Lecture 7

Simultaneous interpreting (синхронний переклад)

Broadly speaking, simultaneous interpreting (SI) is the mode of interpreting in which the
interpreter renders the speech as it is being delivered by a speaker into another language
SIMULTANEOUS INTERPRETING 382 with a minimal TIME LAG of a few seconds.
When interpreting in simultaneous mode between spoken languages, interpreters generally
work in soundproof booths with SI equipment that prevents acoustic overlap between the
original speech, listened to via headphones, and its simultaneous interpretation spoken into a
microphone. SI can also be practiced without a booth, as „whispered interpreting‟ (chuchotage),
for only a few listeners, or using a mobile system (bidule). Interpreters also work simultaneously
between a spoken and a signed language (referred to as „bimodal interpreting‟) or between two
signed languages. Since these are expressed in a visual-gestural modality, there is no acoustic
overlap, but visual contact between the interpreter and the deaf listener(s) or speaker is crucial.
In terms of processing, the simultaneous mode also includes SIGHT
INTERPRETING/TRANSLATION, in which written texts are rendered into spoken or signed
languages in real time. This is also done in the combined mode known as SIMULTANEOUS
WITH TEXT, where the interpreter processes the speaker‟s speech at the same time as
following the written text of that speech. Another hybrid form is SIMULTANEOUS
CONSECUTIVE, where the interpreter, working consecutively, records the original speech
with a digital device and then interprets the recording simultaneously rather than relying (only)
on notes. The widespread use of SI in international organizations in the post-war era,
particularly following its initial public success at the NUREMBERG TRIAL, has led SI to be most
closely associated with CONFERENCE INTERPRETING. Although this connection is
undeniably present, interpreters had probably been rendering speeches simultaneously long
before the institutionalization of simultaneous conference interpreting, both in the form of
whispered interpreting and as the typical working mode in SIGNED LANGUAGE
INTERPRETING. On the other hand, the close linkage established in the twentieth century
between SI and TECHNOLOGY continues, paving the way for the new opportunities and
challenges that have arisen with REMOTE INTERPRETING, and leadingto new international
standards for technical infrastructure. From early on, researchers were interested in the
element of simultaneity in SI, whereby a person listens to a speech as it is being delivered,
processes it and (re)produces it in another language. As early as 1930, Sanz drew attention to
the specific processes in SI, and since the 1970s psycholinguists have taken an interest in SI as a
source of insights into aspects of bilingual human language processing. Despite the progress
that has been achieved, the complexity of the task and the difficulty of isolating one single
sensitive variable from the many which interact within the overall processcontinue to challenge
PSYCHOLINGUISTIC APPROACHES and, more generally, COGNITIVE APPROACHES
to research on SI. The main areas of investigation to date include MODELS of the cognitive
process; the function of key components such as WORKING MEMORY; the neural
correlates of bilingual processing in SI; the dynamics of
INPUT VARIABLES and output features in professional performance; and the STRATEGIES
used by interpreters to cope with cognitive and linguistic challenges. Modeling the interpreting
process has been a major concern of research on SI. Some of the models developed have
attempted to sequence and explain the interconnections between micro-level cognitive
processes (e.g. Gerver 1976; Moser 1978), while others have focused on the interplay of the
main operational tasks (Kirchhoff 1976a, 1976/2002; Lederer 1981; Darò & Fabbro 1994; Gile
1995a). The „pragmatic‟ dimension, which was initially neglected, was also incorporated into a
number of the later models, acknowledging the importance of input-related factors as well as
situational and psychological dimensions (Cokely 1992a; Setton 1999). Working memory and the
allocation of attentional resources have figured centrally in process models and have also been
studied separately. Although good working memory has been considered an essential
requirement for SI, research to date has not offered much conclusive evidence of the posited
connection (see Timarová et al. 2014). There is some evidence that it is not superior working
memory capacity per se which leads to expert performance, but rather an acquired ability to
manage competing demands on limited cognitive resources in SI (Liu et al. 2004); other studies
emphasize that experts distinguish themselves from novices by the more meaningful relations
they have formed between items in MEMORY (Köpke & Signorelli 2012). In addition to the
storage function, researchers are increasingly interested in exploring the executive functions of
working memory, such as „controlled attention‟ (Timarová et al. 2014) or the management of
COGNITIVE LOAD. The cognitive mechanisms of INTERFERENCE and suppression have
also received attention due to their relevance in SI, where various resources need to be shared.
Based on studies exploring both unimodal and bimodal SI, researchers have attempted to
examine the impact of phonological interference and the interplay between SI and suppression
in lexical access, metaphor comprehension and syntactic parsing (e.g. Gernsbacher & Shlesinger
1997; Isham 2000).

