, governments of the
member states pleading before the two courts have the right to use their official
languages.
The linguistic profile of the audience is therefore variable: the division must
make up its interpreting teams according to the language of the case, the
language of the member state whose government is an intervener, the
interpreting requirements of the judges sitting on the case, and the language of
the visitors.

2.2.4. The Directorate-General for Interpretation of the European

Commission

The Directorate-General for Interpretation is the European Commission’s

interpreting service. DG Interpretation manages the allocation of Commission
meeting rooms and provides support for the smooth running of meetings in
many languages that are held there. It also organises conferences for
Directorates-General and departments of the Commission. Its mission therefore
is to make possible multilingual communication at the core of Community
decision-making. DG Interpretation provides interpreters for some 11,000
meetings every year and is the largest interpreting service in the world.
The Council of the European Union, The Committee of the Regions, The
European Economic and Social Committee, The European Investment Bank, and
the agencies and offices in the member states also all get their interpreters from
DG Interpretation. At present, the Council accounts for around 46% of the
interpreting services provided, followed by the Commission with around 40%.
The remaining 14% are spread over the two Committees, the European
Investment Bank and various agencies and other bodies.

31
 2.3. Translation in the European Union

A multilingual organisation like the European Union must rely on professional

linguists to keep it running smoothly. In particular, the work of translators of
written text enables the European Union to meet its legal obligations in terms of
informing and communicating with its citizens. Most importantly, the Union
passes laws directly binding on its citizens and companies, and so they and their
courts must have a version of the laws they have to comply with or enforce in a
language they can understand.
There are translation services producing language versions of written
documents for all of the EU’s various institutions and bodies. These services are
located at the Council of the European Union, the European Commission, the
Court of Justice, the European Central Bank, and the Translation Centre for the
Bodies of the European Union. Each of these contributes to enhancing the
transparency, legitimacy and efficiency of the Union. Their specific tasks and
working methods vary depending on the special functions of each institution.
The translation services recruit their permanent staff through open
competitive examinations. These competitions begin with written tests on
translation and on the European Union and general knowledge. Candidates also
have to do translations into the language of the competition from two other
official languages. This is followed by an oral test, aiming to assess the
candidate’s general level of knowledge, expressive and interpersonal abilities.
The competition takes about eight to ten months. The successful candidates are
placed on a reserve list. When there is a vacancy in a translation service, it
selects translators from this reserve list, and invites them for further interviews
and medical examinations. The reserve lists are normally valid for about one
year (Wagner, Bech and Martínez 2002: 30).
The general conditions for eligibility to work as a translator in the EU
institutions are as follows (Wagner, Bech and Martínez 2002: 32). Candidates
must be citizens of a member state, must hold a university degree in languages
or in a specialised field such as economics, law, science etc. They must have a
perfect command of the target language and a thorough knowledge of at least
two other official EU languages.
Translators in the EU institutions are mostly required to translate into their
mother tongues. The organisation of the translation services is based on teams of
translators who work into the same target language. Such teams are called units
or divisions and are managed by a head of unit or division. They also include
some administrative staff, who support the translators’ work. In each translation
service, requests for translation from clients are received and distributed by a
planning office. These offices provide the link between the translation services
and the requesting services.

32
 2.3.1. The Translation Directorate of the European Parliament

The Translation Directorate, based in Luxembourg, is made up of individual

language units and a data-processing, language and documentation support unit.
Each language unit is responsible for one official language of the European
Union.
Translators translate into their mother tongue documents relating to the
various activities of Parliament, i.e. texts of a political, legal, budgetary and
administrative nature, etc., from at least two other official languages of the
European Union. They must have a good command of information technology
(IT) and office systems tools.

2.3.2. The Translation Department of the Council

Under the Treaty, every citizen is entitled to write to the Council – or indeed to
any of the Union’s institutions and bodies – in one of the official languages and
receive a reply in that language. However, for practical reasons, there have
always been limits on multilingualism at the Council. For communication within
the institution, all of whose officials and other staff are expected to know two
Union languages in addition to their mother tongue, the most widely understood
languages are used; the same goes for work involving civil servants and experts
from the member states, who in general also speak at least one foreign language.
These limits are dictated by both practical considerations and budgetary
constraints, in the interests of keeping operating expenditure down.
The Translation Department plays a key role in the Council’s application of
the rules governing the languages used by the European Union’s institutions. Its
main task is to provide all the translations necessary so that the documents which
the Council has to discuss are available to it in all the official and working
languages. Those translations must be of suitable quality and available to their
intended users in time.
In 2005 the Translation Department translated into each of the old official
and working languages some 5000 documents, or about 50,000 pages; the new
languages are gradually approaching the same figures. The total number of
pages of translation exceeded 800,000. The work is done almost entirely in-
house; outside translation is resorted to only where necessary. There are more
than 700 translators in the Translation Department, with a support staff of more
than 400 officials and other staff. The Department is organised into language
units corresponding to the official and working languages of the institutions.
Translators specialise to some extent in specific technical or policy areas, but are
involved in the translation of all types of documents intended for Council bodies.
The Translation Department cooperates closely with the lawyer linguists of
the Quality of Legislation Directorate, which is part of the Council’s Legal

33
Service and whose main task is the legal finalisation of the legislative texts that
the Council adopts.

2.3.3. The Directorate-General for Translation of the European

Commission

The Directorate-General for Translation (DGT) is the largest single department

in the European Commission. The DGT provides the Commission with
translation services and linguistic advice for its written communication in all
official languages. It works only for the Directorates-General and Services of the
Commission. It translates written text into and out of the official languages of
the EU, and some others when needed.
The texts translated by the DGT include legislation, policy documents,
speeches, background papers on legal, technical, financial, scientific and
economic issues, correspondence, press material – whatever the Commission
and its departments need for their work. Legislation and documents of major
public importance or interest are produced in all of the official languages, but
that accounts for only about one third of the DGT’s work. Other documents,
such as communications with the national authorities and correspondence, are
translated only into the languages needed.
The DGT can also summarise long documents in a target language if the
customer does not need a full translation, or can edit original documents drafted
in a language other than the mother tongue of the writer. It develops and
maintains a wide range of computer-assisted translation tools, including
translation memories in every official language, a huge terminology database
and a machine translation system. It helps the other Commission departments to
maintain websites in many languages.
The DGT uses various methods of translation, involving in-house and
freelance translators. A freelance translator is an independent translator who
works for a translation agency on a temporary basis, in particular projects.
The traditional method of translation is when an in-house translator with
the target language required translates the document, and a second in-house
translator revises it.
The external method is when the translation is created by a freelance
translator and revised by an in-house translator with the target language
required.
The two-way method is when an in-house translator whose principal
language is the source language of a document and who has an excellent
knowledge of the target language translates the document out of his or her
principal language.
The relay method is when an in-house translator whose principal language is
the source language of a document translates it into a relay language, usually

34
English or French, known by another translator who then puts it into the
requested target language. This method is used for uncommon language
combinations such as from Estonian into Greek.
The three-way method is when an in-house translator translates from and
into languages of which neither is his or her principal language, as, for example,
when an Italian translator puts a text from Arabic into English.
At the DGT, there are about 1750 full-time translators and some 600 support
staff involved in management, administration, research and development,
communication and planning. The Directorate-General for Translation is based
in Luxembourg and Brussels. The Directorate-General for Translation also sends
materials out to freelance translators throughout Europe.
In 2006 output from the DGT was 1,541,518 pages. Of this, 80% was done
in-house and the rest by freelances. A page is 1,500 typed characters, not
including spaces. The workload tends to rise each year, because of the constant
expansion in the Commission’s areas of activity. Because of the sudden sharp
increase in the DGT’s workload when nine new official languages were added in
2004, the Commission adopted a strategy of translation demand management:
limits were imposed on the length of documents which could be sent for
translation and priorities were set as to the types of documents which could be
handled, with legislation given top priority. These restrictions are expected to
apply until enough translators into the new languages can be recruited to provide
the same level of service as for the previous eleven official languages. For 2007,
translation costs at the Commission were estimated to be €302 million.

2.3.4. The Translation Directorate of the European Court of Justice

The translation service shared between the Court of Justice, the Court of First
Instance and the Civil Service Tribunal is called the Translation Directorate of
the European Court of Justice. At the beginning of 2006, the Translation
Directorate was staffed by 796 employees, representing 45% of the staff of the
institution.
Translation in the Court is carried out under mandatory language rules and
covers all the language combinations of the official languages of the European
Union. The volume of work currently stands at approximately 700,000 pages per
year. The texts to be translated are all legal texts, highly technical in nature.
They are written by lawyers for lawyers. The service employs only linguists who
are fully qualified lawyers and have a thorough knowledge of at least two
languages other than their mother tongue.
The Translation Directorate includes an Organisation and Methods Section,
which has many fields of activity: training and information, organisation and
methods of work, new tools to aid translation, etc; 20 language divisions,
producing all translations into the language concerned from all the other

35
languages; and a General Services Division, which is responsible for tasks in
support of the whole of the Directorate: management and distribution of
translations, management of freelance work, pre-processing of texts for
translation, documentary and terminological research, management of the
archives, scanning of references for preliminary rulings, training and
dissemination of information, etc.
The language divisions all do essentially the same work. For every new case,
they translate the document initiating proceedings (the order for reference) or the
information notice summarising the dispute, drawn up in the relevant registry
(Court of Justice or Court of First Instance) and published in the Official Journal
of the European Union. When their language is also that of the case, they are
required to translate a certain number of documents which are essential for the
case-file: observations of the member states, applications for leave to intervene
submitted by third parties, reports for the hearing and the Opinion of the
Advocate General in cases before the Court of Justice. In every case they also
translate the judgment, its summary and the information notice for the judgment,
published in the Official Journal.
Since French is the language of the deliberations, the French Division plays a
special role. It systematically translates the document initiating proceedings as
well as all the pleadings. In cases before the Court of Justice, it also translates
(like the other divisions) the Opinion of the Advocate General. It never translates
judgments, as they are drawn up in French.

2.3.5. The Translation Division of the European Central Bank

The Translation Division of the European Central Bank (ECB) is a central

service unit, which provides linguistic expertise and support in the 23 official
languages to the ECB. The services include translation of documents from
English, revision of texts, proofreading of texts intended for publication and
compilation of relevant terminology.

2.3.6. The Translation Centre for the Bodies of the European Union

Established in 1994, the Translation Centre’s mission is to meet the translation

needs of the other decentralised European agencies and offices. The Centre is a
full member of the Interinstitutional Translation Committee for Translation and
Interpretation and contributes to activities in the following fields: resources,
terminology, administration, external translation, new working methods and
techniques, etc.
The Centre is established in Luxembourg. It is an agency governed by
European public law. It has its own financial resources constituted by the

36
transfers from the agencies, offices and institutions in exchange for services
provided.
The number of pages translated by the Centre amounted to 505,000 in 2005.
The rapid growth in activities since its establishment in 2005 has gone hand in
hand with an increase in the Centre’s staff, numbering over 180 staff members in
2009.
The Centre’s work covers a wide variety of subjects: applications for
Community trade marks and models, opinions for the European Medicines
Agency on medicinal products for human and veterinary use, reports on
organised crime and human trafficking, information on combating racism and
xenophobia, data on drugs and drug abuse, applications for new plant varieties,
initiatives for education and training actions for the Central and Eastern
European countries and the Mediterranean area, reports on the state of the
environment in Europe, documentation on good practice regarding health and
safety at work, reports on living and working conditions in the EU and the
candidate countries, documents concerning joint aviation requirements, etc.
The Centre has language teams for all Community languages. Translators are
normally not specialised in one specific domain but translate documents for most
of the clients. If the Centre cannot cover a language combination in-house or if
the subject is of a very technical nature, it can call upon a network of freelance
translators that it has built up over the years.

2.3.7. The Interinstitutional Committee on Translation and

Interpretation

The Interinstitutional Committee on Translation and Interpretation (ICTI) is a

forum for cooperation between the language services of the European Union’s
institutions. It was established in 1995 on the initiative of the heads of
administration of the various institutions, and now includes representatives of all
the translating and interpreting services.
From the beginning, the Committee has been responsible for coordinating
and carrying out interinstitutional projects in fields which include making
savings in the use of resources, complementarities in relation to management,
pooling research into new working methods and techniques and the successive
enlargements of the Union.
One of the visible results of this cooperation is the interinstitutional
terminology database called IATE. IATE is used by the EU’s various institutions
to collect, distribute and jointly manage EU-specific terminology. IATE is
designed to house all of the EU’s terminology resources in one central location
on the web, making this information more readily available and standardising its
presentation. IATE combines all the individual term databases previously used
by the EU’s various translations services. It is run jointly by all the main EU

37
institutions (European Commission, Parliament, Council, Court of Justice, Court
of Auditors, Economic and Social Committee, Committee of the Regions,
European Central Bank, European Investment Bank and the Translation Centre).

Figure 4. Sample search in IATE

2.4. Language technology

Language technologies are both an essential tool for safeguarding Europe’s rich
cultural heritage and a source of future economic growth. By reducing the cost
of working in multiple languages, such technologies do not just help to preserve
linguistic diversity but also reduce the costs for European businesses and other
organisations in working together in Europe’s single market.
Information technology is also playing an ever increasing role in translation
work. The integration of the different actors and tools in the translation process
is one of the major challenges for translation services today. Computerised
workflow systems allow all phases of this process to be managed and monitored.
These include reception of the translation job, automatic pre-translation and

38
provision of reference materials, translation, quality control and the transmission
of the final product to the customer.

2.4.1. Language technology at the DGT of the European Commission

DGT makes various computerised resources and tools available to translators,

who use them to fit the translation needs and their personal preferences. The
main document formats used are Word, Excel, HTML and PowerPoint.
Irrespective of their working methods, all translators’ needs are basically the
same: appropriate terminology (dictionaries, glossaries, terminological
databases, etc.); reference documents (paper, electronic archives, aligned texts,
etc.); a facility enabling them to re-use previously translated texts (copy-pasting
from other applications, electronic archives, translation memories, etc.); and
central and local assistance. Secretaries have evolved from pure typists into
translation assistants. Secretaries and translators work hand in hand, with pre-
and post-processing being handled by secretaries, and translators focusing on the
real translation work. At central level, assistance is provided by the Helpdesk
and Pre-processing Team, and at local level, within the language units.
To perform its tasks, the DG for Translation has a wide variety of language
resources at its disposal: terminology in many different forms (multilingual
libraries, Internet access, etc.). At desktop level, terminology searches are
mainly performed via IATE (interinstitutional terminology database) and Quest
(one-stop access to a series of general-interest terminology databases);
processing of sentence fragments by TMan, an internal subsentence-level
replacement tool; translation memories held centrally by Euramis, thus enabling
genuine data sharing; texts as such to be retrieved from DGT’s internal archiving
system (called DGTVista) or from any other source; machine translation, which,
at the European Commission, is used not only as an information browsing tool
but also as a genuine translation aid, and can thus be regarded as a fully fledged
language resource.

2.5. Conclusions

Translation has come a long way since its beginnings to become the most
complex and biggest linguistic enterprise of the present day: the translation
services of the European Union. The translator is no longer an individual
working mostly in isolation but is part of a network of specialists who can rely
on a huge pool of experience and resources made available by modern
information technologies. Naturally, changes in the practice of translation have
also been accompanied and fostered by changes in the theory of translation. The
following chapters provide an introduction into the fundamental questions of
translation theory.

39
 2.6 Study questions

After reading the chapter, answer the following questions. If you cannot
remember, go back and revise the related part.

Which are the most important institutions of the EU and what is their
function?
Which are the official languages of the EU?
What is the founding principle of the EU and how is it manifested in the
EU’s language policy?
What forms of interpreting are used in the EU?
What language regimes are employed in conference interpreting? What
language combinations are possible in interpreting?
Which EU institutions operate interpretation services?
What makes the European Parliament unique in linguistic terms?
How does the interpretation service work in the European Parliament, in the
European Court of Justice, and in the European Commission?
Which EU institutions operate translation services? How are the translation
services organised in the EU institutions?
Which EU institution operates the largest translation service?
What methods of translation are used in the DGT of the European
Commission?
How does the Translation Directorate of the European Court of Justice
operate?
Which translation service is responsible for working with the decentralised
EU agencies and offices? What is the ICTI?
What are the main advantages of modern language technologies?
What language resources and tools are used at the DGT?

40
 3. BASIC CONCEPTS AND PROBLEMS OF TRANSLATION

This chapter introduces some basic notions of translation and discusses some of
the pivotal problems that have been the concern of translation scholars since
ancient times, and presents arguments for the claim that the study of translation
is best carried out in a general communicative framework.

3.1. General questions

In this section we shall tackle three questions that are at the heart of any thinking
about translation. The first is an existential one concerning the often-bemoaned
impossibility of translation, the second is a qualitative one concerning the need
for assessing translational decisions, and the third one regards the form and
range of a theory of translation.

3.1.1. The problem of translatability: translation, language, culture

Traduttore traditore, the translator is a traitor, says the Italian proverb and,
evidently, there is some truth in it. Anybody who has ever tried their hand at
translation will remember the painful experience of not being able to render
some word or expression in a ‘perfect’ way. There are always elements in a text
which defy the translator and which, in the end, get more or less lost in the
process. This is what makes some people think that in a sense translation is an
impossible endeavour. Yet we know that people have actually practised
translation for thousands of years.
According to Nida (1959), the main source of problems in translation is the
mismatches between the grammatical and lexical categories of the languages
involved. Some information is necessarily lost in the translation, for instance,
when the original’s language obligatorily expresses something that the
translation’s language does not. This is the case with English and Hungarian
third person singular pronouns, where there is a gender contrast in English,
which does not exist in Hungarian.
For example, the English sentence “She smiled at the teacher” can be
translated into Hungarian as “(Ő) rámosolygott a tanárra”. The Hungarian
sentence, however, is neutral in terms of the gender of the subject and,
moreover, the personal pronoun in subject position is normally deleted, unless it
bears emphatic stress. The result then is a translation which contains less
information that the original.
According to information theory, when information encoded in one code
system is decoded and then encoded in a different code system, some of the

41
information will get lost. Such inevitable information loss is called entropy.
Since translation is also (partly) a process of re-coding, the notion of entropy
entails that translation necessarily involves losses. The more codes involved, the
greater the entropy. This would constitute a strong argument against the use of
relay languages in translation.
However, allowing for a wider view of translatability, such problems can be
solved, for example, by a paraphrasing procedure since, as Roman Jakobson
writes, “languages differ essentially in what they must convey and not in what
they can convey” (Jakobson 1959: 234). What this optimistic formulation means
is that anything that can be expressed in one language can also be expressed in
some way in another language. But not everybody shares this optimism.
Catford (1965) distinguishes two reasons for the untranslatability of a
particular element: one linguistic and one cultural. Linguistic untranslatability
is when an element is untranslatable because of the differences between the
language of the original text and that of the translation.
Obvious problems of linguistic untranslatability arise when the effect of the
source text is based on some formal property of the source language, as in these
questions about the English language:

Why is it that writers write, but fingers don’t fing, grocers don’t groce,
and hammers don’t ham? If the plural of tooth is teeth, why the plural
of booth isn’t beeth? If teachers taught, why didn’t preachers praught?
If a vegetarian eats vegetables, what does a humanitarian eat?

Another kind of problem is when the source text builds on meaning relations
within the language. In the following example the translation problem is caused
by the fact that the English expression “to be out” is polysemous, while in
Hungarian we use different words to express these meanings.

When the stars are out, they are visible, but when the lights are out,
they are invisible. →
Mikor a csillagok kigyúlnak, akkor láthatók, de mikor a fények
kihúnynak, akkor nem láthatók.

A different case is cultural untranslatability, when some element cannot be

translated because of differences between the concepts and ideas that are
available in two different cultures. For instance, what would be the English
translation of this Hungarian sentence?

Nem engedett a negyvennyolcból.

Perhaps it could be translated as something like this:

42
 He did not make any concessions.

But surely native Hungarian speakers can tell (provided they also speak English)
that the English sentence does not mean exactly the same as the original: some
elements of meaning are lost.
Linguistic and cultural untranslatability, however, seem to be two aspects of
the same thing, given the intricate relationship between language and culture,
where culture is understood in a wide, anthropological sense, explained in
Tomalin and Stempelski (1993), as including its members’ ideas, beliefs,
customs, language, and material as well as non-material products. The best-
known (and most controversial) formulation of this relationship is the so-called
Sapir–Whorf hypothesis, named after the linguists Edward Sapir and Benjamin
Lee Whorf, from whom it originated.
According to the linguistic determinism hypothesis, the language we speak
effectively determines the way we make sense of the world of phenomena, the
way in which we understand reality. Whorf, writing in 1940, in his article
‘Science and linguistics’, expounded this idea in the following words:

We dissect nature along lines laid down by our native languages. The
categories and types that we isolate from the world of phenomena we
do not find there because they stare every observer in the face; on the
contrary, the world is presented in a kaleidoscopic flux of impressions
which has to be organized by our minds – and this means largely by
the linguistic systems in our mind. We cut nature up, organize it into
concepts, and ascribe significances as we do, largely because we are
parties to an agreement to organize it this way – an agreement that
holds throughout our speech community and is codified in the patterns
of our language. The agreement is, of course, an implicit and unstated
one, but its terms are absolutely obligatory; we cannot talk at all except
by subscribing to the organization and classification of data which the
agreement decrees. (Whorf 1997: 213–214)

One year later, in ‘Language, mind, and reality’, he returned to this question and
argued again that every language is a pattern-system with its own forms and
categories, ordained by the culture of which it is part and, by using the language,
the individual not only communicates, but also perceives and analyses
phenomena, channels his reasoning, and “builds the house of his consciousness”
(Whorf 1997: 252).
The hypothesis in this form, often referred to as linguistic determinism,
entails that speakers of different languages dwell in different ‘houses of
consciousness’ and are thus prevented from understanding each other, to a level
proportionate to the degree of the differences between their languages. And this

43
would, of course, amount to accepting the claim that translation is, depending on
the level of relatedness between the languages involved, impossible.
However, the extreme form of the hypothesis has been rejected by most
scholars, on the ground of empirical observations to the effect that the mutual
understanding of people from different cultures is far from hopeless, even if
made difficult in certain cases by conceptual differences. One specific form of
this process of understanding is the learning of foreign languages, the very
existence of which serves as a proof of the possibility of conceptualising in
terms of languages other than one’s own. In Hungary, for instance, we do not
have the kind of beer that the English call ‘stout’, yet we can grasp the meaning
of the concept by explanation or demonstration.
Thus the general view has come to be that whereas translation may be made
difficult by linguistic and cultural differences, it is never as difficult as to be
utterly impossible. As Mounin writes, “translation can never be completely
finished, which also demonstrates that it is never wholly impossible either”
(Mounin 1963: 279). There are then scholars, like Anthony Pym, who take it for
granted that translation is possible, and use the notion of translation in defining
culture: “It is enough to define the limits of a culture as the points where
transferred texts have had to be (intralingually or interlingually) translated”
(Pym 1992: 26). Naturally, at this point much depends on what we think
translation is, but as long as we regard translation as some form of mediation
between languages, and cultures, this point seems acceptable.

3.1.2. Translation quality assessment

“Translation criticism is an essential link between translation theory and its

practice” (Newmark 1988: 184). Essential, writes Newmark, and rightly so,
since criticism is connected to three activities which are doubtless among the
chief practical concerns of the study of translation: the evaluation of existing
translations, the search for grips in making decisions during the translation
process, and the improvement of translator education practices. Holmes also
names translation criticism as one of the crucial areas of the applied branch of
translation studies (Holmes 1988a: 78).
One major question regarding quality assessment is whether it is possible to
carry it out with the necessary degree of objectivity. If not, it will amount to no
more than a mere list of subjective impressions. Objectivity can only be
achieved if quality assessment is based on explicit and systematic criteria. As to
the nature of these criteria, two questions arise: (a) Should they be prescriptive
or descriptive, rather? And (b), should they include only the translation or,
rather, a comparison of the translation with the original? In general, it could be
said that the majority of translation scholars today prefer a descriptive and
comparative approach. However, the usefulness of comparing the translation to

44
the original has been brought into question by some scholars. Toury, for
example, is afraid that such a comparison of translation to original may lead to a
mere enumeration of translation errors and an undue reverence for the original
(Toury 1978: 26). Such an approach, scholars like him would argue, fails to
acknowledge the fact that the translation is bound not only to the original text
but also to the conditions of the situation in which it is created. Therefore, the
quality assessment of a translation can be done in at least two ways: in terms of
its closeness to the original text or in terms of its acceptability in the context of
the receiving culture.
To describe these relationships, Shveitser (1993: 51) uses the terms
equivalence and adequacy, and defines them in the following way. The notion of
equivalence answers the question whether the translated text corresponds to the
original, while adequacy answers the question whether a given translation, as a
process, meets the requirements of given communicative conditions.
This distinction, originating in this form from Komissarov (1980), entails that
equivalence and adequacy are not necessarily linked with each other. A
translation can be (to some degree) equivalent to the original and not adequate in
the given situation. Or, vice versa, it can be adequately performed while not at
the same time equivalent to its source. The questions of equivalence and
adequacy will be given a more detailed treatment in Sections 3.2 and 3.3.

