Translating documentaries: from Neanderthals to the Supernanny

Abstract This article analyses the main features of documentary translation using a corpus of 20 documentaries translated by the author from English into Catalan over a two-year period (2006–07). After a short review of the literature on documentary translation, the main characteristics of this type of translation in terms of topics, speakers and transfer modes are described. Special emphasis is put on language register transfer in non-fictional programmes, taking into account mode, field and tenor. Finally, further challenges such as the translation of proper nouns, comprehension problems or mistranscriptions are presented.


Introduction
Documentaries, fiction films, TV series and cartoons: these are the four categories generally used by product managers in Spanish dubbing studios when offering a job to freelance audiovisual translators. Despite this simplification, the variety of products to be included under the label 'documentary' is increasingly wider and boundaries with fiction are getting fuzzier. Fiction films and factual films are generally produced differently, but still no clear-cut limits can be drawn: whilst authentic materials find their way into fiction films (Forrest Gump, JFK), false data are also found in documentaries (mockumentaries). Yet both translators and the audience expect documentaries to contain trustworthy information; as explained by Bordwell and Thompson (2008, p. 338), '[e]very documentary aims to present factual information about the world, but the ways in which this can be done are just as varied as for fiction films'.
Documentaries have attracted the attention of many researchers in Cinema Studies (Barnouw, 1983;Beattie, 2004;Corner, 1996;Nichols, 1991Nichols, , 2001Plantiga, 1997;Renov, 1993), who have ventured into a wide field of research, encompassing an enormous variety of products: compilation films (with images from archival sources), interview or talking-heads documentaries (with testimonies), direct-cinema documentaries (recording ongoing events), nature documentaries (recording wildlife), portrait documentaries (based on the life of a person) and synthetic documentaries (mixing several of the previous options). In terms of form, documentaries can follow a narrative, present a categorical form in which information is delivered in a simple fashion or show a rhetorical structure presenting a persuasive argument (Bordwell & Thompson, 2008).
This article aims to describe the main features of documentary translation using a corpus of 20 documentaries translated by the author from English into Catalan over a two-year period (2006Á07). After a short review of the literature, the main characteristics of this type of translation in terms of topics, speakers and transfer modes are presented, following a qualitative approach. Special emphasis is put on language register transfer taking into account mode, field and tenor. Finally, further challenges are briefly described and some conclusions are reached.
The translation of documentaries in AVT Some authors have dealt with documentary translation from the field of AVT, although the literature is not extremely extensive. Franco (2000a) aims to demonstrate that documentary translation is a specific practice because some aspects are typical of the factual genre (material, translation models and transfer modes) and also because of cultural aspects. Her article reports on the initial results of a PhD thesis (Franco 2000b) which analysed 22 documentaries about Brazil produced in western Europe and voiced-over from Brazilian Portuguese into French and German. The comparison between the French and German versions of the same documentary demonstrated that cultural issues play an important role in defining the content and form of voiced-over versions. In Franco (2001a) the treatment of culture-specific items in translated documentaries is further examined, and she concludes that a certain level of exoticism is welcome in this type of product. The cultural component in the discourse of science is also analysed by Hoorickx-Raucq (2005), who uses both scientific publications and TV documentaries translated from English into French to suggest that cultural markers are not simply related to terminology or structural organisation but also to cognitive processes.
Various case studies have also been published regarding documentary translation: Remael (1995Á96) examines the translation of a BBC documentary for the Flemish channel BRTN. In a later article, Remael (2007) considers how three types of AVT (introductory comments, narration and subtitling) are combined in three English-language documentaries broadcast in 2002 and 2003 on VRT-Canvas, the Flemish public television. Remael takes into account norms related to the AVT genre proper and to the channel's linguistic policy, and norms determined by personal preference/socialisation. On the other hand, Kaufmann (2004) deals with the linguistic uniformity found in the French translation of a documentary about poor immigrants in Israel, due to the language policy of the television network ARTE. According to the researcher, these norms delete the language variety of the immigrants and alter the original meaning.