Simultaneous with text (синхронний переклад з текстом оригіналу)

„Simultaneous with text‟ (SI with text) refers to SIMULTANEOUS INTERPRETING (SI) of
a speech read from a source text that is available to the interpreter, who uses it as part of the
input to follow the speaker. This dual input (auditory and visual) makes SI with text a cognitive
task distinct from SI done from speech alone, and from SIGHT
INTERPRETING/TRANSLATION (done from written text), whether or not the interpreter
has seen the text before interpreting. For this reason, „SI with text‟ is a more accurate label for
this task than other terms found in the literature such as „sight interpreting‟, „oral sight
translation‟ or „simultaneous sight translation‟. Although SI with text can in theory be
performed as a mixture of SI and sight translation, going from „pure‟ SI without any reference
to the text to „pure‟ sight translation without any reference to the sound (Gile 1997), it is best
classified as a form of SI, since the oral input has priority: the interpreter must „check against
delivery‟ and, in the event of deviation, follow the speaker‟s actual words rather than the text
which has been provided. Interpreting a speaker reading from a prepared text that has not been
supplied to the interpreter is widely recognized as stressful and unsatisfactory (Mackintosh
2002); indeed, contracts initially exempted conference interpreters from this task, though this is
now hardly applicable in practice. SI with text is considered more acceptable, but nevertheless
highly complex. Written text tends to be more informationally and linguistically dense than
impromptu or semi-prepared speech, and may be presented at a high speed and/or without the
natural PROSODY that facilitates oral communication. The interpreter‟s auditory input may
thus be too fast to manage without the written text; yet relying only on the latter poses the risk
of falling behind and missing the speaker‟s deviations. Like sight translation, SI with text is also
assumed to involve a higher risk of linguistic INTERFERENCE – one of the few aspects of
this mode that have been the object of empirical research (Lamberger-Felber & Schneider
2008). In an experimental study with twelve experienced professionals interpreting three
authentic English conference speeches either with or without the written text available in the
booth, Lamberger-Felber (2001, 2003) sought to compare SI and SI with text (either with or
without time to prepare) with regard to target-text ACCURACY and completeness, in
particular OMISSIONS. Though complicated by high individual variability of performance and
perceptions of task difficulty, the results showed that the percentage of correctly rendered
proper names and NUMBERS was distinctly higher for SI with text (98% with time to prepare,
92% without) than when the interpreters did not have the written text. In a similar vein,
Coverlizza (2004) compared interpreters‟ SI performance on certain items in read-out
presentations, either without text or with the text provided ten minutes in advance, and found
that participants performed better with the text on strings of adjectives, numbers, long lists and
an anecdote, but were more fluent when working without a text. Lambert (2004) found
significantly improved SI-with-text performance in student interpreters when the text was
available with ten minutes to prepare. In addition to the difficulty of the written text, the time
available for PREPARATION stands out as a key factor determining the effectiveness of SI
with text. Acceptable QUALITY can be achieved for all but the most „written‟ material, if
provided to the interpreter with sufficient time for advance preparation. This may include
reading and annotating it in various ways to direct the eye, for example by underlining or
highlighting numbers or unfamiliar names, or to guide restructuring (especially for typologically
contrasting language pairs), noting idiomatic target-language expressions to combat the risk of
interference, or making notes in the margin to capture the gist of a dense paragraph and
facilitate COMPRESSION (see Seleskovitch & Lederer 2002; Setton 2006b). STRATEGIES for
using the written text while interpreting from auditory input may thus vary with the time
available for preparation, but also depend on individual preferences linked to the interpreter‟s
abilities in fast reading, text scanning and eye–ear coordination (or the availability of help from
a boothmate). If the text arrives too late to prepare, the interpreter may prefer to set it aside
and work purely by ear. However, such preferences and strategic decision-making and their
effects have yet to be investigated. Based on available research findings as well as professional
experience, a strong case can be made for the inclusion of SI with text in interpreter training
programs, with the focus on preparation techniques and training eye–ear coordination (Setton
& Dawrant 2016). Moreover, the skills acquired for SI with text can also be applied to „mixed-
media‟ presentations (e.g. PowerPoint) and, most recently, to speeches delivered with real-time
captioning on a screen. Even more so than SI with text, however, such dual-input modes are
still relatively unresearched.

