This article was downloaded by: [University North Carolina - Chapel Hill]

On: 26 June 2013, At: 07:11

Publisher: Routledge
Informa Ltd Registered in England and Wales Registered Number: 1072954 Registered office: Mortimer House,
37-41 Mortimer Street, London W1T 3JH, UK

Perspectives: Studies in Translatology

Publication details, including instructions for authors and subscription information:
http://www.tandfonline.com/loi/rmps20

Subtitling: Diagonal translation

a
Henrik Gottlieb
a
University of Copenhagen, Denmark
Published online: 28 Apr 2010.

To cite this article: Henrik Gottlieb (1994): Subtitling: Diagonal translation, Perspectives: Studies in Translatology, 2:1,
101-121

To link to this article: http://dx.doi.org/10.1080/0907676X.1994.9961227

PLEASE SCROLL DOWN FOR ARTICLE

Full terms and conditions of use: http://www.tandfonline.com/page/terms-and-conditions

This article may be used for research, teaching, and private study purposes. Any substantial or systematic
reproduction, redistribution, reselling, loan, sub-licensing, systematic supply, or distribution in any form to
anyone is expressly forbidden.

The publisher does not give any warranty express or implied or make any representation that the contents
will be complete or accurate or up to date. The accuracy of any instructions, formulae, and drug doses should
be independently verified with primary sources. The publisher shall not be liable for any loss, actions, claims,
proceedings, demand, or costs or damages whatsoever or howsoever caused arising directly or indirectly in
connection with or arising out of the use of this material.
 101
SUBTITLING: DIAGONAL TRANSLATION

Henrik Gottlieb, University of Copenhagen, Denmark

Abstract
Subtitling of televisedforeign-language material not only changes language; it also switches
from the spoken to the written mode, and it presents itself 'in real time', as a dynamic text type.
Hence, due to the complex, 'diagonal' nature of subtitling, the subtitler must possess the musical
ears of an interpreter, the stylistic sensitivity of a literary translator, the visual acuteness of a film
cutter, and the esthetic sense of a book designer.
The author discusses subtitling in the context of language transfer, and suggests nine basic
Downloaded by [University North Carolina - Chapel Hill] at 07:11 26 June 2013

fields to consider when creating - and evaluating - interlingual subtitles for television and video.

The nature of subtitling

Subtitling is an amphibion: it flows with the current of speech, defining the
pace of reception; it jumps at regular intervals, allowing a new text chunk to be
read; and flying over the audiovisual landscape, it does not mingle with the hu-
man voices of that landscape: instead it provides the audience with a bird's-eye
view of'the scenery.
Dubbing, which flows but never jumps, is synchronous par excellence, while
the staccato, chunky nature of subtitles places these somewhere between the
printed page and the spoken dialog, in terms of mode of reception. In the bottom
cell of the table below, the horizontal axis indicates viewing time, the vertical
gives the amount of verbal information translated at any given time:1

Table 1:
Text typo BOOK SUBTITLED TV DUBBEDTV
(TRANSLATION)

Mode Of Printed page, 1-2 lines Phoneme-for-phoneme

presentation normally open text or teletext revolting
unillustrated in original in edited
audiovisual context audiovisual context
Reception time
definer Reader Subtitler Dubbing team

Synchrony
between (No original
receptor and present)
original

Real Urne Realtime

 102 1994 Perspectives: Studies in Translatology 2 (1)

As can be seen, subtitling keeps 'jumping ahead' of the dialog, giving fast
readers the dubious privilege of anticipating lines before they are uttered. Slow
readers, on the other hand, keep lagging behind, whereas dubbed dialog is per-
ceived simultaneously by all viewers.
In Denmark, where dubbing is only found in children's programs,2 subtitling
has become one of the dominant written text types in public life: by 1992 the
weekly time spent reading TV and video subtitles amounted to nearly four hours
(228 minutes) per capita. Printed translations only account for a little less than
two hours a week.3 Whether original Danish printed texts still hold the lead, as
in pre-electronic times, is doubtful. Even if people still spend as much (leisure)
Downloaded by [University North Carolina - Chapel Hill] at 07:11 26 June 2013

time reading printed sources as they did in 1987 when the latest major survey
was conducted,4 the weekly consumption of non-electronic writing will not reach
the 5-hour mark.
Subtitling is an overt type of translation,5 retaining the original version, thus
laying itself bare to criticism from everybody with the slightest knowledge of the
source language. At the same time, subtitling is fragmentary in that it only repre-
sents the lexical and the syntactic features of the dialog. The prosodie features
are not truly represented in subtitles: exclamation marks, italics, etc., are only
faint echoes of the certain ring that intonation gives the wording of the dialog.
Added to this, subtitling has to do without well-known literary and dramatic de-
vices such as stage direction, author's remarks, footnotes, etc. The audience has
to turn to the original acoustic and visual clues in trying to grasp the meaning be-
hind the words of the subtitles.
As opposed to subtitling, dubbing - the ever-present rival - offers a discrete,
covert mode of translation, replacing the entire dialog track, and sometimes even
the accompanying music & effects-track, with a target-language version. This
integral translation gives people an all-in-one representation of the dialog, not
forcing its audience to add a third cognitive effort (reading) to the two basic ef-
forts: watching and listening. And by being covert, the well-performed dubbing
gives its audience an impression of being presented with the original, or at least
a good copy, sealed off from reproach, as the original soundtrack is deafened.*
If dubbing is such a perfect illusion, not burdening the translator with any
critical remarks from second-guessers, why advocate subtitling? A quick and lo-
gical answer would be: "For its honesty." However, although its authenticity can-
not be denied, neither can its intrusion on the image and, consequently, its dis-
 Henrik Gottlieb: Subtitling: Diagonal Translation 103

traction of the audience. Even people with a good knowledge of the source lan-
guage find it difficult not to read the subtitles as they keep flashing on and off
the screen. Belgian studies have suggested that "reading behavior is more or less
automatically elicited when a subject faces a text with the same message auditori-
ly available." (d'Ydewalle, van Rensbergen and Pollet 1985: 20)
To emphasize this point, in a study focusing on vertical subtitles,7 even sub-
jects with no prior experience with subtitled TV programs "did in fact use the
subtitles considerably. Therefore, we feel confident in concluding that reading
subtitles is not due to habit formation. When there is a choice between the speech
and text channels, the subjects read the subtitles." (d'Ydewalle, Praet, Verfaule
Downloaded by [University North Carolina - Chapel Hill] at 07:11 26 June 2013

and Van Rensbergen 1991: 660)

So, in order to focus viewer attention on the image proper, whenever the
soundtrack contains discourse of an informative nature, revoicing should be con-
sidered. This is especially relevant in programs or sequences with off-screen nar-
ration. In documentaries, the TV picture is often so crammed with information,
that for this reason alone subtitling is virtually impossible. But when we talk
about human interest stories, TV fiction and feature films, that is expressive
genres focusing on people, as opposed to objects or abstract phenomena, only
subtitling will provide the authenticity needed.

