In intersemiotic translation, like in any kind of translation in general, instead of pretend that it is possible to translate/communicate everything, against the evidence, it is advisable to take the loss into account from the start and, consequently, to work out a translation strategy that rationally enables us to decide what are the most distinctive components of the text and, conversely, those that can be sacrificed in favor of the translatability of another aspect of the text. Clüver says that a translated text is inevitably not equivalent to the prototext and, at the same time, it contains something more or something less with respect to the prototext.

In substance, what Clüver says can be linked to what we have already said about the dominant (see, in particular, unit 19). Before approaching a text, the translator must make a series of decisions aimed at pinpointing the dominant of the text, not only in an intrinsic sense but also in function of the cultural context in which the original is located, within the culture where it was originated, and the cultural context onto which the original is projected, within the receiving culture.

  When we translate one text into another, the series of decisions made may not be completely apparent to the reader of the translation if, for example, there is not a translator's afterword or another sort of metatext explaining the reason of her choices. Sometimes even the translator is unaware of her own choices, because she made them irrationally: she approached the translation, if you can accept the expression, without consciousness. In this case, the translation strategy is simply random.

  One of the reasons for such irrationality is that, at any rate, both at the beginning and at the end of the translation process we have a text. If we do not analyze thoroughly its differences, what was lost in the passage from the original may escape our notice. Denotative, connotative aspects, images, sounds, rhythms, syntactic structures, lexical coherence, intratextual, intertextual references, and so on: some of these components may not be found in the translated text, but this is not necessarily apparent.

  Conversely, when one of the two texts in an intertextual translation is not verbal, the choice between the parts to be translated and those that must be sacrificed is far more apparent. Indeed, the intersemiotic translator, willing or not, is forced to divide the original text into parts. It does not matter how: denotation/connotation, expression/content, dialogues/descriptions, intertextual/intratextual references etc. Then she must disassemble the prototext into these parts, find a translating element for each of them, and reassemble them recreating the coherence and cohesion, which - as we have just observed - is the essence of a text.

  Let us take filmic translation for example. Torop expressed an interesting opinion about it:

  The difference that Torop highlighted concerns the written word and the pronounced word. In films, the former is used rarely whereas dialogue is given much space. A filmic composition can be divided into different elements: dialogue between characters, the physical setting, possible voice-overs, the musical score, the editing, the framing, lighting, coloration, close-up or not, perspective, the composition of the frame and, in the case of human voice, also the timbre and the intonation. In order to carry on the filmic translation of a verbal text, a rational subdivision of the original is inevitable for deciding to what elements of the filmic composition to entrust the translation of given stylistic or narratological elements of the prototext.

  Here are some examples. Let us suppose we have to deal with the filmic translation of a novel. The author of the screenplay may decide to take the dialogues of the textual original and to use them exactly as they are in the film. This is what happened in most of Nick Dear's film version of Pride and Prejudice made for the BBC ³.

  There are other aspects of the prototext that can, however, be rendered in various ways. Let us go on about the Austen's film version; in the text, when Elizabeth Bennet receives a letter, the narrator obviously recounts it, whereas in the film we can see the actress, Jennifer Ehle, open the envelope, and read the letter. Part of the text is read in voice-over (by the actress herself), while other parts serve as a sound background for other scenes (basically, we have a flashforward of the images with respect to the sound). Other parts are seen from the sender's point of view as he writes them, as it happens with a letter from Darcy, with a quick flashback.

  Other parts of the film translation are necessarily interpreted in a freer way. Let us take, for example, the music that accompanies many dances. In the absence of a specific suggestion by the author, the screenwriter was obliged to choose arbitrarily - from the repertoire of English dance music between the 18th and the 19th centuries - those used for the soundtrack. Jane Austen's style, her way of narrating, probably represents another inevitable translation loss.

  We have said that intersemiotic translation implies a sort of subdivision of the original into various elements and the identification of components able to translate said elements within the coherence of the translated text.

  The same is equally valid for textual and intertextual translation.

  At a more general level, we can say that to translate means to rationalize. If, in the original, some ambiguous or polysemic elements are found, first of all the translator must read them, identify them, interpret them and then she must translate what is translatable, in a rational way. This may lead us to think that, in a translated text, it is easier to distinguish the various elements of the work, that there are fewer ambiguous passages, or that the polysemy of words is diluted.

  When the translated text, however, is built in an extra-literary code, like with translation into film, we observe a paradox: in a film it is more difficult to tell the expression plane from the content plane; with respect to a written text, it is more difficult to analyze the series of images, sounds and words in motion. For this reason, in a film (as well as in poetry), an artistic device based on a more irrational principle may appear more marked than in written prose.

  Is, then, the filmic version of a novel simultaneously more rational and less rational than the original? Notwithstanding the intrinsic prevalence of the irrational component in films or in painting or in music - namely, in the product of the translation process - it is evident that the process is, at any rate, more rational. To say the least, in the translator's mind it is clear what element translates a given component, and what components are left untranslated (lost).

  Let us take the case of Prokof´ev's Peter and the Wolf. In some versions for children, there is even a sort of "preface" explaining what musical instruments correspond to which characters of the fable so that, once the interpreter has made his choice, the latter is evident to the user too. In this case, the translation criteria are apparent, as well as what was transposed into music, what could be found in the fable and what was remaining as a translation loss.

  One last aspect of intersemiotic translation that we would like to take into consideration is that of translatability. Since the original text and the translated text, or metatext, are not easily comparable in terms of specific criteria, the concepts of "translatability" and "accuracy" can only be considered in conventional terms. Textual translation follows the principle according to which an original can possibly have many different translations, all of them potentially accurate; such potentiality is even more developed in intersemiotic translation, to such an extent that any attempt to retranslate a text into its original language - hoping to recreate, as a result, the original text - is inconceivable. As Torop says, it is not possible to recognize a text that has been reverse-translated, because it will result in a new text ⁴.

Bibliographical references

AUSTEN J. Pride and Prejudice. Screenplay by Nick Dear, London, BBC, 1995.

CLÜVER, C. On intersemiotic transposition. Poetics Today, vol. 10, n. 1, Tel Aviv, The Porter Institute for Poetics and Semiotics, 1989.

LOTMAN JU. Izbrannye stat´i v trëh tomah. vol. 1. Stat´i po semiotike i tipologii kul´tury. Tallinn, Aleksandra, 1992. ISBN 5-450-01551-8.

TOROP P. La traduzione totale. Ed. by B. Osimo. Modena, Guaraldi Logos, 2000. ISBN 88-8049-195-4. Or. ed. Total´nyj perevod. Tartu, Tartu Ülikooli Kirjastus [Tartu University Press], 1995. ISBN 9985-56-122-8.