Speech-to-text interpreting ()
Speech-to-text interpreting, which is also known as captioning or as real-time speech-to-text
services, is increasingly being used in some countries, such as the United States, to provide
communication access for individuals who are deaf or hard of hearing (D/HH), mostly in
educational settings. In speech-to-text interpreting, the provider of the service, who is often in
the classroom or meeting next to the D/HH individual(s), produces text as it is being spoken
by a speaker, such as a teacher, and displays it on a portable device so that the individual can
understand what is happening. The process is typically intra-lingual – that is, the text interpreter
listens to the spoken words and produces the text display in the same language. Speech-to-text
interpreting services are occasionally provided to individuals with other needs, such as a hearing
student with a learning disability. In the two common speech-to-text interpreting service
options, the text interpreter uses either (a) a standard typing (QWERTY) keyboard or (b) a
stenographic machine. A little used option is (c) automatic speech recognition (Steinfeld 1998;
Stinson et al. 2008). These services may be provided in the classroom or meeting location, or
remotely. If remotely, the speaker typically wears a bluetooth microphone and the spoken
message is delivered using either a cellular phone or Voice over Internet Protocol (VoIP) via a
cellular or broadband Internet connection (i.e. Skype, Google Hangout). These services also
support display of text on a variety of devices (standard laptops, smartphones, etc.). The text
interpreter may produce copies, typically electronic, of the saved text (Stinson et al. 2014).
Typing-based services (C-Print, Typewell) use a standard QWERTY keyboard (Stinson et al.
2008). A trained service provider uses the word-abbreviation feature in the system‟s software to
speed up typing. The software transforms abbreviations into full words on the computer
screen. In addition, text interpreters learn strategies for identifying important points and for
condensing information. These systems cannot provide word-for-word transcription, because
they cannot keep up with the speed of speech; however, the systems do capture almost all of
the meaning of what is being said (Elliot et al. 2001). Research has compared comprehension
and retention of information after viewing a lecture with C-Print speech-to-text support and
after viewing a comparable lecture with interpreter support. Students retained more or an equal
amount of information with C-Print as with interpreter support (Marschark et al. 2006; Stinson
et al. 2009). With a steno-based system, a trained stenographer uses a 24-key machine to encode
spoken words phonetically into a computer that converts them into the written characters of
the language concerned (often English) and displays the resulting text in real time. The
stenographer depresses several keys simultaneously, instead of sequentially as in conventional
typing. Equipment includes special software that translates entries into complete words, as well
as a laptop. Generally the text is produced verbatim. Research has found that D/HH students
demonstrated significantly better comprehension of a spoken message with this technology
than when only following the video and audio of the message (Steinfeld 1998). A study that
compared student lecture comprehension with steno-, C-Print, and interpreting services did not
find significant differences in comprehension for the three methods (Marschark et al. 2006).

Whisper interpreting is similar to simultaneous interpreting but the interpreter does not use a
headset or microphone, rather the interpreter sits next to the person (or group of people) who
require interpreting and whispers or speaks softly while interpreting in the target language. This
form of interpreting is much harder on the interpreter‟s voice.
This is often used for a business meeting where just one person requires interpreting, or for
example, in a courtroom where someone in the back of the room requires interpreting to
understand what is being said.

Consecutive interpreting (послідовний усний переклад)

Interpreters, whose presence was attested in ancient EGYPT, have for centuries served rulers,
conquerors, military and religious leaders, as well as traders and explorers. They would typically
have worked between two parties in consecutive mode, speaking after someone had spoken in
 another language. However, no special label was used for this age-old practice before the
twentieth century, when it came to be referred to as liaison interpreting or, more
commonly now, DIALOGUE INTERPRETING. The term „consecutive interpreting‟,
used as a broad label for the „default‟ mode of interpreting, came into use after the
1920s, to mark the difference between traditional interpreting and what was then the newly
tested „telephonic‟ (i.e. simultaneous) mode (Baigorri-Jalón 2014). It was in the League of
Nations, the first international organisation to employ its own staff interpreters, that
„classic‟ consecutive interpreting flourished. Gustave Camerlynck, André Kaminker and
Paul MANTOUX are famous examples of interpreters who worked professionally but
without any special training, as did all interpreters at the time. Comments on their work
are found in MEMOIRS byinterpreters (e.g. Schmidt 1949; Jacob 1962) and politicians
(e.g. Lansing 1921; Madariaga 1974), and interpreters themselves also provided an
account of their work in an early study bySanz (1930).