3.1.3. Translation theory

The criteria for translation quality assessment should be provided by a theory of

translation. In this respect we are first faced with a meta-theoretical question:
What should this theory be like? Holmes (1988a) argues that the study of
translation is an empirical discipline and, as such, has two major objectives: “to
describe particular phenomena in the world of our experience and to establish
general principles by means of which they can be explained and predicted”
(Hempel 1967: 1, cited in Holmes 1988a: 71). In other words, the basic
requirements are that its statements should be (a) empirically testable, (b)
predictive, and (c) general.
Accordingly, translation studies, as the discipline has come to be known
after Holmes (1988a), aims at (a) describing the phenomena of translating (the
process) and translations (the result of the process), and (b) establishing general
principles by means of which the above-mentioned phenomena can be explained
and predicted. These two branches of the discipline are called by Holmes
descriptive translation studies and theoretical translation studies (Holmes
1988a:71). He also distinguishes a third branch, which he calls applied
translation studies, including four areas: translator training, translation aids,
translation policy, and the one that has been discussed in the previous section,
translation criticism.

45
 Holmes also points out that a comprehensive theory of translation must be
“the product of teamwork between specialists in a variety of fields – text studies,
linguistics (particularly psycho- and socio-linguistics), literary studies,
psychology, and sociology” (Holmes 1988b: 101). This view of an
interdisciplinary translation theory is shared by many scholars, but has also
come under criticism in the past years, most notably by Gutt (1991), whose
arguments will be presented in the concluding section of this chapter.
What is certain, however, regardless of such disputes, is that a theory of
translation should be an attempt to answer two basic questions. Firstly, what is
translation? Secondly, how does translation happen? Or as Bell (1991: 2)
reformulates it, for reasons that will soon become clear: What is a translator?

3.1.3.1. What is translation?

Naturally, a theory of translation must be able to delimit its domain of interest: it
must be able to say what translation is. First of all, to avoid confusion, we need
to clearly separate three meanings of the term: translation as a process
(translating), the product of the process (a translation), and translation as an
abstract concept, encompassing both of the above (Bell 1991: 13).
Jakobson (1959) distinguishes three types of the translation process
according to what codes are involved: (a) intralingual translation, (b) interlingual
translation, or translation proper, and (c) intersemiotic translation.
Intralingual translation means rewording a text within the same language.
This is what happens, for instance, when an old text such as the famous Halotti
beszéd és könyörgés, from the end of the 12th century, which was written in Old
Hungarian, is translated into (relatively) modern Hungarian. Here is a short
excerpt from the original text, followed by its translation.

Látjátuk feleim szümtükhel, mik vogymuk: isá, por ës homou

vogymuk. Mënyi milosztben terömtevé elevé miü isëmüköt Ádámot,
ës aduttå valá neki påråadicsumot házoá. Ës mënd pårådicsumben
valou gyimilcsëktül mondá neki élnië. Hëon tilutoá űt igy fá
gyimilcsétűl. Gye mondoá neki, mérët nüm ënëik: iså, ki napon ëmdöl
az gyimilcstűl, hålálnek håláláål holsz. Hådlåvá holtát terömtevé
Istentűl, gye feledevé. Engedé ürdüng intetüinek, ës ëvék az tilvot
gyimilcstűl. Ës az gyimilcsben hålálu evék.

Látjátok, feleim, szemetekkel, mik vagyunk: íme, por és hamu

vagyunk. Mennyi malasztban teremté kezdetben [az Úr] a mi ősünket,
Ádámot, és adta vala neki a paradicsomot házzá. És mind[en]
paradicsomban való gyümölcsöktől monda neki élnie. Csupán tiltá őt
egy fa gyümölcsétől. De mondá neki, mért ne ennék: „Bizony, [a]ki
napon eszel az[on] gyümölcstől, halálnak halálával halsz”. Hallá holtát

46
 teremtő Istentől, de feledé. Engede az ördög intetének, és evék az[on]
tiltott gyümölcstől. És az[on] gyümölcsben halált evék.

Interlingual translation means rendering a text into another language. (For the
sake of simplicity, let us here disregard the problem of how to distinguish one
language from another, Serb from Croat, for example. A “working hypothesis”
could be the following formulation, attributed to the sociolinguist Max
Weinreich, whereby a language is defined as a dialect with an army and navy.)
To take the previous example again, as we know, the Old Hungarian Halotti
beszéd és könyörgés was itself a translation from Latin.
Intersemiotic translation means the interpretation of signs by means of
signs of another sign system. Examples of this third type are reading the time off
a watch or turning a sentence into a Morse signal. In a more abstract sense,
however, any form of language use involves some intersemiotic transfer, where
the various kinds of representations in the mind, from chemical to verbal, are
translated into each other until a linguistic utterance is finally produced. (On this
point, also see Sections 3.4 and 4.1.2.)
According to the relationship that holds between the information content of
the original text and that of the translated text, the following three basic types
may be distinguished (Hervey and Higgins 1992: 17): (a) Gist translation,
where the translation merely provides a summary of the original. This can be
useful when the readers of the translation are only interested in the most
important points of the original, as in translations of the “one hundred famous
novels” type. (b) Rephrasing is the basic type in which the original is
supposedly translated without any modification of the content. This is obviously
the dominant underlying assumption, for instance, of readers of literary works of
art in translation. And (c), in exegetic translation the translation provides more
information than the original: it is furnished with explanations or
supplementations which are considered necessary by the translator for the
intended reader to make full sense of the translation.
According to the medium of the process we also distinguish between oral
translation, or interpreting, and written translation, which will be referred to,
from now on, as translation proper. These two types of activity, although not
essentially unlike, are sufficiently different in a number of ways and have
consequently been the subject of two different sub-branches of the general study
of translation. In this book we will mostly be concerned with written translation
between different languages.
What we have to do now, then, is explicate the nature of the process – but, as
will turn out, this is not as trivial a problem as it might seem at first glimpse. The
question of the nature of translation (as a process as well as the product of this
process) is discussed in more detail in Sections 3.2 and 3.3, and in Chapter Four.

47
 3.1.3.2. The process of translation
On the face of it, what happens is that the translator receives the original text, or
source text (ST), written by the source author in the source language (SL) for
the source reader, and turns it into a translation, or target text (TT), written in
the target language (TL) for a target reader. The translator’s role, then, is that
of a mediator between the writer of the ST and the reader of the TT. More
specifically, a translator is a “bilingual mediating agent between monolingual
communication participants in two different language communities” (House
1981: 1). The translator’s role is different from that of a monolingual
communicator in that whereas the latter’s task is to produce a text within the
given SL circumstances (in a primary communication situation), the translator
first has to make sense of the ST and then produce a TT, based on the ST, within
the bounds of the given TL circumstances, which may be very different from the
SL circumstances (in a secondary communication situation). We can thus
divide the process into two stages: the analysis of the ST and the synthesis of
the TT, both of which are mental processes taking place in the mind of the
translator.
A first approximation of these mental processes is found in Robinson’s
‘shuttle model’ (1997), according to which translation is not, as has been
commonly assumed, primarily a conscious analytical process but one that
involves two equally important mental states: a subliminal ‘flow’ state and a
conscious analytical state, between which the translator shuttles back and forth,
as required by the specific circumstances arising during the translation process.
‘Flow’ is understood in the sense of Csikszentmihalyi (1995) as a state of
mind in which an activity is performed solely for the sake of the pleasure of
doing it. This sort of autotelic, or self-rewarding, experience “produces a very
specific experiential state, so desirable that one wishes to replicate it as often as
posible” (Csikszentmihalyi 1995: 29).
In this model, the professional translator spends most of the time during work
in the subliminal state of mind, doing translation in a rapid and mostly
unconscious, or subconscious, manner. Their actions in this mental state are
guided by habit, which is the result of the sublimation of recurring experience
sifted through highly conscious analytical thinking. Work in this subliminal
mental state involves the use of procedural memory (Robinson 1997: 95).
Procedural memory is the part of memory that is specialised in performing
habitualised activities.
When the translator is baffled by a problem that cannot be solved in this
rapid, unconscious processing mode, their mind will switch over to the analytical
state, employing conceptual memory for working out a solution to the problem
at hand by checking and examining things through painstakingly conscious
analysis. Conceptual memory is the part of memory that is specialised in the
storage and handling of concepts.

48
 The particular solution chosen may be the result of intuitive guesswork or
may be based on analogy with earlier experiences of a similar kind, which
through recurring several times have begun to form into an inductive pattern in
the translator’s mind. Thus a vital element of this model is the view of the
translator as a life-long learner, where learning takes place as the result of
repeated exposure to novel problems. Constant learning is seen as a necessary
condition for sustaining the translator’s long-term interest in their work. In
Csikszentmihalyi’s terms, to maintain the flow experience “one must increase
the complexity of the activity by developing new skills and taking on new
challenges” (Csikszentmihalyi 1995: 30).
Learning comes about as the outcome of a cyclical movement, in conscious
and unconscious ways, through the stages of instinct (an intuitive readiness for
the task), experience (engagement with real-world challenges) and habit
(internalised, or sublimated, modes of reacting to challenges). This view of
learning draws on the ideas of Charles Peirce, as explicated in his Collected
Papers (Peirce 1931–66: 5.477–5.480). In these terms the translation process is
seen as one where the translator brings in an unfocused aptitude for working
with language and translating, which is then tested and polished by a countless
variety of experiences with linguistic and other translational problems and, over
time, this experience with specific problems is turned into subliminal
behavioural patterns, that is habits, which enable the translator to work in very
fast, subliminal, largely unconscious ways until a novel problem arises that falls
outside of the range of habitualised solutions. A crucial part of the translator’s
professional skill or competence is the ability to switch from subliminal
processing to conscious analytical processing in such situations, which is also
one of the habits that the translator has to internalise.
In Peirce’s terms, inductive pattern building begins with an intuitive leap
from a heap of unordered data to a hypothesis. This initial intuitive leap he calls
abduction. The hypothesis is then tested inductively through recurring
experience and is subsequently turned into a generalisation by deduction. It is
this deductive generalisation which finally becomes sublimated by the translator
as a habitual action and out of the network of such internalised habits emerges
the translator’s own theory of translation. This personal translation theory may
remain partly, or even largely, unconscious but in the end this is what is
responsible for what we call translational competence (Robinson 1997: 105).
The shuttle model of translation also builds on Weick’s idea of the
enactment–selection–retention cycle in learning. In translation, the enactment
stage corresponds to making specific translational decisions, selection
corresponds to editing the translation and reflecting on these decisions, while
retention corresponds to sublimating the results of the previous stages.
Translating, editing and sublimating work in a cyclical manner: new experiences
can always change old habits and in turn these modified habits will give rise to

49
new solutions. The main point is that the translator’s competence is not a stable,
unchanging construct but a constantly evolving, dynamic system of knowledge
which, with every new cycle in the learning process, attains higher and higher
degrees of complexity, thereby making possible the sustenance of ‘flow’ in the
translator’s mind.

3.2. Equivalence-based approaches to translation

What a theory of translation needs to explain then is exactly what cognitive

mechanisms are at work in the human mind during the above-described
processes, during textual comprehension and in using the results of the textual
analysis process in synthesising a text in a different language. The possible form
of such a theory will receive its shape during the discussions that follow in the
rest of this chapter and in Chapters Four and Five.

3.2.1. The concept of equivalence

Translation, as the etymological source of the word, the Latin expression

tanslatio linguarum (transfer of language) suggests, is a transfer process and, as
such, must involve the preservation of certain elements: these elements,
remaining unaltered, are called the invariants of the process. In general, what is
supposed to remain invariant in translation is some sort of meaning, but the
question is not as simple, as we will see later. This notion of invariance is, in
turn, bound up with the notion of equivalence, which “presupposes a
relationship between text A and text B, or segments thereof, in which a given
invariance has been preserved” (Shveitser 1993: 50). All throughout the history
of thinking about translation, the notion of equivalence has been used as a device
for explaining the phenomenon of translation. This has been done in two
characteristically different ways. In the one, it is used as a definitive condition,
in a normative manner: for a text to classify as a translation it has to be
characterised by some form of equivalence to the original. Examples of this type
of definition are not difficult to find in the literature. Here I shall cite just a few
of the more typical of them.

Translation [is] the replacement of textual material in one language

(SL) by equivalent textual material in another language (TL). (Catford
1965: 20)

Translation [is] the transformation of a text originally in one language

into an equivalent text in a different language retaining, as far as
possible, the content of the message and the formal features and
functional roles of the original text. (Bell 1991: XV)

50
 Translation is the expression in another language (or target language)
of what has been expressed in another, source language, preserving
semantic and stylistic equivalences. (Dubois 1973, cited in Bell 1991:
5, Roger Bell’s translation)

Translation is the replacement of a representation of a text in one

language by a representation of an equivalent text in a second
language. (Hartman and Stork 1972: 713, cited in Bell 1991: 6)

In the other way, equivalence between the source and target texts is not looked
upon as a prerequisite but as a result to be discovered after the translation has
taken place. Thus Gideon Toury distinguished between two uses of the term
equivalence: a descriptive use, denoting the relationship between two actual
texts (or utterances, in the wide sense), and a theoretical use, denoting an ideal
relationship between the source text and the target text to be constructed (Toury
1980: 39).
In the first sense, equivalence is an evaluative measure, which needs to be
established after the translation has been produced, and the question is what
kinds and degrees of equivalence can be observed to exist between the two texts.
In the second sense it is a definitional category: a target text is said to be a
translation of the source text if and only if it is equivalent to the original in some
previously specified way. As Barkhudarov (1993: 46) writes in this vein: “The
translator’s goal is always to achieve equivalence, no matter what type of text is
translated”.
In either case, equivalence implies an assumed relationship between the
source and the target texts. In the second sense it is a relationship which is to be
reconstructed by approaching as closely as possible an assumed ideal. In the first
sense, the equivalence of the two texts is constructed during the translation
process.
What is important to stress here is that equivalence is looked upon as a
relation between two actual texts rather than between two languages, since the
(partial) incompatibility of linguistic systems has long been recognised by
linguists as well as by translation scholars. That is, in Saussure’s terms, it is a
matter of parole rather than of langue. Catford (1965), for instance, uses the
terms textual equivalence and formal correspondence to distinguish between
the two types of relation.
In Catford’s formulation, a textual equivalent is a target language text or
expression which, on a particular occasion of use, is observed to be the
equivalent of a given source language text or expression. Textual equivalences
can be established by a formal procedure called commutation, which involves
changing a segment in the source text and observing the corresponding change
in the target text. This reveals how source and target text segments are related to

51
each other. On this basis we can define a textual translation equivalent as that
segment of the target text which is changed when and only when a given
segment of the source text is changed (Catford 1965: 28). Consider this English
sentence:

The dog’s barking is driving me mad.

A possible Hungarian translation would be the following:

A kutya ugatása az őrületbe kerget.

If we now substitute the word yelping for barking, the translation will also
change:

The dog’s yelping is driving me mad. →

A kutya csaholása az őrületbe kerget.

This way we have established that the English word “barking” has Hungarian
“ugatás” as its equivalent in this short text.
On the other hand, a formal correspondent is any target language category
“which can be said to occupy, as nearly as possible, the same ‘place’ in the
‘economy’ of the TL as the given SL category occupies in the SL” (Catford
1965:27). For instance, we can say that the English definite article the has the
Hungarian article a/az as its formal correspondent. But notice that formal
correspondents do not always become textual equivalents. Consider the
following Hungarian sentence and its English translation:

Az ember a természet legnagyobb csodája. →

Man is the greatest miracle of nature.

In this case both examples of the Hungarian definite article have the zero article
as their equivalents in the translation. This is simply because the articles are used
in different ways in different contexts in the two languages.
Catford’s model of translation is actually more concerned with formal
structural differences between languages than with communication between
languages and, as a result, his examples are sentences out of context instead of
real contextualised utterances. These shortcomings are rightly criticised by
Hatim and Mason (1990: 26), who note that in this way translation theory
becomes a mere branch of contrastive linguistics and translation problems are
reduced to a matter of non-correspondence between formal categories in
different languages.

52
 3.2.2. Normative approaches

Since the time of Cicero and Horace (1st c. BC) there has been in the Western
tradition of thinking about translation a strong emphasis on prescribing to
translators how to and how not to translate. In the Classical Roman tradition
there was a difference made between two kinds of translation: faithful
translation and free translation, which is basically a difference between
rendering the text slavishly word-by-word or in a less rigid form, which is truer
to the spirit of the original. The latter form is the one advocated by both Cicero
and Horace and, as a matter of fact, Cicero used the expression ‘to write like a
translator’ (that is, in a word-for-word fashion) in a pejorative sense. Later in the
4th century AD, St. Jerome introduced the notion of sense-for-sense translation,
which defines a middle-ground between the two extremes, placing free
translation outside of the boundaries of translation proper. This trichotomy has
survived in translation theory up to the present day in various forms, such as the
one between, for instance, literal translation, communicative translation and
adaptation (cf., e.g., Newmark 1988). Advocates of this school share the
common view of translation equivalence as a notion which is basically built
upon the notion of meaning, although they may differ in the question of what
kind or kinds of meaning. (The chief theoretical concern then, of course, is to
define what meaning is.) This view is expressed in the following definitions of
translation:

Translation is a process by which a spoken or written utterance takes

place in one language which is intended and presumed to convey the
same meaning as a previously existing utterance in another language.
(Rabin 1958: 123)

Translation [is] rendering the meaning of a text into another language

in the way that the author intended the text. (Newmark 1988: 5)

A different route of thought is taken by scholars who try to characterise

equivalence in terms of function of some sort. Nida (1964), for example,
concerned primarily with Bible translation, considers the linguistic sign, contrary
to Chomsky (1957), not as a carrier of linguistic meaning in the first place but as
an entity fulfilling a certain function in a given society. In this respect his
approach draws on Wittgenstein, who writes that “the meaning of a word is its
use in the language” (Wittgenstein 1958: par. 43). Nida puts the emphasis on
there being a dynamic equivalence between the translation and the original, by
which he basically means that the manner in which the target reader responds to
the target text must be the same as that in which the source reader responded to

53
the source text. This he calls the principle of equivalent effect. We find the
antecedent of this in Tytler’s definition of a good translation:

That in which the merit of the original work is so completely

transfused into another language as to be as distinctly apprehended,
and as strongly felt, by a native of the country to which that language
belongs, as it is by those who speak the language of the original work.
(Tytler 1791: 8–9, cited in Bell 1991: 11)

What is important to Nida, as a result of his religious concerns with the

transmission of the word of God, is that the target text should function in the
same way as the source text, and consequently the more formal aspects of the
text (lexical and grammatical meaning) are relegated to a secondary rank of
significance. Equivalence is thus treated here as a functional rather than a formal
category but, as is clear from the above formulation, it is still used as a defining
(or normative) condition, focused on the source text and source reader. The main
problem with this formulation is that there are no guidelines along which the
reader’s reaction could be determined since “every reading of a text is a unique,
unrepeatable act and a text is bound to evoke differing responses in different
receivers” (Hatim and Mason 1990: 24). This is especially true of creative
literature, as “a literary work of art is always open-ended, and the greater the
degree of this openness, the better it is for that work itself” (Eco 1975: 12, my
translation). In the light of this, the translator can actually be expected “to
preserve, as far as possible, the range of possible responses; in other words, not
to reduce the dynamic role of the reader” (Hatim and Mason 1990: 11).
In House’s (1981) model, on the other hand, the basic requirement of a
translation is that it should be characterised by an overall functional
equivalence to the original, where function means the “use which the text has in
the particular context of situation” (House 1981: 37) and the quality of the
translation is measured by the degree to which its function and textual profile
correspond to those of the source text. The textual profile of a text includes
eight linguistic characteristics: geographical origin, social class, time, medium,
participation, social role relationship, social attitude and province (field of
knowledge), and these determine the norm against which the profile of the target
text is compared. House suggests that the quality of the translation can be
measured in terms of the number and degree of mismatches on these parameters,
but even she recognises that, because of the complexity of possible matrices of
mismatches, a hierarchy of errors needs to be set up, which, however, in any
individual case will depend on the objective of the evaluation (House 1981:
209), and thus, as Gutt points out, “House’s model for quality assessment
provides a basis only for systematic comparison – but not for value judgements”
(Gutt 1991: 13, italics as in original).

54
 3.2.3. Descriptive approaches

Breaking away from the restrictions that a normative approach imposes on the
study of translation, a broadly descriptive paradigm was formed in Germany in
the 1970s under the name of Übersetzungwissenschaft, that is, the science of
translation. While obviously influenced by Nida’s ideas, this school reveals a
shift of focus from normativism to descriptivism and empiricism. The leading
figure of the school was Wolfram Wilss, who worked within a basically
Chomskyan framework where translation is supposed to be carried out by the
creation of syntactic, semantic and pragmatic equivalence, made possible by the
existence of universal categories in deep structure. Wilss (1977) aimed to
construct a general science of translation in which descriptive studies of
empirical translation phenomena may be carried out. This theory was also meant
to provide a framework for translation criticism and quality assessment, and for
the training of translators, the applied branches of the study of translation.
Based on the recognition that theorising often involves unnecessary
subjective judgements and attitudes, the translation studies approach of the 70’s
and 80’s, initiated by James Holmes in Holland, reverses the order of theory
building and empirical research. As Holmes argues in his programmatic paper
The Name and Nature of Translation Studies, first presented at the Third
International Congress of Applied Linguistics in 1972, translation studies has
two main objectives: “(1) to describe the phenomena of translating and
translation(s) […], and (2) to establish general principles by means of which
these phenomena can be explained and predicted” (Holmes 1988a: 71). The
third, applied branch of the discipline, in his view, tackles four areas: translator
training, translation aids, translation policy and translation criticism. He also
voiced the need to include several levels of focus, from the product of translation
(the target text) through its function (in the context of the target culture) to the
process of translation (as a mental operation taking place in the mind of the
translator). This, of course, entailed a very complex multidisciplinary effort,
involving linguists, literary scholars, sociologists, psychologists, etc. As a result,
the very complexity of the task has become a major stumbling block in the way
of creating a unified theory.
Although Holmes insisted that the focus of investigation should be the
translation process, in time translation studies has become identified, as Gentzler
(1993: 93) writes, mostly with the product-oriented branch, focused on the
empirical description of translated texts. Researchers were intent on establishing
one-to-one relationships between source and target text segments, based on
functional notions of equivalence. The general belief was that translators possess
a subjective capacity enabling them to produce an equivalent of the source text
in the target language, and that the target text, as a result, exerts some influence
on the cultural and social conventions of the target culture (Gentzler 1993: 107).

55
 3.2.4. Problems with the notion of equivalence

In recent years, the notion of equivalence has come under criticism on basically
two grounds: problems with defining it properly and with its use as an evaluative
device. The first point has partly to do with the circularity of the concept: it is
supposed to define translation but then its very existence and nature is dependent
on the translation process (Pym 1992: 37).
Thus some scholars think of equivalence as something given in translation.
For example, the equivalence postulate (Toury 1980: 113) states that the
existence of equivalence between the ST and the TT is a corollary of the very
existence of the target text. This is based on his definition of a translation as
any text that is considered as such in the target culture.
But even then, there remains the second problem of how to characterise the
nature of this relationship holding between the two texts. Basically, discussions
centre around two oppositions: that of form versus content, and of meaning
versus function.
As for the latter, several proposals have been put forward by different
authors. Some of them are based on a conception of linguistic meaning of some
sort, all failing to account for the totality of meanings that a given text may be
intended to convey. Others build on some conception of function, but disregard
the fact that the same text may fulfil several functions at the same time. Yet
others experiment with a mixture of the previous two approaches. Unfortunately,
because of the immense complexity of textual meanings and functions, it seems
impossible to provide a sound definition of equivalence along these lines.
As regards the question of content versus form, it is an unfortunate fact that
Tytler’s (1791) requirements of a translation, namely that (a) it should be a
complete transcript of the ideas of the original, (b) its style and manner of
writing should be the same as those of the original, and that (c) it should
preserve the ease of the original composition, cannot be satisfied because the
different requirements are often in conflict with one another. The
correspondence of the target text to the source text in any respect is necessarily
less than perfect. As Bell notes, “the ideal of total equivalence is a chimera”
(Bell 1991: 6). And this means that something always gets lost in translation.
One may try to be faithful to the content as well as to the form of the original in
the translation but the often conflicting nature of the two will generally exclude
the possibility of being completely true to both. Moreover, since the style is also
part of the message, when a formal element is lost, an element of content will get
lost along with it too. Alexieva (1993) suggests that the term equivalence should
be taken to mean, instead of ‘complete identity’, something like ‘optimum
degree of approximation’ but then, again, the vagueness of the expression
optimum degree raises doubts as to the usefulness of this formulation as a tool
for evaluation.

56
 Thus it seems that equivalence is a comparative rather than a classificatory
concept, in Carnap’s (1950) sense, where a classificatory concept provides a
necessary and sufficient condition for class membership and a comparative
concept is one that is used to make comparative judgements. While such
judgements allow statements about sameness and degrees of difference, as Gutt
notes, they “do not in and of themselves constitute value judgements: they can
be turned into value judgements – but only on the further assumption that the
more ‘equivalent’ a translation is, the better it is” (Gutt 1991: 14, italics as in
original). Obviously, if it happens that I resemble Ernest Hemingway more
closely than another person, it still does not mean that I am a better writer than
he is, or that I am better than he is, in any sense. A closer correspondence of my
facial features to those of Hemingway makes me better only if closeness of
correspondence is regarded valuable for some reason – for example because we
are participants of a look-alike contest. That is, equivalence is not a concept that
can in itself be used to evaluate translations – it needs to be placed within a
larger frame of values. The question, then, is what this frame should be.
In Nida’s (1959) view, already discussed briefly in Section 3.1 above, the
source of information loss (or entropy) in translation is the non-correspondence
of lexical and grammatical categories between different languages. But this view
obviously misses the point that linguistic items need to be considered in actual
contexts, not in isolation. In other words, we need to look at, first and foremost,
the communicative value they have in a context, rather than at the abstract value
they have in the linguistic system. Consequently, a text, for our purposes, is to
be considered primarily as a communicative act, which takes place in a social
and cultural context, coming about as the result of the writer’s intention to
achieve a particular communicative purpose. In the process of translating, two
such intentions may be identified: that of the original writer and that of the
translator. Since any text is a function of the socio-cultural context in which it is
born, a translated text, too, is to be viewed as a communicative transaction in
context. Thus the question of evaluating a translation is clearly relatable to the
communicative factors that condition the process of its creation (of which the
source text is but one): a good translation is one that is successful in achieving
its communicative purpose in the given circumstances. This is the approach that
will be looked at more closely in the following section.