From the point of view of practitioners, Mir (1999) and Mateu (2005), both professional translators working for the Catalan television company TVC, highlight terminology, innovation, flexibility, proper nouns, language and style as some of the more challenging issues. Matamala (2009) describes further difficulties and provides an analysis of the different techniques used in all the documentaries broadcast over a week in 2003 on TVC. Aaltonen (2002) also points out some translation problems using interviews with a translator plus think-aloud protocols, and concludes that some basic concepts must be considered by translators such as the audience, time and space limitations, stylistic and lexical choices, and faithfulness to the original.
A more general overview of documentary translation is found in Espasa (2004), who considers two myths related to the genre: (1) a documentary is not a film, and (2) documentary translation is not specifically audiovisual. In order to characterise it as specifically audiovisual, Espasa takes into account discursive aspects such as field, mode, translation mode and textual function, as well as audience design.
As for the AV transfer modes used in documentaries, voice-over plays a remarkable role, and this is why reference to documentaries can be found in publications dealing with voice-over, such as Franco (2001b), who analyses terminological and conceptual issues, Orero (2004Orero ( , 2005Orero ( , 2006Orero ( , 2007, forthcoming), who deals with TV interviews and synchronisation aspects, and, more specifically, Franco, Matamala and Orero (forthcoming), who identify the main problems found in the translation of documentaries by means of voice-overs and off-screen dubbings. Subtitling is also used and two authors who have studied the specificities of this transfer mode in documentaries are Ferreira (2002) and Kaufmann (2008). Ferreira chooses a French scientific documentary subtitled into Portuguese to pinpoint the particular features of subtitling nonfiction for TV, whilst Kaufmann explains the main features of a documentary film, concentrating on the definition of the source text, which deeply affects the subtitling process.
Finally, a promising study was devised by Herrero (2005), who wanted to analyse documentaries from a TS point of view. Aims, corpus and methodology were outlined but, unluckily, no further reference to this project is found in the literature.
All in all, the number of contributions regarding the translation of documentaries has increased in recent years but still further research needs to be carried out in order to grasp the specificities of this genre within AVT.

A corpus-based study
The aim of this article is to offer a corpus-based study of the main challenges translators have to overcome when translating documentaries. The corpus is made up of the documentaries translated by the author for the Catalan television channel (TVC) over two years (2006,2007). If more than one episode for each series was translated, only the first one was chosen. The result is a list of 20 documentaries (Table 1) whose original and translated versions have been analysed to portray the main features of documentary translation by means of a qualitative approach. It must be stressed that these documentaries were translated by the author without knowing that they would be analysed in the future.
Translating documentaries: a wide array of topics The first issue worth mentioning is that documentaries can deal with almost any topic and, therefore, audiovisual translators must be ready to find information about almost anything. In fact, as pointed out by Castro (1997), audiovisual translators must deal with all sorts of products: from sitcoms full of slang to wildlife documentaries. It could be said that audiovisual translators Á contrary to what generally happens in specialised translation Á are not specialists because of the topic but because of the transmission channel (audiovisual) and because of the transfer modes associated with it (dubbing, subtitling, voice-over, etc.). Looking at the documentaries in our corpus, different categories can be identified in terms of subject matter: . nature (eight), including Kenyan wildlife, giant mantas in San Benedicto, wildlife in Southern Asia, arthropods or elephant inseminations, among others; . history, archaeology and current affairs (seven), including Alexander the Great, Neanderthals, UN peacekeeping forces or the submarine Kursk; . art (one), and more specifically land sculptures; . religion (one), Christianity in this case; . child care (one) in an episode by the famous Supernanny, and . other topics (two) such as drug trials in India or problematic adoptions.
Translating documentaries: many voices, many AV transfer modes Different voices can be heard in documentaries: a narrator (generally off-screen), talking heads (people interviewed), people talking spontaneously and archival or TV footage. In some docudramas people (generally actors) staging re-enactments of real events are also to be found. Table 2 summarises how many of these characters appear in each documentary as well as their presence in the whole documentary. This percentage has been calculated by counting the total number of words of the documentary and distributing it among the four selected types of voices.