Following the successful use of SIMULTANEOUS INTERPRETING (SI) at the

NUREMBERG TRIAL and its subsequent adoption by the United Nations (Baigorri-
Jalón 2004), consecutive interpreting was gradually displaced by the new mode of
CONFERENCE INTERPRETING. Indeed, consecutive interpreting is the exception
rather than the rule in thework of most conference interpreters today. Since consecutive
interpreting, broadly defined, may involve the rendering of source-language utterances
lasting anywhere from a few seconds to several minutes or more (e.g. González et al.
1991/2012), there is no hard and fast distinction between „classic consecutive‟ and
„short consecutive‟. Prototypically, the former implies (monologic) speeches with a
duration of five or six – or sometimes even as long as twenty – minutes; the latter usually
deals with utterances as short as a single phrase or sentence. In practice, these two extremes
can be said to be increasingly converging, though there is a lack of systematic empirical data
in this regard (as is, indeed, the case for many other aspects of consecutive interpreting).

FIDELITY IN INTERPRETATION AND TRANSLATION

Plan

Introduction: Professional Translation: An Act of Communication

1. Verbalizing a Simple Idea
2. Principles of Fidelity
1
3. The Rules of Reformulation
4. Secondary Information: an Obstacle and a Help
5. Interpretation vs. Translation from the Secondary Information Perspective

Professional Translation: An Act of Communication

The professional translation is essentially a serviced activity with a communication

function, performed in a professional setting with a professional aim in mind and
constrained by this setting.
Translation activity (interpretation and translation) is found in several contexts.

The most widespread and best known type school translation, insofar as virtually everyone
experiences it in school. Its aims are to improve and/or test students' passive and active
knowledge of a foreign language: translating into the foreign language shows and improves
writing ability in that language, or at least the ability to write foreign language texts following
lexical and syntactic choices induced by the source- language text; translating from the foreign
language improves and demonstratescomprehension of words and linguistic structures in that
language.

Another case is the translation of a text in a foreign language into one's own languagefor
easier personal use at a later stage: for instance, one may wish to translate into one's own
language parts of a user's manual if it is written in a language one does notread easily and if
one expects to have to consult the manual rather often.
Yet another case is the translation of a piece of literature, in particular poetry, for the
translator's pleasure. The pleasure can be associated with the fact that translation involves a
careful study of the text, or with the creative translation process itself.

As for interpretation, the most widespread non-professional variety is occasional help given
to friends, relatives, or tourists during visits, sightseeing, shopping, etc.

Professional Translation differs from all these in two basic respects:

• Professional Translation is aimed at a Receiver (reader or listener) other than the
Translator him-or herself; in professional Translation, the Receiver is essentially
interested in the Text, in whatever "message" it carries, and/or in the Sender (author
or speaker), not in the Translator or in the Translation process.
• Professional Translation is done on request and for a financial reward.

2
Professional Translation is therefore professional act of communication, and as such,
it is subjected to professional rules, as well as to particular rules relating to
communication.

Professional translation acts on a text written by an author, generally for source language
readers as distinct from the translation-readers. One may add that in some cases it can be
written directly for translation-readers—a user's manual may bewritten especially for clients
in an overseas market—and in other cases both for source-language readers and for
translation-readers—as is the case of a user's manual for products aimed at both the domestic
and overseas markets. Another point is that professional translation is paid for by a Client,
defined here as the person, company, department, or other organization which orders the
translation and pays the translator,and that this Client need not be the author or the reader of
the translation. In the most general case, the configuration of communication actors in which
professional translation occurs can be the following one:
Author —> Source-language reader
Translator —> Target-language reader
Client

It defines two acts of communication, one going from the author to the source- language
reader, and the other from the translator to the target-language reader.

The classical model of Translation is:

Author —> Translator —> Reader

Interpretation differs from translation in that the Sender (the speaker) generally speaks
either to the target-language listeners only or to both source-language listenersand target-
language readers.

Speaker —> Source-language listener

Interpreter —> Target-language listener Client

It should be noted that in interpreting, unlike translation, all parties concerned are aware of
the communication situation, including possible difficulties associated with the interlingual
and sometimes intercultural transfer. Since generally all parties wish to communicate, more
cooperation can be expected from them than in translation, where they are aware of a text
rather than of a communication situation. This includescooperation from speakers, who may
try to speak more slowly, enunciate moreclearly, choose certain terms and structures and
avoid others, and clarify terms and concepts . Cooperation may also be forthcoming from
listeners, especially in

3
consecutive, where they can help the interpreter with word equivalents and generally listen
sympathetically, though this is not always the case. In other words, althoughthe interpreter
essentially works alone, he or she may be helped through on-line interaction with both Sender
and Receiver, while in translation such interaction is rather rare.