Subtitling speech acts: words in the balance

Every act of verbal transmission forces the transmitter to make priorities. Dif-
ferent media and different types of discourse naturally impose different con-
straints on the transmitter, or - to put it more optimistically - leave the transmitter
with different sets of clues for solving the particular issues at stake. Thus, sub-
titling does not differ from literary translation, for instance, in that it constrains
the translator, but simply because the constraints involved in an audiovisual
context are different from those of the patient, but impotent paper.8
In rendering what human voices are trying to express, be that in literature,
films or TV fiction, there are no absolutes, no canonized solutions. But this fact
should not be taken as an easy excuse for claiming that any phrasing is as good
as the next. In subtitling, as in all types of translation, no word should be acci-
dental, and even good ideas should be tested against alternative solutions. As I
can see from my students' work, there is always more than one solution to a sub-
titling problem, but even more 'solutions' that miss the target. And in order to
 104 1994 Perspectives: Studies in Translatology 2 (1)

hit the target, all relevant linguistic, esthetical and technical means should be util-
ized, and both dialog, film and viewers must be considered.
In subtitling, the speech act is in focus; verbal intentions and visual effects
are more important than atomized lexical elements. This gives the subtitler a cer-
tain amount of linguistic freedom.' But the adequate rendering must seem self-
evident to the viewers: the audience is not served with memorials of the plight
of the subtitler. Only the results count, not the hours spent translating a sequence
that some theorists might classify as 'untranslatable'.

The transition from spoken dialog to written subtitles

Translation and interpreting, the two traditional counterparts in interlingual
Downloaded by [University North Carolina - Chapel Hill] at 07:11 26 June 2013

communication, are horizontal, one-dimensional types of verbal transmission.

Both carry verbal messages across from one human language to another, and both
stick to the semiotic nature of the source message: speech remains speech, and
writing remains writing:

Table 2a:
SOURCE LANGUAGE TARGET LANGUAGE

j.
SPEECH SDSak 1 INTERPRETING /^> Speak

WRITING Text 1 TRANSLATION J ~^>» Text

Subtitling, on the other hand, can be either vertical or diagonal.

Vertical subtitling takes speech down in writing, whereas diagonal subtitling,
being two-dimensional, 'jaywalks' (crosses over) from source-language speech
to target-language writing:

Table 2b:
SOURCE LANGUAGE TARGET LANGUAGE
SPEECH Speak Speak

WRITING Text Text

 Henrik Gottlieb: Subtitling: Diagonal Translation 105

Naturally, in this paper, diagonal subtitling is in focus. Being an 'oblique' trans-

lation, this type of verbal transmission was, in its infancy, considered an ugly
duckling or even a non-translation.
One of the pioneers of translation studies postulated:
"Translation between media is impossible (i.e. one cannot 'translate' from the spokenio the
written form of a text or vice-versa)." (Catford 1965: 53)
I quite agree that in the everyday sense of the word, you cannot translate
from one medium to another. A novel, for instance, cannot be 'translated' into
a movie. But by expanding the concept of translation, as Roman Jakobson, a
contemporary of Catford, did (see Jakobson 1966) the term intersemiotic
translation can be applied to the transfer between semiotically different entities,
Downloaded by [University North Carolina - Chapel Hill] at 07:11 26 June 2013

as for instance screen adaptation. However, subtitling - vertical or diagonal - is

fnfrasemiotic; it operates within the confines of the film and TV media, and stays
within the code of verbal language. The subtitler does not even alter the original;
he or she adds an element, but does not delete anything from the audiovisual
whole.
The problem, however, is that the graphemic subtitles should correspond with
the phonemic dialog that the subtitles should double. And the incompatibility of
the oral and the written sub-codes can indeed act as a hindrance to the intended
correspondence. In a handbook for British ('vertical') subtitlers,10 the dream of
harmonious brotherhood between speech and writing is ruptured:
"The attempt to achieve perfect subtitling has some affinity to the search for the Holy Grail.
The differing design features of written and spoken languages dictate that a perfect corre-
spondence between the two cannot obtain." (Baker, Lamboume and Rowston 1984: 6)
If we settle for something slightly less than perfect, we first have to locate
the differences between the two verbal sub-codes involved as well as their differ-
ing contexts.
The features distinguishing speech from normal written texts are:
1) The interlocutors are in direct contact with each other; via their dialog they
share a situation. This produces an implicit language where things are taken
for granted. Written sources usually need to explicate and extend the mes-
sage, as the reader is unknown, or at least not present.
2) Spoken language has different esthetic norms, namely, another placing of cer-
tain stylistic features on the axes correct-incorrect, and formal-informal.
In spontaneous speech (which may be 'artificial', as in feature films) the subtitler
will often find:
3) Pauses, false starts, self-corrections and interruptions.
4) Unfinished sentences and 'grammatically unacceptable' constructions.
 106 1994 Perspectives: Studies in Translatology 2 (1)

5) Slips-of-the-tongue, self-contradictions, ambiguities and nonsense.

6) Situations where people all talk at the same time, a feature very difficult to
render in writing.
Finally, it is characteristic of certain real or ficticious persons that:
7) Their language contains dialectal or sociolectal features that the established
orthography is unable to cope with.
8) Their language contains idiolectal features, that is idiosynchrasies only used
by the speaker.
9) Their pronunciation of certain words may be so indistinct that these words
defy identification.
Downloaded by [University North Carolina - Chapel Hill] at 07:11 26 June 2013

Thus, in any diagonal - and thus interlingual - subtitling, one must, on top
of translating utterances from one language to another, transfer the dialog from
one sub-code (the seemingly unruly spoken language) to another (the more rigid
written language). If this shift of sub-code was not performed, as a fundamental
part of the subtitling process, the audience would be taken aback by reading the
oddities of spoken discourse. But as the dialog is always re-coded on the way to
the bottom of the screen, people only react if the other dimension of diagonal
subtitling - the translation proper - seems imperfect.
However, evaluating subtitles as interlingual translation is not easy either.
Because of the complex, polysemiotic nature of film and TV, a comparison be-
tween subtitles and (transcribed) dialog will not suffice for making adequacy
judgments. In the case of book translations, a simple verbal text comparison will
work, if factors such as difference in time, place and readership are considered.
But when dealing with subtitling, the synthesis of three synchronous semiotic
channels (image, dialog and subtitles) should be compared with the original two-
channel discourse. Severed from the audiovisual context, neither subtitles nor dia-
log will render the full meaning of the film. So in judging the quality of subtitles,
one must examine the degree to which the subtitled version as a whole manages
to convey the semantic gestalt of the original.