The structure of the selected documentaries varies in terms of speakers but some similarities are to be found: . There is generally one narrator, which gives coherence to the whole film. In all cases but two (Baby, be mine and Peacekeepers) the narrator (or commentator) is a man, and gender is always kept in the translation. Only in one instance (Killer wave) do two narrators coexist: one adopts a historical perspective and talks scientifically about tidal surges in England, and more specifically about the tidal surge in 1607 that left 2000 dead in the country, whilst the other adopts the point of view of a narrator who describes the re-enactment of hypothetical situations that could have happened at the time. For instance, this second narrator reads 'In Llanweir, in Monmouthshire, four miles from the sea, the servants of Mistress Vann prepare her lunch' while actors imitating these servants are shown on screen. The presence of a narrator is superior to other voices in 11 documentaries, with percentages higher than 50%. . There is one documentary (Equator) where only the voice of a narrator is heard; in most documentaries the words of narrators are accompanied by talking heads. The importance of these talking heads in terms of percentages varies from 3.07% (Supernanny) or 14.48% (Angel avalanche) to 77.45%, in documentaries such as Essentials of faith, built upon testimonies. . Real speech is also captured in most cases, giving place to a typical narratological structure: a narrator presents the situation, a talking head gives her opinion and some shots of real action with real speech are included. This incidental speech is not found in two documentaries and generally displays low percentages, always below 10%, except for three productions where lots of images from real-life situations are included. A clear example of this is Supernanny, as will be shown next. . In Supernanny the presence of a narrator is incidental and in very few instances do talking heads express their opinions. Most of the documentary is built on excerpts of real action, properly edited and with interspersed commentaries by the Supernanny. . Some excerpts from archival or TV footage are included in five documentaries, with percentages ranging from 0.39% to a maximum of 5.62% in Peacekeepers.
All in all, taking into account this corpus, 51.73% of documentary speech corresponds on average to narration, 37.57% to talking heads, 10.18% to incidental speech and 0.52% to archival or TV footage. Each of the previous voices is transferred using a specific AV modality according to the guidelines of the TV network: in Catalan TV off-screen narrators are substituted by a target version narration (offscreen dubbing), talking heads are revoiced by means of a voice-over, and incidental talk is transferred using either voice-over or subtitles. Voice-over is generally used when the language of the original coincides with the language of the narrator, whereas subtitles are used for other languages. This means that an audiovisual translator dealing with documentaries to be translated for Catalan TV has to master all these transfer modes, with a special emphasis on voice-over and off-screen dubbing Á in fact, in the selected corpus 95% of documentaries combine off-screen dubbing with voice-over and few subtitles. Each transfer mode entails its own challenges: . Voice-over: adapting the length of the text (voice-over isochrony), synchronising text and body language (kinetic synchrony), synchronising text and visuals (action synchrony), and rephrasing the language to create a comprehensible discourse which nonetheless retains the register (Franco, Matamala & Orero, forthcoming); . Off-screen dubbing: adapting the length of the text (isochrony), synchronising text and visuals (action synchrony) and creating a readable translation which keeps the register of the original (Franco, Matamala & Orero, forthcoming); . Subtitling: summarising the information and synchronising it with the visuals, among other difficulties.
Translating documentaries: language register Language varieties can be classified according to different parameters. Following the proposal by Hatim and Mason (1990), based on Halliday, language can change depending on two elements: the user (defined by geographical, temporal, social, [non]-standard and idiolectal features) and use (also called register). In this section I will analyse the variations in register found in the corpus, taking into account three variables that determine register: mode (channel of communication), field (the subject matter) and tenor (the participants and their relationships). Needless to say varieties related to the user would also yield some interesting results, but such an analysis would go beyond the purpose of this article.

Mode
Documentaries are audiovisual products which, in terms of mode, are mostly oral with some visual elements: the audio part, originally in English with occasional foreignlanguage excerpts Á for example, some interviewees speak in Russian in Baby, be mine Á is generally conveyed orally in the target language (except for a few subtitles). As for the visuals, they remain unaltered except for all meaningful written language, which is generally transferred into a written caption in the target language. Common examples are the names and affiliations of talking heads, which are generally rendered in a caption (example 1), but this could also include any sign on screen such as newspaper headlines (example 2) or date and location information (example 3).
(2) 'Power is at the bottom of the sea' (Kursk).