On the other hand, if the proportion of target-language listeners in the audience isvery
small, the interpreter may suffer from interference instead of benefiting from cooperation,
especially in consecutive. The reason is that source-language listeners often perceive
interpretation as a necessary evil; when the delegates who actually need interpretation are
few in number and unimportant for the others, interpretation can be perceived as an
unnecessary evil that entails loss of time as well as technical constraints (speed of delivery,
seating arrangements, the mandatory use of a microphone, etc.). Delegates may therefore put
pressure on the interpreter to be very brief or summarize, and will not cooperate in other ways.

The issue of fidelity

The issue of fidelity is probably the most basic and widely discussed component of
Translation quality. Setting aside the question of translation competence theoreticianshave
wondered whether translation can be "fully" faithful while retaining the editorial(or literary)
merit of the source-language text. The most obvious problem with fidelity stems from the
well-known fact that languages are not isomorphic: in other words, there is no one-to-one
correspondence between them as regards lexical elements ("words") or linguistic structures
associated with rules of grammar, stylistic rules, etc. In particular, there is no automatic
equivalence between words in the source and target languages and apparently similar
structures may have different uses and different connotations. It is necessary to know that
sometimes different changes of the construction such as “adding” and “deleting” do not
amount to a breach of fidelity.

Daniel Jile in his book “Basic Concepts and Models for Interpreter and Translator Training”
describes an experiment which he carried out more than 30 times over more than 13 years in
various countries and with languages as different as Arabic, Chinese, English, Finnish,
French, German, Hebrew, Italian, Japanese, Spanish, Swedish, and Tahitian, with consistent
results as outlined below, demonstrating wide applicability.

1. Verbalizing a Simple Idea

A simple drawing suggesting an elementary informational Message is presented tothe

participants, who are told what situation is simulated and asked to formulate the

4
Message in their own words in their native tongue. The sentences thus obtained are then read
aloud.

At this point, it may be appropriate to recall and stress that in this book, which deals
essentially with the Translation of Informational Texts, the word Message is defined
not as the statement produced, i.e. the verbal materialization of a communicative
intention, but as the information that the Sender wants to get across the Receiver and
around which the verbal statement will be constructed.

In this example the drawing shown to participants depicted a road as seen from insidea car,
with a road sign showing "Paris 50 km". The participants were told the following: “You are
sitting in the car next to the driver. At a certain point in time yousee the road sign. Please
write down exactly what you would say in your mother tongue to the driver to tell him what
the sign says."

The first consistent empirical result is the following: Given exactly the same Message
presented under exactly identical conditions at the same point in time, individuals
sharing the same mother tongue tend to write different sentences to express it.

This result was partly due to differences in the way the students understood the Message
presented to them, and the analysis of the subject shown. In all cases, the statements did seem
to reflect essentially the same Message, as is the case of the six sentences listed below, a
subset of those collected from native English speakers in one replication of the experiment
done in the United States in October 1988, at the 29th convention of the American Translators'
Association.
(1) Fifty kilometers to Paris.
(2) Still fifty kilometers to go.
(3) We'll be in Paris in fifty kilometers.
(4) Fifty kilometers longer.
(5) We'll be there in fifty kilometers.
(6) Paris is fifty kilometers from here.

5
The finding that the same Message as defined here is formulated in different ways has
a corollary: If different sentences can correspond to the same Message, and if this is
applicable in both the source language and the target language, then in translation,
different sentences in the target language may reflect the same Message as the one
initially generated in the source language.

This reasoning is based on the essential point that the statement is built with the purpose of
transmitting it. However, this legitimacy may rightly be challenged, as fidelity is generally
taken to apply not only to the Message but also to its packaging. That is, fidelity applies not
only to the Message, but also to its author through the ways he expresses it. It turned out
that sentences expressing the same Message couldbe strikingly different. Thus we can say
that the sentences differ from each other andfrom the drawing in their information content.

These differences may be divided into:

• Information gains: Information given in one sentence which is not found in another
or in the drawing, such as the arrival in Paris being a future-event, or the existence of
at least one more person besides the speaker who is concerned by the statement;
• Information losses: Information not given in the sentence under consideration
although it is present in a sentence it is being compared to or in the drawing, such as
the explicit mention of Paris.
It is clearly seen that the Message is conveyed fully in all sentences. But it should be kept in
mind that the situation is based on the fact that the Sender and the Receiver share the same
knowledge.