3.3. Communicative-functional approaches to translation

In the past twenty or so years, as a reaction to the above-outlined problems with

using the notion of equivalence to the original text as the basic concept in
translation, the whole idea has been abandoned by some theorists in favour of an
approach that lays the emphasis on the communicative purpose or function of
the target text in the given target situation. The view of translation as a

57
communicative event entails that the translator is seen as a bilingual
communicator in an intercultural situation, who needs to be competent not only
in two languages, in the Chomskyan sense, but also in two cultures, in Hymes’s
(1972) sense, who defines communicative competence roughly in the following
way: Communicative competence is the language user’s ability to produce
utterances which, apart from being grammatical, are also appropriate and
acceptable in the given socio-cultural circumstances.
Probably the most notable theories of this sort are Even-Zohar’s (1978)
polysystem theory, several aspects of which were adopted later by translation
studies scholars, Holz-Mänttäri’s (1984) theory of translatorial action and Reiss
and Vermeer’s (1984) skopos theory. All of them view translation primarily as a
process of intercultural communication, in which the translator’s foremost task is
not to reproduce the original along some notion of equivalence but to produce a
target text which is characterised by functional adequacy, which means that it
functions in an adequate, or appropriate, way in the given target situation. Thus
the concept of functional adequacy replaces equivalence here as a result of a
shift of attention from the source text as the standard of comparison to the role of
the target text in the target-culture situation.

3.3.1. Polysystem theory

The opposite of the views of the translation studies approach of the 70’s occurs
in the ideas of polysystem theory, inasmuch as it holds that it is the norms and
conventions of the target culture that crucially influence the presuppositions and
thus the decisions of the translator. The term polysystem was coined by Even-
Zohar, an Israeli scholar, who used it to refer to the hierarchical system of
subsystems within a culture. One of the components of this all-encompassing
socio-cultural polysystem is the polysystem of literature, which is characterised
by a constant competition between the central and peripheral genres for the
dominant, or canonical, positions, thus bringing about what we call literary
evolution (Even-Zohar 1990: 91). Even-Zohar stresses the necessity of including
translated literature in the literary polysystem for the apparent reason that
translations may exert a significant impact on the evolution of the polysystem in
the appropriate circumstances (Even-Zohar 1978: 15). The placing of translation
within this wider cultural context entails that it is viewed as a dynamic process
controlled by the norms of the target polysystem and that the adequacy of a
target text is a function of the existing cultural situation.
Toury (1978) defines a translation in a rather wide sense as any text that is
accepted as such in the target culture. It was he who introduced the notion of
norms in his model, based on Even-Zohar’s polysystem approach to literature.
The expression norm here is not meant in an evaluative sense. Translational
norms are the constraints that regulate translation behaviour within a given

58
socio-cultural context, based on descriptive generalisations about what
translators typically do in particular circumstances.
The translator’s role is seen not as a mere rendering of linguistic expressions
into another language but as involving the fulfilment of a social function, the
norms of which are specified by the community in which the translation takes
place. Consequently, the translator has to acquire these norms in order to be able
to function in an appropriate manner, and thus the acquisition of these norms is a
prerequisite to becoming a competent translator, capable of producing adequate
translations.
According to Toury, a translation can never be completely equivalent to the
original since it is the product of a different cultural context, and nor can it be
entirely acceptable to the target system since it will necessarily contain
informative and formal elements that are unfamiliar to the receiving culture.
Thus the translator is torn between the two opposing requirements of
equivalence and acceptability and the translation can never conform to both
perfectly. To be able to determine the position of the translation between the two
extremes we need, in Toury’s words, an “invariant of comparison”, which he
describes, strangely enough, as Gentzler notes (1993: 132), as a universal
literary and linguistic form and not, as would be expected within a theory of this
sort, as something that is culturally conditioned. This position seems somewhat
self-contradictory since one cannot at the same time adopt the view that each
literary culture is different in its structure and norms from every other one and
that behind each culture we find the same universals. Even so, polysystem
theory cannot be denied the merit of directing attention, instead of the earlier
exclusive focus on the source text, to the cultural context of the translation and,
consequently, of redefining the role of the source text as that of “a stimulus or
source of information”, in Baker’s (1993: 238) words (which, as will turn out
later, is a major point in the theory of translation advocated in this book).

3.3.2. Skopos theory

One of the factors in this context is the function which the translation is expected
to perform in the target culture, and it is exactly this function that is in the focus
of Reiss and Vermeer’s (1984) theory of skopos, according to which translation,
as a form of human action, is determined by its purpose, referred to technically
by the Greek term skopos. So the translation is determined not by the source
text, or circumstances of the source situation, but by the skopos of the target text
in the target situation. Translation, then, is defined here as the process of
producing a target text which is functionally adequate in the given target
situation.
As the skopos of the target text is partly a function of the target reader, it
follows that (a) the skopos of the target text may be different from that of the

59
source text, and (b) one and the same source text may be translated in various
ways, according to the needs and circumstances of the target reader. A corollary
of the first observation is that equivalence between the source and target texts is
only a limiting case of the more general case of adequacy to the skopos of the
translation. The second one, in turn, entails that before the translation process
can begin, it is necessary that the skopos of the translation, based partly on a
consideration of the target reader, is clearly specified.
It is this latter point that has given rise to some of the more serious objections
to the theory, especially as it relates to literary translation. A commonplace of
literary theory now is that a work of literature is always open-ended, in Eco’s
(1975) sense; that is, it has a very complex hierarchy of functions, and to specify
the skopos of a literary text would amount to limiting the number of its possible
interpretations. Therefore a similarly complex hierarchy of purposes would need
to be set up for the translation but, as Gutt points out, this raises “the question of
what that further dimension or principle is that determines the hierarchical
ordering of purposes” (Gutt 1991: 17), and this question is left open by skopos
theory.

3.3.3. The theory of translatorial action

An approach in a similar vein is Holz-Mänttäri’s (1984) theory of translatorial

action. In this model, drawing on action theory and communication theory,
translation is regarded as part of a wider translatorial action, which involves
the translator in various purposeful activities, ranging from negotiating with
clients through researching to producing the translation: a text which can
function appropriately in the specific target situation. Translatorial action, in
turn, is part of a more complex communicative action that is to take place in a
transcultural situation. Thus, eventually, translation is conceived of as an action
whose purpose is to enable the achievement of functionally adequate
communication between cultures.
The main aim of the model is to provide a specification of the different
factors that have a bearing on translatorial action and thereby helping the
translator, an expert in transcultural communication, in making appropriate
decisions. Since, as is also emphasised by skopos theory, an action is determined
primarily by its purpose, the result of the action, the target text, must also be
evaluated in terms of its adequacy for this purpose: its functional adequacy.
The source text has no, or very little, part in the evaluation process: it merely
serves as source material for the action, which can be radically modified if the
purpose of the action requires so. This is because the translator is committed not
to the source text but to the overall aim of producing an adequate target text
within the superordinate process of transcultural communication. Unfortunately,
while this requirement is clear enough, the model fails to explicate the notion of

60
adequacy properly, and thus cannot provide the grips it promises: the translator
is simply presumed to be an expert who has the requisite expertise enabling him
or her to make the decisions leading to the desired result.

3.3.4. Problems with the communicative-functional approaches

The polysystem approach defines translation in a loosely empirical manner as

anything that is accepted as such in the target culture. This is a culture-specific
notion which could in itself be accepted if it was not contradicted by the way in
which Toury formulates his measure of evaluation as a universal
linguistic/literary form.
Skopos theory, on the other hand, relies on a notion of textual functions,
which are ordered in hierarchical complexes, but fails to make clear the criteria
by which such an ordering could be carried out and consequently falls short of
providing a useful evaluative device.
The action theory of translation avoids this problem by positing that the
translator is committed to the purpose of the communicative interaction and the
task is to produce a text that adequately meets that purpose – however, it is left
to the intuition of the translator to decide what adequacy means in this frame and
thus the subjective element remains as sadly emphatic as before.
One further common problem with these theories is that while they very
appropriately direct attention to the target situation and the target reader, they
relegate the source text to a secondary rank of source of information and seem to
undervalue its importance in the process of translation. It is dangerous, however,
to separate the target side from the source text to such an extent. Undoubtedly,
any act of communication is carried out with a goal in mind and the way this
goal is achieved is conditioned by the circumstances of the communication (the
secondary context), but translation is a special case in that the translation is tied
not only to the target situation but also to the original text. One may question
extreme views on the measure of required source text bias such as Nabokov’s,
according to whom in (poetry) translation only the strictest form of literalism is
acceptable (cited in Steiner 1975/1992: 254), but one would never want to deny
that some degree of linkage must necessarily be present between the source text
and the target text for it to qualify as a translation. The difficulty lies in
determining the necessary degree of this bond. And it is here that these theories
are not explicit enough.

3.4. Conclusions

We have seen that both the equivalence-based (that is source text-oriented) and
communicative function-based (that is process-oriented) theories fail on more or
less the same two grounds: their definition of translation either runs into

61
circularity or is left open at one point, and they do not provide explicit evaluative
criteria which could be used during the decision making process and in the
assessment of translation quality.
These problems can be traced back, as Gutt (1991) points out, to the fact that
the theories outlined in the above sections are all based on a descriptive-
classificatory approach, which can categorise phenomena (be it textual features
or communicative functions), make statements about the correspondences
between phenomena in terms of these categories, but beyond that “it has no
other principle to offer,” because “the value, significance, importance etc. of a
phenomenon do not lie in its inherent properties, but in its relation to human
beings” (Gutt 1991: 20). What is needed, then, is a shift away from this
paradigm towards an explanatory theory that avoids the pitfalls of such a
classification.
As in recent years it has become generally accepted that translation is best
conceived of as a process of (intercultural) communication, it seems that the
most appropriate theory of translation will be formed within a general theory of
communication. This is not a new contention, even if its truth has been the
subject of debates until very recently for reasons of the inadequacy of particular
communicative theories for translation. As far back as 1975, Steiner wrote: “Any
model of communication is at the same time a model of translation” (Steiner
1975/1992: 47), since “inside or between languages, human communication
equals translation” (Steiner 1975/1992: 49). (The meaning of this statement will
become clearer in Section 4.1.2.) Consequently, if the phenomenon of
translation can be explained within the bounds of a general communication
theory, there is no need for an entirely separate theory of translation. As for the
above-cited objection in relation to specific theories, “the fact that a particular
approach to communication is inadequate does not necessarily mean that any
communicative approach is inadequate” (Gutt 1991: 22). The real question is,
which theory of communication is the one that is most suitable for the study of
translation. This question will be attended to in Chapters Four and Five.

3.5. Study questions

After reading the chapter, answer the following questions. If you cannot
remember, go back and revise the related part.

What are the possible sources of translation problems and how is this
question related to the Linguistic Determinism Hypothesis?
What are the major problems of translation quality assessment?
What are the major goals, characteristics, and branches of the study of
translation?
How can we describe the process of translation?

62
 In what ways has the concept of equivalence in translation been defined?
What basic approaches to the concept of equivalence can be distinguished
and what are the problems with these approaches?
What major communicative-functional approaches to translation can be
distinguished and what are the problems with these approaches?

63
 4. THE BASICS OF COMMUNICATION

This chapter will briefly present the cognitive and communicative framework in
which the study of translation can be best carried out, to prepare the way for a
more specific characterisation in Chapter Five of translation as an interpretive
form of communication, as developed by Gutt (1991).

4.1. The translator’s mind

In the following discussions we shall assume that the human mind is structured
in a way that is outlined in Smith and Tsimpli (1995). This model of the mind is
the result of their investigation of the case of a uniquely atypical autistic person,
Christopher, whom they group with those referred to in the psychological
literature as idiot-savants or, less pejoratively, simply as savants.

4.1.1. The case of a linguistic savant

Savants are characterised by some serious intellectual (and often physical)

impairment accompanied by a surprising talent in mathematics, arts or music.
Christopher’s case can be said to be unique in that in spite of his general
impairment he has extraordinary linguistic abilities – he claims to be familiar
with some fifteen to twenty languages. Cases like his illustrate the phenomenon
called double dissociation.
Double dissociation means that a person with impaired intelligence may
none the less have normal (or even enhanced) linguistic abilities and, conversely,
someone with otherwise unaffected intelligence can exhibit language
impairment (like some deaf people or aphasics).
However, whereas Christopher’s native language competence is not
essentially unlike that of any normal individual, some areas of language use are
beyond his capabilities. His vocabulary and morphology seem to be intact but he
has difficulty in processing other aspects of linguistic input. This is clearly
manifest in his translations, which, although remarkably prompt, reveal a
significant lack of sensitivity to contextual and linguistic constraints: he has
difficulty in integrating formal linguistic and contextual information. This gap
between lexical and morphological processing, on the one hand, and
grammatical and pragmatic processing, on the other, provides interesting
implications as to the place of the linguistic faculty within the architecture of the
mind.

64
 4.1.2. Smith and Tsimpli’s theory of the mind

Smith and Tsimpli’s cognitive theory rests on three pillars: Fodor’s (1983)
modularity hypothesis, Anderson’s (1992) theory of intelligence, and Sperber
and Wilson’s (1986) theory of relevance.
The modularity hypothesis assumes that the human mind has a
compartmental structure, in which the different components are distinguished by
their functional properties. Basically, there are two types of such components:
perceptual input systems and central cognitive systems. Perceptual input
systems are responsible for providing environmental information (in the wide
sense of the word) for the central cognitive systems, which are responsible for
thought and storing knowledge in the memory, partly in the form of
propositions. The (input) systems are modular and genetically determined. They
are also domain-specific and informationally incapsulated, which means that the
operations carried out in the modules are not subject to central control, and the
flow of information is only from the modules toward the central systems, which
are assumed to be non-modular in nature.
One such input system is responsible for language. The language module
carries out phonological, syntactic and semantic operations, and provides, as its
output, input for pragmatic interpretation done by the central systems, which
integrate the given linguistic representation with contextual and general
information.
Specific
Processors
Central system Visuo-
Spatial Input Modules
Basic
Processing Verbal- Vision
Mechanism Propositional
Audition

Language

Mind

KNOWLEDGE
‘Conceptual’ BASE
Modules

Figure 5. Anderson’s model revised

Adapted from Smith and Tsimpli (1995)

65
Anderson’s (1992) model (see Figure 5) was conceived of in this Fodorian
framework, attempting to spell out the properties of the central systems in some
more detail and ascribing to them a quasi-modular structure. In this model, input
from the modules is either stored in the memory (or Knowledge Base) or is
processed by the Basic Processing Mechanism (BPM), which is the faculty that
implements thinking, and is held to be responsible for differences of intelligence,
due to differences in its speed between individuals. Before any information
reaches the Knowledge Base, however, it has to be evaluated for its relevance.
Quite obviously, we do not want to store in memory information that is
irrelevant to us. This process of evaluation is done by mechanisms that Anderson
calls specific processors (SP), having universal problem-solving capacity. Two
such mechanisms are proposed: a visuo-spatial and a verbal-propositional one,
the former devoted mainly to simultaneous, while the latter to successive
processing. These are responsible for producing problem solving algorithms,
which are then implemented by the BPM. Thus the speed of thinking is crucially
dependent on the speed of the BPM, whatever the efficiency of the SPs.
Smith and Tsimpli (1995) take Anderson’s model as a point of departure but
propose some significant revisions (see Figure 6). The most important of these
are the following (Smith and Tsimpli 1995: 188). (a) They assign a range of
linguistic functions to the Verbal-Propositional SP. (b) They postulate an
Executive function and see the BPM as a faculty acting as a constraint on the
operation of the other systems. (c) They propose that the Language Module
intersects with the Central System, with the Morphology component acting as an
interface between them.
Linguistic processing in this model takes place in the following way. In
comprehension, linguistic input from the Vision or Audition Module is analysed
by the Language Module and turned into the language of thought (LoT) by
means of the Verbal-Propositional SP. (Thus we could say that human language
is translated here into mental language. Remember the statement that we cited in
Section 3.4 from Steiner (1975/1992: 49): “inside or between languages, human
communication equals translation”.) This representation may then be integrated
into the Knowledge Base.
The Verbal-Propositional SP is connected with the Language Module
through the BPM, to make possible the interaction that is needed for the
pragmatic processing of the output of the Language Module (including the
pragmatic enrichment of linguistic input, disambiguation, reference assignment
etc.). The interface between the grammar and the conceptual lexicon, which is
part of the central system, is constituted by the morphological component. The
morphological component carries out the mapping of concepts and conceptual
structure onto words and argument structure.

66
 Central System

Input
LoT’
Visuo- Modules
BPM Spatial
SP
LoT
Central
Modules Mind
Vision
Verbal-
Propositional
SP
Face

Audition
Etc. EXECUTIVE

M I
Lexicon O N
R T Conceptual
P E Lexicon
Syntax H R Taste
O F
L A
O C
KNOWLEDGE BASE
G E
Y
Etc.

Language Module

Figure 6. Smith and Tsimpli’s model of the mind

Adapted from Smith and Tsimpli (1995)

Speech production happens in the inverse direction: first a representation in the

language of thought is generated via a specific processor from the Knowledge
Base. This is then transformed into a linguistic representation via the interface
and the BPM. Thus the initiation of thought comes from outside the specific
processors, which have the task of encoding this stimulus by generating an
algorithm to be implemented by the BPM. This is what, for Anderson, thinking
consists in (Anderson 1992: 17–18). This process implies an Executive function,
which may be located either between the Knowledge Base and the specific
processor or in the BPM itself (as suggested in Smith and Tsimpli 1993: 327).
The Executive is a horizontal faculty (in the sense of Fodor 1983), which means
that information can flow through it in both directions. Its function is to check
representations in the BPM and, if found correct, to pass them on to the output

67
procedures. The result of these output procedures is the performance of language
in the form of speech or writing.

4.1.3. Christopher’s translation problems and the mind

Normally, translation as a communicative effort involves the interaction of

lexical, linguistic and central processes. Christopher’s translation is remarkable
in that, for him, it seems to be a purely linguistic exercise, independent of central
control. He is characterised by outstanding lexical and morphological skills but
his syntactic and pragmatic abilities are severely limited by problems of
processing load (Smith and Tsimpli 1995: 188). The most obvious difference
between him and normal individuals is that he is unable to integrate his lexical
abilities into his sentence structure processing, which is aggravated by his
ignorance of contextual relevance. In other words, he stops at the word
recognition stage and does not proceed to the integration stage, which would
involve the use of contextual information and inference (Smith and Tsimpli
1995: 85–6, 163). For Christopher, translation simply means the translation of
words. What we find behind his translation problems, in the end, is a
dissociation of the various components of his mind, preventing the successful
performance, beyond the lexical level, of translation as a communicative
activity.

4.2. Communication: basic definitions

Communication may be defined as a process involving two information-

processing devices, where one of them modifies the physical environment of the
other, which as a result constructs representations similar to those stored in the
first device (Sperber and Wilson 1986: 1).
Communication can thus be seen as the transfer of information between the
two devices. Since we are mainly interested here in human communication, from
now on we shall identify these devices with individuals, using the terms
communicator and audience. Information is meant in the broad sense, relating
not only to facts (true descriptions of the actual world) but to all kinds of
assumptions presented by the communicator as factual. An assumption is a
thought treated by the individual as a representation of the world. A thought, in
turn, is a conceptual representation (as opposed to sensory representations or
emotional states) (Sperber and Wilson 1986: 2). A conceptual representation is
a mental structure whose components are concepts, which are semantically
related to each other in various ways.
In a relevance-theoretic framework the meaning of a concept is made up of a
logical entry and an encyclopaedic entry, stored in memory. The logical entry
contains the truth-functional properties of the concept. It may be empty, partially

68
filled or fully definitional, depending on the nature of the concept. For an
example, let us consider the following assumption containing the concept CAT.
(From here on, concepts and conceptual representations, including assumptions,
will be marked by small capitals.)

THE CAT SAT ON THE MAT.

From this it logically follows that

AN ANIMAL SAT ON THE MAT.

This means that the logical entry of CAT contains the property [animal]. These
properties are sometimes referred to as semantic features. Such semantic features
are necessary components of meaning in the sense that if this feature is negated,
it will lead to a contradiction. Consider the following assumption:

IT IS A CAT BUT IT IS NOT AN ANIMAL.

This is obviously a contradiction, which then means that [animal] is a semantic

feature of CAT. A fuller characterisation of CAT would include the following
features: [animate], [non-human], [animal], [mammal], [feline], etc. These are
the features that define the creature that we call a cat. Compare this with the
concept BEAR. If it is to be different from CAT it has to have at least one feature
that is different from those of CAT. In fact, its logical entry contains the
following features: [animate], [non-human], [animal], [mammal], [ursine] etc.
So what makes a cat different from a bear is (at least) the difference between the
features [feline] and [ursine].
The encyclopaedic entry, on the other hand, contains various kinds of
representational information about the concept, such as cultural or personal
beliefs or associations. The encyclopaedic entry for BEAR may contain the
following assumtions: A BEAR IS A DANGEROUS ANIMAL. Or: A BEAR SLEEPS
THROUGH WINTER. Such assumptions may be regarded as universally shared
pieces of information about bears, but they are not necessarily characteristic of
each and every bear. Beyond these, the encyclopaedic entry can also contain
culturally bound assumptions such as THERE ARE NO WILD BEARS IN HUNGARY.
The fact that words may carry such personal or cultural associations has of
course been recognised for a long time. For example, the famous Hungarian
writer, Sándor Márai, wrote the following in 1938:

Bizonyos szavak múlhatatlanul elindítanak bennem folyamatos láto-

másokat; elég egy újságcikkben megpillantanom ezt a szót: „garázda”,
s valamilyen németalföldi csoportjelenetet látok vastag testekkel, fél-

69
 homályos szobákkal, s jóízű és nyers drámaisággal, melyet megéltem
egyszer, s félelmes és vonzó emlékem maradt az egészről. Ez a szó:
„pincegádor”, őszi kertet idéz föl, fonnyadt almaszagú kertet, minden
tárgyi részletével, amint lézeng és úszik valahol az ökörnyálas, sárga
fényben. Bizonyos szavak különös emlékeket kapcsolnak, melyeknek
semmi közük a szó értelmi tartalmához. A szavak mást is jelentenek,
mint amire a szótár tanít. (A négy évszak, Budapest: Helikon, 2000, p.
201)

Let us now examine how concepts build up assumptions. An assumption, as was

defined at the beginning of this section, is a structured set of concepts. The
content of an assumption is the function of the logical entries of the concepts
that it contains, and the context in which it is processed is, at least partly, drawn
from the encyclopaedic entries of these concepts (Sperber and Wilson, 1986:
89).
Consider the following utterance as an example: “The woman was terrified to
see a bear approaching her little daughter.” When the audience hears the word
bear, it will activate in his mind the encyclopaedic assumption A BEAR IS A
DANGEROUS ANIMAL. This becomes part of the context and in this context the
hearer will be able to understand why the woman was terrified. On the other
hand, if we take this example: “The woman was terrified to see a rabbit
approaching her little daughter,” no such assumption is activated and the hearer
will not be able to easily interpret the utterance. (Unless it has to do with the
rabbit in Monty Python’s Holy Grail. Now, you will not be able to interpret this
last sentence, unless you have seen the film and the reference to Monty Python’s
Holy Grail has activated a certain assumption in your mind.)
As we have seen in 4.1.2, in the model of the mind that we are building on
(Fodor 1983, Anderson 1992, Smith and Tsimpli 1995), the mind consists of
various specialised systems, which are of two basic types. The input systems
process perceptual information (or, to be more precise, stimuli), turning sensory
representations into conceptual representations of a uniform format. These
representations are then processed by the central systems, which integrate
information from the input systems and information drawn from memory, and
carry out inferential tasks (Sperber and Wilson 1986: 71–72).
Since conceptual representations are subject to inferential processing, they
must have logical properties (apart from other, non-logical ones). These logical
properties of a conceptual representation will be called its logical form. To take
part in logical operations it is sufficient that the logical form should be well-
formed; it is not necessary, however, that it should be semantically complete.
When a logical form is semantically complete, and can thus be assigned a truth
value, we will call it propositional. Otherwise it will be called non-
propositional (Sperber and Wilson 1986: 72).

70
 Non-propositional logical forms can be used in inferential operations.
Consider, for instance, sentence (1) below.

(1) He likes chocolate.

It contains a pronoun which, until bound by an entity in the world (through a

process called reference assignment), makes the sentence semantically
incomplete. Yet it can be used to deduce at least the following implications:

(2) HE LIKES SOMETHING.

(3) SOMEBODY LIKES CHOCOLATE.
(4) SOMEBODY LIKES SOMETHING.

