RSS feed

On the subtitling of visualised metaphors

Jan Pedersen, Institute for Interpreting and Translation Studies, Stockholm University


There has been a great interest in metaphors in translation in recent years, but few scholars have taken into consideration the extra complexities that are involved when translation takes place in audiovisual media. This paper seeks to remedy that, by qualitatively investigating how visualised metaphors in the TV series Yes, Prime Minister, and other audiovisual sources, have been translated into Swedish subtitles. When a metaphor is visualised, the vehicle of the metaphor is visible on screen, and this means that there is ambiguity between the literal and the figurative sense of the metaphor, with one sense coming through one discourse channel and the other coming via another of the discourse channels that makes up the polysemiotic text. This may cause serious translation crisis points and put severe constraints on the options available to the subtitler if unintended intersemiotic tension is to be avoided. This paper puts forward the theories necessary for handling these complexities and also offers some advice as to what strategies can be useful for solving these translation problems. The results indicate that it is a fairly rare translation problem, and one which can often be solved using conventional strategies. However, occasionally, it creates nearly unsolvable obstacles that cannot be solved without intersemiotic tension.


Visualised metaphors, polysemiotics, intersemiotic tension, audiovisual translation, subtitling, translation crisis point.

1. Introduction

Metaphors constitute one of the great traditional translation problems that have been scrutinised very thoroughly in Translation Studies in recent years (cf. e.g. Lindqvist 2002; 2005, Dickins 2005, Samaniego Fernández et al. 2003, Monti 2006, Mohanty 2010). It is particularly metaphors based on culture, rather than the allegedly universal metaphors that cause problems: what does a translator do when no equivalent cultural notion or experience exists in the target culture? There have been many suggestions for translation strategies that can be used – and are used ­– to solve these problems when a sensu stricto (Van Den Broeck 1981: 77) solution would be infelicitous, most of which are based on substitution of the metaphor or a reduction to sense in one way or another.

What has not been studied when it comes to metaphor translation, however, are the further challenges and layers of complexity caused by the medium of audiovisual translation (AVT). Whereas isosemiotic translation (e.g. literary translation) only has to consider one channel of discourse (i.e. the written word), AVT has to deal with four (Gottlieb 1997: 143). In an audiovisual text (e.g. a film or a TV programme), there is not only the verbal channels of discourse, i.e. spoken language (the verbal audio channel) and written language (the verbal visual channel, e.g. in the form of captions and headlines), but there are also two non-verbal channels: one audio (e.g. music and/or sound effects), and perhaps more strikingly a non-verbal visual channel, which includes everything you see on screen.

The complexities of AVT become foregrounded whenever a metaphor is visualised, so that the verbal audio and the non-verbal visual channels interact very closely. These complexities become even more acute when such a visualised metaphor does not have an equivalent expression in the target language and there is semiotic tension between discourse channels. In these cases the translation strategies that would normally be applied (substitution or reduction to sense) are infelicitous, because how can you replace the metaphorical sense of ‘being stuck between a rock and a hard place’ by a different target image or reduction to sense, if a person is seen dangling between said objects? This paper will look into these issues and give a few tentative suggestions for creative solutions to the problem, using a qualitative analysis of material from the English sitcom Yes, Prime Minister (1986–1988) and The Simpsons Movie (2007).

2. Metaphors and other figures of speech

Aristotle described the essence of metaphors as “understanding and experiencing one kind of thing in terms of another” and in its basic sense, that definition still holds today. There have been many elaborations of this simple definition, for instance by Dickins (2005: 228), who states that “‘Metaphor’ is defined […] as a figure of speech in which a word or phrase is used in a non-basic sense, this non-basic sense suggesting a likeness or analogy (whether real or not […]) with another more basic sense of the same word or phrase.” In conceptual metaphor research, you often find the following structure for describing metaphors ‘X IS Y’ where X is the thing described and Y is the concept used to describe X, as in ‘ARGUMENT IS WAR’ (cf.  Lakoff and Johnson 1980). The terms for X and Y vary a great deal among metaphor researchers, depending on their disciplinary standpoint and other factors. They can be called ‘tenor’ and ‘vehicle,’ ‘source’ and ‘target,’ ‘focus’ and ‘frame,’ ‘image’ and ‘object,’ and other terms also exist for these central concepts in metaphor research. In many ways, the cognitive terms ‘source’ and ‘target’ are intuitively attractive, but since this is a paper in the field of Translation Studies, these terms are already entrenched for use in terms like ‘source text’ (ST) and ‘target text’ (TT), so it might be confusing to use them here. Rather than going into a lengthy discussion about the conceptual justification for using one term rather than another, let us simply say that in this paper, Goatly’s (1997) classic terms will be used: ‘topic’ is the thing that is described (ARGUMENT in Lakoff and Johnson’s example) and ‘vehicle’ is the thing that is used to describe it (WAR) and the likeness between the two is called ‘grounds.’ This last term is also important, since metaphors do not map all the sense of the vehicle onto the topic. If they did, they would quite simply not be metaphors. As Lakoff and Johnson put it, “the metaphorical structuring involved here is partial, not total. If it were total, one concept would actually be the other” (1980: 13).