Subtitling people: 9 pedagogical pillars

Subtitlers should start by asking themselves:

1) What am I going to subtitle?

In films and TV programs, the auditive channel transmits two semiotically diffe-
rent soundtracks: the dialog track and the 'music & effects track'. Deciding what
 Henrik Gottlieb: Subtitling: Diagonal Translation 107

elements to subtitle is not limited to identifying speech sounds as either 'dialog'

or 'music & effects'. Not every spoken utterance needs to be put down in writ-
ing: depending on the visual context, repetitive exclamations and certain formula-
ic phrases - such as greetings - may be left untranslated. On the other hand, ver-
bal music & effects elements may have a semantic function beyond that of a
backdrop to action and dialog. If, for example, a fragment of a song heard in the
background is relevant to the plot of the film, those lines should be subtitled, pro-
vided that there is no dialog in that particular sequence.
Even the visual channel often contributes with verbal elements that need sub-
titling. In his excellent handbook of an art, Swedish subtitler Jan Ivarsson distin-
Downloaded by [University North Carolina - Chapel Hill] at 07:11 26 June 2013

guishes between 'displays' and 'captions'. While the term 'displays' refers to
(fragments of) texts recorded by the camera - letters, newspaper headlines, ban-
ners, etc. - 'captions' are defined as "texts that are added to the film or tape after
shooting", that is superimposed texts such as location signs or toptitles stating the
name and occupation of interviewees, etc (Ivarsson 1992: 104)."
A guideline in deciding whether to subtitle a given verbal element is the
point-of-view of the film. In fiction especially, it is important to 'know whose
side you are on'. An example: in the Danish video version of Kevin Costner's
Dances with Wolves, an American pro-Indian Western from 1990, the protago-
nist's first encounter with the Native Americans is subtitled in this way:
Kostner (carrying a woman he found outside the Sioux village):
"She's hurt." Hun er sâret.
As the crowd of Indians do not react:
"She's hurt." Hun er sâret.
Then a young Sioux approaches, waving his hands:
"Na osh niuw! Na osh niuw!" Du er ikke velkommen her.
Kostner calmly continues:
"No, she's hurt." Nej, hun er sàret
"Diah! Diah yayo!" Forsvind!
"Na osh niuw, fustra! Diah yayo!" Du er ikke velkommen her. Forsvind.
[The Lakota I have attempted to transcribe phonetically, and according to the Danish titles
the first Lakota remark means "You're not welcome here!", the second simply "Scram!"]
In this sequence, the Danish subtitler apparently decided that he wanted to
place his audience in the position of the 'omniscient director': the Danish video
audience 'understands' what the chief says, while the film hero does not know
a word of Lakota. An equally acceptable strategy would have been to let the au-
dience side with the protagonist, sharing his uneasiness about the situation, by not
subtitling the remarks in Lakota. The deciding factor may have been the Ameri-
can movie version, subtitling the Lakota for its domestic audience.
 108 1994 Perspectives: Studies in Translatology 2 (1)

2) Ant I hearing what is actually said?

Film and TV manuscripts are always perfect when you do not need to consult
them. The things you can hear immediately will be there. But when it comes to
difficult passages, you often find 'inaudible' or '???' written in the script, or even
worse, a misinterpretation of what is being said. Besides, lots of programs come
without scripts anyway. So, subtitlers must rely on their own ears or do some in-
telligent guessing, though sometimes a native speaker or an expert has to be sum-
moned. As using such 'external ears' is a costly and time-consuming matter, a
subtitler worth his salt needs to have the aural competence of an interpreter. Ex-
cellent translation skills and fluency in the target language alone do not make for
Downloaded by [University North Carolina - Chapel Hill] at 07:11 26 June 2013

a good subtitler.
And even good subtitlers are sometimes cheated by their ears: in the TV spe-
cial The Making of "Dances with Wolves", broadcast by Danmarks Radio May
23 1991, an old Sioux lady acting in the film tells the interviewen
"We're kind of a quiet people. But we never, never ever were portrayed that way
till this movie." This statement fits well with the image of the film: as we already
know, Dances with Wolves shows native Americans as human beings, not savage
redskins. But this Danish subtitle alienates its audience:
Vi er et stilfxrdigt folk, men vi har
ikke vaeret os forraderiet bevidst.
(We are a quiet people, but we have not been aware of this betrayal.)
Explanation: The Danish subtitler expects all Americans to pronounce their
Vs. This pronunciation is a high-prestige feature of US English, but in informal
conversation, Americans of all social groups often do not pronounce their 'r's in
unstressed syllables, as in this case. The subtitler may have heard the unvoiced,
aspired quality of the initial bilabial, thus defined as a 'p'. But apparently the
'r'gument was felt to be stronger, leading to a subtitle that indeed "betrayed" the
dialog.

3) Do I know the exact meaning of the words in this context?

Different languages do not consist of word-lists covering one another in a 1:1
relationship. And not even within a specific language does a word always mean
the same: some words cover a row of different entities in the world around us,
depending on their extralinguistic context. But a number of words vary in mean-
ing, depending on their, intralinguistic context. Idioms are typical examples, each
idiomatic expression synthesizing a meaning that cannot be inferred directly from
the meanings of its elements. Idioms are often found in slang, but all text types
 Henrik Gottlieb: Subtitling: Diagonal Translation 109

offer a range of idiomatic expressions. This is well-known by all translators, and

so is the fact that familiar words sometimes pop up in 'strange' contexts, such
as jargon or dialects. In such cases, a conscientious effort is usually put into solv-
ing the problem. But sometimes trivial words are the stuff that problems are
made of.
In the following example, lack of congruence between the semantic fields of
the source and target languages is what breaks the camel's back: in a German po-
litics-and-love story from 1991, Marx & Coca-Cola (DR-TV April 18, 1993),12
a West German businessman is delayed 24 hours on his way home through ex-
East Germany. He is running out of gas, as his Saab 9000 gets stuck in a muddy
Downloaded by [University North Carolina - Chapel Hill] at 07:11 26 June 2013

Mecklenburg field. As our protagonist finally manages to get through to Ham-

burg on his car phone, he tells his mother:
"Hoffentlich hast du dir nicht zu viele Gedanken gemacht. Mir geht's gut. Ich hatte nur 'ne
Panne."
[Hope you haven't worried too much. I'm all right I just had some trouble with the car.]