The fact that (1) can logically imply assumptions (2–4) follows from the very
notion of logical implication, which is a logical-syntactic, that is purely formal,
relation between assumptions. In this sense, logical implication may be
considered a deductive rule, that is, one that applies to assumptions in virtue of
their logical forms. However, a deductive rule is different from other formal
computations in that it is also a truth-preserving operation: the conclusion stands
in a semantic entailment relation with the premise. Thus, an assumption that is
logically implied by another is also necessarily semantically entailed by it,
whereas the reverse relation is not necessarily true (Sperber and Wilson 1986:
84). So, for instance, while (1) may semantically entail (5) below, (5) in no way
is logically implied by (1):

(5) HE EATS A LOT OF CHOCOLATE.

We see then that incomplete logical forms can be of use in information processing
at the intermediate stages. They can be used to derive valid inferences and can be
stored in memory as assumption schemas. An assumption schema is an
incomplete logical form, stored in memory, which can be completed into a
propositional form in an appropriate context. Since only fully propositional forms
can represent an actual state of affairs in the world, assumption schemas must be
semantically completed, on the basis of contextual information, into full-fledged
assumptions before they can be said to contribute to the individual’s knowledge of
the world (Sperber and Wilson 1986: 73). When an assumption is entertained as a
true description of the world, we call it a factual assumption.
Thus an individual’s Knowledge Base, or encyclopaedic memory, contains
non-propositional representations (assumption schemas, for example) as well as
propositional representations (assumptions) and, moreover, these assumptions
may be associated with various attitudes. How might these propositional

71
attitudes be represented in memory? For instance, let us suppose that the
memory contains the following basic assumption, treated as a fact.

THE EARTH IS ROUND.

In turn, this assumption may be embedded in another, second-order factual

representation such as the one below:

I KNOW THAT THE EARTH IS ROUND.

Likewise, any basic factual assumption P with a certain propositional attitude

may be represented through a second-order factual assumption, as in the
following examples:

I DESIRE THAT P.
I PROMISE THAT P.
I REGRET THAT P.

It can be assumed, then, that the individual’s representation of the world consists
of a stock of factual assumptions, some of which are basic, others of the second
order, representing attitudes to embedded propositional or non-propositional
forms (Sperber and Wilson 1986: 74–75).
The adequacy of the individual’s representation of the world, however, is
only partly dependent on what assumptions he holds: it is also, in part, a
function of how confident he is in the truth (or likelihood) of these assumptions:
“Improvements in our representation of the world can be achieved not only by
adding justified new assumptions to it, but also by appropriately raising or
lowering our degree of confidence in them, the degree to which we take them to
be confirmed” (Sperber and Wilson 1986: 76). This degree of confidence in the
truth of an assumption is a non-logical property of assumptions and is
metaphorically called their strength. The strength of an assumption comes about
as a by-product of its processing history and is comparable to its accessibility
from memory.
The easier an assumption is to recall from memory, the more likely that it
will be used in the individual’s efforts to better his understanding of the world.
In this sense, some assumptions can be said to be more manifest to the individual
than others. This notion can be extended to include not only assumptions in
memory but facts in general. Sperber and Wilson (1986: 39) define manifestness
as follows: “A fact is manifest to an individual at a given time if and only if he


Note: For the sake of simplicity, we shall follow Sperber and Wilson’s convention here to refer
to the communicator as she, and to the audience (the information-processing individual) as he.

72
is capable at that time of representing it mentally and accepting its representation
as true or probably true”. On this notion of manifestness is based the definition
of cognitive environment, which is the set of facts that are manifest to the
individual. If a fact is manifest to an individual, it is perceptible or inferable by
him. Thus, his total cognitive environment is a function of his physical
environment and his cognitive abilities, including his perceptual, linguistic and
inferential abilities along with information that he has memorised. It contains not
only the facts that he is aware of at the given moment but also the facts that he is
capable of becoming aware of.
It follows that since the physical and cognitive abilities of two people are
never entirely identical, they can never share their total cognitive environments;
they can, however, share a subset of them. A shared cognitive environment of
two people is the intersection of their individual cognitive environments. A
shared cognitive environment in which it is manifest which people share it is
called a mutual cognitive environment. In this cognitive environment every
manifest assumption is, by definition, mutually manifest (Sperber and Wilson
1986: 41–42).
Mutual manifestness is of course a much weaker condition than mutual
knowledge, which, as Sperber and Wilson (1986: 15–21) have shown, cannot
exist. It is not strong enough to guarantee a symmetrical co-ordination in the
choice of the code and context of communication between the communicator
and the audience. This problem can, however, be avoided by assuming that it is
the responsibility of the communicator to make correct assumptions about the
appropriate codes and contextual information the audience will have at their
disposal in the comprehension process and that break-downs in communication
are likely to occur when the communicator fails to make the correct assumptions
(Sperber and Wilson 1986: 43).
Eventually, communication means that the communicator provides a stimulus
(verbal or non-verbal), whereby she alters the physical environment and,
consequently, the cognitive environment of the audience. A stimulus is a
phenomenon (a perceptible object or event) which is to have some cognitive
effect in the audience. The communicator provides the stimulus with the
intention that it will make manifest to the audience certain assumptions that she
would like to share with him, thereby altering his cognitive environment and, at
the same time, extending their mutual cognitive environment, which is of crucial
importance in bettering their possibilities of future interaction. If she fails to
achieve this aim, it is either because she has made the wrong assumptions about
what assumptions are manifest to the audience and, further, what assumptions he
is actually making, or because she has failed to provide an adequate stimulus.
Communication is a risky enterprise. As Sperber and Wilson (1986: 45) put it:
“failures in communication are to be expected: what is mysterious and requires
explanation is not failure but success”.

73
 4.3. Coded and ostensive-inferential communication

The following sections will describe how the traditional view of communication
as a process of encoding and decoding can be integrated into a wider view of
communication based on ostention and inference.

4.3.1. Coded communication

Sperber and Wilson (1986) suggest that humans use two different modes of
communication: coded communication and ostensive-inferential communication.
In coded communication, as described by Shannon and Weaver (1949), the
communicator (or source) encodes her message, which is a representation
internal to the information-processing device, into a signal, a physical
representation, which is sent through a channel to the audience (or destination),
who receives the message by decoding the signal. Unless the signal is seriously
distorted by noise along the channel, the success of communication is
guaranteed on condition that the source and the destination share the same code.
A code is defined as a system which pairs messages with signals. In this model
(see Figure 7), which is often referred to as the semiotic model of
communication, communication is made possible by the existence of an
underlying shared code, without which the audience would have no way of
deciphering the meaning of the signal.
Traditionally, it has been assumed that linguistic communication, like any
other form of communication, can be explained in terms of the semiotic model.
Saussure, for example, believed that semiotics (in his formulation, semiology)
will come up with such general laws that will prove applicable to language as
well. He describes language as “a system of signs expressing ideas” (Saussure
1995: 15), that is, as a code.

message signal received signal received message

source encoder channel decoder destination

noise

Figure 7. The Code Model (Shannon and Weaver 1949)

Now while it is true that language is a code pairing phonetic representations

(signals) with semantic representations, it is easy to see that the latter cannot be

74
equated with the messages that are actually intended to be communicated.
Consider the following example:

Gerwyn: Shall we go out for a drink?

Jean: It’s raining.

It is clear that Jean’s message is not simply that the weather is bad but rather that
in these meteorological conditions she does not feel like going out. Could it be
said that Jean’s answer somehow encodes this message? Not in any reasonable
way. What it does encode, more or less, is merely that there is rain falling. How
then will Gerwyn be able to understand that Jean does not want to go out?
Sperber and Wilson (1986) maintain that the gap between the semantic
representation that the utterance encodes and the actual message is bridged by “a
process of inferential recognition of the communicator’s intentions” (Sperber
and Wilson 1986: 9). In this, they say nothing more or less than what Grice
(1975) and his followers did. However, they go further than that and offer an
ostensive-inferential model of communication that not only makes it explicit
what happens when the gap is bridged by the audience but, most importantly,
also explains how it is done. The next section will give an overview of this
model.

4.3.2. Ostensive-inferential communication

The terms ostensive and inferential describe two complementary aspects of the
same process of communication. It is an ostensive process because it involves
the communicator in producing a stimulus that points toward her informative
intentions. It is also inferential because the audience will use the stimulus
provided by the communicator in an inferential process of comprehension as
evidence for what those intentions may be. Ostension, then, means making
manifest to the audience an intention to make something manifest (Sperber and
Wilson 1986: 49), whereas an inferential process is one which “starts from a set
of premises and results in a set of conclusions which follow from, or are at least
warranted by, the premises” (Sperber and Wilson 1986: 12–13).
This mode of communication can thus be defined in the following terms. In
ostensive-inferential communication the communicator produces a stimulus
which makes it mutually manifest to communicator and audience that the
communicator intends, by means of this stimulus, to make manifest or more
manifest to the audience a set of assumptions {I} (Sperber and Wilson 1986:
63). This set of assumptions is what we can call, to use a familiar term, the
communicator’s message.

75
 4.3.2.1. Ostension
Ostension, as can be seen from this definition, provides two layers of
information. First there is an upper layer of information, which is the
information that the communicator wants to communicate something and,
second, there is a lower, basic, layer of information: the one that the
communicator actually intends to communicate to the audience. As will be seen,
the upper layer has just as important a role in the communication process as the
basic one. This basic layer of information, which is “to make manifest or more
manifest a set of assumptions {I}” is called the informative intention of the
communicator. The upper layer of information, “which makes it mutually
manifest to communicator and audience” that the communicator has this
informative intention, is referred to as the communicative intention of the
communicator. In other words, the communicative intention is a second-order
informative intention, which is fulfilled as soon as the informative intention is
recognised, even if it is itself not fulfilled (Sperber and Wilson 1986: 29–30).
Eventually, what makes true communication different from other forms of
information transfer is exactly the presence of this communicative intention:
here we have a means of distinguishing, in the technical sense, between
informing, where the communicative intention is missing or is covert (not
mutually manifest) and communication, where this intention is overt (mutually
manifest). As Sperber and Wilson (1986: 61) put it: informing modifies the
cognitive environment of the audience, while communication modifies the
mutual cognitive environment of the communicator and audience.
This can be important in a number of ways. First, in certain situations the
communicative intention may provide the clue to unlocking the informative
intention. Such cases are discussed in, for instance, Grice (1975), where the
notion of conversational implicature crucially rests on the requirement that the
communication is done overtly. Second, by extending the mutual cognitive
environment of the communicator and her audience the communicative intention
also fulfils a social function: it opens up new possibilities of further
communication.
But most importantly, ostensive behaviour provides the audience with
evidence to the effect that it will be rewarding for him to engage in the
communication process. Information processing, as any other cognitive process,
involves a certain amount of effort that the individual is bound to invest in the
hope of gains. Based on the assumption that the human mind is an efficient
information processing device, it seems that any reasonable individual will take
part in an act of communication only if he can expect that the gains of the
process will be worth the effort he makes. Sperber and Wilson (1986) suggest
that ostensive behaviour provides evidence of this in the form of a guarantee of
relevance. It implies this guarantee because humans naturally look for
information that seems relevant to them, that is, information that will improve

76
their overall representation of the world. Their main thesis is that “an act of
ostention carries a guarantee of relevance, and that this fact [...] makes manifest
the intention behind the ostention” (Sperber and Wilson 1986: 50). They call this
the principle of relevance and maintain that it is the key to explaining how
inferential communication works.

4.3.2.2. Inference
Inference can be defined as the process of validating an assumption on the basis
of other assumptions. It contributes to the fixation of belief, just as for example
perception does. Sperber and Wilson (1986) maintain that inferential
comprehension is a central thought process with free access to conceptual
memory. This implies that comprehension is a global process in the sense that
“any conceptually represented information available to the addressee can be
used as a premise in this inference process” (Sperber and Wilson 1986: 65).
They also assume that inferential comprehension is non-demonstrative. In non-
demonstrative inference the audience has no way of decoding or deducing
what the communicator’s intentions are; all he can do is construct a hypothesis
about them and then confirm this hypothesis on the basis of the ostensive
evidence that the communicator provides. Spontaneous non-demonstrative
inference, they suggest, is not, overall, as much a logical process as a suitably
constrained heuristic one. “Hypothesis formation involves the use of deductive
rules, but is not totally governed by them; hypothesis confirmation is a non-
logical cognitive phenomenon: it is a by-product of the way assumptions are
processed, deductively or otherwise” (Sperber and Wilson 1986: 69). Moreover,
in non-demonstrative inference no logical rules other than deductive rules are
used, that is such rules that apply to assumptions purely in virtue of their logical
form and “take account of the semantic properties of assumptions only insofar as
these are reflected in their form” (Sperber and Wilson 1986: 85).
Comprehension involves explicating (analysing) the content of assumptions
and thus the deductive rules used in this process must be interpretive in this
sense. Therefore, Sperber and Wilson (1986) hypothesise that in spontaneous
deductive processing the mind only has access to so-called elimination rules,
which yield conclusions from which a specified concept that is part of the
premises is erased. Two examples, taken from Sperber and Wilson (1986: 86–
87), are given below:

And-elimination
(a) Input: (P and Q)
Output: P
(b) Input: (P and Q)
Output: Q

77
 Modus ponendo ponens
Input:
(i) P
(ii) (If P then Q)
Output: Q

It will be easy to understand how these rules work, if we substitute the following
assumptions for P and Q:

P: IT IS RAINING.
Q: THE STREETS ARE WET.

Then and-elimination produces the following results.

Input: IT IS RAINING AND THE STREETS ARE WET.

From the truth of the input two things follow:

Output1: IT IS RAINING.
Output2: THE STREETS ARE WET.

Modus ponendo ponens will produce the following result.

Input1: IT IS RAINING.
Input2: IF IT IS RAINING THEN THE STREETS ARE WET.
Output: THE STREETS ARE WET.

The conclusions that these rules yield are called non-trivial, exactly because
they are explications of the content of the premises, and these are the only kind
of conclusions that the mind generates when processing a set of input
assumptions.
As can be seen in the examples above, elimination rules are of two formal
types. And-elimination is an analytic rule, which takes one single assumption as
input. Modus ponendo ponens is a synthetic rule, which takes two separate
assumptions as input. A set of input assumptions is said to analytically imply an
assumption when it is one of the conclusions in a deduction which only involves
the use of analytic rules. A set of input assumptions is said to synthetically
imply an assumption when it is not an analytic implication of the premises.
Then, computing all the analytic implications of a set of premises is a
necessary and sufficient condition for understanding the content of the premises,
whereas the synthetic implications of the premises offer inferential information
that is the result of a combination of the input assumptions but could not have

78
been generated separately from the individual premises. In other words: when
the deductive device draws the analytic implications of an assumption, it
computes the intrinsic meaning of it. On the other hand, when the device
computes the synthetic implications of the assumption in combination with other
assumptions, it gains an understanding of what that assumption means in the
context of other assumptions that the device already has access to. This process
is called the contextualisation of an assumption and the resulting synthetic
implications are referred to as contextual implications. As Sperber and Wilson
(1986: 108) put it:

A central function of the deductive device is thus to derive,

spontaneously, automatically and unconsciously, the contextual
implications of any newly presented information in a context of old
information. Other things being equal, the more contextual
implications it yields, the more this new information will improve the
individual’s existing representation of the world.

But contextual implications are not the only way to improve a representation of
the world. A given set of old assumptions can also be modified in the following
two ways. When a new assumption provides further evidence for, and thus
strengthens, some old assumption, its effect is called contextual strengthening.
When a new assumption provides contradictory evidence against and
consequently erases from the context an old assumption, its effect is called
contextual elimination.
We have thus described three ways in which the contextualisation of new
assumptions in a context of old assumptions may achieve some contextual
effect: by adding new assumptions to the context in the form of contextual
implications, by strengthening some old assumptions or by erasing others.
Otherwise, if a contextualisation does not modify the context (because the new
assumptions are found too weak and are consequently erased) or all it does is
simply add some new assumptions to it, it will have failed to achieve any
contextual effect (Sperber and Wilson 1986: 117).
In an intuitive sense, an assumption which does not provide any contextual
effect in a given context of other assumptions may be said to be irrelevant in that
context. In the next section we shall take a closer look at the notion of relevance,
which has a central role in the theory of ostensive-inferential communication
under discussion here.

4.3.2.3. Relevance
We have informally said that an assumption (or, in general, an ostensive
stimulus) is relevant inasmuch as it leads to an improvement of an existing
representation of the world. But, as we have seen, any such improvement comes

79
at the cost of a certain amount of effort that the individual has to exert in the
inferential process of comprehension. An efficient information-processing
device (such as a human being) will embark on this enterprise only when the
gains (the contextual effects) can be hoped to come about at a reasonable cost
(processing effort). This can be thought of as a manifestation of the principle
that game theory calls the mini-max strategy: participants in an exchange will
naturally strive to achieve a maximum of gains at a minimum of costs.
Otherwise, when the gains seem too small or the cost too large, the device will
disrupt the process and the communication breaks down. These two aspects of
the process can be built into a comparative definition of relevance in the
following manner:

Relevance
Extent condition 1: an assumption is relevant in a context to the extent
that its contextual effects in this context are large.
Extent condition 2: an assumption is relevant in a context to the extent
that the effort required to process it in this context is small.
(Sperber and Wilson 1986: 125, italics as in original)

The effort required to process a stimulus in a context is the function of several

factors. Wilson (1992: 174) explicates the three most important of these: the
complexity of the stimulus, the accessibility of the context, and the inferential
effort needed to compute the contextual effects of the stimulus in that context.
The human mind is assumed to have the ability to assess the relevance of
assumptions through monitoring the physico-chemical changes that are caused
by the mental processes involved in comprehension. This monitoring process is
mostly intuitive, which means that relevance is not normally represented
conceptually in the mind. It arises as a by-product of mental processes and if it is
ever represented, it is represented in the form not of quantitative, but rather of
comparative judgements and gross absolute judgements such as “irrelevant” or
“quite relevant” (Sperber and Wilson 1986: 132).
An individual will only pay attention to a stimulus when he can expect that it
will prove relevant to him. Thus, when a communicator produces a stimulus
with the intention to convey a certain set of assumptions, she will have to, in a
way, implicitly promise the audience that, on the one hand, the stimulus will
lead to the desired effects and, on the other, it will not take more effort than is
necessary for achieving these effects. This requirement is at the heart of
ostensive-inferential communication and is called the principle of relevance:
“Every act of ostensive communication communicates the presumption of its
own optimal relevance” (Sperber and Wilson 1986: 158). By optimal relevance
we shall mean that the processing of a stimulus leads to contextual effects that

80
are worth the audience’s attention and, moreover, that it puts the audience to no
unnecessary processing effort in achieving those effects.
Importantly, the principle of relevance is not something that participants of a
communicative act must be aware of and must obey in the way that Grice’s co-
operative principle and maxims are mutually acknowledged and observed (or
flouted) by them – it is a general characteristic that automatically applies to
every act of ostensive-inferential communication, without exception.
The task of the audience is to make hypotheses about the informative
intention of the communicator and to pick the one that the communicator can be
thought to have believed to confirm the presumption of optimal relevance. Such
an interpretation is called consistent with the principle of relevance (Sperber
of Wilson 1986: 166). Now, since the order in which the interpretive hypotheses
are processed also affects their relevance, the principle of relevance has the very
important consequence that the audience is entitled to accept as correct the first
interpretation which is tested and found consistent with the principle of
relevance.
There is still one open question here. The success of the communication
process crucially depends on whether the stimulus is sought to achieve relevance
by the audience in the context envisaged by the communicator. How the
audience is able to choose the proper context is the topic of the next section.

4.3.2.4. Context
The addressee of a communication will naturally aim at maximising the
relevance of the ostensive stimulus (and any assumption that it makes manifest
to him) by trying to strike a balance between the effects it yields and the effort it
takes to process. When such a balance is achieved, the stimulus is said to have
been optimally processed (Sperber and Wilson 1986: 144). A stimulus can be
optimally processed only if the audience is able to construct the context in which
the communicator intended her stimulus to achieve relevance.
The construction of this context can be imagined in the following way
(Sperber and Wilson 1986: 139–142). It is assumed that the initial context is
made up of assumptions left over in the memory of the deductive device from
the immediately preceding deduction. These are assumptions which have arisen
as synthetic implications or have been used as premises of these implications, or
have been strengthened in the process. All the other assumptions are erased and
transferred to some short-term memory store. This immediately given context
can be extended in different ways.
First, the context may be extended by adding to it assumptions derived or
used in previous deductions, which can still be retrieved from short-term
memory. Second, the extension may be made by adding encyclopaedic
information associated with concepts which are already present in the context or
in the assumption that is being processed. This means retrieving assumption

81
schemas or assumptions from encyclopaedic memory. Third, it is also possible
to add to the context information about the directly observable physical
environment through the perceptual faculties. These possibilities together
determine a maximal context but it would be absurd to think that all these
extensions necessarily and automatically take place. What this would entail is a
major increase of processing effort and, consequently, a substantial decrease in
relevance.
How, then, is the actual context of the deductive process determined? It is
selected in accordance with the principle of relevance. When the audience
presumes that the ostensive stimulus is consistent with the principle of
relevance, he will look for a context that will support this presumption: a context
that enables him to maximise the relevance of the given stimulus. Obviously,
since context selection also involves mental effort, optimal relevance can be
achieved when this effort is smallest: with a minimal context. Thus, the context
of the deductive process is not seen as given, or predetermined; it is constructed
during the process as a result of the audience’s natural search for relevance. It
may be interesting to note here that the idea of gradual context construction is
not new at all: it is expressed, for example, by the principle of local
interpretation, which instructs the addressee of an utterance not to build a
context any larger than necessary to arrive at an interpretation (Brown and Yule
1983: 59). What is new is that the process of context construction is explained
here as a natural consequence of a general principle of communication.

4.4. Linguistic communication

The sections that follow will explain the fundamental characteristics of

communication through language.

4.4.1. Language and inferential communication

A grammar, in the Chomskyan sense, is a code pairing phonetic representations

with semantic representations. A language, then, can be seen as a set of well-
formed formulas, made up of combinations of items drawn from the lexicon of
the language, which are generated by this code, or grammar. In other words, a
language is a code-governed representational system.
In our cognitive framework, linguistic information is processed by a
specialised input system: the Language Module, which automatically decodes
any linguistic stimulus perceived by the individual, producing as output mental
formulas that we call semantic representations. These representations,
however, are still too abstract in the sense that they are, normally, semantically
incomplete: they are not full-fledged assumptions but, rather, assumption
schemas, which need to be inferentially developed before they can represent a

82
definite state of affairs in some (actual or possible) world (see Figure 8).
Linguistic communication thus involves two layers: a coding/decoding process
and an ostensive/inferential process. The former is subordinate to the latter in
that the output of linguistic processing serves as the input for the central process
of inference, in which the audience uses semantic representations as a source of
hypotheses and evidence about the communicator’s informative intentions.

contextual
assumptions

semantic INFERENCE utterance

representation meaning

Figure 8. The role of inference in linguistic communication

The audience of a linguistic communication has a very complex inferential task.

First, he has to assign a unique propositional form to the utterance. This may
involve a number of procedures, including disambiguation, reference assignment
and enrichment. Disambiguation means selecting one of the possible semantic
representations assigned to the sentence by the grammar. Reference assignment
means binding variables like pronouns by giving them particular values.
Enrichment involves the specification of vague expressions. An utterance,
however, does not merely express a proposition: it expresses it in a certain
mood. The second (but by no means the last) task of the audience is thus to
identify the communicator’s propositional attitude on the basis of the
linguistically encoded mood of the utterance.
The commonplace case when an utterance communicates its own
propositional form, that is when this propositional form is part of the intended
interpretation, is called an ordinary assertion. In such a case, when an
assumption communicated by the utterance is a development of a logical form
encoded by the utterance, we call this assumption an explicature. The situation
is different, of course, with figurative or non-assertive utterances: here the
propositional form of the utterance is not an explicature because it is not part of
the intended interpretation. When an assumption is communicated otherwise, we
call it an implicature (Sperber and Wilson 1986: 182). Then, all the intended
analytic implications of a logical form encoded by the utterance are explicatures,
and all the intended contextual assumptions (implicated premises) with all the
intended contextual implications (implicated conclusions) of that logical form

83
are implicatures. What an utterance communicates consists, thus, in its
explicatures and implicatures.

4.4.2. Descriptive and interpretive language use

In an ordinary assertion, all the analytic and contextual implications of the

proposition expressed by the utterance are exactly the same as those that the
communicator actually intended to communicate: this proposition is a literal
interpretation of a thought of the communicator’s. The resemblance between a
proposition and a thought, of course, is normally less than literal. A proposition
expressed by an utterance will normally be only a partial interpretation of a
thought, where they will share some, but not all, of their analytic and/or
contextual implications. We will say that a propositional form and a thought or
more generally, since a thought is a mental representation with a propositional
form, two propositional forms interpretively resemble one another in a context
to the extent that they share their analytic and contextual implications in that
context (Wilson and Sperber 1988: 138).
Compare, for instance, the following two propositions (supposing that they
are about the same person).

(6) JAMES IS A FATHER.

(7) JAMES HAS A DAUGHTER.

Clearly, both (6) and (7) analytically imply, among other things, (8a) and (8b),
whereas (8c) is only implied by (7):

(8) (a) JAMES IS A PARENT.

(b) JAMES HAS A CHILD.
(c) JAMES HAS A FEMALE CHILD.