A metaphor is a figure of speech where the link between topic and vehicle is implicit. This means that we understand the formula ‘X IS Y’ as X IS (in some salient sense similar too) Y, and the salient part is the grounds. Here is an example from the material used in this study. Prime Minister James ‘Jim’ Hacker and his Permanent Secretary Sir Humphrey Appleby are discussing the layout and functions of various rooms of 10 Downing Street, and Sir Humphrey remarks:


Number Ten is a railway junction.


(1:4 “The Key” 6.46)

By which he obviously does not mean that trains go through the Prime Ministerial home, but rather that the building is in certain salient ways like a railway junction. This is implicitly understood by speaker and hearer alike.

Traditionally, metaphors were considered as poetic devices that were used expressively to illustrate something in a text. Remnants of this attitude could be seen as late as in 1976, when Dagut clearly stated that creative metaphors were the only real metaphors, and other kinds of metaphors were more or less disregarded as cases of polysemy. This all changed with the publication of Lakoff and Johnson’s highly influential monograph Metaphors We Live By in 1980, and the common assertion is nowadays that metaphors in a broad sense are a very common, necessary and integral part of language. The main reason for this is that metaphors have the function of describing or understanding that which is abstract and/or complex by using concepts that are simpler and/or more concrete, by means of similarity or (as in example (1)) analogy. Working from a Translation Studies perspective, Newmark (1988: 104) states that metaphors have two functions or purposes: 1) referential and cognitive; they are used to describe and identify, 2) pragmatic and aesthetic; they are used to appeal to our senses. Ideally these two purposes should combine, according to him.

In Lakoff and Johnson’s view, polysemy is an aspect of metaphor and this means that the term metaphor has a very broad use. This view has the advantage of enhancing our understanding of linguistic and cognitive processes, but from a textual point of view it has the disadvantage of making the concept of metaphor somewhat too broad to be manageable, and a subdivision of the concept is thus necessary. Many such subdivisions have been suggested, e.g. by Dagut (1976), Van den Broeck (1981), Toury (1995), Lindqvist (2002; 2005) or Dickins (2005). There seems to be some consensus among these scholars that the metaphor concept works on a continuum as regards their dynamic force, i.e. how alive the metaphoric element is in the minds of the language users. This continuum ranges from dead metaphors, such as using spatial vehicles for temporal topics or dead anthropomorphing metaphors such as the foot of the mountain, to creative metaphors that are coined on the spot, so to speak. There are many ways to further subdivide the middle part of the continuum, and as in most matters linguistic, it is often a question of fuzzy borders between the categories. In this paper, we will be using Dickins’s (2005) adaptation of Newmark’s (1988) categories of metaphors. This adaptation has the advantage of being less subjective than Newmark’s original, while still being translation-oriented. Dickins’s main distinction is between lexicalized and non-lexicalized metaphors. The lexicalized metaphors are further subdivided into dead, stock and recent metaphors (the last category being comprised of metaphors that are fairly recently coined, yet where the vehicle does not tend to be foregrounded in the users’ minds). The non-lexicalized categories are schematically adapted metaphors, which are new metaphors, but which depend on established metaphors for their interpretation, and original metaphors, which are completely new.

The kind of metaphor that is central to this paper is what could be called revitalised metaphor. Such metaphors come into existence through a process where “dead or lexicalised metaphors may become ‘live’ symbols again” (Van den Broeck 1981: 76). In other words, lexicalised metaphors (in Dickins’s terms), which – through their inclusion in that category – do not normally have an active vehicle, may have their vehicles reactivated. This is done through foregrounding the vehicle, which is normally backgrounded to the point of not being active at all in most language users’ minds. Or as Van den Broeck (1981: 83) puts it: “[f]oregrounding […] re-awakens the symbolic force of the dead metaphor, so that in a sense it becomes a ‘live’ metaphor again;” in Dickins’s terms a live metaphor would be a non-lexicalised one. Revitalised metaphors cause ambiguity between the literal and the figurative sense of a metaphor, as the vehicle is played off against the topic (cf. Van den Broeck 1981: 82-83).  Let us concretise the concept of revitalised metaphors through the use of an example from Yes, Prime Minister. The PM, Jim Hacker, is in metaphorical hot water about the size of the grant to the National Theatre, which might lead to public embarrassment for him. In this scene, the PM, his political advisor Dorothy Wainwright, and his private secretary Bernard Wooley discuss the matter. The PM is quite distraught:

(2) Jim Hacker:

This is a real hot potato. If I don't do anything, it could turn into a banana skin.

Bernard Wooley:

Prime Minister, a hot potato can’t become a banana skin.