Field
Field is definitely one of the main challenges in documentary translation. As demonstrated, documentaries cover a wide array of topics and professionals need to be able to develop strategies to translate subjects as far-ranging as Neanderthals and the Supernanny. In most documentaries language is not highly technical because specialists are generally aware that they are addressing a lay audience but, nonetheless, terminology is quite particular and frequent. Finding an equivalent is not an extremely difficult task thanks to the many resources available to translators (dictionaries, terminological databases, corpora, specialists, etc.), as in examples 4 to 8.
However, terminology might pose problems related to equivalence, usage (in vitro or in vivo terminology), absence of terminology, and ambiguity, among other things.
. Equivalence: in some instances understanding the original is easy, especially because images offer a clear referent, but finding the right equivalent can be a challenging task, as in example 9.
(9) Dr. Rubin: Gavin and I built a PVC cage that would fit over one of the skiff's sides that we could attach a hydrophone to. That allows us to actively track the animals around the island so we can use the tags in an active way by following them in the skiff with a hydrophone so it gives us a daily pattern of the animal's movements. (Blue realm) Although no equivalent could be found in dictionaries or terminological databases for the term 'tag' in this particular context (probably referring to global positioning tags), the translator could see the actual 'tags' and decided to translate it by using the word 'dispositius' (literally, 'devices').
. Terminology in vitro vs. terminology in vivo: this distinction proposed by Cabré (1999) refers to the fact that the translator might have to choose whether to use the equivalent proposed by linguistic authorities and found in terminological or lexicographical works (in vitro) or the unit really used by specialists in the field, which might be a loanword (in vivo).
This is the problem encountered in Angel avalanche when talking about snowboard(s) and snowboarders. According to the Catalan terminological body Termcat (www.termcat.cat), 'snowboarding' should be translated by the expression 'surf de neu', a 'snowboard' should be transferred as a 'planxa de neu' and 'snowboarder' as 'surfista' or 'surfista de neu'. However, most snowboarders in Catalonia use the English term. Hence, should the translator use the term in vitro or the term in vivo? Each option offers both advantages and disadvantages but, taking into account that the word was used by a narrator and not by a surfer, official proposals were used, with a footnote addressed to the language editor where the translator showed her doubts and opened the door to a later change if it was considered necessary.
. Absence of (or inability to find) terminology in the target language: confirming that a term has no equivalent in the target language is another problem found in the corpus.
Strategies used to overcome this problem are various but in Equator: Battle for the light two can be seen: creating new terminology or using a more general translation, as in 10 and 11.
(10) In Borneo, a rhinoceros hornbill and a rare helmeted hornbill arrive early, inspecting the ripeness of its tiny fruit. (11) But this snake ignores the file-eared frog . . . It prefers a much easier meal of frog spawn.
In the first example ('helmeted hornbill', Rhinoplax vigil) the translator innovated and created her own proposal ('calau de casc'), based on analogy with other Romance languages. In the second instance ('file-eared frog', Polypedates otilophus) the translator was unable to find an equivalent but, considering the reference to this animal was not particularly relevant in the documentary, a generalizing strategy was chosen and 'frog' ('granota') was the proposed translation.
. Ambiguity: another terminology-related problem is the ambiguity of the equivalent.
In Monumental vision, for example, a land sculpture called 'Ratio' is built up: (12) Narrator: Andrew then went searching for the location to build his stairway sculpture which he called 'Ratio'.
One of the possible Catalan translations for the English word 'ratio' is 'raó ' but the word 'raó ' in Catalan has another meaning which is more frequent: 'reason'. Taking into account that 'raó ' could entail an ambiguity problem, the translation read: (13) Narrator: L'Andrew es va disposar a buscar el lloc on construiria la seva escultura d'escales, que va anomenar Raó en el sentit matemàtic. [Andrew went searching for the location to build his stairway sculpture which he called 'Ratio' in its mathematical meaning.] . Obscure equivalent to the target audience: finally, an equivalent can be incomprehensible for a lay audience.
This was the case of the term 'storm surge' or 'storm tide', which appeared in terminological databases as 'sobreelevació meteorològica', an obscure equivalent for most viewers Á and for the translator herself. However, after considering that the whole documentary dealt with this topic and it would become self-explanatory, it was decided to leave the Catalan equivalent.