One fundamental reason for this reaction seems to be that given a text to translate, the
Translator, who is generally not familiar with the precise circumstances under whichit was
generated, may not know which is the Message and which is the Secondary Information. As
a result, there is a tendency to translate all the information, so as not to miss any relevant
component. This, however, does not account for the fact that even participants who were
present through out the first phase and who therefore knew precisely what the Message was
and under what circumstances it was put into words, translated each sentence differently. Still
more interesting, this applies not only to student-translators and to amateurs, but also to many
experienced interpreters and translators. One possible explanation is the experimental setup,
which is intrinsically unnatural and may induce behavior that is not identical to that which
would occur in a clear-cut translation situation in the field. However, another possiblereason
may be the existence of an implicit operational rule, which seems to have beeninternalized
even in professional translators, that translation should preferably reformulate not only the
original Message, but all the information contained in the source-language text. This may be
due to the fact that translators do not really focus on the distinction between Primary and
Secondary Information unless translation poses difficult problems and forces them to make
choices. From observation, it seems

6
that conference interpreters take a wider margin of freedom, possibly because of difficulties
in speech production and cognitive processing of language and data, which result in more
forced choices of this kind.

2. Principles of Fidelity

In determining principles of fidelity for interpretation and translation, it seems appropriate to

start not with the finished linguistic product, but with the setting of information.

In this respect it is important to remember that in informative communication such as is found

in conference interpretation and technical and scientific translation, the Sender formulates the
discourse as the carrier of a Message for the purpose of achieving an aim such as informing,
explaining, and/or persuading. For the Sender the communication is successful if this aim is
achieved. Generally, the Translator “represents" the Sender and the Sender's interests, and
therefore “does a good job” if the Translation contributes to the success of the Sender's
endeavor.
On the other hand, there seems to be a consensus that Translators cannot rewrite or
reformulate the speech in a completely different way which they believe will achieve the
Sender's objective more efficiently than the Sender’s words. That is, the Translator must
contribute toward successful communication while following what is essentially the same
"route" as the one the Sender chose in the source language to lead the Receiver along.

The Message

There is therefore a minimum fidelity principle as regards reformulation of information.

The absolute fidelity rule is that the Message or Primary Information should always be
re-expressed in the target-language Text. The situation is not so clear-cut with respect
to Secondary Information.

The Primary Information consists of the following types: Framing Information,

Linguistically Induced Information and Private Information.

Framing Information

Framing Information is selected by the Sender for the purpose of facilitating the reception of
the Message by the Receiver. But the Receivers of the original speech or text may not have
the same preexisting knowledge and values as Receivers of its Translated version. Framing
Information which is appropriate for the original Receivers may not be suitable for target-
language Receivers, in which case reformulating the FI in the target language may defeat the
aims of communication by making the information too explicit or not explicit enough for the
target-language Receiver.

7
For instance, when translating an American's statement about "Cairo, Egypt" (as opposed to
"Cairo, Illinois") into "Le Caire, en Egypte," the French translatorworking for French readers
makes a strange or at best a humorous statement, probably to the detriment of the real
Message: most French readers only think of "Cairo, Egypt" and are totally unaware of the
existence of "Cairo, Illinois," so that theexplicit mention of Egypt may distract their attention
from| the actual Message. Conversely, when translating an Australian text quoting prices in
"dollars," it may be useful to add Framing Information by specifying " Australian dollars " so
that these are not mistaken for U.S. dollars.

It follows that fidelity to the Sender's interests may require eliminating some of the original
FI and adding some FI for the benefit of the target language Receiver.

It must be stressed that the selection of Framing Information is not strictly | determined by
objective circumstances. Depending on the Sender's style, the FI may vary. The selection of
FI therefore reflects the personality of the Sender to a varying but non-negligible extent. From
discussions with Translators, teachers of translation, and users of translation services, it would
seem that there is also a general consensus that Translation should also reflect the Sender's
personality. This view, which is onlyseldom made explicit in writings about non-literary
translation, is| consistent with the general principle of fidelity; after all, the Sender's Message
and interests also basically reflect personal choices, and therefore "personality." However,
the consensus regarding the translation of non-literary texts also seems to incorporate a low-
priority rating of fidelity to the author "personality," as opposed to informational fidelity.