Jim Hacker:


Bernard Wooley:

Well, if you don’t do anything, a hot potato just becomes a cold potato.

(2:7 “The Patron of the Arts” 21.47 –22.02)

Jim Hacker used (and mixed) two stock metaphors in (2), clearly being totally unaware of their status as metaphors, their vehicles being inactive in his mind. Because of the inactive vehicles, the fact that he mixes his metaphors is (at least to him) not a problem, or as Dickins puts it:  “Where the metaphors in question are dead or stock metaphors, mixed metaphor is very often not particularly noticeable” (2005: 253). Bernard, on the other hand, does notice them and revitalises the metaphors by foregrounding the vehicles. In this case, this is a source of amusement, as a kind of pun. In real life, revitalising metaphors is often the forte of children, who may be unaware of the figurative use of a term, and take it literally. It is of course also the forte of people who like a good pun, like the script writers of Yes, Prime Minister. One way of revitalising metaphors is by visualising them, and this is the main issue of this paper.

3. Metaphors in translation

Metaphors are potential sources of translation crisis points (TCPs; cf. Pedersen 2007; 2008), i.e. points in translation where the translator has to abandon his or her automated processes and resort to strategic behaviour (cf. Lörscher 1991). They are thus potential translation problems, and as such, they have attracted a fair amount of attention in the last few decades. For instance, Dickins, Hervey and Higgins (2002: 150-155) suggest that in translation  between Arabic and English there is what they call ‘leftward shifting’ on the metaphor continuum, in that original ST metaphors may be replaced by TT stock metaphors, ST schematic metaphors may be replace by TT stock metaphors and so on. They also find that metaphors are frequently replaced by similes. A similar observation is also made by Lindqvist (2002; 2005) who found loss of figurative language in the translations of fiction that she investigated. Similarly, Dagut (1976: 30) claims that there is a loss of dynamic force when a ST metaphor is replaced by a ‘weaker’ form of language (either a paraphrase or a less original metaphor). There is thus evidence to suggest that Toury’s law of growing standardisation (1995: 267 ff) is at work when it comes to metaphors: TTs seem to be less ‘exuberant’ (to use Newmark’s 1988:112 term) in the use of figurative language than STs.

Returning to the issue of the strategic behaviour of translators when rendering metaphors, we will now look at what strategies are at their disposal for doing so. Lindqvist (2002: 155-156) combines the strategies suggested by Toury (1995) and Van Den Broeck (1981) into the following list:

  1. Översättning sensu stricto (translation sensu stricto; i.e. a more or less word-for-word translation);
  2. bildersättning (i.e. substitution, where one figure of speech in the ST is replaced by another in the TT);
  3. bildförlust (i.e. paraphrase, where the metaphor is replaced by a non-metaphorical expression);
  4. bildvinst (i.e. compensation, where a non-metaphorical expression is replaced by a metaphor);
  5. bildstrykning (complete omission)
  6. bildtillägg (compensation whereby a metaphor is added where there was nothing in the ST).

To this list, one might add strategies such as ‘specification,’ where a metaphor is rendered but also explained and ‘retention,’where a metaphor is transferred without translation.

Even though metaphors have been studied fairly extensively in recent years, they have been largely neglected in the field of audiovisual translation (AVT) and particularly in subtitling. This paper aims at rectifying this in a small way, so we will presently look at metaphors in subtitling, but first, it is necessary to establish some basic facts about subtitling as a mode of translation.

4. The conditions of subtitling

It could be said that there are three particular conditions that set subtitling apart from more traditional forms of translation. These are 1) time and space constraints; 2) shift in semiotic mode from speech to writing; and 3) the polysemiotic context.

The time and space constraints are well-known to anyone who has come into contact with subtitling. The space constraints consist of a limit of the number of characters that can be fitted into a line of subtitling, normally between 36 and 40 these days (cf. Pedersen 2011: 19), including punctuation marks and blank spaces, and a normal maximum of two lines. The time constraints are supposedly viewer-centred and tend to be (for television) centred on the ‘12 characters per second-rule,’ which means that for each second of exposure time, an average viewer can manage to read 12 characters while also processing the on-screen images. Actual reception studies of reading speed are few and far between (but cf. Tveit 2004: 58–64), and it is sometimes suggested that viewers can easily handle a higher reading speed, but these are the contemporary norms for television at least in traditional subtitling countries in Europe. This means that on average about a third of what is said on television gets lost in the subtitles (Pedersen 2011: 21). This is a pure quantitative loss, mind you, since the actual information loss is much lower, as the message is condensed by the subtitler. For the present study, the time and space constraints mean that there is not very often a great deal of room for explanations and long paraphrases of metaphors.