It could be argued that the previous difficulties are not exclusive to audiovisual translation, but I would like to point out some features which make them even more relevant: first of all, visuals determine the translation. In other words, the translation has to match the images shown on screen and this might restrict the array of possible translation strategies that can be used. For example, omission of a term might not be an option if it refers to an object which is shown on screen. In spite of this setback, as already stated, images are also of great value when deducing the meaning of unknown terms because they add information.
Second, the fact that the product is received audiovisually and not in paper format affects the work of the translator, who is aware of the generally ephemeral nature of the documentary. In this sense, translators cannot expect the audience to stop and look up a word in a dictionary Á as could be done in a written translation Á nor can they add explanatory footnotes to enhance the understanding of the target text. These constraints might therefore determine the translator's choice, as it is necessary to create a readily comprehensible product.
Finally, depending on the transfer mode chosen, there are even more specific restrictions: in off-screen dubbing and voice-over the translation will be received orally. This means that an oral input (original audiovisual product) will be captured in a written form (translation script); it will then be delivered orally by a voice talent and will finally be received audiovisually by the audience. In contrast, in subtitling, an oral input will be converted into a written output (subtitles) which will coexist with the original oral output. This interplay of channels (oral/written) affects the terminology chosen by the translator: just to give an example, it is completely different to read the scientific name of an animal in a written publication and to listen to it in an audiovisual product. Hence, strategies in written and audiovisual media vary.

Tenor
Finally, concerning the tenor, documentaries are usually associated with formal language. However, documentaries are an evolving filmic product which has gone beyond the typical structure of a formal off-screen narrator. In the selected corpus, examples from different types of tenor can be seen, ranging from the formal narrator in 14 to the colloquial language in 15: (14) One extraordinary group of animals rivals us as masters of earth. They ruled the ancient sea, and they were the first creatures to venture onto land. With highly adaptable limbs and bodies, they invaded every realm of the planet . . . equipped by nature to survive almost any challenge. (Blue realm) (15) Child: Anna is a crybaby, Anna is a crybaby.
Dad: Stop saying that. Jo: Brian is incredibly passive. Dad: Knock it off. Jo: He doesn't say boo to a mouse really. They don't listen to him. Dad: Ellie, don't tease me. Just give it to me. (Supernanny) The job of the translator is to render each register faithfully, taking into account the limitations of the transfer mode chosen.
In voice-overs, although colloquial features in lexis are generally kept, it is not always the case with syntax, which is generally corrected. As explained by Matamala (2009): 'even when the interviewee does not speak good English and makes mistakes, these are corrected in the translated version, since the information given by the speaker and not formal issues are what really matters. On the contrary, in fiction films, these elements might be important in order to define the character and create a credible (not real) oral language. Can we imagine a Woody Allen acting with a freeflowing discourse without hesitation? Could we stand the fact that a dubbing actor would hesitate that much in revoicing an interview in which Allen talks about his latest movie?' Example 16 shows the rephrasing of the original undertaken by the translator, which is especially remarkable when one realises that the transcription provided did not use punctuation at all in this excerpt. Moreover, the original texts present some complex sentences ('it has woven itself in the fabric of Christianity worldwide') and some concepts difficult to convey in Catalan such as 'empowerment'. Although the term 'apoderament' has been proposed lately by terminological authorities to translate 'empowerment', at the time the documentary was translated no valid option was found and the translator decided to reformulate the approximate meaning of the word.
In off-screen dubbing, there are apparently fewer constraints because it is used for narrators, which tend towards formality. However, I have observed a trend towards informality in narration which has altered some preconceptions about the translation of commentators. This includes examples ranging from semi-formal narrators with features of colloquial language to totally informal narrators who want to present themselves as intrepid adventurers and connect with a young audience. When confronted with these mixed tenors, translators have to find the appropriate resources in their own language to produce the same effect. Finally, although less used in our corpus, capturing the tenor in subtitling also presents some challenges, especially in Catalan. Contrary to what is done in other TV networks and languages, where the written language of subtitles does not reproduce most morphological features of colloquial language, Catalan TV proposes to keep them in controversial issues such as pronouns, for instance. Therefore, depending on the tenor of the original, Catalan subtitles will depict formal or informal pronouns in a written form.