If Translation is also to reflect the Sender’s “personality” beyond the Message content, the
Sender's personal choice of FI should in principle be followed by the Translator as well.
However, in the translation of informational texts and speeches, the role of which is to convey
information for the purpose of achieving an objective, the impact of the information with
respect to the objective definitely seems to take precedence over fidelity to the Sender’s “
personality”. Therefore, if the Sender's original FI does not seem appropriate for the Receivers
of the target-language product, there is some justification for the Translator's changing it until
it does.

Linguistically Induced Information(LII)

The case of Linguistically Induced Information is different, in that LII is not even selected by
the Senders of their own free will. They are often free to choose one of several options, but
must choose one. In the source language, the LII, which generallycontains some redundant
elements and some non-relevant ones, is by definition natural and well integrated into the
discourse. However, in the target language, the reformulation of such LII can be awkward or
even distort the Message.

8
Incidentally, the LII requirements of the target language, plus the fact that translatorscannot
always discriminate between the Message and LII and tend to translate the latter in order to
be sure not to leave any part of the Message untranslated, often combine and generate target-
language texts which contain more information than thesource-language text. In other words,
the target-language text includes not onlyLinguistically Induced information from the source-
language, but also its own LII. In view of the fact that translators also tend to retain the
Framing Information of the original text, translations tend to be longer than source-language
texts.

It is the source language, and not the Sender, that basically determines the LII in the source
text. This would imply that there is no reason to try to reproduce the LII in the target language,
which is different by definition. On the other hand, Senders may be free to choose from two
or more options, and their linguistic style does partly determine the LII which will ultimately
be carried in the source-language text. Since such LII does reflect an individual personality
to some extent, there is some reason to try to reproduce it in the target-language text. However,
when aiming at optimal communication efficiency, reproducing LII should have a much lower
priority than reproducing Framing Information.

Personal Information

Personal Information differs from the other two types of Secondary Information in that it is
neither imposed by linguistic rules nor selected for framing purposes, but is by definition a
pure reflection of the Sender's personality as manifested linguistically. It should therefore be
followed if possible, but not if the cost in terms of communication efficiency (readability,
clarity, strength of the target-language product, etc.) is even moderately high. In particular,
Personal Information indicating, through grammatical and other errors or regional
expressions, that Senders are usinga language other than their own, or that they come from a
particular area in their country, should generally not be reconstructed in the target language,
since they are not relevant and may distract the Receiver's attention from the Message. As for
the case when Personal Information generates a negative image of the Senders, forinstance
by showing that they are not well educated, Sender loyalty would imply that such information
is not to be reconstructed in the target language either.

3. The Rules of Reformulation

The order of ideas identified as part of the Message in the source text must be followed in the
target text: if the Sender starts a source-language text by presenting idea A, then illustrates it
with examples B, C, D, Translators should follow the same order.

9
In translation, within a sentence, structural changes are generally accepted by the parties
concerned, and are, therefore, legitimate. In the interest of efficiency ofcommunication, long
sentences may be segmented into shorter ones, and sentences that follow each other in the
source-language text may be merged in the target language.

In interpretation, more extensive stylistic and informational changes may be acceptable. The
reason is that in a written text, authors are supposed to have had the time to review and correct
their prose until it reflects their thought as they want to. In oral discourse this is not the case,
and more elements may be escaping the speaker’s control. The interpreter should be
concentrated on the Message and make the linguistic choices. For instance, sentences that
speakers do not complete because of speech-production difficulties should be completed in
the target language, and sentences that the speaker repeats because he has lost the train of
thought do not haveto be repeated.

If the translator feels that the particular choice of words or linguistic structures may have been
made deliberately for impact, this choice should be followed whenever possible. This is
frequently the case with word repetitions, humorous distortion of words or grammar, etc.

4. Secondary Information: an Obstacle and a Help

Secondary Information is one of the most frequent sources of fidelity problems and decision-
making requirements in Translation of informational Texts. The questions that arise regularly
are whether to reformulate in the target-language Text informationthat might be detrimental
to communication, and whether to introduce new Secondary Information to help
communication become more effective.

From experience, in informative translation many of these questions are answered

spontaneously by the Translator without any conscious decision making. For example, when
Translating from English into Japanese, the English singular/plural LII is most often
suppressed spontaneously when the information is irrelevant, because Translators are aware
of the fact that trying to reformulate it may make the Japanese text clumsy, or even distort the
Message. Similarly, in a speech made in English by a female speaker, interpreters working
into Hebrew will not hesitate to introduce the LII indicating the sex of the speaker not made
explicit by the English speech.