The shift from source language (SL) speech to target language (TL) writing, or what Gottlieb metaphorically calls “semiotic jaywalking” (2001: 16), means that the message gets modified from ST to TT. Many oral features, such as pauses, repetitions, false starts, etc. that are very common to spoken discourse, get ‘cleansed’ to  adapt to the code of written language. Gottlieb calls this “intrasemiotic redundancy” (2001: 56), and it is the source of most of the condensation that is necessitated by the time and space constraints. It seems rather unlikely that this condition should affect the present study, as it is unclear whether metaphors are more common or used differently in speech or in writing.

The third condition of subtitling is crucial to the present study, however. This is the polysemiotic (or multimodal) context of the subtitled text, and it is one that subtitling shares with all other forms of AVT. As isosemiotic (cf. Gottlieb 1997: 146) translation, such as literary translation only has to deal with one semiotic channel of discourse which remains the same in the ST and the TT, diasemiotic translation (Gottlieb 2004: 219 ff) has to deal with at least four semiotic channels. These four are:

  • Verbal audio: the dialogue and its paraverbal elements
  • Non-verbal audio: (background) music and sound effects
  • Verbal visual: displays and captions
  • Non-verbal visual: composition and montage

(Gottlieb 1997: 143, my translation)

Gottlieb discusses these four channels in connection with his concept of intersemiotic redundancy (2001: 20), which in a nutshell says that the other semiotic channels compensate for information loss stemming from condensation of the verbal audio channel. This is because the information conveyed via the four channels is often redundant, i.e. overlapping, and this can thus be used as a source of verbal condensation. This arguably represents a somewhat simplistic view of the interplay between these semiotic channels, however, and I have thus further developed this idea into what I call polysemiotics (Pedersen 2007: 163–164), which has a much broader scope. This is because the semiotic channels interact in more complex ways than merely copying the information present in the other channels and thus creating redundancy (even though they do that as well). The semiotic channels do not always work together to form semiotic cohesion (as described by Díaz Cintas and Remael 2007: 49), in which the channels complete each other and add more and new information to form a fuller narrative than each channel would provide on its own. The semiotic channels can also pull in different directions to cause semiotic tension. A famous example of this is a scene in Quentin Tarantino’s Reservoir Dogs (1992). In this very bloody scene, a gangster brutally tortures a fettered policeman and cuts off his ear, all to the merry music of “Stuck in the Middle with You” by Stealers Wheel in this display of what Zabalbeascoa undoubtedly would call “audiovisual irony” (2003: 305). Here, the visual channels and the audio channels are clearly pulling in completely opposite directions. There is no intersemiotic redundancy here. Instead, it is a clear case of semiotic tension, or even of semiotic dissonance. There are also many other ways in which the semiotic channels can interact, with two of them completing each other, one providing redundancy and one being inactive, for instance. Creators of audiovisual texts normally use semiotic cohesion as a means of telling their story in as efficient a way as possible, but they may also use semiotic tension to unsettle their viewers, or to play with them. The use of semiotic tension is right at the heart of this paper, as that can very much be the result of subtitling visualised metaphors, as we shall see.

5. Visualised metaphors in subtitling  

Metaphors are verbal images conveyed through language. As they are images, they can also be visualised in various ways and for various purposes. One of these purposes is for comic effect, and that is what we are looking at in this paper. The comic effect appears when a lexicalised metaphor that is used in its everyday figurative sense is revitalised and the viewer is reminded of the literal sense of the vehicle in the metaphor in what might be called an ‘intersemiotic pun.’

We saw an instance of this sort of pun (albeit without visualisation) in example (2) where Bernard Wooley interpreted ‘cold potato’ literally and thus chose the sense of the vehicle rather than the sense of the topic in that metaphor. What happens then is that the topic and the vehicle are activated simultaneously in the viewer’s mind, where normally only the topic would be activated. We then get the ambiguity between literal and figurative sense that Van Den Broeck talks about (1981: 82-83), and this ambiguity is thus the butt of the pun. In example (2) the pun was triggered by Bernard’s pedantry about mixed metaphors (‘hot potato’andbanana skin’). Mixed lexicalised metaphors do not normally cause any serious problems of understanding, however, because they are still coherent as long as they have shared or overlapping entailments (Lakoff and Johnson 1980: 89-96), which would be the case in example (2). Both the ‘banana skin’ and the ‘hot potato’ metaphors are lexicalised (dead or stock) metaphors, and both metaphors have the entailment of something unpleasant which is best left untouched. That is why it only becomes a pun when Bernard points out the literal sense of the first metaphor; we need someone who is pedantic about mixed metaphors to activate the vehicle.

Let us return to the visualised metaphors which may be used for intersemiotic punning. In example (3) the Prime Minister is annoyed with the National Theatre which puts on plays criticising the government, while also asking for higher grants from the government, and he, Bernard Wooley and Sir Humphrey are in the Cabinet Office discussing this:

Jim Hacker:

They insult me and then expect me to give them more money.

Sir Humphrey:

Yes, it is a rather undignified posture. But that is what the artists always do: crawling towards the government on their hands and knees, shaking their fists.