But the Danish subtitle goes: Du har vel Udce vaeret nerv0s?
Jeg punkterede bare.
[You haven't worried, I hope.
I just had a puncture.]
In itself, this imprecise translation (he did not have a puncture) is no disaster.
But it leaves us with a false impression of our man: he has fallen in love with
the East German girl who pulled him out of the mud, and now we are.let to be-
lieve he is lying! (His fiancee is sitting next to his mother, waiting for him.) But
it was indeed a 'Panne' he had. The problem is that although the word often re-
fers to a puncture, within the field of traffic it covers 'any technical problem with
your car that forces you to stop' (my definition). In Danish we do not have a si-
milar hyperonym: the nearest you can get is 'vr0vl med bilen' (trouble with the
car). But then the subtitler could have stuck to facts, writing for instance "Jeg 10b
t0r." (I ran out of gas.)
When the subtitler has decided what to translate, and how to interpret the di-
alog to be translated, or, in other words, has performed a correct pragmatic, pho-
nemic, and semantic decoding of the dialog - pillars 1 through 3 - the viewers
may hope to be presented with:

4) A congenial segmentation of dialog

The Dutch television subtitler Helene Reid (Reid 1990: 100 [and following
pages]) distinguishes between grammatical, rhetorical, and visual segmentation
 110 1994 Perspectives: Studies in Translatobgy 2(1)

of the dialog in films and TV. The segments thus established are the units to be
shaped into (target language) subtitles. Whereas a grammatical segmentation is
guided by semantic units in the dialog, a rhetorical segmentation follows the
rhythm of speech, and a visual segmentation follows cuts and camera moves.
It should be noted, however, that quite often cutting follows the rhythm of
speech, and most of the time speakers pause in grammatically acceptable places.
The subtitler is in trouble only when, for some reason, changes in image and dia-
log are not simultaneous. In most cases, then, the subtitler will prefer a rhetorical
segmentation. This is the prevalent method anyway, as the subtitle is the alter ego
of the corresponding dialog segment. As a rule, subtitlers tend to 'cut' where the
Downloaded by [University North Carolina - Chapel Hill] at 07:11 26 June 2013

speaker pauses to breathe. And for physiological reasons, these pauses are typic-
ally found at intervals of 5-6 seconds, exactly the two-line subtitle standard fol-
lowed by most broadcasters. So if we define a dialog segment as 'the words
spoken between two breathing intervals', a TV subtitle will, by and large, cover
such a segment. However, a substantial number of the utterances to be subtitled
are shorter than the intake of human breath. That is why only a minor part of
professional subtitles are shaped as full-length two-liners.
Every well-formed subtitle should read as a self-contained entity, but in utter-
ances covering more than one segment, a consistent thread must be preserved.
The titles should be cohesive if the dialog is. But the subtitler should not spill the
beans by prematurely revealing a point, such as the answer to a dramatic ques-
tion. The target-language audience should get the points as things happen, not be-
fore they happen. Naturally, this problem is caused by the fact that many viewers
are able to read the whole subtitle faster than the actors deliver the original dia-
log, a problem not relevant with dubbing (see table 1).
An example of how to tackle this obstacle successfully is found in the
subtitling of an episode of the popular TV series LA. Law (DR-TV October 5,
1989):
Dialog: Subtitles: Back-tianslaäon ftom Danish:
My client is Min klient er My client is
an immature 75 year-old - et umodent bam pâ 75 âr - an immature child of 75 -
and so is his opponent, og det gaelder and that goes for
Mr. Henderson. ogsâ modparten. the opponent, too.
The attorney's remark is uttered quite slowly - it takes seven seconds - so for
the subtitler it would have been the easiest thing to include the entire utterance
in one subtitle, with enough reading time for everybody:
Min klient er en umoden 75-ârig,
og det er modparten ogsâ.
 Henrik Gottlieb: Subtitling: Diagonal Translation \\\

But in the film, as the first part of the sentence is uttered, there is a cross-cut
from the infuriated client to his maliciously amused opponent, whose jaw drops
as the attorney pronounces the second part of the sentence. Thus the segmenta-
tion into two subtitles - even separated by a 15-frame interval" - with the se-
cond subtitle not on the air until the moment the inflated ego of the opponent is
punctured in the courtroom. This segmentation means that the viewers have 4 se-
conds to read the initial title, with only 2 seconds and 10 frames left for the con-
clusive one - slightly less than the norm of 10 characters per second. But an all-
in-one solution would have been devastating to the dramatic progression of the
sequence.
Downloaded by [University North Carolina - Chapel Hill] at 07:11 26 June 2013

5) A loyal, yet idiomatic translation

As mentioned previously, certain TV program genres call for highly informa-
tive, 'exact' subtitles, whereas other types of program allow for greater linguistic
liberties in the description of people and situations. Thus, in subtitling 'facts',
such as documentaries and news stories, finding the right terms are paramount,
while style is of primary importance in fiction. Subtitling facts is somewhat like
adding captions to a picture, as in a newspaper, whereas subtitling fiction is
recreating lines, as in drama. So, when dealing with factual programs, the litmus
test would be "Is this what we normally call the objects or events we see here?".
In fiction, interpreting people, we would have to say "Is this what this kind of
person would say in this situation?". But often a TV program is a mix of genres,
demanding that the subtitler be loyal to the latest voice on the screen - without
putting off the viewer. In some instances, loyalty toward the image and/or the
viewers forces the subtitler to make compromises.
An example where feedback from the original dialog was considered irrecon-
cilable with an 'exact' translation is found in the subtitling of the first episode
of the youth series Beverly Hills 90210 (TV2 March 23, 1992):
Dialog: Subtitle:
Boy: Do you mind if I sit down? - Vil du heist sidde alene?
Girl: No, go for it! - Nej, saet dig bare.
Back-translation:
- Would you rather sit alone?
- No, do sit down.
Here, the auditive feedback from the word 'no', recognized by even non-Eng-
lish speaking Danes, forces the subtitler to turn the question upside down. In Da-
nish, two more idiomatic renderings would be:
 112 1994 Perspectives: Studies in Translatology 2(1)

