Discourse Analysis as an approach to subtitling an audiovisual text

As the new media are advancing at an incredible pace all over the world, there is an increasing need for specialists to carry out the transfer of its content into different languages. This thriving activity known as media translation or/and audiovisual translation has attracted the attention of many scholars who dealt with the linguistic, cultural and technical aspects of this practice, and mainly tried to suggest strategies and approaches to guide the translator, as this transfer is sometimes carried out by amateurs and nonprofessional translators. Therefore, the aim of this paper is to investigate to what extent discourse analysis can serve as an approach to subtitling an audiovisual text and demonstrate how it may limit the loss resulting from the strategy of omission implied by subtitling that generates a poorly worded, standard and unstructured text in the target language.


Introduction
It goes without saying that the re-expression phase involved by the process of translation has to be preceded by the analysis of the text being translated; as far as the audiovisual text is concerned, the translator has to analyze the linguistic, cultural and technical specificities of this multimodal text, define the role of each of its components and mainly spotlight the elements that ensure its cohesion and coherence.However, this operation is usually overlooked due to the technical, linguistic and textual constraints implied by the practice of subtitling.Therefore, our research is intended to highlight the role of discourse analysis in the translation process of an audiovisual text and show how it can limit loss.For this, we structured this paper in four sections.In the first section, we will get a glimpse at audiovisual translation, focusing particularly on the specificities of subtitling.In the second section, we will present some principles underlying discourse analysis and that may serve audiovisual translation.Last, we will demonstrate the role of the concepts of coherence and cohesion in relation to subtitling.

Audiovisual translation
As a new discipline in Translation Studies, audiovisual translation is mainly concerned with the transfer of the media content including television, cinema, theatre, the internet, the opera and all multimedia products.These products rely on a set of codes to convey information as image and sound, therefore, the meaning in an audiovisual text is conveyed by speech, image and sound, Chaume (in ORERO, 2004: 18), says in this regard that "[the audiovisual text] is a semiotic construct comprising several signifying codes that operate simultaneously in the production of meaning".Accordingly, the translator has to take into account the interaction between the verbal, paraverbal and nonverbal elements present in the original product.Linguistic variation, intonation, kinesics all are of paramount importance and have therefore to be rendered in the target text.
Two main methods are adopted in the field of audiovisual translation: dubbing and subtitling.Theorists in audiovisual translation distinguish between different kinds of subtitling according to different parameters (linguistic, technical, etc.).At the linguistic level, DIAZ-CINTAS and RAMAEL (2007) distinguish between three types of subtitles: intralingual, interlingual and bilingual.The first type is mainly designed for the deaf and hard-of-hearing, and is also used for teaching purposes (teaching foreign languages), the translation in this type in carried out within the same language so as to enable SDF to have access to the product, they usually integrate colours and explicit all the paralinguistic elements on the screen.In the second type, the message is conveyed from one language to another, implying different changes.
By and large, DIAZ CINTAS & REMAEL (2007: 14) define subtitling as: "a translation practice that consists of presenting a written text, generally on the lower part of the screen, that endeavours to recount the original dialogue of the speakers, as well as the discursive elements that appear in the image (letters, inserts, graffiti, inscriptions, placards, and the like), and the information that is contained on the soundtrack (song, voices off)." Accordingly, subtitling doesn't only involve the transfer of the original "text" or dialogue into another language, all the information carried in the image and the soundtrack, in addition to voice qualities have to be conveyed.However, given the technical, textual, linguistic and cultural constraints imposed on the subtitler, rendering all the linguistic and paralinguistic specificities of the source text remains an unattainable ideal.This is why subtitling is referred to as a reduced version of the original discourse.
Moreover, the shift from spoken discourse to written discourse involved by subtitling raises difficulties and generates some semantic and stylistic loss.Indeed, all features of orality, discourse markers, slang words, polite formulas, swearwords, taboo language, etc. tend to be omitted for technical or cultural considerations, which results in a standard text, GORIS (in HATIM & Mason, 1997: 21) says in this regard: "(…) standardization in subtitling (…) imposes the elimination of dialectal language (…) The social differentiation of the language is not maintained.The vulgar terms and expression are eliminated, and even the popular elements are "corrected".
This tendency to omit some linguistic features of the source discourse and its emotive function is likely to distort the intended meaning and hamper comprehension.Therefore, the subtitler has first to analyze the structure and content of source discourse and define the function of each element before carrying out the re-expression operation; this is why we have thought it convenient to rely on Discourse Analysis as an approach to subtitling.