Jim Hacker:

Beating me over the head with their begging bowl.

Bernard Wooley:

I’m sorry to be pedantic, Prime Minister, but they can’t beat you over the head if they’re on their knees. Unless of course they have very long arms…

Jim Hacker:

Get off the floor, Bernard.

Bernard Wooley:

Yes, I’m sorry.


Swedish subtitles:

Back translation:

De kommer krypande till staten
med knutna nävar.

-Slår mitt huvud
med         tiggarbössan.
-Jag vill inte vara petig…

…men det går inte om de kryper.
Om inte armarna är jättelånga.

-Res på dig!
-Jag ber om ursäkt.

They come crawling to the state
with clenched fists.

-Hit my head with the begging bowl.
-I don’t want to be pedantic…

…but that’s not possible if they’re crawling.
Unless their arms are very long.

-Get off the floor!
-I apologise.

 (2:6 “Patron of the Arts”: 6.09 – 6.05)

The reason for the PM’s final remark is that Bernard has gone down on the floor and gesticulates to illustrate his point, as can be seen in image A.

Very long arms

Image A: Very long arms1.

In image A, we can see how Bernard has gone down on his knees to illustrate the first metaphor and reaches high into the air to illustrate the second. The two conceptual metaphors involved in this joke could be expressed as ‘FINANCIAL NEED IS BEGGING’ and ‘CRITICISING IS A PHYSICAL ATTACK.’ ‘Crawling’and ‘shaking fists are lexicalised manifestations of these, whereas the PM’s final metaphor is original, if adapted from these conceptual metaphors. In other words, these metaphors are either stock or adapted stock metaphors, which is why Bernard’s pedantry is once again funny. It could also be argued that his illustration of them makes it even funnier, as this sort of behaviour is not what you would expect from a man in his position (pun intended) in an institutional setting such as the Cabinet Office. By illustrating his point, he has visualised the metaphor.

From a polysemiotic viewpoint we now have the following situation: the topic of the three metaphors (‘crawling,’ ‘shaking fists,’ ‘beating with begging bowl’) are all active in the verbal audio channel. The vehicle is foregrounded in the verbal audio channel as well, through Bernard’s verbal objection, but it is also foregrounded in the non-verbal video channel. From a subtitling standpoint, it has the implication that the subtitler must take the non-verbal visual channel into consideration when choosing his or her translation solution. If they can be translated sensu stricto, the translation problem is easily solved, but if they cannot, using a paraphrase or generalisation would only bring about the gist of what Jim Hacker and Sir Humphrey are saying; they would not explain Bernard’s odd behaviour. Luckily, the metaphors involved are transcultural, i.e. known in both the source and target culture (cf. Pedersen 2011: 106–110), or adapted from such metaphors. They can thus be rendered sensu stricto, which the Swedish subtitler has done.

A more challenging situation arises when a visualised lexicalised metaphor is monocultural, i.e. not shared by the source and target languages. The subtitler then needs to create semiotic cohesion between the channels, without having recourse to the same image in the target language, and if the vehicle is foregrounded in one channel and the topic in another, that potentially creates unintended semiotic tension. The subtitler may even have to decide to go either with the topic or the vehicle. There is a potential conflict of norms here. The polysemiotics tell the subtitler one thing, which would normally be to “follow the picture,” in order not to create unwanted semiotic tension, which would leave the viewers confused, and in doubt of the ‘correctness’ of the subtitles. This would mean that the focus would be on the visualised vehicle. However, as I have shown elsewhere (Pedersen 2008), subtitling is a pragmatic form of translation), which implies that it is the pragmatic sense of an utterance that takes precedence, not the surface form, if a hierarchy of priorities has to be established (see Zabalbeascoa 1994;1996). It is what the speaker wants to put across, not how he or she puts it across that matters, if there is need for priority. This would give the focus to the topic of the utterance.

Elsewhere (Pedersen 2010: 17–18), I have argued that there is a contract of illusion between the subtitlers and the viewers. The viewers’ part of the contract is to suspends their linguistic disbelief. This means that they pretend that what the subtitles say is what the characters on screen are saying (in reality, of course, it is not; it is condensed, translated and in another medium etc.), and that means that the viewers understand what the characters are saying. For their part of the contract, the subtitlers make their subtitles as unobtrusive as possible, leading to the automatised consumption of subtitles that viewers in subtitling countries are accustomed to. However, if there is unintended semiotic tension between two semiotic channels, this contract is void. The viewers become aware that they are reading subtitles, and very likely suspect that there is something wrong with them.

We have such a potential situation in example (4), where Jim Hacker is being urged to do something ground-breaking to sort out Britain’s education problems, and Jim becomes very excited:

(4) Jim Hacker:

Do you think I could? Grasp the nettle? Take the bull by the horns?

Bernard Wooley:

Prime Minister, you can't grasp the nettle, if you take the bull by the horns.