A) "Mâjeggodtsiddeher?" + "Ja,selvf0lgelig!"
["May I sit down? + "Yes, of course!"]
B) "Har du noget imod, jeg sauter mig her?" + "Nej, nej."
("You don't mind my sitting down?" + "No, no."]
Version (A) is brief and well-suited for the screen, but a 'no* is turned into
a 'yes'. That does not work. Version (B) produces a 'no', but takes up too much
space, the maximum number of characters being 35 per line. And unlike the LA.
Law example above, there is no time for cutting the dialog into two titles. So we
end up with a slightly awkward solution, which does, however, convey the mean-
ing adequately.
Downloaded by [University North Carolina - Chapel Hill] at 07:11 26 June 2013

6) A minimized loss of information

Fully benefiting from the polysemiotic nature of television, the necessary
quantitative reduction of the dialog can be reached with only minor qualitative
losses in stylistic and/or denotative information (see Gottlieb 1992). Supported
by the original soundtrack and image, providing helpful intonative and visual aids
to understanding the dialog, subtitles will usually allow for an adequate transfer
of the original message. However, even in dialog with little support from the
image, a quite substantial condensation is often possible.
In Truffaut's Day for Night from 1971 (TV2 August 7, 1993) the following
remark is uttered by a person talking to the camera:
"Eh bien, je crois que l'histoire est traitée sur le mode tragique,
et comme dans toute tragédie chaque personnage va jusqu'au bout de son destin."
Time is short - the rather long sentence is pronounced as one dialog segment,
lasting exactly five seconds, and it is both preceded and followed by speak - so
the Danish subtitle goes:
Filmen er bygget op som en tragédie;
derfor m0der nver person sin skxbne.
[The film is composed as a tragedy; so each person meets his destiny.]
By deleting the redundant initial 'well', the typical spoken discourse modera-
tor 'I think', and the repetitive 'and as in all tragedies', the volume of the utter-
ance is halved, yet the meaning is preserved. Had this fast-spoken remark con-
tained fewer redundant elements, or substantial culture-specific material, decima-
tion of the semantic content would have been inevitable.14
 Henrik Gottlieb: Subtitling: Diagonal Translation 113

7) A 'user-friendly' text composition

In subtitling, time is always at a premium. As the viewers are not in charge
of their pace of reading, the subtitler has to facilitate their work. The audience
should be able to read the subtitles and enjoy the film.

Line breaks
An important factor in this respect is the layout of the individual subtitles.
Should one or two lines be preferred? And should the entire top line be used
before shifting to the bottom line? This is one of the few aspects of TV subtitling
ever tested empirically. At Leuven University, Belgium, the viewing time of one-
liners versus two-liners was investigated. But as several factors are involved -
Downloaded by [University North Carolina - Chapel Hill] at 07:11 26 June 2013

such as the rate of turntaking among the speakers, and the complexity of the
television image - results are ambiguous (see Muylaert et al. 1983; Praet et al.
1990). However, it seems to be the case that, other things being equal, (fewer)
larger subtitle units are read faster than (more) smaller ones. But within the unit,
line segmentation is important to viewer reception. Semantically motivated line
breaks enhance reading speed: the fewer eye fixations on the text, the better.
The following subtitle, from the American documentary Dream Deceivers
(TV2 March 16, 1992), will illustrate this:
Dialog:
You could literally see his mind clickin' and bells and lights goin' up.
I think he ... "They're right!"
Subtitle (as broadcast) Constructed subtitle:
Man kunne se, at det gik op for ham, Man kunne se, at det gik
at vi havde ret. op for ham, at vi havde ret.
[You could see that it dawned upon him that we were right.]
Apart from the interesting shift away from direct speech in the subtitle, the
point is: you can read it in just three fixations. The constructed subtitle - of a lay-
out still commonly found - takes four fixations to read. Adding to the time delay
is the fact that the phrasal verb 'gik op for' (dawned upon) is separated into the
presumably literal 'gik' ('walked' or 'went') and 'op for' ('up for'). So for sever-
al reasons unmotivated line breaks should be avoided, a micro-level parallel to
the dialog segmentation issue dealt with earlier.

Text volume and syntax

In principle, the subtitler can opt for as much text as time and space would
allow. But that might mean depriving the audience of the time necessary to 'read'
 114 1994 Perspectives: Studies in Translatology 2 (1)

the facial expressions and the movements on the screen. The time left for non-
verbal viewing should match the time spent reading. But of course, although most
viewers manage to keep this balance with programs subtitled according to the 10
characters-per-second norm mentioned earlier, slow readers never have enough
time for watching the film. On the other hand, subtitled TV may be the only
thing that keeps them from lapsing into virtual illiteracy.
Though in subtitling the volume of text is reduced, the sentence structure is
usually maintained, due to the feedback effect introduced earlier. However, gram-
matical differences may block the optimal rapport between subtitle and speak. In
subtitling from German (into Danish or English, for instance) the main verb in
Downloaded by [University North Carolina - Chapel Hill] at 07:11 26 June 2013

a dependent clause will almost inevitably show up on the screen 'prematurely',

often one title ahead of the German verb spoken. Forcing a foreign syntax onto
the domestic-language subtitles is not recommendable, if at all possible, and it
certainly does not make reading easy either. This may encourage the subtitler to
recompose the whole sentence. But then again, we may end up a with a heavily
syncopated subtitle, and, as in the following case, a time problem.
In this excerpt from a retrospective TV portrait of Malcolm X (DR-TV April
4, 1993), the civil rights activist James Farmer says:
[The Malcolm X that] Subtitles:
(1) most of the young blacks 1. Det mesie af det, de unge sorte
are loving, adoring, elsker, tilbeder og nassten forguder, -
worshipping, almost, 2. - er Malcolm, som nan var,
(2) is the pre-Mecca Malcolm, inden han tog til Mekka.
(3) and they're not even aware that 3. De ved slet ikke, at der var
there was a post-Mecca Malcolm, en Malcolm efter Mekka, -
(4) and that his views were changing. 4. - og at han asndrede holdning.

