MA in Translation & Interpreting – Helen J Lever

3.0 CHAPTER TWO – SUBTITLING

This chapter will look at the concept of subtitling. Definitions will be

provided of the main concepts and an explanation of the general theory

behind subtitling will be given. Discussion will then move on to a more

detailed analysis of the constraints associated with subtitling and the

implications in terms of translation. Finally, the impact of these constraints

on cultural features will be examined, paying particular attention to the

elements of culture outlined in the previous chapter, namely dialect and

accent.

Subtitles can be defined as ‘transcriptions of film or TV dialogue, presented

simultaneously on the screen’.1 The types of subtitling available can be

divided into two categories: intralingual and interlingual subtitling.

Intralingual subtitling deals with the production of subtitles that remain in

the same language as the original and are used for the deaf or hard of

hearing, or for language learners. Henrik Gottlieb describes it as vertical,

‘in the sense that it involves taking speech down in writing, changing mode

but not language.’2 Interlingual subtitling, on the other hand, refers to both

a change in mode and language, going ‘from one language into another

language, and from spoken dialogue into a written, condensed translation

which appears on the screen.’3 Therefore, for the purposes of this work,

“subtitling” will refer to the interlingual subtitling described above.

- 30 -
 MA in Translation & Interpreting – Helen J Lever

When looking at subtitling, from the point of view of translation, many

constraints exist that will effect the translation in question and cause the

translator to modify the source text to a certain degree. Firstly, there are

synchronisation constraints, which affect a number of aspects of language

transfer. The very fact that the medium changes from oral to written will

cause problems, given that it generally takes longer to process information

received through the eyes than through the ears. So, if, in the spoken text, a

large amount of information is provided, in the subtitles it will be necessary

to modify the way this information is presented without losing any of it, in

order to simplify the delivery and account for the change of medium.

Similarly, if the length of certain utterances is too great or the structure too

complex, it will be necessary to make some modifications in order to

facilitate understanding through the written format. A long, complex

sentence, although easy to comprehend if heard, would be very difficult to

follow if it were left unaltered, given the timeframe available to read

subtitles. For example, if we take the following narration as the spoken

version of a text:

“In beautiful autumnal colours, the stunning models glided around the

splendid arena, oozing sophistication and elegance, doing rightful

justice to the imaginative designers.”

the subtitled written would more than likely be rendered in a more simplistic

fashion, such as :

- 31 -
 MA in Translation & Interpreting – Helen J Lever

“In the arena the designers’ outfits were modelled with style and

elegance.”