Jim Hacker:

Oh really, Bernard?

Bernard Wooley:

All I meant was: if you grasp the nettle (bends down to grasp imaginary nettle) you could take the bull by one horn with the other hand but not both horns, because your hand isn't big enough (flexes hand to grasp imaginary horns). And if you did take the bull by one horn (holding invisible horn with one hand), it would be rather dangerous, because… mp mp (nods hand to illustrate charging bull). It was, it was just a mixed metaphor and since we were discussing education, I…

(2:7 “The National Education Service” 10.08–10.40)

Swedish Subtitles:

Back translation:

Kan jag spänna bågen
och ta tjuren vid hornen?
Ni kan inte ta tjuren vid hornen,
om ni spänt bågen.

-Säger du det?

-Spänner ni bågen med en hand…
…kan ni ta ett horn med andra,
men ni kan inte ta tag i bägge.
Tar ni tjuren vid ett horn,
kan det vara farligt för…
Det var en katakres, och eftersom
vi diskuterar utbildning, så…

Can I draw the bow
and take the bull by the horns?
You cannot take the bull by the horns, if you’ve drawn the bow.

-You don’t say?

-If you draw the bow with one hand…
…you can take one horn with the other, but you can’t take both.
If you take the bull by one horn, it can be dangerous because…
It was a catachresis, and since
we’re discussing education, I…

Dangerous bull

Image B: Dangerous bull

In image B, we see how Bernard is making much of his point through the non-verbal visual channel, by showing how the bull might attack you with its free horn, if you just hold on to the other. In the clip, he also moves his head, as to nudge you with his imaginary horn. This can be contrasted by his actions in the non-verbal visual channel in example (3) where the input of the non-verbal visual channel was actually intersemiotically redundant, as they conveyed the same information as the verbal audio channel, in that they were mere illustrations. In example (4), Bernard could have chosen to verbalise his mime by saying something along the lines of “…dangerous, because the bull might attack you with its remaining horn” or the like. Instead, this is just conveyed through the non-verbal visual channel, for comic effect. This means that the polysemiotics are complementary here, and not redundant.

The two rather hackneyed metaphors at stake here are thus complimentarily active in two channels, the verbal audio, which foregrounds the topic and the vehicle, and the non-verbal visual, which only foregrounds the vehicle, in this example. One of the metaphors (take the bull by the horn) is transcultural, i.e. common to both languages involved, while the other (grasp the nettle) is monocultural (cf. Pedersen 2011: 106–110), i.e. the image is not used in Swedish. Nevertheless, unless there is to be unintended semiotic tension, the subtitler will have to come up with an image that can be visualised in the same way. The Swedish subtitler has come up with the image ‘draw the bow,’ which is a Swedish metaphor roughly meaning ‘going out on a limb and making an effort’2. The problem in this context is, of course, that it also takes two hands to draw a bow, but the subtitler solved that by adding ‘with one hand,’ and it could actually look like Bernard is standing on a bow and drawing it with one hand while taking the bull by one horn. Thus, using creative substitution, and hoping for a little good will from the viewers, the Swedish subtitler has here avoided unintended semiotic tension in this example.

In the examples we have seen so far, which include the sum total of the visualised metaphors used in the two original series of Yes Prime Minister, the translation solutions chosen have been fairly felicitous. The scarcity of visualised metaphors may be due to the nature of the material, but it may also be the case that explicitly visualised metaphors are generally rare, because the material was actually specifically chosen because it did contain this feature (yet only two of more than 550 extracted metaphors were in fact visualised). The reason for the felicitous solutions may be that the subtitlers involved are highly competent, and that may be the case, but it is definitely due to the nature of the metaphors involved. They are either transcultural or adapted from transcultural metaphors, which means that they do not cause any serious TCPs. This in turn may be due to the cultural proximity of the two languages involved, and there is a increasing tendency for Swedish to adopt Anglophone figures of speech; for instance, it is now possible, if marked, to say ‘vad det kokar ner till’ (‘what it boils down to’) in Swedish, which was completely unheard of just ten years ago. It is highly likely that it would be more difficult to felicitously translate visualised metaphors in another language combination, e.g. English and Arabic (cf. Dickins 2005). However, the positive, if meager results of this investigation do in no way negate the possibility that that TCPs may be caused by visualised metaphors between English and Swedish. This can be illustrated with an example from The Simpsons Movie (2007). Here, Homer Simpson, for various comic reasons, is stuck on a wrecking ball, which bounces between several objects that cause him grievous bodily harm (again for comic reasons). The wrecking ball scene ends with him swinging between the two objects seen in image C.