10
On the other hand, serious problems arise when additional information is required because
the target-language rules are not known to the Translator and are not given in the source-
language Text. For example, in a conference, a speaker may refer to somebody as "Monsieur
X," giving the interpreter working into English Linguistically Induced Information relating
to the gender of X but failing to indicate whether he should be referred to as "Dr. X," "Prof.
X," etc., as may be appropriate in English in the relevant context. Failure on the part of the
interpreter to refer to X by the proper title may affect communication. Similarly, since the
singular/plural discrimination in Western languages is generally difficult to escape,
translation from Japanese into a Western language poses problems when the Translator must
decide whether the objects or other entities the author refers to are in the singular or plural,
but does not have the information. The Translator has to make a decision and take a chance
on the possibility of an erroneous decision.

Secondary Information is often much more valuable to the Translator than to the Receiver.
For the latter, part of it is already known. The most difficult problems with respect to fidelity
and the resolution of ambiguity arise when target-language rules require information not
provided by the source-language Text. Experience shows thatthe frequency of such problems
depends largely on the specific language pair involved.

For instance, in the translation of informational texts and speeches between English and
French, such problems are few: the occasional forms of address as illustrated above, the use
of the passive form in English, which cannot always be replicated in French and which poses
problems to the Translator who does not know the agent of the verb, etc. On the other hand,
LII-generated problems are numerous in translation between Japanese and Western languages
in particular because of the following two differences.

- Western languages generally discriminate between singular and plural and between various
points in the past, present, and future, whereas Japanese does not necessarily do so. This does
not cause difficulties when translating into Japanese, because such Linguistically Induced
Information simply disappears in the target-language product; but when translating from
Japanese, problems resulting from the lack of background information are sometimes difficult
to solve.

-Western languages tend to indicate explicitly the subject and object of verbs, which is not
the case in Japanese. When translating from Japanese into a Western language, problems
sometimes arise because the target language requires information about the subject and/or
object of the verb and none is available.

5. Interpretation vs. Translation from the Secondary Information Perspective

In translation, numerous difficult decisions regarding fidelity have to be made, which lead to
iterative corrections of the target-language text. The question therefore arises as to whether
it is possible for interpretation, with its practically instantaneous and

11
virtually correction-free production process, to be reasonably effective in producing faithful and
linguistically acceptable target-language speech. Two basic facilitating factors can be identified in
the conference interpretation environment as opposed to the translation environment.
In international conferences, speakers and listeners are assembled in the same place at the
same time, and speakers know they are talking to target- languagelisteners as well as to
delegates who understand their own language. Generally, they also know more about their target-
language listeners than authors do about their target-language readers. Therefore, Framing
Information is more likely to be suitable for target-language listeners in interpretation than in
translation. Moreover, the diagrams and slides shown during the speech, as well as the body
language of the speaker, provide cues beyond those included in the linguistic part of the
interpreters' speech, and help them achieve more effective communication.
In international conferences, the Receivers (the delegates) process the speaker's words by ear.
The well-known evanescence of the spoken word is associated not onlywith semantic rather than
verbal memory of speech, but most probably also with less "word-bound" processing of speech:
listeners seem to devote their attention to a form of processing, they are more concerned with
general propositions than with linguisticstructures because the speed of delivery (about 100 to 200
words per minute) which may limit the amount of processing that can be done on the speech they
hear. Whenever they do concentrate on nuances, it is precisely because such nuances are an
important part of the Message, e.g. in diplomatic and political speeches. Therefore, itis probable
that listeners tend to focus on Primary Information and the effect of changes introduced by the
interpreter in secondary information becomes less of a problem than in written translation.

When reading rapidly, readers probably also tend to concentrate on Primary information, but
the reading process is not as linear as the listening process: readers may focus on a particular word
for a longer time or reread a text segment after going through it a first time, and their perception
of the Message may be more word-bound than in speech processing because of this non-linearity,
likely resulting in a greater amount of Secondary Information. Moreover, the selection of
Secondary Information is not carefully thought out and corrected as can be the case in written
texts. Speakersare less in control of the Secondary Information content of their speeches which
loosens the constraints on fidelity in the reproduction of Secondary Information by interpreters.
This may explain why delegates who speak both the source and the target languages and even the
speakers themselves, often come up to interpreters afterconsecutive interpretation and ask them
how they manage to do such a perfect “word- for-word” translation of speeches, when the
interpreter has in fact exploited a certain degree of freedom and made many changes in Secondary
Information.

12