Thus, we can say that the two propositions interpretively resemble each other,
while this resemblance is less than complete. An important point here is that
interpretive resemblance is a matter of degree: it may range from the case of
complete, or literal, resemblance through partial to marginal resemblance.
This notion of interpretive resemblance can be extended to cover
resemblance between utterances, but certain problems need to be considered, as
Gutt (1991: 39–44) points out. To begin with, some expressions (greetings like
hello or proper nouns like James) have no logical forms and thus cannot be
developed into full-fledged propositions. They can, however, still resemble each
other in their interpretations. If James says hello to me, I can interpret his
utterance through a mental description of it in the form of JAMES HAS GREETED
ME INFORMALLY, which does have a propositional form.

84
 Secondly, two utterances with identical propositional forms (active-passive
pairs, for instance) may communicate different things because the form of the
utterance can also constrain the way in which it is interpreted in the given
context. Conversely, the same utterance may convey different interpretations
because interpretations are relevance-dependent and relevance is itself
dependent on the context, and thus the interpretation of an utterance is a function
of the context in which it arises.
These problems are avoided if we generalise the definition of interpretive
resemblance in the following way: “two utterances, or even more generally, two
ostensive stimuli, interpretively resemble one another to the extent that they
share their explicatures and implicatures (Gutt 1991: 44)”. This definition allows
for stimuli that either have a propositional form or do not and it is also context-
sensitive inasmuch as explicatures and implicatures arise in context.
Now an utterance is a public representation of a thought of the speaker’s. A
representation with a propositional form can represent things in two ways:
descriptively or interpretively. Descriptive representation is when the
propositional form of a representation is true of a state of affairs. Interpretive
representation is when there is a resemblance between the propositional form
of a representation and the propositional form of some other representation
(Sperber and Wilson 1986: 228–229). Thus, in a strict sense, every utterance is
an interpretation of a thought (which, in turn, may be a descriptive
representation of a state of affairs). It does not necessarily have to be a literal
interpretation - and in most cases it is not. The audience can judge the closeness
of the interpretation on the basis of the principle of relevance. A thought, in turn,
may be in an interpretive relationship with another thought or in a descriptive
relationship with some state of affairs.
Sperber and Wilson (1986: 231) argue that all the basic tropes and
illocutionary forces can be explained in this way. Metaphor involves an
interpretive relation between the propositional form of an utterance and a
thought; irony involves an interpretive relation between the speaker’s thoughts
and thoughts or utterances attributed to others; interrogatives and exclamatives
involve an interpretive relation between the speaker’s thoughts and desirable
thoughts; assertion involves a descriptive relation between the speaker’s
thoughts and a state of affairs; and requesting and advising involve a
descriptive relation between the speaker’s thoughts and a desirable state of
affairs.
Sperber and Wilson (1986: 233) claim that since a communicator aims at
optimal relevance and since there is no reason to assume that the optimally
relevant expression is always the most literal interpretation of a thought, literal
language use is to be considered a limiting case rather than the norm. Very often,
for example, it is in the interest of the speaker to intentionally avoid literalness
and in this way extend the range of possible implicatures of her utterance,

85
thereby reducing the amount of responsibility she has to take for what she says.
Poetic effects, like metaphor, are also explained in this framework as the result
of a wide range of relatively weak implicatures: the richer the variety of weak
contextual effects and the greater the audience’s responsibility for constructing
them, the more beautiful, surprising and effective a poetic expression is. Poetry
builds on weak impressions, while scientific language, for instance, tends to
build on the use of explicatures. Differences in style can thus be explained, at
least partly, as differences in the way that relevance is achieved: by greater or
lesser reliance on implicature (Sperber and Wilson 1986: 224).

4.5. Study questions

After reading the chapter, answer the following questions. If you cannot
remember, go back and revise the related part.

What are the basic properties of the mind’s architecture on Fodor’s and
Anderson’s cognitive approach?
How does verbal processing take place on the cognitive approach? How can
Christopher’s translation problems be explained?
What is communication? How can we define the concepts of information,
assumption, thought, concept, proposition, cognitive environment, and
manifestness?
How are the semiotic and the ostensive/inferential models of communication
related?
What layers of information can be distinguished in communication? What is
the difference between informative and communicative behaviour? What is the
importance of ostension in communication?
What is logical inference? What is the role of inference in communication?
How can spontaneous inference be characterised? What effects can
contextualisation have?
What does relevance mean in communication? What is the principle of
relevance? What important consequences does it have?
What is the context of an utterance? What is included in the initial context?
How can it be extended? How is the actual context selected by the audience?
What are the fundamental characteristics of linguistic communication? What
layers of linguistic communication can be distinguished?
What inferential tasks does the audience have in linguistic communication?
In what ways can an assumption be communicated by an utterance?
What does interpretive resemblance mean between propositional forms?
What does interpretive resemblance mean in general?
In what ways can an utterance represent a thought? In what ways can a
thought represent another thought or reality?

86
 5. TRANSLATION AS INTERPRETATION

In Chapter Three we saw that neither the equivalence-based nor the

communicative function-based approaches can provide a complete and
consistent answer to the question of what translation really is. While we
obviously have an intuitive knowledge of what makes a translation a translation,
if we want to be able to provide a systematic and explicit characterisation of the
nature of the translation process, it will have to be done in terms of a suitable
theory of communication. Such a theory, Sperber and Wilson’s relevance theory,
was outlined in Chapter Four. This chapter, in turn, will show how the nature of
translation can be explicated within the bounds of this theory, as presented in
Gutt (1991).

5.1. Translation as interpretation: a first approximation

Today it is commonly accepted that translation is more than just mere

manipulation of language or linguistic utterances – it is a form of interlingual or,
in a wider sense, intercultural communication. The essential question here is
how translation is different from other forms of communication.
One specific characteristic of communication through translation is, of
course, that it involves, apart from the original or source communicator, an extra
communicator who mediates between the source communicator and the target of
the translation process, the target audience. In this respect, translation seems
similar to the situation where in a noisy place somebody has to render the words
of the person standing on their right side to the one standing on the left (even
though they may all share a common language). What makes the rendering of
the message necessary here is the noise that blocks the channel of
communication between the source communicator and the target audience.
There are then other situations in which the signal, even if familiar to the
audience, does not seem to make sense. This reminds us of the case of
sophisticated texts on complex topics (linguistics or communication, for
instance), where the readers, although familiar with the language, will be
puzzled at what the language conveys and will ask for the help of somebody
who can explain or interpret the text for them.
Translation, however, is different from these cases in that the interference of
a mediator is necessitated not simply because the signal needs to be amplified or
because the audience cannot make sense of it but mainly because the signal itself
is unintelligible to the audience. Thus translation may be seen as a form of
communication where the translator, a mediator, interprets the source

87
communicator’s message for the target audience because for some reason they
cannot process the source signals.
Unfortunately, this definition of translation as an interpretive communicative
process is still too wide and imprecise in that it allows for the inclusion within its
bounds of phenomena which are not normally thought of as instances of
translation, such as reading to a child. However, it puts into focus the notion of
interpretation, which may serve as the starting point of the quest for a more
rigorous definition of translation.

5.2. Interpretation

We saw in Chapter Four that interpretation is an inferential process whereby the

audience infers from the stimulus the intended meaning of the communicator.
For this to happen, the audience must use the set of contextual assumptions
envisaged by the communicator, otherwise the communication may fail. Let us
call the situation when this condition is fulfilled a primary communication
situation, and the second where the audience uses a more or less different set of
contextual assumptions a secondary communication situation (Gutt 1991: 73).
(These notions are more explicit formulations of the ones introduced in section
3.1.3.2.) As Heltai (2003: 158) shows, a secondary communication situation
occurs when any element of the communication situation has changed. Since in
translation some elements, including the communicator, the audience, the code
and the channel, are in most cases different, translation will normally take place
in a secondary communication situation. Any of these elements can affect the
communicability of a message, but the element which has the decisive role in
this respect is the context.
A secondary communication situation will often result in a misinterpretation.
An observation appears in Seleskovich (1977) to the effect that problems in the
process of interpretation arise when the translator lacks the necessary knowledge
of the world and/or of the cognitive context (of the text), which could enable her
to work out the non-verbal sense of the text on the basis of its linguistic
meaning. Gutt (1991: 73) lists four kinds of misinterpretation which may arise
when a linguistic utterance is interpreted against a context different from the one
that was actually intended by the communicator:
(a) The use of wrong contextual assumptions can lead to the choice of the
wrong semantic representation. For example, we can think of a situation where
the speaker says “all that I’m after today”, but the hearer, using a different
context than that used by the speaker, analyses this sequence of speech sounds
into the homophonous “all the time after today”. The result is a sentence which
has a completely different linguistic meaning.
(b) An incorrect context may lead to the derivation of a wrong propositional
form. Imagine the following conversation:

88
 A: He is so cute.
B: Peter? Don’t be ridiculous.
A: No, no, I meant John.

What happens here is that B assigns an incorrect referent to the pronoun in A’s
utterance and consequently arrives at this mistaken proposition:

PETER IS CUTE.

However, what A actually wanted to communicate was this proposition:

JOHN IS CUTE.

(c) Incorrect contextual assumptions can prevent the identification of a

propositional form as an intended explicature or as only a source of implicatures.
Imagine that somebody says: “I wonder if you have the time.” The proposition
expressed by the utterance is roughly this:

THE SPEAKER WONDERS IF THE HEARER HAS THE TIME.

Quite obviously the speaker does not intend to communicate this proposition,
which is clear from the fact that a “yes” would not be an adequate answer here.
This proposition is not an explicature, then, but only serves as a contextual
premise on the basis of which the hearer will work out that the speaker intended
to communicate a request for the time.
(d) A wrong context can also lead to the derivation of implicatures not
intended or, vice versa, to the loss of implicatures actually intended by the
communicator. Consider this example:

A: Will you come to the party?

B: John will be there.
A: So you won’t.
B: No, of course I will!

A misunderstands what B’s answer implies because she uses the wrong context.
She combines B’s answer with the assumption that B does not want to meet
John, but actually B meant it to be combined with the assumption that he wants
to meet John.
Returning to translation, we have noted that it often (if not always) occurs in
secondary communication situations. An important question here is whether a
given message can be communicated in such a situation and to what extent. By
the term message we mean the set of assumptions intended to be communicated
by the communicator.

89
 According to relevance theory an act of communication can only be
successful if it achieves relevance in a given set of contextual assumptions, and
relevance is defined in terms of contextual effects and processing effort. Now
relevance is clearly context-dependent: a given set of assumptions to be
communicated that yields an appropriate number of contextual effects in one
context may fail to do so in a different context. A context that is different from
the one in which a message was originally meant to be communicated is called a
secondary context. Thus in a secondary context the communication may break
down. Alternatively, it can break down because the effort needed to work out
these contextual effects in a secondary context may be unreasonably great,
leading to a loss of interest in the communication on the part of the audience. As
Bell (1991: 213) writes, this is the point, the threshold of termination, “where
the reader has got enough out of the text and/or feels that, in cost-benefit terms,
there is little point in continuing”.
It is then a gross oversimplification of matters to say that a given message
can always be communicated through translation: it is only possible if the
secondary context makes it possible to communicate that message. And this is
exactly what Steiner says when he writes: “Not everything can be translated
now. Contexts can be lost, bodies of reference which in the past made it possible
to interpret a piece of writing which now eludes us” (Steiner 1975/1992: 262,
italics as in original). Some things may defy translation at a given moment but
through changes of context and language may become translatable in the future.
Translators, too, have long been (even if only intuitively) aware of this fact.
This is manifest in translations which are addressed to an audience essentially
different from the original one, for instance when a great classic of American
literature like The Last of the Mohicans by James Fennimore Cooper was
rendered into Hungarian by Ádám Réz in such a way that long politico-historical
descriptive passages were eliminated for the obvious reason that the translation
was done for children, who would not be interested in these or, rather, would not
be prepared to interpret such descriptions, all of which might result in the child
reader losing interest and putting the book down. Thus in such a case it may be a
wise decision on the part of the translator to leave out these parts, in order to
ensure that the communication as a whole would be successful.
In sum, the primary question in translation is not in what way a given
message can be communicated in the target language but whether it is
communicable at all in the context of the receiving culture, in the given
communicative situation, in consistency with the principle of relevance. All
other considerations follow from the answer given to this fundamental question.
Thus, as Gutt (1991: 180) writes, the translator, first of all, needs to clarify for
herself whether the original informative intention is communicable in the given
circumstances or it needs to be modified, and only then can she start thinking
about the question of exactly how her message may be formulated.

90
 5.3. Translation as interlingual interpretive use

If, as is often the case, the same informative intention cannot be conveyed in a
secondary context, then it will need to be altered in order to make it
communicable, while ensuring at the same time that only such changes are
effected as absolutely necessary to achieve this purpose. Translation can then be
seen as the act of communicating in the secondary context an informative
intention that interpretively resembles the original one as closely as possible
under the given conditions. This entails that the principle of relevance in
translation is manifested as a presumption of optimal resemblance: the
translation is “(a) presumed to interpretively resemble the original […] and (b)
the resemblance it shows is to be consistent with the presumption of optimal
relevance” (Gutt 1991: 101). In other words: the translation should resemble the
original in such a way that it provides adequate contextual effects and it should
be formulated in such a manner that the intended interpretation can be recovered
by the audience without undue processing effort.
The following example, taken from Péter Esterházy’s Hrabal könyve
(Magvető Kiadó, Budapest, 1990, p. 10) and its English translation by Judith
Sollosy (Quartet Books, London, 1993, p. 4) will perhaps elucidate what optimal
resemblance means in translation.

Volt cukrászda, két konkurens kocsma, melyet mindenki a régi nevén

hívott, a Serház meg a Kondász…

There was a café of sorts and two rival taverns, which everyone called
by their old names, the Beerhall and the Kondász…

The problem here is that the Hungarian word ser is associated with an
encyclopaedic assumption to the effect that the expression is old-fashioned, it is
not used any longer, and evokes the atmosphere of “the golden days” of the
Austro-Hungarian Monarchy. Since in this part of the book the writer describes
the layering upon each other of the past and present, this assumption definitely
has some contextual importance here. However, the English word beer does not
carry a comparable assumption and this part of the context is thus lost in the
translation. On the other hand, it has a near synonym in English, ale, which does
carry an encyclopaedic assumption, awaking images of the past, of a drink
brewed in the traditional way, without adding hops. Moreover, the related
compound alehouse is further loaded with the encyclopaedic assumption that the
expression is outdated, old-fashioned, and its use in the translation would thus
have resulted in the closest possible interpretive resemblance to the original.
We have arrived at a definition of translation which seems to provide all the
necessary conditions to guide the translator:

91
 They determine in what respects the translation should resemble the
original – only in those respects that can be expected to make it
adequately relevant to the receptor language audience. They determine
also that the translation should be clear and natural in expression in the
sense that it should not be unnecessarily difficult to understand. (Gutt
1991: 102)

These conditions, among other things, seem to explain why it is preferred that
the translator should translate into her mother tongue (or her “language of
habitual use”, as is sometimes allowed). The translator, on the one hand, has to
be able to predict what assumptions might be present in the audience’s cognitive
environment and this is most likely when they share a common culture. And, on
the other hand, she has to possess an ease of expression in the target language
which is normally possible only in the mother tongue. That is, in most cases the
translator will be familiar with the target cultural context and the language to an
extent sufficient to enable her to satisfy the above conditions only when she is a
native speaker of the target language.
The above definition also accounts for another interesting problem, namely
that although the degree of resemblance between translation and original can
always be increased, for some reason it often seems undesirable. We can now
explain why this is so: exactly because the increase in resemblance may be
accompanied by an increase of processing effort which might outweigh the gains
in contextual effects. (We can think of The Last of the Mohicans again.) The two
factors, contextual effects and processing effort need to be carefully balanced by
the translator, who has to accept the fact that losses in contextual effects are
sometimes unavoidable in order to keep the processing effort at a reasonable
level, thereby ensuring the overall success of the communication. Relevance, it
needs to be kept in mind, is always a joint function of contextual effects and
processing effort.
It seems in order that we clarify two points here. First of all, how should we
understand the phrase “the translation optimally resembles the original”? The
terms translation and original are certainly not meant here as the translated and
the original text (a text, in the narrow sense, being a sequence of printed marks)
but as the communicative acts, or utterances, of which the two texts are written
records.
The second question concerns the specificity of translation (as a form of
interpretive language use) compared to monolingual communication. The main
differences can be summarised as follows.
In monolingual communication the communicator communicates (that is,
provides evidence for) her own thoughts. On the other hand, a translator
communicates (provides evidence for) assumptions that she has worked out on
the basis of a written representation of the source communicator’s thoughts. The

92
translator does not have direct access to these thoughts – she can only infer
them.
A monolingual communicator communicates to an audience with whom she
shares a common code (that is, the language). In translation, the assumptions that
are communicated through the source text in the source language need to be
communicated in a different language. Translation thus involves a change from
one code to another, which is built on a conceptual system that is likely to be, at
least partially, different from the one used by the source language. And this point
leads us to the question of context.
A monolingual communicator, ideally, shares a mutual cognitive
environment with the audience, which provides for a primary context. A
translator, on the other hand, often has to operate in a secondary communication
situation, where the source communicator and the target audience do not have a
shared cognitive environment, and this means that the translator needs to
communicate in a secondary context.
Thus, to sum up, the uniqueness and the difficulty of translation lies in the
fact that (a) it involves second-order interpretation, (b) it may (and most often
does) necessitate a shift between conceptual systems and (c) it may (and often
does) take place in a secondary context. We will shortly examine an example in
the next section that highlights the related problems in (b) and (c).

5.3.1. Direct and indirect translation

A limiting case of translation as interpretive use is when the interpretive

resemblance between the translation and the original is purportedly complete:
that is, when the translation “purports to allow the recovery of the originally
intended interpretation interlingually” (Gutt 1991: 163). In a way, this is similar
to direct quotations, which may be employed to allow the recovery of the
original interpretation intralingually, on condition that they are processed in the
original context. Generally, two stimuli may give rise to the same interpretation
if and only if they are processed in the same context, because any interpretation
is causally dependent on the interplay between stimulus and context.
This kind of direct translation, consequently, is only possible if the
translation is processed in the original, or primary, context, otherwise the
contextual differences will result in differences in contextual effects.
Technically, the following definition can be adopted:

A receptor language utterance is a direct translation of a source

language utterance if and only if it purports to interpretively resemble
the original completely in the context envisaged for the original. (Gutt
1991: 163)

93
Naturally, in interlingual (intercultural) situations it is very rare that the original
context should be available in the target culture. It is possible perhaps in
circumstances where different language communities have shared the same
geographical, political, and economic environment for a long enough time to
eliminate major cultural differences but in most cases a secondary
communication situation will exclude the possibility of direct translation. This,
then, implies that the default case is not direct but indirect translation, which
covers various grades of interpretive resemblance. Consider, for an illustration
of the exposition above, the following example, taken from an interview with
Clint Eastwood by Ginny Dougarry (The Times Magazine, 28 March 1998, p.
19):

In the early Fifties, during his two-year stint in the US Army, he had a
casual relationship with a schoolteacher in Carmel. When he attempted
to end the affair, she turned violent. Did it frighten him? “Yeah, it gave
me the spooks,” he says. “It wasn’t a homicide – someone trying to kill
me. But it was someone stalking me and threatening to kill themself.”

What is interesting here is the use of the pronominal form themself, a strange
hybrid composed of a plural and a singular form. The use of the singular self is
obvious: Clint Eastwood is talking about one person here. But then why is he
using the plural them instead of the singular feminine her, when the referent is
clearly a female person? What is the extra meaning carried by themself as
opposed to the expected herself?
To understand that, we need to be familiar with the use of the third person
plural pronoun they as a gender-neutral substitute for the pronouns he and she.
Here is an example from a book called How to Be a Brit.

Not long ago I kept seeing Post Office vans with the attractive slogan:
‘Everyone should have a phone of their own.’ In a letter to the
Guardian I remarked: ‘But I think nearly everyone do already.’ (Mikes,
G. How to Be a Brit. Penguin Books, London, 1984, p. 197)

Although in this text the author makes fun of this use of the pronoun they, its use
in colloquial English in this function is fairly wide-spread today. Of course, the
Post Office slogan could also have been formulated as “Everyone should have a
phone of his own” but that too would have made some people frown. In certain
contexts the use of generic they is more natural and acceptable than the generic
use of the masculine form he. This use of the third person singular masculine
pronoun (and of other masculine forms like chairman) has been in decline since
the 1960s. The most likely reason for this is the need, generated by the growing
feminist movement, to remove from the language any uses that would seem to

94
discriminate against women. As a result, the use of generic they has also gained
ground, because its gender-neutral nature makes it a politically correct form. So
instead of saying this: “If anyone is interested in translation, he or she will be
happy to read this book”, we can say: “If anyone is interested in translation, they
will be happy to read this book”. The use of they will make the sentence shorter
and will also enable us to escape the suspicion of being sexist.
But this use of they is not a new phenomenon. Here is just one example from
Shakespeare’s Hamlet:

’Tis meet that some more audience than a mother, since nature makes
them partial, should o'erhear the speech.

As regards the generic reflexive themself, it too had been used for some time
before it disappeared after the 16th century. According to Burchfield (1996: 777)
it re-emerged in the 1980s as a result of the need for gender-neutral pronouns. It
is not a standard form but occurs from time to time.
One online dictionary provides the following definition for themself:

Themself is a reflexive pronoun that is sometimes used informally in

speech instead of himself or herself when the sex of the person is not
known or not relevant: Any member of the party would try to distance
themself from this policy. Its use, however, is not acceptable in
standard English and should be avoided. (msn.encarta.com)

Thus anyone familiar with the present-day American cultural context will realise
that this strange hybrid form probably occurs in the text as a result of a
somewhat exaggerated effort by Clint Eastwood to use a gender-neutral pronoun
in order to comply with expectations of political correctness. Thus his words
convey the following implicatures:

(1) ANYONE CAN BE IN SUCH A MENTAL STATE, A MAN NO LESS THAN

A WOMAN.
(2) CLINT EASTWOOD IS POLITICALLY CORRECT.

The question is whether it is possible to produce a direct translation of this text

segment into Hungarian. In other words: is it possible to translate it in a way so
that assumptions (1) and (2) will be implicated by the target text?
We run into two problems here. One is a problem of the code: the Hungarian
language does not make a gender distinction in the pronoun system. (There is
one exception: the diminutive pronoun magácska is only used to address
women.) The other is a problem of context: the English text builds on
assumptions that are not available for the Hungarian reader.

95
 Assumption (1) may seem easy to save at first sight since the Hungarian
pronoun maga (or its emphatic form önmaga) has the same logical content as
English themself: it is a third person singular gender-neutral reflexive pronoun.
Then the translation would be something like this:

Nem egy gyilkosról van szó – nem olyasvalakiről, aki meg akart ölni
engem. De ez a valaki állandóan a nyomomban járt, és azzal
fenyegetőzött, hogy megöli önmagát.

The pronoun önmaga is as neutral in terms of gender as possible. The problem is

that exactly because there is no gender contrast encoded in Hungarian, this is
actually the form that can be expected here. Consequently, there is no element of
surprise, there is no stimulus to drive the hearer towards extra implications. The
translation will not activate assumption (1) and thus the result is a piece of
indirect translation. In English there is a linguistic variable that requires a
decision from the communicator, but there is no such variable in Hungarian. The
translation loss occurs here, partly, due to a linguistic difference.
On the other hand, although the Hungarian pronoun is equivalent to the
English one logically, there is a significant difference between them in terms of
their encyclopaedic contents. The encyclopaedic entry of the English word
contains this assumption:

(3) THE GENDER OF THE REFERENT IS IRRELEVANT.

The effect of the form themself is based on exactly this assumption. (1) can be
deduced in the following way. From the previous text segments we know or can
infer:

(4) CLINT EASTWOOD BROKE UP WITH HIS GIRLFRIEND.

(5) BECAUSE OF THE BREAK-UP, THE GIRLFRIEND WAS IN A
TORMENTED STATE OF MIND.

Assumption (5) analytically implies the following:

(6) BECAUSE OF THE BREAK-UP, SOMEBODY WAS IN A TORMENTED

STATE OF MIND.

Since in the last sentence of the interview Eastwood is talking about the girl, at
the end of the sentence the reader expects the pronoun herself. Thus when the
pronoun themself occurs instead it leads them to try and infer some extra
contextual effect. The context of the inference is partly provided by assumption

96
(3), which is activated by the pronoun. In the context of (3) assumption (6) can
easily lead the audience to draw the conclusion in (1), repeated here:

(1) ANYONE CAN BE IN SUCH A MENTAL STATE, A MAN NO LESS THAN

A WOMAN.

Importantly, without assumption (3) it is not possible to infer (1), and the
Hungarian pronoun does not carry assumption (3). This is not a difference
between the linguistic systems but a difference in the availability of a
background assumption relating to the use of a linguistic expression: a pragmatic
(contextual) factor. The translation loss is thus also caused by the fact that in the
translation this culture-specific assumption is not activated. The target language
expression preserves the logical content of the original but fails to preserve its
relevant encyclopaedic content.
Would it be possible to somehow bridge the lack of (3) in the Hungarian
translation? Perhaps we could try to explicate assumption (1) at the end of the
text, in an extra sentence:

Nem egy gyilkosról van szó – nem olyasvalakiről, aki meg akart en-
gem ölni. De ez a valaki állandóan a nyomomban járt, és azzal fenye-
getőzött, hogy megöli önmagát. Bárki kerülhet ilyen lelkiállapotba, egy
férfi éppúgy, mint egy nő.