Rock and a hard place
Image C: A rock and a hard place (The Simpsons Movie 1:01:45–1:01:46)

As we can see in image C, Homer is thus literally stuck between a rock and (a saloon called) a hard place, repeatedly being slammed into both of them. The intersemiotic pun here is completely based on the visual channels, the verbal visual (the name of the saloon) and the non-verbal visual (the rock and Homer being stuck between them). In the source text, there is no semiotic tension here: all active channels reinforce the intersemiotic pun of the revitalised literal meaning of the metaphor (and saying) ‘stuck between a rock and a hard place.’ Furthermore, there is no tension between the topic and the vehicle here either: the vehicle is visualised, in that it is literally true that he is stuck (or at least dangling) between a rock and a hard place, but so is the topic, i.e. the metaphoric sense of him being in trouble stuck between two unpleasant alternatives. In fact, it could be argued that there is no metaphor here at all, as the vehicle and the topic are the same (see the argument made by Lakoff and Johnson quoted above).

Even though the metaphor (or possibly non-metaphor) here is less complex than the previous examples, the translation is not. The metaphor ‘stuck between a rock and a hard place’ is in this case monocultural, in that the image does not exist in Swedish. There are similar ones, but none that would fit the images without causing intersemiotic tension. A translation is required here, even though there is no dialogue, as the verbal visual (the salient name of the saloon) requires translation. If we go through the strategies available to the translator as listed above, we find that translation sensu strictu would not be felicitous at all; and it would be polysemiotically impossible to use substitution or omission as that would require altering the source text, which is not normally possible, unlike when e.g. localising a website (e.g. Charalampidou forthcoming) or in some cases of game localisation (O’Hagan and Mangiron 2013). A paraphrase is equally impossible, as the literal sense is literally all over the screen. What actually did happen in the subtitled version released on DVD in Sweden is that the subtitler quite simply gave up: no subtitle accompanied the image. Since there is in fact a written text in the source text, and this remains legible in the target text, this means that the solution would have to be categorised as a form of retention, which is infelicitous here, as it leaves the viewers without guidance. It is thus unlikely that retention has been chosen for the benefit of the viewers, but rather as an admittance of failure. I have arranged seminars where the subtitlers at Sveriges Television, the public service broadcasting company in Sweden, and also the staff and teachers at the Institute of Interpretation and Translation Studies at Stockholm University in May 2013 discussed this particular issue.  Possible solutions were discussed at length, but no one could come up with a felicitous solution for it. Is it thus possible that we have here encountered a truly unsolvable TCP? I have previously stated my doubts about there being such a thing (Pedersen 2005). The solution to this problem may be found outside the norms of subtitling, however, if one were to apply one of the solutions commonly found in fansubbing, and added an explanation of the joke. There is a great deal of time in which to do so; the scene goes on for approximately ten seconds, and explanatory notes are occasionally, if extremely rarely, found in Swedish subtitles, which means that it would not be totally unheard of. Granted, an explained joke is not very funny, but neither is an incomprehensible one. Also, choosing this option in mainstream subtitling would break the contract of illusion, as the viewers’ attention is drawn to the fact that they are watching subtitles. At the end of the day, this example clearly shows that visualised metaphors (and non-metaphors) are possible sources of very dire TCPs.

6. Conclusion

This article has endeavoured to show the complexities that arise from the subtitling of visualised metaphors. These translation problems may not be very common, but when they do occur, they may produce near-unsolvable TCPs. The article has introduced the notion of intersemiotic tension and discussed that this – if unintended – may be something to avoid if one is keen on producing felicitous subtitles that do not break the contract of illusion between the subtitler and the viewer. The key to producing felicitous subtitling solutions to TCPs stemming from visualised metaphors is to take the whole polysemiotic message into account when producing subtitles, including both the verbal and the non-verbal channels. If this is done, there is a great chance that the contract of illusion can be saved and unintended intersemiotic tension can be avoided. By grasping the metaphorical nettle and not shying away from taking the polysemiotic bull by the horns, the subtitler may thus avoid the banana skin of the infelicitous translation solution and emerge from in between the rock that is the topic and the hard place that is the vehicle of that hot potato, the visualised metaphor.