Discourse analysis
Given the multidisciplinary nature of Discourse Analysis, providing an exhaustive and precise definition of this field and its object of study is no easy matter, as each discipline deals with a given aspect of discourse.Indeed, according to Gillian BROWN & George YULE (1983), psycholinguistics, for example, deals with language comprehension, as for philosophical linguistics, it mainly addresses semantic relationships between the segments of discourse.
However, it is worth noting that discourse analysis is not concerned with the study and description of random sentences or the structures of language, it rather involves the analysis of these structures in context and the relationship between the discourse and the situation and the speaker (Dominique MAINGUENEAU, 2004), because language is not merely a set of sentences, it is rather an act of communication.The discourse analyst has therefore to take into account the co-text, the situational context and the genre specificities, as each discourse requires a given analysis principles.

The role of context
Broadly speaking, there are two types of context: the co-text and the situational context.The situational context refers to the circumstances in which the utterance occurs such as the interlocutors, the space and time parameters.
Many theorists have set a list of the features and elements of situational context, among which we can mention, FIRTH (in BROWN & YULE, 1983) who distinguishes between the features of the participants, the features of objects and the effect of the verbal action.According to him (idem), features of the participants include their verbal and nonverbal action as the phonological, lexical and syntactic material, and the prosodic material gathering intonation, pauses, timber, tone and all voice features in addition to kinesics and all the elements that accompany, emphasize or contradict the verbal message.
HYMES (in BROWN & YULE, 1983), on the other hand, has highlighted some features of the situational context as follows: -The addressor: Having a prior knowledge regarding the producer of the message is likely to help the receiver understand the intended meaning.It is worth mentioning that in films, there are two addressors or more, namely the producer, screenwriter, or sound engineer and the characters that represent the voice of the producer.
-The addressee or the audience: in subtitling, the translator has to take into account the reading speed of the audience, and provide short subtitles, as they give more time for the viewer to process the information and help him keep an eye on the image.
-The topic: according to BROWN & YULE (1983), the topic doesn't lie in sentences; it is rather in the speaker.In the case of audiovisual texts, dialogue isn't enough to determine the topic, it has to be accompanied by image and sound that introduce the space and time in which the events occur.
-The setting: it refers to the space and time parameters.
-The channel: it refers to the means by which the participants establish and maintain contact.
-The code: If the code used is language, the discourse analyst will have to study the lexical and syntactic features, and all the sociostylistic variations.However, some of these characteristics tend to be lost in subtitling, especially when translating into Arabic.Indeed, when it comes to linguistic variation, it is hard for the subtitler to chose among the wide range of dialects available, this is why using Modern Standard Arabic is almost the best option.
-The message-from: it refers to the genre: a poem, a fairy tale, a film, a play, etc.
-The event: The actions described in the discourse.Therefore, an accurate and coherent rendering of the message of the original discourse requires a prior analysis of all the said specificities.