This also relates to the reading speed of the viewer, which needs to be

taken into consideration. Clearly, establishing the profile of the target

audience will be crucial in assessing the viewer’s potential reading speed

and is an important factor for the translator. In this aspect, literacy and age

are both influencing factors. For example, there is likely to be a great deal

of difference between an adult’s and a child’s reading speed, and, as such,

the subtitles would need to be altered accordingly, for example, by

simplifying the vocabulary for a child. Also, if the subtitled version of a

broadcast were to be shown in a country with a low level of literacy, it

would be necessary to ensure the subtitles were fairly basic. So, ‘an

accurate assessment of audience literacy and of the viewers’ knowledge of

the subject matter is therefore essential to the success of subtitling’.4

Another constraint, in terms of synchronising the subtitling, is the

correlation with the visuals on-screen and the sound track. In other words,

the subtitles must link up with what the viewer is seeing and hearing on the

screen, so they are able to maintain their coherence. This is especially

important if the translator is dealing with a visual gag or comedy element, in

which case it would not be correct to spoil the surprise before the joke has

happened. The same would apply in a “whodunit” type of scene, where it is

vital to the story for the viewer not to be aware of a clue or perhaps the

identity of a culprit until the appropriate moment. Frame changes represent

another constraint to the translator. If a subtitle appears over a frame

- 32 -
 MA in Translation & Interpreting – Helen J Lever

change, it is likely that the subtitle will be lost to the viewer, as their

attention is likely to be diverted away from the subtitle, due to the change in

scene. Therefore, the translator must ensure that the subtitles have

disappeared from the screen by the time the frame changes. However,

sometimes this is not always possible. As Josephine Dries points out, ‘a

good subtitler will wish to avoid a subtitle staying on screen across a shot

change, and certainly across a scene change. When cuts follow each other

too quickly, the minimum reading time (in most countries 1 ½ seconds, in

some 1 second) sometimes forces the subtitlers to break this law’.5 A final

constraint, which must be accounted for, involves the medium through

which the subtitles are to be shown, which, in turn, will affect factors such

as font size and audience age range. In cinemas, for example, the size of the

screen is much larger and therefore, will allow more flexibility with regard

to the font size of the subtitles, making the reading of them easier and more

comfortable for the viewer, whereas, on television, the translator is

restricted to a smaller screen space, thereby limiting the size of the font.

Also, the type of audience likely to be attracted by these two mediums will

differ, in which case the translator’s task is made more difficult as they are

less likely to be able to assess the reading speed and literacy of the audience.

However, as Ivarsson and Carroll point out, ‘whatever the medium, the

principle of maximum legibility always applies’.6

Secondly, throughout this discussion, it is clear that there are two main

factors relevant to achieving successful subtitling: legibility and readability

or simplicity. Strategies have been developed focusing on these two

- 33 -
 MA in Translation & Interpreting – Helen J Lever

elements, in order to provide suitable guidelines for subtitlers. First of all,

let us look at legibility. To achieve legibility, subtitlers must focus on the

appearance of the subtitles and the timing. With regard to the appearance of

the subtitle, for languages that read left to right, the positional options are

for text to be either centred or left-aligned. Historically, subtitles had

always been centred, mainly due to the poor quality of television sets, which

tended to be fuzzy round the edges. So, in order not to lose any of the text,

they were positioned in the centre of the screen. However, as technology

has improved and, so, presents fewer restrictions, today’s convention

favours more the left-aligned approach, as it is suggested that it is easier for

the eye to focus on a new line if it begins in the same spot on the screen. In

other words, the viewer is likely to read more quickly if his or her eyes are

not continually having to dart around the screen to find the beginning of the

speech. The position of the subtitle is something that can be accounted for

when a programme is in its production stage, especially if the producers are

aware that the production is going to be exported and, therefore, subtitled.

In her guidelines for production, Josephine Dries suggests that ‘since the

subtitles usually cover the lower 15% (two lines) of the screen, care should

be taken that the mouth of the speaker is slightly higher on the screen than it

would normally be. Viewers focus on the moving part of the picture, i.e. the

mouth in a close-up, which should not be covered by the subtitle’.7

As well as the position of the subtitles, the number and length of the lines

appearing on screen at any one time must be taken into consideration. The

general consensus is that two lines of subtitles should be the maximum to

- 34 -
 MA in Translation & Interpreting – Helen J Lever

appear on the screen at the same time. In fact, this point is included in the

“Code of Good Subtitling Practice” developed by Ivarsson and Carroll for

discussion at the ‘Languages and the Media’ conference in Berlin, October

15-16, 1998. Point 19 states ‘the number of lines in any subtitle must be

limited to two’8. However, this is considered as the maximum limit not the

preferred option. It is suggested, therefore, that ‘if it is possible to write the

text of any one subtitle in a single line, it is usually better to do so’. 9 Of

course, the medium for which the subtitles are intended will also have an

impact on this, as the size of the cinema screen may justify having two lines

rather than one. This, then, leads on to the issue of the line breaks. If there

are two lines of subtitles, where should the break between them occur?

Generally, one line of subtitle should have between 35-40 characters, which

implies that a lengthier subtitle will need a break at some point. As such,

one line is likely to be longer than the other. Here, there is a connection

with the position of the subtitle on the screen, as discussed earlier. If the

subtitles are centred, it does not matter which of the lines is longer, as the

distant the eye must travel to look at the beginning of the second line will be

the same either way round. However, if the subtitles are left-aligned, the

eye will have much further to move if the first line is much longer than the

second, in which case it is better for the first line to be the shorter of the

two. This is clearly illustrated, by Ivarsson and Carroll, in the following

diagram:10

- 35 -
 MA in Translation & Interpreting – Helen J Lever

e.g. Centred:

He said that he would not be able to come

until tomorrow.