Back-translation:
1. Most of what the young blacks love, adore and almost worship,
2. is Malcolm as he was before he left for Mecca.
3. They don't even know that there was a Malcolm after Mecca,
4. and that he changed his views.
In subtitles no. 2 and 3, the central elements have changed places, turning a
(distinctly pronounced) A-B-C order into a C-A-B permutation. In no. 2, this in-
versed structure is expanded into a complex sentence. But due to the syntactic
differences between English and Danish,15 this is a situation where a more con-
cise, acceptable rendering is hard to find. For obvious reasons, the time available
for reading the bulky subtitle no. 2 is far too short, but the wording is perfectly
idiomatic.
 Henrik Gottlieb: Subtitling: Diagonal Translation 115

Of course, inflated constructions are usually time-consuming, but brevity, on

the other hand, is not always beautiful. Sometimes, longer synonyms read better
than shorter ones. Often, word frequency is what counts. A simple example: in
Danish, 'blindtarmsbetœndelse' is read faster (and understood by more people)
than 'appendicitis'. The Danish lay term literally translates as 'blind intestine in-
flammation', so it is self-explanatory. As opposed to English, which relies heavi-
ly on Greco-Latin semantic building blocks, the Scandinavian languages and Ger-
man contain many 'hard words' composed of Germanic morphemes, reserving
several linguistic imports (such as the Danish medical term 'appendicitis') for
professional lingo.
Downloaded by [University North Carolina - Chapel Hill] at 07:11 26 June 2013

8) Elegant and precise cueing

In the previous chapters we have heard that a good subtitler (apart from be-
ing a good translator) needs the musical ears of an interpreter, the no-nonsense
judgment of the news editor and the designer's sense of esthetics. In order to pre-
sent the subtitles in a synchronous manner, the subtitler must also have the steady
hand of a surgeon and the timing of a percussionist.
Modern films and TV programs are often cut so fast, presenting the audience
with such a wealth of audiovisual information that the subtitles must be very tight
too. Fortunately, action-loaded scenes are (still) predominantly non-verbal. Hav-
ing to add subtitles to such a background would really be a case of overkill. As
for the more verbose scenes and genres, these usually operate with fewer, often
rhetorical, cuts that give the subtitler the minimum of time needed for the sub-
titles to be read. And in order not to work against the rhythm of the original,
hereby delaying the viewing process, subtitles must be presented in synchrony
with the dominant auditive (phrasing) or. visual (cutting) signals. In subtitling for
the cinema, and frequently for video, too, cuts are always respected, even at the
cost of precious reading time. So if the audience cannot read the subtitles as the
dialog is spoken before the next cut, it is just too bad. To the film industry, es-
thetics is what counts, not reading time for subtitles.
In public service television, however, there are exceptions to the rule of syn-
chrony. If an important, fast-spoken utterance is followed by silence and the ca-
mera remains focused on the speaker, a TV subtitler would often keep the sub-
title on the air for a second or two, giving viewers the time needed to read it.
Sometimes, even with a cut right after the utterance - still followed by silence,
 116 1994 Perspectives: Studies in Translatology 2(1)

of course - the television subtitler may prefer to let the title stay on the screen
for an instant. Deciding to let a subtitle linger on less than a second after a cut
is not recommendable, though. That would leave the viewer with a flimsy im-
pression, not much extra reading time, and the feeling that the subtitler just does
not manage a precise cueing. As mentioned previously, any modern subtitling
systems allow you to operate with an exactitude of 40 milliseconds, equaling one
frame. Yet to demand that every subtitle be timed with such precision is probably
being too pedantic. However, most viewers would notice a displacement of, say,
4-8 frames. Displacements of this magnitude tend to leave a spasmodic impres-
sion, but programs marred by conflicts between the audiovisual flow and the sub-
Downloaded by [University North Carolina - Chapel Hill] at 07:11 26 June 2013

title rhythm have become rare now that subtitlers realize that poor cueing can
ruin good subtitles.
Needless to say, people do not go the movies or turn on their TV set to
watch subtitles. These are but a means to an end, and the ideal subtitling is the
one that gives the audience the feeling of understanding the film, forgetting what
language it was in. 'Transparent' cueing is a necessary prerequisite in obtaining
this wonderful illusion.

9) Meticulous proofreading and -listening

TV critics and viewers alike often focus on this last aspect of subtitling. Mis-
spellings, bad punctuation and typing errors are often attacked, and rightly so.
Just as imprecise cueing impedes the viewers' steady reception of a program, or-
thographic errors act as tripwires, distracting the audience. But just as words and
commas need correct treatment, so do other humble carriers of information. For
instance, some subtitlers, as other people with a literary background, have a prob-
lem with numbers.
In one episode of the TV comedy Empty Nest (DR-TV November 11, 1992),
we encounter two numeric errors: "Wednesday the third", pronounced slowly and
distinctly, becomes "Onsdag d. 23" Later in that episode "A dollar, fifteen cents
an hour" is rendered as "$1,50 mere i timen".
The fact that we are dealing with fiction does not allow for this strange kind
of poetic license. The informative aspect being secondary in the comedy genre
should not lead to sloppyness with facts, no matter how insignificant they seem.
But figures are tricky, even to native speakers, so in this case the script may have
been wrong. Then again, this problem only stresses the need for 'prooflistening'.
 Henrik Gottlieb: Subtitling: Diagonal translation 117

something that literary translators do not have to do, and interpreters - to their
regret - are unable to do.

Pivot translations: rationality at a price

A new potential source of errors in the world of subtitling is found in the in-
creasing use of pivot translations, defined as 'translations produced not from the
original, but from a existing translation in another language'. Satellite transmis-
sions across language barriers make pivot subtitling a financially attractive meth-
od of subtitling major-language material for a series of smaller speech communi-
ties, often simultaneously broadcast - via satellite."
Pivot subtitling usually works this way: a TV film is subtitled from language
Downloaded by [University North Carolina - Chapel Hill] at 07:11 26 June 2013

1 (typically English) into language 2 (the process described earlier in this article)
the end product being a floppy disk, cued and ready for broadcast. This disk is
then copied and sent to the people in charge of subtitling in, say, languages 3 and
4. The subtitles in these languages are then created by simply overwriting the
pivot subtitles, or - in the case of more conscientious subtitlers - by deleting the
pivot subtitles and then putting new wine into old bottles, by fitting 'local' titles
into 'foreign' dialog segmentation and time cueing (see table 3).
Table 3
NORMAL SUBTITLING PIVOT SUBTITLING

Rim (original version) Rim (original version)

: ;
^ ^ Pivot
Rim Rim subtitles
(with L3 subtitles) (with L4 subtitles)

Rim Rim
(with L3 subtitles) (with L4 subtitles)