It is easy to see that this is also problematic, in a way. In the English original, (1)
is only an implicated assumption, but in the translation it becomes an explicated
assumption. Thus, although assumption (1) is communicated by both the original
and the translation, it is communicated in a different way. For what is
communicated explicitly, the communicator has direct responsibility. For what is
communicated implicitly, the communicator does not take responsibility – most
of the responsibility is transferred onto the audience. So in the translation Clint
Eastwood takes direct responsibility for the truth of (1). But notice that he did
not actually say what is in (1). He merely directed the audience towards
deducing it by constraining their choice of context. The translation is thus
misleading since it makes the reader believe that Clint Eastwood said something
that he actually did not. The result, again, is a piece of indirect translation.
It would also be possible to explicate, instead of (1), encyclopaedic
assumption (3) at the end:

Nem egy gyilkosról van szó – nem olyasvalakiről, aki meg akart en-
gem ölni. De ez a valaki állandóan a nyomomban járt, és azzal fenye-
getőzött, hogy megöli önmagát. És az nem számít, hogy férfi vagy nő
volt az illető.

97
This however leads to the same problem and the translation again will be
indirect.
What then about assumption (2)? Here another problem turns up having to do
with a difference of cultural contexts. The notion of political correctness and the
phenomenon of politically correct language use are largely unknown to the
average Hungarian. If we thus assume that the Hungarian target reader knows
nothing about what political correctness means in America and we do not
consider it possible to introduce this notion within the limits of the given
translation task, then assumption (2) will get completely lost, and the result will
be another instance of indirect translation, since the interpretive resemblance
between the original and the translation is less than complete.
Let us now suppose that the intended target reader is familiar with the
concept of political correctness; this would mean that the translation operates in
the same context as the original does. Then it would theoretically be possible to
try and look for a solution that will convey assumption (2) in the Hungarian text.
What could be this solution? First of all, we need to notice that assumption (2)
can be worked out by the audience in the following context:

(1) ANYONE CAN BE IN SUCH A MENTAL STATE, A MAN NO LESS THAN

A WOMAN.
(7) CLINT EASTWOOD COMMUNICATES ASSUMPTION (1).

From this it follows (somewhat simplifying the deduction):

(8) CLINT EASTWOOD DOES NOT MAKE A DIFFERENCE BETWEEN

PEOPLE ON THE BASIS OF THEIR GENDER.

This can be combined with the following contextual assumption:

(9) ONE WHO DOES NOT MAKE A DIFFERENCE BETWEEN PEOPLE BASED
ON THEIR GENDER IS POLITICALLY CORRECT.

The result will be this contextual implication:

(2) CLINT EASTWOOD IS POLITICALLY CORRECT.

Thus to preserve (2) we need assumption (1) as a contextual premise. But we

have seen that (1) can only be rendered indirectly, if the translator makes it an
explicature. Thus we run into the same problem that was explained above and
the translation is bound to be indirect.
So we find that the implicated assumptions in (1) and (2) either get lost in the
Hungarian translation or can only be communicated indirectly. This is primarily

98
because of a linguistic difference concerning pronominal gender contrast
between English and Hungarian, and partly because of a cultural difference
concerning the availability of some encyclopaedic background assumption. This
is then an example that serves to simultaneously illustrate the problems that
Catford (1965) calls linguistic and cultural untranslatability. What makes the
example especially interesting is how the two factors are actually interrelated
with each other. In this particular case, it is impossible to separate them from
each other and the two factors together cause the translation to become indirect.
We might of course think of other solutions to get around this problem and
produce a direct translation. For this the intended target reader should have
access to the original context. It would be possible, for instance, for the
translator to explain the situation to the reader in a footnote or endnote. This
way, by expanding the target reader’s cognitive environment, the translator
could provide for the availability of the necessary contextual assumptions. This
would however require some cognitive effort from the reader and it is
questionable whether the reader of a popular magazine, who is reading for fun,
would be willing to make this effort. It is unlikely that the magazine reader
would read a footnote in small letters or a lengthy introduction. So this solution
would probably make the preservation of the assumptions in (1) and (2) too
effort-consuming and would thus threaten the optimal relevance of the
translation. Therefore, it seems that the best the translator can produce here is a
somewhat indirect translation but, regarding that the text makes sense even if (1)
and (2) are lost in the translation, no serious damage occurs. There are cases
when the translator has to accept the fact that certain things cannot be translated
under the given circumstances.
It could, on the other hand, be a good idea to expand the reader’s cognitive
environment by means of an introduction or a translator’s foreword in the
translation of a (serious) novel, providing pieces of contextual background
information necessary for the interpretation of the text that the translator
assumes the target reader does not have ready access to. Somebody who
endeavours to read a whole book is probably also willing to exert some extra
effort to learn new things.
The tasks of the translator may thus include the education of the reader, but
such education is only possible when the reader is willing to be educated.
Translators need to see clearly in this matter and have to adapt their strategy to
the circumstances of the given communication situation. If possible, they can try
to expand the readers’ cognitive environment by explaining things but then the
readers must also be informed about what sort of task they have in reading the
translation. The reader must be clear about the possible gains and the efforts
required. It is always the balance of gains and efforts that ensures the relevance
of the translation for the reader.

99
 5.3.2. Conditions and corollaries of direct translation

The notion of direct translation sheds light on some important points. First of all,
complete interpretive resemblance can only be aimed at if the translator herself
is capable of performing a thorough interpretation of the original (Gutt 1991:
164). If this condition is not fulfilled, then the translation cannot even purport to
be direct, in the true sense of the term. What this entails is the requirement for
the translator to be thoroughly familiar with not only the two languages but also
with the other elements of the communication situation, including the two
cultures (cultural contexts) in question.
Second, direct translation may serve as a useful means of familiarising the
target audience with the source culture by communicating to them the original
informative intention. On the other hand, the originally intended interpretation,
as we have seen, is only communicable in the original context, which entails that
the target audience needs to have, or seek, access to all of this contextual
background information. This means that the translator has to look for ways to
provide such information and it also points to the fact that direct translation in
many cases requires some extra effort on the part of the audience as well, in the
hope of gaining a full understanding of the original message.
Thirdly, as Gutt (1991: 183) points out, in such circumstances a crucial
requirement, in order that the communication does not fail, is that the audience
be explicitly made aware by the translator of the intended degree of resemblance
between the original and the translation in a translator’s foreword or otherwise.
Finally, the translator, as any communicator, has to make sure that her
communicative intentions are in accordance with the expectations of the
audience. If she thinks that the intended target audience will not be able or
willing to exert the extra effort demanded by a direct translation then she will be
bound to choose another approach to the given translation task, in order to
ensure the success of the communication (Gutt 1991: 185).

5.4. Conclusions

The notion of translation as interpretive language use is based on the view that
translation is a form of communication and, as such, can be accounted for in
terms of the relevance theory of communication. This implies that the theory of
translation is a natural part of the theory of communication. Any translation
principle, rule or guideline is an application of the principle of relevance and “all
the aspects of translation […], including matters of evaluation, are explicable in
terms of the interaction of context, stimulus and interpretation” through this
principle (Gutt 1991: 188).
Of course, the importance of the context had already been realised by the
communicative-functional approaches to translation as well. Polysystem theory,

100
skopos theory, and the action theory of translation all emphasised that a
translation is always the product of a specific context, including various factors
such as cultural conventions, the circumstances and expectations of the target
reader, or the intended purpose of the translation, and thus the content of the
translation is effectively determined by these factors. As a consequence, in these
theories the source text is relegated, from the status of an absolute measure of
evaluation, to that of a mere stimulus or source material and the success of the
translation is measured by its functional adequacy in the target context. In this,
these approaches can be regarded as the forerunners of the relevance-theoretic
account of translation. What we can see as a major advantage of relevance
theory is that, contrary to them, it actually explicates what adequacy means in a
context: a translated text is adequate in a context inasmuch as it achieves optimal
resemblance in it. (And of course we could even redefine the notion of
equivalence as a limiting case of adequacy, as the instance of complete
interpretive resemblance of the translation to the original.) This sort of
explicitness is not offered by any of the previous three theories.

5.5. Study questions

After reading the chapter, answer the following questions. If you cannot
remember, go back and revise the related part.

In what sense can we say that translation is a form of interpretation?

How does the context influence interpretation? What kinds of
misinterpretations may arise in a secondary communication situation? What is
the primary question in translation?
How can translation be defined with the help of the notion of interpretive
resemblance?
How does this definition provide the necessary conditions for translation?
What are the consequences of these conditions?
How is translation different from monolingual communication?
What is direct translation and how is it similar to direct quotation? In what
circumstances is direct translation possible and when is it not?
What requirements follow from the notion of direct translation concerning
the translator?
What may be the advantages of direct translation and how can it be
implemented in a secondary communication situation?

101
 6. CULTURE AND TRANSLATION

6.1. Introduction

We have seen that a secondary communication situation may make the

translation of certain elements very difficult or even impossible. A secondary
communication situation is especially likely to arise in cases when the source
communicator and the target reader are members of distant cultures. As Lőrincz
(2007: 117) observes on the basis of her examination of English and Russian
translations of Sándor Petőfi’s poems, the fact that even the best of these
translations are not completely equivalent to the originals results from the
difficulties of taking the connotations and cultural allusions of the original texts
from the source language across to the target readers.
Cultural distance causes problems in the translation of culture-specific
expressions, or cultural realia expressions, as they are called by Klaudy (1994,
2003), for example. These expressions put the translator in a situation where she
has to make a decision. She has two basic choices, or strategies: either to accept
that these culture-specific contents will get lost in the translation or to try and
extend the target reader’s cognitive environment so that these contents may be
preserved. This is what the present chapter will explore. In particular, our aim
here is to answer two questions. (a) How is the translator’s strategy manifested
in the operations which are actually selected to tackle culture-specific
expressions? (b) How can we explain the choice of a particular operation in a
particular context?
Since we are concerned in this chapter only with problems that arise as a
result of cultural differences and operations that may serve as possible solutions
to such problems, there will be no full account given here of translational
operations or techniques. For a detailed description of these techniques, the
reader is referred to Klaudy (1994, 2003) who provides an elaborate
classification of lexical and grammatical transfer operations, as she calls them.
Section 6.2 will clarify what we mean by the terms culture and culture-
specific expression. Section 6.3, in turn, introduces, along with a short
explanation of translation strategies, the four basic translation operations, which
will be discussed through examples in Section 6.4.

6.2. Culture and culture-specific expressions

Being interested in translation as a process cutting across cultures, we need to

clarify here what is meant by culture and culture-specific expressions. Some
scholars, like Pym, use the notion of translation in defining culture: “It is enough

102
to define the limits of a culture as the points where transferred texts have had to
be (intralingually or interlingually) translated” (Pym 1992: 26). Translation can
thus be seen as an indicator of the existence of cultural differences. In our
present cognitive framework, these are best regarded as differences in the shared
cognitive environments of groups of individuals or, rather, the mutual cognitive
environments of groups of individuals. A mutual cognitive environment, as we
defined it in Section 4.2, is a shared cognitive environment in which it is
manifest which people share it (Sperber and Wilson 1986: 41). Culture, then, in
the wide sense, may be defined as consisting in the set of assumptions that are
mutually manifest for a group of individuals. Then, cultural differences are
differences between two sets of such mutually manifest assumptions. What we
need to pin down more precisely is the actual nature of these differences.
There are obviously assumptions which all humans are likely to hold, due to
the existence of phenomena which are universally observable, such as PEOPLE
HAVE TWO LEGS or THE SUN RISES IN THE EAST. Other phenomena are not
universal in this sense and will give rise to assumptions that, provided they are
shared by a whole community of individuals, may be said to be culture-specific.
The culture-specificity of an assumption thus means that the assumption figures
in the mutual cognitive environment of one community but is not present in the
mutual cognitive environment of another.
Trivially, any assumption about the language system of a community, and the
meanings it can express, may be culture-specific and, in this sense, any
expression in a language can be culture-specific. This is exactly the fact that
accounts for what we call linguistic untranslatability (see Section 3.1.1).
However, for our present purposes it seems more useful to exclude from our
objects of examination assumptions about the language system. Beyond this, any
culture-specific assumption will be our concern, and any expression in a
linguistic form which activates any such assumption will be relevant for us in a
non-trivial sense. These are what we will call culture-specific expressions (or
culturally bound, using Duff’s (1981) words).
At this point, let us briefly examine why it is better to refrain from using the
phrase cultural realia and to talk about culture-specific expressions instead.
According to Klaudy (2003: 205) the term cultural realia “may refer to a thing
or concept specific to a given cultural/linguistic community, or to the name that
we assign to that particular thing.” The latter sense is obviously an extended
meaning, which may be used for ease of expression but which also obscures the
difference between language and what language may be used to express. Other
authors extend the meaning of the term in another direction. Vlahov and Florin
(1980: 51), for instance, distinguish three groups of cultural realia: geographical,
anthropological and socio-political, including categories like geographical
objects, plant and animal species, foods, drinks, clothes, occupations, tools,
music, instruments, festivals, customs, nicknames, measures, administrative

103
units, organisations, institutions, social movements, social classes, political
symbols, military units and ranks etc. The list is clearly not complete, and it need
not be, but it shows that the term may be understood in a very wide sense to
include all possible aspects of a culture. Yet it obscures another crucial point,
which is this: What we are interested in here is not the complete inventory of a
culture but, rather, what makes a culture different from another. The term
culture-specific, in the sense outlined above, highlights much better the main
idea that in translating from one cultural context into another we are concerned,
first and foremost, with the differences between these contexts. In this sense, in a
given situation anything that carries some special meaning for the intended
audience may become culture-specific, and the question of culture-specificity
can be resolved only with regard to the relationship between two languages
(Valló 2000: 44). Or, rather, with regard to the relationship between two
cognitive environments. For this reason, it is not necessary to make a complete
list of categories that define a culture. What is more important is that we need to
be able to assess how specific assumptions in a context contribute to the
relevance of an utterance and how the possible lack of such assumptions in
another context leads to the choice of a certain translation operation.

6.3. Translation operations and translation strategies

In this section we will examine what solutions the translator can employ to
bridge the distance between the cognitive environments of the source and target
readers in cases where the interpretation of an expression requires access to a
background assumption that is missing from the target readers’ cognitive
environment.

6.3.1. Translation operations

The term operation is defined in the Longman Exams Dictionary (Pearson

Education Ltd., Harlow, 2006) as “a set of planned actions or activities for a
particular purpose”. An operation is thus carried out by means of one or more
actions, which we can call procedures or techniques. A procedure, in turn, is
defined in this dictionary as “a way of doing something, especially the correct or
usual way”.
On the basis of the scope of an operation used in translation, Klaudy (2003:
1968) distinguishes between two main types: lexical and grammatical
operations. Since in translation cultural differences cause problems mainly
(though not exclusively) on the lexical level, in the discussion to follow we shall
be concerned with lexical operations. However, we will not use Klaudy’s system
of operational categories here. Instead, we will introduce categories which
follow directly from the theoretical framework explained in the previous two

104
chapters. Thus the various treatments that culture-specific expressions are
subject to in the process of translation can be categorised into four basic
translation operations. Our translation operations are defined by the four
possible configurations in which the logical and encyclopaedic meanings of an
expression may be conveyed in translation. These configurations can be
illustrated in the following way: (1) [+L, +E], (2) [+L, −E], (3) [−L, +E] and (4)
[−L, −E], where L stands for logical meaning and E, for encyclopaedic meaning.
(1) [+L, +E] will be called total transfer (TT). In TT both the relevant
logical content and the encyclopaedic content of the original is preserved. One
way to achieve TT in the case of culture-specific expressions is to employ a
procedure called transference. Transference, as Newmark (1988: 81) puts it, is
“the process of transferring a SL word to a TL text as a translation procedure”.
This is essentially the same as Catford’s definition: “an operation in which the
TL text, or, rather, parts of the TL text, do have values set up in the SL: in other
words, have SL meanings” (Catford 1965: 43, italics as in original). In simple
words, this is when we decide to incorporate the SL expression unchanged into
the TL text because this makes possible the recovery in the target text of some
assumptions, even though at the cost of an increased level of processing effort,
which would not otherwise be accessible in the target cultural context. In some
cases TT can also be achieved through literal translation, but in a secondary
context a literal translation will commonly result in the loss of some
encyclopaedic assumptions.
(2) [+L, −E] is called logical transfer (LT). In LT only the relevant logical
content of the original is preserved. This can be implemented by a simple literal
translation, in the proper sense, which means the translation of the logical
content of an expression. This kind of translation proper, then, is the procedure
of using a “dictionary equivalent” of the original. In relevance-theoretic terms
this means rendering the SL expression by a TL expression which, by preserving
the logical content of the original, gives rise to the same relevant analytic
implications in the target text as the original did in the source text. For example,
we can literally translate Hungarian Szabad Nép, the title of a newspaper, into
English as Free People. Such a solution, however, will not normally activate in a
secondary context the encyclopaedic assumptions associated with the original,
since in a secondary context we are dealing with the lack of certain relevant
background assumptions. This means that some contextual effects may get lost
in the translation. Such losses in translation are acceptable when the preservation
of a contextual effect would require too much processing effort from the target
reader.
(3) [−L, +E] is called encyclopaedic transfer (ET). In ET only the relevant
encyclopaedic content of the original is preserved. ET can be performed, for
instance, by a procedure called substitution. By substitution we will refer to
those cases when the source language expression is replaced in the translation by

105
a target language correspondent that is different in terms of logical content but
carries with it encyclopaedic assumptions which are the same as, or similar to,
those of the original. For example, the Hungarian expression bagóért eladni
valamit can be substituted by English sell something for a song. In a relevance-
theoretic framework we could say that an expression that is substituted in this
way, by directly activating relevant contextual assumptions in the target context,
is the one that requires the least processing effort from the target audience and
any other solution, increasing the amount of processing effort, would need to be
justified by a substantial gain in contextual effects.
Another way to implement ET is to use explicitation. Explicitation is the
procedure of making explicit in the target text an assumption which is only
implied by the source text. For example, the Hungarian expression HÉV, which
is an acronym for helyiérdekű vasút (‘local railways’), can be explicated in
English as a commuter train. Similarly, Hungarian szabadságharc can be
explicated in English as the war of liberation in 1848.
ET also subsumes a procedure called transcription, where the graphological
units of the source language expression are replaced by target language
graphological units, based on conventionally established correspondences. In
such cases of graphological substitution the target language form makes explicit
the phonological value of the original expression. One example is when in the
Hungarian translation of Fennimore’s The Last of the Mohicans the name
Chingachgook is rendered into Hungarian as Csingacsguk. The inclusion of
transcription within ET is justified by the fact that its application is motivated
mainly by considerations of optimising processing effort.
A similar procedure is what we call transliteration. Transliteration is the
procedure of substituting the graphological units of one writing system by
graphological units of another writing system. This happens, for instance, when
Hungarian Moszkva is substituted for Russian Mocквa.
(4) [−L, −E] is called zero transfer (ZT). In ZT neither the logical content
nor the encyclopaedic content of the original is preserved in the translation. This
can be achieved, for instance, by modifying the meaning of the original.
Modification is the procedure of choosing for the source language expression a
target language substitute which is semantically unrelated to the original. In
relevance-theoretic terms this means replacing the original with a target
language expression which involves a substantial alteration of the logical and
encyclopaedic content of the source language expression. For instance, such
modification occurs when the Hungarian expression közért, meaning something
like ‘grocery shop’, is rendered into English by the market. The logical meaning
is obviously different, needing no explanation. As for the encyclopaedic content,
one has to know that the Hungarian expression came into use during the socialist
era, and this piece of information is part of the encyclopaedic entry of the word.
The English expression, however, carries no such encyclopaedic assumption.

106
 ZT can also be performed by simply leaving out a segment of the source text.
This procedure is called omission. For instance, this happened when in the
English translation of Péter Esterházy’s book Hrabal könyve the translator
decided to omit the expression önkéntes rendőr.
Modification and omission are clearly procedures which are aimed at
minimising processing effort, even if it means losing some relevant assumptions
and, consequently, contextual effects.
We thus have a simple set of four general translation operations, which can
be realised through various procedures. Total transfer is an operation which
preserves both the relevant logical and encyclopaedic content of the original
expression, logical transfer preserves the logical but not the encyclopaedic
content, encyclopaedic transfer preserves only the encyclopaedic content and,
finally, zero transfer preserves neither. In general, the use of total transfer is
motivated by an attempt to maximise contextual effects. Zero transfer is
motivated by an attempt to minimise processing effort. Logical transfer and
encyclopaedic transfer can be aimed at optimising the relevance of the
translation by striking a balance between contextual effects and processing
effort.

6.3.2. Translation strategies

A translation strategy is a general plan of how a given translation task will be

carried out. With regard to cultural differences, two translation strategies are
distinguished in the literature as the two essential ways to render a source text
into a foreign culture. They are commonly termed foreignising (also called
alienating) and domesticating (or naturalising). In Venuti’s words,
domesticating is an assimilationist approach, conforming to the dominant
values of the target culture, while foreignising is “motivated by an impulse to
preserve linguistic and cultural differences by deviating from prevailing
domestic values” (Venuti 1998: 241). In this sense, a domesticating translation
will typically alter, or even cancel out, assumptions which are absent from, or
alien to, the target cultural context, thereby minimising the processing effort that
the target reader needs to exert in interpreting the target text. On the other hand,
a foreignising translation will aim at preserving such assumptions, thereby
making it possible for the target reader to access the originally intended
interpretation, even at the cost of a higher level of processing effort. The higher
costs of processing in such cases will ideally be counter-balanced by the increase
in contextual effects. It would appear, then, that the implementation of a
foreignising approach will significantly involve the use of total transfer, whereas
a domesticating strategy will crucially involve zero transfer of source language
expressions into the target language text.

107
 6.4. The use of the four operations in implementing strategies

The translator’s strategy concerning the given translation task may thus be traced
down through regularities in the translator’s use of the different operations. For
the purposes of the discussion, we will use examples from Péter Esterházy’s
Hrabal könyve (Magvető Kiadó, Budapest, 1990) and Judith Sollosy’s English
translation (The Book of Hrabal. Quartet Books, London, 1993). The various
culture-specific expressions in this book can be categorised into nine classes,
which are the following: administrative culture, social culture, intellectual
culture, material culture, persons, topography, situation schemas, history and,
finally, units and measures. Although these categories have no theoretical
relevance, as we saw in Section 6.2, they will be useful in describing what
happens in this translation.
The most frequent operation in the translation is encyclopaedic transfer,
followed by total transfer, logical transfer and, finally, zero transfer. This would
suggest that the translator’s choice was a basically foreignising approach to the
translation. What is interesting still is that (a) encyclopaedic transfers and logical
transfers occur in a relatively great number in the material and intellectual
culture categories, (b) the persons and the topography categories are dominated
by the use of total transfer, and (c) the situation schemas category is
characterised by the excessive domination of encyclopaedic transfer. However,
to check out what these facts actually mean, we will need to look behind the
numbers and see what we can learn from the individual examples.

6.4.1. Total transfer

The use of total transfer in the translation dominates two categories, those of
expressions referring to persons and topographic features. These expressions,
being among the most numerous categories in the text, can be identified as the
prime indicators of the cultural and physical setting of the story, and are thus
mainly transferred, with very few exceptions, to provide for the accessibility of
the appropriate background assumptions concerning the setting of the story. In
one extreme case even the common noun head of a street name is transferred
(Váci utca), presumably because it marks one of the best-known places in
Budapest and is supposed to figure as a unit in the target reader’s cognitive
environment. The exceptions are either simple mistakes, as with the
modification of three personal names (Bólyai, in the translation, for Bolyai,
Odon Suck for Sück Ödön Mihály and Dansco for Dancsó) or are due to the
relevance of the logical content of the expression, as with the logical transfer of
three topographic expressions (Inner City for Belváros, or Black Forest for
Fekete-erdő). Other transferred expressions can be found in the categories of
administrative culture (ÁVÓ, discussed below), history, material and intellectual

108
culture, social culture and units and measures, all contributing to the
preservation of the original spacial, temporal and cultural setting for the story,
serving thus as tools of foreignising.

6.4.2. Logical transfer

Logical transfer is a means of preserving the logical content of the original

expression, in order to ensure that the translated utterance gives rise to the same
analytic implications as the original, irrespective of what contextual assumptions
may be activated by the original. This can be the most obvious solution when the
source expression activates some relevant encyclopaedic assumptions which,
however, cannot be preserved in the translation in an effort-effective way. Thus,
in the English version the pronoun you is used for both maga and önök, which
share the same logical content but are loaded with different stylistic values in
terms of the formality of the relationship between speaker and listener. Other
examples have to do with assumptions about the socialist past of Hungary. For
instance, the Hungarian expression társadalmi munka becomes community work
in the English text; and Hungarian tanács, which roughly corresponds to ‘local
authority’ in the source text, is rendered into English as council. Similarly, in the
English text we have counter-revolution for the original ellenforradalom, which
for Hungarian readers also carries the ecyclopaedic assumption that it refers to
the events of 1956.
On the other hand, there are also cases where translation proper is used to
introduce completely new concepts into the target reader’s cognitive
environment. This happens, for example, when Hungarian paprikás krumpli (a
typically Hungarian dish) is rendered as paprika-potato in the translation.
There is another interesting example (a part of which we have already taken a
look at in the previous chapter) which shows that logical transfers, when
combined in delicate ways, may also serve to preserve culturally induced
implicatures through the extension of the context.