He said

that he would not be able to come until tomorrow.

Left-aligned:

He said that he would not be able to come

until tomorrow.

He said

that he would not be able to come until tomorrow.

In either case, it is very important that the line break does not affect the

meaning of the subtitle, by splitting the sentence in such a way as to alter

the viewer’s understanding of what is being said.

The typeface or font used for the subtitles needs, also, to be considered.

Very ornate fonts are usually quite difficult to read whatever the medium,

and, so, would be especially difficult in a subtitle format, particularly on the

small screen. Also, text written in uppercase tends to be harder to read than

text written in the standard lowercase and uppercase structure. Therefore, it

is necessary, in order to enhance legibility, to make sure the subtitles are

presented in a basic, simple font, in the standard lowercase / uppercase

format. The use of italics is often introduced to denote a distant voice, a

voice on the telephone or off screen, quotes or internal voices such as those

- 36 -
 MA in Translation & Interpreting – Helen J Lever

in dreams or flashbacks. As such, it is important that they are reserved for

this specific use, in order not to confuse the viewer.

The colour of the subtitle on the screen will also affect its legibility, in

comparison to the background on which it is set. Clearly, if the subtitle is in

white and there is a lot of white or light-coloured background, the subtitle

will not stand out sufficiently to be seen, let alone read. Likewise, the same

would apply for black subtitles on dark or black backgrounds.

Unfortunately, it is unlikely that the same level of darkness will appear in

the background throughout the film. Therefore, one possibility is to place a

box behind the subtitles to contrast against the colour of the text. However,

the technology available to the subtitler or production team will determine

whether or not this is a feasible option. As point 3 of the technical aspects

of the “Code of Good Subtitling Practice” states, ‘in video applications,

character clarity can be enhanced by a drop shadow or semi-transparent or

black box behind the subtitles.’11

Finally, punctuation is a feature of the appearance of the subtitle that needs

to be addressed, as it can easily distract the viewer and distort the meaning

of the utterance. As such, it is best to keep punctuation to a minimum and

to strike a fair balance between providing enough punctuation for the viewer

to understand the meaning and providing too much for it to become a

distraction. As an example, let us look at the following two sentences:

“He said, “I can’t come ‘til 6 o’clock!”

“He said he can’t come until 6.00 pm”

- 37 -
 MA in Translation & Interpreting – Helen J Lever

If we take into consideration the amount of time the subtitle is to remain on

the screen, reading speed etc. then it is clear that the second sentence will be

easier to read as the unnecessary punctuation has been removed and,

therefore, does not cause a distraction. In some cases, however, it is

necessary to include punctuation, for example, when two different people

are speaking, which is often denoted by a dash “ – ” separating the two

speakers. In essence, Luyken sums it up by saying ‘distractions such as

complex sentences, abbreviations, unnecessary punctuation, incomplete

sentences and ambiguities (unless reflecting ambiguity in the source) must

be avoided’.12

The second area relevant to achieving legibility is that relating to timing.

Unfortunately, it is very difficult to set fixed timings for subtitling, as there

are a number of factors that can impact on reading behaviour, such as ‘the

level of literacy of the target audience, the viewer’s degree of familiarity

with the programme language, the genre of the subtitled programme and the

pace of the visuals’.13 However, firstly, in terms of the length of time one

subtitle should remain on the screen, guidelines have been proposed

suggesting that somewhere between one and seven seconds would be

appropriate. Again, this will depend on the factors mentioned but also on

the length of the subtitle, whether it is one or two lines, a short or a long

utterance. In relation to this, Ivarsson and Carroll have suggested the

following:

‘As a general rule, the minimum time for even a very short subtitle on

a television screen is at least one and a half seconds and the maximum

- 38 -
 MA in Translation & Interpreting – Helen J Lever

time for a full two-liner should not exceed five to six seconds. A full

one-liner should be kept on the screen for about three seconds, one and

a half lines for about four seconds.’14

They, also, go on to comment that if a subtitle is left on the screen for too

long, the viewer is then tempted to read it more than once, thereby

destroying the sense of a flowing dialogue. On the other hand, if the subtitle

is a short one, maybe a simple “yes” or “no”, that ‘could theoretically be

taken off the screen after a mere half-second, the risk is high that the eye

will not register it at all. For this reason, even the briefest subtitle should

remain on the screen for about a second and a half’.15

After establishing a suitable length of time for the subtitles to appear on

screen, another concept relating to timing concerns the amount of time that

should be left between subtitles. Again, Ivarsson and Carroll explain how

‘experience shows that if subtitles are placed back to back without a

minimal pause between them, the eye often does not register that a new

subtitle has appeared, but remains fixed on some other part of the screen’.16

Clearly, therefore, some sort of gap is needed for the viewer to

subconsciously acknowledge that the subtitles have changed and new

dialogue has been presented. Ivarsson and Carroll suggest that ‘to eliminate

this problem it is recommended to leave approximately four frames or the

equivalent of about one sixth of a second between subtitles, even when the

dialogue is ongoing’.17 This, inevitably, creates difficulties if the speech is

rapid or if a dialogue between two people contains very short utterances,

- 39 -
 MA in Translation & Interpreting – Helen J Lever

which can undoubtedly be uttered faster than they can be read. It is possible

to overcome this by borrowing time from the gaps that occur naturally

within a programme. However, the main techniques used to combat this

problem relate to ways of achieving readability.

All the constraints previously outlined will come together under the heading

of readability and influence the decisions made by the translator. Clearly, it

is here that the translator must consider ways to present the text so that it is

readable, taking into consideration constraints such as reading speed and

genre, but also bearing in mind the space available on the screen, the time

the subtitle will be shown, the frame changes and all the other limitations on

their translation abilities. In order to do this, there are two techniques used

to achieve readability; one is by way of reduction and the other, by way of

simplification, and it is here that we can look more closely at the

implications of subtitling constraints on translation. Firstly, reduction can

be achieved by simply using fewer words to render the same meaning.

Given that we are dealing with interlingual subtitling, this technique could

also be useful for any reference that lacks an equivalent in the foreign

language. For example, if the term “metropolitan police force” were to

appear in a dialogue, it would be unlikely to have an equivalent in other

languages. Therefore, unless it was a key element of the story, it would be

better to simply translate it as “police”, whilst at the same time adhering to

the technique of reduction.

- 40 -
 MA in Translation & Interpreting – Helen J Lever

Another form of reduction is to simplify the syntactical structure of the

spoken sentence. For example, if the spoken dialogue is given as “What

time can we expect your son to arrive?” the subtitle can feasibly become

“What time is your son arriving?” Or, sentences that begin with “Are we

likely to…” can be simplified by rendering them as “Will we…”. This may

well be a process that is imposed on the translator by the structure of the

foreign language into which they are translating. But equally such

sentences may represent a stylistic way of speaking, which would need to be

rendered in the foreign language. Therefore, the translator would need to

achieve a balance between maintaining the style but, at the same time,

working within the constraints.

Merging short dialogues together can be another successful technique used

to achieve readability. Let us look at the following dialogue:

- You need to fill in this form.

- OK.

- Then you have to take it to the issue desk.

- Right.

- You must have your passport with you, too.

- That’s fine.

It would be perfectly feasible to render this dialogue in the equivalent

written form for subtitles as:

- 41 -
 MA in Translation & Interpreting – Helen J Lever

- You must fill in this form, take is to the issue desk and take your

passport.

- OK.

In this way, the viewer can fully grasp the meaning of the conversation, and

the instructions issued within it, without being distracted by the extra

affirmations within the dialogue. Clearly, the more natural a conversation is

the more likely it is for such extra utterances to occur and, as such, the

translator will almost be in a position of having to summarise the

conversation. For the purposes of subtitling, this would be quite acceptable

as the viewer will still have access to the original voices, albeit in a foreign

language, so can subconsciously add a “yes” or a “no” to their

understanding of the dialogue.