As with any pivot translation, pivot subtitling is typically used where the
pivot language is closer to the target language, or more well-known in the target
culture, than the source language is. But in the process of subtitling, where the
translation proper only accounts for a part of the work and money invested, a
prospective pivot language need not be particularly close to the target language
in question: as long as the segmentation and cueing can be 'borrowed' from the
 118 1994 Perspectives: Studies in Translatology 2 (1)

pivot subtitles, time and money are saved. However, the practice of pivot subtitl-
ing implies four potential pitfalls, not normally found in subtitling:
1) Repetition of translation errors present in the pivot subtitles;
2) Transfer of pivot-language features not acceptable in the target language;
3) Transfer of segmentation incompatible with target language syntax;
4) Transfer of subtitle layout and cueing inferior to existing national standards.17
Pivot subtitling is also found in the home video market, but rather than stay-
ing a somewhat grotesque exception to normal practices, it is gaining ground: the
largest TV channel in Scandinavia, TV3, transmitting all programs from London
via the Scansat satellite, utilizes this method in no less than 80% of its Danish
Downloaded by [University North Carolina - Chapel Hill] at 07:11 26 June 2013

programing.18 In all cases, the subtitler receives a Swedish disk in due course,
and then it is up to him or her to produce a Danish version. As the subtitler is
not paid to do any cueing, the Swedish segmentation, often different from Danish
practice, is followed.
But just as teletext gives the owners of commercial TV stations the opportu-
nity to buy clusters of subtitlings at bargain prices, teletext technology opens an
alternative scenario: personal subtitling. By this term I mean a scenario in which,
for the first time, the viewer is able to choose, not just between different target
languages, but between different styles or levels of subtitling. Via his remote con-
trol unit each viewer watching a foreign-language program would then select, for
instance, 1) a no-subtitling option, 2) a fast, uncondensed subtitling, 3) a normal-
speed subtitling, 4) a subtitling for slow readers, 5) a pictogram-supported subtitl-
ing for the deaf and hard-of-hearing, 6) a subtitling in a domestic minority lan-
guage, or finally, 7) a foreign-learner subtitling in the source language."
Depending on the individual program and its expected target audience, differ-
ent clusters of options could be offered, should TV companies consider five or
six subtitlings per program too costly. But as dubbing is 15 times as expensive
as subtitling,20 even the full range of versions could be offered for just over one
third of the language transfer price paid today in Europe's dubbing countries.
Apart from sheer conservatism, the only obstacle to changing the present si-
tuation (with major speech communities dubbing all foreign-language programs
and minor countries using 'one-size-fits-all' subtitles) is the fact that, even today,
almost half of the existing TV sets lack the teletext facilities necessary for receiv-
ing 'tailored' subtitles. But this is just a matter of time; except for portables, all
TV sets sold in (Western) Europe nowadays come with teletext. And with new
 Henrik Gottlieb: Subtitling: Diagonal Translation 119

high-resolution character generators, future TV will display subtitles with a letter

quality and a richness in fonts not found in traditional subtitling.
Hopefully, with improved teletext technology, personal subtitling will set new
standards of language transfer in future television.

Notes
1. Revoicing is a hyperonym covering all types of translation adding new verbal material to the
soundtrack of a film or television program. The major types of revoicing are dubbing, comment-
ary and voice-over. The difference between the two latter types is that whereas in informative pro-
grams commentary replaces off-screen narration by a target-language version, voice-over (found
in both fiction and news programs) merely adds a one-voice translation to the original soundtrack,
still partly audible.
Downloaded by [University North Carolina - Chapel Hill] at 07:11 26 June 2013

2. As one of the minor speech communities in Europe, Denmark subtitles all foreign-language
programs aimed at an adult audience. A mapping of Europe in terms of favored type of screen
translation (see Gottlieb 1994) shows the watershed between on one side (larger) dubbing coun-
tries such as Germany, Spain, Italy, and France, and on the other side subtitling communities,
such as Scandinavia and the Low Countries.
3. The figures are based on earlier studies, showing that during 22 per cent of the time spent
watching TV, Danes were presented with (interlingual) subtitles on the screen (see Gottlieb 1991:
108-109). According to the official Danish Cultural and Media Statistics 1980-1992 (106; 128),
in 1992 the private TV consumption amounted to 17 hours and 16 minutes a week per Dane
(above the age of 13). Adding to this, members of video households (46% of all homes) watched
video 3 hours a week. (Source: the Danmarics Radio weekly, DRåben no. 6, 1993.)
4. In 1988, the Danish National Institute of Social Research (Socialforskningsinstituttet) published
a number of studies of daily life in Denmark. One of these books, dealing in particular with how
much time people spend on their leisure-time activities, has served as a useful database (Andersen
1988).
5. For the terms 'overt' and 'covert' I am indebted to Juliane House (see House 1981: 189; 194)
and R. Battarbee (1986).
6. Actual dubbings are often problematic, however (Vöge 1977). Thorough discussions of the
problems of dubbing are found in Götz & Herbst (1987) and Zabalbeascoa (1993). A more critical
stance is found in Danan (1991), warning of the potential and actual ethnocentrism involved in
dubbing.
7. For a definition of vertical sublitling see the following page.
8. The formal and textual constraints of subtitling are discussed at length in Gottlieb 1992: 164-
166.
9. When dealing with non-fiction material, in which terminology plays a greater role, this freedom
is markedly reduced: some instructive programs and documentaries are so informative that finding
an adequate expression is no longer the question; the problem here is finding a way to fit these
often lengthy and time-consuming terms into the subtitles.
10. In the U.K., as in America and many dubbing countries, vertical subtitling (for the deaf and
hard-of-hearing) is far more common than diagonal subtitling, which is aimed at (in such coun-
tries) "the under 50's, the better educated and more affluent, as well as students and other intel-
lectual minorities,... and those with an interest in the original language of production". (Luyken
et al. 1991: 189)
11. This handbook, Subtitling for the Media, was the first book to deal with (diagonal) subtitling
in an international context.
 120 1994 Perspectives: Studies in Translatology 2 (1)