Volt cukrászda, két konkurens (1a) kocsma, melyet mindenki a régi

nevén hívott, a (2a) Serház meg a Kondász (az öreg Kondász még élt,
asztala volt a sarokban, és (3a) pintenként rendelte a sört, amiről a
gyakran cserélődő csaposok ritkán tudták, mennyi, hát, fiacskám, egy
(4a) korsó meg egy vágás! […]). (Esterházy, p. 10, italics as in
original)

The English translation reads in the following way:

There was a café of sorts and two rival (1b) taverns, which everyone
called by their old names, the (2b) Beerhall and the Kondász (old man

109
 Kondász was still kicking, he had his own table in the corner and
ordered beer by the (3b) pint, an unknown quantity for the succession
of ever new barkeepers, it’s a (4b) pitcher and a dash, son! […]).
(Sollosy, p. 4, italics as in original)

Remember that the problem here is that the Hungarian word ser in (2a) Serház,
the original of (2b), is associated with an encyclopaedic assumption to the effect
that the expression is old-fashioned, it is not used any longer, and evokes the
atmosphere of “the golden days” of the Austro-Hungarian Monarchy. However,
the English word beer does not carry this assumption and this part of the context
is thus lost in the translation. Thus, while (2b) is a close enough rendering of the
original in terms of logical content, part of the context is lost. However, going
back to (1b), we see that the target expression, tavern, compared with the
original expression, (1a) kocsma, meaning something like ‘a cheap pub’, gains in
encyclopaedic content in just the opposite way: it activates assumptions relating
to the past, whereas the original does not. Thus the translation in (1b) serves the
purpose of compensating for the loss of contextual assumptions later in (2b). The
same can be observed in relation to (3b) and (4b), where the English word pint
misses the encyclopaedic assumption carried by the Hungarian original in (3a)
that it is an obscure measure not used any more, but the word pitcher brings in
encyclopaedic assumptions about long-gone days, not activated by the
Hungarian original korsó, which simply means ‘beer mug’.

6.4.3. Encyclopaedic transfer

When a particular source language expression activates a culture-specific

assumption, it can be substituted by a target language expression which is
different in logical content but carries with it the relevant encyclopaedic content.
The substitution may effect a partial change of logical content, as in the case of
English bologna substituted for Hungarian parizer, which is a kind of cheap cold
meat. In other cases there is a complete change of logical content, as with
English (union) dues substituted for Hungarian (szakszervezeti) bélyeg, meaning
‘(union) stamp’.
Substitution is dominant in the situation schemas category, which is no
surprise. Situation schema expressions include phraseologies, idioms, proverbs,
conventional metaphors and the like, all being characteristic ways of how
members of a culture categorise the wide range of possible situations. This is
obviously an area of cognition where cultures tend to be very different.
Moreover, these schemas are so deeply rooted in the thinking of people and are
so easily activated in the proper context that to exchange them for different ones
would surely result in a great amount of extra processing effort. In these cases, it
is not the logical content of the expression which carries the relevant information

110
but the encyclopaedic assumptions which are activated by the expression and for
this reason, almost all examples of such expressions in the original are
substituted by ones native to the English cultural context. This is how we get
English simple as a pie as a substitute for the Hungarian expression
pofonegyszerű, meaning ‘simple as a slap’. In the same way, expressions
activating assumptions relating to social relations and attitudes can be
substituted: we have in the English target text the word dear standing for
Hungarian fiam, which literally means ‘my son’ but is used to address (in certain
social circles) a spouse.
Also often substituted are expressions relating to topographic objects which
have their own names in the target culture. Thus we have English Danube
standing for Hungarian Duna. What is important in such cases is that the
reference remains invariant, and since the reference here is determined not by
the logical entry, which may be empty in the case of proper names, but by the
encyclopaedic entry, it will take less processing effort to recover the referent
through an expression whose encyclopaedic content is readily accessible for the
target reader.
For the same reason, substitution is prevalent with the full names of persons
in the translation. In Hungarian, the order of names is family name first, first
name second, and since English readers are not supposed to have access to this
assumption, the reversed order is substituted in each case. Thus we get Laci
Bárány in the English text for the Hungarian original Bárány Laci, with the
constituent elements of these names transferred.
In a somewhat similar fashion, when a name in the original activates an
encyclopaedic assumption which is not likely to be present in the target cultural
context, the relevant assumption can be provided by the translator in the form of
an explicitation, combined with transference. This is how the Hungarian name
Károlyi Mihály becomes former prime minister Károlyi in English or Petőfi is
rendered into English as the poet Petőfi. This solution serves to spare the target
reader from some extra processing effort.
Another such example is provided by the expression ávó, short for
államvédelmi osztály, meaning ‘state defence department’, which occurs in three
different renderings in the target text. The first occurrence is an explicitation,
combined with transference (secret police ÁVÓ), the second is an explicitation
(secret police), and the third a simple transfer (ÁVÓ). This then suggests that
encyclopaedic transfer can be used in ingenious ways to lead the readers toward
the source culture by smuggling into their cognitive environments assumptions
which originate in the source culture.
One further interesting example is provided by the following sentences.

(5a) Az, ami az amerikaiaknak a blues, az a magyaroknak a keserves –

(6a) erre a felismerésére büszke volt, ezért, és nem másért, szerette

111
 jobban a kurucokat a labancoknál. (Esterházy, p. 158, italics as in
original)

(5b) The keserves, or lamenting song, means for the Hungarians what
the blues does for Americans. (6b) He was proud of this discovery, and
for this and for no other reason did he prefer the Kurucz to the
Labancz. (7) The anti-Habsburg Kurucz soldiers knew how to cry into
their wine, not like those pro-Habsburg Labancz. (Sollosy, p. 139,
italics as in original)

The Hungarian expressions keserves, kuruc and labanc are first all transferred
(although because of the slight spelling change it could be argued whether the
latter two are substitutions, rather), but then the translator, feeling a need to
explicate some background assumptions, added the expression lamenting song in
(5b) and added sentence (7), which does not occur in the original but makes
explicit an encyclopaedic assumption implicit in (5a) and (6a). Clearly, the
explicitations took place here because the assumptions that they make accessible
are necessary for working out the relevance of (6b), and since the target readers
do not have access to these assumptions as part of the encyclopaedic entries of
keserves, kuruc and labanc, the translator probably thought the readers need help
in order that the necessary processing effort is not unreasonably high.

6.4.4. Zero transfer

Zero transfer seems to occur for two main reasons. It may be an obvious solution
when a concept is missing from the target culture and the preservation of the
logical content would entail an increase of processing effort not justified by the
gains in contextual effects. For instance, the Hungarian expression önkéntes
rendőr, meaning ‘voluntary policeman’ is omitted in the translation, because in
the target culture there is no comparable institution and the concept is not vital in
terms of the development of the story. Thus the translator decided that the loss in
contextual effect is more tolerable here than the potential increase of processing
effort which would result from the preservation of the expression. In other
instances we find that the translator renders the original by an expression
activating a completely different concept, but which, being familiar to the target
reader, requires less processing effort. Thus we have in the English translation
shoe repair shop for Hungarian harisnyaszemfelszedő, meaning ‘stockings
mender’, or the market for Hungarian közért, meaning ‘a kind of grocery shop’.
The other typical case is when some encyclopaedic assumption is not present
in the target cultural context and the relevance of the utterance can be ensured in
the most cost-effective way by modifying both the logical and the encyclopaedic
content of the original. For example, the English text has the expression Silly

112
Billy for Hungarian Bunkócska te drága, which is a reference to a Russian song
that was well-known in the Hungary of the socialist era.

6.5. Conclusions

As for the use of the different operations in implementing strategic decisions, the
examples above seem to prove that total transfer, motivated by an attempt to
preserve the contextual effects of the original, serves as the essential means of
foreignising, while zero transfer basically serves the purposes of domesticating,
since its use is sanctioned primarily by the need to minimise the level of
processing effort. The other two operations, logical transfer and encyclopaedic
transfer, are such flexible tools in the hands of a creative translator that can be
used to work in either direction.
What a closer look at the examples suggests is that the particular target text
examined here is fairly balanced in the sense that while it reveals a strong
overall inclination toward the foreignising strategy, this is not accomplished in a
rigid manner and it gives way to domesticating procedures when their use seems
more cost-effective. Naturally, in a secondary communication situation the ideal
of direct translation, which could only be achieved, if at all, through an
uncompromising foregnising strategy, is not a realistic aim. It makes a lot more
sense to accept that the differences between the cultural contexts will inevitably
lead to losses in translation and to try and do the best one can in such a situation:
compromise and let go of certain informative intentions of the original in favour
of other more directly relevant ones that can be saved.

6.6. Study questions

After reading the chapter, answer the following questions. If you cannot
remember, go back and revise the related part.

What is a culture-specific expression in a trivial and in a non-trivial sense?

Why is it better not to use the term cultural realia?
What translation operations can be distinguished on the basis of the semantic
structure of concepts?
What basic choices does the translator have to make in relation to culture-
specific expressions? What are the basic translation strategies in relation to
cultural differences?
How can these strategies be implemented by the use of the four basic
translation operations?
How can the basic translation operations be realised by the use of the various
translation procedures?

113
 7. TRANSLATION COMPETENCE: WHAT MAKES A GOOD
TRANSLATOR

The Longman Dictionary of Contemporary English defines competence as “the

ability and skill to do what is needed.” Thus a simple general definition of
translation competence could be the following. Translation competence is the
ability of a person to produce an acceptable translation of a source text on the
basis of an internalised pool of knowledge and skills relating to translation.

7.1. The minimalist concept of translation competence

However, as Pym (2003: 489) argues, this kind of general competence in

translation arises as the function of two more specific skills, which uniquely
define translational competence. Translation competence in the minimalist
sense is the union of the following two skills: (a) The ability to generate a series
of possible target texts for the same source text, and (b) the ability to select from
among these only one viable target text, quickly and confidently.
In the end, translation is a problem-solving process, involving two cognitive
tasks: generation and selection. Of course, we need to keep in mind that text
generation in translation is different from monolingual text generation. In
translation the target text is generated on the basis of a mental representation that
is the result of the interpretation of a source text. In monolingual text generation,
on the other hand, the text is generated from a mental representation that is not
based on the interpretation of an external source.
Pym provides the following example of how this problem-solving occurs in
translation. He gave his students the task of rendering into English this Spanish
sentence:

Los alumnos que hayan estudiado en el extranjero y deseen iniciar

estudios en las universidades españolas deberán convalidar u
homologar sus estudios.

This was translated into

Students who have studied outside Spain and wish to enter a program
at a Spanish university must convalidar or homologar their foreign
studies.

The problem here is with the two verbs in italics, for which no English
equivalents exist. The English term accreditation covers the meaning of both,

114
the Spanish terms being more specific. Basically, homologación means the
accreditation of degrees and diplomas, while convalidación means the
accreditation of actual study courses. How can an English translator use
accreditation to preserve the two different meanings? The solution rests on
considering the needs of the future target readers, the foreign students. If they do
not need their degrees or courses to be accredited, then the two Spanish terms
can simply be merged under the umbrella term accreditation. On the other hand,
if the foreign students do need accreditation, then they will also certainly need to
know the terms in Spanish, along with the description of the bureaucratic
processes that they designate, so in this case the original terms will have to be
preserved in the translation and explained in English.
This example nicely illustrates how a translator needs to be able to produce
alternative solutions and choose between them on the basis of a careful
consideration of the relevant factors of the secondary communication situation.

7.2. Complex translation competence

Looked at from a wider angle, translation is an even more complex activity, also
involving textual interpretation in context, text analysis, task analysis,
researching etc. Thus, in this light, being competent as a translator involves
several different kinds of competences. The various sub-competences needed in
translation combine together to form a complex translational competence.
According to Klaudy (2003: 121), there are five such sub-competences:

(a) linguistic competence,

(b) subject-related competence,
(d) intercultural competence,
(e) transfer competence and
(f) communicative competence.

7.2.1. Linguistic competence

Linguistic competence involves knowledge of the source and target languages.

In the Chomskyan sense, it can be defined this way: Linguistic competence is
the language user’s ability to create and understand an infinite number of
grammatically correct new expressions.
It goes without saying that professional translation presupposes at least a
college-level degree, or an equivalent level of knowledge, in one’s languages.
But even that will not qualify one to translate between these languages in any
direction. We must make a distinction between the language that we translate
from and the language that we translate into. Normally, one translates from the
foreign language into the native tongue. This is because while one’s ability to

115
read in the foreign language may be excellent, this is not necessarily paired with
the same level of writing ability. One is seldom so familiar with a foreign
language as to be able to write with the necessary sophistication that is expected
in professional communication. It is not impossible, of course, but it normally
takes a lot more study and experience than one can acquire at college or
university. To achieve that level, in most cases one needs to spend an extended
period of time in the target language country or countries. In the European
Union, as we saw in Chapter Two, the normal practice is to translate into the
mother tongue.
Importantly, knowledge and command of a language is not something that
one can acquire for once and for all. Languages are like living organisms: they
keep changing day by day. As Owens (1996) so aptly puts it, they are in a
constant state of flux, never ceasing to change. Her example is the word gay.
Thirty or so years ago it simply meant ‘joyous’. Today this meaning has become
secondary and to most people the word means a particular kind of sexual
orientation. But such changes of meaning are only one example. New words also
appear on a daily basis. Who would have known such words as anime, blog, e-
mail etc. a few decades ago?
Linguistic structures, and their use, can also be seen to change, although a lot
more slowly. For instance, probably under the influence of English, the lingua
franca of our age, the use of the Hungarian suffix -hat/-het seems to be
changing. For instance, a newspaper headline said the following about a German
footballer’s possible transfer to England: “Ballack a Chealsea-hez igazolhat”.
The first interpretation that comes to mind is that Ballack has permission to
transfer to Chealsea. But then it turns out from the article that the actual
interpretation intended by the author is that it is possible that Ballack will
transfer. This is because the article was obviously a translation from English,
and in journalistic Hungarian the English auxiliary may, even in its possibility
meaning is more and more translated by -hat/-het, which used to be used in this
sense less often before English came to have such a dominant role.
Thus one thing that a good translator cannot stop doing is learning his or her
languages. And this applies to the mother tongue as well. Even one’s knowledge
of the mother tongue is never perfect. A good translator is an eternal language
learner, and probably the best way of perfecting our knowledge of languages
above a certain level is reading all the time, as much as possible, and as widely
as possible.

7.2.2. Subject-related competence

Knowledge of two languages, however, is only part of the knowledge that is

necessary for translating well. It is a common misconception that anyone who
speaks two languages well is automatically a competent translator. If you have

116
ever tried yourself in translation, you will remember the first painful experience
of, for instance, not being able to find the appropriate expression in the target
language, although you have thought you have a good command of both
languages involved. Or not being able to understand the source text, for that
matter. Probably every novice translator acquires such memories. One of the
first translation tasks the author of this book undertook was the translation of a
text on dental science from English into Hungarian. It was a failure, to a large
degree, on two accounts. First, the source text did not make sense. Although it
was written in English, it was written about a subject that I was not the least
familiar with. And there was also the problem that I could not produce an
appropriate Hungarian target text, because I was also a complete stranger to the
medical register of Hungarian. Yet this experience proved useful in the end,
because it taught me a very important thing: you have to be clear about your
limits concerning the subjects that you can translate. This will save you from the
humiliation of having to return a task unfinished to the commissioner. And,
consequently, it will save you from having a good professional reputation
seriously damaged.
This essential familiarity with the subject matter is called subject-related
competence. Subject-related competence is the ability of a person to
understand the concepts and conceptual relationships of a given field of
knowledge.
Here is one example of how lack of knowledge in a subject field may result
in a serious translation error. This excerpt was taken from a product description
leaflet.

After the implementation of the European Directive 2002/96/EU in the

national legal system, the following applies […]

It was translated into Hungarian in the following way:

Az európai irányelvek 2002/96/EU ajánlása szerint, a megjelölt

időponttól kezdve, minden tagállamban érvényesek a következők […]

It is quite obvious that the translator did not have a clue about the meaning of the
term European Directive and how a directive functions in EU law. He or she
tried to guess the meaning but, as often happens, the guess was mistaken and the
whole sentence went astray.
For another example, consider the following excerpt from the film Ghost Dog
(directed by Jim Jarmusch, Plywood Productions, 1999). Even if someone is a
highly competent user of English, they will not necessarily be able to understand
this piece of text:

117
 Our bodies are given life from the midst of nothingness. Existing
where there is nothing is the meaning of the phrase, “Form is
emptiness.” That all things are provided for by nothingness is the
meaning of the phrase, “Emptiness is form.” One should not think that
these are two separate things.

Without some background knowledge of Buddhist philosophy the translator will

not be able to grasp the concepts and relations that the text builds on and will
consequently produce a mistaken translation like the following, taken from the
Hungarian version of the film, shown under the title Szellemkutya:

A testünk a semmi közepéből keletkezett. Létezni a semmiben, ezt

jelenti „a forma üresség” megállapítás. Amikor minden azért van, hogy
a semmibe jusson, életre kel az „ürítsd ki a formát” kifejezés. Tévedés
azt hinni, hogy ez két különböző dolog.

A more precise, although rough, translation of the original would be this:

Testünk a semmi közepéből kapja életét. Ott létezni, ahol semmi sincs,
ezt jelenti „a forma nem más, mint üresség” kifejezés. Hogy minden
dolog a semmiből származik, ezt jelenti „az üresség nem más, mint
forma” kifejezés. Nem szabad azt gondolnunk, hogy ez két egymástól
független dolog.

Ideally, in specialised translation, one is not only a linguist but also a specialist
of the given field. This would require that a legal translator should also have a
degree in law, a medical translator in medicine, etc. This, however, is not always
practical. As a matter of fact, if this was the case, it would mean that a lot of
time and money is wasted on training lawyers and doctors who will then never
practice their professions. So in real life, specialised translators are not very
frequently trained specialists – but they do have to possess the necessary amount
of knowledge in the field that enables them to grasp the meaning of specialised
texts, on the one hand, and to produce acceptable target texts in the given
register, on the other.
Thus it is not necessary, as Tarnóczi (1966: 273) writes, for translators to
have the detailed knowledge of a specialist, but they must be widely read in the
subject and well-informed enough to be able to understand the specialised text.
A good translator may not be familiar with the minutiae of any subject but has
an overall familiarity with the essence of several areas of knowledge (Tarnóczi
1966: 274). What this requires from the specialised translator, again, is
willingness to learn and expand.

118
 The term specialised translation is sometimes used interchangeably with
technical translation. However, for the sake of clarity, it will be useful to make a
distinction between these two genres of translation. Technical translation has a
narrower meaning, which is included in the meaning of specialised translation.
Technical translation is the translation of texts written in the context of
scientific or technological disciplines. Specialised translation is the translation
of texts in specialised fields such as law, economics, politics, commerce,
finance, government etc.
What is common in technical translation and specialised translation is that
both involve texts using specialised language, including a special vocabulary of
technical terms. A technical term is an expression which has a clearly and
unambiguously defined meaning within a given field of knowledge.
The distinction between technical and specialised translation is useful
because it helps the translator to remember an important difference between the
two genres. This difference is that while science and technology are truly
international, other fields of knowledge, such as law or administration, are
always linked to a particular culture. Thus, for instance, no two legal or political
systems may be perfectly mapped onto each other.
Consider for example the British English words solicitor and barrister. They
both refer to lawyers who can represent people in law courts, solicitors in lower,
and barristers in higher courts. The problem is that there is no such distinction in
the Hungarian legal system and consequently there are no Hungarian equivalents
for these terms. We can only use the superordinate term ügyvéd for both, with an
explanatory note, if necessary.
As another example to illustrate how differences between legal cultures may
be manifested in language, we can take the word impeach as an example, which
has different meanings in Britain and the United States. The British meaning is
‘to charge somebody with a crime against the state’. The American meaning is
‘to charge the holder of a public office with misconduct’. Thus, when choosing a
proper target language correspondent, the translator needs to be aware of the
difference between the cultural contexts.
Or think of the different political function the President of Hungary has
compared to that of the President of the United States of America. Or compare
the presidential system of America with the system of constitutional monarchy
in Great Britain. Similar problems will also arise at the lower levels of
administration. For instance, how should we translate into Hungarian the British
English term Lord Mayor? Certainly not as főpolgármester. The Lord Mayor of
the City of London has a very different function from that of the főpolgármester
of Budapest. One possible way to render the expression is to transfer it into the
target text in the original form, supported by an explicitation of its relevant
encyclopaedic content.

119
 Consequently, while we can say that technical translation is non-cultural in
the sense that it does not involve problems arising from the differences between
the cultural contexts of the source and the target communication situations, in
specialised translation such problems are rather frequent and are to be expected.
This way, specialised translation also requires that the translator should be
familiar with both cultural contexts.

7.2.3. Intercultural competence

Intercultural competence is the translator’s ability to notice the differences

between the source culture and the target culture and to successfully bridge those
differences with a view to the target audience.
In most forms of translation, knowledge of the two cultures involved is just
as important as knowledge of the languages that members of the two cultures
speak. Translation in this sense is not merely linguistic mediation but a process
of communication between two different cultures. And failure to notice the
differences between these cultures may lead to misunderstanding and confusion.
Just as you may experience culture shock when you travel to an unfamiliar
country, where you feel at a loss because things are so different, a translation
that does not consider the cultural differences will leave the target reader baffled
about those things. In order to avoid such “culture shocks”, the translator must
act not only in his or her capacity as a linguistic aide but also as a “travel guide”.
Here again, it is important to see that to successfully perform this function the
translator needs to keep up with the changes that occur in both cultures. In this
age of technological boom, objects, ideas, opinions, events, fashions, life styles
etc. change so rapidly that if you were to fall into a twenty-year coma, you
would find yourself in an alien world when you wake up. To keep abreast of the
changes, a translator has to read newspapers and magazines, watch TV, browse
the Internet and use any other sources of information on a regular basis. By
doing so, of course, translators can keep themselves up-to-date not only about
the cultural changes but also about the resulting changes in language. Language
and culture, as we know, are so intertwined with each other that we cannot learn
one without the other.

7.2.4. Transfer competence

Transfer competence is the ability of a translator to develop mental habits to

overcome problems resulting from the differences between the two languages
and cultures involved. The ability to transfer texts implies that the translator has
acquired not only knowledge of the two languages but also a contrastive
knowledge of them, and is aware of their differences. This kind of knowledge is
not usually considered part of bilingual competence. This knowledge is the basis

120
of the mental procedures that through practice turn into habits in the mind,
stored in procedural memory. Such internalised translation procedures arise as a
result of regular work with a particular language pair. If the translator works in
both directions between his or her A and B language, or also has other
languages, these language-pair specific transfer procedures may give rise to
language-pair independent general transfer procedures. This is one pivotal
capacity that distinguishes the professional translator from the average bilingual
speaker: the translator has the ability of easy and effortless use of transfer
procedures, while the average speaker does not.

7.2.5. Communicative competence

Translation, as we have seen, is a special form of communication and thus

translators are also communicators. To function well as translators, they need to
be competent as bilingual communicators. Communicative competence is the
ability of a person, over and above the basic language decoding and encoding
skills, to quickly and easily produce such utterances that are pragmatically,
functionally and socially appropriate in a given context.
Since translation is a bilingual form of communication that involves the
reading of the source text and writing of the target text, it follows that the
translator has to be a good reader as well as an effective writer. Here is one more
point that distinguishes the translator from an average bilingual person: not
everybody is a good reader or an effective writer, not even in their mother
tongue. Reading and writing for effective communication are two pivotal skills
for a translator.
People tend to forget that translation begins in the mind, when you read the
source text. Without a detailed understanding of the original, however, there is
no good translation. Therefore, the proper methods of reading are also part of the
communicative competence of the translator.
Another general mistake is to suppose that every native speaker of a language
is a competent writer in that language. This is obviously false. And not only
because not everyone is born with the facility to write in a clear, fluent and
intriguing manner, but also because you can only write well if you are confident
in the grammar, punctuation and spelling of your language. The language
checking function of the word processing program can help, of course, but it will
not solve every problem for you. (In fact, it often creates problems itself.) Thus
it is absolutely necessary for a translator to be thoroughly familiar with the
grammatical rules and norms of the target language and to have reliable
reference works available.
Communicative competence in general can be seen as covering three sub-
competences called pragmatic, discourse, and sociolinguistic competence.
Pragmatic competence is the communicator’s ability to adapt the intended

121
message to the context that is available to the audience in such a way as to make
possible the optimal processing of the linguistic stimulus. Discourse
competence is the ability of a person to perceive and produce linguistically
cohesive and cognitively coherent texts designed to achieve a communicative
goal. Sociolinguistic competence is the ability of a person to judge the
appropriateness of utterances in particular social situations.

7.2.6. Knowledge of translation theory

Various other authors distinguish further sub-competences needed in translation.