-Cohesion and coherence
As we have said previously, discourse analysis deals with language in use and discourse as a logical flow of ideas and a set of structured segments linked by connectors.This semantic and grammatical connectivity is respectively referred to as coherence and cohesion.
According to HALLIDAY and HASAN (1976: 4): "The concept of cohesion is a semantic one; it refers to relations of meaning that exist within the text, and that define it as a text.Cohesion occurs where the interpretation of some element in the discourse is dependent on that of another.The one presupposes the other, in the sense that it cannot be effectively decoded except by recourse to it.When this happens, a relation of cohesion is set up, and the two elements, the presupposing and the presupposed, are thereby at least potentially integrated into a text." By and large, cohesion refers to the lexical and grammatical devices connecting the segments of a text.It is set by connectors or markers of conjunctive relations expressing addition, opposition, contrast, etc.It can also be based on co-reference.Co-reference indicates the textual or extratextual items used to refer to an object.
According to the same authors (idem), the cohesive resources include: -Reference: Broadly speaking, reference describes the fact of using language to refer to the world, yet BROWN and YULE (1983) believe that reference doesn't merely describe this relationship between language and the world, it also involves the use of words by the speaker to perform actions.In written language or traditional texts, it indicates the resources used to refer to a participant or object.These resources include demonstratives, the definite article, pronouns, comparatives, and phonic adverbs as here, there, now, then -Ellipsis: Ellipsis happens when "something which is present in the selection of underlying (systematic) options is omitted in the structure-whether or not the resulting structure is in itself incomplete" (HALLIDAY & HASAN, 1976: 144).It refers to the omission of a clause or small items, implied by context, as the use of short answers in English and leaving out to-infinitives e.g: "did you post the letter": " I forgot to" -Substitution: it refers to the resources used to replace an item or clause already said in order to avoid repetition as the use of so e.g : "do you think she will pass her driving license?" "I think so." -Conjunction: it refers to the items combining two clauses and explicating their semantic relationship.There are four types of conjunctions: additive (and, furthermore, besides, etc), adversative (yet, though, but, however, etc), causal (because, as, since, etc) and temporal (first, next, finally, etc) -Lexical cohesion: lexical cohesion occurs when the semantic relationships within a text are ensured by words.There are two types of lexical cohesion: reiteration and collocation.Reiteration occurs when a word is replaced by its synonym, near-synonym, or hyponym.
It appears from what has been said that cohesive devices play a major role in producing a unified and coherent text and creating some texture, HALLIDAY and HASAN (1983: 2) say in this regard: "A text has texture and this is what distinguishes it from something that is not a text".Therefore, omitting these markers is likely to generate an unstructured set of sentences.
However, for technical considerations (the subtitles appear in two lines, each one should include 32 to 41 characters, and they have to stay on screen for 5 or seconds), a lot of devices tend to be omitted in subtitling such as relative pronouns, interjections, cohesive devices, phatic markers, and discourse markers that are some of the salient features of spoken discourse.
Discourse markers are words or expressions linking the units of discourse and are most apparent in spoken language as y'know, I mean, well, etc, and have therefore to be rendered in subtitles, as each marker carries different meanings according to the context, well for example can express a wide range of situations and have different functions, this is why it doesn't always have one equivalent in other languages, i.e in French it isn't always translated by bien and in Arabic it isn't always rendered by ‫,حسنا‬ the same remark goes for y'know that can be used to "conclude an argument, introduce a story preface, evoke a new referent" (SCHIFFRIN, TANNEN & HAMILTON, 2001: 66) according to the situation.
Also, adverbs and adjectives tend to be omitted in subtitles, even if they sometimes perform a function.In "I don't think you will succeed" and "frankly, I don't think you will succeed" for example, there is a slight difference.
The omission of the said devices does alter text's cohesion and coherence, and it mainly hampers audience's comprehension, this why subtitled and dubbed movies don't enjoy the same success reached by the original product.
Yet, some justify this omission strategy in subtitling by the fact that the audiovisual text is polysemiotic and therefore the meaning isn't carried only via text or subtitles, it can be rendered by image or/and sound, and sometimes the text, image and sound do carry the same meaning, which is known as redundancy, so this reduction is justified by redundancy.Paralinguistic and nonlinguistic cues do complement the linguistic message.CRUSE (in DIAZ-CINTAS & REMAEL, 2007: 161).
"Speech naturally involves linguistic, paralinguistic and nonlinguistic signs.Paralinguistic signs cannot be interpreted except in relation to the language they are accompanying.On the other hand, non-linguistic signs are interpretable and can be produced without the co-existence of language.Non-linguistic signs or natural signs such as facial expressions, postural and proxemic signs, gestures, and even some linguistic features 'are likely to be the most cross-culturally interpretable".
Therefore, it is this redundancy in films that makes subtitlers delete some cohesive devices.CHAUME (in ORERO, 2004) speaks in this regard about semiotic cohesion to refer to the interaction between the verbal and nonverbal cues.The missing information in the dialogue is filled by the image, moreover, there is interaction between the dialogue and kinesics, i.e gestures, as everything in films fulfills a function and carries a meaning.

Conclusion:
The adoption of discourse analysis in the process of subtitling an audiovisual text has proved useful, as it provides the subtitler with key notions to be applied during the transfer of the media content; discourse analysis is then directly linked to the reduction strategies implied by subtitling.Indeed, subtitling is different from other translation practices, since a large array of parameters come into play, as the change of mode from spoken to written language in addition to technical constraints that limit the translator's freedom and sometimes creates unstructured and incoherent random sentences in the target language, which is likely to alter the intended meaning and hinder information processing.
Therefore, analyzing the structure, function and context of the audiovisual product, defining the role of each of its components, and determining the elements ensuring its cohesion and coherence are instrumental.
However, considering the polysemiotic nature of the audiovisual text, one approach isn't likely to provide a thorough analysis.Indeed, an integrated approach borrowing concepts from disciplines as semiotics, film studies, cultural studies and translation theories like the Skopos theory and the polysystem, is more advocated.