Translating documentaries: further challenges
In the documentaries that make up our corpus further challenges have been detected, showing the complexity of a task which is not always recognised. The first difficulty encountered is proper nouns, for reasons as varied as the ones listed below: . Proper nouns in other alphabets need to be transcribed taking into account the target language conventions and not the English ones. For example, Russian names in Baby, be mine or Kursk. . Proper nouns referring to geographical locations whose transcription has to be verified in atlases. For instance, lakes in Kenya (Kenyan trilogy) or Afghanistan (Smell of paradise). ? . Names of universities: a common difficulty is knowing how to translate the name of the institution where talking heads work, generally universities. For example, the captions 'Victor Hurowitz, Ben-Gurion University' or 'Israel Finkelstein, Tel Aviv University' (King Solomon's . . . ) are apparently easy to translate, but one must know whether they refer to a city or not because in Catalan a preposition is used or not used based on this ('Universitat de Tel Aviv' and 'Universitat Ben-Gurion'). . Book and film titles: in this case, the translator has to check whether the movie or film has been translated and provide the official translation. This happened in Baby be mine ('The Rescuers') or Essentials of faith ('Doubts and loves', by Richard Holloway).
Undoubtedly the translation of proper nouns is also problematic in other types of translations (Ballard 1993, Elman 1986, Espinal 1989, Franco Aixelà 2000, Moya 2000. Nevertheless, as pointed out by Mayoral (2000, pp. 103Á104), it presents some distinct features in AVTwhich, in my view, can be partly transferred to the translation of documentaries: the translator is part of a working chain and might not take the final decision on the translation of proper nouns; the audience for audiovisual products is very wide and their knowledge of foreign languages and culture is not homogeneous; the translation of proper nouns can affect the commercial value of a product; the transmission of the local flavour depends on the work, the audience and the genre, and intertextuality can affect the translation of proper nouns.
Regarding the different transfer modes used, I would add some more distinctive elements: for example, in off-screen dubbing and voice-over (oral modes) the pronunciation of proper nouns might be a challenge and it is the task of either the translator or the language adviser to include notes on how to pronounce certain foreign proper nouns. In contrast, in subtitling (written mode) pronunciation is not relevant, but spelling is of the essence and guidelines given by the client on the translation, transcription and transliteration of proper nouns must be followed in order to produce a coherent translation.
Comprehension problems also seem to have had a special significance in documentaries, especially when real speech is included or non-standard accents are used. This is a specificity of AVT since, although delivering a written product, translators work with a source text which is an audiovisual product Á which might have been captured into a written transcript or not. In this sense, it could be said that audiovisual translators are closer to interpreters than to translators of written texts, although the lack of immediacy and the possibility of going backwards and forwards to check an incomprehensible item are a substantial difference.
In Supernanny the original transcripts include words and even sentences that are rendered with an '[unintelligible]', generally because there is too much noise (usually children screaming). In the selected corpus, three solutions are found, as shown in example 17: (1) not offering a translation when it is not relevant and the sentence is extremely short; (2) creating a comprehensible sentence with the intelligible words, either ignoring the unintelligible parts or using a more general word which fits in the context, and (3) as a last resort, inventing something which fits in the context. This is also especially arduous when other languages are used, as in example 18 from Ancient voices: Alexander, the god king. The transcription given to the translator reads as follows: (18) Liana: They say it is not written, the word poison is written here Á IOY.. Tony: But this is the Latin personal name, Servius. Servius Sulpicius. This is . . . Further minor challenges that are not described in this article are inconsistencies between the written script and the audio, songs that have to be translated, Bible and book excerpts, among others.

Conclusions
In conclusion, this article has described the main features of documentary translation using a corpus of 20 documentaries. It has been shown that each documentary presents its own characteristics but, still, some recurrent problems are found, either related to the transfer mode used for each speaker or related to register. Terminological problems occupy a prominent position, be it because certain equivalents are not found in the target language or because more than one equivalent is possible, among other reasons. Finally, further difficulties encountered in the corpus are described, with many examples dealing with proper nouns, unintelligible excerpts and mistranscriptions. All in all, I hope to have shed some light on the fascinating field of documentary translation, which, as demonstrated earlier in the section 'The translation of documentaries in AVT', deserves to be further investigated. Wider corpora encompassing more languages could be used to analyse the translation of a genre which transmits knowledge, disseminates terminology and sometimes even invents with new terms.