Finally, the subtitler can take advantage of the visual element linked to the

subtitles in order to delete certain parts of the dialogue that are, by and

large, redundant. ‘Any information which may be gleaned from the actor’s

performance ought not to be reproduced in the text.’18 Again, this would

particularly help if a cultural reference required a lengthy or complex

explanation, but could quite easily be understood through the visuals. For

instance, in a scene where two people are in a bakery discussing whether to

buy a Danish pastry or a Bakewell tart, there would be no need for the

subtitler to attempt to explain these references or try to translate them

precisely, as the visuals would provide a more than satisfactory description.

- 42 -
 MA in Translation & Interpreting – Helen J Lever

Simplification, however, rather than making changes regarding the amount

of text appearing on the screen, deals with the type or quality of text that

appears on the screen. Firstly, the very fact that the medium changes from

spoken to written when subtitles are introduced, will inevitably cause the

loss of such aspects of speech as intonation. This can, to a certain extent, be

rendered in the written format by way of punctuation. However, in

subtitles, as has been mentioned, the inclusion of punctuation can affect the

legibility of the text. Therefore, it is likely that these aspects will incur

greater loss when a text is subtitled, rather than, say, translated in the book

form, which allows for the inclusion of more punctuation and more

explanation, for example, in the form of adding “she said in a surprised

tone”. A common feature of simplification is, therefore, that marked

language will be made more neutral or standard. In films and television

programmes, this can have an important effect on the characters portrayed,

as their speech could well be an intentional feature used to establish certain

facts about the characters. This can be seen in a grammatical context, for

example, where very formal or informal speech is used, perhaps, to convey

the relationship between two characters, or where fairly ungrammatical

speech is used to show that a character has had a poor education. These

aspects, if maintained in the subtitles, would hinder the readability, and,

therefore, the common practice is to convert them into more standard

varieties of the language. So, a character with little education who uttered

“I isn’t stupid”, would probably appear in the subtitled version as saying

“I’m not stupid”, which would then suggest a higher level of education than

was originally intended by the author, or, indeed, than other aspects of the

- 43 -
 MA in Translation & Interpreting – Helen J Lever

character would suggest. It would also lose the comic element of somebody

trying to maintain that they are more intelligent or educated than people

think, but doing so by using a phrase that is grammatically incorrect, in

relation to standard variety norms. This would then create certain

inconsistencies between the visuals and the dialogue.

Another type of language that tends to be standardised in the same way is

that used within specialised genres, such as literary or technical language,

which require some sort of prior knowledge, particularly if being read. This

would occur, for instance, in the subtitling of medical or forensic

programmes. It would be impossible for the subtitle viewer of, say, ER to

follow what was happening if all the technical information were rendered in

the subtitles. As it is, the majority of the listening viewers will no doubt not

understand exactly what is being discussed, but grasp the gist. As such, it is

necessary to ensure that the subtitles provide the same effect. Clearly, the

disadvantage is that the reader will not have the impression of

professionalism, technical expertise and realism if the staff of the ER talks

about “blood pressure” rather than “BP”. Luyken points out that another

reason for this standardisation could be to do with the customs of the

country where the programme is to be aired. In the field of politics, he

states that ‘highly emotional and emotive speeches might be the oratorical

fashion in one country but be utterly repugnant to viewers in another.’19

Strong language is another type that tends to be toned down in subtitling or

even not translated. The reasons for this could, in fact, be out of the

- 44 -
 MA in Translation & Interpreting – Helen J Lever

subtitler’s hands, if censorship prohibits the rendering of swear words.

However, it could also be the subtitler’s own personal preference to tone

down the bad language used within a script. The problem that occurs if

strong language is toned down, again, relates to characterisation. There are

countless films in which the “bad guys” use such strong language, and, in

fact, it could be argued that the more swear words they use the nastier they

are portrayed. So, in cancelling this effect out by neutralising the language

in the subtitles, the traits of the character are altered, thereby perhaps

making them appear less threatening and evil than they are intended to be.

Another problem that exists, however, in this context relates to the

acceptability of swear words in the foreign language and the implications of

their use. For example, in Spanish, swear words are more frequently used

than in English and, as such, have far less impact on the viewer than its

English equivalent. Also, because the strength of different swear words

varies, it may be difficult for the subtitler to gauge and therefore, the

equivalent chosen may reflect either a stronger or weaker version of the

original, again, affecting the overall impression for the viewer.