12. In this and the following TV examples, 'DR' stands for Danmarks Radio, the non-commercial
Danish TV broadcasting authority, while 'TV2' is the up-and-coming, partly commercial Danish
rival. Both networks are national, public service TV stations, covering 75% of the total Danish
television viewing (see Danish Cultural and Media Statistics 1980-1992: 107).
13. The "moving" television image is made up of 25 different frames per second, and the duration
of each subtitle, in seconds and frames, is determined by the subtitler's segmentation and may be
time-coded with an exactitude of one frame. On television and video, the timing of the subtitles
is normally decided and performed solely by the subtitler him/herself, as discussed in section 8.
14. The important aspect of how to subtitle culture-specific elements is treated exhaustively in Ne-
dergaard-Larsen 1993.
15. If the target language generally uses a more complex syntax than the source language, a heavy
reduction may be necessary. See Kovačič 1994, dealing with the language pair English-Slovenian.
16. The European Institute for the Media recommends pivot subtitling done ' the other way round' :
"The larger target languages should again be seen as 'pivot languages' to which increased Lan-
Downloaded by [University North Carolina - Chapel Hill] at 07:11 26 June 2013

guage Transfer should be directed from the less widely spoken languages." (Luyken et al. 1991:
192)
17. To be fair, well-made pivot versions may clear the ground of semantic land mines, thus yield-
ing better target-language results than might otherwise have been achieved, presuming the target-
language subtitlers are lesser talents. But without this presumption, pivot subtitlings as safety nets
are 'unsafe at any speed.'
18. Programs are subtitled via teletext into Swedish, Danish and Norwegian. This figure was
given to me in the spring of 1993 by the manager of Dansk Video Tekst, which produces all Da-
nish TV3 subtitles. As for Norway, I presume the use of pivot subtitling for TV3 is of the same
magnitude as in Denmark.
19. The potential of teletext for viewer-specific subtitling is explored in Jørgensen 1992.
20. According to a major research report, published by the European Institute for the Media
(Luyken et al. 1991: 105), the average cost of dubbing a one-hour TV program is 11,000 ECU,
as opposed to 740 ECU for subtitling the same program.

Works cited
Andersen, Dines. 1988. Danskernes dagligdag 1987. De 16-74-driges tidsanvendelse i vintermåne-
derne 1987. Socialforskningsinstituttet. Rapport 88: 4. 2 vols.
Baker, Robert G. & Andrew D. Lambourne & Guy Rowston. 1984. Handbook for Television
Subtitlers. Independent Broadcasting Authority.
Battarbee, K. 1986. Subtitles and soundtrack. In: Gambier, Yves (Ed.) Trans. School of
Translation Studies, Turku University. 144-165
Catford, J.C. 1965. A Linguistic Theory of Translation. London: Oxford University Press.
Danan, Martine. 1991. Dubbing as an expression of nationalism. Meta, XXXVI. 606-614
Danish Cultural and Media Statistics 1980-1992. 1993. Copenhagen: Danmarks Statistik.
Gottlieb, Henrik. 1991. Tekstning - synkron billedmedieoversættelse. Copenhagen: University of
Copenhagen. Gold medal prize thesis.
Gottlieb, Henrik. 1992. Subtitling - A new university discipline. In: Dollerup, Cay & Anne
Loddegaard, Anne (Eds.). Teaching Translation and Interpreting. Amsterdam & Philadelphia:
John Benjamins Publishing Company. 161-170.
Gottlieb, Henrik. 1994. Subtitling: People translating people. In: Dollerup, Cay & Annette
Lindegaard (Eds.) Teaching Translation and Interpreting 2. Papers from the second "Language
International" Conference, Elsinore, June 4-6, 1993. Amsterdam/Philadelphia: John Benjamins.
 Henrik Coltlieb: Subtitling: Diagonal Translation 121

Götz, Dieter & Thomas Herbst. 1987. Der frühe Vogel fängt den Wurm: Erste Überlegungen zu
einer Theorie der Synchronisation (englisch-deutsch). AAA - Arbeiten aus Anglistik und Amerika-
nistik XII. 13-26.
House, Juliane. 1981 [2. edition]. A model for translation quality assessment. Tübingen:
Gunter Narr.
Ivarsson, Jan. 1992. Subtitling for the Media: Handbook of an Art. Stockholm: Transedit.
Jakobson, Roman. 1966. [Original edition 1959]. On Linguistic Aspects of Translation. In:
Brower, Reuben A. (Ed.) On Translation. New York: Oxford University Press. 232-239.
Jørgensen, Tom Bernbom. 1992. Målgruppeorienteret tv-tekstning. In: Nedergaard-Larsen, Birgit
& Henrik Gottlieb (Eds.) Tv-tekster: Oversættelse efter mål (= DAO 3). Copenhagen: Center for
Translation Studies: University of Copenhagen. 25-73.
Kovačič, Irena. 1994 [forthcoming]. Relevance as a factor in subtitling reductions. In: Dollerup,
Cay & Annette Lindegaard (Eds.) Teaching Translation and Interpreting 2. Papers from the
second "Language International" Conference, Elsinore, June 4-6, 1993. Amsterdam/Philadelphia:
John Benjamins.
Downloaded by [University North Carolina - Chapel Hill] at 07:11 26 June 2013

Luyken, Georg-Michael & Thomas Herbst & Jo Langham-Brown & Helene Reid & Hermann
Spinhof. 1991. Overcoming language barriers in television. The European Institute for the Media
Muylaert, W. & J. Nootens & D. Poesmans & A.K. Pugh. 1983. Design and Utilization of
Subtitles on Foreign Language Television Programmes. In: Neide, P.H. (Ed.) Theorie, Methoden
und Modelle der Kontaktlinguistik. Bonn: Dümmler. 201-214.
Nedergaard-Larsen, Birgit. 1993. Cultural factors in subtitling. Perspectives: Studies in
Translatology, 1993:2. 207-241.
Praet, Caroline & Karl Vervaille & Peter de Graef & Johan van Rensbergen & Géry d'Ydewalle.
1990. A one line text is not half a two line text. In: Groner, R. & Géry d'Ydewalle & R. Parham
(Eds.) From eye to mind: Information acquisition in Perception, Search and Reading. North-
Holland: Elsevier Science Publishers. 205-213.
Reid, Helene. 1990. Literature on the Screen: Subtitle Translating for Public Broadcasting. In:
Westerweel, Bart & Theo D'haen (Eds.). Something Understood. Studies in Anglo-Dutch literary
translation. Amsterdam: Rodopi. 97-107.
Vöge, Hans. 1977. The Translation of Films: Sub-Titling versus Dubbing. Babel, XXIII (3). 120-
125.
d'Ydewalle, Géry & Johan van Rensbergen & Joris Pollet. 1985. Reading a message when the
same message is available auditorily in another language: The case of subtitling. (= Psychological
Reports of Leuven University 54).
d'Ydewalle, Géry & Caroline Praet & Karl Verfaule & Johan Van Rensbergen. 1991. Watching
Subtitled Television. Automatic Reading Behavior. Communication Research 18, # 5, October
1991.
Zabalbeascoa, Patrick. 1993. Developing Translation Studies to better account for Audiovisual
Texts and other New Forms of Text production. Spain: University of Lleida. Doctoral dissertation