  • Charalampidou, Parthena (forthcoming). “Verbo-pictorial metaphor and website localisation: A Cultural approach to the transfer of persuasive multisemiotic texts to the Greek locale.” To appear in Selected Papers from the 2012 Stockholm Metaphor Festival.
  • Dagut, Menachem (1976). “Can ‘Metaphor’ Be Translated.” Babel. 22(1), 21-23.
  • Díaz Cintas, Jorge and Aline Remael (2007). Audiovisual Translation: Subtitling. Manchester: St. Jerome.
  • Dickins, James (2005). “Two models for metaphor translation.” Target 17(2). 227-273.
  • Dickins, James, Sándor G.J. Higgins and Ian Higgins (2002). Thinking Arabic Translation. Berlin and New York: Mouton de Gruyter.
  • Goatly, Andrew (2007). The Language of Metaphors. London: Routledge.
  • Gottlieb, Henrik (1997). Subtitles, Translation & Idioms. Copenhagen: Center for Translation Studies, University of Copenhagen.
  • (2001). Screen Translation: Six studies in subtitling, dubbing and voice-over. Copenhagen: Center for Translation Studies, University of Copenhagen.
  • (2004). “Subtitles and International Anglification.” Special issue of Worlds of Words 3(1), 219-230.
  • Lakoff, George and Mark Johnson (1980). Metaphors We Live By. Chicago and London: The University of Chicago Press.
  • Lindqvist, Yvonne (2002). Översättning som social praktik: Tony Morrison och Harlequinserien på svenska. Acta Universitatis Stockholmiensis. Stockholm: Almqvist & Wiksell International.
  • (2005). Högt och lågt i skönlitterär översättning till svenska. Uppsala: Hallgren och Fallgren.
  • Lörscher, Wolfgang (1991). Translation Performance, Translation Process, and Translation Strategies: A Psycholinguistic Investigation. Tübingen: Gunter Narr.
  • Mohanty, Niranjan (2010). “Intranslatability and the translator’s task.” Perspectives: Studies in Translatology 4(2), 163-171.
  • Monti, Enrico (2006). “Dwelling upon Metaphors: The Translation of William Gass’s Novellas.” Nordic Journal of English Studies. 5(1), 117-132.
  • Newmark, Peter (1988). A Textbook of Translation. Harlow: Longman.
  • O’Hagan, Minako and Mangiron, Carmen (2013). Game Localization.Translating for the global digital entertainment industry. Amsterdam: Benjamins.
  • Pedersen, Jan (2005). “How is culture rendered in subtitles?” In Nauert, S (ed.) Challenges of Multidimensional Translation. Proceedings of the Marie Curie Euroconferences MuTra: Challenges of Multidimensional Translation - Saarbrücken 2-6 May 2005, (consulted 17.10.2014)
  • Pedersen, Jan (2007). Scandinavian Subtitles: A Comparative Study of Subtitling Norms in Sweden and Denmark with a Focus on Extralinguistic Cultural References. Doctoral Thesis. Stockholm University: Department of English.
  • (2008). "High Felicity: a speech act approach to quality assessment in subtitling." Chiaro, Delia, Christina Heiss, and Chiara Bucaria (eds) (2008). Between Text and Image. Updating research in screen translation. Amsterdam/Philadelphia: Benjamins, 101-116.
  • (2010). “Audiovisual Translation – In General and in Scandinavia.” Perspectives: Studies in Translatology, 1-22.
  • (2011). Subtitling norms for television: an exploration focusing on extralinguistic cultural references.Amsterdam/Philadelphia: John Benjamins.
  • Samaniego Fernández, Eva, Marisol Velasco Sacristán, and Pedro A. Fuertes Olivera (2003). “Translations we live by: Metaphor Translation and Cognitive Innovation.” Paper presented at 8th International Cognitive Linguistics Conference (University of La Rioja, 20-25 July 2003).
  • Toury, Gideon (1995). Descriptive Translation Studies – And Beyond. Amsterdam and Philadelphia: John Benjamins.
  • Tveit, Jan Emil (2004). Translating for Television: A Handbook in Screen Translation. Bergen: Kolofon.
  • Van den Broeck, Raymond (1981). “The limits of translatability exemplified by metaphor translation.” Poetics today 2(4), 73-84. (Consulted 3.11-2014)
  • Zabalbeascoa, Patrick (2003). “Translating Audiovisual Screen Irony.” Pérez González, Luis (ed.) (2003). Speaking in Tongues: Languages across Contexts and Users. Valencia: Edicions Universitat de València, 305-319.
Audiovisual sources
  • Reservoir Dogs. 1992. Quentin Tarantino. USA: Miramax.
  • The Simpsons Movie. 2007. David Silverman. USA: Twentieth Century Fox.
  • Yes Prime Minister. 1986-1988. Anthony Jay and Jonathan Lynn. UK: BBC.
    • Series 1; episode 4: “The Key”
    • Series 2; episode 6: “The Patron of the Arts”
    • Series 2; episode 7: “The National Education Service.”


Jan Pedersen pictureDr Jan Pedersen is Director of Studies at the Institute for Interpreting and Translation Studies at the Department of Swedish and Multilingualism at Stockholm University, Sweden, where he also teaches audiovisual translation. He has worked as a subtitler for many years and is the president of ESIST, the European Association for Studies in Screen Translation, and Associate Editor of Perspectives: Studies in Translatology.



Note 1:
The proprietors of the original audiovisual material in the present study (the BBC and Fox) were contacted to get permission to include frames from the material in this article for scientific analysis. They either did not respond or demanded a fee that was far beyond the means of the present study. So, for copyright reasons, the actual frames from the original audiovisual material have been replaced by an artist’s rendition. I am indebted to Lisa Rondahl Pedersen for the sketches.
Return to this point in the text

Note 2:
Originally coined by Swedish poet Verner von Heidenstam.<
Return to this point in the text