Similarly, another area where neutralisation commonly occurs is with the

use of language reflecting social and cultural origins of the speaker; in other

words, the use of dialects and accents. This tends to become the standard

variety of a language in the subtitles, as it is very difficult for viewers to

read a written dialect. The fact that most written publications are produced

using the standard variety, as discussed in the previous chapter, means that

viewers are accustomed to reading that variety. Therefore, if they are then

- 45 -
 MA in Translation & Interpreting – Helen J Lever

faced with a different variety, offering, perhaps, different spelling, grammar

or vocabulary, it becomes more difficult and will take longer to read. So,

for example someone in a film who says something along the lines of:

“’e aint done nuffin’ yet!”

is more likely to be portrayed in the subtitles as having said the equivalent

of:

“He hasn’t done anything yet!”

This, again, can have an unwanted effect on the characters being portrayed,

as the social implications associated with the way a person talks cannot be

rendered in the written form. This is also especially difficult in the context

under discussion, namely, interlingual subtitling, as the social implications

associated with a dialect of one language, may not be the same as the social

implications associated with the dialect of another language. As such, by

rendering the subtitles in a dialect form of the target language, the subtitler

may be associating an inappropriate set of social features with the character

in question. However, if the storyline demands that viewers understand

some sort of social differentiation between characters, it will be necessary to

compensate for it in some way, and, as such, it may be better to accept that,

whilst the two sets of social features are not identical, the idea of

differentiation is being implied.

- 46 -
1
Baker, M. (ed.) (1998), Routledge Encyclopaedia of Translation Studies, London, Routledge, pp 244-245
2
Baker, M. (ed.) (1998), Routledge Encyclopaedia of Translation Studies, London, Routledge, p 247
3
Dries, J. (1995), Dubbing and subtitling: guidelines for production and distribution, Düsseldorf, European Institute for
the Media, p 26
4
Luyken, G-M. (1991), Overcoming language barriers in television: dubbing and subtitling for the European audience,
Manchester, European Institute for the Media, p 56
5
Dries, J. (1995), Dubbing and subtitling: guidelines for production and distribution, Düsseldorf, European Institute for
the Media, p 33
6
Ivarsson, J. & Carroll, M. (1998), Subtitling, Simrishamn, TransEdit, p 39
7
Dries, J. (1995), Dubbing and subtitling: guidelines for production and distribution, Düsseldorf, European Institute for
the Media, p 32
8
Ivarsson, J. & Carroll, M., Code of Good Subtitling Practice, Language Today, April 1998, p 20
9
Ivarsson, J. & Carroll, M. (1998), Subtitling, Simrishamn, TransEdit, p 76
10
Ivarsson, J. & Carroll, M. (1998), Subtitling, Simrishamn, TransEdit, p 77
11
Ivarsson, J. & Carroll, M., Code of Good Subtitling Practice, Language Today, April 1998, p 21
12
Luyken, G-M. (1991), Overcoming language barriers in television: dubbing and subtitling for the European audience,
Manchester, European Institute for the Media, p 56
13
Ivarsson, J. & Carroll, M. (1998), Subtitling, Simrishamn, TransEdit, p 65
14
Ivarsson, J. & Carroll, M. (1998), Subtitling, Simrishamn, TransEdit, p 65
15
Ivarsson, J. & Carroll, M. (1998), Subtitling, Simrishamn, TransEdit, p 64
16
Ivarsson, J. & Carroll, M. (1998), Subtitling, Simrishamn, TransEdit, p 64
17
Ivarsson, J. & Carroll, M. (1998), Subtitling, Simrishamn, TransEdit, pp 64-65
18
Luyken, G-M. (1991), Overcoming language barriers in television: dubbing and subtitling for the European audience,
Manchester, European Institute for the Media, p 56
19
Luyken, G-M. (1991), Overcoming language barriers in television: dubbing and subtitling for the European audience,
Manchester, European Institute for the Media, p 158