Perhaps the two most important of these are what we call instrumental
competence and extra-linguistic competence. Instrumental competence is the
pool of skills relating to professional translation practice, including knowledge
of the tools of the translator, research skills, typing skills etc. Extra-linguistic
competence is the ability of a person to apply knowledge of the mental and
physical world in translation. This means all sorts of knowledge stored in
conceptual memory, including knowledge of the subject or the cultures involved,
as we discussed above, and also including knowledge of translation theory.
There is a tendency to underestimate the value of theory to practice in
translation. While it is true that translation is a practical activity and is best
learned through practice, this does not mean that theory cannot help. Building is
also a very practical activity and yet no one denies the usefulness of theoretical
studies in the training of architects. Perhaps this parallelism is not perfect, since
translation is not an exact science, but in what follows we will try to summon
some arguments to underpin this statement.
Machine translation systems have become more and more sophisticated and
useful in recent years and are used to translate ever greater amounts of text.
However, the quality of machine translation, although it has improved
immensely since the beginnings, is still far from satisfactory in all respects. Even
a moderately competent human translator can still produce a better translation
than the best machine translation system. Yet novice translators often translate in
a way that makes their translations very much like those produced by machines.
Here is one interesting example from Chesterman (1996). The Finnish source
text was used as the general text in a Licensed Translator Exam. It is about how
Finnish municipalities tried to overcome an economical crisis. The supposed aim
of the target text was to appear in a fictive local government publication. One
segment of the original text reads as follows:

Sekä tilastoihin ja kunnissa tehtyihin haastattelututkimuksiin

perustuvat selvitykset että puhtaasti valtikunnallisiin tilastoihin
perustuva palvelukatsaus osoittavat terveydenhuollon tehostuneen…

122
An awkward interlineal English translation would be the following:

Both (the) statistics-(based) and in municipalities carried-out interview

studies-based reports and (the) purely national statistics-based service
review show (that the) health service improved.

A communicatively acceptable English translation of this can be the following:

Municipal statistics and interview studies, and also national statistics,

show that Finnish health care has become more efficient.

Now read the following eight English versions. Seven were produced by people
taking the translation exam and one by a machine translation system called
TranSmart. Can you find out which translation was done by the machine?

(1) Reports based on both statistics and interviews in municipalities,

and service reviews based on national statistics show that health care
has become more effective.

(2) Both the surveys based on statistics and interviews carried out in
municipalities and the survey of services which is based on purely
nationwide statistics show that Finnish health care has become more
efficient.

(3) Studies based on statistics and interviews done in the municipalities

and a service review based on national statistics both indicate that the
health services have improved.

(4) Studies based on statistical and interview research conducted in

municipalities, as well as a service survey based purely on national
statistics, demonstrate that the efficiency of health care has improved.

(5) A study on services which is based on statistics as well as

interviews made in the communes and on purely nationwide statistics
shows that health care has become more efficient.

(6) The effectiveness of public nursing has improved. This is shown

both by reports based on statistics and local interviews and by a survey
based on strictly national statistics.

(7) The health service has become more efficient according both to
statements based on statistics and interview researches done in the

123
 municipalities and on a service review based on purely national
statistics.

(8) Both the reports which are based on the statistics and on the
interview studies that have been done in the municipalities and the
service review which is purely based on national statistics show that
the public health service has improved.

If it is not easy to spot the translation done by the machine, which is (8), it is
only partly because the machine translation system is relatively well-developed.
The other reason is that the examinees were novice translators, who in several
respects seem to translate like a machine: word-by-word, grammatical
construction-by-grammatical construction. Such target texts, while making
sense, are not very good translations because owing to their closeness to the
original source language formulation they are not particularly easy to process for
the target language reader.
Why do these novice translators translate like a machine? According to
Chesterman, it is because they, like the machine, do not have access to the
theoretical concepts that professional translators do. It seems that lack of access
to these theoretical concepts affects the quality of their translation in a negative
way. In other words: there is a relation between knowledge of translation theory
and professional translation expertise. Baker (1992: 2) says that the main value
of theory is that it prepares us for dealing with unpredictable phenomena, gives
us confidence, and makes possible professional development on the basis of a
formalised pool of knowledge shared by a professional community.
Chesterman offers a short list of simple theoretical concepts in possession of
which these novice translators would have been able to produce better
translations of the original Finnish sentence.
The examinees would probably not have translated the original in a word-by-
word fashion if they had had the concept that we call deverbalisation.
Deverbalisation is the process when translators distance themselves from the
surface structure of the source text to arrive at the assumptions intended by the
source communicator and then express this intended meaning in the target
language. The aim of deverbalisation is to avoid the mechanical reproduction of
source-text structures. The important question to ask yourself as a translator is
this: If I were to express the same idea in the target language without a source
text around, would I express it in that way or is there a more natural way to do it
in the target language? Translations (6) and (7) above exhibit some signs of
devarbalisation: in both versions the concluding short nominal subclause of the
original has been fronted and in (6) it has been turned into an independent
sentence. These changes probably took place in order to make the processing
easier for the target reader.

124
 Another useful concept is a technique called transposition. The term was
introduced by Vinay and Darbelnet (1958/1995) but other authors use the same
term with a different meaning and apply various other names instead. We shall
use the name applied by Klaudy (2003), who calls it replacement. Replacement
is a translation procedure whereby a source language item belonging to one
word-class is replaced by a target language item belonging to a different word-
class.
According to Klaudy (2003: 427), the most frequent word-class replacement
in translation from Indo-European languages into Hungarian is replacement of a
noun with a verb (called verbalisation), while in translation from Hungarian into
Indo-European languages it is replacement of a verb with a noun (called
nominalisation). For instance, when in English a person is described as a laugh a
minute, in Hungarian we may say: állandóan viccel. Such replacements are very
frequent in any language pair, partly because two linguistic systems never
completely overlap and partly because the conventions of usage, including
idioms, are normally different.
In the Finnish text above, the word kunnissa is the plural inessive form of the
word kunta, meaning ‘municipality’. Literally, kunnissa means ‘in (the)
municipalities’. We can observe that almost all the translations applied the literal
equivalent of the original, except for sample (6), which rendered it using the
adjective local, which, being shorter and simpler, makes the processing of the
expression less effortful for the target reader, thereby increasing its relevance.
One could argue, of course, that such a simplification also reduces the precision
of the translation by omitting a certain amount of information. This is true, but
the question is whether the omitted piece of information would make the
translation more relevant or not. If the answer is yes, then the omission is a
mistake. If the answer is no, then it is not a mistake. When the omitted
information can be inferred from the context, we call it implicitation.
Implicitation is a translation procedure whereby a piece of information that is
explicit in the source text is made implicit in the target text, relying on the
context for conveying the meaning (adapted from Vinay and Darbalnet 1995:
344).
Implicitation can be rightfully applied when the information that is made
implicit would not make the translation more relevant. Not everything that is
explicit in the source text contributes to its relevance. For instance, when an
English original speaks about an unmanned space probe, in Hungarian it is
absolutely adequate to simply say űrszonda, because it necessarily implies that
the device is unmanned. Thus, as a matter of fact, if the English expression were
to be translated into something like legénység nélküli űrszonda, it would lead to
a decrease of relevance, because the longer solution would require more
processing effort from the target reader.

125
 To return to the Finnish example, in this particular case, the sentence puts
forward three main points. The first is that the conclusion is based on two
sources of evidence: one being municipal and the other national. The second is
that the municipal source is based on statistics and interviews. The last one is
that the national source is based on statistics. Based on this, Chesterman offers
this solution:

(9) Municipal statistics and interview studies, and also national

statistics, show that…

The elements omitted are the following:

(a) sekä (both)

(b) tehtyihin (carried out)
(c) perustuvat (based on)
(d) selvitykset (reports)
(e) puhtaast (purely)
(f) perustuva (based on)
(g) palvelukatsaus (service review)

Is there any information lost that would make the translation more relevant? The
determiner sekä is omitted and is replaced by the adverb also in the translation,
which fulfils the same function perfectly. Furthermore, target readers can infer
that the municipal interview studies were carried out by someone and that the
statistics and interview studies were presented in the form of a report, and they
can also infer that the national statistics were based on some kind of survey of
the health services and that national statistics are not municipal, so the word
purely is completely unnecessary. Thus the implicitation of all of these elements
does not make the translation less relevant than the original. In fact, the brevity
of the target text enables easier and faster processing, which means less
processing effort and thus greater relevance.
The opposite of implicitation, explicitation, can also enhance the relevance of
a translation. As was explained in Section 6.3.1, explicitation is a translation
procedure whereby a piece of information that is implicit in the source text is
made explicit in the target text.
For instance, instead of the original terveydenhuollon, meaning ‘health care’,
in sample translation (2) we find Finnish health care. Taking into consideration
that the intended target audience of an English translation of the original text is
likely to be a foreign reader, the addition of the word Finnish, by making the
reference to Finland explicit, also contributes to the relevance of the target text.
The question may arise whether such changes in the translation, affecting
both the content and the style of the original, are legitimate or count as a betrayal

126
of the original. The answer largely depends on the kind of text that is being
translated and the aim of the translation. Of course, in the case of a literary work
of art, such changes are normally ill-advised, because in serious literature
faithfulness, or equivalence, to the original is of primary importance. On the
other hand, in the case of an informative piece of writing as the example
discussed here such changes are to be regarded absolutely legitimate so far as
they produce a clearer and thus more relevant translation.

7.3. The final question

The final question is this: Can we teach someone how to translate? The
optimistic answer is yes, of course. A realistic answer is: Well, some aspects of
translation obviously can be taught, while other aspects cannot. The important
thing to remember is that translation is a very complex process involving a wide
range of activities and requisite skills. These skills, just like in any other human
activity, are partly grounded in a native talent, or intuitive capacity, partly in
experience, and partly in systematic knowledge. Thus the three keywords are
instinct, experience and knowledge. One thing that education cannot influence is
whether a person has an inborn aptitude for a particular activity. However,
experience and knowledge are areas that certainly can be developed by
education.
There are some who deny the usefulness of education in creative activities
like translation. For instance, writing about artistic work in general, in an article
Gyula Illyés (cited by Tarnóczi 1966: 250) says that such work is a matter of
inspiration and it is impossible and useless to raise consciousness about it by
learning. Others, however, acknowledge the importance of learning and
practicing. Tarnóczi (1966: 246) defines a translator in these words: The
translator is a person who unifies the necessary abilities and qualifications and
keeps them permanently alive and improving by practicing translation.

7.3.1. Instinct

Communication is not a science – it is also a matter of intuition and personal

style, because it is a creative activity. Thus different translators will produce
different translations of the same source text, even in the same circumstances, all
of which may be good in their own right. In fact, the surprising thing would be if
two different translators would ever produce identical translations of the same
source text. But all this is just natural if we consider that no two persons ever
have identical cognitive environments. And since the translator’s cognitive
environment has a crucial bearing on the translation process and thus on the
shape of the target text, the conclusion must be that in translation differences are

127
the rule and identity is the exception. All this points again to one of the principal
facts in translation: there are no “perfect and eternal” solutions in translation.
Intuition probably plays a lesser role in specialised translation than in literary
translation, since in specialised translation the focus is on the information
conveyed, and it is conveyed by rather conventionalised linguistic means,
including grammatical and stylistic devices as well as standardized terminology.
On the other hand, as Robinson (1997: 117) notes, in literary or advertising
translation intuition is very important in order to come up with an interesting or
striking or effective solution. In such translations creativity and efficiency are
more important than accuracy, and intuition has a critical role to keep the
imagination flowing.
Robinson (1997: 118) gives the following very illustrative example. When
the Armenian-American poet Diana derHovanessian was working together with
an Armenian scholar to translate a collection of contemporary Armenian poetry
into English, there was an English word for mountain-climbing that she felt was
the right solution in the translation, because it was poetically appropriate. Her
Armenian collaborator insisted that it had different connotations than the
original Armenian word used by the poet. However, the translator’s intuition
was later confirmed by the poet himself, who said that he wished he had used the
Armenian equivalent of the English word because it was better than the one that
actually occurred in the poem.

7.3.2. Experience

Experience has a vital part in the growth of the various kinds of competence
needed in translation. Education can contribute to students’ experience of
transfer procedures, languages, subject matters, cultures, or effective
communication techniques by providing them with lots of opportunity for
practice. Providing students with realistic practical tasks and giving feedback on
them is one major role of education in translation.
But the scope of class-room experience is rather limited. A good translator
needs to accumulate experience of as many aspects of the world, and as much of
it, as possible. Baker (1992: 3) quotes Lanna Castellano, a respected translator,
who said of translation: “Our profession is based on knowledge and experience.
It has the longest apprenticeship of any profession. Not until thirty do you start
to be useful as a translator, not until fifty do you start to be in your prime.” Or,
as Árpád Göncz remarks in an essay (1991: 100), translation is not a “young
genre”. What he means by this is that there are things that you will know only
when you have lived long enough to experience various aspects of life.
As Robinson (1997) points out, the translator’s experience must include not
only the languages he or she works with, along with books and dictionaries on
them, but also experience of the people who use these languages, of how they

128
use them, of what people do, experience of the social networks and practices that
people engage in and, eventually, of the cultures that nurture people and their
languages. As translators gather more and more experience of various aspects of
the world, they begin to see patterns and order in the world, which in time
become internalised and become part of their competence as a translator.
Ideally, deductive translational rules or principles should arise inductively as
a result of the translator’s own practical experience of overcoming a series of
individual problems (Robinson 1997: 120). For this inductive process of
building patterns through practice to be effective it is necessary, on the one hand,
that the translator is exposed to as much practical translation experience as is
possible and, on the other hand, that it involves a constant readiness to notice
and reflect on regularities and patterns as well as unusual phenomena. In other
words, what is needed is an awareness of problems and an ability to weigh
alternative solutions against each other.
The importance of this reflexive ability is nicely illustrated by what Cioran
(1999: 46) writes in an essay: “I have known narrow-minded and even stupid
writers, but all the translators I have met, without exception, are intelligent
people, often more intelligent than those that they translate. (There is more
reflection involved in translation than in writing.)”

7.3.3. Knowledge

Acquiring general rules and principles of translation through one’s own first-
hand experience is enormously time and energy consuming, and necessarily
restricted in scope. No one in their life time can gain experience of everything
first hand. Therefore, we also have to rely on knowledge that other people before
us have gathered. This is what education is all about: the transmission of
practical and theoretical knowledge to ever newer generations of translators with
the aim of helping to solve practical problems of translation and promoting good
quality in the practice of professional translation.
Undoubtedly, in the teaching of translation practical tasks are of utmost
importance. But, as we have seen, practice is not complete without reflection and
reflection is systematised by theory. It will be useful to remember that, as Baker
(1992: 2) points out, the value of theory lies in the fact that studying and
reflecting on what we do and how we do it in translation (a) prepares us for
dealing with the unpredictable, (b) gives us confidence, which comes from
knowing that our decisions are based on concrete knowledge rather than
intuition, and (c) makes possible further development on the basis of a
formalised pool of knowledge shared by a professional community.
Naturally, theoretical knowledge of translation will not develop into practical
skills without practice. Thus a realistic rationale behind training in translation,
says Robinson (1997: 122), is the following. Given a background of general

129
principles and regularities of translation and plenty of chances to test those
principles in practice through actual translation tasks, learners will progress
much more rapidly toward professional translation competence than they would
without the help of others’ experience and knowledge. A similar conclusion
follows from the minimalist translation competence concept of Pym (2003),
formulated in slightly different words: Training in translation should ideally be
(a) relatively non-analytical, which means that it must not focus on the
application of rules; (b) it should always be done in context; and (c) it should be
example-oriented. Practice must have priority, but it needs to be complemented
and supported by theory.

7.4. Study questions

After reading the chapter, answer the following questions. If you cannot
remember, go back and revise the related part.

What is the minimalist concept of translation competence?

What is the complex view of translation competence? Which are the five sub-
competences included and what is their role in translation?
How can theoretical concepts be useful for the practicing translator?
Can translation be taught? What aspects of translation can be taught and what
aspects cannot be taught?
What is the importance of theoretical knowledge in the development of
translational competence?

130
 REFERENCES

Alexieva, B. 1993. A cognitive approach to translation equivalence. In: Zlateva, P. (ed.),

101–109.
Anderson, M. 1992. Intelligence and Development. A Cognitive Theory. Oxford:
Blackwell.
Baker, M. 1992. In Other Words. London: Routledge.
Baker, M. 1993. Corpus Linguistics and Translation Studies. In: Baker, M., Francis, G.
and Tognini-Bonelli, E. (eds.) Text and Technology. Philadelphia/Amsterdam:
John Benjamins Publishing Company, 233–250.
Baker, M. (ed.) 1998. Routledge Encyclopedia of Translation Studies. London:
Routledge.
Barkhudarov, L. 1993. The problem of the unit of translation. In: Zlateva, P. (ed.), 39–
46.
Bart, I. 1981. A mérce. In: Bart, I. and Rákos, S. (eds.) A műfordítás ma. Budapest:
Gondolat.
Bell, R. T. 1991. Translation and Translating. London: Longman.
Brower, R. A. (ed.) 1959. On Translation. Harvard: Harvard University Press.
Brown, G. and Yule, G. 1983. Discourse Analysis. Cambridge: Cambridge University
Press.
Burchfield, R. W. 1996. New Fowler’s Modern English Usage. Oxford: Oxford
University Press.
Carnap, R. 1950. Logical Foundations of Probability. London: Routledge and Kegan
Paul.
Catford, J. C. 1965. A Linguistic Theory of Translation. Oxford: Oxford University Press.
Chesterman, A. 1996. Psst! Theory can be useful! Kääntäjä 10: 4–5.
Chomsky, N. 1957. Syntactic Structures. The Hague: Mouton.
Cicero, M. T. 1949. De invitatione, De optimo genere oratorum, Topica. Transl. H. M.
Hubell. Cambridge, MA: Harvard University Press.
Cioran, E. M. 1999. Füzetek. Transl. Pál Réz. Budapest: Európa Kiadó.
Csikszentmihalyi, M. 1995. The Flow Experience and Its Significance for Human
Psychology. In: Csikszentmihalyi, M. and Selega Csikszentmihalyi, I. (eds.)
Optimal Experience: Psychological Studies of Flow in Consciousness. New York:
Cambridge University Press, 15–35.
Dubois, J. et al. 1973. Dictionnaire de linguistique. Paris: Larousse.
Duff, A. 1981. The Third Language. Oxford: Pergamon Press.
Even-Zohar, I. 1978. The Position of Translated Literature Within the Literary
Polysystem. In: Holmes, J. S., Lambert, J. and van den Broeck, R. (eds.), 117–
127.
Even-Zohar, I. 1990. Polysystem Studies. Tel Aviv: The Porter Institute for Poetics and
Semiotics.
Fodor, J. A. 1983. The Modularity of Mind. Cambridge, Mass.: MIT Press.
Gentzler, E. 1993. Contemporary Translation Theories. London: Routledge.
Göncz, Á. 1991. Gyaluforgács. Budapest: Pesti Szalon Kiadó.

131
Grice, H. P. 1975. Logic and Conversation. In: Cole, P. and Morgan, J. L. (eds.) Speech
Acts. Syntax and Semantics 3. New York: Academic Press, 41–58.
Gutt, E-A. 1991. Translation and Relevance. Oxford: Basil Blackwell.
Hartmann, R. R. K. and Stork, F. C. 1972. Dictionary of Language and Linguistics.
Amsterdam: Applied Science.
Hatim, B. and Mason, I. 1990. Discourse and the Translator. London: Routledge.
Hell, Gy. 2004. A fordításról – Cicero nyomán. Modern Filológiai Közlemények VI.2:
28–36.
Heltai, P. 2003. Message Adjustment in Translation. Across Languages and Cultures 4
(2): 145–185.
Hempel, C. G. 1967. Fundamentals of Concept Formation in Empirical Science.
Chicago: University of Chicago Press.
Hervey, S. and Higgins, I. 1992. Thinking Translation. London: Routledge.
Holmes, J. S. 1988. Translated! Amsterdam: Rodopi.
Holmes, J. S. 1988a. The Name and Nature of Translation Studies. In: Holmes, J. S.
1988, 67–80.
Holmes, J. S. 1988b. The Future of Translation Theory: A Handful of Theses. In:
Holmes, J. S. 1988, 99–102.
Holmes, J. S., Lambert, J. and van den Broeck, R. (eds.) 1978. Literature and
Translation: New Perspectives in Literary Studies. Leuven: Acco.
Holz-Mänttäri, J. 1984. Translatorisches Handeln: Theorie und Methode. (Annales
Academiae Scientiarum Fennicae B 226). Helsinki: Suomalainen
Tiedeakatemia/Finnish Academy of Science.
House, J. 1981. A Model for Translation Quality Assessment. Tübingen: Gunter Narr
Verlag.
Hymes, D. 1972. On Communicatice Competence. In: Pride, J. B. and Holmes, J. (eds.)
Sociolinguistics: Selected Readings. Harmondsworth: Penguin Books, 269–293.
Jakobson, R. 1959. On Linguistic Aspects of Translation. In: Brower, R. A. (ed.), 232–
239.
Jerome. 2004. Letter to Pammachius. In: Venuti, L. (ed.), 21–30.
Klaudy, K. 1994. A fordítás elmélete és gyakorlata. Budapest: Scholastika.
Klaudy, K. 2003. Languages in Translation. Budapest: Scholastika.
Komissarov, V. N. 1980. Lingvistica perevoda. Moscow: Mezhdunarodnye otnosheniya.
Lőrincz, J. 2007. Kultúrák párbeszéde. Eger: Líceum Kiadó.
Mounin, G. 1963. Les problémes théoretiques de la traduction. Paris: Gallimard.
Newmark, P. 1988. A Textbook of Translation. New York, London etc.: Prentice Hall.
Nida, E. A. 1959. Bible translating. In: Brower, R. A. (ed.), 11–31.
Nida, E. A. 1964. Toward a Science of Translating. Leiden: E. J. Brill.
Owens, R. 1996. The Translator’s Handbook. London: Aslib.
Peirce, Ch. S. 1931–1966. Collected Papers of Charles Sanders Peirce. Edited by
Charles Hartshorne, Paul Weiss and Arthur W. Burks. 8 vols. Cambridge, MA:
Belknap Press of Harvard University Press.
Pym, A. 1992. Translation and Text Transfer: An Essay on the Principles of
Intercultural Communication. Frankfurt: Peter Lang.
Pym, A. 2003. Redefining Translation Competence in an Electronic Age. In Defence of
a Minimalist Approach. Meta XLVIII: 481–497.

132
Rabin, C. 1958. The Linguistics of Translation. In: Smith, A. H. (ed.) Aspects of
Translation. (Studies in Communication 2). London: Secker and Warburg, 123–
145.
Radó, Gy. 1998. Hungarian tradition. In: Baker, M. (ed.), 448–455.
Reiss, K. and Vermeer, H. J. 1984. Grundlegung einer allgemeinen Übersetzungstheorie.
Tübingen: Niemeyer.
Robinson, D. 1997. Becoming a Translator. London and New York: Routledge.
Robinson, D. 1998. Babel, tower of. In: Baker, M. (ed.), 21–22.
Saussure, F. 1995. Course in General Linguistics. Transl. Roy Harris. London:
Duckworth.
Seleskovich, D. 1977. Take Care of the Sense and the Sounds Will Take Care of
Themselves or Why Interpreting is not Tantamount to Translating Languages.
The Incorporated Linguist 16: 27–33.
Shannon, C. E. and Weaver, W. 1949. The Mathemical Theory of Communication.
Urbana: University of Illinois Press.
Shveitser, A. 1993. Equivalence and adequacy. In: Zlateva, P. (ed.), 47–56.
Smith, N. and Tsimpli I-M. 1995. The Mind of a Savant. Oxford: Basil Blackwell.
Sperber, D. and Wilson, D. 1986. Relevance. Oxford: Basil Blackwell.
Steiner, G. 1975/1992. After Babel. Oxford: Oxford University Press.
Tarnóczi, L. 1966. Fordítókalauz. Budapest: Közgazdasági és Jogi Könyvkiadó.
Tomalin, B. and Stempelski, S. 1993. Cultural Awareness. Oxford: Oxford University
Press.
Toury, G. 1978. The Nature and Role of Norms in Literary Translation. In: Holmes, J.
S., J. Lambert and van den Broeck, R. (eds.), 83–100.
Toury, G. 1980. In Search of a Theory of Translation. Tel Aviv: Porter Institute.
Tytler, A. 1791. Essay on the Principles of Translation. London: Dent.
Valló, Zs. 2000. A fordítás pragmatikai dimenziói és a kulturális reáliák.
Fordítástudomány II.1: 34–49.
Venuti, L. (ed.) 2004. The Translation Studies Reader. New York and London:
Routledge.
Venuti, L. 1998. Strategies of Translation. In: Baker, M. (ed.), 240–244.
Vinay, J. P. and Darbelnet, J. 1995. Comparative Stylistics of French and English. A
Methodology for Translation. Transl. Juan C. Sager and M. J. Hamel.
Amsterdam: John Benjamins.
Vlahov, S. and S. Florin. 1980. Neperevodimoye v perevode. Moscow: Mezhdunarodniye
Otnosheniya.
Wagner, E., Bech, S. and Martínez, J. M. 2002. Translating for the European Union
Institutions. Manchester, UK and Northampton, MA: St. Jerome Publishing.
Whorf, B. L. 1997. Language, Thought, and Reality. Selected Writings of Benjamin Lee
Whorf. Cambridge, Mass.: The MIT Press.
Wilson, D. 1992. Reference and Relevance. UCL Working Papers in Linguistics Vol. 4.
London: Department of Phonetics and Linguistics, University College London,
167–191.
Wilson, D. and Sperber, D. 1988. Representation and relevance. In: Kempson, R. (ed.)
Mental Representations: The Interface Between Language and Reality.
Cambridge: Cambridge University Press, 133–153.

133
Wilss, W. 1977. Übersetzungwissenschaft. Probleme und Methoden. Stuttgart: Ernst Klett.
Wittgenstein, L. 1958. Philosophical Investigations. Transl. G. E. M. Anscombe.
Oxford: Basil Blackwell.
Woodsworth, J. 1998. History of Translation. In: Baker, M. (ed.), 100–105.
Zlateva, P. (ed.) 1993. Translation as Social Action. London and New York: Routledge.

134